PILE FOUNDATION ANALYSIS AND DESIGNSERIES IN GEOTECHNICAL ENGINEERING Edited by 7. William Lambe Robert V. Whitman Professors of Civil Engineering ‘Massachusetts Institute of Technology BOOKS IN SERIES: Soil Testing for Engineers by T. William Lambe, 1951 Soil Mechanics by T. William Lambe and Robert V. Whitman, 1968 Soil Dynamics by Robert V. Whitman (in progress) Fundamencals of Soil Behavior by Jarnes K. Mitchell, 1976 Elastic Solutions for Soil and Rock Mechanics by H. G. Poulos and E. H, Davis, 1974 Soil Mechanics, $1 Version by T. William Lambe and Robert V. Whitman, 1978 ‘The aim of this series isto present the modern concepts of soil engineering, which is the stence and technology of soils and their application to problems in civil engineering. ‘The word “Soil” is interpreted broadly to inctade all earth materials whose properties and behavior influence civil engineering construction Soil engineering is founded upon many basic disciplines: mechanics and dynamics; physical geology and engineering geology; clay mineralogy and colloidal chemistry; and mechanics of granular systems and fluid mechanies. Prin ciples from these basis disciplines are backed by experimen: tal evidence from laboratory and field investigations and from observations on actual structures. Judgment derived from experience and engineering economics are central to soil enginzering. The books if tis series are intended primarily for use in university courses, at both the undergraduate end graduate levels, The editors also expect that all of the books will serve as valuable reference material for practicing engineers. T. Wiliam Lambe and Robert V. WhitmanA ZDNet PILE FOUNDATION ANALYSIS AND DESIGN G. POULOS E. H. DAVIS The University of Syqpayses-n mireCopyright © 1980 All rights reserved. Published simultaneously in Coad. Reproduction or tansation of any part of this work beysind that petmited by Sections 107 ana 108 of the 1976 United States Copy- Fight Act without the permission of the copyright owner is unlav= ful. Requests for pexmission of further infesmation should be aildvessed ta the Permissions Department (BSP Sas PSeSeseSeseSesesesesesegEsesesasaS: bE Rt +e + At ARAM RET —O— Be Bo Ara a roe fe Skt ak oF RB 10TH BG his &# e seme A Bie kw RAH IOT HE GM Hie he Rem eRe FORIRTWIL : SALW M2261 3d Fada, SeSeSiaseseSenesesesatesasasaqasacasagaayPREFACE ‘This book deals with methods of analysis that may be use ful in design of pile foundations. Many excellent text books are concerned with the more practical aspects of pile foundations, such as the factors influencing the selection of | the type of pile, the techniques of installation, and practical details of construction and maintenance of piles. No attempt has been made to duplicate this type of inform. tion. The aims of the present book are to 1. Present a consistent theoretical approach to the predic: tion of pile deformation end load capacity 2. Present parametric solutions for a wide cange of cases, 3. Demonstrate how such solutions can be used for design purposes, 4. Review the applicability of these approaches to practi cal problems, In any theory, a certain amount of idealization is neces sary to obtain a tractable mathematical solution; this is especially so when dealing with problems involving sol. In dealing with the deformations of pile foundations i this book, we have generally considered the soll as an elastic material, with allowances made for pilesoil sip and soil yield where appropriate. Although real soils possess few, if any, of the attractive attributes of an ideal homogeneous isotropic elastic material, they nevertheless can often be treated as elastic over a limited range of stress, provided that the “elastic” parameters are determined for this stress range. When used in this manner, with due discretion and a measure of engineering judgment, elastc-based theory has hhad considerable success in predicting the deformation of both shallow and deep foundations. Although other simple soll models have also been successfully used for various aspects of pile analysis (For example, the theory of subgrade reaction as applied to laterally loaded piles), elastic theory provides a unified basis for the analysis of all types of foundation; it also makes possible identification of the Parameters that exercise a significant influence on pile performance. Since elastic theory allows consideration of stress transmission through a mass, it can be used to analyze the Interaction between two or more piles and, therefore, to examine the behavior of groups of piles. ‘The material contained in this book is organized as follows: 1. The behavior of piles under vertical loads (Chapters 2106) 2. The behavior of piles under lateral Toading (Chapters 7 and 8) and under combined vertical and lateral loading (Chapter) 3, The behavior of piled rafts (Chapter 10) 4, Piles subjected to vertical of lateral soil movements (Chapters 111013), 5. Miscellaneous topics such as pile buckling, dynamic loading, and pile load tests (Chapters 4 to 16) Although the text deals with a relatively wide range of topics, it is by no means exhaustive, Furthermore, since geotechnical analysis is advancing ata very rapid rate, there may well be cases in which the analytical techniques we describe may have been superseded by more versatile methods capable of modeling real soil behavior more realistically. Nevertheless, we feel that the techniques and solutions presented in this book can be usefully applied to ‘most practical problems and provide a basic series of results against which the results of more sophisticated analyses ray be checked, Some worked examples are given to illustrate the appli cation of the solutions to practical problems. Because units are by no means standardized as yet, some of the examples are worked in’SI units, some in British units, and a few in the Continental metric system, We thank the: many people who have conteibuted to this book and in particular Dr. N. S. Mattes, of the Elec: tricity Commission of New South Wales, who obtained a considerable number of the elastic solutions presented, Dr. J. R, Booker and Dr. P. T. Brown of the University of Sydney, who provided a great deal of assistance with various aspects of the theoretical analyses, Mr. P. J. N. ells, who provided valuable information on the subject of piles,wi PREFACE. to rock, and Dr. T.J. Wiesner, who obtsined some of the solutions presented in Chapter 10. The Civil Engineering Graduates Association of the University of Sydney gave financial support for the postgraduate .course on pile foundations that formed the basis of this book. Grateful acknowledgement is given to Professor J. W. Roderick, former Head of the Department of Civil Engineering, who made the facilities of the Department available to us, to C. J. Peiti, B. Crook, J. Kilpatrick, $.Picken,J. Knight and B, Rocke who undertook the typing and assembly of the ‘manuscript, and R, Brew and H. Papallo who prepared many of the diagrams. HG. Poulos EH. DavisCONTENTS 1 GENERAL PRINCIPLES 11 Introduction 1.2 Structural Approach 1.3 Basie Theory Required 13.1 Failure Theory 1.3.2 Blastic Theory 1.3.3 Changes in Soll Type 1.3.4 The Role of Idealization 14 Examples of Theoretical Pile Calculations 14.1 Vertically Loaded Foundations on Deep Clay 1.42 Vertically Loaded Foundations on Clay over Gravel 1.4.3 Foundations Subject to Rotation 2. EFFECTS OF INSTALLATION OF PILES 2.1 Introduction 2.2. Effects of Pile Driving ia Clays 2.2.1 Influence on Soil Shear Strength and Pile Capacity 2.2.2 Pore Pressures Developed During Driving 2.2.3 Dissipation of Excess Pore Pressures 2.2.4 Displacements due to Driving 2.3. Effects of Pile Driving in Sands 23.1 Single Piles 23.2 Pile Groups 24 Bifects of Installing Bored Piles 24.4 Clay Soils 24.2 Sands 3 ULTIMATE LOAD CAPACITY OF PILES 3.1 Introduction 3.2. Ultimate Load Capacity of Single Piles 3.2.1 General Expression 3.22 Piles in Clay 3.2.3 Piles in Sand 33 Pile Groups 3.3.1 Pile Groups in Clay 10 B rf 4 1s 1s i 18 18 18 18 19 24 30 30 3.322 Pile Groups in Sand 34 Piles to Rock 34.1 Point Bearing Capacity 3.42 Pile Rock Adhesion 3. Use of In-Situ Tests 3.5.1 Static Cone Penetromeser 3.5.2 Standard Penetration Test 3.6 Special Types of Pile 3.6.1 Large Bored Piers 3.6.2 Under-Reamed Bored Piles 3.6.3 Screw Piles 3.7 Uplift Resistance 3.7.1 Singie Piles 3.7.2 Pile Groups 3.8 Load Capacity of Bent Piles 4 LOAD CAPACITY BY DYNAMIC METHODS. 4.1. Introduction 4.2 Pie Driving Formulas 4.2.1 Derivation of General Formula 4.2.2 Practical Driving Formulae 4.23 Realiability of Dynamic Formulae 4.3 Pile Driving Analysis by the Wave Equation 4.3.1 The Wave Equation 4.3.2 Smith’s Idealization 4333 Basic Equations 4.3.4 Values of Soil Parameters 43.5 The Effect of Set-Up" 44. Typical Solutions from Wave Equation Analysis 4.4.1, Resistance versus Set Curves 4.4. Pile Stresses 4.5. Reliability of Wave Equation 4.6 Pile Impedance 5 SETTLEMENT ANALYSIS OF SINGLE PILES 5.1 Introduction 5.2. Theoretical Methods of Analysis 5.2.1 Load Transfer Method 3s 38 38 40 41 4l 43 a 43 “4 44 45 45 48 49 32 2 33 33 54 s4 58 39 59 61 63 65 66 66 68 68 0 n n n nwil CONTENTS: 5.2.2 Analysis Based on Elastic Theory 5.2.3 Finite Element Analysis 5.2.4 Comparison between Solutions from Mindlin Approach and Finite Element Analysis 523. Theoretical Solutions for Settlement and Load Distribution 5.3.1 Stress and Load Distribution in Pile 5.3.2 Load Transferred 10 Pile Tip 5.333 Settlement of Pile 5.34 Settlements in a Soil Mass Resulting From a Pile 5.3.5 Immediate and Final Settlements 54 Simplified Method for Constructing Load- Settlement Curve to Failure 5.5. Determination of Soil Parameters SS.1 Laboratory Triaxial Tests 5.5.2 Pile Loading Tests 5.5.3. Empirical Correlations 5.5.4 Typical Values of K 5.6 Some Compsrisons between Observed and Prodicted Pile Settlements, 6 SETTLEMENT OF PILE GROUPS 6.1 Introduction 62. Analysis of Group Settlement 6.2.1 TwoPile Interaction Analysis 6.22 Interaction Factors 623 Analysis of General Groups 6.3. Theoretical Solutions for Freestanding Groups 6.3.1 Settlement of Floating and End- Beating Groups 63.2 Load Distribution in Groups with Rigid Cap 63.3 Groups with Equally Loaded Piles 6.38 Approximation of Group as 2 Single Pier 64 Settlement of Groups Caused by Compres- sible Underlying Strata 65 Preparation and Use of Design Charts 646 Surface Settlements Around a Group 6.7 Observed and Theoretical Group Behavior 62.1 Settlements 672 Load Distribution 6.73 Group Behavior Predicted feom Single Pile Test Results 7 ULTIMATE LATERAL RESISTANCE OF PILES 7.1 Introduction " 83 83 84 84 85 86 94 96 99 101 102 102 102 103 109 109 nie 110 no nv 120 120 126 128 129 132 133 135 135 135 139 141 143 143 7.2 Single Piles 7.2.1 Conventional Statical Approach 72.2 Broms’s Theory 7.2.3 Plane Strain Solutions 7.24 Piles with Significant Base Resis- tance 728 Socketed Piles 7.26 Piles Subjected to Inclined Loading 7.2.7 Battered Piles 73 Pile Groups 73.1 Groups of Vertical Piles 7.3.2 Groups Containing Battered Piles 74 Use of Pies to Increase Slope Stability 7.5. Methods for Increasing the Lateral Resis- tance of Piles 8 LOAD-DEFLECTION PREDICTION FOR LATERALLY LOADED PILES 8.1 Introduction 8.2. Subgrade-Reaction Analysis 82.1 Basic Theory 8.22 Solutions to Linear Theory 8.23 Modulus of Subgrade Reaction 8.2.4 Nonlinear Analysis 8.3. Elastic Analysis for Single Piles 8.3.1 Basic Theory 8.3.2 Solutions for Floating Pile in Uni form Soil 8.3.3 Soluvions for Floating Pile in Soil with Linearly Increasing Modulus 83.4 Solutions for Socketed Files 84 Analysis of Pile Groups 84.1 Introduction 8.4.2 Elastic Analysis of Interaction Between Two Piles 8.4.3 Solutions for TwoPile Interaction 8.444 Elastic Analysis of General Pile Groups 8.4.5 Elastic Solutions for Square Groups 8.4.6 Approximate Prediction of Load- Deflection Curve for a Group 8.5. Determination of Soit Modulus 8.6 Comparisons Between Theoretical and Ob- sesved Load Deflection Behavior 9 GENERAL ANALYSIS OF PILE GROUPS 9.1 Introduction 9.2. Simple Statical Analysis 9.3. Equivalent-Bent Method 9.3.1 Principle of Method \3 144 146 152 153 153 134 156 157 157 159 160 161 163 163 lea 64 166 m 175 7 7 182 192 199 209 209 au 216 217 21 223 2s 233 233 233, 24 2494 9s 10 PILE. 101 102 103 104 105 9.3.2 Determination of Equivalent Bent 9.3.3 Torsional Response of Piles Elastic Analysis of Pile Behavior 9.4.1. Analysis of Single Battered Pile 94.2. Analysis of Pile Groups 9.4.3 Parametric Studies of Pile Groups ‘Comparison of Methods of Pile-Group Analysis -RAFT SYSTEMS Intcoduetion Analysis Elastic Solutions for Square Groups 10.3.1 Influence of Poisson's Ratios ¥, 10.3.2 Influence of Pile Arrangement 10.3.3 Systems Containing Large Numbers of Piles 10.3.4 Effect of Pile Compressibiity and Raft Flexibility Simplified Analysis for Load Settlement Curve to Failure Other Analytical Approaches 11 NEGATIVE FRICTION ON END-BEARING PILES 11.1 Introduction na 113 na Field Studies on Instrumented Piles 11.2.1 Observed Downdrag Forces 11.2.2 Development of Downdrag with Time 1123 Effect of Pile Driving on Negative Friction 11.24 Methods of Reducing Negative Frievion Analysis of Downdrag Forces 113.1 Introduction 11.3.2 Analysis of Final Downdrag Forces 11.3.3 Development of Downdrag with Time 11.34 Modifications to Elastic Analysis ‘Theoretical Solutions for Single Pile 11.4.1 Final Maximum Downdrag Force 11.4.2 Rate of Development of Downdrag Force 11.43 Pile Settlement 114.4 Rate of Development of Settlement L145 Effect of Pile Crushing 235, 237 237 237 242 243 248 250 250 250 253 237 257 287 258 259 262 265 265 269 269 269, 269 269, 271 271 272 273 278 274 274 28 279 282 283 3 lus 116 CONTENTS ix 11.46 Pile in Soil Subjected to Variable Loads 11.4.7 Data on Pile Soil Parameters Pile Groups ‘Comparisons Between Measured and Pre- dicted Pile Behavior PILES IN SWELLING AND SHRINKING SOILS 121 122 123 124 125 126 12.7 Introduction Existing Methods of Analysis Analysis Based on Elstic Theory 123.1 Basic Analysis 123.2 Pile Soil Slip 12.3.3 Compression Failure of Pile 12.34 Tension Failure of Pile 123.5 Nonuniform Soil 12.36 Variation with Time ‘Typical Solutions for Pile Movement and Load 124.1 Purely Elastic Pile Soil Interface 12.4.2 Solutions Incorporating Pile Soil Slip 12.4.3 Effect of Tensile Failure of the Pile 124.4 Differences Between Piles in Swelling and Consolidating Soils Design Curves Application of Theoretical Analysis to Practical Problems 12.6.1 Prediction of Soil- Movement Profile 12.6.2 Pile-Soil Interface Strength 12.63 Soil Modulus Observations of Pile Behavior and Compati sons with Theory PILES IN SOIL UNDERGOING LATERAL MOVEMENT 13 32 133 134 Bs Introduction Analysis Typical Results 13.3.1 Effect of Relative Pile Flexibility 13.3.2 Effect of Boundary Condicions 13.3.3 Effect of Soil Movement Distribu. tion 13.34 Effect of Magnitude of Soil Moves ment 13.3.5 Effect of Pile Diameter 13.36 Effect of E, and p, Distributions Application of Analysis to Practical Problems Comparisons with Field Measurements 284 285 288 289 294 294 295 296 296 297 297 297 297 298, 298 298, 299 303 304 304 306 306 307 au 31 312 314 315 316 316 316 318 318, 319 319contents 14 BUCKLING OF SLENDER PILES 1s 4a 142 43 144 145 Introduction Fully Embedded Piles 14.2.1 Basie Subgrade Reaction Theory 14.2.2 Solutions for Constant ky 14.2.3 Solutions for Linearly Varying ky Partially Embedded Piles 14.3.1 Theoretical Approach 14.3.2 Solutions for Constant ky 14.3.3 Solutions for Linearly Varying ky Effect of Practical Complications 144.1 Axial Load Transfer Along Pile 14.42 ‘Initial Imperfections 14.43 Inelastic Buckling 144.4 Group Effects ‘Analysis Using Elastic Theory 14.5.1 Analysis 14.5.2 Typical Solutions 14.5.3 Comparison with Subgrade-Reaction Solutions DYNAMIC LOADS ON PILES 1s 152 Introduction Estimation of Dynamic Loads 15.2.1 Machine Loads 15.2.2 Wave Forces 15.2.3 Earthquake Forces Pile Response to Axial Loads 15.3.1 End-Bearing Piles 15.3.2 Floating Piles or End-Bearing Piles with Load Transfer 15.3.3 Pile Groups Pile Response to Lateral Loading 15.4.1 Equivalent Cantilever Systems 15.4.2 Finite Difference Analysis 15.4.3 Novak's Analysis 15.44 Pile Groups 336 336 337 338 338 338 339 339 341 345 347 348 348 351 351 155. Pile Response to Earthquake Forces 16 PILE LOAD TESTS 16.1. Introduction 16.2 Maintained Loading Test 162.1 Procedure 16.2.2 Interpretation of Load Tests 16.3 Constant-Rate-of Penetration Test 164 Method of Equilibrium 16.5. Sources of Error in Settlément Measure- ‘ments in Pile Load Tests 16.5.1 Errors Resulting from Use of Re. ference Beam 16.5.2 Errors Resulting from Jacking ‘Against Anchor Piles 16.5.3 Enrots Resulting from Jacking Against Ground Anchors 16.6. Lateral Load Tests 16.7. Torsional Testing APPENDIX A. INTEGRATION OF MINDLIN'S EQUATIONS FOR PILE SETTLEMENT ANALYSIS APPENDIX B ELASTIC EQUATIONS USED FOR LATERALLY-LOADED PILE ANALYSIS REFERENCES AUTHOR INDEX SUBJECT INDEX 353 354 354 355 355 356 358. 358 359 359 360 363 365 365 366 369 a7 383 389PILE FOUNDATION ANALYSIS AND DESIGNGENERAL PRINCIPLES 1.1 INTROD TION ‘The use of pills is man’s oldest method of overcoming the difficulties of founding on soft soils, Although it dates back to prehistoric fake villages, until the late nineteenth cen: tury, the design of pile foundations was based entirely on experience, or even divine providence. Modern literature on piles can be said to date from the publication of Piles and Pile Driving, edited by Wellington of the Engineering News (later to become the Engineering News-Record) in 1893 in whick che widely known Engineering News pile-driving formula was proposed. Since this frst atiempt ata theoret: ical assessment of the capacity of a pie, a great volume of field experience and empirical data on the performance of pile foundations has been published. In recent years, the increasing demang on the foundation engineer to predict reliably the behavior of his pile designs has stimulated more-sophisticated theoretical research into the interaction between a pile or piles and the embedding soil, so that a Targe volume of empirical knowledge is now balanced by a comparable theoretical understanding This balance between empicicism and theory is a com- ‘mon feature of progress in many engineering fields, Any engineering design or solution to a practical problem can bbe imperfect in two ways. It can be imperfect because it is inadequate: that is, parts of the structure fail or deform excessively; in the present context, the design involws too few, too-slender, or too-short piles, But it can also be imperfect because it is more than adequate (too many, oolong, or too-substantial piles) and therefore is an excessively costly design, Design based on empiricism alone tends to focus attention on the former, because recorded experience gerierally only distinguishes between unsatis: factory and trouble-free pecformance and rarely between economical and uneconomical design. Only by under- standing the behavior of the engineering structure in an analytical as well as empirical sense can engineers reason- ably expect to achieve designs that are neither inadequate inor overadequate. In other words, to obtain the full benefit of experience of actual engineering behavior, it is important (o fiave a sound theoretical understanding of the problem. OF course, it is equally important that engineering theory should be based initially on oxperience and extended or ‘modified in the light of further experience, 1.2 STRUCTURAL APPROACH It is only too convenient to divide the design of major buildings into two components: the design of the structure and the design of the foundations. The structure for its own2 GENERAL PRINCIPLES reasons alone i assumed to produce certain column loads, and tie foundations are merely required to carry these predete mined loads. In reality, for complicated structures, the load: on the foundations determine their movernent, but this movement affeots the loads imposed by the struc: ture; thete is inevitably, iteration between structure and foundation. tn fact, the whole complex of structural frame, foundation components (footings, piles, pile caps, raft, etc. and soll or rock forming the founding material, together comprise one interacting structural system, The interaction between a pile and its embedding soil, and that between one pile and another pile, provide subsets of the lager set of all interacting structural components. If an overall structural approach is to be successful, we need to know much more about a particular pile than that it can be classified as, say, a 50-ton pile, We need to know its load-settlement behavior up 0 failure, possibly its behavior under lateral load and moment, and how its behavior is modified by adjacent piles. This is analogous to saying that we need the complete load-deformation characteristics of beams and columns, not just their load capacities, before we can analyze complete. strvctural frameworks. Most of this book is concerned with bringing the ana- lytical treatment of the losd-deformation and the failure behavior of pile foundation systems to the same level of sophistication as similar analytical treatments available for systems of structural frames. With this achieved, it is a relatively simple matter with modem computer programs to combine the structural and foundation systems into ‘one—but that matter is outside the scope of this book. 1.3 BASIC THEORY REQUIRED Piles embeded in soil provide a reinforcement to the soil, increasing its load capacity and modifying its deformation ‘behavior in much the same way as the steel reinforces the concrete in reinforced ot prestressed concrete members, Unfortunately, although a sufficiently accurate analysis of the effects of reinforcement in concrete members can usually be obtained by adaptation of the simple theory of bending, the extended-continuum nature of the emived- ding soil around piles makes the corresponding analysis of the reinforcement effect of piles much more difficult 13.1 Failure Theory In the present state of knowledge, it is generally only possible to consider failure as something that ogca7s mainly at the interface between the sides of the pile and the soil, ignoring the detalls of failure within the sol, although for the pile base, ordinary bearing-capacity theories may be applicable. Thus for vertical failure, the shear stress at the shaft-soil interface attains a limiting value (possibly varying with depth and soil type), and for horizontal failure result ing from lateral load or moment, the normal stress at the interface attains a limiting value (again, possibly varying with depth). In such a simplified approach, any reduction in failure load for a particular pile because of the presence of a nearby pile cannot be taken into account, except that the failure load for a group of closely spaced piles ‘can sometimes be calculated from bearing-capacity theory for buried footings on the assumption that the piles and soil between them act as one solid block. This load can be taken as the answer if itis less than the um of the failure loads for the piles, calculated individually, 13.2 Elastic Theory Soil and rock are not ideal elastic swaterials in that stress and strain afe not linearly related, strains are not fully recoverable on reduction In stress, and strains are not independent of time, However, at east it can be said that strains in soil increase as styesses increase. Furthermore, the assumption of anything mote complicated than 4 linearly elastic material for the soil in the pile-soil contin- tuum situation would generally Jead to unduly complicated theory lacking useful generality. The use of linear elastic theory is therefore expedient and should be sufficiently accurate for engineering purposes, provided that elastic “constants” are employed that are appropriate to the patticular problem. That is, they have either been back- figured from field tests on piles in similar situations, or determined from laboratory tests employing stress changes similar to the average changes in the soil mass in the partic lar case. ‘The basic elastic sesponse of the soil from which the solutions for elastic piles in elastic soil can be derived is given by Mindlin’s set of equations for the stresses and displacements throughout an elastic half-space resulting from horizontal or vertical point load applied at a point beneath the surface, As will be explained in subsequent chapters, this basic response can be integrated to give the pile soil interface stresses in such a way that the displace- ments of the pile and soil are compatible, Modifications to take account of failure at some parts of the interface are then relatively easy to make. Alternatively, the elastic response can be assumed to be that of a series of unconnected springs, that is, a Winkler medium or the subgrade reaction assumption, In spite of what is said by some of the protagonists of this approach, it most Gundamentally be inferior to the elastic continuum approach of the Mindlin equations, since it ignores thevery real interconnection among elements throughout the soil mass. However, it does have the advantages of com. putational simplicity and perhaps moreeady adaptation to complications such as change in soil type. On the other hand, it can never take into account the important matter of interaction between adjacent pies, 1.3.3 Changes in Soil Type Appropriate idealization of actual subsurfate conditions frequently involves consideration of one or more distinct layers of material of different properties. Piles in soft clay are often driven to a stiffer stratum of sand, which may in turn overlie a different clay before encountering bedrock, Both the failure theory and the deformation theory should therefore be capable of coping with such changes in properties from layer to layer. The modification cof the failure theory for this matter presents litle diff culty, but the modification of the deformation theory may require a number of simplifying assumptions. 1.34 The Role of Idealization Engineering theory can only give the behavior in an ideal situation, It must start from assumptions with regard to the properties of the materials and their disposition. In complicated problems such as that of pile behavior, the engineering theory itself is often not fully rigorous, since approximations have to be made to obtain numerical answers even for the ideal situation. Thus, there is a judge: ment to make about the extent to which a particular set of numerical answers is an accurate enough answer to the ideal problem. This judgement is largely the job of the research engineer. There is also a judgement to be made on the practical side, of the accuracy with which the idealized situation fits the real situation, This judgement is largely the job of the practicing engineer. It is important that the difference becween these judgements should be recognized and that they should never be made as one. For example, the fact that predicted behavior is not verified by subse- quent observation can mean either that the prediction was based on inaccurate theory, or that it was based on an ‘unrealistic idealization, or even both, 1.4 EXAMPLES OF THEORETICAL PILE CALCULATIONS Asa foretaste of the types of prediction possible from the theory to be given in subsequent chapters, the results of caleulations are given for differentexample pile founda tions, and for comparison, a surface pad footing, GENERAL PRINCIPLES 3 Working Loods ene [tours |2towns | aeones | a Le! | eacare J fseo —_Shrs%0 1500 eee a Js coset | Ey 700K" 78 Ee seoket | Vsoa uu Cores foctng CS) gee re le FIGURE L.1 Example 1. Foundations or deep cay 14.1 Ver lly Loaded Foundations on Deep Clay The’ foundations considered are shown in Fig. 1.1. All carry the same load and have the same factor of safety TABLE 1.1 BEHAVIOR OF EXAMPLE FOUNDATIONS ON DEEP CLAY bea Pad Single Pile* 4Fite Fxample Tooting Pile pad Group Factor of Sufeny 222 tof Pad or cap 100 nM Shaft sw 8S « Pile base Boot 3 Percent of Pad or eap 100) seers working Shaft 2 38 Toad taken Pile base g 2 2 by Settlements Immediate 09 in? 231m. 08m at Percent of OR 36 OF working immed. resulting load fom yield Contolidation —4.2in. OL ip, Oj. 0.2m Total final 53in, 1.0in 27in_ 1.0%n, 6 See Fig. 1 Elastic shortening of ple as 9 columo = 0,04 in. © sti shortening f pile groupas columns = 0.11 in4 GENERAL PRINCIPLES against ultimate failure. The calculated behavior is given in ‘Table 1.1, from which it can be seen that the surface pad footing, an unlikely choice for a comparatively heavy load ‘on a soft clay, settles what would probably be an excessive amount. Furthermore, a high proportion of the settlement is inrecoverable, so that variations in load might produce further settlement. The single large-diameter pile and the fourpile group have similar behaviors and may even involve setilements that are more than satisfactorily small. The ‘ease of a pad with one small-diameter pile is unusual but represents an interesting intermediate case between pad only and piles only. At the working load, the pile is carrying its full faluredoad but nevertheless succeeds in reducing the settlement well below that of the pad on its own, 2 Vertically Loaded Foundations on Clay over Gravel The cases of a surface pad footing and a pile driven to a stiff gravel base under the soft clay are illustrated in Fig 1.2. Again, each foundation caries the same load and has the same factor of safety. The results of calculation ate given in Table 1.2, The behavior of the pad footing is unaffected by the gravel, since the clay has @ depth of Tne times the footing breadth. The pile, being “end bearing,” can be of smaller diameter than before; in fact, TABLE 1.2 BEHAVIOR OF EXAMPLE OUNDATIONS ON CLAY OVER GRAVEL? 2 2 Example Pad footing __"End-bearng” Pe Factor of Safety 2 (cones stuength rovers) Percent Pad 100) Foiture on Shaft 1B taken by: Pile base 87 escent Pad 100 : working Shatt 3 Toad taken Pile base 6 by Settlements Immediate 4.1, 03° in at Percent immed. 565% 0 working resulting load from yield Consol Lin, ° Tonalvinal 5.3 05 in © See Fi. 12. Ehustic shortening of pile 8 column = 0.25 in Working Loads pee per 29, soft Clay yf cut OS ket Ey +100 kst . £6 00 ket a ¥s03 (1 Pos Foetng (0)"Era Bearing FIGURE 1.2 Example 2. Foundations on cay over eave itis the strength of the concrete of the pile which deter- imines its diameter, rather than soil properties. The settle- meat of the pile is now even smaller, as would be expected, although still greater than the straight column compression, It is interesting to note that although the pile is classified as “end-bearing,” a third of the load is in fact taken by side shear on the shaft. For more slender piles, the pro- portion of the load taken by the shaft can be even higher. ‘The fact that “end-bearing” piles are far from 100 percent cend:bearing has been verified in the field in several instances, 1.43 Foundations subject to Rotation For the same vertical Iuad as in the previous examples, but also with a horizontal load and 4 moment applied, ‘a pad and pier are compared in the third example (Fig. 1.3), The results of calculations are given in Table 1.3, In order to carry even a relatively small moment, the pad footing has vad to be enlarged to an inordinate size, and in view of the movements and rotations at the working loads, it woulé unlikely to be considered a satisfactory solution, The Iength and diameter of the pier have been selected so that the factor of safety against failure resulting from the vertical load is the same as that against failure resulting from the horizontal load end moment.Working Losds 240 ws 340 Kips Fra ep tt 700 kip 81 hi 2) spe GENERAL PRINCIPLES 5 ‘TABLE 1.3 BEHAVIOR OF EXAMPLE FOUNDATIONS SUBJECT To ROTATION 3 3b i Example Pad Footing Single Pier c J Factor of Safety 2 2 a | Sauare Settlement Mnmediate 27a Lin | Consol. asin, 0.2% fe) Pad Fasting {rea Touafieal —_3.3in Lin Soft cloy 50 Horizontal Immediate 14 in 08 pera i Gisplacement Consol o3in A2in | Total final 17m Lin es eh Per ET 6 80 kst Rotation Immediate 0.45" os" virag Consol og oot" Tota final 055° 0.20" FIGURE 13 Foundations subject to tation Percent immediate movement 56% o resulting From loca yield see Fig.EFFECTS OF INSTALLATION OF PILES 2.1 INTRODUCTION Piles may be classified in a number of ways, the common methods being (@) By the material of which they are formed. (b) By the method of installation. ‘The commonly used materials for piles are concrete, steel, and timber; discussions on the relative practical merits of cach type of pile for various applications mey be found in many references, for example, Chellis (1962), Tomlinson (1975), Bowles (1977), and Whitaker (2970). In terms of installation method, piles may be classified as (2) Deiven piles, (b) Bored or castin-situ piles, (c) Driven and castinsitu piles (a) Screw pites, Detailed descrivsions of these methods and equipment used in installation may also be found in the above four references. ‘The method of installation of a pile may have @ pro- found effects on its behavior under load. It may also determine the severity of effect on nearby structures, including undesirable movements, vibrations, or even structural damage, Much of the available data on installa- tion effects is concerned with driven piles, since pile driving gererally creates more disturbance than other methods, Relatively little is known of the effects of constructing bored piles In this chapter, the effects of pile driving in clays are examined, with particular emphasis on the pore pressures developed around the pile and the resulting influence on the surrounding soil Piles driven into sand are then con: sidered, and finally, a brief review, largely qualitative, of the effects of installing bored piles is given It should be emphasized that this chapter is concerned with the extent to which installation changes the proper- ties of the goil surrounding piles from those existing prior to installation, which are presumably determinable by normal methods of site investigation, sampling, and labora- tory or insitu testing, Furthermore, it is concerned with the manner in which such changes, at least with clays, ‘may subsequently become modified the longer the loading of the installed pile is delayed, The change in bearing capacity of a loaded pile as a clay consolidates under the stresses produced by the load on the pile is not dealt with here, but is considered in Chapter 3, It is important to maintain a clear distinction between these two matters2.2 EFFECTS OF PILE DRIVING IN CLAYS ‘The effects of pile driving in clays have been classified into four major categories by de Mello (1969) (@) Remolding or partial structural alteration of the soil surrounding the pile. (©) Alteration of the stress state in the soil in'the vicinity of the pile (©) Dissipation of the excess pore pressures developed around the pie. (@ Long-term phenomena of strength-regain in the sol, Some data is available on al the above effects, although the state of knowledge, particularly in relation to (d), is gener- ally limited, 2.2.1 Influence on Soil Shear Strength and Pile Capacity Early investigations into the effects of pile driving on the properties of clays were made by Housel and Burkey (1948) and Cummings, Kerkhoff, and Peck (1950). Based fon the evidence from load tests to failure carried out on piles at different times after theic installation, it can be inferred that the undrained strength of a clay is initially decreased considerably because of driving, but that signif- icant regain of strength occurs with elapsed time between driving and pile testing. Generally, it may be expected that the driving of piles into clay will initially cause some (or even considerable) loss in undrained strength of the clay because of remolding at constant water content. Subse quently, the strength will usually increase because of a combination of two factors: thixotropic regain of ‘un- drained strength as the structural bonds destroyed by remolding are at least partially restored. and increase resulting from local consolidation of the clay produced by dissipation of excess pore-water pressures that arise from the increase in stress in the soil surrounding the pile. Conceivably, there would be situations in which the consolidation was negative (Le., a swelling with time), thus producing a weakening in addition to that caused by remolding—for example, for stiff, overconsolidated clays, Although investigations into the extent of the dis- turbance around a pile caused by driving have produced somewhat conflicting results, the available evidence (de Mello, 1969) suggests that immediately after driving, the amount of remolding decreased from about 100% at the pilesoil interface to virtually zero at about 1.5 to 2.0 iameters from the pile surface, Investigations by Orrje and Broms (1967) of concrete piles in a sensitive clay EFFECTS OF INSTALLATION OF PILES 7 of. 77 bx Howe! Io uton Gu £60 gs ang 20 ia sa 100 se 4 5 0 100 «1000 FIGURE 2.41 Increase: of load pacity with time (Soderberg, 1962) showed that the undrained strength had almost returned to its original value after nine months, except when piles were spaced at less than about 4.0 diameters, in which case little strength-cegain (and in some cases a further loss) was noted with time. Other than for thixotropic regain, the rate of increase of soil strength subsequent to pile driving is reluted to the rate of dissipation of excess pore pressure. Data presented by Soderberg (1962) showed that the increase in ultimate load capacity of « pile (and hence, shear strength of the soil) was very similar in character to the rate of dissipation of excess pore pressure with time (see Fig. 2.1). Some estimate of the “setup” time m.y be obtained from a knowledge of the excess pore pressures developed around the pile, and the rate of dissipation of these pore pressures. 2.2.2 Pore Pressures Developed during Driving A number of measurements of the excess pore pressure developed in a soil because of pile driving have been made; forexample, Bjerrum et al.(1988), Bjerrum and Johannessen (1960), Milligan et al. (1962), Lambe. and Hom (1965), Lo and Stermac (1965), Orge and Broms (1967), Hanna (1967), Koizumi and Ito (1967), D’Appolonia and Lambe (1971). Results of measurements of pore pressure at the pile face in many of these papers have revealed that the excess pore pressures may become equal to or even greater than the effective overburden stress. However, the induced excess pore pressures decrease rapidly with distance from the pile and generally dissipate very rapidly A summary of some measurements of the variation with radial distance of the excess pore pressures around a single driven pile are given in Fig. 2.2. The excess pore pressure ‘hu is expressed dimensionlessly 28 Au/o'yo, where o'yo is the vertical effective stess in-situ prior to’ diving, while5 EFFECTS OF INSTALLATION OF PILES Lage © wnacebue (Lo ana termae 198) 18 Short ivr (La and Sermae 1086) 1 nb River (Lo aa Stermos, 13651129 f pan fas ber (Lo ana Stem, 1965197 deh there Clay (Bjrrom and Johannsen, '9601-17 5 apm) 1b Movie ley (jem and Jotonnesen, 1980-170. deh 19 Farm in fara 9691-160 doth) Berton Blum iy (D'Apsotoni ara Lambe 1970) 1 Vawved Clay fSoderronandfeligen 1961-120 dosh) 'F Vareed Gly (Soden and tin, 1961-125 f depth 1 Vaved Clay (Stem sea Reign, 138/130 Fe ep etsy FIGURE 22 Summary of measured pore pressures the radial distance r from the pile is expressed in terms of the pile radius a There is a considerable scatter in the points in this figure resulting largely from differences in soil type, the larger pore pressures being associated with the more sensitive sols. In the viein'ty of the pile, very high excess pore pres- sures are deveioped, in some cases approaching 1.5 to 2.0 times the in-situ vertical effective stress. Data presented by Aithart et al, (1969) suggests that near the pile tip, even greater pore pressures may be developed, amounting to 3 t0 4 times the insitu vertical effective stress. Beyond 7/a of about 4 for normal clays, and about 8 for sensitive clays, a rapid decrease in pore pressure with distance ocauts, and beyond [a = 30 the excess pore pressures are virtually negtiible. ‘A further consequence vf the development of pore pressures around a pile during driving has been reported by Fellenius and Broms (1969), who found thet sir icant negative frietion and downdrag forces were induced in a pile because of reconsolidation of the soil around the pile, This aspect is discussed further in Chapter 11 Estimation of Pore Pressures ‘A. umber of methods have been developed to predict, the excessporespressure distribution around a driven pile, For cases in which it is sufficient to estimate only the maximum pore pressure developed near the pile surface, Lo and Stermac (1965) derived an expression from the ‘consideration of failure of a radial zone of soil around the pile. D'Appolonia and Lambe (1971) derived an alternative form of Lo and Stermac’s equation, namely, en where uly, = maximum excess pore pressure K," = insite coefficient of earth pressre at ost u_= undead shear strength Ay ~ porepressure coefficient A at fallure Che inital vertical effective stres in soi Comparisons, seported by Lo and Stermac (1968) and, Lo (1968), between measured pore pressures and those calculated from Eq. (2.1), showed generally good agree- ment, Within the failure zone of the soil surrounding the pile, the pore pressures were a maximum and constant, and driving of adjacent piles only increased the pore pres- sure slightly, Outside the failure zone, the pore pressure decreased rapidly with distance and was negligible at a radial distance of about 16 diameters from the pile (see Fig. 2.2). Driving of adjacent piles developed pore pressures in this outer zone that added up directly until the m: ‘mum value was reached, Thus, the maximum pore pressures induced by driving a number of piles in a pile foundation may be predicted simply as the value of Auzq, from equa- tion (2.1). The data presented by Lo and Stermac (1965) suggests & radius of the failure zone of about 4-pile radii, Which is consistent with the extent of remolding around the pile, discussed in the previous section. ‘Theoretical methods of estimating the distribution of excess pore pressure with distance from the pile have been developed by Nishida (1962) and Ladanyi (1963). The former method is based on an elastoplastic analysis, while the latter is an adaption of the theory of expansion of a cylindrical cavity in a mass for use with the measured undrained stress-strain behavior of a soil. Although the latter method is versatile and relates to real soil behavior, it requires considerable computation. Furthermore, it relies fon the details of laboratory stress-strain curve, the accuracy ‘of which is liable to be affected by such factors as sampling disturbance and the inital stress sondition of the sample, |AS a rapid, practical means of estimating the excess. pore-pressure distribution, the following procedure is suggested:(@) The Lo and Stermac expression (Eq. 2.1) is used to ‘obtain the maximum pore pressure Au, from the face Of the pile to a distance R from the face. On the basis of Fig. 22 and also the analysis of Nishida (1962), R varies from 24 to 4a for insensitive clays, to 8¢ for sensitive clays. (b) Beyond the distance R, the excess pore pressure is ‘assumed to vary inversely as the square of the distance r from the pie, that is, su = ain (2) @2) ‘The inverse variation is predicted from clastic theory, as.utilized by Ladanyi (1963) and Nishida (1962). (©) For pile groups, the pore pressure distributions around Individual piles may be superposed, except that the pore pressure cannot exceed Au, as found by Lo and Stermac (1965). The excess pore pressures around 2 pile in sensitive clay as calculated by the above procedure, agree well with the average observed curve in Fig, 2.2. 2.2.3 Dissipation of Excess Pore Pressures A relatively simple solution for the rate of dissipation of excess pore pressures around a driven pile has been pro- posed by Soderberg (1962a). {tis assumed that dissipation ‘occurs radially only, the vertical dissipation that may occur EFFECTS OF INSTALLATION OF PILES 9 near the top and tip of the pile being ignored. The relevant equation of consolidation is then gol OG) co) ch is the two-dimensional coefficient of consolidation for horizontal drainage u_ isexcess pore pressure ‘The above equation may readily be written in finite difference form (€., #¢¢ Gibson and Lumb, 1953), and solved for.the appropriate drainage condition at the pile and initial pore-pressure distribution, Solutions for the excess pore-pressure dissipation at the pile face, for an impermeable ple, were obtained by Soderberg, who found that the form of the intial pore.pressure disttbution had «relatively small influence on this solution. ‘A reasonable measure of the rate of strength or adhe- sion-regan after driving appears to be to consider the rate of consolidation within a limited area inthe vicinity of the pile, Such solutions are shown in Fig. 2.3, assuming an initial excess pore-pressure distribution as suggested in Section 2.2.2 and a faluze zone having ratios of radius R to pile radiusa of 3 and 5. The degree of consolidation within a radius °F is shown for both a permeable and an impermeable pile tis interesting to compate these theoretical solutions with an empirical relationship suggested by Radugin (1969). oe 100 1900 o2t So taiure \ zone a) e/ 304 - osrek as o6| g oe Fl to FIGURE, 2.3 Theoretical solitons fot rate of consolidation near 2 driven pile.10 EFFECTS OF INSTALLATION OF PILES 0 200 02 aa Pecan a=6in) 7 os Degree of consnidation Op Rate of Adhesion Increase vl Empirical curve (Rewagin, 1569) FIGURE 2.4 Comparison between empirical and theoretical solutions for Assuming R/a = 5, cn = 0.04 sq in/min (typical of a medium clay), and 2 permeable concrete pile of Gin. radius, Fig. 2.4 compares the theoretical degree of consoll- dation versus time curve and Radugin’s empirical curve, assuming the rate of consolidation is the same as the rate (of adhesion increase in Radugin’s relationship. There is some difference between the shape of the curves, but they are generally in sufficient agreement to suggest that the simple consolidation analysis provides 3 reasonable ‘estimate of the rate of increase of load capacity. From a practical point of view, solutions such as those in Fig, 2,3 ae of most use in giving an estimate of the time that should elapse after driving before a load testis carried out, if a reliable estimate of the ultimate undrained load capacity and Joad-settlerment behavior is to be obtained, 'A more rigorous analysis of the stress changes, excess pore pressures, and subsequent consolidation around a driven pile in clay has been presented by Wroth et al (1979). The pile-driving process is modeled 8 the creation of a Jong cylindrical cavity by radial soil movement. Values of stress and pore-pressure change have been obtained using 2 finite-element analysis incorporating a work- ‘hardening soil model (the Cam-clay model). It is concluded that the total and effective stresses adjacent to the pile just after driving may be related directly to the original undrained strength of the soil, and are essentially indepen, dent of the overconsolidation ratio: The final stress state after consolidation is similar to that in an oedometer (Ko) test, except that the radial stress is now the major principal stress. tof adhesion increase. 2.24 Displacement Caused by Drivi Pile driving generally causes a heave of the clay surrounding the pile, followed by consolidation of the clay. This move- ment caused by pile driving may have a significant effect fon adjacent structures and may also cause the piles driven eailige in a multiplepile installation to rise during the siving of the lates piles. Under these circumstances, redriving of the earlier piles is often considered necessary, ‘or may Jead to a decision to use bored rather than driven, piles. The limited data available on the magnitude of the heave is rather conflicting, although much of the conflict may arise from differences in soil types in the various investigations. The ratio of the (otal volume of initial heave 10 the total volume of driven piles within a founda- tion has been found to be about 100% by Adams and Hanne (19703 for steel H-piles in a firm till, $0% for piles in clay by Hagerty and Peck (1971), 60% by Avery and Wilson (1950), and 30% by Orie and Broms (1967) for precast concrete piles in a soft, sensitive, silty clay. The Iatter investigators found that the heave near the edge of the foundation was about 40% of the value at the center Outside the edge of the group, only very small heaves ‘were noted by Adams and Hanna, and Orrje and Broms. ‘Adams and Hanna measured radial and tangential move- ‘ments as well as vertical heave, and found that the maxi mum radial movement was about 15 in., and the maximum tangential displacement about 0.4 in.-both these values being considerably less than the average vertical heave ofabout 4.5 in, As with vertical heave, very small lateral movements occurred beyond the edge of the group. Measurements of the-movement of an existing building 4s a result of driving of plles for the foundations of @ new building were reported by Lambe and Horn (1965). It was found that, at the near corners of the existing building, a heave of about 0.3 in. occurred during driving, but that by the end of construction, a net settlement of about 0.35 in. had occurred. Despite the fact that the piles were preaugered to within about 30 ft of their final ‘elevation, high excess pore-pressures (maximum of about 40 {t of water) were measured near the comer of the existing building, even before a substantial building load was carried by the ples Figure 2.5 shows some measurements of heave and settlement of buildings caused by pile driving, 2s reported by D’Appolonia and Lambe (1971). The settlement data plotted are for net settlement one to three years after the end of construction. Larger movements than those me sured by Lambe and Horn were found, although the piles ‘were again preaugered to within 20 to 30 ft of the final tip elevation. From measurements of displacements resulting from pile driving in clays, Hagerty and Peck (1971) concluded that the soil displacements are less for piles driven in EFFECTS OF INSTALLATION OF PILES. 11 sensitive clays than for corresponding piles in insensitive clays, and that when piles penetrate alternating strata of fine-grained soil and granular materials, the observed surfacesheave may be much less than that which would hhave occurred in insensitive clay soils. It was also found that if the sequence of pile driving involved driving piles first along the perimeter of the foundation, the heave of the soil surface in the central area of the foundation is increased and that of the surrounding area correspondingly decreased. Observations also were made of lateral move- ments, and it was found that driven piles tended to be displaced away from subsequent driving, with movements continuing for a considerable length of time after com: pletion of driving. Where large differences in elevation existed within the foundation area, pile driving often displaced the soil preferentially toward the areas in which the lower elevations occurred. Estimation of Displacements Lambe and Horn (1965) proposed method of estimating the heave and subsequent sectlemest of the surface of soil neat a pile resulting from driving of the pile. Although the method was found to predict movements considerably larger than those measured, it appears to be a logical pro: cedure and worthy of further application. The method Distance trom Neorest Pile Cop (Ht) 0 0 % aa) yoos * | elma lice oh lhe de 5 Sages geese ce eo : | : Senatatemerie sn Hl tt i mo) ei ow asi | Z € 008s | Relrigaranen 045 | i FIGURE 2.5 Moversens of nearby buildings caused by Pledriving operations (D'Appolonia ané Lambe, 1971).12 EEFECTS OF INSTALLATION OF PILES ettectn EE Sn (aise a step Deposit formation Steps tay Undroinad.loodng and consolation ° ‘doef usd’ 8s gopuesnel step 2 Pie deovng_ suiotion Step 3 Dissipation suosequent to. drivng FIGURE 246 Test proveduee for displacement ealelation (Lambe and How. 1968) (© Canada, 1965 by Univesity of Toronto Press.) Js based on the stiess-path approach advocated by Lambe (1964). It consists of estimating the effective stress-path for an “aversge” element in the compressible soi layer, running a laboratory test on a sample of this sol such that the loading folloxs the stess path estimated for the field element, and using the laboratory-measured value of vertical stain to estimate the building heave and settle- sent. There are essentially three steps in the procedure, ‘which is shown diagrammatically in Fig, 2.6 1. Consolidation of the sample (usually under K, con: ditions) 10 the in-situ stresses in the layer prior to pile ziving, Where the clement under consideration is beneath fan existing building, as it was in the case described by Lambe and Horn, this first stage imolves simulation of the lundrained loading caused by constrection of the building followed by consolidation, after initial K, consolidation ‘of the element to the field stress state. 2. Increasing the total lateral stress at constant total vertical stress until the pore pressure in the element equals that measured by the field piezometers or that calculated fon the basis of Section 2.2.2. The vertical strin measured | paction around the pies was studied by means of radio- raphy techniques, 1t was found that in an initially very loose sand (telative density D, = 17%), soil movement extended 3 t04 pile diameters from the side ofthe pie and 2.5 (0 3.5 diameters below the ple tip. In a medium dense | + 6 03040 Varta! sine FIGURE 2.8 Displacements atound driven pile in sand (after Robinsky and Morzison, 1964). (Reproduced by peemisson of the National Research Council of Canada from the Canadian Geo- technical Journal, Vol. 1, 1968, p. 81.) FIGURE 2.9 Stains around deivon ple in sand (after Robinsky and Morrizon, 1968), (Reproduced by permission of the National Reseach Council of Canada from the Canadian Geotechnical Jour nal, Vol. 1, 1964, p. 81.)14 EFFECTS OF INSTALLATION OF PILES. Observey Rests: Estimated Rosusargie of treten &* O ‘Mopr perce! stress rot = =O Test Nom,Woreaster, Mass Penetration Resstance 0 FIGURE 2,39 Obseccd vs. computed compaction of sand near pile (after Meyerhof, 1959). around a pile driven in sand; atypical result compared with observitions is shown’ in Fig, 2.10. These results are in broad agreement with those of Rebinsky and Morrison (1964), but according to Meyethof, the amount of com- paction near the tip is greater, and that near the top of the shaft is less ‘A simpler method of estimating the effects of driving ‘pile in loose sand in the vicinity of the tip is that pro- posed by Kishida (1967). On the basis of field and mode! test-esults, he assumes that the diameter ofthe compacted zone around a pile is 7. Within this zone, he further assumes that the angle of friction ¢! changes linearly with distance from the original value of 6 at a radius r = 3.50 to a maximum value of ¢3 at the pile tip, as shown in Fig. 21 ‘The relationship between 6," and gs" is taken to be y= Hit 40" . 2 6) FIGURE 144 Etlect of diving on ¢ (Kishi, 1967), When ¢, = 40° in Eq, (2.6), there is no change in relative density dye to pile driving. 232 Pite Groups When groups of piles are driven into a loose sand, the soil around and between the piles becomes highly compacted, and if the ple spacing is sufficiently close (less than about Sx daniel), the ultimate load capacity of the group may be greater than the sum of the capacities of the indi vidual piles—that is, the efficiency of the group is greater than 1. On the other hand, if the sand is so dense that pile driving causes loosening rather than compaction, the group efficiency may be less than 1 ‘An estimate of the effects of driving a gsoup of piles into loose sand may be made by application of the approach suggested by Kishida (1967) for single piles, assuming that superposition of the effects of single piles is applicable. tn applying Eq. (2.6) the value of @ is the changed value caused by previous piles. By application of this approach, a rough estimate may be made of the effect on ultimate load capacity of the order of installation of the piles. It has been found in practice that piles driven later have a great Woad capacity than those driven earlier. Some field measurements of the amount of compc- tion caused by the driving of a group in a granular sol, in which standard penetration tests have been cattied out before and after driving of groups have been reported by Philcox (1962). The (est results ate shown in Fig. 2.12. In case (2), the standard penetration number, WV, neat the center of a foursile group, was more than doubled by driving. In case (b), the inetease in N for a point a little away from the center ofa ninepile group shows a relatively smaller increase (average about. 75%). Cases (6) and (4) show that the increase in NV becomes less as the point considered becomes moro distant from the center of the g10up. Another point apparent ftom Figs, 2.12 and 2.12¢ is that the effeet on driving on N, and hence on soil density, is greater below the tip than along the shaft, Im order to gelate the increase in 1V to the increase in 9, Kishida (1967) suggests that # and NV may be related by the following expression: = JRO 1s" en ‘The differences between the degree of densification at various points within a group, as shown in Fig, 2.12, suggest that the load capacity of piles near the center of the group may be greater than those eas the edge of the group, and that the load distribution, even at working, loads, may be-nonuniform, with lanjer loads being carried by the center piles-as predicted by Kishida’s approach.EEFECTS OF INSTALLATION OF PILES 13, ow mw | eo T i : Fed : z Fase oe a Tas seen Toit Siem Tao) Fen secan TAO™ 307 meri antl ave. Rowan Kerberiey Rd austin dag RoW FIGURE 2.12 A comparison of W Values before and after diving piles Phi Structural Engineers, London.) 1962), (Reproduced by permission of the Institution of ‘This behavior, which is in contrast to that usually doverved the soil. The adhesion has been found to be less than the for groups in clay, has been observed in tests carried out by _undrained cohesion before installation, mainly ‘iecause of Hanna (1963) and Beredugo (1566). As suggested by softening, of the clay immediately adjacent te the soil Kishida (1967), the effects of differing compaction may surface. This softening may arise from three causes: also explain the dependence of pileoad distribution on the order of driving piles in sand (@) “Absorption of moisture from the wet concrete (©) Migration of the water ftom the boty of the clay toward the less highly stressed zone around the borehole. edeencne Oc uer atin BOREL Eee (©) Water poured into the boring to facilitate operation 24.1 Clay Soils of the cutting too. ‘The effects of installing bored piles in clay have been stud- Factor (¢) may be eliminated by good drilling technique, ied largely in relation to the adhesion between the pile and and (b) can be minimized by carrying out the drilling and16. EFFECTS OF INSTALLATION OF PILES +0000 000 cox S000 coo é Bae 1000 Feadgton 00 Reva Green Erber soo Fry soa Sauna i Veta Gateay st Sta a a FIGURE 2.13 Relation between shear strength and wa concreting operations as rapidly as possible. Some effects from factors (2) and (b) are considered to be inevitable by ‘Skempton (1959), but their seriousness will depend largely ton the technique employed, whether or not casing or ‘driling fluid is used to support the walls during construc- tion, and the time taken for construction of the pile Palmer and Holland (1966) contend that softening in ver- consolidated London clay is minimized if drilling and con- creting is cartied out within one or two hours. Meyethof and Murdock (1953) measured the water contents of the clay immediately adjacent to the shaft of| a bored pile in London clay and found an increase of nearly 4% at the cantact surface, although at a distance of 3 in, from the shaft, the water contents had not altered. ‘This inerease should be a maximum value, as the hole was drilled by hand and took two to three days to complete ‘An estimate of the effect of the increase in water content ‘ean be made if information is available on the relation between shear strength and water content. Such a relation- ship for London clay has been presented by Skempton (1959) and is reproduced in Fig, 2.13. These results show that an increase in water content of only 1% results in a 20% change in the ratio eafcy, of undrained adhesion q t0 ofiginal undrained strength cy, while fora 4% increase jn, water content, ca/ey is reduced to about 0.3. Values of Caley for bored piles are discussed in detail in Section 32. mF aa ater content, 5 ‘content for London cay (L3quié Limit, 70-85} (Skempton, 1958). {A further effect of installing bored pile is that the clay just beneath the pile base may be disturbed and softened by the action of the boring tools. The effects of this distuxbance may result in increased settlements, espe- cially for belled piers, in which the base caries a major proportion of the load; hence, it is important to clean out the base thoroughly. However, as stated by Skempton (1959), base disturbance and softening should have a negligible effect on the ultimate bearing capacity of the base becasse of the comparatively large mass of clay involved when the base penetrates the clay. In contrast, the shearing process developed in the clay along the pile shaft is probably restricted to the narrow softened zone. Construction problems may also arise with bored piles, and a number of these have been described by Pandey (1967) in relation to the foundations for a heavy industrial building, including the Following (a) Caving of the boretiole, resulting in necking or misalign. rant of the pile. (b) Aggregate separation within the pile. (©) Buckling of the pile reinforcement, Such structural defects may be difficult to detect, since a load test may not reveal any abnormal behavior, especially if the load is only taken to the design load. Barker and Reese (1970) investigated the influence of | drilling fluids on the performance of bored piers. Theyconcluded that when proper construction techniques are employed, dsiling mud has no detrimental effects on the load-carrying characteristics of a borgd shalt. The con: crete properties and the concrete-placement procedure are the two most critical factors involved in the construc- tion process. Elimination of the effects of the drilling mud is accomplished when it is completely displaced by the concrete, seaulting in a vigorous scouring of the bore. hole wall by the rising concrete. Should drilling mud be tapped between the conerete and borehole wall, it would virtually climinate the development of any shear-load transfer in the vicinity of the trapped drilling mud, The use of casing in placing the conerete involves a greater risk of trapping drilling mud than does the procedure of placing. the concrete under the drilling mud by the use.of a tremie fr concrete pump. Barker and Reese suggested that reduc- tion factors of 0.6 for clay and 0.8 for sand and silts should be applied to the shear strength in the design of drilled shafts. However, no reliance should be placed on load transfer developing within three shaft diameters of the surface or one diameter of the base 24.2 Sands ‘There is relatively little quantitative information on the effects of installing bored piles in sands or cohesionless soils. Such piles usually requite casing or drilling fluid 10 support the walls of the hole and sinking of the hote, and subsequent withdrawal of the casing while coneceting the shafts likely to disturb and loosen the soil to some extent, Also, some loosening is liable to occur at the ‘bottom of the pile 2s a result of baling or “shelling-out” the hole, and when this is done wader water, the upward surge on withdrawal of the baler or shell can loosen the EEFECTSOF INSTALLATION OF PILES. 17 soil for several feet below and around the pile, ‘Thus, in calculating the load capacity of a bored pile in sand, Tomlinson (1975) suggests that the ultimate value of angle of shearing resistance @ should be used, unless the pile is formed in a dense gravet when the “surging” effect may not take place. If heavy compaction can be given to the concrete at the base of the piles, then the disturbed and loosened soil may be recompacted and the value of # for the dense state used, However, if the shaft is dstnscted by the reinforcing cage, such compaction may not be possible Tests on bored piles in sand have been reported by Touma and Reese (1974) and Clemence and Brumund (1975). Touma and Reese found evidence of the arching that occurs around the pile with driven piles (see section 3.2) and that results in the development of limiting values of skin friction and base resistance at depth. It was also found that the skin resistance, for piles penetrating tess than 25 ft, could be correlated with the integral around the pile periphery of 0), tan @' (where oy = effective over burden pressure), using a reduction factor of about 0.7 ‘There were indications that smaller reduction factors are appropriate for greater penetcations, From a largescale test on a bored pier in sand bearing on a simulated rock. base, Clemence and Brumund (1975) found that 20 to 30% of the design axial load in “‘end-bearing” drilled piers was carried by the pile skin. A roughly linear increase in skin {tiotion with depth was measured, except near the lower part of the pier, where a shacp increase in skin friction was noted, presumably because of the confining effect of the rock base, It was found possible to use the results of direct shear tests for the soikpier interface materials to predict the limiting skin friction, except near the tip, where the calculated skin frietion was lower than that measured,ULTIMATE LOAD CAPACITY OF PILES 3.1 INTRODUCTION There are two usual approaches to the calculation of the ultimate load capacity of piles: the “static” approach, which uses the normal soil-mechanies method 10 calculate the load capacity from measured soil properties; and the “dynamic” approach, which estimates the load capacity of driven piles from analysis of pile-driving data. The first approach will be described in detail in this chapter and the second in Chapter 4, Jn this chapter, a general expression for the ultimate load capacity of a single pile is given and its application to piles in'clay and sand is desesibed. Approaches for groups of piles in clay and sand will then be ontlined. Other topics include the design of piles to rock, the use of insitu tests such as the standaed penetration test and the static cone to estimate pile-load capacity, the caleulation of uplift resistance of piles and groups, and the load capacity of bent piles 3.2 ULTIMATE LOAD CAPACITY OF SINGLE PILES 3.2.1 General Expression Fhe net ultimate load capacity, Py, of a. single pile is ‘generally accepted to be equal 10 the sum of the ultimate 18 shaft and base resistances, less the weight of the pile; that Pu = Poy + Pou -W Gy where Poy = ultimate shaft resistance Poy = ultimate base resistance Wo = weight of pile Poy can be evaluated by integration of the pile-soil shear strength rq over the surface area of the shaft. te isiven by the Coulomb expression Ta = Ga +p t49 by 62 where ‘re = pile-soil shear strength adhesion normal stress between pile and soil = angle of friction between pile and soil + tt is an Implicit assumption of Eq. 3.1 that shaft and base Sucty correct, but there i htc dovibt that It is eoreot enoughoq is in turn frequently related to the vertical stress oy, as an = Keay Ga) where Ke = coefficient of lateral pressure Thus, re © cat oyKs tan by G4) and fa = SE Cre . Sh Clea ayKy tan tole (35) C = pile perimeter L. = length of pile shaft It Is usually accepted that the ultimate resistance Phy ccan be evaluated fiom bearing-capacity theory as Pou = ApleNe + a,5q + 05yaNy) G6) where Ay = area of pile base e = cohesion of soil = vertical stressin sail at level of pile base tunit weight of soil pile diameter = bearing capacity of factors, which are primarily functions of the angle of intevnal friction @ of the soil, the relative compressibility of the soil and the pile geometry Rigorous solutions for the bearing capacity of surface footings using the methods of classical plasticity are now wellestablished (Prandil,, 1923; Sokolovskil, 1965; Cox, 1962; Lundgren and Mortensen, 1953; Davis and Booker, 1971), and the only doubts regarding the practical validity of these solations ie in the possible effects of the differ ‘ences between the behavior of real soil and that of the ideal ‘material assumed in the theory. At the present time, there are few if any classical plasticity solutions that are relevant to a buried footing, and therefore, for the calculation of ULTIMATE LOAD CAPACITY OF PILES 19 base resistance of piles, reliance has tobe placed oa approx: imate theoretical or semiempirical methods. With regard to: sands, these methods have been reviewed by Vesic (1967), who found that the solution of Borezantzev et (1961) generally fitted experimental results best, From Eqs. 3.1), (3.5), and (3.6), Py = fh leg +oyK stan 4) dt + AgleNe + 0n5Ng + 0S7dN,)- W an Equation (3.7) is a general expression for the ultimate Youd capacity of a single pile, If the undrained or short= term ultimate Toad capacity is 10 be computed, the soll parameters ¢, %, ¢, and 7 should be values appropriate to undrained conditions, and oy and 0,5 should be the total stesses. If the long-term ultimate load capacity of piles in sand is required, the soil parameters should be drained values, and of and o}p the effective vestical stresses, The vertical stresses are usually taken to be the overburden stresses, and for clays, this is probably true enough, even close to the piles, However, fr sands, there is now cleot evidence implying that the vertical stress near the pile may be less than the overburden. This matter is dis cussed in greater detail in Section 3.23. For steel Hpiles, two modes of failure of the shaft are possible: (a) the development of the limiting pile soil shear stsength along the entire surface area of the pile; and (b) the development of the limiting pile-soil shear strength along the outer parts of the flanges, plus the development of the full shear strength of the soil along the plane joining the tips of the flanges—that is, the soil within the outer boundaries of the pile effectively forms part of the pile shaft, Therefore, when using Fa. (3.7), the ultimate skin cesistance, Pay, should be taken as the lescer of the two values 3.2.2 Piles in Clay 4.221 UNDRAINED LOAD CAPACITY For piles in clay, the undrained load capacity is generally taken to be the ctitical value unless the clay is highly coverconsolidated. (Burland, 1973, contends, however, that an effective tress-drained analysis is more appropriate, as the rate of pore-pressure dissipation is $0 rapid that for normal rates of load application, drained conditions generally preval in the soil near the pile shaft) Ifthe clay is saturated, the undrained angle of friction oy is 2210, and 4 may also be taken as zer0. In addition, Ng = 1 and N= 0 for 9= 0,50 that Eq, (3.7) reduces to20. ULTIMATE LOAD CAPACITY OF PILES y= fl cate rAsleaNetog)-¥ whee ¢, = undrained cohesion of soil at level of pile base undrained pile-soil adhesion Forther simplification is possible in many cases, since for piles without an enlarged base, Ayo, = W, in which case P= Sh ceade + Ascude G10) Undrained Pile Soil Adhesion ce ‘The undrained pile-soil adhesion cg varies considerably ‘with many factors, including pile type, soil type, and method of installation. Ideally, cy for a given pile at a given site should be determined from a pile-loading test, but since this is not always possible, resort must often be rade to empirical values of cq. Many attempts have been rade {0 correlate ce with undrained cohesion cy, notably ‘Tomlinson (1957, 1970), Peck (1958), Woodward et al (1961), Coyle and Reese (1966), Vesic (1967), Morgan and Poulos (1968), MeClelland et al. (1969), MeCleltand (1972), and McClelland (1974). For driven piles, typical relationships between colew and ¢y, based on the summary provided by McClelland (1974), are shown in Fig. 3.1. It is generally agreed that for soft clays (cy < 24 KPa), calc, is 1 (or even greater); however, for driven piles in stiff clays, a wide scatter of ‘TABLE 3.1 DESIGN VALUES OF ADHESION FACTORS FOR PILES DRIVEN INTO STIFF COHESIVE SOILS? Peneteation Case __Soll Conditions Ratio? aleu 1 Sands or sandy sols <20, 125 overlying tit cohesive sails >00 See Fig. 3.2 MW Soft lays or ses <200>8) 040 ‘overlying sift cohesive sie > 070 ML Suffcohesive soils <20(>8) 040 out overlying stata 220 See Fig. 33 After Fominson (1979). ene Depth of penetsstion Insti elay NOTE 1+ Adhesion fctore not applicable to H-section pile. NOTE 2: Shaft adnesion in overburden soit for cases | and most be calculated separstely ‘apid dissipation of excess pore pressures ‘due to driving may result i 2 locally overconsolidated condition, land hence a value of cy even greater than cy forthe unaffected so = 25 50 15 1125 150. 18 T sh 4 ook 4 oa 4 2p 7 L L L 1 L L 1 05 0 15 35 25 30 Undies Coton knee FIGURE 4 Adhesion factors for driven piles in clay (after MeCletland, 1974).ULTIMATE LOAD CAPACITY OF PILES 21 ners rear sera et &Nn? sn 7 “ en a T eo’ T ee on Pedometer a z Meal? 239 0 Precast concrete 310 J ° 1000 Undid en oe \vedrton ear ae 180 20 os 1000 72000 30 Lg 2000) roa et FIGURE 3.2 Adhesion futors for cat (sands and gravels overlying stiff to very sif cohesive sil) (Tomlinson, 1970). rng Ie) kN? 200 20 Prec oonerete = Design curve for 5 senensnenesia 370] 5000 Lndrsine ear stent ey? FIGURE 3.3 Adhesion factors for cas I (sito very stif clays without overlying stata) (Tomlinson, 1970). values of Cafey is evident. This scatter is often attributed to the effects of “whip” during driving. A sxe complete investigation of adhesion for driven piles in stiff clay has ‘been made by Tomlinson (1970), who found that ¢a/cy may be markedly influenced by the sol stata overlying the clay, as well as by the value of. Tomlinson has sug- gested the adhesion factors shown in Table 3.1 and Figs, 3.2 and 3.3 for cy > 1000 Ibjsq ft (48 KPa). The most notable feature of Tomlinson’s results are the hgh values of “a/ey for case I, where sand or sandy gravel overlies the clay, because of the “carrying down” of a skin of the ‘overlying sol into the clay by the pile. There appears to be little data on appropriate values of ¢g for driven piles founded in very sensitive clays, and the extent to which “setup” compensates for remolding can at present only be determined by a load test For bored piles, the available data on “ajc, is not as extensive as for driven piles, and much of the data that is available is related to London clay: Table 3.2 gives a sum rary of adhesion factors, ante of which 1s expressed it terms of remolded strength, ¢r, a5 well as the undisturbed uundrained strength, cy. Results obtained by Skempton (1959) and Meyerhof and Murdock (1953) suggest that Aan upper limit of ¢ is 2000 Ib/sq ft (96 kPa).22. ULTIMATE LOAD CAPACITY OF PILES ‘TABLE 32 ADHESION FACTORS FOR BORED PILES IN CLAY Altnesion Soit type Factor Value Reference London city ally 025-0.7 Golder and Leonard ‘Average, — (1954) 04s ‘Tomlinson (1987) ‘Skempton (1959) Sensitive clay ele 1 Golder (1987) Highly expansive ley OS Mohan and Chanda lay a6 0 o1 02 o 2 125] ns FIGURE 54 Feetinal capacty coeffielent 2 #5, pile pe sion (Vijayvergiya and Foe A somewhat different approach to the calculation of the ultimate shaft capacity P,y has been adopted by Vijay- vergiya and Focht (1972) for steel-pipe piles. From an examination of a number of loading tests on such piles, they concluded that P,,, can he expressed as follows: Pog = OIm + 2em)As 6.100) fm = mean effective vertcal“stess between ground surface and ple tip erage undrained shear strength along pile, 03. os os tecston soc owan 2 us fons aa Giwnnd Gee 4 ems 3 Sonne Tomton Nevorems 2pm” venie o teStlong ite 3 Wectenee Dertomite ba MSC Houston To cic is Sontineics + See emt Estron 2 Mannon (aes ueeeee 1972),theron Ne Ne (Rectangle) Lp > Bel au [ovoero" v0] po | , aoe FIOURE 3.5 Dearinz-capacity factors for foundations in eay (@ = 10) (after Skemp, 1951), Ay = pile surface area 2 = dimensionless coefficient In effect, the average pile-soil adhesion factor is then dg = aft + 2) 6.108) A was found to be a function of pile penteation and is potted in ig. 3.4 Equation (3102) has been used extensively to predit the shaft capacity of heavily faded pie piles for offshore stectues Bearing Capacity Factor Ne ‘The value of We usually used in design is that proposed by ‘Skempton (1951) for a circular area, which increases from 6.14 for a surface foundation to @ limiting value of 9 for length 2 4 diameters (Fig. 3.5). The latter value of Ne = 9 has been confirmed in tests in London clay (Skempton, 1959) and has been widely accepted in practice. However, differing values have been found by other investigators; {or exainple, Sowers (1961) has found 5 < Ne <8 for model tests, and Mohan (1961) has found 5.7 < Ne < 8.2 for expansive clays. The variations in the value of ‘Ne may well be associated with the influence of the stress- ULTIMATE LOAD CAPACITY OF PILES 23 strain behavior of the soil. From an analysis of the expan- sion of a cavity in a mass, Ladanyi (1963) found that for insensitive clays, 14 << N. < 9.3, depending on the stress-strain behavior of the stil. This analysis broadly ‘confirmed the earlier analysis of Bishop et al, (1945), which gave the following result for a circular base (as quoted ‘oy Ladanyi), Neo +h a o(&)) 3.2.22 DRAINED LOAD CAPACITY Guy For piles in stiff, overconsolidated clays, the drained loud capacity, rather than the undrained, may be the critical value, and Vesic (1967, 1969) and Chandler (1966, 1968) have advocated an effectivesteess approach in such cases, If the simplifying assumption is made that the drained pile-soil adhesion <, is zero and that the tenus in ke, (3.7) involving the bearing capacity factors N- and N, can be ignored, the drained ultimate load capacity from Ec, (3.7) may be expressed as = [hcoue, anes " = Poot, tn 6 a (13) effective vertical stress at depth = ‘effective vertical stress at level of pile base = drained angle of friction between pile and soil Burland (1973) discusses appropriate values of the combined parameter 8 = K, tan §y and demonstrates tht 4 Tower limit for this factor for normaly consolidated clay canbe given #8, B= (1-sin@’) tan ¢! G.3) where (gf = effective stess friction angle for the clay For values of 9! in the range of 20 to 30 degrees, Ea (3.13) shows that 8 varies only between 0.24 and 0.29. ‘This range of values is consistent with values of = Ky tan 6% deduced from measurements of negative fiiction on piles in soft clay (see Figs. 11.26 and 11.27). Meyethof (1976) also presents date that suggests similar values of A; however, there is some data to suggest that f decreases24 ULTIMATE LOAD CAPACITY OF PILES Avcage shin esitance (Hsin? o 4 2 Fae 5 20 0 Flot too, mei nant a Stediom dens sand 1G [ieee ee verge skin estan tose?) v0 07 04 FIGURE 36 Variation of skin 1esistance with pile length (Veli, 1967, \ with increasing pile length, and that for long piles (in Xcess of about 60m), 8 could be as ow as0.15, For piles in stiff clays, Burland suggests that taking K, = Ko and oj = the remolded friction angle, gives an upper limit to the skin friction for bored piles and lower limit for driven piles, Meyethof (1976) presents data indicating that K, for driven piles in stiff clay is about 15 times Ko, while K, for bored piles is about half the value for ériven piles, For overconsolidated soils, Kp can be approximately estimated as Ky = (1-sin@)VOCR where OCR = overconsolidation ratio It is infered that 9 can be taken as tion angle of the clay Im the absence of contrary data, 0} and ofy may be taken as the effective vertical overburden stresses, Values of Nig muy be taken to be the samme as for piles in sand; these values ae ploted in Fig, 3:12 G.14) the drained fri 3.23 Piles in Sand Conventional methods of calculation of the ultimate load capacity of piles in sand (Broms, 1966; Nordlund, 1963) Pot eine isin ®) NE oR a lt safc a oe res te a cten 4 at powers) o og . NS so —4 tooo tne) (can os \ vot HL Jicose [Medium dense sand \ vet [ee ‘ \ vf fi ia | ‘ \ ] : wh, Joi |, ; cae ee eee Point resistance [omit FIGURE 3.7 Variation of point resistance with pile length (ese, 1967).assume that the vertical stresses oy and oy in Eq. (3.7) are the effective vertical stresses caused by overburden, However, extensive research by Vesic (1967) and Kerisel (1961) has revealed that the unit shaft and base resistances of pile do not necessarily increase linearly with depth, ‘but instead reach almost constant values beyond a certain depth (Figs. 3.6 and 3.7). These characteristics have been confirmed by subsequent research (eg., BCP Comm. 1971; Hanna and Tan, 1973). Vesic also found that the ratio of the limiting unit point and shaft resistances, f/f, of a pile at depth in a homogeneous soilanass appears to be independent of pile size, and is a function of relative density of the sand and method of installation of the piles. Relationships between foif, and angle of internal friction (¢'), obtained by Vesic, are shown in Fig. 3.8 ‘The above research indicates that the vertical effective stress adjacent to the pile is not necessarily equal to the effective overburden pressure, but reaches a limiting value at depth. This phenomenon was attributed by Vesic to arching and is similar to that considered by Tereaghi (1943) in relation to tunnels, There are however other hypotheses, such as arching in @ horizontal plane, which ‘might explain the phenomena shown in Figs. 3.6 and 3.7 400 300} 100|-— a0} — 2p —— 4 whet | | FIGURE 38 Variation of foifg with ¢ (Vesie, 1967) ULTIMATE LOAD CAPACITY OF PILES 25 FIGURE 3.9 Simplified distribution of yenieal stress adjacent to pile ia san. Sone design approaches have effectively incorporated Vesi’s findings by specifying an upper limit to the shaft and base resistances. Eor example, McClelland et al, (1969), have suggested, for medium-dense clean sand_the following design parameters: ¢ = 30°; Ky = 0.7 (compression loads) of 05 (tension loads), with a maximum value of shaft resistance f, of } ton{ft? (9€ kN/m*); and Ny = 41, with a maximum base resistance fs of 100 ton/ft? (9.6 W/m?) However, such approaches take litte account of the nature of the sand and may not accurately reflect the variation of pile capacity with pile penetration, as the limiting resistances generally will only become operative at relative ly large penetrations (of the order of 30 to 40m). In order to develop a method of ultimate load pre. diction that better represents the physical reality than the conventional approaches, and yet is not excessively complicated, an idealized distribution of effective vertical stress o}, with depth adjacent to a pile is shown in Fig 3.9. 0} is assumed to be equal to the overburden pressure to some critical depth ze, beyond which a remains con. stant. The use of this idealized distribution, although simplified, leads to the two main characteristics of behavior observed by Vesic: namely, that the average uiimate skin resistance and the ultimate base resistance become con- stant beyond a certain depth of penetration. If the pile-soil adhesion ¢, and the term eNe are taken as ze70 in Bq, (3 7), and the term O.S1d N, is neglected as being small in zelation to the term involving Ng, the ulti mate load capacity of a single pile in sand may be expressed as follows: Pu = [EP uCouKs tan eae + Apoiag-W B15) where26 ULTIMATE LOAD CAPACITY OF PILES effective vertical stress along shaft effective wverburden stress for z ze effective vertical stress at level of pile base cortection factor for tapered pile (= 1 for uniform diameter pile) (On the basis of the test results of Vesic (1967), values of Ky tan $4 and the dimensionless critical depth z-/4 have been evaluated. Vesic’s results are presented in terms of the relative density D, of the sand, but the results may also be expressed in tems of the angle of internal friction @', by using a relationship such as that suggested by Meyer: bof (1956) gf = 8 + 15D, G16) Relationships between K, tag and g', and zejg and g', are shown in Fig. 3.10. In a layered-soil profile, the critical depth 2. refers to the position of the pile embedded in the sand. It should be emphasized that these relation- ships may be subject to amendment in the ight of further test results. For example, at present, the dependence of K, tang) on the pile material is not defined, Vesic's tests ‘were casted out on steel tube piles, but the values of Ky tangy derived from these tests appear to be applicable to cther pile materials. However, in the light of future test 1 For erivan plas ® = %4 oe For bored ites. =O, -3 results, it may be possible to derive different relationships for different pile materials. For bored or jacked piles, the values of Ky tan 9 in Fig. 3.105 are considered to be far too large, and it is sug. gested that values derived from the data of Meyerhof (1976) are more appropriate for design. These values are shown in Fig. 3.106, and have been obtained by assuming 0.756'. The values for bored piles are reasonably consistent with, although more conservative than, those recommended by Reese, Touma, and O'Neill (1976). Also shown are values of K, tan 9 for driven piles, derived from Meyerhofs data; these later values are considerably smaller (typically about one half) of the values given in Fig, 3.108, Some of this difference may lie ia the method of interpretation of the data of Vesic and others by Meyer- hof, which leads to smaller values of K, tan $4 associated With lager values of fd The bearing capacity factor Ng is plotted against $ in Fig. 3.11, these values being based on those derived by Berezanteev et al. (1961), Vesic (3967) has pointed out that there is a great vaviation in theoretical values of Ng derived by different investigators, but the values of Berezantzey etal. appear to fit the available test data best. ‘The solutions given by Berezantzey et a. indicate only a small effect of relative embedment depth L/d, and the curve in Fig. 3.11 represents an average ofthis small range. ‘The cures given by Moyerhof (1976) show a larger effect 10 (Fig. 3.100, Fig. 3.108) (Fig.3100), @s@; (Fig 3100) where Dis angle of internal miction prior to fel eg/aws @ (2) Ky ton @y ve @ (Griven Pies) Ce) Valves of Kater Oy Bored ‘ Meyserot Be) oo Ea ae 10] zof & as at 15 b- 1 A 1 at Bete bed FIGURE 3.10 Values of ld and K, tan a for piles in sand,140 For oriven piles, @ For bored piles, B= 2-3 where @; «angle of ternal friction prior to metalation of pie 1000 Na 100 rr 5 FIGURE 3.11 Relationship between Ng and ¢ (atter Berezanteoy tal, 1961). of L/d; however, the curve of Fig, 3.11 also lies near the middle of Meyerhofs range Values of the taper correction factor Fy are plotted against ¢ in Fig, 3.12 and have been derived from the re- sults of the analysis developed by Nordlund (1963). Im applying the results in Fig. 3.10 to Fig, 3.12, its suggested that the following values of ¢ be used to allow for the effects of pile installation —1__1_] ° oF 16 15 20 Pl taper arate ot oper factor Fy (afer Nowdlund, 1963). FIGURE 3.12 ULTIMATE LOAD CAPACITY OF PILES. 27 {a} Driven Piles (@) For the determination of Nz, the values of @ beneath the pile top should be taken asthe final value subsequent to driving, as given by Kishida (1967) on fit a7) where oh ‘ernal friction prior to installation of (8) For the determination of Ky tan 95 and zc/d, the value of @ slang the pile shaft should be taken as the mean of the values prior to, and subsequent to, driving; that is, Be t 9 = Fe + 10 8.18) (b) Bored Pies (@ For the determination of Ng and ze/d, i is suggested that the value of ¢ be taken a4 - 3, 10 allow for the possible loosening effect of installation (see Section 2.4) (0) For Ky tan 95, Fig, 3.10e should be used, taking the value of 4 directly. The above suggestions may also require modification in the light of future investigations. Furthermore, if jeting, is used in conjunction with driving, the shaft esistance nay decrease dramatically and be even less than the vaine for corresponding bored pile McClelland (1974) has reported tests in which the use of jeting with external return flow followed by driving reduced the ultimate shaft capacity by about 0%, wile installation by jetting alone reduced the ultimate shaft capacity to only about 10% of the value for a ple installed by driving only ‘Another case in which caution should be exercised is when piles are to be installed in calcareous sands. Such sands may show friction angles of 38° or more, but have been found to provide vastly inferior support for driven piles than normal silica sands. in such cases, McClelland (1974) suggests limiting the skin resistance to'02 tons/ft? (19 kN/m#) and base resistance to 50 tongtt? (4800 kN/m?). In such cigcumstanees, drilled and grouted piles nay provide e° more satisfactory solution than wholly driven ples, In many practical cases, only standard penetration- test data may be available, The value of 4 may be esti-28 ULTIMATE LOAD CAPACITY OF PILES nia of internal tretien prior to mstaiation of ils where. Re Values of L/e / 2° ot 2 ao aa FIGURE 3.13 Dimensionless ultimate shatluad capacity for pile in unifarm ne mated from a correlation such as that given by Peck, Hansen, and Thorburn (1974), or by the following em- pirical relationship suggested by Kishida (1967) G.19) where WN = standard penetration number A. more detaled discussion of the relationship between 24 and Ni, and also 6', and relative density D,, is given by de Mello (1971). For the case of a driven pile in a uniform layer of sand, dimensionless values of the ultimate shaft load and ultimate base load may be derived using Eq. (3.15) and Figs. 3.10, 3.11, and 3.12. la Fig. 3.13, the dimensionless ultimate sheft load Py, (7d is plotted against ¢ for various values of L/d; 7 is the effective anit weight of the soil above the critical depth zc. The marked increase in ultimate shaft load with increasing L/d and ¢ is clearly shown. The dimensionless ultimate base load Pyy/yd4y Is plotted 8-40 —— For berad pies, 0 +8}, -3 For ewan pies, ner riot to rsystetion of pile Fea, woo La e388 a ae FIGURE 3.14 Dimensionless ultimate baseload capacity for pile in uniform Sand against ¢ in Fig. 3.14, The value of L/d does not generally have a marked effect on the ultimate bast load unless 6 i celatively large, that is, for dense sands, ‘The use of a high value of @ for very dense sands (say, > 40°) simultaneously for the shaft and the base, should ‘so be treated with caution, since the full base cesistance may well only be mobilized after @ movement sufficient for the operative value of 9 along the shaft to be signif cantly less than the peal If the pile is founded in a relatively thin, firm stratum underlain by a weaker layer, the ultimate base load may be governed by the resistance of the pile to punching into ‘he weaker soll. Meyerhof (1976) shows that if the weaker layer is situated less than about £0 base diameters below the base, a reduction in base cepacity can be expected, he suggests that in such cases, the ultimate point resistance ‘can be taken to decrease linearly from the value at 10d above the weaker layer to the value at the surface of the weaker layer. ‘The suggested epproach of ultimate load calculation bas been applied to 43 reported load-tests on driven piles. The details of the parameters chosen for the calculations are given in Table 3.3, and the comparison between cal- culated and measured ultimate loads is shown in Fig. 3.15 ‘The mean ratio of calculated to observed ultimate loads is 0.98, with a standard deviation of 0.21. It should be pointed out that the ultimate load of all ples considezed in the comparison is less than 300 tons. The use of thisTest reported by Nordlund (1869) ‘Power plant site-stes it Mojave River bniae % Por Mena 9 OS views Fo test reported by Ves (1867) “S Uitimate tin fonds 1 Utemate base loads 8 ls tot eeprte by Dest (1979) lal ere by Tavs (1970) # Coteutated 9, U8 tne) a 8 $8 $8 8 8 s 2000 Messred (U.S. tos) FIGURE 3.15 Compasison between calculated and measured ulti ‘mate Toad capacity of driven ples in sand, approach for ples of much larger capcity-those used for offshore. structures for example-should be treated With caution. Indeed, for relately short, lage-diameter piles, the average values of shaft resistance given by thi foproach ae ‘considerably larger than thote normally opted for design purposes (for example, the values ag gested by McClelland etal, 1969). These high valet trse because of the combination of high values of fy tan ¢2 Fig 3.103) with a relatively lag critical depth, In sch sates, a more conservative estimate of shaft resi tance may be desirable for design, based on the ves of K, tan $4 derived from Meyerhof (1976) and shown in Fig 3l0e ‘To dlustate the application of the suggested method of calelation, the following example detail calultions fortwo of the ple ets epored by Norlued (1963). Mlustrative Example The piles considered are Piles B and A from the Power Plant Site, Azea I, Helena, Ark. Pile B was a closed-end steel-pipe pile, 24.4 m long and 0.32 m in diameter, driven into fine sand grading to coarse and having an average ULTIMATE LOAD CAPACITY OF FILES 29 TABLE 3.3 SUMMARY OF COMPARISONS BETWEEN CALCULATED AND OBSERVED LOAD CAPACITY OF PILES IN SAND. Reference coe Remark Nordlund (1963) PowerPlant __Valovof 9) suggested Site-Areas by Nordlund sed ran Mojave River_——_Upper 14 fof sand Bdge assumed to hae lower GB") than fower depths @! = 40") because of etng ding instalation PortMana Values of sugested by Nordlund used Values of @', suggested by Nordund used Buffalo Bayou Iovechange __Vertal ste due to so above excavation level ignored O Steet Asabove; pile tested Viaduct asasqare ple Vesic 1967) Pes From reported W values, MiG & H2 following values of @ chosen: O12 f,9, 3371230 it,¢, "38 BOF tee = 42" Desi (373) vies2.310 #4 sttmed tobe 33°, onstant with depth ‘Taveras (1971) 26 constant with depth Piles 126, 6% assumed to be 33°, standard penetration value, V, of about 16. The water table was 3.4 m below the surface. On the basis of the available data, the following values were adopted. (2) Bulk unit weight above water table = 17.3 kN/m?. (b) Submerged unit weight below water table = 78 KN/m? (c) Angle of internal friction angle prior to installation: 5° (0 -2.4 m) 2° (2.4- 18.3 m) 30° (18.3 -20.8 m) 33° © 208m) Considering first the ultimate skin resistance, the values of 9 given by Eq. (3.18) are as follows: @ = 28,75 0-24 m) @ = 34° (24-183 m)30 ULTIMATE LOAD CAPACITY OF PILES 32.5° (183-208 m) 34.78" (© 20.8 m) From fig. 3.0, the values of K, tang are 1.00 (0-24 sm), 1,30 (24-183 m), 118 (18.320 8 m), 138 (20.8 m) IF itis asumed that the critical depth is Tes than 2.4 m ‘below the surface, then for @ = 28.75°, z¢/d = 3.0, from Fig. 3.10a; that is, zp = 5.0% O32 = 1.56 m. Thus, the assumption is justifies. ‘M the ecitcal depth, the effective overburden strest Ole = 156X173 = 2699 KN/m? Because the pile has uniform diameter, F,, = 1 For the ultimate base resistance, the value of @ given by Eq. (3.17) is 36.5°, From Fig. 3.11, the value of Ny is 98 n= ncaa0f (282) x rss42.09% (24-120) 100 +3699 83-20 % 1,30 + 26.99 X (20.8 - 18.3) awe 2690 44-209) 135 032 + 26.99% 4X 8164213 1029 KN (115.6 0) ‘his compares with ehe meusured value of 1312 RN (125 0 File A was a Raymond Standard pile, 10 m tong, with 4 head dhamneter of 0.55 m and a tip diameter of 0.20 m. The pile taper w is 1°. From Fig. 3.12, for w = 1", Fey 3.35 (0-24 in), and Ry = 4.1 (2418.3 m), The values of Kytang are as for pile B. Assuming again that the csitisal depth is above 2.4m, ze/d = 5.0 48 before, and taking an average value of d of O-S1, 2~ = 2.53 m, that is, seater than 24 m, However, the difference is negligible land lence 2 willbe taken as 285m. At this level Ole = 288 %173 = 44.12 Nf? Aram oh = 24X 173 5 41.52 RNIm? Since the pile tip is, founded in the second stratum, 9 from Eq. (3.17) is 36° and the corresponding value of Ny is 8 Substituting into Eq, (3.157 and using, for simpli the misan diameter of the pile in the upper 2.4 m and lower 7.6 m, Fo) XX 01 X 3.35 X 1.00 24" ‘ai g2edt.1y ES a, = (22? 0.15 + 44.12 x (10-2.58)] 0.207 XX 033K 41 X 1,304 44,12X wx OS x88 2243 KN (252.20) ‘Tne measured ultimate load for this pile was 2400 kN 2700). 3.3 PILE GROUPS In examining the behavior of pile groups, it is necessary ‘to distinguish between two types of group: (2) A free-standing group, in which the pile cap is not in contact with the underlying soil, (©) A “piled foundation,” in which the pile cap is in con- tact withthe underlying soi For both types, itis customary to relate the ultimate fos capacity of the group to the load capacity of a single pile through an efficiency factor n, where _= ultimate load capacity of group sum of ultimate load capacities of individual piles G.20) 3.3.1 Pile Groups in Clay 3.31. FREESTANDING GROUPS For freestanding groups of fiction or Moating ples in elay, the efficiency is unity at ‘elatvely lange spacings, but decreases as the spacing decreases. For point-bearng piles, the efficiency is usally considered to be unity for all spacings—that is, grouping has no effect on load capacity, although in theory the efficiency could be greater than unity for closely.spaced piles that are point-bearing, for example, in dense gravel. For piles that derive their load capacity from both sideadhesion and end-bearing, Chellis (1962) recommends that the group effect be taken into consideration forthe sideadhesion component only. Several empirical efficiency formulas have been used to tay and relate group efficiency to pile spacings, among which are the following:(8) Converse-Labarre formula, Dn + (met G21) number of sows number of piles in a row E = arctan d/s, im degrees = pile diameter 5 = center-tocenter spacing of piles (b) Fela’s rule, which reduces the calculated load capacity of each pile in a group by 1/16 for each adjacent pile, that is, no secount is taken of the pile spacing. (©) A rule of uncertain origin, in whitch the calculated load, capacity of each pile is reduced by a proportion J for each, adjacent pile where =h T= yale @.2) A comparison made by Chellis (1962), between these and other empirical formulas shows @ considerable variation in values of 1 for @ given group, and since there appears to be little field evidence to support the consistent use of any empirical formula, an alternative means of estimating ‘group efficiency is desirable, One of the most widely used means of estimating group-load capacity is that given by Terzaghi and Peck (1948), whereby the group capacity is the leser of (a) The sum of the ultimate capacities of the individual piles in the ‘group; or (b) the bearing capacity for block failure of the group, that is, for a rectangular block By X L,, Py = BiLyeNe + 2B, + LV LE 8.23) where ¢ = undrained cohesion at base of group L_ = length of piles Ng = bearing capacity factor corresponding to depth 1 (see Fig. 3.5) © = average cohesion between surface and depth L Model tests on free-sanding groups varried out by Whitaker (1957) confirmed the existence of the above two types of, failure. For 2 given length and aumber of piles in a group, there was a critical value of spacing at which the mechanism of ‘failure changed from block failure to individual pile failure. For spacings closer than the critical value, failure ULTIMATE LOAD CAPACITY OF PILES. 31 ‘was accompanied by the formation of vertical slip planes Joining the perimeter piles, the block of clay enclosed by the slip planes sinking with the pile relative to the general, surface of the clay. For wider spacing, the ples penetrated individually into the clay.‘The critical spacing was found to increase as the number of piles in the group increased Although Whitaker's tests confirmed the existence of the above two types of failure, the transition between the ultimate group capacity as given by individual ple Faure and that given by block failure was not as abrupt as the TTeraaghi and Peck approach suggests. In order to obtain a more realistic estimate of the ultimate load capacity of 2 group, the following empirical relationship is suggested { 4 + ony | en boans Lem 42030 0 segeme | ‘ature f I o 3 FIGURE 3.16 Example of ‘lationship between number of ples and ultimate load eapacity of eroup.32 ULTIMATE LOAD CAPACITY OF PILES to Pe 24) where Py = ultimate Joad capacity of group P; = ultimate load capacity of single pile number of piles in group ultimate load capacity of block (Eq. 3.23) Eq, (3.24) may be reexpressed as follows: 25) where n= group efficiency Figure 3.16 iostrates an example of the relationship between the ultimate load capacity of a group of specified dimensions and the number of piles in the group, cal: culated using Eq. (424). This figure shows the transition betweon singlepile failure and block failure as the aumber of piles increases. In the design of such a group, it is 10 08 06 7 x 26a obvious that virtually no advantage is gained by using more piles than is requited to cause failure of the group as a block; in the example in Fig, 3.16, increasing the number of piles beyond about 80 produces very little increase in ulti- ‘mate load capacity ‘A considerable number of model tests have been carried ut to determine group efficiency factors in homogeneous clay—for example, Whitaker (1957), Saffery and Tate (1961), and Sowers et al, (1961). A summary of some of these tests has been presented by de Mello'(1969) and is reproduced in Fig. 3.17, From this summary, it may be seen that higher effiiency factors occur for (2) Piles having smaller lengthto-diameter ratios (0) Larger spacings (6) Smaller numbers of ples in the group. For spacings commonly used in practice (25d to 4d), ris on the order of 0.7 to 0.85, and little increase in 7 cccurs beyor these spacings, except for large groups of relatively long jes. Figuies 3.18 and 3.19 show comparisons between the measured efficiecy-spacing relationships from the tests ‘of Whitaker (1957) and those calculated from Eq. (3.25). The agreement is generally quite good and the method of x24 00) { pessssseees Soaenpiaamec FIGURE 3.17 Rolatonships for freestanding groups of 2* so 9* piles of lengths 12d te 484, from mod tet (after de Mello, 1969).ULTIMATE LOAD CAPACITY OF PILES. 23 oof Cateulted 0 Expeinena (uae, 1857) bee FIGURE 3.18 Experimental and caleulated group efficiency effect of group size calculation appears to predict with reasonable accuracy the effects of group size, pile spacing, and pile length. It has often been assumed that all piles in a group are ‘equally loaded. However, if the group supports a rigid cap, the load distribution within the group is generally not uniform, the outer piles tending to be more heavily- foaded than the piles near the center. Whitaker (1957) 91 10 osL_L1 4 ie ope ee GT ee w 44 » 16 bie eo wo roy TT os ost tt 5 2 o 2 «6 8 woe 48 ‘eMesure Wher, 5957) = caleuated FIGURE 3.19 Enfect of pile length on group ficiency. hhas measured the load carried by the piles in model free- standing groups in clay by introducing a small load gauge at the head of each pile, The results for a 3? group of piles at three different spacings are shown in Fig. 3.20, in which the average percentage of load taken by each pile is plotted against the group load as a percentage of the group load at failure, At spacings of 2d and 4d, the comer piles take the greatest load (about 13 to 25% more than the average load) while the center pile takes the least (about 18 to 35% less than the average). At a spacing of 8d, virtually no proup action was observed and the load distribution was aniform, The load distribution for a 5? group, ata spacing of 21, is shown in Fig. 3.21. The corner piles eached their ‘maximum load at about 80% of the ultimate group load, and carried a constant load thereafter. At failure, the corner piles carted about 28% more than the average load, while the center and lightestloaded pile carried about 44% less. Therefore, there appears to be a tendency for the load distribution to become increasingly nonuniform as the number of piles in the group increases. A theoretical method for calculating the load distribution prior to ultimate failure is described in Chapter 6, and this method also confirms the trends displayed by Whitaker’s tests, 33.1.2 PILED FOUNDATIONS ‘The ultimate load capacity of a piled foundation (ie. a pile group having a cap cast on or beneath the surface of the soil) may be taken as the lesser of the following two values: (a) The ultimate load capacity of @ block containing the piles (Eq. 3.23) plus the ultimate load capacity of that portion of the cap outside the perimeter of the block.34 ULTIMATE LOAD CAPACITY OF PILES i : i i 3 i 100 0 Load on group a3 pocenage the ea AO FOAO 22080 AO 8040, FIGURE 3.20 Load distibution in 3 po szoup (Whitaker, 1970) i wea Lanta ronson §0 8050 90 "Oe Aumanect nies a £5 £6 % 8O > Mammotoiank ee 0°00 50 Ow smmmotaino 10%'0'O 'O + Memes ‘0°00 oo FIGURE 3.21 Load distribution in 5* pile group at 24 spacing ‘Whitaker, 1970), aera ot ie Acre ofits 8 ene ile (b) The sum of the ultimate Joad capacity of cap and the piles, acting individually, that is, for group of n piles of diameter d and length L, supported by a rectangular cap of | dimensions Be X Le. Py = wiGAst Ap cb Nel *Necte * (Bele ~ ned?|4) 626 where &% = average adhesion along pile y= undrained cohesion at level ct pile tip undained cohesion beneath ple cap Ne_ = beating capacity factor for pile (se Fig. 3.5) Nec = bearing capacity factor for rectangular cap Be X dog (Le > Be) = 5.14 (1 + 0.2 Belli) (Skempton 1951) ‘The fitst value will apply for close second will apply at wider spacings when individual action Whitaker (1960) carried out tests on model piled foun- ‘ations in clay and found that at close spacings, block ‘ure occurred, and that when the cap did not extend beyond the perimeter of the group, it added nothing to the efficien- cy of the group. At greater spacings, the efficiency-versss- spacing relationship was found to be an extersion of the relationship for block failure, with the efficiency exceeding unity bacanse of the effect of “he cap. Good agreement wasoak os} ok Ettisieney. 9 ost oa ost___ 1 1 2 3 4 ‘Shacing feo, 8 = Test on Heetanding grows FIGURE 3.22 Efficiency of pled groups (Whitaker, 1970), obtained between the model test results and the predicted cfficiency from the block failure equation (Fig. 3.22) ‘The load-settlement behavior of piled foundations containing a relatively small number of piles to reduce seltlemeat is considered in detail in Chapter 10. 3.3.13 ECCENTRIC LOADING Model tests on groups with small eccentriciies of load have been carried out by Salfery and Tate (1961), who found that for eccentricities up to two thirds of the spacing, the group efficiency is not noticeably affected. Meyerhot 1.963) also reported that model tests on piled foundations showed that the load eccentricity had no effect en Toad capacity for eccentricities up to half the group width This behavior is explained by the fact that the reduced base resistance is offset by mobilization of lateral ress tance. The group capacity can therefore be ealeulated as for symmetrical vertical loading, except that for groups whose width is on the same order asthe pile length, Meyer hhof (1963) suggests thatthe shat resistance can be ignored and the base resistance caleulated in a fashion similar to eccentrially loaded spread foundations, that is, using a reduced effective base area ULTIMATE LOAD CAPACITY OF PILES 35 3.3.2 Pile Groups in Sand 3.3.21 FREESTANDING GROUPS ‘There is less information on pile groups in sand than en stoups in clay, but it has been fainly well established that group efficiencies in sand may-often be greater than 1 A summary of some of the avalable data on larger piles is given in Table 3.4 ‘A summary of some tests on niodel piles, presented by Lo (1967), is reproduced in Feg. 3.23. The data shown Jn this figure are reasonably consistent with the data in Table 3.4. Results of tests on somewhat larget_model piles, in groups of four and nine, caried out by Vesic (1969), are shown in Fig. 3.24. Vesic measured the point load separately from the shaft resistance, and in the Fight of his measureménts, he concluded that when ihe efficiency of closely spaced piles was greater than unity, tis increase was inthe shaft eather than the point resistance. The broad conclusion to be drawn from the above data is that unless the sand is very dense or the piles are widely spaced, the overall efficiency is likely to be greater than 1. The maxinum efficeney is reached at a spacing of 2103 diameters and generally ranges between 1.3 and 2 2.3.22 INFLUENCE OF PILE CAPS ‘As can be seen in Fig, 3.24, the pile cap can contribute significantly to the load capacity of the group, particularly in the case of the smaller fourpile groups. However, it seems likely that mobilization of the bearing eapacity of the full area of the cap requires considerably greater mnove- ment than that required to mobilize the capacity of the piles themselves, This is the implication of tests by Vesic, and for practical purposes, the contribution of the cap can be taken to be the beating capacity of a stip footing of half-width equal to the distance from the edge of the cap to the outside of the pile. 3.3.23 ECCENTRIC LOADING The influence of eccentric loading on the load capacity of pile groups in sand has been studied by Kishida and Meyerhof (1965) in a series of model tests. These tests showed that small eccentricities of luad have no signif: icant influence on the bearing capacity of freestanding groups and piled groups because the applied moment is, resisted mainly by the earth pressure moment on the sides of the group. At larger eccentricities, the load cape city decreases rapidly because of smaller point resistance of the group by a reduction of the effective base area. In estimating load capacity, Kishida and Aeyerhof suggest that the moment caused by a load at eccenteicitysdo0:8 aivnbs 10) seas8 (arent) Tepe me » yeae 4 ot te au sI0N s61 Wey wor 6 ut $ {soronpaatesay $91 t "nd wad et z et os we “001 (6561 0}94689 sod pazon 1> surg z wot zanswestiet se san Sond usta < e ort woes ors spoues (seen) sora wo noisy PT P a ia omer sau wo8toy mtd on NYS NIS2NO¥9 TIId FTVIS-IOUVT NO VALVE LSAL4O ANYHKAS ¥E4TAVE 36TpPOWRspues uF ADuDFaNIs daodl Jo SoneR palMseayy exe AWADIL srnwerp ay w Buses Le 1 sous sro ag a Selvigg seeing Noun crea sulin38 ULTIMATE LOAD CAPACITY OF PILES aa a FIGURE 3.25 Beating capacity of moael pile groups under eccentric load in sand: (a) freestanding ple etoups; (6) pled foundations ‘Kishida and Meyerhof, 1965). (Copyright Canads, 1965 by University of Toronto Press) is balanced by the moment eaused by lateral forces on the sis of the group until it reaches the maximum value corresponding to the coefficient of passive earth pressure. Within this limit, the eccentricity of load is assumed to have no effect on the point resistance, When the moment Ve is greater than can be resisted by side pressure on the outer piles, the extra is considered to be taken by an ecventric base resistance for the case of block failure; or, for individual pile falure, by the development of uplift resistance of some piles. The total bearing capacity then decreases with further increase in eccentricity Comparisons between the theoretical and measured effect of load eccentricity on load capacity are shown in Fig, 3.25 for the tests carried out by Kishida and Meyerhof (1965), and there is fair agreement for tests in both loose aril dense sans 3.224 LOAD DISTRIBUTION IN GROUP ‘The most detailed data available on load distribution within groups in sand is that reported by Vesic (1969), who made axial load measurements in individual piles during group placement, as well as during loading tests. For the fourpile groups tested, the measured load distek bution was almost uniform, as expected; the maximum deviation from the average was 3 to 7%. For the nine. pile groups, significant nonuniformity of load was mea- sured, The center pile carried about 36% more load than the average, while the comer piles carried about 12% less and the edye piles 3% more, Other tests on similar ‘groups showed a similar trend, with the center piles catry- ing between 20% and 50% more than the average. These results are in contrast to the load distribution in groups in clay, where the center pile caries the least load and the ‘corner piles the most. ‘The influence of the order of driving piles in @ group on the load distribution has been studied by Beredugo (1966) and Kishida (1967). They found that when the Joad on the group was relatively small, piles that had been installed earlier carried Jess load than those that have been installed later; but when the failure Toad of the group was approached, the influence of driving order diminished, and the position of the pile in the group became the domi- nant factor. At this stage, the piles near the center took the most load and the corner piles the least, as in Vesic's experiments, Beredugo also investigated the effects of repeated loading and found that there was a progressive reduction of the influence of driving order, and that for the third and subsequent loadings, the pile position was the domi: nant factor at all loads up to the ultimate of the group. 3.4 PILES TO ROCK 3.4.1 Point-Bearing Capacity There are a number of possibl_ approactes to the estime- tion of point-bearing capacity of piles to rock, including(@) The use of bearing-cepacity theories to calculate the ultimate point-beating capacity Py. (b) The use of empirical data to determine the allowable ppoiat pressure pe (©) The use of insitu tests to estimate either ultimate ‘point capacity pp, oF allowable point pressure Pye Bearing-Capacity Theories Pells (1977) has classified theoretical approaches into three categories 1, Methods that essentially assume rock failure to be “plastic” and either use soil mechanics-type bearing. capacity analyses oF modifications thereof to account for the curved nature of the peak faliuse envelope of rock, 2. Methods that idealize the zones of failure beneath 2 Footing in a form that allows either the brittleness-stength ratio or the brittlenessmodular ratio to be taken into account, 3. Methods based on limiting the maximum stress beneath the loaded area to a value less than required to initiate fracture. These methods assume essentially tit once the maximum strength i exceeded at any point in a brittle material, total collapse occurs. For a typical sandstone having an effective friction angle g” in excess of 45°, effective cohesion c' of about one-tenth of the uniaxial strength, guy, and a ratio of Young's modulusto-uniaxial strength of about 200, Pells shows that the various theories predict an ultimate point bearing capacity ranging between 4.9¢um (incipient failure theory based on the modified Griffith theory) to S6qurm (classical plasticity theory). Various model tests on intact rock carried out by Pells and others indicate ultimate capacities ranging between 4 and II times qum. Pell ‘TABLE 3.5 TYPICAL PROPERTIES OF ROCK (PECK, 1969) ULTIMATE LOAD CAPACITY OF PILES. 39 saws attention to the fact that the load-penetration curve for rocks of medium strength or less (S100 MPa) has a large “plastic” component, despite the brite nature of the rock. The curve divides into two portions, with what appears to be a change of slope associated with the forma tion of a crushed zone beneath the footing. The displace- iments required to mobilize the full bearing capacity of stich rocks are very large, and it seems that a factor of safety of 3 to 4s required (o limit the displacements to less than 2% of the footing diameter. Very brittle rocks (dum > 150 MPa), do not exhibit this “plastic” load-penetration ‘The presence of jointing in the rock will tend to reduce the ultimate bearing, capacity. The presence of closely spaced continuous tight joints may not reduce the bearing capacity much below that for the intact rock material. {F they are open vertical joints with a spacing less than the width or diameter ofthe pile point, the point is essentially supported by unconfined rock columns and the bearing capacity may be expected to be slightly less than the aver. age uniaxial strength of the rock. If the joint spacing is much wider than the footing width, Meyerhot (1953) sug- gests that the crushed zone beneath the footing spits the blosk of rock formed by the joints, Sowers and Sowers (4970) present a theory for this case that generally ind cates a beating capacity slightly greater than the uniaxial strength. Thus, in summary, theoretical considerations sug est that the vitimate bearing capacity is unlikely 10 be re duced much below the uniaxial strength of the intact rock, even if open vertical joints are present. Use of Empirical Data Allowable bearing pressures on rock have often been specified by various building codes and authorities, either based on a description of the rock, of in terms of the ws Rock Compressive Shear Fuv'psy Poisson's Ratio Type Swen um Strength ro) (es) iets ab. Pets Lab. Basalt 28,000-67,000 083s 36-59) 030-032 0:26-0:28 Granite 10,000-38,700 2000-4260 S.6-11.6 Sante 925-027 017-029 Quarta 16,000-84,800 Be 8s 36-125 925-030 007-017 Limestone 450-28,400 1200-2980 Eeerey 0240.27 Marble “7900-27,000 1280-6830, Sandstone +4900-20,900 284-2998 1+ 56 10 90 028-0.30 0.07-0.17 State 1950-31,000 1999-3550 10 25 5384 030-032 024-025 Shale 500-6500 026-027 020-025 Concrete 2000-500 40-1000 25-40 25-40 ous 01s40 ULTIMATE LOAD CAPACITY OF PILES ona 2 | a ba 1 ; | ih i : a ‘0 Vz 30 ry 0 0 ‘Urcontna strength MPa) FIGURE 3.26 Achieved end-bearng pressures in Field tests on piles to rock (Thorne, 1977). ‘uniaxial compressive strength duym- Some typical values of dum and other rock properties are summarized in Table 3.5. Typically, allowable pressures, be, ranging between 0.2 and 0.5 times dum have been stipulated. An ‘example of stipulated bearing pressures related to rock types is provide by Ordinance No. 70 in New South Wales, Australia, in which values of pq range between 430 kN/m? for soft shale to 3210 kN/m? for hard sandstone free from defects to a depth of 900 mm, ‘Thorne (1977) has coliected data on recorded values of bearing capacity, as shown ‘in Fig. 3.26. These values vary from 0:3qym {0 about 4gum, and most casts do not involve failure. The few recorded failures ae in swelling shales and in fractured rocks, itis clear from these results that the fracture spacing has an effect on the bearing capa- city, although the data is insufficient to quantify this fect. On the basis of the available data, an allowable point- bearing pressure on the order of 0.3¢um would appear to bbe quite conservative forall but swelling shales. Reference to the theoretical solutions show that such values generally imply a factor of safety of atleast 3in fractured or closely- jointed rocks and 12 or more for intact rocks. The Use of In-Situ Tests ‘A mumber of methods of in-situ testing of rock have been developed in recent years. Plateload tests have frequently been used but may be expensive if the rock is strong and large loads are required. Freeman et al. (1972) have described the use of the Ménard Pressuremeter to estimate the allowable point-bearing capacity, poe, of piles in rock, and suggest that py may be taken as the value where the pressureversusvolume relationship starts to become nonlinear, Satisfactory designs of caissons in sound shale bedrock using the above approach have been reported by Freeman et al., and design pressures considerably larger than those specified by empirical relationships or building codes have been used. 3.422 Pile Rock Adhesion When piles are socketed or driven into rock, some load ‘ransfer to the embedded portion of the shaft will usually occur. Theoretical solitions for load transfer are discussed in Section 5.3, and also by Ladanyi (1977). The distribu- tion of applied load between side-adhesion and end-bearing at working loads, a8 given by theory, has been supported by insitu measurements at a number of sites (Pells, 1977), ‘There is not a great amount of data on ultimate values of pilecock adhesion, but Thorne (1977) has summarized some of the available data, and this summary is reproduced in Fig. 3.27. These results show that a number of failuresULTIMATE LOAD CAPACITY OF PILES 41 * noes Foie 7 T | | = Sraney Canacie shale} 3 Neva > ? Sintec West 2 . ‘asnting Mg st g ae & } fons | esa ~ Rerpa eamam mm | ‘Melbourne ie 25 Ma encrete SES Sitarnia cn ote ene = ‘shale | © UK siltstone/mudsione I AUK shal ° cae a Oe Uncotine sent Ma) FIGURE 3.27 Adhesion stained in isk tests on piles in rock (Thorne, 1979, have occurred, even in relatively unjointed rocks, at values 3.6 USE OF IN-SITU TESTS on the order of O.lqum. It should be noted that in many instances, concrete strength will be the limiting factor, and in the few instances in which information is available fon concrete strengths, failure has occured at an average shear-stress of between 0.05 to 0.2 times the ultimate com. pressure strength of concrete, f:. However, the tests of Jaspar and Shtenko (1969) indicated thet considerable caution must be exercised with piles in expansive shales that are likely to be affected by water; an adhesion of only about 11 psi (75 kPa) was measused in these tests Freeman et al. (1972) suggest a design value of allowable pile-rock adhesion of 100 to 150 psi (700 to 1000 kPa), depending on the quality of the rock, With such a value, they recommended that the full calculated end-bearing capacity be added to obtain the total design-load capacity. On the basis of the limited information available, it would appear reasonable to use as a design value an allow. able adhesion of O.05f; or O.0Squm, whichever is the lesser value, These values should not be applied to highly fractured rocks, for which values of adhesion between 75 and 150 KPa may be more appropriate. It must be empha- sized that care should be exercised t0 remove all remolded soil from the socket zone. Furthermore, for uplift loads, a reduction of the above loads (e.g. by about 30%} appears to be desirable. 3.5.1 Static Cone Penetrometer The bass of the tests the measurement af the sestance to penetration of a 60° cone with a base area of 10 sq crm. Two types of cone ae commonly used the standard point, with which only point redstance can be meanued; ind’ the fctionjacket point, which allows both point tesitane and local skin esitance to be meesred (Bepe mann, 1983 and 1963), In purely cohesive sols itis genealy accepted thatthe cone point resistance, Cg, is cated tothe undrained cobe Cua = eae 27 ‘As discussed in the previous section, the factor Ne may vary widely both theoretically and in practice, and values of Ne ranging from 10 to 30 have been suggested. The ‘major causes of this variation are the sensitivity of the soil, the relative compressibility of the soil, and the occurrence Gf adhesion on the side of the cone, The variation in the rate of strain between the cone test and other testing methods also has an effect on the deduced value of Ne,42. ULTIMATE LOAD CAPACITY OF PILES but the use of a constantpenetration rate minimizes ‘variations from this cause. For design purposes, a value of Nc = 15 to 48 appears reasonable (Begemann, 1965; Thomas, 1969; Blight, 1967; Thome and Burman, 1968) Van der Veen (1957) suggested that the ultimate ress tance of a pile point, of diameter dp, could be derived from the corresponding cone-penetration curve by taking the average cone resistance over a distance buy below the pile point and aig above the point. Average values: of 2 3.75 and b= 1 were suggested by Van der Veen The adhesion measured by the friction jacket may safely be taken as the skin friction for driven piles in clays {Begemann, 1963). Alternatively, but less desirably, the cohesion may ‘be estimated from the point resistance and an appropriate reduction made to obtain the pile-soil axthesion (see Section 3.2.1), For piles in sand, various attempts have been made 10 relate the conte-point resistance to the angle of friction and relative density of the sand (Meyerhof, 1956; Shultze and Mezler, 1965; Plantema, 1957), but it has been found that cone resistance is very sensitive to changes in density. For practical use, the previously mentioned suggestion of Van der Veen (1957) nay be adopted; namely, that the ‘ultimate point resistance of the pile be taken as the average Cone ana pl cin extance caste so 1010 cone resistance Czy within a distance 3.75 dy above and dy below the pile tip, where dy is the diameter of the pile tip Fullscale tests carried out by Vesic (1967) showed that the point resistance of the pies tested is comparable With that of the penetrometer, but the shaft resistance of the piles was approximately double that measured by the penetrometer, Thus, the ultimate load capacity is given by Py = Cade + Feds 8.28) wh Cea = measured cone-point resistance at bast Je = average shaft friction along pile, as measured on the friction jacket For driven stee! H-piles, Meyerhof (1956) suggested that the above shaft resistance should be halved, ‘A comparison between the pile and penetrometer resistances for the tests reported by Vesic (1967) is shown in Fig. 3.28, The upper and lower limits of the penetro- miter values are shown, Correlation with static cone tests Pile skis retnce and ‘svblod cone sat rsatsnce 6.= 2h tone 200 o4 oe 12 URE 3.28 Variation of point and skin resistances with depth (Vasic, 1967),was found by Vesic to be better than with the results of standard penatration tests (see velow), For cases in which separate measurements of fretion- jacket resistances are not made, Meyerhof (1956) suggested that for deven concrete or tier ples, the ultimate skin fection f, could be estimated from the cone point resis- tance Cea a8 follows fu > 0005Cka 62) For driven steel Hepiles, Meyerhof suggested that the above value be halved. Some comparisons (Mohan et al, 1963; Thorne and Burman, 1968) indicate that Eq. (2.29) underestimstes the skin friction by a factor of about 2 if Oa is less than about 35 keflem?* Iu sans, it is necessary t0 make a distinction between ‘he skin friction for downward and upward loading. Modi fications for uplift resistance are diseussed in Section 3.7. 3.5.2 Standard Penetration Test Meyerhof (1956) has correlated the shaft and base resis tances of 2 pile with the cesults of a standard penetration test. For displacement piles in saturated sand, the ultimate load, in U.S. tons, is given by Ky Py = Ady + 30) where Np = standard penetration mumber, N. at pile base WN = average value of V along pile shaft For small displacement piles (eg., steel piles), VAs Pu = ANpAy +75 31) where Ap = net sectional area of toe (sq £1) Ag = 8108s surface area of shaft (5q ft) (area ofall sur faces of flanges and web for H-piles) {In Eq. (3.30), the recommended upper limit of the unit shaft resistance (W/S0) is 1 tonjft? and it Eq. (3.31), 05 toni? ‘The above equations have also been used with some saceess in stiff clays (Bromham and Styles, 1971). ULTIMATE LOAD CAPACITY OF PILES 43 3.5.3 Pressuremeter Test ‘The use of the pressuremeter in foundation design has bees developed extensively in France in recent yeats. ts appli- cation to the estimation of pile load capacity has been summarizéd by Baguelin et al (1978) who present curves relating ultimate base capacity Co the pressuremeter limit pressure, for both driven and castinsitu piles. Relation. ships are also presented between ultimate skin resistance and limit pressure for steel or concrete pies in granular and cohesive soils, and for east-insitu piles in weathered rock. The following upper limits on the ultimate skin resistance are suggested by Baguelin et al for pressuremeter limit pressures in excess of 1500 kPa; ‘concrete displacement piles in eranular soll 122 kPa concrete displacement ples in cohesive soil, or ste! displace- ‘ment piles i granular soil 82 kPa steel displacement piles in cohesive soil 62 kPa nonslisplacement ples in any soil 40 kPa 3.6 SPECIAL TYPES OF PILE 3.6.1 Latpe Bored Piers Large-diameter bored piles have come into increasing use in recent yeats as an alternative to pile groups. They have been constructed up to 10 ft in diameter and in lengths exceeding 100 ft, often with an underreamed or belled base. Such piles have found extensive use in London clay, and much of the research on large bored piers is based on their behavior in London clay. Empirical methods of esign have been developed oa the basis of extensive expe- rience and research, One of the earliest investigations was in model tests om piles with enlarged bases, reported by Cooke aad Whitaker (1961). These tests revealed that, ‘whereas settlements on the order of 10 to 15% of the base diameter were required to develop the ultimate base capa- city, the full shaft resistance was developed at very small settlements, on the order of 0.5 to 1.0% of the shaft iameter. (The theory given in Chapter 5 supports these findings.) A considerable amount of field-est evidence has subsequently been obtained (Whitaker and Cooke, 1966,44 ULTIMATE LOAD CAPACITY OF PILES Burland et al, 1966), and the behavior of fullscale large bored piers has been found to be similar to that of the model piles. Because of the different degrees of shaftand base load mobilization at a given pier settlement, it may be advisable to determine the working load on a large piee by applying separate factors to the ultimate shaft and base resistances; for example, Skempton (1966) suggested. 2 safety factor of 1.5 for shaft resistance and 3.0 for base resistaice, for piers with an enlarged base of diameter 6 ft or less..4n many cases, the working load for bored piers, especially those with enlarged bases, will be deter. mined by settlement considerations rather than ultimate capacity (Whitaker and Cooke, 1966, Burland eta, 1966) Settlement theory is discussed in Chapter 5. 3.6.2 Underreamed Bored Piles Underreamed piles have been extensively used in India, both 35 load-bearing and anchor piles in expansive clays For anchor piles, a single enlarged bulb is often used, while for load-bearing, one or more bulbs may be used A single vtderreamed pile can be (eated in similar man ner to a pile with an enlarged base, except that the bulb ray be situated above the bese of the pile. Mohan etal (1967) suggest that the base and shaft resistance be added to give the ultimate load capacity. Thus, referring to Fig 3.29, for pile in clay, w Pa = ody + #fldo? ~ a \caNeo coNeb 8.32) ieee A i ocd k >| * ( (0) single wadereomed (2) Douce andereamed ie ie FIGURE 3.29 Undereamed pies cohesion at pile base cohesion at level of base of bulb value of N at pile base level value of Nt level of base of bulb average pile-soil adhesion surface area of pile shaft d= bulb diameter Values of cg, Neb, and Neg can be obtained from Section 32. For double or multiple underreamed piles with the bulbs suitably spaced, the sol between the bulbs tends to act as part of the pile, so that the full resistance of the soil can be developed on the surface 4-A' of a cylinder with a diameter equal to that of the bulbs and height equal to their spacing. Model tests cared out by Mohan et al (1967) have confirmed this behavior. Mohan et al (1969) Ihave suggested two methods for estimating the load capa city of multiple underreamed piles 1, Summation of the frictional resistance along the shaft above and below the bulbs, shearing resistance of the cylinder circumscribing the bulbs, and the bearing capacity of the bottom bull and base. 2, Summation of the frictional resistance along the shaft sbove the top bulb and below the bottom bulb, and the bearing capacity of all the bulbs and the base, 11 was found that for a typical example ofa pile in London clay, these methods give almost identical results. For othec cases the lesser ofthe two capacities given by the equations should be taken. Mohan et al. (1967) suggest that the optimum spacing ‘of the bulbs in a multiple underreamed pile lies between 1.25 and 15 times the bulb diameter for maximum effi ciency. Asan example of the economy in material that may be obtained by using underreamed piles, they calculated thst a multiple underreamed pile in London clay can develop the samme load capacity as a uniform pile of about four times the volume. 3:63 Screw Piles Screw piles have been used in several countries for mast and tower foundations and for underpinning work. Load tests on model and full-scale serew piles have been reported by Wilson (1950) and by Trofimenkov and Mariupolskii (4965). Wilson (1950) developed a method of analysis of the load capacity of screw piles in both sand and clay,FIGURE 3.30 [dealized screw ple based on the use of elastic theory. In a relatively simple analysis for screw piles in clay proposed by Skempton (1950), the load capacity is taken to be the sum of the Dearing capacity of the sceew and the side resistance along the shaft, assuming no skin friction to be mobilized for @ distance above the serew equal to its diameter, Thus, referring to Fig, 3.30, By = NecoAy + éxalLedg) 633) where G_ ~ average remolded shear strength along the shaft in the length (1 -do) cy = average of undisturbed and remolded shear strength of soil beneath the screw Ap = area of serew IMATE LOAD CAPACITY OF PILES 45, The remolded strength of the soll is used because the clay adjacent to the shaft is likely to be almost fully remolded by the passage of the screw and by the lateral displacement caused by the cylinder. ‘A comparison made by Skempton between measured and predicted load capacities by the above method showed that the predicted ultimate oats were within 15% of the measured values, although always greater. Trofimenkow and Mariupolskii (1965) employed the same basis of ca: culation as the above and also obtained good agreement between measured and calculated load capacity. 3.7 UPLIFT RESISTANCE 3.7.1 Single Piles Piles may be required to resis uplift Forces-for example, in foundations of structures subjected to large overturning moments such a6 tall chimneys, transmission towers, or jetty strucires, Methods of calculating the adhesion to resist uplift ate the same as those ust for bearing piles. For a uniform pile in clay, the ultimate uplift resis tance, Pua is Pau = GaAs + Wp G34) where Wp = weight of pile ‘G_ = average adhesion along pile shaft soueeotca | ee ominson (1957) ‘averse vluss Skenpton 1959] foe loud tet | —r— monn and Ghana (1961 «turer i967" | ba ite : puting es eso and Uri (1968) 7000 sco 2000 cranes est FIGURE 3.31 [National Research Council of Canada fom the Canadian Geotechnical Jo 7500 reg (ost) 3007300 ao00 Jonship between cyfey and untrained shear strength fot pulling toss (Sowa, 1970). (Reproduced by permission of the urnal, Vol 7,970, 7p.482-498.)46 ULTIMATE LOAD CAPACITY OF PILES Relatively fw pulling tstson piles have been reported in the literature, A summary of some of the available results is given by Sowa (1970), who has found that the values of caleu spree reasonably well with the values for ples sub- jected to downward loading (Fig. 331). For piles of uniform diameter in sand, the ultimate uplift capacity may be calculated a6 the sum of the shaft resistance plus the weight of the pile. There is, however, little data available on the skin friction for upward loading, and the available data is to some extent conflicting. For ‘example, ests r2ported by Ireland (1957) on piles driven into fine sand suggest that the average skin friction for uplift loading is equal to that for downward loading, but data summarized by Sowa (1970) and Downs and Chiewrzzi (4966) indicates considerable variations in average skin friction between different tests, although there is 2 ten- dency for the values to be lower than for downward load- ing, especially for cast-in piles. In the absence of other information, a reduction to two thirds of the caleulated shaft resistance for downward loading is recommended. However, a reliable estimate is best determined by carving out a pulling test in-situ If static.cone-penetration tests are used as a basis for estimating ultimate uplift skin resistance, Begemann (1965) suggests that the calculated skin resistance for downward loading be adjusted by a reduction factor dependent on the soil and pie type. He also suggests reduced values of skin resistance be used if the uplift load is oscillating Begemann’s suggestions, however, should be viewed with considerable caution, as they are based on limited data ‘Additional uplift resistance may be odtuined by under- reaming or enlarging the bast of the pile, and in such cases, the pile shaft may have little oF no influence on the uplift capacity. Traditional methods of cesign assume the resistance of the enlarged base to be the weight of = cone of earth having sides that rise either vertically or at 30° from the vertical. Neither of these-methods has proved reliable in practice, however. The 30°.come method is usually conservative at shallow depths but caa give a con- siderable overestimate of uplift capacity at large depths (Turner, 1962). Parr and Varner (1962) showed that the vertical falluresurface approach did not apply to piles in clay, although it could apply to backfilled footings. Alter- native theories for uplift resistance of enlarged bases have been proposed by Balla (1961), MacDonald (1963), and Spence (1965)-these theories differing in the assump- tions regarding the shage ofthe failure surface. Meyerhof and Adams (1968) have developed an approx- imate approach fused on observations made in laboratory model tests. They suggest that the short-term uplift capa- city of a pile in clay (under undrained conditions) is given by the lesser of (a) The shear resistance of a vertical cylinder above the base, multiplied by a factor K, plus the weight of soil and pile, W, above the base (b) The uplift capacity of the base plus W, that is, (dp? «a! ag OY oy ea I Ny = uplift coefficient = N- for downward load Examination of the results of model and field tests Jed Meyethof and Adams to suggest the following values of c Soft clays k Medium clays k Stiff fissured clays k ‘The low values of kin the stiffer clays are partly attributed by Meyerhof ard Adams to the influence of tension eracks arising from premature tension-failure inthe clay. It has been found that negative pore pressures may cour in clays during uplift, particularly with shallow embedment depths. The uplift capacity under sustained loading may therefore be less than the short-term or un: rained capacity, because the clay tends to soften with time as the negative pore pressures dissipate. ‘The long-term uplift capacity can be estimated from the theory for a ‘material with both friction and cohesion, using the drained parameters gg and cg of the clay. For a soil with both cohesion and friction, the follow. ing expressions were obtained by Meyerhof and Adams for the ultimate load capacity, Py of a eircular base: (a) Shallow depths ( L< dy) Pay = nedgl + 5ydp LK, tane+ W @.35) (b) Great depths (1. > Pr = nedpH+ Sydy(2L -HNHK, tang *W 3.36) where soll unit weight* shape factor1 miLfa, with ¢ maximum value of 1+ mld Ku = earth-pressure coefficient (approximately 0.9 0.95 forg values between 25° and 40°) m= coeflicient depending on g = limiting height of failure surface above base W. = weight of sol and pile in cylinder above base ‘The upper limit of the uplift capacity is the sum of the net bearing-capacity of the base, the side adhesion of the shaft, and the weight of the pile, that is, Punmax = q(de -L) (Ne + o}4Nq) (337) Agfa Wh where Ne,Nq * bearingcapacity factors fi ultimate shaftshear resistance o}y = effective vertical stress at level of pile base Meyerhot and Adams suggest that the values of No and Ny Cor downward load can be used in this context, ‘but theoretically this is incorrect, and somewhat lower * fvoyant of total, as appropriate. ULTIMATE LOAD CAPACITY OF PILES 47 values may be appropriate to upward loading. However, the theory for failure of anchor piles with enlarged bases, cor of anchor plates more generally, has yet to be fully developed For use in Eqs, (3.35) and (3.36), values of //dp, 5, and m, obtained from tests results by Meyerhof and Adams, ae shown in Table 3.6. The ultimate uplift capa: city should be taken as the lesser value of that given by Eq. (3.37) and the appropriate equations 335 or 3.36, The results of model tests in clays, reported by Meyer- hof and Adams (1968), ate shown in Fig. 332, Both the undtained and long-term pullout loads are shown, and the TABLE 3.6 FACTORS FOR UPLIFT ANALYSIS e 20 ws 3 3 4s a Hy 2 3 4 5 7 9 Ut m 005 01 O18 02 035 05 06 330 7.60) 2 From Meyethof and Adams (1968), (Reproduced by persion of the National Revearch Council of Canada teom the Canadien Geotechnical Journal, Vol. 5, 1968, pp. 125.284) ‘enter | 20 Mena ie capecy ‘tower ff] sna coy seol-| | 4) | col x 4 | \ | | © “f Fa 4 | 1 Te ail dull f Ul iil }|z | | WLLL ile a! Bie cy cay Tus iy Fae mer Hacer aaa tm ME Mh nin Bin Sete won Soni Seat ciese7 oat eee FIGURE 3.32 Comparison of shorter and longterm pullout tests in cay (Meyerhof and Adams, 1968). (Reproduced bY permission ofthe "National Reseach Council of Canada fom the Canadian Geotechalea Journal, Vol 8, 1968, pp. 225.244.)48 ULTIMATE LOAD CAPACITY OF PILES. coniiderable reduction in load capacity with time can clearly be seen. The extent of the load-capacity decrease becomes greater as the soil becomes stiffer, The predicted long-term capacities of the piles show reasonable agree- ment with the measured values. The above theory can also be used to estimate the uplift capacity of ples in sand. Meyerhof and Adams have compared predicted and measured uplift capacities for buried footings in send and have found fair agreement, although there isa relatively wide seatter of points. 3.2 Bile Groups Meyerhof and Adams (1968) suggest that the ultimate uplift load of a group be calculated as the lesser of (a) The sum of the uplift of the individual footings. (b) The uplift load of an equivalent pier foundation consisting ofthe footings and enclosed soil mass. Meyethof and Adams (1968) have presented some date on the uplift efficiency of groups of two snd four model circular footings in lay. The results indieste that the uplift efficiency increases with the spacing of the foot- ings or bases and as the depth of embedment decreases, bbut decreases as the number of footings or bases in the group increases, The uplit efficiencies are found 10 be in povd agreement with those found by Whitaker (1957) for freestanding groups with downward loa. For uplift loading on pile groups in sand, there appears to be little data from full scale field tests, However, Meyer- hhof and Adams (1968) have carted out tests on small sroups of ircular footings and rough circular shafts, and hhave analyzed the group efficiencies. For a given sand density, the uplift efficiencies of the groups increase roughly Hinearly with the spacing of the Footings shafts, TABLE 3.7 SUMMARY OF REPORTED PILE-BENDING MEASUREMENTS Outof Pite Alignment Reference Pile Type Leigh soil Type Tip ‘Type of Bend Parsons and ‘Composite: lower wore 20 Ff, ayers of aan Gentle sweep over Wilson 551, Lin. pipe, ‘stan I, medium lower length (1958) top 5818. sand, fine sand, silt cormugated pipe with clay layers, avel, bedrock Bjerrum ‘Sea Weetinn sore Cay 12te Gentle sweep 9s Johnson Composite: lower sor 20 tit overlying Bn Gentte seep aver 1962) 0 ft, 104i. ‘medium sand Tower leneth upper 50 f, ‘eatragateld taper pipe Mohs (1963) 10%in. pipe ast 80 ft softs roast Gone sweep stiff sind clay, ‘medium dense sand Nationat recast hexagonal, om 50m soft clay. 10m Hn Gente Swedish Hercules jointed cy, st, sand, Council rock st Tom 1968) Hanna (1967) ‘Stee! Hection Of 34st clay, S054 308 ‘Triple curvature 14.BPr3 soft clay, 64 tsi relatively sharp ay, sh lection changes Stool H.sction ban son Double cutvature, 14 BP 59 relatively sharp ‘Seeatian changes