Preview only show first 10 pages with watermark. For full document please download

Canadian Foundation Engineering Manual 4th







Preface iii
) .... ; .. ..,r_.
f l ' ~ ~ ~ ; . ~ : :
l ~ ~   . _n_ .,.
The Canadian Foundation Engineering Manual is a publication ofthe Canadian Geotechnical Society. It is originally
based on a manual prepared under the auspices of the National Research Council of Canada Associate Committee
on the National Building Code, Subcommittee on Structural Design for the Building Code. A draft manual for
public comment was published in 1975. In 1976, the Canadian Geotechnical Society assumed responsibility for the
Manual and placed it under the Technical Committee on Foundations. This coinmittee revised the 1975 draft and
published in 1978 the first edition of the Canadian Foundation Engineering Manual, which incorporated suggestions
received on the 1975 draft.
The Society solicited comments on the Manual and suggestions for revisions and additions in Seminars across the
country. In 1983, the Society requested that the Technical Committee review the comments and suggestions received
and prepare a second edition of the Manual published in 1985. A third edition was produced in 1992, including
various revisions and additions. Further developments in applied GeoEngineering and Ground Engineering are
included in this fourth edition, published in 2006.
The Manual is truly produced by the membership ofthe Canadian Geotechnical Society. The number of individuals
who have contributed to the manual first, the preparation of the 1975 draft, then the 1978 first edition, the 1985
second edition, the 1992 third edition and this 2006 fourth edition - is very large. Specific individuals who contributed
to the fourth edition were: '
D.E. Becker and 1. D. Moore (Editors)
1. Lafleur (Editor, French Edition)
S.L. Barbour
R.J. Bathurst
S. Boone
R. W.I Brachman
B. Brockbank
M. Diederichs
M.H. El Naggar
1. Fannin
D. Fredlund
I ·r-··
1. Howie
D.1. Hutchinson
J.M. Konrad
S. Leroueil
K. Novakowski
1. Shang
The Manual provides information on geotechnical aspects of foundation engineering, as practiced in Canada, so
that the user will more readily be able to interpret the intent and performance requirements ofthe National Building
Code of Canada (the release ofthis fourth edition coincides with publication ofthe NBCC, 2005) and the Canadian'
, t
iv  Canadian  Foundation Engineering Manual 
Foundationengineeringis notaprecisescience,butistoa extentbaseduponexperienceandjudgement.The
Manual assumes thattheuseris experiencedinandunderstandsthe specializedfield ofgeotechnicaland ground
engineering. TheManualis notatextbook,norasubstitutefor theexperienceandjudgementofapersonfamiliar
1. Acceptable design guidelines for the solution ofroutine foundation engineering problems, as based on
2. Anoutlineofthelimitationsof certainmethodsofanalysis.
3. Informationonpropertiesofsoilandrock,includingspecificconditionsencounteredinCanada.
4. Commentsonconstructionproblems,wheretheseinfluencethedesignorthequalityofthefoundation.
TheManualcontainssuggestedrather'thanmandatoryprocedures.It istheintentionoftheCanadianGeotechnical
Society to continue theprocess ofreview, andto update the Manual as the needarises. While reasonable efforts
Comments and suggestions on the technical contents ofthe Manual are welcome. Such comments should be
Canadian Geotechnical Society
Email: [email protected] 
v  Table of Contents 
Table of Contents 
Preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. iii 
1  Introduction..  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  1 
2  Definitions,  Symbols and Units  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  2 
2.1  Definitions  .................................................................. 2 
2.2  Symbols  .................................................................. , .  5 
2.2.1  The International System of Units (SI) .. , ........................ , . , , . , ..... 6 
3  Identification and Classification of Soil and Rock ................... "  13 
3.1  Classification of Soils ...................................................... , ..  13 
3.1.1  Introduction., ........................................................ 13 
3.1.2  Field Identification Procedures .................. ; ........................ 13 
3.2  Classification of Rocks  .... , ..................... , ........................... 19 
3.2.1  Introduction.'., .... , .. , ............................................... 19 
3.2.2  Geological   . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  20 
3.2.3  Structural Features of Rockmasses ........................................ 20 
3.2.4  Engineering Properties of Rock Masses  .... '............................... 20 
4  Site Investigations ............................................ "  31 
4.1  Introduction  ........................................................... , , . ,.  31 
4.2  Objectives of Site Investigations  . , ........... , ............ , .......... , .......... 31 
4.3  Background Information ....................................................... 32 
4.4  Extent of Investigation.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  33 
4.4.1  Introduction.......................................................... 33 
4.4.2  Depth ofInvestigation .................................................. 34 
4.4.3  Number and Spacing of Boreholes .,' .................... , ................. 35 
4.4.4  Accuracy  ofInvestigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  36 
4.5  In-Situ Testing of Soils ........................................................ 36 
4.5.1  Introduction............................................ '"  ........ '"  36 
4.5.2  Standard Penetration Test (SPT).  , ........ , ....... , ..... , . , ........... , ...  37 
4.5.3  Dynamic Cone Penetration Test (DCPT)  . , .......................... , ...... 44 
4.5.4  Cone Penetration Test (CPT) ............................................. 45 
4.5.5  Becker Penetration Test (BPT)  .......................................... ,  47 
4.5.6  Field Vane Test (FVT) .................................. , ............. , .  48 
4,5.7  PressuremeterTests,{PMT) ...... , ....................................... 50 
4.5.8  Di1atometer Test (DMT) ....... ; ...... , .......................... , ...... 55 
vi Canadian Foundation Engineering Manual
4.5.9  The Plate-Load and Screw-Plate Tests  ..................... : ............... 55 
4.6  Boring and Sampling  ....................................................... 56 
4.6.1  Boring.............................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  56 
4.6.2  Test Pits.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  56 
4.6.3  Sampling ............................................................ 57 
4.6.4  Backfilling........................................................... 62 
4.7  Laboratory Testing of Soil Samples .............................................. 62 
4.7.1  Sample Selection ...................................................... 63 
4.7.2  Index Property Tests  ................................................... 63 
4.7.3  Tests  for  Corrosivity  ................................................... 63 
4.7.4  Structural Properties Tests ............................................... 63 
4.7.5  Dynamic Tests ......................................................... 63 
4.7.6  Compaction Tests  ..................................................... 64 
4.7.7  Typical Test Properties  ................................................. 64 
4.8  Investigation of Rock ......................................................... 70 
4.8.1  General  ............................................................. 70 
4.8.2  Core Drilling of Rock .................................................. 71 
4.8.3  Use of Core Samples ..................................... : ............. 72 
4.8.4  In-situ Testing  ........................................................ 72 
4.9  Investigation of Groundwater ................................................... 73 
4.9.1  General  ............................................................. 73 
4.9.2  Investigation in Boreholes ............................................... 73 
4.9.3  Investigation by Piezometers  ............................................ 74 
4.10  Geotechnical Report ......................................................... 74 
4.11  Selection of Design Parameters  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  75  . 
4.11.1  Approach to Design .... : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  75 
4.11.2  Estimation of Soil  for Design  ................................... 76 
4.11.3  Confirmation of Material Behaviour by Construction Monitoring.  . . . . . . . . . . . . ..  77 
4.12  Background Information for  Site Investigations  ................................... 77 
5  Special Site Conditions . ...................  ..................... 78 
5.1  Introduction  ................................................................ 78 
5.2  Soils  ...................................................................... 78 
5.2.1  Organic Soils, Peat and Muskeg .......................................... 78 
5.2.2  Normally Consolidated Clays ............................................ 78 
5.2.3  Sensitive Clays  ....................................................... 79 
5.2.4  Swelling and Shrinking Clays ............................................ 79 
5.2.5  Loose, Granular Soils  .................................................. 79 
5.2.6  Metastable Soils  ...................................................... 79 
5.2.7  Glacial Till.  .......................................................... 80 
5.2.8  Fill. ......................................... '.'  ..................... 80 
5.3  Rocks  ..................................................................... 80 
5.3.1  Volcanic Rocks  ..... ',' ................................................ 80 
5.3.2  Soluble Rocks ........................................................ 80 
5.3.3  Shales  .............................................................. 80 
5.4  Problem Conditions  ............... : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  81 
5.4.1  Meander Loops and Cutoffs  ............................................. 81 
5.4.2  Landslides ........................................................... 81 
5.4.3  Kettle Holes .......................................................... 81 
5.4.4  MinedAreas  ......................................................... 82 
5.4.5  Permafrost ........................................................... 82 
Table of Contents vii
5.4.6  Noxious or Explosive Gas ............................................... 82 
5.4.7  Effects of Heat or Cold ................................................. 82 
5.4.8  Soil Distortions ....................................................... 83 
5.4.9  Sulphate Soils and Groundwater .......................................... 83 
6  Earthquake - Resistant Design  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  84 
6.1  Introduction  ................................................................ 84 
6.2  Earthquake Size  ............................................................. 85 
6.2.l  Earthquake Intensity ................................................... 85 
6.2.2  Earthquake Magnitude  ................................................. 85 
6.2.3  Earthquake Energy  .................................................... 86 
6.3  Earthquake Statistics and Probability of Occurrence.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  86 
6.4  Earthquake Ground Motions  .................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  86 
6.4.1  Amplitude Parameters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  87 
6.4.2  Frequency Content  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  89 
6.4.3  Duration  ......... ".................................................. 89 
6.5  Building Design  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  89 
6.5.l  Equivalent Static Force Procedure  ........................................ 90 
,6.5.2  Dynamic Analysis ..................................................... 96 
6.6  Liquefaction .......................' ......................................... 99 
6.6.l  Factors Influencing Liquefaction  ........................................ 100 
6.6.2  Assessment of Liquefaction  ......................... '................... 100 
6.6.3  Evaluation of Liquefaction Potential  ..................................... 101 
6.6.4  Liquefaction-Like Soil Behaviour  ......................................... 111 
6.7  Seismic Design of Retaining Walls  ............ '................................. 112 
6.7.l  Seismic Pressures on Retaining Walls  ... , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  113 
6.7.2  Effects of Water on Wall Pressures ....................................... 115 
6.7.3  Seismic Displacement ofRetaining Walls  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  115 
6 ~ 7   4 Seismic Design Consideration  .......................................... 116 
6.8  Seismic Stability of Slopes and Dams  ........................................... 118 
6.8.1  Mechanisms of Seismic Effects  ......................................... 118 
6.8.2  Evaluation of Seismic Slope Stability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  119 
6.8.3  Evaluation of Seismic Deformations of Slopes  .......... '"  ................ 120 
6.9  Seismic Design of Foundation ................................................. 121 
6.9.l  Bearing Capacity of Shallow Foundations ................................. 121 
6.9.2  Seismic Design of Deep Foundations ..................................... 122 
6.9.3  Foundation Provisions ................................................. 122 
Foundation Design  .'.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  123 
7.1  Introduction and Design Objectives ............................................. 123 
7.2  Tolerable Risk and Safety Considerations  ....................... i................. 123 
7.3  Uncertainties in Foundation Design ............................................. 124 
7.4  Geotechnical Design Process ................................................. ,  124 
7.5  Foundation Design Methodology ............................................... 125 
7.6  Role of Engineering Judgment and Experience .................................... 128 
7.7  Interaction Between Structural and Geotechnical Engineers ... : ...................... 128 
7.7.1  Raft Design and Modulus of Sub grade Reaction  ............................ 128 
Limit States and Limit States Design. . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  132 
8.1  Introduction  ............................................................... 132 
viii Canadian Foundation Engineering Manual
, ,
8.2  What Are Limit States?  ...................................................... 133 
8.3  Limit States Design (LSD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  134 
8.4  LSD Based on Load and Resistance Factor Design (LRFD) ....................... , .,  136 
8.5  Characteristic Value  ........................................................ ,  138 
8.6  Recommended Values  for Geotechnical Resistance Factors  ........... ; .............. 138 
8.7  Terminology and Calculation Examples .......................................... 140 
8.7.1  Calculation Examples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  140 
8.8  Working Stress Design and Global Factors of Safety................................ 141 
9  Bearing Pressure on Rock ...................................... ,.  143 
9.1  Introduction  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  143 
9.2  Foundations on Sound Rock ................................................... 145 
9.3  Estimates of Bearing Pressure  ................................................. 147 
9.4  Foundations on Weak Rock  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  148 
9.5  Special Cases  .............................................................. 149 
9.6  Differential Settlement ....................................................... 149 
10  Bearing Capacity of Shallow Foundations on Soil.  . . . . . . . . . . . . . . . . . ..  150 
10.1  Introduction  .............................................................. 150 
10.2  Conventional Bearing Capacity Foundations on Soil.  .............................. 150 
10.3  Bearing Capacity Directly from In-Situ Testing ................................... 155 
10.4  Factored Geotechnical Bearing Resistance at Ultimate Limit States. . . . . . . . . . . . . . . . . ..  157 
11  Settlement of Shallow Foundations  ............................... 158 
11.1  Introduction............................................................... 158 
11.2  Comp'onents ofDefiection  .................................................... 158 
11.2.1  Settlement of Fine-Grained Soils  ...................................... ,  159 
11.2.2  Settlement of Coarse-Grained Soils  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  159 
11.3  Three-Dimensional Elastic Displacement Method.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  159 
11.3.1  Approximating Soil Response as an Ideal Elastic Material  .. . . . . . . . . . . . . . . . ..  159 
11.3.2  Drained and Undrained Moduli.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  160 
11.3. 3  Three-Dimensional Elastic Strain Integration.  . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  160 
11.3.4  Elastic Displacement Solutions ......................................... 160 
11.4  One-Dimensional Consolidation Method ....................................... ,  162 
11.4.2  One-Dimensional Settlement:  e-Iogcr' Method ............................. 165 
11.4.3  Modifications to  One-Dimensional Settlement ............................. 166 
11.5  Local Yield. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  166 
11.6  Estimating Stress Increments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  166 
11.6.1  Point Load ...............  ..................  . ~ ..................... 166 
11.6.2  Uniformly Loaded Strip  .............................................. 167 
11.6.3  Uniformly Loaded Circle  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  168 
11.6.4  Uniformly Loaded Rectangle  ................ , ......................... 169 
11.7  Obtaining Settlement Parameters ........................................ , .....  170 
11.8  Settlement of Coarse-grained Soils Directly from In-Situ Testing.  . . . . . . . . . . . . . . . . . . ..  172 
11.8.1  Standard Penetration Test (SPT)  .......  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  172 
11.8.2  Cone Penetration Test (CPT).  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  173 
11.9  Numerical Methods ......................................................... 175 
11.10  Creep.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  175 
Table of Contents ix
11.11  Rate of Settlement ............................................................ ,  176 
11.11.1  One-Dimensional Consolidation ....................................... 176 
11.11.2  Three-Dimensional Consolidation ...................................... 177 
11.11.3  Numerical Methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  178 
11.12  Allowable (Tolerable) Settlement. ............................................ 178 
12  Drainage and Filter Design.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  181 
12.1  Introduction  ..... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  181 
12.2  Filter Provisions  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  181 
12.3  Filter Design Criteria  ...................................................... :  182 
12.4  Drainage Pipes and Traps .................................................... 183 
13  Frost Action.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  185 
13.1  Introduction  ............. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  185 
13.2  Ice Segregation in Freezing Soil. .............................................. 185 
13.3  Prediction of Frost Heave Rate ................................................ 187 
13.4  Frost Penetration Prediction.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . ..  190 
13.5  Frost Action and Foundations ................................................. 195 
13.6  Frost Action during Construction in Winter ..... "  ............................... 197 
14  Machine Foundations  .......................................... 200 
14.1  Introduction  ........................................................... ···  200 
14.2  Design Objectives  ......................................................... 200 
14.3  Types of Dynamic Loads  .................................................... 200 
14.3.1  Dynamic Loads Due to Machine Operation ......................... , ..... 200 
14:3.2  Ground Transmitted Loading  .......................................... 201 
14.4  Types of Foundations  ....................................................... 202 
14.5  Foundation Impedance Functions  ............................................. 202 
14.5.1  Impedance Functions of Shallow Foundations ............................. 202 
14.5.2  Embedment Effects .................................................. 203 
14.5.3  Impedance Functions of a Layer of Limited Thickness  ...................... 205 
14.5.4  Trial Sizing of Shallow Foundations ..................................... 206 
14.6  Deep Foundations .......................................................... 206 
14.6.1  Impedance Functions of Piles .......................................... 206 
14.6.2  Pile-Soil-Pile Interaction .............................................. 208 
14.6.3  Trial Sizing of Piled Foundations ....................................... 208 
14.7  Evaluation ofSoi! Parameters  ................................................ 209 
14.7.1  Shear Modulus  ..................................................... 209 
.14.7.2  Material Damping Ratio .............................i ................. 209 
14.7.3  Poisson's Ratio and Soil Density  ....................................... 209 
14.8  Response to Harmonic Loading ............................................... 210 
14.8.1  Response of Rigid Foundations in One Degree of Freedom................... 210 
14.8.2  Coupled Response of Rigid Foundations  ................................. 211 
14.8.3  Response of Rigid Foundations in  Six Degrees of Freedom  .................. 212 
14.9  Response to Impact Loading  ................................................. 212 
14.9.1  Design Criteria  ..................................................... 212 
14.9.2  Response of One Mass Foundation ...................................... 213 
14.9.3  Response of Two Mass Foundation  ..................................... 213 
14.10  Response to  Ground-Transmitted Excitation .................................... 213 
x  Canadian Foundation Engineering  Manual 
15  Foundations on Expansive Soils. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  215 
15.1  Introduction  .................................................'.  . . . . . . . . . . ..  215 
15.2  Identification and Characterization of Expansive Soils  ............................. 217 
15.2.1  Identification of Expansive Soils:  Clay Fraction, Mineralogy, Atterberg Limits, 
Cation Exchange Capacity  .................................................. 218 
15.2.2  Environmental Conditions  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  222 
15.2.3  Laboratory Test Methods  ............................................. 222 
15.3  Unsaturated Soil Theory and Heave Analyses  .................................... 225 
15.3.1  Prediction of One-Dimensional Heave ................................... 227 
15.3.2  Example of Heave Calculations  ........................................ 229 
15.3.3  C10sed-Fonn Heave Calculations ....................................... 230 
15.4  Design Alternatives, Treatment and Remediation  ................................. 231 
15.4.1  Basic Types of Foundations on Expansive Soils ............................ 231 
15.4.2  Shallow Spread Footings for Heated BUildings  ............................ 231 
15.4.3  Crawl Spaces Near or Slightly Below Grade on Shallow Foundations  .......... 232 
15.4.4  Pile and Grade-Beam System .......................................... 232 
15.4.5  Stiffened Slabs-on-Grade  ............................................. 233  . 
15.4.6  Moisture Control and Soil Stabilization .................................. 234 
16  Site and Soil Improvement Techniques  . . . . . . . . . . . . . . . . . . . . . . . . . . . .  237 
16.1  Introduction  ........... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  237 
16.2  Preloading................................................................ 237 
16.2.1  Introduction........................................................ 237 
16.2.2  Principle of Pre loading ............................................... 237 
16.2.3  Design Considerations  ............................................... 238 
16.3  Vertical Drains  ............................................................ 239 
16.3.1  Introduction................ : ....................................... 239 
16.3.2  Theoretical Background  .............................................. 240 
16.3.3  Practical Aspects to Consider in Design .................................. 242 
16.4  Dynamic Consolidation  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  245 
16.4.1  Introduction........................................................ 245 
16.4.2  Methodology ....................................................... 245 
16.4.3  Ground Response  ................................................... 246 
16.5  In-Depth Vibro Compaction Processes .......................................... 249 
16.5.1  Introduction........................................................ 249 
16.5.2  Equipment ......................................................... 249 
16.5.3  Vibro Processes ..................................................... 249 
16.6  Lime Treatment.  ........................................................... 251 
16.6.1  The Action of Lime in Soil ............................................ 251 
16.6.2  Surface Lime Treatment .......................... ',' .................. 251 
16.6.3  Deep Lime Treatment ................................................ 251 
16.7  Ground Freezing 
,  t 
........................................................... 252  ,  ) 
16.7.1  The Freezing Process  ................................................ 252  : 
"  1 

16.7.2  Exploration and Evaluation ofFonnations to be Frozen  ..................... 252  .  i 
16.7.3  References .................................................. '"  .... 253 

16.8  Blast Densificatio:Q  ......................................................... 253 
16.9  Compaction Grouting ....................................................... 254 
l6..l0  Chemical Grouting ........................................................ 254 
16.11  Preloading by Vacuum  ................................................  .... 255 
16.12  Electro-Osmotic and Electro-Kinetic Stabilization  ............................... 256 
Table of Contents xi
17  Deep Foundations - Introduction  ................................. 260 
17.1  Definition  ................................................................ 260 
17.2  Design Procedures  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  260 
17.3  Pile-Type Classification  ..................................................... 260 
17.4  Limitations  ............................................................... 260 
18  Geotechnical Design of Deep Foundations .......................... 262 
18.1  Introduction  .............................................................. 262 
18.2  Geotechnical Axial Resistance of Piles in Soil at Ultimate Limit States ................ 262 
18.2.1  Single Piles - Static Analysis  .......................................... 262 
18.2.2  Pile Groups - Static Analysis  ... , ................. , .. , ................. 268 
18.2.3  Single Piles - Penetrometer Methods  .................................... 269 
18.2.4  Single Piles - Dynamic Methods ....................................... ,  272 
18.2.5  Negative Friction and Downdrag on Piles  ................................ 273 
18.2.6  Uplift Resistance .................................................... 276 
18.2.7  Other Considerations  ................................................ 277 
18.3  Settlement of Piles in Soil  ............................ ; ...................... 279 
18.3.1  Settlement of Single Piles ............................................. 279 
18.3.2  Settlement ofa Pile Group  ............................................ 284 
18.4  Lateral Capacity of Piles in Soil  .............................................. 286 
18.4.1  Broms' Method ..................................................... 288 
18.4.2  Pressurenieter Method  ............................................... 288 
18.5  Lateral Pile Deflections  ..................................................... 291 
18.5.1  The p-y Curves Approach  ............................................ 291 
18.5.2  Elastic Continuum Theory  ............................................ 292 
18.6  Geotechnical Axial Capacity ofDeep Foundations on Rock ......................... 295 
18.6.1  Introduction........................................................ 295 
18.6.2  Drilled Piers or Caissons - Design Assumptions  ........................... 295 
18.6.3  End-Bearing  .... '................................................... 295 
18.6.4  Shaft Capacity of Socket. ........................ , ; ................... 297 
18.6.5  Design for Combined Toe  and Shaft Resistance ......... '................... 298 
18.6.6  Other Failure Modes ................................................. 299 
18.7  Settlement of Piers Socketed into Rock ......................................... 299 
1 8 ~ 7   1 Fundamentals  ...................................................... 299 
18.7.2  Settlement Estimated from Pressuremeter Testing .......................... 300 
18.7.3  Settlement from Plate Test Loading  ..................................... 300 
18.7.4  Settlement using Elastic Solutions  ..............'........................ 300 
19  Structural Design and Installation of Piles. . . . . . . . . . . . . . . . . . . . . . . . . .  303 
19.1  Introduction, .............. , ..... , ..... , ... , .............................. 303 
19.1.1  Resistance ofDeep Foundations  ....................................... 303 
19.1.2  Wave-Equation Analysis .............................................. 304 
19.1.3  Dynamic Monitoring  ................................................ 305 
19.1,4  Dynamic Pile Driving Formulae  ....................................... 305 
19.2  Wood Piles  ............................................................... 305 
19.2.1  Use of Wood Piles  .................................................. 305 
19.2.2  Materials  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  306 
19.2.3  Structural Design  ................................................... 306 
19.2.4  Installation ofWood Piles ............................................. 306 
xii Canadian Foundation Engineering Manual
19.2.5  Common Installation Problems ......................................... 306 
19.3  Precast and Prestressed Concrete Piles .......................................... 306 
19.3.1  Use of Precast and Prestressed Concrete Piles  ............................. 306 
19.3.2  Materials and Fabrication ............................................. 307 
19.3.3  Pile Splices  ........................................................ 307 
19.3.4  Structural Design .................................................... 307 
19.3.5  Installation  ........................................................ 308 
19.3.6  Common Installation Problems  ........................................ 309 
19.4  Steel H-Piles .............................................................. 309 
19.4.1  UseofSteelH-Piles  ................................................. 309 
19.4.2  Materials ......................................... , ................ 310 
19.4.3  Splices  ........................................................... 310 
19.4.4  Structural Design  ................................................... 310 
19.4.5  Installation and Common Installation Problems.  . . . . . . . . . . . . . . . . . . . . . . . . . . .  310 
19.5  Steel Pipe Piles.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  311 
19.5J  Use of Steel Pipe Piles  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  311 
19.5.2  Materials .......................................................... 312 
19.5.3  Structural Design  .................................................... 312 
19.5.4  Installation  ........................................................ 313 
19.5.5  Common Installation Problems ......................................... 314 
19.6  Compacted Expanded-Base Concrete Piles  ...................................... 314 
19.6.1  Use of Compacted Concrete Piles  ...................................... 314 
19.6.2  Materials .......................................................... 314 
19.6.3  Structural Design  .................................................... 314 
19.6.4  Installation  ........................................................ 315 
19.6.5  Common Installation Problems ......................................... 315 
19.7  Bored Piles (Drilled Shafts)  .................................................. 315 
19.7.1  Use of Bored Piles (Drilled Shafts) ...................................... 315 
19.7.2  Materials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  316 
19.7.3  Structural Design.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . . . . . . . . . . . . . . . . . .  316 
19.7.4  Installation......................................................... 316 
19.7.5  Common Installation Problems ......................................... 317 
20  Load Testing of Piles........................................... 318 
20.1  Use of a Load Test  ......................................................... 318 
20.1.1  Common Pile Load Test Prqcedures ..................................... 318 
20.1.2  Load Tests during Design ............................................. 321 
20.1.3  Load Test during Construction ......................................... 321 
20.1.4  Routine Load Tests for Quality Control (Inspection) ........................ 321 
20.2  TestArrangement  .......................................................... 322 
20.2.1  Static Load Test. .................................................... 322 
20.2.2  Statnamic Test ...................................r. • • • • • . • . • . . • • . • • •. 322 
20.2.3  Pseudo-Static Load Test  ...............  .............................. 322 
20.3  Static Load Testing Methods  ........ : ..............•......................... 323 
20.3.1  Methods According to  the ASTM Standard  ............................... 323 
20.3.2  Other Testing Methods  ............................................... 324 
20.4  Presentation ofTest Results  .................................................. 325 
  Static Load Test Results  .............................................. 325 
20.4.2  Rapid Load Test Results .............................................. 325 
20.5  Interpretation of Test Results  ................................................. 325 
20.5.1  Interpretation of Static Load Test Results  ................................ 325 

Table of Contents xiii
20.5.2  Interpretation of Rapid Load Test Results ................................. 328 
21  Inspection of Deep Foundations .................................. 331 
21.1  Introduction  .............................................................. 331 
21.2  Documents  ............................................................... 331 
21.3  Location and Alignment ..................................................... 332 
21.3.1  Location........................................................... 332 
21.3.2  Alignment  ......................................................... 332 
21.3.3  Curvature.......................................................... 333 
21.4  Inspection of Pile Driving Operations  .......................................... 335 
21.4.1  Introduction........................................................ 335 
21.4.2  Driving Equipment .................................................. 335 
21.4.3  Piles  ............................................................. 336 
21.4.4  Driving Procedures .................................................. 336 
21.5  Inspection of Compacted Concrete Piles  ........................................ 337 
21.5.1  Introduction  ....................................................... 337 
21.5.2  Equipment  ........................................................ 337 
21.5.3  Installation  ........................................................ 337 
21.6  Inspection of Bored Deep Foundations  ......................................... 338 
21.6.1  Preliminary Infonnation .............................................. 338 
21.6.2  BoringlDrilling  ..................................................... 338 
21.6.3  Concreting  ........................................................ 338 
21.6.4  General  ........................................................... 339 
·22  Control of Groundwater  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  340 
. 22.1  Methods for the  Control and Removal of Groundwater  ............................ 340 
22.2  Gravity Drainage  .......................................................... 340 
22.3  Pumping From Inside the Excavation  .......................................... 340 
22.3.1  Pumping From Unsupported Excavations  ................................ 341 
22.4  Pumping From Outside the Excavation ......................................... 342 
23  Geosynthetics................................................ 346 
23.1  Introduction  .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  346 
23.2  Geotexti1es  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  348 
23.2.1  Hydraulic Properties of Geotextiles, Geonets and Drainage Geocomposites  ..... 350 
23.2.2  Filtration and Separation .............................................. 351 
23.2.3  Dynamic, Pulsating and Cyclic Flow .................................... 352 
23.2.4  In-Plane Drainage ................................................... 353 
23.3  Geogrids  ................................................i ................. 353 
23.4  Strength and Stiffuess Properties ofGeotexti1es and Geogrids  ....................... 353 
23.5  Geosynthetics in Waste  Containment Applications  ................................ 354 
23.6  Geomembranes .................................. : ....... .'  ................. 356 
23.6.1  Other Geomembrane Applications  ...................................... 358 
23.6.2  Selection  .......................................................... 358 
·23.6.3  Seaming.. , ........................................................ 359· 
23.6.4  Installation......................................................... 359 
23.7  Geosynthetic Clay Liners .................................................... 359 
23.8  Wans .................................................................... 359 
23.9  Slopes and Embankments over Stable Foundations ................................ 359 

xiv Canadian Foundation Engineering Manual
23.9.1  Internal Stability .................................................... 359 
23.9.2  External  Stability .................................................... 361 
23.10  Embankments on Soft Ground ...................... " ........................ 361 
23.10.1  Bearing Capacity ........ '"  .................................. , ..... 362 
23.l0.2  Circular Slip Failure  ................................................ 364 
23.10.3  Lateral Embankment Spreading  ....................................... 365 
23.11  Reinforced Embankments on Soft Foundations with Prefabricated Vertical Drains (PVDs) 365 
23.12  Embankments on Fibrous Peats  .............................................. 365 
23.13  Unpaved Roads  over Soft Ground  ............................................ 367 
23.13.1  Reinforcement Mechanisms and Geosynthetic Requirements  ................ 367 
23.13.2  Design Methods  for Unpaved Roads over Cohesive Soils ................... 367 
23.13.3  Unpaved Roads over Peat Soils  ................................... , . , .  370 
23.14  Paved Roads,  Container Yards  and Railways ......... , , . , ....................... 370 
23.14.1  Geotextiles for Partial Separation  ..................................... 370 
23.14.2  Geosynthetics for Granular Base Reinforcement .......................... 371 
23.15  Construction Survivability for Geosynthetics  ................................... 372 
24  Lateral Earth Pressures & Rigid Retaining Structures ................. 374 
24.1  Coefficient of Lateral Earth Pressure, K ......................................... 374 
24.2  Earth Pressure at-Rest  ...................................................... 374 
24.3  Active and Passive Earth Pressure Theories ...................................... 374 
24.3.1  Active Earth Pressure  ................................................ 375 
24.3.2  Passive Earth Pressure ................................................ 377 
24.3.3  Graphical Solutions for Determination of Loads due to Earth Pressures .... , .... 380 
24.4  Earth Pressure and Effect of Lateral  Strain  ...................................... 381 
24.5  Wall Friction .................................. '.' .......................... 382 
24.6  Water Pressure  ............................................................ 383 
24.7  .Surcharge Loading  ...... ',' ................................................ ,  383 
24.7.1  Uniform Area Loads  ................................................. 383 
24.7.2  Point or Line Loads .................................................. 384 
24.8  Compaction-Induced Pressures  ............................................... 385 
24.9  Earthquake-Induced Pressures  .............................................. "  386 
24.10  Frost-Induced Loads ....................................................... 388 
24.11  Empirical Pressures for Low Walls ............................................ 388 
24.12  Design of Rigid Retaining Walls  ............. , ............................... 390 
24.12.1  Design Earth Pressures .............................................. 390 
24.12.2  Effects of Backfill Extent  ............................................ 390 
24.12.3  Backfill Types ..................................................... 391 
25  Unsupported Excavations.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..  394 
25.1  General .................................................................. 394 
25.2  Excavation in Rock  ........................................................ 394 
25.3  Excavation in Granular Soil ................................ : ................. 394 
25.4  Excavation in Clay  ......................................................... 395 
25.4.1  Behaviour of Clays in Excavated Slopes  ................................. 395 
25.4.2  Short-Term Stability ................ "  ............................... 395 
25.4.3  Long-Term Stability  ................................................. 396 
25.4.4  Construction Measures  ............................................... 396 
Table of Contents xv
26  Supported Excavations & Flexible Retaining Structures ............... 397 
26.1  Introduction  .............................................................. 397 
26.2  Earth Pressures and Deformation .............................................. 399 
26.3  Earth Pressures and Time ... '................................................. 400 
26.4  Effects of Seepage and Drainage  .............................................. 401 
26.5  Surcharge Pressures  ........................................................ 401 
26.6  Frost Pressures  ............................................................ 401 
26.7  Swelling/Expansion Pressures  ................................................ 401 
26.8  Cantilevered (Unbraced) Walls ................................................ 403 
26.8.1  Cantilevered Walls  Loading Conditions  ................................ 403 
26.8.2  Cantilevered Walls  Determination of Penetration Depth .................... 404 
26.8.3  Cantilevered Walls - Determination of Structural Design Bending Moments ..... 404 
26.9  Single-Anchor and Single-Raker Retaining Structures  ............................. 405 
26.9.1  Loading Conditions .................................................. 405 
26.9.2  Penetration Depth and Structural Bending Moments ........................ 405 
26.10  Multiple-Anchor, Multiple-Raker and Internally Braced (Strutted) Retaining Structures .. 407 
26.10.1  Loading Conditions ................................................. 407 
26.10.2  Effect ofAnchor Inclination .......................................... 408 
26.10.3  Braced Retaining Structures  Loading Conditions  ........................ 409 
26.10.4  Coarse-Grained Soils  ............................................... 410 
26.10.5  Soft to  Firm Clays .................................................. 410 
26.10.6  Stiff to Hard Clays  ................................................. 410 
26.10.7  Layered Strata ..................................................... 410 
26.11  Stability of Flexible Retaining Systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  411 
26.11.1  Excavation Base Stability ............................................ 411 
26.11.2  Overall Stability ofAnchored Systems .................................. 412 
26.11.3  Overall Stability ofAnchored Systems .................................. 415 
26.11.4  Structural Design ofVertical Members .................................. 415 
26.12  Horizontal Supports -Anchors, Struts and Rakers  ............................... 416 
26.12.1  Struts ............................................................. 416 
26.12.2  Rakers  and Raker Footings ........................................... 418 
26.12.3  Buried Anchors .................................................... 419 
26.12.4  Soil and Rock Anchors  ......................................... "  ... 420 
26.13  Other Design and Installation Considerations  ................................... 428 
26.13.1  Installation of Sheeting .............................................. 428 
26.13.2  Horizontal Spacing and Installation of Soldier Piles  ....................... 428 
26.13.3  Installation of Secant or Tangent Pile (Caisson) Walls ...................... 428 
26.13.4  Installation of Concrete Diaphragm (Slurry) Walls  ........................ 428 
26.13.5  Lagging Design and Installation ....................................... 429 
26.13.6  Excavation Sequences ............................................... 430 
26.13.7  Design Codes and Drawings .................. '"  ..... '"  ........ "  ... 430 
26.14  Alternative Design Methods ................................................. 430 
26.15  Movements Associated with Excavation  ..................... .f • •••••••••••••••• 432 
26.15.1  Magnitude and Pattern of Movements  .................................. 433 
26.15.2  Granular Soils ..................................................... 437 
26.15.3  Soft to Firm Clays .................................................. 437 
26.15.4  Stiff Clay ......................................................... 437 
26.15.5  Hard Clay and Cohesive Glacial Till  ................................... 438 
26.15.6  Means of Reducing Movements ....... , ............................... 438 
26.16  Support for Adjacent Structures .............................................. 438 
xvi Canadian Foundation Engineering Manual
27  Reinforced Soil Walls.  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  440 
27.1  Itroduction  ............................................................... 440 
27.2  Components  .............................................................. 441 
27.2.1  Reinforcement.  ..................................................... 441 
27.2.2  Soil Backfill ........................................................ 442 
27.2.3  Facing  ............................................................ 443 
27.3  Design Considerations: ...................................................... 444 
27.3.1  Site Specific Design Input.  ............................................ 444 
27.3.2  Design Methodology and Approval  ..................................... 444 
27.3.3  External, Internal, Facing and Global Stability ............................. 445 
27.3.4  Wall Deformations  .................................................. 448 
27.3.5  Seismic Design ..................................................... 448 
References ........................................................ 450 
Index  ............................................................ 485 
1  Introduction 
Chapters 2 to 5 of the Canadian Foundation Engineering Manual cover fundamental matters common to all aspects
of foundation engineering, such as notations, definitions of terms and symbols, the classification of soil and rock,
and discussion of special site conditions. During the preparation of this 4th  edition of the Manual by members of
the Canadian Geotechnical Society, a companion document has been under development to focus explicitly on site
characterization. Since the Manual is being published before that companion document, Chapter 4 continues to
include details of site characterization and subsurface investigation in soil and rock. It is likely that a future edition
of the Manual will be modified and cross-reference the Characterization Guidelines.
Chapters 6 to 8 contain general discussions offoundation design, dealing with earthquake resistant design in Chapter
6, a more general discussion of foundation design in Chapter 7, and specific treatment of Limit States Design
methodologies in Chapter 8. The evolution of geotechnical engineering practice has not yet come to a point where
the whole Manual can be converted to a limit states (LSD) or load and resistance factor design (LRFD) framework.
Again, this will be left as a major contribution in a subsequent edition of the Manual when the status of foundation
engineering practice has moved more comprehensively towards the adoption of LSD or LRFD design concepts.
Chapters 9 to 11 deal with strength and deformation ofshallow foundations on rock and soil. Chapters 12, 13, 14 and
15 deal with specific considerations associated with drainage, frost action, machine foundations and foundations on
expansive soils, respectively. Chapter 16 contains a discussion oftechniques for ground improvement in association
with foundation design and construction.
Chapters 17 to 21 deal explicitly with the design of deep foundations. Chapter 22 has a brief discussion associated
with control qf groundwater. Chapter 23 contains a comprehensive discussion ofthe design and use ofgeosynthetics
to solve geotechnical engineering problems. Chapters 24 to 27 deal with earth retaining structures, unsupported
excavations, and supported excavations and flexible retaining structures, including reinforced soil walls.

Canadian Foundation Engineering Manual
Definitions,  Symbols and Units 
The following is a partial list ofdefinitions ofsome ofthe terms commonly used in foundation design and construction, 
which are referred to in this Manual. Other terms are defined or explained where they are introduced in the text.  For 
additional terms, see Bates and Jackson (1980). 
Adfreezing - the adhesion of soil to  a foundation unit resulting from the freezing of soil water.  (Also referred to as 
'frost grip'.) 
Basal heave - the upward movement of the soil or rock at the base of an excavation. 
Bearing pressure, allowable  - in working stress design it is the maximum pressure that may be applied to a soil 
or  rock by the  foundation  unit considered  in  design  under  expected loading  and  subsurface  conditions  towards 
achieving  desired  performance  of  the  foundation  system.  In  limit  stress  design,  allowable  bearing  pressure 
commonly corresponds to serviceability limit states for settlement not exceeding 25  mm towards achieving desired 
performance of the foundation. 
Bearing or contact pressure - the pressure applied to a soil or rock by a foundation unit. 
Bearing pressure for settlement means the bearing pressure beyond which the specified serviceability criteria are 
no longer satisfied. 
Bearing surface - the contact surface between a foundation unit and the soil or rock upon which it bears. 
Capacity  or  bearing  capacity  or  geotechnical  capacity  - the  maximum  or  ultimate  soil  resistance  mobilized 
by a  loaded foundation unit,  e.g.,  a footing,  or a pile.  (The structural  capacity of a  foundation unit is  the ultimate 
resistance of the unit itself as based on the strength of the building materials). 
Deep foundation - a foundation unit that provides support for a structure by transferring loads either by toe-bearing 
to soil or rock at considerable depth below the structure or by shaft resistance in    ~ soil or rock in which it is placed. 
Piles and caissons are the most common type of deep foundation. 
Downdrag - the transfer of load (dragload) to a deep foundation unit by means of negative skin friction, when soil 
settles in relation to the foundation unit. 
Dragload  - the  load transferred to  a deep foundation unit by negative skin friction  occurring when  the soil settles 
in relation to the foundation unit. 

3 Definitions, Symbols and Units
Dynamic method of analysis - the determination of the capacity, impact force, developed driving energy, etc, of a
driven pile, using analysis of measured strain-waves induced by the driving of the pile.
Effective stress analysis - an analysis using effective stress strength parameters and specifically accounting for the
effects of pore water pressure.
Excavation the space created by the removal of sailor rock for the purpose of construction.
Factored geotechnical bearing resistance (of a foundation unit) - the factored resistance of a foundation unit, as
determined by geotechnical formula using unfactored ( characteristic) soil strength parameters to calculate ultimate
capacity (resistance) that is multiplied by an appropriate geotechnical resistance factor, or, the ultimate capacity (as
determined in a field-test loading) multiplied by an appropriate geotechnical resistance factor.
Factored geotechnical bearing resistance means the calculated ultimate (nominal) bearing resistance, obtained
using characteristic ground parameters, multiplied by the recommended geotechnical resistance factor.
Factored Geotechnical Resistance at ULS - the product of the geotechnical resistance factor and the geotechnical
ultimate (nominal) sailor rock resistance.
Factored load - nominal (characteristic) or specified load multiplied by the appropriate load factor.
Factored geotechnical pull out resistance (i.e. against uplift) means the calculated ultimate (nominal) pull out
resistance, obtained using characteristic ground parameters, multiplied by the recommended geotechnical resistance
  resistance (of a foundation unit) the factored geotechnical or structural resistance of the unit.
Factored geotechnical sliding resistance means the calculated ultimate (nominal) sliding resistance, obtained
using characteristic ground parameters, multiplied by the recommended geotechnical resistance factor,
Factor of safety - in working stress design, the ratio of maximum available resistance to the resistance mobilized
under the applied load.
Fill- artificial (man-made) deposits consisting of soil, rock, rubble, industrial waste such as slag, organic material,
or a combination ofthese, which are transported and placed on the natural surface of soil or rock. It mayor may not
be compacted.
Foundation - a system or arrangement ofstructural members through which the loads from a building are transferred
to supporting sailor rock.
Foundation unit - one of the structural members of the foundation of a building such as a footing, raft, or pile.
Frost action - the phenomenon occurring when water in soil is subjected to freezing, which, because of the water-
ice phase change or ice lens growth, results in a total volume increase, andlorthe build-up of expansive forces under
confined conditions, and the subsequent thawing that leads to loss of soil strength and increased compressibility.
Frost-susceptible soil - soil in which significant ice- segregation will occur resulting in frost heave, or heaving
pressures, when requisite moisture and freezing conditions exist.
Geotechnical Reaction at SLS - the reaction of the sailor rock at the deformation associated with a SLS
4 Canadian Foundation Engineering Manual
Geotechnical Resistance at ULS - the  geotechnical ultimate resistance  of soil  or  rock corresponding to  a failure 
mechanism (limit state) predicted from theoretical analysis using unfactored geotechnical parameters obtained from 
test or estimated from assessed values. 
Grade - the average level of finished ground adjoining a building at all exterior walls. 
Groundwater free water in the ground. 
Groundwater, artesian - a confined body of water under a pressure that gives a level of hydrostatic pore pressure 
(phreatic  elevation)  higher than the  top  surface  of the  soil  unit in which the pore  water pressure exists.  Flowing 
artesian corresponds to the condition when the phreatic elevation is higher than the ground surface. 
Groundwater level (groundwater table) - the top surface of free water in the ground. 
Groundwater, perched - free water in the ground extending to a limited depth. 
Hydrostatic pore pressure - a pore water pressure varying as pressure  in a non-moving free  standing column of 
Ice-segregation - the growth of ice in lenses, layers, and veins in  soil, commonly,  but not always, oriented normal 
to the direction of heat loss. 
Lateral pressure (load), design - the maximum pressure (load) that may be applied in the horizontal direction to a 
soil or rock by a foundation unit. 
Load, service - the load actually applied to a foundation unit and which is not greater than the design load. 
Load factor - the factor used to modifY (usually increase) the actual load acting on and from a structure, as used in 
ultimate limit states design. 
Negative shaft resistance - soil  resistance  acting  downward  along  the  side  of a  deep  foundation  unit due  to  an 
applied uplift load 
Negative skin friction - soil resistance acting downward along the side of a deep foundation unit due to downdrag 
Overconsolidation ratio (OCR) - the  ratio  between  the  preconso1idation  pressure  and  the  current  effective 
overburden stress. 
Peat - a  highly  organic  soil  consisting· chiefly  of fragmented  remains  of vegetable  matter  that  is  sequentially 
Pier - a deep foundation unit with a large diameter to length ratio, usually,  a large diameter bored pile or caisson 
Pile - a slender deep foundation unit, made of materials such as wood,  steel, or concrete, or combinations thereof, 
which  is  either  premanufactured  and  placed  by driving,  jacking, jetting,  or  screwing,  or  cast-in-place  in  a  hole 
formed by driving, excavating, or boring.  (Cast-in-place bored piles are often referred to as  caissons in Canada). 
Pile head - the upper end of a pile. 
Pile toe - a  premanufactured separate  reinforcement  attached  to  the  bottom  end  (pile  toe)  of a  pile  to  facilitate 
driving,  to protect the pile toe,  and/or to improve the toe resistance of the pile. 
Definitions. Symbols and Units 5
Pile toe· the bottom end of a pile.
Pore pressure ratio the ratio between the pore pressure and the total overburden stress.
Rock· a natural aggregate of minerals that cannot readily be broken by hand.
Rock shoe· a special type of pile shoe.
Rock quality designation (RQD) - a measure of the degree of fractures in rock cores, defined as the ratio of the
accumulated lengths (minimum 100 mm) of sound rock over the total core length.
Safety factor - a factor modifYing ·reducing· overall capacity or strength as used in working stress design. The
safety factor is defined as a ratio of maximum available resistance to mobilized resistance or to applied load.
Safety margin - the margin (dimensional) between mobilized resistance, applied load, or actual value and maximum
available resistance or acceptable value, e.g., the margin between the mobilized shear stress and the shear strength, .
or the margin between calculated settlement and maximum acceptable settlement.
Shaft resistance - the resistance mobilized on the shaft (side) of a deep foundation. Upward acting is called positive
shaft resistance. Downward acting is called negative shaft resistance (See also negative skin friction).
Shallow foundation· a foundation unit that provides support for a building by transferring loads to soil or rock
located close to the lowest part of the building.
Site investigation (characterization)· the appraisal ofthe general subsurface conditions by analysis ofinformation
gained by such methods as geological and geophysical surveys, in-situ testing, sampling, visual inspection, laboratory
testing of samples of the subsurface materials, and groundwater observations and measurements.
Slaking - crumbling and disintegration of earth material when exposed to air and moisture.
Soak-sensitive soil - soil which, when saturated, or near saturated, and subjected to a shearing force, will lose all or
part ofits strength. The dominant grain size fraction in this soil is usually medium and coarse silt. Soak-sensitive soil
is frost-susceptible soil and, if ice-segregation occurs, when thawing it will become very soft and slough easily.
Soil - that portion of the earth's crust which is fragmentary, or such that some individual particles ofa dried sample
can be readily separated by agitation in water; it includes boulders, cobbles, gravel, sand, silt, clay, and organic
Specifications - project specific requirements indicating applicable codes, standards, and guidelines. Normally,
Performance Specifications stipulate the end-results without detailing how to achieve them, whereas Compliance or
Prescriptive Specifications detail mandatory methods, materials, etc. to use.
Total stress analysis - an analysis using undrained soil parameters and not the influence of pore water
2.2 Symbols
Wherever possible, the symbols in the Canadian Foundation Engineering Manual are based on the list that has been
prepared by the Subcommittee on Symbols, Units, and Definitions of the International Society of Soil Mechanics
and Foundation Engineering (ISSMFE, 1977, and Barsvary et aI, 1980).
6 Canadian Foundation Engineering Manual
2.2.1 The International System of Units (51)
In the SI-System, all parameters such as length, volume, mass, force, etc. to be inserted in a formula are assumed to
be inserted with the value given in the base unit. It is incorrect to use formulae requiring insertion of parameters in
other dimensions than the base units, because this would require the user to memorize not just the parameter, but also
its "preferred" dimension, which could vary from reference to reference. For instance, in the well-known Newton's
law, F rna, force is to be inserted in N, mass in kg, and acceleration in mls
• Thus, a force given as 57 MN must
be inserted as the value 57 x 10
, In other words, the multiples are always considered as an abbreviation ofnumbers.
This is a clear improvement over the old system, where every formula had to define whether the parameter was to be
input as lb, tons, kips, etc. Therefore, unless specifically indicated to the contrary, all formulae given in the Manual
assume the use of parameters given in base SI-units.
The term mass in the SI-System is used to specifY the quantity of matter contained in material objects and is
independent of their location in the universe [Unit =kilogram (kg); the unit Mg to indicate 1000 kg should not be
used, as gramme (g) is not a base unit; nor should the unit tonne be used].
The term weight is a measure of the gravitational force acting on a material object at a specified location, [unit =
newton O\l"); standard gravity at sea level = 9.81 mls
, In practical foundation engineering applications, the gravity
constant is often taken as equal to 10 mls2].
The term ~ n   t weight in the SI System is the gravitational force per unit volume [Unit N/m
The term density refers to mass per unit volume [Unit kg/m3].
Stress and pressure are expressed as the force per unit area (N/m2= Pa). The unit kilopascal (kPa) is commonly used
il2- Canadian practice.
A prime denotes effective stress (e.g., cr')
A bar above a symbol denotes an average property (e.g., u)
A dot above a symbol denotes a derivative with respect to time, also referred to as rate (e.g., !:i).
For symbols indicating force, an upper case letter is used for total force, or force per width or linear length, and a
lower case letter is used for force per unit area, i.e., pressure, stress, shear resistance.
Normally, when the abbreviating symbols are not used, the units Newton, metre, kilogram, and second are spelled
without plural endings (e.g., 50 kiloNewton, 200 metre, etc.)
Table 2.1 contains a list ofterms, symbols, SI units, and recommended multiples for Canadian practice. The numeral
I in the unit column denotes a dimensionless quantity.
For a complete table, see Barsvary et al. (1980).
7  Definitions,  Symbols and  Units 
TABLE 2.1  List ofTerms, Symbols, S.I.  Units, and Recommended Multiples
m  (km,  mm, ).lm) 
m  (km, mm, ).lm) 
m  (km, mm, ).lm) 
m  (km, mm, ).lm) 
m  (km, mm, ).lm) 
m  (km, mm, ).lm) 
m  (km, mm, ).lm) 
m  (km, mm, ).lm) 

, ha, cm
, mm


,  mm


Pa, N/m

(MPa, kPa) 
Pa, N/m

(MPa, kPa) 
N  (MN, kN) 
degree Celsius eC) 
J, Nm (kJ, kNm) 
Nm (MNm, kNm) 
Unit and  Recommended 



Breadth, width 
Planar coordinates 
Polar coordinates 
Gravity acceleration 
Unit weight 
Pressure, stress 
Shear stress 
Force, load 
Energy, work 
Moment of Force, torque 
Safety factor 
2.718  (base of natural logarithm.) 
Logarithm base 10 
II  - Physical  Properties 

Density and Unit Weights
Unit weight 
Density of solid particles 
Unit weight of solid particles 
Density of Water 
Unit weight of water 
L,  I 
H, h 





S  b  I 
ym  0



8 Canadian Foundation Engineering Manual

Dry density 
Dry unit weight 
Saturated density 
Saturated unit weight 
Void ratio 
Water content 
Degree of saturation 
Relative density{formerly specific gravity) 
Liquid limit 
Plastic limit 
Shrinkage limit 
Plasticity index 
Liquidity index 
Consistency index 
Void ratio in loosest state 
Void ratio  in densest state 
Density index (formerly relative density) 
Grain Size
Cirain diameter 
n percent diameter 
Curvature coefficient 
Hydraulic Properties
Hydraulic head or potential 
Rate offiow 
Flow velocity 
Hydraulic gradient 
hydraulic conductivity (permeability) 
Seepage force per UJ;lit volume 
- Physical Properties
S b I
ym 0

P sat 








III - Mechanical Properties

T 5 b I
e  ym 0
In-Situ Tests
Cone tip-resistance 
Local side-shear 
Standard penetration test (SPT) index 
Dynamic cone penetrometer blow count 
Pressuremeter limit pressure 
Pressuremeter modulus 
Unit and Recommended


m(mm, f!m) 
m(mm, f!m) 
Unit and Recommended
Pa (kPa) 
Pa (kPa) 
blows/O.3 m 
blows/O.3  m 
Pa (kPa) 
Pa (kPa) 
"                 ___"'d 
9  Definitions,  Symbols and  Units 
III  - Mechanical Properties 
Term  -----,
Effective cohesion intercept 
Apparent cohesion intercept 
Effective angle of internal friction 
Undrained shear strength 
Residual shear strength 
Remoulded shear strength 
Uniaxial compressive strength 
Tensile strength 
Point load settlement index 
Consolidation (One-Dimensional)
Coefficient of volume change 
Compression index 
Recompression il}dex 
Coefficient of secondary consolidation 
Modulus number 
Recompression modulus number 
S:velling index 
Permeability change index 
Coefficient of consolidation (vertical) 
Coefficient of consolidation (horizontal) 
Time factor; vertical drainage 
Time factor; horizontal drainage 
Degree of consolidation 
Preconso 1  idation Pressure 
Pore pressure 
Pore-water pressure 
Pore-air pressure 
Total normal stress 
Effective normal stress 
Shear stress 
Principal stresses (major, intermediate and minor) 
Average stress or octahedral normal stress 
Octahedral shear stress 
Linear strain 
Volumetric strain 


su'  C










IV - Stress and Strain 

b  I

0' l' 0'2' 0'3

Pa (MPa, kPa) 
Pa (MPa, kPa) 
Pa (MPa, kPa) 
Pa (MPa, kPa) 
Pa (kPa) 
Pa (MPa, kPa) 
Pa (MPa, kPa) 


/s (cm2/s) 
/s (cm2/s) 
Unit and Recommended 
M  It"  I
u  .pes 
Pa (kPa) 
Pa (kPa) 
Pa (kPa) 
Pa (MPa, kPa) 
Pa (MPa, kPa) 
Pa (MPa, kPa) 
Pa  kPa) 
Pa (MPa, kPa) 
Pa (MPa, kPa) 

10 Canadian Foundation Engineering Manual
IV  . Stress and Strain 
b  I
Term  ym  0
Shear strain 
Principal strains (major,  intermediate,  and minor) 
£1'  £2'  1::3
v Poisson's ratio 
E Modulus of linear deformation 
Elastic axial deformation 
11 Displacement 
G Modulus of shear deformation 
K Modulus of compressibility 
M Tangent modulus 
M Secant modulus 

m Modulus number 
Stress exponent 
Coefficient of Friction 
Coefficient of viscosity 
V - Design  Parameters 
E'arth Pressure

S  b  I 
ym  0
Earth pressure thrust, total:  active and passive 
Pa'  Pp 
Earth pressure, unit: active and passive 
Pa' Pp 
Angle of wall friction 
Coefficient of active and passive earth pressure 
Coefficient of earth pressure at rest  K 
Coefficient of earth pressure acting against a pile shaft  K

Breadth of foundations  B 
Length of foundation  L 
Depth of foundation beneath ground  D 
Total length of a pile  L 
Embedment length of a pile 

Diameter of a pile  B, b 
Applied load 

Applied vertical load 
Applied horizontal load 
Applied (axial) pressure 
Settlement  s,  S 
Eccentricity of load  e 
Inclination of load 
Modulus  of subgrade reaction 
Bearing capacity coefficients 
'  N
,  Ny,  Nt 
Unit and Recommended 
Pa (GPa, MPa, kPa) 
m  (mm,  /lm) 
m(mm, /lm) 
Pa (GPa, MPa, kPa) 
Pa (GPa, MPa, kPa) 
Pa (GPa, MPa, kPa) 
Pa (GPa, MPa, kPa) 
Unit and  Recommended 
N  (kN, MN) 
Pa (kPa, MPa) 

N  (MN,kN) 
N  (MN,kN) 
Pa (MPa, kPa) 
Definitions, Symbols and Units 11
v -Design Parameters
S b I
ym 0
Unit and Recommended
M It" I
U Ipes
Vertical height of slope H m
Depth below toe of earth slope to hard stratum D m
Angle of slope to horizontal
Dip of planar rock joint
qJ degrees
Depth to water table z
Pore-pressure ratio r
Maximum dry density
Maximum wet density
Optimum dry density
Water content at optimum dry density w
12 Canadian FoundationEngineeringManual
Greek Letter Notations
Alpha A a secondary(subscript)
Beta B angleofslopetohorizontal

Gamma r y shearstrainunitweight
Delta 11 0 angleofwallfriction
Displacement deflection
Epsilon E B strain 
Zeta Z 
Eta H viscositycoefficient
Theta e e 
Iota I 1 
Kappa K 1C 
Lambda A A.
Mu M
Nu N v Poisson'sratio
Xi ,::.
Omicron 0 0
Pi IT 1t 3.14 
Ro P p density 
Sigma 2: (J pressure,stress 
Tau T 1: Shearstress,strength 
Ypsilon Y u
Phi <D
Chi X

Omega n Q) 
Identification and Classification of Soil and Rock 13
Identification and Classification of Soil and Rock
3. Identification and Classification of Soil and Rock
3.1 Classification of Soils
3.1.1 I ntrod u ction
Soil is that portion of the earth's crust that is fragmentary, or such that some individual particles of a dried sample
may be readily separated by agitation in water; it includes boulders, cobbles, gravel, sand, silt, clay, and organic
matter. There are three major groups of soils:
Coarse-grained soils - containing particles that are large enough to be visible to the naked eye. They include
gravels and sands and are often referred to as cohesionless or non-cohesive soils.
Fine-grained soils - containing particles that are not visible to the naked eye. They are identified primarily on the
basis of their behaviour in a number of simple indicator tests. They include silts and clays. Clays are often referred
to as cohesive soils.
Strictly defined, coarse-grained soils are soils havmg more than 50% ofthe dry weight larger than particle size 0.075
mm (see Subsection, and fine-grained soils are soils having more than 50% of the dry weight smaller than
particle size 0.075 mm.
Organic soils - containing a high natural organic content.
3.1.2 Field Identification Procedures
The following procedures and tests may be carried out in the field to identify and describe soils. Coarse-Grained Soils or Fractions
Coarse-grained soils are easily identified in the field because the individual particles large enough to be visible to
the naked eye. The smallest particles that may be distinguished individually are approximately 0.1 mm in diameter
(approximately the size of the openings of the No. 200 sieve (0.075 mm) used in the laboratory identification test).
Coarse-grained soils and silts are identified on the basis of grain size diameter as follows:
Silt - particles of size 0.002 - 0.060 mm
Sand - particles of size 0.06 -2.0 rom
Gravel - particles of size 2 - 60 mm
Cobbles - particle's of size 60 - 200 mm
Boulders - particles >200mm
14  Canadian Foundation Engineering  Manual 
Thesilt, sand,andgravelfractions arefurtherdividedintofine,medium,andcoarseproportions,asfollows:
Silt: Fine 0.002- 0.006mm
Medium 0.006- 0.020mm
Coarse 0.020- 0.060mm
Sand: Fine 0.06- 0.20mm
Medium 0.20- 0.60mm
Coarse 0.60- 2.00mm
Gravel: Fine 2.0- 6.0mm
Medium 6.0- 20.0mm
Coarse 20.0- 60.0mm
Otherphysical properties ofsoils that may influence engineeringcharacteristics should also be identified. They
• Gradingdescribesparticlesizedistribution.Asoilthathasapredominanceofparticlesofonesizeis'poorly
graded',whereassoilthathasparticlesofawiderangeofsizeswithnodominatingsizeis 'wellgraded'.
• Shapeandsurfaceconditionsof grains:particlesmaybeplaty,elongated,orequidimensional,andtheymay
• A qualitative term describing the compactness condition ofa cohesionless soil is often interpreted from
theresultsofaStandardPenetrationTest(SPT). Thistestisdescribedinmore detailinSubsection4.5.2.
Compactness and penetration values are often related according to Table 3.1, which was proposed by
TerzaghiandPeck(1967). Noticethattheterm"compactnesscondition"replacestheearlierterm"relative
TABLE 3.1  Compactness Condition ofSands from Standard Penetration Tests
(blows  per 0.3 m) 
0 4 
4 10 
10 30 
Other relationships between the SPT N-index and the compactness condition attempt to take into account the
magnitude ofthe overburden pressure at the sampling depthto be taken into consideration. Three sets ofsuch
correlations are now available: themost commonlyused setwasproposedbyGibbs andHoltz(1957), but ithas
beenmodifiedbySchultzeandMelzer(1965). '
1. the SPTN-indexmustbeindependentof theoperatorandtheboringmethod;
2. the correlation between the SPT N-index and the compactness condition must be accurate to within 

Identification and Classification of Soil and Rock 15
3. the same correlation between the SPT N-index and the compactness condition must'be used by all.
N one of these conditions is fully satisfied. It must be recognized, therefore, that the SPT is a very subjective test,
and different operators can report substantially different N-values without the differences necessarily corresponding
to actual variables in soil condition. A recent improvement in the testing method has been the adoption by some
countries of a free-failing trip-hammer. Fine-Grained Soils or Fractions
These procedures are to be performed on the soil fraction passing sieve No. 40, the openings ofwhich are about 0.4
mrn in diameter. For field classification purposes screening is riot required because the coarse particles that interfere
with the tests are simply removed by hand. Dilatancy (reaction to shaking)
After removing particles larger than No. 40 sieve size, prepare a pat of moist soil with a volume of about 10 cm. If
necessary, add enough water to make the soil soft but not sticky. Then, place the pat in the open palm of one hand
and shake horizontally, striking vigorously against the other hand several times. A positive reaction consists of the
appearance ofwater on the surface of the pat, which changes to a livery consistency and becomes glossy. When the
sample is squeezed between the fingers, the water and gloss disappear from the surface, the pat stiffens, and finally
cracks or crumbles. The rapidity of appearance of water during shaking and of its disappearance during squeezing
assist in identifying the character of the fines in a soil. Very fine, clean sands give the quickest and most distinct
reaction, whereas a plastic clay has no reaction. Inorganic silts, such as a typical rock flour, show a moderately quick
3.1,.2.2(2) Dry Strength (crushing characteristics)
After removing particles larger than No. 40 sieve size, mould a pat of soil to the consistency ofputty, adding water
if necessary. Allow the pat to dry completely by oven, sun, or air drying, and then test its strength by breaking
and crumbling between the fingers. This strength Is a measure of the character and quantity of the clay fraction
contained in the soil. The dry strength increases with increasing plasticity:
High dry strength is characteristic for inorganic clays of high plasticity. Typical inorganic silt possesses only very
slight dry strength. Silty fine sands and silts have about the same slight dry strength, but can be distinguished by the
feel when powdering the dried specimens. Fine sand feels gritty, whereas typical silt has the smooth feel offlour. Toughness (conSistency near plastic limit)
After removing particles larger than the No. 40 sieve size, a specimen of soil about 10 cm in volume is molded to the
consistency ofputty. Iftoo dry, water must be added and, if sticky, the specimen should be spread out in a thin layer
and allowed to lose some moisture by evaporation. Then the specimen is rolled out by hand on a smooth surface or
between the palms into a thread about 3 mm in diameter. The thread is then folded and rolled repeatedly. During
the manipulation, the moisture content is gradually reduced and the specimen stiffens, until it is no longer malleable
and crumbles. This indicates that the plastic limit has been reached. After the thread has crumbled, the pieces should
be lumped together and a slight kneading action continued until the lump crumbles. The tougher the thread near the
plastic limit and the stiffer the lump when it finally crumbles, the more active is the colloidal clay fraction in the soil.
Weakness of the thread at the plastic limit and quick loss of coherence of the lump below the plastic limit indicate
either inorganic clay oflow plasticity, or materials such as kaolin-type clays and organic clays (which occur below
the A-line in the plasticity chart; see Figure 3.1.
:;.',: :
16 Canadian Foundation Engineering Manual

'" 0.. 
w 50
WL =30
50  60  70  80  90  100
FIGURE 3.1 The plasticity chart (after Casagrande, 1948)
Highly  organic  clays  have  a  weak  and  spongy feel  at  the  plastic  limit.  Other physical properties  of fine-grained 
soils, which may  influence their engineering characteristics,  should also  be identified.  Typical such properties are 
as fQ1lows: Consistency of Cohesive Soil at Natural Water Content
TABLE 3.2 Approximate Consistency ofCohesive Soils
Very soft 
Very stiff 
Field Identification
Easily penetrated several centimeters by the fist 
Easily penetrated several centimeters by the thumb 
Can be penetrated several centimeters by the thumb with  moderate effort 
Readily indented by the thumb but penetrated only with great effort 
Readily indented by the thumb nail 
Indented with difficulty by the thumbnail 
The  consistency notations  given qualitatively  in Table 3.2 are  similar to those defined by values  of shear strength 
in  Table  3.3,  below.  However,  the field  identification methods  in  Table 3.2  are not  suitable  for  the  quantitative 
determinations of soil strength. Discontinuities
Discontinuities of the undisturbed soil should be identified, such as bedding, the presence ofjoints, cracks, fissures, 
or slickensides, and evidence of weathering or cementation, and thickness, orientation,  and distortion. 
- .... - - - - - - - - - - -     ' { - . ~ - - - - - - - - - - - - ~ ~ - - - - - - - - ~ - - - - ~ . ~ ~ ~ ~ - - ~ ~ - - - - - - , . .     ~ - - ~ -
Identification and Classification of Soil and Rock 17  Colour 
Colour may be described by the Munsell system (Goddard,  1979).  Odour 
Odour, if any, can provide evidence of the presence of organic material.  Organic Soils 
These are readily identified by colour, odour, spongy feel and frequently by fibrous texture. 
3.1.3  Laboratory Identification Tests  Grain-Size Tests 
In the laboratory, grain-size tests are carried out according to a test method, which includes procedures for analysis 
of coarse-grained soils (i.e., fractions larger than 0.075 rom) by sieving, and the analysis offine-grained soils by the 
hydrometer test (ASTM D422). 
The results ofthe grain size test are used to classify the soil beyond the rough separation into fine grained and coarse 
grained. The classification is  based on amounts by weight within the respective grain-size fractions, as follows: 
noun  gravel, sand, silt, clay  > 35 % and main fraction 
"and"  and gravel, and silt, etc.  >35 %
adjective  gravelly, sandy,  silty, clayey, etc.  20 %-35 % 
"some"  some sand, some silt, etc.  10%-20% 
"trace"  trace sand, trace silt, etc.  1 %  -10 % 
A  soil with 30 %  clay, 45  %  silt,  18  %  sand,  and 7  % gravel would thus be named "clayey silt,  some sand, trace 
gravel." However,  the  clay  fraction  in such a  soil  forms  the  dominant matrix,  and a  soil of this  composition will 
behave geotechnically much like a clay soil. Some classification systems base the description on the plasticity chart. 
For example, if the Ip  and w
for the soil were to plot above the A-Line, the description would be silty clay, some 
sand trace gravel.  Atterberg  Limits 
The  range  of water content,  called plasticity  index,  Ip  wL - W p'  over which  a  fine-grained  soil  is  plastic,  is  an 
important indicator of its probable engineering behaviour.  The Atterberg limits, w p  =plastic limit and w L =liquid 
limit, defining these water contents are determined in accordance with the standard ASTM methods (ASTM D423 
'and ASTM D424, respectively).  The liquid limit can also be determined by the Swedish fall-cone test (Garneau and 
Lebihan,  1977).  The preparation of soil samples for these tests should be determined according to  Procedure B of 
the ASTM Standard Method for "Wet Preparation of Soil  Samples  for Grain-Size Analysis and Determination of 
Soil Constants" (ASTM D2217)., 
The liquid limit, w L' is used to classify clays and silts as to degree of plasticity, as  follows: 
Low  degree ofplasticity  w
Medium  degree of plasticity  30  <w <50
High  degree of plasticity  50  <w
In the plasticity diagram, Figure  3.1, the liquid limit is  combined with the plasticity index.  Experience has  shown 
that soils with similar origin and properties plot in specific areas  in the diagram,  which makes the diagram a very 
useful  tool for identifying and classifying fine-grained soils. 
18 Canadian Foundation Engineering Manual Classification by Undrained Strength
Fine-grained  soils  can be  classified in broad terms  in relation to  undrained strength  going  from  very  soft  to  hard 
consistency (see Table 3.3).  Originally, the table was based on results from unconfined compression tests.  Today, 
however,  field  and  laboratory  vane  tests,  laboratory  fall-cone  tests,  shear-box  tests,  unconfined  compression  and 
triaxial  and  other  test  methods  may  be  used.  This  implies,  of course,  that  the  classification  is  somewhat 
arbitrary, as different tests do not give the same values of strength.  It also implies that when the  consistency values 
are given, the testing method should be identified. 
Commonly, the  consistency  and undrained shear strength of clay soils  is  correlated to the  SPT N-Index values  as 
shown in Table 3.3  (Terzaghi  and Peck 1967). It is  noted that this  correlation needs to be used with caution as the 
correlation is only very approximate. 
TABLE 3.3 Consistency and Undrained Shear Strength ofCohesive Soils
Very soft 
Very stiff 
Undrained Shear Strength
<  12 
12 - 25 
50 - 100 
100 - 200 
Spt N-Index
2  4 
8 - 15 
15  - 30 
>30 Classification by Sensitivity
Sensitivity  is .  an  important  characteristic' of fine-grained  soils.  It is  defined  as  the  ratio  of intact  to  remoulded 
undrained shear strength, and is measured in the laboratory by means ofthe Swedish fall-cone test or in the field by 
means of the vane test. 
Classes of sensitivity may be defined as follows: 
low sensitivity  S  < 10 
medium sensitivity  10  < St <40 
high sensitivity  40 <St Density Index 10
The density index, ID, of cohesionless soils is defined as 
1 1
ID '"  Pdmax X Pd - P4m1n
Pd Pdmax - Pdmm
Pdmln Pdmax
Identification and Classification of Soil and Rock 19
The reference densities (P
min and P
max) or the void ratios (e
and e
), corresponding to the loosest and the
densestconditionofthematerialunderconsideration, arenot definedinthestrictsenseofthewordbecause they
areessentiallyrelatedtothemethodusedformeasuringthem. Intoday'spractice,alargenumberof methodsarein
use,buttheASTMD4253 andD4254StandardMethodaregenerallypreferred.
The in-situ dry density, P (or voidratio, e,) ofthe soil can be measured directly by the "sandcone" or 'rubber
bymeansof anappropriatesamplingmethod,anundisturbedsampleofthecohesionlessmaterialmightberetrieved
fordirectmeasurementofits density.
A sampleofthe soilis usedtodetermine inthe laboratorythe minimumandmaximumdensities bymeans ofan
appropriate testing method, preferably the ASTMD2049 Standard. From these values, the density index canbe
calculated. Tobeof practicalvalueindesign,themeasurementof allinputdensitiesmustbe:
• independentofthetestingmethod; 
independentof theoperator;and 
of asuitableaccuracy. 
Recent investigations have shown that these conditions are not fully satisfied. The density index is therefore to
be regarded as very approximate, to be used only in conjunctionwith experience andconsiderable engineering
theearlierterm"relativedensity." References
and'FoundationEngineering. This standard is similarin manyrespects to theUnified Soil Classification System
(USCS) usedintheUnitedStates(Casagrande, 1948). Sometimesusedin Canadais the systemoftheAmerican
Association ofState Highway and Transportation Officials (AASHTO), which differs inthe definitions ofsoil
classesfrom the uses.A comprehensivecompilationofdifferent classificationsystemswaspublishedbyHoltz
Standards for the testing and laboratory,classification ofsoils in Canada follow closely the standards ofthe
AmericanSocietyforTestingandMaterials(ASTM).Standards ofparticularimportancefortheidentificationand
classificationofsoilandrockarefound intheAnnualBookofASTMStandards,Section4,Construction,Volume
3.2 Classification of Rocks
3.2.1 Introduction
Rockisanaturalaggregateof mineralsthatcannotbereadilybrokenbyhandandthatwillnotdisintegrateonafirst
formedbysetsofdiscontinuitiesorpossiblyevenbyasinglediscontinuity(Wyllie, 1992).It isusual,therefore, to
investigatethestructuralgeology ofasitethoroughly,andto distinguishbetweenthepropertiesoftheintactrock
andthepropertiesof themuchlargerrockmass,whichincludestheeffectsof therockdiscontinuities.
Theinfluenceof thediscontinuitiesuponthe materialstrengthdependsuponthescaleofthefoundationrelativeto
In thefollowingtext,referenceismadetothestandardsforrocktestingdevelopedandpublishedbytheInternational
20 Canadian Foundation Engineering Manual
Society for Rock Mechanics (ISRM). 
3.2.2 Geological Classification
Rock is classified with respect to its geological origin or lithology as  follows: 
Igneous rocks, such as granite, diorite and basalt, which are formed by the solidification of molten material, 
either by intrusion of magma at depth in the earth's crnst, or by  extrusion oflava at the earth's surface; 
•  Sedimentary rocks, such as sandstone, limestone and shale, which are formed by lithification of sedimentary 
soils; and 
Metamorphic  rocks,  such as  quartzite,  schist  and  gneiss,  which  were  originally  igneous, metamorphic  or 
sedimentary rocks,  and which have been altered physically and sometimes chemically or mineralogically, 
by the application of intense heat and/or pressure at some time in their geological history. 
3.2.3 Structural Features of Rockmasses
Geological  structures  generally  have  a  significant  influence  on  rockmass  properties,  increasing  the  rockmass 
deformability  and reducing  the  rockmass  strength,  as  compared to  the  deformability  and strength  of intact  rock. 
In  some  cases,  discontinuities provide  planes  of weakness  along  which slip  or excessive  deformation  can  occur, 
leading to structurally controlled failure of the mass.  Some important definitions follow: 
An aggregate ofblocks ofsolid rock material containing structural features that constitute mechanical discontinuities. 
Any in-situ rock with all of its inherent geomechanical discontinuities. 
Rock material or intact rock
The  consolidated  aggregate  of mineral  particles  forming  solid  material  between  structural  discontinuities.  The 
pieces may range from a few millimeters to several meters in size. 
Structural discontinuities
All  geological features  that separate solid blocks of the  rockmass,  such  as joints, faults,  bedding planes,  foliation 
planes, cleavage planes, shear zones  and solution cavities. These features  are weaker than the intact rock,  thereby 
reducing  the  strength  of the  rockmass  and  increasing  its  deformability.  A  list  of the  different types  of rockmass 
discontinuities and their characteristics is given in Table 3.4. 
Major discontinuity or Major structure
A structural discontinuity .that is sufficiently well developed and continuous such that shear failure  along it will not 
involve shearing of    ~ intact rock. 
3.2.4 Engineering Properties of Rock Masses
The  quality  of a  rockmass  for  foundation  purposes  depends  mainly  on  the  strength  of the  Intact  rock  material 
and on the spacing, persistence,  aperture,  roughness, filling,  weathering and orittntation of the discontinuities,  The 
presence  of groundwater within the  discontinuities  may  also  alter the  strength of the rockmass.  The  influence  of 
discontinuities  on rockmass characteristics and strength is  further discussed by Hoek, Kaiser and Bawden (1995). 
Engineering properties ofrockmasses can be determined from methods for estimating joint strength, for estimating 
rockmass strength and deformability or from rockmass classification systems. Strength of Intact Rock
Strength is the maximum stress level that can be carried by a specimen.  Rocks can be classified on the basis of their 
intact strength using values ranging  from  extremely weak to  extremely strong as  defined by the approximate  field 
strength criteria set out in Table 3.5. The strength.grades are related to Uniaxial Compressive Strength,  O"d'  and to 
Identification and Classification of Soil and Rock 21
the Point Load Strength Index. Uniaxial Compressive Strength is determined from tests on'prepared cylindrical
samples of intact rock as per the ISRM Standard (1979). Alternatively, strength can be determined from pieces of
core or from irregularly shaped, unprepared samples of rock, using the Point Load test as per the ISRM Standard
(1985). Additional strength testing on core can be by triaxial tests (ISRM, 1978; ASTM D2664-86) or by tensile
strength tests (Brazilian Test, ISRM 1978; Direct Tension Test, ISRM 1978; ASTM D2936-84).
TABLE 3.4 Rockmass Discontinuity Descriptions (after Hunt, 1986)
Disconti nuity Definition Characteristics
A separation in the rockmass, a break.
A fracture along which there has been no
observable relative movement.
Signifies joints, faults, slickensides, foliations
and cleavage.
Most common defect encountered. Present in
most formations in some geometric pattern
related to rock type and stress field.
Open joints allow free movement of water,
increasing decomposition rate of mass.
Tight joints resist weathering and the mass
decomposes uniformly.
A fracture along which there has been an
observable amount of displacement.
Fault zones usually consist of crushed and
sheared rock through which water can move
relatively freely, increasing weathering.
Faults generally occur as parallel to sub-parallel
sets of fractures along which movement has
taken place to a greater or lesser degree.
Foliation planes
Pre-existing failure surface: from faulting,
landslides, expansion.
Continuous foliation surface results from
orientation of mineral grains during
Stress fractures from folding.
Shiny, polished surfaces with striations. Often
the weakest elements in a mass, since strength is
often near residual.
Can be present as open joints or merely
orientations without openings. Strength and
deformation relate to the orientation of applied
stress to the foliations.
Found primarily in shales and slates; usually
very closely spaced. '
Bedding planes Contacts between sedimentary rocks.
Often are zones containing weak materials such
as lignite or montmorillonite clays.
Intensely sheared zone.
Openings in soluble rocks resulting from
groundwater movement, or in igneous rocks
from gas pockets
Strong laminations: original mineral constituents
and fabric crushed and pulverized.
In limestone, range from caverns to tubes. In
rhyolite and other igneous rocks, range from
voids of various sizes to tubes.
. 22 Canadian Foundation Engineering Manual
TABLE 3.5 Classification ofRock with Respect to Strength (after Marinos and Hoek, 2001)
Specimen can only 
Fresh basalt, chert, diabase, 
> 10  be chipped with a  > 250 R6 
gneiss,  granite, quartzite 
geological hammer 
Specimen requires  Amphibolite, sandstone, 
basalt, gabbro, gneiss,  many blows of a
100  250 R5  Very strong  4  10
geological hammer to  granodiorite, peridotite, 
fracture  it  rhyolite,  tuff 
Specimen requires 
more than one blow of  Limestone, marble, 
2-4 R4  Strong  50 - 100 
a geological hammer  sandstone, schist 
to fracture  it 
Cannot be scraped or 
peeled with a pocket 
Medium  knife, specimen can  Concrete, phyllite, schist, 
25  -50 R3  I 2 
be fractured with a  strong  siltstone 
single blow from  a 
geological hammer 
R2 Weak  *** 5  25 
Can be peeled with 
a pocket knife  with 
difficulty, shallow  Chalk, claystone, potash, 
indentation made by  marl, siltstone, shale, 
a firm blow with the  rocksalt 
point of a geological 
Rl  Very weak  5  *** 
Crumbles under firm 
blows with point of a 
geological hammer, 
can be peeled with a 
pocket knife 
Highly weathered or altered 
. rock, shale 
0.25  - 1  ***  Indented by thumbnail  Stiff fault gouge 
Grade according to ISRM (1981). 
All rock types exhibit a broad range of uniaxial compressive strengths reflecting heterogeneity in 
composition and anisotropy in structure. Strong rocks are characterized by well-interlocked crystal fabric and 
few voids. 
Rocks with a uniaxjal compressive strength below 25 MPa are likely to yield highly ambiguous results 
under point load testing. 
Some natural materials, which geologically may be referred to as rock, should be treated from an engineering point 
ofview as  soils.  Some   x ~ m p l   s of materials that fall into this category include: 
Identification and Classification of Soil and Rock 23
Soft or weakly cemented rocks with unconfined or uniaxial compressive strength < 1 MPa;
Any material that can be dug by hand with a shovel;
Cemented sands and gravels, in which the cementing is discontinuous; and
Rocks such as: marl and volcanic tuff, highly altered or crushed rocks, rocks with closely spaced continuous
joints, and residual soils containing rock fragments.
The strength of sedimentary rocks derived from clay and silt sized paliicles, such as shale or mudstone, generally
degrades when exposed to repeated cycles of wetting and drying. The slake durability test can be used to determine
whether the rock will degrade, and if so, how rapidly this will occur. Standards for the slake durability test are
provided by the ISRM (1979).
Characteristics of Discontinuities
The structural integrity of a rockmass will be affected by the presence of discontinuities. Major, discrete, through-
going structures such as shears, faults or other major weakness zones will dominate the rockmass behaviour where
they are present. Ubiquitous (present everywhere) structure will also affect the behaviour ofthe rockmass. Systems
of extension joints and minor shear structures will have formed under historical stress fields, which were relatively
consistent over a local region. As a result, there are usually several distinct groups of similarly oriented structures
within a rockmass, termed joint sets or joint families. Ungrouped joints are defined as random. Both discrete and
ubiquitous features should be measured, characterized and analysed.
Full characterization of a rockmass requires measurement of a number of characteristics of the discontinuities,
including discontinuity orientation (Section, discontinuity strength (Section and discontinuity
spacing (Section Guidance for description of discontinuities in rockmasses is provided by the ISRM
(1978) .
. Discontinuity Orientation
Discontinuities are considered to be adversely oriented ifthey provide minimal or limited resistance to sliding under
the applied load. Joint orientation can be found from logging drill cores, surveying boreholes and/or from mapping
surface exposures of the rockmass.
To determine joint orientation from core logging, measurements must be made on oriented core. It is essential that
t4e orientation of the borehole be recorded. It is then necessary to take two measurements of orientation for each
joint or discontinuity: alpha, the minimum angle between the maximum dip vector ofthe discontinuity and the core
axis, and beta, the dip direction of the plane, measured clockwise from north or the reference line for the core. The
true orientation of the discontinuities with respect to north can then be calculated following procedures defined by
Priest (1985).
Jo:i;nt orientation can also be found by surveying the drill holes. Surveys can be conducted by inspecting the holes
with borehole cameras, periscopes or probes. Generally, the orientation of each feature can be determined by ,
the angle the feature makes with the hole, and the length of the inscribed circle or oval created by
dIscontinuity around the perimeter of the borehole. The calculation of the true orierttation of the feature depends
Upon both the orientation and the diameter of the drill hole.
•.• of joint characterjstics can also be carried out on exposures of the rockmass on outcrops, or in other
where the rock is exposed, such as shafts, trenches and adits. In these locations, the dip and dip
direction of each discontinuity can be measured directly on surface exposures of each structure, using a geological
  It is important when mapping rockmass exposures that the length of the sampling window, or scan line,
IS of sufficient length to sample enough features to provide a statistically valid basis for analysis. A minimum of
local measurements are normally required to define the structure in a localized zone of rock (Hutchinson and
Diederichs, 1996). .
24  Canadian  Foundation Engineering Manual 
Priest and Hudson (1976) suggest that between 150 and 350 measurements should be-taken at a number of sample 
locations, selected to provide data about different lithologies, or about highly variable discontinuity characteristics_ 
When establishing a mapping program it is  important to consider the following  issues: 
Increased numbers of measurements improve the data precision as well  as  coniidence in the output 
Increased length of sampling or scan lines leads to  increased precision in the data. 
•  Measurements  taken  from  scan  lines  of similar  orientation  will  be  subject  to  data  bias_  Therefore  it  is 
advisable  to  orient  successive  scanlines  in  different  orientations  where  possible,  and  to  COlTect  for  bias 
(Terzaghi,  1965). 
Where the  rockrnass  quality and  nature are variable,  it is important to separate  the  data into  sub-sets,  on the basis 
of distinct geological conditions, if possible. For example, where discontinuities have been measured in a rockrnass 
comprising two  distinct and substantial lithologies, the structural analysis should be calTied out on the full data set, 
and then on sub-sets divided on  the basis oflithology, to determine if the structural patterns are different. 
As  noted  previously,  it  is  important  to  distinguish  between  discrete  and  Ubiquitous  structures  in  analysis  of the 
rockrnass  stability  and  strength.  The  ubiquitous  structures  can  generally be  grouped  into  one  or more  sets  with 
similar  orientation.  Random joints  may  also  be  present  in  the  rockrnass.  The  visual  examination  and  statistical 
grouping of structural data into sets is  best accomplished using a stereonet.  The  outcome  of this work is generally 
a representative  (mean) orientation for each cluster or set ofjoint data.  Further information regarding information 
plotting and data analysis on stereonets is provided by Hoek et al  (1995) and Priest (1993). Discontinuity Strength
Discontinuity  strength  can  be  defined  using  several  distinct  formulations.  These  include  the  strength  criterion 
•  proposed by Barton and Choubey (1977) and further discussed by Hoek at a1  (1995), as well as the simplified Mohr-
Coulomb analysis,  requiring input parameters of friction,  <p,  and cohesion, c, discussed by Wyllie (1992). 
The  most  accurate measurement of discontinuity  strength  is  made by performing direct shear tests,  which  can be 
calTied  out in the laboratory or in-situ< on undisturbed samples.  Guidelines  for performing these  tests  are  given by 
Wyllie (1992) and the  ISRM (1974). 
The strength of a discontinuity depends upon the roughness, persistence, and aperture, as well as  upon the presence 
of any  iniilling  or water.  Each  of these  parameters,  defined below,  should  be  measured  during  any  geotechnical 
mapping program. 
Roughness of a  discontinuity  adds  to  its  resistance  to  shear,  especially  when  the  asperities  on  one  side  of the 
discontinuity interlock with those on the other side. 
The importance of surface roughness declines as  the aperture, filling thickness and previous displacement along the 
discontinuity increase.  Roughness  is  generally measured by comparing observations  to published surface profiles 
providing an estimate  of the  Joint Roughness  Coefficient (JRC)  (Barton,  1973;  Barton and Choubey,  1977; Hoek 
et aI,  1995). 
Roughness  can be divided into  small-scale and larger-scale roughness.  The small-scale roughness,  measured over 
a sample distance  of up to  10  cm,  is  defined as  rough,  smooth or polished (slickensided).  Roughness  at the metre 
scale is termed stepped, undulating or planar.   
Joint persistence is an estimate of the length of each individual joint. Joints may range from non-persistent or not 
continuous,  through  to  highly  persistent or  fully  continuous.  Joints,  which  are  highly  persistent (long),  are  more 
likely to  combine with other structures to  form  large free blocks of rock, than are short joints . 

Joint  aperture  is  the  perpendiCUlar  distance  separating  the  adjacent,walls  of an  open  discontinuity,  which  may 
Identification and Classification 01 Sail and Rack 25
be water  filled.  Other fillings  of the  discontinuity  should  be  described separately, as  discussed  in  the  next point. 
Aperture provides an indication ofthe secondary permeability ofthe rockmass as well as some idea of its looseness. 
Unfortunately, apertures that can be observed directly are usually disturbed by blasting, excavation and weathering. 
Observations ofthe less disturbed rockmasses exposed within boreholes, using a borehole camera or periscope, can 
be very useful. 
Where possible,  the joint aperture  should be  measured using  feeler gauges, or a  measuring  tape,  and  classified as 
shown in Table 3.6.  Impression packer testing can also  be  used to  provide a measurement of the aperture as well. 
TABLE 3.6 Classification 0/Joint Aperture
Joint Aperture
<0.5 mm 
0.5  to  10mm 
> 10mm 
Where permeability  of the joints is  of importance,  in-situ permeability  testing  should be  carried  out.  During  the 
mapping  program,  observations  of any  evidence  of current  or  previous  water  flow  along  the  joints  should  be 
recorded. Classification ofthe joints based on these observations can be made using Table 3.7. 
TABLE 3.7 Classification o/Discontinuities depending upon Water Flow
Class Description
1  Water flow  not possible 
2  No evidence of water flow 
3  Evidence ofwater flow  (e.g.  rust staining) 
4  Dampness 
5  Seepage 
6  Flow (volume per unit  of time) 
Joint infilling is the material separating the adjacent rock walls of discontinuities.  It may be formed by the in-situ 
weathering  or alteration of the rock adjoining  the  discontinuity,  or it may be transported.  It may be described by 
the  methods  used  for  the  field  identification  of soils  (see  Section  3.1).  The width of the  filled  discontinuity,  the 
mineralogy ofthe infilling, and the roughness ofthe discontinuity walls will all affect the strength and deformability 
of the  discontinuity  and should  be  examined  and  described.  Water flow  can be described  in terms of the  classes 
shown in Table 3.8. 
TABLE 3.8 Classification 0/Filled Discontinuities depending upon Water Flow Proposed by the ISRM (1981) 
Class Description
1  Filling is dry and has low permeability 
2  Filling is damp; no free water is  present 
3  Filling is wet;  drops of free water are present 
4  Filling shows outwash; continuous flow  of water is present 
5  Filling is locally washed out and there is considerable water flow  along channels 
!  ~ ; ; L ; : ~ ) h

26  Canadian Foundation  Engineering Manual Discontinuity Spacing
Discontinuity spacing is important because closely spacedjoints result in a smaller block size, increasing the
potentialfor internalshiftingandrotationastherockmassdeforms,andtherebyreducingstability.
DiscontinuityspacingisdefinedbyPriest(1993) as thedistancebetweena pairofdiscontinuitiesmeasuredalong
a lineofspecifiedlocationandorientation(orscanline). Hedefines threemaintypes ofdiscontinuityspacingsas
1. Totalspacingisthespacingbetweenapairofimmediatelyadjacentdiscontinuitiesmeasuredalonga line
2. Set spacing is the spacing between a pair ofimmediately adjacent discontinuities from a particular
discontinuityset,measuredalongalineof anyspecifiedorientation.
3. Normalsetspacingisthesetspacingmeasuredparalleltothemeannormaltotheset.
Thespacingofdiscontinuitiescanvaryfromextremelywidetoextremelyclose,asshowninTable3.9. Inthiscase,
the distance betweenadjacentdiscontinuities is measured overa samplinglengthnotshorterthan3 meters. The
samplinglengthshouldbegreaterthantentimestheestimateddiscontinuityspacing,if possible(ISRM, 1981).
TABLE 3.9 Classification ofRock with Respect to Discontinuity Spacing (ISRM, 1981)
Spacing Classification
Very close 
Spacing Width (m)
0.02to 0.06
0.06to 0.20
0.20to 0.6
0.6 to 2.0
2.0 to6.0
RockQualityDesignation,orRQD,originallyproposedbyDeereetal.(1967)isanindirectmeasureof thenumber
offractures within a rockmass. The method provides a quick and objective technique for estimating rockmass
qualityduringdiamonddrillcorelogging,as showninTable3.10.
RQD(%)=  LLengthof corepieces>10 cm x 100 
Totallengthof corerun 
TABLE 3.10 Classification ofRock with Respect to RQD Value
Rqd Classification Rqd Value ('Yo)
50to 75 
75to 90 
Identification and Classification of Soil and Rock 27
If the core is broken by handling or during drilling (Le., the fracture surfaces are fresh, irregular breaks rather than
tural joint surfaces), the fresh, broken pieces should be fitted together and counted as one piece. Some judgment
?anecessary in the case of thinly bedded sedimentary rocks and foliated metamorphic rocks, and the method is not
18 precise in these cases as it is for igneous rocks or for thickly bedded limestones or sandstones. The system has
applied successfully to shales, although it is necessary to log the cores immediately upon removing them from
the core barrel, before air-slaking and cracking can begin.
The procedure obviously penalizes rockmasses where recovery is poor. This is appropriate because poor core
recovery usually reflects poor quality rock. Poor drilling equipment and techniques can cause poor recovery. For
this reason, double-tube core barrels of at least NX (54 mm in diameter) must be used, and proper supervision
of drilling is imperative. It is noted that the original definition for RQD Index was based on N size core.
Philosophically, RQD provides a crude estimate ofthe percentage of the rockmass which can be expected to behave
in a fashion similar to a laboratory sample (typically 10 cm long). Rockmass with a low RQD « 50%) has few
intact blocks larger than 1 0 cm. In such rockmasses, joints and fractures dominate the rock's response to stress. The
strength and stiffness of the rock, as determined in the lab, has little relevance here. On the other hand, rockmasses
with RQD > 95% possess strength and stiffness much closer to the values. obtained in the lab. Joints may still
dominate behaviour, especially in the low stress environments of most foundations. A semi-empirical technique for
evaluating rockmass strength and deformability is discussed in the following section.
A great deal of work has been done to correlate RQD with joint frequency, rockmass stiffuess and other properties.
The interested reader is referred to Deere and Miller (1966), Deere and Deere (1988), Cording and Deere (1972),
Coon Merritt (1970) and Bieniawski (1979). Jointed Rockmass Strength and Deformability
The strength of the rockmass will depend on such factors as the shear strength of the surfaces of the blocks defined
by; cliscontinuities, their continuous length, and their alignment relative to the load direction (Wyllie, 1992). If the
loads are great enough to extend fractures and break intact rock, or if the rockmass can dilate, resulting in loss of
between the blocks, then the rockmass strength may be diminished significantly from that of the in-situ
Where foundations contain potentially unstable blocks that may slide from the foundation, the shear strength
parameters of the discontinuities should be used in design (Section, rather than the rockmass strength.
Direct measurements of rockmass deformability are best conducted in-situ for foundations carrying substantial
lol'tds, for example major bridge footings. The tests available include borehole jacking tests, plate load tests and
jacking tests for the rockmass modulus. Direct shear tests are used to determine the shear strength of the
fractures. Further details regarding these tests and the use of the data so derived are provided by Wyllie (1992).
He also notes that the test results should be checked against values calculated from the performance of other
foundations constructed in similar geological conditions.
and deformation properties ofjointed rockmasses can be estimated using the Hoek-Brown failure criterion
  and Brown, 1997) from three parameters (Hoek and Marinos, 2000; Marinos and Hoek, 2001):
The uniaxial compressive strength of the intact rock elements contained within the rockmass (see Section
A constant, mi, that defines the frictional characteristics of the component minerals within each intact rock
The Geological Strength Index (GSI) which relates the properties of the intact rock elements to those of the
overall rockmass (Table 3.12).
The generalized Hoek-Brown failure criterion is defined as: (J I 0"; +(J Ci(mb 0"; + sr (3.1)
O"d J
28  Canadian  Foundation Engineering  Manual 
where  (J'1 and (J'3 are the maximum and minimum effective stresses at failure 
(J .  is  the uniaxial compressive strength of the intact rock pieces 
is the value of the Hoek-Brown constant m  for the rockmass, and 
(GSI 1001
\: 28  ) 

is the Hoek-Brown constant for the intact rock (Table 3.11) 
S  and a are constants which depend upon the rockrnass characteristics. 
(GSI -100]
For GSI > 25, a =  0.5, and  s  =  exp \  9  ) 

0.65- GSI
II  For GSI < 25, s  0,  and  a
The defonnation modulus for weak rocks (cr < 100 MPa), can be estimated from the following equation (Marinos 
and Hoek, 2001):. r-- It 0/  )
fa ci   (3.2)
Marinos and Hoek (2001) caution that this criterion is only applicable to 'isotropic' rockrnasses, wherein the strength 
of the whole mass controls its behaviour. In  anisotropic rockmasses, such as a strong, blocky sandstone, where the 

blocks are separated by clay coated and slickensided bedding surfaces, the rockmass behaviour is controlled by the 
The Hoek-Brown constant, m
, can be detennined from triaxial testing of core samples, using the procedure discussed 
by Hoek et al  (1995), or from the  values given in  Table  3.11.  Most of the values provided in  the table have been 
•  derived  from  triaxial  testing  on  intact  core  samples.  The  ranges  of values  shown  reflect  the  natural  variability 
in  the  strength  of earth  materials,  and depend  upon  the  accuracy  of the  lithological  description  of the  rock.  For 
example,  Marinos  and Hoek (2001)  note  that  the  tenn granite describes  a  clearly defined rock type  that  exhibits 
very similar mechanical characteristics, independent of origin. As a result, mj for granite is defined as 32 ± 3. On the 
other hand, volcanic breccia is  not very precise in tenns of mineral  composition, with the result that mj is given as 
19 ±  5, denoting a higher level of uncertainty. The ranges ofvalues depend upon the granularity and interlocking of 
the crystal structure.  The higher values are associated with tightly interlocked and more frictional characteristics. 
Values for  the Geological Strength Index (GSI), which relates  the  properties  of the  intact rock elements to  those 
of the overall  rockmass,  are provided in Table 3.12.  A  similar table,  developed for  heterogeneous rockmasses,  is 
provided by Marinos and Hoek (2001). Rockmass Classification
A  number of classification systems have been  developed to provide the basis for engineering characterization  of 
rockmasses. An excellent overview of these techniques is provided by Hoek et al.  (1995). Most of the classificatioIY 
systems incorporating a number of parameters (Wickham et aI.,  1972; Bieniawski,  1973, 1979,  1989; Barton et aI., 
1974),  were  derived from  civil engineering case  histories  in which all compohents  of the  engineering geological 
character  of the  rockmass  were  considered.  More  recently,  the  systems  have  been  modified  to  account  for  the 
conditions affecting rockmass stability in underground mining situations. 
While no single classification system has been developed for or applied to foundation design, the type of infonnation 
collected  for  the  two  more  common  civil  engineering  classification  schemes,  Q  (Barton  et  aI,  1974)  and  RMR 
(Bieniawski, 1989) should be considered. These techniques have been applied to empirical design situations, where 
previous  experience plays a  large  part in the design  of the  excavation in  the  rockmass.  Empirical techniques  are 
not used in foundation engineering, where a  more concentrated expenditure of effort and resources is  required and 
possible,  due  to  the much smaller spatial extent of the work, and the relatively high external  loads  applied  to  the 
Identification  and  Classification of Soil  and  Rock  29 
TABLE 3.11 Values a/Hoek-Brown Constant mJor Intact Rock, by Rock Group (after Marinos and Hoek, 2001)
Siltstone (4±2)
Breccia * 
( 18±3) Marl
Foliated ** 
Values in parentheses are estimates.
*  Conglomerates and breccias may have a wide range ofvalues, depending on the nature ofthe cementingmaterial
and the degree of cementation. Values 'range between those of sandstone and those of fine-grained
**  These values are for intact rock specimens tested normal to bedding or foliation. Values of m; will be
significantly different if failure occurs along a weakness plane.
30  Canadian Foundation Engineering Manual 
TABLE 3.12 GSI Estimatesfor Rockmasses, from Hoek and Marinos (2000)
From the letter codes describing the stucture 
and suface of the rock mass (from Table 4),  pick 
the appropriate box in this chart. Estimate the 
average value of the Geological strength index 
(GSI) from the contours. Do not attempt to be too 
precise. Quoting a range GSI from 36 to 42 is 
more realistic than stating that GSI = 3S. 
BLOCKY· very well interlocked 
undisturbed rock mass consisting 
of cubical blocks formed by three 
orthogonal discontinuity sets 

VERY BLOCKY· interlocked, 
partially disturbed rock mass with 
multifaceted angular blocks formed 
by four or more diseontlnuity sets 


BLOCKYIDISTURBED • folded  .... 
and/or faulted with angular blocks 
formed  by many intersecting 

discontinuitty sets 


DISINTEGRATED - poorly interlocked, 
heavily broken rock mass with a 
mixture of angular and rounded rock 
.  pieces 

Site Investigations 31
Site Investigations
4. Site Investigations
4.1 Introduction
A site investigation involves  the appraisal and characterization of the general subsurface conditions by analysis  of 
information  gained by such methods  as  geological and  geophysical  surveys,  drilling  boreholes,  and sampling,  in-
situ testing, laboratory testing of samples of the subsurface materials, groundwater observations, visual inspection, 
and local experience. 
The site investigation is one of the most important steps in any foundation design,  and should be carried out under 
the direction  of a person with knowledge  and experience  in planning and  executing such investigations.  Drilling 
crews should be experienced specifically in borings for  geotechnical explorations.  A valuable guide is provided by 
ASCE (1976). 
4.2' Objectives of Site Investigations
An engineer  requires  sufficient knowledge  of the  ground conditions  at a  site  to  estimate the  response  of the soils 
or rocks  to changes induced. by the  site works.  Peck (1962) noted that the three factors  of most importance  to  the 
successful practice of subsurface engineering were: 
Know ledge of precedents 
A working knowledge of geology 
Knowledge of soil mechanics. 
A  knowledge  of precedents in  similar  ground conditions  helps  to  ensure  that no  surprises are  encountered in the 
design  and  construction of the  works;  knowledge  of geology  should  enable  the  engineer to  anticipate  the  range 
of possible  variations  in  ground  conditions  between the  locations  of any  borings;  and knowledge  of soil  or rock 
mechanics should minimize the chances of inadequate performance of the ground during and after construction. 
A  site  characterization  should  be  carried  out  for  all  projects.  The  level  of detail  of any  characterization  should 
be  appropriate to  the proposed site  use  and to the  consequences  of failure  to  meet the  performance requirements. 
The engineer should be able  to  prepare  a  design that will  not exceed ultimate  and serviceability limit states  (see 
Chapters  7  and  8  for  further  discussion).  This  means that there  should be no  danger of catastrophic collapse  and 
deformations and other environmental changes should be within tolerable limits. Depending on the particular nature 
of the proposed development, the site characterization mayor may not involve field exploration. 
Once  the  scope of work has been 'established for  the proposed engineering works,  the site characterization should 
comprise three components: 
Desk Study and Site Reconnaissance 
Field Exploration 
32 Canadian Foundation EngineeringManual

The first component is the most critical. It consists ofa review ofexisting infonnation about the site including
the geology. Attentionto detail in this phase in conjunctionwith asite reconnaissancetoreview existing surface
ofthisphaseof theworkwilldependontheexperienceoftheengineerintheparticulargeologicalenvironmentand
Uponcompletionofthisphase,apreliminarysub-surfacemodelofthesiteshouldhavebeenestablished, enabling
consideration offoundation design issues and preliminary selection offoundation options. The engineer may
proceedtoplananappropriatefield exploration.
Theprimaryobjectivesoffieldexplorationaretodetennine(j.S accuratelyasmayberequired:
•  thenatureandsequenceofthesubsurfacestrata;
•  thegroundwaterconditions atthesite;
thephysicalpropertiesofthesoilsandrockunderlyingthesite; and
otherspecific infonnation, when needed, such as the chemical composition ofthe groundwater, and the
characteristicsof thefoundations ofadjacentstructures.
Siteinvestigationsshouldbeorganizedto obtainallpossibleinfonnationcommensuratewithprojectobjectivesfor
athoroughunderstandingofthesubsurfaceconditions andprobablefoundationbehaviour.Additional infonnation
ontheobjectives,planningandexecutionofsiteinvestigationsisprovidedbyBecker(200I). ,
'phase and shouldprovide sufficientcharacterizationofmaterialproperties to allow estimationofthe response of
the sitetotheproposedengineeringworks. Inmany cases,themacrostructure ofthe groundsuchasjointingand
fissuring willcontrolthesite and foundationperfonnanceduring andafterconstruction. Anunderstandingofsite
geologywillallowtheengineertoanticipatesuchcasesandfield explorationshoulddetenninethepresenceofany
layers orzones likelyto cause difficultyduring construction or operationofthe facility. Forexample, thinweak
layers may be critical for stability or thin penneable layers may be critical in excavations. The selection ofan
Upon completion ofthe stratigraphic logging and material classification, appropriate design parameters can be
•  Experiencewithsimilarfoundations insimilargroundconditions,
Correlation with the known properties of soils or rocks from other sites with similar classification
4.3 Background Information
the type ofstructure to be built, its intended use, characteristics ofthe structure, intended construction
method,startingdate, andestimatedperiodofconstruction;
• the behaviour ofexisting structures adjacent to the site, as well as infonnation available through local
• theprobablesoilconditionsatthesitebyanalysisofgeologicalandgeotechnicalreports and maps,aerial
Site Investigations 33
4.4 Extent of Investigation
4.4.1 Introduction
The extent of the ground investigation is determined by the soil type and variability of soil and groundwater, the
type of project, and the amount of existing information. It is important that the general character and variability of
the ground be established before deciding on the basic principles of the foundation design of the project.
The combination of each proj ect and site is likely to be unique, and the following general comments should therefore
be considered as a guide in planning the site investigation and not as a set of rules to be applied rigidly in every
The greater the natural variability of the ground, greater will be the extent of the ground investigation required
to obtain an indication ofthe character ofthe ground. The depth of exploration is generally determined by the nature
of the project, but it may be necessary to explore to greater depths at a limited number of locations to establish the
overall geological conditions.
The investigation should provide sufficient data for an adequate and economical design of the project. It should
also be sufficient to cover possible methods of construction and, where appropriate, indicate sources of construction
materials. The lateral and vertical extent of the investigation should cover all ground that may be significantly
affected by the project and construction, such as the zone of stressed ground beneath the bottom of a group of piles,
and the stability of an adj acent slope, if present.
The boreholes should be located so that a general geological view of the whole site can be obtained with adequate
details of the engineering properties of the soils and rocks and of groundwater conditions. More detailed information
should be obtained at the location of important structures and foundations, at locations of special engineering
difficulty or importance, and where ground conditions are complicated, such as suspected buried valleys and old
landslide areas. Rigid, preconceived patterns of boreholes should be avoided. In some cases, it will not be possible
to locate structures until much of the ground investigation data has been obtained. In such cases, the program of
investigations should be modified accordingly. In the case oflarger projects, the site investigation is often undertaken
in stages. A preliminary stage provides general information and this is followed by a second stage and, if required,
additional stages as the details of the project and foundation design develop.
Reference is made to boreholes as the means of site investigation. However, in some cases, boreholes can be replaced
by, or supplemented by, test pits, test trenches, soundings or probe holes. Regardless of the type of investigation, it
is essential that the locations and ground levels for all exploration points be established, if necessary, by survey.
Information and recommendations on the extent of site investigations, both depth and number of boreholes, can be
found in various references. The references that have served as the basis for some ofthe comments presented in this
section include ASCE (1976), British Standards Institution, BS 5930 (1981) and Navfac DM 7.01 (1986).
Robertson (1997) suggested the risk-based approach to characterization shown on Figure 4.1. For low risk projects
  to medium sized jobs with few hazards and limited consequences of failure);f it is only necessary to classify
the soils visually and, perhaps, by index testing to allow selection of design parameters. Design may then be based
qn presumptive bearing pressUres. For medium risk projects, some form of in-situ testing will be necessary. The
in-situ testing conventionally consists of penetration testing from which some estimate of the soil properties can be
obtained by correlation. Design methods are also available where in-situ test results are used directly to select design
values of bearing pressure. Where the consequences of unexpected ground response result in an unacceptable level
of risk, a much more elaborate field and laboratory program should be carried out.
Suggestions for the depth of boreholes and spacing of boreholes are considered in the following sections. The
suggestions for minimum depth of boreholes can be more definitive since there is a logical analytical basis. The
minimum depth is related to the depth at which the increase in soil stress caused by foundation loads is small and

34  Canadian Foundation Engineering Manua! 
will not cause any  significant settlement.  The suggestions  for  spacing of boreholes  are  however,  more  difficult to 
make and less definitive since much depends on the soil variability, type ofproject, performance requirements, and 
foundation type selected. 
4.4.2 Depth of Investigation
The  site  investigation  should  be  carried  to  such a  depth that the  entire  zone  of soil  or  rock  affected  by  changes 
caused by the  structure or  construction will be  adequately explored.  The following recommendations are provided 
as guidelines: 
A commonly used rule of thumb for minimum depth of boreholes is to extend the boreholes to such a depth 
that the  net increase  in  soil  stress  under the  weight  of the  structure  is  less  than  10%  of the  applied  load, 
or less than 5 % of the  effective stress  in the  soil at that depth, whichever is  less.  A  reduction in the depth 
can be  considered  if bedrock or dense  soil is  encountered within the  minimum depth.  In the case  of very 
compressible  normally  consolidated clay  soils  located at  depth,  it  may  be  necessary  to  extend boreholes 
deeper than determined by the  10 % and 5 % rules. 
The net increase in soil stress should appropriately take into account the effect of fill or excavation that may 
be required for  site grading. 
The soil stress increase should take into account adjacent foundations since they may increase the soil stress 
at depth, and the corresponding minimum depth of boreholes. 
Boreholes  should  extend  below  all  deposits  that  may be  unsuitable  for  foundation  purposes  such  as  fill 
ground, and weak compressible soils. 
The minimum borehole depth beneath the  lowest part of the foundation generally should not be  less than 6 
m, unless bedrock or dense soil is encountered at a shallower depth. 
If rock is found the borehole  should penetrate at least 3 m  in more than one borehole to  confirm whether 
bedrock or a boulder  has  been found.  Three meters may not be  adequate  for  some geological conditIons; 
e.g.,  where large slabs  of rock may occur as  rafts in till deposits. No guidance can be given in  such cases 
but where doubt arises, consideration should be  given to drilling deeper boreholes. 
•  In the case of end bearing piles on rock, the boreholes should be deep enough to establish conclusively the 
presenc.e ofbedrock as considered previously. Furthermore, the boreholes or selected number ofboreholes 
should be extended to  a sufficient depth  to minimize the  possibility of weaker strata occurring below the 
bedrock surface which could affect the performance of piles. In addition, when weathered rock is  present, 
the boreholes should extend to a sufficient depth into the unweathered rock. 
Since  the  foundation  type  and  design  is  not  always  finalized  at  the  beginning  of the  site  investigation, 
it  may  be  prudent  to  drill  holes  deeper  than  originally  estimated  to  allow  some  variation  during  project 
Not all boreholes need to be drilled to the  same depth since shallower intermediate boreholes may provide 
adequate information for more  lightly loaded foundations.  Also, the level of detailed sampling and  in-situ 
testing may vary considerably from borehole to borehole, depending on the design needs. 
Pile-supported  rafts  on  clays  are  often  used  solely  to  reduce  settlement.  In  these  cases,  the  depth  of 
exploration is governed by the need to examine all strata that could contribute significantly to the settlement. 
A  commonly used approximation in settlement calculations for piled rafts is to assume that the entire load 
is  carried on an imaginary raft located at a depth equal to two-thirds of t   ~ pile length. The borehole depth 
should extend to the level at which the soil stress increase from the imaginary raft is small and will not cause 
significant  settlement.  In practice,  on many  occasions,  this  would  lead  to  an  excessive  and unnecessary 
depth of exploration so the engineer directing the investigation should terminate the exploration at the depth 
where the relatively incompressible strata have been reached. 
•  Fill ground,  and weak  compressible  soils  seldom contribute to the  shaft resistance  of a pile and may  add 
downdrag  to  the  pile  load.  The  entire  pile load,  possibly with the  addition  of downdrag,  will have  to  be 
borne by the  stronger  strata lying below the weak materials.  This will increase the stress  at the bottom of 
the piles and consequently the corresponding depth of boreholes. 
•  For driven  pile  foundations  the  length  of the  piles  is  not  known  with  any  accuracy  until  installation  of 
test piling or construction begins.  Selection of the  depth of boreholes  should make  an allowance  for  this 
Site Investigations 35
uncertainty.  General guidance can be provided from previous experience in the area. 
If any structure is likely to be affected by subsidence due  to mining or any other causes, greater exploration 
depths than those recommended above may be  required. 
Ground Investigation
I n-situ testing
& Disturbed samples
• In-situ testing
e.g. SPT. CPT (SCPTu).
• Possibly specific tests
e.g. PMT. FVT
• Index testing
e.g. Atterberg limits, grain
size distribution, em;n!emax, Gs 
Preliminary Site Evaluation
e.g. geologic model, desk study,
risk assessment
Ground Investigation
Same as for low risk
projects, plus the following:
Additional specific
in-situ tests
Basic laboratory
testing on
selected bulk
Preliminary ground
Same as for low risk
projects, plus the following:
• Identify critical
Additional in-situ tests
High quality
undisturbed samDles
High quality laboratory
testing (response)
• Undisturbed samples
• In-situ stresses
• Appropriate stress path
• Careful measurements
FIGURE 4.1. Generalizedjlow chart to illustrate the likely geotechnical site investigation
based on risk (after Robertson, 1997)
4.4.3 Number and Spacing of Boreholes
Determination of the  minimum  depth  of boreholes  has  a  logical  basis  which  is  related  to  the  depth  at which  the 
increase  in  soil  stress  caused by  the  foundation  loads  is  small  and  will not  cause  any  significant settlement.  The 
basis  for  determining  the spacing  of boreholes  is  less  logical,  and  spacing  is  based more  on the  variability of site 
conditions, type ofproject, performance requirements, experience, and judgment. More boreholes and closer spacing 
is generally recommended for sites which are  located in  less developed areas where previous experience  is sparse 
or non-existent. The following comments are given for planning purposes.  The results of the site investigation may 
indicate more   o ~ p l e x foundation soil conditions which may require additional boreholes. 
For buildings smaller than about  1000  m

in plan area but larger than about 250  m
,  a minimum of four boreholes 
where  the  ground  surface  is  level,  and  the  first  two  boreholes  indicate  regular  stratification,  may  be  adequate. 
Five boreholes are generally preferable (at building comers and centre), and especially if the site is not level.  For 
buildings  smaller than  about 250  m
,  a minimum of three  boreholes  may  be  adequate.  A  single ,borehole  may  be 
sufficient for  a  concentrated foundation  such  as  an  industrial process  tower base  in  a fixed  location with  the  hole 
made at that location, and where the  general stratigraphy is known from nearby boreholes. 
36 Canadian Foundation Engineering Manual
The use of a single borehole for even a small project should be discouraged and not considered prudent except
for special circumstances as noted above, otherwise three boreholes is the minimum. The results of one borehole
can be misleading, for example, drilling into a large boulder and misinterpreting as bedrock. Many experienced
geotechnical engineers know from direct experience or have personal knowledge that the consequences of drilling a
single borehole can be significant. In practical terms, once a drill rig is mobilized to the site, the cost of an additional
one or two boreholes is usually not large.
The preceding comments are intended to provide guidance on the minimum number of boreholes for smaller
structures where the perfonnance of the foundations are not particularly critical. Drilling of than the suggested
minimum number of boreholes should have a sound technical basis.
The determination of the number of boreholes and spacing for larger, more complex, and critical projects fonns a
very important part of the geotechnical design process, and cannot be covered by simple rules which apply across
the entire country. Establishing the scope of a geotechnical investigation and subsequent supervision requires the
direction of an experienced geotechnical engineer.
4.4.4 Accuracy of Investigation
Subsurface investigations should call for a variety of methods to determine the soil properties critical in design.
In particular it is good practice, whenever possible, to use both field and laboratory tests for soil strength and
compressibility determinations. The accuracy of the stratigraphy, as determined by geophysical methods such as
seismic reflection or refraction, or resistivity measurements, should always be checked by borings or other direct
4.5 In-Situ Testing of Soils
4.5.1 Introduction
The physical and mechanical properties ofsoils are determined either by in-situ or laboratory testing or a combination
of both. Both approaches have advantages, disadvantages, and limitations in their applicability.
The measurement of soil properties by in-situ test methods has developed rapidly during the last two decades.
Improvements in equipment, instrumentation, techniques, and analytical procedures have been significant.
In-situ test methods can be divided into two groups: logging methods and specific methods.
Commonly, the logging methods are penetration-type tests which are usually and economical. When based on
empirical correlations, logging methods provide qualitative values ofvarious geotechnical parameters for foundation
design. Specific methods are generally more specialized and often slower and more expensive to perform than
the logging methods. They are normally carried out to obtain specific soil parameters, such as shear strength or
deformation modulus.
The logging and the specific methods are often complementary in their use. The logging methods are best suited
for stratigraphic logging with a preliminary and qualitative evaluation of the soil parameters, while the specific
methods are best suited for use in critical areas, as defined by the logging methods, where more detailed assessment
is required of specific parameters. The investigation may include undisturbed sampling and laboratory testing.
The logging method should be fast, economic, continuous, and most importantly, repeatable. The specific method
should be suited to fundamental analyses to provide a required parameter. One of the best examples of a combination
of logging and specific test methods is the cone penetrometer and the pressuremeter.
Reviews of in-situ testing techniques and their applicability have been published by several   e.g., Mitchell
et al. (1978), Campanella and Robertson (1982), and Lunne, et al. (1989). Common in-situ techniques are listed
in Table 4.1.
Site Investigations 37
4.5.2 Standard Penetration Test (SPT)
The introduction in the United States in 1902 of driving a 25-mm diameter open-end pipe into the soil during the
wash-boring process marked the beginning of dynamic testing and sampling of soils. Between the late 1920s and
early 1930s, the test was standardized using a 51-mm O.D. split-barrel sampler, driven into the soil with a 63.5-kg
weight having a fall of760 mm. The blows required to drive the split-barrel sampler a distance of300 mm, after
an initial penetration of 150 mm, is refelTed to as the SPT N value. This procedure has been accepted internationally
with only slight modifications. The number of blows for each of the three 150-mm penetrations must be recorded.
The Standard Penetration Test (SPT) is useful in site exploration and foundation design. Standard Penetration Test
results in exploratory borings give a qualitative guide to the in-situ engineering properties and provide a sample of
the soil for classification purposes. This information is helpful in determining the extent and type of undisturbed
samples that may be required.
TABLE 4.1 Summary ofCommon In-Situ Tests
Type of test
applicable to
Properties that
can be determined
Remarks , References*
Test (SPT)
Soft to firm
Qualitative evaluation
of compactness.
Qualitative comparison
of subsoil stratification.
(See Section
ASTM D 1586-84 Peck et
al. (1974) Tavenas (1971)
Kovacs et a1. (1981)
ESOPT II (1982)
ISOPT (1988)
Schmertmann (1979)
Skempton (1986)
Test (DCPT)
Sand Clay
Qualitative evaluation
of compactness.
Qualitative comparison
of subsoil stratification.
(See Section
ISSMFE (1977b, 1989)
Ireland et a1. (1970)
ISOPT (1988)
Test (CPT)
Sand, silt,
and clay
Continuous evaluation
of density and strength
of sands. Continuous
evaluation of undrained
shear strength in clays.
(See Section
4.5.4.) Test is
best suited for
the design of
footings and
piles in sand;
tests in clay are
more reliable
when used in
with vane tests
Sanglerat (1972)
Schmertmann (1970,
ISOPT (1988)
ISSMFE (1 977b, 1989)
ASTM D3441-79
Robertson and
Campanella (1983a, b)
Konrad and Law (1987a,
Test (BPI)
and cobbly
Soft soils
Qualitative evaluation
of compactness
(See Section
Anderson (1968)
Harder and Seed (1986)
Sy and Campanella
(1992a, b)
Field Vane
Test (FVI)
Sands and
Undrained shear
See Section
4,5,6) Test
should be used
with care,
particularly in
fissured, varved
and highly
plastic clays.
ASTM D 2573-72
Bjerrum (1972)
Schmertmann (1975)
Wroth and HVghes (1973)
Wroth (1975)
38 Canadian Foundation Engineering Manual
Not Properties that
Remarks References*
Type of test
applicable to can be determined
1 Menard (1965) Eisenstein
Soft rock,
Pressure- .\ Soft sensitive
Bearing capacity and i (See Section
dense sand,
i Morrison (1973)
clays loose silts meter Test
compressibility 4 5
gravel, and
1 . .
) Baguelin et al. (1978)
(PMT) : and sands
I. Ladanyi (1972)
• Empirical correlation
i Marchetti (1980)
for soil type, Ko, Flat
Campanella and (See Section
Sand and
overconsolidation Gravel Dilatometer
Robertson (1982, 1991) 4.5.8)
ratio, undrained shear
Test (DMT)
Schmertmann (1986)
strength, and modulus
(See Section
4.5.9) Strictly
Plate Bearing Deformation modulus. applicable only
Test and Modulus of subgrade if the deposit is Sand and
ASTM D 1194-72
reaction. Bearing uniform; size
Screw Plate clay
capacity. effects must be
! considered in
other cases.
head tests
in boreholes
have limited
accuracy. Hvorslev (1949) Sherard
Evaluation of
et al. (1963) Results reliable Permeability Sand and
coefficient of
to one order Olson and Daniel (1981) gravel Test
Tavenas et al. (l983a, b)
are obtained
only from long
term, large scale
pumping tests.
of magnitude
* See corresponding Sections ofthis chapter for a more complete list of references.
Details of the split-barrel sampler and procedure for the Standard Penetration Test are described in ISSMFE (1989)
and ASTM D1586. The split-barrel sampler commonly used in the United States often differs from such samplers
used elsewhere in that the inner liner is not used. As a result, the inner diameter of the sampler is greater than
specified, and since the soil friction developed inside the sampler is reduced, the N value may be underestimated
by up to 20 %.
F or all of its wide use and simple procedure, the results of the SPT are greatly affected by the sampling and drilling
operations. In addition, it is generally recognized that in granular soils of the same density, blow counts increase
-- with increasing grain size above a grain size of about 2 mm.
Improper drilling and sampling procedures which can affect the Standard Penetration Test (SPT) N value are listed
in Table 4.2.
F or the foregoing reasons, it is readily apparent that the repeatability ofthe Standard Penetration Test is questiopable.
In addition, relationships developed for SPT N value versus an exact density should be used with caution. The
Standard Penetration Test is, however, useful in site exploration and foundation design and provides a qualitative
Site Investigations 39
guide to the in-situ properties of the soil and a sample for classification purposes. The evaluation of the test results
should be undertaken by an experienced geotechnical A detailed discussion of the possible errors in SPT
results has been presented by Schmertmann (1979) and Skempton (1986).
TABLE 4.2 Procedures that may affect the SPT N Value
Inappropriate test procedure Potential consequence
Inadequate cleaning of the borehole
Not seating the sampler spoon on undisturbed soil
Driving ofthe sampler spoon above the bottom of the
SPT is not entirely undertaken in original soil; sludge may
be trapped in the sampler and compressed as the sampler
is driven; increase the blow count; (this may also prevent
sample recovery)
Incorrect N-values obtained
N-values are increased in sands and reduced in cohesive
Failure to maintain sufficient hydrostatic head in the
borehole throughout the entire drilling, sampling, and
testing procedure
The water level in the borehole must be at least equal
to the piezometric level in the sand, otherwise the sand
at the bottom of the borehole may become quick and be
transformed into a loose state, rising inside the casing.
Overdrive sampling spoon. Higher N-values usually result from overdriven sampler.
Sampling spoon plugged by gravel.
Plugged casing
Higher N-values result when gravel plugs sampler, and
resistance of an underlying stratum of loose sand could be
highly overestimated.
High N-values may be recorded for loose sand when
sampling below the groundwater table if hydrostatic
pressure causes sand to rise and plug casing.
Overwashing ahead of casing.
Low N-values may result for dense sand since sand is
loosened by overwashing.
Drilling method. '
Drilling techniques such as using a cased hole compared to
a mud stabilized hole may result in different N-values for
some soils.
Not using the standard hammer drop
Free fall of the drive hammer is not attained
Energy delivered per blow is not uniform (European
countries have adopted an automatic trip hammer, which
currently is not in common use in North America)
Using more than 1 Yi tums of rope around the drum andior
using wire cable will restrict the fall of the drive hammer.
Not using correct weight of drive hammer
Drive hammer does not strike the drive cap concentrically
Not using a guide rod
Driller frequently supplies drive hammers with weights
varying from the standard by as much as 5 kg
Impact energy is reduced, increasing the N-values
Incorrect N-values obtained
40 Canadian Foundation Engineering Manual
Inappropriate test procedure
Potential consequence
If the tip is damaged and reduces the opening or increases
Not using a good tip on the sampling spoon
the end area, the N-value can be increased
Use of drill rods heavier than standard Heavier rods result in incorrect N-values
Extreme length of drill rods
Loose connection between rods, top rod, and drive cap
Not recording blow counts and penetration accurately
Incorrect drilling procedures
Experience indicates that at depth over about 15m, N-
values are too high, due to energy losses in the drill rods;
use of a down-the-hole hammer should be considered
Insufficient tightening of drill rods results in and drive cap
poor energy transmission and increased N-values
Incorrect N-values obtained
The SPT was originally developed from wash boring
techniques; drilling procedures that seriously disturb the
soil will adversely affect the N-values, e.g., drilling with
cable-tool equipment. The use of wash boring with a side
discharge bit or rotary with a tricone drill bit and mud flush
is recommended.
Using drill holes that are too large
Holes greater than 100 mm in diameter are not
recommended; use of large diameter-holes may decrease
the blow count, especially in sands.
Inadequate supervision
Frequently a sampler will be impeded by gravel or cobbles,
causing a sudden increase in blow count; this is often
not recognized by an inexperienced observer (accurate
recording of drilling, sampling, and depth is always
Improper logging of soils Not describing the sample correctly
Using too large a pump
Too high a pump capacity will loosen the soil at the base of
the hole causing a decrease in blow count
Numerous studies have shown considerable variations in the procedures and equipment used throughout the world
for this supposedly standardized test. However, the SPT, with all its problems, is still the most commonly used in-
situ test today. As a result considerable research on individual aspects of the standard penetration test equipment
and procedures have been carried out in North America and Japan in an effort to better understand the factors
affecting the test (Schmertmann, 1979; Kovacs and Salomone, 1982; Y oshimi and Tokimatsu, 1983). Considerable
improvements in the understanding of the dynamics of the SPT have occurred in recent years (Schmertmann and
Palacios; 1979, Kovacs et aI., 1981; Kovacs and Salomone, 1982; Sy and Campanella, 1991a and b). Skempton
(1986) and Decourt (1989) present thorough reviews of SPT corrections and correlations with soil properties.
On the basis ofthe studies referred to above and other investigations, several corrections for adjusting or standardizing
the field standard penetration test value, N, are considered in the following paragraphs. While the corrected N values
may be required for design purposes, the original field N values should always be given on the borehole logs. These
corrections or adjustments to N values can include:
Correction for the actual energy delivered to the drill rod. Energy levels vary significantly, depending on
the equipment and procedures used.
Site Investigations 41
Correction for the influence of the overburden stress on N values. 
Correction to  account for the length of the drill rod. 
Correction to account for absence or presence of a liner inside the split-spoon sampler. 
Correction to  account for the influence of the  diameter of the borehole. 
Energy  measurement  during  recent  studies  has  shown  that  ERr'  the  energy  delivered  to  the  rods  during  an  SPT 
expressed  as  a  ratio  of the  theoretical  free-fall  potential  energy,  can  vary  from  about  30  % to  90  % (Kovacs  and 
Salomone,  1982;  Robertson  et  al.  1983).  The  energy  delivered  to  the  drill  rod  varies  with  the  hammer  release 
system,  hammer  type,  anvil  and  operator  characteristics.  The  type  of hammer  and  anvil  appear  to  influence  the 
energy transfer mechanism. 
In  view  of the  variation  of energy  input during  the  SPT for  various  situations,  there  is  clearly  a  need  to  be  able 
to  adjust or normalize  the N  values to  allow  comparison on a common basis.  Schmertmann and  Palacios,  (1979), 
have  shown  that  the  SPT  blowcount  is  approximately  inversely  proportional  to  the  delivered  energy.  Kovacs  et 
al.  (1984),  Seed  et  al.  (1984)  and  Robertson  et  al.  (1983)  have  suggested  that  an  energy  level  of 60  % appears 
to  represent  a reasonable  historical  average  for  most SPT  based empirical  correlations.  Seed et al.  (1984)  clearly 
specify that for liquefaction analyses the SPT N values must be corrected to an energy level of 60 %. 
N-values  measured  with  a known  or estimated rod  energy  ratio,  ERr'  in percent,  can  be  normalized to  an  energy 
level  of 60  %,  that is  to N
, by the following conversion: 
60  60 
Based on data summarized by  Skempton (1986)  and Seed  et al.  (1984), recommended  generalized energy ratios, 
ERr'  in  percent,  are  given in  Table  4.3.  These values represent broad global correlations  and should be used with 
TABLE 4.3  Generalized SPT Energy Ratios
(Based on Seed et al., 1984; Skempton, 1986)
Country  Hammer  Release  Er

(%) Err/60 
North and South 
2 turns of rope 
2 turns of rope Trip 
55  to  83 
0.92 to  1.38 
2  turns of rope 
2 turns of rope 
Safety  2 turns  of rope  50  0.83
Trip  60  l.0 Automatic 
Italy  Donut  Trip  65  1.08 
42 Canadian Foundation Engineering Manual
TABLE 4.4. Approximate Corrections to Measured SPT N-Values (after Skempton, 1986)
Correction Factor Item Correction Factor Value


Rod Length (below anvil): 

4-6 m 
Standard Sampler US 
Sampler without liners 
Borehole diameter: 
65  - 115  mm 
The  International  Reference  Test procedure  (ISSMFE,  1989)  recommends  that  in  situations  where  comparisons 
of SPT  results  are  important,  calibrations  should be  made  to  evaluate the efficiency  of the equipment in  terms  of 
energy transfer.  Table 4.3  provides  only a guide to  anticipated  average  energy levels.  The recommended method 
of SPT  energy  measurement  is  specified  in  ASTM D4633-86  and  ISOPT (1988).  For  projects  where  SPT results 
are important, such as  liquefaction studies, or where major project decisions rely on the SPT, energy measurements 
should be made. 
The SPT N values vary with the  confining stress, and consequently, the overburden pressure.  An overburden stress 
correction  is required  to normalize the  field blowcounts to  a constant reference vertical  effective normal  stress  as 
done  for  liquefaction  studies.  This  correction eliminates  the  increase  in  blowcount at  constant density due to  the 
increase in confining s t r e ~ s  
A  variety  of methods  of correcting  for  overburden  pressure  have  been  suggested  by  various  investigators  and 
several of these have been summarized by Liao  and Whitman (1986).  Liao  and  Whitman (1986)  also  proposed  a 
correction factor 
which  is  very  similar  to  the  other acceptable  correction  factors  and  is  simple  to  use.  The  correction  factor  used 
elsewhere  in  this  Manual,  however,  is  that  proposed  by  Peck  et  al.  (1974)  and  is  described  in  the  following 
A commonly used  overburden reference effective stress level, particularly for liquefaction studies,  is  1.0 tsf or 1.0 
,  and the corresponding value in  SI units, is approximately 96  kPa. If the N-value at depth corresponding to 
an  effective overburden stress of 1.0 tsf (96 kPa)  is  considered, the correction factor C to  be applied to the field N 
values for other effective overburden stresses is given approximately by 
.  [1920J
eN =0.7710g
~ (4.2) 
where  C
overburden correction factor 
effective overburden stress at the  level ofN-value in kPa 
The  equation for  C
is  not  valid  for  a,: less  than  about  0.25  tsf (24  kPa)  since  for  low  overburden pressures  the 
equation for  C
leads to unreasonably large correction factors.  To overcome this problem, Peck et al.  (1974)  have 
proposed using the chart given as Figure 11.8 (Chapter  11) which is  a plot of  versus effective overburden stress 
Site Investigations  43 
(pressure). For values of overburden pressure more than 24 kPa, the cOlTection factor C

on Figure 11.8 cOlTesponds
to that obtained from the equation for C". To avoid excessively large values of C\, for small effective overburden
pressures, the plot on Figure 11.8 has been arbitrarily extended to a C;\ value of 2.0 at zero effective overburden
pressure. Although the maximum value of of2.0 has been suggested, it is probably prudent in practice not to use
values larger than about 1 unless justified by special studies.
The normal practice in liquefaction studies is to normalize the N-values to an energy ratio of 60 %, and also for an
effective overburden pressure of 1.0 tsf (96 kPa), (see Seed et aI., 1984) This normalized value, known as (N 1)60' is
given by the following equation:
N) N( ERr )(c )  (4.3)
I 60 l 60   N
where =  N value COlTected and normalized for energy ratio of60 % and normalized for effective
overburden pressure of 1.0 tsf or 96 kPa (SI units)
field blowcount
rod energy ratio normalized to 60 % (Table 4.3)
overburden stress correction
Further corrections to N values can also be made, when appropriate for the effects of rod length, sampler type and
borehole diameter. Approximate correction factors are given in Table 4.4. Wave equation studies (Schmertmann
and Palacios, 1979) show that the theoretical energy ratio decreases with rod length less than about 10m. The
approximate correction factor, Cr, is given in Table 4.4. Note, however, that when applying Seed's simplified
liquef3ftion procedure, the (N
)60 value should be COlTected by multiplying with a rod length COlTection factor of
0.75 for depths less than 3 m as recommended by Seed, et al. (1984).
Studies by Schmertmann (1979) also found that removing the liner from an SPT sampler designed for a liner
improved sample recovery but reduced the measured blowcounts by about 20 %. The corresponding correction
factor in Table 4.4 is C '
Although good modern practice has the SPT undertaken in a borehole with a diameter between 65 mm and 115 mm,
many countries allow testing in boreholes up to 200 mm in diameter. The effect of testing within relatively large
diameter boreholes can be significant in sands and probably negligible in clays. Approximate correction factors for
the borehole diameter, Cd' are given in Table 4.4.
In addition to the foregoing, there are some other factors which may require consideration and possible correction
for specialized applications. These factors include grain size, overconsolidation, aging and cementation (Skempton,
1986). Also, special consideration may be required ifheavy or long rods (greater than about 20 m) are used. Energy
losses and damping may result in N-values that are too high,
While using normalized (N
)60 values together with other corrections as appropriate has merit, many ofthe standard
penetration N-value empirical relationships given in this Manual were developed before it was common practice to
correct field N-values. The question then arises as to whether, and in what manner the N-values should be cOlTected
and the following comments are provided for guidance.
A review of the procedures recommended for correcting N-values by authors offoundation engineering text books
indicates that there is some difference of opinion. Das (1990) and Fang (1991) both recommend the use of the
overburden pressure cOlTection for the Standard Penetration Test. Bowles (1988) perhaps provides one ofthe more
comprehensive evaluations ofN-value corrections. He states that since there are several opinions on N corrections,
then the following three basic approaches are possible:
1. Do nothing which, with current equipment and conditions, may be nearly correct for some situations.
44  Canadian Foundation Engineering  Manual 
2.  Adjust only for overburden pressure. 
3.  Use  the  equation  for  (N1)60  and  when  appropriate  apply  cOHections  for  rod  length,  C
,  sample  liner,  C

and  borehole  diameter,  Cd'  This  is  probably the  best method  but  requires  equipment  calibration  for  ER. 
This procedure may become mandatory to  allow  extrapolation of N data across  geographic regions  where 
different equipment is  used. 
In  view  of the  absence  of general  agreement  on  the  application  of N-value  corrections,  the  following  guidelines 
are  given for  use  in  this  Manual.  The  N  values  should  be  corrected to  the  (N )60  values,  together  with  any  other 
conections as appropriate, when used for liquefaction studies. The N-values should also be corrected as specifically 
identified in the various chapters of this Manual but such corrections may not include all the possible factors. 
In the absence of any specific recommendations in this Manual on corrections to the N-values prior to using empirical 
relationships,  it is  difficult to provide  specific guidance.  Often no cOHections  are  used and this may be reasonably 
appropriate in  Canadian practice for some conditions as  suggested by the following  comments. 
energy  efficiency  of much  of the  Standard  Penetration  Test equipment currently  in  use  in  Canadian  practice 
is  very similar to that used when the various N-value empirical relationships were  developed, that  is  (ERr)  was 45 
to 60 percent so the energy con'ection may be small. The rod length correction C

is  applicable for rod lengths less 
than  10  m.  However,  most  existing  empirical  correlations  with  SPT  N-values  did  not  incorporate  C

and  hence 
this  correction  may  not  be necessary  in  many  cases.  In usual  Canadian practice,  the  sampler liner correction,  C

and  the  borehole  diameter  correction,  Cd'  are  both  1.0  so  no  correction  is  required.  Consequently,  for  the  usual 
Canadian practice, the most  likely  correction to  field  N-values  for  use  in  the  N-value  empirical relationships that 
may be considered is  the overburden correction factor,  C
,  which may apply in cases where overburden pressure is 
a significant factor. 
The  overburden  correction  factor,  however,  is  not  always  used  in  current  practice,  and  the  significance  of this 
omission  will  depend  on  the  type  of problem  and  empirical  relationship  for  N  being  considered.  Ignoring  the 
correction  factor for N-values at shallow  depths  will be conservative.  Ignoring the overburden correction factor  at 
greater depths  may be  unconservative  if the  empirical relationship  being  considered does  not  extend to  the  same 
depth range, or makes no allowance for influence of depth. 
4.5.3 Dynamic Cone Penetration Test (DCPT)
The dynamic cone penetration test is a continuous test which utilizes a dropping weight to drive a cone and rod into 
the ground. The number ofblows for each 300-mm penetration (200 mm in European practice) is recorded. A variety 
of equipment is used in different areas. The Dynamic Probe Working Party ofthe ISSMFE Technical Committee on 
Penetration Testing has published suggested international reference test procedures  in the Proceedings of the First 
International Symposium on Penetration TestingiISOPT-1I0RLANDO/March 1988. This reference contains useful 
discussions of the test. 
Usually  in  North American  practice,  the  rods  consist  of the  same  44.4  mm  diameter rods  used for  the  Standard 
Penetration Test (SPT), and the drive weight and height of fall is the same as in the SPT. A variety ofcones are used. 
They may be fixed or disposable (to reduce resistance on withdrawal) and usually are pointed. The diameter ofthe 
cones used range from  50 mm to  100 mm and maybe short or sleeved, depending on the soil strata and the  desired 
information. Some experience has suggested that short cones should be avoided and that a cone with 45° taper from 
a 30 mm diameter blunt tip to a 60 mm diameter with a minimum 150 mm long sleeve reduces rod friction compared 
to  a short (unsleeved) cone. 
In cohesive soils if a dynamic cone is used to delineate the boundary between stiff to firm  clay and soft to very soft 
clay,  experience  has  shown that very  large  cones,  100  mm or larger,  with  a sleeve that  is  2.5  times the  diameter, 
could provide a better resistance contrast between the strata. 
The  dynamic  penetrometer is  subject to  all of the  disadvantages  of the Standard  Penetration Test  and should not 
Site Investigations 45
be used for quantitative evaluation of the soil density and  other parameters.  One major problem with the Dynamic 
Cone  Penetration Test  is  rod  friction  which builds up  as  the  probe  depth increases.  At depths  beyond  15  m to  20 
m,  the effect of rod friction tends to  mask the cone tip resistance, making interpretation of test results difficult.  Rod 
friction  can be  minimized by  use  of an  outer casing which  "follows" behind the  cone,  or by periodic drilling and 
continuing the  Dynamic Cone Penetration Tests  fl:om  the  bottom of the  drill hole.  In  some areas,  local experience 
and  calibration  with  information  from  sampled  drill  holes  have  made  the  dynamic  cone  penetration  test  a useful 
in-situ technique.  The main  advantage  of the  dynamic  cone    test is  that it is  fast  and  economical, and  a 
continuous resistance versus  depth profile  is  obtained that  can provide a visual relationship  of soil type or density 
4.5.4 Cone Penetration Test (CPT)
Many static cone penetrometers were  developed and used in Europe before gaining acceptance in North American 
practice (Table 4.5). The main reasons for  the increasing interest in  cone penetration tests  (CPT) are the simplicity 
of testing, reproducibility of results, and the greater amenability oftest data to rational analysis.  A cone point with 
a  10  cm
base  area  and  an  apex  angle  of 60° .has  been  specified  in  European  and  American  standards  (ISSMFE, 
1989, and  ASTM D3441). A friction  sleeve with  an  area of 150  cm

is  located immediately above  the  cone point. 
Mechanical  cone penetrometers  (Begemann,  1965)  have  a telescopic action,  which requires  a double  rod  system. 
With the  electrical cone penetrometers,  the  friction  sleeve  and cone  point advance  continuously with a single  rod 
Not withstanding  that the  mechanical penetrometers offer the  advantage of an  initial  low  cost for  equipment and 
simplicity  of operation,  they  have  the  disadvantage of a slow incremental procedure,  ineffectiveness in soft soils, 
requirement of moving parts,  labour-intensive  data handling  and  presentation,  and  limited  accuracy.  The  electric 
cone penetrometers have built-in load-cells that record continuously the point-pressure, qc'  and the local side shear, 
\. The  load-cells  can  be  made  in a  variety of capacities  from  50  to  150 kN for point  resistance  and 7.5  to  15  kN 
for  local  side shear,  depending on  the  strength of the soils  to  be penetrated.  Typically,  an  electric  cable connects 
the  cone-and-sleeve  load-cells  with  the  recording  equipment  at  the  ground  surface  although  other  data  transfer 
technologies are available. 
TABLE 4.5 Types o/Cone Penetration Tests (Adaptedfrom Schmertmann, 1975)
Type ,
Static or 
quasistatic  mechanical jacking  cone 
Rotation of a  Sweden
Weight- sounding 
weighted helical  variable  Finland
cone  Norway 
Hydraulic or 
20 mrnls  Worldwide 
The  electric  cone  penetrometer  offers  obvious  advantages  over  the  mechanical  penetrometer,  such  as:  it  is  a 
more  rapid  procedure,  it provides  continuous  recording,  higher  accuracy  and  repeatability,  there  is  the  potential 
for  automatic  data  logging,  reduction,  and  plotting,  and  additional  sensors  can  also  be  incorporated  in  the  cone 
point.  Electric cones carry a high initial cost for equipment and require highly skilled operators with knowledge of 
electronics.  They also require adequate back-up in technical facilities  for calibration and maintenance. 
The  most  significant  advantage  that  electric  cone  penetrometers  offer  is  their  repeatability  and  accuracy.  An 
important application of the cone-penetration test is to  determine accurately the soil profile. Extensive use  is made 
of the  friction  ratio,  i.e.,  the  ratio  between the  point-pressure  and  the  side  shear,  as  a  means  of soil  classification 
(Begemann,  1965,  Schmertmann,  1975,  Douglas  and  Olsen,  1981).  It has been  shown  over the  past several years 
- - -
46 Canadian Foundation Engineering Manual
that stress normalization of cone point resistance and friction ratio is correct from a fundamental perspective, and
its use provides a much better correlation with retrieved samples. It must, however, be kept in mind at all times that
the CPT provides an indication of soil type behaviour, which is different from explicit soil type in some instances,
but is what the geotechnical engineer ultimately requires for design purposes. Robertson (1990) presents stress
nonnalized soil classification charts.
A significant development in the electric cone-penetration testing has been the addition of a pore-pressure gauge at
the base of the cone. Pore-pressure measurement during static cone-penetration testing provides more information
on the stratification and adds new dimensions to the interpretation of geotechnical parameters, especially in loose
or soft, fine-grained deposits (Konrad and Law, 1987a). The continuous measurement of pore pressures along with
the point resistance and side shear makes the electric cone penetrometer the premier tool for stratification logging
of soil deposits (Campanella and Robertson, 1982; Tavenas, 1981).
The excess pore pressure measured during penetration is a useful indication ofthe soil type and provides an excellent
means for detecting stratigraphic detail, and appears to be a good indicator of stress history (Konrad and Law,
1987b). In addition, when the steady penetration is stopped momentarily, the dissipation of the excess pore pressure
with time can be used as an indicator of the coefficient of consolidation.
Finally, the equilibrium pore-pressure value, i.e., the pore pressure when all excess pore pressure has dissipated, is
a measure of the phreatic elevation in the ground.
Cone resistances and pore pressures are governed by a large number of variables, such as soil type, density,
stress level, soil fabric, and mineralogy. Many theories exist to promote a better understanding ofthe process of a
penetrating cone, but correlations with soil characteristics remain largely empirical.
Empirical correlations have also been proposed for relating the results of the cone penetration test to the Standard
Penetration Test (SPT), as well as to soil parameters, such as shear strength, density index, compressibility, and
modulus (Campanella and Robertson, 1981; Robertson and Campanella, 1983 a, b).

......-';...- ---",,-
_... -"'" ......
---- --- -'
- - ";\.1'-1\0\'\ ........... --C-
.... --... ----
.... _ ... '#...- - RANGE OF RESULTS
....... --: ___ f"'- Burland and Burbidge, 1985
Robertson et aI. 1983 .

--- --- -
• € €
.... ...-
__ -
0.01 0.1 1.0
FIGURE 4.2 Variation ofqclN ratio with mean grain size (adapted from Robertson et al., 1983;
and Burland and Burbidge, 1985). The dashed lines show the upper and lower limits ofobservations.
Site Investigations 47
The use of the CPT to estimate equivalent SPT values is a common application for foundation design. The major
advantages of the CPT over the SPT are its continuous profile and the higher accuracy and repeatability it provides;
subsequently if a good CPT-SPT correlation exists, very comprehensive equivalent SPT values can be obtained.
The relationship between the CPT, represented by the tip resistance, qc' and the SPT, represented by the blow count
N, has been determined in a number of studies over the past 30 years eMeigh and Nixon, 1961; Thornbum, 1970;
Schmertmann, 1970; Burbidge, 1982; Robertson et aI., 1983; Burland and Burbidge, 1985). The relationship
between CPT and SPT is expressed in terms of the ratio q/N (kN/m2 per blows per 0.3 m); q/N data from available
literature is summarized on Figure 4.2 against the mean particle size of the soils tested.
4.5.5 Becker Penetration Test (BPT)
The Becker hammer drill was developed in 1958 in Alberta, Canada, initially for seismic oil explorations in difficult
gravel sites. The drill is now widely used in North America in mining explorations and in geotechnical investigations
for drilling, sampling and penetration testing in sand, gravel and boulder formations. The drill consists of driving a
specially designed double-walled casing into the ground with a double-acting diesel pile hammer and using an air
injection, reverse-circulation technique to remove the cuttings from the hole. The Becker drill system is more or less
'standardized', being manufactured by only one company, Drill Systems, in Calgary, Alberta. The hammer used in
the Becker system is an international Construction Equipment, Inc. (ICE) Model 180 double-acting atomized fuel
injection diesel pile hammer; with a manufacturer's rated energy of 11.0 kJ. The casings come in 2.4 m or 3.0 m
lengths and are available in three standard sizes: 140 mm O.D. by 83 mm I.D., 170 mm O.D. by 110 mm LD. and
230 mm O.D. by 150 mm ID. The main advantage of the Becker hammer drill is the ability to sample or penetrate
relatively coarse-grained soil deposits at a fast rate. More details of the hammer drill can be found in Anderson
The Becker casing can be driven open-ended with a hardened drive bit for drilling and sampling, in which case
compressed air is forced down the annulus of the casing to flush the cuttings up the centre of the inner pipe to the
surface. The continuous cuttings or soil particles are collected at the ground surface via a cyclone which dissipates
the energy of the fast-moving air/soil stream. The drilling can be stopped at any depth and the open-ended casing
allows access to the bottom of the hole for tube sampling, standard penetration test or other in-situ tests, or for rock
coring. Undisturbed sampling or penetration testing conducted through the casing in saturated sand and silt may
not be reliable, since stoppage of drilling and air shutoff result in unequal hydrostatic conditions inside and outside
the casing, causing disturbance or "quicking" of the soil formation below the casing level. This is manifested in the
field by soil filling up the bottom section of the casing when drilling is stopped. On completion ofdrilling, the casing
is withdrawn by a puller system comprising two hydraulic jacks operating in parallel on tapered slips that grip the
casing and react against the ground.
The Becker casing can also be driven close-ended, without using compressed air, as a large-scale penetration test
to evaluate soil density and pile driveability. In this mode, commonly referred to as the Becker Penetration Test
(BPT), the driving resistances or blowcounts are recorded for each OJ m of penetration. Because of the larger
pipe (or sampler) diameter to particle size ratio, the BPT blowcounts have been considered more reliable than SPT
N-values in gravelly soils. As a result, numerous attempts have been carried out in the past to correlate the BPT
blowcounts to standard penetration test (SPT) N-values for foundation design and liquefaction assessment. Most of
these BPT-SPT correlations, however, have limited or local applications, since they. do not take into account two
important factors affecting the BPT blowcounts: variable hammer energy output and shaft resistance acting on the
Becker casing during driving.
Like all diesel hammers, the Becker hammer gives variable energy output depending on combustion conditions
and soil resistances. Harder and Seed (1986) have proposed a practical method using hammer bounce chamber
pressure measurements to correct the measured BPT blow counts to a reference "full combustion rating curve"
before correlating with corrected SPT N-values. The method is rig or hammer specific and requires a BPT -SPT
correlation be established for each drill rig. A more fundamental method of correcting BPT blowcounts based on
transferred energy is proposed by Sy and Campanella (1 992a). This energy method, however, requires measuring
force and acceleration near the top of the casing during the BPT, similar to dynamic monitoring of pile driving
48 Canadian Foundation Engineering Manual
(ASTM D4945-89).
The Becker Penetration Test also simulates the driving of a displacement pile and can be used for pile driveability
evaluations (SDS Drilling Ltd.; Morrison and Watts, 1985; Sy and Campanella, 1 992b).
4.5.6 Field Vane Test (FVT)
The field vane test is the most common method of in-situ determination of undrained shear strength of clays. The
vane is best suited for soft-to-firm clays; it should not be used in cohesion less soils.
The vane equipment consists of a vane blade, a set of rods, and a torque measuring apparatus. The vane blade
should have a height-to-diameter ratio of 2; typical dimensions are 100 by 50 mm. The effect of soil friction on the
measured torque should be eliminated or be measurable. The torque-measuring apparatus should permit accurate,
reproducible readings, preferably in the form of a torque-angular deformation curve. Specific details of the vane
shear test and equipment can be found in ASTM D2573. The vane may be rectangular or tapered.
The vane-test performance and interpretation are subject to some limitations or errors, which should be taken into
account when using the test results. The insertion of the vane blade produces a displacement and remolding of
the soil. Experience shows that thicker blades tend to produce reduced strengths. For acceptable results, the blade
thickness should not exceed 5 % of the vane diameter.
The failure mode around a vane is complex. The test interpretation is based on the simplified assumption of a
cylindrical failure surface corresponding to the periphery of the vane blade (Aas, 1965). The undrained shear
strength can be calculated from the measured torque, provided that the shear strengths on the horizontal and vertical
planes are assumed equal, by the following relation:
2T (4.4)
-3(H /D+a/2)
Su undrainedshear strength
T maximum applied torque
H =- vane height
D = vane diameter
a factor which is a function of the assumed shear distribution along the top and bottom of the failure
a = 0.66 if uniform shear is assumed (usual assumption)
a = 0.50 if triangular distribution is assumed (i.e., shear strength mobilized is proportional to strain)
a 0.60 if parabolic distribution is assumed
For the assumption of a = 0.66, which is the usual assumption, and a vane height to vane diameter ratio of 2.0, the
above equation becomes: .
s = ~   :
II 3.66D3
The above equations are for a rectangular vane. For a tapered vane refer to the ASTM D2573, and for a vane with a
45 degree taper, HID = 2.0, a = 0.66, and vane rod diameter d, the undrained shear strength is given by the following
Site Investigations 49
The vane shear test actually measures a weighted average of the shear strength on vertical and horizontal planes.  It 
is  possible to  determine the horizontal  and  vertical  either plane by  the  test in  similar 
soil conditions using vanes of different shapes or helght/dlameter ratlOs. It has been found  that, In general, the ratlo 
of horizontal/vertical  shear strength  is  less  than unity  and  when this  is  applicable,  the  field  vane  value  of su'  is  a 
conservative  estimate of the shear strength along the  vertical  plane.  Becker et al.  (1988) provide an  interpretation 
where vane strength is  essentially controlled by horizontal stresses on  the  vertical plane. 
III  100 


I /'

1.2  1---+--+--+-+--+-+--1--.--1----1 
i.o  J---t-.r+--I-+---I----i-__I,---r---+----l 
0.2  '---'----'----l_.l.-.......L...--L---"_...L--L---l
J-L 0.8 
0.6  J---+--I---I-""'-o+::--F""'I--",.....--
0.6  J---...,J--...,......,J----'J---J--__I=-__I 
20  40  60  80  100  120  0  0.2  M 0.6  0.8  1.0 
Plasticity Index
FIGURE 4.3 Vane correction/actor (after (a) Bjerrum, 1972,  and (b) Aas, et al., 1986)
The  measured field  vane shear strength may require  a correction and  Bjerrum (1972,  1973) proposed a correction 
factor,  Jl,  which relates the corrected vane strength, (s)COIT'  to the field vane shear strength, (S)field'  as follows: 
(s)COIT  Jl  (s)field 
where  Jl  varies  with  plasticity  index  as  shown  on  Figure  4.3(a).  Aas  et  aL  (1986)  undertook  a  substantial  re-
evaluation  of the  Bjerrum  chart  to  include  overconsolidation  ratio  (OCR)  and  aging  to  produce  a  revised  chart, 
Figure 4.3(b)  where  Jl  is  given as  a function of the  ratio  (s)field,lcr
',  and cr
'  is the  effective overburden pressure. 
Figure  4.3(b)  is  used by entering  the  top  chart  with PI and  (s)field,!  cr  '  to  establish  whether the  clay  is  within  the 

normally  consolidated  (NC) range between  the  limits  'young'  and  'aged', or overconsolidated  (OC).  The  bottom 
chart of Figure 4.3(b) is then used to  obtain,  Jl  for  the ratio (s)field,!  cr 
' and the corresponding NC or OC curve. Aas 
et al.  recommend a  maximum design value for  Jl  of 1.0 for  (s)field,! cr;  less  than 0.20 since  Jl  is  rather sensitive for 
low values of (s)field,! cry'.  Refer to  Aas  et aL  (1986) for further details. 
Although the correction for plasticity index is used by many practitioners, Leroueil et aL  (1990) and Leroueil (200 I) 
provide information that this correction may not be necessary for soft clays. The vane shear value can also be used 
to  estimate the OCR (Mayne and Mitchell,  1988). 
The vane shear strength is  usually plotted against  depth to  provide a strength profile.  It is  a good practice to  carry 
out,  in  parallel  with  vane  tests,  other in-situ tests  such as  static  cone,  or  piezocone-penetration tests,  which yield 
continuous profiles and to  correlate these results with the vane test values.  ASTM STP  1014 (1988) contains papers 
50 Canadian Foundation Engineering Manual
on the testing and interpretation of vane shear tests.
4.5.7 Pressuremeter Tests (PMT) Introduction
Pressuremeters are used to measure the in-situ defonnation (compressibility) and strength properties ofa wide variety
of soil types, weathered rock, and low to moderate strength of intact rock. Two major types of pressuremeters have
been developed which are cunentiy in use in Canada; the pre-bored pressuremeter and the self-boring pressuremeter.
The Menard-type pressuremeter is a well-known type of pre-bored pressuremeter. Each type of pressuremeter has
advantages and limitations largely governed by the type of material to be tested and the method of geotechnical
analysis to be canied out. All types ofpressuremeter tests are sensitive to the method ofprobe installation and testing,
and highly trained staff who possess a thorough understanding of the equipment and test procedures are required to
obtain reliable results. Pressuremeter testing is generally canied out by specialized drilling and/or testing contractors
although some engineering consultants and public agencies have their own equipment and trained personnel.
The pressuremeter test was first developed by Louis Menard in 1956. The Menard-type pressuremeter test procedure
basically consists of horizontal expansion of a membrane mounted on a relatively long probe placed in a slightly
oversized, pre-bored hole. Lateral displacement of the membrane and borehole wall is achieved by injecting either
liquid or gas into the probe at selected pressure increments. Displacements are measured either in tenus of the
volume ofliquid injected into the probe or more directly by callipers or displacement transducers for the gas inflated
systems. Pressures are measured either with a surface gauge or pressure transducer in the probe. The pressuremeter
test allows the detenuination of the load-defonnation characteristics of the tested soil.
The Menard-type tests are sensitive to the degree of soil disturbance caused by drilling the borehole. In order to
minimize the soil disturbance, the self-boring pressuremeter was developed independently in France (Baguelin et
aI., 1972) and in England (Wroth and Hughes, 1973). The principles of the test are similar to the Menard-type test,
however, a small rotating cutting head is located in the tip ofthe probe. The probe is advanced by pushing the probe
into the soil.
Displaced soil enters the cutting head where it is removed using water or a bentonite slurry pumped through a
double rod assembly. Self-boring pressuremeters can be equipped with a pore-pressure transducer mounted on the
exterior of the probe. The membrane is inflated using either liquid or gas in a manner similar to the Menard-type
pressuremeter. Similarly, lateral displacements of the borehole wall during the test can be measured either by the
volume of injected liquid, or more commonly, with displacement transducers, and the test pressures are measured
with a surface gauge or pressure transducers located in the probe.
Relatively small, full displacement pressuremeters have also been combined with static cone penetrometers (Hughes
and Robertson, 1985; and Withers et aI., 1986) in order to provide a multipurpose tool for site investigations. Menard-Type Pressuremeter Tests
The following discussion will deal with pressuremeters of the Menard design because they are the most common in
engineering practice today. This discussion may not be entirely applicable to other pressuremeter designs.
The standard Menard pressuremeter consists ofa probe connected to a pressure-volume control unit with stiff tubing.
Probes are generally available in three diameters consistent with commonly utilized drill hole sizes (A, Band N).
The probe consists of a metal cylinder covered with an inflatable membrane and protective sheath comprised of
a series of metal strips. The probe is separated into three independent cells; the two end cells are guard cells used
to reduce end effects on the middle cell to produce predominantly radial strains in the soil interval tested. Lateral
displacements are measured only in the middle cell. All cells are nonnally filled with water or antifreeze although
some systems use gas to inflate the guard cells. Pressure is applied to the fluid in a series of increments by a gas
Site Investigations 51
control system acting on a reservoir in the control unit. Volume changes in the reservoir are measured by graduated
transparent tubes on the control unit. A more complete description of the Menard system is presented in Baguelin
et al. (1978).
Other pressuremeter probes without the two end cells have been introduced. The test results from such probes may
need to be corrected before use in common pressuremeter design methods.
Borehole  Preparation 
It is extremely important to minimize disturbance of the borehole wall during the drilling process. Appropriate
drilling procedures are described by Baguelin et aI. (1978). Normal drilling and sampling techniques are generally
intended to minimize disturbance of the collected samples and may not be suitable for pressuremeter testing.
Drilling methods should be selected to prevent collapse of the borehole wall, minimize erosion of the soil, and
prevent softening of the soil (Finn et aI., 1984). When pressuremeter tests are conducted in a soil type where limited
local experience in pressuremeter testing is available, several methods of drilling should be evaluated to determine
the optimum method. General guidance regarding the initial selection of drilling methods for various soil types is
presented in Table 4.6.
Test Procedure 
Typically, Menard-type pressuremetertests are carried out as stress controlled tests by applying a series of increasing
pressure increments. The maximum pressure expected during the test should be divided into a minimum often equal
pressure increments. Each pressure increment is maintained for a one minute period with volume or radial strain
measurements recorded at intervals of 15, 30, and 60 seconds. All pressure increments should be maintained for
the same time period. Tests are generally considered to be complete when the volume of the liquid injected during
the test is equal to the initial volume of the borehole. In hard .soils and rocks it may not be possible to inject this
volume and the test is terminated at the maximum pressure for the system. If the sides of the borehole are enlarged
excessively either by improper sizing of the drilling equipment or erosion of the borehole wall, the maximum
inflation volume of the probe may be reached prior to injection of the required volume.
TABLE 4.6  Methods 0/Borehole Preparation/or Menard-type Pressuremeter Tests
Soil Type  Drilling Methods 
Finn to Stiff Clay Pushed tube with internal camfer
! Pushed or driven tube with internal camfer
Stiff to Hard Clay . Core drilling with mud or possibly foam flush
Continuous flight auger
! Pushed or driven tube with internal camfer
Core drilling with mud or possibly foam flush (very stiff to hard silts)
Pushed or driven tube with internal camfer (with mud below the water table)
Core drilling with mud flush (dense to very dense sands)
Very difficult to avoid disturbance. A driven slotted casing is sometimes used, however
disturbance is significant due to lateral displacement of the soil
Glacial Till
Weak or Weathered Rock
Sound Rock
Core drilling with mud (very dense finer grained tills with high silt and/or clay content)
Driven thick-walled tube with internal camfer (medium dense to dense finer grained
tills as above)
Driven slotted casing (applicable only to medium dense tills - very high soil
Core drilling with mud or possibly foam flush
Core drilling with water, mud or foam flush
52 Canadian Foundation Engineering Manual
Strain controlled tests are possible for instruments which measure displacements of the borehole wall directly with
either callipers or transducers. Computer controlled load application greatly simplifies the test procedure; however,
the availability of the equipment is limited. Strain rate selection is important for clays, particularly in the plastic
stress range (Anderson, 1979; Windle and Wroth, 1977).
Test Interpretation
The results ofa standard Menard-type pressuremeter test corrected for volume and membrane resistance are shown in
Figure 4.4 as the Pressuremeter Curve. The pressure must be corrected for the hydrostatic pressure in the measuring
circuit above the water table. In the first stage ofthe test, the volume increases rapidly with small changes in pressure
as the probe is inflated against the soil. The volume at the point where the curve becomes approximately linear is
termed v
' which is equal to the difference between the volume of the hole and the initial volume of the probe. The
corresponding pressure at this point is called Po; however, this pressure does not represent the true in-situ pressure
in the ground because of stress relief during the formation of the hole. At higher pressures the volume increases
slowly with pressure. The creep volume change in this pressure range is small and approximately constant, which
indicates pseudo-elastic behaviour of the soil. The slope of the volume - pressure curve in this range is related to the
shear modulus of the soil as discussed below. The pressure corresponding to the end of the constant creep volume
measurements is called the creep pressure Pr At higher pressures the volume and the creep volume increase rapidly
indicating the development of soil failure around the probe. The pressure - volume curve tends to an asymptotic
limit corresponding to the limit pressure PI'
The theoretical basis for the pressuremeter test is the radial expansion of a cavity in an infinite elastic medium
which was developed first by Lame (1852). Details of the cavity expansion theory are presented in Baguelin et al.
(1978) and Mair and Wood (1987). The equation for the radial expansion of a cylindrical cavity in an infinite elastic
medium is:
G the shear modulus
V the volume of the cavity
p pressure in the cavity
The pressuremeter test produces only shear stresses in the soil; no compressive stresses are involved although
the test would appear to be entirely compressive. The modulus value calculated from the pressuremeter test is,
therefore, a shear modulus (G
). While the slope of the pressuremeter curve, b.p/b.V is constant from Vo to vf' the
volume V is not. Therefore, the shear modulus G is dependent on the volume of the cavity selected, which for the
pressuremeter test is, by convention, selected at the midpoint ofthe pseudo - elastic portion of the pressure - volume
curves (Figure 4.4). The corresponding shear modulus is defined as G • The shear modulus is calculated using the
= (ve + Vm ) (p/v)
where Vc the initial volume of the probe prior to infiation
(vo + vr )/2 (Figure 4.4)
The term plv is the slope of the pressure - volume line in the pseudo-elastic range
The test results are most often presented in terms of an equivalent Young's modulus (E) assuming an isotropic
elastic soil using the equation:
E = 2G (1 + v)

Site Investigations 53
Poisson's ratio 
The standard Menard approach is  to assume a Poisson's ratio of 0.33 and the resulting modulus is  called the Menard 
modulus (EM) where 
(4.11 ) 
When the previous equation for (G
) is  substituted, then the Menard modulus (E:v)  is  given by: 
EM 2.66(vc  + vm)  (p/v) (4.12) 
Other  values  of Poisson's  ratio  may  be  more  appropriate  depending  on  the  soil  or  rock  type  and  the  drainage 
conditions,  i.e.,  fine  grained  vs.  coarse-grained  soil  and  undrained  vs.  drained  loading.  General  guidance  on  the 
selection of appropriate Poisson's ratios is  presented in Mair and Wood (1987). 
Similar  interpretation  techniques  are  used  for  tests  which  have  been  cyclically  unloaded  and  reloaded.  A  shear 
modulus can be calculated for either portion of the load cycle.  The volume vm  used in the calculation is  the average 
volume over the load cycle. The shear modulus computed from the cyclic portion of pressuremeter tests is generally 
considered to  be  representative  of the  "elastic" stiffness  of the  soil provided the  strains are  small  (Wroth,  1982). 
Shear modulus is sensitive to effective stress and strain level and the use ofthe test results in design should consider 
these factors. 
The Menard limit pressure is defined as  that pressure at which the volume is  equal to twice the initial volume ofthe 
hole, that is  2(vo  + vJ Various methods  are available to determine the limit pressure,  as  described by Baguelin et 
aI,  (1978). In cases where the borehole is oversized or the oversized or the soil shear strength is very high, the limit 
pressure  may not be attainable  during the  test.  In these  cases  the  limit pressure may be estimated from  the creep 
pressure (Pr) using the following  empirical relationship: 
0.5< Prl PI < 0.75  (4.13) 
The ratio  of the pressuremeter modulus to  the limit pressure tends to  be a constant  characteristic of any given soil 
type.  Typical values are shown in Table 4.7. 
TABLE 4.7 Typical Menard Pressuremeter Values
Type Of Soil
Soft clay 
Firm clay 
Stiff clay 
Loose silty sand 
Sand and gravel 
Old fill 
Recent fill 
300  800 
600 - 2500 
100 - 500 
200 - 1500 
1200 - 5000 
1000 - 5000 
400  1000 
50 - 300 

54 Canadian Foundation Engineering Manual
Z 500
/0 \ >- o
::;; 400 o I -
co: >
<l o
:.::; CURVE 0 I
30 ,"-
Cl 300
u 0 I
"" U
:: Vf --.
Z. :
200 __ !/'  
! I I
o 1I-_.r:.:::=J-t:,.. -::-=:Je:.:-::-=z.-=::-::.z.- 0
o 100 200 :300 400 SOO 600 700 800 900 1000 1100
Po P
FIGURE 4.4 Typical pressuremeter and creep curves - Menard type pressuremeter
Use of Menard-Type Pressuremeter in Foundation Design
In France the Menard-type pressuremeter test results have been empirically correlated to foundation design and
perfoDnance for many soil types. If these design methodologies are used, the tests must be carried out in accordance
with standardized test procedures. Foundation designs must be limited to soil conditions similar to those used to
, develop the empirical correlations ..
The pressuremeter test is a useful tool for investigation of firm to hard clay, silt, sand, glacial till, weathered rock,
and low to moderate strength intact rock. The test can also be used for frozen soil and soil containing gas in the
pores. The Menard-type pressuremeter is not recommended for general use in clean gravelly sailor soft clay. Self-Boring Pressuremeter Test (SBPMT)
In an effort to minimize soil disturbance in relatively soft soils, the self-boring pressuremeter test (SBPMT) was
developed (Baguelin et aI., 1972; Wroth and Hughes, 1973).
The sell-boring pressuremeter is similar to a Menard-type pressuremeter as it consists essentially of a thick-wall
tube with a flexible membrane attached to the outside. The instrument is pushed into the ground and the soil
displaced by a sharp cutting shoe is removed up the centre of the instrument by the action of a rotating cutter or
jetting device just inside the shoe of the instrument. The cuttings are flushed to the surface by drill mud, which is
pumped down to the cutting head.
Once the instrument is at the desired depth, and following the dissipation of excess pore-water pressure, the
membrane surrounding the instrument is expanded against the soil. The expansion at the centre of the instrument
is measured by displacement transducers. Pore pressure cells can be incorporated into the membrane to monitor
changes in pore-water pressures.
The self-boring pressuremeter can be installed into relatively soft soils and the test results can be interpreted using
analytical methods. A summary of the methods of interpretation is presented in Mair and Wood (1987).
Site Investigations 55
The Menard-type pressuremeter test and the self-boring pressuremeter test should considered as two distinct
and separate tests. The Menard-type pressuremeter test is usually interpreted using empirical correlations related to
specific design rules. In very stiff soils or rocks, where a pre-bored hole can be made with o ~   y elastic unloading of
the soil, the Menard type pressuremeter data can be analysed from a more fimdamental baSls.
4.5.8 Dilatometer Test (DMT)
The flat plate dilatometer test is used in certain regions of NOlth America (Marchetti, 1980) for foundation design.
The tool can be classified as a logging tool that is easy to use and provides a range of empirically predicted soil
Detailed requirements for equipment, test procedure, accuracy of measurements and presentation of test results
were recommended by ASTM Subcommittee D.18.02 (Schmertmann, 1986).
A good review ofthe dilatometer test is provided by lamiolkowski et al. (1985) and Robertson (1986). An overview
of the dilatometer test and interpretation of in-situ test results is given by Lunne et al. (1989). Details of equipment
developments are presented by Mitchell (1988).
The flat plate dilatometer is 14 mm thick by 95 mm wide, with a flexible steel membrane 60 mm in diameter
on the face of the blade. The pressure for lift-off of the diaphragm, the pressure required to deflect the centre of
the diaphragm 1 mm into the soil, and the pressure at which the diaphragm returns to its initial position (closing
pressure) are recorded at each depth. Readings are made every 200 mm in depth and the dilatometer, which has
a sharpened blade, is advanced at a constant rate of 20 mm/s, with a cone penetrometer rig or similar pushing
apparatus. Correlations have been developed between dilatometer readings and soil type, earth pressure at rest,
overconsolidation ratio, undrained shear strength, and constrained modulus. However, correlations should be used
with .caution and verified by local experience before use in any specific case.
4.5.9 The Plate-Load and Screw-Plate Tests
Plate-load tests have been a traditional in-situ method for estimating the bearing capacity of foundations on soil, and
for obtaining the soil modulus for the purpose of estimating the settlement of foundations on soil or rock. Plate-load
tests involve measuring the applied load and penetration of a plate being pushed into a soil or rock mass. The test is
most commonly carried out in shallow pits or trenches but is also undertaken at depth in the bottom of a borehole,
pit or adit. In soils, the test is carried out to determine the shear strength and deformation characteristics of the
material beneath the loaded plate. The ultimate load is not often attainable in rock where the test is primarily used
to determine the deformation characteristics.
[ The test is usually carried out either as a series of maintained loads of increasing magnitude or at a constant rate
of penetration. In the former, the ground is allowed to consolidate under each load before a further increment
is applied; this will yield the drained deformation characteristics and also strength characteristics if the test is
continued to failure. In the latter, the rate of penetration is generally such that little or no drainage occurs, and the
test gives the corresponding undrained deformation and strength characteristics. The degree of drainage is governed
by the size of the plate, the rate of testing, and the soil type. !
The results of a single plate-load test apply only to the ground which is significantly stressed by the plate and this is
typically a depth of about one and a half times the diameter or width of the plate. The depth of ground stressed by a
structural foundation will, in general, be much greater than that stressed by the plate-load test and, for this reason,
the results of loading tests carried out at a single elevation do not normally give a direct indication of the allowable
bearing capacity and settlement characteristics of the full-scale structural foundation. To determine the variation of
ground properties with depth, it will generally be necessary to carry out a series of plate tests at different depths.
These should be carried out such that each test subjects the ground to the same effective stress level it would receive
at working load. Because of the difficulty in undertaking a series of tests at different depths, screw-plate tests which
are described later, may be considered.
56 Canadian Foundation Engineering Manual
One of main limitations of the plate-load test lies in the possibility of ground disturbance during the excavation to
gain access to the test position. Excavation causes an unavoidable change in the ground stresses, which may result
in ilTeversible changes to the properties which the test is intended to study.
For example, in stiff fissured over- consolidated clay, some swelling and expansion of the clay due to opening of
fissures and other discontinuities will inevitably occur during the setting-up process, and can considerably reduce
the values of the deformation moduli.
- In spite ofthis effect, the moduli determined from plate-load tests may be more reliable and often many times higher
than those obtained from standard laboratory tests. In a project that involves a large deep excavation, the excavation
may cause disturbance to the ground beneath, with a consequent effect on the deformation characteristics. In such a
case, it will be necessary to allow for this unavoidable disturbance when interpreting the results of loading tests.
(' Plate-load testing procedures are described in ASTM D-1194-72, (1987) and British Standards Institution Code
. of Practice, BS 5930 (1981). It is recommended that dial gauges, reading to an accuracy of 0.02 mm, be used for
deformation measurements. Interpretation of the test results are given in BS 5930 (1981) and Navfac DM 7.01,
It may not be possible or practical to perform plate-load tests at depth in the soil. An alternative method developed
in Europe is the screw-plate test, which uses a flat-pitch auger device that can be screwed to the desired depth in the
soil and loaded in a similar manner to a plate-load test. The horizontally projected area over the single 360
flight is taken as the loading-plate area.
A variety of loading procedures for the screw-plate test can be applied depending on the soil type and data required.
Constant rate ofload or deformation can be applied and load versus deformation plotted to obtain the modulus and
strength of the soil. Some success has been reported (Janbu and Senneset, 1973) in obtaining consolidation data
from the screw-plate test. It may also be possible to estimate the pre consolidation pressure in a sand deposit from
the test (Dahlberg, 1974).
Plate-load tests and screw-plate load tests are only a part of the necessary procedure for soil investigation for
foundation design, and should be undertaken in conjunction with other methods. These tests should be calTied out
under the direction of experts thoroughly conversant with foundation investigations and design.
4.6 Boring and Sampling
The properties of soils can be detennined from laboratory tests on samples recovered from boreholes. The quality
of the samples depends mainly on the boring method, the sampling equipment, and the procedures used to retrieve
4.6.1 Boring
Many different methods may be used to advance a borehole in soils. The more common boring methods are
summarized in Table 4.8, which has been adapted from Navfac DM 7.01 (1986). The method of advancing a casing
and washing the inside with water (washboring) is one of the most commonly used in Canada. It results in a good
quality borehole, provided the washing is done properly, i.e., using a limited water pressure and washing to, but
never beyond, the bottom of the casing. In loose sands and silts, material may rise up in the casing during washing;
bentonite mud should be used instead of water in such cases. Auger boring, including hollow stem auguring, and
rotary drilling are also commonly used methods of drilling boreholes in Canada.
4.6.2 Test Pits
Test pits excavated by a backhoe can often provide valuable information on soil characteristics at shallow depth.
Care should be exercised in excavating such pits, especially in loose sands, soft clays, or close to the water table.
General comments on test pits and test trenches are summarized in Table 4.9, which has been adapted from Navfac
DM 7.01 (1986).
Site Investigations 57
4.6.3 Sampling
For the purpose of this Manual, four classes of samples based on degree of disturbance have been defined as listed
in Table 4.10. Mechanical properties, which serve as bases for the design of foundations, can be measured only on
samples of Class 1. Such samples should usually be retrieved for the design of foundations on clays. Problem soils,
as referred to in Chapter 5, may require special sampling procedures as indicated therein.
Common samplers for disturbed and undisturbed soil samples and disturbed rock cores are summarized in Tables
4.11 and 4.12, both have been adapted from Navfac DM 7.01 (1986).
TABLE 4.8 Types ofBorings
Auger Boring
Flight Auger
Rock Core
Procedure Utilized
Hand or power op'erated augering with
periodic removal of material. In some cases
continuous auger may be used requiring
only one withdrawal. Changes indicated by
examination of material removed. Casing
generally not used.
Power operated. Hollow stem serves as a
Chopping, twisting, and jetting action of a
light bit as circulating drilling fluid removes
cuttings. Changes indicated by rate of
progress, action of rods, and examination of
cuttings in drill fluid. Casing may be needed
to prevent caving.
Power rotation of drilling bit as circulating
fluid removes cuttings from hole. Changes
indicated by rate of progress, action of
drilling tools, and examination of cuttings in
drilling fluid. Casing usually not required
except near surface.
Power chopping with limited amount of
water at bottom of hole. Water becomes
a slurry that is periodically removed with
bailer or sand pump. Changes indicated
by rate of progress, action of drilling tools,
and composition of slurry removed. Casing
required except in stable rock.
Power rotation of a core barrel as
circulating water removes ground-up
material from hole. Water also acts as
coolant for core barrel bit. Generally hole
is cased to rock.
Appli cability
Ordinarily used for shallow explorations above water table
in partly saturated sands and silts, and soft to stiff cohesive
soils. Can clean out hole between drive samples. Fast when
power-driven. Large diameter bucket auger permits hole
examination. Hole collapses in soft and sandy soils below
. water table.
Access for sampling (disturbed or undisturbed) or coring
through hollow stem. Should not be used with plug in
granular soil. Not suitable for undisturbed sampling in sand
and silt below groundwater table.
Used in sands, sand and gravel without boulders, and soft to
hard cohesive soils. Usually can be adapted for inaccessible
locations, such as on water, in swamps, on slopes, or within
buildings. Difficult to obtain undisturbed samples.
Applicable to all soils except those containing large gravel,
cobbles, and boulders (where it may be combined with
coring). Difficult to determine changes accurately in some
soils. Not practical in inaccessible locations for heavy
truck-mounted equipment (track-mounted equipment is
available). Soil and rock samples usually limited to 150
mm diameter.
Not preferred for ordinary exploration or where undisturbed
samples are required because of difficulty in determining
strata changes, disturbance. caused below chopping bit,
difficulty ofaccess, and usually higher cost. Sometimes used
in combination with auger or wash borings for penetration
of coarse gravel, boulders, and rock formations. Could be
useful to probe cavities and weakness in rock by changes
in drill rate.
Used alone and in combination with other types of boring
to drill weathered rocks, bedrock, and boulder formations.
58  Canadian  Foundation  Engineering  Manual 
TABLE 4.8  Types ofBorings (continued)
Applicability  ,
Rotarytypedrillingmethod,where coring
deviceis an integralpartofthedrill rod
stringwhichalso servesas acasing. Core
Wire-Line Efficientfordeepholecoringover30monlandand
the drillrod. Theinnerbarrelis releasedby
TABLE 4.9  Use, Capabilities and Limitations of Test Pits and Trenches
Exploration Method  General Use 
Pitsand Shafts
.Bulksampling, in-
BackhoeExcavated generallylessthan5m limitedto depthsabove
TestPitsandTrenches .deep, canbeupto groundwaterlevel,limited
rates,depthof bedrock
10mdeep. undisturbedsampling.
andgroundwater. '
IE  "'-.- ....---
. 1· qUlpmentaccesscanbe
F t  . . as ,moreeconomlCa  .
Pre-excavationforpiles ·  dIfficult. UndIsturbedsamples
and shafts,landslide
· t typo II  andblocksamplescanbe
DrilledShafts lameers lca y ..
investigations,drainage obtaInedWIth someeffort
0  Sl d ' 1" . .
2.0m.  . otte ,casIng ImltsVIsual
Capabilities  Limitations  ' 
inaccessible areas,
IE'  ,
lessmechanical limitedtodepthsabove
disturbanceof groundwaterlevel.
Fast, economical, •Equipmentaccess,generally
and sometimeslandslide
faulting, subsurface
Site Investigations 59
TABLE 4.10 Classification ofSoil Samples
4 Disturbed
A- Blocksamples
B- Stationarypistonsampler
Openthin-walledtube sampler
a 'splitspoon'
. A,B,C,D,E,F,G,H,I,J,K
A- Stratigraphy B- Stratification C- Organiccontent
D- Grainsize distribution E- Atterberglimits F- DensityIndex
G Watercontent H- UnitWeight I Permeability
J Compressibility K- Shearstrength
1.  Blocksamplesare bestwhen dealingwith sensitive, varved, orfissured clays. Whereverpossible,block
samplesshouldbetakenin suchsoils.
2.  SamplesofClass I arebeststoredinaverticalpositioninaroomwithconstanthumidityandataconstant
temperature. Therelativehumidityshouldnotbeless than80%.
3.  Testingshouldoccurasquicklyaspossibleaftersampling. Wheneverpossible,testingshouldbeperformed
4.  Becauseofinevitables.tressrelief,samplesof allclassesmaybedisturbed. Thedisturbancedependsonthe
S.  Water-contentsamplesshouldbetakenfromfreshlycutfacesofapitasthepitisadvanced. Smalldiameter
spiralaugersaresuitable for obtainingwater-contentsamplesofcohesive soils,ifcareistakento remove
freewaterfromthesample,aswellas allsoilscrapedfromupperlayersinthewallof theborehole. Water-
TABLE 4.11
is standard.
sizesup to 100
ID available.
Common Samplers for Disturbed Soil Samples and Rock Cores
Best results in soil
or rock types
in whichsamplercan
be driven. Gravels
Methods of
Causes of
disturbance or
low recovery
SPTis madeusing
hammerfalling 762
mm. Undisturbed
withliners. Some
60 Canadian Foundation Engineering Manual
TABLE 4.11 Common Samplers/or Disturbed Soil Samples and Rock Cores (continued)
Causes of
Methods of
disturbance or Remarks
low recovery
25 mm OD tubes
150 mm long.
Retractable Maximum of F or silts, clays, fine
Plug six tubes can be and loose sands.
filled in single
75 mm to 406 For most soils above
Continuous mm diameter. water table. Will not
Helical Can penetrate to penetrate hard soils
Flight depths in excess or those containing
Improper soil
types for sampler.
Hard soils,
cobbles, boulders.
Light weight, highly
portable units can be
hand carried to job.
Sample disturbance is
Rapid method of
determining soil
profile. Bag samples
can be obtained. Log
and sample depths
must account for lag
I sample at surface.
Generally 150
mm to 200 mm
ODwith 75 mm
to 100 mm ID
Same as bucket. Rotation.
Soil too hard to
A special type of
flight auger with
hollow centre through
which undisturbed
of 15 m. cobbles or boulders. between penetration
of bit and arrival of
hollow stem.
Up to 1067 mm
Up to 1220
i mm diameter
common. Larger
available. With
extensions, depths
i greater than 25 m
are possible.
Standard sizes
samples or SPT cal).
be taken.
Rapid method of
Hard soils, determining soil
Same as flight auger. Rotation.
cobbles, boulders. profile. Bag samples
can be obtained.
For most soils above
water table. Can Several types of
dig harder soil than buckets available,
above types, and including those with
Soil too hard to
can penetrate soils Rotation. ripper teeth and
with cobbles. and chopping buckets.
small boulders when Progress is slow when
equipped with a rock extensions are used.
38 mm to 75
Diamond mm OD,22 mm
Core to 54 mm core.
Barrels Barrel lengths 1.5
m to 3.0 m for
Hard rock. All barrels
can be fitted with
insert bits for coring
soft rock or hard soiL
Site Investigations 61
TABLE 4.11 Common Samplers for Disturbed Soil Samples and Rock Cores (continued)
Primarily for strong,
sound and uniform
Fractured rock.
Rock too soft.
Dlill fluid must
circulate around core
rock must not be
subject to erosion.
Single tube not often
used for exploration.
fractured friable and
soft rock.
Improper rotation
or feed rate in
fractured or soft
Has inner barrel or
swivel which does not
rotate with outer tube.
For soft, erodible
rock. Best with
bottom discharge bit.
Triple Tube Same as Double Tube.
Same as Double
Differs from Double
tube by having an
additional inner split
tube liner. Intensely
fractured rock core
best preserved in this
TABLE 4.12 Common Samplers for Undisturbed Samples
Method of Causes of
Dimensions Remarks
penetration disturbance
75 mm OD
73 mm ID
most common.
Available from
50 mm to 127 mm
. OD. 762 mm
sample length is
For cohesive
or soft soils.
Gravelly soils
• will crimp the
Pressing with
fast, smooth
stroke. Can
be carefully
Erratic pressure
applied during
hammering, gravel
particles crimping ,
• tube edge, improper
soil types for
Simplest sampler for
undisturbed samples.
Boring should be clean
before lowering sampler.
. Little waste area in
sampler. Not suitable for
hard, dense or gravelly
75 mm OD
most common.
Available from 50
mm to 127 mm
OD. 726mm
sample length is
For soft to
medium clays
and fine silts.
Not for sandy
Pressing with
steady stroke.
Erratic. pressure
during sampling,
allowing piston
rod to move during
press. Improper
soil types for
Piston at end of sampler
prevents entry of fluid
and contaminating
material. Requires heavy
drill rig with hydraulic
drill head. Generally less
disturbed samples than
Shelby. Not suitable for
hard, dense or gravelly
l soil.
62 Canadian Foundation Engineering Manual
TABLE 4.12 Common Samplers/or Undisturbed Samples (continued)
75 mmOD
most common
available from
50 mm to 100
mm OD, 914 mm
sample length.
Samplers from 89
mm OD to 197
mm OD. (60.3
mm to 160mm
size samples.) 60
mm sample length
F or silts-clays
and some
sandy soils.
Can be used
for stiff to
hard clay, silt
and sands
with some
Hydraulic or
compressed air
Rotation and
clamping of
drill rods, en-atic
operating sampler.
Poor drilling
Needs only standard drill
rods. Requires adequate
hydraulic or air capacity
to activate sampler.
Generally less disturbed
samples than Shelby.
Not suitable for hard,
dense or gravelly soil.
Inner tube face projects
beyond outer tube
which rotates. Amount
of proj ection can be
adjusted. Generally
takes good samples: Not
suitable for loose sands
is standard. soft rock.
and soft clays.
Sampler 105 mm Differs from Denison in
OD, uses 75 mm
: Same as
Shelby Tubes.
610 mm sample
Same as
Same as Denison.
that inner tube projection
is spring controlled.
Often ineffective in
length. cohesionless soils.
Highest quality
Hand cut
block or
Sample cut by
sampling in
cohesive soils,
residual soil,
Change of state
of stress by
Requires accessible
excavation. Requires
dewatering if sampling
below groundwater.
rock, soft rock.
4.6.4 Backfilling
Backfilling of boreholes and test pits should be done carefully. The quality and compaction of the backfill material
should be sufficient to prevent hazard to persons or animals, and should prevent water movement or collapse,
particularly in drilling for deep excavations or tunnels. In the case ofa contaminated care is required to minimize
possible flow through the boreholes to water supply aquifers.
4.7 Laboratory Testing of Soil Samples
It is beyond the scope ofthis Manual to cover in detail all laboratory testing techniques now in use in soil mechanics.
However, the more common tests are summarized in Tables 4.13, 4.14, 4.15, 4.16 and4.17 to provide some guidance
on standard (ASTM) or suggested test procedures, the variations that maybe appropriate, and the type and size of
samples required. These tables have been adapted from Navfac DM 7.01 (1986). Testing procedures references
given in the above tables are summarized for convenience in Table 4.18. The index property tests in Table 4.13 are
Site Investigations 63
considered in more detail in Chapter 3. Other comments for general guidance are given in the following paragraphs
and these comments are essentially those given in Navfac DM 7.01 (1986).
4.7.1 Sample Selection
Samples to be tested should be representative of each significant stratum, or be an average of the range of materials
present, depending on the design and project requirements. A thin stratum can be significant if it has engineering
features such as being weak or cemented. Ifit appears difficult to obtain representative samples because of variations
in the stratum, it may be necessary to consider subdivision of the stratum for sampling, testing, and design pm-poses.
In general, tests on samples of mixed or stratified material, such as varved clay, should be avoided. Usually such
results are not indicative of material characteristics and better data for analysis can be obtained by testing the
different materials separately.
Undisturbed samples for structural properties tests must be treated with care to avoid disturbance. An "undisturbed"
sample found to be disturbed before testing normally should not be tested. Fine-grained cohesive samples naturally
moist in the ground should not be allowed to dry before testing, as irreversible changes can occur; organic soils
are particularly sensitive. Soils with chemical salts in the pore water may change if water is added, diluting the salt
concentration, or if water is removed, concentrating or precipitating the salt. Organic soils require long-term low
temperature (60°C) drying to avoid severe oxidation (burning) of the organic material.
4.7.2 Index Property Tests
Index properties are used to classify soils, to group soils in major strata, and to correlate the results of structural
properties tests on one portion of a stratum with other portions of that stratum or other similar deposits where only
index test data are available. Procedures for most index tests are standardized (Table 4.13). Either representative
disturbed or undisturbed samples are utilized.
Tests are selected after review of borehole data and visual identification of samples recovered. In general, the test
program should be planned so that soil properties and their variation can be defined adequately for the lateral and
vertical extent of the project concerned.
4.7.3 Tests for Corrosivity
The likelihood of soil adversely affecting foundation elements or utilities (concrete and metal elements) can be
evaluated on a preliminary basis from the results of the tests referenced in Table 4.13. The tests should be run on
samples of soil which will be in contact with the foundations and/or utilities in question and typically these will
be only near-surface materials. Usually the chemical tests are run only if there is reason to suspect the presence of
those ions.
4.7.4 Structural Properties Tests
Tests for structural properties should be planned for particular design problems. Rigid standardization oftest programs
is inappropriate. Typical tests for determining structural properties are given in Table 4.14. Perform tests only on
undisturbed samples or on compacted specimens prepared by standard procedures. In certain cases, completely
remolded samples are utilized to estimate the effect of disturbance. Plan tests to determine typical properties of
major strata rather than arbitrarily distributing tests in proportion to the number of undisturbed samples obtained. A
limited number of high quality tests on carefully selected representative undisturbed samples are preferred.
4.7.5 Dynamic Tests
Dynamic testing of soil and rock involves three ranges: low frequency (generally less than 10 hz) cyclic testing,
resonant column high frequency testing, and ultrasonic pulse testing (Table 4.15). The dynamic tests are used
to evaluate foundation support characteristics under cyclic or transient loadings such as machinery, traffic, or
64 Canadian Foundation Engineering Manual
earthquakes. For earthquake loading, a primary concern is often liquefaction. Young's modulus (E
)' shear modulus
(0), and damping characteristics are detennined by cyclic triaxial, cyclic simple shear, and resonant column tests as
shown on Table 4.15. Table 4.16 shows the range of strain levels for which each test is applicable.
From the resonant frequency of the material in longitudinal and torsional modes, Poisson's ratio can be computed
from test data. Foundation response to dynamic loading and the effect ofwave energy on its sun'oundings is studied
in the light of these test results. The ultrasonic pulse test also evaluates the two moduli and Poisson's ratio, but the
test results are more reliable for rocks than for soils. Dynamic tests can be run on undisturbed or compacted samples
and the number of tests will depend on project circumstances.
4.7.6 Compaction Tests
In exploring for bon-ow materials, the number of index tests or compaction tests may be required in proportion to
the volume of bon-ow involved or the number of samples obtained. The requirements for compacted soil sample
tests are given in Table 4.17.
Structural properties tests are assigned after bon-ow materials have been grouped in major categories by index and
compaction properties. Select samples for structural tests to represent the main soil groups and probable compacted
condition. The number of compaction tests will depend on project requirements and bon-ow variability.
4.7.7 Typical Test Properties
Various con-elations between index and structural properties are available showing the probably range of test
values and relation of parameters. In testing for structural properties, correlations can be used to extend results to
similar soils for which index values only are available. Correlations are of varying quality, expressed by standard
deviation, which is the range above and below the average trend, within which about two-thirds of all values occur.
These relationships are useful in preliminary analyses but must not supplant careful tests of structural properties.
The relationships should never be applied in final analyses without verification by tests of the particular material
TABLE 4.13 Requirements/or Index Properties Tests and Testing Standards
Reference for
standard tests
Variations from standard
test Procedures, sample
Size or weight sample for
Moisture content
of soil
ASTM D2216 (1)
None. (Test requires natural
moisture content.)
As large as convenient.
Moisture, ash, and
organic matter of
peat materials
Dry unit weight
Specific gravity:
(relative density)
ASTM P2974 (1)
Determine dry a sample of
measured total volume. (Requires
undisturbed sample.)
As large as convenient.
Material smaller
than No. (4.75
mm) sieve size
ASTM P854 (1)
Volumetric flask preferable;
vacuum preferable for de-airing.
25 g to 50 g for fine-grained soil;
150 g for coarse- grained soils.
Site Investigations 65
TABLE 4.13  Requirements/or Index Properties Tests and Testing Standards (continued)
Variations from standard 
test Procedures, sample 
sieve  size 
ASTM  e127 (1) 
Use fraction passing No.  40 
(0.425  mm) sieve;  material should  i
,  not be dried before testing. 
Atterberg Limits:
Liquid Limit  ASTM D423 (1)  I None. 
Plastic Limit  ASTM D424 (1) 
Ground glass plate preferable for 
Shrinkage Limit  (4) 
In some cases a trimmed 
specimen of undisturbed material 
may be used rather than a 
remolded sample. 
Sieve analysis  ASTM D422 (1) 
Selection of sieves to be utilized 
may vary for samples of different 
ASTM D422 (1) 
Fraction of sample for  hydrometer 
analysis may be that passing 
No.  200 (0.075  mm) sieve.  For 
fine-grained soil entire sample 
may be used.  All material must 
be smaller than No.  10 (2.0 mm) 
,  sieve. 
100  g to  500 g 
30 g 
500  g for  soil with grains to  9.5  mm; 
to 5,000 g for soil with grains to 
65  g for fine-grained soil;  115  g for 
sandy soil. 
Sulphate content 
Chloride content 
ASTM D  1293  (1) 
Several alternative procedures in  i Soil/water solution prepared, see 
reference.  reference. 
Several alternative procedures in  Soil/water solution prepared, see 
I reference. 
Reference is  for pH of water. 
,  For mostly solid substances, 
• solution made with distilled water 
and filtrate tested;  standard not 
66  Canadian Foundation  Engineering Manual 
TABLE 4.13 Requirements for Index Properties Tests and Testing Standards (continued)
Size or weight sample for
Written standard not available. 
Follow guidelines provided 
by manufacturers of testing 
In-situ test procedure.  (6)
Resistivity (field) 
(a)  Number in parenthesis indicates reference number in Table 4.18 
(b)  Samples  for  tests  may  either  be  disturbed  or  undisturbed;  all  samples  must  be  representative  and  non-
segregated; exceptions noted. 
(c)  Weights of samples for tests on air-dried basis. 
TABLE 4.14 Requirements for Structural Properties
Reference for
Size of weight of sample for
Variations from suggested
test (undisturbed, remolded,
test procedures
tests (a)
or compacted)
Constant Head 
permeable soil) 
(2), (4) 
Sample size depends  on maximum 
grain size, 40 mm diameter by 350 
mm height for  silt and fine  sand. 
Variable Head  (2),  (4) 
Generally applicable to fine-
grained soils . 
Similar to constant head sample. 
.  Limited to soils containing less 
Constant Head  than  10% passing No.  200  (0.075 
mm)  sieve size.  For clean coarse- (  coarse-grained 
(l ),(4) 
grained soil the procedure in 
reference (4)  is  preferable. 
Capillary head for certain fine-
Capillary Head 
(2)  grained soils may have to  be 
detennined indirectly. 
Consolidation  (2) 

Collapse Potential  I 

To  investigate secondary 
compression, individual loads 
may be maintained for  more than 
24 hours. 
!. Sample diameter should be at least 
ten times the  size of the  largest soil 
200 g to 250 g dry weight. 
Diameter preferably 63  mm 
or larger.  Ratio of diameter to 
thickness  of 3 to  4. 
Diameter preferably 63  mm 
or larger.  Ratio of diameter to 
thickness  of 3 to 4.l 
Two  specimens for each test, 
with diameter 63  mm  or larger. 
Diameter to height ratio  3 to 4. 
Site Investigations 67
TABLE 4.14  Requirementsfor Structural Properties (continued)
Size of weight of sample for  Reference for 
Variations from  suggested 
test (undisturbed, remolded,  suggested 
test procedures 
or compacted)  tests (a) 
Shear Strength:
Generally  12  mm thick, 75  mm by
Limited to tests on cohesion less 
75  mm or ASTM D3080 
soils or to consolidated shear tests  Direct Shear 
100 mm by  100 mm in  plan, or (1), (2) 
on fine-grained soils. 
e  uivalent. 
Alternative procedure given in 
Similar to triaxial test samples. 
:  Reference 4. 
  __+-_______j--____________-1  be less than 3 and greater than 
Consolidated- 2.  Common sizes are:  71  mm 
undrained  (2),(3),(4)  I Consolidated-undrained tests  may  diameter,  165  mm high.  Larger 
.-.::..-------+---------li run with or without pore pressure  :  sizes are appropriate for gravelly 
Consolidated- (2),(3),(4)  i measurements, according to basis  materials to  be used in earth 
drained  • for design.  embankments. 
Unconsolidated- i Ratio of height to  diameter should 
ASTM D2850 (1) 
" Block of undisturbed soil at least 
Vane  Shear  ...i '-;-:-;:--L_th_r_e_e_t_im_e_s_d_i_m_e_n_s_io_n_s_o_f_v_an_e_,_
(a)  Number in parenthesis indicates reference number in Table 4.18. 
TABLE 4.15  Requirementsfor Dynamic Tests
Reference for 
Test  suggested 
tests (a) 
Cyclic Loading
Simple Shear 
Torsional Shear  (10)  Can use hollow specimen. 
Same as  for  triaxial test for 
Resonant Column Can use hollow specimen,  structuraLproperties; lengths 
sometimes greater. 
Ultrasonic Pulse
D40151(1 )(11) 
Same as  for  triaxial test for 
:  structural properties 
Soil  (12) 
Same as  for  triaxial test for 
structural properties, 
ASTM D2845 
Prism, length less than five 
times  lateral dimension; lateral 
dimension at least five  times 
length of compression wave. 
68 Canadian Foundation Engineering Manual
(a)  Number in parenthesis indicates reference number in Table 4.18 
(b)  Except for  the  ultrasonic  pulse  test  on  rock  and  resonant  column  tests,  there  are  no  recognized  standard 
procedures for dynamic testing.  References are to descriptions of tests and test requirements by recognized 
authorities in those areas. 
TABLE 4.16 Capabilities o/Dynamic Testing Apparatus
Cyclic  Attenu-
Modulus  •  Modulus  i ~ Damping 
Shearing Strain Amplitude (%) i Shear  I Youngs  I
Stress  anon 
10-4  10.



1 G ! E •
Resonant column (!§ample)  X  X X
Resonant column (hollow samule) 
X X X 
X X ~ 9 .   . ' i c pulse  X 
Cy:clic Triaxial  X X  X 
Cyclic Simole Shear  X  X X 
Typical Motion Characteristics 
X • Indicates the properties that can be detennined. 
Properly  Strong 
Close in
Designed  Ground 
Machine  Shaking  . 
Earth  k  ExplosIon 
I' qua  e I
TABLE 4.17 Requirements/or Compacted Soil Sample Tests
Standard Proctor 
2.49 kg hammer, 
305 mm drop 
ASTM D698 (1)
Preferable not to reuse samples 
for  successive compaction 
Each determination (typically 4 or 5 
determinations per test): 
Method A:  3.0 kg 
Method B: 6.5 kg 
Method C:  4.5 kg 
Method D:  10 kg 
Site Investigations 69
TABLE 4.17 Requirementsfor Compacted Soil Sample Tests (continued)
ModifiedProctor Preferablenottoreusesamples
MethodA: 3.5 kg
4.54 hammer,
457 mmdrop
of Cohesion lessSoils
for successivecompaction
MethodB: 7.5 kg
MethodC: 5.5 kg
MethodD: 1l.5kg
Varies from4.5 kgto 60
11.5 dependingon gradation. Ratio
4.5 kg to7 kg dependingon
ResistanceR-value ASTMD2844(l)
4.5 kg to7kgdependingon
ExpansionPressure AASHTOTl90(7) proceduresofTable4.14may
Permeabilityand soils. Altematively,testing 7kgofmaterialpassingNo. 4(4.75
compression proceduresof Table4.14may mm)sievesize.
(a) NumberinparenthesisindicatesreferencenumberinTable4.18
(b) Forothersourcesofstandardtestprocedures,seeTable4.6.
(c) Weightofsamplesfortestsgivenonair-driedbasis.
TABLE 4.18 References Cited in Tables 4.13, 4.14, 4.15and 4.17
1.  American Society for Testing and Materials "Annual Book ofASTM Standards, Part 19 - Natural
BuildingStone,SoilandRock,Peat,Mosses,andHumus;Part14- ConcreteandMineralAggregates;
Part4 - StructuralSteel"; :ASTM,Philadelphia,Pennsylvania.
2.  Lambe,T.W. (1951). "Soil forEngineers"; JohnWiley,NewYork.
3.  Bishop, A.W. andDJ. Henkel(1962). "TheMeasurementofSoil Properties intheTriaxialTest";
4.  Office ofthe Chiefof (1970). "Laboratory Soils testing"; Department of the Army,
70 CanadianFoundationEngineeringManual
TABLE 4.18 References Cited in Tables 4.]3, 4.14, 4.15and 4.17(continued)
5. AmericanSocietyof AgronomyandtheAmericansocietyforTestingandMaterials(1965). "Methods 
6. NationalBureauofStandards. "UndergroundCorrosion";CircularC450,UnitedStatesGovernment
7. AmericanAssociationofStateHighwayandTransportationOfficials(1978)"StandardSpecifications
for TransportationMaterials andMethods ofSamplingand testing"; PartII,AASHTO, Washington,
8. Jennings, lE. and K. Knight (1975). "AGuide to Construction on or with Materials Exhibiting
AdditionalSettlementDueto CollapseofGrainStmctures"; SixthRegionalConferenceforAfricaon
9. Silver,MarshalL. (1976). "LaboratoryTriaxialTestingProcedurestoDeterminetheCyclicStrength
ofSoils"; prepared under contract to US NRC, (contract No. WRC-E(11-1)-2433), Report No.
10. Wood,R.D. (1978). "MeasurementofDynamicsoilProperties";ASCEGeotechnicalDivisionSpecial
11. Dernevich, V.P., B.O. Hardin and D.J. Shippy (1978). "Modulus and Sampling ofSoils by the
ResonantColumnMethod";ASTM, STP654.
12. Stephenson,R.W. (1978). "UltrasonicTestingforDeterminingDynamicSoilModulus";ASTM,STP
13. BureauofReclamation (1974). "Permeability and SettlementofSoils"; EarthManual, Designation
4.8 Investigation of Rock
4.8.1 General
orintotherocksurfaceorfor excavationsinrock. Thesiteinvestigationtechniquesusedinrockshouldreflectthe
designdatarequired. Pertinentinformationtobedeterminedshouldinclude:
Geologicalcharacteristicsofthesitetoprovideanoverviewofthesiteandprovidethebasisfor correlation
betweenboringsandoutcropmapping. Reviewofexistingpublisheddataisuseful.
Rock, ifpresenton the surface, should be mapped and the outline ofthe rock surface and rock surface
elevations recorded. Outcrops should be mapped using conventional mapping techniques. Geophysical
techniques suchas seismicrefractiontechniquesmaybeusefulfor detectingtop ofrocksurfacescovered
withoverburden. Geophysicaltechniquesshouldbeconfirmedbyboreholeswhenpositionofrocksurface
Rockat depthshouldbe investigated using boreholes. The recovered rock samples shouldbe classified
anddescribedasnotedbelow. Useofdownholegeophysicscanaddvaluabledatato asingleboreholelog.
The nature oftheseams washedawaybydrilling maybedeterminedand, in some instances, engineering

Site Investigations 71
propeliies can be correlated by geophysical logs or borehole camera logging.
Extent and character of alteration and weathering, and an assessment of the sensitivity or resistance to
weathering or chemical reaction. (Includes slaking, swelling or acid drainage generation).
Characteristics and orientations (including folds and fold axes) of discontinuities such as bedding planes,
faults, joints, foliations or cleavage planes.
Strength and compressibility of the rock mass.
Permeability and groundwater levels.
In permafrost rich areas of the Canadian north, care must be taken to determine ice content within
discontinuities. Ice rich lenses could melt and cause settlements. Special drilling procedures with cooled
drilling fluids may be warranted.
4.8.2  Core Drilling of Rock 
When information is required at depth in rock, boring may be required. Attention to the overall geological setting
may indicate if detached bedrock may be present. The borehole or boreholes should be carried well below the first
encountered top of rock to confirm the presence of bedrock.
Boreholes for the investigation of rock can be advanced by many different methods, as discussed in detail by
Franklin and Dusseault (1989). These may include:
rotary core drilling with double or triple core barrels, with or without wire line, with air or water flush;
rotary tricone drilling with air or water flush; and
percussion drills, down-the-hole-harnmers, etc.
Rock drilling provides rock core of various diameters, typically ranging from NQ to HQ sizes for geotechnical
investigations. Cores recovered using triple tube wire line core barrels are the least disturbed and are useful
for assessing discontinuity characteristics. Oriented core may be used to determine spatial relationships of the
discontinuities. Core recovered using wireline double tube systems provide pieces of sequential core but often
the discontinuities are disturbed and the true nature may be difficult to determine. Sheared zones may be badly
disturbed. Tricone drilling provides cuttings ofthe rock material which do not allow any assessment ofdiscontinuity
characteristics. Infilling material is often lost. Percussion drill and down-the-hole hammer drills are excellent for
production drilling.
Drilling of soft bedrock may require the use of a Pitcher sampler or Christenson spring loaded bit. Soft seams of
sheared material may still be lost during drilling. Large diameter (1 m) holes augured or churn drilled to depth, then
mapped and sampled from a mobile cage, have been successfully used to identify zones of weakness and to recover
samples for direct shear testing.
When a drilling program is designed it is often prudent to seek the advice of an experienced drilling contractor,
particularly with respect to drill suitability. Where it is important to recover high quality cored rock or if testing
down hole is an integral part ofthe program, an hourly rate, testing rate, or some combined basis for payment should
be sought, as opposed to a rate per length drilled.
Care must be taken to ensure maximum possible core recovery. Changes in drilling noise, vibrations, pressure on
drill bit, colour, pressure and flow of drilling water, and all other drilling operations should be carefully recorded.
72  Canadian Foundation Engineering Manual 
Care should be taken when drilling through overburden to  bedrock to  ensure that bedrock has  in  fact  been reached 
and that a floating large slab of rock in a till  or colluvium or residual soil has not been misinterpreted. The borehole 
should be  drilled a minimum of 3 meters  into bedrock,  in more than one  borehole,  to  confirm whether bedrock or 
a boulder has been found.  For some geological conditions, such as when floating rock slabs are  possible, the depth 
of drilling should be  increased. 
In-situ  testing  in  the  borehole  is  recommended  whenever  possible.  The  rock  exposed  along  a  borehole  will  be 
disturbed  by  drilling,  but  the  position  and  orientation  of the  discontinuities  will  not  be  affected.  Testing  using 
downhole  geophysical  techniques  and  observation  using  a  borehole  camera  or  probe  can  provide  velY  useful 
information about the integrity of the rock mass. 
4.8.3  Use of Core Samples  Identification and Classification
Information about identification and classification of rocks  is  presented in Chapter 3 of this  manual.  Core  logging 
procedures should include collection of this data. Particular attention should be paid to the identification of the rock 
discontinuities,  including their nature  and origin,  geometry  and weathering.  Colour photographs  of the  rock core, 
, presented in the con'ect stratigraphic sequence and with the core depths indicated, are a useful record and can assist 
office studies.  Laboratory Testing of Core Samples
Laboratory tests (described in Chapter 3) are useful for determining the strength and deformability of the intact rock 
elements.  Such results may not be representative of the actual rock mass, since they are performed on samples free 
of discontinuities.  The relative importance of the rock characteristics  versus the  rockmass  characteristics depends 
upon the size of the foundation and the effect ofthe discontinuities. The range ofpossible discontinuity conditions is 
considered in the Geological Strength Index approach (GSI) discussed in Chapter 3. In this method, a combination 
ofthe surface conditions ofthe discontinuities and the rockmass structural state provide a factor to modifY the intact 
. rock strength to more representative rockmass strength. This evaluation relies upon an assessment of the intact rock 
strength  and the rockmass  conditions.  Where large structures are  to  be founded  on  or in rock,  insitu tests  such  as 
described in the next section should be conducted. 
4.8.4 In-situ Testing
In-situ testing ofrockmass deformation characteristics should be carried out for design oflarge structures supported 
in and on rocks.  A variety oftests, as discussed by La and Hefny (200 I)  are summarized here: 
•  Plate  load  test.  This  is  the  most  common  in-situ  rock  mechanics  test  method.  Standards  for  testing 
procedures and interpretation are given by ISRM (1979a, b) and ASTM (D4394-84 and D4395-85). In  this 
simple test, a load is applied to a prepared flat surface ofthe rock mass through a plate and the deformation 
is measured.  The  deformation modulus  is then calculated from  this  data.  The  main disadvantages  of this 
technique include the expense ofpreparing the site for the testing, only a small volume ofrock is tested, and 
the common presence of a disturbed zone around the excavation usually leads to conservative results. 
Large :flat jack test.  In this  simple test  (ISRM,  1986),  large hydraulic :flat jacks are  inserted into  a nanow 
slot cut  into  an  exposed  rock surface.  Pressure  applied to  the :flat jacks results  in  measured  normal  rock 
deformation. The rock mass deformation modulus can be determined from this data. The advantages of this 
test include the fact that a large  volume of the rock mass is  influenced by the test,  and that it is  performed 
in  a relatively undisturbed zone  of the  rock mass.  The disadvantages  include the  need for  skilled drilling 
personnel,  the weak theoretical  background for  the  interpretation,  seating problems  when  conducting the 
test, and the  fact that most :flat jacks are generally non-recoverable. 
Site Investigations 73
Dilatometertest. Dilatometertests maybe carriedusingeitherflexible orstiffequipment. Intheflexible
type, (ISRM, 1987) a uniformly distributed pressure is applied to the borehole wall by hydraulically
expandingaflexible membrane. The resulting hole expansion is detenninedby measuring the volume of
is determinedfromtherelationshipbetweentheappliedpressureanddeformation.
In the stifftype, (ISRM, 1996; ASTM, D4971-89) unidirectional pressure is applied to the borehole wall by
two opposed curvedsteel platens, each covering a 90-degree sector. The advantage ofthe easily performed and
inexpensive dilatometer test is the ability to perform the test at different depths and locations. As a result, the
onlyasmallvolumeofrockis influencebythetest.Thereforethemodulusobtainediscomparabletothelaboratory
modulus, butnottotherockmassmodulus.
Lo and Hefny (200I) and ASCE (1996) describe other in-situ tests involving tunnelling, dynamic testing using
4.9  Investigation of Groundwater 
4.9.1  General 
Groundwater is a·critical factor in foundation design and construction. Many foundation problems are directly
or indirectly relatedto groundwater, hence groundwaterconditions, both physical andchemical, shouldbe given
carefulattentionduringall stagesofasoilsinvestigation.
• theexistenceof groundwater- normal,perched,hydrostatic, orartesian;
• theexactlevelofthegroundwatertable,andof thelowerlimitofperchedgroundwater; 
• thevariationof thesecharacteristicsoverthesiteandwithtime,and
• thechemicalcompositionofthegroundwater.
A thorough evaluation ofgroundwater measurement, instrumentation selection, installation and observations is
beyondthescopeofthisManual.Otherreferencesshouldbe consultedsuchasDunnic1iff(1988).Considerablecare
goodsystemwillprovidetherequiredinformationfor designwhileapoorsystemcangivemisleadingresults.
4.9.2  Investigation  in  Boreholes 
Fieldrecords shouldbemade duringdrillingof allboreholeobservationsrelatedto groundwaterandtheserecords
conditions. The water level should be measured during drilling and after the completion ofthe borehole. All
informationshouldberecordedontheboringlog, alongwiththedepthof theborehole andthedepthofthecasing
The groundwaterobservationsmade in openboreholes should be treatedwithcaution. Groundwaterobservations
ofthesematerialsandthelongerperiodsof timerequiredbeforethewaterlevelintheboreholereachesequilibrium.
One ofthe more common methods for measuring groundwater levels is to install an open observation well for
the full depth ofthe borehole. The observation well usually consists ofa pipe with a perforated section at the
bottom. The pipe extends to the groundsurface andis backfilledfor the entire hole withsand, with a sealat the
74 Canadian Foundation Engineering Manual
ground surface. The major disadvantage is that different soil strata may  be under different hydrostatic pressure, and 
the  groundwater level recorded may  be  inaccurate and  misleading.  Furthermore the  continuous  sand  backfill  may 
allow  cross-connection of water  in  different  strata  and  this  could result in  misleading  observations.  Most  of the 
disadvantages  of the  open borehole or observation well can be overcome by  installing open standpipe piezometers 
that  are  sealed into specific  strata and these are discussed below. 
4.9.3 Investigation by Piezometers
In  all  cases  where  groundwater  conditions  are  important  in  design,  or  are  difficult,  or  where  direct  borehole 
observation is  not applicable, the groundwater conditions should be investigated by the installation and observation 
ofpiezometers (pore-pressure meters). In designing such installations, attention should be paid to the stratigraphy (for 
location of the  piezometer tips)  and the soil type (for selection of the type of piezometer). Time lag is a particularly 
important parameter  in  the  selection  of piezometer  type,  and  proper  installation  is  critical  to  the  performance  of 
piezometers.  In particular,  when installed  in  a borehole, piezometers should be  isolated from the borehole by, for 
instance, sealing with bentonite  a small distance  above  and below the  piezometer tip  (which should be surrounded 
by clean sand). 
The simplest and generally considered to be the most reliable piezometer is  the  open standpipe piezometer installed 
in  the  borehole  at  the  depth  required  with  sand  backfill  placed  around  the  porous  end  within  the  depth  of the 
stratum being observed. This stratum is isolated by placing bentonite seals above  and below the sand backfill. The 
borehole  above the upper bentonite seal should be backfilled with a special sealing grout.  For further  details  refer 
to Dunnicliff (1988). 
If the  foundation  strata in which  the  piezometers  are  to  be  located  are  of low  permeability  and  the  time  lag  for 
open standpipe piezometer measurements  is  excessive,  or if piezometers are required in locations  inaccessible for 
reading the vertical open standpipe piezometers, then different types ofpiezometers will be required. Other types of 
piezometers can be grouped  into  those that have  a diaphragm between the transducer and the porewater and those 
that do not.  Instruments in the first group are piezometers with pneumatic, vibrating wire, and electrical resistance 
strain gage transducers.  Instruments in the  second group are open standpipe and  twin-tube hydraulic piezometers. 
Refer to  other sources  such as Dunnicliff (1988) for further details. 
4.10 Geotechnical Report
Data from  site  investigations  are  usually referred to  frequently  and for  many different purposes during the  design 
period,  during  construction,  and  often  after  completion  of the  project.  Appropriate  reports  should  therefore  be 
prepared for  each site investigation.  They  should be clear, complete, and  accurate.  The following  outline  may be 
used  as  a guide in arranging data in such reports: 
Terms of reference of the  investigation 
Scope of the investigation 
Procedures and equipment used in the investigation 
Proposed-structure or structures 
Geological setting 
Topography, vegetation, and other surface features 
Soil profile and properties 
Groundwater observations 
Existing adjacent structures 
Foundation studies, including alternatives 
Recommended field  instrumentation and monitoring 
Recommended construction procedures,  if appropriate 
Recommended field  services 
Conclusions and recommendations 
Site Investigations  75
Limitations of the investigation
Graphic presentations
Map showing the site location, including north arrow
Detailed plan of the site showing contours and elevations, and location of proposed structures, boreholes,
and adjacent stmctures and features of importance
Boring logs, including all the necessary pertinent information on soil, rock, and groundwater
Stratigraphical and geotechnical profiles
Groundwater profiles
Laboratory data
Special graphic presentations
4.11 Selection of Design Parameters
4.11.1 Approach to Design
There are four distinct categories of calculation methods in geotechnical design as follows (Hight and Leroueil
1. Empirical
Direct use of in-situ or laboratory test results, relying on correlation
with performance data and experience
2. Semi-empirical Indirect use of in-situ or laboratory test results, combining field
experience and simple theory
3. Analytical Theoretical models based on elasticity, plasticity, etc.
4. Numerical Complex soil models based at least in part on real soil behavior
The complexity ofsoil behaviour has resulted in a need for empiricism and so a substantial number ofcurrent design
methods in geotechnical engineering practice fall in categories I and 2. This has led to the development of a large
number of design methods, each applicable to one specific design case. Charts are frequently available to aid in
design. Because design methods were developed using properties determined in a particular manner, it is important
to follow design approaches in their entirety as the previous success of the approach may rely on compensating
errors. One area in which this is particularly important is pile design. Pile installation alters soil properties. The
magnitude of the change in soil properties depends on the installation method and on the initial conditions. This
effect of changes in ground conditions as a result of foundation constmction must be specifically considered during
site characterization and selection of design parameters.
Historically, design has involved separate consideration of strength and deformation. Limit equilibrium has been
used to design against failure and linear elasticity   the non-linear theory of consolidation has been used to estimate
deformation. In the limit equilibrium approach, the mobilized strength at failure will likely vary along the particular
failure surface under consideration and will differ from peak strength. Site variability and soil strength anisotropy
become important when selecting the design strength.
Advances in numerical modeling have given engineers the capability to model soil response to all stages of site
development. Constitutive models have been developed which account for some or all of the above aspects of
material behaviour. These models have been implemented in numerical models in commercially available computer
programs. The determination of appropriate input parameters requires judgment and a good understanding of soil
behaviour. It is critical that any model used in design should be calibrated by comparison to case histories of similar
foundation elements or systems in similar soil conditions.
76  Canadian Foundation Engineering  Manual 
4.11.2 Estimation of Soil Properties for Design
To characterize the engineering behaviour ofthe soil or rock at a site, the following parameters are critically
• In-situstresses
Overconsolidationratio to allowdefinitionofyieldstresses
Potential for strain weakening or swelling, that is, the existence ofsoil structure or expansive clay
Thepresenceofanyjointsorothermacrostructurethatmay dominatetheengineeringbehaviour. 
Once the materials have been identified, estimates ofcharacteristic behaviour can be based on one orall ofthe
Previousexperience in materialswithsimilarclassificationpropertiesandofsimilargeologicaloriginand
• Sitespecificlaboratorytesting 
Prototypetesting,e.g. footing orpileloadtests. 
Comparison to similar materials
Forestimatesofsoilpropertiesbasedontheknownbehaviourof similarmaterial,itisnecessarytohaveameansof
identifyinghow closelythematerialsatthesiteresembleothers for whichdatahavebeenpublished. Examplesof
SanFranciscoBayMud, Ottawasand, FraserRiversand, Toyourasand,Leighton Buzzardsand,Tieinosand,etc.
ClassificationpropertiessuchasAtterbergLimitsorsoilgradationscanbeusedto assistthe engineerto makethis
In-situ testing
Theresultsofin-situtestingcanalsobeusedas anindexofsoilbehaviour.Traditionally,propertycharacterization
has beenbasedon blow countsorSPTN-valuesmeasuredduringsplit-spoonsampling. Morereliable techniques
such as thepiezometercone penetration testing (CPT) are now available. It is important to note that the loading
Ifthe in-situ test parameters are to be correlated to engineering behaviour, the soil being investigated
shouldresemblevery closelythe soilusedto developthe correlation. Thisrequires similarityofdrainage
The in-situ test must be carried out in exactly the same way as it was during the development ofthe
interpreting the data must have a strong understanding ofsoil behaviour and must exercise extreme
diligence intheselection, specificationandobservationofthe in-situ tests. This is particularly important
whenattemptingtoapplycorrelationsdevelopedinonegeologicalregimeto soilsorrocksinanother.
Laboratory testing
All sampling causes some soil disturbance. The effect ofsample disturbance on the soil behaviour obtained in
Site Investigations 77
preparation oftest specimens. In general, disturbance leads to a reduction in stiffness andpeakstrength ofsoils
when tested atstressesrepresentativeofin-situconditions. Disturbancemayalsomakeitdifficultto delineate the
yieldstressofthesoil. In sands,ithasbeenobservedthatattemptstoobtainundisturbedsamplesbymethodsother
thanin-situfreezing andcoring, typicallyresultinsamplesofloosesandthat aredenserthanthe in-situcondition
andsamplesofdensesandthatarelooserthanthe in-situcondition.
Prototype testing 
Materialbehaviour canalso becharacterizedbyload testinga prototype ofaparticularfoundation element. The
soilelementsaffectedbythetestwillexperiencearangeofstressesandstrainsanditis importantto ensurethatthe
zoneofsoilinfluencedbythetestis representativeofthesoiltobeloadedbytheactualfoundation. Thestrainrates
imposedduringthe testsmustalso beconsideredin relationto those in effectunderapplication ofworking loads
4.11.3  Confirmation of Material  Behaviour by Construction Monitoring 
Muchcanbelearnedfrommonitoringof soilbehaviourduringconstmctionandduringtheservicelifeofstmctures.
4.12  Background Information for Site Investigations 
andterritorial). Avarietyofinformationis alsonowdirectlyavailableontheinternet, orgovernmentand private
sourcescanbelocatedusinganinternetsearch. Examplesofvaluableresourcesareas follows:
•  SatelliteandUnusualImagery
•  LandUseandPlanningSurveys
78 Canadian Foundation Engineering Manual
Special Site Conditions 
5. Special Site Conditions
5.1  Introduction
The following sections give brief descriptions of the types of soil, rock, or conditions that require precautionary
measures to achieve satisfactory design and performance. Early recognition of these types ofsoil, rock, or conditions
is essential to allow sufficient time for adequate investigations and the development of designs. An excellent
overview of the various soils in Canada is provided by Legget (1965 and 1976).
5.2  Soils
5.2.1  Organic Soils, Peat and Muskeg
Soils containing significant amounts of organic materials, either as colloids or in fibrous form, are generally
weak and will deform excessively under load. Such soils include peat and organic silts and clays typical of many
estuarine, lacustrine, or fluvial environments. Such soils are usually not satisfactory as foundations for even very
light structures because of excessive settlements that can result from loading the soil.
Many parts of Canada, especially in northern regions, have muskeg deposits that pose many significant and
challenging geotechnical design and construction problems. The interested reader is referred to MacFarlane (1969)
and Radforth and Brawner (1977) for detailed information and discussion concerning this special site condition.
5.2.2  Normally Consolidated Clays
Organic clays soft to medium consistency, which have been consolidated only under the weight of existing
conditions, are found in many areas. Typical of these are the clays of the Windsor-Lake St. Gaff region and the
varved clays in the northern parts of Manitoba, Ontario, and Quebec. Imposition of additional load, such as a
building, will result in significant long-term settlement. The magnitude and approximate rate of such settlement can
be predicted from analyses based on carefully conducted consolidation tests on undisturbed samples. Such studies
should be made before any significant structure is founded on or above these clays, in order to determine whether
settlements will be acceptable, considering the characteristics and purpose of the structure.
Driving piles through normally consolidated plastic clays may cause heave or displacements of previously driven
piles or adjacent structures. The bottom of excavations made in such soils may heave, and adjoining areas of
structures may move or settle, unless the hazards are recognized and proper precautions taken to prevent such
Special precautions may be necessary in sampling and testing varved clays. Any analysis should take into account
the important differences in properties between the various layers in the clays.
Special Site Conditions 79
5.2.3 Sensitive Clays
Sensitive clays are defined as having a remolded strength of 25 % or less of the undisturbed strength. Some clays
are much more sensitive than this, and clays having a remolded to undisturbed strength ratio of I to 20, or even I
to 100, are known.
Typically, such clays have field water contents equal to or greater than their liquid limits, and such relations may
indicate their presence. Extensive deposits ofsensitive clays occur in some areas, for example, the Champlain clays
of the St. Lawrence and Ottawa River Valleys. Where such clays have been preconsolidated by partial desiccation,
or by the weight of materials subsequently eroded, foundations may be placed on the clays, provided that the
foundation load produces shearing stresses under the foundations that are well within the shear strength of the clay,
or else excessive settlement and possibly catastrophic failure will result. Disastrous flow slides have developed in
the Champlain clays in a number ofplaces, and the hazard must always be considered. Deep excavations in sensitive
clays are extremely hazardous, because of possible severe loss in shear strength, resulting from strains within the
soil mass beneath and adjacent to the excavation.
Determination of the physical properties necessary for evaluating the significance of sensitive clays to a proposed
structure requires taking and testing of both undisturbed and remolded samples of the clays, and thorough analysis
of the possible hazards involved. Because ofthe extreme sensitivity of such clays to even minor disturbances, taking
and testing undisturbed samples require sophisticated equipment and techniques, and should be attempted only by
competent personnel experienced in this type of work.
5.2.4 Swelling and Shrinking Clays
Swelling and shrinking clays are clays that expand or contract markedly upon changes in water content. Such clays
occur widely in the provinces of Alberta, Manitoba and Saskatchewan, and are usually associated with lacustrine
deposits. Shallow foundations constructed on such clays may be subject to movements brought about by volume
changes, because of changes of the water content in the clays Deep foundations supporting structural floors can be
damaged if the enclosing clay is .confined. Special design provisions should be made, which take into account the
possibility of movements or swelling pressures in the clays (see Chapter 15).
5.2.5 Loose, Granular Soils
All granular soils are subject to some compaction or densification when subjected to vibration. Normally this is
of significance only below the permanent water table. Sands above the water table, as a rule, will be only slightly
compacted by most building vibration, because of friction developed between the grains from capillary forces.
Usually for sands in a compact to dense state, settlements induced by vibration will be well within normal structural
tolerance, except for very heavy vibration, as from forging hammers or similar equipment (discussed in detail in
Chapter 14). However, ifthe sands are in a loose to very loose state, significant settlement may result from even minor
vibrations or from nearby pile driving. In some cases, earthquakes have brought about the liquefaction ofvery loose
sands, such as occurred in Niigata, Japan. In this event, structures supported above such soils may be completely
destroyed. Loose sands will settle significantly under static load only. Such settlements may exceed allowable
tolerance. Consequently, loose sands should be investigated carefully, and their limits established; densification or
compaction of such deposits may be essential before structures can safely be founded above or within them.
5.2.6 Metastable Soils
Metastable soils include several types of soil, abnormally loosely deposited, which may collapse on saturation.
Such collapses will cause severe or even catastrophic settlement of structures founded in or above these soils. Loess
is the most common metastable soil.
Because metastable soils are strong and stable when dry, they can be misleading in investigations and extreme care
should be taken to ensure identification and proper foundation design wherever such soils occur. The open, porous
80 Canadian Foundation Engineering Manual
structure, which is the usual means of identification, may be completely collapsed by the boring techniques. Where
such conditions may be anticipated, borings should be done by auger methods, and test pits should be dug, from
which undisturbed samples may be taken to determine accurate in-place densities.
5.2.7 Glacial Till
Till is unsorted and unstratified glacial drift deposited directly by and undemeath glaciers. Its soil grains are usually
angular and all size fractions are normally present (Legget, 1962 and 1979; Legget and Karrow, 1983). Basal till
(consolidated under the full weight of the glacier) is normally very dense, whereas ablation till (deposited from the
glacier during ablation) may not be dense. Till is generally a good foundation material, but problems have arisen
with the presence of soft layers and large boulders. Till may be to excavate. Fine-grained till is generally
susceptible to frost.
5.2.8 Fill
An engineered fill placed under careful control may be an extremely dense material, more uniform, more rigid, and
stronger than almost all natural deposits. When not placed under controlled conditions, it may be a heterogeneous
mass of rubbish, debris, and loose soil of many types useless as a foundation material. It may, of course, also be
some combination intermediate between these extremes.
Unless the conditions and quality control under which a fill was placed are fully known, the fill must be presumed
unsatisfactory for use under foundations. Investigations must establish its limits, depths, and characteristics
5.3 Rocks
5.3.1 Volcanic Rocks
Parts of the Canadian Cordillera and the Western Interior Plains have extensive deposits of geologically young
volcanic rocks. Some tuffs within these volcanic sequences have high porosities, low densities, and low shear
and compressive strengths. These materials weather rapidly, in some places, to smectites (swelling clay minerals;
5.3.2 Soluble Rocks
Rocks such as limestone, gypsum, rock salt, and marble are subject to high rates of solution by groundwater, and
may contain solution channels, caverns, and sinkholes, which may cave to the earth's surface. These conditions
present special foundation problems (Calembert 1973).
5.3.3 Shales
, -; Shales are the most abundant of sedimentary rocks and commonly the weakest from the standpoint of foundations.
Two special problems with certain shale formations have been identified in Canada.
In Western Canada, the Bearpaw Formation and other shales ofCretaceous age have been found to swell considerably
when stress release or unloading leads to the absorption ofwater by the clay minerals, in combination with exposure
to air. Bearpaw shales also have a low frictional resistance, which may create slope stability problems for both
excavations and construction on or near natural slopes in Bearpaw shales. Special advice should be sought if
Bearpaw or comparable shales are encountered along deep river valleys.
In Eastern Canada, volumetric expansion of some shale formations, caused by the weathering of iron sulphide
minerals (mainly pyrite), accelerated by oxidizing bacteria, has occurred in a few localities. Conditions leading to
mineralogical alteration seem to be related to lowering ofthe groundwater table and to raising ofthe temperature in
the shale, particularly when the shale is highly fractured, These conditions enhance bacterial growth and oxidation
Special Site Conditions 81
heat loss from the building spaces to the supporting shale. Shales often weather rapidly when exposed to air in
excavations. Specialmeasuresarewarrantedtoavoidprolongedcontactwithair.
Astheeffectofchemicaldegradationoffoundationrockontheperformanceofthe structuremaybecomeobvious
difficultiesatthetimeofsiteexplorationandthetakingof remedialmeasuresduringdesignandconstructionphases
5.4 Problem Conditions
5.4.1 Meander Loops and Cutoffs
Meandering streams from time to time develop chute cutoffs across meander bends, leaving disused, crescent-
shapedwater-filledchannels,calledoxbowlakes, whichlaterfillwithverysoft,organicsiltsandclays.Frequently,
thesecrescent-shapedfeatures canbedetectedinaerialphotographsorfrom accuratetopographicmaps.Thesoils
filling these abandonedwaterways canbe weakand highly compressible. It is necessary, therefore, to determine
theirlimitsandtoestablishthe depthsofthesoft,compressiblesoils.
5.4.2 Landslides
oldlandslidesorunstable soils in apotentiallandslidestate are moredifficulttodetect. Theymaybesignalledby
ofsensitive clays increases significantly the risk oflandslides. The stability ofsuch an area maybe so marginal
thatevenminordisturbances such as asmallexcavationnearthetoe ofaslope, orslightchanges in groundwater
conditionsordrainage,mayactivateaslide.It issimplertotakeprecautionsto avoidtriggeringalandslidethanto
stoponeinmotion,butitisbetterstillto avoidthelandslideorpotentiallandslideareaaltogether.
Thebanksof activelyerodingriversarealwaysinastateofmarginalstability.Thisisparticularlytrueof theoutside
bendsofsuchrivers,becauseactivecuttingisusuallyinprogress,especiallyduringperiodsof highwater.Ongoing
sloughingofaslopeisoftenanindicationofincipientfailure (EdenandJarrett, 1971).
•  proceduresanddesignsthatwillimprovethestability.Boththesteepnessandheightofslopesareimportantfactors
influencingthe stability. Steepeninganaturalslope, orexcavatingnearthetoe, orplacingfill atthetop ofslopes,
eithertemporarily orpermanently, will adversely affectthe stability ofthe slope andmay resultin slope failure.
Proper design analysis is required whenever such construction works are contemplated. In particular, the design
mustconsidertheaspectsof aseasonallyvaryinggroundwaterregime,aswellastheeffectoffreezingandthawing
require additionaldrainageplacedhorizontallyinthesidesoftheslopes.
5.4.3 Kettle Holes
During the deposition ofglacial outwash by the retreating continental ice sheets, large blocks ofice commonly
becamestrandedortrappedinthe outwash deposits. Upon melting, these blocks left depressions in the outwash
mantle, many ofwhichweresubsequentlyfilledwithpeatorwith softorganicsoils. Suchdepressions, knownas
kettle holes, range in diameter from a few meters to several hundred meters. Usually, the depths ofkettle holes
do not exceed 40 % oftheirminimum lateral dimensions; the depths are limited to the angles ofrepose ofthe
surroundingmaterials.Kettleholesarenormallyeasilyidentifiedas shallowsurfacedepressions.Insomelocalities,
82 Canadian Foundation Engineering Manual
5.4.4 Mined Areas
Sites above or adjacent to mined areas may be subject to severe ground movements and differential settlements,
resulting from subsidence or caving. For coal mines and other types of mines in horizontal strata, the zone of
disturbance generally does not extend laterally from the edge of the mined areas for a distance more than half the
depth of the mine below the surface. There is little control of the solution process that occurs in potash or salt mines,
and subsidence may extend several hundred meters beyond the edges of the mine or well field. Some evidence
indicates that the solution may extend farthest up the dip of the strata.
Investigations must be extremely thorough and all possible data on old mines should be obtained wherever such
differential settlement conditions are suspected. While good maps for active or recently closed mines may be
available, the accuracy and reliability of maps on plans for long abandoned mines are frequently poor. Furthermore,
there are many mined-out areas, especially in the older mining regions, for which no records are now available.
5.4.5 Permafrost
Permafrost is the thermal condition of the earth's crust and surficial deposits, occurring when temperature has been
below the freezing point continuously for a number of years. Half of Canada's land surface lies in the permafrost
region, either in the continuous zone where the ground is frozen to great depths, or in the discontinuous zone
where permafrost is thinner and there are areas of unfrozen ground (Brown 1970, Johnston 1981, AnderSland and
Anderson 1978).
The existence of permafrost causes problems for the development of the northern regions extending into the Arctic.
Engineering structures are, of course, greatly affected by the low temperatures. Ice layers and pore ice give soil a
rock-like structure with high strength. However, heat transmitted by buildings often causes the ice to melt, and the
resulting slurry is unable to support the structure. Many districts in northern Canada have examples of structural
damage caused by permafrost. In construction and maintenance of buildings, normal techniques must, therefore, be
modified at considerable additional cost. Expected changes in global climate are exacerbating these problems.
The accumulated experience from careful, scientifically planned and conducted investigations makes it technically
possible to build practically any structure in the permafrost area (Rowley et aI., 1975). Design and construction in
permafrost should be carried out only by those who possess special expertise.
5.4.6 Noxious or Explosive Gas
Noxious or explosive gases, of which methane is the most common, are occasionally encountered in clay or silt
deposits and in landfill sites containing decaying organic matter. They constitute a hazard to workers constructing
caissons or deep excavations. Gases may be found in shale or other sedimentary rock deposits in various areas of the
country. These may be a special hazard in deep excavations, or where borings have encountered such gases, which
have discharged into the construction area. The history of the local area of discharge of gas from borings, even if
only for short periods of time, should be especially noted and suitable precautions taken.
A particular problem may exist in tunnels or drainage systems where the oxidation of iron sulphides by bacteria can
deplete the free oxygen supply in poorly ventilated areas so much that persons entering may be asphyxiated. Such
areas should be thoroughly purged with clean air before anyone enters, and adequate ventilation must be assured
while people are present.
5.4.7 Effects of Heat or Cold
Soils should be protected against contact with surfaces that will be extremely hot or cold. Desiccation of clay soils
beneath furnaces or alongside ducts carrying hot gases will cause differential settlements Therefore, insulation and
ventilation is necessary around high-temperature structures.
Special Site Conditions 83
To prevent the potential collapse of retaining walls in the winter due to ice lens formation, the walls must be back-
filled with non frost-sensitive material for a distance equal to maximum frost penetration. The extent of the backfill
may be reduced by means of insulation behind the wall. Proper drainage must also be provided.
5.4.8 Soil Distortions
Soils distort both laterally and vertically under surface loadings. Lateral distortion is generally not significant,
but severe lateral distortions may develop in highly plastic soils toward the edge of surface loadings, even where
the loads are not sufficient to cause rupture or mud waves. These lateral distortions may affect foundations, or
structure-supporting piles, or pipe trenches located in or adjacent to areas subject to high-surface loading such as
along the edge of fills or a coal pile. Lateral distortions are a special hazard if sensitive clays are present. In such
soils, shearing strains accompanying the distortions may lead to significant loss of shear strength or possibly even
to flow failures or slides.
Both lateral and vertical displacements may develop when displacement-type piles are driven. Cohesive soils are
especially subject to such displacement. Previously driven piles or existing foundations may be displaced, or the soil
movements may result in excessive pressures on retaining walls, on sheeting for excavations, or on buried pipes.
Heaved piles may be redriven and used. If there is significant lateral displacements the piles may be kinked or
bowed beyond the safe limit ofuse. These hazards must be evaluated in the investigation program. Provision should
be made in design and construction procedures to ensure that other structures or piles are not damaged or displaced
by the driving ofadjacent piles. Preboring through the cohesive strata should be required ifthere is risk of disturbing
existing structures or previously driven piles.
5.4.9 Sulphate Soils and Groundwater
Sulphates in the soil and groundwater can cause significant deterioration of Portland cement concrete. Because
contact ofconcrete with sulphates invariably is due to sulphate solution in the groundwater, isolation ofthe concrete
by interception or removal ofsulphate-laden waters will prevent deterioration ofthe concrete. An alternative so lution
is to use sulphate-resistant cement in the concrete.
The presence of sulphates in the groundwater does not automatically justify the use of sulphate-resistant cement.
High-quality watertight concrete is less susceptible to deterioration by sulphates than lower quality concrete.
Furthermore, the use of sulphate-resistant cement does not necessarily make the concrete sulphate-proof.
  . ~ .
84  Canadian  Foundation  Engineering Manual 
Earthquake  - Resistant Design 
6 Earthquake - Resistant Design
6.1 Introduction
Earthquake shaking is an important source of extemalload that must be considered in the deSIgn of civil engineering 
structures because of its  potential for disastrous  consequences. The  degree of importance of earthquake loading at 
any given site is  related to a number of factors  including: 
•  the  composition the probable intensity and likelihood of occurrence of an earthquake; 
the  magnitude  of the  forces  transmitted  to  the  structures  as  a  result  of the  earthquake  ground  motions 
(displacement, velocity and acceleration); 
•  the amplitude, duration and frequency content of strong ground motion;  and 
•  and behaviour of the subsoils. 
Hazards associated with earthquakes include ground shaking, structural hazards,  liquefaction, landslides, retaining 
structure  failures,  and  lifeline  hazards.  The  practice  of earthquake  engineering  involves  the  identification  and 
mitigation  of these  hazards.  With  the  advancement  of our knowledge  regarding  earthquake  phenomena  and  the 
development of better earthquake-resistant design procedures for  different structures, it is  possible to  mitigate the 
effects Qf strong earthquakes and to reduce loss of life,  injuries and damage. However, it is extremely difficult, and 
in many cases impossible, to produce an earthquake-proof structure. Depending on the type ofstructure and its use, 
the foundation conditions, and the  costs involved, a structure can, generally, only be  designed to be more resistant 
(not immune) to  an earthquake. 
Many important developments in the field ofearthquake engineering have occurred in the last four decades. Advanced 
structural seismic  analysis methods, comprehensive experimental procedures for the  assessment and evaluation of 
the behaviour of different types  of soil,  and considerable  data on  the performance  of different structures  and  soil 
profiles  during earthquakes are available to help designers in producing earthquake-resistant designs. Geotechnical 
earthquake engineers have to  address a number of issues when designing safe structures in a seismic environment. 
They have to establish design ground motions, assess the seismic capacity and performance offoundations, consider 
the  interaction  effects  between  structures  and  the  supporting  ground,  and  evaluate  the  effects  of the  earthquake 
excitation on the strength parameters of the soil.  Each of these issues represents a category of problems that varies 
according to the type of structure under consideration. 
The purpose of this chapter is to present some of the key concepts and procedures used by geotechnical earthquake 
engineers  to  design  safer  structures  in  a  seismic  environment.  References  that  give  detailed  accounts  of the 
procedures will be provided as  needed. However, situations that involve a high risk of seismic hazards, and bridges, 
tall buildings or dams resting on soft foundation soils, generally require detailed dynamic analysis by engineers very 
knowledgeable in  earthquake engineering.  Some  of the seismological concepts and terminology will be given first 
to  enable the  geotechnical engineer to understand the basis  of both earthquake characterization and seismic design 
Earthquake - Resistant Design 85
6.2 Earthquake Size
The  size  of an  earthquake  can  be  described based on its  effects  (Earthquake  Intensity);  the  amplitude  of seismic 
waves (Earthquake Magnitude); or its  total released seismic energy (Earthquake Energy). 
6.2.1 Earthquake Intensity
Earthquake  intensity  is  the  oldest  measure  and  uses  a  qualitative  description  of the  earthquake  effects  based  on 
observed damage and human reactions. Different scales of intensity include the Rossi-F orel scale (RF); the Modified 
Mercalli Intensity scale (MMI)  that represents conditions in California; the Japanese Meteorological Agency scale 
(JMA) used in Japan;  and the Medvedev-Sponheuer-Karnik scale (MSK) used in Central and Eastern Europe. 
6.2.2 Earthquake Magnitude
Most scales  of earthquake  magnitude  are  based on  some  measured quantity  of ground  shaking and  are  generally 
empirical.  Most  of these  magnitude  scales  are  less  sensitive  in  representing  stronger earthquakes  (referred  to  as 
Richter Local Magnitude (Richter  1958): Defines a magnitude scale for shallow, local (epicentral distance less than 
600 km) earthquakes in  southern California. 
ML = log A  (6.1) 
A  the maximum trace amplitude  (in  microns) recorded on a Wood-Anderson seismometer located 100 km 
from the epicentre of the earthquake. 
Surface Wave Magnitude: A worldwide magnitude scale based on the amplitude ofRayleigh waves with a period of 
about 20 s. It is used to describe the size of shallow (focal depth < 70 km), distant (epicentral distance>  1000 km) 
or moderate to large earthquakes.  It is given by 
Ms = log A +  1.66 log i1 +2.0  (6.2) 
where  epicentral  distance 
A = maximum ground displacement (microns) and i1  = h'  £  x 360°. 
eart  Clfcum  erence 
Body Wave Magnitude: A worldwide magnitude scale based on the amplitude of the first few cycles of p-waves. It
is  used for deep focus  earthquakes  and is  given by 
Mb  logA-logT+0.Oli1+5.9  (6.3) 
A p-wave amplitude in microns, T p-wave period (about 1 s), and 
epicentral  distance  x  3600. 
earth circumference 
Moment Magnitude Mw:  This is the only magnitude scale that is not subject to saturation because it does not depend 
on ground shaking-levels. It is based on the seismic moment and is given by 
M = log Mo  -10.7
w  1.5 
in which 
Mo = the seismic moment in dyne-cm =~   r D ,
86 Canadian Foundation Engineering Manual
~   = the mpture strength of the material along the fault, Ar = the rupture area and D the average amount of
These quantities can be estimated from geologic records for historical earthquakes or from the long-period
components of a seismogram (Bullen and Bolt 1985).
6.2.3 Earthquake Energy
The total seismic energy released during an earthquake is estimated by
log E = 11.8 + 1.5 Ms ( 6.5)
E is expressed in ergs. This relationship is also applicable to moment magnitude.
6.3 Earthquake Statistics and Probability of Occurrence
The rate of OCCUlTence of an earthquake with a magnitude equal to or greater than M for a given area and time may
be estimated by (Gutenberg and Richter 1944)
logloN(M) = a bM (6.6)
N(M) is the number of earthquakes ~ M (commonly per year) and a and b are constants for a given seismic
zone and are established by fitting the available earthquake data. Fitting Equation 6.6 to incomplete data may
indicate, incolTectly, higher OCCUlTence rates for larger earthquakes. It is also worth noting that Equation 6.6
does not always hold.
The probability of OCCUlTence of at least one earthquake with a magnitude ::::: M in a given time can be
calculated by
P = 1 - e -NI
N is the rate of OCCUlTence per year and t is the time period in years under consideration.
The seismic loads used in the National Building Code of Canada (NBCC 2005) are based on a 2 per cent probability
of exceedance over 50 years (a 2475-year earthquake). This means that over a 50-year period there is a 2 per cent
chance that the ground motions given in the NBCC (2005) will be exceeded.
6.4 Earthquake Ground Motions
The ground motions produced by earthquakes at a particular site are influenced by many factors and can be quite
complicated. They are a function of the distance from the earthquake's causative fault, and the depth, mechanism
and duration of the fault mpture causing the earthquake as well as the characteristics of the soil profile at the site.
In practice, three translational components, the vertical and two perpendicular horizontal directions of ground
motion are recorded. The significant characteristics of the ground motion (known as ground motion parameters) for
engineering purposes are: the amplitude; frequency content; and duration of the motion.
To evaluate the ground motion parameters, measurements of ground motions in actual earthquakes are required.
Instmments used to accomplish these measurements are seismographs that produce seismograms (velocity response)
and accelerographs that produce accelerograms (acceleration response).
Earthquake - Resistant Design 67
6.4.1 Amplitude Parameters
The ground motion is commonly described with a time history of the acceleration, velocity or displacement. The
amplitude is generally characterized by the peak value of acceleration (measured). Peak values of velocity and
displacement can be calculated by integrating the acceleration time history. Alternatively, when using the response
spectrum approach, the peak values of velocity and displacement can be computed approximately by
a(ro)=rov(ro)=ro 2U(ro) (6.8)
U, v and a are the transfOlmed displacement, velocity and acceleration obtained by subjecting the measured
acceleration time history to a Fourier transfonn, and co is the predominant circular frequency of the
ealthquake. Peak Acceleration
The peak horizontal acceleration (PHA) is obtained as the maximum resultant due to the vector sum oftwo olthogonal
components. It is unlikely that the maximum acceleration in two orthogonal components occur simultaneously,
however, and the PHA is taken in practice as the maximum measured horizontal acceleration. Horizontal
accelerations are used to describe ground motions and their dynamic forces induced in stiff structures. The peak
vertical acceleration (PVA) is less important for engineering purposes and can be taken to be approximately as two
thirds of PHA. Ground motions with high peak accelerations and long duration are usually destructive. Peak Velocity
The peak horizontal velocity (PHV) better characterizes the ground motions at intennediate periods, 0.4 s > T >
0.2 s. For flexible structures, the PHV may provide a more accurate indication of the potential for damage during
earthquakes in the intennediate period range. Peak Displacement
Peak displacements are associated with the lower frequency components ofthe ground motion. They are difficult to
detennine accurately and, as a result, are less commonly used as a measure of ground motion. Seismic Regions of Canada
Ground motion probability values are given in tenns of probability of exceedance, that is the likelihood of a given
horizontal acceleration or velocity on finn soil sites, being exceeded during a particular time period. The 2005
National Building Code of Canada (NBCC 2005) presents the seismic hazard for Canada in tenns of a probabilistic
based unifonn hazard spectrum, replacing the probabilistic estimates of peak ground velocity (PGV) and peak
ground acceleration (PGA) In the earlier codes. Spectral acceleration at 0.2, 0.5, 1.0 and 2.0 second periods and
peak acceleration fonn the basis of the seismic provisions ofNBCC (2005),
Eastern and western Canada are treated slightly differently because of the different properties of the crust in these
regions, Figure 6.1 shows the earthquakes and the regionalization used and identifies in a general way the low-
seismicity central part ofCanada defined as "stable Canada." The different physical properties of the crust in eastern
and western Canada and the different nature of the earthquake sources in south-western Canada required the use of
four separate strong ground motion relations as detailed by Adams and Halchuk (2004). Seismic hazard to the west
of the leftmost dashed line on Figure 1 has been calculated using western strong ground motion relations; eastern
relations are used for the remaining regions.
88 Canadian Foundation Engineering Manual

. 2.7·3.9 It 5.0 - 6.4
.. 4.0 . 4.9 •   6.5
FIGURE 6.1 Map ofCanada (showing the earthquake catalogue usedfor the 4th Generation model together
with dashed lines delimiting the eastern and western seismic regions and the "stable Canada" central region.)
The spectral acceleration parameters are denoted by Sa(T), where T is the period and are defined later in Tables
6.1B and C (Section for different soil conditions. The PGA values are also presented for use in liquefaction
analyses. The NBCC (2005) explicitly considers ground motions from the potential Cascadia subduction earthquake
located off the west coast of Vancouver Island. While the amplitudes of such an earthquake are expected to be
smaller than from local crustal earthquakes, the duration of shaking will be greater which has implications for
liquefaction assessment.
. Seismic hazard values were. calculated for a grid extending over Canada and used to create national contour maps
such as Figure 6.2. Figure 6.3 shows the Uniform Hazard Spectra (UHS) for a few major cities to illustrate the range
and period dependence of seismic hazard across Canada.
12 12 20 40 60 80 100 120 %g
FIGURE 6.2 Sa(O.2) for Canada (median values of5 % damped spectral acceleration
for Site Class C and a probability of2 %/50 years)
Earthquake. Resistant Design 89
0.01  H--------i------t---
0.1  0.2  0.5 
Period (seconds)
0.005 '-'-----"'-------"'------------' 
1  2 
FIGURE 6.3  Uniform Hazard Spectra for median 2 %/50 year ground motions on Site Class Cfor key cities
6.4.2 Frequency Content
The dynamic response of structures is very sensitive to the frequency content of the loading. Earthquake excitations 
typically contain a broad range of frequencies.  The frequency  content  describes  the  distribution of ground motion 
amplitudes  with respect  to  frequency,  which can be  represented by a  Fourier Amplitude  Spectrum (i.e.,  a plot of 
Fourier amplitude versus frequency)  or a Response Spectrum. The predominant circular frequency,  (0,  in Equation 
6.8  is  defined as  the frequency  corresponding to the maximum value of the Fourier amplitude spectrum. The value 
of (0 can be approximated by the number of zero crossings per second in the accelerogram mUltiplied by 21(. 
6.4.3 Duration
The duration  of shaking  significantly  influences  the  damage  caused  by  an  earthquake.  The  liquefaction of loose 
saturated sand depends on the number of stress reversals that take place during an earthquake. Earthquakes oflonger 
duration are most likely to  cause more damage. 
The duration is  evaluated from the accelerogram. Different methods are specified to  evaluate the duration of strong 
motion  in  an  accelerogram.  The  duration  can be  defined  as  the  time  between  the  first  and last exceedances  of a 
threshold acceleration (usually 0.05 g), or as the time interval between the points at which 5 % and 95  % ofthe total 
energy has been recorded. 
6.5 Building Design
It is almost impossible to design buildings that remain elastic for all levels of earthquakes.  Therefore, the intention 
of building  codes  and provisions  is  not  to  eliminate  earthquake  damage  completely.  Rather,  structures  should be 
designed to resist: 
1.  a moderate level earthquake, which has a high probability ofoccurring at least once during the expected life 
of the structure, without structural  damage, but possibly with some non-structural damage; and 
2.  a  major  level earthquake,  which has  a  low  probability  of occurrence, without  collapse, but possibly with 
some structural damage. 
90 Canadian  Foundation Engineering Manual 
In  general,  there are  two  procedures  to  the  earthquake-resistant design of buildings:  a  static  analysis procedure  in 
which  the  earthquake  loading  is  characterized by equivalent  static  forces  and  dynamic  analysis  procedures.  The 
dynamic analysis procedures  include linear analysis using either the Modal Response Spectrum Method where the 
earthquake  loading  is  characterized  by design response  spectra or the  linear time-history  analysis,  and  nonlinear 
time-history analysis. 
6.5.1 Equivalent Static Force Procedure
The static approach specified in the NBCC (2005) is used for structures satisfying the conditions of sentence 
ofthe code (e.g., regular building with a height  than 60 m and natural lateral period less than 2 s).  The procedure 
involves calculating a design seismic base shear proportional to the weight of the  structure.  The equivalent lateral 
seismic  force  procedure  of the  NBCC  (2005)  specifies  that  a  structure  should  be  designed  to  resist  a  minimum 
seismic base shear, V,  given by 
except that V shall not be less than, V =  S(2.0)M)E  W I(RdRo)
T.  is  fundamental  period  of the  structure,  SeT)  the  design  spectral  acceleration,  expressed  as  a  ratio  to 
gravitational acceleration,  for  a period of T,  Mv  Factor to  account for  higher mode effect on base shear,  as 
defined in NBCC  Sentence,  I = Earthquake  importance  factor  of the  structure,  as  described  in 

NBCC Sentence,  W= weight of the structure,  Ductility related force  modification factor and 
R  =  Overstrength related force modification factor. DeSign Spectral Acceleration, S (T)
The design spectral acceleration values ofS(T) is determined as follows (linear interpolation is used for intermediate 
values ofT): 
SeT)  =  F.S.(O.2) forT::: 0.2  s  (6.10) 
F  S  (0.5) or F S (0.2) whichever is  smaller for T  0.5 
v  a  a  a 
FvS.(l.O) forT =  1.0 s 
F  S  (2.0) for T =2.0 s 
v  • 
FvS .(2.0)/2 for T 2: 4.0 s 
S.(T) =  the 5 % damped spectral response acceleration values for the reference ground conditions (Site Class 
C in NBCC Table 4. 1.8.4.A), and F. and Fv are acceleration and velocity based site coefficients given inNBCC 
Tables and using linear interpolation for intermediate values of S.(0.2) and S.(1.0). Foundation Effect
The soil conditions at a site have been shown to exert a major influence on the type and amount of structural damage 
that can  result  from  an  earthquake.  As  the  motions  propagate  from  bedrock  to  the  surface,  the  soil  layers  may 
amplify the motions in selected frequency  ranges around their natural frequencies.  In addition, a structure founded 
on  soil,  with  natural  frequencies  close  to  thos.e  of the  soil  layers,  may undergo  even  more intense  shaking due  to 
the  development of a  state  of quasi-resonance between  the  structure  and the  foundation  soil.  The  natural  circular 
frequency of a soil  layer in horizontal direction, ro
'  is given by 
(6.l1 ) 
(i) =--
u  2h

is the shear wave velocity  of the soil  layer and h  is  its thickness. 
Earthquake. Resistant Design 91
Direct  calculation  of the  local  site  effects  is  possible  using  suitable  mathematical  models  such  as  lumped  mass 
approaches  and  finite  element  models  with  realistic  soil  properties  and  assuming  vertically  propagating  shear 
waves  or Rayleigh waves from  the  bedrock during the earthquake.  In these analyses, the source mechanism of the 
earthquake and the geology of the travel path are incorporated in the bedrock input motion. 
The seismic provisions ofthe NBCC (2005) incorporate site effects by categorizing the wide variety of possible soil 
conditions  into seven types  classified  according to  the average  properties  of the  top  30  m  of the  soil profile.  This 
classification  is  based  on the  average  shear  wave  velocity,  V
'  standard  penetration  resistance,  N
,  or undrained 
shear strength, su'  as  shown in Table 6.1A. The factors Fa  and F  v given in Tables 6.1B  and 6.1 C reflect the effect of 
possible  soil amplification (or de-amplification)  and soil-structure interaction resonance into  the  estimation of the 
seismic design forces  for buildings having no unusual characteristics. 
While  the  site  coefficients  Fa  and  Fv  provide  a  simple  way  of introducing  surface  layer  effects  for  conventional 
buildings,  a  fuller  evaluation of amplification should be  completed for  areas  of significant seismic activity and/or 
non-conventional buildings. 
Quasi-resonance conditions are of particular impoFtance when the predominant period ofthe input rock motion (or 
firm  ground)  is  close to the fundamental period of the less-firm surface layers since this results in amplifications of 
two to five.  In this case, the firm ground or underlying rock accelerations must be modified for potential amplification 
by less-firm surface layers. The site coefficients are fairly  realistic except for  this  case. 
TABLE 6.1A Site Classification/or Seismic Site Response
(Table 4.1.B.4.A. in NBCC 2005)
Not applicable  Not applicable  B  Rock  760 < v's 1500 
Very Dense Soil 

and  Soft Rock 
Stiff Soil 
Soft Soil 
360 <  V, < 760  N60>  50  su>  100kPa 

<360  15::::  N60  ::::  50  50 < su  s100kPa 
V, <180  N60 < 15  s  < 50kPa 
Any profile with more than 3 m of soil with the following characteristics: 
Plastic index Ip  > 20 
Moisture content w 2: 40%,  and 
Undrained shear strength  < 25  kPa 
Site Specitic Evaluation Required 
Note (I)  Other soils include: 
a)  Liquefiable  soils,  quick  and  highly  sensitive  clays,  collapsible  weakly  cemented  soils,  and  other  soils 
susceptible to failure or collapse under seismic loading. 
b)  Peat and/or highly organic clays  greater than 3 m in thickness. 
c)  Highly plastic clays (Ip> 75) with thickness greater than 8 m. 
d)  Soft to medium stiff clays with thickness greater than 30 m. 
92 Canadian Foundation Engineering Manual
TABLE 6.1 B Values ofFa as a Function ofSite Class and S/0.2)
(Table 4.1.B.4.B in NBCC 20OS)
B 0.8
c 1.0
1.0 1.0
1.2 1.1 1.1
1.4 1.1 0.9
(2) (2) (2)
TABLE 6.1 C Values ofF" as a Function ofSite Class and S/1. 0)
(Table 4.1.B.4.C in NBCC 2005)
c 1.0
D 1.4
E 2.1
1.0 1.0 1.0
1.3 1.2 1.1
2.0, 1.9 1.7
(2) (2) (2)
Note (2) F and F for site Class F are determined by performing site specific geotechnical investigations and dynamic
a v
site response analyses.
The seismic design procedures outlined in the NBCC (2005) are based on the assumption that the structures are
founded on a rigid base that moves with the ground surface motion. Real foundations possess both flexibility and
damping capacity that alter the structural response. The flexibility of the foundation increases the fundamental
period of a structure and the damping dissipates energy by wave radiation away from the structure and by hysteretic
damping in the foundation, thus increasing the effective damping of the structure. These effects are referred to as
soil-structure interaction and are not considered explicitly in the code. For most buildings considered by the code,
neglecting soil-structure interaction results in conservative designs. However, neglecting soil-structure interaction
effects may not be conservative for tall structures and/or structures with substantial embedded parts and should be
considered explicitly in a dynamic analysis. Importance Factor, IE
Some structures are designed for essential public services. It is desirable that these structures remain operational
after an earthquake (defined as post disaster in the code). They include buildings that house electrical generating and
distribution systems, fire and police stations, hospitals, radio stations and towers, telephone exchanges, water and
sewage pumping stations, fuel supplies and schools. Such structures are assigned an IE value of 1.5. The importance
factor I 1.3 is associated with special purpose structures where failure could endanger the lives of a large number
of people or affect the environment well beyond the confines of the bUilding. These would include facilities for the
Earthquake" Resistant Design  93 
manufacture  or storage of toxic material, nuclear power stations, etc.  Force Reduction Factors, Rd  and System Overstrength Factors, Ro 
The values of Rd  and Ro and the corresponding system restrictions shall conform to NBCC Table (Table 6.2). 
When  a  particular value of Rd  is  required,  the  associated Ro  shall be used.  For combinations  of different types  of 
SFRS acting in the same direction in the same storey, RdRo  shall be taken as the lowest value of RdRo  corresponding 
to  these systems. 
TABLE 6.2 SFRS Force Modification Factors (R), System Overstrength Factors (R)
and General Restrictions (1)
(Table  in NBCC) 
Forming Part of Sentence  (1) 
Steel Structures Designed and Detailed According to CSA S16 

Ductile moment resisting frames 
····················"·--··"P.··.·, .. ···_-, .. ,·_- .. ,, ... -.... ,  ...... 
Moderately  ductile  moment  resisting 
..  ····1 
ductility  moment  resisting 
2.0  1.3  NL  NL  60  NP  NP 
Moderately  ductile  concentrically 
braced frames 
•  Non-chevron braces 
•  Chevron braces 
•  Tension only braces 
.. """, ....
I·····..  ·· 
., .............. , .."., 
Limited  ductility  concentrically braced 
•  Non-chevron braces 
•  Chevron braces 
•  Tension only braces 
Ductile eccentrically braced frames  4.0  1.5  NL 
.......  ...  ....................  ···  .. · .... ···1···  ....  ···· 
Ductile frame plate shearwalls  5.0  1.6  NL 
..................  . ....................... .. 
Moderately ductile plate shearwalls  2.0  1.5  NL 
... '  ... ,,",", ....  "  .....• 1.... , .•."" .... ""... , ......... H ..... '  .. .. 
Conventional  construction  of moment 
frames,  braced frames  or shearwalls  1.5 
1.3  NL  NL  15  15  15 
Other steel SFRS(s) not    "'1.0  I 15+"15"\ NP  I NP  NP 
Concrete Structures Designed and Detailed According to CSA A23.3 
Ductile moment resisting frames  4.0  1.7  NL  NL  NL  NL  NL • 
Moderately ductile moment resisting
40 60  40 NL NL 2.5  1.4
NL NL NL  NL NL 4.0  1.7 Ductile coupled walls  • 
"  .. , .. - ....,  ........... ,......................... ,  ...  , 
_,L: ..
94 Canadian Foundation Engineering Manual
Restrictions (2)
NL  NL Ductile shearwalls 
...... " ........... "".. ..
3.5  .! 1.6...."  ..• ""...."".......-+.. 
Mo{lerate!lv ductile shearwalls  NL  60 
•  • •  Moment resisting frames  15  NP 
.  •  Shearwalls  40  30 
""""............ " ..... " ...
Other  concrete 
Timber Structures Designed and Detailed According to CSA 086
•  Nailed shearwalls-wood based panel  3.0
Cases Where
•  Shearwalls  wbod based and gypsum 
panels in combination 
Braced or moment resisting frame with 
ductile connections 
•  Moderately ductile 
•  Limited ductility 
Other wood or  gypsum  based SFRS(  s) 
Not listed above 
NL  30  20 
NL  20  20 
Masonry Structures Designed and Detailed According to CSA 5304.1

Moderately ductile shearwalls
u  ........................ ,.••• "" •••••• ."  ................ _._.  , .. ", ••••••• " •••• .,.,... 
Limited ductility shear walls 
Conventional Construction 
•  Shearwalls 
•  Moment resisting frames 
Unreinforced masonry 
Other  masonry  SFRS(s)  not  listed 
..... " ..... , ....
60  40 
• ••••••.•••.• "."."n... 
40  30 
........... , .............
.................... ,.
Notes to Table 6.2: 
(1)  See NBCC Sentence 
(2)  Notes on restrictions: 
NP in table means  not permitted. 
Numbers in table are maximum height limits in metres. 
NL in table means system is permitted and not limited in height as an SFRS. Height may be limited elsewhere 
in other Parts. Higher Mode Factor Mv and Base Overturning Reduction Factor J
The seismic lateral force acting on a building during an earthquake is due  to the inertial forces  acting on the masses 
of the  structures  caused by the  seismic  motion  of the  base.  The motion  of the structure  is  complex,  involving  the 
Earthquake. Resistant Design  95 
superposition  of a  number  of modes  of vibration  about  several  axes.  Table  of the  NBCC  (2005)  (Table 
6.3)  assigns.Mv  and  J values  to  different  types  of stmctural  systems,  which  are  established  based  on  design  and 
conshuction  experience,  and the  perfonnance  evaluation  of stmctures  in  major  and  moderate  ealihquakes.  These 
values account for the  capacity of the  stmctural  system to  absorb energy by damping and inelastic action through 
several cycles of load reversaL 
TABLE 6.3  Higher Mode Factor M"and Base Overturning Reduction Factor JIi.J1
(Table  in NBCC) 
FOlming Part of Sentence 
Type of Lateral 
ReSisting Systems 
"  fr 1
I  Mornent reslstll1g  ames  or 
"coupled walls" (3) 
<  8.0  Braced frames 
Walls, wall-frame systems, 
other  (4) 
Moment resisting frames  or 
"coupled walls" (3) 
J ForT  S 
1.0  1.0  1.0 
1.0  LO  1.0 
l.0  1.2  1.0 
1.0  1.2  1.0 
::::  8.0  !  Braced frames  1.0  1.5  1.0  0.5 
Walls, wall-frame systems, 
1.0  2.5  1.0  0.4
other systerns(4) 
(I)  For  values, of Mv  between  periods  of, 1.0  and  2.0  s,  the  product  S(TJM shall  be  obtained  by  linear 

(2)  Values of J between periods of 0.5 and 2.0 s shall be obtained by linear interpolation. 
(3)  Coupled wall  is  a wall  system with coupling  beams  where  at  least 66  %  of the  base overturning moment 
resisted  by the wall system is  carried by the  axial  tension  and  compression  forces  resulting from  shear in 
the  coupling beams. 
(4)  For hybrid systems, use values corresponding to walls or carry out a dynamic analysis. Distribution of Base Shear 
The base shear is the  sum of the inertial forces  acting on the  masses of the  stmctures caused by the seismic motion 
of the base. The motion of the shucture is complex, involving the superposition ofa number of modes of vibration 
about several axes. 
For  shuctures  with fundamental  periods  less  than  0.7  s,  the  addition  of the  spectral-modal  responses  results  in  a 
lateral inertial  force  distribution that is  approximately triangular  in  shape,  with the  apex  at the  base.  For buildings 
having longer periods, higher forces are induced at the upper portion of the stmcture due to increasing contributions 
to top storey amplitudes by all the contributing modes. The redistribution offorces is accounted for by applying pali 
of the base shear as  a concentrated force,  Ft'  to  the top  of the structure. 
The total  lateral  seismic force,  V, shall  be  distributed  such that a portion,  , shall be assumed to  be  concentrated 
at the top  of the building,  where F{ is  equal to  0.07    but need not exceed 0.25  V and may be considered as  zero 
where  T
does not exceed 0.7 s;  the remainder,  V- F
shall be distributed along the height of the building, including 
the top level,  in accordance with the formula. 
96 Canadian Foundation Engineering Manual
F, (V )W,h, I(Ewjhj)
is the inertial force induced at any level x which is proportional to the weight W, at that level. Overturning Moments
The lateral forces that are induced in a structure by earthquakes give rise to moments that are the product of the
induced lateral forces times the distance to the storey level under consideration. They have to be resisted by axial
forces and moments in the vertical load-carrying members. While the base shear contributions ofmodes higher than
the fundamental mode can be significant, the corresponding modal overturning moments for the higher modes are
smalL As the equivalent static lateral base shear in the NBCC (2005) also includes the contributions from higher
modes for moderately tall and tall structures, a reduction in the overturning moments computed from these lateral
forces appears justified. This is achieved by means ofthe multiplierJ as given in NBCC (2005) Table (Table
6.3). If, however, the response of the structure is dominated by its fundamental mode, the overturning moment
should be calculated without any J-factor reductions. Alternatively, a dynamic analysis should be used to calculate
the maximum overturning moment. Torsional Moments
The inertial forces induced in the structure by earthquake ground motions act through the centre of gravity of the
masses. If the centre of mass and the centre of rigidity do not coincide because of asymmetrical arrangement of
structural elements or uneven mass distributions, torsional moments will arise. The design should endeavour to
make the structural system as symmetrical as possible and should consider the effect of torsion on the behaviour of
the structural elements.
6.5.2 Dynamic Analysis
. For critical buildings and buildings with significant irregularities, the dynamic analysis approach is recommended
to improve the accuracy of calculation of the seismic response including the distribution of forces in the building.
The dynamic analysis approach includes response spectrum methods and time domain response methods. Response Spectra
The response spectrum describes the maximum response of a Single Degree Of Freedom System (SDOF) to a
particular input motion and is a function of the natural frequency and damping ratio of the SDOF system, and the
frequency content and amplitude of the input motion. The response may be expressed in terms of acceleration,
velocity or displacement.
The maximum values of acceleration, velocity and displacement are referred to as the spectral acceleration, S ,
spectral velocity, Sv' and spectral displacement, S d' respectively. They can be related to each other as follows:
(6. 13a)
Sv=lul ~ f   O S d
(6. 13c)
(fJo is the natural circular frequency of the SDOF system. These response spectra provide a meaningful
characterization of earthquake ground motion and can be related to structural response quantities,
Earthquake   Resistant Design  97 
E is the maximum earthquake elastic energy stored in the structure,  V is  the structure elastic base shear 
ma;:  mar 
and m is  the mass  of the  structure.  The base  shear,  however,  would be less  than  that calculated by Equation 
6.14b  for  structures  that experience  inelastic  material  behaviour (e.g.,  cracking  of concrete  and yielding  of 
steel)  during  an earthquake.  However, this reduction is  only  allowed for structures that have the  capacity to 
deform beyond the yield point without major structural failure  (ductile structures). 
Most structures are not SDOF systems and higher modes may contribute to the response. This effect can be accounted 
for approximately using the higher mode factors  given in Table 6.3. Design Spectra 
Spectral shapes from real records are usually smoothed to produce smooth spectral shapes suitable for use in design. 
Most of the  design  spectra commonly used are  based  on  the  Newmark-Hall approach. As  an example, Figure  6.4 
shows the  design  spectra (normalized to  the  maximum ground acceleration)  developed  by  Seed and Idriss  (1982) 
and recommended for use in building codes. The National Building Code ofCanada (NBCC 2005) includes similar 
smooth  design  spectra (which 'are  not based on Newmark-Hall  approach,  but  obtained  directly from  probabilistic 
seismic  hazard  assessments  based  on  spectral  amplitudes,  i.e.,  uniform  hazard  spectrum).  Acceleration  levels  of 
probable earthquakes can be used to scale the spectral shapes to  provide design spectra of particular projects. 
sos = FaSs 
sD1  =F


= 0.2 Ts 
T  SD1 
s  =--


FIGURE 6.4  Example ofa design spectra Site Specific Response Spectra 
Site-specific response spectra are developed with due  consideration of the following  aspects: 
I,  Seismotectonic characterization:  includes evaluation of seismic source, wave attenuation from the  sources 
to  the site and site evaluation. 
2.  Assessment ofseismic exposure: involves probabilistic analysis of data from possible significant earthquake 
98  Canadian  Foundation  Engineering Manual 
sources including nearby, mid-field and far-field events to establish the events with the most likely significant 
contribution to ground motions at the site. 
3.  Ground  motion  characterization:  encompasses  selection  and  scaling  of rock  input  motion  records  from 
earthquakes with magnitude, epicentral distance, types offaulting and site conditions similar to those of the 
design events for the specific site. The site specific motions are then determined from the rock input motion 
using ground response analysis  (e.g., Schnabel et aL,  1972); and 
4.  Design ground motion specification: includes the specification ofsmooth response spectra, and the selection 
of sets of representative  ground motion time histories suitable  for use in dynamic analysis of the structural 
For critical  structures,  spectra  are  usually  developed  for  two  different  levels  of motions,  namely  operating  level 
events and major level events. Operating level events are moderate earthquakes with a high probability ofoccurrence. 
Structures are designed to  survive these  events without significant damage and to  continue to  operate.  Major level 
events  are  severe  earthquakes  with  a  low probability  of occurrence,  and  significant damage,  but not collapse,  is 
therefore  acceptable.  Furthermore,  the seismic design of critical structures usually involves  dynamic time-history 
analyses using a number of ground motion records representative of operating level  and design level events. Soil-Structure Interaction Effects
The NBCC considers buildings sitting on firm ground (360 mls < V, < 760 mls. However; in most cases, buildings 
are constructed with flexible foundations embedded in soil layers. The soil-structure interaction (SSI) influences the 
seismic response of structures and should be investigated for cases involving  critical or unconventional structures. 
The soil-structure interaction modifies the dynamic characteristics of the structure: 
1.  It reduces the natural frequency ofthe soil-structure system to a value lower than that ofthe structure under 
fixed-base conditions (structures found on rock are considered to be fixed-base). 
2.  It increases  the  effective  damping  ratio  to  a  value  greater  than  that  of the  structure  itself.  SSI  also  has 
some important effects on the ground input motion and the seismic response of the structure.  For example, 
large  foundation  slabs  can reduce  the  high  frequency  motions  and hence reduce the  input motions  to  the 
structure,  and uplift  of foundation  slabs can  reduce  forces  transmitted  to  the  structure.  Furthermore,  SSI 
reduces  the  maximum  structural  distortion  and  increases  the  overall  displacement  by  an  amount  that  is 
inversely proportional to soil stiffuess. Thus it tends to reduce the demands on the structure but because of 
the  increased flexibility  of the  system, the overall displacement increases.  These effects  can be  important 
for  tall,  slender structures or for  closely spaced  structures that may be subject  to  pounding when relative 
displacements become large. 
In a seismic soil-structure interaction analysis, a structure with finite dimensions interacts dynamically through the 
structure-soil interface with a soil of infinite dimensions. A detailed analysis for this problem may be desirable and 
can be accomplished effectively using the finite  element (or finite  difference) method.  Methods for the analysis of 
soil-structure interaction can be divided into two main categories:  direct methods and multistep methods. 
Direct Method: entire  soil-foundation-structure  system  is  modelled and  analysed in  one  single  step.  Free-
field  input motions  are specified along the base and sides of the model and the resulting response ofthe interacting 
system is  computed.  It is  preferable  that the base  of the  mesh is placed at the top of the bedrock.  The governing 
equations of motion for this  case are 
in which 
{u}  are the relative motions between nodal points in the soil or structure and the top ofthe rock and  {ub(t)}
Earthquake. Resistant Design 99
arethespecifiedfree-field accelerationsattheboundarynodalpoints.
The ground motion, U
, is prescribedforthe surface ofbedrock orfirm ground. When the near surface soils are
not firm ground, as is most oftenthe case, the corresponding free-field motionofthe model, fib' is appliedat the
appropriatedepthas outcropmotionandthe surfacemotionis predictedaccordingly.Thesurfacemotionpredicted
reflectsthesoilconditionsatthesite. Inthisprocess,nonlinearityof soilbehaviourshouldbeaccountedforinorder
to avoidumealisticamplificationoftheresponse.
Equation6.15 is solvedinthefrequency domainusingFFT, timedomainusingtheWilsoneormodifiedNewton-
Raphsonmethod,orin terms of modalanalysis.
Several software packages are available now that have the capability to analyse the soil-structure interaction
ofanalysisis recommendedforcriticalstructuresorwhenperformance-baseddesignis considered.
Multi-step Method: Inthismethod, emphasis is placedon thenotations ofthe kinematic andinertialinteraction.
Thisis accomplishedby isolatingthe two primarycauses ofsoil-structureinteraction. Becausethis methodrelies
onsuperposition, it is limitedto the analysis oflinear(orequivalentlinear) systems. The analysis isdescribedas
Kinematic interaction: In the free-field, an earthquake will cause soil displacements in both the horizontal and
verticaldirections. Ifafoundation on thesurface of, orembeddedin, asoildepositis sostiffthatit cannotfollow
the free-field deformationpattern, its motion will be influencedbykinematic interaction, evenifit hasno mass.
Kinematicinteractionwilloccurwheneverthestiffnessof thefoundationsystemimpedesdevelopmentof thefree-
field motions. Kinematicinteraction can also induce different modes ofvibration in a structure. Forexample, if
propagatingwavescan,in asimilarmanner,inducetorsionalvibrationofthefoundation.
Themulti-stepanalysisproceedsas follows:
1. Akinematicinteractionanalysis,inwhichthefoundation-structuresystemisassumedto havestiffnessbut 
nomass,isperformedandthefoundationinputmotionis obtained. 
2. The foundation input motion is applied to obtain an inertial load on the structure in inertial interaction 
Massive failures occurred during the Alaska (1964) and Niigata (1964) earthquakes showed the importance of
damagecausedbygroundfailure andtheneedfor an analysisofthesuitabilityofthesiteselectedforthestructure
beforeits design and construction. Whilein certaincases ofgroundfailure itis possibleto designsafestructures
by properly designing their foundations, in other cases some mitigating measures must be taken such as soil
Seismic liquefaction refers to a sudden loss in stiffness and strength ofsoil due to cyclic loading effects ofan
earthquake. The loss arises from a tendency for soil to contract under cyclic loading, and ifsuch contraction is
prevented or curtailed by the presence ofwater in the pores that cannot escape, it leads to a rise in pore water
the strength andstiffnessalso dropto zero andthesoilbehavesas aheavyliquid. However,unless thesoilisvery
looseitwilldilate and regainsomestiffnessandstrength,as it strains.Thepost-liquefactionstrengthis called the
residualstrength andmaybe 1to 10timeslowerthanthestaticstrength.
100  Canadian  Foundation  Engineering  Manual 
Ifthe residual strength is sufficient, it will prevent a bearing failure for level ground conditions, but may still
resultin excessivesettlement. Forsloping ground conditions, ifthe residual strength is sufficient itwill preventa
flow slide, butdisplacementscommonlyrefen-edto as lateral spreading, couldbe excessive. Inaddition, evenfor
level groundconditionwhere there is no possibility ofaflow slide and lateralmovements maybetolerable, very
significantsettlements mayoccurdue to dissipationofexcessporewaterpressures duringand afterthe periodof
waves. The soilresponsedepends onthe mechanicalcharacteristicsofthe soillayers,thedepthofthewatertable
andthe intensityanddurationofthe groundshaking. Ifthe soilconsistsofdeposits ofloosegranularmaterials it
settlementsofthe groundsurface.Thiscompactionofthesoilmayresultinthedevelopmentofexcesshydrostatic
pore water pressures ofsufficientmagnitude to cause liquefaction of soil, resulting in settlement, tilting and
Liquefactiondoesnotoccuratrandom,butisrestrictedto certaingeologicandhydrologicenvironments,primarily
recently deposited sands and silts in areas withhigh ground water levels. Generally, the youngerandlooser the
to liquefactionincludeHolocenedelta,riverchannel,floodplain,andaeoliandeposits,andpoorlycompactedfills.
Liquefaction has been most abundant in areas where ground water lies within 10mofthe ground surface; few
instances ofliquefactionhaveoccun-edinareaswithgroundwaterdeeperthan20m. Densesoils, includingwell-
compactedfills, havelowsusceptibilitytoliquefaction.
6.6.1 Factors Influencing Liquefaction
Thefollowingfactors influencetheliquefactionpotentialofagivensite:
1.  Soil type: saturated granular soils, especially fine loose sands and reclaimed soils, with poor drainage
conditionsaresusceptibleto liquefaction.
2. Relativedensity: loosesandsaremoresusceptibleto liquefaction, e.g., sandwithDr> 80%isnotlikelyto
3. Confiningpressure: theconfiningpressure,cr ' increasestheresistanceto liquefaction.
4. Stress due to earthquake: as the intensity ofthe ground shaking increases, the shear stress ratio, (T/cr )'
increasesandtheliquefactionis morelikelyto occur.
5. Durationofearthquake: as the duration ofthe earthquakeincreases,the numberofstress cycles increases
leadingtoanincreaseintheexcessporewaterpressure, andconsequentlyliquefaction.
6. Drainageconditions:poordrainageallowsporepressurebuild-upandconsequentlyliquefaction.
6.6.2 Assessment of Liquefaction
Liquefactionassessmentinvolves addressingthefollowingconcerns:
evaluation ofliquefaction potential, Le., will liquefaction be triggered in significant zones ofthe soil 
couldabearingfailure orflow slideoccurandif not, 
Theseeffectscanbe assessedfromsimplifiedordetailedanalysisprocedures.
Earthquake - Resistant Design 101
Simplified analysis ofliquefaction triggering involves comparing the Cyclic Stress Ratio, CSR caused by the design 
earthquake with the Cyclic Resistance Ratio, CRR that the soil possesses due to  its density. 
6.6.3 Evaluation of Liquefaction Potential
Liquefaction potential  can be  evaluated  if the  cyclic  shear  stress  imposed by the  earthquake  and  the  liquefaction 
resistance  of the soil are characterized. Methods used to  evaluate the  liquefaction potential can be categorized into 
two main groups:  methods based on past perfonnance and analytical procedures. Liquefaction Potential Based on Past Performance
Based on the damage survey and field observations after earthquakes, the liquefaction potential can be identified from 
the perfonnance of similar deposits. An example for  this  approach is  the  method developed based on observations 
from  the  Niigata  Earthquake  (1964).  In  this  method,  the  standard  penetration  resistance,  N,  and  the  confining 
pressure are used to characterize the liquefaction resistance of soil. Based on this approach, it may be suggested that 
sands with N > 20 are not susceptible to  liquefaction. The earthquake magnitude, M, and the epicentral distance of 
liquefied sites are used to characterize the cyclic loading from the earthquake. Based on observations from previous 
earthquakes,  it may be  suggested that earthquakes  with magnitudes  less  than 6 and/or epicentral distances  greater 
than 500 km may not induce liquefaction. Analytical Procedure
A number of approaches have been developed over the years to evaluate the liquefaction potential. The most common 
of these, the cyclic stress approach, is briefly presented. 
Following the procedure proposed by Seed and ldriss (1971), the initial liquefaction is defined as  the point at which 
the increase in pore pressure, u  , is equal to the initial effective confining pressure [i.e.,  when u  = cr :cJ. 
excess  excess  J
The cyclic stress approach involves two  steps and their comparison: 
1.  Calculation  of cyclic  shear  stresses  due  to  earthquake  loading  at  different  depths  expressed  in  tenns  of 
cyclic stress ratio, CSR. 
2.  Characterization of liquefaction resistance of the soil deposits expressed in tenns of cyclic resistance ratio, 
These two steps are  described as follows. Characterization of Earthquake Loading
The cyclic stress approach is  based on the assumption that excess pore pressure generation is  fundamentally related 
to the cyclic shear stresses. The earthquake loading is characterized by a level ofunifonn cyclic shear stress, derived 
from ground response analysis or from a simplified procedure, applied at an equivalent number of cycles. 
Ground response analyses  should be  used to predict time histories  of shear stress  at different depths  within a  soil 
deposit. An equivalent unifonn shear stress is then calculated as 0.65  of the peak shear stress obtained. 
Seed's  Simplified  Equation:  For small projects, the simplified procedure proposed by Seed and Idriss (1971)  can 
be  used  to  estimate the  cyclic  shear stress  due  to  the  earthquake for  level sites,  in tenns  of the  cyclic  stress  ratio, 
CSR, Le.: 
102 Canadian Foundation Engineering Manual
a == the peak ground surface acceleration for the design earthquake, g := gravity acceleration, (j" := total
overburden pressure, cr:, = the initial effective overburden pressure and rd = stress reduction value at
the depth of interest that accounts approximately for the flexibility of the soil profile. The stress reduction
coefficient, r
, can be approximated by
r,,= 1.0 - 0.00765z for z S 9.15m (6.l7a)
rd=1.174-0.0267z for 9.l5m<zs23m (6.17b)
z is the depth below the ground surface in metres.
Ground response analysis using equivalent-linear total stress programs: Liquefaction Triggering is traditionally
assessed by conducting an     ground response analysis using the 1 D program SHAKE.
The analyses can also be conducted in 2D using the program FLUSH and others.
The induced cyclic stress ratio (CSR) (0.65 of the peak value of 'cyJa'vo) from the ground response analysis is
equated to the cyclic resistance ratio (CRR) to obtain a factor of safety against liquefaction triggering as indicated
in equation 6.18. Input for the ground response analysis would be the firm ground time histories. As indicated in
equation 6.18 below, corrections are typically made for magnitude (K
), confining stress (K,) and sometimes static
bias (K):
Factor of Safety against liquefaction = (CRR x K x K x K )/CSR (6.18)
C1 In a
Ground response analysis using non-linear total-stress program with hysteretic damping: In the equivalent
linear analyses, the same damping is used for both small strain and large strain cycles throughout the duration of
shaking. In reality, small strain cycles will have significantly lower damping than high strain cycles. This shortfall
can be addressed by using a constitutive model with hysteretic damping. Such models have been developed to run
within FLAC and other programs and can be used to assess liquefaction triggering in both 1 D and 2D approximations.
The CSR would typically be set equal to 0.65 of the peak value and factor of safety against liquefaction would be
calculated using equation 6.18. The method should be calibrated using measured responses from actual earthquakes
prior to use. Other advantages of the method are that it can be readily used in 2D analyses and therefore used with
sloping ground surface. Structural elements can be included and soil-structure effects modeled if desired.
2D total stress models which track the dynamic shear stress history within each element and trigger liquefaction if
a specified threshold is reached are also available.
Ground response analysis using non-linear effective stress programs: These procedures can be used to assess
both liquefaction triggering and the consequences of liquefaction. Seismic Hazard, Choice of Magnitude and Records
This section deals with the earthquake hazard, the magnitude ofthe earthquake to be used in liquefaction assessment,
and suggestions on earthquake records to be used.
Use the spectra given in the NBCC (2005) for firm ground conditions for the I :2475 hazard (for Vancouver, use the
Cascadia subduCtion hazard). If the 1 :475 hazard is needed this can be scaled from the 1 :2475 spectrum or found
in Geological Survey of Canada web site.
Magnitude for use in Liquefaction Assessment
Deaggregation of the hazard for Vancouver for the 1 :2475 probability gives magnitudes ofM6.5 to M6.9 depending
Earthquake· Resistant Design 103
on whether the mean or median values, and the Sa(.2) or Sa(l) deaggregation is considered. Using the 80th
percentile on the deaggregation results gives a range of M7.Oto M7.3. The results for the 1:2475 Sa(0.2) and
SaC 1.0) deaggregation is shown in the Table 6.4 below. The maximum recorded crustal earthquake in the Vancouver
region has been M7.3, but the hazard calculations assume an upper bound ofM7.7 as being possible. It is suggested
to use the Sa(l) deaggregation because it gives larger values and the period of 1 second is closer to the first period
of many soft sites than is the period of 0.2 seconds.
The 80th percentile deaggregation value should be used because the seismic hazard is substantially influenced by
the upper tail ofthe seismic hazard, as the larger ground motions have a much higher probability of causing damage.
Therefore, for Vancouver, a magnitude ofM7.25 should be used in assessing liquefaction for the I :2475 hazard, and
M8.2 should be used for the Cascadia subduction earthquake. If the I :475 hazard is considered, use M6.5.
TABLE 6.4 Earthquake Magnitude/or Vancouver Evaluatedfi'om Deaggregation
. Measure
Mean 6.52 6.90
Median 6.51 6.82
80%ile 6.93 7.30
Selection of earthquake records
The Geological Survey of Canada is assembling a suite ofrecords for both the 1 :2475 and 1 :475 probabilistic hazard
and for the Cascadia subduction earthquake. However, it is not easy to find a suite of records that give a good fit
to the spectrum and have the appropriate duration and/or number of cycles. Some useful guidelines for choosing
records are:
The records should have a spectrum close to the UHRS, and should have duration consistent with the magnitude.
The record should be scaled so that the spectrum matches the design spectrum in the period range of interest
(related to the fundamental period of the site), or the records should be spectrum matched to the design spectrum.
The record should have a number of large cycles, for example the NCEER assessment criteria assume that a M7
earthquake record has 10 significant full cycles greater than 0.65 PGA. Characterization of Liquefaction Resistance
The Cyclic Resistance Ratio, CRR, is a measure ofthe soils ability to resist liquefaction and the development oflarge
strains, and depends mainly on the soil type and density or state. There are two approaches to the characterization
of liquefaction resistance, namely methods based on the results of laboratory tests, and methods based on in-situ
Laboratory tests: Different laboratory tests are performed mostly on isotropically consolidated triaxial specimens
or on Ko- consolidated simple shear specimens. In these tests, liquefaction failure is defined as the point at which
initial liquefaction was reached or at which some limiting cyclic strain amplitude (commonly 5-20 %) was reached.
The measured cyclic stress at the onset of liquefaction failure is the liquefaction resistance and is frequently given
in terms of the cyclic resistance ratio, CRR = 'tCy/cr'vo'
Comments on testing methods: Undisturbed samples retrieved using specialized sampling techniques (such as
ground freezing) should be used in the tests. The simple shear test is the most common test although it is difficult
to eliminate its problems. The torsional shear test is sometimes used to ensure uniform distribution of the shear
stress but it is very costly and difficult to obtain a hollow sample. Shaking table tests suffer from the lack of suitable
104 Canadian Foundation Engineering Manual
confining pressure.  Cyclic triaxial tests  are  also  used,  however,  they  impose  different  loading conditions  than the 
soil experiences during an earthquake and their cyclic stresses need to be corrected. 
Cyclic  simple  shear tests  are  considered most  representative  of field  conditions  during  earthquake  loading,  The 
results of such a test for loose Fraser River sand are shown in Figure 6.5. The effective stress path shows the nonnal 
effective  stress reducing with each cycle  of shear stress from  an  initial value  of lOO  kPa to essentially zero after 6 
cycles.  Figure  6.5  also  shows the shear stress Vs  shear strain  response,  where  it  may  be seen that strains  are  very 
small,  less  than  0.1  %,  for  the  first  5  cycles  and  become  very  large,  10  %,  on  the  6
cycle,  when  liquefaction  is 
triggered. The applied stress ratio for this sample was 0.1  and caused liquefaction in 6 cycles. The CRR is generally 
specified as the  stress ratio to cause liquefaction in  15  cycles, and from additional tests carried out on this  material 
(CRR\s  0.085. 
Vertical Effective Stress, cr',CkPa) 
•  Point ofy=3.75% 
o'",=lOOkPa;  D,,=40% 
(Le.  Assumed  triggering poinl 
',.,JI1',,=0.1 0;    =0.0
20  of liquefaction  for 
comparison  purposes) 

!l- '"

t: '"
5  r.Il 1 
-30  '----------------------' 
•  Point ofy=3.75%  0,,=40% 
(i.e.  Assumed triggering  'q.J<:/",=0.10; ,.10'", =0.0 
point ofliquetaction tor  20 
comparison purposes) 
10 5 
-20  . 
Shear Strain, y(%) 
FIGURE 6.5 Stress path andshear stress-strain response ofloose Fraser River
sand, cyclic simple shear tests (Wijewickreme et al. 2005)
The liquefaction response shown in Figure 6.5 is typic(j.l for loose sands where the application of an additional cycle 
of load triggers  an  abrupt change  in behaviour from  stiff to  soft.  The  soft post-liquefaction response is  controlled 
by dilation. The drop in shear stiffness upon liquefaction can be in the range of 100 to  1000 times. The strength or 
strength ratio  available after liquefaction,  called the residual  strength  can be significant,  and  from  Figure  6.5,  the 
strength ratio is  at least 0.1  for  loose Fraser river sand.  However, field  experience indicates that the  strength ratio 
can be significantly lower than values obtained from undrained tests. The reason for this may be due to upward flow 
of water associated with generated excess pore water pressures. This may cause some elements to  expand  lose 
their dilation effect, particularly those beneath layers of lower permeability. 
For silt  and clay material  the  response  to  cyclic  loading  and  liquefaction  can  be  quite  different  than  for  sand  as 
shown  in  Figure  6.6.  This  figure  shows  effective  stress  path and  shear stress-strain response  for  loose  normally 
consolidated Fraser River silt under cyclic simple shear loading. The effective stress path shows the normal effective 
stress reducing with each cycle from  its  initial value  lOO  kPa, but not dropping below  10 kPa.  After the initial few 
cycles,  loading is associated with an increase in effective stress resulting from dilation.  Only the unloading shows 
strong contraction effects. The shear stress-strain response shows a gradual increase in strain with number of cycles, 
and there is no abrupt change in shear stiffness from stiff to  soft. There is also no  indication of a strength reduction 
below the applied stress ratio of 0.2, thus the post-liquefaction or residual strength ratio is at least 0.2 for the tested 
silt.  The stiffness reduces with each cycle, and after  11  cycles is  10  to 20 times softer than the first cycle. 
Earthquake· Resistant Design  105 
OCR  =1.0 
CSR  = 0.20 
ro e
~ W  36.2% 
:: 1 


0 20 40 60 80 100 120
Vertical  Effective Stress, crv (kPa)  Shear Strain, y (%) 
FIGURE 6.6 Stress path and shear stress-strain response ofFraser River silt,
cyclic simple shear tests (Sanin and Wijewickreme, 2006)
These test results indicate that fine-grained normally consolidated silts and clays of low plasticity can be far more
resistant to liquefaction than loose sands.
Test results together with field experience suggest that the liquefaction response of coarse-grained soils, gravels,
sands and non-plastic silts should be handled differently than fine-grained silts and clays. While it might seem
desirable to recover undisturbed samples (it is possible to do so in fine-grained soils) and obtain a direct measure of
liquefaction resistance from cyclic testing, it is very difficult and expensive to obtain undisturbed samples in coarse-
grained soils. It is therefore recommended that CRR for coarse-grained soils be based on penetration resistance in
accordance with NCEER (2001). For fine-grained soils, it is recommended that CRR be based on Atterberg limits
. and/or direct testing.
In-situ tests: The soil parameters determined from in-situ tests are used as liquefaction resistance parameters.
Standard penetration resistance: The corrected SPT resistance is plotted vs. cyclic resistance ratio for clean sand
(Figure 6.7) sites where liquefaction was or was not observed in earthquakes ofM: 7.5 to determine the minimum
cyclic stress ratio at which liquefaction could be expected. CRR for other magnitudes may be obtained by multiplying
the CRR for M =  7.5 earthquakes by a correction factor, Kw as recommended by NCEER (2001), i.e.:
The data used in Figure 6.7 are for cyclic resistance ratios associated with overburden pressure, cr o =  100 kPa. For
higher overburden pressure values, the cyclic resistance ratio must be corrected using a correction factor K" given
eRR '
Values for Kef may be taken from the average curve of Seed and Harder (1990) (Figure 6.8).
~   J
106 Canadian Foundation Engineering Manual

-- ;>


Percent Fines =35  15  s: 5 
.. S()+ 
.. 12 
, I
[ I
I  1 
[ I
01 1

I f eRR curves for 5,15,  and 
35  percent fines, respectively 
(J 0.2
Modified Chinese Code Proposal (clay content = 5%) ® 
Marginal  No 
Liquefaction  Liquefaction  Liquefaction 
Pan • American data  8 III
Japanese data  •  Q  (1)
; Chinese data.t.  A.
__ __ ____
o  10  20  30  40  so 
Corrected Blow Count, (N
FIGURE 6.7 CRR! vs (N) 60 (Youd et al. 2001)
Cone penetration test: The tip resistance from the cone penetration test (CPT) is used as  a measure ofliquefaction 
resistance.  CPT-based  liquefaction  curves  have  been  developed  based  on  correlation  with  laboratory  test  and 
theoretically derived values of CPT resistance (Figure 6.9). In CPT-based liquefaction evaluations, the tip resistance 
is normalized to a standard effective overburden pressure of 96 kPa by 
(6.21) or 
VERTICAL EFFECTIVE STRESS. (Iv' (a!m  units, e.g.  Isf) 
FIGURE 6.8 Recommended curves for estimating KJor engineering practice (Youd et al. 2001)
Earthquake. Resistant Deslgn  107 

(J) .2
co (.)
0::: c
(I) (fl 
.!::;(I) 0.2
.!::? .!::?
>. >. 0.1
0.25 <  Dso(mm)  < 2.0 
Fe (%)  <  5 
.... CPT Clean  Sand 
Base Curve 

No Liquefaction 


i  • ;,.....

o 0
0 0<t3

...... eot2

Field Performance  Uq.  No Liq. 
NCEER (1996) 
Stark & Olson (1995)  0
Suzuki eL al (1996b) 

Workshop  A

0 50 100 150 200 250 300
Corrected CPT Tip  Resistance,  qc1N 
FIGURE 6.9 Curve recommendedfor calculation ofCRRfrom CPT data along with empirical liquefaction
datafrom compiled case histories (reproducedfrom Robertson and Wride 1998) (Youd et al. 2001)
Shear wave velocity: The measured shear wave velocities can be used to assess the liquefaction resistance, (usually
in addition to assessment using SPT or CPT). Measured shear wave velocities are normalized to a standard effective
overburden pressure of 96 kPa by
s  vO  "/ 
n =  3 to 4. The normalized shear wave velocity is plotted vs. CRR in Figure 6.1 0, which can be used to
evaluate the liquefaction potential directly, or is used to evaluate the CRR, which is used in turn to evaluate
the liquefaction potential.
0.6 r-----r---r----,---,.,..,..--..,.----,
Data Bases on: 
Mw- 5.9108.3; adjusted  by
Mw= 7.5
dividing  CSR by (MwI7.5)-2.56.,352055 Fines 
Uncemented,  .  m  I  I Content (%)
Holoceneage sOils 
Average values of'"  • 
VS1and 8mBx  A I I 
A  0 
fines Content 
Jr..A6to 34%  • 
100 200 300
Overburden Stress-Corrected  Shear Wave 
Velocity,  VS1, mfs 
FIGURE 6.10 Liquefaction relationship recommended for clean, uncemented soils with liquefaction data
from compiled case histories (Reproducedfrom Andrus and Stokoe 2000) (Youd et al. 2001)
1 108 Canadian Foundation Engineering Manual
The base eRR obtained from these figures will be given the symbol eRR,. The CRR for a general condition is  given  . 1 
eRR = CRR  * K  * K  * K  (6.23)
1  111 fJ a 
K  is  a conection factor for earthquake magnitudes other than M7.S,
Km is  a conection factor to account for  effective overburden stresses other than  100 kPa, and 
K  is  a conection factor  ground slope. 

The recommended K  (MSF) curve is  shown in  Figure 6.11,  and for an M7 earthquake K  = 1.25. 
m  m 
The recommended K  curves  depend  on  relative  density  as  was  shown  in  Figure  6.8.  NCEER does  not make  a 

recommendation regarding K".  The default value is  unity,  K" 


<U 1.5 


-+- Seed and  Idriss.  (1982) 
  ........ Idriss 
x  Ambraseys (1985) 
¢  Anango (1996) 
•  Arango (1996) 
__ Andrus and Stokoe 
...  Youd  and Noble,  PL <20% 
t:. Youd  and  Noble, PL <32% 
...  Youd  and Noble.  PL <50% 
5.0  6.0  7.0  8.0  9.0 
Earthquake Magnitude, Mw 
FIGURE 6.11 Magnitude Scaling Factors derived by various investigators
(Reproduced/rom Youd and Noble 1997)  (Youd et al. 2001) Evaluation of Initiation of Liquefaction
The evaluation is easily performed graphically. First, the variation of cyclic stress ratio, CSR, with depth is plotted. 
The variation ofthe cyclic resistance ratio, CRR, with depth is then plotted on the same graph.  Liquefaction can be 
expected at  depths  where the  loading exceeds the resistance or when the factor of safety against liquefaction, FS
is  less than 1, where: 
L CSR Residual Strength for Gravel, Sands, and Non Plastic Silts
Field experience during past earthquakes  indicates that residual  strengths can be much lower than values obtained 
from  undrained tests on undisturbed samples. This may be due to upward flow  of water associated with generated 
excess  pore  water pressures.  This  may cause  some  elements  to  expand to  a  higher void  ratio,  and  hence  a  lower 
critical  state  strength.  Based on back analysis  of field  case histories,  Seed and Harder (1990) proposed upper and 
lower bounds on residual strength as shown in Figure 6.12.  It may be noted that there are no  data points associated 
with large movements or flow slides for SPT blowcounts greater than  16. 
1 ••

Earthquake - Resistant Design  109 
Olson and Stark (2002) present residual strength in terms of strength ratio, Figure 6.13.  Their values range between 
about 0.05  and 0.1  for  SPT blow counts  in the range 2 to  12.  They also  developed residual strength ratios  in  terms 
of CPT tip resistance.  Their relationship is  shown in Figure 6.14. 
•  Measured  SPT and  Critical Strength Data 

o  Estimated SPT and Critical Strength Data 
o  Construction Induced  Liquefaction -
-<  f-

Estimated  Data 







FIGURE 6.12 Recommended relationship between su,,. and N"60, CS  (Seed and Harder 1990)
•  Back-calculated liquefied strength ratio and measured SPT 
e  Back-calculated liquefied strength ratio and converted $PT 
from measured CPT 
o  Back-cak:ulated liquefied strength ratio and estimated SPT 
tJ.  Estimated liquefied strength ratio and measured, converted 

or estimated SPT 
(Number  beside  symbol  hdieates average  fines  content) 
Davies  and  Campanella  (1994) 
Proposed  relationship ---:7---'\ 
o  4  6  10  12  16  18 
Normalized  SPT  blowcOlllt,  (N1)60 
FIGURE 6.13 A comparison ofliquefied strength ratio relationships based on normalized SPT blowcount
(Olson and Stark 2002)
110  Canadian Foundation  Engineering  Manual 
or estimated CPT 
!  {)IUnber beside  symbol  ndIcates  average  fines  content) 
Normalized CPTtip resistance,  Qc1  (MPa) 
•  Back-calculated liquefied strength ratio  and  measured  CPT 
e  Back-calculated liquefied strength ratio and converted CPT 
from measured SPT 
o  Back-calculated liquefied strength ratio and estimated CPT 
D.  Estimated liquefied strength ratio and measured.  converted 
9  10 
FIGURE 6.14 A comparison ofliquefied strength ratio relationships based on normalized CPT tip resistance
(Olson and Stark 2002)
It is recommended that for  zones predicted to  liquefy, the residual strength be estimated as  follows: 
1.  For normalized  SPT blowcounts  less  than or equal  to  15,  use mean values  from  Seed and  Harder  (1990) 
and/or Stark and Olson (2002). 
2.  For normalized SPT blowcounts greater than or equal to  25, use drained strength values. 
3.  For normalized SPT blow count values between 15  and 25, use a linear variation of residual strength. 
Although  liquefaction can  be  triggered  in  dense  sands having normalized SPT blowcount values greater than 25, 
the drained strength values can be used, as  dilation upon straining will cause the pore pressures to drop to their pre-
earthquake values or lower.  CRR for Silts and Clays 
It has been noted that some fine-grained soils that classify as non-liquefiable according to commonly used empirical 
"Chinese  Criteria"  (Wang  1979;  Koester  1992;  Finn  et  al.  1994)  have  in  fact  experienced  liquefaction  during 
earthquakes (Boulanger et al.  1998,  Bray et al.  2004).  Some data from  laboratory cyclic shear testing of silts  also 
confirmed the limitation ofChinese Criteria as a tool to identify potentially liquefiable soils (Sanin and Wijewickreme 
2004; Boulanger and Idriss 2004). 
As  an  alternative,  Boulanger  and  Idriss  (2004)  recommend  that  fine-grained  soils  be  classified  as  "sand-like" 
(susceptible to liquefaction) iflp < 7,  and "clay-like" ifIp  7.  However, some limitations in this approach have been 
noted from  cyClic  direct simple shear tests  conducted on specimens from  a  cha1)l1el  fill  silt  from  the  Fraser River 
Delta (Sanin and Wijewickreme 2005). 
Earthquake - Resistant Design  111 
Based on the field perfonnance of fine-grained soil sites in Adapazari following the 1999 Kocaeli (Turkey)
earthquake, combined with data from laboratory cyclic shear testing, Bray et a1. (2004) have proposed alternate
empirical criteria to delineate liquefaction susceptibility of fine-grained soils. It is recommended that the Use of
Chinese Criteria be discontinued, and Bray et a1. (2004) criteria (Figure 6.15) be used to detennine liquefaction
susceptibility of fine-grained soils:
a) w/wL  2: 0.85 and Ip::: 12: Susceptible to liquefaction or cyclic mobility*;
b) W/WL 2: 0.8 and 12 < Ip < 20: Moderately susceptible to liquefaction or cyclic mobility*;
c) W/WL  < 0.8 and Ip 2: 20: No liquefaction or cyclic mobility, but may undergo significant defonnations if
cyclic shear stresses> Static undrained shear strength (s).
*This classification may be revised on a site-specific basis using data from laboratory cyclic shear testing of
good quality field samples [e.g., samples obtained using thin-walled tube samples with sharpened (i.e., <50)
cutting edge and no inside clearance].
til 20
Fraser River 
channel fill 


0.0 0.5 1.0 1.5 2.0

FIGURE 6.15 Bray et al. (2004) criteriafor liquefaction assessment affine-grained soils  Residual Strength for Silts and Clays 
It is recommended that the residual strength (Sr) for silt and clay zones be detennined as per guidelines given
a) w/wL 2: 0.85 and Ip::: 12: Sr =remolded shear strength (Sremolded), unless appropriate testing ofundisturbed
samples can show greater strength;
b) w/wL  2: 0.8 and 12 < Ip < 20: Sr =  0.85s
' where Su static undrained shear strength;
c) W/WL < 0.8 and Ip 2: 20: Sr Su-
This approach essentially employs the liquefaction potential detennined using the recommended Bray et a1. (2004)
criteria as the basis for the detennination of Sr. This assumes that the full static undrained strength (su), or most part
of it, is available as the residual strength after cyclic loading, unless the soil is susceptible to liquefaction.
6.6.4  Liquefaction-Like Soil Behaviour 
The liquefaction potential ofloose, saturated sands is well recognized as described above. Similar abrupt structural
changes, however, could be caused by earthquakes also in some highly sensitive clays such as the Canadian Leda
clay or the Norwegian quick clay.
112 Canadian Foundation Engineering Manual
6.6.5 Induced Ground Movements
There  are  several  empirical  and  approximate  procedures  for  estimating  ground  movement  for  situations  where 
liquefaction may be triggered. 
The  lateral  spreading  equation  of Youd  gives  ground  displacement  as  a  function  of simple  site  properties,  soil 
profile properties, and earthquake magnitude and distance. Post-liquefaction settlement is discussed in the following 
subsection. Post-Liquefaction Settlements for Coarse-Grained Soil
Post-liquefaction settlements  occur during and after earthquake shaking.  For level  ground  conditions,  the  amount 
can be  computed from the volumetric reconsolidation strains induced as  the excess pore pressures dissipate. Based 
on field  experience during past earthquakes, the amount of strain depends on SPT blowcount and the  CSR applied 
by the  design  earthquake.  The  curves proposed by Cetin et  al.  (2004)  are  shown in  Figure  6.16  and  indicate  that 
volumetric reconsolidation strains can range between about  10  % for  very loose sand to  1 % for  very dense sands. 
These curves are recommended. 
The  settlement calculated from  this chart is induced by consolidation of the liquefied soil only.  Footings and other 
structures  founded  over or within  liquefied  soil will  also  deform due to shear strain within the  liquefied soil.  This 
shear strain typically occurs during the period of strong shaking whereas the consolidation settlements often occur 
following the period of strong shaking. The shear strain deformations are additional to the consolidation settlements 
and can be of similar or greater magnitude. 
Cetill et ai.  (2002) 
6  10  15  20  40  46 
FIGURE 6.16 Recommended relationships/or volumetric reconsolidation strains as afunction 
0/equivalent uniform cyclic stress ratio and N/, 60, cs/or Mw  7.5  (Wu  2002) 
6.7 Seismic Design of Retaining Walls
! '
The dynamic response ofretaining walls is quite complex. Walls can translate and/or rotate, and the relative amounts  I
of translation  and  rotation  depend  on  the  wall  design.  The  magnitude  and  distribution  of dynamic  wall pressures 
during  an  earthquake  are  influenced  by the  mode  of wall  movement.  The  maximum soil  thrust  acting  on  a wall 
generally occurs when the wall has  moved toward the  backfill.  The minimum soil thrust occurs when the  wall has  I
moved away from  the  backfill. The shape of the  earth pressure distribution on  the  back of the  wall  changes  as  the 
wall  moves.  The position  of the  resultant  of the  dynamic pressure is  highest when the wall has  moved toward the 
soiL  Dynamic  wall  pressures  are  influenced  by  the  dynamic  response  of the  wall  and backfill,  and  can  increase 
significantly  near  the  natural  frequency  of the  wall-backfill  system.  Permanent  soil  displacements  also  increase 
Earthquake - Resistant Design  113 
at  frequencies  near  the  natural  frequency  of the  wall-backfill  system.  Because of the  complexity  of the  problem, 
simplified models that make various simplifying assumptions  are used for the seismic design of retaining walls. 
6.7.1  Seismic Pressures on  Retaining Walls 
Seismic pressures  on  retaining  walls  are  usually  estimated using simplified methods.  Some  of these  methods  are 
given here.  Active  Earth  Pressure Conditions  M·O Method 
This  method  is  based on  a pseudostatic  analysis of seismic earth pressure on retaining structures  and has  become 
known as the Mononobe-Okabe (M-O) method. The M-O method is a direct extension of the static Coulomb theory 
to pseudo-static conditions. 
For dry cohesionless backfill, the  total  active thrust can be expressed in a  form similar to  that developed for  static 
conditions, i.e.: 
the dynamic active earth pressure coefficient, K
, is 


8cos(o  +8 +\jf{1+
soil  angle  of internal  friction,  e=  slope  of backfill  with  horizontal,  P  slope  of the  back face  of the 
retaining wall with vertical,  (5  =  angle of friction of wall-backfill interface, 
\If tan-I (1    ,and kh and kv are seismic coefficients in the horizontal and vertical directions, 
respectively, for <p-P2::\If.  The seismic coefficient in the horizontal direction,  k
, is defined as a ratio of the peak 
ground acceleration in the horizontal direction to the gravity acceleration, g, Le.: 

The seismic coefficient in the vertical direction, k
, is  defined similarly. 
The total active thrust, PAP  can be  divided into a static component, P
and a dynamic component, ilP

in which, 
KA = the coefficient of static active earth pressure (from Coulomb theory), i.e.: 
114 Canadian Foundation Engineering Manual
($ -0)
=----'[:---;:=======::;2]' KA
ecos(8  + 0)  1 +  I sin(8  + $ ) sin($ - )  -
+O)cosW  -0) 
The total active thrust may then be considered to  act at  a height, h, from the base of the wall, 
h =P,(H /3)+ M
PA£ (6.31 )
Passive Earth Pressure Conditions M-O Method
The total passive thrust on a wall retaining a dry cohesion less backfill is given by 
the dynamic passive earth pressure coefficient, KpE'  is 

cos'Vcos26cos(8  -6 +'11 
The total passive thrust can also  be divided into static and dynamic components: 
Pp is the static passive thrust, given by 

ecos(8  -e{l+             
Note that the dynamic component, llPPE'  acts in the opposite direction of the static component, Pp'  thus reducing the 
available passive resistance. 
Discussion: The M-O analysis provides a useful means of estimating earthquake-induced loads on retaining walls. 
A positive horizontal seismic coefficient causes the total active thrust to  exceed the static active thrust and the  total 
passive  resistance  to  be less  than  the  static  passive resistance.  Since  the stability  of a particular wall  is  generally 
reduced by an increase in  active thrust and/or a decrease in passive resistance,  the  M-O method produces  seismic 
loads that are more critical than the static loads. 
The M-O analysis has some limitations.  The determination of the  seismic coefficient is difficult; the analysis is not 
appropriate for  soils that experience significant loss of strength during earthquakes,  and it over predicts  the actual 
total passive thrust,  particularly for (5  >
Earthquake. Resistant Design 115
6.7.2 Effects of Water on Wall Pressures
The  water  exerts  loads  on waterfront retaining walls  both  during  and  after  earthquakes.  The  water  outboard of a 
retaining  wall  and within the  backfill  can  exert  dynamic  pressures  on  the  wall.  The  total  water pressures that act 
on retaining  walls  in  the  absence  of seepage within the  backfill  can be  divided  into  two  components:  hydrostatic 
pressure  that  increases  linearly with depth  and  acts  on the  wall  before,  during  and  after  earthquake  shaking,  and 
hydrodynamic pressure that results from the dynamic response of the water itself. Water Outboard of Wall
The hydrodynamic pressures on a retaining wall are usually  estimated from Westergaard's  solution for the case of 
a vertical rigid darn retaining a semi-infinite reservoir of water that is excited by harmonic, horizontal motion of its 
rigid base.  Westergaard  computed the  amplitude  of the  hydrodynamic  pressure  at  a  depth  Zw  below water surface 
H  depth of the water. The resultant hydrodynamic thrust is  given by 
7  2
H (6.38)
P" == 12kh"Y w
The total actual thrust due to the water is  equal to the sum of the hydrostatic and hydrodynamic thrusts. Water in Backfill
The presence ofwater in the backfill behind a retaining wall can influence the seismic loads on the wall in a number 
of ways.  It alters the  inertial forces  within the backfill and develops hydrodynamic pressures within the backfilL 
For low permeability soils, the inertial forces  due to earthquake shaking will be proportional to the total unit weight 
ofthe soil. In this case, the M -0 method can be modified to account for the presence ofporewater within the backfill 
Yb =unit weight of backfill and  r
An equivalent hydrostatic thrust based on a  fluid  of unit weight  Y Y +  r"Y must be  added to  the  soil thrust. 
eq w b
Soil thrusts from partially submerged backfills may be computed using an average unit weight based on the relative 
volumes of soil within the active wedge that are above and below the phreatic surface. 
For high permeability soils, the inertial forces will be proportional to the submerged unit weight of the soil.  In this 
case, the porewater pressure acting on the wall is  given by the Westgaard solution, i.e., Equations 6.37 and 6.38. 
6.7.3 Seismic Displacement of Retaining Walls
The serviceability of retaining walls is related to permanent deformations that occur during earthquakes. Therefore, 
analyses that predict permanent wall deformations provide a  more useful indication of retaining wall performance. 
116 Canadian Foundation Engineering Manual Deterministic Approach
This method is developed for the seismic design of gravity walls based on allowable wall displacements. In this
method, the yield acceleration, defined as the acceleration that is just large enough to cause the wall to slide on its
base, is calculated by (Richard and Elms, 1979) n
in which ,
is calculated using the M-O method with kh = , and Wis the weight of the retaining wall.
g ··.
'· .. 1'·
G,. = [tanG>b
The permanent displacement can then be calculated from .
2 3
0.087 m   x ~ m   x
V = the peak ground velocity and a the peak ground acceleration.
max max . Statistical Approach
Whitman and Liao (1985) used a statistical approach to evaluate the permanent displacement of retaining walls due
to earthquake excitation. They studied the results of sliding block analyses of 14 ground motions and found that the
permanent displacements were lognormally distributed with mean values
(6.42) Finite Element Analysis
The finite element analysis can be used to compute the earthquake-induced deformations of retaining walls. A
rigorous' analysis should be capable of accounting for nonlinear, inelastic behaviour of the soil and of the interfaces
between the soil and the elements of the wall. Some considerations have to be included in the analysis with respect
to the boundaries and elements size.
6.7.4 Seismic DeSign Consideration
The design of retaining walls for seismic conditions is similar to the design for static conditions. Seismic design
procedures make use ofsimplifying assumptions to allow the use of available procedures for static conditions . Gravity Walls
Gravity walls are customarily designed using one of two approaches: a seismic pressure-based approach or a
permanent displacement-based approach.
Design Based on Seismic Pressures: The M-O method is commonly used along with an inertial force with the same
pseudo-static acceleration applied to the active wedge as is applied to the wall itself. Pseudo-static accelerations are
generally considerably smaller than anticipated peak accelerations (values between O.OSg and O.lSg are used). The
wall must be designed to avoid sliding, overturning and bearing capacity failure. The pseudo-static forces along
with static analysis procedures are used in this approach.
Design Based on Allowable Displacements: This approach allows the designer to cotisider the consequences
of permanent displacement for an individual wall when selecting an allowable displacement for design. Design
procedures based on Richard-Elms (1979) and Whitman-Liao (l985) methods for estimation of permanent I
displacement as discussed in Sections and I
Earthquake - Resistant Design  117 
TheRichard-Elmsprocedureissummarizedas follows:
1. Selectanallowablepermanentdisplacement, dall'
2. Calculatetheyieldaccelerationrequiredtoproducetheallowablepermanentdisplacementas
3. Calculate P
using the M-O method with the yield acceleration from step 2 as the pseudostatic
4. Calculate the wall weight required to limit the permanent displacement to the allowable permanent
W =  P
cas(S +e) - PAE  sin(8 + e )tan<P
tan<Pb- a,.1 g  (6.44)
5. ApplyafactorofsafetytotheweightofthewalLAfactorofsafety,FS == 1.1 to 1.2is suitable.
definedprobabilitiesofexceedence.Theyieldaccelerationinthiscaseis calculatedas
( - 2 J 
a =-n 
y  9.4 am.xd
M =  modelerror=3.5.Then,the sameprocedureasRichard-Elmsisfollowed. Reinforced Soil Walls
Duringanearthquake,areinforcedsoilwallis subjectedtoadynamicsoilthrustatthe backofthereinforcedzone
and to inertial forces withinthe reinforced zone in addition to static forces. Thewall mustbedesignedto avoid
externalinstability(sliding,overturningandbearingcapacityfailure) andinternalinstability(pulloutfailureofthe
External Stability: A reinforced earthwall can be treated like a gravitywall. Theexternal stabilityofan earth
reinforcedwallcanbeevaluatedas follows:
1. Determinethepeakhorizontalgroundsurfaceacceleration,a .
2. Calculatethepeakaccelerationatthecentroidofthereinforcedzonefromtheequation

(1.45- G;ax)amax   6 . 4 ~ )
3. Calculatethedynamicsoilthrustfrom
a  'Y 
MAE  =0.375-
_  -
=unitweightof backfill.
4. Calculatethe inertialforceactingonthereinforcedzone
- Gc'Y  ,HL
lR -

118  Canadian  Foundation  Engineering  Manual 
5.  Add P £ and 50 % of P and check the  external stability. FS  (Seismic)::::  75  % FS  (Static). 
Internal Stability: Internal stability is  evaluated as  follows: 
1.  Determine the pseudo-static inertial force  acting on the potential failure  zone, 
p _  QcW,
1,1 - I',

is the weight of the failure  mass (Figure 6.17). 
2.  Determine  the  share  of each  reinforcement  layer  from  PIA' according  to  its  resistance  area  (this  is  the 
earthquake-induced tensile force  for each reinforcement layer). 
3.  Determine the total tensile force  for each layer as  the sum of the dynamic and static components. 
4.  Check that the reinforcement allowable tensile strength> 75  % of the total tensile force  for  each layer. 
S.  Check the length of the reinforcement so  that the FS against pullout failure> 7S  % FS  (static conditions). 
(8)  (b)
FIGURE 6.17  Critical potential failure surfaces for evaluation ofinternal seismic stability
ofreinforced earth walls: a) inextensible reinforcement; b) extensible reinforcement
6.8  Seismic Stability of Slopes and  Dams 
Slopes, embankments and dams may be damaged or may even fail due to earthquake induced shaking ofthe ground. 
Landslides often occur in earthquakes and dam failures have also been reported. There is no doubt that earthquakes 
can pose a  serious threat to  the stability of slopes  and can induce significant damage.  The  damage manifests  itself 
in the  form of slides, slumping,  cracks and permanent deformations. 
6.8.1  Mechanisms of Seismic Effects 
The mechanism leading to slope failures  can be attributed to two factors: the earthquake induced forces and stresses; 
and the radical structural change of the soil that may be brought about by these seismic stresses. 
The first effect is present even in soils that do not experience any basic change as  a result of the shaking such as  stiff 
clay,  gravel or dense,  coarse sand.  In this case,  some movement, could be substantial, of the slope occurs when the 
total  stress  exceeds the  strength available.  On the other hand, fine,  loose,  saturated san(ds  may undergo  a complete 
change  of character when they liquefY.  Liquefaction may occur in  a sizeable bulk of soil  or only  in narrow  seams 
and lenses  of liquefiable material  enclosed in relatively impermeable deposits. The liquefaction potential of loose, 
saturated sands is well recognized but similar abrupt structural changes could also be caused by earthquakes in some 
Earthquake. Resistant Design 119
highly sensitive clays such as  the Canadian Leda clay. 
6.8.2 Evaluation of Seismic Slope Stability
The stability  of slopes  is  influenced by many factors,  and  a complete slope  stability  evaluation must  consider the 
effects  of each  factor.  Geological,  hydrological,  geometrical  and  material  characteristics  are  needed  to  reliably 
perform both static and seismic slope stability analyses. 
The seismic stability of a slope is strongly influenced by its static stability because slopes with low factors  of safety 
against failure  under static conditions need low  additional dynamic stresses to reach yield. Therefore, the  factor of 
safety of any slope under static conditions must be significantly greater than 1.0 to  accommodate seismic demands. 
The  acceptable value of the  factor of safety depends on the uncertainty in  the  model  used for the analysis, the soil 
parameters and the magnitude and duration of seismic excitation, in addition to the potential consequences of slope 
An  analysis  of seismic  stability  of slopes  has  to  consider  the  effects  of dynamic  stresses  induced by  earthquake 
shaking; and the change in the strength and stress-strain behaviour ofthe slope materials due to the seismic loading. 
These effects  may lead to  yield and plastic deformations  due  to  inertial  or weakening effects.  The inertial effects 
occur when the earthquake-induced dynamic stresses reach the shear strength ofthe soil (that may remain constant), 
producing  slope  deformations.  The  weakening  effects  occur  when  the  soil  is  weakened  due  to  the  earthquake 
loading (liquefaction or softening) and cannot remain stable under earthquake-induced stresses. When the available 
shear  strength  becomes  smaller than  the  static  shear stress  required  to  maintain  equilibrium,  flow  failures  occur. 
Deformation failures  occur when the  shear strength of a soil  is  reduced below the  earthquake-induced (dynamic) 
shear stresses. 
The  potential  of a  flow  slide  is  commonly  evaluated  by  conventional  static  slope  stability  analyses  using  soil 
strengths based on end-of-earthquake conditions. 
In a typical analysis, the following procedure is used: 
1.  the liquefaction potential is  calculated at all points on a potential failure surface; 
2.  Residual  strengths  are  assigned  to  the  failure  surface  portions  with  factor  of  safety  against 
liquefaction < 1; 
3.  IfFS against liquefaction> 1, strength values are based on the effective stresses at the end ofthe earthquake; 
4.  Using  these  strength  values,  conventional  limit  eqUilibrium  slope  stability  analyses  are  performed  to 
calculate an overall FS  against flow  sliding.  If the  overall FS is  less than  1, flow  sliding is  expected. 
A number of techniques have been developed for the analysis of seismic inertial effects on slopes. These techniques 
differ in the way the earthquake motion and the  dynamic response of the slope are modelled. 
The  knowledge  of seismic  forces  makes  it  possible  to  examine  the  stability  of the  embankment  approximately 
using the so-called pseudo-static approach and to establish the deformations that seismic forces  produce. However, 
experience has  shown that pseudo-static  analyses  can be  unretiable  for  soils  that  build up  large  pore pressures  or 
show more  than  15  % degradation of strength due to  earthquake shaking.  Pseudo-static  analyses  produced factors 
of safety well above  1 for a number of dams that later failed during earthquakes. These cases illustrate the inability 
of the pseudo-static methods to  evaluate the seismic stability of slopes. 
Because  of the  difficulty  in  the  assignment  of appropriate  pseudo-static  coefficient,  the  use  of this  approach  has 
decreased.  Methods  based on evaluation of permanent slope  deformation  are  being  used  increasingly  for  seismic 
slope stability analysis. 
120  Canadian Foundation  Engineering  Manual 
6.8.3 Evaluation of Seismic Deformations of Slopes
In practice, the dynamic response ofearth dams and embankments is usually computed using equivalent linear
analyses. These analyses are conducted in tenns oftotal stresses and thus the effects ofthe seismic porewater

behaviour. Therefore, these analyses can only predict the distribution ofaccelerations and shear stresses in the
embankmentand semi-empirical methods areusually usedto estimatethepennanentdeformations andporewater
pressures using the acceleration and stress data (Seed et al. 1975). Adetailedreview ofthesemethods is givenin
Finn(1993). Newmark Sliding Block Analysis
The serviceability ofa slope after an earthquake is controlled by defonnations. Therefore, analyses that predict
slopedisplacementsprovideamoreusefulindicationof seismicslopestability.
Newmarkmethod(Newmark 1965)isthemostcommonapproachusedtopredictseismicslopedisplacement.
Inthis method, thebehaviourofaslope under earthquake-induced accelerations is given by the displacement of
a block resting on an inclined plane (Figure 6.l8a).Ata particular instantoftime, the horizontal acceleration of
theblockwill induce ahorizontal inertial force, khW (Figure 6.l8b).As k/r increases,thedynamicfactor ofsafety
decreases, andtherewillbesomepositivevalueof k" thatwillproduceafactorof safetyof1.0.
Thiscoefficient,termedtheyieldcoefficient,ky' correspondstotheyieldacceleration,a
=kyg. Theyieldcoefficient
is givenby
<p is the angle offriction ofthe slope material (assuming purely frictional soil) and   is the slope angle.
Whenaslope is subjectedto apulseofaccelerationthatexceeds its yield acceleration,itwillundergosome
UsingtheNewmarkapproach,thetotalrelativedisplacement,drel' of theslopecanbegivenby

A istheamplitudeofarectangularpulseaccelerationgreaterthantheyieldaccelerationand/).t isits duration.
Equation 6.51 shows clearly that the total relative displacement depends strongly on both the amount by
Usingtherectangularpulse solution,Newmarkrelatedsingle-pUlse slopedisplacementtopeakbasevelocity, v ,
fJ-a \
d - Y I  (6.52)
,el-2 -A
Newmark found that a reasonable upperbound to the permanent displacements produced by severalearthquake
motionnormalizedtopeakaccelerationsofO.5gandpeakvelocitiesof0.76mlswas givenby
Earthquake. Resistant Design 121,
. ..
= Wcos 
FIGURE 6.18  a) Analogy between potential landslide and a block resting on inclined plane;
b) Forces acting on a block resting on an inclined plane
6.B.3.2 . Nonlinear Analysis 
Nonlinear methods of analysis were also developed to calculate the seismic response of slopes accounting for
the effects of the intrinsic nonlinear behaviour of the soil. Although some of these procedures include elaborate
representation ofthe basic behaviour ofthe soil, their reliability and suitability are limited due to the complexity and
the need for some soil parameters that are not usually measured in field or laboratory testing. Finn (2000) reviewed
the main nonlinear procedures used in current practice and outlined their advantages and limitations.
Seismic DeSign  of Foundation 
The soil-structure interaction effects that take place during the seismic excitation govern the seismic response of
foundations. Except for cases where liquefaction occurred, or sensitive clays lost their strength under cyclic loading,
foundations failures during earthquakes are rare. The strength and stiffness of the foundation elements in regard
to transient dynamic loading are a function of the rate of loading. In general, the stiffness, and for most soils, the
strength, increase with the rate of loading.
Bearing Capacity of Shallow Foundations 
The effect of the inertia forces within the soil mass is to generate shear stresses that would reduce the capacity.
Several studies have shown that the reduction in the bearing capacity due to soil inertia is not more than 15 % to
20 % for k :5 0.3 (Shi and Richards, 1995). Therefore, the main seismic consideration in the design offoundations
would be the effects of eccentric and/or inclined loading conditions due to the induced horizontal inertial seismic
loads from the superstructure.
... . .. 
122  Canadian  Foundation  Engineering  Manual 
To  account  for  the  effects  of horizontal  seismic forces  on  the  bearing capacity  of a footing,  the  resultant inclined 
eccentric load is considered in the calculation of the bearing capacity ofthe footing.  In this case, a reduced effective 
footing  width  and  load  inclination  factors  are  used  in  the  analysis  as  described  in  Chapter  10  of this  manual. 
Because of the short duration of the seismic loads, a smaller factor of safety can be  adopted for the seismic  design 
of foundations. 
6.9.2 Seismic Design of Deep Foundations
The response of deep foundation to earthquake loading is  quite complex. The main factors  that govern the  seismic 
behaviour of deep  foundations  are  the  interactive soil-pile forces  and  the  loss  of the  soil support to  the  piles.  For 
piles in a group, the pile-soil-pile interaction effects add to the complexity of the problem. 
The proper evaluation of the  seismic response characteristics of pile groups requires dynamic analyses that require 
the  use  of computer programs.  The  main  features  that  should  be  considered  in  these  analyses  are  the  nonlinear 
behaviour of the  soil  adjacent to  the piles, the slippage and  separation that occur at the  soil-pile  interface and  the 
energy dissipation through different damping mechanisms. These analyses can be used to  calculate the response of 
the foundation  system to  the  seismic  loading,  and the  capacity of the  foundation  can be evaluated based on some 
ultimate displacement considerations. 
6.9.3 Foundation Provisions
The National Building Code  of Canada,  NBCC  (2005)  includes  the  following  provisions  to  ensure  matching  the 
foundation seismic capacity with the capacity of the seismic force resisting system (SFRS). 
1.  Foundations  shall  be  designed  to  resist  the  lateral  load  capacity  of the  SFRS,  except  that  when  the 
foundations  are  allowed  to  rock,  the  design  forces  need  not  exceed  0.5  RdRo times  those  determined  in 
2.  The design of the foundations  shall be such that they are capable of transferring the earthquake loads and 
effects  between the building and the ground without yielding and without exceeding the capacities of the 
~ o   l and rock. 
3.  For cases where Il.S. (0.2) is  equal to  or greater than 0.2, the following requirements shall be satisfied: 
a.  Piles or pile caps,  drilled piers, and caissons shall be interconnected by continuous ties in not less than 
two directions. 
b.  Piles,  drilled  piers,  and  caissons  shall  be  embedded  a  minimum  of  100 mm  into  the  pile  cap  or 
c.  Piles, drilled piers, and caissons other than wooden piles shall be connected to the pile cap or structure 
for a minimum tension force  equal to 0.15  times the factored compression capacity of the pile. 
4.  At  sites  where  IEF,Sa  (0.2) is  equal  to  or  greater  than  0.35,  basement  walls  shall  be  designed  to  resist 
earthquake lateral pressures from backfill or natural ground. 
5.  At sites where Il.S. (0.2) is  greater than 0.75, the following requirements shall be satisfied: 
1.  A pile,  drilled pier, or caisson shall be designed and detailed to accommodate cyclic inelastic behaviour 
when the design moment in the  element due to earthquake effects  is  greater than 75  % of its  moment 
2.  Spread footings founded on soil defined as Site Class E or F shall be interconnected by continuous ties 
in not less than two directions. 
6.  Each segment of a tie  between elements shall be designed to  carry by tension or compression a horizontal 
force  at  least  equal  to  the  greatest  factored  pile  cap  or  column vertical  load  in  the  elements  it  connects 
mUltiplied  by  a factor  of 0.15  Il.S.(O.2), unless  it  can be  demonstrated that  equivalent restraints  can  be 
provided by other means. 
7.  The potential for  liquefaction and the  consequences,  such as  significant ground  displacements  and loss  of 
soil  strength  and stiffness, shall be evaluated based on Ground Motion Parameters and shall be taken into 
account in the design of the structure and its  foundations. 
Foundation Design 123
Foundation Design 
7 Foundation Design
7.1 Introduction and Design Objectives
The basic purpose of foundations (shallow and deep) is to safely and adequately transfer load effects, from and
acting on any given structure, to the ground. The term ground is general; it includes both soil and rock. Foundation
design essentially involves two basic considerations
The foundation unites) must not collapse (i.e., not induce overall shear failure of the supporting ground);
• Post-construction settlement of the foundation unites) must be within tolerable limits.
As discussed in Chapter 8, the first consideration involves Ultimate Limit States (ULS), and the second consideration
involves Serviceability Limit States (SLS).
The primary objectives of engineering design are safety, serviceability, and economy. Safety and serviceability
can be improved by increasing the design margins or levels of safety to reduce the probability of failure. However,
this generally increases costs. Considerations of overall economy in design involve balancing the increased cost
associated with increased safety (and improved performance) against the potential losses (costs, lives and other
factors) that could result from unsatisfactory performance or failure. The basic design criterion is that the resistance
of the system must be greater than the imposed load effects, while achieving an acceptable or required level of
safety and performance.
7.2 Tolerable Risk and Safety Considerations
Design must assure an acceptable risk or a required level ofsafety; but how does one rationalize what is an acceptable
or tolerable level of risk?
The probability of failure that is associated with a given design needs to be compatible with the level of risk that
people (i.e., society) are willing to accept in specific situations or from natural and constructed works. This is
referred to as tolerable risk. Tolerable risk refers to a willingness to live with a risk so' as to secure certain benefits,
and in the confidence that risk is being properly controlled or managed.
The specified desired level of safety for design is defined by relevant jurisdictional codes of practice (e.g., the
National Building Code of Canada (NBCC), the Canadian Highway Bridge Design Code (CHBDC) and others).
Codes generally describe recommended good engineering design practice by defining a set of requirements, or
provisions, that are aimed at achieving a minimum level of technical quality, and the desired or specified level of
safety. Codes can be viewed as documents for the quality assurance of the design of engineering structures and
facilities. Codes are legal documents and, as such, compliance with the code is required by law. A code represents a
legal means to facilitate sound, rational design decisions to be made by engineers. It assists the engineer in making
124 Canadian Foundation Engineering Manual
the"right"decisionsthatleadto sufficientlysafestructures.Agoodcode doesnotnecessarilyleadtono failures,
butleadstodesignsituationswherethenumberoffailures areacceptableortherisklevelistolerable.
7.3 Uncertainties in Foundation DeSign
Significantandvaryingdegreesof uncertaintyareinherentlyinvolvedinthefoundationdesignprocess.Allowances
1. Uncertaintiesinestimatingtheloadeffects
2. Uncertaintiesassociatedwithinherentvariabilityofthe ground
3. Uncertaintiesinevaluationofgeotechnicalmaterialproperties
4. Uncertaintiesassociatedwiththe degree towhichtheanalysisrepresentstheactualbehaviour/responseof
thefoundation, structure,andthegroundthatsupportsthestructure.
The above uncertainties involve bothstructural andgeotechnical aspects and other considerationsthat contribute
to the overall risk. Standard design philosophies and proceduresgenerallytake uncertainty into accountthrough
the application ofspecified safety factors to manage risk satisfactorily. Inworking (allowable) stress designthis
is handledby the overall global factorofsafety; whereas in limitstates design, theuse ofseparatepartial factors
on loads andresistances are used (referto Chapter 8). Natural groundvariability and evaluationofgeotechnical
withthe depositionandformationofsoilandrock.
Incontrast,grosserrorsincludinghumanerrors oromissionsthatoccurinpracticearenotquantifiedortakeninto
accountthroughsafetyfactorsindesign.Theseerrorsareusuallyhandledby, ormitigatedthrough,qualitycontrol
and quality assuranceprograms, andindependentthirdpartyreviews on largerprojects. It is notedthat gross or
humanerrorsareprobablyresponsibleformostofthe failuresthatoccur.
7.4 Geotechnical Design Process
The geotechnical design process, as it relates to foundation design, is schematically summarized in Figure 7.1.
requirements basedontheclient'sneeds).Abasicdesignissue, fromtheperspectiveofgeotechnicalengineers,is
of technicalquality,engineersrefertoajurisdictionalcodeofpractice.Thepurposeof codesistoassistengineersin
Froman interpretationoftheresultsfromtheinvestigation,geotechnicalengineersformulateageotechnicalmodel
ofthe site in terms ofstratigraphy, soil and groundwaterconditions, and engineeringproperties. Codes andtheir
Thegeotechnicalparameters are dependentonmanyfactors andare subjecttosignificantinherentvariabilityand
uncertainty.Thereisnouniqueanswerto questionsassociatedwiththe shearstrengthanddeformationparameters
Theselectionofcharacteristicdesignvaluesofsoilandrockpropertiesneedsto accountforthefollowingissues:
Foundation Design 125
Extent  or  zone  of influence  in  the  ground  that  contributes  to  overall  behaviour  and  perfonnance  of the 
ground under load effects for possible limit states or failure modes 
•  Effect of construction activities on in-situ ground properties and characteristics 
Influence of workmanship on constructed or improved ground 
Scale effects  and possible differences between the results  of discrete small  sized laboratory and field tests 
relative to the overall ground mass due to factors  such as: 
- presence of fissures, joints and other planeslzones of weakness 
- testing rate effects 
- stress path effects 
- brittleness or ductility (stress-strain response) 
Other factors  considered to  be relevant for the site and project. 
In summary, the selection of the characteristic value for design should appropriately take into account all factors that 
influence the  property or parameter under consideration.  The selection of suitable  characteristic  values,  therefore, 
requires engineering judgment and experience. Additional discussion on characteristic values for design is presented 
in Chapter 8. 
The selection of the procedure used to  detennine ultimate geotechnical resistances will be  influenced by the scope 
of the site investigation and the complexity ofsubsurface conditions at the site. The calculation procedure or design 
equation for geotechnical resistance is usually based on theories of elasticity, plasticity and other relevant theoretical 
frameworks.  In  addition,  ultimate  bearing  capacity  and  many  geotechnical  design  parameters  are  frequently 
selected on the basis ofempirical correlations to in-situ tests such as the Standard Penetration Test (SPT), piezocone 
penetration  test  (CPT),  pressuremeter test  and  other in-situ tests.  These  correlations  involve  inherent uncertainty 
and may be site specific.  Such empirical correlations need to be applied judiciously and with caution.  Some people 
suggest that the geotechnical community should reduce, if not avoid, reliance on these types of correlation models. 
Nevertheless, traditional, empirical correlations are expected to remain in use and will continue to be an integral part 
of design practice for  some time.  This is  because the geotechnical professional  heritage  is  embodied in empirical 
correlati ons. 
A  sound,  basic  design  approach  requires  a  thorough  understanding  of the  key  design  issues,  of the  geological 
setting and  geotechnical  conditions,  and  of the  interaction between them.  In most cases,  a  good understanding of 
these factors  is  as  important, if not more so, as  the analytical/numerical methods used for analysis and calculation. 
It is  important to  initially capture the  essence of the  problem,  and then proceed with  appropriate,  simple analysis 
followed by an increasing level of sophistication and complexity, as required or as the project demands. 
For  the  calculation  model  and  load  effects,  codes  specify  safety  factors  aimed  at  producing  a  design  with  an 
acceptable  risk  or level  of safety.  The  safety factors  specified  help  to  account for  and  to  mitigate uncertainties  in 
the design process, such as  those related to  loads, material properties, design equations, and inherent variability in 
the  ground conditions at the  site.  For large,  complex and  special projects  that involve  a  high  degree  of risk  (e.g., 
long-span  bridge)  a  comprehensive  site  investigation  may  be  able  to  provide sufficient  data for  the  geotechnical 
parameters for  strata at various depths to  be  described in tenns of a mean and standard deviation.  If sufficient data 
are available to describe adequately both loads and resistances, a complete or fully probabilistic method, involving 
reliability theory,  may be used for design and for risk management. 
As  shown in Figure  7.1,  the geotechnical  model  of the  site,  calculation model,  and load  effects are  considered  in 
the geotechnical analysis  of load carrying capacity and settlement of the foundation.  The results from the analysis, 
when appropriately  tempered or modified by engineering judgment and experience,  are  then used in the  decision 
making process as to what constitutes the most appropriate type and size of foundation unit for the building. 
Foundation DeSign Methodology
A detailed flow  chart for  the  design  of foundations  is  shown  on Figure 7.2.  In  many cases,  the flow  chart can be 
simplified depending on the project requirements. However, the figure illustrates the key factors and interaction that 
126  Canadian Foundation Engineering  Manual 
affect the design and selection of the most appropriate choice of foundation for a given site and project.
oDO [100

8 = ?
FIGURE 7.1 Components offoundation design and role ofcodes ofpractice
(after Ovesen 1981, 1993 and Becker 1996a).
An important aspect ofthe flow chart (that is inherent to limit states design methodology) is the distinct and explicit
separate treatment of ultimate and serviceability limit states. Although the traditional working (allowable) stress
design approach also considers both ultimate capacity and settlement, the separation or distinction between them
was not clear or evident. For example, the traditional global factor of safety of three in working (allowable) stress
design often is used to limit settlements to acceptable values, while at the same time to account for uncertainties
associated with applied loads and ultimate geotechnical bearing capacity. The separate and distinct treatment
of ultimate capacity and settlement (serviceability) are key aspects and form the kernel concept of limit states
design that is all too frequently missed, or not well understood or appreciated by foundation engineers. Additional
information and discussion on limit states design is provided in Chapter 8.
- -- - - -- - -- - -
- - - - -- - - -
- - -- - - - - - - - - - --
- - - - -
FoundationDesign 127
regarding proposed
I type, function

i Formulatespecnicationof I
Assemble information
structure: geology,
topography, climatic
Field Investigation
....  - -
Selectionoffoundation  - -
- - --
:.- - :.-=- -=-+ -:. ~ - - - - - - - - - - -,- --I
protection,slopestability, 1
offootingsinplan Factorsaffectingdepth
oflooting: frost

erosion, topography,
., I
soil conditions,water
level,swelling ......
- insitutests
- groundwater
evaluation01 each
t ,
: investigations
I ~   I k } properties
----+---- -
- -
- .--- -
... I
.J.. I
...l ___
ICheckwith predicted
pertormance design

r - -
I yes

CheckU bearingcapacity
...... no
: footing
- obServerock/soii
frequentlygovern conditions
. on eachlooling
1 no
footings isexcessive
predictedperformance I
- monitorbehaviour
foundation design
yes I
yes _,.__
-- -- --- ..
- dewatering
FIGURE 7.2 Flow diagram for design offoundations (from NBCC (2005) - User sGuide)
•  Appropriate field and laboratory investigation to define the geotechnical model and characteristic design
128 Canadian Foundation Engineering Manual
performance.  The  key  geotechnical  ultimate  limit  states  (ULS)  are  bearing  capacity,  sliding,  overturning, 
uplift  and  excessive  foundation  deformation that  would  cause  a  ULS  condition  to  occur in  the  structure. 
For  serviceability  limit  states  (SLS),  the  main  consideration  is  deformation  (in  terms  of settlement  and 
horizontal displacement, vibration effects and others). 
Checking (through appropriate analysis)  of each  identified limit state to  ensure that they  either would not 
exist, or are within acceptable levels of risk (probability of occurrence). 
7.6 Role of Engineering Judgment and Experience
Engineering judgment and experience are,  and always  will be,  an  essential part of geotechnical engineering;  they 
are vital for managing safety (risk) of geotechnical structures. There will always  be a need for judgment, tempered 
by experience, to be applied to new technologies and tools.  Many aspects of geotechnical design are heavily reliant 
on engineering judgment and experience. 
The spirit of the  limit states design concept, as it was originally conceived, is particularly important in geotechnical 
engineering. The proper identification of potential modes of failure  or limit states of a foundation,  which is the first 
step in design, is not always a trivial task. This step generally requires a thorough understanding and appreciation of 
the interaction between the geological environment, loading characteristics, and foundation behaviour. 
Reasonable analyses can be made using relatively simple models if the  essence of geotechnical behavior and soil-
structure interaction is captured in such models. There must also be a sufficient data and experience base to calibrate 
these  models.  Empirically  based  models  are  only  applicable  within  the  range  of specific  conditions  reflected  or 
included  in  the  calibration  process.  Extrapolation  beyond  these  conditions  can  potentially  result  in  erroneous 
predictions of performance. 
In summary,  engineering judgment and experience play  an  integral role in geotechnical engineering analysis  and 
design.  Uncertainties  in  loads,  material strengths  (resistance),  models,  identification of potential failure  modes  or 
limit states, and geotechnical predictions all need to be considered collectively in controlling or ensuring an adequate 
level of safety in the design. The role of the geotechnical engineer through his or her judgment and experience, and 
that of others, in appreciating the complexities of geotechnical behavior and recognizing the inherent limitations in 
geotechnical models and theories  is  of considerable importance. The management of safety (risk)  in  geotechnical 
engineering design is distributed amongst the many aspects of the overall design process, including experience and 
7.7 Interaction Between Structural and Geotechnical Engineers
Geotechnical resistance  and reaction are  a coupled function  of applicable  geotechnical parameters and of applied 
loading  effects.  Consequently,  close  and  effective  communication  and  design  interaction  between  structural  and 
geotechnical engineers need to  take place to assure compatibility with the various design criteria, and achievement 
of desired  performance  and  economy.  Although  this  interaction  and  effective  communication  should  occur  for 
all  classes  of problems,  it  is  especially  important,  if not  essential,  for  more  complex  soil-structure  interaction 
considerations  where  the  design  procedure  involves,  or  is  based on  a  modulus  of sub grade  reaction  (vertical  or 
horizontal).  Examples  include  horizontal  deformation  and  capacity  of piles,  retaining  walls  and  raft/floor  slab 
foundations. Additional discussion is presented in Section 7.7.1. 
Some  codes,  such  as  the  Canadian  Highway  Bridge  Design  Code  (CHBDC),  formally  require  that  appropriate 
design interaction and communication occur between geotechnical and structural engineers.  This legal requirement 
of such design interaction is  an important precedent and step towards safe, economical design of foundations. 
7.7.1 Raft Design and Modulus of Subgrade Reaction
In  the  design  of a  raft  foundation,  structural  engineers  usually  ask  for  the  value  of the  coefficient  (modulus)  of 
sub grade  reaction  of the  supporting soil.  Because of local  variations  in  soil type  under  the  raft,  disturbances  that 
Foundation Design 129
take place during excavation and placement of steel reinforcing, and limitations of the theory, only approximate
indications of the magnitude of the coefficient of sub grade reaction can be given. In addition, because the stresses
from the raft affect the soil to considerable depth below bearing level, longer-term consolidation settlements may
develop; these settlements also may vary, depending on the differences in soil compressibility existing at different
points under the raft. Such considerations need to be taken into account by the geotechnical engineer when assessing
appropriate values for sub grade reaction.
Unlike strength and compressibility, the modulus of sub grade reaction is not a fundamental soil property. Rather,
it is a common design approach used by structural engineers to model the interface between the foundation soil
and concrete footing (Le., soil-structure interaction). The modulus of sub grade reaction is a number required by
structural engineers' to model the deformation and stiffness response of a footing (raft) on soil. The modulus of
sub grade reaction is deflned as:
k=q/8 (7.1 )
k modulus of sub grade reaction
q applied bearing or contact pressure on footing
8 settlement of footing under applied pressure q
The modulus of subgrade reaction, though simple in its definition, is a very difficult parameter to evaluate properly
because it is not a unique fundamental property that is readily measured. Its value depends on many factors, including
size and shape of footing (raft), type of soil, relative stiffness offooting and soil, duration of loading relative to the
hydraulic conductivity of the loaded soil, and others. The value of modulus of sub grade reaction can also vary from,
one point to another beneath a footing or raft (e.g., centre, edge or comer) and can change with time, in particular
for soils with low hydraulic conductivity such as clays.
Field plate load tests are commonly used to determine numerical values for the modulus of subgrade reaction. A
database ofn1,lmerical values and types of soil has been developed. Because the modulus value can change with size
of footing, a one foot (300 mm) square footing has been adopted as the standard basis for comparison purposes,
and frequently serves as the starting point for design. The technical literature cites typical values for the modulus of
subgrade reaction, kyl' (for   square plate) for a variety of soil types. Typical ranges in kYl are summarized
in Table 7.1. Appropriate design values for modulus ofsub grade reaction generally decrease ifthe size of the loaded
plate (or footing) is larger than one foot (30Q.I'lun) by one foot (300 mm). The manner in which the value of modulus
of subgrade reaction decreases with increasing footing size varies with the type of foundation soiL Additional
information is provided below, as well as by Terzaghi (1955), NavFac (1982) and Winterkom and Fang (1975).
TABLE  7.1  Typical Ranges In Vertical Modulus OfSub grade Reaction
Soil Type
Granular Soils (Moist or Dry) (2)
kV1 (MPa/m) (1)
Loose 5 - 20
Compact Sand 1 20-60
Dense. 60 -160
Very Dense
160 300 (3)
Cohesive Soils
Soft <5
Firm 5 -10
10-30 Stiff
30 - 80 Very Stiff
80 200 (3)
130  Canadian  Foundation  Engineering Manual 
1.  For a 1  (300 mm) x  1 ft.  (300 mm) plate 
For granular soils, kVb  = ky!  (\:lr 
For strip footing  on cohesive soil, kYb  = kjb 
If the  loaded area on cohesive soil is of width b and length mb, kYb 
b  l.5rn 
kYb  modulus for actual footing  dimension b 
b  =  foundation width 
When using the above expressions, care must be taken to  ensure that the units are consistent.  These equations were 
initially  derived for  b in  units  of feet.  Therefore,  when  using b  in  meters,  the  expression (b+ 1)  needs  to  become 
(b+O.3)  and (m+O.S)  becomes (m+0.15). 
2.  If below groundwater table, these values should be multiplied by 0.6. 
3.  Higher values to be used only if assessed on basis of adequate test results and settlement calculation. 
Values for modulus ofsubgrade reaction can be derived from the results ofplate load tests using elastic displacement 
theory as  represented by: 
I  an influence factor that is dependent on geometry of footing and thickness of compressible soil relative 
to footing width 
b  width of footing 
v  =  Poisson's ratio (v = 0.5 for undrained condition and typically about 0.3 for drained conditions for most 
soils)  . 
E  =  modulus of deformation (Eu  if examining undrained condition and E' for  drained condition) 
Rearrangement of Equation (7.2) gives 
Therefore, ifvalues ofE are known for the soil within the zone of influence, beneath a footing of width b, reasonable 
estimates  can  be  made  for  the  modulus  of sub grade  reaction,  k
,  using  Equation  7.3.  Typical  values  for  E  are 
provided in references such as Bowles (1988), NavFac (1982) and many others. 
It is generally considered that the use of settlement calculation is  a more  rational method of assessing modulus  of 
sub grade reaction than is the use of adjusting typical values ofkV! for a one-foot square plate.  The value of modulus 
of sub grade reaction for the footing or raft under consideration is the applied pressure at a given location divided by 
the settlement calculated at that location for the applied pressure (i.e., k=q/8). 
It is  emphasized  that values  of kYb  as  determined  from  extrapolation of plate  bearing tests  or  from  kVI  should be 
used with judgment and care.  The  deformation response  of a  smaller sized plate may not be  representative of the 
response  of the  larger sized  actual  foundation  because  the  zone  of influence  extends much  deeper for  the  actual 
foundation. This aspect is especially important in ground with variable stratigraphy and engineering properties with 
depth,  in particular for the  case  of softer soil  at  depth  to  which the  zone  of influence  for  a  small  plate  would not 
Foundation Design 131
extend. The results from the test plate would not reflect the response of the soft layer at depth. Further, the results
from plate load tests on clays and clayey silts may be unreliable because the time associated with the testing may not
permit complete consolidation (drainage of excess porewater pressure) of these fine-grained soils. An assessment
of whether an undrained or drained condition prevailed during the test must always be made. For design, the test
results obtained would need to be adjusted (corrected) as appropriate.

132 Canadian Foundation Engineering Manual
Limit States and Limit States Design

Limit States and Limit States Design
8.1 Introduction

The geotechnical engineering profession in Canada and elsewhere throughout the world is in the process ofevaluating 
and incorporating limit states design (LSD) into codes ofpractice for geotechnical aspects offoundation engineering. 
A benefit  of LSD  for  geotechnical  aspects  of foundation  design  is  that  it provides  a  consistent  design  approach! 
philosophy  between  structural  and  geotechnical  engineers.  Information  on  the  background  and  development  of 
LSD for structures and for foundations is provided by Allen (1975), MacGregor (1976), Meyerhof(1982, 1984 and 
1995), Duncan et al.  (1989), Green (1989,  1991  and  1993),  Ovesen and Orr (1991), Becker (1996a and b),  Green 
and  Becker  (2000)  and  Becker  (2003).  In  addition,  the  proceedings  of international  workshops  and  symposia, 
including DGS  (1993), LSD 2000 and IWS  Kamakura 2002 provide substantial information and discussion. 
To date (i.e., early 2000's) geotechnical engineering practice largely continues to use traditional working (allowable) 
stress (WSD) design for foundations. However, most structural design is carried out using LSD concepts.  Therefore, 
a significant degree of inconsistency exists  in the design interaction between structural and geotechnical engineers, 
which could lead to different levels ofsafety and to errors. There is no basic reason why limit states design principles 
cannot be applied to the design offoundations. Ground (soil or rock) can be treated as an engineering material, albeit 
one  that  may  exhibit considerable  variability  and  deformability.  Models  can  be  developed  to  show  how  ground 
resists  forces  and  deformations,  and  how  ground  can  induce  load  on  structures.  The  principles  of engineering 
mechanics and of deformable bodies can be applied in conjunction with analytical procedures to  analyse foundation 
units for serviceability and ultimate limit sates. 
Both structural and geotechnical  engineers  have the common mandate of achieving a specified level of safety and 
minimizing repair and loss  of function  during the  life  of a  structure.  The  design should also  be  efficient  from  an 
economic viewpoint. Economic advantage can be realized if all members-components of the structure are designed 
to a consistent appropriate level ofsafety or reliability. This objective is enhanced ifboth geotechnical and structural 
aspects  of foundation  design are based on the same design approach and concepts. Therefore, a strong motivation 
for  the use  of LSD  in  foundation  engineering  is  the need, benefit and  importance of a  compatible design process 
between  structural  and  geotechnical  engineers.  However,  there  are  important  technical  benefits  associated  with 
the use  of LSD  for  geotechnical aspects  of foundation  design.  LSD  has significant merit and advantages  over the 
traditional  WSD  approach  for  foundation  design  (Becker  1996a).  LSD  can  be  viewed  as  a  logical  extension  to 
WSD. It is  considered that LSD will eventually become the general state ofpractice by geotechnical engineers  for 
foundation design. 
To  date,  some existing  Canadian Codes such as  the Canadian Highway Bridge Design Code (CHBDC 2000), the 
Ontario  Highway Bridge Design Code (OHBDC  1983  and  1992), the National Building Code of Canada (NBCC 
2005), the Canadian Standard Association (CSA) S472 Standard for Foundations in the Offshore Code (CSA 1992) 
have introduced or require LSD  for  foundations.  Green and Becker (2000) and  Becker (2003)  provide a  status  of 
LSD in Canada for geotechnical engineering design practice. 
Limit States and Limit States Design 133
8.2 What Are Limit States?
Limitstates are definedas conditions underwhich astructure orits componentmembers no longerperformtheir
intendedfunction.Wheneverastructureorpartof astructurefailstosatisfyoneof itsintendedperformancecriteria
itissaidtohavereachedalimitstate.Alimitstateisassociatedwithunsatisfactoryperformance. '
Ultimatelimitstates (ULS)areprimarilyconcernedwithcollapsemechanisms ofthe structureand, hence, safety.
exceeding the load carrying ability ofthe ground that supports the foundation (i.e., ultimate bearing
largedeformationofthe foundationsubgradethatleadstoanulsbeinginducedinthestructure,and
lossof overallstability.
Because oftheir relationship to safety, ULS conditions are designed for a low probability ofoccurrence that is
consistentwiththe desiredorspecifiedlevelofsafetyandreliability.
Serviceability limit states (SLS) represent conditions or mechanisms that restrict or constrain the intended use,
function oroccupancy ofthe structure under expected service orworking loads. SLS areusuallyassociatedwith
movementsordeformationsthatinterruptorhinderthefunction(i.e.,serviceability)ofthestructure. Forfoundation
•  excessivemovements(e.g., settlement,differentialsettlement,heave,lateralmovement,cracking,tilt),
unacceptablevibrations, and
•  localdamageanddeterioration.
SLShaveamuchhigherlikelihoodorprobabilityofoccurrencethanULS. SLSmaybeviewedasthosethingsthat
ThedistinctionbetweenULSandSLSmaybebetterappreciatedbythefollowingexample. Abuildingthatdoesnot
collapseunderspecifiedloading hasperformedsatisfactorily againstanULS condition. However, ifdeformation
has occurredto the extentthatthe ownercannot open doors to the buildingorifthefloor andwalls are severely
damagetothebuildingno longerallowsit to performitsintendeduseoroccupancy(serviceability).
Allowable movements offoundations and structures depend on soil-structure interaction, desired serviceability,
harmful cracking, vibration, and distortion restricting the use ofa given structure. Empirical damage criteria
are generally related to relative rotation (i.e., angular distortion, deflection ratio,·or tilt ofthe structure). For
superstructures,thesecriteriadifferfor framebuildings(bareorc1added), load-bearingwalls(saggingorhogging),
isprovidedbyBurlandetaL (1977)andFeld(1965).
Forcommontypes ofbuildingsandforsomeothertypesof engineeringstructures,tentative safelimitshavebeen
suggestedas aguide (Bjerrum 1963 andMeyerhof1982).Appropriate guides are also given in otherparts ofthis
Manual. However, theseguidelines shouldnotstandinthewayofdirectcommunicationandinteractionbetween
. ...
134 Canadian Foundation Engineering Manual •. 

The  loads that are  applied to  a foundation consist of pennanent (dead)  and transient (e.g., live,  snow, wind)  loads.  1 
The  full  values  of live  (transient)  load  effects  do  not  necessarily  need  to  be  used  in  a  calculation  or  analysis,  of  j
magnitude of foundation settlement. Full or complete values of permanent load effects always need to  be included; 
however, whether the total magnitude of live (transient) load effects needs to be used depends on the consolidation 
characteristics of the soils that exist within the zone of influent below the foundation. 
For  cohesionless  soils,  settlement  estimates  should  be  based  on  the  maximum  (dead  and  live)  loads  with  an 
allowance for any  dynamic effects. For fine-grained soils that have relatively low rates of consolidation settlement, 
the duration of the transient load effects is usually not sufficient for a substantial portion of consolidation settlement 
to take  place under transient loading.  In  these cases,  ignoring the  transient  load effects or using only a proportion 
of the total transient load in  a settlement analysis may be appropriate. The appropriate proportion of total transient 
load  effects  for  a  given  circumstances depends  primarily  on the  duration of the  applied  transient  loading  relative 
to  the  coefficient  of consolidation  of the  foundation  soils.  Although  relevant  Codes  of Practice  may  specify  or 
provide guidance as  to  suitable proportions for  use  in  settlement analyses,  this  task is usually  left to the discretion 
of geotechnical engineers.  Settlement estimates  for  cohesive  soils therefore,  could be based on dead loads, plus  a 
reduced load for  live and other transient loads. 
The  effects  of elastic  displacements,  shear  distortions  and  permanent  hysteresis  effects  that  may  be  induced  by 
transient loading effects should be considered and included in settlement analysis, as  appropriate. 
8.3 Limit States Design (LSD)
In essence, limit states design (LSD) involves the identification of all possible limit states or "failure" mechanisms, 
and the  subsequent checking that  the  probability or likelihood of occurrence  of each  limit state  identified will be 
within  an acceptable or specified level of safety or reliability.  The term "failure" is used here in the  general sense 
of unsatisfactory performance. It does not necessarily mean rupture or collapse. The applicable,  acceptable level of 
safety or reliability is usually defined by the target reliability index that is  specified by governing codes. 
Each  potential  limit  state  identified  is  considered  separately,  and  through  the  design  process  its  occurrence  is 
demonstrated to  be sufficiently improbable (or eliminated) or to be acceptable. 
ULS conditions are checked using separate, partial factors on loads and On nominal (ultimate) geotechnical resistance. 
The values of these partial factors  are  specified by applicable  codes and manuals (guidelines)  for state-of-practice. 
The magnitudes of the partial factors  are usually based on calibration to WSD (including engineering judgment) or 
on reliability theory,  or a combination of both (Becker  1996a and b, Green and Becker 2000, Kulhawy and Phoon 
2002,  and Phoon et al.  2003). The magnitude of the  specified partial factors  serve as  a means of risk management 
towards achieving the desired or target level of safety/reliability. 
The SLS conditions are checked using working or service loads and unfactored geotechnical properties. In essence, a 
partial factor ofone is used on all specified loads and on the characteristic values for deformation and compressibility 
properties of the ground. Geotechnical characteristic values are generally based on conservative ( cautious estimate) 
mean values obtained from in-situ and laboratory tests. In this sense, the methodology of SLS calculation is virtually 
identical between LSD and WSD approaches. 
The  explicit  distinction between safety (ultimate)  and  deformation  (serviceability)  analyses/calculations,  and the 
classification of performance that flows  from this distinction, reflect the kernel concept of limit states design.  This 
distinct and explicit separate treatment of ULS  and SLS  is  the essence  and most important fundamental aspect of 
limit states design. 
Although the traditional working (allowable) stress design approach considers ultimate capacity and settlement, the 
separation or distinction between them was not clear or evident. For example, the traditional global factor ofsafety 
of three  in  working  (allowable)  stress  design  often  is  used to  limit settlements to  acceptable values,  while  at the 
same time to account for uncertainties associated with applied loads and ultimate geotechnical resistance ( capacity). 
Limit States and Limit States Design 135
The separate, distinct treatment of ultimate capacity and settlement (serviceability) is the key aspect of limit states
design that is all too fi'equently missed, or not well understood or appreciated by geotechnical engineers.
The historical development of geotechnical LSD has been described and summarized by Ovesen and Orr (1991),
Meyerhof (1995), Becker (1 996a) and others. The approaches to LSD have developed differently in NOlth America
and in Europe, mainly in the manner for calculating factored geotechnical resistances at ULS.
In the factored strength (European) approach, specified partial factors are applied directly to the geotechnical strength
parameters of cohesion and angle of internal friction. The resulting factored strength parameters are then used in
traditional equations/formulae for the direct calculation of factored geotechnical resistance at ULS for design. This
is the approach advocated and required by Eurocode 7 (ENV 1991, 1994, 1997, Eurocode 7 (1987 and 1990)).
In North America, a factored resistance methodology, such as load and resistance factor design (LRFD), has become
the standard approach. In this method, an overall specified resistance factor is applied to the calculated or assessed
ultimate geotechnical resistance for each applicable limit state. The ultimate resistance is firstly calculated from
"rear' or unfactored ( characteristic) strength parameters using traditional equations or formulae; the calculated
ultimate resistance is then multiplied by a single, specified geotechnical resistance factor to obtain the factored
geotechnical resistance at ULS for design.
Figure 8.1 summarizes the comparison of these two LSD approaches. The advantages and disadvantages of the two
approaches are a subject of debate by geotechnical engineers throughout the world. The interested reader is referred
to Becker (l996a) for a detailed discussion. For the purposes ofthis manual, the factored resistance approach is used
because, as stated in Chapter 7, it forms the basis of many existing codes of practice currently in use in Canada and
the United States.
It is noted that this LSD approach does not alter the methods for calculating ultimate geotechnical resistance
(capacity). The calculations are performed according to the same traditional and classical methods that are familiar
to all geotechnical engineers using working (allowable) stress (WSD) design. The key difference is the manner in
which the design value is obtained and used. In WSD, a single global factor of safety is used; whereas in LRFD,
several partial load and' resistance factors are employed. The only difference in the execution of calculation for LSD
design values is that the ultimate geotechnical resistance (capacity) is multiplied by a different (both in terms of
rationale and magnitude) factor of safety.
(factored strength approach)
...  .... 
Factored Factored
Unfactored (I R d d) Resistance
"""II- Characteristic
strength .e. e uce > (I.e. Increased)
---II-- Strength ----JIIo-- for DesIgn, Load Effects, S
Load Effects,
Parameters Po t Rd
rame ers,
Ic, q, I ltc, t [email protected]]1] .... 
WHERE   C f / ~ f ) <   C / ~ )
_ S
136 CanadianFoundation Engineering Manual
... ...
Strength .......... (nominal) (i.e..Reduced) >
(i.e. Increased)
Unfactored Unfactored  
Load Effects
(e $) Rn for Design
Parameters Resistance. Resistance
- - <fiR - _ n
@IJ .. X FACTOR, .......... 
- - - Sn
FIGURE 8.1 Comparison oflimit states design approaches for ultimate limit states
(after Ovensen and Orr, 1991; Becker, 1996a)
LSD BasedonLoadandResistanceFactorDesign(LRFD)
Significant and varying degrees ofuncertainty are inherently involved in foundation and other geotechnical design. 
Therefore in recent years, there as been a trend towards the use ofreliability-based design and probabilistic methods 
in  geotechnical  engineering  design.  However,  complete  probabilistic  design  is  difficult  to  apply  reliably  and 
appropriately, in particular in most practical geotechnical design situations, generally because of lack ofstatistically 
viable  information.  Complete probabilistic  methods  are  also  time-consuming  and expensive,  which  makes  them 
practical  or  suitable  for  large,  special  projects  only.  Because  of these  difficulties,  simpler,  yet probabilistically 
baseq design procedures have  developed.  LRFD  is  an example where the partial factors  have been based on 
or calibrated using probability and reliability concepts.  For the consideration of ultimate  limit states,  the  separate 
consideration of loads, materials and performance provides the opportunity for the design to be more responsive to 
the  differences between types  of loads, material types,  fundamental  behaviour of the  structure  and of the  ground, 
and consequence of different modes of unsatisfactorily performance (i.e., limit states). 
The basic design equation is: 

is the factored geotechnical resistance 
<D  is the geotechnical resistance factor 

is  the nominal  (ultimate) geotechnical resistance determined through engineering analyses 
(e.g., bearing capacity) using characteristic (unfactored) values for geotechnical parameters 
or performance  data  (e.g.,  pile  load  test);  it  represents  the  geotechnical  engineer's  best 
estimate  of resistance,  that  has  appropriately  taken  into  account  all  factors  influencing 
is  the  summation  of the  factored  overall  load  effects  for  a  given  load  combination 

is  the  load factor corresponding to a particular load; it accounts for uncertainties in loads 
Limit Slales and Limit States Design 137
is  a  specified  load component of the  overall  load effects  (e.g.,  dead load due to weight of
structure or live load due to wind); and 
represents various types of loads such as  dead load, live load, wind load, etc. 
The values for load factors (a), geotechnical resistance factors (<D)  and load combinations are specified by applicable 
codes (e.g., NBCC, CHBDC, AASHTO, etc.). 
The  load  factors,  a, are  usually  greater  than  one;  they  account  for  uncertainties  in  loads  and their probability  of 
occurrence.  The  resistance  factors  (or  performance  factors  as  they  are  sometimes  called),  <D,  are  less  than  one 
and account for  variabilities  in geotechnical  parameters  and analysis uncertainties  when  calculating  geotechnical 
ultimate (nominal) resistances. 
The design equation can be visualized by inspecting the interaction ofthe probability distribution curves for resistance 
and  load effects,  as  shown schematically on Figure 8.2.  It should be noted that the resistance  and load effects  are 
assumed to  be independent variables, which is  approximately true for the case of static  loading.  The characteristic 
or nominal values for load effects (S) and resistance (R,,)  do not necessarily need to be taken as the mean values of 
Sand R, respectively.  The nominal or characteristic values for design are related to the mean values as follows: 
~ = Rand 
Sn=  S 
kR  is  the ratio of mean value to nominal (characteristic) value for resistance; and 
ks  is  the ratio of mean value to specified (characteristic) value for load effects. 
The factors ofkR and ks  are used to define characteristic values of design based on the mean values ofthe resistance 
and load distribution curves, respectively. Typically, kR  values are equal to or greater than one (i.e., Rn ~ R) and kg 
values are less one (Le.,  Sn ;;::S). The terms kR  and kg  are  also referred to  as  bias factors by some researchers. The 
bias factor is  one if the mean value is used as the characteristic value. 
Sn= S/ks 
  S ) ~
S Sn  Rn  R
LSD FORMAT: cD  R n ;;;: IX  Sn
FIGURE 8.2 Load and resistance factor design (LRFD)
138 Canadian Foundation Engineering Manual
In practice, values for a and <I> are specified in codes. They are based on target values of reliability or acceptable
probabilities of failure selected to be consistent with the current state-of-practice. In general, different values of a
and <I> are provided for different limit states. Although values of a may differ between codes in various countries,
load factors are typically in the range of 0.85 to 1.3 for dead loads and in the range of 1.5 to 2.0 for live and
environmental loads. A load factor ofless than 1.0 for dead loads is used when the dead load component contributes
to the resistance against overturning, uplift, and sliding. Typical values of<D range from about 0.3 to 0.9, depending on
ground type, method of calculating resistance, and class of structure such as foundation type or retaining structure.
a.s Characteristic Value
It is important to note that the load and resistance factors are interrelated to each other. That is, the value of 0:. is
dependent on the value of <1>, and vice versa. The values of a and <I> depend on the target reliability index for design,
the variability of the parameters that affect loads and resistances, and the definition of their characteristic values.
Load and resistance factors have been derived (calibrated) based on characteristic values that have been defined in a
specific manner. Therefore, consistent sets of these factors must be used in design as per their intended purpose and
specific evaluation. It is inappropriate to use a set ofresistance factors (that have been derived for specific values
ofload factors) and directly use them with other load factors that have been taken from an unrelated source, or vice
versa. For consistent and rational design in practice, the selection of a given characteristic value for geotechnical
resistance needs to be made in the same manner as that used to derive the specified resistance factor. That is, if the
mean value was used in the derivation of the resistance factor, the mean value of a given geotechnical property
should be used in the calculation of geotechnical resistance. The use ofthe mean value or a value slightly different
from the mean is frequently used in reliability analysis for the determination (calibration) of load and geotechnical
resistance factors.
The key statistical parameters (i.e., the ratio ofthe mean value to characteristic value and the coefficient ofvariation)
for geotechnical resistance depend on many factors, including site investigation method, quality and quantity of
testing (laboratory and in the field), construction quality control, and method of analysis.
The selection ofnominal or characteristic strength for design varies with local state-of-practice and with the training,
intuition; background, and experience ofthe individual geotechnical engineer. Frequently, the mean value or a value
slightly less than the mean is selected by geotechnical engineers as the characteristic value for design purposes.
Eurocode 7 proposes a "cautious estimate" of the mean value for the characteristic value.
The geotechnical engineer selects representative (characteristic) geotechnical parameters based on the results of
appropriate investigations (field and laboratory). Representative in this sense refers to the geotechnical engineer's
best estimate of the likely values of parameters required for design. As discussed in Chapter 7 (Section 7.4), the
selection of the characteristic value, for a given design situation, should appropriately take into account all factors
that have influence pn the parameter or property for the volume of ground (zone of influence) under consideration.
The selection of appropriate characteristic values is assisted by engineering judgment and experience. In addition
and as mentioned above, the geotechnical engineer should be cognizant of the interrelationship between resistance
and load factors and characteristic value when selecting characteristic geotechnical parameters for design purposes.
A cautious estimate of the mean value for the affected volume of ground (zone of influence) is generally considered
to be a logical value to use for the characteristic value.
Recommended Values for Geotechnical Resistance Factors
The recommended resistance factors are specified in applicable codes and manuals ofpractice. Although the values
recommended by various codes tend to be similar, there are some specific differences. For example, the values in
the NBCC (2005) and CHBDC (2000) are shown in Table 8.1 and Table 8.2, respectively. The reliability index
associated with these resistance factors ranges from 2.8 to 3.5, a range that is generally consistent with values
commonly specified for the design of structures.
Limit States and Limit States Design 139
TABLE 8.1 Geotechnical Resistance Factors jor Shallow and Deep Foundations  NBCC (2005) 
Shallow foundation
Vertical  bearing  resistance  from  semi-empirical  analysis  using  laboratory  and 
in-situ test data 
(i)  based on friction  (c  0) 
(ii)  based on cohesion/adhesion (tan   0) 
Deep foundation
Resistance to  axial load 
(i)  semi-empirical analysis using laboratory and in-situ test data 
(ii)  analysis using static loading test results 
(iii)  analysis using dynamic monitoring results 
(iv)  uplift resistance by semi-empirical analysis 
(v)  uplift resistance using loading test results 
Horizontal load resistance 
TABLE 8.2 Geotechnical Resistance Factors 
Shallow Foundations
Bearing Resistance 
Passive Resistance 
Horizontal Resistance (Sliding) 
Ground Anchors (Soil or Rock)
Static Analysis Tension 
Static Test Tension 
Deep Foundations - Piles
Static Analysis  Compression 
Static Test  Compression 
Dynamic Analysis Compression 
CHBDC (2000) 
.  0.4 
Factor, (J)
Dynamic Test Compression (field measurement and analysis)  0.5 
Horizontal Passive Resistance  0.5 
The AASHTO  Code  (1997  and  1998)  specifies  many  more  resistance  factors  than  is  provided  by  CHBDC  and 
NBCC.  For each class offoundation, AASHTO specifies resistance values that are to be used for different methods 
of calculation  and  geotechnical  data.  For example,  a  different value  is  given  if the  geotechnical  data  is  based on 
Standard Penetration Testing (SPT),  Pieco-cone  Penetration Testing  (CPT),  or laboratory testing. As  a result,  the 
number of specified resistance  factors  in the  AASHTO  Code  exceeds  that  of CHBDC by  more  than  an  order  of 
. .... 
140 Canadian Foundation Engineering Manual
Although there is a merit in what the AASHTO Code has done, the approach for both the CHBDC and NBCC was
to keep the process simple, at least during the initial stages of transition between working stress design and limit
states design. For the NBCC and CHBDC, it was felt that it is more important that the fundamental principles of
limit states design for foundations be conveyed to and understood by geotechnical practitioners, as well as structural
engineers designing the foundations. The initial transition should be as gradual and smooth as possible. Providing
a myriad of partial factors that cover a large range of methods used in practice may not be conducive to better
understanding and acceptance of the design method by geotechnical engineers, who are accustomed to using only a
few values of global factor of safety. Refinements and level of sophistication and details can come later when more
experience with limit states design for foundations has been gained. In addition, the existing geotechnical database
in terms of bias factor, coefficient of variation and other statistical parameters need to be further developed and
better understood before levels of refinement such as those included in AASHTO can be reliably developed for
Canadian codes.
8.7 Terminology and Calculation Examples
The various codes tend to use slightly different terminology for LSD design values. When designing based on a
given code, the geotechnical engineer needs to be cognizant of the specific terms and definitions that are specified
by that code. For example, the NBCC Commentary L Foundations (2005) uses the following terms for expressing
recommended geotechnical criteria for the design of the building structure, including its foundations.
Bearing pressure for settlement means the bearing pressure beyond which the specified serviceability criteria are
no longer satisfied.
Factored bearing resistance means the calculated ultimate (nominal) bearing resistance, obtained using characteristic
ground parameters, multiplied by the recommended geotechnical resistance factor.
Factored sliding resistance means the calculated ultimate (nominal) sliding resistance, obtained using characteristic
ground parameters, multiplied by the recommended geotechnical resistance factor.
Factoredpull out resistance (i.e., against uplift) means the calculated ultimate (nominal) pull out resistance, obtained
using characteristic ground parameters, multiplied by the recommended geotechnical resistance factor.
CHBDC (2000) uses the following definitions.
Factored Geotechnical Resistance at ULS the product of the geotechnical resistance factor and the geotechnical
ultimate (nominal) soil or rock resistance.
Geotechnical Reaction at SLS the reaction of the soil or rock at the deformation associated with a SLS
Geotechnical Resistance at ULS - the geotechnical ultimate resistance of soil or rock corresponding to a failure
mechanism (limit state) predicted from theoretical analysis using unfactored geotechnical parameters obtained from
test or estimated from assessed values.
8.7.1 Calculation Examples
The following two examples demonstrate the simple calculation ofdesign values for factored geotechnical resistance
The basic equation for factored geotechnical resistance is <D R where <D is the geotechnical resistance factor and R
n n
is the ultimate (nominal) geotechnical resistance.
Limit States and Limit States Design 141
Shallow Foundation
An ultimate bearing capacity of 800 kPa has been calculated using the classical bearing capacity equation. For
LRFD, the factored bearing resistance at ULS is 400 kPa (Le., 0.5 x 800, where cD = 0.5 from Table 8.1).
Deep Foundation
A static pile load test has shown an ultimate axial capacity of 2,500 kN. The factored axial geotechnical resistance
at ULS is 1,500 kN (i.e., 0.6 x 2,500, where cD = 0.6 from Table 8.1).
Working Stress Design and Global Factors of Safety
Working stress design (WSD) was one of the first rational design methods used in civil engineering. It has been the
traditional design basis since it was first introduced in the early 1800's. WSD is also referred to as allowable stress
or permissible stress design. The basis of the design is to ensure that throughout the structure, when it is SUbjected
to the working or service applied load, the induced stresses are less than the allowable stresses. A single, global
factor of safety is used, which collectively considers or lumps all uncertainty associated with the design process
into a single value, with no distinction made as to whether it is applied to material strength and resistances or to
load effects.
The assessment of the level of safety of the structure is made on the basis of global factors of safety, that were
developed from previous experience with similar structures in similar environments or under similar conditions.
The values of the global factor of safety selected for design reflect past experience and the consequence of failure.
The more serious the consequence of failure or the higher the uncertainty, the higher the global factor of safety.
Similar values of global factor of safety became customary for geotechnical design throughout the world. The
ranges of customary global factors of safety are shown in Table 8.3.
TABLE 8.3 Ranges ofGlobal Factor ofSafety Commonly Usedfor Foundation Design
Failure Type Item Factor of Safety, FS
Shearing Earthworks 1.3 to 1.5
Earth retaining structures, excavations 1.5 to 2
Foundations 2 t03
Seepage Uplift heave 1.5 to 2
Exit gradient, piping 2 to 3
Ultimate pile Load tests 1.5 to 2.0
Loads Dynamic formulae 3
Note: Data after Terzaghi and Peck (1948, 1967).
A global factor of safety represents a relationship between allowable and applied   Although this concept
is simple and useful, it is also accompanied by difficulties and ambiguity. Problems arise when factors of safety are
used without first defining them and understanding why they were introduced. A single global factor ofsafety would
have unambiguous meaning if carefully prescribed standard procedures for selecting capacity, for defining loads,
and for carrying out the analysis or calculations were always used in design. However, these steps are usually not
well defined, or followed uniformly or consistently by all geotechnical engineers. In practice, different engineers
will use different approaches and select different values of strength for design, even for the same site. For example,
some engineers may use a mean value for strength, while others will use a much more conservative value such as
minimum or lower bound values in measured strength. Therefore, for the same numerical value of global safety
factor, the actual margin of safety can be very different. Further, the value of the global factor of safety tells us very
little quantitatively as to the possibility or probability of failure.
142 Canadian Foundation Engineering Manual

The global factor of safety (FS)  can be defined in many ways.  The traditional FS  is  defined as the ratio  of ultimate 
resistance (R) to  the applied load (SJ 
For  FS  =  1,  a  limiting  condition  theoretically  exists  where  the  resistance  equals  the  load  effects  (i.e.,  a  state  of 
The  limitations  of WSD  and  the  use of a single  global  factor  of safety have been discussed  by  Green  (1989)  and 
Becker (1996a). Despite all its apparent limitations, the global FS and WSD approach is a simple approach that has 
worked well  in geotechnical engineering  design.  WSD has  been the traditional  design method for  over  100 years. 
Consequently, an extensive database and experience have been assimilated over the years towards the development 
of good engineering practice. Improvements and refinements have been incorporated as the need arises. It would be 
foolish and inconceivable to ignore this substantial database and experience gained in WSD. It is noted that despite 
the  shortcomings ofWSD, the development oflimit states design (in some codes using partial factors)  has  utilized 
the WSD experience for calibration purposes to produce designs with comparable levels of safety as those  existing 
in previous design codes based on WSD. 
Figure  8.3  shows  the  relationship  between  global  safety  factor,  resistance  factor  and  reliability  index  based  on 
statistical assumptions for  variability in bearing resistance (coefficient of variation equal to 0.3  and a ratio of mean 
to nominal value of 1.1) typical for shallow and deep foundations.  An advantage of Figure 8.3  is that the reliability  I
index may be more readily appreciated by  geotechnical  engineers who  have considerable experience in using the 

traditional values of global safety factor.  This may assist in bridging the gap, during the transitional stage, between 
the use of working stress design and limit states design. 
..... ,'
  j ~ ......,.,
I J 
'" 0.3
\ ..-""
1  \  ~
~ I
0.8  ~
0.7  ()
0.6  tj
0.5  ~
G:i 5 

~ 4

~ 3 

1.5  2.0  2.5  3.0  3.5  4.0 
FIGURE 8.3 Relationship between FS, <1>,  and for bearing capacity kR = 1.1, V = 0.3 (from Becker, 1996b).
Bearing Pressure on Rock 143
Bearing Pressure on Rock
Bearing Pressure on Rock
9.1 Introduction
Rock is usually recognized as the best foundation materiaL Generally, bearing capacity failure and factored bearing 
resistance at ultimate limit states are rarely an issue for sound, intact rockmasses. However, design engineers should 
be aware  of the  dangers  associated with unfavourable rock  conditions, since  overstressing a rock foundation may 
result in large settlement or sudden failure.  Such failure may be due to  either deformation or failure of intact, weak 
rock  or  due  to  sliding  failure  along unfavourably  oriented  structural  planes  of weakness.  A  foundation  on  rock 
should. be designed with the same care as  a foundation on soil. 
Failure  of rock  foundations  may  occur as  the result of one  of several  mechanisms  as  shown in Figure  9.1  (from 
Franklin and Dusseault,  1989). The failure modes are described as: 
•  Bearing  capacity  failures,  which  occur  when  soil  foundations  are  overloaded  (Figures and  b),  are 
uncommon  in  rock.  However,  such  failures  may  occur  beneath  heavily  loaded  footings  on  weak  clay 
Consolidation failures are quite common in weathered rocks were the footing is placed within the weathered 
profile  (Figures and  e).  In  this  case,  unweathered  rock  corestones  are  pushed  downward  under  the 
footing  load, because of a combination of low shear strength along clay-coated  lateral joints and voids  or 
compressible fillings  in the horizontal joints. 
A  punching failure  (Figure  9.ld) may occur where  the  foundation  comprises  a  porous rock type,  such as 
shale,  tuff and porous  limestones  (chalk).  The  mechanism  comprises  a  combination  of elastic  distortion 
of the solid framework between the voids  and the  crushing of the  rock where it is  locally highly stressed 
(Sowers  and  Sowers,  1970).  Following  such  a  failure,  the  grains  are  in much  closer  contact.  Continued 
leaching and weathering will weaken these rock types, resulting in further consolidation with time. 
•  Slope failure may be induced by foundation loading of the ground surface adjacent to a depression or slope 
(Figure 9.1f). In this case, the stress induced by the foundation  is sufficient to overcome the strength of the 
slope material. 
•  Subsidence of the  ground surface may result from collapse of strata undercut by sub-surface voids.  Such 
voids  may be  natural  or mining  induced.  Natural voids  can be  formed  by solution weathering  of gypsum 
or rocksalt and are commonly encountered in limestone terrain (Figure 9.1g). When weathering is focused 
along intersecting vertical joints, a chimney-like  opening called a pipe will  form,  which may extend from 
the base of the soil overburden to a depth ofmany tens of meters. When pipes are covered by granular soils, 
the finer silt and sand components can wash downward into the pipes, leaving a metastable arch of coarse 
sand and gravel which may subsequently collapse (Figure 9.1h). 
144 Canadian Foundation Engineering Manual
FIGURE 9.1 Mechanisms offoundation failure (from Franklin and Dusseault, 1989; adapted from Sowers,
1976): a) Prandtl-type shearing in weak rock; b) shearing with superimposed brittle crust; c) compression of
weathered joints; d) compression and punching ofporous rock underlying a rigid crust; e) breaking ofpinnacles
from a weathered rock surface;.f) slope failure caused by superimposed loading; g) collapse ofa shallow cave;
and h) sinkhole caused by soil erosion into solution cavities
The methods proposed in this Manual for the determination of design bearing pressure on rock are suitable for
various ranges of rock quality. The design bearing pressure is generally for serviceability limit states for settlements
not exceeding 25 mm. The bearing pressure assessment is for relatively sound rock not subject to the special
conditions shown in Figures 9.1 b through h. Guidance on the applicability of the proposed methods is provided in
Table 9.1.
Bearing Pressure on Rock  145 
TABLE 9.1  Applicability 0/Methods/or the Determination a/Design Bearing Pressure on
Rock depending upon Rockmass Quality
Rockmass Quality  Basis of Design Method 
Sound rock 
Rockmass  with  wide or  very wide discontinuity 
Core strength (see Section 9.2) 
Rockmass  with  closed  discontinuities  at  moderately
close, wide and very wide spacing
Core strength (see SectIOn 9.2) 
Low to very low strength rock 
Rockmass  with  close or  very closely spaced
Very low strength rock 
Rockmass with very closely spaced discontinuities 
Soil mechanics approach 
Note: Italicised tetIns are defined in  Section 3.2. Preliminary estimates are provided in Table 9.3  and Section 9.3. 
In all cases, field tests may also be used to  assess the capacity and load-deformation characteristics ofthe rockmass, 
as  discussed in Chapter 4. 
Foundations on  Sound Rock 
For the purpose of this section, a rock mass is  considered sound when the spacing of discontinuities is in excess of 
.  . 
When the  rock  is  sound,  the  strength  of the  rock  foundation  is  commonly  in excess  of the  design requirements, 
provided the discontinuities are closed and are favorably  oriented with respect to  the  applied forces.  Geotechnical 
investigations should, therefore, concentrate on the following  foundation aspects: 
•  identification  and  mapping  of all  discontinuities  in  the  rock  mass  within  the  zone  of influence  of the 
foundation,  including the determination of the aperture of discontinuities; 
•  evaluation ofthe mechanical properties ofthese discontinuities, such as frictional resistance, compressibility, 
and strength offilling material; and 
identification and evaluation of the strength of the intact rock material. 
Such investigations should be carried out by a person competent in this work, and following  the guidelines set out 
in Chapters  3 and 4. 
The  final  determination of the  design bearing pressure on  rock may be governed by the  results  of the  analysis  of 
the influence  of the discontinuities  on the  behaviour of the  foundation.  As  a guideline,  in the  case  of a rock  mass 
with favourable  characteristics (e.g.,  the rock surface is perpendicular to  the foundation,  the load has no  tangential 
component,  the  rock mass  has  no  open  discontinuities),  the  design bearing  pressure  may  be  estimated  from  the 
following approximate relation: 
q  -K  xq  (9.1)
a  sp u-core 
146 Canadian Foundation Engineering Manual
average   compressivestrengthofrock(asdetemlinedfromASTMD2938).
anempiricalcoefficient,whichincludesa factor ofsafetyof3 (intermsofworkingstress design)
andrangesfrom0.1 to0.4(seeTable9.2andFigure9.2).
TABLE 9.2 Coefficients ofDiscontinuity Spacing,  Ksp 
0.3 to 1
Thefactors influencingthemagnitudeofthecoefficientareshowngraphicallyinFigure9.2.Therelationshipgiven
inthefigureisvalidforarockmasswithspacingof discontinuitiesgreaterthan300mm,apertureofdiscontinuities
lessthan5rnm(orlessthan25rnm, iffilledwithsoilorrockdebris),andfor afoundationwidthgreaterthan300
mIn. Forsedimentaryrocks,thestratamustbehorizontalornearlyso.
00.2 0.<I 0.6 0.S 1.0 1.2 1.4 1.6 1.8 2.0 
RA TIO ell? 
FIGURE 9.2: Bearing pressure coefficient Ksp 
Thebearing-pressure coefficient, K
' as givenin Figure 9.2, takes into account the size effectand thepresence
ofdiscontinuities andincludes a nominal safetyfactor of3 againstthe lower-bound bearingcapacityofthe rock
foundation. Thefactorof safetyagainstgeneralbearingfailure(ultimatelimitstates)maybeuptotentimeshigher.
Foramoredetailedexplanation, seeLadanyietal. (1974). FranklinandGruspier(1983)discussaspecialcaseof
foundations onshale.
, ;'-
Bearing Pressure on Rock 147
Estimates of Bearing Pressure
Universally applicable values of design bearing pressure cannot be given. The design bearing pressure is generally
for serviceability limit states for settlement not exceeding 25 mm, or by the settlement criteria, as described in
Chapter 11. Nevertheless, it is often useful to estimate a bearing pressure for preliminary design on the basis of the
material description. Such values must be verified or treated with caution for final design. Table 9.3 gives presumed
preliminary design bearing pressure for different types of soils and rocks.
TABLE 9.3 Presumed Preliminary DeSign Bearing Pressure
Types and Conditions of Rocks and Strength of Rock
Remarks Design Bearing
Soils Material
Pressure (kPa)
148 Canadian Foundation Engineering Manual
TABLE  9.3  Presumed Preliminary Design Bearing Pressure (continued)
Types And  Conditions Of 
Rocks And  Soils 
Very stiff to  hard clays or 
heterogeneous mixtures such as  till 
Stiff clays 
Firm clays 
Soft clays and silts 
Very soft clays and silts 
Peat and organic soils 
Strength Of Rock 
Design  Bearing 
Pressure (kPa) 
not applicable 
Not applicable 
Not applicable 
Fine-grained soils  are 
susceptible to  long-term 
consolidation settlement 
due to  imposed loads 
and are often susceptible 
to  severe swelling or 
shrinking due to  changed 
moisture conditions.  If 
the Plasticity Index (I )
exceeds 30 and the clay 
content exceeds 25  %,  the 
long-term performance 
of the foundation  may 
be significantly affected 
by swelling or shrinking   
ofthe subsoils, and a 
complete assessment 
of these possibilities  is 
necessary as  discussed in 
Chapter  15 
1.  The  above  values  for  sedimentary  or  foliated  rocks  apply  where  the  strata  or  the  foliation  are  level  or nearly 
so,  and,  then,  only  if the  area  has  ample  lateral  support.  Tilted  strata  and  their  relation  to  nearby  slopes  or 
excavations should be assessed by a person knowledgeable in this field of work. 
2.  Sound rock conditions allow minor cracks at spacing not closer than  1 m. 
3.  To  be assessed by examination in-situ, including test loading if necessary. 
4.  These rocks  are apt to  swell on release of stress, and on exposure to  water they are apt to  soften and swell. 
5.  The  above  values  are  preliminary  estimates  only  and  may  need  to  be  adjusted  upwards  or  downwards  in  a 
specific case. No consideration has been made for the depth of embedment of the foundation.  Reference should 
be made to  other parts of the Manual when using this table. 
Foundations on  Weak Rock 
Conditions are frequently encountered where the rock material is very weak, has very closely spaced discontinuities, 
or is heavily weathered or fragmented. It is common practice in such cases to consider the rock as  a soil mass and to 
design the foundation on the basis of conventional soil mechanics. However, the strength parameters necessary for 
such a design are difficult to evaluate. For more details on the estimation of strength and deformation parameters of 
rock masses, see the discussion in Chapter 3. Additional detail may also be found in Barton et al.  (1974), Bieniawski 
(1976), and Hoek and Brown (1980). Table 9.3  provides suggestions for preliminary estimates. 
Bearing Pressure on Rock  149 
Special Cases
Bearing Capacity of Jointed and Layered Rockmasses
. for a foundation onrockthat isjointedis dependent onthejointspacing andaperture, theareato be
desIgn .h t h . . h ...
d the location ofthe load WIt respect 0 t eJomts. T esecharactenstlcs dIctate whetherthe rockwill
'c. c" '. allompression unconfined compression, or splitting. Where a weak compressible layer is present in the
undergo C' . . . . . •
,. d t' nrockmass,thehardrocklayercanfallmflexureorpunchmg.IftheratIO ofthethIcknessof thehardrock

foun a 10
atIOn IS sma,
11 th
'111'k I
Iftheratiois large,andthe
layerto . h k '11 J:: '1  b fl . .
:  I  trengthoftherockIS small,t e roc WI lal y exure. ThIS analYSIS can alsobeusedfor designs with
s .
i' :hardrocklayers overvOIds.
:;i; ,  . gcapacitycalculationsfor thisrangeofconditions areproposedbyLo andHefny(2001) andbytheASCE
. Bearm 'b'I' 'd'
lorservlcea ,I consl eratIOns can takenas theultimatebearing
.  '. ill  .d dby the factor ofsafety. Generally, the mmimum factor ofsafety IS 3 for a structure load compnsmg the
. d:adloadandfull live load. Forfactored  bearingresistance atultimate limitstates,theultimate
bearingcapacityismultipliedbythegeotechmcalreSIstancefactorof0.5 as perTables 8.1 and8.2(Chapter8).
Foundations on Karstic Formations
. Sinkholes are oftenthe cause of onkarstic for:mations. These cavities, causedbythe chemical
. actionbetweenlimestoneandaCIdIc water, are trregular and dlfficultto predict. Sinkholes maydevelop at any
:me;therefore,investigationsarenecessarythroughoutthelifeof thestructure.
Sinkholes canbedetectedusinga numberofgeophysicaltechniques, including ground-penetratingradar(GPR),
electromagnetic conductivity measurements (EM), and by drilling core samples. Sinkhole remediation can be
performedby: concrete undergroundbr.idging;.loadeccentricity; replacementofcollapsedmaterial
withconcrete.FormoremformatlOnonthedetectIOnof sInkholesandremedIalmeasures,seeWyllie(1992).
Differential Settlement
Differential settlement occurs when adjacent footings are SUbjected to unequal settlements. Settlement (0) for
footings onelasticmediumcanbecalculatedbythefollowing equationfromLoandCooke(1989):

Maximumdifferentialsettlementshouldbecalculatedandtestedforduringdesignstagesto avoidredesignofthe
footings. Settlementinrockwithseams orfaults canbeestimatedbyplateloadtestingasdiscussedinChapter4.
150 Canadian Foundation Engineering Manual
Bearing Capacity of Shallow
Foundations on Soil
10 Bearing Capacity of Shallow Foundations on Soil
10.1 Introduction
One  possible  ultimate  limit  state  of a  shallow  foundation  involves  the  case  where  the  applied  loads  exceed  the 
resistance ofthe ground beneath the foundation. The geotechnical resistance at this ultimate limit state is termed the 
ultimate bearing capacity of the ground that supports the foundation.  The ultimate bearing capacity depends on the 
strength of the ground,  ground conditions (e.g., thickness  and presence of weak layers,  depth to bedrock),  and the 
nature of applied loading (e.g., vertical, horizontal and inclined forces;  moments). Methods to estimate the ultimate 
bearing  capacity  of shallow  foundations  on  fine- and  coarse-grained  soils  are  presented  in  this  Chapter.  Other 
possible ultimate  limit states  for  shallow foundations  may include  sliding,  overturning and general slope stability 
and their  influence  on foundation  design need to be assessed for  each individual project.  The  serviceability  limit 
state of the foundation is  considered separately from  the ultimate limit state,  as  presented in Chapter  11.  Shallow 
foundations  are those constructed on or embedded near the ground surface such that the distance from the ground 
surface to the underside of the foundation is not greater than the width (or least plan dimension) of the foundation. 
10.2 Conventional Bearing Capacity
·10.2.1 Bearing Capacity Equation
The  ultimate  bearing  capacity  (i.e.  the  geotechnical  bearing  resistance  at  the  ultimate  limit  state)  of a  shallow 
foundation  on  uniform  soil  as  shown  in  Figure  10.1  with  shear  strength  parameters  c  and  <P may  be  calculated 
ultimate bearing capacity (denoted as  Rn in limit states design-see Section 8.4), 
Ne' Nql Ny
dimensionless bearing capacity factors  (see  10.2.3), 
Se' Sq'Sy dimensionless modification factors for foundation shape, inclination, depth and tilt and ground 
slope (see  10.2.4), 
vertical stress acting at the  elevation of the base offoundation (see  10.2.2), 
width of foundation or least plan dimension of the foundation, 
soil cohesion (see  10.2.2), 
y soil unit weight (see  10.2.6). 
Unless otherwise noted, any consistent set of units may be used for the parameters in Equation  10.1. 
Equation 10.1  expresses the ultimate bearing capacity of a foundation experiencing general shear failure as the sum 
of:  the  shear resistance of a weightless material with cohesive strength parameter c (Nc term), the  shear resistance 

Bearing Capacity of Shallow Foundations on Soil 151
of a frictional  but weightless material with angle of friction $' on addition of a surcharge qs at the foundation level 
(N term),  and the  shear resistance  of a  frictional  material with angle of friction  and weight y but no  surcharge 
(Ny term). 
Shear strength parameters c and $' are normally selected within depth B beneath the base of the foundation. 
+z  01 
*id!di1u.t*IBI. I qs
... '"  " __ ....  t ..t..t..t."".. 
Ground .lZ.
t( B )I
Uniform ground
with c, ¢,  r
FIGURE 10.1 Definition ofgeometry andparameters for ultimate bearing capacity ofa shallow foundation
10.2.2. Undrained and Drained Conditions
The values  of c  and $' for  use in the general bearing capacity equation (Equation  10.1)  depend  on the  type of soil 
and whether short-term (undrained) or long-term (drained) conditions are being examined. The short-term stability 
of a foundation involving fine-grained soils can be calculated by taking c equal to  the undrained shear strength, s u '
and $  O.  The long-term stability of a  foundation can be obtained with c equal to the effective cohesion intercept, 
c/, and $' equal to  the effective angle  of internal friction of the soil, $'.  In most cases,  short-term stability controls 
design,  especially for soft to very ,stiff clays. 
The surcharge qs for use in the general bearing capacity also depends on whether undrained or drained conditions are 
being considered. For undrained conditions qs is the total vertical stress acting adjacent to the base ofthe foundation; 
whereas, for drained conditions it is equal to the vertical effective stress and consequently will be influenced by the 
position of the groundwater level (see Section  10.2.6). 
10.2.3 Bearing Capacity Factors
Bearing capacity factors  have  been derived based on modified plasticity solutions  for uniform ground conditions. 
Bearing capacity factors Nc and N have been reported by Meyerhof (1963), Hansen (1970) and Vesic (1975) to  be 
equal to: 
Nc (N
- 1)  cot$  (10.2) 
= tan (  45°+ 
Several  formulations  of the  bearing  capacity  factor  Ny are  available  (Terzaghi,  1943;  Meyerhof,  1963;  Hansen, 
1970; Vesic,  1975) but tend to overestimate N when compared with the more rigorous plasticity solution of Davis 
and Booker (1971). An approximate value of N suitable for $' > 10

obtained from Davis and Booker (1971) is: 
N = (lOA)
for a smooth interface between the foundation and the ground, while for a rough interface it equals: 
$  (l0.5) 
where <i> is  in degrees. 
152 Canadian Foundation Engineering Manual
For the case of undrained stability (c su' 4>' = 0) the bearing capacity factors become:
= (2 + n) (10.6)
N = 1 and (10.7)
0 ( 10.8)
Bearing capacity factors N ' N , and Ny for uniform ground conditions are presented in Table 10.1 and plotted in
c q
Figure 10.2.
TABLE 10.1 Bearing capacity factors Nc and N
from Meyerhof(1963) and Ny from Davis and Booker (1971)
11 3.9
N" rough
N smooth
I )'
0 0
1.3 0.8
20 15 1.7
6.4 3.0
16 7.1 3.6 2.0
22 17 7.8 4.2 2.4
23 8.7 2.8
19 9.6 3.3
11 3.8
26 22 12 4.5
24 13 9.7 5.3
28 26 15 6.2
16 28. 14 7.3
30 18 30 16 8.6
31 21 33 19 10
32 23 22 35 12
33 26 27 14
42 29 31 17
35 46 33 37 19
36 38 51 44 23
37 43 56 52 27
38 61 49 61 32
68 56 73 37
75 64 44 86
Sma1l (2001) and Poulos et al. (2001) present useful summaries ofbearing capacity factors for soils with an increase
in strength with depth, finite depth, fissured clays, layered soils, and foundations near slopes.
Bearing Capacity of Shallow Foundations on Soil 153
100 100
80 80

"'0 60 60
40 40
20 20
0 0
0 10 20 30 40 0 10 20 30 40
rp' (degrees) rp' (degrees)
FIGURE 10.2 Bearing capacity factors Nc and Nqfrom Meyerhof(1963) and N/rom Davis and Booker (1971)
10.2.4 Modification Factors
The bearing capacity factors were derived for the case of strip footing on a level base subjected to loading
perpendicular to the foundation. Deviations from these conditions can be accounted for, where appropriate, by
factors to modify the bearing capacity factors for the effects of foundation shape (S os' SqS and Sf )' load inclination
(s ., S . and S .), foundation depth (S d' S dand S d)' surface slope (S P' S Pand S p)and foundation tilt (S 15' S rand S,)
c, ql yl C q Y c q Y c qu 7u
where expressions for the various modification factors are given in Table 10.2 based on Vesic (1975).
TABLE 10.2 Modification Factors for General Bearing Capacity Equation (based on Vesic, 1975)
Factor Sc Sq Sy
B'N 'B' B'
Foundation shape, s S = 1-0.4- Sqs=l+ L,tan¢
Y-I' L' L' Nc
Inclined loading, i [1] ¢=O, SCI
[ H r+!
S - I-
(1- V + B,Zccot¢r Ii - V + B'L'ccot¢
I . Sq;
¢> 0, Sci = Sq;
1 + OAk Foundation depth, d[2)
Syd 1 Sqd = 1+2tan¢(l sin¢Yk
¢> 0, S =S -
cd qd Ne tan¢
154  Canadian Foundation Engineering  Manual 
TABLE 10.2 Modification Factorsfor General Bearing Capacity Equation 
(based on  Vesic,  1975) (continued) 
S  =l_L  I
Surface slope, p[3] 
cfJ  tr + 2 
SyfJ  '" (1- tanp)"  [4] 
S  - S  _  1  - SqfJ  \'
cfJ- qfJ  Nctan¢  , 
¢ =0,  S  Base  inclination,  8[5J 
cD  tr + 2 
[1]  V = vertical force; H  horizontal force;  m  depends on direction of inclined loading 8 relative  to long side of 
the foundation:  If force  inclined in B  direction  m  m

= (2+BIL)/(l +BIL),  if inclined in  L  direction 
(8=0°)  m  =  =(2+LlB)I(1+LlB),  and if inclined at angle 8 to L  direction m=me =  mLcos
8  + m
8. m

[2]  k  DIB if DIB'S:l;  k=tan·1(DIB)  if DIB >1. 
[3]  p= inclination below horizontal ofthe ground surface away from the edge ofthe foundation (see Figure  lOA); 
for p< 11:/4; Pin radians. 
[4]  For sloping ground case where  0 N  =  -2sin,B must be used in bearing capacity equation. 

[5]  b= inclination from the horizontal ofthe underside ofthe foundation (see Figure  lOA); for b < n:l4; b in radians. 

H  ..+"- "-

2  2 
FIGURE 10.4 Definition ofparameters for shallow foundation with ground slope pand base tilt b 
10.2.5 Eccentric Forces and Moments
If the foundation  is  subjected to  vertical forces  that act  eccentric  to  the  centroid of the foundation,  the size  of the 
foundation used in the bearing capacity equation should be reduced: 
B' =B  2e

l' =  B -2e  (10.13)

B,L  actual foundation  dimensions, 
B',L'  reduced dimensions for use in bearing capacity equation,  and 
,  e

eccentricities of force  in directions Band L  from the centroid. 
This is an approximate but reasonable approach to account for eccentricities provided that the resultant loading acts 
within the middle third of the foundation  (i.e.  e < B/6).  Values ofB' and L' are  to  be used in all bearing capacity 
calculations. The term k for depth modification factors S d  and S  and the term m for load inclination factors S  .,  S  . 
c  qu' 
Cl  ql
and S . as  shown in Table  10.2 remain in terms of Land B. 
Foundations  that  are  subject to  moments MB  and  ML  in  the  Band L  directions  and vertical  load  V acting  through 
the centroid can be  treated as an equivalent loading system with vertical load  V acting at eccentricities e and e as
B  L 
shown in Figure  10.3. 
Bearing Capacity of Shallow Foundations on  Soil  155 
ML ~


~ B ~

~ e L

IE--L'----7Iot  IE-B "--'" 
ML  Me 
=- eB=-
V  V 
FIGURE 10.3 Shallow foundation subjected to  moments and vertical force 
10.2.6  Influence of Groundwater 
The position of the groundwater level will influence the selection of Y  and qs  for use in the general bearing capacity 
equation when considering drained conditions as  summarized in Table  10.3. 
TABLE 10.3 Unit  Weight and Surcharge for Drained Conditions  in  the  General Bearing Capacity 
Equation depending on Depthfrom Surface  to  the  Groundwater Level z (as  defined in Figure 10.1). 
The foundation  is located at depth D  beneath  the ground surface 
Depth from surface of groundwater level 
Unit weight l' for N,  term 

, Ysub 
Surcharge term 'I, 
sllb  Ybu/kD 
z-D&  )
Ysub  + ~ bulk  -Ysub 
bll1k  Ybu1kD 
The bulk unit weight Y
should be selected based on the minimum water content of the soil above the water table. 
Effective stresses can be introduced into the N  term by using the submerged unit weight Y b'  which is  equal to: 
.  r  • 
Y  is  the saturated unit weight and Y  is the unit weight of water. w 
In all cases in Table  10.3, qs  is the vertical effective stress adjacent to the foundation  at its base. 
10.3  Bearing Capacity Directly from In-Situ Testing 
10.3.1  Standard Penetration Test (SPT) 
There  is  no  direct  relationship  between  standard  penetration  test  (SPT)  resistance  N  and  the  ultimate  bearing 
156 Canadian Foundation Engineering Manual
capacity. Shear strength parameters for use in the general bearing capacity equation can be estimated from empirical 
correlations with SPT-N (e.g.,  Hatanaka and Uchida,  1996;  Terzaghi et al.,  1996). Empirical design charts relating 
the design bearing pressure for foundations on sand to  SPT  -N are available; however, since these are also based on 
limiting settlement of the  foundation  they are presented in Section  11.8.1.  Such  empirical  correlations need to be 
treated with caution and adjusted as appropriate by experience. 
10.3.2 Cone Penetration Test (CPT)
Shear  strength  parameters  for  use  in  the  general  bearing  capacity  equation  can  be  estimated  from  empirical 
correlations with cone penetration test (CPT) results(e.g., Lunne et ai.,  1997). Empirical methods are also available 
to estimate the ultimate bearing capacity directly from  CPT tip resistance qc'
For coarse-grained soils: 
K$ = empirical  factor relating ultimate bearing capacity and average CPT tip resistance for  coarse-grained 
soils, and 
qc = average tip resistance over a depth B beneath the foundation. 
Values  of Kq, depend  on soil density  and foundation shape  and range between 0.16  to 0.3  (Lunne et al.,  1997).  A 
value of  ~ = 0.16 can be used for most cases, recognizing that limiting settlement will generally control foundation 
For fine  grained soils and undrained conditions: 
K = empirical  factor  relating  ultimate  bearing  capacity  and  average  CPT  tip  resistance  for  fine-grained 
soils, and 
all other parameters are as previously defined.  Factor Ksu ranges from 0.3 to 0.6 depending on foundation shape and 
embedment, and soil stress history and sensitivity.  A value ofKsu 0.3  can be conservatively used for most cases. 
These empirical correlations need to be treated with caution and adjusted where appropriate based on experience. 
10.3.3 Pressuremeter and Dilatometer Tests
In-situ tests such as the pressuremeter test (PMT) and flat dilatometer test (DMT) can be used to obtain shear strength 
parameters for use in the general bearing capacity equation (e.g., Lunne et ai.,  1989; Marchetti et aI.,  2001). 
10.3.4 Plate Load Test
A plate load test, if loaded to failure,  can be used to assess the ultimate bearing capacity. In this test a reduced-scale 
foundation is subjected to load and the deflection is recorded. The plate load test involves the actual ground material 
beneath the foundation and can be useful to obtain soil parameters and to verify the method of analysis. The general 
bearing capacity equation can be used to  interpret results  if ground conditions are homogeneous with depth.  Scale 
effects are important as the results will depend on the size of the reduced-scale foundation relative to the underlying 
sequence  of soil strata.  Appropriate  engineering judgment must be exercised prior to  any  extrapolation to  larger 
foundations.  An additional disadvantage is the costs required to conduct the tests.  As a result,  plate load tests may 
only be appropriate for medium to higher risk projects.  The plate load test is also useful in the evaluation of ground 
Bearing Capacity of Shallow Foundations on Soil 157
stiffness (e.g., see Sections 7.7.1 and 11.7)
10.4 Factored Geotechnical Bearing Resistance at Ultimate Limit States
Geotechnical resistance at the ultimate limit state is reduced (multiplied by the appropriate geotechnical resistance
factor (see Tables 8.1 and 8.2 in Chapter 8) to provide the factored geotechnical bearing resistance for foundation
10.4.1 Net Ultimate Bearing Pressure
The ultimate bearing capacity qu is the total stress that can be applied at foundation level. If an excavation is made
for the foundation, stresses in excess of the original overburden stress at the foundation level contribute to bearing
failure. The net bearing capacity is defined as:
net bearing capacity,
qu ultimate bearing capacity, and
total overburden stress removed at foundation level.
There is no possibility of bearing failure if the applied load at the foundation level is equal to that of the excavated
soil. This is the basis for the design of what is termed full-compensated (or floating) foundations.
10.4.2 Allowable Bearing Capacity
In a working stress design (WSD) approach (see Chapter 8) all uncertainty is accounted for in one parameter called
the global factor of safety against ultimate bearing capacity FS. The allowable bearing capacity, ~ I I   that can be
applied at the foundation level is:
The value of FS against ultimate bearing capacity of a shallow foundation is normally taken equal to 3 (see Section
8.8 in Chapter 8).
For shallow foundations on the ground surface or neglecting the effect of the excavated ground, the allowable
bearing pressure becomes:
10.4.3 Factored Geotechnical Bearing Resistance
Using the load and resistance factor design (LRFD) approach (see Chapter 8), uncertainty in loads acting on the
foundation and the resistance of the foundation are treated separately. Loads acting on the foundation are increased
using appropriate factors for live and dead loads, while the geotechnical resistance is decreased using a geotechnical
resistance factor <P.
For the bearing resistance of shallow foundations the geotechnical resistance factor <P may be taken to be 0.5 (see
Tables 8.1 and 8.2 in Chapter 8).
158  Canadian Foundation Engineering Manual 
Settlement of Shallow Foundations 
11 Settlement of Shallow Foundations
11.1 Introduction
The settlement  of a foundation  must be within tolerable  or  acceptable limits to  satisfy the  specified serviceability 
limit  states  criteria  (see  Chapter  8)  for  a  given  project.  Methods  to  estimate  the  possible  magnitude  of ground 
settlement, the rate of settlement and the maximum allowable settlement are presented in this Chapter. 
The settlement of shallow foundations depends on the magnitude of the applied forces,  geometry ofthe foundation, 
type of ground conditions, ground stiffness and in some cases ground strength. 
The rate of settlement depends  on the  rate  of loading relative  to  the rate  of excess pore pressure  dissipation.  For 
saturated soils,  if the rate of loading exceeds the rate  of dissipation, pore pressures in excess of steady-state values 
will  be  generated.  Settlement  of the  foundation  will  then  increase  with  time  until  the  excess  pore  pressures  are 
dissipated.  Thereafter,  creep  settlement  can continue with  time.  Soil  type,  permeability,  drainage  conditions  and 
magnitude ofloads influence how quickly excess pore pressures can dissipate. 
Maximum allowable settlements (Le.,  the serviceability limit states criteria) largely depend on type and end use of 
the structure, nature of the ground conditions and risk of the project. 
11.2 Components of Deflection
Vertical deflections of a shallow foundation may arise from: 
•  undrained shear distortions  that occur with no change in void ratio (or volume);, 
drained  settlements arising  from change in void ratio (or volume) and shear distortions that occur from an 
increase in effective stresses; and 
creep settlements arising from  change in void ratio (or volume) that occur at constant effective stresses. 
Undrained  distortions  occur  from  shear  strains  when  the  rate  of loading  is  fast  relative  to  the  time  required  for 
excess  pore pressures  to  dissipate  (i.e.,  under conditions  of undrained  loading). Bince undrained distortions  arise 
from  shear strains they occur for situations other than one-dimensional loading and become more prominent as the 
size of the  loaded area decreases  relative to the thickness  of the  compressible layer.  Drained settlements  are  time-
dependent displacements associated with primary consolidation (i.e., decrease in void ratio) of the foundation  soils 
as the effective stresses increase. Drained settlements may occur rapidly for coarse-grained soils (e.g., sand, gravel), 
or very slowly for  fine-grained  soils  (e.g.,  silt,  clay).  Creep  settlements are time-dependent settlements  associated 
with  secondary  consolidation.  Reference  to  the  total  final  settlement  of a  foundation  in  the  subsequent  sections 
neglects  creep settlement.  For most practical cases  creep  settlements  may be  added to the total final  settlement as 
discussed in Section 11.10. 
Settlement of Shallow Foundations 159
Settlement of Fine-Grained Soils
For most  foundation  applications,  fine-grained  soils  typically  experience  both  undrained  distortions  and drained 
settlements,  and  possibly  secondary  compression.  Undrained  distortions  can  be  a  sigpificant  proportion  of the 
total settlement for overconsolidated clays, but often can be  small relative to  the  drained settlements  for  normally 
consolidated clays. 
The total final  settlement, STF' and the settlement at time t, ST!) are equal to  (Davis and Poulos,  1968): 
Si  undrained distortion 
SCF = final consolidation settlement, and 
U = degree of consolidation settlement. 
Methods  to  estimate  the  total  final  settlement  and  undrained  distortions  for  fine-grained  soils  are  presented  in 
Sections  11.3.3,  11.3.4,  and  11.4.2.  The  selection  of appropriate  parameters  discussed  for  use  in  these  methods 
is  discussed  in  Section  11.7.  Methods to  estimate the degree  of consolidation settlement are  presented in  Section 
11.2.2 Settlement of Coarse-Grained Soils
For  most  foundation  applications,  coarse-grained  soils  do  not  experience  undrained  distortions  since  they  are 
sufficiently permeable  to  dissipate  excess  pore pressures  rapidly,  relative to  the  rate of applied loading.  Coarse-
grained  soils  experience  drained  settlements  from  compression  of the  soil  skeleton  (i.e.,  decreases  in  void ratio) 
for increases in effective stress.  Since excess pore pressures dissipate rapidly, the settlement at time t is essentially 
equal to the total final settlement STF' Coarse-grained soils may also experience creep or additional deflections from 
cyclic loading. 
The  total final  settlement of foundations  on  coarse-grained  soils  can be calculated using the  elastic  displacement 
method  described  in Section  1l.3  (with  the  selection of appropriate parameters  discussed  in  Section  11.7)  or  by 
direct methods related to in-situ testing as described in Section  11.8. 
11.3 Three-Dimensional Elastic Displacement Method
11.3.1 Approximating Soil Response as an Ideal Elastic Material
When  subject  to  increases  in  stress  by  loading  from  foundations,  soil  materials  exhibit  nonlinear  and  inelastic 
stress-strain  response,  such  that  increments  in  stress  are  not  linearly  proportional  to  increments  in  strain,  and 
permanent strains remain upon unloading. Additionally, the  stress-strain response may be  dependent on the stress 
path.  Estimates  of settlement (serviceability  limit state)  are  made  at service  loads  (Le.,  working stresses) that are 
usually well below the ultimate limit state. 
F or  such  conditions,  the  issue  of soil  nonlinearity  may  be  resolved  by  selecting  secant  (or  average)  stiffness 
parameters for the appropriate stress (and/or strain) increment ofthe ground loaded by the foundation. Thus, despite 
the  fact  that soils  are not usually  elastic  materials,  elastic displacement theory  can  be  used to  obtain estimates  of 
foundation settlement for most practical cases. 
In elastic displacement theory the soil is treated as a saturated two phase material that is normally assumed to have a 
homogeneous and isotropic elastic soil skeleton with Young's modulus E' and Poisson's ratio v' and incompressible 
160 Canadian Foundation Engineering Manual
pore  water  (solutions  exist  that  explicitly  consider  soil  anisotropy  e.g.,  Rowe  and  Booker,  1981 a,b).  It is  the 
responsibility of the geotechnical engineer to  evaluate these elastic parameters in the context of the true modulus of 
deformation of the ground loaded by the foundation.  Guidance on the selection ofYoung's modulus E and Poisson's 
ratio v are provided in Sections  11.3.2  and  11.7. 
11.3.2 Drained and Undrained Moduli
The total final  settlement can be estimated from elastic displacement theory by using the  change in effective stress 
(once  all  excess  pore  pressures  have  dissipated)  and  drained  modulus  E' and  v'. Undrained  distortions  can  be 
calculated using the change in total stresses  and undrained modulus E and Poisson's ratio of v = 0.5  to satisfy the 
tI  u
conditions of zero volume change. 
11.3.3 Three-Dimensional Elastic Strain Integration
The total final settlement and undrained distortion can be calculated by summing the vertical strains !::..s,  arising from 
loading on the  foundation.  This approach may be useful for some problems where different layers are encountered 
or  ground properties  vary  beneath  the  foundation.  These  calculations  can be  easily  conducted  using  spreadsheet 
computer programs. 
The  increase  in  vertical  strain  is  related  to  the  increase  in  stress  using  three-dimensional  elasticity in x, y and z
coordinates  (where z  is  in  the vertical  direction).  Similar expressions  can be written for  polar coordinates  for use 
with circular foundations  (e.g., see Poulos and Davis,  1974). The total final  settlement can be calculated using: 

increment in vertical strain from the  increase in effective stresses of sublayer i, 


!::..a'  increment in effective stresses in x, y and z directions of sublayer i,
x' y' z
E',v'  secant drained Young's modulus and Poisson's ratio for the appropriate stress  increment 
and layer i, 
n =  number of sub layers , and 
oh  thickness of sub layer i.
The  increment  in  effective  stress  can  be  found  using  available solutions  for  stress  distribution with depth for  the 
appropriate  loaded region  (see  Section  11.6).  The  number  of sublayers  should be  selected to  provide  a sufficient 
integration of vertical strain with depth and also to  capture different ground conditions beneath the foundation. 
The undrained distortion can be calculated in a similar manner using: 

increment in vertical strain from the increase in total stresses of sub layer i, 
!::..a  '  !::..a  ,  !::..a  increment in total stresses in x,  y and z directions of sub layer i,
x y z 
v  secant undrained Young's modulus and Poisson's ratio for the appropriate stress 
increment and layer i, and all other parameters as previously defined. 
11.3.4 Elastic Displacement Solutions
Elastic  displacement  solutions  for  various  foundation  shapes,  soil  homogeneity,  finite  layer  depth,  mutlilayered 
soils, foundation roughness, foundation stiffness, and drainage conditions have been provided by Poulos and Davis 
Settlement of ShallOW Foundations 161
(1974).  Results from these solutions are presented in a graphical manner. This can be useful to illustrate the influence 
of key  parameters  on  foundation  settlement  (e.g.,  size  of loaded  area  relative  to  the  thickness  of compressible 
deposit).  They can also provide a useful check on the results from  more elaborate analyses. 
Elastic displacement solutions are presented in the subsequent sections in terms ofYoung's modulus E and Poisson's 
ratio  v  and can be  used to  find  the  total final  settlement using E=E' and v=v',  and  the  undrained distortion using 
E=E  and v=v =0.5. 
If U
a)  Flexible Strip Foundation 
The settlement beneath the centre of a flexible strip foundation  on the  surface of a uniform layer of isotropic elastic 
material of thickness h and subject to uniform vertical pressure q is equal to: 
q average pressure applied to  the ground by the foundation, 
B width of strip foundation, 
E drained or undrained modulus  of ground, 
Is  influence factor for a strip foundation given in  Figure, and 
h distance from ground surface to an incompressible base. 
b)  Circular Foundation 
The settlement beneath  the  centre  of a  circular  foundation  on the surface of a homogeneous  and isotropic  elastic 
material of thickness h  and subject to uniform vertical pressure q is equal to: 
( 11.6) 
q  average pressure applied to the ground by the foundation, 
B diameter ofthe circular foundation, 
E drained or undrained modulus of ground, and 
Ie  =  influence factor for  a circular foundation given in Figure  11.1b. 
2.5  1.0 
h B
h B

0  2  4 6  8  10 
0  2  4 6  8  10 
FIGURE 11.1 Influence factors for the settlement beneath the  centre offlexible:  (a)  strip foundation 
ofwidth B  and (b)  circular foundation  ofdiameter B,  on a uniform isotropic compressible 
material ofthickness h. Modifiedfrom Rowe and Booker (J981a,b) 
162 Canadian Foundation Engineering Manual
For the more general cases involving non-uniform ground stiffuess, foundation rigidity and burial beneath the
surface, as defined in Figure 11.2, the settlement beneath the centre of a shallow circular foundation resting on an
isotropic elastic material of finite thickness whose stiffness increases linearly with depth and is subject to uniform
vertical pressure q can be estimated using (Mayne and Poulos, 1999):
S = q BIG IF  IE  (l-v 2)
q  average pressure applied to the ground by the foundation,
B  diameter of the circular foundation,
IG  influence factor for nonuniform ground stiffness given in Figure 11.2a,
IF  influence factor for foundation stiffness given in Figure 11.2b,
IE  influence factor for foundation embedment given in Figure 11.2c,
v Poisson's ratio, and
Eo  drained or undrained modulus at the ground surface.
The influence factor for nonuniform ground stiffness is plotted against the dimensionless term:
k  is the increase in modulus with depth.
The influence factor for foundation stiffness is defined in terms of the dimensionless foundation flexibility ratio KF
which is equal to:
K = EF  (2t)3 
( 11.9)
EF-is the Young's modulus ofthe foundation material (e.g., concrete, steel), t is the thickness of the foundation,
and E is the average modulus of the ground within depth B beneath the foundation.
Although developed for circular footings, this method can be used for square and rectangular footing (provided the
length is less than three times the breadth) with an equivalent diameter used for B such that the total force applied
to the foundation is the same.
11.4 One-Dimensional Consolidation Method
11.4.1 Oedometer Test
The stiffuess parameters for many practical settlement calculations involving fine-grained soils can be obtained
from one-dimensional consolidation laboratory tests, referred to herein as the oedometer test. In principle, the
settlement of coarse-grained soils could also be assessed using the oedometer test; however, this is not normally
practical given the difficulties in obtaining undisturbed samples of coarse-grained soils.
Settlement of ShallOW Foundations 163
0.01 0.1 10 100
,..... "-
B k
fI z
0.001 0.01 0.1 1 10 100
=~   ~ )
0 5 10 15 20
~ -
fI= 0.5
FIGURE 11.2 lrifluence factors for the settlement beneath the centre ofa uniformly loaded circular
foundation ofdiameter B on a finite compressible layer ofthickness h whose stiffness increases
linearly with depth. Influence factors for: (a) nonuniform ground stiffness, (b) foundation rigidity
and (c) foundation embedment. Modifiedfrom Mayne andPoulos (1999)
164 Canadian Foundation Engineering Manual
Specific details on the procedures to conduct and interpret results from the oedometer test can be found elsewhere
(e.g., ASTM D2435; Holtz and Kovacs, 1981). In this test, soil samples retrieved from the field are subjected to
increments in total vertical stress under conditions of zero lateral strain. Excess pore pressures that generate in
the sample from an increment in total stress are allowed to dissipate (normally for a 24 hour period see ASTM
D2435 for possible deviations) prior to placement of an additional total stress increment. It is often assumed that the
increase in effective stress is equal to the increase in total stress at the end of each increment. The change in void
ratio (obtained from the change in height of the sample) is recorded for each stress increment. The void ratio at the
end of each increment is plotted versus the logarithm of the effective stress on the sample as illustrated in Figure
for sample
Effective stress d
(log scale)
FIGURE 11.3 Oedometer test results showing definition ofparameters to calculate one-dimensional settlement
The following parameters can be defined in reference to Figure 11.3:
eo initial void ratio of the sample corresponding to the initial (or in-situ) vertical effective stress (J'lo'
preconsolidation pressure which corresponds to the previous maximum vertical effective stress
experienced by the sample,
compression index, and
C recompression index (this portion of the plot is present only if J   ~ > O"J
The preconsolidation pressure is related to the stress history of the deposit where normally consolidated soils
have 0" approximately equal to (J" and overconsolidated soils have 0" greater than (J" . The magnitude of the
pop 0
preconsolidation pressure may also be presented in terms the overconsolidation ratio, OCR, where:
OCR = -p (11.10)
The preconsolidation pressure can be estimated using the empirical and graphical Casagrande procedure (for specific
details see Holtz and Kovacs, 1981). This approach is normally sufficient for estimation of foundation settlement
provided there is a defined change in slope of the e-log(J" plot. An alternate approach may be necessary for soils
with a more gradual change in slope of the e-log(J" plot (e.g., Becker et aI., 1987). Geologic information about the
site can also be used to assist with the estimation of the preconsolidation pressures.
Sampling disturbance decreases the preconsolidation pressure obtained from the oedometer test and also increases the
calculated settlements (Leroueil, 1996). Empirical methods exist to modify the measured laboratory curve to account
for changes in sample compressibility arising from sampling disturbance (e.g., see Holtz and Kovacs, 1981).
Settlement of Shallow Foundations 165
Slopes  C

and  C
are  dimensionless  parameters.  Although  they  represent  the  compressibility  of a  particular  soil 
sample with a single value over a certain stress range, this does not imply that its stiffness is constant over that stress 
range.  Rather the value of C

combined with the logarithmic scale captnres the strain-hardening behaviour of soils 
(i.e., they become stiffer as  the effective stresses increase). 
11.4.2 One-Dimensional Settlement: e-Ioga' Method
For cases where the loaded area ofthe foundation is large relative to the thickness ofthe compressible deposit, lateral 
strains  may  be sufficiently  small  such that the  foundation  settlement can be approximated with one-dimensional 
strain  models.  Since  one-dimensional  strain  conditions  are  imposed  during  a  conventional  oedometer test,  one-
dimensional settlement is  denoted herein as oedometer settlement Saed' 
One-dimensional settlement from an increase in initial vertical effective stress a'a to final  vertical effective stress 
is  obtained by summing the increase in vertical strains with depth. The increment in vertical strain is  obtained from 
the change in void ratio !::.e for an increase in effective stress from laboratory oedometer data viz: 
n[  ]  n[-l1e  ] 
Soed  = l1e z 8h i = 1+ eo  8h i

increment in vertical strain from the increase in vertical effective stresses of sub layer i, 
!::.e = change in void ratio from the increase in vertical effective stresses(i.e.,  at) of sublayer i, 
e  = initial void ratio of the  sample corresponding to the initial (or in-situ)  vertical effective stress a

o  o 
sublayer i, 
n  = number of sublayers, and 
8h = thickness of sub layer i. 
The negative sign in front ofthe /::"e term is to account for the decrease in void ratio for an increase in effective stress. 
The change in void ratio depends on the stress history ofthe soil and magnitnde ofthe final vertical effective stresses 
relative  to  the preconsolidation pressure.  Final  vertical  effective  stresses  can be obtained  using  elastic solutions 
(Section 11.6) and incorporating changes'in water levels beneath the foundation. Ifthe soil is normally consolidated, 
then the change in void ratio is  equal to: 
M  -C,   J 
If the soil is  overconsolidated and  < a'p'  then the change in void ratio is equal to: 
l1e  =-C
lOgJO(cr  I

while if overconsolidated and  (J'p'  the change in void ratio is  equal to: 
8e  -C" !OglO( J-c, !OglOl:tJ 
Alternatively, one-dimensional settlement can be  expressed in terms of the coefficient of volume decrease, my':
n  n 
Soed  =IJl1e 

8h1==  :lJmv I1cr; 8hJi 
i=l  ;=1
The coefficient of volume  decrease  is  the  slope obtained from  a plot of effective  stress  (plotted on a linear scale) 
versus vertical strain obtained from an oedometer test.  An appropriate secant value of mv should be selected for the 
effective stress increment expected beneath the foundation since mv is  dependent on stress  level and stress history. 
Calculation of one-dimensional settlement is a special case of the more general three-dimensional elastic settlement 
presented in Section 11.3 where lateral strains are neglected (Le., v  0) and a one-dimensional constrained modulus 
(11m)  is  used for the elastic modulus. 
166 Canadian Foundation Engineering Manual
11.4.3 Modifications to One-Dimensional Settlement
For foundations with one-dimensional conditions there are no undrained distortions Sj and the total final settlement 
will be equal to the one-dimensional settlement STF= Soed' This would be applicable for foundations where the loaded 
area is large relative to the thickness of the compressible deposit. 
Modification to S  may be required for foundations with other than one-dimensional conditions (e.g., foundations 
where lateral  strains will occur).  For normally consolidated clays, Soed provides a good approximation for  the final 
consolidation  settlement  SCF' whereas  for  stiff overconsolidated  clays  Soed is  a  good  approximation  to  the  total 
final  settlement  (Burland  et  aL,  1977;  Poulos,  2000).  Thus  the  following  modifications  are  required  to  one-
dimensional settlement theory for applications to  two- and three-dimensional condition. 
For normally consolidated clays: 
F or stiff overconsolidated clays. 
STF =Soed
S. (11.18)
11.5 Local Yield
The  undrained  distortion  of heavily loaded foundations  on weak soils  may  be  larger than those  calculated using 
elastic displacement theory because oflocal ground yield (shear failure) beneath the foundation.  The consolidation 
settlement and the rate of settlement are not greatly affected by local yield (Small et al.,  1976; Carter et al.  1979). 
Based  on  the  results  provided  by  D' Appolonia  et  al.  (1971),  local  yield  may  have  an  influence  on  undrained 
distortions for foundations with a global factor of safety against bearing capacity ofthree or greater (FS? 3) if: 
(1 ~ K   a' <s (11.19) 
o  0 u
coefficient oflateral earth pressure, 
initial vertical effective stress beneath the base of the foundation,  and 
undrained shear strength ofthe soil within depth B beneath the base of the foundation (B is the least 
plan dimension of the foundation).  . 
For  cases  that satisfy  Equation  11.19,  the  effects  oflocal yield  on undrained  distortions  can be quantified using 
modification factors for strip foundations reported by D' Appolonia et aL  (1971) or by using numerical methods (see 
Section 11.9).  Local yield can be neglected for cases that do not satisfy Equation 11.19. 
11.6 Estimating Stress Increments
Increments in total stress beneath a loaded region can be estimated using elastic theory. The following solutions for 
the  stress  increments in a  homogeneous,  isotropic,  semi-infinite  elastic  medium when subject to different loaded 
areas  were  obtained  from  Poulos  and Davis  (1974).  This  reference  also  provides  a  useful  compilation  of elastic 
stress distribution solutions for other loading conditions and nonuniform ground conditions. 
11.6.1 Point Load
The stress increments at a point with coordinates rand z beneath a point load of magnitude P on the ground surface 
(Figure  11.4) are:  3P 3
cr  =
z 2n:
Settlement of Shallow Foundations 167
z + (1- 2v )R] 
a =---
r 2n R2 
2nR2  R  R + z 
'  a,.,  a
and Or. are the vertical, radial, tangential and shear stresses induced by the point load.
__________ ________
FIGURE 11.4 Vertical force P acting on  the ground surface 
11.6.2 Uniformly Loaded Strip
The stress increments beneath an infinitely long strip of width 2b  subject to uniform vertical pressure q  on the
ground surface are:
a z =!L P. + sina cos(a + 28 )] (l1.2Ia)

ax =!Lp.  sino. cos (a +28)] (l1.2Ib)

a =2Qva
Y  n 
t xz  = !Lsina sin(a + 28 ) (l1.21d)

a, a ,a and 0 are the v.ertical, horizontal, axial and shear stresses, and a.  and b are angles in radians as shown
z  x  y xz
in Figure II.Sa. Positive angles are counter clockwise from the vertical. Contours of vertical stresses from
Equation ll.21a are plotted in Figure 11.Sb.
168  Canadian Foundation Engineering Manual 

x  ~
........ a'01\ .... N 
'"'....  ,
'.  , 
  ' : ~
0.5 1.0 1.5 2 2.5 3
FIGURE 11.5  Vertical stress (J"z  beneath a uniformly loaded strip ofwidth 2b on 
the ground surface subject to  vertical pressure q 
11.6.3 Uniformly Loaded Circle 


0.5 1.0 1.5 2 2.5 3
FIGURE 11.6  Vertical stress (J"z  beneath a uniformly loaded circle ofdiameter 2a on 
the ground surface subject to  vertical pressure q 
Settlement of Shallow Foundations 169
The stress increments beneath the centre of a circular area (i.e., along r 0) with radius a subject to uniform vertical
pressure q on the ground surface (Figure 11.6a) are:
q[ 2(l+v)z
+ {
0' r =0' e = - (1 + 2v ) - { ~
~ ] 2 + /
(Jz' (JI" and (JIJ are the vertical, radial and tangential stresses.
The vertical stresses beneath a uniformly loaded circle are plotted in Figure 11.6b.
11.6.4 Uniformly Loaded Rectangle
The stress increments beneath the comer of a rectangle oflength L and width b subject to a uniform vertical pressure
q on the ground surface (Figure 11.7a) are:
=- tan
LbZ[ 1 1)]
q [ -t(LbJ
- +-- --2 +-2
2n zR3 R3 R
-I ( Lb J LbZ]
0' = tan - ---
x 2n
. ZR3 Rl2R3
( I1.23f)
R2 =(b
+   y ~
R3 =(p + b
+ Z2),Y:;
170 Canadian Foundation Engineering Manual
and a a and a are the vertical, horizontal and axial stresses and
and r are the shear stresses.
z'.x' y
Alternatively, the vertical stress beneath the corner of a uniformly loaded rectangle is given by:
is an influence coefficient plotted in Figure 11.7b. The stress at points other than beneath the corner of the
rectangle can be obtained from linear superposition (i.e., addition and/or subtraction of influence coefficients).
For example, the stress beneath the centre of a rectangle with dimensions 2L by 2b is equal to four times the
stress beneath the comer of a rectangle with length L and width b.
m :::.L/z
n =b/z
0:= q
m = 0.1
~ c r  
o.00 t=::::::;o;;;;;,.o:::=::c:i:iI:::.-...I.-I.....I..J...u..LJLl...-....J-.I...&...1..I...I..UI
0.01 0.1 1 10
FIGURE 11.7 Vertical stress a
beneath the corner ofa uniformly loaded rectangle of
width b and length L on the ground surface subject to vertical pressure q
11.7 Obtaining Settlement Parameters
Selection of ground stiffness or compressibility parameters is an important step to estimate the settlement of shallow
foundations. For example, this may involve obtaining estimates of drained and undrained moduli (E " Vi, and E ) for
use in the elastic displacement methods presented in Section 11.3. Compressibility parameters for use in the uone_
dimensional e-loga' method (C
, and Cc) are normally obtained from the oedometer test (Section 11.4.1).
The soil parameters for input into any settlement calculations should not be viewed as constants but rather dependent
on many factors including: ground conditions, geologic setting, type of foundation (i.e., shallow or deep) and nature
of loading. Engineering judgement is required in the selection of stiffness parameters, and consequently, they should
always be selected by a qualified and experienced engineer.
Settlement of Shallow Foundations 171
Settlement parameters may be estimated using several different methods ranging from empirical correlations with
penetration tests (e.g., SPT, CPT), to laboratory tests on high-quality samples from the field (oedometer, triaxial
testing), to field testing directed at obtaining parameters for shallow foundations (e.g., plate load tests, measurements
of shear wave velocity). Becker (2001) provides a summary of available field (in-situ) testing methods. Often
settlement parameters are assessed using different methods to provide a bound on the parameters and to check for
consistency between values.
The extent ofthe testing involved in the selection of settlement parameters can be based on the risk ofthe foundation
project. For foundations projects oflow-risk (e.g., those involving few hazards with a low probability of occurrence
and limited consequences), parameter selection may be largely based on assessed values based on past local
experience or on empirical correlations with standard penetration test (SPT) blow count and/or cone penetration test
(CPT) tip resistance. Whereas for medium and higher risk projects, in addition to the use of empirical correlations,
laboratory testing on high-quality samples and/or specialized field (in-situ) testing are normally warranted.
Values of drained Poisson's ratio are not normally measured for most foundation projects, but rather estimated from
published values for similar soil and strain levels. Mayne and Poulos (1999) suggest that a Poisson's ratio between
0.1 < v' < 0.2 can be used for both fine and coarse-grained soil for the strain levels expected beneath shallow
Field penetration results from the Standard Penetration Test (SPT) blow count and/or Cone Penetration Test (CPT)
tip resistance are often available from the site investigation. Both of these penetration tests do not provide direct
measurements of soil stiffuess since they do not simulate the stress path or strain level of shallow foundations.
Empirical correlations exist between SPT results and modulus for coarse-grained soils (e.g., Berardi and Lancellotta,
1991) and may be used as an initial guide. CPT data is generally more reliable and reproducible compared to
the SPT. Available correlations between CPT tip resistance and modulus for coarse-grained soils can be found
elsewhere (e.g., see Baldi et aI., 1989). Modulus can also be inferred from field results from the flat dilatometer
test (DMT) or the pressuremeter test (PMT), e.g., see Lunne et aI. (1989), Marchetti et al. (2001). Regardless ofthe
testing procedure, such correlations should not be extrapolated to ground conditions different from those that they
were derived for (e.g., soil type, fines content, stress history, etc.). All empirical correlations need to be treated with
caution and adjusted as appropriate by experience.
Both drained and undrained moduli can be obtained from laboratory triaxial testing on high quality samples with
values selected over the appropriate stress range. The challenges of obtaining undisturbed samples ofcoarse-grained
soils often preclude laboratory testing on these materials for most foundation projects. Drained modulus for use in
three-dimensional calculations can also be estimated from the constrained modulus (D'=lIm) from oedometer test
results. For an isotropic elastic material the drained Young's modulus for three-dimensional conditions E' is related
to the one-dimensional constrained modulus D' by:
(l+v')(1 2v')D'
For fine-grained soils, correlations have been developed relating undrained modulus to undrained shear strength and
have been summarized by Lade (200 I).
For both fine- and coarse-grained soils, values of drained and undrained moduli can also be related to the small
strain shear modulus G . The small strain shear modulus is the same for static and dynamic loading, characterizes
both drained and undrained deformations, and is relatively insensitive to OCR of both sands and natural clays
(e.g., see Poulos et aI., 2001; Burland, 1989). The small strain shear modulus can be obtained from the shear wave
velocity Vs and total mass density of the soil P
= Pr Vs
Shear wave velocity can be measured in the field from seismic cone penetration test (sePT, e.g., see Lunne et
aI., 1997), or from cross-hole wave tests (ASTM D4428). Shear modulus G decreases from G
as shear strains
172 Canadian Foundation Engineering Manual
increase. Consequently, adjustments to modulus depending on level of stress or strain of the foundation can then
be made (e.g., see Fahey and Carter, 1993; Lehane and Fahey, 2002). Poulos et al. (2001) provide a summary of
findings on shear modulus dependence on strain level and propose a simple framework to incorporate these into
practical estimations of foundation settlement.
For fine-grained soils, undrained modulus E1/ can be obtained from shear modulus using:
E 3G (11.26)
For stiff overconsolidated clays, the drained modulus E' can be found from G using the relationship for an ideal
elastic material:
E'=2(1+v)G (11.27)
Equations 11.26 and 11.27 can be used to relate the drained and undrained moduli for an overconsolidated clay. For
soft compressible clays, the ratio of drained to undrained moduli may be much smaller than that derived from elastic
theory, with the ratio becoming smaller as the soil becomes more compressible.
11.8 Settlement of Coarse-grained Soils Directly fr9m In-Situ Testing
11.8.1 Standard Penetration Test (SPT)
a) Method of Peck et al. (1974)
Peck et al. (1974) provided an empirical chart that relates the design bearing pressure for a foundation on sand
with the results from Standard Penetration Test (SPT) N resistance, foundation width and foundation embedment
as given in Figure 11.8. The design bearing pressures from Figure 11.8 are expected to produce settlements smaller
than 25 mm. This figure can be used to estimate preliminarily geotechnical bearing resistance at serviceability limits
not exceeding 25 mm of total settlement.
SPT-N values need to be adjusted for depth (overburden pressure effects) using the relationship in Figure 11.8d
before using Figures 11.8 a-c.
A representative value of SPT-N should be used to a depth ofB beneath the foundation. This approach was developed
from field data gathered prior to the 1970s, thus N probably is for an energy ratio of 50-55 %. This approach was
also developed for conditions where the groundwater level is located deep beneath the foundation elevation. If the
groundwater level rises to the ground surface, no more than halfthe pressure values indicated in Figure 11.8 should
be used. For intermediate positions of the groundwater level (i.e., 0 < z :.; D+B) the design bearing pressure from
Figure 11.8 can be multiplied by the factor C w' given by:
z is the depth to the groundwater level and D is the depth to the underside of the foundation, both relative to
the ground surface.
Estimates ofdesign bearing pressure from Figure 11.8 are generally viewed as being conservative. Tan and Duncan
(1991) found that the results using the method of Peck et al. (1974) were not very accurate as they overestimated
settlements by an average factor of2.7 when compared with 76 cases involving shallow foundations on sand (with
B < 10 m). Although inaccurate, Tan and Duncan (1991) also found this approach to be reliable, as settlements were
underestimated in only 20 % of the 76 cases. Consequently with appropriate engineering judgement, the approach
of Peck et al. (1974) may be suitable for foundation design oflow risk projects and assessing geotechnical bearing
resistance at serviceability limit states not exceeding 25 mm of total settlement.
Settlement of Shallow Foundations  173 
b) Methodof BurlandandBurbidge(1985)

- 14  q
S  drainedsettlement(mm),
) beneaththefoundationadjustedtoenergy
B widthoffoundation(m),and
q averagepressureappliedtothegroundbythefoundation(kPa).
Inthisapproach,theSPT-Nvalueisadjustedtoanenergyratioof60% (e.g.,seeTerzaghieta1., 1996)butitisnot
necessarytomodifythevalue ofNforoverburdeneffects. Whenselectingthe averagepenetrationresistance, N
N'60 15+!(N6o-15) (11.30)
Ifthe thickness ofthe compressiblecoarse-grainedlayeris less than BO.
) the actual thickness can be substituted
forBO.7) inEquation 11.29.ForoverconsolidatedsandsBurlandandBurbidge(1985)foundthatthesettlementwas
approximatelyone-thirdof thatfornormallyconsolidatedsands.
TanandDuncan(1991)foundthatthemethodof BurlandandBurbidge(1985)wasmoreaccurate(overestimated
settlementsbyafactorof1.5)butlessreliable(underestimatedsettlements50% of thetime)thanthatof Pecketal.
(1974)whencomparedwith76casesinvolvingshallowfoundations onsand.
600 0

Q.)- 100
>  Q.)
III  o ....

>  UJ
0- 200
U  ....
N60 c:
.;: :t:
0.0 0.5 1.0 0.0 0.5 .1.0 1.5 0.5 1.0 1.5 2.0
Width of foundation,  B  (m) 
FIGURE 11.8 Designbearingpressureforfoundations onsandforsettlementnotexceeding25 mm 
basedonSPT-Nresultsfor: (a) DIB=l, (b)DIB=O.5, and(c) DIB=0.25. SPT-Nvaluefromfield 
to bemodifiedbyfactorCNgivenin (d)forusein(a)-(c). ModifiedfromPecketat. (1974) 
11.8.2 Cone Penetration Test (CPT)
Conepenetrationtest(CPT)tipresistanceqc canbeusedto estimatefoundationsettlementincoarse-grainedsoils
using the approach ofSchrnertmann et al. (1978). This approach uses a simple approximation ofelastic strain
distribution with the drained modulus obtained from the correlations with the CPT tip resistance. The sand is
dividedintoanumberoflayers(n) ofthicknessfj.z downtoadepthbelowthebaseofthefoundationequalto2Bfor
asquarefootingand4B forastripfooting(lengthoffooting,L >  lOB). Arepresentativevalueofqc is assignedto
eachlayer. Thesettlementisthengivenby:
174 Canadian Foundation Engineering Manual

S =C



t3.q  2: 
factor to  allow for  strain relief from  embedment, 

1-0.5J£  (11.32)
factor to  account for creep and cyclic loading,  C

( t  \
I + 0.210g
01.33) -j

factor to  account for foundation shape, 
1.03  0 . 0 3   ~ )  2':  0.73, for strip foundations 
1.0, for circular and square foundations, 
Ilq  net foundation pressure = q - q:, 
q  average pressure applied to the ground by the foundation, 
initial vertical effective stress at foundation depth D, 
t  time since load application in years, 
I  strain influence factor (see Figure  11.9), 

Ilz.  thickness of layer i,
modulus of the sand for layer i  , 
3.5Qc  for strip footings  (LIB>  10), or  (11.35a) 
=  2.5Qc  for square or circular footings  (LIB = 1), and  (l1.35b) 
average CPT tip resistance for each layer. 
The triangular distributions used to approximate the vertical strains with depth are given in Figure  11.9. 
The peak value ofthe strain influence factor (I )  occurs at a depth (zDP)  ofBI2 beneath square or circular foundations 
and a depth of B beneath strip foundations, and has  a value given by: 
q ~ is the initial vertical effective stress  at the depth corresponding to the peak value of strain. 
A useful refinement to  this  method would be to  use  the  actual strain distribution beneath the foundation  given  by 
elastic  theory  in  Section  11.3  instead  of the  triangular  approximation,  which  could  be  readily  programmed  in  a 
spreadsheet for easy calculation (e.g., see Mayne and Poulos,  1999). 
The modulus values  obtained with  the  correlation with  qc  are  reasonable  for  recent normally  consolidated sands. 
Estimates  of sand modulus  from  qc  can  also  be  obtained  from  Baldi  et  al.  (1989)  as  a  function  of the  degree  of 
loading, soil density,  stress history, cementation, age, grain shape and mineralogy. These correlations suggest ratios 
of E'lqc from 2 to 4  for recent,  normally consolidated sands;  4 to  6 for  aged (>1000 years), normally consolidated 
sands; and 6 to 20 for overconsolidated sands. 
Settlement of Shallow Foundations 175

B Zo ,
....................t ..t. qp

0.0 0.2 0.4
0.8 1.0
FIGU,RE 11.9 Influence factor Izfor estimating settlement offoundation on sand
using Schmertmann smethod (modifiedfrom Schmertmann et al. 1978)
11.9, Numerical Methods
.JIletbods.provide the opportunity to. model complex. ground conditions (ifknown). It is also possible to model
ofthe foundation (e.g., bending moments, shearforces, deflections) for use in structural design. More elaborate
ofsoilnonlinearityoryielding; however,thisrequiresknowledgeoftheconstitutiverelationship andtheabilityto
Numericalmetho,dsmaypemore appropriateformediumandhigh-riskfoundations wherethereis sufficientdata
available tojustify more elaborate analysis. Additionally the finite elementmesh orfinite difference grid should
thereisnegligiblechangeinthenumericalsolution. Considerationmustalsobegiventotheselectionof boundary
11.10 Creep
", For SQils,laboratoryandfielddatasuggestthatcreep(Le.,secondarycompression)displacementsoccur
sim4lt,aneously with primaryconsolidation(Leroueil, 1996). Formostpracticalcases involvinglow compressible
C!(1+e)<0.25, duringprimary doesnotneedtobeexplicitlycalculated.Consequently,
',.creepsettlements are addedto the,total final settlementto accountfor displacements ofthe foundation whenthe
effective  (i.e.,attheendof primaryconsolidation).
,Foundationdisplacementsfromsecondarycompressionattimet canbeestimatedfrom:
Sse = Ho I
176 Canadian Foundation Engineering Manual
C secondary compression index in terms of void ratio, 
duration of primary consolidation, and 
thickness of compressible layer. 
Values  of C  may  be  obtained  from  the  oedometer  test.  Often  a  reasonable  estimate  for  normally  consolidated 
. a
inorganic clays and silts is equal to  0.04C
'  with values for  other ground types reported in Terzaghi et ai.  (1996). 
For highly  compressible  clays  with  C /(1 +e ) >  0.25,  viscous  effects  may  contribute  to  foundation  displacements 
during the time frame  of primary consolidation and may be estimated as discussed by Leroueil (1996). 
11.11 Rate of Settlement
The rate ofsettlement may be of importance for foundations on fine-grained soils and depends on-how quickly excess 
pore pressure can dissipate. Generally the rate of settlement depends on the type ofsoil,hydraulic conductivity ofthe 
soil, and drainage boundary conditions. The rate of settlement is  quantified by the average degree of consolidation 
U for use in Equation  11.2  and may be obtained using one- or three- dimensional consolidation theories depending 
on the foundation  conditions. 
11.11.1 One-Dimensional Consolidation
One-dimensional consolidation theory of Terzaghi (for details see Terzaghi et aI.,  1996) assumes that pore pressures 
can  dissipate  only  in  a  vertical  direction  (Le.,  there  is  no  lateral  flow).  It may  be  used  to  estimate  the  rate  of 
settlement for foundations where the assumption of one-dimensional drainage maybe reasonable (e.g., foundations 
where the surface load is large relative to the layer thickness). 
The average degree of consolidation U obtained from Terzaghi's one-dimensional theory is plotted in Figure  11.10 
versus dimensionless time  factor,  T,
one-dimensional coefficient of consolidation, 
t time, and 
H drainage path of the consolidating layer. 
The one-dimensional coefficient of consolidation is  normally obtained from oedometer results for load increments 
taken over the  appropriate  stress  range  (for the  graphical  procedures  see  Holtz and Kovacs,  1981)  and  may  also 
be  estimated  from  in-situ cone  penetration tests  (CPT)  with  pore pressure  measurements  (e.g.,  see  Lunne  et al., 
The drainage path H relates to the boundary conditions above and below the consolidating layer.  Conditions of two-
way drainage exist  if excess  pore pressures can dissipate at the top  and bottom of the  consolidating layer and the 
drainage path would be equal to  one-half of the thickness of the consolidating layer: One-way drainage conditions 
exist if the excess pore pressures can only dissipate to one of the layer boundaries and  is  equal to the thickness of 
the  consolidating layer.  Figure  11.10 may be used for conditions  involving two-way drainage with initial  linearly· 
distributed  excess  pore  pressures  and  for  one-way  drainage  where  the  initial  excess  pore  pressures  are  uniform 
throughout the consolidating layer. 
The average  degree of consolidation in Figure  11.10  was  obtained assuming that the  foundation load was  rapidly 
applied and then held constant. An estimate of the influence of gradual loading on the rate of consolidation is  given 
by Terzaghi et al.  (1996). 
Settlement of Shallow Foundations 177
0.001 0.01 0.1 1
FIGURE 11.10 Average degree ofconsolidation for one-dimensional conditions.
Modified from Tergazhi et al. (1996)
11.11.2 Three-Dimensional Consolidation
For many practical foundations, lateral flow ofwater will occur and consequently Terzaghi's one-dimensional
consolidation solution will underestimate the rate ofsettlement with time. Other factors being equal, smaller
foundations willsettlefastergiventheabilityofexcessporepressurestodissipatelaterallyandvertically.
The approximatesolutionsofDavisandPoulos(1972.) maybeusedtoestimatethedegree ofsettlementfortwo-
andthree-dimensional drainage. Alternatively, usingthe solutions ofDavisandPoulos (1972),the coefficient of
c =R c  (11.39)
ve f v
one-dimensionalcoefficientofconsolidation(e.g.,obtainedfrom odeometerresultsoverthe
modifiedcoefficientofconsolidation(foruseinFigure 11.10),
R = modificationfactortoaccountforthree-dimensionaleffects.
_  FactorRf(i.e., c
/ c)isplottedinFigure 11.11 andispresentedforbothstripandcircularfoundationsandforthree
PT permeabletopsurface,
PB permeablebottomsurface,
IF impermeablefoundation, and
IB impermeablebase.
Square foundations can be approximated as a circle. An approximation for rectangular foundations is given by
DavisandPoulos(1972). Modifications to accountfor anisotropicpermeabilityofthe consolidatingsoilare also
      .. ---
178 Canadian Foundation Engineering Manual


1 ()




()  10 

FIGURE 11.11 Equivalent coefficient ofconsolidation c 
for use in  one-dimensional rate ofsettlement 
analysis to  account for three-dimensional effects for:  (a)  strip foundation ofwidth B  and (b)  circular foundation 
ofdiameter B  on a uniform layer ofconsolidating soil ofthickness h.  Modified from Poulos  (2000) 
11.11.3 Numerical Methods
The rate ofsettlement can also be obtained by employing numerical methods. This approach may only be appropriate 
for  medium and higher risk projects  where  there  is sufficient data  to  warrant more  elaborate  amilysis.  Numerical 
methods can solve the equations ofBiot (1941) to calculate both changes in stresses and pore pressures in response to 
applied loads. More realistic constitutive models can be used for the soil to characterize effects such as the decrease 
in hydraulic conductivity with decreases in void ratio during consolidation, as well. as possible viscous effects of the 
soil.  Either finite  element or finite  difference numerical  approximations may be  employed. It is  important to have 
sufficient refinement of finite  element mesh or finite difference grid and sufficiently small time increments to  avoid 
numerical errors. 
11.12 Allowable (Tolerable) Settlement
Foundation  deflections  need  to  be  limited  to  allowable  levels  to  ensure  adequate  serviceability  of the  structure. 
Figure  11.12 illustrates the types of limiting deflections that need to be considered to avoid damage to the structure. 
An overlying structure experiences no additional structural loads from a uniform vertical deflection ofthe foundation 
(Figure  11.12a).  However, limits on the  total settlement of the  structure are required to prevent damage to  services 
2  3 4  5 
1  2 3 4  5 
Settlement of Shallow Foundations  179 
connected to  the  building (e.g.,  gas  lines, water and sewer pipes).  Differential settlements refer to  the case where 
one  portion  of the  foundation  settles  more  than  that  at  other locations.  Differential  settlements  will  occur from 
differences in loads applied to the foundation and/or from the natural variability ofthe ground beneath the foundation 
(e.g.,  from  variations  in  thickness,  presence  and stiffness  of a  compressible layer,  depth to  bedrock).  For framed 
strUctures  (Figure  ll.l2b), limiting differential settlements are defined  in  terms of an allowable angular distortion, 
which is equal to  the  differential  settlement divided by the  distance over which the  differential settlement occurs. 
For unreinforced load bearing walls and panels (Figure  1l.l2c and d),  allowable settlement to  limit cracking of the 
wall is  expressed as  a  deflection ratio,  which is  equal  to  the  relative  sag or hog divided by the  length of the  wall. 
Overall and local tilt of the  structure may also need to be limited (Figure  11.l2e). 
Span between  Span between 
(a)  columns  (b)  columns 
IE  )oj 
IE  )It 
  ,--- - -....i

... ........t S  t......... 

I  I  I 
Wall length 
r--- -
I .......... 
----;   ..  "'--1 

.  l 
c::- _  ....t  - __.I  S
.. -..-'" 
I  I  I 

Angular distortion  =
Deflection  ratio  =L l\S h 
.............-....................... _............ __........ _......._-: 
______  Wall  length 
Local tilt 
FIGURE 11.12  Illustration oftypes of tolerable settlements for shallow foundations. 
Dashed lines  indicate undeflected position ofstructure.  Modifiedfrom Burland and Wroth  (1974) 
Tolerable limits on foundation deflections listed in Table 11.1  may be used for low risk projects and as an initial guide 
for higher risk projects.  For higher risk projects, consideration should be given to (Boone,  1996): the configuration, 
flexural and shear stiffness ofthe building sections; nature of the ground deflection profile; location of the structure 
relative to the deflection profile; and possible slip between the foundation and the ground.  The values cited in Table 
11.1  and in  Boone  (1996)  provide realistic  estimates of tolerable  settlement.  They should not, however,  preclude 
specific structural assessment of tolerable settlement of a given building or structure.  Communication between the 
structural  and  geotechnical  engineer  is  encouraged  to  address  adequately  appropriate  serviceability  limit  states 
180 Canadian Foundation Engineering Manual
TABLE 11.1 Guidelines for Limiting Settlement ofFramed Buildings and Load Bearing Walls
(adaptedfrom Poulos et al., 2001)
Type Of Damage Criterion Limiting Value
Structural damage Angular distortion 11150 - 1/250
Cracking in walls and partitions Angular distortion
111000 111400: end bays
Visual appearance Tilt 1/300
Connection to services Total settlement
75 mm: sands
135 mm: clays
Cracking by relative sag
Deflection ratio
112500: walliengthlheight= 1
111250: walllength/height=5
Cracking by relative hog' Deflection ratio
1/5000: walliengthlheight=l
1/2500: walllengthlheight=5
, For unreinforced load bearing walls.
Drainage and Filter Design 181
Drainage and Filter Design
12 Drainage and Filter Design
12.1 Introduction
must provide, overthe service life ofthe structure, a means for the collectionand discharge ofwaterthatwould
otherwise impairits performance. Thedetrimental effectsofwateronsubsurfacefacilities are manifestedinways
• the ingressandpresenceof waterinlocationsthatwereintendedtobedry;
• the impactofdissolvedsalt,whichis corrosivetoPortlandcementconcrete;and,
a reduction ofshearstrength in the soil as the effective stress diminishes inresponse to increasingpore
Drainage pipes are used to collect and remove subsurface water. The pipes must have structural, hydraulic and
while adequately conveying the inflow. Perforated or slotted drainage pipes, into which water seeps, must be
12.2 Filter Provisions
Filtermaterials,forexampleoneormore specifiedgradationsofcoarse-grainedsoil, oralternatively ageotextile,
Accordingly,thefiltrationprocessitselfispredicatedonthedevelopment, overtime,ofastableinterfacebetween
basesoilandfiltermaterial. Geotextilefilters areaddressedseparatelyinChapter23.
Agradedgranularfiltershouldsatisfythe followingperformancerequirements:
1. The voids ofthe filter should be small enough to restrict particles ofthe base soil from penetrating or
washingthroughit, fulfillingacriterionof"soilretention."
2. Thefiltermaterialshouldbemoreperviousthanthebasesoil,fulfillinga"permeabilitycriterion."
3. Thefiltershouldbe sufficientlythickto ensurearepresentativegradationthroughout.
4. Thefilter shouldnotsegregateduringprocessing,handling,placing,spreadingorcompaction.
5. Thefiltermaterialshouldbephysicallydurable, andchemicallyinert.
6. Thefiltershouldnotbesusceptibletointernalinstability,wherebyseepageflowactstoinducemigrationof
thefinerfractionof thegradation.
7. Thefiltergradationshouldbecompatiblewiththesize,locationanddistributionofopeningsinthedrainage
182 Canadian Foundation Engineering Manual
12.3 Filter Design Criteria
Perfonnance requirements are addressed by a series of design criteria. The criteria are empirical, having been
established from interpretation of experimental observations, with occasional consideration of theoretical analysis
and practical constraints. They are founded on observations of steady unidirectional flow and, accordingly, are
appropriate to such conditions in the field. In describing the base soil, its grain size distribution should be determined
by wet sieving and without the use of a dispersing agent: the fines fraction so obtained is believed representative
of that encountered by the filter (GEO, 1993). Reddi (2003) provides a concise summary of filter requirements in
drainage applications, and many of the related design criteria, including a series of worked examples.
12.3.1 Retention Criterion
The pore size distribution of the filter is strongly influenced by its grain size distribution. A pore size that is
sufficiently small will restrict the passage of finer grains through the filter. Retention of the base soil is therefore
achieved through specifying a maximum value for the ratio of a characteristic grain size of filter CD,) to grain size
of base soil (d
)' Laboratory testing of Bertram (1940), Karpoff (1955) and Sherard et al. (l984a) confirm the
general suitability of a criterion first advocated by Terzaghi in the design of drains for embankment dams, where:
In a minor variation to the criterion, these studies have led to the recommendation (GEO, 1993) in current practice
that filters comprising sands and gravels (D
1arger than about 1.0 mm) satisfy:
Either criterion provides a suitable margin of safety against inadequate retention, the onset of which has been noted
to occur at a ratio of D15 /d
in excess of ten.
For base soils comprising clays, Sherard et aL (l984b) recommended a sand filter with a DIS of 0.5 mm. For sandy
clays and silts, the filter criterion D
< 5 is reasonable and conservative.
12.3.2 Permeability Criterion
A pore size that is sufficiently large will promote unimpeded flow of water from the base soil, through the filter.
Adequate permeability of the filter is therefore achieved through specification of a minimum value for the ratio of a
characteristic grain size of filter (D IS) to grain size of base soil (diS)' Terzaghi first advocated a ratio for base soils,
Recognizing that permeability is, to some extent, a function of the square of the Djd
ratio, a relative permeability
of about 25 is implied by the recommendation for current practice (GEO, 1993) that:
Drainage and Filter Design 183
12.3.3 Other Design Considerations
The  following  suggestions  are  made,  based  on  experience  reported  in  the  literature,  to  address  additional 
considerations arising from the requirements of a filter: 
The filter should be sufficiently thick to ensure a representative gradation in the region ofinflow. Accordingly, 
the minimum thickness  is  strongly  influenced by the  size  of the  larger grains.  While no  specific  criterion 
exists,  it  is  suggested the filter  be  at least  300  mm thick,  to  ensure  a reasonably  consistent distribution  of 
The filter  should not segregate  adversely  during  processing,  handling,  placing,  spreading or compaction. 
Experience shows that susceptibility to  segregation increases with the range in grain size, and the maximum 
particle size. The phenomenon is therefore limited by imposing an upper limit on the coefficient ofuniformity 
)' and it  is  suggested that C < 20 and D < 50 mm. 
u 100 
The filter material should be physically durable, and chemically inert. Accordingly, consideration should be 
given to  the mineralogy ofthe filter material,  and its  compatibility with the pH of the subsurface water. 
The  filter  should  not  be  susceptible  to  internal  instability,  whereby  seepage  induces  a  migration  of the 
finer  fraction of the  gradation. Experience shows that internal  instability is  most likely in s'Oils  that have  a 
gently inclined  gradation  in  the finer  fraction  of the  grain  size distribution,  and  in  soils  exhibiting  a  gap-
gradation. Kenney and Lau (1985,  1986) postulate a boundary to  internal instability based on the shape of 
the gradation  curve  over  its  finer  fraction:  the  increment  of mass  fraction  (H),  over a  designated range of 
grain size (D to  4D) beyond a point on the grading curve (F),  defines a ratio  H/F that is  deemed indicative 
of potential  instability when  HlF > 1. It complements  an  earlier  approach  (Kezdi,  1979) based  on  a split 
gradation analysis, and the principle of soil retention of the finer fraction by the coarser fraction. 
The  filter  gradation  should  be  compatible  with  the  size,  location  and  distribution  of perforations  in  the 
drainage pipe.  For steady unidirectional flow,  experience  suggests D85  should not  exceed the  diameter  of 
circular openings, and D70  should not exceed the width of slot openings. 
12.4 Drainage Pipes and Traps
Drainage  pipe must be installed  at a  slope  that is  sufficient  to  induce  a  flow  velocity capable  of transporting any 
fine  grains  that  wash  in through  the  openings  of the pipe.  The  minimum  slope is  1  %.  It is  important  that  traps 
be  installed,  which cause the flow  to  change and result in  deposition  of suspended  solids  at  locations that  can be 
accessed for purposes of inspection and cleaning.  The use of valves may be necessary to  ensure flow  occurs in the 
desired direction, and to prevent the possibility of a back-flow in the drainage system. 
12.4.1 Construction of Subsurface Drains
Key elements in the configuration ofa perimeter drainage system for a shallow foundation are illustrated schematically, 
for three scenarios, in Figure 12.1. It is important to slope of the base of the trench away from the footing,  to slope 
the wall ofthe trench such that minor sloughing is avoided during placement ofthe drain, and to direct surface water 
away from  the trench itself.  Intended use  of the  structure determines the need for  damp-proofing the  outside  face 
of the wall. A  geotextile may be used to separate the foundation soil from that of the filter and backfill:  experience 
does not support wrapping geotextile around the drainage pipe, due  to the concentration of flow.  It is important to 
locate the  invert of the  drainage  pipe below the  top  surface  of the basement floor  slab.  Where concern exists  for 
integrity of the footing, and the efficiency of its bearing action, the invert of the drainage pipe should not be located 
below the elevation of the footing.  . 
184 Canadian Foundation Engineering Manual
FIGURE 12.1 Typical Sections Showing Arrangement ofSubsurface
Perimeter Drains around Shallow Foundations
(1) perforated or slotted pipe placed about 300 rnm below the upper level of the basement floor slab;
(2) unperforated drain pipe connected to appropriate trap and backwater valve before connecting to a sewer. The
trap shall have provisions for inspection and cleaning;
(3) filter material that is compatible with the grain size characteristics of the fine-grained foundation and backfill
soils, as well as with the perforations of the pipe;
(4) filter material continuously or intermittently placed next to the foundation wall to intercept water from.window
wells and from low areas near the building (see also 6);
(5) damp-proofing on wall- optional depending on the quality of the concrete wall;
(6) optional use of sheet drain, or synthetic filter blanket, next to the foundation wall to replace the soil filter
according to (4);
(7) foundation and backfill soils, which may contain fine-grained and erodible materials; and
(8) "topping-off' material sloping outward to lead off the surface water. It is usually desirable to use low
permeability soil to reduce the risk of overloading the pipe.
. .---'"
Frost Action 185
Frost Action 
13 Frost Action
13.1 Introduction
The Canadian climate results in freezing of the near-surface ground for several months each winter almost
everywhere in Canada. The depth of seasonal frost penetration ranges from minimal to several meters, depending
upon local climate, soil conditions and snow cover. Ground freezing frequently results in volumetric expansion of
the soil which causes heaving of structures located above or adjacent to the freezing soil. Thaw during the following
spring will release the excess water, usually causing loss of strength or complete collapse of the soil structure. This
natural seasonal process can be very damaging to infrastructure, such as roads and buried pipelines, and may also
cause serious problems for buildings (Crawford, 1968; Penner and Crawford, 1983).
This chapter provides a description of the phenomenon of frost heave, its causes and a brief summary of current
predictive capabilities. Guidance is provided for simplified prediction offrost penetration and selection of mitigative
design measures. The comments are not intended to deal with structures on a permafrost foundation. A thorough
understanding ofthe nature and distribution offrozen soil is required to predict soil behaviour in permafrost regions.
The reader is referred to a comprehensive treatment of this more complex topic such as found in Brown (1970),
Andersland and Anderson (1978), Johnston (1981) and Andersland and Ladanyi (2004).
13.2 Ice Segregation in Freezing Soil
Water in soil pores begins to freeze as the temperature is lowered through OCC. Figure 13.1 illustrates the progressive
reduction ofunfrozen water content as the relative proportions of water and ice change at sub-zero temperatures for
sand, silt and clay. Continued formation of ice in the soil pores at progressively decreasing temperatures confines
the remaining water to progressively smaller pore spaces. A pressure differential between the ice and water phases
draws water from the unfrozen soil into the freezing soil. Fine-grained soils, which freeze over a broader range of
temperature, are particularly susceptible to moisture migration along a pressure gradient, resulting in growth of ice
lenses. The resulting heave rate and magnitude depend upon soil type, overburden pressure, groundwater conditions,
freezing rate, and other factors. The extent of ice lensing that can occur in a clay soil is illustrated in Figure 13.2.
Where restraint in the form of a building is present, heaving pressures develop that mayor may not be able to
overcome the restraint. Heaving pressures may be very high, depending upon the restraint offered by the surrounding
structure and soil; values equivalent to 1800 kPa were measured on a 300 mm diameter plate (Penner and Gold,
186 Canadian Foundation Engineering Manual
-1 -2 "':3 -4 -5
FIGURE 13.1 Unfrozen water contentfor a range offrozen soils (qfter Williams and Smith, 1989)
FIGURE 13.2 Sample offrozen clay showing ice segregation
Frost Action 187
The  rate  of heaving  in  a  frost  susceptible  soil  is  limited  by  the  rate  of heat  extraction  from  the  freezing  fringe 
where water is  migrating to feed  growing ice lenses. This complex heat and moisture flow phenomenon is normally 
uncoupled to simplify engineering predictions.  Penetration of the freezing  isothenn with  time and temperature  is 
predicted  first by ground thermal  analyses  without consideration to  the  impact of moisture  redistribution  and  ice 
lensing. The predicted extent of frost penetration and knowledge ofthe thermal gradients that exist within the frozen 
soil are then used as inputs for prediction of heave magnitudes due to ice segregation. 
Engineering methods for predicting ground thermal conditions and frost heave have evolved significantly in the past 
decade such that practical solution techniques are now available. The remainder of this chapter summarizes current 
practice in this evolving field together with some practical considerations for mitigating frost heave damage. 
13.3 Prediction of Frost Heave Rate
13.3.1 Ice Segregation Models
Several  hydrodynamic  models  have  been  developed  to  express  the  coupled  heat  and  moisture ·flow  that  cause 
frost  heave.  These models have been reviewed by Nixon (1987,  1991) to evaluate their applicability for practical 
engineering predictions. 
Ice lenses  grow  within the  frozen  fringe  where  the  temperature  is  less  than  ODe (Miller,  1978).  The temperature 
of the growing ice lens is  related to  the overburden pressure (Konrad &  Morgenstern,  1982). Ice also forms  in the 
larger pores between the active ice lens and the ODe isotherm, requiring water to flow through the fringe ofpartially 
frozen soil to feed  the growing lens. The rate of lens growth is dependent upon the finite hydraulic conductivity of 
the partially frozen  fringe  and the rate  of heat extraction at the ice lens. All hydrodynamic models therefore relate 
the velocity of water through the freezing fringe to the temperature gradient, and to the permeability ofthe partially 
frozen soil. The heave rate can be computed from the rate of change of the velocity of water in the frozen soil. 
A practical method for predicting frost heave magnitude for geotechnical engineering applications was developed by 
Konrad and Morgenstern (1980). Their semi-empirical formulation does not rely on measurement ofthe permeability 
of frozen  soils  or other physical parameters that characterize  the  movement of water through the  freezing  fringe. 
They relate the water velocity directly to  the thermal gradient in the frozen  soil. The constant of proportionality is 
termed the segregation potential (SP).  The SP parameter is dependent upon overburden pressure but is  considered 
to  be  independent  of the  rate  of cooling  in  the  freezing  fringe  at  low  cooling rates.  The  SP parameter must  be 
determined  from  a  series of step temperature  freezing  tests  carried out at various  overburden pressures.  The tests 
must reasonably simulate the freezing rates or thermal gradients expected in the field. 
The heave rate  (dh/dt) under field conditions can be predicted from: 
dh/dt = SP G
+ 0.09 n dX/dt (13.1) 
SP is  the segregation potential determined from freezing tests 
is  the thermal gradient in the frozen soil at the freezing fringe,  determined from geothermal 
dX/dt is the rate of advance of the frost front  determined from geothermal simulations, 
n is the soil porosity reduced to account for the percentage of in-situ porewater that remains frozen 
within the anticipated range of ground temperatures. 
A summary of published data relating the  SP parameter to  overburden pressure for various  soils was presented by 
Nixon (1987), and is  shown in Figure  13.3. 
188 Canadian Foundation Engineering Manual

« LJl 200


<3 E
w E 20
(fl 10
o 50 100 150 200 250 300
FIGURE 13.3 Published segregation potential (SP) parameter data (after Nixon, 1987)
13.3.2 Frost Susceptibility
Frost susceptibility of soils refers to the propensity of the soil to grow ice lenses and heave during freezing. At
present, there are no precise criteria for classifying soils according to their frost susceptibility. A common guideline,
developed by Casagrande (1932) based on observation and experience, relates frost susceptibility of soils to the
percentage offine fraction less than 0.02 nun.
The Casagrande guide has been extended by the U.S. Corps of Engineers to a widely used classification system,
shown in Table 13.1. Soils are listed in four categories, F 1 to F4, in approximate increasing order offrost susceptibility
and loss of strength during thaw.
Where frost susceptibility and heave are critical parameters in foundation design, laboratory frost heave testing
should be carried out. There are no current standards for heave tests; thus, it is important to develop a test program
that meets the requirements of the project. This may range from simple confirmation of frost susceptibility and
heave rate to determination of specific parameters such as segregation potential (SP) that can be used in a frost
heave prediction model.
Frost heave tests are carried out in an insulated freezing cell where precise control can be maintained over
temperatures. A sub-zero temperature is applied to the upper or lower sample cap. The other end ofthe sample may
be uncontrolled, insulated or maintained at some positive temperature. The end temperatures might be controlled
either as a step temperature change or a time-dependant "ramped" temperature change. The ramped temperature
change is chosen if a near-constant freezing rate is desired. The volume of free water drawn into the sample at the
unfrozen end cap is measured with time and related to the volumetric increase or sample heave rate. An interpretation
of frost heave test data in terms of segregation potential is described by Konrad and Morgenstern (1981).
Frost Action 189
TABLE  13.1  u.s. Corps ofEngineers Frost Design Soil Classification
Fl Gravelly soils  3 to  10  GW,  GP,  GW-GM, GP-GM 
F2 a) Gravelly soils  10  to  20  GM,  GW-GM,  GP-GM 
b) Sands  3 to  15  SW,  SP,  SM,  SW-SM, SP-SM 
F3 a)  Gravelly soils  >20  GM,GC 
b) Sands, except very fine  silty sands  >15  SM,SC 
c) Clays, PI >12  CL,CR 
F4 a) All silts  ML,MR 
b) Very fine  silty sands  >15  SM 
c) Clays, PI <12  CL, CL-ML 
d)  Varved  clays  and  other  fined- CL and ML; CL, ML, and SM;  CI,  CR, 
grained, banded sediments  and ML;  CL, CH, ML, and SM 
13.3.3 SP from  Soil  Index Properties 
A comprehensive  study conducted  by Komad (1999)  established that the segregation potential  parameter (SP)  of 
saturated fine-grained soils can be adequately related to  a few basic soil index properties. For a soil freezing under 
zero applied  overburden pressure, a reference value  of the  segregation potential, SP0'  is  best empirically related to 
the mean grain size of the fines  fraction «0.075 nun), dso(FF),  the specific surface area of the fines fraction,  Ss'  and 
the ratio  of water content to  the  liquid  limit,  W/WL as  illustrated by Figure  13.4. For a ratio W/WL close to  0.7, the 
empirical relationship for clayey silts is: 
where  dso(FF)  is  expressed in  !lm. 
In  well-graded  soils  or gap-graded  soils,  SP  is  directly  proportional  to  the  relative  fines  content,  i.e.  the  ratio  of 
actual fines  and the amount of fines  needed to fill  a11  the  pore space between the coarser-grained particles.  Details 
on a complete frost-susceptibility assessment methodology is  given in Konrad (1999). 
Frost susceptibility assessment was  recently extended to non-clay  soils  such  as  tills  and  crushed rock  by Konrad 
190 Canadian Foundation Engineering Manual
Oi 200
w/wL = 0.7 ± 0.1
o Rieke et aI., 1983
0.65 < w/wL < 0.8
&W/WL> 0.7
• w/wL '" 0.7
dso(FF) (1J.lT1) .
FIGURE 13.4 Frost susceptibility assessment was recently extended to
non-clay soils such as tills and crushed rock by Konrad (2005)
13.4 Frost Penetration Prediction
13.4.1 Ground Thermal Analyses
The dominant mechanism of heat transfer in soils is thermal conduction. Heat flow in the ground follows Fourier's
Law of conduction with a term to account for the release or absorption oflatent heat of water during phase change.
Heat transfer by mechanisms other than conduction may only be a factor in porous soils where groundwater
flow is occurring. Water velocities generally must exceed 10-
cm/s before convective heat flow starts to become
Analytical methods, or closed-form mathematical solutions of the well-known Laplace equation, can provide an
approximation of seasonal frost penetration for simple conditions. Prediction of transient ground temperature
changes for problems with complex stratigraphy and variable boundary conditions requires solution by numerical
methods. Numerical models in common use are either finite difference or finite element solutions. A comprehensive
review of numerical methods for ground thermal regime calculations has been provided by Goodrich (1982). Two
numerical models in common use in Canada are described by Nixon (1983) and by Hwang (1976).
Numerical methods are required for geotechnical design calculations other than simple prediction of the maximum
depth of frost penetration. The usual range of problems involves layered systems, temperature-dependent thermal
properties, and time-dependent boundary conditions such as ground surface heat exchange. A realistic simulation
of the temperature-dependent liberation or absorption oflatent heat during freezing or thawing, associated with the
changes in unfrozen water content shown in Figure 13.1, is also an essential feature in any numerical simulation.
Frost Action 191
Numerical  methods  are  very  flexible  and  can  reasonably  simulate geotechnical complexities  in  either one  or  two 
dimensions. However, they require familiarity with an appropriate computer program and experience deriving input 
parameters.  The  results  are  normally  expressed  as  temperature  isotherms  on  a  two-dimensional  plot  for  various 
times  of interest to  the  designer.  The  results  can also  be  expressed as  a propagation of the  freezing  isotherm with 
time or as a transient thermal gradient which may be input to a subsequent prediction offrost heave in an uncoupled 
analysis  of heat and moisture flow. 
13.4.2  Simplified Solutions for Maximum Frost Penetration Neglecting Frost Heave
Frost  penetration  is  proportional  to  the  square  root  of time  for  a  step  change in  ground  surface  temperature.  The 
most  useful  form  of the  relationship  is  the  modified  Berggren  equation  as  described  by Aldrich  (1956),  Sanger 
(1963) and Johnston (1981),  and shown as  Equation  l3.3: 
X depth of frost penetration 
surface freezing  index which can be estimated from the air freezine  index times a ground 
surface interface factor "n" 
Thermal conductivity of the frozen soil 
Volumetric latent heat of the soil 
A dimensionless coefficient (Figure 13.8) 
The surface  freezing  index expresses the average negative  surface temperature  and the time over which it applies. 
The empirical n-factor can be  used  to  determine  surface freezing  index  from  the  air-freezing  index.  Published n-
factors  for  various  types  of surfaces  are  shown in Table  13.2. The air-freezing  index is  a  summation of the  daily 
mean  degree-days  for  the  freezing  period.  A  long-term mean  (30 year)  air freezing  index can  be  estimated from 
monthly mean air temperature data published by Environment Canada. Typical variation in air freezing index within 
Canada is shown in Figure 13.5. 
TABLE 13.2 Values ofn-Factorsfor Different Surfaces (from Johnston, 1981)
Surface type Freezing-n
Spruce trees, brush, moss over peat - soil surface  0.29  (under snow) 
As  above with trees cleared  soil surface  0.25  (under snow) 
Turf  0.5  (under snow) 
Snow  1.0 
(most probable range) 
0.6  1.0 
(0.9  0.95) 
Asphalt pavement  0.29 - 1.0  or greater 
(most probable range)  (0.9 - 0.95) 
Concrete pavement  0.25  0.95 
(most probable range)  (0.7 -0.9) 
192 Canadian Foundation Engineering Manual
Winter  air  temperatures  vary  substantially  from  year  to  year  everywhere  in  Canada.  Therefore,  it  is  seldom 
appropriate to  use the  long-tenn mean air-freezing index for design purposes. 
Common  practice  is  to  choose  some  return  period  or  recurrence  interval  and  to  estimate  the  most severe  winter 
likely to  occur within that period. The US  Corps of Engineers method, as described by Linell et aL  (1963), is to use 
either the most severe winter of the previous ten years or the average of the three most severe winters in the previous 
30 years. 
A  simple  relationship  between  design  freezing  index,  taken  as  the  coldest  winter  over  the  last  lO-year  period, 
and  mean  freezing  index  was  developed  by  Horn  (1987)  by  curve  fitting  data  for  20 cities  across  Canada.  The 
relationship is  given as: 
Id  = 100 + 1.29 I  (13.4)
Design Freezing Index (DC-days) 
Mean Freezing Index (OC-days) 
SA$H)  ON 1lf'. PERIOD 1m TO l'i61l  
FIGURE 13.6 Thermal conductivity offrozen coarse-grained soil (after Kersten, 1949)
Frost Action  193 
This  relationship  is  recommended  for  the  design  air  freezing  index  in  the  absence  of an  in-depth  evaluation  of 
historical climate data.  The surface freezing  index for the modified Berggren equation then becomes: 
The  thermal  conductivity  of soil  can  be  estimated  from  relationships  to  soil  index  properties.  The  relationships 
developed by Kersten (1949) for frozen coarse and fine-grained soils are shown in Figures 13.6 and 13.7, respectively. 
Frost penetration  depths  based  on Kersten's  relationships  for  coarse-grained  soils  may  under  predict  frost  depth 
significantly for unsaturated soils. 
The thermal conductivity of coarse-grained soils is also dependent on soil  mineralogy. The thermal conductivity of 
quartz is about four times that of other common soil minerals. The Kersten correlation is  only appropriate for sands 
that have neither a very low nor a very high fraction of quartz particles. A more thorough treatment of soil  thermal 
properties  and their variability with index properties and soil constituents has been provided by Farouki  (1986). A 
generalized thermal  conductivity model  for  soils  and construction  materials  is  also  provided by  Cote  and Konrad 
2. 2    __,---,I-----1--L-+ 

Vi  1.8 

o  1.6 
1.4 +--L.----+--r----.""---r'''--r''-_/_ 
o  5  10  15  20  25  30  35 
FIGURE 13.6  Thermal conductivity offrozen coarse-grained soil (after Kersten, 1949)

1.8 -j 

..t 1.71 
1.6 J 


1.2 -1 

(W/m  K) 
1.1    --'T---"--/-
o  5  10  15  20  25  30  35  45  50 
FIGURE 13.7  Thermal conductivity offrozenfine grained soil (after Kersten, 1949)
194 Canadian Foundation Engineering Manual
The volumetric latent heat term of the soil (L
) can be estimated from the relationship: 
Yd  Is  the dry unit weight of the soil 
w  Is  the  gravimetric water content of the soil expressed as  a fraction 
L  Is the latent heat of fusion  of water to ice which can be taken as  334 kJlkg. 
The above relationship for  latent heat of the  soil, when used in the modified Berggren equation, assumes that all of 
the  water in the  soil  freezes  at O°C.  This  will  result in under prediction of the  freezing  depth in fine-grained  soils 
which  freeze  over  a  range  of temperature,  as  described  in  Section  13.2. Alternatively,  the  volumetric  latent  heat 
term can be corrected to account for unfrozen water using the relationships of Figure  13.1  if an average frozen soil 
temperature can be estimated. 
Lambda (A,) is a dimensionless coefficient that is a function of the temperature gradient, the volumetric latent heat of 
the soil and the volumetric heat capacity ofthe soil. The coefficient can be determined from a relationship developed 
by Sanger (1963) shown in Figure  13.8. The dimensionless parameters thermal  ratio    ~ ) and  fusion  parameter (11) 
can be determined from:  . 
~ = - - -
!l = Lt
MAAT Is  the mean annual air temperature coq for the site determined from  Canadian Climate Normals 
t Is  the duration of the freezing period (days) 
Is  the ground surface freezing  index (OC-days) 
C Is the volumetric heat capacity of the frozen soil 
Is the specific heat of dry soil which can be taken as  0.71  kJlkg °C 
Is the specific heat of ice which can be taken as  2.1  kJ/kg °C 
W  Is  the gravimetric water content of the soil 
For many practical field  freezing  situations, A, is  close to  unity.  Omitting it from the freezing  equation results  in  a 
slight over prediction offrost depth. 
Frost Action 195

w 0.6 

0.01  0.1 1.0 10.0 
FIGURE 13.8 Lambda ()J coefficientfor modified Berggren equation (after Sanger, 1963)
13.4.3 Frost Susceptible Soils
While  frost  depth  in  non-frost  susceptible  soils  is  readily  estimated  with  the  modified  Berggren  equation,  the 
calculation of frost  depth  in frost  susceptible soils  must  account for  the  release  of latent  heat  associated with the 
formation of ice lenses. 
An extension  to  Stefan's  approach  yields  enough  accuracy  for  practical  considerations.  Using  the  segregation 
potential to quantify the rate ofice formation with Stefan's assumptions gives the modified Stefan equation (Konrad, 
SP Is the value of the segregation potential in m
L Is the volumetric latent heat of water,  i.e.  334 MJ/m

Is the latent heat of soil (Equation  13.6) 
Is  the  thermal  conductivity  of the  frozen  soil  from  Kersten's  relationship  given  in  Figures  13.6 
and 13.7 
Is the ground surface freezing  index (OC  - days) 
13.5 Frost Action and Foundations
The conventional approach for protection ofbuilding foundations against frost action is to locate shallow foundations 
at a depth greater than the design depth of frost penetration. The modified Berggren equation, described in Section 
13.4.2,  may  be  used  to  determine  the  design  depth  of frost  penetration.  This  procedure  can  be  used  to  establish 
the  minimum  depth  of soil  cover  over  an  exterior  footing.  The  depth  of perimeter  foundation  walls  for  heated 
structures  may  be  reduced  somewhat to  account  for  heat  loss  from  the  bUilding.  Alternatively,  foundation  depth 
for  protection  against  frost  action  may  be  specified  in  local  building  codes  or is  frequently  determined  by  local 
196 Canadian Foundation Engineering Manual
experience. However, caution should be exercised where a significant depth of the footing cover is comprised of
dry, coarse-grained soil as frost depths could exceed local experience.
13.5.1 Adfreezing
Soil in contact with shallow foundations can freeze to the foundation, developing a substantial adfreeze bond.
Backfill soil that is frost susceptible can heave and transmit uplift forces to the foundation. Spread footings normally
have sufficient uplift resistance from their expanded base to resist heave, but the structural design of the wall-
footing connection must be sufficient to transmit any load applied through adfreeze. Average adfreeze bond stresses,
determined from field experiments, typically range from 65 kPa for fine-grained soils frozen to wood or concrete to
100 kPa for fine-grained soils frozen to steel (Penner, 1974). Design adfreeze bonds for saturated gravel frozen to
steel piles can be estimated at 150 kPa (Penner and Goodrich, 1983). The most severe uplift conditions can occur
where frost penetrates through frost stable gravel fill into highly frost susceptible soils surrounding a foundation.
These conditions result in a heaving situation with maximum adfreeze bond stress and have been known to jack H-
piles driven to depths in the order of 13 m (Hayley, 1988).
It is good practice to backfill against foundations with non-frost susceptible soil. Provision should be made for
drainage around the foundation perimeter, below the maximum depth of frost penetration. The granular backfill
should be capped with less permeable soil and a surface grade provided to shed runoff before it enters the backfill.
13.5.2 Thermal Insulation
It may not always be feasible to place foundation-bearing surfaces below the design depth of frost penetration.
Conditions such as high groundwater level or particularly deep predicted frost penetration may make excavation
impractical. For these and other cases, thin soil cover may be supplemented with insulating materials. Rigid board
insulation, fabricated from extruded polystyrene, is the most common material for subsurface use. This closed
cell insulation is manufactured with high compressive strength and a smooth exterior skin to resist deterioration
by absorption of moisture. Polystyrene insulation deteriorates rapidly in the presence of hydrocarbons; therefore,
alternative materials should be used where the possibility of oil spills exist. A design methodology for insulated
foundations has been presented by Robinsky and Bespfiug (1973). Summaries of their design charts for heated
and unheated structures are shown in Figures 13.10 and 13.11, respectively. These charts can be used as a guide
for estimating the required thickness of insulation. However, actual design conditions should be checked using a
geothermal analysis of the type described in Section 13.4 .1.
Insulation sheets should be placed with minimum soil covers of 300 mm and extend at least 1.2 m out from the
building. Deeper placement is warranted in high traffic areas. A sheet of vertical insulation should be fastened to
the exterior wall above the horizontal insulation up to the insulated exterior wall. Common practice is to place the
required thickness of insulation in two layers with staggered joints and to increase the thickness by 50 to 100 percent
at the comers. The surface of the insulation should be sloped such that groundwater contacting the impervious
sheets is directed away from the building.
13.5.3 Other Design Considerations
Unheated or partially heated appurtenances to a primary structure are frequently the source of frost heave
displacements. Decks, porches and unheated garages often require insulation. Where these structures may be at
greater risk of frost heave, they should be separated from the primary structure.
Buildings without basements are often supported on cast-in-p1ace concrete piles with perimeter grade beams.
Perimeter concrete grade beams fonned and cast on the ground are particularly susceptible to damage by frost
action. Uplift forces that may develop under grade beams can be transmitted back to piles resulting in tension failure
if reinforcement is not provided. It is common practice to provide cardboard void formers below grade beams where
there is a risk of frost action. A minimum thickness of 100 mm is necessary, with greater thicknesses suggested
where conditions are anticipated to be severe. Synthetk insulation should not be used as a void former because of
Frost Actiol1 197
its high compressive strength. It is also common practice to make reinforcing in grade beams symmetrical on the top
and bottom such that some uplift load can be tolerated without risk of cracking. Tension reinforcement must then be
provided in cast-in-place concrete piles with adequate tie-in reinforcement at the connections.
E =-=-= SANDY
E 80
1"0=1.7 Mg/m
{<'_ -
MC=10% n Ii,V
J.-' ..r. (f)
(I') === SILTY OR CLAYEY ~ Y '),'2. \ I ~
1"D=1.4 Mg/m
'V 60
u MC=30% if"\7   \ ~ rn
\J /' r:? L=,.......:-- - -
~ ,<>v\J /,""'-
Z 40
\ -;:::.\.
2 44
/ 'vh
/ \Jry1/ 'aLD
/ ~
L=1.22 m
o 400 800 1200 1600 2000 2400 2800
FIGURE 13.10 Design curves/or minimum insulation requirements/or heated structures
(adaptedfrom Robinsky and Bespflug, 1973)
13.6 Frost Action during Construction in Winter
Construction in winter is routine in Canada. Special care must be taken to prevent frost action affecting foundations
after construction and before heat is applied. Frost heaving and damage frequently occur on construction sites in
early winter before temporary heating begins.
13.6.1 Shallow Footings, Pile Caps and Crawl Spaces
Interior footings, which are often placed just below basement floors, are particularly vulnerable to frost action
beneath the footings even when straw is used as temporary insulation over the floor surface (Crocker, 1965). Under
these circumstances, basement floors may heave causing distortion of partitions or permanent structural damage.
Concrete pile caps cast on the ground surface are also vulnerable to frost heave during winter construction. Freezing
of supporting soils can lift caps relative to the piles resulting in undesirable deflections during construction as the
building load resets the cap. It is important, therefore, that foundations at shallow depths in buildings designed to be
heated be adequately protected during the construction period either by temporary heating or adequate insulation.
Buildings in which crawl spaces are provided between the foundation and the first floor level are also vulnerable to
frost action. Temporary heating is often only installed above the first floor for the sake of progress of the work and
the crawl space is forgotten. Temperatures drop to those prevailing outside and frost heaving occurs. The sample of
198  Canadian  Foundation  Engineering Manual 
frozen soil shown in Figure  13.2 was obtained beneath the concrete raft of a seven storey building with crawl space, 
which was heaved more than 50 mm during construction. 
=1.7  Mg/m

1D=1.4  Mg/m

400  800  1200  1500  2000 
E 20U 


u  120 


(f'l  40 
850-2200  'C-DAYS

L  - 2.44  m 
Z  =  .300  mm 


FIGURE 13.11 Design curves for minimuminsuiation requirements for unheated structures
(adaptedfrom Robinsky and Bespjiug, 1973)
13.6.2 Excavation Walls and Supports
Dangerous conditions may develop in the walls of excavations supported by sheet piling or soldier pile and lagging 
systems  if they remain  open  without heating during winter construction.  Cold  air,  being  denser than warmer air, 
flows  into below ground openings and accelerates heat extraction from the soil behind the retaining structures. 
The  direction  of heat  flow  under these  conditions  is  primarily  horizontal,  producing  a  preferred  ice  lens  growth 
direction  that  is  parallel  to  the  walls.  This  can  result  in  large  outward  pressures  against  the  wall  increasing  the 
loads transferred to  the supporting members, which may lead to  overstressing (Morgenstern and Sego,  1981). The 
horizontal components of loads on anchors and rakers may  increase considerably.  Horizontal struts spanning from 
wall to wall may be subjected to stress  increases with contributions from both walls. Additional loads may develop 
when struts expand from the heat of the sun. 
The development ofpotentially dangerous conditions must be recognized and mitigative measures taken. Deflection 
of walls  and  supporting  systems  should  be  monitored  for  early  detection  of potential  stress  increases  associated 
with frost action. This monitoring should be performed even where increased factors of safety have been used in the 
design to accommodate the expected stress increases. 
Where  observations  indicate that  excessive heaving pressures  are  developing against the walls,  appropriate  steps 

Frost Action 199
must be taken to prevent overstressing ofthe support systems. For anchored flexible walls, where inward movements
of 25 mm to 50 mm may be tolerable, stresses on the individual tiebacks may be reduced by "slacking off" on
the locking system. Other support systems, such as rakers and horizontal struts, are more difficult to adjust and
avoidance of excessive stresses may require a supply of heat to the walls to thaw frozen ground. Where subsurface
conditions are such that excessive frost action may be expected and where significant wall movements cannot be
tolerated, heating systems should be installed to prevent frost action from occurring.
200 Canadian Foundation Engineering Manual
Machine Foundations 
14 Machine Foundations
14.1 Introduction
Geotechnical engineers encounter problems related to machine foundations when designing foundations for
machinery and vibrating equipment or designing foundations for vibration-sensitive equipment subjected to
vibrations from external sources. In both cases, the foundation design is usually governed by serviceability limit
states performance considerations, not strength requirements.
14.2 Design Objectives
The main objective of the foundation design for vibration-sensitive equipment is to limit the response amplitudes
to the specified tolerance in all vibration modes. The tolerance is usually set by the machine manufacturer to ensure
satisfactory performance of the machine and minimum disturbance for people working in its immediate vicinity.
The response of foundations subjected to dynamic loads depends on the type and geometry of the foundation, the
flexibility of the supporting ground and the type of dynamic loading.
The dynamic response analysis essentially involves the calculation of the vibration characteristics of the machine-
foundation-soil system (i.e. the natural frequencies and the vibration amplitudes due to all sources of vibration).
The required complexity of the response analysis depends on the type of the foundation system. For flexible
foundation systems (e.g tabletop or mat foundations), dynamic finite element analysis may be necessary. For rigid
foundations resting directly on the soil or supported by pile groups, simplified analytical and/or numerical methods
are commonly used and are given here.
14.3 Types of Dynamic Loads
14.3.1  Dynamic Loads Due to Machine Operation
A machine causes distinct dynamic forces depending on its manufacturing purpose and the type of motion the
machine parts describe, whether it is of a rotating, oscillating or an impacting nature. The machine dynamic forces
can be periodic, transient or random.  Periodic Loading
Rotating and reciprocating machines produce centrifugal periodic (harmonic) forces due to unbalanced rotors.
An unbalanced mass me rotating with an eccentricity e and circular velocity OJ produces a centrifugal force
P = me e ai. Examples of machines with predominantly rotating parts are fans, centrifugal separators, vibrators,
lathes, centrifugal pumps, electrical motors, turbines and generators. Arya et al. (1979) provide tables for typical
values of eccentricity for rotating machines and unbalanced forces and couples for different crank arrangements.
Oscillating parts of machines produce bi-harmonic inertia forces and centrifugal forces associated with the motion
Machine Foundations 201
of the piston, the flywheel and the crank mechanism. Examples of machines with predominately oscillating parts
are piston engines, reciprocating compressors and pumps, presses, crushing and screening machines. The machine
manufacturer usually specifies the characteristics of the dynamic force from reciprocating machines.
Transient Loading
Impacting parts of machines develop intermittent dynamic forces that are transient in nature. Transient loading
is characterized by a non-periodic time history of limited duration. The load time history could be smooth as the
one produced by hammer blows or more irregular similar to that generated by crushers and shredders. This type of
loading is represented either by an analytical expression or by a set of digital data. Random Loading
Some machines such as mills, pumps and crushers produce fluctuating forces that are random in nature. A random
force and its effect is most meaningfully treated in statistical terms and its energy distribution with regard to
frequency is described by a power spectral density (power spectrum). Detailed information on dynamic loading is
given in Barkan (1962), Richart (1975) andArya et al. (1979).
14.3.2 Ground Transmitted Loading
In the case ofvibration-sensitive equipment, the vibration problems may stem from external sources such as ground-
transmitted vibrations from traffic, trains and blasting activities. Vibration criteria supplied by the manufacturer of
vibration-sensitive equipment are typically specified in terms of "floor vibrations." Before the facility is built,
though, floor vibration cannot be measured directly but, rather, must be predicted by analytical means. Seismic
excitation at the site due to ground-transmitted vibration could be, in many cases, an important factor for designing
the facility, or even in deciding whether or not it will be built. To assess the level of seismic excitation at the site due
to traffic, trains or blasting activities the ground vibration has to be monitored. Ground vibration is usually evaluated
in terms of ground acceleration measurements. Vibration Monitoring Equipment
Components of the ground vibration monitoring equipment include sensors, mountings for the sensors, and data
acquisition systems. The monitoring system should be designed to provide the required sensitivity, minimize data
sampling errors, and achieve the robust performance necessary for the anticipated environmental conditions.
Sensors:  ground accelerations can be measured using seismic accelerometers with appropriate sensitivity and
suitable operational temperature range. The mounted natural frequency of the sensor should be higher than the
maximum excitation frequency of interest to minimize measurement bias in the frequency range of interest.
Mounting arrangement: The sensors are usually mounted on especially designed posts. The posts should be rigid
and light. The length ofthe post should be smaller than the minimum wavelength of soil vibrations for the maximum
frequency of interest. It should also enable the simultaneous attachment of accelerometers in three mutually
orthogonal directions, with two oriented horizontally and the third vertically. The sensors must be protected from
interference from other factors such as wind, rain, snow and electromagnetic fields.
Data  acquisition  system:  The digital data acquisition system should be compatible with the sensors used in
measuring the vibration. Proper analog filtering should be used to ensure that no frequency interference occurs.
The sampling frequency has to be higher than the highest frequency component of interest. Representation of Ground-Transmitted Excitation
The ground-transmitted excitation can be represented as acceleration time history or in terms ofacceleration Fourier
transform. The time history will show the maximum acceleration experienced at the location ofthe foundation, while
202 Canadian Foundation Engineering Manual
the Fourier transfonn will show the frequency content and the distribution of excitation energy with frequency.
14.4 Types of Foundations
Machine foundations are designed as block foundations, wall foundations, mat foundations, or frame foundations.
Block foundations are solid blocks of concrete with sizable thickness, wall foundations are block foundations
with cavities, and mat foundations are foundations with a limited thickness compared to their surface dimensions.
Block foundations, the most common type, and wall foundations behave as rigid bodies. Mat foundations of small
depth may behave as elastic slabs. Sometimes the foundation features a joint slab supporting a few rigid blocks for
individual machines. The foundations can rest directly on soil (shallow foundations) or on piles (deep foundations).
The foundation type results in considerable differences in response.
14.5 Foundation Impedance Functions
The response of soils and foundations to dynamic excitation is frequency dependent and, thus, is a function of
the stiffness and damping parameters of the foundation. Therefore, the evaluation of the appropriate stiffness and
damping parameters (impedance functions) for the foundation soil or pilelsoil system is a step in the analysis.
The foundation block can be represented in the dynamic analysis as a lumped mass with a spring and dashpot. The
block has a mass, m, and is free to move in six directions, i.e., it has six degrees of freedom, three translational and
three rotationaL These are the displacements along the Cartesian axes x, y and z and rotation about the same axes.
The response of the mass depends on the spring and the dashpot that represents the supporting soil medium or pilei
soil system. The spring represents the elasticity of the soil and the dashpot represents damping caused by energy
dissipation. This section presents a general introduction to this subject and a summary of approaches and fonnulae
that can be used to evaluate the stiffness and damping of shallow and deep foundations.
14.5.1 Impedance Functions of Shallow Foundations
Shallow foundations are often idealized by a massless circular disc. For circular bases the complex stiffness K
(also. called the dynamic impedance function) associated with direction i is obtained by the detennination of the
relationship between a harmonic force acting on a massless disc resting on the surface of the halfspace and the
resulting displacement of the disc. This complex stiffness can be expressed in tenns of the true stiffness constant,
, and damping constant, c
' as
(14.1 )
in which is tr; static stiffness, a
=~ R = dimensionless frequency, R is the disc radius, Vs =   =shear wave
velocity ofthe soil, G and pare the soil shear modulus and mass density, respectively, and k' and c'. are stiffness and
damping constants nonnalized as follows: k! =5 , d = !s c.. In the case of an isotropic homogeneous halfspace
I k I k.R I '
the approximate static stiffness constants for the vertical translation, v, horizontal translation (sliding), u, rocking, 'V,
and torsion, 11, are shown in Table 14.1, in which v is Poisson's ratio and G is the soil shear modulus. The constants
k/ and C/ are frequency-dependent and may be approximated using the treatment outlined by Wolf (1995):
f.1 Zo Vs 2 d' ( )
k;(ao)=l- rc RVao an ci ao R V
where V is the pertinent wave velocity as given in Table in which Vp is the dilational wave velocity
~ ~ 2 G 1-v. The other parameters are given in Table 14.2.
P 1-2v
Machine Foundations 203
TABLE  14.1  Static Stijfrlesses ofa Disc Resting on the Surface ofa Homogeneous Haljspace
k --
Horizontal  Rocking  Torsion 
k k =-
\jf 3(1-v) u 2-v
TABLE  14.2  Parameters ofApproximate Solution for Footings Resting on Swface ofSoil Haljspace
Vertical  Horizontal  Rocking  Torsion 


. 1t
V Vp v::; 1/3 Vs Vp v::; 113 Vs
j 2Vs v 113 2Vs v> 1/3
ZoIRj 1t 9n
9n V
8 (2 -v)
n   -(I-v) -
( r
32 V, 4 \v,J j
0 0 0 v ::; 113 0 v::; 113 )l
1.2(V -l.IP1oR 2.4(V
\ 3/
v 1/3 v 2113
To account for the material damping, the stiffness and damping constants including the soil hysteretic damping, f3,
are given by
I 20 Vs I
and (a
) =-R-+-k;
V a
For shapes which differ from circular, the real noncircular base is replaced by an equivalent circular base with a
suitable radius. The radius of the equivalent circular foundation is usually determined by equating the areas of
the actual base (Ao) and equivalent base for vertical and horizontal translations, the moments of inertia (10) for
rotation in the vertical plane (rocking) and the polar moments of inertia (10) for torsion about the vertical axis. For
rectangular bases having dimensions a and b, the equivalent radii are given in Table 14.3.
TABLE  14.3  Equivalent Radii for a Rectangular Footing having Dimensions a and b
Vertical  Horizontal  Rocking  Torsion 

14.5.2 Embedment Effects 
Embedment is known to increase both stiffness and damping but the increase in damping is more significant. The
response of embedded footings can be approximated by assuming that soil reactions acting on the base are equal
to those of a surface footing and the reactions acting on the footing sides are equal to those of an independent layer
overlying the halfspace (Figure 14.1). Novak and Beredugo (1972) and Beredugo and Novak (1972) used plane
204  Canadian Foundation  Engineering Manual 
strain  solutions  for  side  reactions  and  a  halfspace  solution  for  base  reactions,  and  the  notations  in  Figure 14.1  to 
derive  the  stiffness  and  damping  constants  given in Table  14.4.  The parameters  C  defining  the  base  stiffuess  and 
damping and S defining the side stiffuess and damping in Table  14.4 are frequency  dependent.  However, it is often 
sufficient to  select suitable constant values to represent the parameters over a limited frequency range. 


V Backfill '"  co

Os,  Ps

- Yc
-- C
2R  0, p, v 


FIGURE 14.1 Notationsfor embeddedfoundation
TABLE 14.4 Stiffness and Damping Constants for Embedded Footings
Vertical  R2 + S  IiJP,  G,)
v2 v2 p  G 
- R2 JPG[ycC
Machine Foundations 205
Such constant values are suggested in Table 14.5. The lack of confining pressure at the surface often leads to
separation of the soil from the foundation, which reduces the effectiveness of embedment. To account for the lack of
confining pressure at the surface which leads to separation of the soil from the foundation, an effective embedment
depth, D, smaller than the true one, may be used. An extensive set of tables and charts for stiffness and damping
constants of embedded footings of arbitrary shapes is given by Gazetas (1991).
TABLE 14.5 Stiffness and Damping Parameters (j3 0)
Side Layer
i SVI =2.7
Svl =2.7
SV2 = 6.7
Sv2 = 6.7
= 7.5
= 5.2
= 6.8
Co = 5.0
SUI = 4.1
SU2 = 10.6
SU2 = 9.1
CUI = 5.1
CUi =4.7
= 3.2
S'I'I =2.5
STj1 = 10.2
= 1.8
S , 1.8
STj2 = 5.4
C'I'l =4.3
Co/I =3.3
C'I'2 = 0.7
CTj2 = 0.7
14.5.3 Impedance Functions of a Layer of Limited Thickness
The stiffness of a layer oflimited thickness is higher than that of a halfspace but its geometric damping decreases or
even vanishes ifthe excitation frequency is lower than the first natural frequency ofthe soil layer. For a homogeneous
soil layer, the first vertical and horizontal natural frequencies, CO and co ' respectively, are:
v u
OJ =1r Vs ~ 2   2 - v) and OJ = 1r ~ (14.4)
v 2H I-2v u 2H
The damping parameters at frequencies lower than COy and CO may be calculated by:
S 2/3 SUi and S =2/3 Svl (14.5)
u2 a v2 a
o 0
The stiffness and damping of a footing embedded in a layer of limited thickness, H, can be defined in a manner
similar to Eq. 14.1. However, the static stiffuesses, kv' of circular foundations may be given by (Elsabee and
Morray, 1977; Kausel and Ushijima, 1979):
4GR R .
kv - (1 + 1.28-)(1 + 0.470 )[1 + (0.85
0.280) D/ ]
I-v H
1 IH
- 8GR 1 R 2 5 D
k =-(1+--)(1+-0)(1+--)
u 2 v 2H 3 4H
_ 8GR
1 R D
krp = 3(1- v) (1 +(5 H)(l + 20)(1 + 0.7 H)
206  Canadian  Foundation  Engineering Manual 
=(0.48 - 0.03)R k/l (14.6d) 
_ 16GR
k'l= 3  (1+2.678) 
These  empirical  expressions  for  the  stiffness  are  refel1"ed  to  the  centre  of  the  base  and  are  valid  for 
0= D/R 5  1.5,  D/H 5  0.75  and RJH 50.5. The dynamic  stiffness  and  damping  can  be  calculated taking  k'and c' 
equal to the halfspace functions (Equations  14.1-14.3).  For frequencies below the first layer natural frequencies, it 
would be safe to  ignore geometric damping completely (first tenn and damping formula in Equation 14.3).  Similar 
formulae for foundations on shallow layer can be found  in Gazetas (1991). 
14.5.4 Trial Sizing of Shallow Foundations
The  design  of a  shallow foundation  for  a  centrifugal  or reciprocating  machine starts  with trial  dimensions  of the 
foundation  block.  The  trial  sizing  is  based on  guidelines  derived  from  past  experience.  The following  guidelines 
may be used for the trial sizing of the foundation block: 

1.  Generally,  the  base  of the  foundation  should  be  above  the  groundwater  table.  It should  be  resting  on 
competent native soil (no backfill or vibration-sensitive soil). 

2.  The mass of the block should be 2 to  3 times the mass of the supported centrifugal machine, and 3 to 5 times 
the supported reciprocating machine. 
3.  The top of the block should be 0.3 m  above the elevation of the finished fioor. 
4.  The thickness of the block should be the greatest of 0.6 m, the anchorage length of the anchor bolts and 1/5
the  least dimension of the footing. 
5.  The width should be  1 to  1.5  times the vertical distance from the base to the machine centerline to increase 
!  damping  in rocking mode. 
6.  The length is  estimated from  the mass  requirement  and  estimated thickness  and width of the  foundation. 
The length should then be increased by 0.3  m  for maintenance purposes. 
7.  The length  and  width  of the  foundation  are  adjusted  so  that  the  centre  of gravity  of the  machine  plus 
equipment lies  within  5  % of the  foundation  dimension  in  each  direction,  from  the  foundation  centre  of 
8.  It is desirable to increase the embedded depth ofthe foundation to increase the damping and provide lateral 
restraint as welL 
9.  If resonance  is  predicted  from  the  dynamic  analysis,  increase  or  decrease  the  mass  of the  foundation 
to  change  its  natural  frequency  (try  to  undertune  for  rotating  machines  and  overtune  for  reciprocating 
14.6 Deep Foundations
The  dynamic  stiffness  and  damping  of a  pile  group  are  affected  by  both  the  interaction  between  the  piles  and 
surrounding soil, and the interaction between individual piles. Therefore, the calculation of the stiffness for a group 
ofidentical piles may be performed in two steps. First, the stiffness ofthe single pile is calculated. Second, the group 
effect is accounted for using "interaction factors." 
14.6.1 Impedance Functions of Piles
The pile length, bending and axial stiffness, tip and head conditions, mass, batter and the surrounding soil properties 
and their variation with depth and layering, affect the  dynamic stiffness of a pile. The impedance functions  of piles 
can be described as 
Machine Foundations 207
The stiffness constants, ki' and the constants of equivalent viscous damping, c
' for individual motions of the pile
head suggested by Novak (1974) are shown in Table 14.6.
TABLE 14.6 Stiffness and Damping Constants Jar Single Piles
Horizontal Rocking Coupling
i EpI
. kll=ylul
kc=yIcl c

ccp=VfCP2 CC
RV fC2
cll =v1112
s s s s s
These constants are a function of the pile's elastic modulus, E cross-sectional area, A, and its moment of inertia
and torsional stiffness I and G/, respectively. R is the radius of circular piles and equivalent radius for non-circular
piles. The symbol in Table 14.6 represents dimensionless stiffness and damping functions whose subscript
1 indicates stiffness and 2 indicates damping. These functions depend on the following parameters: the relative
stiffness ofthe pile and soil, E/G; dimensionless frequency, a ; the slenderness ratio, LIR, in which L pile length;
material damping of both the soil and pile; the variation of soil and pile properties with depth; and the tip and
head conditions. However, E /G, the soil profile and, for the vertical direction, the tip condition have the strongest
effect on the stiffness. The stiffness and damping     are given for a few basic cases in Table 14.7, for
horizontal response, for a dimensionless frequency, a = 0.3. For other cases, see Novak and E1 Sharnouby (1983) .
TABLE 14.7 Sti.ffoess and Damping Parameters ojHorizontal Response
(LiR > 25Jar homogeneous soil and LlR > 30Jar parabolic soil profile)
(Reproducedfrom Novak and El Sharnouby J983 with permission ojASCE)
with Depth
with Depth
to G,ol1
-0.0217 0.0021
1-0.0429 0.0061
0.25 -0.0668 0.0123
-0.0929 0.0210
250 i -0.1281 0.0358 i -0.1786
10000 0.2207 -0.0232 0.0047 0.0024 0.1634 -0.0358 0.0119
2500 0.3097 -0.0459 0.0132 0.0068 0.2224 -0.0692 0.0329
0.4 1000 0.3860 -0.0714 0.0261 0.0136 0.2677 -0.1052 0.0641
500 0.4547 -0.0991 0.0436 0.0231 0.3034 -0.1425 0.1054
250 0.5336 -0.1365 0.0726 0.0394 0.3377 -0.1896 0.1717
.10000 0.1800 i -0.0144 0.0019 0.0008 0.1450 -0.0252 0.0060
12500 0.2452 -0.0267 0.0047 0.0020 0.2025 -0.0484 0.0159
0.25 1000 0.3000 -0.0400 0.0086 0.0037 0.2499 -0.0737 0.0303
500 0.3489 -0.0543 0.0136 0.0059 0.2910 -0.1008 0.0491
250 0.4049
1-0.0734 0.0215 0.0094 0.3361 • -0.1370 i 0.0793
10000 0.1857 -0.0153 0.0020 0.0009 0.1508 -0.0271 0.0067
2500 0.2529 -0.0284 0.0051 0.0022 0.2101 -0.0519 0.0177
1000 0.3094 -0.0426 0.0094 0.0041 0.2589 -0.0790 0.0336
500 0.3596 -0.0577 0.0149 0.0065 0.3009 -0.1079 0.0544
250 0.4170 -0.0780 0.0236 0.0103 0.3468 -0.1461 0.0880

208  Canadian  Foundation  Engineering Manual 
14.6.2 Pile-Soil-Pile Interaction
When piles in a group are closely spaced, they interact with each other because the displacement ofone pile to
contribute to the displacement ofothers. To obtain anaccurate analysis ofdynamic behaviourofpilegroups itis
necessary to use asuitable computerprogram. However, asimplifiedapproximateanalysis, canbeformulated on
thebasisofinteractionfactors, a,introducedbyPoulos(1971)for staticanalysisandextendedtothedynamiccase
by Kaynia and Kausel (1982) who presented charts for dynamic interaction. For a homogeneous halfspace, the

interactionfactorsbetweentwopilesmaybegivenby (DobryandGazetas, 1988 andGazetasandMakris, 1991)

(14.8b) I 
wherea anda areverticalandhorizontalinteractionfactors, respectively, Sid =pilespacingto diameterratio, 8
v u 1
istheanglebetweenthedirectionofloadactionandtheplaneinwhichpileslie, and V =theso-calledLysmer's 1
analogvelocity= 3.4v,. .


Tocalculatethedynamicstiffuessof apilegroupusingtheinteractionfactors approach,theimpedancefunctions of
singlepilesandtheinteractionfactors are calculatedfirst, thenthe group impedancefunctions are computed. The j 
stiffnessanddampingconstantsofindividualpilesarecalculatedusingexpressionsgivenin Table 14.6orformulae

(1982). Theimpedancefunctions ofapilegroupof npilesarethengivenby


i=1  )=1 

where   ,KrG and KeG are the vertical, horizontal, rockingandcouplinggroup stiffness,respectively. In
Eq. 14.9 kv isthestaticvert!falstiffnessofthe singlepile, [e v] =[atl where a; = complex interaction factors
pilesi and}, = kv !Ky
,andKy isthecomplexverticalimpedancefunctionofthesinglepile. Similarly,
kh isthestatichorizontalstiffnessofthepile[e h] = where a;=  complex interaction coefficients for the
horizontaltranslationsandrotations.Theformulation ofthe[a]h canbefoundinEINaggarandNovak(1995).
14.6.3 Trial Sizing of Piled Foundations
The designofadeepfoundation fora centrifugalorreciprocatingmachinestarts withtrial dimensions ofthepile
cap, andsizeandconfigurationofthepilegroup(StepNo.3in the designprocedure).Thetrialsizingis basedon
guidelinesderivedfrompastexperience.Thefollowingguidelinesmaybeusedfortrial sizingthepilecap:
1. Thepilecap(block)massshouldbe 1.5to2.5timesthemassof thecentrifugalmachineand2.5 to4times 
themassof thereciprocatingmachine. 
Machine Foundations  209 
2.  The top of the cap should be 0.3  m above the elevation ofthe finished fioor. 
3.  The thickness of the block should be the  greatest of 0.6 m, the anchorage length of the anchor bolts and  1/5 
the  least dimension of the  block. 
4.  The width should be  1 to  1.5  times  the vertical distance from the base to  the  machine centerline to increase 
damping in rocking mode. 
5.  The  length  is  estimated from  the  mass  requirement  and  estimated thickness  and width of the  block.  The 
length should then be increased by 0.3  m for maintenance purposes. 
6.  The length and width of the block are  adjusted so  that the  centre of gravity of the  machine plus equipment 
lies within 5 % of the block dimension in each direction,  from the  block centre  of gravity. 
7.  It is desirable to  increase the embedded depth of the foundation to increase the damping and provide lateral 
restraint as well. 
The  following  guidelines may be used for the trial configuration of the pile group: 
1.  The  number  and  size  of piles  are  selected  such  that  the  average  static  load  per pile::;:  Yz the  pile  design 
2.  The  piles  are  arranged  so  that  the  centroid  of the  pile  group  coincides  with  the  centre  of gravity  of the 
combined structure and machine. 
3.  If battered piles  are used to provide lateral resistance  (they are better than vertical piles in  this  aspect), the 
batter should be  away from the pile cap and should be symmetrical. 
4.  Ifpiers are used,  enlarged bases are recommended. 
5.  Piles and piers must be properly anchored to the pile cap for adequate rigidity (as cormnonly assumed in the 
14.7 Evaluation of Soil Parameters
The soil parameters required for the  dynamic analysis include the shear modulus, G,  the material damping ratio, D, 
Poisson's ratio,  v,  and mass  density,  p.  Some  of the  procedures  that can be used to  evaluate these  parameters  are 
given here. 
14.7.1 Shear Modulus
The  shear strains  developed in the  supporting soil medium due  to the dynamic loading  from  machine foundations 
are usually  of a much  smaller magnitude than the  strains  produces  by  static  loading.  The  value  of the  soil  shear 
modulus at smaller strains is much higher than its value at larger strains.  Therefore, the soil shear modulus used for 
the computation of the foundation  impedance functions should be evaluated for smaller strain laboratory field tests 
(see Richart et aL  1975 for details on experimental procedures).  In the absence of measured values, the correlations 
in Table  14.8 can be used to  evaluate the shear modulus. 
14.7.2 Material Damping Ratio
Soil material damping is  a measure of energy lost due to friction between soil particles during the dynamic loading. 
Material damping ratio  can be  obtained from  resonant  column testing  and the  Spectral Analysis  of Seismic Wave 
procedure (SASW). The material damping is typically 0.03  to 0.05  for  sand and saturated clay. 
14.7.3 Poisson's Ratio and Soil Density
The  dynamic behaviour of foundations  is  less  sensitive to  the values  of v  and  p.  Typical values  for  v are given in 
Table  14.9.  The  soil  mass  density  values  should always  be  calculated  from  the  total  unit  weight rather  than  the 
buoyant unit weight.  Total weights are used in dynamic problems because both the solid and liquid phases vibrate. 
210 Canadian Foundation Engineering Manual
TABLE 14.8 Some Correlations for Soil Shear Modulus
Soil Type
e)2   *
(kPa) Sand (round-grained) 

1+ e
3230(2.97  e)2 (j l/2 
* (kPa)
Sand (angular-grained)  G

I.  (1968) 
1+ e
Hardin and Black 
Hardin and Black 
G  35000No.
(a )0.4  • (pst)
max  60  0 
.  Seed et al.  (1986) 
Clay (moderate 
G  = 3230(2.97  e)2 (jol!2(OCR)K  *. (kPa) 
max  l+e
Hardin and Dmevich 
. a = (Cf]  +(j 2  +Cf

)  = effective octahedral stress 

••  OCR is over consolidation ratio and K = function of the soil plasticity index, PI,  and is  given by 
PI (%)  o 20  40  60  80  100 
K o  0.18  0.3  0.410.48  0.5 
TABLE 14.9 Typical Values ofSoil Poisson's Ratio
Soil Type
Saturated clay  0.45,.0.50 
Unsaturated clay  0.35-0.45 
Silt, Medium dense sand - Gravel 
Dense sand - Gravel  0.4-0.5 
14.8 Response to Harmonic Loading
The machine foundation can vibrate in any or all six possible modes due to the excitation from the vibrating 
machine it supports. For ease ofanalysis, some ofthese modes can be considered separately (e.g. vertical or torsional) 
and design is carried out by considering the displacement due to these modes separately. 
14.8.1 Response of Rigid Foundations in One Degree of Freedom
The  response  of the  foundation  in one  degree  of freedom  (1  DOF) to  a  harmonic  load with  an  amplitude,  P,  and 
frequency ill, can be  given by 
where m,  k  and c are the mass, stiffness and damping of the  foundation,  and  <l> The stiffness  and 
damping constants k and c are established as described in Sections 14.5  and 14.6 and are frequency dependent hence 
the response has to  be calculated using Equation 14.10. However, if they can be considered frequency  independent 
, Machine Foundations 211
in  the frequency range of interest, Equation 14.10 can be rearranged and the real amplitude can be written as 
P 1
v ==  vst £ (14.11)
k,  [1  (:/],+4D'(:')' 
in which the natural frequency, COo  ==  kv , the damping ratio,  D= v  = the static displacement and f: Ji:;'
m 2  kv
kv sl 
==  dynamic amplification factor given by 
IDa  "COO 
For a harmonic excitation, the maximum displacement is  given by 
14.8.2 Coupled Response of Rigid Foundations
The coupled motion in the vertical plane represents an important case because it results from excitation by moments 
and horizontal forces acting in the vertical plane.  The horizontal sliding, u(t), and rocking, lfI(t), describe the coupled 
motion.  For a simple rectangular footing with dimensions a and b, the mass moment of inertia of the system is 
m b
_1 (a
+ b
)+mJ (Ye __)2+
; (14.14)
12  2 
where rnj  is  the mass  of the footing,  m
is  the mass of the machine, Yc  and Ym  are distance  from C.G. to  foundation 
base  and  machine  centre  of gravity,  respectively.  The  stiffness  constants,  kuu' ku'l" and  k'l''I' are  described  as  the 
stiffness 'constants  for  translation and rotation  at  the  base  of the  footing,  transformed  to  the  Centre of Gravity  of 
the system, CG. If the stiffness constants referred to  the centre of the base are ku  and k",  (calculated as  described in 
sections  14.5  or  14.6), the stiffuess constants referred to CG are 
The response of the foundation  system in  the  coupled motion to an excitation loading given by a horizontal force, 
pet), and a moment, M(t) can be evaluated using the modal analysis. First, the natural frequencies and modes of free 
vibration are  i.e. 

ID 1,2  (14.16) 
With these two natural frequencies,  w.  0 = 1,2), the two vibration modes are 

a . =_1 

}  "f}.
k mco 

uu- j 
Then, the footing translation and rocking are 
2  2 
u(t) =Lqjujsm(cot+<j»  and  ",(t)= Lqj'" jSm(IDt+<j» 
j:l  j=l
212 Canadian Foundation Engineering Manual
in which q. and 4>. are
-1 r2D/1) /i) 'j'
-tan 2 2 (14.19)
'- (0 j-(i)
PUj+lvf\v j' M
and D
= 2w 1 .    
If the damping in the system is small, the results from modal analysis are very close to the results obtained from the
direct approach.
14.8.3 Response of Rigid Foundations in Six Degrees of Freedom
When the rigid foundation is of general shape, the response is in six degrees-of-freedom, all of them, possibly,
coupled. The stifihess constants are described at the base of the footing, then transformed to the reference point,
Centre of Gravity CG. The stiffness and damping are described in terms of impedance functions KiF Considering
the dynamic equilibrium of forces and moments for the system will result in six linear algebraic equations that can
be solved for the vibration amplitudes.
14.9 Response to Impact Loading
Shock producing machines generate dynamic effects that differ from those of rotating machines and the design of
their foundations, therefore, requires special consideration. Different foundation arrangements are used to support
impact-producing machines. The foundation block is most often cast directly on soil. When the transmission of
vibration in the vicinity and adjoining facilities is of concern, the block may be supported on vibration isolating
14.9.1 Design Criteria
The design ofa hammer foundation must ensure satisfactory performance ofthe hammer and minimum disturbance
tothe environment. These objectives are met by limiting the vibration amplitudes, settlement, physiological effects.
and stresses to the given tolerances. Performance Criteria
The manufacturer should specify the limits on the vibration amplitudes. The physiological effects are related to
vibration velocity and acceleration rather than displacement. The vibration velocity can be calculated approximately
as vm:::: vm(Oo where vm the maximum displacement and ruo = the natural frequency of the foundation. For data on
human perceptibility collected see Richart et al. (1970). Stresses in all parts of the foundation have to remain within
allowable limits. Dynamic stress is repetitive and fatigue effects have to be accounted for by using a factor of safety
greater than 3 in the design.
The adequacy of the mass for a hammer foundation is best proven by detailed analysis of stresses and amplitudes.
Some guidelines have been suggested for the preliminary choice of the weight of the foundation block. Assuming
the anvil weight 20 Go' where Go is the weight of the head, the weight of the block, G , can be estimated by
= 75G
C )2 (14.20) G
where Co = the maximum velocity of the head and C =5.6 mis, (Rausch, 1950).
r Vibration Effects on the Environment
Vibration propagates from the footing into the surroundings in the form of ground motion. The vertical amplitude of
the ground motion, v
' at a distance r from the foundation vertical axis can be evaluated approximately as
Machine Foundations 213
- fio -a("-ra)
VI" - Vo -e
where Vo  footing amplitude, ro  . the distance ofthe footing edge from its vertical axis and a  empirical coefficient 
ranges from 0 to 0.05 mi. The horizontal amplitude may be considered equal to the vertical  one.  The response of a 
structure located near the hammer foundation can be predicted using the methods of structural dynamics. 
14.9.2 Response of One Mass Foundation
When the  anvil  is  rigidly  mounted  on the  foundation  block  and  the hammer blow  does  not act eccentrically,  the 
foundation response can be  analysed using a  one degree  of freedom model.  The  response  corresponding to  initial 
velocity of the system, , can be written as 
where co' 0  =COo.Jl- D2 ,  D ~ The initial velocity of the system,  C,  can be obtained from the consideration 
of  the  collision  between  the  head  and  the  foundation.  The  peak  force  transmitted  into  the  ground  is 
+ (CCO 
)2  and the peak stress is  a  F / Ab ' where v=peak displacement and Ab = the base area. 
14.9.3 Response of Two Mass Foundation
When the  anvil rests  on  an elastic  pad, a  hammer foundation  should be considered as  a  two  mass system.  In this 
model, m is the mass of the anvil and m is the mass of the footing; kl and c are the stiffness and damping constants 
1 2 1
of the  pad  and  k2 and  c
are  the  stiffness  and  damping  of the  soil  or  piles  supporting  the  footing.  Stiffness  and 
damping of foundations  can be  evaluated using the  approaches  described  in  sections  14.5  and  14.6.  The  stiffness 
and damping constants of a pad can be given by 
R kp
k   ~ andc  2tJ  - (14.23)
PdP Pro 

where E ,A ,d, andp'  are Young's modulus, area, thickness and material damping of the pad, respectively, and COo
= naturiI frbquency  of the block calculated with k . With these values, the natural frequencies can be calculated as 
The damped response can be evaluated using the approach developed by Novak and El-Hifnawy (1983). 
14.10 Response to Ground-Transmitted Excitation
The basic  response to harmonic  loading  in  1 DOF  is  given by Eq.  14.10.  For ground-transmitted excitation,  the 
forcing function,  pet), is  given by  {- mii(t) }where ii(t) is the absolute ground acceleration time history measured 
at  the  location  of the  future  foundation.  In  this  case,  there  are  two  approaches  to  solve  for  the  response  of the 
foundation.  In the first approach, the Duhamel integral of ii(t) is used to calculate the relative displacement of the 
foundation,  i.e. 
214 Canadian Foundation Engineering Manual
The response ofthe machine-foundation system is influenced by both its natural frequency and the frequency content 
of loading. The traffic loading is transmitted to the foundation as  a combination of seismic waves propagating in the 
ground at different frequencies.  Therefore, alternatively, a Fourier analysis can be used to  calculate the response of 
the foundation to  the transient load in the frequency  domain.  In this  type of analysis,  the  load is represented by the 
sum of a series of harmonic  components obtained by subjecting the  load time history to  a  Fast Fourier Transform 
(FFT).  In the FFT, the forcing function is  given at an  even number, N, of equidistant points in the time domain, and 
N/2 frequency  components are  obtained.  Thus,  increased  accuracy  can only be obtained by  increasing the number 
of data points. 
The response of the system can be related to  the loading by 
where x
and  w" are  the  amplitude  and frequency  of that harmonic  component  and  H(w,) is  the  modulus  of the 
complex transfer function, H(w,,) given by 
The principle of superposition gives the total response as  8(t) L8n (t).
1    ~

Foundations on Expansive Soils 215
Foundations on Expansive Soils
15 Foundations on Expansive Soils
15.1 Introduction
Expansive soils are defined as any soil that has the potential to undergo significant volume change as a result of
changes in water content. The magnitude of volume change considered to be significant is defined in terms of
the serviceability limit states performance of affected surface structures such as shallow foundations, utilities, or
Light structures such as the house shown in Figure 15.1, are generally constructed with limited knowledge of the
soil conditions. However, the buildings often suffer subsequent distress because of volume changes (deformation)
in the soils below the structure.
Heaved walk
and step
Flat if telepost
adjusted correctly
Heaved lawn
FIGURE 15.1 Ground movements associated with the construction of
shallow footings on an expansive soil (Hamilton, 1977)
Vertical ground movements generally occur as a consequence of unloading associated with the excavation for the
basement of the house, or a change in the normal evaporation and evapotranspiration regime at the ground surface.
An example of structure distress can often be seen in floor slabs that are meant to function as "floating" slabs but
seldom 'float' (Figure 15.2). This is just one of many ways in which light structures suffer distress due to volume
changes in expansive soils.
216 Canadian Foundation Engineering Manual
FIGURE 15.2 Typical cracking pattern around a basement slab that was meant to peiform as a "floating slab"
Light structures most commonly experience distress associated with expansive soils; however, swelling pressures
of expansive soils can be high, causing movement to multi-story structures. Relatively short piles below a light
structure, along with a structural floor slab, provide a common solution to many expansive soils problems (Figure
Main floor
Extended active
FIGURE 15.3 Illustration ofa short pile and structural floor slab below a light structure such as a house
The potential for a soil to be expansive is largely controlled by the mineralogy and percentage of the clay-size
fraction, while changes in water content are also dependent upon changes in environmental conditions at the ground
surface. Environmental conditions result in the wetting and/or drying of the soil in response to moisture transfers
across the soil-atmosphere boundary.
Foundations on Expansive Soils 217
Changes in the water content in the soil may be the result of natural causes such as  climatic fluctuations or the result 
of human activity such as  surface irrigation, runoff from paved areas or leakage from buried utilities. 
Expansive soils problems are encountered in almost every country of the world and have been found to be extremely 
costly to accommodate fully in original design or remedial design.  Expansive soils have been referred to as the 'hidden 
disaster' in the United States and cause more damage to structures, (particularly light buildings and pavements), than 
all  other natural  hazards  including earthquakes  and floods  (Jones  and Holtz,  1973).  It has been estimated that the 
average  annual  losses  due to  structural distress  associated with expansive soils  in  the United States  is  in the order 
of $7 billion (Krohn and Slosson,  1980). While the amount of damage in Canada may be considerably less,  it is  still 
substantial (Fredlund,  1979). 
Problems related to  structures on expansive soils  are  accentuated since structures  incurring the  most damage have 
generally  had the least  engineering design prior to construction. Engineers  are  often reluctant to become involved 
in the  study  of expansive  soils  problems  because  the  consulting  fees  are  generally  small relative to  the  potential 
risk of litigation. There is need to establish accepted standard ofpractice or "protocol" for geotechnical engineering 
practice as  it relates to expansive  soils. 
This chapter in the  Canadian Foundation Manual does not provide a complete description and analysis ofproblems 
related  to  the  behavior  of expansive  soils.  Rather,  the  goal  of this  chapter  is  simply  to  provide  infonnation  on 
factors  controlling heave in expansive soils and to present an outline of a simple method based on one-dimensional 
oedometer test results,  to estimate the magnitude of potential heave. 
There  are  several  important questions  that need to  be  addressed  in order to  evaluate  the  impact that  an expansive 
soil may have on foundation performance: 
How can a potentially expansive soil be identified? (i.e., soil characterization). 
What environmental conditions can cause changes in water content in an expansive soil? (i.e., environment 
What methods can be used to predict the magnitude of volume change or heave that might be experienced 
subsequent to  completion of construction?  (Le., predictive model), 
•  What design and remedial measures can be taken to minimize damage to light engineered structures?  (i.e., 
design methods). 
The first two questions focus primarily on the identification and characterization of expansive soils. These methods 
are  described in Section  15.2. The next question shows  that there  is  need to have a predictive method based O,n  an 
appropriate theoretical framework to relate changes in void ratio to changes in stress state.  This predictive method 
,  is  outlined  in  Section  15.3.  The  final  section  of this  chapter  provides  a  general  discussion  of issues  related  to 
foundation design and remediation measures that can be taken when dealing with expansive soils. 
15.2 Identification and Characterization of Expansive Soils
The geographic regions in Canada where expansive soils problems may occur can be delineated by first identifying 
those areas containing soils with the prerequisite mineralogy and lithology (Quigley,  1980).  Secondly, the climatic 
conditions must lend themselves to the potential for large  changes in water content. Expansive soils  are  comprised 
of clay soils that contain a significant fraction of active clay minerals. Glacial and post-glacial processes laid down 
most  of the  clay-rich  deposits  of concern  in  the  construction  and  performance  of surface  structures  in  Canada. 
These clay-rich soils are  found either in glacial lacustrine deposits or in glacial tills  (Figure  15.4). 
~ ~  
218 Canadian Foundation Engineering Manual
o 1000 
FIGURE 15.4  Distribution ofmarine andfreshwater glacial andpostglacial lakes of Canada (Quigley, 1980)
Many of the lacustrine deposits in Eastern Canada have illite or chlorite mica as the dominant minerals. Soils
consisting of these minerals are generally considered to be non-swelling, although there may be large shrinkage
upon drying if the initial void ratios are high. The Champlain Sea (or Leda) clays of the Ottawa Valley and St.
Lawrence Lowlands are one of a number of such clays. In the Western provinces, the montmorillonite shales from
Cretaceous formations in the Interior Plains provide the active clay minerals that give rise to expansive soils.
Most of these deposits are found in lacustrine clay deposits that were once large glacial lakes. The clay deposits
surrounding Lake Agazziz near Winnipeg and Lake Regina near Regina are examples of these deposits.
The natural environment, as well as anthropogenic changes in the environment, can produce significant changes
in the water content of the surficial soils. In the more humid parts of Canada, clays sensitive to shrinkage have not
previously been SUbjected to drying to the extent now occurring as a result of construction and the introduction of
non-native vegetation. The surficial clays in Western Canada have historically been subjected to arid or semi-arid
climatic conditions. The development of surface structures, such as light residential housing, inevitably leads to a
change in moisture fluxes across the ground surface as the result of irrigation, leakage from underground utilities,
or vegetation.
A general description of the soil and environmental conditions that can lead to significant volume changes in
the near-ground-surface soils are described in the next section. This information can be used for the preliminary
identificatiop of potentially expansive soils areas. This is followed by a description of how to measure appropriate
soil properties for use in a heave analyses, as described in Section 15.3.1.
15.2.1  Identification of Expansive Soils:  Clay Fraction,  Mineralogy, Atterberg  Limits, 
Cation Exchange Capacity 
A potentially expansive soil contains a relatively high percentage of highly active clay minerals. The expansion of
the diftUse double layers within the clay fraction results in changes in water content. Methods of identifying the key
features ofpotentially expansive soils are described in this section.
Standard hydrometer analyses can be used to identify the 'clay-sized' fraction that is less than two microns in
diameter (ASTM D-422). However, not all particles of this size fraction are clay minerals. It is recommended that
Foundations on  Expansive Soils  219 
the mineralogy of the  'clay fraction' be measured. The most common method ofidentif)dng and quantifying the clay 
mineralogy is  from an X-ray diffraction analyses. All of the clay mineral types are in close  proximity on the  X-ray 
trace  and  consequently  it is  important  to  use the  correct  opening for the  X-rays,  (i.e.,  a  narrow  slit),  along with  a 
qualified technician when interpreting the test results.  Of primary importance is the quantification ofthe amount of 
montmorillonite (or Smectite) clay mineral in the clay fraction of the soil sample. 
The  Atterberg  Limits,  (i.e.,  Plastic  Limit,  Liquid  Limit,  and  Shrinkage  Limit),  can  be  measured  as  part  of a 
geotechnical investigation. The difference between the Plastic Limit and Liquid Limit is referred to as  the Plasticity 
Index.  The Plasticity Index is related to the  percentage of clay-sized particles  and the mineralogy of the clay-sized 
Van  der Merwe (1964)  provided a correlation between the  Plasticity Index, the percent of clay-sized particles,  and 
the potential  for  swelling as  shown in Figure  15.5.  Swelling potential ranged from  low  to  very  high.  The highest 
potential  for  swelling  occurred when  the  soil  had  a  high  percentage  of clay-sized particles  and  a  high  Plasticity 
Index. A soil can be described as  having a high potential for  swelling but the expansiveness of the soil will only  be 
revealed when the initial water content of the soil is low. 

.s::  40 

x  30 
>.  20

Very High 

-'"  Low 
o  10  20  30  40  50  60  70 
Clay fraction  of whole sample,  (%<2u) 
FIGURE 15.5 Classification ofpotential severity ofan expansive soil based
on the plasticity andpercent clay-sized particles (van de Merwe, 1964) 
A useful  index  that  can be  computed  from  the  Plasticity Index  and the  percentage  of clay fraction  (%clay)  is  the 
Soil Activity CAc): 
Ac  Plasticity Index / (%clay)  (15.1) 
Skempton (1953) classified clays as  'inactive' whenAc was less than 0.75;  'normal' whenAc was between 0.75  and 
1.25 and 'active' when Ac was greater than  1.25. It is clays in the 'active' range that cause the greatest difficulty with 
respect to  swelling (and shrinking).  Nelson and Miller (1992)  listed typical values for the Activity of various  clay 
minerals:  kaolinite, 0.33  to 0.46;  illite, 0.9; Ca-montmorillonite,  1.5; Na-montmorillonite, 7.2. Figure  15.6 uses the 
Activity  of the  soil  and  the  percent clay-sized  particles to  classify  the  potential  for  swelling  of compacted  clays 
(Seed et al,  1962). The amount of swelling that can be anticipated with clayey soils can range from less than 1.5  % 
to more than 25  % depending upon the activity of the  soil and the amount of clay-sized particles. 

220 Canadian Foundation Engineering Manual
Swelling Potential; 25%
____Swelling Polenlial = 5%
o 10 20 30 40 50 60 70 80 90 1 00
Percent clay sizes (Finer than 0.002 mm)
FIGURE 15.6 Classification ofpotential swell for compacted clays based on the 
Activity ofthe soil (Seed et aI,  1962) 
Table 15.1 wasfirstproposedbyHoltzandGibbs(1956) andrelatescolloidalcontent(wherecolloids are defined
to be particles lessthan 0.001 mmindiameter),PlasticityIndexandShrinkageLimitto thepotential forvolume
tableis notmeantto be usedas abasisforpredictingheave, butratherto provideapreliminaryassessmentofthe
potentialforvolumechange.It isusefultoaugmentthetablewithobservationsfrom localexperience.
TABLE 15.1 Potentialfor Expansion as Estimatedfrom  Classification  Test Data *
10- 20
Low < 10 < 15
> 15
* AfterHoltzandGibbs (1956). 
** Drytosaturatedconditions- underasurchargeof6.9 kPa(1 psi). 
*** Particleslessthan0.001 mmindiameter 
Figure 15.7 illustrates the generalpatternofpercentswell for a compacted, highlyplastic soil (Holtz and Gibbs,
1956).Whiletheamountof swellingmayvaryfromonesoiltoanother,thepatternof totalswelluponwettingfrom
variousdensityandwatercontentconditionsshouldshowthesametrend. Theresults illustratethatcompactionof
asoilatahigh densityincreasesthe amountofswellinguponwetting. Also, compactionatwatercontents above
Foundations on Expansive Soils  221 
--:g 1600 
.  iii 

10  15  20  25 30 35 40 
'\  I  I  I 
- Volume change - percent 
Surcharge load  =7 kPa 
I  I 
  Saturation  curve 
(Gs ==2.749 

)  I 
ir Standard AASHTO 
i curv 
10o,y I " "
VI  " 
/  I  \' .", "'I:
/ ·6% I  \  \  "-
"I'  I ,\' 
,  "-


" ....  I
\  '0% 
I  "-
- --

.,,'" j%
-_ ...

---- -----
Initial water content,  Wo (%)
FIGURE 15.7 Pattern ofpercent swell for a soil compacted at various water contents
and densities (Holtz and Gibbs, 1956)
Cation Exchange Capacity, CEC, is a measurement of the quantity of positively charged dissolved ions required
to satisfy the negative charge imbalance on the surface of clay particles, and is commonly quoted in terms of
millieqivalents per 100 grams of dry soil (Mitchell, 1993).
CEC is related to clay mineralogy and the amount ofclay-sized particles present in the colloidal fraction. High values
of CEC mean that there is a high surface activity in the clay fraction and consequently a greater potential for volume
change. CEC measurements are routinely available in most agriculture soil testing laboratories. Typical values in
meqllOOg of soil, for the three basic clay minerals are; kaolinite =  3 to 15; illite =  10 to 40; and montmorillonite
(smectite) =  80 to 150 (Mitchell, 1993).
For a soil with a given CEC, the potential expansion during wetting can also be affected by the valance ofthe cation
adsorbed on the exchangeable sites as well as the chemistry of the pore-fluid. Most agriculture soils laboratories
can measure the chemistry of the salts present in the clay using a 'saturation extraction' technique (Klute, 1986).
This involves adding water to a dry soil until free water is observed to form on the clays. The sample is then
centrifuged and the chemistry of the 'extract' is measured. The greatest potential expansion will occur when the
adsorbed cations are monovalent (e.g., Na+) and when the pore-fluid is dilute. The presence of divalent cations and
concentrated solutions can cause volume change due to swelling to be suppressed (Mitchell, 1993).
A soil property called the coefficient oflinear extensibility, or COLE, has been routinely measured by the U.S. Soil
Conservation Service, National Soil Survey Laboratory in the United States. The test measures the lineal strain of
an undisturbed, unconfined specimen when it is dried from one-third of an atmosphere of suction (i.e., 33 kPa), to
oven-dried conditions. The specimens are brought to equilibrium at one-third of an atmosphere and coated with a
flexible plastic resin. The COLE value of many soils has been related to the swelling properties of soils and has
been quite extensively used in the United States (McKeen and Neilsen, 1978; McKeen and Hamberg, 1981; Nelson
and Miller, 1992).
222 Canadian Foundation Engineering Manual
15.2.2 Environmental Conditions
Expansive soils are generally clay-rich sediments deposited in glacio-lacustrine lakes that have undergone
extensive drying since deposition. The drying is the result of evaporation from the soil surface and transpiration by
vegetation. The soils must be located in an environmental condition in which potential evapotranspiration exceeds
A useful index to quantify soil moisture deficiency was developed by Thornthwaite (1948) and is called the
Thornthwaite Moisture Index (TMI). The TMI categorizes climate primarily on the average precipitation conditions
and potential evaporation conditions. Negative values for the TMI indicate that the climate is arid, and consequently,
expansive soil may undergo significant seasonal swelling upon wetting (O'Neill and Poonnoayed, 1980). The
climate categories and the associated dimensionless Thornthwaite Moisture Indices are shown in Table 15.2.
TABLE 15.2 Climate Classification According to the Thornthwaite Moisture Index (1948)
Climate Classification
Extremely Humid >+40
Humid +20 to +40
Sub-humid oto +20
Semi-arid -20 to-40
Arid < -40
Computational methods that more accurately compute the actual evapotranspiration from the ground surface have I
been developed (Wilson et aI, 1991), The analysis involves the solution of a coupled heat and moisture mass
transport model. The model has been applied to specific sites (e.g., for soil-cover designs) as opposed to being used
to develop climatic maps.
15.2.3 Laboratory Test Methods
The one-dimensional oedometer (i.e., consolidation apparatus) has been used in many countries ofthe world to test
and obtain physical soil properties for expansive soils. The objective ofthe laboratory test is to assess the in situ stress
conditions and measure soil properties that can be used for the prediction ofvertical heave (Fredlund and Rahardjo,
1993). Although the consolidation test was originally developed as a laboratory simulation of compressible soft
clays, it can be also be used to provide valuable information on expansive soils. There are numerous test procedures
that have been proposed in the literature but the two most common tests are the Constant Volume swell test (CV
test), and the Free Swell test, (FS test). The test procedures for both ofthese tests can be found in ASTM designation
D- 4546-90. Both tests are conducted in a manner similar to a consolidation test with the primary difference related
to the procedure for the setup and the commencement of the test. Constant Volume Swell Test Procedure
The Constant Volume swell test is conducted on an undisturbed soil specimen that is trimmed into a consolidation
ring. The specimen is placed in the oedometer and seated under a nominal load. The specimen is then inundated
with water and as it attempts to swell, the load on the specimen is increased to prevent any volume increase or
swelling. When the specimen no longer exhibits a tendency to swell, the applied load is further increased in a series
of increments in a manner similar to that of a conventional consolidation test. Once the recompression branch or the
'virgin' branch ofthe consolidation curve has been established, the specimen is unloaded in a series of decrements in
order to establish the swelling index. The loading decrements are usually twice as large as the loading increments.
The Constant Volume swell test provides two important measurements that are required for predicting heave; namely,
Foundations on Expansive Soils 223
estimate of the swelling P" (or the corrected swelling pressure, P:) and the swelling
mdex, C
' Although the swellmg pressure of a SOlllS sometImes construed to be a soil property, the swelling pressure
is more conectly a measure of the in-situ stress state of the expansive soil. The undisturbed soil sample was taken
from its in-situ condition where it was subjected to the overburden stress (total stress). As well, the soil was subjected
to the effect of negative pore-water pressures (or matric suction). The total stress and matric suction combine on the
total stress plane to provide an indication of the initial state of stress in a soil. If the change in stress state is known
along with the swelling index, the volume change associated with stress state changes can be computed.
Consider the stress path followed in the laboratory when a soil specimen is tested using the Constant Volume test
procedure subsequent to sampling (Figure 15.8). Once the soil specimen is submerged in water, the specimen
attempts to swell while the matnc suction is dissipated. However, the total stress on the specimen is increased to
keep the specimen from increasing in volume. Gradually, the matnc suction within the soil specimen is reduced to
zero and the volume of the specimen has been maintained constant by the increase in the total stress. Figure 15.8
shows that the swelling pressure represents the sum ofthe in-situ overburden stress and the matric suction of the soil
translated onto the total stress plane. As such, the swelling pressure is dependent upon the in-situ matric suction.
"" f!
'(Ua - UW)in'ilU.... • Token load

Ideal stress - deformation path
Actual stress - deformaUon path
1 P, (uncorrected swelling pressure)
P', (corrected M t· ct' ( )
swelling a nc su lon, u. - Uw
pressure) "if
« " '-S'
,,<!- 4>'
,;;,,,>0v   • Assume nO volume change
tJ> g; dUring sampling
«-" <f).f"

FIGURE 15.8 Ideal and actual stress state versus void ratio path/ollowed
when performing a one-dimensional oedometer test
The measured swelling pressure will, however, be under-estimated unless the effect of "sampling disturbance" and
"apparatus compressibility" are taken into account (Fredlund, 1969). The interpretation of the Constant Volume
swell test must include a correction for the compressibility of the consolidation apparatus, the compressibility of
filter paper (if filter paper was used during the test), and the seating of the porous stones and the soil specimen.
Desiccated swelling soils have a low compressibility and the compressibility of the apparatus can substantially
affect the measurement of the swelling pressure as well as the slope of the rebound curve (Le., swelling index).
The compressibility correction can be measured by substituting a steel plug for the soil specimen and measuring
deflections accruing to the apparatus under each load increment. This correction is relatively consistent for a
particular consolidometer and its accessories. It is recommended that filter paper not be placed above and below the
soil specimen because of the magnitude of its compressibility. Figure 15.9 illustrates data from a Constant Volume
swell test, with and without a correction applied for compressibility.
Sampling disturbance will result in a measured swelling pressure that is lower than the in-situ value. (This
phenomenon is similar to the observed effect of "sampling disturbance" on the measurement of preconsolidation
pressure in a consolidation test on soft clays.) In the oedometer test, it is not possible for the soil specimen to return
to its precise in-situ stress state after sampling without displaying some curvature on the void ratio versus effective
stress plot (i.e., when going from the swelling pressure to the recompression curve or onto the virgin compression
224 Canadian Foundation Engineering Manual
curve). The procedure for determining the 'corrected swelling pressure' begins by correcting the laboratory data to
account for compressibility of the apparatus. The correction for 'sampling disturbance' is then applied in order
to establish the "corrected swelling pressure."
Pressure, P,
Test data adjusted
for oedometer compressibility
Unadjusted test data
Compressibility of
Log (0' - u,)
FIGURE 15.9 Adjustment ofone-dimensional oedometer laboratory test data to
account for the compressibility ofthe apparatus (Fredlund, 1983)
In 1936, Casagrande proposed an empirical construction that could be applied to saturated compressible soils in
order to determine more accurately the preconsolidation pressure. The empirical construction was, in essence, a
means to compensate for the effects of 'sampling disturbance.' A similar procedure to account for the effect of
"sampling disturbance" on the swelling pressure was proposed by Fredlund (1987) and is illustrated in Figure 15.10.
The slope of the rebound curve is used as part of the empirical construction procedure (rather than the slope of the
virgin compression curve). The final plot ofvoid ratio versus logarithm oftotal stress gives the plot shown in Figure
pressure, P's
-'- eo -'-'f__           stress state)
FIGURE 15.10 Constant Volume swell test results showing the empirical procedure to correct
the "swelling pressure" for the effect ofsampling disturbance (Fredlund, 1987)
Foundations on Expansive Soils 225
The "corrected swelling pressure," P:, is estimated as shown and the swelling index, C" is obtained from the slope
of the rebound curve. The 'corrected swelling pressure' and the swelling index are used as input data to the heave
analyses. Free Swell Test Procedure
The preparation of the soil specimen for the Free Swell test is similar to that described for the Constant Volume
test. Once the soil specimen has been prepared, a token load is applied to the specimen. Water is then added to
the oedometer pot and the specimen is allowed to swell freely until an equilibrium condition is attained. The soil
specimen is then loaded by doubling the load on the specimen and allowing equilibrium to be attained under each
applied load. Using this test procedure, the swelling pressure is defined as the load required for the void ratio
to return to its original value as shown in Figure 15.11. It is not necessary to apply a 'correction' for sampling
disturbance when using this test procedure. The effects of sampling disturbance are taken into account through
the test procedure. The swelling pressure measured from the Free Swell test and the 'corrected swelling pressure'
obtained from the Constant Volume test are generally quite similar (Fredlund, 1983).
FIGURE 15.11 Typical plot ofdata from a Free Swell oedometer test on an expansive soil
15.3 Unsaturated Soil Theory and Heave Analyses
The volume change experienced in an expansive soil should be understood in terms of the changes occurring in
the stress state of the soiL In other words, it is better to describe the expansion (or shrinkage) of a soil in terms of
changes in the stress state rather than in terms of water content.
When a soil becomes unsaturated it is necessary to use two independent stress state tensors to define the complete
stress state of the soil (Fredlund and Morgenstern, 1977). These two stress tensors are referred to as the "net
normal" stress tensor and the 'matric suction' tensor, and are defined as follows:
226 Canadian Foundation Engineering Manual
1: xy
l(cr, -u.)
1: yx
(cr y u

'" ]


0 ' 0 ,  ° total normal stresses in the X-, y-,  and z- directions, respectively, 
x y 
't  ,  't '  'tzx  =  shear stresses in the X-,  y-,  and z- planes, respectively, 
xy  yz 
=pore-water pressure, and 
un  =pore-air pressure. 
Matric  suction  is  defined  as  the  difference  between the  pore-air  pressure  and  the  pore-water  pressure,  (i.e.,  (u

u  ))' Changes in the environment (e.g., rainfall on the ground surface or evaporation of moisture from the ground 
surface), produce Ii change in the matric suction in the  soil, with time.  In other words, the matric suction tensor is 
changed. Likewise, changes brought about by construction (e.g.,  excavation of soil or the placement offill), cause 
changes in the net normal stress tensor. Independent soil properties are associated with each ofthe two stress tensors 
and consequently the stress tensors must be handled in an independent manner. 
The  osmotic component of soil  suction does not need to  be taken into  consideration unless  the  salt content of the 
soil is specifically ,changed in the problem under consideration. In general, this  is  not necessary because changes in 
the salt content in the laboratory and in-situ are similar. 
In an  expansive  soil,  the  volume  of the  soil  increases  as  a  result  of a  decrease  in  matric  suction.  Similarly,  the 
volume of the soil decreases as  a result of an increase in matric suction. The volume of the soil can also decrease in 
an independent manner as a result of changes in the external loading. Analytical procedures related to the prediction 
of heave  should be visualized  and understood  in terms  of changes  in the  stress  state  of the  soil.  It is  particularly 
important  to  visualize  the  expansive  soils  problem  in  terms  of two  independent  stress  state  variables  because 
changes  in  the  pore-water pressure are always three-dimensional  in  character while  external  loading  imposed  by 
man's design  are  more commonly one-dimensional  or two-dimensional in  character.  For example, a  vertical total 
load will produce a tendency for an outward movement in the lateral direction while an  increase in matric suction 
will have a tendency for inward movement in the lateral direction. 
Numerous testing procedures and analytical procedures have been proposed in the research literature for predicting 
the  amount  of heave  that  can be  anticipated  in  an  expansive  soil  under  various  soil  and  design  configurations. 
Generally,  the  success of each of the  methods  is  somewhat limited by incomplete  appreciation  of,  or inability to 
predict,  the  changes  in  environmental  conditions.  The  present  state-of-the-art  in  predicting  maximum  probable 
heave is satisfactory for most engineering purposes; however, the prediction ofthe rate at which the volume changes 
may take place is considerably more difficult because it depends upon the availability of water to the soil. 
The  rate  of heave  is  also  related  to  the  coefficient  of permeability  of the  soil.  Field  rates  of heave  are  strongly 
influenced by the macrostructure of the  soil,  which is  difficult (if not impossible) to  assess  from  a  laboratory test. 
The unpredictable availability of water from surface and subsurface sources is  also difficult to  predict. 
Foundations on Expansive Soils 227
Field shrinkage rates are affected by the efficiency with which moisture can be removed from the subsoil.
Evapotranspiration proceeds in a fairly predictable manner when the water content of the soil is high, but is less
predictable at lower water contents because of plant-root extensions, plant wilting, soil cracking and other factors.
15.3.1 Prediction of One-Dimensional Heave
The prediction ofheave (or swelling) can be canied out in a manner similar to that used when calculating consolidation
or settlement of a soft clay layer (Fredlund et aI, 1980). The prediction of heave requires an understanding of the
initial and final stress states and the defonnation modulus of the soil. The Constant Volume swell test provides the
necessary infonnation to assess the initial stress state (i.e., the COlTected swelling pressure), while the swelling
index, C, is taken as the defonnation modulus.
The swelling index, C" generally ranges from 1 0 to 20 percent of the compressive index, C s' for a particular soil.
Figure 15.12 shows approximate values for the swelling index values that have been conelated with the liquid limit
and the rebound void ratio of a soil (NAVFAC DM-7, 1971). The estimated values of the swelling index are useful
for obtaining an estimate of the swelling.

o 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Void ratio from Which rebound occurs
FIGURE 15.12 Correlation ofswelling index, Cs, with the Atterberg Limits and
in situ void ratio for an expansive soil (NAVFAC DM-7, 1971)
The equation of a straight line on a semi-logarithm plot can be used as the basic equation for the prediction ofheave.
The equation conesponds to the in situ stress paths projected onto the net nonnal stress plane (Figure 15.13).
Final stress conditions
Actual stress path
Analysis stress
Matric suction,(ua-U,.)
FIGURE 15.13 Actual stress path in situ and the stress path used in the analyses/or total heave
228 Canadian Foundation Engineering Manual
The stress path followed during the swelling of the  soil corresponds to  the rebound curve (i.e.,  CJ from  the initial 
stress  state  to  the  final  stress  state.  The  equation  for  the  rebound portion of the  swelling curve can be written  as  I
b.e  =C

6.e change in void ratio (i.e.,  e
- eo) 
e  void ratio 

final void ratio 
swelling index 
final stress state 
initial stress state or the "corrected" swelling pressure, P: 
The initial stress state, P , can be visualized in terms of the overburden pressure plus the matric  suction equivalent 

(see Figure  15.14): 
P =(a  -u)+(u  u)  (15.4)
o  y a  a  w 

total overburden pressure 
(ay  u

)  net overburden pressure 
(ua-u)  = matric suction. 
The pore-air pressure in the field can be assumed to remain at atmospheric conditions. The initial stress state, Po' can 
always be taken as the 'corrected swelling pressure,' P:. The final stress state, PI. must take into account total stress 
changes and the final pore-water pressure conditions. 
a  ± b.a  - u  (15.5)
y Y wf 
change in total stress due to excavation or the placement offill 
= final pore-water pressure. 
An estimate ofthe final pore-water pressures must be made as part ofthe assessment ofthe final stress state (Hamilton 
1969). Several possibilities can be considered as reasonable long-term pore-water pressure states. First, it could be 
assumed that hydrostatic conditions above and below an estimated water table would be reasonable. Assuming that 
this water table rises to ground surface is  the most conservative assumption and will produce the greatest estimate 
of heave.  Second,  it could also  be assumed that soil suctions  throughout the soil profile  will dissipate to  zero but 
that no  positive pore-water  pressures  will  develop.  Third,  it  could be  assumed  that  under  long-term  equilibrium 
conditions, the pore-water pressures will remain at a slightly negative value. This assumption produces the smallest 
prediction of heave. It has been observed that all ofthese assumptions related to final pore-water pressure conditions 
generally produce similar estimates of heave since most of the heave occurs in the uppermost soil layer where the 
matric suction change is largest. 
The selection ofthe final pore-water pressure boundary conditions can vary from one geographic location to another 
depending upon climatic conditions.  For example, the equilibrium suction below an asphalt pavement surface has 
been related  with  the  Thomthwaite  Moisture  Index.  On  many  small,  engineered  structures,  however,  it  is  often 
artificial causes such as leaky water lines and poor drainage that control the final pore-water pressures in the soil. 
The heave of an individual soil layer can be written in terms of a change in void ratio  as  follows: 
Foundations on Expansive Soils  229 
!3.h. heave of an  individual layers 


thickness of the layer under consideration 

change in void ratio of the  layer under consideration (i.e.,  eli e)
e. initial void ratio of the soil layer,  and  . 
final void ratio of soil layer. 
The change in void ratio, !3.e
,  in Equation 15.6 can be computed using Equation  15.3 to  give the following form: 
=  final  stress state in  the soil layer, and 
initial stress state of the soil layer. 
The total heave from several soil layers, !3.H, is equal to the sum of the heave for each soil  layer. 
15.3.2 Example of Heave Calculations
Figure  15.14  illustrates the calculations  required to  predict the potential  heave  from  a  2-meter layer of expansive 
soil.  The  initial  void  ratio  is  1.0,  the  total  unit  weight  is  18  kN/m
and  the  swelling  index,  C"  is  0.1.  Only  one 
oedometertest was performed on a soil sample taken from a depth of 0.75  m and the measured, 'corrected' swelling 
pressure was 200 kPa.  It is assumed that the 'corrected' swelling pressure is  constant throughout the 2-meter layer 
and that the ground surface will be covered with an impermeable layer such as asphalt. The suction in the soil below 
the asphalt will decrease with time due to the discontinuance of evaporation and evapotranspiration from the ground 
surface. It is assumed that the final  pore-water pressures will eventually go to zero at all  depths. 
swelling  Total 
pressure (kPa)  pressure  (kPa) 
o  200
Layer 1  m  kNlm

2 meters 
--------1----· y 
- -- 13.5
Layer 2  0'1 m  y=  ' 
swelling  rc  -
-------- '"---- Undisturbed  I 
Assumed P'. 
Layer 3  1.0m 
sample for  I 
--- -.27.0
y  y  test 

Assume:  1) Surface is covered with an impermeable layer 
2) Fina! pore water pressure equals zero 
1 + eo  POI 

will equal overburden pressure 
Po will  equal corrected  swelling pressure 
Calculations:  5 
Layer 1 ah  :: 500  x  log _4_.- = 41,2  mm 
1  1 + 1.0  200 
0,1  13.5  293 
Layer 2 ah, "  500 x 1 + 1.0  log -200- .  mm 
Log pressure 
Layer 3 ah, =1000  Xi  log 
=43.5 mm 
Total  Heave 114.0 mm 
FIGURE 15.14 Example illustrating heave calculations for 2-meter layer of
expansive soil when matric suction becomes zero
2S0 Canadian Foundation EngineeringManual
""1' i
The 2-meterlayeris subdivided into three layers, the top layerbeing the thinnest. (Normally more than 3 layers
wouldbeusedto obtainan accurate solution).Theamountofheave in each layeris computedbyconsidering the
mid-pointofeachofthethree layers. Theinitialstress state, is equalto the 'corrected'swellingpressure atall
depths.Thefinalstressstate,PI isequaltotheoverburdenpressure.Equation 15.7isusedtocalculatetheheavefor
eachlayer. ThecalculationsshownonFigure 15.14revealatotalheaveof114mm.About36%ofthetotal heave
15.3.3 Closed-Form Heave Calculations
The calculationofthe amount ofheave depends primarilyontheswellingpressure and theswelling index ofthe
Equations 15.7 and 15.8. The assumption is made thatthefinal soil suctionwill be zero. Figure 15.15 shows the
generallayoutofthe geometryunderconsideration,dividedintoanumberofequallyspacedlayers.
. 2
(I - 1)h 
(active ..L 
1-1 1---._-\
!!f." \\\\\\\
P'S pgh
FIGURE 15.15 Idealized geometry profile usedfor the "closed-form" solution
for the amount ofpoten tial heave ifthe soil suction becomes zero
Theexpansiveclaylayeris assumedto startatthegroundsurface.Thesoilis assumedtobecomewet(Le.,thesoil
suctiongoesto zero) to adepthwheretheoverburdenbecomesequaltotheswellingpressure.Thetotalheavefor
anexpansive soil can then becomputed as shown in Figure 15.16. Thetotal heave increases significantly as the
swellingpressureof thesoilincreases.However,itmustbepossibleforwatertoentertheentiresoilprofileinorder
500 1000
Correctedswellingpressure, P',(kPa)
FIGURE 15.16 Closed-form calculation oftotal heave when the soil suction becomes zero
"!i!iiIC .
Foundations on Expansive Soils 231
15.4 Design Alternatives, Treatment and Remediation
Following are  some  general  guidelines  regarding  the  design  of foundations  on  expansive  soils  and  the  control  of 
the  'active  zone.'  The  basic  concept behind the  design of a  foundation  system  on  expansive  soils  involves  giving 
detailed attention to control of the environment (e.g.,  moisture movement) or to  isolation of the structure from  soil 
movement. In general,  it is not prudent to attempt to resist movement imposed by swelling soils.  Rather, it  is better 
to  attempt  to  control  the  environment (i.e.,  moisture  control)  surrounding the  structure.  Suggestions for  moisture 
control are given following a description of possible foundation designs for expansive soils. 
15.4.1 Basic Types of Foundations on Expansive Soils
There are three general foundation alternatives for expansive soils: 
shallow spread footings 
a pier and beam system, or 
stiffened slab-on-grade. 
Shallow  spread  footings  are  the  most  common  type  of foundation  for  light  structures.  Generally  there  is  little 
engineering design associated with these foundations and consequently these structures suffer distress when placed 
on  expansive soils.  It is  often difficult to  convince  owners  that additional funds  should  be initially  invested in  an 
adequate  foundation  that is  placed on expansive soils.  Generally,  an  initial investment in  engineering consultation 
will prove to  be a wise investment after a few years. 
15.4.2 Shallow Spread Footings for Heated Buildings
Shallow foundations may be economical and give adequate service for certain structures on soils oflow-to-moderate 
volume-change potential in humid to sub-humid regions. The foundations should be reinforced to minimize effects 
of seasonal edge movements  and non-uniform bearing surfaces, such  as  over service trenches.  The spread footing 
foundation  should  perform  satisfactorily  provided  there  are  no  deep-seated  or  long-term  effects  such  as  major 
changes in the water table  (Le.,  pore-water pressure  conditions)  or vegetation  conditions.  Shallow spread  footing 
foundations will not likely perform well under severe environmental conditions. 
Good engineering design practices include giving consideration to the following issues: 
Positive surface drainage should be provided away from the structure by carefully selecting the slab surface 
and the  outside grade elevations; 
Placing the slab on a granular, free-draining fill; 
Ensuring stable and uniform moisture conditions under and around the foundation; 
Excluding  deep  root  penetration  under  the  foundation  and  protecting  against  undetected  leakage  from 
underground piping; 
Preventing the back-up of water through poorly backfilled trenches; and 
Providing adequate perimeter insulation around the foundation to eliminate steep thermal gradients through 
reactive soils under and around the foundations. 
Other precautions worth consideration as part of the superstructure design include: 
Utilization of flexible  framing,  cladding and partitioning construction; 
Provision ofadjustable-length interior columns and slip joints in non-load bearing partitions to accommodate 
differential movements;  and 
Providing  free-spanning  of floors  and  roofs  between  load-bearing  exterior  walls  and  frames,  wherever 
232 Canadian Foundation Engineering Manual
15.4.3 Crawl Spaces Near or Slightly Below Grade on Shallow Foundations
In  addition to  the recommendations given above, crawl-space designs require that special attention be  given to  the 
following issues: 
Provision of adequate drainage slopes to sump areas and drainage-tile beds within the crawl space; 
•  Provision of adequate ground cover in the crawl space to  control evaporation of moisture from the soil; 
Provision of adequate heat supply and insulation to prevent frost penetration below footings and to control 
extreme thennal gradients in soils below and around foundation units. This is necessary to prevent excessive 
accumulation of moisture or the drying in the underlying soils;  and 
Provision of adequate ventilation of the  crawl  space throughout all  seasons to prevent condensation  on  or 
within structural materials in the crawl  space. 
The  magnitude  of total,  differential,  and  tilt movements  of shallow  foundations  will  depend on the  many  factors 
related to the active zone and the reactivity of the soils on the site.  Even for soils of relatively low volume-change 
potential,  some  differential  movement  of the  perimeter  spread  footing  units  relative  to  central  units  will  occur. 
Relative  movement  should  be  anticipated  and  provision  should  be  made  for  convenient  length  adjustment  of 
columns supporting central beams  and floors.  Central  load-bearing partitions carried directly on strip  footings  are 
not recommended  unless  an  effective means  can be incorporated for adjusting the  elevation of the superstructure 
below the main floor level. 
The magnitude of total and differential movements experienced by structures on shallow foundations is influenced 
by the net unloading of the soils. This is the case even with a typical full-basement excavation and a lightweight one 
or two-storey building. Although central footings may be designed to carry equal structural loads and to have similar 
dimensions to ensure similar stress increases in the underlying soils, the net area-unloading effect of the excavation 
has  a much  deeper influence.  Consequently,  deep-seated heaving tends  to  affect central footings  much more  than 
perimeter footings. The provision of adjustable columns in important for these situations. 
Serious attention must also be given by designers to stacks, chimneys, heating ducts, furnaces, and other equipment 
placed  on  ground-supported  basement floors.  On moderate-to-high  volume-change  soils,  differential  heaving  of 
basement floors  will  likely become  excessive  and  objectionable  to  many occupants  over a  period of a few  years 
after construction. This problem can best be addressed at the design stage by providing a structural basement-floor 
system  that  spans  between  foundation  supports.  It  is  also  possible  to  provide  an  adjustable flooring  system  that 
can easily be maintained by the occupant or owner. All shallow foundations may be subject to tilt deformations or 
localized settlement caused by non-uniform soil reaction to moisture changes or localized influences, such as  deep 
tree roots,  leaks, or other localized sources ofwater. 
Grade beams and basement walls, which also serve as retaining walls for clay backfills ofmoderate- to high-swelling 
potential, should be designed to resist horizontal earth pressures in accordance with an  equivalent fluid-pressure. 
15.4.4 Pile and Grade-Beam System
A  pile  and  grade-beam  foundation  system  generally  provides  a  superior  foundation  to  that  of a  spread  footing 
system.  The piles  are  generally  of the  cast-in-place  concrete  type  but other types  of piles can  also be  used.  The 
piles need to be extended below the depth of seasonal ground movement.  A grade-beam system supports the loads 
between the piers.  A  structural  floor  slab  system tied  into  the  grade-beam  generally  performs  well.  However,  a 
floating  slab resting  on the  grade beams can also prove to  be  a satisfactory  system.  Compacted sand or gravel  is 
generally placed below the floor  slab but sometimes the floating slab is placed directly on the soil Figure  15.17. 
Foundations on  Expansive Soils  233 
Fill grade 
,.·lrf1l (
.  Reinforced 
Full length steel 
Partition walls 
(suspended from floor joists 
or supported on  compressible 
"Floating floor slab" 
Sand or gravel fill  __. 
or may not be  used) 
Void  space beneath grade 
beams between piles 
Concrete piers drilled  into 
firm  bedrock or to depth 
below level of active zone 
FIGURE 15.17 Typical layout for a drilled pile and grade beam foundation system (Nelson and Miller, 1992)
Good engineering design practices  include giving consideration to the  following issues: 
•  The piles need to be extended well below the depth of seasonable movement and  have  sufficient  depth  to 
resist uplift resulting from the expansion of the soil, 
The piles may be straight shafts  or may be belled at the bottom, as  deemed most suitable for the structure 
under consideration, 
The piles need to be reinforced to resist the potential uplift forces  associated with the expansion of the soil 
in the upper portion of the profile, 
•  Consideration may be given to the possibility of using a material along the upper  portion  of the  pile  that 
reduces the adhesion of the soil to the pile in the swelling portion of the profile, 
The grade-beams need to be tied into the grade-beams, 
A space must be left below the grade-beams (i.e., between the locations ofthe piles) in order to accommodate 
potential upward swelling of the soil below the grade-beam. The amount of space that must be left  below 
the  grade-beam varies depending upon the soil conditions but will commonly be in the order of 150 mm 
(6 inches) or more, and 
Precautions previously mentioned related to surface drainage need to be  respected  for  pile  and  grade-
beam systems as well. 
15.4.5  Stiffened Siabs-on-Grade
Stiffened  slabs-on-grade  (Figure  15.18)  are  not  a  common  type  of foundation  system  in  Canada because  of the 
adverse weather conditions.  Frost penetration further accentuates the potential for foundation movements, over and 
above that due to expansive soils.  In situations where a stiffened slab-on-grade might be considered as  a potential 
foundation type, a competent structural and geotechnical engineer should be retained to design a system that can be 
ensured to perform satisfactorily. 
234 Canadian Foundation Engineering Manual
1'- - - - - .....l'- - - - - - ...., ,"-' - - - - ..... 1 I - -11- - - .....
I I !! I I II
A ~ ~ T B B -n--- I I II
1______11 _______ ,1 ______I 1 ___'1___ _
------ ..,.------" ------,
I 11 I I I
I :1 ,I I
, I )1 I
, I, II 1
1 I" ___ /1 _____ _
Exterior Beam (frame) Interior Beam
FIGURE 15.18 Typical layout for a reinforced slab-on-grade (Nelson and Miller 1992)
The Building Research Advisory Board Recommendations for the design ofresidential stiffened slabs takes hogging
and sagging ground movement conditions into consideration. A qualified structural engineer must be retained to
design the necessary reinforcement that must be included in the slab.
15.4.6 Moisture Control and Soil Stabilization
Measures that ensure a control on the movement of moisture in and out of the foundations soils should be made
a part of the foundation design. Numerous procedures have been used in various parts of the world. Some of
the procedures have proven to be successful in soine countries while not providing a successful solution in other
. countries. It is important that a qualified geotechnical engineer be retained to ensure that moisture control and soil
stabilization techniques are assessed and applied in an appropriate manner for the situation at-hand. There are a
number of details that can be added to the design to ensure the successful performance of the foundation system.
Figure 15.19 shows a concrete apron placed around a foundation, tied into the foundation system. Low permeability
aprons have been found to perform quite well in reducing differential heave in expansive soils.
" "
Shallow subdrain .  
desirable at outer
edge of moisture barrier
FIGURE 15.19 Details such as a (concrete) membrane tied to the foundation system can
assist in controlling infiltration (Nelson and Miller 1992)
Foundations on Expansive Soils 235
Some other possible soil stabilization techniques are as follows:
Soil Stabilization: Many expansive soils can be rendered essentially inert through the addition of lime. Lime
stabilization designs can be considered; however, in most situations it will be sufficient to use a lime modification
procedure. Lime modification usually requires that only 6 to 8 percent lime be mixed with the soil. Soil testing is
required in each situation to determine the amount of lime that should be added. A decision regarding the appropriate
amount of lime to add can be based the reduction in the plasticity index as a.result of adding lime. It should be noted
that the addition of lime to the soil may not be a potential option in many situations because of the toxic nature of
Remove and Replace: In some situations where the expansive soil is relatively shallow, it may be possible to
excavate and replace the expansive soil. The cost-effectiveness of this option will usually control whether or not it
is an option that should be considered.
Mixing for Homogenization: It is the highly heterogeneous nature of expansive soils deposits that give rise to
differential ground movements that are essentially equal to the total ground movements. However, the excavation of
a soil deposit followed by the subsequent recompaction of the soils will result in reduced and more uniform ground
movements. The use of the mixing and recompaction of a soil deposit should be used under the supervision of a
qualified geotechnical engineer. A laboratory-testing program should be undertaken to verify that excavation and
recompaction will produce the anticipated results.
Pre-wetting: The swelling potential ofa soil can theoretically be eliminated by soaking the soil prior to construction.
However, this practice may not produce satisfactory results. It would appear that this practice has been used
successfully in some parts of the world but there are probably more situations where it has been unsuccessful. The
problem appears to arise with the difficulty in obtaining a uniform wetting of the soil. If the soil is cracked near to
the ground surface, it appears that the expansive soils in the upper portion of the profile swell closed and then it is
not possible for further wetting to occur in a reasonable period of time.
After construction has been completed, the moisture in the soil often undergoes a slow redistribution process with
the result that the structure suffers distress. The pre-wetting technique should only be used after thorough study
under the supervision of a qualified geotechnical engineer.
Chemical Stabilization: There are chemicals other than lime that can be used to stabilize an expansive soil. These
chemicals may be salts, enzymes or other chemicals. It is important that the effect of the addition of any particular
chemical on the behavior of the soil be thoroughly studied and appraised by a qualified geotechnical engineer prior
to it use.
Surcharging: Placing a load, such as an inert fill, overtop of an expansive soil may significantly reduce the potential
for volume change. The amount of load applied will depend upon the swelling pressure of the soil. The greatest
amount of swelling generally occurs near to the ground surface. Consequently, an inert fill can be quite effective
in reducing swelling even though it will not likely eliminate the total amount of swelling. The amount of load that
would need to be applied for a particular situation should be assessed by a qualified geotechnical engineer.
Capillary Barriers: A capillary barrier is a more coarse-grained material such as a silt, sand or gravel, placed
over the expansive soil. Normally a coarse-grained material is thought of as a highly permeable material that will
merely allow water to reach the expansive soil with the result that swelling will occur. However, the coefficient of
permeability can be extremely low when a coarse-grained material has a low degree of saturation. The water storage
capacity of a finer/coarser series ofsoil layers can be made quite large. This form ofcapillary barrier can be extremely
effective in reducing the amount of water that reaches the expansive soil by storing infiltrating moisture near the
surface where it can be released back into the atmosphere by evapotranspiration. The capillary barrier needs to be
designed such that it has the appropriate air entry value and storage properties for the situation-at-hand. The des.ign
of the capillary barrier must be consistent with the climate and drainage conditions at the site. Capillary bamers
have been effectively used around light-engineered structures to reduce the amount of distress to the structure.
236 Canadian Foundation Engineering Manual
Each of the above potential solutions to handling an expansive soil should be reviewed and studied by a qualified
geotechnical engineer. The need for input from a qualified geotechnical engineer cannot be over-stated because case
histories reveal that often steps taken to remedy the expansive soils problem merely aggravate the situation.
Site and Soil Improvement Techniques 237
Site and Soil Improvement Techniques
16 Site and Soil Improvement Techniques
16.1 Introduction
A number of techniques can be used to improve the strength and compressibility of subsoils that are too weak to
support conventional shallow foundations. These include pre loading, vertical drains, dynamic consolidation, vibro
processes, lime treatment, ground freezing, blast densification, compaction grouting, chemical grouting, vacuum
preloading, and electrical strengthening methods.
A state-of-the-art report by Mitchell (1981) presents a comprehensive review of soil improvement techniques that
complements the techniques highlighted in this section. Additional information is provided by Bell (1993) and
Moseley (1993).
16.2 Preloading
16.2.1 Introduction
The preloading technique was developed in the 1940s, mainly in connection with highway construction. Since that
time, it has been applied to a wide variety of projects, including buildings, storage tanks, airfields, flood control
structures and land reclamation projects (Johnson, 1970a,b). The technique has been used to improve all types of
natural cohesive soils (including peat), deposits of loose sand and silt, and fills, including waste materials. It is
uneconomical and impractical for structures with heavy, concentrated loads.
16.2.2 Principle of Preloading
Ground treatment by preloading implies placing a load on top of the ground to be treated well in advance of the
construction of the proposed structure. The magnitude of the pressure exerted by the load would usually be greater
than the maximum pressure imposed by the proposed structure.
Temporary surcharge loads (defined as loads in excess of the final loading) are frequently employed to decrease the
time required for preloading. Such surcharge is needed when the pre loading is intended to minimize the effects of
secondary compression.
Methods of applying the preload are by earth fills, water loading, vacuum under impervious membranes, and
groundwater lowering by well-points or deep wells.
After removal of the preload, including any surcharge, a slight rebound is to be expected. However, the rebound
is, usually, very small and negligible. Construction of the final structure may start over the precompressed
soils immediately after removal of the preload. The principle of the pre compression technique is illustrated in
Figure 16.1.
238  Canadian Foundation  Engineering  Manual 


a:;J --l 


a  >=  0"'"
0  00 
.....  :J 

o...  new C"-
0...  «  "- ne 
I- Z 

0  u 
-'  V'! 
«  z 
Z  0 
u:::  U 


1.1.. "" 
o  t:; 
o  Z 
Z  0 
w  U 




"" I-
W  Z

:J  a5 
«  I-
t- W 




a::  2-


;:::  ot:.


0  ::l 
:;J  ::::J 
t>< «  Q!.  Cl  U
u..  U  I- :;J 
I-- ;:::  VI  I- 0<
u  z  Z  t><
Z  <.!l "-
u.J  :;)  0- W 
Z  0  '" 

;:::  I-
'GJ  u 
u.J  "" 
I  I.Ll
.....  V> ..... 

-' Cl  -' -'  I--
I- :;)  Z 
I-- I-
l- V)
«  I--
I- I-

w  w  0  0  u.J 
VI  U  -' 
2- 0- V) :J  2-
FIGURE 16.1 Principle ofpreloading technique
16.2.3 Design Considerations
16·.2.3.1 Evaluation of Settlement
In the  planning  of a  preloading  program,  the  magnitude  and  duration of consolidation under preload need  to  be 
evaluated. This can be done using the methods in Chapter 11  of this Manual. Settlement calculation requires that the 
stratigraphy and properties of the subsoil be determined through a soil investigation of the  site. Parameters such as 
in-situ shear strength, preconsolidation pressure, compression index, swelling index, coefficient of consolidation or 
modulus number, stress exponent, and permeability may be required. In coarser soils having a permeability exceeding 
about  1 x  10-

mis, settlement will occur rapidly,  and a preloading time  of a  few months  is normally sufficient.  In 
more  impervious soils, vertical drains (Section  16.3) may be  employed to  accelerate the consolidation. Stability
Ifthe foundation soils are weak, the design of a preloading program must also consider stability. This may require 
stability berms and the use of it controlled rate of loading to  enable gain in strength of the  foundations soils during 
preloading (Tavenas  et aI.,  1978). Vertical drains  also  serve the purpose of accelerating the  increase  in strength of 
the foundation soils. Size of Preloading Area
The preloading area must exceed the limit ofthe final structure in such a way that the stresses induced at any depth 
in the foundation soils by the preload under the  edge of the proposed structure are uniform and at least equal to  or, 
preferably,  greater than the  final  stresses  at  that location.  In addition,  it  is  desirable  to  extend the preload  area to 
allow for possible future  extension of the proposed structure. 
Site and Soil Improvement Techniques 239 Instrumentation
A  proper  instrumentation  program  should  be  mandatory  for  all  pre loading  schemes  to  provide  a  continuous 
monitoring  of the  results.  The  instrumentation  should  be  designed  to  monitor  in  representative  locations  and 
depths  the  magnitude  and  duration  of settlement  during  preloading,  during  removal  of the  preload,  and  during 
the  construction period of the  structure.  Monitoring  of the  final  structure  for  several years  after  construction  is  a 
recommended practice. 
In pervious  soils,  instrumentation  may  be  limited  to  a  number  of settlement  points  installed  at  final  foundation 
level to monitor the overall settlement of the compressible subsoil. In  clayey soils, the instrumentation may include 
settlement gauges  at  variable  depths  below  ground  surface  and piezometers  to  monitor  the  rate  of pore  pressure 
dissipation and degree of consolidation. Inclinometer may be added to measure horizontal deformations at the edge 
of a preload fill  and to monitor settlement distribution with depth,  if required. Foundation Design for the Final Structure
Once the  preloading  technique  has  been  applied  on  a  compressible  ground  to  make  it  capable  of supporting the 
foundations for the final structure, design ofthese foundations may be done using usual procedures as recommended 
for  spread footings  or rafts.  In  sizing footings,  particular attention must be  paid to  those  shallow foundations  that 
rest at the surface of a thin layer offill over soils with little or no  confining pressure around the footing. Advantages and Disadvantages
The preloading technique  offers  several advantages  over other ground improvement methods, in particular,  when 
time  restrictions  are  not  critical  and  materials  used  to  apply  the  preload  are  available  at  low  costs.  The  main 
advantages are:  . 
Post-construction  settlement  is  reduced  to  relatively  small  values,  in  particular  for  foundations  over 
heterogeneous soils. 
The preload material may be re-used as  general backfill of a site after the  completed preloading. This may 
represent an important economic factor in the selection of a ground improvement method. 
The pre loading technique is a 'quiet' one, free of vibrations or noise usually accompanying other techniques 
of ground improvement,  and should be considered when environmental restrictions are imposed. 
The main disadvantages of the preloading technique are: 
settlements may take longer than expected, causing delays that may be economically unacceptable; 
disposal  of fill  material  used  for  preloading  may  represent  a  costly  item,  unless  it  can  be  reused  on  the 
site; and 
future  extensions  of the  proposed structure  need to  be  considered  in the  preloading program,  which may 
impose an undesirable initial investment for  the foundations  of the  future structures. 
16,3 Vertical Drains
16,3.1 Introduction
Settlements in clayey soils take a long time to develop. The time required depends on two main factors - linearly on 
the permeability ofthe soil, and exponentially on the  drainage path, i.e., the thickness of the  settling soil layer (see 
Section  11.11). The  time can be reduced appreciably if the  drainage path is  shortened by means  of vertical drains. 
The spacing between the drains  controls the length of the  drainage path.  For instance,  drains  installed at a spacing 
that is  a  tenth  of the  thickness  of a soil  layer that is  drained  on  both sides  could  accelerate  the  settlements  about 
25  times.  Furthermore,  as  the  permeability of the  soil  in  the  horizontal  direction is  generally  several times  larger 
than the  permeability in the vertical direction  and the  drainage  when using vertical  drains  occurs  in  the horizontal 
240 Canadian Foundation Engineering Manual
direction, the time for completion of the primary settlement is further shortened.
The potential benefits ofusing vertical drains became obvious very soon after Terzaghi in 1926 published his theory
of consolidation. Thus, vertical drains have been used in engineering practice for more than 50 years. At first,
vertical drains were made of columns of free draining sand (sand drains) installed by various means. In about 1945,
premanufactured band-shaped drains were invented (Kjellman, 1948) and, since about 1970, the technical and
economical advantages of the premanufactured band-shaped drains have all but excluded the use of sand drains.
16.3.2 Theoretical Background
For the analysis of the acceleration of pore pressure dissipation in fine-grained soils (drainage) and subsequent
settlement (consolidation), the theory developed by Barron (1948) and Kjellman (1948) is used (Hansbo, 1979).
The theory is summarized in the Barron and Kjellman formula as follows (see Figure 16.2):
D2 [D ] 1 (16.1 )
t - In--O.75 In--=-
d l-U
time from start of consolidation
zone of influence of a drain
equivalent diameter of a drain
average degree of horizontal consolidation
coefficient of horizontal consolidation
The zone of influence of a drain is the diameter of a circle having the same area as the area influenced by the drain,
i.e., if in a given large area of size A there are n drains placed at some spacing and in some grid pattern, each drain
influences the area Nn. Thus, as shown in Figure 16.2, for drains with a centre- to-centre spacing, clc, in a square
and triangular pattern, the zone of influence D, is 1.13 clc and 1.05 clc, respectively.
In the case' of sand drains, the equivalent diameter, d, is often taken as equal to the nominal diameter of the sand
drain. In the case ofband shaped drains, there is no agreement on what to use as the equivalent diameter ofthe drain.
One approach used is simply to equalize the outside surface area of the bandshaped drain with a circular sand drain
of the same surface. However, this approach does not recognize the difference between the usually open surface
of the pre manufactured drain and the rather closed surface of the sand drain, nor the differences between various
makes of band shaped drains.
Strictly speaking, the equivalent diameter of a bandshaped drain should be termed 'the equivalent circle diameter'
I ;
to separate it from 'the equivalent sand drain diameter'. The equivalent circle diameter is the diameter of a circle
having the same free or unobstructed surface to the surrounding soil as the drain. It has been suggested that the Ii
equivalent circle diameter of a sand drain is the sand drain nominal diameter multiplied by the porosity of the sand f
in the drain. The porosity of loose, free-draining sand is normally about 0.4 to 0.5. Thus, the equivalent circle f·
I '
diameter of a sand drain is about halfof the nominal diameter. However, the consolidation time is not very sensitive
to variations ofthe equivalent diameter. The spacing is important, however, as is also the total length of drains used
at a site.
, ..
Site and Soil Improvement Techniques 241
2 H
Tv:=: - O. 1 - 'g (1 - 0)
u 1 -
t :=: T •
1 D 3 1
T :=: - ( In - - -) In
h 8 d 4 l-U
o SQUARE GRID ) D fir .c/c :::: 1. 13 c/c
6 TRIANGULAR GRID ---7) D --=rr . c/c = 1.05 c/c
FIGURE 16.2 Principle o/vertical drains
For bandshaped drains of, commonly, 100-mm width, values proposed as the equivalent circle diameter have ranged
between 30 mm and 80 mm, and full-scale studies have indicated that the performance ofsuch drains have equalled
the performance of sand drains of;WO mm to 300 mm in nominal diameter.
The degree of consolidation at a certain time, U, is defined as the ratio between the average increase of effective
stress, ~ o   , in the soil over the applied surcharge causing the consolidation process, i.e., L'!o' /q. In practice, it is
determined from measurements of either pore pressure increase or settlement and, alternatively, defined as I minus
the ratio between the average pore pressure increase in the soil over the total pore-pressure increase resulting from
the applied surcharge, i.e., I - u/ U , or, the amount of settlement obtained over the final amount of settlement at
completed consolidation, L'!S/S. The consolidation ratio is generally based on pore pressure increase, because pore
pressures can be determined at the start of a project, whereas the value of the final settlement is not obtained until
after the project is completed. However, as pore pressures and pore-pressure dissipation vary with depth and, in
particular, with the distance to the drains, pore-pressure observations can be unreliable measures of the degree of
The horizontal coefficient of consolidation, c
' is critical for the design of a vertical drain project, because the
dissipation time calculated according to the Barron and Kjellman formula (given above), is inversely proportional
to the c value. The c value is not usually determined in a soils investigation program, but the c. value is. In a
h h
homogeneous soil layer, the horizontal coefficient of consolidation, c
' is generally about two to six times greater
than the vertical coefficient of consolidation, cy' The extent of the mobilization of a coefficient higher than the c
value depends on the disturbance to the soil caused by the installation of the drains. For sand drains, in particular
displacement-type sand drains, a c value that is greater than the c
value can rarely be mobilized.
The coefficient of consolidation varies widely in natural soils. In normally consolidated clays, the c,
value usually ranges from 1 x 10-
to 30 X 10-
/s. In silty clays and clayey silts, the c, value can range from
5 x 10-
to 50 X 10-
242 Canadian Foundation Engineering Manual
The coefficient of consolidation is nonnally detennined from laboratory testing of undisturbed soil samples or
preferably, in-situ by detennining the pore-pressure dissipation time in pore-pressure sounding (piezocone CPT):
The actual value to use requires considerable judgement in its selection, and it cannot be determined more Closely
than within a variation range of three to five times. This means that an engineering design of a project requires
supporting data for selection of the c value to avoid having to employ a very conservative approach.
16.3.3 Practical Aspects to Consider in Design Sand Drains
The sand used in a sand drain must be free draining, which means that the portion of fine-grained soil in the sand
must not exceed 5% by weight and preferably be less than 3%.
As indicated by Casagrande and Poulos (1969), the installation of full-displacement sand drains (driven drains) in
soils that are sensitive to disturbance is not advisable. Jetted sand drains will eliminate much of the undesirable
effect associated with driven sand drains, but at the cost of creating a muddy site, and, potentially the destruction of
the drainage blanket on the ground.
Furthennore, before pouring the sand into the water-filled, jetted hole, the water must be flushed clear so that fines
suspended in the water do not mix with the sand, rendering it non-free-draining. It is more difficult to control the
risk of fine-grained soil sloughing off, or being flushed off the side of the hole and mixing with the stream of sand
during the pouring procedure changing the sand into the non free-draining kind.
Sand drains are apt to neck and become disrupted during the installation work, or as a consequence of lateral
movements in the soil.
Despite the stated disadvantages, sand drains can be useful where large amounts of water are expected, in soils less
sensitive to disturbance by the installation, and where the ratio of length to nominal diameter is not greater than 50,
and the ratio of spacing to the nominal diameter is larger than 10. The Premanufactured Bandshaped Drain
The bandshaped drain consists in principle of a channelled (grooved or studded) core wrapped with a filter. The filter
serves the purpose of letting water freely through while preventing fine soil particles from entering the channels.
The channels lead the water up to the ground surface, or to above the groundwater table, or down to a draining layer
The filter must be able to receive water not only from clay soil, but also from coarser soils, such as silty, fine sand
typically found in lenses, or layers in most fine-grained soils. Furthermore, while the drain is receiving water over
its full length, it must be capable of discharging this water through a very short distance of its length, as discharge
through the end of the drain is a rather special case. Consequently, the filter must have a permeability coefficient no
smaller than that offine sand, i.e., approximately I x 10-
The pre manufactured drain is often manhandled on the construction site: it is dragged on a truck floor and on the
ground, it is left in the sun and in the rain, it gets soaked and is then allowed to freeze, it is stepped on, etc. This
puts great demands on strength, in particular wet strength, on the filter and the glue used to hold the longitudinal
filter seam together. A rip or tear in one spot of the filter of an installed drain can reduce the drain perfonnance
The drain core must provide a free volume (free cross-sectional area) large enough for the water flow not to be
impeded, i.e., the well resistance must be small. The water flow in a chain used for accelerating settlement is very
small and the required free volume is small. Typically, the water flow is smaller than about 5 litres per day or 3
/min, which is about what a dripping tap produces.
Site and Soil Improvement Techniques 243
The drain core must be flexible enough to deform both by folding (due to lateral soil movement) and by axial
when the soil settles around the drain. The settlement, or strain in compression, can be greater than
%. While a drain cannot be sufficiently soft that it compresses this amount axially, it must be able to 'microfold'
because of the imposed compression strain, without breaking or blocking the passage of water, i.e., creating an
excessive well resistance.
At the same time, it must be strong enough to resist large, lateral, soil pressures without collapsing and effectively
.. closing the longitudinal drainage path in its channels. For instance, at a depth of 20 m in a clay soil underneath an
·ernbankment 10 m high, the effective lateral soil stress is 200 kPa to 300 kPa, and it is desirable that the drain be able
to resist this pressure without developing excessive well resistance.
General Aspects
Ifthe settling soil contains thin layers, bands, or lenses of permeable soil, this will have little effect on the vertical
drainage - the case ofno drains. On the other hand, when vertical drains are used, the permeable layers will drain the
consolidating soil and lead the water toward the drains. Such bands or lenses (even ifvery thin) can be quite effective
in channelling water. Normally, therefore, as stratified or banded clays occur in many places, the assumption of the
Barron and Kjellman formula ofhomogeneous soil is commonly not valid. Furthermore, the consolidation time will
not be governed by the spacing of the drains but by the distance between the permeable layers (on condition, of
course, that the horizontal layering has not been broken down by the drain-installation procedure used and that the
filter permeability is not too small).
Figure 16.3 illustrates the acceleration of settlement by means of vertical drains underneath an embankment on
compressible soil. The upper sketch indicates the back pressure in the drain created by use of a filter with too low
coefficient ofpermeability forcing water to rise in the drain inside the filter to a height above the water table, where
balance is achieved between the inflow and the discharge of water.
The lower sketch in Figure 16.3 illustrates the similar condition created by the ponding of rain water and melting
snow in the depression created by the initial amount of settlement. The ponding is due to insufficient horizontal
drainage on 'the ground surface. In a design of a vertical drain project, the expected amount of settlement must be
calculated and a drainage scheme designed that ensures a horizontal gradient from the treated area at all times.
The build-up of back pressure will have a temporary effect on the time development of the settlement. Anyone
unfamiliar with this phenomenon will observe a flattening out of the time-settlement curve and draw the false
conclusion that all of the primary settlement has been obtained. However, eventually the back pressure will
disappear, and the settlement, delayed due to the back pressure, will recur.
The acceleration of settlement by means of vertical drains is only efficient where the applied surcharge creates a
final effective stress in the soil that exceeds the preconsolidation pressure in the soil. This requirement often governs
the installation depth of drains.
Often, it is unnecessary to install the drains beneath an embankment of width B beyond a depth that is greater than
BI2 to B.
The minimum width of the installation of the vertical drains should extend to the foot ofthe embankment. To reduce
the magnitude of differential settlement causing bowing of the surface, it is recommended that drains be installed to
a distance outside the embankment equal to about half the height of the embankment.
The theoretical analysis is sensitive to the parameters used as input, in particular to the coefficient of consolidation.
Unless prior experience is available from or a nearby site where a similar installation took place, any design using a
theoretical analysis, whether the Bamon and Kjellman formula or a more involved one, has only qualitative value.
When data are available from similar sites, the design analysis can be assisted by a thorough site investigation, which
is aimed at establishing the presence of bands or lenses of permeable soil at the site. This requires pore-pressure
244 Canadian Foundation Engineering Manual
sounding  (Chapter  4)  and  continuous  sampling  of undisturbed  soil  with  subsequent  laboratory  identification and 
DRAINS  -.. 
- -
". ,

G.W.  = groundwater level; W = water level in drains. 
FIGURE 16.3 Typical cases ofback pressure in vertical drains
In most cases,  spacing of the drains  will have to  be estimated and the project monitored by means of observations 
of the development of settlement and pore pressures over time. The settlement should be  monitored not just at the 
ground surface, but also as to its distribution with depth. Piezometers need to be carefully installed in relation to the 
drains. Naturally, the  data will be oflimited value unless coupled with a thorough site investigation. 
Often a predesign testing program is carried out to determine the parameters to use in the design and the spacing and 
type of drain to use. Typically, more than one spacing are used. Equally important is to  arrange for a reference in the 
form  of an  area with no  drains,  so  that the positive effect of the drains can be correctly established. The literature 
contains many comparisons between a theoretical calculation of what the time development of the settlement would 
have  been without  use  of vertical  drains  and  observed  development  underneath  a  drained  area.  Whereas  such  a 
comparison  has  the  'advantage'  of always  'proving'  the  desired  positive  effect  of the  drain  installation,  it  is  of 
limited engineering value. 
When  monitoring  the  effect  of a  drain  installation,  it  is  important  that  the  observation  period  be  long  enough, 
preferably  up  to  the  end  of the  primary  consolidation.  Experience  has  shown  that  large  potential  errors  can  be 
associated  with  a value  of achieved  degree  of consolidation determined before about  75  % of the  settlement has 
been obtained.  Other Uses of Drains
The use of premanufactured band-shaped drains is not limited to strengthening of clay soils and the acceleration of 
settlement underneath embankments, fill  areas such as  airports,  and reclaimed land.  Other applications have been 
to  release  pore pressures  in tailings  dams  and  in  slopes,  and to relieve pore  pressures behind retaining  walls.  The 
drains have also been used in combination with load application by the vacuum method described in Section 16.11 
Site and Soil Improvement Techniques 245
(see also Holtz and Wager, 1975).
Dynamic Consolidation
Dynamic consolidation, also known as dynamic compaction or heavy tamping, is a method of ground improvement
that was developed in the early 1 970s by Louis Menard.
In essence, the technique consists of the application of high intensity impacts over the surface of the ground to be
treated by means of a free-falling, heavy steel or concrete weight. The strain waves generated by these impacts
travel to considerable depths and rearrange the soils into a denser, more compact state.
Dynamic compaction is used to increase bearing capacity and decrease total and differential settlement within a
specified depth of improvement to allow the use ofshallow footings for different types of civil-engineering projects,
including runways, coal facilities, dockyards, etc. It has been used to reduce the liquefaction potential ofloose soils
in seismically active regions. Some unique applications include compaction under water, displacing unsuitable
materials such as peat, and collapsing sinkholes and abandoned mine workings.
The main advantages offered by the process are its low cost, rapidity of execution, applicability to a large variety of
constructed fills and loose natural soils, and usefulness for improving sites underlain by peat and landfills.
16.4.2 Methodology
Dynamic compaction involves the use ofheavy steel or concrete block tamper weighing typically 100 leN to 200 leN
and dropped in free-fall from heights ofup to 30 m using heavy crawler cranes. Under such conditions, compressible
soils have been compacted to depths of as much as 15 m. With special equipment, it is possible to drop heavier
weights and improve soils at greater depths.
The distribution of the impacts and the sequence of the application are critical in achieving successful compaction,
particularly if deeper zones are to be treated. The impacts are normally applied in increments, each complete
coverage of the working surface being referred to as a phase. The early phases, also called the high-energy phases,
are designed to improve the deeper layers with impact points at a spacing dictated by the depth of the compressible
layer. Generally, the phase is followed by a low-energy phase with contiguous impacts (hence, the name 'ironing
phase'), which is mainly designed to densify the surficial layers.
Although the process is effective on saturated coarser grained soils and can be used even on sites where the water
table is near the surface, it is nevertheless complicated and possibly ineffective in fine-grained soils by the creation of
increased porewaterpressures during compaction. This phenomenon will reduce the effectiveness ofthe subsequent
phases, unless it is recognized and measures taken to promote and accelerate the dissipation of pore pressures. If
not, remoulded soil conditions can develop.
During its execution, the process must be continuously monitored, first, to evaluate the degree of soil improvement
being achieved, and, second, for environmental considerations such as potential damage to nearby structures and
annoyance to people. Earthworks to level the site after each phase and to replace uncompactable materials with
suitable soils are also part of the operation. Final verification testing to ensure that the specification requirements
have been fulfilled must be performed upon completion of the treatment.
246 Canadian Foundation Engineering Manual
16.4.3 Ground Response Ground Deformation
The  impact  of falling  weight  upon  the  earth  compacts  the  natural  soils  and  collapses  voids  in fill  soils,  causing 
deformation in both vertical and horizontal directions. The induced settlement is  significant inasmuch as it provides 
an indication of the efficiency of the process. The magnitude of settlement depends on the initial compactness ofthe 
soils, the applied energy per unit area, and the adequacy of the compaction plan.  Generally, the induced subsidence 
amounts to between 5 % and 7 %  ofthe thickness of the loose soils being treated.  Several sites are reported to have 
subsided as much as  2 m as  a result of the treatment. 
The horizontal deformations, although accompanied by some degree ofcompaction, are important mainly because of 
the potential displacement of adjacent structures. In the case of fine-grained soils, noticeable swelling of the ground 
surface generally occurs, as  much as  0.3  m  in certain instances, but diminishing quickly to become undetectable at 
a distance  of 4 m  to  5 m from the treated area. Ground Vibration
The impact of the tamper on the ground generates compression, shear, and Rayleigh waves. Rayleigh waves, which 
travel at surface, generate vibrations that may affect nearby structures as well as people living and working in them. 
These vibrations  normally have a  frequency of about 5 Hz to  8 Hz,  and the shock accompanying each blow of the 
falling  weight is  felt for about one second. 
Peak particle velocities are  generally used to define damage criteria for building structures and annoyance levels to 
persons. The peak particle velocities increase significantly with the densification of the soil mass. Pore Pressures
Where the water table is within the depth of influence ofthe process, the densification is accompanied by an increase 
ofporewater pressures. In the case of sands and gravels, these pressures dissipate quickly. In less pervious soils, the 
induced pore pressures may take days or even weeks to dissipate fully. Soil Improvement
The engineering properties ofsoils densified by the process are improved to a depth and degree that depend largely on 
the proper assessment of the several variables and parameters characterizing each project.  The variable parameters 
include the weight of the tamper,  its height of fall  and impact surface  area,  the  grid spacing, the  number of phases, 
the total compactive energy, and the time delay between phases.  The non-variable, or given, parameters include the 
existing soil types, the initial soil conditions, groundwater levels, and the environment. 
A convenient approximation of the maximum depth of influence is (d  ) in metres given by the following empirical 
=a...)WH  (16.2) 
W  weight oftamper (N) 
H  height of fall  (m) 
0.  a factor usually taken as  5 x  10-

t07 X  10-

(dimension  m / N ) 
Improvement achieved by means of dynamic consolidation has been observed to increase with depth to a maximum 
at a specific depth and then diminish with depth until reaching a depth, d , below which the soil properties remain 
unchanged. The specific depth is  approximately between one third and one-half of the maximum depth. 
Site and Soil Improvement Techniques 247 Control Testing
Quality-control measures must be undertaken to ensure that improvement does indeed occur and that the engineering
characteristics of the soil have been attained as specified.
Control testing may be divided into three types: production, environmental and specification. Production control
includes quality-assurance aspects, such as logging the impacts, elevation survey of the working surface,
and monitoring the changing soil characteristics during treatment using in situ geotechnical testing methods.
Environmental control consists ofmeasuring ground vibration levels and carrying out boundary surveys to minimize
the effects of the tamping operations on adjacent properties. When compacting in close proximity to existing
structures, it may include instrumentation designed to detect potential movement and deformation. Specifications
or verification controls are carried out after the treatment is completed to certify that the objectives of the treatment
have been attained.
The most frequently used in-situ geotechnical methods for production or specification control have been the
pressuremeter, the standard penetration test (SPT), the static cone penetrometer (CPT), and the dynamic cone
penetrometer. Geophysical surveys have proven useful in soils that are difficult to test with conventional methods,
such as rockfills. Other types of field measurements include observation of pore pressure, measurement of peak
particle velocities, and subsurface settlement. Field vanes, dilatometer and plate test loading are also used. Limitations of the Process
The safe use of dynamic consolidation presupposes the knowledge and understanding of its limitations. The main
limitations are depth, soil type and soil conditions, the environment, the engineering requirements, and the climate.
Reviewing these various factors, the following guidelines are suggested: Depth
Using conventional lifting equipment, it is possible to treat free-draining granular soils to depths of 15 m and fine
silty sands and silts to depths of 10 m to 12 m. Greater depths of improvement have been achieved with special
equipment, but the effidency of the process beyond depths greater than 15 m remains unproven, except for coarse-
grained material, and any such application should be approached with caution. Soil Types and Conditions
The efficiency of the process for improving clays and clayey alluvials remains unproven. Such applications should
be considered only for projects where the potential economy is sufficiently important to justifY a full-scale field
Because oftheir loose state and the presence of numerous voids, most types of constructed fills, including clay fills,
can be successfully improved by the process. However, secondary settlement due to loss of volume accompanying
the decomposition of organic matter remains a phenomenon that is difficult to assess.
The application of the treatment may be complicated if the water table is closer than 2 m below ground surface.
Remedial measures will generally consist of raising the grade with imported materials but may also consist of
pumping to lower the groundwater level. High pore pressures generated in fine-grained soils can adversely influence
the results of the process. Environment
The application of the process is accompanied by noise, vibrations, gusts of air, and permanent soil deformations.
It may therefore be disadvantageous when used in an urban area.
Noise generated by the impact is generally muffled and not objectionable. By contrast, the roar of the lifting crane
246 Canadian Foundation Engineering Manual
engine can be quite loud and may have to be abated by suitable equipment. Typically, the noise level (impact +
engine noise) will reach 130 dB at a distance of 12 m, decreasing below 100 dB at 100 m.
Air gusts will displace materials around the edges of impact points, sometimes projecting chunks of earth and
mud to considerable distances creating a risk of damage to property and injury to persons. Suitable precautions are
Vibrations generated by the process are not nonnally damaging, unless peak particle velocities exceed 50 mmls, but
will at a much lower value cause annoyance to persons nearby. It should be stressed, however, that the reaction of
people to vibrations is generally unpredictable.
When working in close proximity to existing structures, the intensity of vibrations can be reduced by the use of a
lighter tamper, or a lower height offall, or a combination of both. The observance of the following fonnula should
ensure a safe operation.
D distance from impact (m)
W= weight of tamper (N)
H= height offall (m)
Notwithstanding the above, an experienced specialty contractor may work safely as close as 3 m from underground
services and 6 m from sound structures.
Soil defonnation that is pennanent can occur as much as 6 m away from the limits of compaction. At the ground
surface swelling occurs, which could raise and crack pavements and sidewalks; at depth, lateral displacement occurs,
which could affect underground structures. Pennanent horizontal displacements of 20 mm at a distance of 4 m, and
of 6 mm at a distance of 6 m, have been recorded. Engineering Requirements
Although many types of soils will be improved by the process, the attainable engineering characteristics vary
considerably. As a general guideline, the limits in Table 16.1 are proposed, where the presumed design bearing
pressures are given. The design bearing pressures correspond to serviceability limit states not exceeding 25mm
total settlement.
TABLE 16.1 Presumed Design Bearing Pressures for Soils Treated by Dynamic Consolidation
Type of Soil
Fine-grained alluvials, silty fills
Design bearing pressure (kpa)
100 - 150
Heterogeneous fills 100 - 200
Fine silty sand, hydraulic fills 200
Coarse sand, gravel 300
Well-graded gravel, rockfill 400-500 Climate
Adverse weather conditions such as heavy rainfalls, snow, and frost do not preclude the use of the process; they
may however, have a considerable influence on its costs. Dynamic compaction has been carried out in Canada in
temperatures as low as -15°C to -25°C.
Site and Soil Improvement Techniques 249
16.5 In-Depth Vibro Compaction Processes
16.5.1 Introduction
Improving soils by using depth vibrators started in the early 1930s, when the concept was developed that deep
deposits of soil could be compacted by means of a machine that would carry the source of vibration to the desired
depth (Steuermann, 1939). Since then, depth vibrators have been used extensively throughout the world for the
densification of granular deposits. The process uses elongated vibrators, and, when combined with water jetting, it
is generally known as Vibro compaction.
In the early 1960s, the use of more technologically advanced depth vibrators led to their use for the improvement
of fine- grained soils and fill materials by replacing the fines in the soil, which are washed out during the jetting,
with coarse materials, which then are compacted by the process. This application of the process is called Vibro
replacement or, somewhat incorrectly, the 'stone column' method, although it is essentially no different from the
Vibro compaction method.
The Vibro process provides an economically attractive and technically feasible basis for the treatment of soils that
exhibit (in their natural state) insufficient strength to support anticipated building loads.
16.5.2 Equipment
The essential element of the depth vibrator is a heavy tubular steel body, approximately 300 mm to 400 mm in
diameter and 4.5 m to 5.0 m long, within which are housed eccentric discs that rotate on an axial shaft. In order
to reach depths greater than 5 m, the vibrating unit (called the vibrator) is connected to simple extension tubes of
approximately the same diameter. The complete assembly is suspended from a conventional crane.
Two types of vibrator are in use: an electrically driven vibrator with a frequency of 60 Hz, and ahydraulically driven
vibrator with variable frequency. The power demand of the electrically powered vibrators is generally about 35 kJ
to 50 kJ, although vibrators of up to 100 kJ effect are also available.
16.5.3 Vibro Processes Vibro Compaction of Loose Cohesion less Soils
The vibrator is allowed to penetrate the soil under its own weight (approximately 30 kN to 60 kN, depending upon
the total length of the unit) with the help of water or air jetting from the nose cone, and the induced vibrations. After
penetration to the required depth, the water flow is reduced and the vibrator is withdrawn in small incremental lifts
ensuring uniform compaction of the soil from depth to grade. Vibro compaction will cause a reduction in volume
of the soil up to 10 %, often leading to substantial reductions in the level of the site surface. If the elevation of the
site is to be maintained, granular material (either imported or from other areas of the site) can be added around the
vibrator. As illustrated in Figure 16.4, the added material gravitates down around the vibrator to the base ofthe hole,
where it is compacted and integrated into the natural subsoil by the action of the vibrator.
Since the vibrations produced at depth emanate from a point close to the bottom end of the vibrator, and since these
vibrations radiate in the horizontal plane, there is little difficulty in achieving uniform densification with increase
in depth. The radial densification of granular soils (even though the vibrations are produced in the horizontal plane)
is limited, however. In well-graded sands, centre-to-centre spacings approaching 3 m to 3.5 m may be sufficient to
achieve a density index in the order of 70 %. Closer spacings can produce density indexes of approximately 90 %
(D' Appolonia, 1953).
The centre-to-centre spacings used for individual sites depend not only on the degree of compaction required, but
also on the material to be densified. While correlations between spacings and compactness condition achieved have
been undertaken, these can only be related to sites having identical soil conditions and where the same type of
250 Canadian Foundation Engineering Manual
vibrator is used. Normally, a spacing of about 1.5 m is required for fine sand. Most of the compaction takes place
within the first five minutes at any given treatment depth. Vibro Replacement in Soft Cohesive Soils and Inorganic Fills
The equipment used in this process is identical to that for Vibro compaction. The vibrator sinks under its own weight,
assisted by vibration (and water or air as a flushing medium) into the ground until it reaches the predetermined
depth. Water is generally used as a flushing medium in fully saturated soils, and compressed air is used in partially
saturated soils. For work in saturated fine sand or silt, it is essential that the water level in the hole is kept at a level
that is at least equal to and preferably higher than the groundwater table throughout the advancement of the hole
and installation of the graveL
During the penetration of the vibrator, the water flowing up along the side of the vibrator washes out the fines in the
soil, leaving the coarser material in the hole. The lost material is replaced incrementally from depth to grade with
charges of coarse-size fill, usually well-graded gravel of size between 10 mm and 80 mrn. The vibrator is repeatedly
withdrawn and reinserted to ensure a uniform result. With each charge of gravel, the vibrator displaces the backfill
horizontally into the native soils, while at the same time, compacting underneath its bottom edge. Repetition of this
procedure forms an irregular, cylindrical gravel column between the bottom of penetration and working grade (see
Figure 16.4).
i ~
(a) ( b) (c)
: ~
  ' ~
.... ",....
FIGURE 16.4 The principle ofthe Vibro process
The diameter of the compacted column ranges normally between 0.6 m and 1 m and depends mainly on the strength
of the native soil, the sort of flushing medium used to create the hole, and on the time spent to compact and displace
the backfill. Columns are generally installed in a square or triangular grid pattern. The spacing between the column
centres ranges from about 1 m to 3 m and is mainly determined by the degree of improvement required to achieve
the following four basic objectives:
1. to limit total settlements;
2. to reduce differential settlements;
3. to achieve higher bearing capacity; and
4. to increase shear strength. !
Various theories have been proposed for the design and failure criteria of compacted columns (Hughes et aI., 1975).]
One of these theories considers the column as an axially loaded member of frictional material supported by theJ.'
passive resistance of the surrounding native soil. Accordingly, the ratio of .pplied s!re" on the column to p."ive ...... '
Site and Soil Improvement Techniques 251'
restraint is  a maximum  at the depth of maximum effect.  The resulting settlement depends  upon the induced radial 
strain in the soil when passive resistance develops. Hughes et aI., (1975) indicated that the vertical displacement of 
the stone column, within the range of service stresses, is smaller than half the maximum radial strain in the column. 
Settlement of a  treated  area may  be  estimated by  determining  a  stiffness  - 'elastic' modulus  or  modulus number 
(stress  exponent  usually  equal  to  0.5)  - of the  untreated  soil  and  the  columns  in  combination,  and  performing  a 
settlement calculation as presented in Chapter 11. 
16.6 Lime Treatment
16.6.1 The Action of Lime in Soil
Soil  improvement  by means  of mixing  lime  into fine-grained  soil  is  probably  the  oldest  of all  site-improvement 
methods.  It was used, for instance, in Roman roads 2,000 years  ago. When unslaked lime is mixed into moist soil, 
the following  four reactions take place: 
ion exchange; 
cementation (pozzolanic reaction);  and 
Hydration reduces the water content and raises the temperature of the soil. In the process, the shear strength of the 
soil increases. The hydration starts immediately and is finished within a short time. 
Ion  exchange  also  starts  immediately  and  finishes  early.  As  a  result  of this  process,  water-stable  aggregates  are 
formed, which have low compressibility and high permeability compared to the original soil. 
The  pozzolanic  reaction  is  comparatively  slow  and  continues  for  a  long  time.  The  resulting  cementation  of the 
soil particles  results  in  a  considerable  shear-strength  increase  and  reduction  of compressibility.  Carbonation  is  a 
reaction between the lime and air and results in a reduced strength. When the mixing of lime takes place below the 
groundwater table, its influence is minimized. 
The amount of lime necessary to achieve a maximum improvement of strength and compressibility is about 3 % to 
6 % of dry lime per dry weight ofsoil. The lime has to be mixed thoroughly with the soil and quickly, or the reaction 
will be incomplete. 
16.6.2 Surface Lime Treatment
Surface  lime treatment  consists  of spreading  lime on  a soil  to  be stabilized and mixing  it with a  rotary tiller.  The 
optimum  water content and  the  liquid  limit of the  soil  will  increase  and  the  lime-treated fine-grained  soil can be 
compacted using a sheepfoot roller or similar equipment. The method is used for wet and soft sites where the soil is 
very silty and difficult for construction equipment to travel on.  The lime treatment and compaction creates about a 
0.2-m-thick layer of soil, which, in addition to being strengthened, has become more pervious. It should be  noted, 
however, that unslaked lime is dangerous to inhale, and powdered lime spread on the ground may constitute a health 
16.6.3 Deep Lime Treatment
Lime can be mixed into the soil by means ofspecial equipment, which will produce a column oftreated soil (Broms 
and Boman, 1977, 1979; Holm et aI.,  1983). The lime column will be capable ofsupporting point loads much greater 
than those that the untreated soil can support. When lime columns are placed in a grid pattern over an area, they will 
have  the  combined beneficial  effect  of both increasing shear strength  (bearing  capacity) and  reducing  settlement, 
particularly the differential settlement.  Furthermore, because of the increased permeability of the  lime-treated soil, 
the lime columns may act as vertical drains and accelerate the settlement. 
252 Canadian Foundation Engineering Manual
Lime columns have been used to support embankments and spread footings; they have been used in trenches both
to retain the trench walls and to support sewage pipes placed in the trench; they have been used in combination with
pile foundations for buildings, where the piles support the structure and the lime columns the ground floor, as well
as the immediate area outside the building; and they have been used to stabilize areas damaged by landslides.
16.7 Ground Freezing
16.7.1 The Freezing Process
Controlled ground freezing for mining and construction applications has been in use for more than a century. This
method may be used in most soil or rock formations, but it is better suited to soft ground than to rock conditions and
is not suitable in coarse gravel, boulder soils, or expansive soils. Freezing may be used for any size, shape, or depth
of excavation, and the same physical plant can be used from job to job, despite wide variation in these factors.
Freezing is normally used to provide structural underpinning, or temporary support for an excavation or to prevent
groundwater flow into an excavated area. As the low permeability frozen earth barrier is constructed prior to
excavation, it generally eliminates the need for compressed air, dewatering, or the concern for adjacent ground
subsidence during dewatering or excavation. However, lateral groundwater flows may result in failure ofthe freezing
program, if not properly taken into consideration during the planning process. Furthermore, though subsidence may
not be of concern, ground movements resulting from frost expansion of the soil during freezing may occur under
certain conditions, and this potential hazard must be considered in the planning.
Freezing can be completed rapidly if necessary, or desirable, although the freezing rate is directly related to overall
costs, and rapid freezing is relatively more costly than slower freezing.
Frozen ground behaves as a visco-plastic material (exhibits creep), whose strength properties depend primarily on
the ice content, the duration of applied load, and the temperature ofthe ground.
The refrigeration plant and refrigerant or coolant distribution system may represent as much as 45 % to 60 % of the
direct costs 'of a freezing project. Furthermore, the direct costs, as well as the time required to complete adequate II
freezing, depend to some extent on the type of freezing approach used.
The thermal energy required to freeze ground is directly proportional to the water content of the soil. For coarse-
grained soils, the energy requirements are relatively low, provided no lateral groundwater flow occurs. Infine-grained
silt and clay soils, the energy requirement will normally be higher. As a rule of thumb, the energy requirements
in megaj oules per cubic meter (MJ/m3) of soil frozen will be between 9 and 12 times the percentage of the water
16.7.2 Exploration and Evaluation of Formations to be Frozen
One of the most important factors in devising a freezing system is to thoroughly explore the subsurface formation
to be frozen. If the nature of the subsurface structure is not well known, an adequate and efficient freezing system,
no matter how well it is designed, may not accomplish its purpose.
In order to determine what freezing facilities should be provided to stabilize the subsurface structure, the
characteristics of the materials to be frozen should be ascertained as accurately as possible. This requirement cannot
be stressed too strongly.
The exploration is generally accomplished by the drilling of boreholes, taking samples of material from all zones,
and observing water conditions below ground. A sufficient number ofexploratory holes should be drilled so that the
entire mass to be frozen is covered, and a complete record of the exploratory operations should be kept.
The exploration should supply the following information:
Site and Soil Improvement Techniques 253
•  nature ofmaterials:rock-shale-c1ay-anhydritesetc. atal1 depths;
•  w<;l.ter contentinal1 strata;
evidenceandamountofoccludedairor inwaterorairinaquifers;
•  evidenceofhorizontalwaterflowsinaquifersthroughmasstobefrozen.
16.7.3 References
16.8 Blast Densification
In favourable circumstances,deep compactionbyblastingcanbe an effective and economic means ofachieving
bearingcapacity,andreducesusceptibilitytobothstatic- andseismic-inducedliquefaction.
advancing a cased hole byjetting, vibration or other means; a borehole 150 mm in diameter is usually
sufficient;uncasedholeswithheavydrillingmudforhole support,andsubsequentlystemming,havealso
installing explosives at appropriate depths as drilling casing is withdrawn, or left in disposable plastic
Blast densification can often offer considerable economic advantages, as the majorpiece ofsite equipment is a
suitabledrillingrig. Suchequipmentisoftenmorereadilyavailableatremotesitesforlessmobilizationcoststhan
Areal impedimentto the applications ofblastdensification is the lackofcontractors who arepreparedto bid to
a predetermined specification for site improvement. Guidelines for blast densification are largely empirical and
trials are usuallyrequiredto determinethe optimumconfigurationofcharge size, depthanddetonationsequence.
Blastdensification,as in othertechniquessuchasvibroflotation,worksbylocalizedliquefactionofloosesaturated
Criteriaare available to estimatethe charges requiredto achieve full liquefaction (thatis, a zero effective stress
condition). Blasting is particularly effective ifa loose sand layer is overlain by dense sand which provides a
containmenteffect.Chargesaresetwithinthelooselayerandastand-offdistanceestablishedso asnottoweaken
thesurface dense layer. Charges are oftensetat the one-third to two-thirdspoints within the loose zone. Charge
densitiesof10 g/m3 to30g/m3 ofsoilto bedensifiedare commonlyused.
Aninterestingfeatureofblastdensifications,andonewhichis alsoobservedinvibroflotation,isthatwhilesurface
requireseveralweekstoreflectgroundimprovementintennsof thein-situtestmeasurements.
254 Canadian Foundation Engineering Manual
Safety, particularlyin urban environments, is often a perceived, ratherthan areal concern. The chargeperdelay
is oftenrelativelysmall,allowingoff-sitegroundvibrationsto beheldto acceptable limitsusingthe sametypeof
criterioncommonto piledrivingordynamiccompaction.
16.9 Compaction Grouting
Compactiongroutingis the injectionofvery stiff, lowslump (0.25 cmto 3.0cmslump)mortartype grout under
theuse ofinjectionusingsiltysoilswithouttheadditionofcement.Thisrequiresthatthesiltysoilsbewellgraded
andthatpumpingtrialsbe carriedout. The technique is usedforstrengthening loose disturbedorsoftsoils under
existing structures, for reduction ofsettlementduring softgroundtunnelling, compactionofsoils for earthquake
liquefactionresistance,andforsinkholefilling anddensification.
depth. Groutispumpeduntil oneofthe followingcriteriaisreached
• refusalatapredeterminedpressure,
• amaximumgroutvolume(or'take')isreached,and 
The resultant injection consists ofa homogenous grout bulb ora series oflinked grout bulbs, which are formed
around the endofthe groutpipes. The injection ofthe groutdisplacesthe in-situsoil and compacts the adjacent
QAJQC controlduringconstructionisachievedby:
slumptestswherecementis used
• controlmixwhennon-cementmixesareused
16.10 Chemical Grouting
Chemicalgroutingisthepermeationofsandsandgravelswithfluidgroutsto producesandstone likemassesorto
"fill"thevoids andtherebyreducewaterflow. Grouttypesconsist ofsodiumsilicate, acrylates, polyurethaneand
microfinecement. Moretoxicchemicalshaveenvironmentalrestrictions,whichwouldprecludetheiruse.
Sleeve port grout pipes are installed in a predetermined pattern (vertical, horizontal orhorizontal) in a grouted
borehole. Groutis injectedthroughtheportsatspecificdesignedintervalsandratestofullytreatthearea.Avariety
oftheprocessistermed"tubeamanchette"in Europe.
The process is suitable for cohesionless soils particularly clean sands and gravels although some effect can be
achievedinsiltysands. Theapproachisparticularlyusefulintrundling,utilitysupport, andgroundwatercontrol.
QAJQC duringconstructionisachievedbymonitoring: 
geltime records, 
• flow rates, 
shearwavevelocitymeasurementsusingcrosshole techniques, 
waterpressuretesting, and 
Site and SoillmprovementTechniques 255
intrusive testing methods such as Standard Penetration Tests.
16.11 Preloading by Vacuum
The principles ofusing vacuum for preloading of soft clayey soils were first introduced in the early 1950s (Kjellman
1952). When a vacuum is applied to a soil mass, it generates a negative pore water pressure. If the total stress
remains unchanged, the negative pore pressure increases the effective stress in the soil and this leads to consolidation.
A schematic of the vacuum preloading method is shown in Figure 16.5.
The working platform consists of a sand layer through which vertical.drains are placed in the soil. The vertical
drains must be above any sand layer to sustain the vacuum pressure. A flexible geomembrane (polyethylene) liner
covers the area and keys into an anchor trench that provides a watertight seal. A perforated pipe system is placed
beneath the liner to collect water. Specially prepared vacuum pumps capable of pumping water as well as air are
connected to the collection system. It is essential that the area be consolidated is totally sealed and isolated from any
surrounding permeable soils to avoid the loss of vacuum. Leaks must also be avoided. Since pinholes or cracks in
the sealing membrane are difficult to locate and repair, the membrane should be covered with water, which will also
minimize potential damage from foot traffic and wildlife. When the required preloading pressure is higher than the
capacity of the vacuum pumps, a surcharge fill may be used in conjunction with the vacuum method, as shown in
Figure 16.5. The fill must be free from stones or sharp objects. If a fill is placed on the membrane liner during the
vacuum period, it may be necessary to add a leak detection system under the liner to help locate leaks.
Vacuum pump
Water collection pipe
•_____ Sand layer
Soft clav
FIGURE 16.5 Schematic ofvacuum preloading system (modified from Shang et al. 1998)
The vacuum method has the following characteristics (Shang et al. 1998): (1) a vacuum pressure of more than 80
kPa (600 mm Hg) can be achieved in practice using available vacuum equipment, which is equivalent to a fi1l4m
to 4.5 m in height; (2) the lateral deformation of soil is inward due to the suction generated by the vacuum (instead
of soil" squeeze-out" encountered in a surcharge prcloading process, tensile cracks·may develop adjacent to the
treated area); and (3) there is no need to control the rate of vacuum application to prevent a bearing capacity failure
because applying a vacuum pressure leads to an immediate increase of the effective stress and hence strengthening
of the soil.
Despite a relatively good understanding of the principles of the vacuum method (Holtz and Wager 1975), the
technique was not used widely in geotechnical engineering practice until the early 1980s, due mainly to high cost.
The technology gained the attention of the Asian geotechnical community in the late 1980s (Qian et al. 1992) due
to advances in geosynthetics and the shortage of land along shorelines. Prefabricated vertical (wick) drains that
are effective, cost efficient and easy to install compared to sand drains, have made the cost of the vacuum method
256 Canadian Foundation Engineering Manual
A number of projects have been undertaken in the Netherlands, France, Malaysia, Sweden and China. One of the
largest projects was the East Pier Project in Xingang Port, Tianjin, China (Shang et al. 1998). The soil improvement
project was conducted on 480,000 m of reclaimed land using the vacuum preloading method. After 29 months
including a pre10ading period of 135 to 247 days, the average consolidation settlement reached 2.0 m, corresponding
to increases in undrained shear strength of two to four folds, as shown in Figure 16.6. The study showed that the
vacuum method was an effective tool for the consolidation of very soft, highly compressible clayey soils over
a large area. The technique is especially feasible in cases where there is a lack of suitable materials for use as a
surcharge and extremely low shear strength. Access to a power supply for the vacuum pumps is necessary.
16.12 Electro-Osmotic and Electro-Kinetic Stabilization
Electro-osmosis is a technique used for the consolidation and strengthening of soft, saturated clayey soils. When
a direct current (DC) voltage is applied to soil via electrode poles, the soil pore water will be attracted towards the
direction of the negative terminal ( cathode) due to the interaction of the electric field, the ions in the pore water and
the soil particles. If drainage is provided at the cathode and prohibited at the anode, consolidation will be induced
by electro-osmosis, resulting in the lower soil water content, higher shear strength and lower compressibility. In
addition, electrochemical reactions associated with an electro-osmotic process alter the physical and chemical
properties of the soil and lead to a further increase in shear strength (Mitche111993).
Casagrande (1937, 1959) first applied the technique of electro-osmosis to strengthen and stabilize soft silty clays in
the middle 1930s. Since then, successful field tests have been reported that used electro-osmosis to strengthen silty
clays and soft sensitive clays; to stabilize earth slopes and to reinforce steel piles installed in clayey soils (Bjerrum
et al. 1967; Casagrande 1983; La et al. 1991). Electro-osmotic consolidation has been considered for projects
requiring a rapid improvement in the properties of soft clayey soils.
(a) (b) (c) 
Water Contenl, % Void Ratloe
00 00 0 2 0 20 60
:     : - ...., .. 'T'T" :

.12 l
FIGURE 16.6 Soil properties before and after vacuum preloading consolidation, 
East Pier Project,  Tianjin,  China  (modifiedfrom Shang et al.  1998) 
Site and Soil Improvement Techniques 257
When  an  open  cathode  and  sealed  anode  condition  is  present,  a  negative  pore  water  pressure  is  generated  upon 
the application of a  direct current (DC)  electrical field.  In one dimension,  the pore pressure generated by electro-
osmosis, u  (x,  t ~ (0)  (kPa), at a distance x (m) to the cathode is  given by (Esrig  1968): 
Ueo{X) = -
YwU{x) (16.4) 
ke (m2/sV) = electro-osmotic permeability 
k" (mls)
hydraulic conductivity 
9.81 kN/m
= unit weight of water 
U(X) (V) electrical potential at  distance x to  the cathode 
Additional  information on  vacuum preloading  can  be  found  in  Thevanayagam  et al.  (1994),  Thevanayagam  and 
Nesarajah (1996). 
The equation shown above indicates that the pore pressure induced by electro-osmosis is negative and proportional 
to  the  electrical potential (i.e.,  it  has  a  maximum  magnitude  at  the  anode  and zero  at  the  cathode).  The  negative 
pore  pressure  results  in  an  increase  in  the  effective  stress  in  the  soil,  leading  to  consolidation,  as  described  in 
the  conventional  consolidation  theory.  Knowing the pore  pressure  generated  by  electro-osmosis,  the  time  rate  of 
electro-osmotic consolidation can be estimated by  conventional consolidation theory. 
The electro-osmotic permeability, ke'  governs the water flow in a soil mass under an electrical gradient in the shnilar 
way as  the hydraulic conductivity  governs  the flow  in soil  under a  hydraulic  gradient.  When both the  anode and 
cathode are open to  drainage and the hydraulic gradient is  set to  zero,  ke  can be determined by measuring the flow 
velocity across a soil plug using an empirical relation (Mitchell  1993) 
qe  water flow  vector due to  an electrical gradient (mls)
E  electric field intensity vector, defined as 
E=-V'U  (16.6) 
The power consumption per cubic metre of soil mass per hour is  calculated from: 
p  unit power consumption (kW 1m
K =  electrical conductivity of the soil  (l/Qm) 
Equation (16.7) indicates that the power consumption of electro-osmotic treatment increases with the soil electrical 
conductivity and applied electric field.  Table 16.2 summarizes the typical ranges of soil and electrical properties that 
are suitable and have been used for  electro-osmotic consolidation. 
TABLE 16.2 Design Parameters and Common Soil Properties in Electro-osmotic Consolidation
Parameter Unit Typical Range
, Hydraulic Conductivity  mls  10.

ke  Electro osmotic Permeability  m
/sV - 10.

K,  Electrical Conductivity of Soil  simens/m (lInm)  0.01-0.5 
E, Electric Field Intensity  Vim 20-100 
c  , Coefficient of Consolidation 
p, Hourly Power Consumption 

,  : ~
258 Canadian Foundation Engineering Manual
A two-dimensional electro-osmotic consolidation model was developed by Shang (1998) that can take the effects
of both preloading and electro-osmotic consolidation into account. The most predominant electrochemical effects
during an electro-osmotic process include the development of a pH gradient, the generation of gases and heating.
The pH of soil water will increase rapidly to as high as llor 12 at the cathode and decrease to almost two at the
anode. Consequently, metallic anodes will corrode. Oxygen gas is generated at the anode and hydrogen gas at the
1 cathode due to hydrolytic reactions. The electrical current also generates heating. The seriousness of these effects
is directly related to the applied voltage and current. It is usually desirable to minimize heating effects to reduce
power consumption. It has been found that applying polarity reversal and intermittent (pulse) current can reduce pH
gradient and corrosion and increase the electro-osmotic permeability of the soil, thus improving the efficiency of
electro-osmotic treatment (Shang et aL 1996).
The evaluation of electro-osmotic consolidation on a specific soil can be conducted according to the following
Determination of parameters
In addition to conventional soil properties such as the grain size, preconsolidation pressure, shear strength, water
content, hydraulic conductivity, kv' and coefficient of consolidation c
' the parameters required for a treatability
analysis include the electro-osmotic permeability, k
; electrical conductivity, K; intensity of electric field, E; and
power consumption, p. All these parameters can be determined from laboratory tests prior to field application. Table
16.2 lists the typical ranges of the major parameters for soils that are suitable for electro-osmotic treatment.
Electrical Operation System in Field Applications
The electrical operation system can be designed based on the parameters obtained from laboratory tests and from
the geotechnical investigation of the site. Typically, the electrode poles consist of metallic rods or pipes installed
vertically into the ground with prefabricated vertical drains installed at the cathode, as shown in Figure 16.7. The
depth of the electrode insertion should be equal to the thickness of the soil layer to be treated. The upper portion of
the electrodes in contact with the ground surface crust or top drainage layer should be electrically insulated using a
dielectric coating to avoid short-circuiting due to the presence of surface water (Lo et al. 1991).
The material, layout and spacing of electrodes and the applied voltage are ofutrnost importance to a successful field
application. Among the most commonly used conducting metallic materials used, the best results were reported
using electrodes made of iron and copper rather than aluminium (Sprute and Kelsh 1980, Mohamedelhassan and
Shang 2001). Electrodes made of carbon-coated steel rods and graphite have been used in laboratory studies to
prevent electrode corrosion.
The typical spacing between anodes and cathodes reported in the literature ranged from I m to 3 m (Casagrande
1983; Lo et al. 1991). In general, an approximately uniform electric field gives the best results (Casagrande 1983).
To achieve an approximately uniform electric field, the spacing between electrode rods ofthe same polarity should
be much less than spacing of the opposite polarity.
Surcharge q
r + l I I
~   r a l n I
+ I---
Calhode H
FIGURE 16.7 Schematic ofelectro-osmotic consolidation (modified from Shang 1998)
Site and Soil Improvement Techniques 259
Power supply capacity can be estimated based on the soil's electrical conductivity and electrode layout. It has been
found that a more dramatic voltage drop takes place at the soil-electrode contacts at a higher applied voltage, which
made the treatment less efficient (Casagrande 1983; Shang et aL 1996). It was also observed that the voltage drop at
the soil-electrode interface is affected by the electrode materials (Mohamedelhassan and Shang 2001). Therefore,
a lower voltage applied across smaller anode-cathode spacing is desirable to generate the required electric field and
special attention should be made for the electrode materials and configurations. However, the cost of electrodes
and installation should also be considered. The final design will be based on a balance of the cost of electrodes and
electrode installation as well as the treatment efficiency. For additional information, seeArrnan (1978), Broms (1979),
Mitchell (1981,1993), US. Navy (1983), Van Impe (1989) Hausmann (1990), Micic et aL (2003a, 2003b).
Electro-kinetic stabilization is a hybrid between electro-osmosis and chemical grouting. The infusion of certain
stabilization chemicals into silty and sandy soils is made more efficient by the application of an electrical potential
difference to the soil mass. The procedure is more effective in silty soils that are otherwise difficult to grout ordinarily.
Information on this technique can be found in Brams (1979) and Mitchell (1981). More recently, electrokinetic
assisted chemical stabilization has been applied to offshore calcareous soils (silts and sands) for stabilization of
petroleum platforms (Mohamedelhassan and Shang, 2003, Shang et aL 2004a and 2004b).
260 Canadian Foundation Engineering Manual
Deep Foundations - Introduction 
17 Deep Foundations - Introduction
17.1 Definition
A deep foundation is a foundation that provides support for a structure by means oftoe resistance in a competent soil
or rock at some depth below the structure, and/or by shaft resistance in the soil or rock in which it is placed. Piles
are the most common type of deep foundation.
Piles are usually installed to support compression, uplift, or lateral loads from a structure. Although capacity aspects
may be emphasized in design, the foremost reason for using piles is to reduce deformation, normally settlement.
Piles are also used to densify granular soils and so stiffen the soil andlor change the natural frequency of soil under
foundations for machinery, and are essential in situations in which water may scour foundation soils.
Piles can be pre-manufactured or cast-in-place; they can be driven, jacked, jetted, screwed, bored, or excavated.
They can be made ofwood, concrete, or steel, or a combination thereof. Bored piles oflarge diameter are frequently
referred to as drilled piers in Canada.
17.2 Design Procedures
The quality of a deep foundation depends on the installation or construction technique, on equipment, and on
workmanship. Such parameters cannot always be quantified nor taken into account in normal design procedures.
Consequently, it is often desirable to design deep foundations on the basis of test loading of actual foundation units
and to monitor construction to ensure that design requirements are fulfilled.
However, only a few projects are large enough to warrant full-scale testing during the design phase, and, in most
cases, tests (proof-tests) are performed only during or even after construction of the foundation. Therefore, it is
necessary to provide the engineer with appropriate design methods. Chapters 17 through 21 of the Manual present
methods applicable to the various types of deep foundations encountered in practice.
17.3 Pile-Type Classification
The classification ofpile types is governed by a number offactors (see Table 17.1), most ofwhich must be considered
before finalizing a design.
17.4 Limitations
Because of the influence of construction procedures on the behaviour of deep foundations, inspection should be
considered as an integral part ofthe design ofdeep foundations and should be carried out by the engineer responsible
for the design.
Deep Foundations - Introduction 261 
TABLE 17.1 Pile-Type Classification
Factor  Subgroup 
Concrete; steel; wood. 3. Material
1. Installation Driven; bored; cast in-situ; excavated; augered.
2. Displacement Displacement; low-displacement; non- displacement.
4. Function
5. Capacity
6. Shape
Shaft bearing; toe bearing; combination.
High; moderate; low,
Square; round, hexagonal; octagonal; H-section; Tapered.
7. Environment Land; marine; off-shore.
8. Inclination Vertical; battered.
9. Length Long; short.
10. Structure Bridges; buildings; platforms; towers; machinery; etc.
262 Canadian Foundation Engineering Manual
Geotechnical Design of Deep Foundations
18 Geotechnical Design of Deep Foundations
18.1 Introduction
The design method used for a particular deep foundation will depend on the soil in which it lays, whether it is
cohesive (clay) or cohesionless (sand), and whether the pile toe bears on soil or rock. In addition, each pile design
should be based on considerations of both ultimate limit states (load capacity) and serviceability limit states
(expected deformations or settlements). In the sections that follow, consideration is given to the geotechnical axial
capacity (Section 18.2) and settlement (Section 18.3) of piles in soil, the lateral capacity (Section 18.4) and lateral
movement (Section 18.5) of piles in soil, and the geotechnical axial capacity (Section 18.6) and settlement (Section
18.7) of piles bearing on rock. Both single pile behaviour for isolated piles and multiple pile behaviour for pile
groups are examined.
18.2 Geotechnical Axial Resistance of Piles in Soil at Ultimate Limit States
18.2.1 Single Piles· Static Analysis
This section ·considers the geotechnical axial capacity of piles embedded in soil. Piles derive their load-carrying
capacity from both toe and shaft resistance. The relative contribution ofeach to the total capacity ofthe pile depends, j
essentially, on the density and shear strength of the soil and on the characteristics of the pile.
The geotechnical axial capacity of a single pile, R, can be estimated by summing the shear stresses along the shaft,
qs' adding the bearing capacity of the pile toe, qt' and subtracting the pile weight, viz.
1 R I,Cqst.z+A,ql Wp
the pile of circumference, C, and embedded length, L, is subdivided into segments of length, b.z, and the pile
toe has area, AI' and pile weight, Wp.
The factored geotechnical axial resistance at ultimate limit states is taken as the ultimate axial capacity (R) multiplied
by the geotechnical resistance factor (<D) of 0.4 for compression and 0.3 for uplift (see Tables 8.1 and 8.2 in Chapter
8). Cohesionless Soils
For cohesionless soil, the unit shaft friction at any depth z along the pile is given by
q =(J' K tan (5 = R(J'
s v s   v
Geotechnical Design of Deep Foundations 263
and the bearing capacity of the pile toe is
=N (5'
a combined shaft resistance factor
K coefficient of lateral earth pressure
(5 vertical effective stress adjacent to the pile at depth z
8 the angle of friction between the pile and the soil
bearing capacity factor
  5 ~
vertical effective stress at the pile toe
The value of Ks is influenced by the angle of shearing resistance, the method of installation, the compressibility,
degree of overconsolidation and original state of stress in the ground, as well as the material, size and shape of
the pile. It increases with the in-situ density and angle of shearing resistance of the soil and with the amount of
displacement. It is higher for displacement- type piles than for low-displacement-type piles such as H-piles. For
bored piles, Ks is usually assumed equal to the coefficient of earth pressure at rest, Ko. For driven displacement-type
piles, Ks is normally assumed to be twice the value of Ko'
The value of 0 depends on the surface roughness of the pile, which depends on the pile material (steel, concrete.
wood), the mean particle size of the soil, the normal pressure at the pile-soil interface and method of installation. It
ranges from 0.5 to 1.0 ~ .
The combined shaft resistance coefficient ~ generally ranges from 0.20 to 1.5 as indicated in Table 18.1 - see
Fleming et al. (1992) for further discussion.
TABLE 18.1 Range ofj3 Coefficients
Soil Type
Driven Piles
Silt 0.2 - 0.30 0.3 - 0.5
Loose sand 0.2 - 0.4 0.3 - 0.8
Medium sand 0.3 - 0.5 0.6 - 1.0
Dense sand 0.4 - 0.6 0.8 -1.2
Gravel 0.4 - 0.7 0.8 - 1.5
O'Neill and Reese (1999) indicate that ~ decreases as the bored (cast-in-place) pile length increases in sands and
gravels. The values in Table 18.1 could be considered average values for rather long piles.
The toe bearing capacity factor Nt depends on soil composition in terms of grain size distribution, angularity and
mineralogy of the grains, natural soil density, and other factors. Typical ranges of values for Nt are given in Table
TABLE 18.2 Range ofNt Factors
Soil Type
Driven Piles
Silt 10 - 30 20-40
Loose sand 20-30 30 - 80
Medium sand 30-60 50 - 120
Dense sand 50 - 100 100 -120
Gravel 80 - 150 150-300
264 Canadian Foundation Engineering Manual
i. The toe response ofbored piles is certainly softer than for driven piles. However, it may be argued that this
is a serviceability issue and not a capacity issue. The toe capacity is only governed by the geological nature
of the deposit near the pile toe rather than the method of installation. Thus, the Nt value for both cast-in-place
and driven piles should be the same and equal to those given in Table 18.2 for driven piles. In the absence of
test loading, a factor of safety of at least three should be applied to any theoretical computation.
ii. Both q and a' may continue to increase with increasing depth, but at a decreasing rate. For practical
design purposes, it is advisable to adopt limiting values of both qs and a: for long piles in cohesionless soils
(Poulos et al., 2001). Jardine and Chow (1996) and Jardine et aL (1998) provide a method to estimate qs in
cohesionless soils based on the use of the cone penetrometer, in which the cone resistance is used to estimate
radial effective stresses after pile installation and accounts for effects of soil dilation at the pile-soil interface
and pile compressibility. The method also accounts for the effect of pile depth. This method should be used
whenever CPT tests can be conducted.} Tapered Piles
For tapered piles, the skin friction at any station along the pile shaft can be calculated by (Wei & EI Naggar 1998):
q = K,K a! tan 0 (18.4)
s s v
The taper coefficient K, is introduced to capture the taper effect and in the case of cylindrical piles K, =1. The taper
coefficient for cohesionless soil is a function of the internal friction of the soil, pile-soil interface friction angle, 8,
taper angle, e, the pile geometry, settlement level and the effective overburden pressure. The taper coefficient K, is
given by (EI Naggar and Sakr 2000):
K = tan(e +0 ) cot(o ) + 4G tanee) tan(e +0 ) coteS )Sy
1+ 2l;; tan(e) tan(e +0 ) (1 + 2l;; tan(e) tan(e +0 ) )Kpv
G =the shear modulus ofthe sand, t; = In (r/r) in which rm the average pile radius and r
is a radius at which
the shear stress becomes negligible and is taken to be equal to 2.5L (I-v) where v = Poisson's ratio of the
soil, and Sr is the pile settlement at the ultimate load as a ratio of its diameter = 0.1. The effective overburden
pressure has a profound effect on K, as it decreases quickly with an increase in a ' For practical tapered piles
length, K may be taken as two.
f (2) Layered Soils under the Pile Toe
For piles bearing on layered soils, the toe capacity should be estimated with due consideration. Meyerhof (1976)
and Meyerhof and Sastry (1978) considered three cases of layered soil profiles. In the first case, where a weak soil
layer over lies a dense sand layer, the full toe capacity is not developed until the pile penetrates six diameters into
the dense sand (Meyerhof and Sastry 1978). The toe capacity can be assumed to decrease linearly from the value
for the dense sand layer to the value for the weak layer for a penetration distance less than six pile diameters. In
the second case, a weak layer underlies a dense sand layer and the toe capacity would be affected if the pile toe is
less than three times the pile toe diameter above the weak layer (Matsui 1993). The toe capacity can be assumed
to decrease linearly from the value for the dense sand layer to the value for the weak layer for a distance less than
three pile diameters.
In the third case, the dense sand layer is sandwiched between two weak layers and their effects must be considered
together. Cohesive Soils
Design methods for piles in fine-grained soils are in some cases of doubtful reliability. This is particularly so for the
Geotechnical Design  of Deep Foundations  265 
bearing capacity of shaft-bearing piles in clays of medium-to-high shear strength. Because of this, pile test loading
should be carried out where economically justified or, alternatively, an adequate factor of safety should be used.
Piles in cohesive soils and bearing on stiff soils may mobilize substantial toe resistance, which, for large-diameter
bored piles, may represent the usable capacity of the pile. Total Stress versus Effective Stress Approach
Until recently it was the general practice to evaluate the capacity of piles in clay from a total stress approach, i.e.,
on the basis of the undrained shear strength, su' of the clay. Empirical correlations between Su and the toe-and-shaft
resistance on a pile have been developed, but these have not proved reliable, particularly for Su in excess of about
25 kPa. Therefore, analysis in terms of effective stresses is more rational, i.e., the same method as used for piles
in cohesionless soils applies in all details. Burland (1973) provides a detailed discussion on relevant values of
Skempton (1951) and Ladanyi (1963) present discussion and values of Nt' The relationship in Subsection
may be used in design with the following values:
= 0.25 - 0.32 (18.6)

3 -10 (18.7)
For tapered piles, Blanchet et al. (1980) suggest using = 0.5 to 0.6. Shaft Resistance in Clays with Su < 100 kPa
A pile driven in clay with undrained shear strength smaller than 100 kPa derives its capacity almost entirely from
shaft resistance. It is still common practice to determine the ultimate shaft resistance of a single pile using total
stress analysis from the formula:
where a adhesion coefficient ranging from 0.5 to 1.0.
Figure 18.1 shows the adhesion coefficient as a function of the undrained shear strength of the clay. However, the
actual resistance depends significantly on the geometry of the foundation, the installation method and sequence,
the properties of the clay, and time effects. The capacity of piles determined from the above formula should be
confirmed by test loading.
1.2        -up-;:'lit-:::"1
\ Tomlinson 1957
. (concrete piles) • Data group 1
_  u  ..  Dalagroup2
,. ca.Datagroup3
" • Sbafts in uplift
t:I  .I  0  Data group 1
• a Data group 2
a.. a
.1\a a 0 
A •

load tests
o L-__ __ ___ __ ___ 
o  50 100 150 200 250 300
Undrained Shear Strength, Su (kN/m
FIGURE 18.1 Adhesion as a function ofundrained shear strength

266 Canadian Foundation Engineering Manual Shaft Resistance in Clays where Su > 100 kPa
A pile driven in clay with an undrained shear strength in excess of 100 kPa derives its  capacity from both shaft and 
toe resistance. However, the shaft resistance of such a pile cannot be predicted with any degree of reliability because 
little is known of the effect of driving on the resistance and on the final effective contact area between clay and pile. 
For preliminary design, the relationship given in Section can be used.  For final  design purposes,  however, 
it is  suggested that the pile capacity be  determined by test loading. 
Large-diameter bored piles (with or without enlarged belled bases, or under-reamed shafts)  are  successfully used 
in  clays  or  cohesive  soils  where  Su > 100  kPa.  Present design  methods have been  derived  from  extensive  studies 
on  bored piles  in  London clays.  Considering  the  special properties of these  soils,  the  generalization  of empirical 
design parameters to  other types  of soils  should be made  with caution.  Bored piles  are  also  used in  argillaceous 
intermediate geomaterials (cohesive earth materials), such as hard clays and clay-based rock (e.g., Queenston shale 
formations).  Hassan  et  al.  (1997)  provide  a  method  to  estimate  q
and  q
that  accounts  for  the  pile-geomaterial 
interface roughness and the initial effective stress at the interface. For bored piles in porous sandstone, the methods 
provided by Seidel (1993) and McVay et al.  (1992) are more suitable. Toe Resistance
The ultimate toe resistance may be estimated from: 
R[ toe resistance 
AI cross-sectional area of pile at toe 
su minimum undrained shear strength of the clay at pile toe 
N, a bearing capacity coefficient that is a  function of the pile diameter, as follows: 
Pile toe diameter  N,
smaller than O.Sm  9 
0.5  m to  I  m  7 
larger than  1  m  6 
In very  stiff clays  and  tills,  where  samples  are difficult  to  retrieve  and the  undrained shear strength  is  not  easily 
measured, a pressuremeter may be used to  evaluate the strength of the soil. Stratified Deposits
The relative contribution ofthe various strata penetrated by a pile to the capacity ofthat pile is primarily a function 
of the  relative  stiffness  of these  layers  andof the  type  of pile.  Static  analysis  for  totafaxial  capacity essentially 
involves calculating contributions ofvarious unit shaft resistance values, qs' associated with the different strata that 
the pile penetrates and the end-bearing associated with the stratum containing the pile toe. 
Furthermore,  it is  important to  install the top  of the  pile  a  distance of at least four  diameters  into  any  stiffer clay 
stratum so that the full  value N, 9 can be used, and to  watch  for  the presence of a  weaker stratum below the toe 
which could reduce the toe resistance. Helical (Screw) Piles
The basic form of a helical pile or anchor for construction applications consists of a helically shaped bearing plate 
or multiple plates attached to a central shaft. Historically, helical piles or anchors have been used in relatively light 
load  applications, with shaft diameters  and  helix diameters typically  less  than  100 mm and 400 mm respectively. 
Geotechnical Design of Deep Foundations 267
Recentlyhowever,throughthedevelopmentofhigh-capacitytorquedrives(inexcessof50,000ft-Ibs) thatareused
forhelicalpileinstallation,largerdiametershaftsandhelixeshavebeenconstructedand installed.
Wheninstalled to properdepth andtorque, thehelicalplatesactas individualbearingelementsto supporta load.
Thehelicalpileis thereforeadeep, end-bearingfoundationthatcanbeusedtoresistbothcompressiveandtension
loads. Installationofhelicalpiles is accomplishedbyhydraulictorquedrivesthatcanbemountedtojustaboutany
typeofmachine (e.g. bed-mounteddrillrigs,rubber-tiredbackhoes,skid-steerloaders,mini-excavators,andtrack-
Thetotalcapacityof thehelicalpileoranchorequalsthebearingcapacityof thesoilappliedtotheindividualhelical
plate(s)and,in someinstances,theskinfrictionoftheshaft.Thisis:
Theevaluationofthesecomponents is describedfurtherbelow.
bythe geotechnicalresistance factor (<l» ofOA forcompressionand0.3 for uplift(Tables 8.1 and 8.2 inChapter
8). Helical Plate(s) Bearing Capacity
Thetotalcapacityofanend-bearinghelicalpileisevaluatedas thesumofthecapacitiesofeachindividualhelical
it totheindividualhelicalplate(s)areas,i.e,

Ah(suNc + yDhNq+ O.5yBN
) (18.11)
Individualhelixbearingcapacity Q
Su undrainedshearstrengthofthesoil
y Unitweightofthe soil
Dh Depthtohelicalbearingplate
B diameterofthehelicalplate
N ,NandN Bearingcapacityfactors forlocalshearconditions
c q y
Thetotalhelicalplatescapacity,Qt' canbe expressedas:
The bearing capacity equationis applicable only when the helical bearingplates are.spacedfar enough apart, at
leastthree times the diameterofthe helix, to avoid overlapping oftheir stress zones. In cases involving
overlapping stress zones, the multi-helix capacity can be determined by computing the bearing capacity ofthe
bottomplate,andthecylindricalshearcapacitydevelopedbetweenthe upperandlowerplate(s).Theformulation
providedbelowwithrevisionofpileshaftdiametertoeffectivehelixdiametermaybeconsidered. Capacity Due to Skin Friction
Theskin friction along thepileshafttypicallyisnotconsideredinthetotalcapacityunlesstheshaftis at least 100
mmindiameter(orequivalentdiameter).Thecapacityduetoskinfriction canbecalculatedas follows:
2GS Canadian Foundation Engineering Manual
Qf Frictional resistance of pile 
D  Diameter of pile shaft 
fs  Sum of friction and adhesion between soil and pile 
LlL Incremental pile length over which ltD and fs  are taken as constant 
f Relationship of Load Capacity to Installation Torque
An  estimate  of the  helical  pile  ultimate  capacity  may  be  achieved  through  monitoring  of  installation  torque. 
Recording  of installation torque  can also  serve as  a  quality  control  step,  identifying piles  that did not achieve the 
expected installation torque and may require load testing.  The relationship between the load capacity and installation 
torque, which was developed based on pullout tests on helical piles, can be described using the following empirical 
KyxT  (18.14) 
R Ultimate capacity of screw pile 
Empirical torque factor 
T Average installation torque 
The value of Kymay range from 31ft to 201ft ifT is recorded in ft-lbs,  or  101m to 331m ifT is recorded in N-m. The 
selection ofKyis dependent upon the soil conditions and anchor design including plate and shaft diameter. For small 
sized square shaft anchors (less than 90 mm diameter), the Ky value has been found to range from 101ft to  121ft with 
101ft  (331m) being  the  recommended default value.  For pipe shaft anchors  (90 mm O.D. pipe), the recommended 
default value is  71ft (231m), with this value decreasing to 31ft (lO/m) for shaft diameters approaching 200 mm. 
Torque monitoring tools are available to provide a  suitable method of production control  during installation. As a 
quality assurance measure, it is recommended that the engineer specifies a required torque during construction. 
Installation torque is primarily a function of the frictional  resistance along the shaft, the frictional  resistance along 
the top  and bottom surfaces  of the  helical  plate(s),  and  the passive resistance  along of the  leading edge(s)  of the 
plate(s).  Although soft zones at depth may not influence the recorded torques, they may adversely impact the load 
carrying capacity ofthe helical pile. As a result, a good understanding ofthe ground conditions around pile(s), within 
and extending beyond the zone that is  expected to be stressed as a result ofloads on the pile(s) is  important. 
18.2.2 Pile Groups. Static Analysis
It is common practice to define the axial capacity ofa pile group relative to the sum of the capacities ofthe individual 
piles  in the group.  Group  'efficiency' is defined as  the ratio  of the group capacity to  this  sum of the individual pile 
capacities. Cohesion less Soils
Driven piles  in cohesionless  soils  develop larger individual  capacities when installed as  a  group  (group  efficiency 
> 1)  since  lateral  earth  pressure  and  sand  density  increase  with  the  driving  of additional  piles.  Therefore,  it is 
conservative to use the sum of the individual pile capacities as  an  estimate of the pile group capacity. 
For bored pile  groups,  the  individual  pile  capacity is  reduced  by  the  addition  of the  extra piles,  since  the boring 
process reduces sand density and lateral earth pressures (efficiency is  < 1). 
For bored pile  groups,  a reduction factor  (Meyerhof (1976)  suggests  0.67  for piles in clean sand) may need to be 
Geotechnical Design of Deep Foundations 269
applied to the sum of individual pile capacities. However, for piles in sands with some fines (e.g., silty sands) and
if the cap is firmly in contact with the soil and spacing of the piles is less than 4.5d (where d = pile diameter) the
group efficiency is 1.0. Cohesive Soils
In addition to the possibility that individual piles in a pile group act independently to support the applied load, a
closely spaced pile group, can act as a 'block' whereby the soil between adjacent piles is dragged down between
them, shaft resistance develops around the perimeter of the group only, and end-resistance develops under the whole
ofthe pile-soil block. A rational approach to estimating pile group capacity is to use the minimum of a) the sum of
the individual pile capacities and b) the capacity ofthe pile- soil block analysed as an equivalent single pile. For this
block capacity calculation, an average unit shaft resistance, qs' must be calculated since for zones on the perimeter
where there is soil-soil contact, q = s and for zones where there is soil-pile contact, q = as. The block perimeter
sus u
is the circumference, C, of the equivalent pile, and the area of the block base is taken as the base area, At' of the
equivalent pile.
18.2.3 Single Piles - Penetrometer Methods Limitations
Field test data are often available in the form of static or dynamic penetration resistance. Clearly, it is appealing
to generate predictions of axial capacity directly from penetration resistance, rather than from more fundamental
soil shear strength parameters. Caution must be exercised however, given that this attempted simplification may
disregard the complexity of both the penetration tests themselves and the axial pile response. Cone Penetration Test
The axial capacity of deep foundations in soils can be computed from the results of a static cone penetration test
(CPT). The test is suitable for a large range ofsoils provided adequate pushing force is available for sufficient depth
qf penetration.
The ultimate geotechnical axial capacity of a single pile can be estimated using the basic equation given in Section
18.2.1 and estimating the unit base resistance, qt' and unit shaft resistance, qs' from
qc cone penetration resistance (units of stress) from CPT
qca = equivalent cone penetration resistance at pile base according to Figure 18.2
kc bearing capacity factor based on soil type and pile type (Table 18.3)
(J. friction coefficient (Table 18.4)
• •
270  Canadian  Foundation Engineering  Manual 
FIGURE 18.2 CPT method to determine equivalent cone resistance at pile base
(after Bustamante and Gianeselli, 1982)
This approach is based on extensive full scale pile load test data from France (Bustamante &  Gianeselli, 1982) and
supported by pile load test data in North America (Robertson et al., 1988; Briaud &  Tucker, 1988). The scaling
effect to account for the difference in size between the cone penetrometer and the pile and the method of installation
is accounted for in the selection of k and a. using Tables 18.3 and 18.4.
The method developed by Lehane and Jardine (1994) and Jardine and Chow (1996) is especially useful in estimating
qs for piles driven in cohesionless soils using the cone penetrometer measurement. The method accounts for effects
of soil dii"ation at the pile-soil interface and pile depth and compressibility. It should be used whenever CPT tests
are conducted. The ultimate axial capacity for design is influenced by the number of CPTs performed, the observed
variability of the test results and the local experience available. Caution should be exercised when designing piles
in sensitive clays.
TABLE 18.3 Bearing Capacity Factors, k

Soil Type
Soft clay and mud <1
Moderately compact clay 1 to 5
Silt and loose sand $5
Compact to stiff clay and compact silt >5
Soft chalk $5
Moderately compact sand and gravel 5 to 12
Weathered to fragmented chalk >5
Compact to very compact sand and gravel > 12
* Note:
Factors kc
0.4 0.5
0.35 0.45
0.4 0.5
0.45 0.55
0.2 0.3
0.4 • 0.5
0.2 0.4
I 0.4
i. Group 1:  Plain bored piles, mud bored piles, micro piles (grouted under low pressure), cased bored piles, hollow auger
bored piles, piers and barrettes.
ii. Group II: Cast-in-place screwed piles, driven precast piles, prestressed tubular piles, driven cast piles, jacked metal
piles, micropiles (grouted under high pressure with diameters < 250 mm).
Geotechnical Design of Deep Foundations 271
The factored geotechnical axial resistance at ultimate limit states is taken as the predicted ultimate capacity multiplied
by the geotechnical resistance function (<1» of 0.4 for compression and 0.3 for uplift (Table 8.1 in Chapter 8).
TABLE 18.4 Friction Coefficient, a.
0.Q35 0.015 < 1 30 90 90 30 0.015 0.015 0.015 Soft clay and mud
0.035 0.035 0.035
0.08 ~ 0.12 1-5 40 80 40 80
Moderately compact clay
(0.08) (0.08) (0.08)
Coefficient a Maximum Limit of qc (MPa)
::;5 60 150 60 120 0.035 0.035 0.035 0.035 0.08
Compact to stiff clay and
Silt and loose sand
0.035 0.035 0.035
~ 0   2 0 >5 60 120 60 120 0.035 0.08
(0.08) (0.08) (0.08) compact silt
::;5 100 120 100 120 0.035 0.035 0.035 0.035 0.08
Moderately compact sand and
Soft chalk
0.08 0.035 0.08
~ 0   2 5-12 100 200 100 200 0.08 0.12
(0.12) (0.08) (0.12) gravel
0.12 0.08 0.12
>5 60 80 60 80 0.12 0.15 ~ 0.2 Weathered to fragmented chalk
(0.15 (0.12) (0.15)
Compact to very compact sand 0.12 0.08 0.12
~ 0.2 > 12 150 300 150 200 0.12 0.15
(0.15) (0.12) (0.15) and gravel'
Note: Bracketed values of maximum limit unit skin friction, qs' apply to careful execution and minimum disturbance
of soil due to construction.
* Category:
Plain bored piles, mud bored piles, hollow auger bored piles, micropiles (grouted under low pressure), cast-in-
place screwed piles, piers and barrettes.
IB Cased bored piles, driven piles.
IIA Driven precast piles, prestressed tubular piles, jacked concrete piles.
IIB Driven metal piles and jacked metal piles.
Driven grouted piles and driven rammed piles.
High pressure grouted piles with diameters> 250 mm and micropiles grouted under high pressure. Standard Penetration Test
The ultimate geotechnical axial capacity of a single pile in granular soils can be estimated from the results of the
Standard Penetration Test (SPT) as suggested by Meyerhof (1976).
272 Canadian Foundation Engineering Manual
R =mNA
R pile capacity 
M an empirical coefficient equal to  400 for driven piles and to  120 for bored piles 
N SPT index at  the pile toe 
A = pile toe  area 
n an empirical  coefficient equal to two for driven piles and to one for bored piles 
average SPT index along the pile 
A pile  embedded shaft area 
Decourt (1995) developed a  more  comprehensive correlation of the  shaft and  toe resistance  of piles with  the  SPT 
value.  He suggested the following expressions: 
qs =(J. (2.SN
+ 10)  (kPa) (lS.17a) 
ql =  b ~ (kPa)
(J. 1 for displacement piles in any soil and non-displacement piles in clays, and 0.5  to  0.6 for non-
displacement piles in  granular soils. 
!!..t;o = average SPT index (normalized to  60 % energy efficiency)  along the pile shaft 
Nb = average of SPT index in the vicinity of the pile toe 
Kb =  is a base factor given in Table  IS.5. 
TABLE  18.5  Base Factor, Kb (Decourt, 1995)
Soil Type 
Sandy silt 
Clayey silt 
Non-Displacement Piles 
The  Standard Penetration Test  has  significant limitations  (see  Chapter 4),  and care must be exercised when using 
the test results.  For this  reason  and when using working stress design, a  minimum factor  of safety of four should 
be  applied  to  the  calculated  capacity unless  local  experience  indicates  otherwise.  For factored  geotechnical axial 
resistance  of ultimate  limit  states,  it is  suggested that the  ultimate  axial  capacity  be multiplied by a  geotechnical 
resistance factor of 0.3. 
18.2.4 Single Piles· Dynamic Methods Introduction 
The  objective  when  dealing  with  the  dynamic  methods  of pile  design  is  to  relate  the  dynamic  pile  behaviour 
(acceleration  or driving resistance)  to  the  ultimate  static  pile  resistance.  Care  should  be  taken  when  using  these 
methods,  since  they may ignore  the  effects  of 'set up'  in  soft clays  (dynamic  methods  usually provide  estimates 
of pile  capacity just after  driving),  downdrag  (see  next  section)  and  serviceability issues  (whether  expected  pile 
settlement is acceptable). 

Geotechnical Design of Deep Foundations 273
Axial Capacity Based on Dynamic Monitoring
The capacity of a single pile can be estimated by means of dynamic measurements. The reliability of this  estimate 
of the  capacity,  under  favourable  conditions,  can  be  almost  equal  to  that  of a  routine  static  loading  test.  The 
measurements  and  the  evaluation  of the  data  must  be  carried  out  by  a  person  competent  in  this  field.  For  more 
details,  see  Chapters  19  and 20. . Geotechnical Axial Capacity Based on Wave-Equation Analysis
The  wave-equation  analysis  (which  is  discussed  in  Chapter  19)  is  a  tool  for  determining  pile  bearing  capacity, 
pile  driveability,  and  for  hammer  selection.  The  wave  equation  requires  accurate  input  of several  hammer  and 
soil parameters that can vary widely from  case to  case.  Hammer-rated energy  can  differ  substantially from  actual 
measurements,  and  the  soil  parameters  are  'model-dependent'  empirical  values  and  not  rational  properties  that 
can be measured independently.  Unless there is  calibration to  field  measurements, the analysis can only be used to 
provide general guidance. Dynamic Formulae
The assumptions made in the dynamic formulae are oversimplified, and the results cannot always be related to actual 
pile capacity.  One  reason is  that the  dynamic  formulae  input hammer-rated energy  and  not the  actually  delivered 
energy  and  this  results  in  considerable  error.  Nevertheless,  when used by competent  persons  and  related to  local 
experience, a dynamic formula can still serve as a guide to  engineering judgement. However, dynamic formulae are 
best replaced by other techniques. 
18.2.5 Negative Friction and Downdrag on Piles
When piles have been installed in or through a clay deposit that is  subject to consolidation, the resulting downward 
movement of the clay around the piles, as  well as  in any soil above the clay layers, induces downdrag forces  on the 
piles through negative skin friction.  The magnitude  of settlement needed to  cause the negative skin friction  is  very 
small.  For instance,  observations by Fellenius and Broms (1969)  and Fellenius (1972)  of negative skin friction  on 
piles in a 40 m thick clay layer indicate that the relative movement required can be smaller than a millimetre.  Such 
small  relative  movements  occur  easily  as  a  result  of the  large  stiffness  difference  between the  pile  and  the  soil. 
Therefore, with time, small movements or strains will occur in any portion of a pile and positive resistance along a 
lower portion of a pile are the norm rather than the  exception. 
The simplest method  for  computing the  negative  skin friction  is  to  assume that it is  proportional to  the  undrained 
shear strength of the soil (Terzaghi & Peck,  1967). 
= as
qn unit negative skin friction 
a  a reduction coefficient ranging from  0.5  to  1.0 
Su the undrained shear strength after the soil has  consolidated under the new load and therefore should 
be estimated from  CU tests representative ofthe expected overburden pressure. 
Field  observations  on  instrumented piles  have  shown that  the  negative  skin  friction  is  a  function  of the  effective 
stress  acting  on  the  pile  and  may  be  computed  in  the  same  way  as  the  positive  shaft  resistance,  as  detailed  in 
Subsection In most clays and silts, the magnitude ofthe negative  skin (shaft)  friction approximates to  a   
factor of about 0.2 to  0.3. 
The total drag 10ad,Qn' for  a single pile is: 
274 Canadian Foundation Engineering Manual
C shaft circumference or perimeter length
D = length of pile embedded in settling soil.
Alternatively, elastic methods can be used. These methods suggest how downdrag relates to settlement (for example
Poulos & Davis, 1972), and provide a means for estimating the maximum downdrag force and its development with
time. Various theoretical solutions are available for single piles (Poulos & Davis, 1980). Design Considering Downdrag
The design must consider the structural axial capacity, the settlement and the geotechnical axial capacity of the
pile. The downdrag increases the structural loads in the pile and thus has to be accounted for when evaluating the
structural ultimate limit state of the pile. The downdrag also increases the pile settlement and therefore should be
accounted for when evaluating the serviceability limit state of the pile. However, the downdrag has no effect on the
geotechnical axial capacity of the pile. It is important to realize that drag load and transient live load do not combine,
and that two separate loading cases must be considered: permanent load plus drag load, but no transient live load;
and permanent load and transient live load, but no drag load. Furthermore, a rigid, strong pile will have a large
drag load, but small settlement, whereas a less rigid and less strong pile will have a smaller drag load, but larger
settlement. Also, no pile subjected to down drag condition will settle more than the ground surface nearest the pile.
As a first step in the design of the pile, the neutral plane must be determined. The neutral plane is located where the
negative skin friction changes over to positive shaft resistance. It is determined by the requirement that the sum of
the applied dead load plus the drag load is in equilibrium with the sum of the positive shaft resistance and the toe
resistance of the pile. The location of the neutral plane governs both the maximum load in the pile and the settlement
of the pile. Neutral Plane
The neutral plane is found as the intersection of two curves. First, as illustrated in Figure 18.3, a load distribution
curve is drawn from the pile head and· down with the load value starting with the applied dead load and increasing
with the load due to negative skin friction acting along the entire length ofthe pile. Second, a resistance distribution
curve is drawn from the pile toe and up, starting with the value of ultimate toe resistance and increasing with the
positive shaft resistance.
PILE HEAO .---r'--------.r-'i>
PILE TOE - - ~ i £ - - - - ~ - - - - I - - l   - -
aJ bJ
FIGURE 18.3 Calculation ofthe location ofthe neutral plane and the settlement ofa pile
or a pile group (after Fellenius, 1984a)
Geotechnical Design of Deep Foundations 275
The determination ofthe load distribution in a pile is subject to uncertainty. Reliable information on the soil strength
is required when determining the load distribution. It is recommended that the theoretical analysis adopting the
method in Section be used. The analysis should be supplemented with information from penetrometer tests,
such as the SPT and the static cone penetrometer. For driven piles, the analysis should be combined with results
from the analysis of dynamic monitoring data. (2) Structural Axial Capacity
The structural axial capacity of the pile is governed by its structural strength at the neutral plane when subjected
to the permanent load plus the drag load; transient live load is not to be included. At or below the pile cap, the
structural strength of the embedded pile is determined as a short column subjected to the permanent load plus the
transient live load, but drag load is to be excluded.
At the neutral plane, the pile is confined, and the maximum combined load may be determined by applying a
safety factor of l.5 to the pile material strength (steel yield ancl/or concrete 28-day strength and long-term crushing
strength of wood).
Ifthe negative skin friction and the positive shaft resistance as well as the toe resistance values are determined, assuming
soil-strength values 'err' on the strong side, the calculated maximum load on the pile will be conservative. Settlement
As illustrated in Figure 18.3b, the settlement of the pile head is found by drawing a horizontal line from the neutral
plane, as determined according to the foregoing method, to intersect with the curve representing the settlement
distribution in the soil surrounding the pile. The settlement of the pile head is equal to the settlement of the soil at
the elevation of the neutral plane plus the elastic compression of the pile due to the applied dead load and the drag
load (FeUenius, 1984a).
One condition for the analysis is that the movement at the pile toe must be equal to or exceed the movement required
to mobilize the ultimate toe resistance ofthe pile. In most soils, this required movement is equal to about 1 % to 2 %
of the pile toe diameter of driven piles and about 5 % to 10 % of the toe diameter for bored piles. Ifthe movement
is smaller than this required magnitude, the neutral plane will move higher up in the settlement diagram and the
settlement will increase correspondingly. If this occurs, the magnitude of the settlement will normally be negligible
and correspond to the elastic compression of the pile.
The settlement calculation should be carried out according to conventional methods (see Chapter 11) for the
effective stress increase caused by dead load on the pile(s), surcharge, groundwater lowering, ancl/or any other
aspect influencing the stress in the soil. The dead load applied to the pile cap should be assumed to act at an
equivalent footing located at the level ofthe neutral plane and the load distributed from this plane. The settlement of
the pile cap is the sum of the settlement ofthe equivalent footing and the compression ofthe piles above the neutral
plane. Note that Figure 18.3 does not show the settlement due to the dead load acting on the equivalent footing at
the neutral plane.
The accuracy ofthe calculation of the distribution of settlement depends on the reliability of the input data, which in
turn depends on the completeness of the site investigation program. It is imperative that representative samples be
obtained from all soil layers, including those below the pile toe, and that the strength and compressibility properties
of the soil be determined in the laboratory. In-situ testing methods such as vane tests and static cone-penetrometer
tests will enhance the laboratory testing.
For the case in which the structure is built before the pore pressures induced by the pile installation have dissipated,
it is necessary to estimate the additional settlement caused by the pore pressure dissipation.
276 Canadian Foundation Engineering Manual Geotechnical Axial Capacity
The last part ofthe design is to check the safety against plunging failure of the pile. In this case, the pile moves dOwn
along its entire length and the downdrag is eliminated. Therefore, the load is the combination of the dead load and
the live load, no drag load, and the case is similar to that of designing the allowable load of a pile not in a downdrag
As stated by Fellenius (l984a), when the capacity has been determined using the static loading test or the dynamic
testing method, a factor of safety of2.0 or larger ensures that the neutral plane is located below the mid-point of the
pile. When the capacity is calculated from soil-strength values, the factor of safety should not be smaller than 3.0. Special Considerations
Downdrag on piles caused by negative skin friction is most often a settlement problem and rarely a capacity problem.
According to the method recommended in this Section, the service load should not be reduced by any portion of the
drag load unless required by insufficient structural strength ofthe pile at the location ofthe neutral plane, or in order
to lower the location of the neutral plane (reducing settlement).
When settlement occurs around a pile or a pile group, piles that are inclined will be forced to bend by the settling
soil. For this reason, it is advisable to avoid inclined piles in the foundation, or, at least, to limit the inclination of
the piles to values that can follow the settlement without excessive bending being induced in the piles. Furthermore,
piles that are bent, doglegged, or damaged during installation will have a reduced ability to support the service load
in a down drag condition. Therefore, a design carried out according to this section postulates that the pile installation
will be subjected to stringent quality control to ensure that the installation is sound. (6) Downdrag in Groups of Vertical Piles
Briaud and Tucker (1997) examined downdrag effects on groups of vertical piles. They indicated that downdrag
effects may be approximated by considering the downdrag stresses on the perimeter of the group, unless the piles
are very widely spaced. Means for Reducing Downdrag
When the pile settlement is excessive and the solutions, such as those of increasing the pile length or decreasing the
pile diameter, are not practical or economical, the downdrag acting on the piles can be reduced by the application
ofbituminous or other viscous coatings to the pile surfaces before installation (Fellenius, 1975a, 1979). For cast-in-
place piles, floating sleeves have been used successfully. Briaud and Tucker (1997) provide some useful provisions
for reducing downdrag forces in piles.
18.2.6 Uplift Resistance
Pile foundations must sometimes resist uplift forces and should be checked both for resistance to pullout and their
structural ability to carry tensile stresses. The ultimate uplift resistance of a pile is equal to the shaft resistance that
can be mobilized along the surface area of the shaft. For bored piles in clay soils, the uplift resistance is commonly
assumed to be equal to that contributing to the bearing capacity of the pile as described in Section 18.2.1 (O'Neill
& Reese, 1999).
For either bored or driven piles in cohesionless soils, qs in the uplift is about 75 % to 80 % ofits value in compression
(EI Naggar & Sakr, 2001; O'Neill, 2001). However, for piles with high residual stresses as a result of the driving
the actual shaft resistance in uplift (pile in tension) is considerably smaller (about half) compared to the apparent
resistance in compression. In such cases, the applied factors of safety should be double those applied in the case
of compression. The uplift resistance of tapered piles in cohesionless soils is comparable to the uplift resistance of
cylindrical piles with the same average embedded pile diameter (EI Naggar & Wei, 2000).
Geotechnical Design of Deep Foundations 277
When piles  are  built primarily to resist uplift  forces,  the  pullout  resistance  can  be  increased by providing  one  or 
. more sections whose diameter is larger than the average pile diameter. Expanded base piles, underreamed and multi-
underreamed piles, and screw- piles are typical. 
The most reliable way of designing piles subjected to uplift loads  is  by means of uplift testing. The tests  should be 
designed and carried out in accordance with ASTM designation D3689. 
The uplift resistance of a pile  group  is  the  lesser of the two following  values: 
the sum of the uplift resistance of the  piles in the group; and 
•  the sum of the shear resistance mobilized on the surface perimeter of the group plus the effective weight of 
soil and piles enclosed in this perimeter. 
18.2.7 Other Considerations Axial Capacity Based on Test Loading
The design of piles based on theoretical or empirical methods, as  described above,  is subjected to some uncertainty 
•  soil properties that cannot be measured with great accuracy and are variable within a building site; 
the  correlation between the  soil parameters  and  the  bearing  capacity of a  pile  includes  a  margin of error; 
the  actual  driving  or  installation  conditions  vary  from  pile  to  pile  and  cannot  be  properly  taken  into 
Therefore, the best method ofassessing the bearing capacity ofpiles is to test-load typical units. General considerations 
on the use ofload tests,  the recommended methods oftesting, and interpreting the  test results  are  given in Chapter 
20. Compacted Concrete (Expanded-Base) Piles
Compacted concrete piles in granular soils derive their bearing capacity from the densification of the soil around the 
base due to the  installation process.  The bearing capacity of such piles is,  therefore, dependent on the  construction 
method,  and  the  capacity  value  used  should  be  supported  by  documented  local  experience  and/or  static  test 
loading. Piles Installed by Vibration
Piles  may  be  installed  in  soils  with  little  cohesion  using  a  vibratory  device  attached  to  the  top  of the  pile.  This 
method has  two  advantages  over conventional driving:  it  is  relatively quiet and produces  less  excessive vibration 
levels.  Installing piles  by vibration is  facilitated  by  weakening the soil  strength along  the  pile shaft (likely due  to 
liquefaction) and no densification effect is realized due to  the  installation. 
The capacity of piles installed using vibration can be established using static analysis  and using the provisions  for 
bored piles. The capacity of these piles  cannot be estimated from driving records and thus, their capacity has to  be 
verified by dynamic analysis  of restrike blows to all or a specified percentage of the piles. Augured Cast-In-Place-Piles
The augured cast-in-place (ACIP) or continuous flight auger (CFA) pile system was developed in the USA in the late 
1940s.  Today, the method is in wide use throughout the world, including Canada. ACIP piles must be installed by an 
experienced contractor who is familiar with the augercast process and local geology and soil conditions.  ACIP piles 
278  Canadian Foundation Engineering Manual 
can be designed as bored piles. At least one pile  load test should be conducted to  confirm the pile capacity. Soil Set-Up and Relaxation
In  some  soils,  the  capacity  of driven  piles  is  subject  to  change  with  time  during  or  following  driving.  In  dense, 
saturated,  fine-grained  soils,  such  as  non-cohesive  silts  and fine  sands,  the  ultimate  capacity  may  decrease  after 
initial  driving.  This  is  known as relaxation.  In  this case, the driving  process is believed to  cause the  soil to  dilate, 
thereby  generating negative  pore  pressures  and  a temporary  higher strength.  When these  pore pressures return to 
normal, the resistance reduces. 
On the  other hand, temporary liquefaction, which causes a reduced resistance to pile penetration, may also  occur in 
saturated fine  sands or silts. The probability of liquefaction is greater in loose sands, but liquefaction can occur even 
in dense material,  if there is a sufficient number of stress cycles, ifthe magnitude of the stress is  large enough, or if 
the confining pressure  is  low.  After the  temporary pore pressures dissipate,  long-term capacity is  indicated by the 
return to a higher resistance  to pile penetration. 
Because  the  resistance  to  pile  penetration may  increase  (due  to  soil  set-up),  or  decrease  (due  to relaxation),  it  is 
essential that re-striking be carried out once equilibrium conditions in the soil have been re-established. The need for 
re-striking should be recognized in the contract specifications. The time for the return of equilibrium conditions can 
be determined by trial and error or from pore pressure dissipation tests performed during a pause in the penetration of 
a cone penetration test where pore pressures are measured (piezo-cone test) (Robertson et al.,  1990). The resistance 
developed in the first five  blows of re-striking is generally indicative of the equilibrium resistance. 
However,  conclusions on soil set-up from re-striking without simultaneous measurement of developed energy and 
stresses are highly unreliable, and test loading may be required to  appraise the final capacity. The effects of soil set-
up should be treated with great caution in large pile groups. Also, soil set-up cannot be quantified by re-striking piles 
that have been driven to  a penetration resistance greater than about 2 mm to 3 mmlblow in initial driving. 
Piles driven into cohesive soils induce some disturbance, which is a function of: 
•  the soil properties, in particular its sensitivity to remoulding; 
the geometry of the pile foundation (diameter ofpiles , number, and spacing of piles in the groups);  and 
•  the driving method and sequence. 
The  disturbance  results  in  a  temporary  loss  of strength  in  some  soils  and  a  corresponding  reduction  of support 
provided by the piles (see Fellenius &  Samson, 1976; Bozozuk et al.,  1978a; Clark &  Meyerhof,  1972aJb). In some 
cases,  such  as  in  soft  sensitive  clays,  complete  remoulding  of the  clay may  occur.  The  effect  of the  remoulding 
diminishes with time following driving, as the soil adjacent to the pile consolidates. This results in an increase in the 
capacity of the pile occurring at a slower rate around a concrete or steel pile as  opposed to a wooden pile. 
Test  loading  of a  pile  in fine-grained  soil  should  not  be  carried  out without knowledge  of these  processes.  It is 
advisable to delay testing for at least two weeks after driving. Porewater Pressures Induced by Driving
Pile driving in clay generates high porewater pressures, the effects of which are to: 
temporarily reduce the bearing capacity ofthe piles (and of adjacent piles); 
•  affect the  process of reconsolidation  of the  clay  around the  pile, thereby making  it necessary to  delay the 
application of the load.  Delays of 30  days and more are not unusual (Blanchet et al.,  1980); 
•  drastically  alter  the  natural  stability  conditions  in  sloping  ground.  (There  have  been  a  few  examples  of 
major landslides triggered by pile-driving operations.) 
Geotechnical Design of Deep Foundations 279
If necessary, stability can be monitored with instrumentation of the clay layer for measurement of porewater
pressures and soil displacements during driving. Alternatively, porewater pressures can be reduced by the use of
proper driving techniques and sequences (preboring is an efficient way to reduce porewater pressures and soil
displacements); and the use of vertical pre-manufactured drains attached to the surface of the piles, or preferably,
installed at the site prior to the pile driving (see Holtz & Bowman, 1974). Heave Due to Pile Driving
When piles are driven in clays, the volume of soil displaced by the pile generally causes a heave of the soil surface.
The heave of adjacent piles may also occur, and could result in a reduction in the capacity of these piles. This
problem is of particular significance when large pile groups are driven. Construction Effects for Bored Pier
In deep large-diameter excavations for cast-in-place piles, or when the concreting is delayed, significant strength
reductions may occur as a result of heave and lateral flow within the excavation. Also, poor slurry construction
techniques that leave a thick layer of slurry between the pile and surrounding soil can have a detrimental effect on
shaft capacity. These factors should be considered during the design. Penetration Resistance
The penetration per blow (the set) decreases rapidly after a resistance of 5 mmJblow for shaft-bearing piles and
3 mmJblow for toe-bearing piles. There is little justification in requiring sets smaller than 3 mmlblow for a end-
bearing pile that may only be warranted if driving is easy in the soil above the bearing stratum, or under special
18.3 Settlement of Piles in Soil
18.3.1 Settlement of Single Piles
Many factors influence the settlement of single piles, so it is difficult to make precise estimates of settlement of
single piles or pile groups. In general, the shaft resistance is mobilized with very little movement, typically 5 mm
to 10 mm, whereas the toe resistance when embedded in soil requires longer movements typically between 5 % and
10 % of the pile diameter. Hence, the actual load-settlement response of a single pile is a function of the relative
contributions of shaft and toe resistance, the ground conditions and the method of pile installation. However, a
number of empirical and theoretical solutions have been developed that can be used to make reasonable estimates
of pile response. Empirical Method
For normal load levels, the settlement of a pile may be estimated from the empirical formula (Vesic, 1970, 1977):
in which
s =s +s (l8.21a)
S ss sl
elastic deformation of pile shaft
settlement of ground in which the pile is embeded
settlement of pile caused by load transmitted along the pile shaft
settlement of pile toe caused by load transmitted at the toe
280 Canadian Foundation Engineering Manual
The elastic deformation of the pile shaft is  given by: 
actual load transmitted to  the pile toe  (due to applied load) 
actual  shaft load (due to applied load)  .. 
depends  on  distribution  of skin  friction  = 0.5  for  uniform  or parabolic  distribution and  0.67  for 
linear distribution 
total length of the pile 
average cross-section area of the pile 
modulus of elasticity of the pile material 
Alternatively, the pile shaft compression can be approximated by: 
QL = applied pile load. 
The settlement components due to soil deformation are given by; 
C  empirical coefficient (typical values given in Table  18.6) 

d pile diameter 
C = 0.93 + 0.16 (L/d)o.5 (18.25)

TABLE 18.6 Typical Values o/Coefficient C( (Vesic, 1977)
Soil Type Driven Plies Bored Piles
Sand (dense to  loose)  0.02-0.09  0.09-0.18 
Clay (stiff to  soft)  0.02-0.03  0.03-0.06 
Silt (dense to loose)  0.03-0.05  0.09-0.12 Elastic Continuum Solutions
Poulos and Davis (1980) provide a comprehensive set ofresults for both floating and end-bearing piles. For example, 
the  settlement of a pile in a deep layer of uniform elastic material is  expressed as: 
Geotechnical Design of Deep Foundations 281
E soil modulus 
settlement influence factor,  Figure  18.4 
compressibility correction factor,  Figure  18.5 
Poisson's ration  correction factor,  Figure  18.6 
The factors  10' RK and Rv are  obtained  using  analysis  based on Mindlin's  solution for  a vertical point load  applied 
within  an  elastic half-space.  They  are  dependent on pile  length  to  diameter  ratio  LId,  base  diameter  area  ratio 
RA (ratio  of pile  section  to  area  bounded  by  outer  pile  circumference),  pile  modulus  ,  and  compressibility 
Further factors  are available to correct the settlement for the effects of end-bearing onto a stiffer soil as well as finite 
thickness of the soil stratum in which the pile is floating.  The nonlinear pile response can also be modelled by taking 
into account pile-soil slip. 
1·  0
10 20  30  40  50 
f---- -.. -- --
" ..-.. 1-. ---,







" .t--l

j [ J 
For  lid'" 100
10 :: 0·0254 
For  3    1 ___ 
I I Ia  I
FIGURE 18.4  Settlement-influence/actor, 10 (after Poulos and Davis, 1980)
282 Canadian Foundation Engineering Manual

100 1.000 1QOoo
FIGURE 18.5 Compressibility correction factor for settlement, RK (after Poulos and Davis, 1980)
1· 00 r------,r----,.-----,.----,.----::...,

o 0·2 0'3 0'5
FIGURE 18.6 Poisson's ratio correctionfactor for settlement, R (after Poulos and Davis, 1980)
Randolph (Fleming et ai., 1992) developed a closed form solution for 10 for piles in a soil with a modulus that
increases linearly with depth, given by
1+ I _8_ tanh (J.LL) !::..]
0 = 4(1 +v) 'itA (1-v)
J.LL d
4 4rcp
tanh (J.LL) !::..]
(I-v) +l;
J.LL d
in which
'Ill did, db is the diameter of the pile toe; EslEb whereE
is the soil modulus at the pile toe and Eb is the
modulus of the bearing stratum underneath the pile toe; p = E1EsL and
s = In{[0.25 + (2.5 p(l-v)   2:}
Geotechnical Design of Deep Foundations 283
].lL - 2[_2J (18.30)
- ~ A o   5 d
When  applying the  above  elastic solutions,  the  immediate  or undrained  settlement (for pile  in  clay)  is  calculated 
with undrained ESL values and v  0.5.  For total final  settlement calculations in  sand or clay,  drained values  of ESL
and v  s are used. 
To  employ  an  elastic  c.ontinuum  solution  of this  type,  the soil profile must be simplified appropriately  and  elastic 
properties for  the  soil must be  estimated,  in  particular the  secant modulus Es for  working  load levels.  Poulos  and 
Davis  (1980)  suggest  average  values  of Es for  driven  piles  in  sands,  a value  of v  of 0.3  (where  no  test  data  are 
available) and, for  driven piles, a value for soil modulus below the pile toe of 5Es to  10E
' For clays, Callanan and 
Kulhawy (1985) indicate that Eslsu ranges  from 200 to  900, with  an  average of 500.  Greater values may occur for 
shorter piles where L <  15d. Poulos and Davis (1980)  also provide an empirical correlation  between Es and Su for 
piles  in  clay. Alternatively, the  pile  settlement theory can be used to  back-calculate representative soil parameters 
using  results  from  field  tests  on  model  or prototype  piles.  Kulhawy  and  Mayne  (1990)  provide  a  great  deal  of 
information regarding the estimation of soil parameters for  foundation design. 
F or layered  soil profiles,  it is  adequate  for most practical purposes to  replace the  layered soil along the pile shaft 
with an  equivalent homogeneous soil,  using a weighted average,  i.e.: 
n is  the  number of layers  and E
and  hi are  the  elastic  modulus  and  thickness  of layer  i, respectively.  The 
modulus of the soil at the pile base may be taken as  the average of the soil modulus within a distance equal 
to 2 ~ below the pile toe. 
It is  important to note that the relevant mechanical properties ofthe soil are modified as a result ofpile installation, 
in particular for driven piles.  Consequently,  the  values of Es used  in design  are not equal to values obtained from 
laboratory tests on intact specimens; typical values derived from experience as mentioned above should be used in 
the absence of local experience. 
Nonlinear Analysis: For  floating  piles  (which  derive  most  of their  resistance  from  shaft  friction),  linear  elastic 
solutions  are  generally adequate.  However,  for  end-bearing  piles  (which  derive  a  substantial  proportion  of their 
resistance from the toe), the load-settlement behaviour is strongly nonlinear even at normal working loads. For such 
cases, Poulos and Davis  (1980)  developed an approximate procedure that involves the  construction of a tri-linear 
load-settlement curve.  In this procedure, the shaft and toe ultimate resistances are  estimated and used to construct 
the load-settlement curve of the pile.  Load-Transfer Method 
Soil  data  are  measured  from  field  and  laboratory  tests  and  presented  in  the  form  of curves  relating  the  ratio  of 
adhesion to soil shear strength and to the soil movement. Coyle and Reese (1966) developed the method to estimate 
load settlement response for the pile.  This method  accounts for  the continuity of the  soil  mass  in  an  approximate 
manner  as  the  curves  are  established  from  field  measurements,  which  inherently  contain  the  continuity  effects. 
The load transfer method is particularly useful in modeling the load-deformation performance of piles that display 
strong  nonlinear  behaviour  such  as  very  long  compressible  piles.  O'Neill  et  al.  (1977)  extended  the  method  to 
model  the performance  of pile  groups.  A disadvantage  of this method  is  the  difficulty  in obtaining load- transfer 
curves at a particular site. 

284 Canadian Foundation Engineering Manual

18.3.2  Settlement of a Pile Group
i  Introduction
In  groups of closely spaced piles,  individual  piles  interact  so  that  loads  applied to  any  particular pile  will  lead to 
the settlement of other piles in close proximity. This interaction leads to an overall increase in pile group settlement 
and the redistribution ofloads on individual piles.  Elastic analysis of the pile interaction can be used to  establish to 
what extent the shear resistance ofthe soil causes an unloaded pile to settle as  a result of loads applied to an adjacent 
pile (e.g., Poulos & Davis,  1980; Randolph,  1987; El  Sharnouby & Novak,  1990).  These solutions can be used to 
predict pile  group  response taking  into  account the  pile  cap  stiffness  and its  influence  on  load distribution  within 
the group. 
It is also useful to approximate the pile group as an equivalent single pier, particularly when there is a large number 
of piles  in the group or the  influence of an underlying compressible stratum  is  to be estimated, (e.g.,  see Terzaghi 
& Peck,  1967;  Poulos  & Davis,  1980).  However,  this  has  generally been found  to  predict settlement that  greatly 
overestimates the actual values (uneconomical pile lengths will then result where settlement governs the design).  Empirical Methods for Piles in Sand
The  settlement  of a  pile  group  is  evaluated  on  an  empirical  basis  and  it  has  been  found  that  the  methods  are 
less  reliable  than  those  used  for  single  piles  because  of the  limited  reference  data  available.  For pile  groups  in 
cohesionless soil, two empirical methods are available: 
Vesic's Method 
The ratio  of the  settlement of the  pile  group  with  width, B, to  that of the  individual  pile with  diameter,  d, (Vesic, 
1970) is: 
Sgroup fB)
Sindividual = Vld)
Meyerhof's Method 
The settlement ofa pile group, Sgroup in millimetres, may be related to the standard penetration N ofthe soil (Meyerhof, 
1976) by: 
q = equivalent net vertical foundation pressure, in kPa, detennined from q = QILB, whereQ is total load 
transferred to piles, and Land B are the length and width respectively of the plan area of the pile group 
B =  pile group width, in metres 
I = an influence factor ranging from 0.5  to  1.0, (refer to  Meyerhof,  1976).  Empirical Method for Piles in Clay
For the evaluation of the  settlement of pile  groups in homogeneous clay, Terzaghi  and  Peck (1967)  assumed that 
the load carried by the  pile group is transferred to the soil through an equivalent footing  located at one third of the 
pile length up  from the pile toe (Figure  18.7). The load is  assumed to  spread into the soil at a slope of2V:1H under 
the assumption that the equivalent footing is the top of the frustum of a pyramid. The settlement calculation for the 
1 equivalent footing then follows the methods described in Chapter 11. The Terzaghi and Peck method usually results 
in  settlement values that greatly overestimate the  actual values. Therefore, where settlement considerations govern 
the  design, the method may result in  uneconomical pile lengths. 
I Field tests and long-tenn settlement observations of piles in the sensitive clays of the St.  Lawrence Valley suggest 
that  the  assumption  of an  equivalent  footing  placed  at  the  lower  third-point  is  not  representative  of the  actual 

-- - -
Geotechnical Design of Deep Foundations 285
settlement  behaviour  of a  pile  group.  Blanchet  et  aI.,  (1980)  report  that  the  settlement  of a  pile  group  was  due 
mainly to  reconsolidation  of the  clay  after  driving  and  to  shear creep  deformation with  little  if any  consolidation 
settlement. However, for large pile groups and pile groups supporting bridge abutments the consolidation settlement 
may become the main source of settlement. 
All piles have a neutral plane located at some level in the soil, where an equilibrium exists between the loads on the 
pile above the neutral plane and the shaft-and-toe resistance below the neutral plane. The loads consist of the service 
load  (dead  load,  only)  and  down drag  due  to  negative  skin  friction.  The  negative  skin  friction  is  caused  by  shear 
creep deformation in combination with the large stiffness difference between the soil and the pile (Fellenius, 1984a). 
Accordingly,  the  settlement calculation  of a pile  group,  or of a  single pile,  in  a soil  not  undergoing  consolidation 
settlement from  causes other than from the service load, follows  the same approach as given for piles in soil where 
consolidation settlement from  other causes does occur in the soil around the piles. 
In clay soils, reconsolidation can take an appreciable time, i.e., more than a year for large pile groups, and the pore-
pressure  dissipation  occurring during the reconsolidation will  cause settlement.  Therefore,  the  settlement analysis 
must include the effect of the reconsolidation of the soil around the piles after the pile driving. 
Q  Q 
I J.  J.  J.J..\..  
,; .



1---B, L----1

1  (B+z)(Ltz) 1  , 

V *
i i
i i i i
I \

/1 I. ..I \\

FIGURE 18.7  Stress distribution beneath a pile group in homogeneous clay using the equivalent
footing concept (after Terzaghi and Peck, 1967)  Interaction Factors Method
Piles  in  close  proximity  interact,  so  that  load PI' on  one  pile  with  settlement S1 results  in  a  settlement  aS
of an 
adjacent pile where a  is  called the 'interaction factor'. 
Total settlement of a pile j in a group  of n piles: 
Sj = 

286  Canadian  Foundation Engineering Manual 
S  the settlement of pile j
S  the  settlement  of  a  pile  under  unit  load,  evaluated  using  one  of  the  procedures  from 

Section 18.3.1 
the load on pile i
=  the  interaction  factor  relating  settlement  of pile j to  load  on pile  i. They  are  found  using  elastic 
theory,  provided  in  Figure  18.8  for  floating  piles  from  Poulos  and  Davis  (1980).  Other solutions 
are available for  end-bearing piles .. 
An accurate analysis of settlement of pile groups, based on elastic theory has to be done using a suitable computer 
program,  i.e.,  Poulos  and  Randolph  (1982);  EI  Naggar  and Novak:  (1990).  The  methods  based  on  elastic  theory, 
however,  should  not  be  used  in  situations  involving  downdrag,  creep  or  significant  deep-seated  settlement. 
Furthermore, it  is  only applicable within the working load level.  Pile Cap Conditions 
Two  simplified pile cap conditions can be examined using the general settlement equation shown above: 
A rigid pile cap, where all piles settle an equal amount but loads on individual piles are not known. 
A flexible pile cap, where the loads on each pile are known and each pile has different settlement. 
The  flexible  pile  cap  problem  is  solved by using  the  settlement  equation  directly.  The  rigid pile  cap  problem  is 
solved using the n  general  equations (one  for  each pile)  and the known total  load applied to the pile group,  which 
is the sum of the individual loads: 
~ o t =  L   ~ (18.35) 
There  are  then  n+l  equations  with  n+l  unknowns  (P
, '---- Prr' S).  In  addition  to  the  group  settlement S,  the  . 
individual pile loads are evaluated. 
18.4  Lateral  Capacity of Piles  in  Soil 
Vertical  piles  resist lateral  loads  or moments  by deflecting  until  the  necessary  reaction in the  surrounding  soil  is 
mobilized.  The behaviour of the  foundation under such loading  conditions depends essentially on the stiffness  of 
the pile and the strength ofthe soil. 
The horizontal load capacity of vertical piles may be limited in three different ways: 
the capacity ofthe soil may be exceeded, resulting in large horizontal movements of the piles and failure of 
the foundation; 
•  the bending moments  and/or shear may  generate  excessive bending or shear stresses in the  pile material, 
reSUlting  in structural failure of the piles;  or 
•  the deflections of the pile heads may be too large to be compatible with the superstructure . 
All  three  modes  of failure  must  be  considered  in  design.  There  is  much  room  for  improvement  of these  design 
methods, and often the best method is  still the one based on well-planned and well-executed lateral test loading. 

Geotechnical  Design of Deep Foundations  287 

o-e   :  10 
2 3 4 5
0-2 0-15
V,,, 0-5
0-05 o
FIGURE i8.8a Interaction factors for jloatingpiles, Lld=10 (after Poulos andDavis, 1980)


0-15  0-05  o
FIGURE i8.8b Interactionfactorsfor jloatingpiles, Lld=25 (after Poulos andDavis, 1980)
FIGURE 18.8c Interactionfactors for floating piles, Lld=50 (after Poulos andDavis, 1980)

288 Canadian Foundation Engineering Manual
0-15 0-05 o 
FIGURE i8.8d Interaction factors for floating piles, Lld= 100 (after Poulos and Davis, 1980)
18.4.1  Broms' Method 
Various static analyses oflateralload capacity have been reported, including those of Brinch-Hansen (1961). Broms 
(l964a,b) has presented solutions  in graphical form (see Figures  18.9  and  18.10) for  uniform clay and sand strata. 
In each case, two types of pile failure  are examined: 
'short' pile failure where the lateral capacity of the soil adjacent to  the pile is  fully mobilized; and 
'long' pile failure where the bending resistance of the pile is  fully mobilized. 
Solutions are based on a number of simplifying assumptions that cover the magnitude of lateral soil pressures and 
their distribution  along  the  pile.  Results  are  given  for:  a  pile  of diameter  d  and  embedded length,  L; lateral  load 
capacity  Hu;  yield moment  of pile,  Myield;  clay  cohesion,  c
; coefficient  of passive  sand resistance,  Kp;  height  of 
lateral load above groundline, e;  and soil unit weight, y. 
Poulos (1985) has extended Broms' solutions to  consider lateral load capacity for piles in layered clay soils. 
18.4.2  Pressuremeter Method 
Considering  the  close  analogy  between  the  behaviour  of soils  around  a  horizontally  loaded  pile  and  around  a 
pressuremeter probe, an empirical method for determining horizontal resistance R}, from  pressuremeter test results 
has been proposed by Menard (1962). According to this method, the ultimate horizontal resistance of a short head-
restrained pile may be expressed by: 
ultimate horizontal resistance of pile 
limit pressure from pressuremeter test 
embedment depth of pile 
pile diameter 

Geotechnical Design of Deep Foundations 289
:TI"'·' _+-/+-1_'  
-11- d

a) ShortPile
o 4 8 12 16 20
Embadmant Langth. LId



a L ==-___ Rastrainad
b)Long Pile


FIGURE 18.9  Ultimate lateral resistance ofpiles in cohesionless soils (after Broms, 1964b)

290  Canadian  Foundation  Engineering Manual 



a)  Short Pile 
o  20 
a  X) 

o  4  8  12  16  20 
(a)  Embqdmqnt  Lrzngth  Lid

:J  100 

I  60 

v.  40 

Rrzstrainqd. ---
Frqrz  haadad 

6        16 ---t---i 

b)  Long Pile 

lH o 

1  d 
34  6  10  20  AO 60  100  300  600 
FIGURE 18.10 Ultimate lateral resistance ofpiles in cohesive soils (after Broms, 1964a)
Geotechnical Design of Deep Foundations 291
18.5 Lateral Pile Deflections
The response of a pile to lateral loads is highly nonlinear and methods that assume linear behaviour (e.g., theory of
subgrade reaction and theory of elasticity) are appropriate only where maximum pile deflections are small (less than
1 % of the pile diameter), where the loading is static (no cycling) and where the pile material is linear (e.g., steel).
In most practical applications, one or more of these conditions are not met and methods that can model the pile and
soil non-linearity are called for.
Thep-y curves (unit load transfer curves) approach (see Reese etal., 1974) is a widely accepted method for predicting
pile response under static loads because of its simplicity and practical accuracy. The method allows the analysis of
a pile's response to lateral static, cyclic or even transient loads (El Naggar and Bentley 2000). The method is briefly
described in the following section.
18.5.1 The p-y Curves Approach
Based on model tests, p-y curves relate pile deflections to the corresponding soil reaction at any depth (element)
below the ground surface. The p-y curve represents the total soil reaction to the pile motion. It represents the
relationship between the static soil reaction, p, and the pile deflection, y, for a given p-y curve at a specific load
level. The p-y curves are established using empirical equations (Matlock, 1970; Reese & Welch, 1975; Reese et
aI., 1975). The shape of the p-y curve can be estimated based on laboratory results and back calculation of field
performance data (Matlock, 1970; Murchison & O'Neill, 1984; Gazioglu & O'Neill, 1984) or based on in-situ test
results (Baguelin, et al., 1978; Briaud, et aI., 1983; Robertson, et al., 1986) or curve fit to measured strain data using
an accepted method such as the modified Ramberg-Osgood model (Desai & Wu, 1976).
The general procedure for computing p-y curves in clays both above and below the groundwater table and
corresponding parameters are recommended by Matlock (1970) and Bhushan et al. (1979), respectively. The p-y
relationship was based on the following equation:
=0.5 L
( )
where P" . Yso
p soil resistance
y deflection corresponding to p
n a constant relating soil resistance to pile deflection
Y50 corrected deflection at one-half the ultimate soil reaction determined from laboratory tests.
P ultimate soil resistance, is the minimum of:
P = 3s + yxd + Js x (18.38a)
U Ii 1I
P =9sd (18.38b)
S II the undrained shear strength
y the effective unit weight of the soil
J = an empirical coefficient dependent on the shear strength. A value ofJ = 0.5 is typically used for soft
clays (Matlock 1970) and J = 1.5 for stiff clays (Bhushan et aI. 1979).
The most commonly used criteria for development of p-y curves for sand were proposed by Reese et al. (1974)
but tend to give very conservative results. Bhushan et al. (1981) and Bhushan and Askari (1984) used a different
procedure based on full-scale load test results to obtain nonlinear p-y curves for saturated and unsaturated sand.
Bhushan and Haley (1980) and Bhushan et al. (1981) developed p-y curves for different sands below and above the
water table. The secant modulus approach is used to approximate soil reactions at specified lateral displacements.
In this approach, the soil resistance in the static p-y curve model can be calculated using the following equation:
p = (k)(x)(y)(F1)(F2) (18.39)

292 Canadian Foundation Engineering Manual
k a constant that depends on the lateral deflection y (i.e., k decreases as y increases)
and relates the secant modulus of soil for a given value ofy to depth (Es=kx)
x is the depth at which the p-y curve is being generated
Fl andF2 are density and groundwater (saturated or unsaturated) factors, respectively, and can be
determined from Meyer (1979)
The main factors affecting k are the relative density ofthe sand (loose or dense) and the level oflateral displacement.
The secant modulus decreases with increasing displacement and thus the nonlinearity of the sand can be modeled
accurately. This analysis assumes a linear increase of the soil modulus with depth (but varies nonlinearly with
displacement at each depth) that is typical for many sands.
The actual soil response is a function ofthe pile installation and soil type. Methods used to estimate the non-linear
p-y curves do not always account for changes in ground conditions due to pile installation. Some techniques have
been proposed whereby a pressuremeter is installed in a manner that simulates the pile installation and the non-
linear p-y curve determined from the subsequent pressuremeter test (Robertson, et aI., 1986).
Various methods for modeling laterally loaded piles that employ the p-y curve method or the strain wedge method
(Ashour et aI., 1998; Ashour & Norris, 2000) are encoded in computer programs that are available on the market and
are efficiently used to analyse the nonlinear lateral response of piles. Most of these computer programs account for
soil and pile nonlinearity and can handle static, cyclic or transient loading. Furthermore, they calculate the bending
moment and shear forces along the pile shaft, which are required for the structural design of the pile. Some of the
available programs are LPILE (Reese & Wang, 1997), SWM (Ashour et aI., 1998) and FLPIER (McVay et aI.,
1992). For cases where the load is transient (impact loading, seismic loading, etc.), PYLAT (El Naggar & Bentley,
2000) can be used.
18.5.2  Elastic Continuum Theory 
Poulos and David (1980) present solutions for the lateral deflection of a single pile floating within an elastic
continuum responding to a lateral load, H, applied at distance, e, above the groundline. These solutions make use of
soil modulus, Es and are presented in Figures 18.11 to 18.13 for Poisson's ratio of the ground v = 0.5. Groundline
displacement, p, and groundline rotation, e, are expressed as:
the pile has embedded length, L, and the influence factors IpH' IpM IOH and 10M are given in Figs. 18.11 to
18.13. These particular solutions are for a uniform soil and elastic pile, and use the pile flexibility factor,
where the pile has modulus, E
, and second moment of area, Ip' The soil modulus used in
EsL4 these solutions should be calibrated for a given pile type, magnitude ofload, and nature ofload
(static, cyclic or trasient) through site-specific loading tests whenever possible.
There are other solutions for a pile that yields and for a non uniform soil profile (Poulos & Davis, 1980). Nonlinear
pile response has been examined by Poulos (1982).
18.5.3  Group Effects 
The solutions presented in the preceding sections can be used to estimate the lateral response of single piles. When
piles are installed as a group, interaction occurs between the individual piles so that the lateral pile deformations are
Geotechnical Design of Deep Foundations 293
increased. This effect can be quantified using theoretical solutions of Poulos and Davis (1980) so that the pile group
response can be estimated. (See also the work of Randolph, 1981 and Sharnouby & Novak, 1985). A number of
computer programs that employ mainly linear elastic pile and soil models are available for the response analysis of
pile groups such as PGROUP, DEFPIG and PIGLET (Fleming et al., 1992).
For laterally loaded pile groups, the direction ofthe applied load relative to the group becomes important, particularly
for groups driven in a rectangular configuration where the rectangle length is substantially greater than the rectangle
The proper evaluation of the lateral performance of pile groups requires an approach that accounts for the soil
nonlinearity, especially near the ground surface. Budhu and Davies (1987, 1988) and El N aggar and Novak (1996)
have examined the nonlinear pile group response. The most common design method for laterally loaded pile groups
is based on the p-y curve approach. In this method, piles within the group are analysed for lateral loading per single
piles except that the p values are multiplied by a reduction factor termed the p-multiplier (Brown et aL, 1988; Brown
et al., 2000; Mostafa & El Naggar, 2002). Computer programs are available to facilitate the analysis and design of
laterally loaded pile groups (FLPIER (McVay et al., 1996); GROUP (Reese & Wang, 1996; PYLATG (El Naggar
& Mostafa, 2001)).
Valuf2s of Lit Vs ;: 0-5
~ O
r---.. i
.......... x ~ 5  
'" -& 25
I ~
~ ih.
I I l
1 10
FIGURE 18.11 Values a/l
and lea - free-headfloatingpile, constant soil modulus
(after Poulos and Davis, 1980)

294 Canadian Foundation Engineering Manual

L.Id- '-"'- Valuas Of t---._- 1-----



V, =0-5
"- --
.. ,--_.
l- I-- ,..

'-'- -
r--- ---

r--- r-
4 __"_ ••
f··_··· .. ·
-, .. --
1-"-"- ---
_... -
f-- --
,,---- -.. --
- r-- -
...-. _....__. '-
1------ r"'--'
- -
, 16
10'" 10
FIGURE 18.12 Values of19M - free-headfloatingpile, constant soU modulus
(after Poulos and Davis, 1980)

FIGURE 18.13 Values oflpH -free-headfloating pile, constant soil modulus
(after Poulos and Davis, 1980)
Geotechnical Design of Deep Foundations 295
Geotechnical Axial Capacity of Deep Foundations on Rock
18.6.1  Introd uction
Deep foundations placed on or socketed into rock normally carry heavy loads.  They may be used when the  quality 
ofthe rock mass at the surface is poor. They may be driven, drilled, or cast-in-place. Carter and Kulhawy (1988) and 
Lo and Hefny (200 I) provide a useful review of analysis and design methods for  piles socketed into rock. 
Piles can be driven onto or into rock. However, the exact area of contact with rock, the depth ofpenetration into the 
rock, and the quality ofthe rock at the foundation level are largely unknown. Consequently, the determination ofthe 
capacity of such foundations using theoretical or semi-empirical methods cannot be made with certainty. Therefore, 
the capacity should be confirmed on the basis of driving observations, local experience and test loading. 
18.6.2  Drilled Piers or Caissons· Design Assumptions
Deep foundations can be drilled, bored or excavated, and cast-in-place. In this case, the area ofcontact with the rock, 
the depth of penetration into the rock, and the quality of the rock at the foundation  level can be verified. Therefore, 
the capacity of these foundations  may be determined with  a reasonable  degree of confidence using various  design 
methods. The following discussion relates to the axial capacity of the socketed piers. The behaviour of foundations 
under lateral load is discussed by Poulos and Davis (1980), Kulhawy and Carter (1992), Carter and Kulhawy (1992) 
and Wyllie (1992). 
In most cases, where cast-in-place deep foundations  are socketed into  the rock, the depth  of the socket is typically 
one to three times the diameter ofthe foundation. Present Canadian practice for the design ofsuch deep foundations 
varies from region to region. Three different design assumptions are  in use: 
1.  The capacity is  assumed to be derived from toe resistance only.  This  assumption  can be considered to  be 
safe,  since the capacity of the rock is available,  regardless  of the  construction procedure.  However,  if the 
bottom of the excavation is not properly cleaned, the capacity may not be mobilized before large settlements 
occur owing to the compression of any debris remaining in the bottom of the socket. 
2.  The  capacity  is  assumed  to  be  derived  from  the  "bond"  between  concrete  and  rock  along  the  surface 
perimeter  of the  socket.  However,  theoretical  considerations  indicate  that  the  load  distribution  is  not 
necessarily uniform, but depends upon the modulus of elasticity of both concrete and the surrounding rock 
(Coates,  1967; Williams et al.,  1980). Furthermore, the magnitude of shaft resistance, or "bond", is highly 
dependent  on the  quality of the  rock surface  on the  walls  of the  socket and on the  roughness  of the  rock 
3.  The capacity is assumed to be derived from both toe resistance and shaft resistance. In this case, consideration 
must be given to the  load transfer behaviour of the pier-socket system.  Verification  of the design  load by 
full-scale test and/or well-documented local experience is recommended. 
18.6.3  End·Bearing  Introduction
Toe  or end-bearing  resistance  is  the  area  of the  socket base  multiplied by  the  bearing  pressure.  The  socket  base 
capacity may be considered to provide the whole socket capacity (Approach 1 above) or to provide one component 
of the socket capacity (Approach 3).  Bearing Pressure from Pressuremeter Results
In  situ pressuremeter tests  may  be  useful  in  the  determination  of rock mass  properties.  The  pressuremeter  limit 

296 Canadian Foundation Engineering Manual

, serves as a strength index ofthe rock mass. The ultimate capacity of a socketed pile in rock, R" is given 
by the following equation: 
limit pressure as  determined from pressuremeter tests in the zone extending two pile diameters above 
and below the pile toe 
at rest horizontal stress in the rock at the elevation of the pile toe 
total overburden pressure at the toe of the pile 
an  empirical  non-dimensional  coefficient,  which  depends  on  the  socket  diameter-depth  ratio  as 
TABLE 18.7 Bearing Capacity Coefficient Kb as a Function ofNormalized Depth
The  allowable bearing pressure in working stress design is usually taken as the bearing capacity, R
' divided by  a 
safety factor of 3. 
The  factored  geotechnical  axial  resistance  at  ultimate  limit  states  is  taken  as  the  ultimate  axial  capacity  (R)
multiplied by the geotechnical resistance factor  (cD)  of OA for  compression and  0.3  for uplift (Tables  8.1  and 8.2 in 
Chapter 8). Bearing Pressure from Strength of Rock Cores
The method described in Chapter 9 of this Manual is applicable to  deep foundations. According to Ladanyi and Roy 
(1971) the effect of depth is  included and the formula becomes: 
aK d (18A2)
c sp
qa allowable bearing pressure 
average unconfined compressive strength ofrock core, from ASTM D2938 
=::  empirical factor,  as given in Section 9.2 and including a factor of safety of 3 
d depth factor  1+ A ~ ~ 3 
Ls depth (length of the socket) 
B =::  diameter of the socket 
For  limit states  design,  it  is  suggested  that the  ultimate  axial  capacity be calculated as  multiplying the  allowable 
value by three. The factored geotechnical resistance at ultimate limit states would then be obtained by multiplying 
the  ultimate  capacity  by  the  geotechnical  resistance  factor  of OA and  0.3  for  compression  and  uplift  conditions 
respectively (Tables 8.1  and 8.2 in Chapter 8). 
The uniaxial compression strength is not representative ofthe in-situ mechanical properties ofthe rock mass because 
of the  absence  of discontinuities in the  laboratory test specimens.  For such  a rock mass, the  conventional bearing 
capacity  equation  may be  used,  provided  relevant  strength  parameters  have  been  evaluated  from  in-situ  tests  or 
- ~ - . ~ - ~ -   . ~ . , . . , . . . , - ......--'
Geotechnical Design of Deep Foundations 297
estimated on the basis of a rock mass  classification system as  discussed in Chapter 3 of this Manual. Note that the 
bearing capacity equation usually furnishes an upper bound capacity value. 
This method is generally not applicable to soft stratified rocks, such as  shales or limestones. 
18.6.4  Shaft Capacity of Socket  Introduction
Analytical  studies  of measurements  from  test  loading  of drilled piers  socketed  into  bedrock have  indicated  that 
socket shear can account for a large portion ofthe total capacity. The ultimate socket shear load, Qs' is approximately 
given by the following relationship: 
=7rBLq (18.43)
s s s s
diameter of the socket 
length of the socket 
average unit shear resistance  along the socket 
This  shaft capacity may be  taken as the whole  of the socket capacity (Approach 2  in  Section  18.6.2) or as part of 
it (Approach  3).  The  mechanism  of shear  strength  development  is  complex,  depending  upon  the  frictional  and 
adhesive  strength of the  rock-concrete  bond,  as  well  as  any  changes  in normal  stresses  acting  between rock and 
concrete due to dilation associated with interface slip or Poisson's ratio effects. Empirical data are currently used to 
assess the shear strength of the rock-concrete interface. The factored geotechnical  axial resistance at ultimate limit 
states is  obtained by mUltiplying the ultimate capacity by an appropriate value of the geotechnical resistance factor 
(Tables  8.1  and 8.2 in Chapter 8).  Conventional Piers
Piers that are excavated and constructed using conventional methods have a relatively smooth concrete-rock interface. 
Horvath (1982), Horvath et al (1983), and Rowe and Annitage (1984) have examined the relationship between unit 
socket shear and the compressive strength of the rock. An approximate relationship has been developed: 
qs unit socket shear 
qu unconfined compressive strength of rock 
b an empirical factor 
atmospheric pressure 
If, the concrete compressive strength,  f "  is lower than the unconfined compressive strength of the rock  q then: 
c  '  u'
Values for the empirical factor,  b, have been proposed, as follows: 

298 Canadian Foundation Engineering Manual
TABLE 18.8 Proposed b Values for Different Design Treatments
Proposed Value for b Comment Reference
i Expected average value, for use in limit states
design approach
I Rowe and Armitage (1984)
0.63 to 0.94
: Conservative lower bound value, for use in a
working stress design approach
Horvath et al (1983)
0.63 Conservative lower bound value Carter and Kulhawy (1988)
The range in proposed values for the empirical factor reflects the wide variability in test results. Lo and Hefuy
(2001) note that the differences between the proposed factors are in part due to the difference in the design approach
proposed by the authors as noted in the preceding table. Other methods for estimating side shear resistance are
discussed by Lo and Hefny (2001).
Given the large variability in the test data used to determine the empirical method discussed above, it is important
that in situ testing for direct measurement of side shear resistance be made for projects where this factor is of critical
importance. Grooved Piers
Grooves can be made in the socket wall to increase the roughness of the pier-rock interface and thus, increase
the shaft resistance. Using the expression from the preceding section, a best-fit to data as assessed by Rowe and
Armitage (1984) is b 1.9 for grooves of depth and width greater than 10 mm, at spacings between 50 mm and 200
18.6.5 DeSign for Combined Toe and Shaft Resistance
If both toe and shaft resistance are to be used for estimating socket capacity, then the proportions of load carried at
the sides and base must be estimated. This requires some analysis of the socket/rock system.
Pells and Turner (1979) have analysed the socket/rock system, assuming the concrete and rock materials respond
as elastic isotropic solids, and that the bond along the rock-concrete boundary must not be broken. A proportion, n,
of load reaching the socket base can be estimated from Figure 18.15. Assuming uniform shear over the shaft of the
socket, socket shear qs is:
(1- n)Q
qs = L b
s s
n proportion of Q that reaches the socket base, from Figure 18.14
Q total load applied at the top of the socket
Ls socket length
b socket diameter
This side shear, qs' must be compared with allowable values to ensure stability. Generally, the base load, nQ, will be
easily supported at the toe, but this can be checked using qa in the formula in Section and the base area.

Geotechnical Design of Deep Foundations 299
70  -


0  7  8 
FIGURE 18.14  Load distribution in a rock socket (after Pells and Turner, 1979)
18.6.6  Other Failure Modes
In addition to  the  failure  modes  discussed,  the  designer may need to  consider the axial uplift capacity of the rock-
socketed pile, its lateral capacity or its  torsional capacity. Alternatively, design of the socket allowing for both shaft 
and  tip  resistance  but permitting  slip  along  the  shaft-rock  boundary  may  be  contemplated.  These  situations  are 
examined by Carter and Kulhawy (1988). 
18.7  Settlement of Piers Socketed into Rock
18.7.1  Fundamentals
The  settlement  for  a  pier  founded  on  sound  rock  is  generally  negligible.  Settlement may  be  significant  for  piles 
on  soft rock.  Significant  settlement  of foundations  on  rock  is  often  associated with  the  presence  of open  joints, 
or seams  of compressible  material.  Because  of the  discontinuous  nature  of a  rock  mass,  settlement  analysis  of 
drilled pier foundations placed on, or socketed into, rock is  difficult. Where such conditions are anticipated, special 
investigations and analysis and/or test loading are required. 
Settlement may also result from the presence of mud or debris between the bottom of the concrete shaft and the rock 
surface. Careful inspection of the bottom of each excavation is  necessary to  eliminate this problem. 
Elastic moduli measured on rock core samples have little relation to  the settlement behaviour of rock masses, since 
the influence ofjoints and other rock discontinuities is neglected. A settlement analysis based on such moduli must 
include arbitrary assumptions on the influence ofjoints and is,  therefore, of limited practical value. 

18.7.2 Settlement Estimated from Pressuremeter Testing
Settlements can be estimated on the basis of in situ pressuremeter tests. To do so, a large number of tests must be
performed to allow for an assessment of the variability of the pressuremeter modulus of the rock mass, including
some measure of the influence ofjoints and other discontinuities. However, the effect of occasional thin horizontal
joints and compressible seams cannot be taken into account using this method, and the results may be misleading if
such joints or seams do occur. On the other hand, in highly fractured rock, pressuremeter tests may be the best to
provide reliable results.
18.7.3 Settlement from Plate Test Loading
The in-situ plate test loading can be used to assess the settlement behaviour of a rock mass under a deep foundation.
The importance of size effects on the results of such tests must be recognized. Ideally, the plate should be of the
same diameter as the foundation unit; however, for practical reasons, this is seldom possible and smaller plates are
normally used. The results obtained from loading smaller plates may be considered representative of the actual
foundation behaviour, provided the diameter ofthe plate is not smaller than half the diameter of the foundation unit,
and always larger than 0.3 m.
The results of plate load tests are frequently variable. The evaluation of the test results can be performed using the
three-dimensional elastic displacement approach (see Section 11.3). To obtain a reliable evaluation ofthe foundation
behaviour, a series oftests has to be carried out (see Rowe, 1982). The cost of such tests and of the resulting design
work is high. It is only justified for large projects, or where the structure to be supported is very sensitive to
18.7.4 Settlement using Elastic Solutions
1 Introduction
In cases where settlement is important, design methods based on elastic solutions have been proposed by Ladanyi
(1977), PeUs and Turner (1979), Horvath et al (1983), Rowe and Armitage (1987) and others. An excellent summary
of elastic design is given by Lo 'and Hefuy (200 I). Determination of Material Properties
The subsurface conditions at the proposed site should be thoroughly investigated. The material properties of
the concrete and rock should be carefully determined using appropriate laboratory and in-situ testing methods,
Representative values of rock mass modulus, E
, and average shaft resistance at yield are necessary.
Based on back analysis of pile load test data, Rowe ahd Armitage (1987) proposed the following approximate
relationship for rock mass modulus for use in settlement calculations:
E rock mass modulus
qu unconfined compressive strength of rock
b an empirical factor
Pa atmospheric pressure
The best fit to available data was obtained for b = 680. Based on a statistical study, Rowe and Armitage (1987)
concluded that the probability of exceeding design settlement could be 30 % if a value of b 475 was used, and 11
% if a value of b = 340 was used.
~ ~ ~             < 4  
Geotechnical Design of Deep Foundations 301
The most direct method of determining value ofrock mass properties for design calculations is to perform field tests
on full-scale or small-scale socketed piers. It is important that the roughness factor ofthe test sockets be comparable
to the actual pier sockets. Carter and Kulhawy (1988) discuss various aspects offield tests and their interpretation. Estimation of Settlement of the Pier
Once the pier dimensions have been determined, the settlement of the pier can be calculated using Figures 18.15
to 18.17. If the calculated settlement values exceed the allowable settlement, the diameter and/or length of the pier
socket should be adjusted. These solutions assume that the rock and concrete remain bonded together along the
socket shaft. Other Methods
Carter and Kulhawy (1988) provide a comprehensive review of methods for prediction of load-displacement
response of axially loaded piers. The nonlinear solutions of Rowe and Armitage (1987) can be used to predict axial
deformations in soft rock, both before, and after, slip occurs along the pier shaft. Alternatively, Carter and Kulhawy
(1988) provide analytical solutions that give pier response before first slip, and with full slip along the rock-pier
FIGURE 18.15  Elastic settlement ojshear socket (after Pells and Turner, 1979)

:$U2 Canadian Foundation Engineering Manual
0 0.6 

0 1.0 
0  0.9 

Dfr 2 

a  4  6  8 
FIGURE 18.16 Embedment reduction/actor/or shear sockets. elastic modulus 0/concrete,
=elastic modulus o/rock (after Pells and Turner, 1979)
L-____  __  ____  ____  ____ 
o  2 6  8  10

r S 


1 . 0 
FIGURE 18.17 Elastic settlement 0/a complete rock socket (after Pells and Turner, 1979)

Structural Design and Installation of Piles 303
Structural Design and Installation of Piles 
19 Structural Design and Installation of Piles
19.1 Introduction
This chapter gives information on the use of different types of deep foundations, including special features of
structural design and considerations regarding the installation of such foundations.
19.1.1 Resistance of Deep Foundations Structural Resistance
The structural resistance of a deep foundation unit, determined in accordance with the National Building Code of
Canada, represents the load that the unit can support as a structural member. In most cases, the bearing capacity
of a deep foundation unit is governed by geotechnical considerations rather than by the structural resistance of
the unit. The installation and inspection of a deep foundation unit are generally less controllable than for a similar
superstructure member. Moreover, the environment ofthe deep foundation unit may be potentially more damaging
structurally than the environment of the superstructure member.
It is important to note that permissible deviations in alignment and location of the unit can be established and
considered in the design when determining the structural resistance. Normally, it is not possible to install deep
foundations closer than 70 rom to the specified position and, therefore, the design should allow for this location
limitation. When the off-location effect is considered, the restraining influence of the pile cap, tie beams and soil
should be included. The effects of moments and lateral loads must also be considered in the design. Geotechnical Resistance of a Driven Pile
The geotechnical resistance of a driven pile is a function of the dynamic response of the pile, the so-called dynamic
impedance Ep A/c
, where Ep is the modulus of elasticity, Ap is the cross- sectional area of the pile and c
is the
speed of the strain wave in the pile. The strength of the pile material has no influence beyond a minimum value,
which mostly is smaller than about 250 MPa. Therefore, the geotechnical resistance of a driven pile differs from that
of the structural resistance. The potential geotechnical resistance of two piles with the same impedance is the same,
whether the piles are of the same material or are different, e.g., steel or concrete, whereas the structural resistance
may differ.
The allowable geotechnical stress ofa driven pile should be limited to a factor times Ep/c
ofthe pile material. In the
absence offield verification of the existence and magnitude of soil set-up or soil relaxation, the factor is suggested
to be 2 (units = mls). Field verification by means oftest loading or dynamic monitoring (Subsection 19.1.3) will
supersede this suggestion.
The value of2 Eplc
for steel piles is relatively constant and equal to 80 MPa. For precast piles, ordinary reinforced
piles, and prestressed piles, the elastic modulus, E
' and the wave speed, c " vary and Ep/c is not constant. However,
1 r

304 Canadian Foundation Engineering Manual
20 Ep/c
is usually within the relatively narrow range of 12 to 15 MPa and 12 MPa is suggested for use in design.
For further discussion see Fellenius (l984b).
The dynamic impedance of a closed-toe steel pile can be substantially improved by concreting the pile before
restriking. The resulting increased dynamic impedance (new combined value of EpA/c
) will enable the pile to be
driven to a higher geotechnical resistance andlor to verify the existence of soil set-up. By finishing the concrete with
a slightly convex upper surface that protrudes above the steel tube, the tube avoids impact of the hammer. Also, it
is advisable to add some reinforcing bars to the concrete within a zone of four pile diameters from the pile head.
Soil set-up can be verified in the field by a load test or by dynamic measurements during restriking. While restriking
alone is a highly recommended method of quality control and will verify soil relaxation, it does not provide
sufficiently reliable information on soil set-up on a pile driven to refusal, unless the pile impedance is increased or a
heavier hammer is used that can develop more force and driving energy per blow (as opposed to the hammer used in
initial driving) and, therefore, move the pile to a penetration larger than about 3 mm to 4 mm per blow. Ideally, when
driving composite piles the design should ensure that the impedance, EpAp/c
, ofthe sections of the pile is the same.
Ifthe impedance differs by more than a factor of2, serious damage or driving difficulties can arise. Composite piles
are concrete piles with long steel H-pile ends, or steel pipe and steel H-piles combined, or two sections of different
size concrete piles combined, etc.
When driving a pile with a follower made ofthe same material as the pile, the areas should be equal. If the pile and
the follower are of different material, e.g., a concrete pile and a steel follower, the impedances should be equal. This
means that the steel area should be about 20 % that of the concrete area. For additional comments, see Fellenius,
19.1.2 Wave-Equation Analysis
The one-dimensional wave-equation analysis is the application oflongitudinal wave transmission to the pile driving
process, which provides a mathematically accurate expression describing the mechanics of strain wave travel along
a pile after it has been hit by the ram of the pile hammer. This method takes into account the characteristics of:
• the hammer (mass of the ram or piston, height of fall of the ram, rated energy and impact velocity);
the driving cap or helmet (mass, stiffness and coefficient of restriction of the hammer cushion or capblock,
and the pile cushion, when used);
the pile (material, dimensions, mass and stiffness); and
• the soil (assumed deformation characteristics represented by quake and damping factors for shaft and toe
The wave-equation analysis can be used to great advantage when assisting in the selection ofhammers and capblocks,
in the design of cushions, in the prediction of driving stresses and bearing capacities, and in the choice of driving
The wave-equation analysis is fundamentally correct. It can provide qualitative information to use in, for instance,
the comparison between two hammers. However, the results of the analysis are- only as accurate as the data used as
input in the analysis.
When no direct measurements or observations are available for reference (calibration), it will be fortuitous if the
results are quantifiably relevant to the real situation. In the absence of calibration data from the analysis of dynamic
monitoring (Subsection 19.1.3), the wave-equation analysis should be limited to use for providing a range of results
established with due consideration to the possible variations of the hammer-pile-soil system.
The factored geotechnical axial compression resistance at ultimate limit states is taken as the ultimate predicted
capacity multiplied by a geotechnical resistance factor of 0.4 (Tables 8.1 and 8.2).
Structural Design and Installation of Piles 305
The wave-equation analysis should be recognized as one of the major advances of the current state-of-the art. Its
use is highly recommended. However, it should be considered as a tool among many others and should be used
by a person well experienced not only in wave-equation analysis, but also in the overall art of pile installation and
pile-soil analysis.
Dynamic Monitoring 
The effect of the hammer impact on a pile in terms of force (stress, strain) and velocity (acceleration) can be
monitored using special instrumentation and analyzing the obtained force and velocity 'wave traces'. Information
can be obtained as to the proper functioning of the hammer, the impact force, the transferred energy and the soil
response to the impact on the pile.
The dynamic monitoring method has been used in Canada since 1976 and is well established. For details on the
instrumentation and method see Goble et al. (1970), Rausche et al. (1972), Fellenius et al. (1978), and Authier and
Fellenius (1983).
The soil response may be related to the pile static capacity by a method called Case Method Estimate (CMES).
This method is fast and produces a value for each impact as the driving proceeds. For more accurate capacity