Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Evaluating shear wave velocity and pore pressure data from the seismic cone penetration test Gillespie, Donald G. (Donald Gardner) 1990

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1990_A1 G54.pdf [ 9.08MB ]
Metadata
JSON: 831-1.0050502.json
JSON-LD: 831-1.0050502-ld.json
RDF/XML (Pretty): 831-1.0050502-rdf.xml
RDF/JSON: 831-1.0050502-rdf.json
Turtle: 831-1.0050502-turtle.txt
N-Triples: 831-1.0050502-rdf-ntriples.txt
Original Record: 831-1.0050502-source.json
Full Text
831-1.0050502-fulltext.txt
Citation
831-1.0050502.ris

Full Text

EVALUATING SHEAR WAVE VELOCITY AND PORE PRESSURE DATA FROM THE SEISMIC CONE PENETRATION TEST By DONALD GARDNER GILLESPIE M.A.Sc., The University of British Columbia, 1981 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF Ph.D. in THE FACULTY OF GRADUATE STUDIES (Department of C i v i l Engineering) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA JULY 1990 © Donald Gardner Gillespie In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of C i v i l Engineering The University of British Columbia Vancouver, Canada Date August 30, 1990 DE-6 (2/88) ABSTRACT Recent developments in cone penetration testing have resulted in the addition of both pore pressure measurements and seismometers. The seismometers allow shear wave velocity testing to be performed at designated intervals. Both of these additions were researched to improve their application and interpretation. The significant factors effecting the pore pressure generated during cone penetration tests are discussed. The importance of various factors i s especially dependent upon permeability, strength, and stiffness. For a l l sands tested, pore pressures lower than sta t i c were recorded behind the t i p and higher than static were recorded on the face of the cone. It i s believed that the large compressive stresses on the cone face result in positive pore pressures. As the cone t i p passes a s o i l element unloading and continued shearing generate pore pressures lower than static in a l l sands. The sign of this pore pressure (higher or lower than static) was therefore considered primarily a function of the test equipment. Pore pressure response and the rate of dissipation of excess pore pressures were found useful in distinguishing fine granular s o i l s and explaining s o i l stratigraphy. In cohesive soils the details of pore pressure measurement were found to be important only in s t i f f s o i l s . Pore pressures at a l l measurement locations were found to increase with s o i l strength i n soft to firm clays but may be negative of static in very s t i f f clays. Pore pressures behind the cone t i p were often negative of static in s t i f f clays. Measurement techniques were refined to improve the accuracy of downhole shear wave velocity measurements. Comparisons of downhole and crosshole measurements were made at three well documented sites validating the technique. At several sites i t was found useful to consider the Gjnax values determined from shear wave velocity and density to distinguish s o i l type. Gjnax c o n e resistance ratios were shown to vary systematically with cone resistance values in sands. A wide range in G^ax t o cone resistance was observed i n clays. The dependence of both cone penetration resistance and G^x to increased stress level or overburden stress i s discussed. i i i TABLE OF CONTENTS Page ABSTRACT i i TABLE OF CONTENTS i i i LIST OF TABLES v i i i LIST OF FIGURES ix ACKNOWLEDGEMENTS x i i i CHAPTER 1. lOTROrjUCTION 1 1.1 Purpose and Scope 1 1.2 In-situ Testing 2 1.3 CPT Testing: Equipment 3 1.4 CPTU Testing: Equipment 4 1.5 CPTU Interpretation and Data Reduction 5 1.5.1 Strength Determination: Sands 5 1.5.2 Strength Determination: Clays 6 1.5.3 Stiffness Determinations: Sands 7 1.5.4 Consolidation Characteristics: Clays 8 1.5.5 Initial State or Relative Density of Sands 9 1.6 Shear Modulus as a Geotechnical Parameter 11 1.7 Types of Seismic Waves 13 1.8 Conventional Methods of Determining Shear Wave Velocity 15 1.9 Seismic Cone Development 17 1.10 Thesis Organization 17 CHAPTER 2. DATA COLLECTION AND REDUCTION PRCCEDURES 19 2.1 j^troduction 19 2.2 The Seismic Cone Penetrometer 19 iv 2.3 CPTU procedures 22 2.3.1 Pore Pressure Measurement System Compliance 22 2.3.2 Pore Pressure Dissipations 27 2.4 Seismic Cone Testing Procedure 27 2.4.1 Sources 29 2.4.2 Velocity Measurements: Data Reduction 29 2.4.3 Shear Modulus 32 2.5 Possible Error Sources in CPTU Testing 32 2.5.1 Zero load Stability 33 2.5.2 Resolution 35 2.5.3 Load Transfer 36 2.5.4 External Dimension Tolerances 36 2.5.5 Pore Pressure Effect 37 2.5.6 Repeatability 38 2.6 Possible error Sources in Velocity Measurements 40 2.6.1 Errors Associated With The Use of Arrivals/Crossovers and Cross Correlations 40 2.7 Conclusions 44 CHAPTER 3. SITE DESCRIPTIONS 46 3.1 Introduction 46 3.2 Ons0y: Site Description 46 3.3 Haga: Site Description 50 3.4 Holmen: Site Description 50 3.5 Drarnmen Clay Site: Site Description 52 3.6 No. 6 Road, Richmond: Site Description 57 3.7 McDonald Farm: Site Description 57 3.8 Pile Load Test Site: Site Description 60 V 3.9 Richards Island Site: Site Description 64 3.10 Schoolhouse Site: Site Description 65 3.11 Swiinming Point Site: Site Description 67 3.12 Langley Sites: Site Descriptions 70 3.13 Brenda Mines: Site Description 72 3.14 Heber Road 2, 4, 6: Site Descriptions i 72 3.15 Wildlife Site: Site Description 81 3.16 New Westminster Site: Site Description 81 3.17 Summary 85 CHAPTER 4. FACTORS AFFECTING PORE PRESSURE MEASUREMENTS 86 4.1 Introduction 86 4.2 Effect of Measurement location 86 4.3 Effect of Cone Design & Mechanical Details 91 4.4 Element Saturation 97 4.5 Effect of Cone Design/Procedure on Dissipation Tests 101 4.6 Conclusions 105 CHAPTER 5. INTERPRETATION OF PIEZOMETER CONE DATA 106 5.1 introduction 106 5.2 Cone Resistance Corrections 106 5.3 Soil Classification From Dynamic Pore Pressures 109 5.4 Soil Classification From Pore Pressure Dissipation 114 5.5 Effect of Soil Sensitivity on Pore Pressure Measurements 115 5.6 Interpretation of CPTU Data For Stress History 117 5.7 Interpretation of CPTU Data For Undrained Shear Strength 123 5.8 CPTU data and Liquefaction Resistance 130 5.9 Conclusions 137 v i CHAPTER 6. FACTORS AFFECTING SHEAR WAVE VELOCITY DATA 140 6.1 Source Characteristics 140 6.1.1 Hammer Beam Shear Source 140 6.1.2 Explosive Sources 141 6.2 Receivers ; 147 6.3 Identification of Shear Waves from Explosive Sources 149 6.4 Soil Layering and Resolution 152 6.5 DcwThole-Crcsshole Comparisons 153 6.6 Conclusions 158 CHAPTER 7. APPLICATION OF DATA 159 7.1 Introduction ,.... 159 7.2 Shear Modulus as an Engineering Parameter 159 7.3 Correlation of Velocity to Liquefaction Potential 160 7.4 Integration of data into CPTU Data 166 7.4.1 Introduction 166 7.4.2 Application of Data in the Interpretation of Clay Strength from Cone Resistance 167 7.4.3 Integration of Velocity Data into Soil Classification 170 7.5 Correlation Between CPTU (qp) and G^QX Data 175 7.5.1 Introduction 175 7.5.2 Normalization of qp, and data 176 7.5.3 Correlation Between and qp in Sand 182 7.6 Application of P Wave Velocity Data 184 7.7 Conclusions 185 CHAPTER 8. SUMMARY AND CONCLUSIONS 188 8.1 Summary of Pore Pressure Data 188 v i i 8.2 Summary of Velocity Data 189 8.3 Recxammendations for Riture Research 191 PJ5FERENCES 194 v i i i LIST OF TABLES Table Number Title Page 2.1 Possible Sources of Error in CPTU Testing 34 5.1 Normalized Cone Resistance and Pore Pressure Response in Sand ... 133 5.2 Pore Pressure Response at Sand Sites 136 ix LIST OF FIGURES Figure Number Title Page 1.1 Gjnax Attenuation with Strain 12 from Seed and Idriss (1970) 1.2 Characteristics of Shear Waves 14 1.3 Conventional Downhole and Crosshole Velocity Testing 16 from Stokoe and Hoar (1978) 2.1 Dual element Piezometer Cone 21 2.2 Cone Saturation Procedure Using A Syringe 23 2.3 Cone Saturation System Using A Vacuum 25 2.4 Comparison of Pore Pressure Response at Langley 26 2.5 Source and Receiver Configuration 28 2.6 Buffalo Gun Source 30 2.7 Schematic of the Interval Technique 31 2.8 Variation in Signals From Repeated Soundings: Holmen Sand 39 from Lunne, Eidsmoen, Gillespie, and Howland (1986) 2.9 Repeatability of Shear Waves 41 3.1 CPTU Profile at Onsoy clay site 47 3.2 Soil Description at Ons0y Clay Site 48 from Eidsmoen, Gillespie, Lunne and Campanella (1985) 3.3 Velocity Profiles at Onsoy Clay Site 49 from Eidsmoen, Gillespie, Lunne, and Campanella (1985) 3.4 CPTU Profile at Haga Clay Site 51 from Lunne, Eidsmoen, Gillespie, and Howland (1985) 3.5 CPTU Profiles at Holmen Sand Site 53 from Eidsmoen, Gillespie, Lunne, and Campanella (1985) 3.6 Velocity Profiles at Holmen Sand Site 54 from Eidsmoen, Gillespie, Lunne, and Campanella (1985) 3.7 Soil Profile at Drammen Clay Site 55 from Eidsmoen, Gillespie, Lunne, and Campanella (1985) X 3.8 CPIU Plot at Drammen Clay Site 56 from Eidsmoen, Gillespie, Lunne, and Campanella (1985) 3.9 Velocity Profiles at Drammen Clay Site 58 from Eidsmoen, Gillespie, Lunne, and Campanella (1985) 3.10 CPTU Profiles at No. 6 Road Site 59 3.11 CPTU Profile at McDonald Farm 61 3.12 CPTU Profile at Pile Load Test Site 62 3.13 Velocity Profile at Pile Load Test Site 63 3.14 CPTU Profiles at Richards Island Site 66 3.15 CPTU Profiles at Schoolhoiase Site 68 3.16 CPTU Profiles at Swimming Point Site 69 3.17 CPTU and Soil Profile at Langley Research Site 71 3.18 CPTU Profile at Brenda Mines Site 73 from Campanella, Robertson, Gillespie, and KLohn (1984) 3.19 CPTU Profiles at Heber Road 2 Site 75 3.20 Velocity Profile at Heber Road 2 Site 76 3.21 CPTU Profiles at Heber Road 4 Site 77 3.22 Velocity Profile at Heber Road 4 Site 78 3.23 CPTU Profiles at Heber Road 6 Site 79 3.24 Velocity Profile at Heber Road 6 Site 80 3.25 CPTU Profiles at Wildlife Site 82 from Campanella, Robertson, and Gillespie (1986) 3.26 Velocity Profile at Wildlife Site 83 3.27 CPTU Profile at New Westminster Site 84 4.1 Pore Pressure Distribution During CPTU 87 from Robertson, Campanella, Gillespie, and Greig (1986) 4.2 Penetration Pore Pressures at Imperial Valley Site 90 from Campanella, Robertson, and Gillespie (1986) 4.3 Detailed Penetration Pore Pressures at McDonald Farm 92 x i 4.4 Effects of Element Osnpressibility, McDonald Farm 95 4.5 Detailed Penetration Pore Pressures at Richards Island 96 4.6 Detailed Penetration Pore Pressures at Iangley Site 98 4.7 Pore Pressure Dissipations, Examples From Two Sites 100 from Campanella, Robertson, and Gillespie (1986) 4.8 Normalized Excess Pore Pressure Distribution 102 from Gillespie, Robertson, and Campanella (1988) 4.9 Pore Pressure Dissipations at McDonald Farm Predicted versus Measured Dissipation Curves 104 from Gillespie, Robertson, and Campanella (1988) 5.1 Correction of Cone Resistance Data, Onsoy Site 108 5.2 Undrained, Partially Drained, and Drained Response at the McDonald Farm Site 110 5.3 Soil Behaviour Type Classification 112 from Campanella, Robertson, and Gillespie (1986) 5.4 Pore Pressure Response at Wildlife Site 113 from Campanella, Robertson, and Gillespie (1986) 5.5 Pore Pressure Parameters Bg vs OCR 116 5.6 Normalized Pore Pressure Difference 119 5.7 Pore Pressure Ratios vs OCR 120 5.8 Normalized Cone Resistance vs OCR 121 5.9 Cone Factor % E V S Bq1 124 5.10 Cone Factor N A U vs Eq± 126 5.11 Cone Factor vs Bq 1 # Bg2 127 5.12 Cone Factor N A U 129 from Campanella, Robertson, and Gillespie (1986) 5.13 Pre and Post Compaction Profiles at the No. 6 Road Site 132 6.1 Damped Geophone Response Profile to Hammer Shear Source 142 6.2 Accelerameter Response Profile to Buffalo Gun Source 143 6.3 Geophone Response at Schoolhouse Site 145 from Campanella, Robertson, Gillespie, Laing and Kurfurst (1986) x i i 6.4 Damped Gecphone Response Profile to Buffalo Gun Source 146 6.5 Shear Wave Arrivals from Three Sources 151 6.6 Comparison of Downhole and Crosshole Velocity, Holmen 154 from Robertson, Campanella, Gillespie and Rice (1986) 6.7 Comparison of Downhole and Crosshole Velocity, Drammen 155 from Eidsmoen, Gillespie, Lunne and Campanella (1986) 6.8 Comparison of Downhole and Crosshole Velocity, Wildlife Site .... 156 from Robertson, Campanella, Gillespie and Rice (1986) 7.1 Shear Wave Velocity as an Index of Liquefaction Potential 162 7.2 Cone Factor N R T V S Gjnaj/qp Ratio 168 7.3 Stiffness - Cone Resistance Ratios in Various Soil Types 171 7.4 Integration of and Pore Pressure Measurements into Cone Interpretation 173 7.5 Influence of stress on G^^x a n c* P/T' c^-aY 1 7 8 7.6 Influence of stress on and qj, Sand 179 7.7 Variation of Gmax/Cil! Ratios with Normalized Cone Resistance 183 x i i i ACKNCWIEDGFlffiNTS The writer wishes to acknowledge the encouraging support and suggestions from Dr. R.G. Campanella during the course of this study. The cooperation and suggestions from Dr. P. K. Robertson also proved invaluable. None of the field work would have been possible without the technical support of Messrs Art Brooks, Glen Jolly and Dick Postgate. The assistance of Jim Greig who wrote data collection and processing programs is greatfully acknowleged as was cooperative work with the Norwegian Geotechnical Institute personnel including Tom Lunne, and Terje Eidsmoen. Data collection was conducted mainly with Mr. John Howie who also had many technical suggestions. Financial support was provided by the British Columbia Science Council through the GREAT grant program. Special thanks go to my family for their encouragement and support throughout these years. 1 CHAPTER 1 INTRODUCTION 1.1 Purpose and Scope The cone penetration test i s an important component of site investigations throughout the world. Starting as a very simple logging tool the test has become increasingly sophisticated both from the point of view of the amount and quality of data collected and the subsequent interpretation. Instrumentation advances i n the last decade have resulted in the a b i l i t y to add multiple sensors. Early electronic cone penetration testing, CPT, included cone resistance, q c, or cone resistance and local sleeve f r i c t i o n , f s , only. Subsequent advances added simultaneous measurement of pore pressure, u, and hence the term CPTU. Pore pressure measurements were made at different locations by different researchers and practitioners and piezometer cone testing became the most common logging tool. Subsequently, additional sensors including temperature, t, and inclination, i , were added to assist with deployment and to improve data quality. One recent important addition to the CPTU was the addition of shear wave velocity measurements made feasible by incorporating a suitable velocity transducer and developing appropriate procedures to collect and reduce data. Using pore pressure and shear wave velocity data gathered at a wide variety of sites this thesis addresses the following issues: 1) What are the factors that affect each of the two measurements? 2) How can the data be interpreted to obtain required s o i l parameters? 3) How can data collected from different sensors be integrated to assist in the overall interpretation? 2 1.2 In-situ Testing The investigation of s o i l types and properties can be performed in either the laboratory or the f i e l d (in-situ). The most important advantage of the laboratory methods relates to the control of stresses and boundary conditions. The laboratory observations are extended into the f i e l d . Often simplifications are required but the most pressing problem, especially in sands, i s the d i f f i c u l t y of obtaining and maintaining samples in an undisturbed state or even returning them to their original stress conditions. A further problem i s the small quantity of sample tested in comparison to the enormous amount of material loaded i n the f i e l d . Considering the natural v a r i a b i l i t y of geologic materials the quantity of material tested i n a normal laboratory investigation i s small. By performing in-situ tests the complications of reproducing unknown f i e l d stress conditions on disturbed samples are eliminated and a much greater volume of sample i s tested. In CPTU testing a continuous cylinder of s o i l i s influenced and determines the profile. This feature results in the most significant application of the cone t e s t — a s a logging tool. The trade off for obtaining such a volume of information i s that the CPTU invokes largely unknown and complicated stresses on the s o i l and i t s interpretation becomes somewhat empirical. Interpretation methods presently available, which w i l l be b r i e f l y outlined in this chapter, are largely derived from either correlations to conventional laboratory testing, f i e l d behaviour or experimental tests on large laboratory chamber tests. Some interpretation methods, 3 notably for the strength and consolidation properties, have a theoretical basis in either bearing capacity or cavity expansion theory. One of the complications of CPTU interpretation i s the enormous stresses imposed at the cone t i p ; in comparison to normal foundation loads these stresses are i n a range that may even impose localized crushing. The effects of these various problems i s further discussed i n this thesis. 1.3 CPT Testing: Equipment The cone penetration test continues to become more commonly used as a tool for s i t e investigation and geotechnical design. Past publications of greatest significance describing equipment design include: ASCE Symposium on Cone Penetration Testing and Experience, 1981; the 2nd European Symposium on Penetration Testing, 1982; Robertson and Campanella, 1984; and ASCE In-situ 86, 1986. The ISSMFE Technical Committee on Penetration Testing prepared an International Test Procedure, which was described at the ISOPT-1 conference in Florida in 1988 (deBeer, 1988). The f i n a l version of this document was due at the ICSMFE 1989. The ASTM document, D3441, provides some limited guidelines for CPT testing. The f i r s t cone tests were performed using mechanical cones, measuring either cone resistance alone or cone resistance and mantle f r i c t i o n . Loads were measured at the surface and though the test was slower to perform the equipment was simple and robust. Guidelines for the use of mechanical cones were described by Sehmertmann (1978). Although this report was written with mechanical cones i n mind the 4 interpretation methods proposed for cone resistance and p i l e design procedures remain val i d today. The f i r s t significant description of an electronic cone, a two channel cone, was made by de Ruiter (1971). This cone set a standard for much of the future cone development. Tip apex angles were generally established as 60° and f r i c t i o n sleeves immediately behind the t i p were 'standardized' with an area of 150 cm2 and a diameter the same as the t i p at 35.7 mm. Electronic penetrometers incorporated b u i l t - i n load c e l l s that recorded separately the cone resistance and f r i c t i o n sleeve. Bonded strain gauges were commonly used because of their accuracy, simplicity and ruggedness. Various configurations were used in different designs, each of them having their own advantages. The UBC cone used i n this thesis (Figure 2.1) incorporates many of these i n i t i a l developments and i s discussed in greater detail in chapter 2. 1.4 CPTU Testing: Equipment The introduction of pore pressure measurements has significantly changed CPT testing (the notation CPTU i s generally used to signify piezometer cone testing). The f i r s t two publications were by Wissa (1975) and Torstensson (1975). Design considerations were discussed by Campanella, Robertson and Gillespie (1981), Campanella, Gillespie and Robertson (1982) and Smits (1982). Design considerations included f i l t e r type and location; various publications and conferences considered the sensitivity of results to the element location. These considerations are discussed in detail in chapter 4. 5 1.5 CPTU Interpretation and Data Reduction The primary role of the CPTU as a logging tool i s often overlooked in the continuous attempt to enhance the interpretation procedure for the determination of specific mechanical parameters. Many examples of CPTU profiles are shown in chapter 3. Various methods of interpreting CPTU data w i l l be discussed b r i e f l y here and those areas in which advancements have been made in this thesis w i l l be pointed out. 1.5.1 Strength Determination: Sands The CPT has become an important tool for the interpretation of strength characteristics i n sands that are especially prone to the problems associated with sampling. A l l methods presently available are suitable for interpreting cone resistance through non-cemented, normally consolidated, mainly quartz or feldspar sands. The methods are empirically derived from chamber testing, experience and bearing capacity theory. Most commonly used methods are those of Lunne and Christofferson (1983), Durgunoglu and Mitchell (1975), Jamiolkowski, et a l . (1988) and Robertson and Campanella (1983). Additional methods derived from bearing capacity theory include Mitchell and Keaveny (1986). The Mitchell and Keaveny bearing capacity method requires an estimate of s o i l stiffness that reduces i t s usefulness in practice. One of the uncertainties in a l l of the methods i s the importance of the ambient stress level, i n particular the lateral stress. Some methods such as Houlsby and Wroth (1989) consider only the horizontal stress while others such as Been, et a l . (1986) consider both horizontal and vertical stresses. The commonly used methods, however, consider only 6 the v e r t i c a l effective stress and these methods are generally found to be entirely satisfactory. The dependence of cone resistance on increased stresses with depth in the f i e l d was investigated i n this thesis in chapter 7 and found to be slig h t l y different than that found in chamber tests. 1.5.2 Strength Determination: Clays Four approaches have been used to derive methods of obtaining strengths in clays. A l l approaches are used to determine the cone factor N J Q I for use in: Su = (g.T - Ov) / NKT where: q r p = total cone resistance Traditional bearing capacity methods such as Janbu and Senneset (1974) and cavity expansion such as Vesic (1977) solutions have been supplemented by theoretical approaches, especially the strain path method, Baligh (1986). These three theoretical methods have largely been used to confirm the empirical correlations that typically use either f i e l d vane, Greig (1986), or embankment failure or laboratory results, Aas, et a l . (1986). Considerable difference in opinion regarding the observed variation in N ^ J J with p l a s t i c i t y characteristics has been reported in the literature, Lunne, et a l . (1976), La Rochelle (1988) and Aas, et a l . (1986). Work i n this thesis was directed towards reducing the observed scatter i n N^ r values and determining the cause of the reported variation. Two separate approaches were taken in this thesis. In the f i r s t case i t was observed that measurement details and accuracy considerations explained much of the observed 7 variation. This work i s explained in chapter 2. In the second case, to address the accuracy problems associated with low cone resistance in soft clays, methods that used the pore pressure measurements were developed. These methods are discussed i n chapters 4 and 5. 1.5.3 Stiffness Determinations: Sands A reliable determination of large strain stiffness properties of sands in-situ i s of great practical interest because of the problems of sample disturbance. D i f f i c u l t i e s are encountered in any theoretical interpretation because modulus values are stress level and strain level dependent. In addition the effects of drainage and the direction of loading are largely uncontrolled during penetration testing. Reference values obtained from foundation performance are also d i f f i c u l t to obtain or uncertain. As a result of these d i f f i c u l t i e s the commonly used methods of determining stiffness values in sands are derived from the results of chamber testing. The commonly used methods are those of Robertson and Campanella (1983) and B e l l o t t i , et a l . (1989). The results of B e l l o t t i , et a l . (1989) shows a surprisingly large dependence of the ratio E s / q T, where E s i s a drained secant modulus on the previous stress history or aging of a sand. This dependence was not found i n this study in the variation of Gjn a x / q T. This topic i s further discussed i n chapter 7. The determination of stiffness parameters from CPT i s s t i l l uncertain although lower bound values may be reliably obtained using the available methods that correlate cone resistance to stiffness. The seismic cone methods discussed in this thesis show that the small strain stiffness characteristics can be 8 reliably obtained from the seismic cone. The state of the art and improvements made are discussed later i n this thesis. 1.5.4 Consolidation Characteristics: Clays Other than the determination of equilibrium pore pressures perhaps the primary advantage of the piezometer cone test i s that i t can be used to determine the consolidation characteristics of clay s o i l s . Theoretical solutions that exist for the rate of decay of generated excess pore pressure that are most commonly used include those of Torstensson (1977), which were confirmed by Randolph and Wroth (1979), and those of Baligh and Levadoux (1986). The Torstensson solutions are based on cavity expansion theory and consider cylindrical or spherical dissipation of excess pore pressure. The Baligh and Levadoux solutions were developed by generating the excess pore pressure with the strain path method using parameters determined sp e c i f i c a l l y for Boston Blue clay. These methods are a l l sensitive to the i n i t i a l excess pore pressure surrounding the cone, the distribution of which i s not known. Gillespie (1981) examined these solutions as did Soares, et a l . (1987). A curve matching solution by Gupta and Davidson (1986) was not radically different than the other methods. A l l researchers have found that the available solutions are reasonable for predicting the coefficient of consolidation. The restrictions found i n this thesis, which are discussed in a later chapter, relate to commonly observed d i f f i c u l t i e s i n implementing the solutions and d i f f i c u l t i e s with procedure. Certain restrictions on the solutions are discussed in later chapters; the solutions are found to be v a l i d in normally to 9 l i g h t l y overconsolidated s o i l s . Attempts have been made to extend the determination of consolidation characteristics. Given that the coefficient of consolidation can be determined i t i s often reasoned that the permeability can be determined using the drained stiffness. Determining a drained modulus from the penetration part of the CPT prof i l e , which i s an undrained test, i s not pursued here. 1.5.5 I n i t i a l state or Relative Density of Sands The state of the art of determining the i n i t i a l state of sands using relative density, or dilation angle, or state parameter i n the sense of Been, et a l . (1986), i s regarded as d i f f i c u l t and controversial. Most interpretation methods are based on chamber tests. Important refences commonly used include those of Schmertmann (1978), Robertson and Campanella (1983), Lunne and Christofferson (1983), Jamiolkowski, et a l . (1985) and Baldi, et a l . (1986). A l l of these methods provide reasonable answers in normally consolidated, incompressible, non-cemented, imaged, medium grain sized sands. These restrictions are sometimes d i f f i c u l t to deal with and some of them, primarily the effects of grain size, are investigated in a later chapter. On the other hand the range of cone resistance experienced in sands i n the f i e l d varies by two orders of magnitude and sensitive, repeatable, accurate indexes of density can usually be obtained. Furthermore the solutions of Baldi, et a l . (1986) can also be used in overconsolidated sands. 10 1.5.6 Liquefaction Potential The state of the art of determining liquefaction resistance of sands from CPT cone resistance has been developed i n two separate ways depending on the data available to various researchers. The method developed by Robertson and Campanella (1985) recognized that the commonly used liquefaction resistance curves for the standard penetration test were based on experience that could not be reproduced in an analogous manner for the CPT. To use the experience developed in the SPT curves Robertson and Campanella related the SPT N value to cone resistance and calculated new curves for the CPT. The relation between cone resistance and SPT N value was largely based on side by side f i e l d tests conducted i n the Fraser Delta, Robertson, et a l . (1983) and Laing (1981); these tests were sufficiently extensive to reduce the importance of the nonrepeatable ,nature of the SPT. In this manner Robertson and Campanella (1985) were able to combine the advantage of the SPT experience with the advantages of the CPT quality and continuity. The Robertson and Campanella curves were va l i d for clean sands only. Later publications by Shibata and Teparaksa (1988), based on f i e l d performance observations and f i e l d tests, showed the sensitivity of the cone resistance liquefaction potential to grain size in the fine s i l t y sand range. In the clean sand range the Japanese and Chinese experience published by Shibata and Teparaksa confirmed the Robertson and Campanella curves. Work in this thesis i s directed towards the problem of determining the v a l i d i t y of the various solutions without the aid of sampling. I t shows that there are situations where pore pressures or shear wave velocity measurements can 11 be used to refine the CPT bearing based interpretation. 1.6 Shear Modulus as a Geotechnical Parameter The evaluation of seismic response and the response of foundations to dynamic loads such as machine loads r e l i e s on the determination of stiffness properties. The shear modulus relates shear stresses and shear strains i n the manner of T = Gy where G = shear modulus T = applied shear stress 7 = resulting shear strain which appears simple but as indicated i n Figure 1.1 the value of G i s highly strain level dependent. In addition to i t s dependency on strain level the effects of stress level are of primary importance in determining the stiffness of any s o i l . Early work describing the variation of stiffness i s discussed i n Hardin and Drenevich (1972). An indepth look at the effects of stress level i s included i n chapter 7 of this thesis. A study by Seed and Idriss (1970) presented quantitatively the variation of shear modulus with increasing strain i n sands and clays in the manner shown in Figure 1.1. The reduction i n shear modulus i s readily apparent in Figure 1.1; also very significant i s the apparent plateau value at shear strains below approximately 10~3 to 10~4. The shear modulus at low strains i s widely accepted as being reasonably independent of strain levels and i s termed G^g^, or dynamic shear modulus. o ) SANDS Figure 1.1 G m a x Attenuation with S t r a i n from Seed and I d r i s s (1970) 13 Elastic theory can be used to show that c a n b e determined from G = pVs 2 where Vs = velocity of a shear wave p= mass density of the s o i l hence shear wave velocity measurements in-situ can be used to determine shear modulus, G^^. In-situ measurement i s especially attractive because of the advantages of a bulk sampling process compared to testing small discrete samples. In addition in-situ testing eliminates the d i f f i c u l t i e s of samples disturbance and reconsolidation. The usefulness of downhole profiles to determine the properties of any particular thin layer i s discussed in chapter 2 but the enormous advantage of testing an entire s o i l column, even with some averaging of properties across an interval, should never be overlooked. 1.7 Types of Seismic Waves Seismic wave can be divided into body waves that may penetrate deeply and surface waves that are generally considered to travel near the surface. Solids support two types of body waves: compressional, or P (primary), or longitudal, or irrotational waves; and shear, or S, waves. Compressional waves have displacement parallel to the direction of propagation while shear waves have displacement perpendicular to the direction of propagation. A simple depiction of a shear wave i s shown in Figure 1.2a. Note that the direction of propogation i s perpendicular to the displacement direction and that there i s no volume 14 wavefronts ^displacement direction propagation direction Figure 1.2 SH waves are autonomous at s o i l interfaces C h a r a c t e r i s t i c s of Shear waves 15 change, just deformation. Shear wave deformation can be resolved into separate components, one parallel to the surface (SH) and one i n the ver t i c a l direction. Seismic exploration techniques commonly generate either P and SV waves or SH waves depending on the source used. A fundamental difference between SV and SH waves i s the unique behaviour of SH waves at a boundary. SH waves are autonomous (Figure 1.2b); i f an SH wave strikes a horizontal geologic boundary part of the energy i s reflected and part i s transmitted but both components remain SH waves. The direction of propagation i s fixed by Snell's law. 1.8 Conventional Methods of Determining Shear Wave Velocity Two methods are commonly used to determine shear wave velocity. These are termed conventional crosshole and downhole methods. Considerable literature exists that compares the two methods. Early references include White (1965) and later geotechnical aspects were considered by Stokoe and Woods (1972). Figure 1.3 il l u s t r a t e s the two methods. The single most important distinction i s the complication and expense of d r i l l i n g more than one hole to perform the crosshole test. Downhole testing i s always performed with SH waves, which are v e r t i c a l l y propagating with a horizontal particle motion. Conventional crosshole testing considers horizontally propagating waves with a vertical particle motion, SV waves, but i t i s also entirely feasible to generate horizontally traveling shear waves in the horizonal plane, SH waves, with a torsional source. , The significance of the difference i s discussed in chapter 7, which looks at the effects of the different stresses on shear wave velocity. -12 ft (3.7m) o.-PLAN VIEW Verticol Velocity ^-Verticol Impulse (Not to Scole) b.-CROSS-SECTIONAL VIEW Receiver Borehole Cost-ln-Ploce Concrete Block • 20ft (6.1m) (0.6m) 2ft U-o.- PLAN VIEW Electricol Trigger., -2ft (06m) Embedded Ang le Iron -Hommer Inclined Hommer Blo» Ll wsss ^ i t r^i ) £i' ( 0 6 m ) Cosing "*"• " Generation of Body Waves -3-D Velocity Transducer Wedged in Ploce (Not to Scale) b-CROSS-SECTIONAL VIEW Figure 1.3 Conventional Downhole and Crosshole Velocity Testing a f t e r Stokoe and Hoar (1978) 17 In the downhole test there are two fundamental methods of obtaining velocity. The velocity may be determined by measuring the increment of shear wave travel time by either a pseudo interval travel method or a true interval travel method. The pseudo interval travel method i s carried out by advancing a single geophone to various depths in a hole and measuring the travel time interval between depths from separate energy events. The true interval technique requires the simultaneous measurement from a single impulse event at separate geophones having a known separation. The relative merits of the two techniques were investigated in depth by Rice (1984) and Laing (1985). An explanation for some of the complications Rice (1984) discovered i s investigated in this thesis. 1.9 Seismic Cone Development The CPT seismic cone used i n this thesis i s a downhole tool. Early development work was reported by Rice (1984) and Campanella and Robertson (1982). The important work of Rice (1984) was the verification that an incremental method could be used in practice although Rice (1984) found d i f f i c u l t i e s with the repeatability of the interval method. Equipment considerations of these problems were made by Laing (1985); a separate explanation i s considered in chapter 2. 1.10 Thesis Organization Chapter 1 includes introductory remarks and a brief state of the art of the interpretation of cone data that indicates the strengths and weaknesses of CPTU interpretation and where work was performed in this 18 thesis. Chapter 2 gives a brief description of the special equipment and procedures used in the testing conducted in this thesis. Limitations of the data gathered from the various sensors are addressed. Chapter 3 gives a brief description of each of the sites from which data were collected during this thesis. Chapter 4 describes the various measurement details that determine the results obtained from the piezometer measurements during penetration and dissipations. Chapter 5 discusses the interpretation of piezometer cone data. Many s o i l parameters are shown to influence the results making the interpretation for any one s o i l parameter d i f f i c u l t . With some restrictions results can be interpreted for some s o i l parameters. Chapter 6 discusses some of the important details of shear wave velocity measurements and different results obtained using different sources and receivers. Chapter 7 outlines the interpretation of shear wave velocity data for parameters in addition to the small strain shear modulus, Gj^jj. The application of shear wave velocity data to assist in the interpretation of other data from other sensors i s investigated. The dependence of both velocity and cone resistance on stress levels i s discussed. Chapter 8 presents the major findings of this thesis and offers recommendations for future research. 19 CHAPTER 2. DATA COLLECTION AND REDUCTION PROCEDURES 2.1 Introduction This chapter outlines important aspects of the data collection process used in a l l phases of the seismic cone penetration test. The equipment associated with seismic cone testing i s discussed followed by a brief description of important procedure details. Finally the uncertainties, errors and resolution of different measurements are discussed. 2.2 The Seismic Cone Penetrometer Three different cone penetrometers were used most extensively in this study. A l l were b u i l t at the University of Britis h Columbia, UBC: 1) UBC no. 4, five channel 1 q^ ., f s , u, i , t 2) UBC no. 6, amplified six channel: q ^ f s , u, i , t, geophone 3) UBC no. 8, amplified six channel: g^ ., f s , u l or u2, u3, i , accelerometer, where qj, = cone resistance of projected area 10 cm2, f u l l scale capacity of 75 MPa (750 bar) f s = f r i c t i o n sleeve resistance of projected area 150 cm2, f u l l scale capacity of 1 MPa (10 bar) u = pore pressure, variable capacity transducers u l = pore pressure f i l t e r on the cone t i p u2 = pore pressure behind the cone t i p u3 = pore pressure behind the f r i c t i o n sleeve i = inclination from ver t i c a l , f u l l scale 11° t = temperature either geophone or accelerometer oriented horizontally 20 The most sophisticated of the penetrometers i s shown in Figure 2.1, the other penetrometers are mechanically similar but lack the upper pore pressure capability. Earlier cones such as no. 4 were mechanically similar but did not include amplified signals. Cone no. 8 allowed the simultaneous collection of pore pressure at different locations, both behind (above) the sleeve and either on the cone face or behind the cone t i p . Porous elements used i n this study were machined from nominally sized 120 micron porous polypropylene. On several occasions more r i g i d ceramic f i l t e r s were used to verify that element compressibility was not resulting in the generation of pore pressures; these tests are referred to in chapter 4. Friction measurements were made with an equal end area f r i c t i o n sleeve that reduced pore pressure effects. The thin walled f r i c t i o n sleeve load c e l l was designed to be loaded in tension, thereby eliminating possible buckling and collapse problems. A horizontally oriented Geospace GSC-14-13 miniature geophone with a standard natural frequency of 28 Hz was used as a seismic receiver in cone no. 6 and a piezoelectric type accelerometer with a natural frequency of 3 kHz. was used i n cone no. 8. Advantages and properties of each of these receivers i s discussed i n detail i n Laing (1985). A complete description of the UBC testing vehicle i s given in Campanella and Robertson (1981) who also discuss standard testing procedures. Only non-standard and important procedures relevant to shear wave velocity and pore pressure measurements w i l l be discussed here. circuit board mounting porous polypropylene Quad ring taper fit O-ring — Quad ring 21 slope sensor seismic pick-up pore pressure transducer tension load cell (friction sleeve) equal end area (friction sleeve) LEMO connector compression load cell (bearing) pore pressure transducer porous polypropylene Figure 2.1 Dual Element Piezometer Cone 22 2.3 CPTU Procedures Complete details of CPTU testing are discussed i n Campanella and Robertson (1981); this section describes only the most important non-standard details. 2.3.1 Pore Pressure Measurement System Compliance Rapid pore pressure response generally requires that the pore pressure measurement system be extremely r i g i d . The most important factor influencing the stiffness of the measurement system i s the degree of saturation. The compliance of the transducer, the transducer seating and seals contribute l i t t l e to the compliance of the overall system. No apparent difference was found between miniature pore pressure transducers with r i g i d s i l i c o n diaphragms and more compliant larger steel diaphragm transducers, provided both systems were saturated. The importance of complete saturation increases as the permeability of the s o i l decreases. Complete saturation i s less important through high permeability so i l s , through which flow i s sufficiently rapid, and most important in low permeability s o i l s . Saturation was always performed with glycerine which was selected for i t s viscosity and complete miscibility with water. Two procedures were used in this study; they w i l l be referred to as the syringe and the vacuum systems. A simple system requiring only a syringe that could easily be repeated in the f i e l d i s shown in Figure 2.2. With this system, pore pressure elements were saturated i n the laboratory and stored under glycerine prior to use. A second system was used with cone no. 8 the duel piezometer cone; this procedure i s illus t r a t e d in 23 Porous F i Glycerine Syringe Glycerine Slip-on Saturation Cup Figure 2.2 Cone Saturation Procedure Using A Syringe 24 Figure 2.3. In this case the cone could be assembled dry then placed inverted in a vacuum chamber. After submerging the cone in glycerine a vacuum was applied u n t i l no bubbles were observed. The element type and f l u i d selected helped to maintain saturation during the period in which the penetrometer was removed from the saturation stage to the ground. No attempt was made to maintain the probe under f l u i d during this stage. The vacuum chamber system used to saturate the duel element cone was very easy to perform in the UBC testing vehicle but may not be as simple in some f i e l d conditions. I t was considered to be the most effective means of saturating the pore pressure measurement system. Partial saturation using the syringe method was indicated by saturating the cone as well as possible using the syringe method then placing the saturated cone in de-aired glycerine in the vacuum chamber and applying a vacuum. Bubbles emerging from the piezometer measurement system indicated the syringe saturation method did not completely remove a l l a i r . For this reason the vacuum system was preferred; the necessity of using a vacuum pump unfortunately makes i t a less attractive method for f i e l d use especially in remote areas. At one particularly d i f f i c u l t s i t e with a desiccated surface crust the vacuum system was found necessary to achieve sufficient saturation. Results contrasting the response obtained are shown i n Figure 2.4 which shows a sluggish response from the syringe method and a more detailed response obtained when the vacuum method was used. Rapid increases in the pore pressure with no changes in stratigraphy indicated by the other sensors reflects a lagging response at rod changes. to vacuum pump 25 Section of Plexiglass Chamber with Assembled Cone Inside Saturation Procedure: - cone clamped with vacuum chamber i n place - glycerine added and cap placed on top - with t i p and f i l t e r removed measurement system saturated under vacuum 5 - 1 0 minutes - vacuum vented - previously saturated f i l t e r and t i p placed onto cone - completed cone placed under vacuum for 30 minutes - vacuum vented f l u i d poured o f f and cone removed ( i f membrane was used i t was added with f l u i d i n place and cone removed through the top) Figure 2.3 Cone Saturation System Using A Vacuum Figure 2.4 Comparison of Pore Pressure Response at Langley 3D-to <7l 27 2.3.2 Pore Pressure Dissipations Pore pressure dissipations were automatically recorded at a l l rod break intervals, typically 1 m, and at other selected locations. During rod breaks, procedures in i t i a l l y adopted required complete load removal at the top of the cone rods. Unloading at the top of the cone rods is accompanied by a relaxation of end bearing (discussed in a later chapter). For this reason adjustments to the data collection system were made permitting load to be partially maintained by fixing the top of the cone rods. This does not, however, maintain the penetration cone resistance and a variable amount of stress relaxation invariably takes place. This procedure can be very important when pore pressure dissipations are recorded on the cone face. The influence of stress relaxation was assessed by recording the analogue q,-. signals during pore pressure dissipations. 2.4 Seismic Cone Testing Procedure Generally at the same depth as the dissipation tests, shear wave velocity measurements were also performed. The configuration of the source and receiver are shown in Figure 2.5. Vertically propagating shear waves with a horizontal particle motion (SH) were generated at either side of the beam source; this created signals that arrived with opposite signs allowing easier identification of arrivals and allowing use of crossover events. Shear waves generated using the hammer beam sources were checked to maintain the source amplitude approximately constant. The importance of maintaining a reasonably consistent amplitude is discussed further in section 2.6. 28 Cone Penetrometer with Seismometer (active axis parallel to shear source) Figure 2.5 Source and Receiver Configuration 29 2.4.1 Sources Both compression, P, and shear, S, wave testing was performed. The preferred method to generated shear waves was with the sledge hammer source as depicted in Figure 2.5. In the offshore environment a seabed device can be costly therefore some assessment of the applicability of explosive sources including shotgun shells (Buffalo gun), Figure 2.6, and seismic caps was made. The explosive sources generate both P and S energy and were found to be highly reproducible provided the source was placed i n a f l u i d f i l l e d hole. Previous studies by Laing (1985) did not find the same degree of repeatability. By placing the source into a f l u i d f i l l e d hole better repeatability was achieved. Complete descriptions and comparisons of sources and receivers i s given in Laing (1985). 2.4.2 Velocity Measurements: Data Reduction Velocity measurements were generally made at 1 m intervals and calculation procedures followed the pseudo interval technique, which i s displayed i n Figure 2.7. The work of Rice (1984) confirmed the v a l i d i t y of the procedure by comparing the procedure to a true interval technique. Use of the reaction beam under the UBC testing vehicle as a source kept the horizontal offset small allowing the technique to be started at a depth of approximately 2 m. Larger horizontal offsets introduce greater travel path uncertainties associated with wave refraction. Very short offsets were obtained when the weight of a d r i l l i n g r i g was used to hold down the source beam at Norwegian and Imperial Valley sites. This did not result i n additional rod noise. •12 GAUGE SHELL a) Schematic of B u f f a l o Gun (from P u l l a n and MacAulay, 1984) Figure 2.6 Buffalo Gun Source Drop Rod B u f f a l o ^ Gun \ Trigger and Flange' Prebored H o l e N ^ 0 S h e l l -Ready to F i r e Detonated b) Operation of B u f f a l o Gun 31 Figure 2.7 Schematic of the Interval Technique 32 The use of a f r i c t i o n reducer which widens the hole behind the cone probably eliminates the possibility any of the surface noise coupling with the CPT rods. For the f i r s t depth interval the f i r s t a r r i v a l time must be used to calculate shear wave velocity. Over subsequent depth intervals the interval technique can be used with any suitable time marker. Relative merits i n the selection of the f i r s t a r r i v a l or f i r s t crossover or other techniques are discussed in section 2.6.1. 2.4.3 Shear Modulus Shear modulus i s calculated from shear wave velocity from elastic theory using Gmax = P * v s 2 where v s = shear wave velocity p = mass density of the s o i l Density may be known or estimated from sample information. At some sites studied as part of this thesis density was well documented; at others water content information was known; at others correlations between cone resistance and relative density were used to estimate density. The information available at different sites i s outlined in chapter 3. 2.5 Possible Error Sources i n CPTU Testing A brief explanation of the sources and magnitude of possible error associated with each of the CPTU measurements i s required as i t determines some aspects of subsequent interpretation. This section summarizes experience gained i n this study with the UBC cones and 33 other cones, as well as results from repeated soundings at several sites. Some of this work i s published in Lunne, et a l . (1986). Typical ranges of error are tabulated i n Table 2.1 and discussed below. 2.5.1 Zero Load Stability The single most significant source of error for a l l CPTU measurements i s lack of s t a b i l i t y of the zero load reading. The wide range of s o i l s generally encountered during cone penetration testing results in the requirement that load c e l l s be designed to withstand very high loads yet s t i l l be usable for measuring the properties of soft s o i l s . Thermal shifts are generally the cause of zero load in s t a b i l i t y . Provided proper calibration tests are performed and temperature changes are known the thermal i n s t a b i l i t y can, i n many cases, be corrected for. Zero load i n s t a b i l i t y varies from load c e l l to load c e l l but i s generally proportional to the capacity of the load c e l l . For this reason zero load s t a b i l i t y i s most important when the measurement i s a small portion of the f u l l scale capacity. As shown in Table 2.1 zero load s t a b i l i t y i s a serious source of error i n bearing and f r i c t i o n measurements through soft s o i l s . To quantify the magnitude of errors associated with zero load in s t a b i l i t y , zero load readings were recorded before and after a l l tests. Any s h i f t was compared to that expected from the measured temperature changes. Provided the observed zero load s h i f t could be explained by the known temperature calibration and temperature changes a l l readings were corrected using the recorded temperature change and temperature calibration. TYPICAL VALUES POSSIBLE SOURCES  OF ERROR IN q c: Zero Load Stability Digital Resolution Load Transfer External Dimensions Total Bearing Effect (qp correction) Absolute Value Relative error Relative error of uncertainty in soft CLAY* in SAND* .5 bar 20 % .5 % .5 bar 20 % .5 % n i l n i l n i l .05 mm n i l n i l no significant error i f pore pressure measured behind tip POSSIBLE SOURCES  OF ERROR IN f c : Zero Load Stability Digital Resolution Load Transfer External Dimensions Pore Pressure Effect Absolute Value Relative error Relative error of uncertainty in soft CLAY .05 bar .01 bar unknown .05 mm 50 % 10 % unknown small small in SAND" 5 % 1 % unknown possibly large n i l POSSIBLE SOURCES  OF ERROR IN u: Absolute Value Relative error Relative error of uncertainty in soft CLAY in SAND Zero Stability .01 bar .4 % 1 % Digital Resolution .01 bar .4 % 1 % * Example Soils Considered: <3c FR u (bar) (bar) % (bar) soft CLAY 2.5 .125 5 2.5 SAND 100. 1.0 1 1.0 Note: - Absolute errors may be smaller - Absolute errors are typical for 10 ton capacity cones Table 2.1 Possible Sources of Error in CPTU Testing 35 2.5.2 Resolution Digital resolution was included i n Table 2.1. A 12 b i t analogue to d i g i t a l converter was used in this study. The resolution becomes significant i s soft clays for bearing and f r i c t i o n measurements but i s in l i n e with the zero load s t a b i l i t y . The computer data acquisition system used i n this study effectively achieved greater than 12 b i t d i g i t a l resolution, depending on s o i l conditions, by using an averaging procedure i n which 20 readings taken at 30 kHz were averaged to a single value. During this sampling interval the cone would have traveled only approximately 0.013 mm. The a b i l i t y to quantify the properties of a s o i l layer within an interbedded sequence i s also a question of resolution. No fundamental or experimental work has been done in this area. Treadwell (1975) and Schmertmann (1978) both suggest that a probe must penetrate 5 to 10 diameters into a s o i l layer and remain 5 to 10 diameters above a \ subsequent layer interface in order that a reading in the layer represents the properties of the layer in question. Clearly the properties of the layer and the relative contrast of the properties of the layers in question determines each situation. Experimental work performed with cones of different diameter in a layered deposit could c l a r i f y this situation. Friction data may be influenced by a smaller depth of s o i l than cone resistance but no work was done to quantify this. It i s d i f f i c u l t to assess the thickness of the zone of s o i l influencing the pore pressure measurement system. Although i t measures a response generated by the cone t i p i t i s influenced by a smaller zone of s o i l than the 36 cone resistance measurement. One of the best ways to evaluate the depth resolution of each of the channels i s to compare the a b i l i t y of each to identify stratigraphic changes. The a b i l i t y of each of the various sensors to identify changes depends on the nature of the contrast and the s o i l conditions present. This topic i s explored further on a site specific basis in later sections. In any case, data were collected on 25 mm depth intervals, an interval smaller than one cone diameter. This depth resolution appeared to be more than sufficient in the sense that many readings were recorded within a l l detected s o i l layers. 2.5.3 Load Transfer Load can be transferred across the gap between the t i p and sleeve when s o i l particles f i l l the gap between the two pieces. In most situations the load transfer i s small, but occasionally, high f r i c t i o n readings are obtained and when the hole i s completed the f r i c t i o n zeros appear offset u n t i l the cone i s cleaned. The situation i s best avoided with proper maintenance between probings. Several probings were repeated in the course of this study because of this problem. '2.5.4 External Dimension Tolerances Friction sleeve data are especially prone to measurement errors. Many CPTU users have l i t t l e confidence in the absolute value of the f r i c t i o n data collected. Load transfer i s often regarded as the source of the error but another possible source of error, which was noted by de Ruiter (1982), i s the use of undersized f r i c t i o n sleeves. This was 37 avoided by the use of properly sized f r i c t i o n sleeves. Lateral stress conditions, which effect the f r i c t i o n reading, are highly influenced by the exact diameter of the sleeve compared to the cone t i p . Care was taken to ensure that the diameter of the f r i c t i o n sleeve was at least as great as that of the cone. 2.5.5 Pore Pressure Effects The importance of total bearing correction has been emphasised in previous publications including Campanella, et a l . (1982). Provided pore pressure measurements are made i n the correct location the correction i s simple. In many cases, however, pore pressure measurements are made on the cone face. In this case an estimate of the appropriate value behind the t i p has to be made. The relationship between pore pressures measured at these two locations i s relatively straightforward in soft clays were the correction i s most important. In s t i f f clays where the t i p resistance i s higher the pore pressure distribution i s more variable but fortunately the correction i s less important. Friction readings are also influenced by an uneven pore pressure distribution. When pore pressures are different on either end of the fr i c t i o n sleeve (which has equal exposed end areas) a net force i s exerted. The addition of pore pressure measurements at either end of the sleeve allowed this correction to be checked. The net force was found to be less than the d i g i t a l resolution for a l l s o i l s encountered in this study except s t i f f clays. Fortunately in s t i f f clays the high f r i c t i o n induced by the s o i l reduces the importance of any correction. 38 The application of equal end area f r i c t i o n sleeves reduces most of the pore pressure effect. 2.5.6 Repeatability An assessment of the overall accuracy and repeatability of the cone penetration test can be made by comparing repeated adjacent soundings. Some of the variation seen may be due to natural s o i l v a r i a b i l i t y but tests performed as part of this study at several sites in Norway show that data collected on different channels are subject to different levels of repeatability, not a l l of which can be explained by s o i l v a r i a b i l i t y . Figure 2.8 shows a reasonable scatter i n the cone resistance traces and there i s no systematic bias of any one sounding compared to the average. The pore pressure data are well reproduced in the profiles. On the other hand sleeve f r i c t i o n readings are widely scattered for the different profiles. Errors at this s i t e are in the order of a factor of two for sleeve f r i c t i o n . The general trend observed at many sites i s that cone resistance can be accurately reproduced except i n soft clay or peat sites where zero load s t a b i l i t y related errors may be a significant portion of the reading. Friction values may be unreliable, and pore pressure data are generally highly reproducible and reliable provided proper care i s taken to saturate the cone in fine grained so i l s , and that pore pressure records are only compared for similar measurement locations. Although f r i c t i o n readings were generally found to be unreliable in a l l but dense sand sites i t was generally found that changes in the f r i c t i o n reading through a pro f i l e accurately indicated the stratigraphic boundaries. The I Figure 2.8 Variation in Signals from Repeated Soundings: Holmen Sand adapted from Lunne, Eidsmoen, Gillespie and Howland (1986) 40 apparent lack of accuracy of f r i c t i o n data must be considered, however, i n assessing the usefulness of interpretation methods incorporating f r i c t i o n data. 2.6 Possible Error Sources In Shear Wave Velocity Measurements Velocity data are shown throughout this thesis. The order of magnitude of the accuracy that can be anticipated i s discussed in this section. A consideration of errors i s important considering that the velocity i s squared when calculating modulus values. 2.6.1 Errors Associated With the Use of Arrivals/Crossovers and Cross Correlations Use of the pseudo time interval method allows the use of any repeatable marker to calculate shear wave velocity. Rice (1984) experimentally determined the repeatability of the crossover time. Rice concluded that an averaging of up to ten blows could significantly improve the accuracy of the method over the use of a single blow at each test depth. Examination of the cause of these d i f f i c u l t i e s in this thesis did not reach the same conclusion. Improvements in the source beam and ensuring complete ground contact greatly improve the repeatability of the source and subsequent marker times. Compared to the crossover event the arrival i s more repeatable but more d i f f i c u l t to define. Figure 2.9 shows five traces recorded at one depth. Two of the traces are nearly identical, the others show a common pattern, they a l l have the same arrival time but those with larger amplitude have delayed crossover times. The high frequency 42 component of both the high and low amplitude waves has the same velocity resulting in identical arrival times regardless of the source amplitude. The subsequent apparent divergence i s l i k e l y due to the different frequency content of the input signal, the higher amplitude signal has greater low frequency (slower component) and subsequently a delayed crossover. This dispersion results i n the lower repeatability of the crossover event compared to the arrival event. The maximum variation i n the crossover times was observed to be i n the range of 0.5 ms with deliberately different amplitude sources generated with the beam and hammer source. The divergence could be dealt with in several ways: either by averaging (perhaps by summing) multiple events; or by maintaining a repeatable source. In this thesis the latte r approach was used with occasional checks by testing multiple strikes at one location. At most sites, provided good shear beam contact was made, the repeatability of any one measurement was within 0.05 ms. This error in the repeatability of the measurement over a time interval of, typically, 5 ms (velocity of 200 m/s) represents an error of +/- 2 % assuming the worst case of the error being nonsystematic. At the Norwegian sites tested, depth control was regulated by the rod stickup at the back of a flexible light weight d r i l l r i g . Depth control was possibly as bad as 25 mm. This error results in velocity errors in the interval above and the interval below and would result in a maximum error of 50 mm in 1 m or a 5 % error i n interval velocity. For the most uniform sites tested such as the Ons0y site, Figure 3.3, this i s the magnitude of the variation observed between soundings. It 43 i s d i f f i c u l t to know how much of the variation observed may be due to s o i l heterogeneity. In any case, the repeatability of measurements at one depth i s very good provided a reasonably consistent source amplitude i s used. When tests were conducted in adjacent soundings, results were repeated within approximately 5 %. Investigations at the Holmen site, Figure 3.6, resulted i n more variation between soundings but more scatter was observed in the cone resistance traces and hence the variation was considered to result from natural s o i l variations. Use of arrival times, observed to be highly repeatable, did result i n the additional problem, however, of defining a suitable marker location. As shown i n Figure 2.9 the arrival event i s very d i f f i c u l t to define. One means of defining the event was to construct a tangent at the point of inflection after the ar r i v a l . The intersection of the tangent to the zero crossover could then be defined as an a r r i v a l . Cross correlation techniques were attempted at several sites to calculate the interval between subsequent arrivals. This method offers the advantage of permitting automation of the process. The d i g i t a l storage oscilloscope used in this study could optimize the shifting and comparison of up to 1024 points. The method suffers the same dependency upon input source repeatability as the crossover technique, though to a lessor degree, as i t considers a l l of the points selected. The optimum window of data to use for cross correlation appeared to be that obtained between the f i r s t a rrival and the f i r s t crossover. Typically data were collected at the rate of 10 microseconds/point and about 1000 points would be contained between the arr i v a l and the crossover. Results from the cross correlation method were very similar 4 4 to those of the arrival and crossover events provided reasonably consistent signal sources were used. 2.7 Conclusions 1) Some details of the data collection process were described. In particular those areas that are non standard practice or not described elsewhere were emphasised. Cone saturation procedures i n particular were described. Where high quality pore pressure data were required in soft clay sites vacuum techniques were preferred. 2) General guidelines showing the accuracy of CPTU and velocity measurements were given. Table 2.1 gives estimates of the absolute and relative error associated with the different measurements. CPTU testing in soft clay i s often confined to cone resistances less than 0.1 % of the cone f u l l scale capacity. At such low loads zero load d r i f t related errors must be considered. Temperature offsets were found, by laboratory calibration work, to cause much of the zero load errors. Temperature measurements were used to correct a l l data points. Cone resistance measurements in sands, and s t i f f s i l t s or clays are much greater than the zero load i n s t a b i l i t y errors and percentage errors are insignificant. Friction data are unreliable in loose s o i l s and soft clays but the relative changes within a pro f i l e are reliable. Pore pressure readings, provided saturation was good, were found to be extremely accurate i n a l l s o i l conditions. Errors associated with shear wave velocity measurements were discussed. Possible sources of error include the depth measurement, repeatability of the signal, and identification of the event. A trade off between ease of 45 identification of the event and event repeatability occurs between the use of the f i r s t a rrival and the f i r s t crossover. In either case velocity measurements, provided a consistent source i s used, are l i k e l y accurate to 5 %. 46 CHAPTER 3. SITE DESCRIPTIONS 3.1 Introduction During this research program testing was performed at a number of sites, this chapter i s a compilation of site descriptions for those sites investigated by the author. Each description includes a summary of the s o i l properties, reference to more detailed descriptions where available, a CPTU plot, and velocity data as available. Occasionally, adjacent profiles were performed and more than one pr o f i l e i s shown. 3.2 Ons0y: Site Description The Ons0y site i s located approximately four kilometers north of Fredrikstad in Southern Norway. The clay deposit consists of approximately 45 m of marine clay. The top 30 m include a weathered crust 1 m thick underlain by 8 m of soft clay with iron spots, organic matter and shell fragments. Water contents are near the liq u i d l i m i t and are close to 60 %. Below this layer i s a soft homogeneous clay s l i g h t l y more plastic than the clay above and having water contents between 65 and 60 % between depths of 8 and 20 m. The s i t e was investigated with the piezometer cone i n 1982 and again with the seismic cone i n 1984. Detailed results are given in Eidsmoen, et a l . (1985). A detailed site description i s given in Lunne, et a l . (1976). Figure 3.1 shows an example CPTU plot. A sediment pro f i l e description i s given i n Figure 3.2 and shear wave velocity data are shown i n Figure 3.3. The well documented nature of this site made i t useful in this thesis. In addition the high a. UJ o 47 PORE FRICTION CORRECTED FRICTION PRESSURE RESISTANCE CONE RESISTANCE RATIO U. kPa FC. kPa QT. kPa RF«FC/QT r/.i DIFFERENTIAL P.P. RATIO AU/Q.7 10 -V 15 MOO 0 20 0 1000 0 1.0 0 -80 0| i i i i | 0 i i i i i i i i i i i 0 i i i i i i o 10 15 10 -15 2Q! « I I. . I 2Q1 ' ' ' » ' 20' 1 1 1 1 1 1 1 ' 1 ' 20' 1 ' 1 1 ' 20 ' ' ' ' ' 10 15 10 15 i i i Figure 3.1 CPTU Profile at Onstfy clay site DEPTH, m SOIL DESCRIPTION WATER CONTENT, •1. 20 40 60 80 BULK DENSITY, F IELD V A N E S H E A R S T R E N G T H , t / m ! 1 2 3 4 SENSITIVITY 5 Weathered _ CRUST CLAY »-0 0 —"i o 8? • 1.83 1.80 1.64 165 167 1.67 -we •* t7 t • 8 6 6 s 10 with iron-sulphide spots • <-w« —( < j8 o 0 ? 167 1.67 + « «> * o <SW-' a« *7« V O n 10 6 5 6 6 15 CLAY o o o °o o o CP ? o 1.64 V 1- o cd ( 0 < • t-+ o 7 7 5 5 5 20 0 p S 1 1.68 + « •+ S 0» + 5 4 4 5 5 25 O + «&-O O • • 4 5 4 4 4 30 • 0 • 4 4 35 Figure 3.2 Soil Description at Onsoy Clay Site from Eidsmoen, Gillespie, Lunne and Campanella (1985) Shear wave velocity (m/sec.) o 100 200 300 5 E J= 10 15 XO + *-* 4 *• * * Y LEGEND: + CPT 1 • CPT 2 x CPT 3 **• * f • X Figure 3.3 Velocity Profiles at Onsoy Clay Site from Eidsmoen, Gillespie, Lunne and Campanella (1985) 50 degree of uniformity made i t especially useful to investigate the repeatability of the seismic cone test and CPTU measurements. 3.3 Haga: Site Description The Haga site i s located about 50 km north of Oslo. The test site i s located on a slope above the Flomma river. The Haga clay i s a medium s t i f f overconsolidated lean marine clay with an OCR decreasing from about 30 near the surface to about 2 at 7 m. The index properties and an example CPTU plot are shown in Figure 3.4. The clay behaviour at this site i s unusual in many ways, which can, in part, be explained by i t s leached marine history. Present salt content i s only 1 g/1 but the clay i s not particularly sensitive. In addition, a sand and gravel layer below the clay layer drained downslope into the river and created a downward hydraulic gradient of about 1. Static pore pressure values are near atmospheric throughout the entire profile. No seismic cone work was done at this site. Additional details are given in Eidsmoen, et a l . (1985). 3.4 Holmen: Site Description Holmen i s an island in the Drammen river just downstream from Drammen, Norway. Below a sand f i l l 2 m thick i s a very uniform sand layer down to 22 m. The sand i s very loose, medium to coarse grained. Between 22 and 30 m, there i s a fine to medium grained compact sand layer. The s i t e was investigated in 1982 with the piezocone and later Figure 3.4 CPTU Profile at Haga Clay Site adapted from Lunne, Eidsmoen, Gillespie and Howland (1986) 52 in 1985 with the seismic cone. Three repeated seismic cone tests were performed in 1985. The data are reported i n Eidsmoen, et a l . (1985). The s i t e i s of interest because of i t s very uniform nature and the fact that comparative testing was possible. Tests with the UBC seismic cone at Holmen offered a good opportunity to verify the downhole shear wave velocity measurement technique and compare results to crosshole and surface wave techniques these comparisons are made in section 6.5. Figure 3.5 shows an example cone plot with some geotechnical parameters. Figure 3.6 shows the detailed velocity measurements from the three boreholes; those intervals having high velocity also had high t i p resistance measurements reflecting natural v a r i a b i l i t y . 3.5 Drammen Clay Site: Site Description The Drammen clay site at Museumsparken in Drammen, Norway, was investigated in 1985. The site i s used i n this thesis because of i t s well documented properties including other reference velocity measurements. Figure 3.7 shows a summary of the s o i l properties; complete descriptions are given in Eidsmoen, et a l . (1985), Lunne, et a l . (1976) and Lacasse, et a l . (1981). The surface of the s i t e i s covered with 2 m of sand underlain by a plastic clay deposit of marine origin with water contents varying between 55 and 60 %. The p l a s t i c i t y index averages 30 %. The plastic clay i s underlain by a lean clay deposit with water contents between 30 and 35 % and a p l a s t i c i t y index of 10 %. Figure 3.8 shows an example CPTU plot from the sit e . Although the s i t e includes two very different clays within the pr o f i l e the only indication of a change in s o i l type on the CPTU pr o f i l e at PORE SLEEVE CORRECTED FRICTION DIFFERENTIAL P.P PRESSURE FRICTION CONE RESISTANCE RATIO RATIO AU/qT U, kPa fs, kPa q T l MPa RF=fs/qT. 0 5 0 0 0 00 o 10 20 30 SAND FILL medium to coarse SAND D5Q=.45-.90 D1fJ=.20-.50 subrounded quartz feldspar fine to medium SAND with s i l t y sand layers D5Q=.45-.90 D1Q=.20-.50 Figure 3.5 CPTU Profiles at Holmen Sand Site from Eidsmoen, Gillespie, Lunne and Campanella (1985) fll 5 4 Shear wave velocity ( /sec) 100 200 300 • + + • x • X f •I* X -He • + x • + + • + • • + LEGEND: + CPT 1 • CPT 2 x CPT 3 + + • • • -— » Figure from 3.6 Velocity Profiles at Holmen Sand Site Eidsmoen, Gillespie, Lunne and Campanella (1985) DEPTH, m SOIL DESCRIPTION WATER CONTENT, V . 30 40 50 60 BULK DENSITY, F IELD VANE S H E A R S T R E N G T H , l / m ! 1 2 3 4 SENSITIVITY 5 FINE S A N D SILTY CLAY • 1.81 1.79 0 o a a b 6 10 (PLASTIC) CLAY 1.71 1.70 1.7t 175 1.76 1.90 1.93 195 K 9 <; c CO + ? o D *> 8 9 7 6 7 15 - S A N D , - GRAVEL c % a c d O 9( 9 + 9 It J + O ^ EC + ) -^Q i + 9 4 3 2 '2.5 & 5" • 20 (LEAN) CLAY i o '-4 1.98 1.99 1.97 1.98 1.93 9 9 a? 07 + 9 ) 7 O + + 2 3 2 2 5 2 25 —\ i " 198 1.94 + + + 3 3 3.5 3.5 3 30 i a * 1.93 1.94 + + 7 + 9 • + 3 3 3 3 3 35 + 3 3 Figure 3.7 Soil Profile at Drammen Clay Site from Eidsmoen, Gillespie, Lunne and Campanella (1985) P O R E P R E S S U R E F R I C T I O N R E S I S T A N C E B E A R I N G R E S I S T A N C E F R I C T I O N R A T I O D I F F E R E N T I A L P . P . S O I L U ( B A R ) FC . ( B A R ) QT ( B A R ) R F = F - C / Q 7 ! / ) P A T I O kV/UT P R O F I L E f i n e SAND s i l t y CLAY p l a s t i c CLAY l e a n CLAY Figure 3.8 CPTU Plot at Drammen Clay Site from Eidsmoen, Gillespie, Lunne and Campanella (1985) 57 11 ui i s the change i n slope of the t i p resistance vs. depth plot. Neither the f r i c t i o n or pore pressure measurements reflect the change in s o i l type. An interpretation of CPTU profiles that r e l i e s on the change of one of the parameters with depth inherently assumes that a l l other factors remain the same and can be misleading i f the s o i l type i s not well known. A velocity pr o f i l e i s shown in Figure 3.9. As with the Ons0y site, velocities measured at the Drammen site compared well with those made with conventional crosshole techniques. 3.6 No. 6 Road Richmond: Site Description The No. 6 Road site i s located in the Fraser river delta deposits at Richmond, BC. The site includes a thick sand layer, extending to a depth of 20 m which i s medium grained with a trace of s i l t . Figure 3.10 shows a CPTU plot, including before and after compaction results. Vibro replacement techniques were used at the s i t e and the CPTU pr o f i l e after compaction was done in the centroid of the stone columns. The site i s of interest because of the pre and post compaction data. 3.7 McDonald Farm: Site Description The McDonald Farm site i s located near the Vancouver International Airport on Sea Island in Richmond, BC. The site has been reported elsewhere, including Campanella, et a l . (1983), Greig (1985), and Konrad and Law (1985). Beneath the clay s i l t crust i s a sand deposit Shear wave velocity (m/sec.) 100 200 5 15 4 • 1 + • • * + • 1 +• + • + LEGEND: + CPT 1 e CPT 2 I Figure 3.9 Velocity Profiles at Drammen Clay Site from Eidsmoen, Gillespie, Lunne and Campanella (1985) 59 (_ QI 4 J OJ E Q_ LU a PORE PRESSURE U (ra. of water) 0 50 7.5-15-22. 5 J— 1 4 IT ft » > s A. SLEEVE FRICTION (bar) 0 2.5 1 • • 1 1 22.5 Depth Incromont CONE BEARING Qc (bar) 7.5-22.5 025 m Before Compaction After Compaction #1 After Compaction #2 a l l pore pressures measured behind t i p Figure 3.10 CPTU Profiles at No. 6 Road Site 60 that extends to 13 m. The sand i s medium grained, but includes s i l t y fine sand layers found at irregular depths around the sit e . From 13 to 15 m, a fine sand layer i s observed across the sit e . This layer has much lower cone t i p resistance than the material above and i s believed to be loose. A very deep clayey s i l t layer starts at 15 m. Some index properties of this layer are included in the CPTU profiles, Figure 3.11. This layer behaves in a normally consolidated manner and contains sufficient clay to behave as a cohesive sediment. The s i l t layer contains numerous thin lenses of sandy material that can be identified by increases in the t i p resistance and drops in the dynamic pore pressure. 3.8 Pile Load Test Site: Site Description The p i l e load test site i s located on Boundary road i n New Westminster, BC, near a major crossing of the Fraser River. A CPTU plot from the s i t e i s shown in Figure 3.12. The site i s covered by 2 m of f i l l . Below the sand f i l l i s an organic s i l t deposit which has a high organic content in the upper few metres. CPT f r i c t i o n measurements at this s i t e clearly indicate the extent of the peat rich layers. Below the s i l t layer there i s a gradational change to a fine sand unit. Similar to much of the Fraser river sand deposits the sand i s interbedded with s i l t layers which are clearly indicated on the CPTU plot by lower t i p resistance, higher f r i c t i o n ratio and a p a r t i a l l y drained pore pressure response. Figure 3.13 shows the velocity measurements at the site. Characteristic of most organic rich deposits the velocity in the s i l t unit i s low. Compressional wave CONE BEARING Ot (bar) ID-(/) L QI 4-> QI E 200 d e s i c c a t e d SILT medium g r a i n e d SAND w i t h s i l t y SAND l a y e r s , o c c a s i o n a l o r g a n i c s l o o s e f i n e SAND X r— D_ LU a 2Di n o r m a l l y c o n s o l i d a t e d ! c l a y e y SILT PI=10 LL=35 sand 10% s i l t 70% c l a y 20% 30-1 Figure 3.11 CPTU Profile at McDonald Farm to SOIL DENSITY SHEAR WAVE VELOCITY (MG/CU |) Q . (M/SEC) 0 BEARING RESISTANCE DYNAMIC SHEAR MODULUS 500 0 " ( B R R S ) 200 0 ( M P f l ) 200 Figure 3.13 Velocity Profile at Pile Load Test Site 64 testing was attempted but i t was not possible to transmit compressional wave energy through the surface s o i l s . Pore pressure dissipation rates at this site are not reported here but appear to have been influenced by partial saturation. Sluggish response of the piezometer and a time lag before dissipation may have been due to either poor piezometer saturation or incomplete s o i l saturation. The presence of organics and the i n a b i l i t y to transmit compression wave energy through the s o i l both indicate that incomplete s o i l saturation was l i k e l y the cause of the sluggish pore pressure response. This application of compressional wave measurements i s discussed further in chapter 7. 3.9 Richards Island Site: Site Description The Richards Island site i s located in the coastal zone of the Beaufort Sea. The site investigation details are reported in Campanella, et a l . (1987). The purpose of the site investigation was, in part, to evaluate the s u i t a b i l i t y of CPTU equipment to delineate stratigraphy in the winter ice accessible coastal zone of the Beaufort Sea. The shallow water depths allow the use of surface based CPTU equipment but determination of shear wave velocities presented d i f f i c u l t i e s and required the application of explosive sources. The s i t e i s characterised by 2 m of sand and s i l t over an overconsolidated s i l t y clay which extends down to 4.5 m. Below the clay unit down to 8 m i s a dense fine sand unit. The sand i s sufficiently fine to generate a pore pressure response during cone penetration. Below the sand unit a s i l t y clay unit extends down to a 65 hard ice bonded material that could not be penetrated. Surface s o i l s near the mouth of the Mackenzie River are often seasonally frozen by subzero saline water overtop of sediments flushed of salt during the freshette of the previous spring. The CPTU profile, Figure 3.14, shows a thin frozen surface layer indicated by the spike in the cone resistance that also generated higher pore pressure. Temperature measurements at the site taken using a thermistor i n the cone and in adjacent bore holes by others, H i l l , et a l . (1986), showed that the entire borehole was at subzero temperatures but did not indicate the presence of ice bonded s o i l s . The s a l i n i t y and grain size of the s o i l results i n an unfrozen behaviour. The high cone t i p resistance layer at 11 m may, however, result from ice bonding of a low sa l i n i t y sand layer. Very low t i p resistance measured just below this thin layer may result from a b r i t t l e fracture mechanism. Other CPTU measurements give no consistent explanation of the type of sediment at 11 m. Sampling of these marginally frozen sediments may not have preserved marginally bonded layers. Details of the pore pressure response at this site are reported in other chapters. 3.10 Schoolhouse Site: Site Description The schoolhouse site i s located near the hamlet of Tuktoyaktuk. Four CPTU's were performed on a line offshore from this s i t e . Soil conditions at each of these sites were characterised by loose to dense fine and medium grained sands. Velocity measurements were made at one of the soundings and are described by Campanella, et a l . (1987). 99 67 Although an average velocity profile was successfully obtained using explosive techniques, detailed velocity measurements were not possible. Figure 3.15 shows results from adjacent CPTU records with pore pressures measured at different locations on the cone. Pore pressure measurements at the site clearly indicated slight variations in grain size not indicated by f r i c t i o n measurements. With an increase in fines content and associated decrease i n permeability response changed from drained to p a r t i a l l y drained. Through those s o i l s sufficiently fine to record a pore pressure response, pore pressures behind the t i p were always less than static (static conditions were measured following pore pressure dissipations to equilibrium). This pr o f i l e was the only one encountered in this study that resulted in negative pore pressures on the cone face. 3.11 Swimming Point Site: Site Description Some details of the Swimming Point site are discussed i n this thesis. Other descriptions of the site are given i n Campanella, et a l . (1987). The site i s located on the Mackenzie River and testing was performed through the winter ice. Two probings were completed in the river channel and two more nearby at the edge of the channel where winter ice was frozen down to the river bottom. The sites at the edge of the channel were important i n evaluating the results obtained from velocity measurements made with explosive sources because partial contact of the ice onto the s o i l allowed the transmission of shear waves generated using the hammer beam source. The s i t e consists of an organic rich, very loose s i l t y sand deposits. Figure 3.16 shows two PORE PRESSURE SLEEVE FRICTION U (m. of vater) (bar) 0 100 0 2.5 0 0-CONE BEARING Oc (bar) Depth Increment i . 025 m Figure 3.15 CPTU Profiles at Schoolhouse Site FRICTION RATIO DIFFERENTIAL P.P. Rf tt> RATIO 4U/Qc 0 5 -.2 0 .8 "" < ' ' ' INTERPRETED PROFILE ID 15 10 15 D O V dense SAND w i t h s i l t y sand l a y e r s note: t h a t a l t h o u g h pore p r e s s u r e s on f a c e a r e g e n e r a l l y p o s i t i v e t h e r e a r e o c c a s i o n a l zones o f n e g a t i v e pore p r e s s u r e 15 Max Depth i 32.65 m <7\ CO Figure 3.16 CPTU Profiles at Swirandng Point Site. 70 adjacent CPTU profiles. Boring logs from sampling at the site performed by Kurfurst (1986) show fine sand throughout the entire pr o f i l e . The CPTU profile also shows s i l t y layers, evidenced by lower t i p resistance and p a r t i a l l y drained pore pressure response. The probings stop on a gravel layer. More detailed CPTU records are shown in chapter 6. 3.12 Lahgley Sites: Site Descriptions Two different adjacent sites in Langley, BC were used in this study. The s o i l conditions are similar, except that the apparent overconsolidation of the surface clay deposits was l i k e l y caused by different mechanisms. At the Langley 232nd interchange site, the surface clays were compacted during construction of a freeway overpass. Clays below the zone of influence of the compaction process behave in a normally consolidated manner. At the Langley research site, excavation for road construction under a r a i l overpass has unloaded the surface clays, which appear to be l i g h t l y overconsolidated. Pore pressure response at each of these sites i s discussed later i n this thesis. Typical of many clay sites that have water tables below the surface, some d i f f i c u l t y was encountered in maintaining saturation of the pore pressure measurement system. Complete descriptions of the s o i l conditions were shown by Greig (1985), who obtained continuous samples at the Langley research sit e . Remarkably different values reported by Greig for the two sites were reinvestigated by the author and have been attributed to measurement errors with the t i p resistance in the soft clays. Figure 3.17 shows a w L QI 4-> QI E 0_ LU Q PORE PRESSURE U (m. of water) 0 100 15 22.5 SLEEVE-FRICTION (bar) 0 .25 7.5 behind t i p behind s l e e v e J 22.5 7.5 15 22.5 CONE BEARING Qc (bar) 50 1 s u Depth f i e l d Gmax (m) vane avera average (MPa) (kPa) 3 23 14 3.5 22 16-4 25 17 1 4.5 25 19 5 26 20 5.5 27 21 6 36 21 6.5 29 24 7 26 25 7.5 30 26 8 30 30 9 29 33 10 37 34 11 37 38 12 33 44 13 38 41 14 35 39 15 42 41 16 38 45 17 38 46 18 35 47 FRICTION RATIO Rf (X) 0 5 -J 1 1 L. Depth Increment : . 025 m •22. 5-1— Max Depth DIFFERENTIAL P.P. RATIO 4U/Qc 0 1 0 7.5-15-22. 5 18. 9 m Figure 3.17 CPTU and Soil Profile at Langley Research Site 72 CPTU pro f i l e from the Langley research site as well as averaged strength and stiffness information. 3.13 Brenda Mines: Site Description Piezometer cone testing was performed at the Brenda Mines tailings dam and pond to evaluate s o i l characteristics and seepage conditions. The s o i l conditions were characterised by either l i g h t l y compacted medium grained sand in the dam, which was dry to moist or loose fine s i l t y sand tailings hydraulically placed in the ta i l i n g s pond. An unusual condition i n the tailings pond was the existence of remnant frozen layers buried during previous winters. More details of the site investigation are reported by Campanella, et a l . (1985). The site, with i t s uniform s o i l properties and very high stresses, i s of interest to the investigation of stress level effects in cone penetration testing. The investigation also highlighted one of the key advantages of the cone penetration test i n granular s o i l s , which i s the rapid determination of equilibrium water pressures. Maximum penetration at the site was almost 70 m in the main body of the dam; this was achieved with approximately 9,000 kg of thrust. This CPTU record i s shown in Figure 3.18. 3.14 Heber Road 2, 4, 6: Site Descriptions The Heber Road site i s located i n the Imperial Valley, California. The site i s of interest because of liquefaction related damage caused by the El Centro earthquake of 1979. The greatest uncertainty in the interpretation of measurements made at this site i s associated with the Friction Ratio Friction, f e ,bar Cone Beoring,qc,bar F R ( f c / q c ) , % 3 2 I —0— 80 160 240 320 4 0 0 O I 2 0 ( Ibar = lOOkPo) 70 3 2 ° < p : 3 4 ° a. a> •o c a> 3 O o» O in O a. e o c >N "O i_ o u a (A •o a> > w a> tn X) o o Z medium dense tailings SAND coarse fraction from cyclone separation hydraulically placed from slurry) bulldozer compacted Figure 3.18 CPTU Profile at Brenda Mines Site from Campanella. Robertson, Gillespie and KLohn (1984) 74 calculated and measured surface accelerations. The E l Centro earthquake was a magnitude 6.6 but early reported c y c l i c stress ratios of 0.75, Youd and Bennett (1983), have been reduced to approximately 0.40, Youd (1983), based upon analysis of acceleration measurements made at nearby sites. The CPTU results were obtained post liquefaction, but are assumed to be representative of conditions prior to the earthquake. The very recent nature of this deposit increases the uncertainty i n this assumption. Based on the interpretation of Youd and Bennett (1983) the site transects an abandon river channel. The entire p r o f i l e i s believed to be very recent. Youd and Bennett give an age of only 300 - 400 years. At a l l three cone locations, a surface sand f i l l extends to a depth of approximately 1 m. Below the f i l l i s a channel sand deposit, which is believed to have l i q u i f i e d during the E l Centro earthquake, Youd and Bennett (1983). This unit, referred to by Youd and Bennett as unit A 2, is of interest i n this study. Although properties of this unit vary greatly and averaged values must be used with caution the cone t i p resistance values, shear wave velocities, and N values a l l consistently indicate a very loose fine sand deposit. Below the sand unit, which is 2 to 5 m deep, are either interbedded clay and sand units or dense sand. The clay i s plastic and heavily overconsolidated. Complete site descriptions are given in Youd and Bennett, (1983). Figures 3.19, 3.21, and 3.23 show the CPT results, and Figures 3.20, 3.22, and 3.24 also include the velocity results. Friction measurements in the loose sand deposit were very low and subject to PORE PRESSURE CONE BEARING DIFFERENTIAL P.P. U (m. of water) Qt (bar) RATIO AU/Qt Figure 3.19 CPTU Profiles at Heber Road 2 Site SOIL DENSITY SHEAR WAVE VELOCITY (MG/CU M) 1.5 Z.O 0 ' 500 0 200 0 • lQOO BEARING RESISTANCE (BARS) DTNAMIC SHEAR MODULUS (BARS) SOIL PROFILE I I I I I I I I I C i ' ' " i I 4 / 1 I i J i ' r i i i i i I w I i i v i i r ' i i v - | ^ i i i i i i i i i i HEBER ROAD 2 . P C 2 . 1 S T C R O S S . 1 0 SO. CM S E I S M I C CONE MARCH 1 2 . 8 4 DG SAND F I L L l o o s e f i n e SAND D50=. 10-. 12 f i n e SAND s t i f f CLAY Figure 3.20 Velocity Profile at Heber Road 2 Site PORE PRESSURE CONE BEARING DIFFERENTIAL P.P. U (m. of «ater> Ot (bar) RATIO AU/Ot Figure 3.21 CPTU Profiles at Heber Road 4 Site SOIL DENSITY SHEAR WAVE VELOCITT BEARING RESISTANCE (MG/CU CU . . (M/SEC) C n„ n (BARS) 500 0 200 0 DYNAMIC SHEAR MODULUS (BARS) 1000 SOIL PROFILE i I I I I I I I I m I I I I I I I I I I . HEBER ROflD 4 . P C 4 . 1 S T C R O S S . 10 S O . CM S E I S M I C CONE MRRCH 1 2 . 8 4 DG SAND F I L L l o o s e f i n e SAND D50=. 10-. 14 s t i f f CLAY Figure 3.22 Velocity Profile at Heber Road 4 Site 0 3 Figure 3.23 CPTU Profiles at Heber Road 6 Site SOIL DENSITY , . 5 ( M G / C U S'.O SHEAR WAVE VELOCITY (M/SEC) BEARING RESISTANCE DYNAMIC SHEAR MODULUS l B R R S ) 200 0 ( B f l R S ) -I—i—i—i—i—i—i—i—i—I ">~i—i—i—i—i—i—i—i—i—i—I —i i i i i i i i r . HEBER ROAD 6 . P C 5 . 1 S T C R O S S . 1 0 S Q . CM S E I S M I C CONE MARCH 1 2 . 8 4 DG Figure 3.24 Velocity Profile at Heber Road 6 Site SOIL PROFILE SAND FILL loose fine SAND D50=.10-.14 dense SAND s t i f f CLAY 03 o 81 s t a b i l i t y related errors discussed in chapter 2. They were useful, however, in delineating the highly variable stratigraphy. 3.15 Wildlife Site: Site Description The w i l d l i f e site, l i k e the Heber Road sites, i s located in the Imperial Valley, California. The w i l d l i f e site also suffered liquefaction damage during the 1979 E l Centro earthquake. The stratigraphic unit known to have l i q u i f i e d was identified by Bennett, et a l . (1981) by noting the colour of the sand extruded in sand boils. Similar to the Heber road site the s o i l s at the site are very recent. Figures 3.25 and 3.26 show CPT and velocity data from the sit e . Additional CPT results are shown elsewhere in the thesis. 3.16 New Westminster Site: Site Description The New Westminster site i s located near the Fraser River in New Westminster, BC. The site i s comprised of hydraulic f i l l that i s believed to be loose. Below a compacted surface f i l l , the s i t e i s b u i l t up with layers of interbedded fine s i l t y sand and sandy s i l t . The fines content i s sufficiently high that the pore pressure response i s l i k e l y undrained. As shown on Figure 3.27, the time for excess pore pressure dissipations, recorded behind the ti p , to reduce to one half of their i n i t i a l value, ( t 5 0 ) , was approximately 50 to 200 seconds in the sand and s i l t unit. With t 5 0 values of approximately one minute, the time required to measure equilibrium pore pressures was approximately ten minutes. Figure 3.27 also shows the pore pressure response measured in three 82 PORE PRESSURE U (•. of .oter) 100 r ~ 0 2.5mm 2.5mm thick element, "B" -5 mm thick element, "c 15* CONE BEARING Ot (bar) 10 15 5 mm 200 FRICTION RATIO Rf (I) 0 10 INTERPRETED PROFILE 10 15 SILT SILTY SAND CLAYEY SILT STIFF SILTY CLAY Max D e p t h . 1 1. 63 m Figure 3.25 CPTU Profiles at Wildlife Site from Campanella, Robertson and Gillespie (1986) SOIL DENSITY 1.5 ( M G / C U ^ . 0 _| I I I L SHERR WAVE VELOCITY (fi/SEC) m _ 0 1 i 1 1 1 1 BERRING RESISTANCE (BARS) DYNAMIC SHEAR MODULUS l n n n (BARS) 200 0 1000 0 1 I 1 1 1 1 1 1 1—1—1—I w ~ l — i 1 1 1 i 1 1 1—r W I L D L I F E . P C I . 1 S T C R O S S . 1 0 S O . CM S E I S M I C CON 1—rn—1—1—1—1 1 1 MARCH 1 2 . 8 4 DG SOIL PROFILE l o o s e i n t e r b e d d e d sandy t o c l a y e y SILTS l o o s e s i l t y SAND D 5 0=.O56-.ll s t i f f CLAY Figure 3.26 Velocity Profile at Wildlife Site CO 85 different places. The pore pressure response, measured on the face of the cone, i s highly influenced by stops in the push. The effect of unloading the rods at the surface diminishes with depth. The most detailed indication of the s t r a t i f i e d nature of the deposit i s obtained from the t i p resistance. 3.17 Summary The s o i l s encountered during this study cover a wide range of types. Different geologic and a r t i f i c i a l origin, density, grain size, temperature and stress levels have been investigated. They provide a basis for the comments and observations given i n later chapters of this thesis. 86 CHAPTER 4. FACTORS AFFECTING PORE PRESSURE MEASUREMENTS 4.1 Introduction There are many factors that effect the measurement of dynamic (penetration) pore pressures in the piezometer cone penetration test. This chapter addresses only those issues related to equipment and procedural details. Chapter 5 addresses the s o i l factors that effect the measurement of pore pressures by investigating the correlation between s o i l parameters and dynamic pore pressures. In that the importance of equipment and procedural details i s a function of the s o i l type and properties, the influence of s o i l properties i s also discussed where necessary in this chapter. 4.2 Effect of Measurement Location The single most important parameter effecting pore pressure measurement i s the location of the pore pressure sensor. There i s considerable literature that addresses the selection of location and possible advantages of standardization, including Campanella, et a l . (1985) and Jamiolkowski, et a l . (1985). No single pore pressure measurement location i s best for a l l s o i l types or applications and data w i l l be shown indicating that the decision should depend upon the site and application. Data have been collected at several sites with pore pressure elements located at the centre of the cone face, behind the t i p , and behind the f r i c t i o n sleeve. Figure 4.1 shows a summary of some of these results for a variety of s o i l types. The importance of the element location i s a function of s o i l type. The data on Figure 4.1 TARANTO CLAY - Cemented(CaC0 3 l5-30%)( Italy) LONDON C L A Y - St i f f , uncemented, fissured (NGI - BRE) Figure 4.1 Pore Pressure Distribution during CPTU from Robertson, Campanella, Gillespie and Greig (1986) CO 88 have been normalized by dividing by the equilibrium water pressure, (u 0)• In normally consolidated insensitive clays and s i l t s , where large positive pore pressures are generated during shear, pore pressures measured on the face of the t i p are usually approximately three times larger than the equilibrium pore pressure and about 15 % greater than the pore pressure immediately behind the t i p . As the overconsolidation ratio increases in clays and s i l t s the excess pore pressure on the face of the t i p increases. This i s because the area around the face of the t i p i s a zone of maximum compression and high shear whereas the area behind the t i p i s a zone of normal stress decrease. Both areas have large shear stresses and associated pore pressures but the large normal stresses dominate the pore pressure response on the face, whereas the shear stresses become more important behind the t i p . In fine granular materials a pore pressure response may also be observed. Provided penetration i s not completely drained, pore pressures measured behind the t i p are generally negative even in loose sands. As shown in Figure 4.1 pore pressures measured on the face can reach very high values because of the high normal forces associated with this location. The data shown in Figure 4.1 indicates that, regardless of s o i l type, the pore pressure distribution along the face i s insensitive to the exact location, either the very t i p , mid height, or some other location on the face. Lunne, et a l . (1986) report that pore pressures are very similar when measured on the cone t i p or at the middle of the cone face. This i s an important finding that allows comparison of 89 pore pressure data collected from different cone designs each having their pore pressure element located somewhere on the face. This i s the primary advantage of measuring pore pressures on the cone face. Data collected in some overconsolidated s o i l s indicate that the large gradient in pore pressure immediately behind the t i p results in a variation in the measured pore pressure for sli g h t l y different measurement locations. For example, Figure 4.2 from the Imperial Valley, California, shows very different pore pressure readings when measurements were repeated with only slight differences i n measurement locations. Although both values were measured behind the t i p , those closest to the shoulder were much higher than those just 1 mm above. This i s an indication of the high pore pressure gradient that exists at the shoulder. This gradient precludes the application of pore pressure measurements made at this location for quantifying the properties of s t i f f clays. Data collected simultaneously behind the t i p and behind the sleeve indicates that pore pressures are generally sli g h t l y lower behind the sleeve than behind the t i p . This i s due to a combination of two factors, the increased shear induced component behind the sleeve, and the pore pressure dissipation that occurs during the time required for the sensor behind the sleeve to travel the distance between the two sensors, Gillespie (1981). The relative importance of the two factors depends upon s o i l type. The s t i f f clay s o i l s that generate negative pore pressures behind the t i p also have large cone resistance values which can be reliably measured. The focus of the interpretation of CPTU results in these soils remains the cone resistance and f r i c t i o n measurements. 90 PORE PRESSURE U <k of water) CONE BEARING Ot (bar) FRICTION RATIO Rf (X) INTERPRETED PROFILE U) L 01 +J Ql E 0_ LL) O 15 200 -5 mm thick element, C 15 I I I I I 10 0 »- • •— 1— 0 SILT 5 1 L s SILTY SAND CLAYEY SILT 10 1 10 STIFF SILTY CLAY 15 15 Depth Increment i .025 m Max Depth • 11.63 m Figure 4.2 Penetration Pore Pressures at Imperial Valley Site from Campanella, Robertson and Gillespie (1986) 91 Figure 4.3 shows data collected in normally consolidated clayey s i l t for three measurement locations. The difference in pore pressure response, in this type of s o i l , i s primarily due to variation in the normal stresses rather than to variation in the shear induced component. Values such as these are characteristic of normally consolidated s o i l s and are not sensitive to exact details of the measurement location as was shown to be the case in overconsolidated so i l s such as those at the Imperial Valley site, Figure 4.2. Pore pressures sensors located behind the sleeve may offer the advantages of those behind the t i p : robustness; sensitivity to stress history; and insensitivity to rod break procedures (maintaining t i p resistance, clamping top of rods, or allowing top of rods to rebound in an unconstrained manner) while being insensitive to i t s exact location. As shown i n Figure 4.3, in normally consolidated s o i l s this measurement location also shows nearly the same stratigraphic detail as other measurement locations. 4.3 Effect of Cone Design and Mechanical Details The extreme sensitivity of pore pressure measurements to the f i l t e r element location has already been discussed. For any given location there are s t i l l several other considerations that influence the r e l i a b i l i t y and accuracy of results. When penetrating high cone resistance s o i l s there i s a possibility that a f i l t e r element w i l l be subjected to squeezing forces from either: - load transferred from the t i p through the element - direct s o i l forces. Figure 4.3 Detailed Penetration Pore Pressures at McDonald Farm 93 Jamiolkowski, et a l . (1985) report that t i p load transfer through the element has been observed with several cone designs. This problem i s easily checked by rapidly loading a f u l l y assembled de-aired cone; i t i s most common on face t i p piezometer cones. Rapidly loading the cone t i p also ensures that the pore pressure transducer i s mechanically isolated. F i l t e r element squeeze can also be important for cones that measure pore pressure at the t i p or on the face. During the i n i t i a l application of load, positive pore pressures may result. I t i s suggested that the f i l t e r element permeability as well as i t s compressibility must be considered together. For example, i f the f i l t e r element i s squeezed during penetration into a high cone resistance unit, then some positive pore pressure w i l l result unless the f i l t e r and s o i l are of sufficient permeability to dissipate this additional pore pressure. Experience gained during this study with a compressible porous plastic element has not shown any d i f f i c u l t i e s related to f i l t e r squeeze, l i k e l y due to the high permeability of both the element and those soils having sufficient cone t i p resistance to induce element squeezing. Penetration into very fine dense sands or compact glacial t i l l s are most l i k e l y to result in element squeeze d i f f i c u l t i e s . To investigate the importance of element squeeze effects and to ensure that positive pore pressures observed at the McDonald Farm site with face pore pressure elements were not the effect of element squeeze, repeated soundings were made with plastic and ceramic porous f i l t e r s . The comparatively soft plastic f i l t e r s were made from nominal pore size 120 micron polypropylene and the ceramic f i l t e r s were 94 machined from a very s t i f f nominal pore size 10 micron ceramic "aerolith". The results of repeated soundings are shown in Figure 4.4. Positive pore pressures are observed with either f i l t e r type and allowing for natural v a r i a b i l i t y the profiles indicate that use of the soft plastic f i l t e r does not result i n induced pore pressures. A variety of ceramic elements were used in this study at several sand sites. In most cases when the sounding was completed the ceramic f i l t e r element was missing and presumed broken. Surprisingly, in sands i t was impossible to t e l l , even at well documented sites with repeated soundings where within the profile the ceramic elements were lost. It appears that sand f i l l s the void l e f t by the broken f i l t e r and forms a new f i l t e r . Careful examination of the pore pressure response can generally identify any f i l t e r element squeeze problems. For example, at the beginning of a push into dense sand pore pressure generated from compressibility of the f i l t e r should dissipate very quickly. Therefore, i f a spike i n the pore pressure i s observed at the beginning of a push into a uniform material element squeeze may be a problem. If, however, within a uniform s o i l large positive pore pressures are maintained, then they can be considered r e a l i s t i c . An example of this type of behaviour i s shown in Figure 4.5 from Richards Island. Within the sand unit a high positive pore pressure i s maintained for several metres. Rapid dissipation of excess pore pressure during rod breaks within this sand unit indicates that the s o i l i s nearly free draining and i t can therefore be concluded that the measured values are the result of the high normal stresses created during penetration. 95 PORE PRESSURE U (m. of water) Depth Increment s .025 m Max Depth : 21.1 m Figure 4.4 Effects of Element Compressibility at McDonald Farm Figure 4 . 5 Detailed Penetration Pore Pressures at Richards Island 97 4.4 Element Saturation Inadequate saturation of the pore pressure measurement system can lead to inaccurate or sluggish pore pressure measurements. Profiling details w i l l be missed and pore pressure decay rates w i l l be inaccurate. Confidence i n the data i s encouraged by both a rapid return to pre rod break values after pore pressure dissipations and a dynamic response that i s also reflected by changes in either the cone resistance or f r i c t i o n measurements. In overconsolidated fine grained s o i l s large negative pore pressures are generated and measured behind the cone t i p . Commonly, through a desiccated surface crust or within dense fine sand layers negative pore pressures approach cavitation values. In these instances a cone that i s i n i t i a l l y saturated may lose saturation and subsequent performance, depending on s o i l conditions, may be affected. Two common examples of s o i l conditions through which saturation problems are commonly encountered i s shown in Figure 4.6 with measurements taken at different locations on the cone in adjacent holes. When pore pressures are measured behind the t i p , cavitation results in the fine sands at 12.5 to 13 m. Subsequent penetration pore pressure measurements are effected. Examination of the records from other sensors shown in Figure 4.6 shows that measurements made on the face are less susceptible to the cavitation problem and are more useful i n these s o i l conditions. At the same site saturation problems were encountered with the f i l t e r element behind the f r i c t i o n sleeve after penetration through the moist surface. After recording negative pore pressures, sluggish PORE PRESSURE U (m. of watar) Depth Increment i . 025 m Figure 4.6 Detailed Penetration Pore Pressures at Langley Site 99 response was observed. Sufficient resaturation at the higher pressures in the underlying soft clay resulted in satisfactory performance. When pore pressures are measured on the face, a zone of high positive pore pressure, as shown in Figure 4.6, the problem of element desaturation during penetration does not usually occur. In very heavily overconsolidated s o i l s the large positive pore pressures may require that penetration be stopped due to overloading of the pore pressure sensor when pore pressures are measured on the cone face. This situation was observed at a number of sites in overconsolidated clays and frozen s o i l s and can be solved by either measuring the pore pressure at a different location or using a higher capacity pore pressure transducer. Pore pressure dissipations are not necessarily diagnostic of saturation problems. High positive pore pressures are observed at the face of the cone in overconsolidated s o i l s and low values are observed behind, this gradient of pore pressure often results in dissipations recorded behind the t i p that increase before they decrease. Examples of two pore pressure dissipations recorded simultaneously in an overconsolidated clay are included in Figure 4.7b. Pore pressures at the cone face start to dissipate as soon as penetration i s stopped. The response seen behind the sleeve, an increase followed by a slower decay, i s due to the redistribution of excess pore pressures from in front of the cone in both an upward direction as well as the radial direction. This response i s in contrast to that generally observed in normally consolidated clay an example of which i s shown i n Figure 4.7a. 3 o -H -P (0 ft a) 1^ d ui (A CD ft a u o ft 10 9 8 -7 -6 -5 4- -3 -2 1 A) Mc DONALD FARM Normally consolidated clayey SILT (OCR = 1) Depth = 19.7 m CfL_ —B •Friction sleeve eaBannnnnnp U L J u u o a 3 D Q O G 6 6 - g 200 400 600 800 Time ( seconds ) 1000 1200 3 o •H -P (0 ft d) M 10 a) ft a) J-I o ft 10 9 8 7 6 5 BJRICHARDS ISLAND Over consolidated s i l t y CLAY (OCR = 8) Depth = 3.87 m • Friction sleeve I— 40 80 120 160 Time ( seconds ) — i r 200 240 Figure 4.7 Pore Pressure Dissipations, Examples from Two Sites from Campanella, Robertson and Gillespie (1986) 101 4.5 Effect of Cone Design and Procedure on Dissipation Tests The interpretation of pore pressure dissipation data can be done for a quantitative assessment of the horizontal coefficient of consolidation. Theoretical solutions exist for most pore pressure element locations. Commonly used solutions include the f i r s t , Torstensson (1977), and a more theoretical and complete solution, Baligh and Levadoux (1986). The effect of element location must be considered i n the interpretation of pore pressure dissipation data. A l l theoretical solutions for the interpretation of consolidation parameters rely on either a calculated or assumed i n i t i a l pore pressure distribution. Pore pressure values measured at different locations on the cone, either simultaneously or in repeated measurements, often bear l i t t l e relation to those assumed i n theoretical solutions. Dissipations recorded behind the t i p through overconsolidated s o i l s often show an increase before decreasing. This i s l i k e l y due to a redistribution of excess pore pressures. An examination of the v a l i d i t y of theoretically derived i n i t i a l excess pore pressures i s shown in Figure 4.8 which shows i n i t i a l excess pore pressures, at the cone surface, normalized by the excess pore pressure observed along the shaft. Figure 4.8 shows that the Baligh and Levadoux (1986) solution for the i n i t i a l excess pore pressure distribution represents an average of values between normally and l i g h t l y overconsolidated insensitive s o i l s . This i s an important requirement for dissipation analysis. In overconsolidated s o i l s where the i n i t i a l excess pore pressure distribution i s more complex, 102 N O R M A L I Z E D E X C E S S P O R E P R E S S U R E A U/(AU), SH 2.5 S Y M B O L D E P T H (f t ) O C R O 4 5 * 5 3 * 0 . 4 A 6 0 * 5 2 * 0 . 3 • 8 5 * 5 1.3*0.1 T E S T R E S U L T S F R O M B A L I G H e t a l . , 1978 • H A N E Y A L A N G L E Y • M C D O N A L D 3 . 0 3.5 4 . 0 A D D I T I O N A L D A T A F R O M U B C R E S E A R C H S I T E S O C R = 4 . 0 O C R = 1.0 O C R = 1.0 7 Predicted Distribution Baligh and Levadoux 1986 60° CONE Figure 4.8 Normalized Excess Pore Pressure Distribution from Gillespie, Rcfoertson and Campanella (1988) 103 dissipation solutions presently available are inappropriate. Uncertainty in the i n i t i a l excess pore pressures around the cone especially in the s o i l several diameters away from the cone makes the development of dissipation solutions d i f f i c u l t . Field measurements, especially around driven piles in overconsolidated s o i l s , may offer some insight into this uncertainty. Examples of pore pressure dissipations recorded at different locations i n normally consolidated s o i l are compared to theoretical solutions i n Figure 4 . 9 . In general theoretical solutions match the observed behaviour reasonably well at normally consolidated sites. Field test procedures can have a important influence on the quality of the pore pressure dissipation data. When the rods are unloaded at the beginning of a dissipation, a large normal stress r e l i e f may occur and pore pressures w i l l drop. Measurements on the face of the cone are much more sensitive to this problem than those behind the t i p or behind the sleeve. An examination of the cone resistance recorded during dissipations w i l l reveal any d i f f i c u l t i e s . Clamping the rods to reduce the total stress r e l i e f has been cited as a possible remedy to the problem, but the length of the rods and their elastic response makes i t impossible to control the stress at the t i p by controlling the displacement at the top. Maintaining the t i p stress (cone resistance) would not solve the problem since presumably the cone would continue moving. The influence of unloading the t i p stress i s especially important when pore pressures are measured on the face of a cone o < <3 1.20 1.00 0 . 8 0 0 . 6 0 -0 . 4 0 0 . 2 0 -0 . 0 0 Range of 10 Test Results Range of Test Results From McDonald Farm and Predicted Dissipation curve from Levadoux and Baligh (1986) c^= 8 mm/sec 1 -i—i—r—i i i |— 10 1 0 0 1 0 0 0 1 3 13 CO co CD Q_ CD O CL CO CO CD O X 1-.20 1.00 0 . 8 0 0 . 6 0 -0 .40 ' 0 . 2 0 -0 . 0 0 j Range of 10 Test Results V Range of Test Results From McDonald Farm and Predicted Dissipation curve from Levadoux and Baligh (1986) c^= 8 mm/sec 1 10 1 0 0 1000 "O CD E 1.20 1.00 0 . 8 0 0 . 6 0 0 . 4 0 0 . 2 0 -0 . 0 0 Range of 10 Test Results V Range of Test Results From McDonald Farm and Predicted Dissipation curve from Levadoux and Baligh (1986) c^= 8 mm/sec 10 Time (sec) 1 0 0 1 0 0 0 Figure 4.9 Pore Pressure Dissipations at McDonald Farm: Predicted versus Measured Dissipation curves recorded on the face of a 60° Piezocone from Gillespie, Robertson and Campanella (1988) 105 where the greatest unloading takes place and i s least important behind the sleeve. No theoretical solution presently available recognizes the drop i n total stress (cone resistance) that occurs when penetration i s stopped. 4.6 Conclusions 1) Pore pressure response i s primarily controlled by the location of the porous element. The significance of the element location depends on s o i l type and no single element location best serves a l l purposes. The variation in measured values was discussed i n this chapter. Specific details of the measurement system including element compressibility were also discussed. 2) Element saturation i s a significant problem following penetration through desiccated s o i l s ; in addition, penetration through saturated fine s o i l s may also cause cavitation and element desaturation. In many cases, pore pressure measurement on the cone face results in the generation of positive pore pressures and eliminates this problem. 3) Pore pressure dissipation rates were found to be highly influenced by measurement location and procedural details. Dissipation theory adequately accounts for the measurement location in normally to li g h t l y overconsolidated s o i l s but the stress r e l i e f associated with stopping penetration and relieving the cone resistance may result in misleading dissipation rates when pore pressures are measured on the cone face. Pore pressures measured behind the f r i c t i o n sleeve are least effected by this problem. 106 CHAPTER 5. INTERPRETATION OF PORE PRESSURE MEASUREMENTS 5.1 Introduction The previous chapter dealt with cone design and test procedure details that effect pore pressure data. It was shown that considerable care must be exercised in order that the data collected can be interpreted in a proper manner. In this chapter the interpretation of CPTU data i s discussed with an emphasis on integrating the pore pressure data into the traditional cone t i p resistance and f r i c t i o n based interpretation methods. It w i l l be apparent that so many s o i l factors effect the pore pressure data that i t s application must be restricted to specific s o i l conditions and subject to measurement details. There are some incentives for standardization of the pore pressure f i l t e r element location; easier comparison of data would be an obvious benefit. Bearing i n mind, however, that there i s no "best" measurement location the selection of measurement location should be based on l o g i s t i c a l considerations and the ultimate application. Some of these applications and necessary measurement details are outlined. 5.2 Cone Tip Resistance Corrections Other than the establishment of equilibrium pore pressures, the simplest and perhaps most important application of pore pressure measurements may be to correct cone resistance measurements to obtain a total stress measurement. Since different cones respond differently to the pore pressure acting behind the t i p they should a l l be corrected to q i j i (in a manner now referred to as net area ratio correction) by 107 q T qc + (1 - a) * u2 where <Jc measured cone resistance a net area ratio (determined experimentally by loading assembled cone hydrostatically) u2 pore pressure measured behind t i p This alone may j u s t i f y measuring the pore pressure behind the t i p . Lunne, et a l . (1986) showed that much of the scatter in q^ measurements, taken with different cones, can be eliminated by net area ratio corrections. Typical data from one s i t e are shown in Figure 5.1. This work cleary showed the importance of pore pressure corrections in soft clay. Aas, et a l . (1986) using the same data set as that discussed in Lunne, et a l . (1986) show that much of the scatter in previously published cone factor, NK, correlations i s due to the different net area ratios of the cones used to collect g^-. values. Measurements of pore pressure taken on the face have sometimes been used for q T corrections, these values are f i r s t reduced by 10 to 20 %. The wide variation in pore pressure distribution around the cone in a l l but normally consolidated soils reduces the accuracy of q T when the total stress correction i s made in this manner. Fortunately, those so i l s with a wide variation in excess pore pressures around the t i p also are generally overconsolidated. These so i l s have a higher cone resistance and q T corrections are not as significant. Pore pressure measurements taken behind the sleeve, u3, could be used to calculate the correct pore pressure, u2, to use for q T corrections using the observations of the pore pressure distribution previously shown in Figure 4.8. LEGEND A Average of Bat B BRE 5kN C BRE 50 kN D DELFT E FUGRO piezocone F FUGRO f r i c t i o n cone G FUGRO subtraction cone M MCCLELLAND 10 cm2 N MCCLELLAND 15 cm2 U UBC No. 4 S UBC seismic cone K v d BERG f r i c t i o n cone L v d BERG piezocone W WISSA Figure 5.1 Correction of Cone Resistance Data, Onsoy Site adapted from Lunne, Eidsmoen, Gillespie and Howland (1986) 109 5.3 Soil Classification From Dynamic Pore Pressures At the standard rate, 2 cm/s, penetration i n medium grained clean sands and coarser materials takes place under completely drained conditions. Penetration into fine sands, s i l t y sands, and s i l t s generally takes place under p a r t i a l l y drained conditions, finer materials are penetrated under completely undrained conditions. An example showing a l l three conditions i s shown in Figure 5.2 which shows the results of two adjacent CPT profiles from McDonald farm. In the sand unit from 2 to 13 m penetration takes place under nearly completely drained conditions. Pore pressures sl i g h t l y above the hydrostatic value are observed with the face t i p pore pressure element location. These values are generally highest through the s i l t y layers within the sand unit where less drainage occurs. A p a r t i a l l y drained response i s also observed from 13 to 15 m in a fine s i l t y sand unit. When penetration i s stopped to add push rods these excess pore pressures are observed to decay in less than 30 seconds. Penetration pore pressures shown in this profile are characteristic of most sands. If the permeability i s sufficiently low, and any pore pressure response i s observed, i t i s generally less than hydrostatic behind the t i p and greater than hydrostatic on the face of the cone. These pore pressures can be useful to characterize the sand but are very small in comparison to the cone resistance. The clayey s i l t unit below 15 m i s penetrated under completely undrained conditions, the pore pressure response shown in Figure 5.2 i s characteristic of normally or l i g h t l y overconsolidated s o i l s . The pore pressure values are highest on the face and decrease with distance up the shaft. A completely undrained response was, i n 110 Figure 5.2 Urrirained, Partially Drained, and Drained Response at the McDonald Farm Site I l l t his case, confirmed by slowing the rate of penetration and observing no change i n the pore pressure measurement u n t i l rates where changed by two orders of magnitude, Campanella, et a l . (1983). The use of pore pressure parameters to cla s s i f y s o i l type has been proposed by Jones and Rust (1982) and Robertson, et a l . (1985). These charts have been developed from data collected behind the cone t i p . Figure 5.3 shows one classification system for the interpretation of s o i l type from cone resistance and pore pressure data. With the great number of factors that affect pore pressure data i t was not found possible to identify a s o i l type solely on the basis of the cone resistance and pore pressure response. Tip resistance and f r i c t i o n measurements were used to classify the material and pore pressure data were used to distinguish detailed stratigraphy and to confirm the cone resistance and f r i c t i o n based interpretation. The intended application of Figure 5.3 i s for f i e l d c l a s s i f i c a t i o n of s o i l type, as such i t does not use normalized resistance values. The detection of small stratigraphic changes within a s o i l profile may be easier with pore pressure data, which i s thought to be dependent on a smaller zone of s o i l than t i p or f r i c t i o n data. It appears that in overconsolidated s o i l s the detection of stratigraphic changes i s best made with the pore pressure element on the cone face. An example from the Imperial valley in California, Figure 5.4, shows much more detail within the clay unit for the pore pressure element on the face than behind the t i p . In normally consolidated s o i l s i t appears that either location gives similar detail. For example Figure 5.2 shows pore pressure measured at 2 locations, where nearly similar detail i s 112 Zone Soil Behaviour Type 1 sensitive fine grained 10 - 500 2 organic material 2 - 2 0 3 clay 10 - 100 4 s i l t y clay to clay 5 - 10 5 clayey s i l t to s i l t y clay 2 - 5 6 sandy s i l t to clayey s i l t 1 - 2 7 s i l t y sand to sandy s i l t . 5 - 1 8 sand to s i l t y sand 0 -.5 9 sand d r a i n e d 10 gravelly sand to sand d r a i n e d 11 very s t i f f fine grained* unknown 12 sand to clayey sand* unknown * overconsolidated or cemented. Figure 5.3 Soil Behaviour Type Classification from Campanella, Robertson, and Gillespie (1986) 113 Figure 5.4 Pore Pressure Response at Wildlife Site from Campanella, Robertson and Gillespie (1986) 114 observed in a l l profiles. Note that the p r o f i l e obtained from the face t i p element i s also highly influenced by the effect of stopping to add additional cone rods. Within an overconsolidated s o i l p r o f i l e , i f the presence of slight variations i n s o i l type i s of interest, i t appears that the face t i p element location i s most useful. At most normally .consolidated sites pore pressure measurements made at different locations show similar detail. 5.4 Soil Classification From Pore Pressure Dissipation In s o i l s that generate pore pressures, either higher or lower than the hydrostatic value, an index of the permeability or grain size can be obtained from the rate of decay of the excess pore pressure. The excess pore pressure distribution in granular s o i l s i s complex, unknown and highly dependent upon s o i l type; therefore, theoretical dissipation solutions similar to those used in fine grained s o i l s are not feasible. An alternative means of interpreting dissipation rates i s to empirically correlate dissipation rates to material type, permeability, percentage fines or some other useful characterization of s o i l type. One expression of the rate of decay, which has been found to be useful i s the time to 50 % equalization of excess pore pressure, t$Q. These values are useful i n distinguishing s o i l s that have similar t i p resistance and f r i c t i o n ratios. Figure 5.3 i s augmented with the addition of typical pore pressure decay rates, expressed as the time for 50 % dissipation, which were compiled from a large variety of sites. A d i f f i c u l t s o i l classification problem commonly encountered i s the 115 distinction of s i l t s and s t i f f clays. Very often, due to accuracy problems with f r i c t i o n data, these two soils can not be distinguished. Application of the rates of decay of the excess pore pressure can then be used to c l a r i f y the interpretation. Guidelines for the selection of s o i l type based upon pore pressure dissipation rates are included in Figure 5.3. These averaged values are compiled from a large number of sites and are restricted to decays rates recorded behind the t i p with 10 cm2 cones. The location of the pore pressure measurement system and procedural details must also be considered in the interpretation of dissipation rates. Very often the drop i n total stress causes an apparent very rapid decay in excess pore pressure measured on the face of the cone. This can be misinterpreted as a dissipation in coarse grained materials. Inspection of the drop in cone resistance and the use of a different measurement location can document or reduce this d i f f i c u l t y . The integration of pore pressure decay rates and shear wave velocity data i s further discussed i n chapter 7. 5.5 Effect of Soil Sensitivity on Pore Pressure Measurements In overconsolidated s o i l s there has been an observed general trend for very high pore pressures on the face and low to negative pore pressures behind the face. Figure 5.5 shows results obtained in materials of variable sensitivity. The data are presented i n terms of the pore pressure parameter Bq, Senneset and Janbu (1984), where B q = (u - u Q) / (q T - OV) and B q l i s calculated using the pore pressure on the face, u l 2.0 -> • cr o =? ,.0 c r CQ 0.0 116 x°<9> °. ****** & * * t Cx L + x1-O + * + o * * * * * o o o I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 2 3 4 5 - 6 7 8 9 1 0 2.0 > o Z> C\J cr CQ * " X oo 0.0 | i i i i | i l O y ^ f q a i ->i ^  |»i I»I i |»i i i i» | i i i i | i i i i | i i i i | i i i i | 1 2 3 4 5 6 7 8 9 1 0 1.5 > to i c r O ZD I CO =2. CO cr CQ 0.5 -f -0.5 LEGEND: * langley Sites, $ Brent Cross:Lunne et al.(1986). O Haney, • Gloucester, St Marcel, Varennes, NRCC, STP, all Konrad & Law(l987) 0 Onsoy,-s!rSt.Alban:Roy et al.(l 982), + Boston Blue Clay:Baligh & Levadoux(1980) x Haga, A Drammen, * McDonald Farm ' * V * 0 < n b ° * * ** * o o* o o * * o 0 o o o I 1 ' ' ' I ' ' ' I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 1 2 3 - 4 5 6 7 8 O C R u ' ' i ' I 9 10 Figure 5.5 Pore Pressure Parameters Bg vs OCR 117 Bg 2 i s calculated using the pore pressure behind the t i p , u2 Bq 3 i s calculated using the pore pressure behind the sleeve, u3 For any given stress history a wide variation i n either of the pore pressure parameters can be seen. The variation in pore pressure parameter Bg with OCR has lead to the suggestion that a pore pressure ratio may lead to a good correlation for stress history. The effect of sensitivity, and to a lesser degree stiffness ratio, i s to obscure this correlation. Whereas low or negative pore pressures are measured behind the face in highly overconsolidated clays (for example, Figure 5.4) penetration in sensitive clays generally results in positive pore pressures regardless of measurement location. As a general trend i t i s observed that for any stress history, Bg tends to increase with sensitivity but this aspect was d i f f i c u l t to quantify because of a lack of sensitivity data collected in a similar manner. 5.6 Interpretation of CPTU Data For Stress History Stress history has been shown to affect the pore pressure response. An evaluation of the correlation between Bg and OCR was performed where good quality f i e l d data were available. Figure 5.5 presents the pore pressure parameters Bg l f Bg 2, Bg 3 against the best estimate of OCR. The data shown in Figure 5.5 indicate that although B q generally decreases with increasing OCR the scatter in the data reduces the usefulness of this approach. Jamiolkowski, et a l . (1985) reported a similar scatter in the Bg-OCR relationship when data from several sites were considered. Neither of the pore pressure element locations appears to give useful indications of OCR from B a data. 118 Other pore pressure parameters investigated include pore pressure ratios and differences, each with different normalizing parameters. The ratios and differences were thought to be useful after observing that pore pressure measurements behind the t i p were much more sensitive to stress history than those on the cone face. Sully, et a l . (1988) proposed the use of a normalized pore pressure differences. The ratios (ul-u3)/o' v and A U I / A U 3 and A U I / A U 2 were investigated i n this thesis. These ratios were observed to be highly dependent upon stress history at some sites. The application of these parameters to predict stress history was tested by plotting the ratios against best estimates of stress history. Figures 5.6 and 5.7 show a normalized pore pressure difference and poire pressure ratio. In a similar manner as the Bq-OCR relations considerable scatter was observed. The effects of parameters such as s o i l sensitivity appear to obscure those of stress history. A more successful approach to determine OCR i s the direct correlation of normalized cone resistance to OCR. The method of normalizing CPT data i s somewhat controversial i n cohesionless soils but extensive testing at deep clay s o i l sites clearly indicates that cone resistance should be normalized with the v e r t i c a l effective stress. Figure 5.8 shows a compilation of results from a variety of clay sites. Only those sites with well described stress history were used in the correlation. Although there i s some variation between sites a reasonable indication of stress history can be obtained from Figure 5.8. The correlation between stress history and normalized cone resistance i s expected from consideration of normalized strength and shear strength calculations from cone resistance. Assuming the shear 20n LEGEND: > to 3 10H oo oo '<* v+* o o * Langley S i t e s $ Brent C r o s s : Lunne et al.(1986) o Haney if S t . A l b a n : Roy e t a l . (1982) * McDonald Farm 1 i i i i I i i i i I i i i i | i i i i | i i i i | i i i 2 3 4 5 6 7 O C R i i i i i i i i i i i 8 10 20 n > 3 l o -ci o °° o o*x o LEGEND: * Langley S i t e s $ Brent C r o s s : Lunne et al.(1986) O Haney •ft S t . A l b a n : Roy et .a l . (1982) * McDonald Farm A Drammen x Haga i i i i I i I i i i i | i i i i i 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 1 1 1 1 I 2 3 4 5 6 7 8 9 1 0 O C R Figure 5.6 Normalized Pore Pressure Difference 5 -4-ro 1 H 0 o o o <n> o + .+ + o * LEGEND: * Langley Sites O Haney * St.Alban: Roy et al.(1982) * McDonald Farm + Boston Blue Clay: Baligh & Levadoux (1980) i i i i i i r i f t i | 2 3 i i i I i i i i I i i i i l i i i i i i i i i i i i i i i i i i i . r~- y-s —f /—I 5 6 O C R 7 8 5 4 H CM => 3 < 0 -1 0 LEGEND: * Langley Sites O Haney •k St.Alban: Roy et al. (1982) * McDonal d Farm A Drammen x Haga + Boston Blue Clay: Baligh 4 Levadoux (1980) 0 Onsoy x xo A oo c x O x O Co I I I I [ I I I I I I I I I I I I I I I I I I I 1 I I I I I I I I I [ I I I I I 1 I I I I 2 3 4 5 6 7 8 9 1 0 O C R Figure 5.7 Pore Pressure Ratios vs OCR O 20-Plasticity Index > 20 (qp - OV) = 4.24 * (OCR) 0 . 8 6 10 -> b i r -CD O c CO <*-» w w CD or CD c o O CD CD N CO E < • X P l a s t i c i t y Index > 20 Plasticity Index < 2 0 ( Q j - C y ) = 2 . 7 3 * (OCR)0-80 LEGEND: * Gloucester: Konrad & Law (1987) * St Marcel : Konrad S Law (1987) * Varennes: Konrad * Law (1987) * NRCC: Konrad « Law (1987) STP: Konrad & Law (1987) P l a s t i c i t y Index < 20 • On soy <f St .A lban: Roy et a l . (1982) + Boston Blue Clay: Bal igh & Levadoux (1980) x Haga * McDonal d Farm "i r OCR 10 Figure 5 . 8 Normalized Cone Resistance vs OCR 122 strength of normally consolidated and overconsolidated clay can be related by ( s u / oVoc = ( S u / CT' v)NC * OCR111 and undrained strength calculated from cone resistance by su = (q T - o v ) / N K then (q T - o v ) / a' v = ( N K * / CT'V)NC) * OCR M Plotting the logarithm of normalized net bearing against the logarithm of OCR results in a slope, m, and the intercept the product of N K and the normalized normally consolidated strength ( s u / o"' v) N C. Figure 5.7 shows an increase in normalized cone resistance with OCR raised to the power of approximately 0.8 and an intercept of 2.75 to 4.25. The dependence of stress history shown by an m value of 0.8 i s reasonable as i s the intercept which i s close to a calculated value of 4 from the product of NJJ of 16 and (su / CT' V)NC O F 0.25. Each of which represent commonly used values, Jamiolkowski, et a l . (1985). Allowance for s o i l type, indexed by plasticity in the manner shown in Figure 5.8, can improve the correlation between stress history and undrained shear strength. Higher plastic clays are expected to have a higher intercept as shown in Figure 5.8 given that both N K and normalized strength increase with plasticity. The variation in normalized strength with plasticity i s well known but the increase in N K with increasing plasticity i s a more controversial finding f i r s t reported by Aas, et a l . (1986) and reversed earlier findings reported by Lunne, et a l . (1976). Konrad and Law (1987) proposed a pore pressure based correlation for OCR that also requires pressuremeter data. The parameter proposed 123 by Konrad and Law (1987) was insensitive to stress history. These data are included in Figure 5.8 and a much better indication of stress history was observed. 5.7 Interpretation of CPTU Data For Undrained Shear Strength Undrained shear strengths have, in the past, been estimated from cone resistance measurements alone using s u = (9c ~ °v) / NK Pore pressure measurements have been incorporated into the equation by using the total cone resistance, q>p . As already discussed, this correction has been found to reduce the scatter obtained from cones of different design, Lunne, et a l . (1986). Pore pressure measurements behind the t i p are best for this application. Another application of pore pressure measurements for the calculation of undrained shear strength was suggested by Senneset, et a l . (1982). They defined the effective cone resistance and suggest determining the undrained shear strength from s u = (9c _ u) / NKE where qc = measured cone resistance u = dynamic pore pressure NKE = effective cone factor Robertson and Campanella (1983) emphasised the use of the total cone resistance which was found to be extremely important in this calculation. Lunne, et a l . (1985) showed that varied from 2 to 12 as a function of stress history which was also reflected in Bg. Figure 5.9, compiled in this study, shows a wide variation i n Nj^ and no 124 30 - i LEGEND: 0 Ons0y A D r a m m e n p las t ic a D r a m m e n lean * M c D o n a l d F a r m + Lang ley R e s e a r c h Si te x Lang ley 232 Upper Si te 3 20H LU 1 0 -X X x t x taS^|i*Q o I i i i i i i i i i r w ^ i 2.0 0.0 1.0 Bqi=(U1-Uo)/(qrav) Figure 5.9 Cone Factor Nj^ vs 125 apparent trend with Bg. For any given site, however, a good correlation between Nj^ g and Bq i s generally observed. The effects of sensitivity and stiffness ratio result in more scatter i f more than one site (soil type) i s considered. The primary d i f f i c u l t y with the effective resistance approach i s that in soft clays the cone resistance and the measured pore pressure are of very similar value. When subtracted, any errors i n either measurement result in a large variation i n the parameter (qij-u) and hence i n the calculated undrained shear strength. Additional approaches that integrate pore pressure measurements into the interpretation of CPTU data for s o i l strength include the use of the Bg parameter to select the most appropriate pore pressure factor N A U, where N A U = (ul - u Q) / s u A reasonable correlation between N A U and Bq! i s shown in Figure 5.10 which indicates that pore pressure based parameters may successfully be integrated into the interpretation of CPTU data for strength calculations. The data in Figure 5.10 limited the drawing of firm correlations at this time. A more reasonable application of pore pressure data was originally proposed by Lunne, et a l . (1985), based on their observation that the Bg parameters correlated with stress history. Lunne, et a l . (1985) hypothesised that the B q parameter might help in the selection of the most appropriate Njrj. Using data from several sites this approach was investigated for pore pressure data collected on both the cone face and behind the t i p . Figure 5.11 shows that there i s s t i l l considerable 30-i LEGEND: 0 Ons^y A Drammen plastic a Drammen lean * McDonald Farm + Langley Research Site x Langley 232 Upper Site Z3 CO o i 5 II < 2 0 -1 0 -W + O 5? ± K A A x * vS * **** + + + + + * *x XX 0 x x X n—i—i—i—i—i—|—i—i—i—i—r 1.0 i—i—r 0.0 2.0 BqHU1-Uo)/(q T-G v) Figure 5.10 Cone Factor N A U vs 13 CO > to r— 2 5 2 0 -1 5 3 1 0 cti u_ CD § 5 O 0 -0 . 0 LEGEND: 0 Ons0y Drammen plastic Drammen lean McDonald Farm Langley Research Site Langley 232 Upper Site Haga o St.Alban (Roy et el. 1982) a * + x it T 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 . 0 B q 1 = ( U 1 - U o ) / ( q r c y v ) 2 . 0 CO cr ll o 2 5 n > 2 0 -1 5 -S 10H CO CD S 5 0 -0 •ft ° 0 X x A * A * x x x x * I V * i i i i i i i i i i i i i i i i i i i i 0 . 1 . 0 2 . 0 Bq 2 = ( U 2 -U 0 ) / (qT-ov) Figure 5.11 Cone Factor Nj^ p vs Bq l 7 Bg2 128 variation i n the cone factor N^p from Bg parameters when different sites are considered. Local experience or correlations developed at one s i t e appear to give very promising results with this approach. An alternative means for the determination of undrained shear strength was suggested by Campanella, et a l . (1985) using the large excess pore pressure generated during cone penetration testing. In soft clays and l i g h t l y overconsolidated clays the cone resistance can be very small, typically less than 1 % of the load c e l l capacity. Small errors in relation to the capacity, especially those related to zero load s t a b i l i t y , may result in large errors i n the cone resistance measurement. In the same clay, however, pore pressure values may be a large percentage of the pore pressure transducer capacity. Lunne, et a l . (1986) in an evaluation of the accuracy of cone penetration test data report that the pore pressure measurements taken i n soft clays were considerably more repeatable than either the cone resistance or the f r i c t i o n measurements. Provided an appropriate theoretical or empirical expression i s used, therefore, the estimate of undrained shear strength w i l l be inherently more accurate using pore pressure data. A semi-empirical solution was proposed by Massarsch and Broms (1981), based on cavity expansion theories but included the effects of overconsolidation and sensitivity by using Skempton's pore pressure parameter at failure, Af. Incorporating the dependence of stiffness ratio, charts were developed to determine N A u, these charts are shown in Figure 5.12. Clearly a knowledge of the shear modulus or the p l a s t i c i t y index, PI, would assist in the estimate of the undrained S a t u r a t e d C l a y s Approximate Af Range Very s e n s i t i v e t o q u i c k 1.5 - 3.0 No r m a l l y c o n s o l i d a t e d 0.7 - 1.3 L i g h t l y o v e r c o n s o l i d a t e d 0.3 - 0.7 H i g h l y o v e r c o n s o l i d a t e d -0.5 - 0 Figure 5.12 Cone Factor N A U from Campanella, Robertson, and Gillespie (1986) 130 shear strength. The addition of shear wave velocity measurements during cone penetration testing promised to offer an independent measure of the shear modulus. This approach i s discussed further in chapter 7. The tendency for low or negative pore pressures measured behind the t i p i n insensitive overconsolidated clays restricts the application of this technique to normally or ligh t l y overconsolidated clays. Figure 5.12 allows for the application of pore pressure data collected on the face in a wider range of s o i l types. Although the charts in Figure 5.12 are based on cavity expansion theories, they are semi-empirical. The charts do, however, provide a rational means of selecting the cone factor N A U . The charts clearly show how the cone factor N A U w i l l vary with OCR, sensitivity and stiffness. An additional approach was proposed by Keaveny and Mitchell (1986) who propose an alternative effective stress determination of strength parameters. The method requires estimates of OCR, Af, and K Q and i s very sensitive to K Q . 5.8 Liquefaction Resistance Several investigators have shown that pore pressure response may be an index of liquefaction susceptibility. Schmertmann (1978) f i r s t hypothesised, and showed data supporting the idea, that loose sands might generate positive pore pressures and dense sands might generate negative pore pressures during cone penetration. As part of this study several investigations were performed before and after compaction. One site investigated, No. 6 Road in Richmond, described i n section 3.6, 131 was compacted by vibro-replacement to a depth of approximately 20 m. At this s i t e CPTU results, shown in Figure 5.13, generally resulted in pore pressures becoming more negative after compaction. It i s not clear, however, with soils such as those shown in Figure 5.13 that behave in a nearly drained manner, i f the change in pore pressure response i s due to a change in permeability or a change i n dilatancy. The effects of compaction would be expected to affect both parameters. If the void ratio change resulted in response changing from drained to par t i a l l y drained, i t may be expected that the pore pressure response could change to a negative value. Alternatively, densification may have resulted in greater dilation. Interpretation of the overall behaviour using other cone channels does not c l a r i f y the issue. The cone resistance increases indicates that density improvements have been made. The proportionally higher f r i c t i o n increases might be explained by lateral stress increases due to compaction. Surprisingly, l i t t l e change in shear wave velocity was observed. Attempts at using the measured pore pressure response in s i l t y s o i l s have been made because of the grain size dependence of cone resistance measurements. The dependence of cone resistance measurements on grain size makes the use of chamber test data obtained using clean medium grained sand d i f f i c u l t . A compilation of cone resistance and pore pressure measurements at various sand sites, some of which are known to have liquefied, i s shown i n Table 5.1. A clear picture of the liquefaction susceptibility can be seen from the cone resistance measurements, a l l sites known to have liquefied show a consistently low normalized cone resistance. Sites that have LEGEND: Before compaction After compaction#1 After compaction#2 Figure 5.13 Pre and Post Compaction Profiles at the No. 6 Road Site SITE SOIL TYPE Sites known to have Liquefied W i l d l i f e post li q u e f a c t i o n Heber Rd. post l i q u e f a c t i o n AVERAGE SIGN OF NORMALIZED DYNAMIC PORE RESISTANCE PRESSURE RESPONSE (q T/fj v') °n face behind t i p 45 55 Nerlerk p r i o r to li q u e f a c t i o n 70 Reference: Sladen et a l . (1985) Molikpaq p r i o r to li q u e f a c t i o n 50 1-65 Reference: J e f f r i e s (1988) + ve - ve + ve - ve free draining free draining Sites known or considered not to have l i q u e f i e d Brenda Dam hydraulic f i l l Fraser r i v e r sand McDonald Farm No. 6 Road; Knight St. Annacis Tuk Hrbr. Richards Island Fraser r i v e r sand Fraser r i v e r sand Fraser r i v e r sand Beaufort sand Beaufort sand Schoolhouse Beaufort sand 150 100 145 110 100 175 150 170 dry to moist + ve - ve + ve + ve + ve - ve - ve - ve free draining + ve - ve + ve - ve Note: Sign of dynamic pore pressure refers to value i n r e l a t i o n to the s t a t i c pore pressure value. Values shown as + ve mean that the pore pressure was higher than s t a t i c . Table 5.1 Normalized Cone Resistance and Pore Pressure Response i n Sand 134 liquefied a l l have normalized cone resistance less than 70. On the other hand, no clear distinction between the sands emerges from the pore pressure response. It appears that even in loose liquefiable sands, pore pressures measured behind the t i p are negative of the hydrostatic value due to the large stress r e l i e f associated with this area. Measurements taken on the face are, with few exceptions, positive of the static value as would be expected from the high compressive stresses associated with this area. As explained in chapter 4 i t appears that the details of the measuring system are often the key to interpretation of pore pressure response. The magnitude of the pore pressure gives a clue to the reason that cone resistance i s lower in s i l t y sands than i t i s in clean sands. A possible explanation was thought to be the higher pore pressure during penetration through s i l t y sands which reduced the effective stress. The magnitude of the pore pressure i s very small, however, compared to the cone resistance and i t s influence on reducing the cone resistance i s small. Repeated soundings at different rates, Gillespie (1981), showed very l i t t l e cone resistance dependence on changes i n rate and hence drainage. A more l i k e l y explanation i s that the increased compressibility of the s i l t y sand i s the dominant factor reducing the cone resistance compared to that of a clean sand. Compressible materials become compressive at low stress levels and have lower cone resistance than less compressible materials. Cone resistance in s i l t y sands appears to be as much a function of compressibility characteristics as strength characteristics. An alternative means of using pore pressures to assist in the 135 interpretation of liquefaction resistance of sands i s to use the pore pressure response to aid the cone resistance based interpretation. Interpretation of cone resistance for liquefaction resistance has been proposed by Robertson and Campanella (1985) for clean sand with D50 greater than 0.25 ram and for s i l t y sand with D 5 0 less than 0.15, by Seed and De Alba (1986) for sand with D50 of 0.25 ram, and by Shibata and Teparaksa (1988) for a range of sands specified by their D 5 0 values. A l l of these methods require that an estimate of grain size be made. The solutions are sensitive to grain size variations because, a l l other factors held constant, with a decrease in grain size below 0.25 mm there appears to be a increase in liquefaction resistance for the same cone resistance. An accurate interpretation of grain size i s best made using samples collected across depth intervals identified to be c r i t i c a l by the CPT profile. An alternative interpretation i s sometimes possible using the pore pressure response. Table 5.2 shows different t 5 0 values at sand sites with known grain size characteristics. The data collected at these sites indicate that a drained response i s only observed at sites with clean sands having a D 5 0 greater than 0.20 mm and less than 5 % fines. Knowing that a response i s drained or pa r t i a l l y drained can therefore be used as a guide to select the most appropriate resistance curve. A l l of the above methods include a resistance curve for sands having a D50 greater than 0.25 mm, a drained CPT response i s a reasonable indication that this curve should be selected. A pa r t i a l l y drained response would indicate that a lower resistance curve could be used. The Robertson and Campanella resistance curve for sands with D 5 0 less than 0.15 mm Site Heber Road 2 Depth t 5 0 Interval (sec) 2-4 m 30-60 DgQ D 1 0 Soil Type (ram) (mm) .10 fine SAND Heber Road 6 3-6 m drained response .20 fine SAND Wildlife 2-7 m drained PC2A response .15-.25 fine SAND McDonald farm 2-13 m 10-30 13-15 m 30-60 .25 2-5% fines .15 10% fines fine-med. SAND fine SAND trace s i l t Holmen 2-22 m drained response 22-26 m 10-30 .45-.90 .15-.30 med.-crs. SAND .20-.50 .06-.15 fine SAND trace s i l t Nerlerk f i l l 0-10 m drained response .25 3-5% fines SAND trace s i l t Molikpaq 1-65 core 4-20 m drained response .25 2-3% fines SAND trace s i l t note: Pore pressures recorded behind tip for a l l traces References: Heber Road sites, grain size: Youd and Bennett (1983) Wildlife site, grain size: Bennett et a l . (1981) Nerlerk site: Sladen et al. (1985) Molikpaq 1-65 core: Jeffries (1988) Table 5.2 Pore Pressure Response at Sand Sites 137 appears to be appropriate in sands through which a p a r t i a l l y drained response i s observed. The resistance curves proposed by Shibata and Teparaksa, based on the most extensive data base, permit a more detailed analysis of liquefaction resistance based on D 5 u i n steps of 0.05 mm and normalized cone resistance. Dissipation rates do not appear to be a sufficiently accurate means of distinguishing the small variations in grain size allowed with the method proposed by Shibata and Teparaksa. In addition, pore pressure dissipation rates are controlled by permeability which i s largely a function of the grain size of the finer portion, often expressed by D 1 0, whereas the resistance curves are shown to vary with D 5 0. Nevertheless, the distinction of a drained response and i t s appropriate resistance curve i s in many cases sufficient. The Robertson and Campanella curves for D 5 0 greater than 0.25 mm and D 5 0 less than 0.15 mm appear to be reasonably easily selected based on the pore pressure response in the f i e l d . More detailed analysis such as that proposed by Shibata and Teparaksa clearly requires grain size information from sampling. 5.9 Conclusions 1) One important application of pore pressure data i s the correction of cone resistance. The correction i s primarily required because of the lack of uniformity i n cone design. This correction has been shown to account for much of the variation in cone readings obtained with different cones at selected soft clay test sites. As a result of this correction a better understanding of N K correlations has been achieved. 138 2) Pore pressure response i s dominated by drainage characteristics and can therefore be used as an indication of s o i l type. Within the range of fine grained so i l s , essentially identical stratigraphic detail i s indicated with a l l pore pressure measurement locations. In overconsolidated soils, however, pore pressure response on the cone face was found to give the most detailed record. 3) Pore pressure dissipation rates were found to be a useful index of s o i l type but cannot be used to identify exact grain size characteristics. A common interpretation problem i s the distinction of soft s i l t s and overconsolidated clays. Friction readings in these so i l s are often insufficiently reliable for cl a s s i f i c a t i o n purposes but pore pressure decay rates can be used to distinguish these s o i l s . 4) A study of the usefulness of Bg pore pressure parameters as an index of stress history was performed. It was found that although Bg generally decreased with increasing stress history other factors, especially s o i l sensitivity, masked the correlation. A more successful approach to determine OCR appears to be the direct correlation of normalized cone resistance to OCR. The variation of normalized cone resistance with s o i l p l a s t i c i t y was shown and was shown in section 5.6 to be consistent with previous publications regarding the variation of the cone factor N K with s o i l p l a s t i c i t y . 5) The possible use of to calculate undrained shear strength was investigated. The use of Bg to assist in the selection of Nj^; was assessed at sites with high quality data. At these sites a wide variation in % E was observed and no apparent trend with Bg was found. It was f e l t that the subtraction of two very similar measurements 139 l i k e l y amplifies small errors associated with CPT cone resistance measurements in soft clay. 6) The use of Bg parameters to assist in the selection of N A U or NKT d ° e s n°t appear promising when several different sites are considered together. Site specific correlations are, however, much better. 7) A pore pressure based means of calculating undrained strength was presented. This method was found very useful in normally or li g h t l y overconsolidated clays. 8) In s i l t y or fine sand s o i l s pore pressures during cone penetration may be higher or lower than the stat i c values. The sign of the response i s a function only of the measurement details. Pore pressures on the face of the cone in granular s o i l s were found to be positive with only very few exceptions. Pore pressures behind the cone t i p were found to be less than static in almost a l l cases. This response results from the very large stress reduction associated with the area immediately behind the cone. 9) Pore pressure response, drained or undrained, can be helpful in determining the most appropriate CPT based liquefaction analysis. 140 CHAPTER 6. FACTORS AFFECTING SHEAR WAVE VELOCITY DATA 6.1 Source Characteristics 6.1.1 Hammer Beam Shear Source I n i t i a l work with the seismic cone penetrometer was conducted by Rice (1984). Laing (1985) investigated different combinations of sources and receivers for use in offshore and onshore environments. The hammer and weighted beam source was used as a standard for comparison. This source i s used extensively both for downhole and surface methods. Rice optimised the length of the source beam to a length between 2 and 3 m. It was found, in this study, that proper coupling to a well leveled ground surface was more important and controlled the repeatability of the source signal. Repeatability i s more important than signal amplitude with the interval technique of calculating shear wave velocity. Secondary considerations were found to include the beam stiffness. The use of metal beams resulted in greater signal amplitude than similar length timber beams. An aluminum or steel beam weighted down by the leveling jacks of a d r i l l r i g was found to be an excellent shear wave source. Rice used a timber beam with steel end caps weighted down with a vehicle and, although this was an adequate source, Rice found that he needed to use the average of atleast ten blows to reduce the variation of the resultant signal. The UBC in-situ testing vehicle with approximately five tons of reaction on i t s rear steel beam leveling pad i s probably an optimum shear wave source beam. Very large shear wave sources such as the "Marthor", Layotte (1984) , use a similar beam configuration with a 1500 kg hammer f a l l i n g 2 m with only slightly greater reaction than that provided by 141 the UBC testing vehicle. The high quality of data obtained using the sledge hammer and steel beam leveling pad configuration are demonstrated in Figure 6.1. These traces result from single strikes of each side of the source beam and hence the opposite a r r i v a l . Deeper penetration, below about 30 m usually required the use of stacking procedures. With the addition of ten blows very easily interpreted results could be obtained to 50 m. The shear waves generated by a horizontal impact source, SH waves, offer the advantage over other waves of not generating noise at s o i l interfaces through the conversion of incident energy into reflected or refracted P or SV waves. This eliminates noise arriving at the same time and masking the shear wave. 6.1.2 Explosive Sources Some experience was obtained using explosive sources, both shotgun shells and seismic caps. The use of explosive sources i s often the only means of obtaining a signal source i n shallow offshore conditions and on land i s the optimum way of generating high energy P waves. Figure 6.2 shows the strong arrivals observed by a horizontally oriented accelerometer generated by a buffalo gun (shotgun shell) source offset 3 m from the cone rods. The horizontal receiver, designed to record shear waves i s not favorably oriented to record P wave energy traveling near vertically. Inspite of this, large signal amplitude could be observed to a depth of over 50 m. Deep penetration was possible i f the source was offset further i n a horizontal direction. Figure 6.1 Damped Geophone Response Profile to Hammer Shear Source 0 Time (mill iseconds) 80 160 240 320 143 Figure 6.2 Acx^lerameter Response Profile to Buffalo Gun Source 144 In an offshore environment explosive sources i n the water near the seabed resulted i n strong P wave arrivals. Figure 6.3 shows the strong P wave arrivals and reflection from a permafrost boundary below. There also appears to be a second P wave l i k e l y generated by a bubble expansion/collapse mechanism. As a shear wave source, explosives gave very mixed results. Laing (1985) could not achieve repeatable results with the buffalo gun. Use of a water charged hole (for example, an auger hole backfilled with water) was found in this study to give highly repeatable results. The degree of repeatability obtained with the buffalo gun source lowered into a water f i l l e d augured hole was as good as that from a shear beam source. The interpretation for shear wave velocities from explosive sources gave very mixed results. Onshore, the generation of shear waves from explosive sources i s enhanced by any asymmetry of the explosion. Shear waves are also generated at sediment interfaces by P-S coupling. One possible interface i s the ground surface which may generate a shear wave when an upward traveling P wave reaches the ground surface. With a l l these possible sources the distinction of any arr i v a l can sometimes be d i f f i c u l t and may be complicated by the pos s i b i l i t y of destructive interference. Well defined shear wave markers were observed from explosive sources onshore under some conditions. Figure 6.4 shows the shear waves generated by this source. Laing (1985) concluded that results below about 12 m were acceptable compared to those obtained from a shear wave source but that at shallower depths P wave interference and ringing of the receiver masked the shear wave T I M E , msec. Figure 6.3 Geophone Response at Schoolhouse Site from Campanella, Robertson, Gillespie, Laing and Kurfurst (1987) Time (mill iseconds) 0 40 80 120 160 I I I - McDonald Farm - H o g e n t o g l e r Cone - B u f f a l o gun s o u r c e Figure 6.4 Damped Geophone Response Profile to Buffalo Gun Source 147 arrivals. Use of damped receivers, discussed i n a later section, was found i n this study to reduced the ringing problem. Offshore, the explosive sources in water only generate shear waves at the s o i l water interface or with subsequent impedance contrasts caused by stratigraphic changes. Shear waves generated i n this manner are not nearly as easily distinguished as shear waves generated by explosive sources onshore. The interpretation of signals collected offshore was only found possible by consideration of a l l the traces collectively and the recognition of a shear wave marker at depths. This marker could then be traced up through the stacked profile. Figure 6.3 shows the d i f f i c u l t y i n selecting shear wave arrivals from this sort of source. If explosive sources are necessary in an offshore environment the symmetry of the explosion must be reduced. An attempt at doing this was made in the Beaufort Sea site investigation, Campanella, et a l . (1987), by placing the seismic cap on a blade shaped device, this proved only marginally effective because of the d i f f i c u l t y in inserting the blade into the dense s o i l at the site investigated. More effective results may be obtained by the near simultaneous explosion of two sources closely spaced on the sea floor. This source was not attempted but similar ideas of reducing the symmetry of explosive sources have been experimented with in the o i l industry. 6.2 Receivers A variety of receivers were used in this research. They included geophones, damped geophones, and accelerometers. It was quite clear 148 that the excellent mechanical coupling between the s o i l , the cone, and the transducer resulted in the excellent results obtained by a l l of these transducers. Some consideration, however, of the type of transducer best suited to certain conditions i s s t i l l j u s t i f i e d . Geophones or velocity transducers are generally used to measure signals having a frequency roughly between two and twenty times their resonant frequency. The 28 Hz natural frequency velocity transducers used were selected primarily because they f i t , with slight trimming, inside the 10 cm^  cone. Typical frequencies of shear waves observed were 40 to 60 Hz and P waves 400 to 600 Hz. The 28 Hz transducers therefore are reasonably well suited to measure either of these two signals. Velocity transducers characteristically have large phase shifts near their resonant frequency. Hence, lower natural frequency transducers would have reduced any concern for phase shifting of the shear waves. I t i s l i k e l y that the interval technique used in downhole testing cancels a l l or nearly a l l of the phase s h i f t . Laing (1985) compared using the downhole interval technique results obtained from accelerometers and geophones in repeated holes and could not find a systematic difference. Damped geophones were found to be useful with explosive sources by reducing the ringing but also reduced the amplitude of both P and S wave arrivals. Damping i s useful when recording signals generated by explosive sources for shear wave velocity. Vertically oriented receivers were used by Laing (1985) in an attempt to obtain P wave velocities from shear sources by optimising the orientation of the transducer with respect to the incoming wave 149 front. It became clear, however, in t r i a l s conducted i n the Beaufort Sea, Campanella, et a l . (1987), that the v e r t i c a l transducer was dominated by waves travelling down the relatively r i g i d CPT rods. Flexible couplings are generally used in conventional downhole testing for this reason but cannot be incorporated within the cone penetration test. At many sites i t i s necessary to lower the cone equipment down casing designed to prevent lateral buckling through f i l l s , water or ice. In these instances, and i n the deep offshore environment, i t can be d i f f i c u l t or impossible to maintain the orientation of a horizontal geophone. To investigate the sensitivity of the signal to the orientation, a cone was incrementally rotated with repeated strikes on the shear source. It was found that within + 45 degrees entirely acceptable results were obtained. This means that by i n s t a l l i n g two orthogonally placed horizontal geophones within the cone, one receiver w i l l always be oriented i n an acceptable manner. If the geophone was in a completely unfavorably orientation almost no shear wave energy was observed. 6.3 Identification of Shear waves from Explosive Sources Previous sections have showed examples of strong shear wave traces generated from explosive sources, (Figures 6.3 and 6.4). It must be recognized that the interpretation of these traces required consideration of traces recorded over a range of depths and was generally performed by identifying shear wave arrivals at the deepest depth and recognizing the same marker recorded at shallow depths. It 150 was not found possible to interpret a shear wave arr i v a l from a single explosive source. Figure 6.5 shows a typical trace from a horizontally oreinted geophone receiver from the buffalo gun source. Also shown are traces recorded at the same location but generated by a shear wave source and by a vertical hammer impact. Although these traces were recorded at the same depth the travel paths had slig h t l y different lengths. Using the shear wave source as a reference, the shear wave arrival can be selected from the traces recorded from the ver t i c a l impact and the explosive source. As with the hammer (horizontal) shear source i t i s observed that the f i r s t departure i s not as strong as the succeeding. In a l l cases where there was a known arri v a l time i t was found that the shear wave f i r s t departure was of opposite sign to the P wave f i r s t departure. This conclusion was used in a l l subsequent interpretation of traces recorded using explosive sources. Figure 6.5 also illustrates several other points. 1) The rounded and d i f f i c u l t to interpret P wave arr i v a l recorded on the geophone compared to those recorded on the accelerometer (Figure 6.2). 2) The large amplitude of the shear wavelets following the i n i t i a l departure. 3) The complete lack of P wave energy from the vertical impact source. The high ratio of shear wave energy compared to P wave energy from the explosive source and the vertical impact source i s due to the combined effects of the rapid attenuation of P wave energy in the unsaturated surface s o i l s and the unfavorable orientation of the receiver to record 152 ver t i c a l l y propagating P waves. In the offshore environment the P wave amplitude was much stronger than the shear wave amplitude. The unsaturated surface soil s commonly encountered onshore l i k e l y contributes to the greater success at using explosive sources onshore than offshore. Conclusions from the t r i a l s conducted in the Beaufort Sea indicate that two factors reduce the effectiveness of using explosive sources i n the offshore environment: 1) The low impedance contrast at the seabed which results i n only small conversion of P to shear energy. 2) The saturated sediments transmit the P wave energy resulting in subsequent P-shear conversion at s o i l interfaces and therefore creating multiple apparent sources. 6.4 Soil Layering and Resolution The effects of s o i l layering must be considered when assessing CPTU results in interbedded deposits. Conventional rules of thumb suggest that layers less than 0.5 m thick w i l l not achieve f u l l cone t i p resistance values representative of the properties of that layer alone. The v a l i d i t y of such a guideline depends on the cone diameter and contrast between the properties of the layers. Velocity profiles were generally collected on 1 m intervals, this interval appears to be warranted. Beeston and McEvilly (1977) discuss the resolution available in downhole velocity measurements and give the minimum distance below which a layer can be detected as D = y 1 * y 2 * t Vi - V 2 where D = resolution distance 153 t = timing uncertainty and V 2 = velocity above and below the contact Typical uncertainties were discussed in chapter 2 . The timing uncertainty i s so small that the depth resolution i s on the order of a few centimetres for normally encountered variations i n velocity. Measurements made at 1 m intervals appear to be j u s t i f i e d . 6.5 Downhole-Crosshole Comparisons One means of validating the results obtained from the seismic cone test and the pseudo interval technique was to compare results to those obtained from conventional crosshole testing. Considerable debate surrounds the comparison of downhole to crosshole tests. Comparisons between downhole CPT and crosshole tests are shown in Figures 6.6, 6.7 and 6.8. Test results obtained with high quality procedures at the Holmen and Drammen sites show excellent agreement between the two techniques. An earlier comparison shown by Rice (1984) and by Robertson, et a l . (1986) at the Annacis s i t e shows a considerable difference between the two test methods. The CPT velocity measurements were confirmed by later testing. The difference between the two measurements was originally speculated as being due to, and an indication of, inherent anisotropy. Later comparisons at other sites which show much less difference between measurements obtained with the two techniques, indicates that measurement error in the crosshole testing most l i k e l y explains the difference. Stokie, et a l . (1986) performed velocity measurements on large cubic specimens i n the laboratory with variable stress conditions. S H E A R W A V E V E L O C I T Y C m / e ) 0 50 100 150 200 250 0-J 1 I I I 5-3 10-I 15-C P T D O W N H O L E A C R O S S H O L E CONE BEARING INTERPRETED Qc (bar) PROFILE very-loose SAND Figure 6.6 Comparison on Downhole and Crosshole Velocity, Holmen adapted from Eidsmoen, Gillespie, Lunne and Campanella (1985) 155 BEARING RESISTANCE 0 (bar) 10 SOIL PROFILE FINE S A N D SILTY C L A Y PLASTIC C L A Y (PI = 27%) L E A N C L A Y (Pl= 10%) SHEAR WAVE VELOCITY V S (m/s) 0 50 100 150 200 250 • CROSS-HOLE J) D O W N - H O L E Figure 6.7 Comparison of Downhole and Crosshole Velocity, Drammen adapted from Eidsmoen, Gillespie, Lunne and Campanella (1985) 156 CO L CD •P CU E CL Ld Q SHEAR WAVE VELOCITY Cm/s) 0 50 100 150 200 250 0 CONE BEARING Qc (bar) INTERPRETED PROFILE SAND SILT s i l t y SAND CLAY s t i f f p l a s t i c CPT DOWNHOLE CROSSHOLE Crosshole Velocity Data from: Nazarian and Stokoe (1984) Figure 6.8 Comparison of Downhole and Crosshole Velocity, Wildlife Site 157 They clearly showed that stress induced anisotropy affects downhole and crosshole measurements to exactly the same degree. They also state that there can be no difference between the two measurements. Their own results, however, indicate that inherent anisotropy plays a considerable role and l i k e l y explains any difference between downhole and crosshole measurements. The small differences shown at Holmen, Imperial Valley and Drammen are l i k e l y due to inherent anisotropy. If anisotropy i s of interest one means of evaluating i t s importance i s the comparison of velocities measured i n a downhole test and in a crosshole test with ve r t i c a l particle motion. Research into the importance of structural anisotropy i s complicated, however, by the d i f f i c u l t y in separating the effects of structural anisotropy and the generally unknown horizontal stress conditions. One important advantage of the downhole CPT velocity measurement i s that by using small diameter cone rods the influence of borehole induced effects including mechanical disturbance and stress relaxation are minimized. In downhole CPT testing several square metres of s o i l influence the measurements. In comparison, the cone rods and s o i l affected by penetration around them i s very small. Incorporating downhole shear wave velocity measurements into the cone penetration test has obvious economic advantages and was found to enhance the interpretation of both the velocity measurements and the CPT interpretation (chapter 7). In addition, several of the d i f f i c u l t i e s normally encountered i n traditional downhole or crosshole testing are eliminated. These include borehole effects such as stress changes and disturbance effects, and travel path uncertainties. 158 6.6 Conclusions 1) The limitations of shear and P wave sources was discussed in this chapter, numerous examples were shown. The high quality of the shear wave traces recorded onshore i s primarily a result of the excellent coupling between the s o i l and the cone penetrometer which has the receiver firmly bedded inside. Secondary considerations of the details of both the source and the receiver were also discussed. The accuracy and repeatability of the interval technique appears to be very high for depths of penetration less than 30 m. Below this depth signal enhancement techniques become more important. Explosive sources were investigated in both the offshore and onshore environment. In the offshore environment, where their use i s warranted, the high degree of symmetry of an explosion underwater results in poor generation of shear waves at the seabed and other s o i l contacts. Interpretation of the traces obtained offshore from explosive sources required consideration of a l l traces at the sit e . Additional research into the development of a repeatable, easily deployed, shear wave source for use offshore i s warranted. 2) The identification of shear wave arrivals from explosive sources was assisted with the observation that the sign of the P wave and shear wave arrivals, as identified by comparison to pure shear sources, was opposite. 3) Downhole-crosshole comparisons were made at a number of reference sites increasing confidence in the CPT downhole technique. 159 CHAPTER 7. APPLICATION OF DATA 7.1 Introduction Previous chapters have dealt with methods of obtaining shear wave velocity data from the seismic cone test. Many details of the testing methods and interpretation procedures were discussed. I t was shown that, provided the CPT i s suitable at the si t e i n question, a very accurate and repeatable measure of the shear wave velocity can be obtained. This section outlines some of the applications of the low strain shear modulus. The most important of these involve the direct use of Gjjjax data either as an important stiffness parameter or by correlation to other s o i l parameters. Other applications are also proposed i n this chapter, including the integration of G^^ data into conventional CPT interpretation for s o i l type and strength parameters. These applications became apparent when i t was found that, at some sites, previously observed correlations between shear wave velocity data and CPT cone resistance data did not seem to hold. 7.2 Shear Modulus as an Engineering Parameter Shear modulus i s a fundamental s o i l parameter widely used in deformation analysis. As discussed earlier, downhole shear wave velocity measurements are made at small strains and reflect the small strain behaviour of the s o i l . For this reason G^^ measurements are an important input into seismic analysis. If deformation analysis i s to be made at larger strains the small strain modulus must be attenuated to lower values, reflecting the non linear behaviour of so i l s . Commonly used modulus attenuation figures include those of Seed and 160 Idriss (1970). In many s o i l investigations large strain measurements are available from laboratory tests and small strain measurements can be made from laboratory or in s i t u tests. Knowing the modulus values at different strain levels i t i s then possible to refine a modulus attenuation curve to account for local conditions. Secondary considerations such as the effects of strain rate are generally much less important than strain level effects. In fact, because of the very small shear strains involved in shear wave measurements, strain rates are of a similar order of magnitude to those used in conventional large strain laboratory undrained tests. 7.3 Correlation of Velocity to Liquefaction Potential Evaluation of liquefaction potential i s perhaps the most important application of shear wave velocity measurements. The correlation of velocity to liquefaction i s especially important in s i l t y sands where chamber test data are unavailable and i n sands with a coarse gravel content for which both penetration and interpretation are d i f f i c u l t . In clean sands, which have well described cone resistance-relative density or cone resistance-liquefaction resistance relations, velocity measurements may be a useful independent means to confirm cone resistance based interpretations. Cone resistance based liquefaction interpretations have been proposed by Robertson and Campanella (1985), Seed and De Alba (1986), Olsen (1984), Zhou (1981) and others. Three interpretation approaches to liquefaction that consider shear velocity include: 1) Correlation of shear wave velocity to liquefaction potential. 161 2) The threshold strain approach. 3) Comparison of shear wave velocity measurements in the f i e l d and the laboratory. With the lack of f i e l d data collected prior to liquefaction, one way of developing an indirect correlation i s to use an SPT N value correlation to shear wave velocity such as that proposed by Suyama, et a l . (1986). Based on the comparison of 1,654 tests i n sands, Suyama, et a l . propose the following correlation between SPT N value and shear wave velocity V s = 97.0 N 0-341 where V s = shear wave velocity in m/s N = blows per 300 mm This relation can be used to combine the measurement accuracy of the shear wave velocity and the high level of experience gained with the SPT. The correlation i s based on Japanese SPT practice which typically achieves 60 % of theoretical energy. The 60 % energy rating i s consistent with the SPT liquefaction curves proposed by Seed, et a l . (1985). Figure 7.1 shows the addition of shear wave velocity for the prediction of liquefaction resistance onto the SPT chart proposed by Seed, et a l . (1985). Also shown on Figure 7.1 i s a means of correcting velocity measurements to 1 atmosphere. This figure assumes that velocity i s proportional to a single stress, the v e r t i c a l effective stress, raised to the power 0.35. This assumption i s discussed i n greater detail in a later section. 0.6 0 10 20 30 40 50 ^ 0 ^ ) 6 0 2 1 3 269 309 341 Shear Wave Velocity ( m / s ) Corrected to 1 Atmosphere using: velj = vel * 0 v e | 0.0 0.6 0.8 1.0 1.2 1.4 1.6 C N and C v e | Figure 7.1 Shear Wave Velocity as an Index of Liquefaction Potential to 163 The CJJ values shown i n Figure 7 . 1 from Seed, et a l . (1985) are close to the C y ^ curve shown in Figure 7 . 1 at low stress levels. The similarity between the stress correction curves i s important because the basis for the entire correlation are the f i e l d observations between SPT N value and shear wave velocity, gathered over a range of stresses and not corrected for stress levels. The greatest departure between the stress correction curves occurs at high stress. The stress correction curve i s very significant varying by a factor of two in the normal range of stresses considered. Some correlations for the shear wave velocity required to resist liquefaction such as Seed, et a l . (1983) do not consider stress effects other than to r e s t r i c t the correlation to the upper 50 feet. This simplification i s not necessary and may be unconservative at high stresses. Occasionally liquefaction must be considered at low stress levels. This situation i s most l i k e l y with water tables near the surface or i n the offshore environment. Insufficient data are available to ju s t i f y the extrapolation of stress corrections to higher values in the low stress region. The effects of stress levels were d i f f i c u l t to evaluate at low stresses because of s o i l variations and suction conditions in the upper desiccated crust. In addition, high loads introduced by the UBC testing vehicle become important at shallow depths and make the equipment used in this thesis inappropriate to investigate stress effects at shallow depths. In the manner of the development of the SPT based liquefaction resistance curve, refinement of Figure 7 . 1 requires additional measurements of shear wave velocities at sites that did and did not liquefy in historical earthquakes. Although Seed, et a l . (1985) show 164 additional resistance curves for s i l t y sands the variation i s primarily due to the sensitivity of penetration resistance to grain size. Shear wave velocity appears independent of grain size; therefore a velocity based liquefaction resistance curve such as Figure 7.1 i s useful over a wide range in grain size. Direct correlations between liquefaction potential and shear wave velocity can be based upon only a limited data set. Bierschwale and Stokoe (1984) based their curve on data collected at sites that did and did not liquefy in the Imperial Valley. The curves presented by Bierschwale and Stokoe compare very poorly to those developed using correlations to SPT N values, i n addition, they appear to be very unconservative. Bierschwale and Stokoe 7s curves indicate that soi l s with shear wave velocity greater than 140 m/s would require maximum acceleration levels of 0.3 g or more to cause liquefaction. Their curves are based upon data that showed liquefaction restricted to sites with velocity less than 125 m/s. Uncertainty i n the acceleration records may par t i a l l y explain the apparently low threshold velocity indicated by Bierschwale and Stokoe. An additional reason that their correlation appears unconservative i s that i t i s based on Imperial Valley data that clearly f a l l s on the liquefaction side and does not constrain the resistance curve. The second approach for the evaluation of liquefaction potential from shear wave velocity measurements results i n a similar comparison between the cy c l i c stress ratio required to i n i t i a t e liquefaction and shear wave velocity but was formulated from a simple theoretical approach by Dobry, et a l . (1981). This approach was founded on the 165 observation that under uniform cycli c undrained loading a threshold strain exists below which no pore pressures develop. Knowing the threshold strain level from laboratory testing and the stiffness from velocity measurements the cy c l i c shear stresses required to induce pore pressure can be calculated. The key assumption i n the approach proposed by Dobry, et a l . i s that the threshold shear strain can be well defined. The method may be more useful on a si t e specific basis where the threshold strain can be better defined from the results of laboratory testing. Comparison to the methods described above i s not made here because of the sensitivity to the assumed threshold strain. The third approach for using shear wave velocity data i n important investigations may be to consider f i e l d and laboratory measured velocities together. At important sites, or based on local experience, laboratory derived shear wave velocity-liquefaction resistance relations can be used to carefully evaluate f i e l d velocity measurements. Shear and compression wave velocities can be measured in the laboratory, in a non destructive manner, following the consolidation stage. By including shear wave transmitters and receivers into the t r i a x i a l test only minor incremental costs are incurred. In addition to development of a shear wave velocity-liquefaction resistance relation the f i e l d shear wave velocity measurements may also allow the evaluation of sample disturbance. If the velocities are not comparable then either sample disturbance or incorrect consolidation stresses have been applied to the sample. Equivalent f i e l d and laboratory velocities offer some assurance that representative samples can be tested. (There remains the pos s i b i l i t y of 166 offsetting effects of stress variation between f i e l d and laboratory and sample disturbance effects.) 7.4 Integration of Data into CPTU Data 7.4.1 Introduction Based on previous experience with the correlation of acoustic and geotechnical parameters i t became apparent that at some sites i t could be helpful to incorporate the velocity measurements into the interpretation of CPTU data. Two observations highlighted this need: the d i f f i c u l t y i n distinguishing overconsolidated clays from normally consolidated s i l t s using CPTU data alone; and the observation that at some sites velocity data were of unexpected values. This section shows several approaches at the integration of seismic data into CPTU interpretation. This i s a logical approach when the data are collected together and i s based on the fact that shear wave measurements reflect only small strain stiffness properties whereas CPTU measurements are a function of many properties. It must be recognized that stiffness properties that affect CPTU measurements are at much larger stresses and deformations than those encountered in velocity measurements. Other CPT interpretation measurements also use the results of separate sensors to assist in the interpretation. For example, permeability, as indicated by pore pressure response, i s used to determine i f a drained or undrained analysis i s most appropriate. Approaches that incorporate pore pressure measurements into the selection on values to calculate shear strength from cone resistance measurements were discussed in chapter 5. In a similar manner, the 167 possible use of shear waves to enhance the interpretation of CPTU data i s discussed. 7 . 4 . 2 Application of G^x Data in the Interpretation of Clay Strength from Cone Resistance Measurements Previous success at improving the selection of N K values were shown by Lunne, et a l . ( 1 9 8 5 ). With s i t e specific correlations i t was shown that the Bg parameter could be used to help evaluate the most appropriate N r. This empirical correlation followed from the observation that both N K and B q values were known to depend on stress history. Given this mutual dependence on stress history i t was reasoned that there may also be a correlation between N K and Bg. In a more direct manner i t may also be expected that N K may also vary with some measure of the stiffness ratio. Theoretical expressions for N ^ , based on cavity expansion theory, include a stiffness ratio term. One such formula from Vesic ( 1972) gives N K as N K = 1 . 3 3 ( 1 + In G/Sy ) + 2 . 5 7 From this cavity expansion theory i t i s expected that N K should increase with the s o i l stiffness ratio. A knowledge of the low strain modulus allows a f i r s t estimate or index of the stiffness ratio to be made. The ratio G/qij i s similar to a stiffness ratio and variation in this parameter had been observed at different sites. The advantage of such a ratio i s that both parameters are measured i n the same prof i l e i n the seismic cone test. Using averaged values of cone resistance for uniform sections with velocity data, results from the sites investigated are shown in Figure 7 . 2 . Very l i t t l e correlation was 2 5 h LEGEND: 0 O n s ^ y a D r a m m e n p las t ic a D r a m m e n lean * McDona ld F a r m + Lang ley R e s e a r c h Si te x Lang ley 232 Upper Si te 3 II r-o 03 CD o O 2 0 -s 1 5 -1 0 -+ • + + + X A >X X X * , * x * 3 f 5 0 1—r " i — i r 0 2 0 4 0 ' I 1 1 6 0 8 0 —'—I 1 0 0 'max Figure 7.2 Cone Factor %£. vs Gmax/qp Ratio 169 observed and an expected increase in with increasing G/qT was not apparent. In the discussion of pore pressures i t was shown that pore pressures depended on sensitivity, strength, and stress history i n fine grained s o i l s . In the same manner, i t appears that values also depend on a variety of parameters. In this study i t was observed that sensitivity may be a key parameter in determining N K values. This value i s poorly established for most of the sites investigated making i t d i f f i c u l t to establish the dependence of N K on sensitivity. Previous attempts at using p l a s t i c i t y index, commonly used as an index of stiffness ratio, to assist i n the estimate of N^ have also shown considerable scatter, infact, the general trend of increasing or decreasing with p l a s t i c i t y i s unclear. Measurement error can be ruled out for most of the sites used in Figure 7.2 as repeated soundings have been performed. Some of the sites are similar in many respects but have very different N K values. No explanation i s presently available. For example, large variations i n N K are reported by Greig (1985) at the Langley sites which are seemingly otherwise very similar. The two Norwegian clay sites investigated i n this thesis, Ons0y and Drammen, which otherwise are very similar, have very different N^ values. It would seem that the contribution of stiffness ratio i n the determination of N K i s small and that s i t e specific N K correlations remain important. Other attempts have been made to incorporate stiffness into CPT interpretations; for example, Konrad and Law (1987) use a larger strain modulus from the self boring pressuremeter. The complexity of a method requiring pressuremeter data to interpret CPT data makes i t 170 unattractive and d i f f i c u l t to evaluate. 7.4.3 Integration of Velocity Data into Soil Classification The correlation between CPTU cone resistance measurements and velocity data was developed i n this thesis to: 1) Provide a framework for estimating values from CPTU data. 2) Distinguish s o i l types. 3) Understand the variation i n velocity with s o i l type. A synthesis of G/qT ratios and normalized cone bearing was made using values averaged over uniform sections from sites where shear wave velocity data, cone resistance, and known s o i l conditions were available. These values are shown plotted i n Figure 7.3. Several observations become apparent from Figure 7.3. There i s a well developed trend between G/qT and normalized cone resistance i n sands. This trend i s discussed i n detail in section 7.5. A wide variation in G/qT i n clays i s apparent, values ranged from 40 to 80. The variation was noted from site to site and does not appear to vary with normalized cone resistance i n the manner observed for sands. Site specific measurements or local correlations for ^ appear more c r i t i c a l in clays than in sands. Ratios of G/qT in overconsolidated clays approach those of a s i l t . Organic content i s most apparent in lowering normalized cone resistance but some decrease in the ratio G/qT was also observed at organic rich sites. Highly organic s o i l s such as peats may have very low G/qT ratios of approximately 10. Some of the peat sites observed in this study may gain their strength and hence cone resistance from their fibrous nature, this strengthening mechanism TZ.T 172 i s apparently less effective as a stiffening mechanism. The effect of overconsolidation in clays i s much more apparent i n cone resistance measurements than in the G/qiji parameter; Figure 7.3 shows a distinction between overconsolidated and normally consolidated clays using the normalized bearing but complete overlap of the G/qtp parameter. This observation lead to correlations between normalized cone resistance and stress history. Two commonly encountered CPTU interpretation situations are demonstrated here. The f i r s t i s illustrated by results obtained at the Swimming Point site, shown in Figure 7.4. At the Swimming Point site traditional CPT interpretation, based only upon cone resistance and f r i c t i o n ratio was used i n a non-subjective, computerized, manner. This method very accurately distinguished several stratigraphic boundaries but predicted much finer s o i l s than those actually present. Additional measurements included dynamic pore pressure, short dissipation records, and shear wave velocity data. Dynamic pore pressures were not found to be especially useful in this case. D i f f i c u l t i e s with dynamic pore pressure measurements and dissipations included: 1) A complex redistribution of excess pore pressure and a decay direction inconsistent with usual theories of pore pressure dissipation analysis. (Normal analysis i s for decreasing pore pressures with time) 2) Very low, occasionally zero, excess pore pressures. The INTERPRETATION INTERPRETATION PORE PRESSURE U of »oter) SLEEVE FRICTION (bo-) k 100 0 0-10-15 2.5 0 0 10 15 CONE BEARING Oc (bar) FRICTION RATIO Rf «) G/q T (sec) 10 19 1 7 11 • ) 12 \ 20 \ 6.5 S 14 ' 5 12 I 30 ) 60 \ 12 *l > 200 ^ 20 15 BASED UPON Qc and Rf BASED UPON Qc,Rf,U,G/Qc and d i s s i p a t i o n s pre-pushed pre-pushed Interbedded s i l t y CLAY and SILT Interbedded SILT and sandy SILT with organic lay e r s s i l t y CLAY clayey SILT with organics sandy SILT to clayey SILT s i l t y SAND s i l t y CLAY with SILT lenses clayey SILT with organics s i l t y CLAY clayey SILT sandy SILT s i l t y SAND CLAY-organic CLAY-organic s i l t y SAND SAND stopped i n GRAVEL stopped i n GRAVEL Figure 7.4 Integration of Gmax and Pore Pressure Measurements into Cone Interpretation LO 174 low values were complicated further by their sensitivity to rod clamping procedures. At the Swimming Point s i t e a partial grounding of ice onto the river bed at the edge of a deep channel allowed good shear wave velocity measurements i n the river bottom. Calculation of the ratio of G^x to cone resistance, (G/qT) ranged from 7 to 14. This range i s consistent with previous observations in loose sand (discussed i n later sections). Reasonably rapid decay of excess pore pressures, where generated, and the G/qiji correlations both indicate that the p r o f i l e was much coarser than interpreted from traditional CPT methods which rely solely upon cone resistance and f r i c t i o n . The two profiles—one interpreted from cone resistance and f r i c t i o n , the other considering cone resistance f r i c t i o n pore pressure and G/q T—are shown with the CPT data in Figure 7.4. Sampling performed by others, Kurfurst (1986), ultimately confirmed the interpretation made by considering a l l data but did not show the same level of detail shown by the CPT. Traditional q T, FR interpretation i s generally much more accurate than experienced at Swimming Point. At this location a high organic content i n the sand resulted in very high f r i c t i o n ratio measurements. The organic content has the combined effect of lowering the cone resistance measurements and increasing the f r i c t i o n measurements, this results in unusually high f r i c t i o n ratios and biases the prediction towards a finer grained s o i l . The profound difference in mechanical properties expected of an overconsolidated cohesive material and a loose noncohesive material makes distinction of these two materials very important. This example illustrates how a l l available information 175 needs to be considered for best interpretation of CPTU profiles. A second area where interpretation of CPTU profiles can be d i f f i c u l t i s the distinction of l i g h t l y overconsolidated clays and normally consolidated s i l t s . These two materials may have very similar cone resistance and have f r i c t i o n measurements well within normal accuracy. It was hoped that i t may be possible to distinguish these s o i l s by comparing their G/qT ratio. It was expected that the overconsolidated clay would have a greater shear wave velocity than a loose s i l t with similar cone resistance. As shown in Figure 7.3, however, these two materials may have similar G/qT ratios. I t appears that, for clay soils, the G/qT ratio decreases with increasing stress history and becomes similar to that observed at s i l t sites. The data collected i n this thesis clearly show that pore pressure dissipation rates s t i l l form the most definite means of distinquishing these two s o i l types from CPTU testing alone. 7.5 Correlation Between CPT (qT) and G^x Data 7.5.1 Introduction Correlations between cone resistance measurements and G^x are discussed because the ratio may reflect important s o i l properties; in addition well established ratios reduce the need for velocity measurements. Results of CPT chamber tests reported by Baldi (1986) indicate that cone resistance values are insensitive to stress history whereas stiffness values were found to be very sensitive to stress history. Stiffness values reported by Baldi included both a one dimensional loading modulus, calculated from response of the chamber 176 sample, and E 2 5 values calculated from loading otherwise similar t r i a x i a l specimens. In both cases the stiffness/cone resistance ratio was very sensitive to stress history. Baldi concluded that there can be no unique relation between cone resistance and void ratio that holds for a l l ranges of stress history. This conclusion implied that, i f an in s i t u stiffness and cone resistance can be obtained, an estimate of stress history may be possible. For this reason, values measured at various sand sites were tabulated. Before they can be compared, however, i t i s necessary to ensure that the effect of varying stress levels i s not influencing the results. 7.5.2 Normalization of q T, and G^x data Ideally, for the comparison of shear wave velocity data and cone resistance measurements, the effects of stress level should be similar. Data available to determine the dependence of cone resistance on stress levels come from chamber testing and f i e l d observations. Considerable debate centers around the relative importance of vertical and horizontal stress. Houlsby and Hitchman (1987) showed from results of calibration chamber tests that only the horizontal stress seemed important. Other chamber test data reported by Baldi (1986) also showed the importance of horizontal stress. Although horizontal stress i s obviously important, i t can only be known in laboratory situations. The chamber test data are often reported i n a form similar to that used by Baldi, where qc = C 0 * ( CT'v ) C 1 * exp (C2DR) The exponent C x reported by Baldi i s 0.55. Similar values were 177 reported by Schmertmann (1976). Although Houlsby and Hitchman argue that the horizontal stress level i s the most important, i f a single stress i s considered, the exponent agrees with that reported by Baldi. Field evidence from this study indicated a s l i g h t l y greater dependence of qiji on stress at normally encountered stress levels. The dependence of cone resistance on stress levels was investigated by plotting the log of the cone resistance against the log of the vertical effective stress and finding the slope of the best f i t line. Examples showing the dependence of both G^x and cone resistance on stress levels are shown in Figures 7.5 and 7.6. The cone resistance data were averaged in the same sections as the depth interval of the velocity data. Uniform sections only were considered. Figure 7.5 shows data from sites with clay behaviour. The Onsoy site data are especially important because of i t s uniform nature. Moisture content profiles at the site are nearly constant with depth. The influence of stress in the deep soft sediments at the McDonald Farm site are shown because of the high stresses involved. Examples from sand sites are shown in Figure 7.6. One site, the No. 6 Road si t e i n Richmond, i s especially uniform. Another site, the Annacis site, has greater variation i n the cone resistance profile, typical of sand sites, and more scatter can be seen. Nevertheless the dependence of both cone resistance and G^x i s similar at each of the sites. A l l sites consistently showed an exponent indicating the stress dependency of cone resistance and of 0.70 + 0.15 in sands and 0.90 + 0.1 in clays. An analysis of this type, which considers f i e l d data, i s advantageous in that i t accounts for the effects of aging. 5.0 co D_ ^ 4.0 ro E CD T3 Jl 3.0 CO D_ CT o 2.0 o Ons<^y S i te Log Gmax = .82 L o g o V + 2.85 Log QT = .88 Log av' + 1.11 1.0 I i i i i i — i i i i | i i i i i i i i i | i — i i i i—r—rn—r 1.0 1.5 2.0 2 Log of Vertical Effective Stress (kPa) 178 6.0 - , £ 5.0 x ccf c | 4.0 c CO ? ' 3.0 cr o 2.0 o 1.0 McDonald Farm Clayey SILT 15 to 45 m Log Gmax = .76 Log av' + 3.10 Log QT = 1.0 Log a v' + .67 1 1 1 1 1 1 1 1 1 1 2.0 2.5 3.0 Log of Vertical Effective Stress (kPa) Figure 7.5 Influence of stress on Gmax and qp, Clay 5.0 4.0 CO Q. x CO E CD ~o CO I 2.0 H CO o 1 . 0 # 6 Road Richmond Log Gmax = .75 Log av ' + 3.26 Log QT = .65 Log av ' + 2.78 i i i i i i i i i | i i i i i i i i i I i i i i i i i i i | i i i i i i i i i | 1.0 1 .5 2.0 2.5 3.0 Log of Vertical Effective Stress (kPa) CO 1 -cr CO o 5.0 4.0 -3.0 2.0 1.0 Annacis N —1 Log QT = .74 Log av ' + 2.29 i i i i i i i i i I i i i i i i i i i I i i i i i i i i i I i i i i i i i i. i I 1.0 1.5 2.0 2.5 3.0 Log of Vertical Effective Stress (kPa) Figure 7.6 Influence of stress on Gmax and qp, Sand 180 Unfortunately unless information i s available i t also assumes that density and lateral stress conditions, for example Kq, remain constant with depth—an unlikely situation. Only at clay sites where accurate moisture content information i s available can an analysis of this type be used with confidence. The observations presented here are based, in part, on sites for which no void ratio data are available; nevertheless the conclusions are supported by observations from a variety of sites. Another report of the stress dependency of cone resistance i s that of Olsen and Farr (1986). Based upon extensive f i e l d observations Olsen and Farr (1986) report a normalization exponent of 0.7 for sands. At higher stress levels i t i s often observed that cone resistance values do not increase substantially with increased stress (depth). Deep CPTs reported by Campanella, et a l . (1984), obtained i n moist tailings, showed l i t t l e increase in cone resistance with depth beyond a ver t i c a l effective stress of approximately 500 kPa. The influence of factors such as crushing, which depend on density, grain characteristics, and the ambient stress level, probably become more important at these greater depths. It may therefore be that the effect of increased stress levels reduces at very high stress levels. The difference in the influence of stress observed in laboratory and f i e l d tests may result from the effects of aging. Other normalization parameters such as mean effective stress, which result in a dimensionless cone resistance, appear to overemphasise the effects of stress level on cone resistance and introduce the uncertain horizontal stress. The dependence of G^x on stress levels was also considered. The 181 dependence of G m a x obtained from downhole shear wave velocity measurements was very clearly shown experimentally by Stokoe, et a l . (1986) to be of the form c — rrt 0.25 * 0.25 °max ~ c l ° v ° h This work was performed using very large (2 m * 2m * 2m) cubic samples of dry sand and controlling the stresses on a l l boundaries. The vertical and horizontal effective stresses act in the directions of propagation and particle motion. Stresses in these two directions were shown by Stokoe, et a l . to be equally important. The magnitude of the stress in the third direction was shown to be unimportant. If the stress dependency reported by Stokoe, et a l . i s converted to a single stress, either mean normal stress or vertical stress, the influence of stress would be a single exponent 0.5. Other values reported include those by Wroth, et a l . (1984) who state that low strain G values depend on mean normal stress to the power 0.33 and that the exponent increases with strain. To investigate the dependence of shear wave velocity or Gmax o n stress level, in this case vertical stress, the log of G^Q^ was plotted against log vertical effective stress where G^x was available at deep uniform sites. Examples of this analysis were included in Figure 7.5 and 7.6. As with cone resistance the effects of changes in void ratio and lateral stress, Ko, conditions were not considered. This assumption i s s t r i c t l y proven only at the Ons0y site . The results from a l l sites showed that ^ was more dependent on stress than seen in laboratory tests. At sand sites the exponent clearly averaged 0.7 and at clay sites a larger exponent averaging 0.9 was observed. As in the case of the normalization of cone resistance i t may be that the 182 effects of aging change the dependence of G^x o n stress levels. Field data consistently indicate that both cone resistance and Gmax data depend on stress level with a single stress level exponent of approximately 0.7 i n sands. The dependence of each i n clays i s greater, approaching 1.0. For a l l s o i l s tested the ratio, G/qT, i s reasonably independent of stress level. 7.5.3 Correlation Between GJ^QX and q T in Sand Seismic cone data collected at sand sites indicate that G^x can be predicted well i n sand deposits provided that density variations are accounted for by using normalized cone resistance. Based upon f i e l d observations, discussed above, cone resistance was normalized by dividing by the estimated vertical effective stress raised to the power of 0.7. The detailed results from several sand sites are shown in Figure 7.7. The large variation in G/qT with density, as indicated by normalized cone resistance, shows that some care must be used before assuming a ratio common to a l l sands. A commonly used ratio of, G/qj. = 7.5, provides a reasonable prediction of G m a x from q T for most sands with the exception of s i l t y sands, loose sands, 1 and very dense sands. The addition of small amounts of s i l t has the effect of lowering the cone resistance but not changing Gjjj^. The addition of s i l t i n the sand affects both the permeability and the compressibility. Changes to either of these two parameters may influence stiffness and strength characteristics affecting the cone resistance. Given that the pore pressures are only a small fraction of the cone resistance, i t appears that the addition 183 30 - i LEGEND: 25 20 oo 0 McDonald Farm x # 6 Road Richmond + P i l e test Site o Holmen, Norway & Knight St. Richmond a Annacis Site A Richards Island x cd E 1 5 H O 10 o o o o 4>0 DO 0 ^ + + oD # b ° 0 < > 0 x 0 1—I—I—I—I—I—I—I—I—I—I—I—1—1—I—I—1—I—I—I—I—I—I I—I—I—I—I I—I 0 100 200 _ 3 0 0 Normalized Cone Resistance (q-r/o-1 v ' 7 ) Figure 7.7 Variation of Gmax/qj Ratios with Normalized Cone Resistance 184 of s i l t increases the large strain compressibility and reduces the cone resistance. In contrast the small strain stiffness appears relatively unaffected by the addition of s i l t . Very loose clean sand deposits such as the Holmen si t e were also observed to have high G/qp ratios. On the other end of the density scale, very dense sands, those with very high normalized cone resistance, have a lower than average G/qT ratio. This i s l i k e l y due to the influence of dilation on the behaviour of dense sands. The effect of dilation i s manifested in the large strains associated with the CPT but i s less important i n the small strain region associated with shear wave velocity measurement. The variation of G/qT shown in Figure 7.7 appears to indicate thatG/q T depends more on density and grain size than on stress history, at least for the sites selected. It appears that measurements, l i k e cone resistance, are insensitive to stress history and fabric effects and that the ratio G^x/gpi i s unlikely to be useful in distinguishing stress history of sands. 7.6 Application of P wave Velocity Data Compression wave velocity data was discussed i n previous chapters. Theoretically, compression wave velocity and ratios of compression to shear wave velocity should also be a useful parameter for characterizing s o i l s . The type of soil s that permit CPT work, however, restricts the range of normally encountered compression wave velocities to a very small range. In saturated s o i l s , compression wave velocities seldom are lower than 1500 m/s. Soils that can be penetrated by the 185 CPT seldom have compression wave velocities that are higher than 1700 m/s. This narrow range, combined with measurement errors, reduces the usefulness of compression wave measurements for correlation to specific geotechnical parameters. One important application of compression wave velocity measurement follows from i t s sensitivity to the s o i l saturation state. The influence of the degree of saturation has been shown by Ishihara (1967). The results of the experimental work performed indicate that i f a s o i l i s not completely saturated the velocity i s much lower. It does not appear to be possible to distinguish the degree of saturation of p a r t i a l l y saturated soil s but i t i s clearly feasible to determine i f a s o i l i s completely saturated. Saturated s o i l s have velocities greater than 1450 m/s while unsaturated soils have velocities less than 1000 m/s. 7.7 Conclusions: 1) Different methods of using shear wave velocity data for the evaluation of liquefaction potential were discussed. These methods are considered to be supplemental to the use of cone resistance methods for evaluating liquefaction potential. Of the methods outlined a direct correlation between shear wave velocity, corrected to 1 atmosphere, and c y c l i c stress ratio i s considered to be most appropriate for many engineering problems. Compared to penetration based relations velocity methods have the advantage of being independent of grain size and may prove most useful in s i l t y sands. The major uncertainty with this method i s the lack of f i e l d data collected prior to liquefaction. The 186 major advantage i s the accuracy and economy of simultaneous collection of cone resistance and shear wave velocity data. Data collected at the Imperial Valley sites unfortunately f a l l clearly on the side of liquefaction and do not provide much constraint to the relation developed from SPT-shear wave velocity correlations. 2) No correlation between G^x, normalized with cone resistance and cone factor N K was found. Dependence of N K on a stiffness ratio term i s expected from consideration of both cavity expansion theory and OCR or Bg based interpretations. It i s hypothesised that other factors, in some cases poorly constrained, especially s o i l sensitivity, dominate the variation i n N K. The use of a low strain stiffness to assist in the estimation on N K does not appear to be warranted. 3) The application of G^x data in s o i l c l a s s i f i c a t i o n was found to be useful as a tool secondary to cone resistance and f r i c t i o n based interpretations. An example was shown and a chart was developed showing the variation i n G/qT for different s o i l types. This chart showed the wide range i n G/qT found at different clay sites but did not indicate a reasonable means of clearly distinguishing overconsolidated clays from loose s i l t y s o i l s . The distinction of these two s o i l types can be d i f f i c u l t because of f r i c t i o n measurement problems. A specific example was shown il l u s t r a t i n g how Figure 7.3 could be used to augment cone resistance and f r i c t i o n based interpretation methods. 4) The dependence of G^x and cone resistance on stress level was investigated by plotting the log of G m a x and q T vs. the log of the ve r t i c a l effective stress and finding the slope of the best f i t relation for each sit e . Both and q T were found to be proportional 187 to the vertical effective stress raised to the power 0.7 + 0.15 in sands and 0.9 + 0.1 i n clays. Both of these findings are contrary to chamber based studies but are consistent with other f i e l d experience. The similar dependence of both q T and o n stress level (or some known index of it) allows comparison of the ratio Gmax/^T at different stress levels. The dependence of cone resistance on stress levels also appears to be dependent on cone resistance, s o i l compressibility, and stress level. For example at high stress levels increased ambient stress has a decreased effect on cone resistance. 5) In sands a ratio Gmax/^T i s often used to estimate G^x from q T, this ratio was explored in Figure 7.7. The determination of an appropriate ratio was shown to vary with normalized cone resistance. No single value was found for a l l sands but excellent correlations were established provided normalized bearing was considered. 6) P wave velocity data should be useful i n determining s o i l parameters but measurement limitations across the very short depth interval used in practice and the short range in anticipated values in saturated s o i l s presently limits the application of P wave data determined with downhole interval techniques to the distinction of unsaturated s o i l s from saturated s o i l s . 188 CHAPTER 8. SUMMARY AND CONCLUSIONS 8.1 Summary of Pore Pressure Data Pore pressure response during cone penetration for any given s o i l i s primarily controlled by the location of the porous element. The significance of the element location depends on s o i l type and no single element location best serves a l l purposes. Measurements made behind the f r i c t i o n sleeve gave similar results to those measured behind the t i p except in soft clays where pore pressures were slightly lower behind the sleeve. The similarity of measurements behind the t i p and sleeve makes the simultaneous measurement of pore pressures on the face of the t i p and behind the sleeve an ideal combination but q T corrections then require some interpolation to obtain the pore pressure between the two measurement locations. This measurement i s required for correction of cone resistance measurements because of the lack of uniformity in cone design and was shown to account for much of the variation i n cone readings obtained with different cones at selected soft clay test sites. Pore pressure response for any given cone design i s dominated by drainage characteristics and can therefore be used as an indication of s o i l type. The distinction can be made between s o i l s acting i n a drained or undrained manner. Pore pressure dissipation rates were also found to be a useful index of s o i l type but cannot be used to identify detailed grain size characteristics. A common interpretation problem i s the distinction of s i l t s and overconsolidated clays where f r i c t i o n readings are often insufficiently reliable for classification purposes. Pore pressure decay rates can be 189 used to distinguish these s o i l s . In s i l t y or fine sandy so i l s pore pressures during cone penetration may be higher or lower than the static values. The sign of the response i s a function only of the measurement details. Pore pressures on the face of the cone were found, in granular s o i l s , to be positive with very few exceptions. Pore pressures behind the cone t i p were found to be less than static i n nearly a l l cases. This response results from the very large normal stresses on the cone face and the reduction to much lower stresses behind the t i p . Various application of pore pressure measurements were investigated. Pore pressure measurements were found, with restrictions, useful for the determination of undrained strength. The most successful approach to determine OCR appears to be the direct correlation of normalized cone resistance to OCR. A dependence of normalized resistance on s o i l p l a s t i c i t y was indicated and reasonable correlations were shown. A l l published cone resistance interpretation methods for liquefaction resistance are highly dependent upon grain size. Pore pressure response, drained or undrained, was shown to be a reasonable way of deciding the most appropriate CPTU based liquefaction analysis in many s o i l s . 8.2 Summary of Shear Wave Velocity Data The high quality of the shear wave traces recorded onshore i s primarily a result of two factors: the excellent coupling between the s o i l and the cone penetrometer, which has the receiver firmly embedded 190 inside; and the use of SH waves. The accuracy and repeatability of the interval technique appears to be very high for depths of penetration less than 30 m. At greater depths signal enhancement techniques become more important. Downhole velocity measurements were compared to conventional crosshole measurements at a number of sites. At three sites where high quality data were available good comparisons were obtained, thus increasing confidence i n the CPT downhole technique. Different methods of using shear wave velocity data for the evaluation of liquefaction potential were discussed. These methods were considered secondary to the use of cone resistance methods for evaluating liquefaction potential. A direct correlation between shear wave velocity, corrected to 1 atmosphere, and the c y c l i c stress ratio required to i n i t i a t e liquefaction offers the advantage of being independent of grain size and therefore may prove useful in s i l t y sands. The major uncertainty with this method i s the lack of f i e l d data collected prior to liquefaction. The major advantage i s the accuracy and economy of simultaneous collection of cone resistance and shear wave velocity data. Data collected at the Imperial Valley sites unfortunately f a l l clearly on the side of liquefaction and do not provide much constraint to the relation developed from SPT-shear wave velocity correlations. The application of data in s o i l c lassification was found to be useful as a supplementary tool to cone resistance and f r i c t i o n based interpretations. An example was shown and a chart was developed showing the variation in G/qT for different s o i l types. This chart 191 showed the wide range i n G/q-p found at different clay sites but did not indicate a means of clearly distinguishing overconsolidated clays from loose s i l t y s o i l s . The distinction of these two s o i l types can be d i f f i c u l t because of f r i c t i o n measurement problems. The dependence of Gjaax cone resistance on stress level was investigated. Both Gm^ and q T were found to be proportional to the v e r t i c a l effective stress raised to the power 0.7 + 0.15 i n sands and 0.9 + 0.1 in clays. The key assumption in this analysis was that void ratios did not vary significantly with depth; this assumption could only be verified at one clay si t e . The similar dependence of both qp and Gjnax on stress level (or some known index of it) allows comparison of the ratio Gjjjax/qrp at different stress levels. At high stress levels increased ambient stress has a decreased effect on cone resistance. In sands a ratio G^ax/qip was shown to vary with normalized cone resistance. The relation shown can be used to predict G^ax from q T. 8.3 Recommendations for Future Research 1) S i l t remains the most d i f f i c u l t s o i l to interpret from CPTU data. Its addition into sands greatly increases the sand compressibility and thereby affects the penetration resistance. An interpretation method that accounts for the effect of compressibility of s i l t y s o i l s i s required. Accounting for the effect of compressibility w i l l not be an easy task; an extensive series of chamber tests would have to vary stress levels, density and s i l t content. Uniform placement of s i l t y sands i n a chamber w i l l not prove easy. Possible effects of s o i l fabric are more l i k e l y to be important 192 in a mixed s o i l than i n the uniform sand samples tested to date in the chamber. Perhaps f i e l d testing and detailed sampling at silt/sand sites may prove more productive than laboratory chamber work. From a lo g i s t i c a l point of view the use of hydraulic f i l l preloads tested after construction and subsampled during the unloading phase may allow density measurements to be more easily made. 2) The accurate identification of so i l s s t i l l r e l i e s on f r i c t i o n data. Equipment improvements that lead to increased accuracy and repeatability of f r i c t i o n sleeve data are required. The sensitivity of fr i c t i o n sleeve data to exact dimensions of the t i p and sleeve has not been well documented. 3) Improvement to dissipation rate based estimates of permeability and consolidation parameters could be made by comparing f i e l d performance to data collected in the site investigation phase. 4) Although some suggestions have been made in this thesis, the distinction of overconsolidated clays and clay s i l t mixtures i s s t i l l a d i f f i c u l t problem. Future approaches should continue to examine alternate interpretations that combine data from the various sensors already available. 5) No application of shear wave data has yet included consideration of damping parameters. The variation of signal amplitude with frequency may offer a means of obtaining damping parameters from downhole testing. Repeatable broad spectrum sources may be required; the use of true interval techniques should be explored as a means of distinguishing damping effects from geometric damping. 6) The possible role of aging effects proved to be a d i f f i c u l t 193 problem at some of the sites encountered i n this study and in the published literature. Some sites appear not to change with time after compaction while others show large changes in s o i l properties with time following compaction. An investigation that includes repeated soundings with tests at different strain levels (shear wave velocity and cone resistance) for various times after compaction could offer some insite into this phenomena. 7) P wave velocity measurements can give an accurate indication that unsaturated s o i l s are present. The degree of saturation i s more d i f f i c u l t to determine. Further work in this area and in the behaviour of p a r t i a l l y saturated soils i s warranted. 8) Stress correction curves are unavailable at low stresses. These curves could be developed with testing in uniform s o i l s but may require specially b u i l t equipment. 194 REFERENCES Aas, G., Lacasse, S., Lunne, T. and Hoeg, K. (1986), "Use of In-Situ Tests for Foundation Design i n Clay", Proceedings of In-Situ 86, Geotechnical Special Publication No. 6, ASCE, Blacksburg, Virginia, pp. 1-30. Baldi, G. (1986) , "Interpretation of CPT's and CPTU's ; 2nd Part: Drained Penetration Testing i n Sands", Proceedings IV International Geotechnical Seminar on Field Instrumentation and In-Situ Measurements, Singapore, pp. 143-156. Baligh, M.M. and Levadoux, J.N. (1980), "Pore Pressure Dissipation After Cone Penetration", Massachusetts Institute of Technology, Department of C i v i l Engineering, Construction F a c i l i t i e s Division, Cambridge, Massachusetts. Baligh, M.M. and Levadoux, J.N. (1986), "Consolidation after Undrained Piezocone Penetration", Journal of Geotechnical Division, ASCE, Vol. 112, No. GT 7, pp. 727-745. Been, K., Crooks, J. H. A., Becker, D.E. and Jeffries, M.G. (1986), "The Cone Penetration Test in Sands: Part 1, State Parameter Interpretation", Geotechnique, Vol. 36, No. 2, pp. 239-249. Beeston, H.E. and McEvilly, T.V. (1977), "Shear Wave Velocities from Downhole Measurements", Earthquake Engineering and Structural Dynamics, Vol. 5, No. 2, pp. 181-190. Be l l o t t i , R., Ghionna, V., Jamiolkowski, M., Lancellotta, R. and Robertson, P.K. (1989), "Shear Strength of Sands from CPT", Paper to the 12th International Conference Society on Soil Mechanics and Foundation Engineering (ICSMFE), Rio de Janeiro, Brazil. Bierschwale, J.G. and Stokoe, K.H. (1984), "Analytical Evaluation of Liquefaction Potential of Sands Subjected to the 1981 Westmorland Earthquake", Geotechnical Engineering Report, GR-84-15, C i v i l Engineering Dept., The University of Texas at Austin. Campanella, R.G. and Robertson, P.K. (1981), "Applied Cone Research", Symposium on Cone Penetration Testing and Experience, Geotechnical Engineering Division, ASCE, Oct. 1981, pp. 343-362. Campanella, R.G., Gillespie, D.G. and Robertson, P.K. (1982), "Pore Pressures During Cone Penetration Testing", Proceedings of 2nd European Symposium on Penetration Testing, ESOPT II, A.A. Balkema, pp. 507-512. Campanella, R.G., Robertson, P.K. and Gillespie, D.G. (1983), "Cone Penetration Testing in Deltaic Soils", Canadian Geotechnical Journal, Vol. 20, No. 1, February, pp. 23-35. 195 Campanella, R.G., Robertson, P.K., Gillespie, D.G. and Klohn, E.J. (1984) , "Piezometer-friction Cone Investigation at a Tailings Dam", Canadian Geotechnical Journal, Vol. 21, No. 3, pp. 551-562. Campanella, R.G., Robertson, P.K. and Gillespie, D.G. (1986), "Factors Affecting the Pore Water Pressures and i t s Measurement around a Penetrating Cone", Proceedings, 39th Canadian Geotechnical Conference, Ottawa, pp. 291-300. Campanella, R.G., Robertson, P.K., Gillespie, D.G. and Greig, J.W. (1985) , "Recent Developments i n In-situ Testing of Soils", Proceedings of XI International Conference Society for Soil Mechanics and Foundation Engineering (ICSMFE), San Francisco. Vol. 2, pp. 849-854. Campanella, R.G., Robertson, P.K., Gillespie, D.G., Laing, N. and Kurfurst, P.J. (1987), "Seismic Cone Penetration Testing in the Near Offshore of the MacKenzie Delta", Canadian Geotechnical Journal, Vol. 24, No. 1, pp. 154-159. de Beer, E.E., Goelen, E. Heynen, H.J. (1988), "Cone Penetration Test (CPT) International Reference test Procedure", Proceedings, International Symposium on Penetration Testing, ISOPT-1. Orlando, Florida, Proc. Vol. 1, pp. 27-52. de Ruiter, J. (1971), "Electronic Penetrometer for Site Investigations", Journal of Soil Mechanics and Foundation Engineering Division, ASCE, Vol. 97, No. SM2, pp. 457-472. de Ruiter, J. (1982), "The Static Cone Penetration Test", State of the Art Report Proceedings 2nd. European Symposium on Penetration Testing, A.A. Balkema, pp. 389-405. Dobry, R., Stokoe, K.H., Ladd, R.S. and Youd, T.L. (1981), "Liquefaction Susceptability From Shear Wave Velocity", Presented at the October 26-31, ASCE National Convention and Exposition, St. Louis. Douglas, B.J., Strutynsky A.I., Mahar, L.J. and Weaver, J. (1985), "Soil Strength Determinations From The Cone Penetrometer Test", in C i v i l Engineering in the Arctic Offshore Proceeedings of the Conference Arctic '85, Bennett and Machemehl Editors, published by American Society of C i v i l Engineers, pp.153-161. Durgunoglo, H.T. and Mitchell, J.K. (1975), "Static Penetration Resistance of Soils", Proceedings of the ASCE Speciality Conference on In-Situ Measurement of Soil Properties, Raleigh, North Carolina, USA., Vol. 1, pp. 151-189. Eidsmoen, T., Gillespie, D.G., Lunne, T. and Campanella, R.G. (1985), "Tests With UBC Seismic Cone at Three Norwegian Research Sites", UBC Dept. of C i v i l Engineering and NGI joint report, 28 pp. 196 Gillespie, D.G. (1981), "The Piezometer Cone Penetration Test", M.A.Sc. thesis submitted to the University of British Columbia, Dept. of C i v i l Engineering, 137 pp. Gillespie, D.G., Robertson, P.K. and Campanella, R.G. (1988), Discussion of "Consolidation After Undrained Piezocone Penetration. I:Prediction", Journal of Geotechnical Engineering, ASCE, Vol. 114, No. 1, pp. 126-128. Greig, J.W. (1985), "Estimating Undrained Shear Strength of Clay from Cone Penetration Tests", M.A.Sc. thesis submitted to the University of British Columbia, Dept. of C i v i l Engineering, 172 pp. Gupta, R.C. and Davidson, J.L. (1986), "Piezoprobe Determined Coefficient of Consolidation", Soils and Foundations, Vol. 26, No. 3, pp.12-22. Hardin, B.O. and Drenevich, V.P. (1972), "Shear Modulus and Damping in Soils: Design Equations and Curves", Journal of So i l Mechanics and Foundation Division, ASCE, Vol. 98, SM7, pp. 667-692. H i l l , P.R., Moran, K., Kurfurst, P.J. and Pullen, S. (1986), "Physical and Sedimentological Properties of Nearshore Sediments i n the Southern Beaufort Sea", Proceedings of the 3rd Canadian Conference on Marine Geotechnical Engineering", Vol. 1, pp. 301-327. Houlsby, G.T. and Hitchman, R. (1987), "Calibration Chamber Tests of a Cone Penetrometer in Sand", Geotechnique, Vol. 38, No. 1, pp. 39-44. Houlsby, G.T. and Wroth, CP. (1989), "The Influence of So i l Stiffness and lateral stress on the Results of in-Situ Tests", Paper to the 12th International Conference Society on Soil Mechanics and Foundation Engineering (ICSMFE), Rio de Janeiro, Brazil. Ishihara, K. (1967), "Propagation of Compressional Waves in a Saturated So i l " , Proceedings of the International Symposium on WAve Propagation and Dynamic Properties of Earth Materials, SMFD, ASCE, Albuquerque, New Mexico, pp. 451-467. Jamiolkowski, M., Ladd, C.C., Germaine, J.T. and Lancellotta, R. (1985), "New Developments in Field and Laboratory Testing of Soils", Theme Lecture at the 11th International Conference Society for Soil Mechanics and Foundation Engineering (ICSMFE), San Francisco, pp. 57-154. Jones, G.A. and Rust, E.A. (1982), "Piezometer Penetration Testing CPTU", Proceedings of the 2nd European Symposium on Penetration Testing, ESOPT II, Amsterdam, Vol. 2, pp. 607-613. Keaveny, J.M. and Mitchell, J.K. (1986), "Strength of Fine Grained Soils Using the Piezocone", Proceedings of In-Situ 86, Geotechnical Special Publication No. 6, ASCE, Blacksburg, Virginia pp. 668-685. 197 Konrad, J.M., Bozozuk, M. and Law, T. (1985), "Study of In-Situ Test Methods i n Deltaic S i l t " , Proceedings of XI International Conference Society for Soil Mechanics and Foundation Engineering (ICSMFE), San Francisco. Vol. 2, pp. 879-886. Konrad, J.M. and Law, T. (1987), "Preconsolidation Pressures from Piezocone Tests in Marine Clays", Geotechnique, Vol. 37, No. 2, pp. 177-190. Kurfurst, P. (1986), Personal communication Laing, N.L. (1983), "An Examination of the SPT", B.A.Sc. thesis submitted to the University of British Columbia, Dept. of Geological Sciences. Laing, N.L. (1985), "Sources and Recievers With The Seismic Cone Test", M.A.Sc. thesis submitted to the University of Britis h Columbia, Dept. of C i v i l Engineering, 108 pp. La Rochelle, P., Zebdi, P.M., Leroueil, S., Tavenas, F. and Virely, D. (1988), "Piezocone Tests in Sensitive Clays of Eastern Canada", Proceedings, International Symposium on Penetration Testing, ISOPT-1. Orlando, Florida, Proc. Vol. 2, pp. 831-841. Layotte, P.C. (1984), "Marthor: an S-Wave Impulse Source", Presented at the Society of Exploration Geophysicists Shear-Wave Exploration Symposium, Midland, Texas. Lacasse, S. and Lunne, T. (1982), "Penetration Tests in Two Norwegian Clays", Proceeding of the 2nd. European Symposium on Penetration Testing, Amsterdam, Vol. 2, pp. 661-669. Levadoux, J.N. and Baligh, M.M. (1986), "Consolidation After Undrained Piezocone Penetration I: Prediction", Journal of Geotechnical Division, ASCE, Vol. 12, No. 7, pp. 707-726. Lunne, T., Christoffersen, H.P. and Tjelta, T.I. (1985), Engineering Use of Piezocone Data in North Sea Clays", Proceedings of XI International Conference Society for Soil Mechanics and Foundation Engineering (ICSMFE), San Francisco, pp. 907-912. Lunne, T., Eide, O. and de Ruiter, J. (1976), "Correlations Between Cone Resistance and Vane Shear Strength in Soft to Medium S t i f f Clays", Canadian Geotechnical Journal, Vol. 13, No. 4, pp. 430-441. Lunne, T., Eidsmoen, T., Gillespie, D. and Howland, J.D. (1986), "Laboratory and Field Evaluation of Cone Penetrometers", Proceedings of In-Situ 86, Geotechnical Special Publication No. 6, ASCE, Blacksburg, Virginia, pp. 714-729. 198 Marcuson, W.F. and Biegonousky, W.S. (1977), "Laboratory Standard Penetration Tests on Fine Sand", Journal of Geotechnical Engineering Division, ASCE, Vol. 103, No. GT 6, pp. 565-588. Massarsch, K.R. and Broms, B.B. (1981), "Pile Driving in Clay Slopes", Proceedings of International Conference on Soil Mechanics and Foundation Engineering, Stockholm. Mitchell, J.K. and Keaveny, J.M. (1986), "Determining Sand Strength by Cone Penetrometer", Proceedings of In-Situ 86, Geotechnical Special Publication No. 6, ASCE, Blacksburg, Virginia, pp. 823-839. Nazarian, S. and Stokoe, K.H. (1984), "In-situ Shear Wave Velocities from Spectral Analysis of Surface Waves", Proceedings of the Eighth World Conference on Earthquake Engineering, Vol. I l l , pp. 31-38. Olsen, R.S. (1984), "Liquefaction Analysis Using the Cone Penetration Test", Proceedings of the 8th World Conf. on Earthquake Engineering, San Fransisco. Olsen, R.S. and Farr, J.V. (1986), "Site Characterization Using the Cone Penetration Test", Proceedings of In-Situ 86, Geotechnical Special Publication No. 6, ASCE, Blacksburg, Virginia, pp. 854-868. Pullen, S.E. and MacAulay, H.A. (1984), "A New Source for Engineering Seismic Surveys", Unpublished Report, Geological Survey of Canada. Randolph, M.F. and Wroth, CP. (1979), "An Analytical Solution for the Consolidation Around a Driven Pile", International Journal for Numerical and Analytical Methods i n Geomechanics, Vol. 3, pp. 217-229. Rice, A. (1984), "The Seismic Cone Penetrometer", M.A.Sc. thesis submitted to the University of British Columbia, Dept. of C i v i l Engineering, 112ppi Robertson, P.K. (1983), "In-situ Testing of Soils with Special Emphasis on i t s Application to Liquefaction Assessment", Ph.D. Thesis, Dept. of C i v i l Engineering, University of British Columbia. Robertson, P.K. and Campanella, R.G. (1983), "Interpretation of Cone Penetration Tests - Part I (Sand)", Canadian Geotechnical Journal, Vol. 20, No. 4, pp. 718-733. Robertson, P.K. and Campanella, R.G. (1983), "Interpretation of Cone Penetration Tests - Part II (Clay)", Canadian Geotechnical Journal, Vol. 20, No. 4, pp. 734-745. Robertson, P.K. and Campanella, R.G. (1984), "Guidelines for Use, Interpretation and Application of the CPT and CPTU", The University of British Columbia, Dept. of C i v i l Engineering, Soil Mechanics Series, No. 105. 199 Robertson, P.K. and Campanella, R.G. (1985), "Evaluation of Liquefaction Potential of Sands Using the CPT", Journal of Geotechnical Engineering, ASCE, Vol. I l l , No. 3, pp. 384-407. Robertson, P.K., Campanella, R.G., Gillespie, D.G. and Greig, J.W. (1986), "Use of Piezometer Cone Data", Proceedings of In-Situ 86, Geotechnical Special Publication No. 6, ASCE, Blacksburg, Virginia, pp. 1263-1280. Robertson, P.K., Campanella, R.G., Gillespie, D.G. and Rice, A. (1986b), "Seismic CPT to Measure In-situ Shear Wave Velocity", Journal of Geotechnical Engineering, ASCE, Vol. 112, No. 8, pp. 791-803. Robertson, P.K., Campanella, R.G. and Wightman, A. (1983), "SPT-CPT Correlations", Journal of Geotechnical Engineering, ASCE, Vol. 109, No. 11, pp. 1449-1459. Roy, M., Tremblay, M., Tavenas, F. and LaRochelle, P. (1982), "Development of Pore Presures in Static Penetration Tests in Sensitive Clay", Canadian Geotechnical Journal, Vol. 19, No. 2, May 1982, pp. 124-138. Schmertmann, J.H. (1976), "An Updated Correlation Between Relative Density DR, and Fugro Type Electric Cone Bearing, g^.", Unpublished Report to WES, Vicksburg. Schmertmann, J.H. (1978), "Guidelines for Cone Penetration Test, Performance and Design", Federal Highway Administration, Report FHWA. TS.78.209, Washington, July 1978, 145 pp. Seed, H.B., Idriss, I.M. and Arango, I. (1983), "Evaluation of Liquefaction Potential Using Field Performance Data", Journal of Geotechnical Engineering, ASCE, Vol. 109, No. 3, pp. 458-482. Seed, H.B., Tokimatsu, K., Harder, L.F. and Chung, R.M. (1985), "Influence of SPT Procedures in Soil Liquefaction Resistance Evaluations", Journal of Geotechnical Engineering, ASCE, Vol. I l l , No. 12, pp. 1425-1445. Seed, H.B. and De Alba, P. (1986), "Use of SPT and CPT Tests for Evaluation of Liquefaction Restistance of Sands", Proceedings of In-Situ 86, Geotechnical Special Publication No. 6, ASCE, Blacksburg, Virginia, pp. 281-300. Seed, H.B. and Idriss, I.M. (1970), "Soil Moduli and Damping Parameters for Dynamic Response Analysis", Report to SWAJA, Earthquake Engineering Research Center, University of California, Berkeley, Report No. EERC 70-10, pp. 1-40. Senneset, K., Janbu, N. and Svano, G. (1982), "Strength and Deformation Parameters from Cone Penetrations Tests", Proceedings of the European Symposim on Penetration Testing, ESOPT II, Amsterdam, pp. 863-870. 200 Senneset, K. and Janbu, N. (1985), "Shear Strength Parameters Obtained from Static Cone Penetration Tests", ASTM STP 883, Strength Testing of Marine Sediments: Laboratory and Marine Sediments, San Diego, pp. 41-54. Shibata, T. and Teparaksa W. (1988), "Evaluation of Liquefaction Potentials of Soils Using Cone Penetration Tests", Soils and Foundations, Vol. 28, No. 2, pp. 49-60. Smits, F.P. (1982), "Penetration Pore Pressures Measured with Piezometer Cones", Proceeding of the 2nd. European Symposium on Penetration Testing, Amsterdam, Vol. II., pp. 877-881. Stokoe K.H. and Hoar R.J. (1977), "Field Measurements of Shear Wave Velocity by Crosshole and Downhole Seismic Methods", Proceedings of the DMSR 77, Karlsruhe, Vol. 3, pp. 115-137. Stokoe K.H., Lee S.H.H. and Knox, D.P. (1986), "Shear Moduli Measurements Under True Triaxial Stresses", in Advances in the Art of Testing Soils Under Cyclic Conditions, Khosla Editor, pp. 166-185. Stokoe K.H. and Woods R.D. (1972), "In-Situ Shear Wave Velocity by . Cross-hole Method", Journal of the Soil Mechanics and Foundations Division, ASCE, Vol. 98, No. SM5, pp. 443-460. Sully, J.P., Campanella, R.G., and Robertson, P.K. (1988), "Interpretation of Penetration Pore Pressures to Evaluate Stress History in Clays", Proceedings, International Symposium on Penetration Testing, ISOPT-1. Orlando, Florida, Proc. Vol. 2, pp. 993-999. Suyama, K., Imai, T., Ohtomo, H., Ohta, K. and Takahashi, T. (1987), "Dilineation of Strusctures in Alluvium and Diluvium using SH-Wave Reflection and VSP Methods", Shear-wave Exploration, Danbom and Domenico Editors, Society of Exploration Geophysicists, Geophysical Development Series, Volume 1, pp. 165-179. Torstensson, B.A. (1975), "Pore Pressure Sounding Instrument", Proceedings of the ASCE Speciality Conference on In-Situ Measurement of S o i l Properties, Raleigh, North Carolina, Vol. 2, pp. 48-54. Torstensson, B.A. (1977), "The Pore Pressure Probe", Norsk jord- og fjellteknisk forbund. Fjellsprengningsteknikkbergmekanikk-Geoteknikk. Oslo, Foredrag, 34.1-34.15, Trondheim, Norway. Tapir. Treadwell, D.D. (1975), "The Influence of Gravity, Prestress, Compressibility, and Layering on Soil Resistance to Static Penetration", Ph.D. Dissertation, University of California, Berkley, 210 pp. Vesic, A.S. (1972), "Expansion of Cavities in Infinite Soil Masses", Journal of the Soil Mechanics and Foundation Division, ASCE, Vol. 98, SM3, pp. 265-290. 201 Vesic, A.S. (1977), "Design of Pile Foundations", National Cooperation Highway Research Program. Report No. 42, Transportation Research Board Washington, D.C. White, J.E. (1965), "Seismic Waves: Radiation, Transmission, and Attenuation", International Series in the Earth Sciences, Shrock Editor, 302 pp. Wissa, A.E.Z., Martin, R.T. and Garlanger, J.E. (1975), "The Piezometer Probe", Proceedings of the ASCE Speciality Conference on In-Situ Measurement of Soil Properties, Raleigh, N.C, Vol. 1, pp. 536-545. Wroth, CP. (1984), "The Interpetation of In-Situ Soil Tests", XXIV Rankine Lecture, Geotechnique 34, pp. 449-489. Youd, T.L. (1983), Personal communication Youd, T.L. and Bennett, M.J. (1983), "Liquefaction Sites, Imperial Valley, California", Journal of Geotechnical Engineering, ASCE, Vol. 109, No. 3, pp. 440-457. Zhou, S.G. (1981), "Influence of Fines on Evaluating Liquefaction of Sands by CPT", Proceedings, International Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics", St. Louis, Vol. 1, pp. 167-172. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0050502/manifest

Comment

Related Items