UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Split spoon penetration testing in gravels Daniel, Christopher Ryan 2000

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

Item Metadata

Download

Media
831-ubc_2000-0379.pdf [ 6.66MB ]
Metadata
JSON: 831-1.0064071.json
JSON-LD: 831-1.0064071-ld.json
RDF/XML (Pretty): 831-1.0064071-rdf.xml
RDF/JSON: 831-1.0064071-rdf.json
Turtle: 831-1.0064071-turtle.txt
N-Triples: 831-1.0064071-rdf-ntriples.txt
Original Record: 831-1.0064071-source.json
Full Text
831-1.0064071-fulltext.txt
Citation
831-1.0064071.ris

Full Text

SPLIT SPOON PENETRATION TESTING IN G R A V E L S by CHRISTOPHER R Y A N DANIEL B .A .Sc , The University of British Columbia, 1997 A THESIS SUBMITTED PW PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF M A S T E R OF APPLIED SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Civil Engineering; Geotechnical Engineering) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A October 2000 © Christopher Ryan Daniel, 2000 In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s f o r an advanced degree a t t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , I agree t h a t t h e L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d by t h e head o f my department o r by h i s o r h e r r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l n ot be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . Department o f C / 'u i ' -^ Ei/it^iiAeeW</>g^ The U n i v e r s i t y o f B r i t i s h Columbia Vancouver, Canada Date ^ c / o ^ e r /3 , 3 o & o A B S T R A C T The widely used "Standard Penetration Test" (SPT) is considered unreliable for gravel deposits because gravel particles can be larger than the opening of the SPT split-spoon sampler and because drilling methods normally employed for SPT in sands are often impractical in gravels. The "Large Penetration Test" (LPT) potentially reduces the effect of the former through the use of oversized split-spoon samplers. This thesis presents a method of predicting SPT from LPT blow counts so that currently available empirical design methods based on the SPT may be used in gravel deposits. The proposed method considers input energy and sampler dimensions and can be used with any LPT system (i.e. any combination of hammer and split-spoon). The results of the proposed method are compared to those of an existing correlation method and to empirical SPT-LPT correlation factors from the literature, including correlations for the "Japanese LPT" (JLPT) and "Italian LPT" (ILPT). In addition, three field research programs were conducted to develop an empirical correlation factor between the SPT and the "North American LPT" (NALPT). Field tests revealed that the proposed correlation method requires an empirical correction factor of 0.82. Review of the JLPT data set revealed that the SPT-JLPT correlation factor might be unreliable. If the JLPT data is excluded and if the empirical correction factor is applied, it is found that the correlation factors predicted using the proposed method range from 83% to 96% of those observed in sands. The equivalent range for the existing correlation method was 39% to 73%. The author attempted to obtain preliminary insight into the problem of grain size effects using the available test data. The observed SPT-NALPT and SPT-ILPT correlation factors appear to decrease with increasing grain size. Data illustrating a fair relationship between the observed correlation factor and the portion of the soil that is too coarse to enter the SPT split-spoon is presented. T A B L E O F C O N T E N T S A b s t r a c t i i L i s t o f T a b l e s vii L i s t o f F i g u r e s viii L i s t o f S y m b o l s a n d A b b r e v i a t i o n s xiii A c k n o w l e d g e m e n t s xv 1. I N T R O D U C T I O N 1 2 . S P T A N D L P T D E T A I L S A N D C O R R E L A T I O N S 3 2.1 S t a n d a r d P e n e t r a t i o n Tes t ( S P T ) 3 2.2 N o r t h A m e r i c a n L P T ( N A L P T ) 5 2.3 J a p a n e s e L P T ( J L P T ) 12 2.4 I t a l i a n L P T ( I L P T ) 15 3. D Y N A M I C P E N E T R A T I O N T E S T I N G E N E R G Y T H E O R Y 20 3.1 K i n e t i c E n e r g y o f D r i v i n g Sys tems 21 3.2 St ress W a v e T h e o r y 22 3.2.1 Characteristics of Stress Waves 22 3.2.2 Application to SPT and LPT Energy Measurement 24 3.2.2.1 Force Squared (FF) Method 25 3.2.2.2 Force Velocity (FV) Method 27 3.2.3 Stress Wave Modelling 28 4. P R O P O S E D S P T - L P T C O R R E L A T I O N M E T H O D 33 4.1 S o i l Res i s t ance C o n s i d e r a t i o n s 33 iii 4.2 Energy Input Considerations 36 4.3 Synthesis of Proposed Method 46 4.4 Application of Proposed Method 50 5. KIDD2 N A L P T F I E L D P R O G R A M 53 5.1 Drilling Method 57 5.2 Quasi-Static Penetration Tests 57 5.2.1 Description of Test Method 57 5.2.2 Test Results 58 5.2.3 Discussion 58 5.3 Dynamic Penetration Tests 63 5.3.1 Description of Test Method 63 5.3.2 Energy Measurement 64 5.3.3 Test Results 68 5.4 Discussion of Energy Data 69 5.4.1 Data Repeatability 69 5.4.2 Calibration Factors 76 5.4.3 Quality Control Using Upper Bounds 80 5.4.4 Quality Control Using Force-Velocity Proportionality 82 5.5 Calibration of Proposed Method 87 6. S E W A R D , A L A S K A N A L P T F I E L D P R O G R A M 90 6.1 Drilling Method 92 iv 6.2 Dynamic Penetration Tests 92 6.2.1 Description of Test Method 92 6.2.2 Energy Measurement 93 6.2.3 Test Results 94 6.3 Discussion 95 6.3.1 Grain Size Analysis Results 95 6.3.2 Blow Count Repeatability 98 6.3.3 Energy Data Quality 100 6.4 Correlation Factor 106 7. K E E N L E Y S I D E D A M N A L P T F I E L D P R O G R A M 110 7.1 Drilling Method 113 7.2 Dynamic Penetration Tests 113 7.2.1 Description of Test Method 113 7.2.2 Test Results 114 7.3 Discussion of Energy Data 115 7.4 Correlation Factor 119 8. DISCUSSION 126 8.1 Standardization of N A L P T Results 126 8.2 Performance of Proposed Correlation Method 127 8.3 Grain Size Effects 129 8.4 Use of SPT-LPT Correlations 136 v 9. C O N C L U S I O N AND R E C O M M E N D A T I O N S F O R F U T U R E 141 R E S E A R C H Bibliography 144 APPENDIX A - STRESS W A V E F O R M U L A E 148 APPENDIX B - KIDD2 F I E L D P R O G R A M T E S T R E S U L T S 157 APPENDIX C - S E W A R D , A L A S K A F I E L D P R O G R A M T E S T 163 R E S U L T S APPENDIX D - K E E N L E Y S I D E D A M F I E L D P R O G R A M T E S T 168 R E S U L T S vi LIST OF T A B L E S 2.1 Results of SPT-ILPT Comparison (Crova et al., 1993). 17 2.2 Summary of SPT and LPT Details. 18 3.1 Soil Parameters Recommended for use with G R L W E A P 32 (GRLWEAP, 1997). 4.1 Hammer, Rod and Sampler Details Used for G R L W E A P Analyses. 41 4.2 Soil Parameters Used for G R L W E A P Parametric Study. 42 4.3 (N / Ru) Values Calculated Using G R L W E A P Parametric Study 43 Results. 4.4 Summary of Proposed Correlation Method Input Data and Results. 50 4.5 Summary of Observed and Predicted Correlation Factors. 50 6.1 Comparison of Uncorrected N A L P T Blow Counts. 98 6.2 Comparison of N A L P T Velocity and Rod Energy Ratios. 103 8.1 Summary of Observed and Standardized SPT-NALPT Correlation 127 Factors. 8.2 Revised Summary of Observed and Predicted Correlation Factors. 128 8.3 Summary of Available Grain Size Information. 131 v i i LIST O F FIGURES Figure .No. ^ ^ee 2.1 Range of Acceptable Dimensions for SPT Split-Spoon Sampler 4 (ASTM 1991a). 2.2 Donut Hammer Lifted Using Rope and Cathead Method (Robertson 6 etal., 1992). 2.3 Typical Longitudinal Section of a Safety Hammer. 7 2.4 Typical Details of a Trip Release Hammer (Clayton, 1990). 8 2.5 N A L P T Split-Spoon Sampler used by USACE. 9 2.6 Cohesionless Sand and Silt SPT-LPT Correlation Graph (Winterkorn 11 and Fang, 1975). 2.7 JLPT Split-Spoon Sampler {after: Kaito et al., 1971). 13 2.8 SPT-JLPT Correlation Data (Yoshida etal., 1988). 14 2.9 ILPT Split-Spoon Sampler and Hammer (Crova et al., 1993). 16 2.10 Comparison of SPT and LPT Test Details. 19 3.1 G R L W E A P Pile or Rod String Model (after: G R L W E A P , 1997). 29 3.2 Idealized Soil Response to Static and Dynamic Loading. 31 4.1 Forces Acting on (a) SPT Split-Spoon and (b) Piezocone During 34 Quasi-Static Penetration (Schmertmann, 1979). 4.2 Energy Expended During Displacement of (a) Ideal Plastic Soil and 37 (b) Ideal Elastic-Plastic Soil. 4.3 Comparison of SPT Blow Counts Predicted Using Ideal Plastic and 39 Ideal Elastic-Plastic Soil Models to G R L W E A P Analysis Results. viii I T _ ^ ^ 4.4 G R L W E A P Analysis Results for SPT, NALPT, JLPT and ILPT. 40 4.5 Sensitivity of Predicted (N / R u ) Values to Soil Parameter Variations. 44 4.6 Inverse Proportionality Relationship Between (N / R u ) and ENTHRU. 45 4.7 Effect of Rod Cross-Sectional Area on Predicted (N / R u ) Values. 47 4.8 Idealized Effect of Rod Cross-Sectional Area on Axial Force Data. 48 5.1 Kidd2 Piezocone Penetration Test Data. 54 5.2 Distribution of SPT, N A L P T and CPTU Test Holes at Kidd2. 55 5.3 Kidd2 Grain Size Distribution Data. 56 5.4 Sample SPT Quasi-Static Penetration Test Strip Chart Output. 59 5.5 Summary of SPT and N A L P T Quasi-Static Resistance Versus 60 Penetration. 5.6 Measured and Predicted Quasi-Static Penetration Resistance Force 61 Versus Depth. 5.7 Comparison of Measured and Predicted Quasi-Static Penetration 62 Resistance Force. 5.8 Sample and Idealized HPA Strip Chart Output. 65 5.9 Sample D E M Software Operating Screen. 67 5.10a Raw and Energy Corrected SPT Blow Counts Versus Depth. 70 5.1 Ob Raw and Energy Corrected N A L P T Blow Counts Versus Depth. 71 5.11 D E M Output Recorded During Hammer Blows Within the 152 mm 73 to 457 mm (6" to 18") Sampler Penetration Range at 9.5 m (31') Depth in SPT9904. ix Figure No. Title Page 5.12 D E M Output Recorded During Hammer Blows Within the 152 mm 74 to 457 mm (6" to 18") Sampler Penetration Range at 18.6 m (61') Depth in SPT9901. 5.13 Comparison of Average Force and Velocity Data Recorded in Two 75 SPT Test Holes at Differing Depths. 5.14 D E M Output Recorded During Hammer Blows Within the 152 mm 77 to 457 mm (6" to 18") Sampler Penetration Range at 9.5 m (31') Depth in LPT9903. 5.15 D E M Output Recorded During Hammer Blows Within the 152 mm 78 to 457 mm (6" to 18") Sampler Penetration Range at 17.1 m (56') Depth in LPT9902. 5.16 Comparison of Average Force and Velocity Data Recorded in Two 79 N A L P T Test Holes at Differing Depths. 5.17 Relationship Between Additional Potential Energy Due to Sampler 81 Set and Blow Count. 5.18 Comparison of D E M and "Corrected" H P A Energy Data. 83 5.19 Average SPT Force and Velocity Data Recorded at 18.6 m (61') in 84 SPT9904. 5.20 Average N A L P T Force and Velocity Data Recorded at 17.1 m (56') 86 in LPT9902. 5.21 Comparison of FV Energy Corrected SPT and N A L P T Blow Counts 89 Recorded at Kidd2. 6.1 Distribution of SPT, N A L P T and DCPT Test Holes at Seward, 91 Alaska Main Test Site. 6.2 Comparison of Percent Gravel in SPT and N A L P T Samples. 96 6.3 Comparison of Mean Grain Size (D 5 0) of SPT and N A L P T Samples. 97 Figure No. Title Page 6.4 Comparison of Uncorrected N A L P T Blow Counts from SEWA9803 99 and SEWA9806. 6.5 D E M Force and Velocity Data Collected During SPT at 18.1 m 101 (59.3') in SEWA9802. 6.6 D E M Force and Velocity Data Collected During N A L P T at 19.6 m 102 (64.3')inSEWA9803. 6.7 Average SPT Force and Velocity Data Recorded at 18.1 m (59.3') in 104 SEWA9802. 6.8 Average N A L P T Force and Velocity Data Recorded at 19.6 m 105 (64.3') in SEWA9803. 6.9 Comparison of FF Energy Corrected SPT and N A L P T Blow Counts 108 Recorded at Seward, Alaska Main Test Site. 6.10 Comparison of FV Energy Corrected SPT and N A L P T Blow Counts 109 Recorded at Seward, Alaska Main Test Site. 7.1 Plan View of Keenleyside Dam (Lum and Yan, 1994). 111 7.2 Grain Size Envelope for Keenleyside Dam Sand and Gravel Fill 112 Material (Lum and Yan, 1994). 7.3 D E M Force and Velocity Data Collected During N A L P T at 18.3 m 116 (60') in DH99-20. 7.4 Average N A L P T Force and Velocity Data Recorded at 18.3 m (60') 117 in DH99-20. 7.5 Comparison of D E M and "Corrected" HPA Energy Data. 120 7.6 Comparison of FV Energy Corrected SPT and FF (AW) Energy 121 Corrected N A L P T Blow Counts Recorded at Keenleyside Dam. 7.7 Comparison of FV Energy Corrected SPT and FF (NW) Energy 122 Corrected N A L P T Blow Counts Recorded at Keenleyside Dam. xi Figure No. Title Page 7.8 Comparison of Equivalent SPT Blow Counts from BPT Data and FF 124 (AW) Energy Corrected N A L P T Blow Counts Recorded at Keenleyside Dam. 7.9 Comparison of Equivalent SPT Blow Counts from BPT Data and FF 125 (NW) Energy Corrected N A L P T Blow Counts Recorded at Keenleyside Dam. 8.1 Comparison of SPT-ILPT Correlation Factors and Mean Grain Size 133 Data from Messina, Italy (Crova et al., 1993). 8.2 Comparison of SPT-NALPT Correlation Factors and Mean Grain 134 Size Data from Seward, Alaska Research Program. 8.3 Comparison of SPT-NALPT Correlation Factors to "Oversized" 135 Portion of N A L P T Samples from Seward, Alaska Research Program. 8.4 Keenleyside Dam SPT-NALPT Correlation Data (FV Energy 137 Corrected), Sand Versus Gravel Data. 8.5 Dynamic SPT Blows versus Penetration Data from Kidd2 and 138 Seward, Alaska Sites. 8.6 Dynamic N A L P T Blows versus Penetration Data from Kidd2 and 139 Keenleyside Dam Sites. x i i LIST O F S Y M B O L S AND ABBREVIATIONS A area A S T M American Society for Testing and Materials A E split-spoon sampler end bearing area A F split-spoon frictional area at 12" penetration A T E equivalent tip bearing area BC Hydro British Columbia Hydro and Power Authority bpf blows per foot BPT Becker Penetration Test BSC British Soil Classification system c stress wave propagation velocity CPT Cone Penetration Test CPTU Piezocone Penetration Test Cdp pile damping value c , SPT-CPT end bearing correlation factor c 2 SPT-CPT friction correlation factor d split-spoon sampler displacement D dynamic component of total soil resistance DCPT Dynamic Cone Penetration Test D E M Dynamic Energy Monitoring system D 5 0 mean grain size E Young's modulus ENTHRU energy transmitted through drill rods ER energy ratio E R A energy ratio used for G R L W E A P or equivalent analysis ER r energy ratio calculated from the stress wave energy ER V energy ratio calculated from the hammer kinetic energy F force F(t) force which varies with time FF Force Squared stress wave energy measurement method FV Force Velocity stress wave.energy measurement method fs measured CPTU friction sleeve stress g gravitational acceleration G R L W E A P Goble, Rausche and Likins Wave Equation Analysis Pro H height HPA Hammer Performance Analyzer ID inner diameter ILPT Italian Large Penetration Test j Smith damping factor JLPT Japanese Large Penetration Test jsi Smith damping factor at segment (i) K c velocity correction factor xiii ksi soil stiffness at segment (i) K , load cell position correction factor K 2 rod length correction factor L length of drill rod between stress wave measurement point and soil-sampler interface LPT Large Penetration Test L H hammer length N uncorrected blow count N A L P T North American Large Penetration Test N 6 0 blow count corrected to 60% standard energy (Nl)60 blow count corrected to 60% standard energy and 100 kPa overburden pressure OD outer diameter PE maximum potential energy of SPT or LPT hammer PDI Pile Dynamics Incorporated q soil quake q c measured CPTUtip stress qt CPTU tip stress corrected for pore pressure effects q-s quasi-static R total (static + dynamic) soil resistance Rf CPT friction ratio R s Sampler-Hammer Ratio R u ultimate static resistance Rui ultimate static resistance at segment (i) S static component of total soil resitance SPT Standard Penetration Test t time T time required for a stress wave to pass a point on a drill rod U S A C E United States Army Corp of Engineers u s e Unified Soil Classification system u 2 CPTU pore pressure measurement directly behind the cone tip V velocity V(t) velocity which varies with time w weight w buoyant weight X position along a bar or drill rod Z rod impedance At incremental time step Ad change in split-spoon sampler displacement ild hammer dynamic efficiency P mass density xiv A C K N O W L E D G E M E N T S The author gratefully acknowledges the financial support of the Natural Science and Engineering Research Council (NSERC) of Canada, which was provided as a Post-Graduate Schedule A Scholarship. In addition, the research could not have been completed without the generous support of the following organizations: • Foundex Explorations Ltd. of Surrey, B.C. donated drilling services and expertise during the Kidd2 investigation; • The British Columbia Hydro and Power Authority (BC Hydro) donated drilling time and field support during the Keenleyside Dam investigation, provided a D E M system during the Seward and Keenleyside investigations and provided access to the Kidd2 site; • The United States Army Corps of Engineers (USACE) organized and provided drilling services and field support during the Seward investigation; • Klohn-Crippen Consultants Ltd. provided the HPA system used during the three field investigations; and, • Conetec Investigations Ltd., provided funding towards my involvement in the Seward investigation. In addition, I would like to thank my advisor, Dr. John Howie, for introducing me to the topic and for his guidance and financial support during the course of my research, my co-advisor Dr. R.G. Campanella, for introducing me to geotechnical research and for his financial support, Dr. Alex Sy (Klohn-Crippen) for his insights on gravel testing and dynamic energy theory, Dr. Liam Finn (UBC) for his financial support and guidance during the Seward investigation, Dr. Joe Koester (USACE) for his efforts during the Seward investigation and Ken Lum (BC Hydro) for arranging the Keenleyside work. Scott Jackson and Harald Schremp provided first class technical support, as always. A l i Amini, Patrick Koerner, Kevin Payne, Rashmi Pishe and Brian Walker provided much needed assistance during the Kidd2 research program. Thanks to Kim, who is consistently the better half and to parents Trevor and Judi. Thanks also to Jay, Tom, Sue, Andrea, Ruben and Brionie, who have been good friends throughout the process. xv M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels 1. INTRODUCTION Section 1.0 Introduction Split spoon samplers are robust geotechnical tools than can be used at relatively little expense to characterize soil stratigraphy through direct sampling of the strata. The samplers are generally driven into the soil at the base of a clean, supported borehole by striking the top of the rods used to lower the sampler with specially designed hammers. Counting the number of hammer blows required to insert the sampler is a natural extension of the characterization process that can be useful when compared to similar data from other sites. For this reason, engineers most commonly perform "Standard Penetration Tests" (SPT) using a standard 5.08 cm (2") outer diameter split spoon and a 63.5 kg (140 lb) hammer with a drop height of 76 cm (30"). Through the use of standardized equipment, the energy available for penetration of the sampler as well as the surface area upon which soil resistance acts is kept constant between tests and variations of the number of blows required for sampler insertion, the "blow count" (N), should be a measure of soil resistance. Many empirical correlations between soil design parameters and SPT blow counts have been published. In fact, data from tests that are generally considered superior to the SPT are often correlated to SPT blow counts in order to utilize these empirical correlations. Thus, it is often assumed that SPT blow counts can be predicted from the results of other in-situ tests. The SPT is considered unreliable for gravel deposits, primarily because gravel particles can be larger than the opening of the SPT sampler (3.5 cm, 1.375") and secondarily because the drilling methods normally employed for SPT in sands are often impractical in gravels. Perhaps as a result, direct empirical correlations between gravel design parameters and SPT blow counts are seldom encountered. Tools such as the dynamic cone penetration test (DCPT), Becker Penetration Test (BPT), seismic methods and, the topic of this thesis, the Large Penetration Test (LPT) have been used for characterization of gravels because they avoid one or both of the above issues. M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 1.0 Introduction LPT is a generic term that has been used by several authors to describe combinations of oversized split-spoon samplers and hammers for use in gravels. Unfortunately, the LPT does not directly address the difficulties of drilling in gravels but the similarity of the SPT and LPT suggests that nothing more than a scaling factor is required to correlate the two types of blow counts, which is a major advantage over DCPT, BPT and seismic tests. The purpose of this thesis is to present and discuss data collected with one LPT system that is generally available in North America, though not commonly used. The data were collected at both sand and gravel sites and SPT blow counts from adjacent boreholes are presented in all cases. In the course of this research, the author developed a preliminary method of predicting SPT blow counts from the blow counts obtained with any combination of hammer and split spoon sampler. The method is presented and calibrated using the few SPT-LPT correlations that have been published to date and the correlations developed herein. Several authors have developed empirical and semi-empirical correlations between the output of a test suitable for gravel and SPT blow counts to allow indirect use of SPT empirical design methods. Such indirect use of SPT empirical design methods requires the additional assumption that design parameters predicted from "equivalent" SPT blow counts will accurately reflect the performance of the gravel deposits. The research undertaken for this thesis was not designed to investigate the validity of this assumption but the author has attempted to glean some preliminary insight into related issues such as "grain-size effects" and this is presented in the Discussion. M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations 2. SPT AND LPT DETAILS AND CORRELATIONS The following sections describe the SPT and three LPT systems with which the author has experience or that have been described in technical journals. The three LPT systems are identified by area of origin. Existing correlations between the LPT's described and the SPT are presented where available. 2.1 Standard Penetration Test (SPT) Split-spoon samplers are hollow cylinders that are split lengthwise to facilitate sample logging and extraction. Figure 2.1 illustrates the range of acceptable split-spoon dimensions for SPT. The field engineer notes the condition of the cutting shoe and whether or not a sample barrel liner and sample catcher were included, as these details may affect the penetration resistance. The sampler is driven into the soil at the base of a clean, supported borehole by striking an anvil attached to the top of the drill rods with a 63.5 kg (140 lb) hammer dropped 0.76 m (30") yielding a maximum possible energy of 473 J (350 ft-lb). The number of blows required for each 152 mm (6") of penetration are recorded and the total blows over the interval 152 to 457 mm (6 to 18") are summed to give the blow count (N) in blows per foot (bpf). Key details of the SPT sampler and hammer are described in A S T M Standard D1586-84 (ASTM 1991a). The A S T M standard is much more specific about the required design of the split spoon than the hammer, stating only that the hammer must have a mass of 63.5 kg (4.35 slug) and must drop vertically 0.76 m (30") before striking the anvil. As a result, hammer designs vary considerably. The four most commonly used types are donut, safety, trip release and automatic hammers. M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations OPEN SHOE HEAD ROLLPIN (2 at y8 in. diameter) A - 1.0 to 2 0 in. (25 to 50 mm) B - 18.0 to 30.0 in. (0.457 to 0.762 m) C - 1.375 ± 0.005 in. (34.93 ±0.13 mm) D - 1.50 ± 0.05 - 0.00 in. (38.1 ± 1.3 - 0.0 mm) E - 0.10 ± 0.02 in. (2.54 ± 0.25 mm) F - 2.00 ± 0.05 - 0.00 in. (50.8 ± 1 . 3 - 0 . 0 mm) G - 16.0* to 23.0* The 1 Vi in. (38 mm) inside diameter split barrel may be used with a 1 &-gage wall thickness split Bner. The penetrating end ol the drive shoe may be slightly rounded. Metal or plastic retainers may be used to retain soil samples. Figure 2.1 Range of Acceptable Dimensions for SPT Split-Spoon Sampler (ASTM 1991a). M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations The donut hammer is perhaps the simplest SPT hammer, consisting of a simple cylindrical 63.5 kg mass falling down a guide rod. The hammer mass is attached to a rope which runs through a pulley situated above the hammer. The hammer can be lifted by manually pulling on the free end of the rope or by wrapping the rope around a rotating cathead and applying tension, as shown in Figure 2.2. In the latter case, the operator drops the hammer after visually checking the drop height by releasing the tension on the rope. It is also possible to use a winch with a clutch release to lift and release the hammer. The safety hammer was developed to protect rig operators from injury by internalizing the point of impact between the falling mass and the anvil rod (Figure 2.3). The same lift and release methods used for the donut hammer may also be used with the safety hammer. Trip release hammers were developed to improve the repeatability of SPT hammer drops by allowing the operator to mechanically set the drop height. The efficiency of the hammer is also improved by eliminating the friction losses inherent in the rope and cathead method. Details of a typical SPT trip release hammer are shown in Figure 2.4. The drop weight of an automatic hammer is lifted by a chain drive mechanism that is hydraulically powered by the rig itself. The primary advantages of automatic hammers are the speed with which the tests can be completed and the minimal physical effort required by the rig operator. 2.2 North American L P T (NALPT) Split-spoon samplers with outer diameters increasing by increments of 12.7 mm (0.5") above 50.8 mm (2") are widely available. These larger samplers are most commonly used for environmental investigations to maximize sample volume. The sampler shown in Figure 2.5 has been used by the United States Army Corps of Engineers (USACE) for M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations Figure 2.2 Donut Hammer Lifted Using Rope and Cathead Method (Robertson et al., 1992). M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations Hammer Cap Block and Hammer Cylinder Anvil Rod 5.63" OD = 5.5" Four orthogonal rod gu ides, 0.44" thick 0.63" 3.87" OD = 2.61" 39.71" OD = 5.5" ID = 4.82" 2.92" OD = 5.5" ID = 2.75" ' 2.93" OD = 5.5" ' ID = 2.75" J KS3 52.82" OD = 2.61" ID = 2.25" i Scale 1:10 NW Pin 4.68" OD = 2.61" • ID = 1.37" 2.82" . OD = 2.2" ID = 1.37" Figure 2.3 Typical Longitudinal Section of a Safety Hammer. M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations - IHH-g MMmbly « n-0 kg > • trip mechanism -140 lb might < 83 5 kg ) • anvil <wWi ahaft: 32-2 kg ) Figure 2.4 Typical Details of a Trip Release Hammer (Clayton, 1990). M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations NWJ Box Thread 13 mm 76 mm 34 mm Drainage Port Ball Valve 47 mm 40 mm Sample Barrel ID = 64 mm (Without Liner) ID = 61 mm (With Liner) OD = 76 mm Opening for Sample Catcher 47 mm 39 mm ID = 61 mm OD = 76 mm Not to Scale Figure 2.5 N A L P T Split-Spoon Sampler used by U S A C E . M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations both environmental and geotechnical characterization of gravel deposits in Alaska. The author participated in an U S A C E correlation research program and subsequently conducted two similar research programs at different sites. Though test details varied somewhat between these three research programs, the "typical" North American LPT (NALPT) hammer weighs 1335 N (300 lb) and is dropped 0.76 m (30") yielding a maximum possible energy of 1015 J (750 ft-lb), which is 2.14 times the maximum possible SPT energy. The sampler has an outer diameter of 76.2 mm (3") and an inner diameter of 61 mm (2.4") including a 1.3 mm (0.05") thick liner. The number of blows for each 152 mm (6") of penetration are recorded and the blows over the interval 152 mm (6") to 457 mm (18") are summed for the blow count (N). The U S A C E uses an empirical correlation proposed by Winterkorn and Fang (1975) to convert blow counts measured with the North American LPT to equivalent SPT blow counts. The correlation is based on the "Sampler-Hammer Ratio" (Rs): R S = 0 D ' - ' D ' (2.,) where: ID is the inner diameter of the open shoe OD and ID are given in inches W = weight of hammer (lb.) H = height of hammer drop (in.) Thus (Rs) is directly proportional to the sampler dimensions (which determine sampler penetration resistance) and inversely proportional to the hammer potential energy. To determine the SPT-LPT correlation factor, the (Rs) value is plotted on the cohesionless sand and silt correlation graph shown in Figure 2.6 and compared to the (Rs) of the SPT M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 2 0 Split Spoon Penetration Testing in Gravels SPT and LPT Details and Correlations 10 r4 5 H or o E £ o X I CL E -o 5 4 Very Compact Burmister Std. Dense Terzaghi S td . f t Energy Standard W H Do Dj Rs Burmister 250 20 3.625 2.930 3.I2X I 0 - 5 Terzaghi 140 30 2.000 1.375 0.895XI0" 5 Sampler Hammer Ro1io,R s,— -W - Weight of Hammer, pounds H = Height of Drop, inches D 0 S Outside Dia. Sampler, inches Dj= Inside Dia. Sampler, inches Dr= Relative Density, % 10 ~ic> so ioo" Driving Resistance, B, Blows/ft 500 I000 Figure 2.6 Cohesionless Sand and Silt SPT-LPT Correlation Graph (Winterkorn and Fang, 1975). -11 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations (denoted "Terzhagi Std." in the figure). The relationship shown in Figure 2.6 can be approximated by the following equation: (AO (AO The (R s) values of the SPT and the typical N A L P T are 0.895 • 10"5 ft2/lb and 1.017 • 10"5 ft2/lb, respectively, and the predicted SPT-NALPT correlation factor is roughly 0.93. 2.3 Japanese L P T (JLPT) Kaito et al. (1971) first proposed the use of the hammer and split-spoon sampler shown in Figure 2.7 for geotechnical characterization of gravel deposits. The Japanese LPT (JLPT) hammer weighs 981 N (220 lb) and is dropped 1.5 m (59.1") yielding a maximum possible energy of 1472 J (1084 ft-lb), which is 3.11 times the maximum possible SPT energy. The sampler has an outer diameter of 73 mm (2.9") and an inner diameter of 50 mm (2") including a 2 mm thick liner. The number of blows for each 152 mm (6") of penetration are recorded and the blows recorded for the interval 152 mm (6") to 457 mm (18") are summed for the blow count (N). Yoshida et al. (1988) compared SPT and JLPT blow counts obtained in a calibration chamber filled with sands and gravels of varying density. No attempt was made to measure the efficiency of the hammers (hammer efficiency will be discussed in Section 3.0). Based on the data shown in Figure 2.8, they proposed the following two correlations: Nsn=2-NLPT (gravel) (2.3) NSPT =1.5- NLPT (sand) (2.4) S / SPT S JLPT u.oe / (2.2) - 1 2 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations -Box Thread Drainage Port Sample Barrel ID = 54 (Without Liner) ID = 50 (With Liner) OD = 70 30 mm 70 mm 73 Not to Scale Figure 2.7 J L P T Split-Spoon Sampler (after: Kaito et al., 1971). M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations Figure 2.8 SPT-JLPT Correlation Data (Yoshida et al., 1988). -14-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations It should be noted that the "gravels" used during the tests are classified as medium to coarse-grained sands using the Unified Soils Classification (USC) system. Following Yoshida et al.'s work, a number of technical papers were published regarding the use of the Japanese LPT (e.g. Tanaka et a l , 1991, Suzuki et al., 1993 and Hatanaka and Uchida, 1996). The approach of these papers has been to develop new correlations between the JLPT blow count and engineering parameters such as cyclic strength, rather than to generate equivalent SPT blow counts from JLPT data. The (R s) value of the Japanese LPT is 0.875 • 10"5 ft2/lb and the corresponding SPT-JLPT correlation factor is 1.02, which is in poor agreement with Yoshida et al.'s empirical results. 2.4 Italian LPT (ILPT) Crova et al. (1993) describe the use of the split-spoon sampler and hammer shown in Figure 2.9 for geotechnical characterization of sand and sandy-gravel deposits. The trip-release hammer weighs 5592 N (1256 lb) and is dropped 0.5 m (19.7") providing a maximum possible energy of 2796 J (2062 ft-lb), which is 5.91 times the maximum possible SPT energy. The sampler has an outer diameter of 140 mm (5.5") and an inner diameter of 100 mm (3.9"), including a 5 mm (0.2") thick liner. The sum of blows required to drive the sampler from 152 mm (6") to 457 mm (18") penetration is the blow count (N). SPT were also performed in both the sand and sandy-gravel deposits. Average rod energy values of 60% and 85% of maximum possible energy were measured for the SPT and ILPT, respectively. Table 2.1 summarizes the results of their investigation. -15-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations - 1 6 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations Table 2.1 Resu ts of SPT-ILPT Comparison (aft, er. Crovaet al. 1993) Deposit Number of tests N SPT . ^ 1 ( 6 0 ) SPT D 5 0 (mm) N ILPT .^1(60). ILPT 1.41 1.14 Po River Sand 35 ± ± 0.2 to 0.6 0.46 0.40 Holocene sand and gravel 97 1.13 ± 0.52 0.89 ± 0.40 1 to 15 Pleistocene 1.38 1.02 sand and 62 ± ± 1 to 5 gravel 0.45 0.36 Crova et al. conclude that the correlation between the SPT and ILPT is close to one i f both blow counts are corrected to 60% of the maximum hammer potential energy and corrected for overburden stress. The (R s) value of the Italian LPT is 2.978 • 10"5 ft2/lb and the predicted SPT-ILPT correlation factor is 0.44, which is on poor agreement with Crova et al.'s uncorrected empirical results in column 3 of Table 2.1. Table 2.2 and Figure 2.10 summarize the details of the LPT systems described in this section. As noted by Crova et al., the efficiency of the hammer can have a significant effect on the measured blow counts and resulting correlation factors. Details of energy transfer during dynamic penetration tests such as the SPT and LPT are discussed in Section 3.0. M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 2.0 SPT and LPT Details and Correlations Table 2.2 Summary of SP and LPT Details Identification SPT N A L P T JLPT ILPT Outer Diameter mm (in.) 50.8 (2) 76.2 (3) 73 (2.9) 140 (5.5) Inner Diameter With Liner mm (in.) 34.9 (1.375) 61.0 (2.4) 50 (2) 100 (3.9) No Liner mm (in.) 38.1 (1.5) 63.5 (2.5) 54 (2.13) 110 (4.3) Hammer Weight N (lb.) 623 (140) 1335 (300) 981 (220) 5592 (1256) Drop Height mm (in.) 762 (30) 762 (30) 1500 (59.1) 500 (19.7) Maximum Potential Energy J (ft-lb) 473 (350) 1015 (750) 1472 (1084) 2796 (2062) % o f SPT 100 214 311 591 R s ft2/lb 0.895 • 10"5 1.017 • 10"5 0.875 • 10"5 2.978 • 10"5 -18-P E in « C OJ II *1 W > re o re c _ • § » O C C E — - 5 re o « Q O |2 * | % .2 « *- 5 ™ J= CQ i - O c •2 >. « to £ 1 * jo c H SJ § o c co = S2 (p) A6jau3 |B!)ua}0d LunLUiXB|/\| o o o o o o o o o o —Lu. (q|-y) A6jau3 |B!Jua}0c| LunujixBiAj o o o o o o o o o o . J u t >» \ ra \ o3 y> LU \ ro \ enti L_ OJ o r> D_ w me ro um b LUI 0) X \ "3 Ma cr O jajsjuung I d l l ±d"ir IdlVN IdS —T— 00 CD CO ( U | ) ja}9LUB!Q O uo CM O O CM O LO O O o LO o (LULU) jajauiBja M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 3.0 Dynamic Penetration Testing Energy Theory 3. D Y N A M I C P E N E T R A T I O N T E S T I N G E N E R G Y T H E O R Y In order to allow meaningful comparison of SPT results acquired with different hammers and operators it is necessary to consistently deliver the same amount of energy from the hammer to the sampler via the drill rods. For this reason, the mass and drop height of the SPT hammer were standardized to ensure that the hammer would have a consistent potential energy before each drop. As the popularity of the SPT for geotechnical site characterization increased, the number of hammer and drill rod systems in use also increased. In the 1970's, studies were published indicating that different hammer and rod systems were not consistently delivering the same energy, even though the potential energy before each drop was, ideally, the same for all hammers. It was recognized that different hammers would have different efficiencies, that is, they would convert different amounts of the initial potential energy to kinetic energy when dropped. It is now believed that, in addition to fall efficiency variations between hammers, details of the hammer, anvil rod and drill rod geometry may affect the amount of the kinetic energy that is transferred to the drill rods. Schmertmann and Palacios (1979) showed experimentally that the measured blow count was inversely proportional to the energy delivered to the drill rods for blow counts less than 50. Seed et al. (1985) and Skempton (1986) suggested that measured blow counts be corrected to the value that would be recorded if a standard amount of energy had been transmitted through the rods. A standard value of 60% of the potential energy of the hammer (60% of 473.4 J = 284 J) was adopted because it was the average value measured at the time for various rigs. The actual energy measured during the test is converted to an energy ratio using the formula: -20-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 3.0 Split Spoon Penetration Testing in Gravels Dynamic Penetration Testing Energy Theory ER = Measured Energy •100% (3.1) 473.4 J and the energy corrected blow count (N 6 0 ) is calculated as follows: N • ER 60% (3.2) Following this initiative, researchers and practising engineers began to "calibrate" hammer and rod systems by measuring the hammer fall velocity or by measuring the energy contained in the stress wave travelling down the drill rods. Typical E N T H R U values have been proposed for most hammer types so that the additional cost of energy monitoring may be avoided. This generalized approach to dealing with energy variations is questionable because it does not consider details of the hammer and rod system. During the course of this research, kinetic and stress wave energy were recorded whenever possible. The theories and practices of measuring kinetic and stress-wave energy are described below. 3.1 Kinetic Energy of Driving Systems The kinetic energy of the hammer reaches a maximum at the instant before the hammer strikes the anvil rod. The magnitude of the kinetic energy can be calculated by entering the peak velocity of the hammer into the general formula for kinetic energy: Kovacs and Salomone (1982) used photovoltaic reflective scanners to record hammer drop height and peak velocity for a large number of donut and safety hammer impacts. They recorded losses of 21 to 25% of the total potential energy during free fall for donut hammers and 26% to 31% for safety hammers. Kinetic Energy = — • Hammer Mass • (Peak Velocity)2 (3.3) -21 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 3.0 Dynamic Penetration Testing Energy Theory Radar systems are more commonly used to measure fall velocity. One such system is the Hammer Performance Analyzer (HPA) developed by Pile Dynamics Inc. (PDI). The HPA consists of an antenna that emits a directable cone of radio waves and detects returning waves reflected off solid objects within the cone. If the object is in motion, there will be a phase shift between the original and reflected waves (the Doppler effect). The data acquisition system calculates the velocity of the fastest moving object from the largest phase shift recorded. This maximum velocity is recorded as a function of time on a strip chart. The energy ratio calculated from the hammer kinetic energy is denoted (ERV). 3.2 Stress Wave Theory Measuring stress wave energy is considered superior to measuring hammer kinetic energy because energy losses occur during hammer impact and stress wave transmission, after the hammer kinetic energy has been measured. This section describes basic theory related to the formation and propagation of stress waves in linear rod systems, two methods for calculating the energy contained in the stress wave and a stress wave modelling program that was used by the author during the course of this research. 3.2.1 Characteristics of Stress Waves Timoshenko and Young (1955) show that the differential equation of motion for a cross sectional element within a prismatic bar during longitudinal vibration is: ^W-^ (3-4) dt2 dx2 where: u = longitudinal displacement of a cross section of the bar x = position along the bar -22-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 3.0 Split Spoon Penetration Testing in Gravels Dynamic Penetration Testing Energy Theory t = time c = stress wave propagation velocity where f() and g() are arbitrary functions representing stress waves that propagate with equal but opposite velocities (c and -c) within the bar. When a stress wave passes a point on the bar, the result is a change in stress and particle velocity. The particle velocity should not be confused with the stress wave propagation velocity (c) defined above. The following sign conventions are generally used to describe the force and particle velocity: • Compressive forces are positive; • Particle velocities are positive when the resulting particle displacement is in the direction of increasing (x). During SPT and LPT, this direction is generally assumed to be along the rod axis towards the sampler (down). Timoshenko and Goodier (1970) showed that the axial force, F(t), within a single stress wave propagating in the direction of increasing (x) may be related to the particle velocity, V(t), as follows: * 5,120 m/s (16,800 ft/s) for steel E = Young's Modulus p = mass density The general solution of Equation 3.4 is: u (x,t) = f(x + ct)-g(x-ct) (3.5) -23-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 3.0 Dynamic Penetration Testing Energy Theory F{t) = --V(t) c (3.6) = Z • V(t) where (A) is the cross-sectional area of the bar and (Z) is called the "impedance" of the bar. This property is called "force velocity proportionality" and, due to the adopted sign convention, must be slightly modified for waves travelling in the direction of decreasing Palacios (1977) provides a thorough description of the application of Equations 3.4 to 3.7 to stress wave propagation within drill rods. Some of Palacios' results are demonstrated in Appendix A . These results will be used in Section 5.0, 6.0 and 7.0 to assess the quality of stress wave data recorded during the course of this research. 3.2.2 Application to SPT and LPT Energy Measurement Measurements of F(t) and V(t) at a point in the drill rods following hammer impact may be used to calculate the magnitude of the energy transferred from the hammer to the rods. The energy ratio calculated using the stress wave energy is called the rod energy ratio (ER r). For a body in motion, the increment of work done over a time interval centred at time (ti) is given by: (x): c (3.7) = -Z-V(t) dW = F{tx)-dx = F(tx)-V(tx).dt (3.8) -24-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 3.0 Split Spoon Penetration Testing in Gravels Dynamic Penetration Testing Energy Theory By integrating the increment of work over the time it takes the stress wave to pass the measurement point (T), the total work performed by the stress wave, which is equal to the transmitted energy (ENTHRU), may be calculated: If more than one stress wave contributes to the total force and velocity at the measurement point (e.g. if downward and upward propagating waves are present due to a reflection below the measurement point), the engineer must subtract the energy of the upward propagating wave from the energy of the downward propagating wave to determine the energy absorbed by the soil during sampler penetration. The principal difference between the two commonly used methods of SPT energy measurement, Force Squared (FF) and Force Velocity (FV), is that the former cannot differentiate between stress wave energy travelling down and up the rods. 3.2.2.1 Force Squared (FF) Method The Force Squared (FF) method of energy calculation is the current industry standard and is described in A S T M (1991b). The method is based on the assumption that there are no upward propagating stress waves at the measurement location until the arrival of the reflection from the sampler/soil interface (at time 2L/c, where L is the length of rods between the measurement location and soil-sampler interface). In this case, the energy of the stress wave passing the measurement location is equally divided between strain and kinetic energy and measurement of either axial force or velocity is sufficient to determine the transmitted energy (Palacios, 1977). It is easier to measure axial strain than particle velocity so the velocity term in Equation 3.9 is eliminated using Equation 3.6, resulting in the following equation for FF energy: T (3.9) -25-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 3.0 Split Spoon Penetration Testing in Gravels Dynamic Penetration Testing Energy Theory FF ENTHRU = A ' ' ^ 2 ' A c ]T F(tf • At (3.10) where: Ki = load cell position correction factor (tabulated in A S T M , 1991b) K2 = rod length correction factor (tabulated in A S T M , 1991b) K c = velocity correction factor (described in A S T M , 1991b) At = time step between data points The summation is carried out from the time that the downward propagating stress wave first passes the measurement point until (2L/c). A high frequency data sampler is used so that F(t) may be considered constant over the time step (At). The measured (FF) energy is extrapolated to the value that would be measured if the rod length were infinite by the correction factors (Ki) and (K2). Clayton (1990) combines (Ki) and (K2) into a single correction factor (K), assuming exponential decay of the stress wave with time. The (Kc) correction factor is based on empirical evidence that the theoretical wave propagation speed (c) is higher than the actual speed so that the (FF) energy summation is halted before the actual arrival of the reflection from the sampler. Modern instrumentation typically consists of four electrical resistance strain gauges bonded to or bolted on a rod of the type used in the rest of the drill string (called a transducer rod). The transducer rod is usually placed in the string of drill rods directly below the anvil rod. Piezo-resistive and piezo-electric load cells placed in series in the drill string may also be used to record force data but it is generally considered desirable to limit the number of impedance interfaces around the measurement location to minimize reflections (see Appendix A for a description of the effect of impedance interfaces on stress waves). - 2 6 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 3.0 Dynamic Penetration Testing Energy Theory 3.2.2.2 Force Velocity (FV) Method As noted in Appendix A , rod strings in the field typically contain rod couplings at 1.5 or 3.0 metre (5 or 10 foot) spacing as well as other impedance interfaces at which the initial downward propagating stress wave will be partially reflected. Depending on the nature of the impedance mismatch, the force velocity proportionality assumption of the (FF) method may be seriously violated. To avoid problems arising from such partial stress wave reflections, Sy and Campanella (1991a) proposed the use of the (FV) method for SPT energy measurement (the method was already widely used for pile-driving applications). The (FV) method allows for the presence of two time-varying stress waves propagating in opposite directions through the drill rods. Using arrows to represent the direction of wave propagation, the total force at the measurement point is (F(t)^ + F(f)t), which is not proportional to the total velocity of (V(t)^ + V(t)T). The product of the total force and total velocity yields: (F(t) I +F(t) t)- (V(t) I +V(t) T)= F(t) I -V(t) I +F(t) T -V(t) t + F(t)i -V(t)t + F(t)t -V(t) I The underlined terms on the right side of Equation 3.11 cancel and, because F(t)tV(t)t must be negative, the total resulting energy measured is that of the downward minus the upward propagating waves. The formula used in practice to calculate F V energy is: FVENTHRU = ^  F(t) • V(t) • At (3.12) Where, again, (At) is small enough that F(t) and V(t) may be considered constant. The FV energy is calculated over the entire time trace and, in practice, the maximum calculated value is used in Equation 3.1 to calculate (Neo)- The FV energy calculated at (2L/c) is equal to the energy of the downward propagating stress wave minus the portion -27-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 3.0 Split Spoon Penetration Testing in Gravels Dynamic Penetration Testing Energy Theory of that energy that has been reflected at impedance interfaces below the measurement point. When the rod length is very low there may still be significant amounts of energy in the hammer and anvil rod when the first reflection from the soil-sampler interface arrives at the measurement point. It is not necessary to derive correction factors like (Ki) and (K2) because the FV energy continues to be that of the downward propagating wave minus the upward, regardless of the source of the upward propagating wave. The same instrumentation used to measure F(t) for the FF method may be used for the FV method. V(t) is usually obtained by integrating data from accelerometers that have been bonded to or bolted on the transducer rod. Accelerometers are much more expensive than the strain gauges used for force measurement so most F V systems include some sort of damping material between the accelerometer and rod to protect them from damage. It is generally believed that accelerometers should be reliable to 5000 g (g = gravitational acceleration) and capable of measuring signal frequencies as high as 1.5 kHz, even with the damping material. Recent studies at UBC, including this research, suggest that significantly higher amplitude and frequency capacities may be required to accurately measure V(t). 3.2.3 Stress Wave Modelling Stress wave modelling for pile-driving applications was first used to help quantify hammer efficiency and to estimate the static load capacity of driven piles. The popular Goble, Rausche, Likins and Associates wave equation analysis program (GRLWEAP) was used during the course of this research to predict the effect of increasing the energy delivered to the soil from SPT to LPT magnitudes. G R L W E A P models piles or drill rods subjected to dynamic loading as series of discrete elements connected by spring and dashpot pairs, as shown in Figure 3.1. The mass of each element and stiffness of each spring are determined from the density, elastic - 2 8 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 3.0 Dynamic Penetration Testing Energy Theory mm® AL "Pi AL m p i = A(AL)p Toe Resistance Shaft Resistance NOTE: ksi = soil stiffness at segment i (function of quake and Ru i); R u i = ultimate static resistance at segment i; j s i = Smith damping factor at segment i; and, c d p = pile damping value. Figure 3.1 GRLWEAP Pile or Rod String Model (after: GRLWEAP, 1997). - 2 9 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 3.0 Split Spoon Penetration Testing in Gravels Dynamic Penetration Testing Energy Theory modulus and cross-sectional area of the pile or drill rod material. Minimal damping is expected in steel drill rods so a low, empirical value from the G R L W E A P manual (GRLWEAP, 1997) is generally used. The same approach is used to model the hammer. The total resistance to pile penetration (R) is divided into static (S) and dynamic (D) components. It is assumed that the static component is elastic-plastic in nature and is present during both driving and subsequent static loading. Input parameters are the ultimate static resistance (R u) and the quake (q), as shown in Figure 3.2. The (R u) and (q) values are simply another way of stating elastic-plastic material properties. The dynamic component (D) of the total soil resistance is usually modelled as Smith damping: where (j) is the Smith damping factor. Figure 3.2 also shows the effects of Smith damping on the static elastic-plastic load-displacement curve. Conceptually, at the beginning of program execution all pile and hammer elements are at force equilibrium, all pile elements are at rest and all hammer elements are moving with an assigned velocity. The program calculates the spring and dashpot compression or extension that occur at existing element velocities over a very small time step. The resulting net forces that act on each element are used to calculate element-specific accelerations using Newton's Second Law: D = j-V.Rt v (3.13) Acceleration = Force Mass (3.14) -30-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 3.0 Dynamic Penetration Testing Energy Theory o Displacement Figure 3.2 Idealized Soil Response to Static and Dynamic Loading. -31 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 3.0 Dynamic Penetration Testing Energy Theory The new accelerations are applied over the next time step, new element velocities and displacements are calculated and the analysis is repeated. Iterative modelling of SPT or pile F(t) and V(t) data allows estimation of (Ru) for pile design purposes. Values of (q) and (j) recommended by G R L W E A P (1997) for pile-driving analyses are listed in Table 3.1. "Skin" and "Toe" refer to side friction and end bearing parameters. Table 3.1 Soil Parameters Recommended for use wit) i GRLWEAP (GRLWEAP, 1997). Soil Type Quake (q) Damping Coefficient (i) Skin mm (in.) Toe* Skin s/m (s/ft) Toe s/m (s/ft) Cohesive 2.5 (0.1) d/120 0.66 (0.2) 0.49 (0.15) Non-cohesive 2.5 (0.1) d/120 0.16(0.05) 0.49 (0.15) * (d) is the pile diameter, (d / 60) may be more appropriate for silts and fine-grained sands. Toe quake should not be less than 1.5 mm (0.05") for pile-driving applications. G R L W E A P can be used to model SPT stress wave data i f the modelled soil resistance is set to zero everywhere except the bottom 30 cm of the pile / drill rods. It is not clear, however, whether the"soil parameters suggested in Table 3.1, which are based on pile-driving experience, are appropriate for modelling SPT data. Goble and Abou-Matar (1992) used an interesting iterative solution technique to back-calculate soil parameters from SPT F(t) and V(t) data but unfortunately used a research soil model available in G R L W E A P , limiting the applicability of their results. Sy and Campanella (1991b) used the parameters recommended in the G R L W E A P manual to predict SPT F(t) and V(t) data that were in good agreement with field data. Abou-Matar and Goble (1997) were able to accurately recreate SPT data from a laboratory set-up but soil input was not required as the sampler was suspended in air. Morgano and Liang (1992) used skin and toe quakes of 2.5 mm (0.1") and 0.5 mm (0.02"), respectively and a value of 0.328 s/m (0.1 s/ft) for both skin and toe damping. The effect of soil parameters on G R L W E A P results is addressed further in Section 4.0. -32-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method 4. PROPOSED SPT-LPT C O R R E L A T I O N M E T H O D The Winterkorn and Fang (1975) correlation procedure described in Section 2.2 correctly recognizes the importance of the sampler dimensions and input energy but is based on a limited database and does not consider hammer efficiency. The correlation factors predicted using the procedure do not agree well with the JLPT and ILPT factors from the literature. The correlation method proposed in this section is based on the Winterkorn and Fang technique and ideas proposed by Schmertmann (1979) and Schmertmann and Palacios (1979). The method is applicable for any combination of hammer and sampler. 4.1 Soil Resistance Considerations Schmertmann (1979) compared the quasi-static (q-s) penetration resistance acting on an SPT sampler to that acting on a standard 10.0 cm 2 cone penetration test (CPT) penetrometer (Figure 4.1). He hypothesized that the force (F) required at the top of the drill rods to push the SPT sampler at the standard CPT penetration rate of two cm/s could be estimated from CPT measurements at the same depth using the formula: where: Ci = SPT-CPT end bearing correlation factor A E = SPT end bearing area (10.7 cmz) C 2 = SPT-CPT friction correlation factor ID = split-spoon inner diameter (cm) OD= split-spoon outer diameter (cm) d = split-spoon penetration depth (cm) R f = f s / q c fs = measured CPT friction sleeve stress (N/cm2) q c = measured CPT tip stress (N/cm ) W' = buoyant weight of SPT drill rods (N) F = [CrAE+C2 •{ID + OD)-7i-d-RJ[qc-W (4.1) -33-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method (a) SPT (b) CPT Figure 4.1 Forces Acting on SPT Split-Spoon and Piezocone During Quasi-Static Penetration (Schmertmann, 1979). -34-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method Schmertmann compared CPT data obtained with a Begemann mechanical cone to (F) measurements obtained in the same soil units and proposed (Cj) and (C2) values of 1.0 and 0.7, respectively. He suggested that the (fs) measurements were too high because of the design of the mechanical cone and predicted that both (Ci) and (C2) would be roughly equal to 1.0 for electric cone data. Sy and Campanella (1991b) used this approach to estimate (R u) values for their G R L W E A P analysis of SPT data. Adopting this approach for the current objective, the author suggests the calculation of an equivalent tip bearing area (ATE) as follows: ATE={CrAE) + (C2-AF-Rf) (4.2) where: A E = split spoon end bearing area = (TI / 4) • (OD 2 - ID 2) Ap = split-spoon frictional area at 305 mm (12") penetration = (ID + OD) • rt • (305 mm) The purpose of the equivalent tip bearing area is to scale down the large frictional area of the split-spoon sampler for meaningful comparison with the end bearing area. The frictional area (AF) is calculated assuming a sampler penetration depth of 305 mm (12") because this is the average sampler penetration depth during the interval that the blow count is recorded. (Rf) data is generally not available at gravel sites because CPT equipment is expensive and easily damaged by coarse particles. (Rf) is typically between 0.002 and 0.005 (0.2% and 0.5%) in cohesionless sands and should be similar for cohesionless gravel deposits. A n average (Rf) value of 0.35% can be used i f no CPT data is available. The manner in which Equation 4.1 accounts for internal and external friction is very simple and requires further investigation. Paik and Lee (1993) conducted calibration chamber tests using an instrumented, double-walled, open-ended pipe pile that allowed -35-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method them to measure the total force acting on the external wall of the pile and the force distribution along the internal wall of the pile during static loading. Their results suggest that the external lateral earth pressure coefficient (KE) may be considered constant along the penetrated length of the pile following driving (maximum depth « 0.75 m (2.5')). In contrast, they conclude that the majority of the internal friction was developed within three diameters of the pile end and that the internal lateral earth pressure coefficient decreases with increasing distance from the pile tip. The latter observations were supported by the results of a similar study by De Nicola and Randolph (1997). Although the calculation of the internal lateral earth pressure coefficient ( K ^ was based on very simple assumptions, it is clear that the internal and external stress distributions were different in these cases and it is probable that internal and external stress distributions during (q-s) penetration of an SPT sampler would also be different. The empirical correlation factor (C 2) is likely the most efficient way to account for these internal stress distributions. Because the majority of internal friction is generated in the open shoe, the inner diameter of the open shoe should be used in Equation 4.2. 4.2 Energy Input Considerations Consider the case of an SPT in which 60% of the maximum potential energy is absorbed by the soil (0.6 • 473 J = 284 J). If an ideal plastic model is assumed (Figure 4.2a), the energy absorbed by the soil is equal to the sampler displacement (Ad) multiplied by the ultimate static resistance (R u) and the blow count can be determined from: where the units of (R u) are Newtons. If an ideal elastic-plastic model is assumed (Figure 4.2b), the energy absorbed by the soil is again equal to the area under the force-displacement curve and the blow count can be determined from: 7V~ = 03 m Ad (0.3 m)-Rt 284 J •u (4.3) = R„ -0.00106 -36-M.A.Sc. Thesis, Chris R. Daniel The University of Brit ish Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlat ion Method Figure 4.2 Energy Expended During Displacement of (a) Ideal Plastic Soil and (b) Ideal Elastic-Plastic Soil. - 3 7 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method N = 0.3m f 284 J q (4.4) V J where the units of (R u) and (q) are Newtons and metres, respectively. These two cases are variations of the G R L W E A P soil model described in Section 3.2.3 for which the dynamic component of soil resistance (D) has been left out. The case including (D) can be modelled using GRLWEAP. Figure 4.3 compares Equations 4.3 and 4.4 to the output of a G R L W E A P SPT analysis using the following soil parameters: • toe quake = 1.25 mm (0.05"); • skin quake = 2.5 mm (0.1"); • Smith toe damping = 0.492 s/m (0.15 s/ft); and, • Smith skin damping = 0.164 s/m (0.05 s/ft). Discontinuous slope breaks in the G R L W E A P data (e.g. Point A in Figure 4.3) are minor effects of the method used by the program to estimate the blow count. Figure 4.3 clearly illustrates that energy considerations must include not only the total energy delivered to the soil but the amount of that energy that is expended overcoming soil elasticity and dynamic penetration resistance. Soils with blow counts greater than 50 are generally not of concern for most applications. In Figure 4.3, the relationship between the predicted blow count (N) and the ultimate soil resistance (R u) is well represented by a single straight line for blow counts less than 50 and the relationship can therefore be quantified by the slope of the line (N / R u ) . The ability to fully describe the results of a G R L W E A P analysis using a single number provides a simple means of comparing analysis results. G R L W E A P analyses of the three LPT systems described in Section 2.0 were performed using the input listed in Table 4.1 and the same soil parameters used for the original SPT analysis. The analysis output plotted in Figure 4.4 shows that the LPT data are also reasonably well represented by -38-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method Figure 4.3 Comparison of SPT Blow Counts Predicted Using Ideal Plastic and Ideal Elastic-Plastic Soil Models to G R L W E A P Analysis Results. -39-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method Figure 4.4 G R L W E A P Analysis Results for SPT, N A L P T , J L P T and ILPT. -40-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method linear relationships. Slope breaks similar to Point A are also present in the LPT data but the deviations appear to be minor compared to the slope variations between the SPT and the LPT's. Table 4.1 Hammer, Rod and Sampler Details Used for G R L W E A P Analysis. Detail Units SPT N A L P T JLPT ILPT Hammer Weight N (lb.) 623 (140) . 1335 (300) 981 (220) 5592 (1256) Hammer Length cm (in.) 53.3 (21) 69.9 (27.5) 36.8 (14.5) 59.9 (23.6) Hammer Diameter cm (in.) 14.0 (5.5) 17.8 (7) 21.1 (8.3) 39.1 (15.4) Drop Height cm (in.) 76.2 (30) 76.2 (30) 150 (59) 50 (19.7) ER V % 60 60 78 * 60 E N T H R U ** J (ft-lb) 284 (210) 610 (450) 1140 (844) 1678 (1237) Rod Length m (ft.) 17.83 (58.5) 17.83 (58.5) 17.83 (58.5) 17.83 (58.5) Rod Area cm2 (in 2) 8.0 (1.24) 9.3 (1.44) 10.1 (1-57) 60.6 (9.4) Sampler Area cm2 (in.2) 8.8 (1.37) 13.9 (2.16) 18.8 (2.93) 59.4 (9.2) Sampler Length cm (in.) 45.7 (18) 45.7 (18) 45.7 (18) 45.7 (18) . Yoshida et al. (1988) to develop SPT-JLPT correlation (Skempton, 1986). For these analyses, ER r was roughly equivalent to ER V in the portion of the rod typically used to measure energy. Sy and Campanella (1991b) performed a parametric study to determine the effect of the hammer length, hammer, rod and soil damping values, drill rod couplings, slack at rod joints, analysis time step, rod element size and soil resistance on G R L W E A P computed SPT F(t) and V(t) data. They observed that the predicted F(t) and V(t) waveforms were most sensitive to the geometry of the hammer and rods above the measurement point and the damping values for the hammer, rod and soil elements. They also observed that the input soil resistance and hammer efficiency were the most significant factors affecting the -41 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 4.0 Split Spoon Penetration Testing in Gravels Proposed SPT-LPT Correlation Method predicted blow count and ENTHRU values. An earlier study by McLean et al. (1975) had also made the latter conclusion. The author conducted a parametric study to determine the sensitivity of the calculated (N / R u ) values to soil parameter variations. Table 4.2 details six sets of soil parameters that were used to generate (N / R u ) values for the SPT and three LPT's. Cases (a) and (c) are the values recommended in the G R L W E A P manual (Table 3.1) and the Case (b) values were used by Morgano and Liang (1992). Cases (d), (e) and (f) demonstrate the effect of varying the toe quake and skin damping values from those recommended for non-cohesive soils by G R L W E A P (1997). (N / Ru) values calculated from the analysis results are summarized in Table 4.3. Table 4.2 Soil arameters Used for G R L W E A P Parametric Study Parameter Units Case (a) (b) (c) (d) (e) (f) Skin Quake (q) mm (in.) 2.5 (0.1) 2.5 (0.1) 2.5 (0.1) 2.5 (0.1) 2.5 (0.1) 2.5 (0.1) Toe Quake (q) mm (in.) 1.25 (0.05) 0.5 (0.02) 1.25 (0.05) 1.25 (0.05) 2.5 (0.1) 0.5 (0.02) Skin Damping (J) s/m (s/ft.) 0.16 (0.05) 0.33 (0.1) 0.66 (0.2) 0.33 (0.1) 0.16 (0.05) 0.16 (0.05) Toe Damping (J) s/m (s/ft.) 0.50 (0.15) 0.33 (0.1) 0.50 (0.15) 0.50 (0.15) 0.50 (0.15) 0.50 (0.15) Skin / Toe Resistance Distribution % 50/50 25/75 50/50 50/50 50/50 50/50 a) G R L W E A P (1997), non-cohesive soil. b) Morgano and Liang (1992), "various sites' c) G R L W E A P (1997), cohesive soil. -42-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method Table 4.3 (N / R u ) Values Calculated Using G R L W E A P Parametric Study Results. Test N / R u Units: bpf / kN (bpf / kin) Average Case (a) Case (b) Case (c) Case (d) Case (e) Case (f) SPT 2.325 (10.340) 2.130 (9.473) 2.851 (12.676) 2.502 (11.125) 2.443 (10.862) 2.252 (10.015) 2.417 (10.749) N A L P T 0.998 (4.436) 0.946 (4.207) 1.218 (5.416) 1.068 (4.749) 1.025 (4.558) 0.981 (4.363) 1.039 (4.622) JLPT 0.655 (2.914) 0.606 (2.693) 0.836 (3.715) 0.718 (3.191) 0.681 (3.030) 0.638 (2.838) 0.689 (3.064) ILPT 0.356 (1.582) 0.328 (1.460) 0.429 (1.907) 0.381 (1.692) 0.372 (1.656) 0.346 (1.538) 0.369 (1.639) The data in Table 4.3 show that changing the input soil parameters significantly alters the calculated (N / R u ) values. Of principal interest for the current application, however, is the ratio of SPT to LPT (N / R u ) values, as it may be assumed that the quake and damping parameters will not be scale dependent, at least over the range of sampler dimensions under consideration. Figure 4.5 compares the SPT / LPT ratios calculated for each LPT using each set of soil parameters and clearly shows that the ratio variations are much more dependent on the type of LPT than the soil parameters used for the analysis. The dominant factor controlling the calculated (N / R u ) value is E N T H R U . Figure 4.6 compares the Case (a) (N / R u ) values to ENTHRU for the SPT and each of the LPT's. The data are well represented by the equations: N_ Ru N_ R„ 2160 ENTHRU 658 ENTHRU (Imperial Units) (Metric Units) (4.5a) (4.5b) which illustrate the inverse proportionality between blow counts and input energy that was originally observed in the field by Schmertmann and Palacios (1979). Equation 4.5 may be used to predict the (N / R u ) value for any LPT system if E N T H R U is known. Alternatively, the ratio of SPT to LPT (N / R u ) is simply equal to the ratio of LPT to SPT ENTHRU, i f inverse proportionality is assumed. -43 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method 8 i 7 -ILPT (average = 6.56) A Jk. 6 -5 -4 -<~> J L P T (average = 3.51) o O 3 -• • • 2 - NALPT (average = 2.33) 1 - i a b c d e Case Identification Figure 4.5 Sensitivity of Predicted (N / R u) Values to Soil Parameter Variations. -44-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method Figure 4.6 Inverse Proportionality Relationship Between (N / R u) and E N T H R U . -45-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method The fact that the four data points do not fall precisely on a single curve suggests that the inverse proportionality assumption may only be valid over small ranges of input energy such as those encountered during SPT's, or that factors other than E N T H R U and the soil parameters can affect the predicted blow counts. The author conducted an additional parametric study to determine the effect of hammer geometry and rod cross-sectional area on (N / R u ). It was found that changing the rod cross-sectional area had a significant effect on the predicted blow count (Figure 4.7). Schmertmann and Palacios (1979) present field data acquired with (AW) and (NW) rods that demonstrate this phenomenon. The steady increase in (N / R u) with rod area shown in Figure 4.7 may be explained by considering the shape of the down-going stress wave (Abou-matar and Goble, 1997). As the rod area increases, the peak force of the stress wave and the rate of force "decay" following the passing of the peak force also increase (Figure 4.8). The soil reaction force, as modelled by GRLWEAP, can only exceed (Ru) i f damping forces are included in the total resistance. As the peak force increases, damping forces will increase, energy transfer efficiency will decrease and the resulting (N / R u ) slope will increase. The increase in the NWJ rod (N / R u ) with decreasing area below roughly 1.5 in in Figure 4.7 occurs because the (N) vs. (R u) relationships become increasingly non-linear as smaller rods are modelled. The resulting best-fit relationships tend to over-estimate (N) for low values of (R u). The data points used to represent the SPT (AW rods) and North American LPT (NWJ rods) are at the edge of the region where this non-linearity becomes important but the resulting error appears to be small relative to the ratio difference between the two systems. 4.3 Synthesis of Proposed Method The (R u) value is very similar to the quasi-static penetration resistance value (F) described by Schmertmann (1979). As a first approximation, the blow count for any of the SPT or LPT systems may be estimated using the equation: NERA=(N/RU)-F (4.6) -46 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method 1 ' — 1 1 i — 1 — ' — ' — i i 2 3 4 Rod Cross-Sectional Area (in2) Figure 4.7 Effect of Rod Cross-Sectional Area on Predicted (N / R u ) Values. -47-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method Figure 4.8 Idealized Effect of Rod Cross-Sectional Area on Axial Force Data. -48-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method where (ER A ) is the energy ratio used during the G R L W E A P or equivalent analysis. Similarly, the correlation factor between the SPT and any of the LPT's may be estimated by taking the ratio of the two: (N6Q)SPT _ (NJK)SPT • FSPT (NERA)LPT (NIRU)LPT-FLPT where the standard SPT energy ratio of 60% has been used. Alternatively, since the (N / R u ) values of the SPT and LPT's basically fall on a single inverse proportionality relationship, the correlation factor may be predicted using the equation: (X6Q)SPT _ (ENTRHU) L P T • FSPT (4.8) (NERA)LPT (ENTHRU)SPT • FL LPT Referring to Equation 4.1, the ratio of the quasi-static penetration resistances (F) may be determined entirely from the geometry of the split-spoon samplers i f it is assumed that: • the buoyant unit weight of the rods is negligible; and, • the same value of CPT tip resistance (qc) may be used for the SPT and LPT (i.e. no significant scale effects). Equation 4.7 and 4.8 then reduce to: )SPT _ (N/RU)SPT '{ATE)SPT (KEPJLPT ( N / R J L P T \ A T E ) L P T ( - } and (N6Q)SPT _ (ENTRHU)LPT-(ATE)SPT WERJLPT (ENTHRU)SPT • (ATE) (4.10) TE JLPT -49-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method which can be used to predict SPT-LPT correlation factors for any LPT i f energy and dimensional information are available. 4.4 Application of Proposed Method Table 4.4 summarizes the input data for Equations 4.9 and 4.10 for each of the LPT's described in Section 2.0 as well as the resulting correlation factor predictions. Table 4.5 summarizes the correlation factors taken from the literature, predicted using the Winterkorn and Fang (1975) method and predicted using the proposed method. Table 4.4 Summary of Proposed Correlation Method Input Data and Results Test A E (in2) A F (in2) A T E * (in2) Equation 4.9 Equation 4.10 N / R u (bpf/kip) 0 ^ 6 0 )sPT (NER )LPT ENTHRU (ft-lb) 0 ^ 6 0 )sPT i^ER )LPT SPT 1.66 127.2 2.10 10.340 - 210 -N A L P T 2.54 203.6 3.26 4.436 1.50 450 1.38 JLPT 3.46 184.7 4.11 2.914 1.81 844 2.06 ILPT 11.81 354.4 13.05 1.582 1.05 1237 0.95 * Assumes (C 2) = 1.0, R f = 0.0035 Table 4.5 Summary of Observed and Predicted Correlation Factors. Test Type Observed Correlation Factors Winterkorn and Fang Equation 4.9 Equation 4.10 Format Material Value (N)SPT 0 ^ 6 0 )sPT (NER )LPT 0 ^ 6 0 )sPT i^ER )lPT N A L P T - - - 0.93 1.50 1.38 JLPT {N)SPT Sand 1.5 1.02 1.81 2.06 Gravel 2.0 ILPT [N\(60)\spt Sand 1.14 0.44 1.05 0.95 Sand and Gravel 0.89 1^1(60) \ L P T Sand and Gravel 1.02 -50-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method In all of the cases, the correlation factors predicted using the Winterkorn and Fang method are lower and therefore more conservative than those calculated using the proposed method. In fact, the SPT blow counts predicted using the proposed method would range from 1.5 to 2.0 times those predicted using the Winterkorn and Fang method. An assessment of which method is correct (or "more correct") should be based on comparison of predicted and observed correlation factors. Such comparisons are somewhat limited by the fact that the SPT-JLPT correlation factor was developed using raw blow counts and the SPT-ILPT correlation factor was developed using blow counts corrected to a standard energy and overburden pressure (98 kPa). The observed SPT-JLPT correlation factors are 1.5 to 2.0 times the value predicted using the Winterkorn and Fang method. The correlation factors predicted using the proposed method are in good agreement with the observed gravel correlation factor but are higher than the sand factor. It should be noted that Yoshida et al. (1988) used Tonbi type hammers for both the SPT and JLPT's that were used to develop their correlation factors. The correlation factors tabulated in Table 4.5 were calculated assuming SPT and JLPT energy ratios of 60% and 78%, respectively. If it is assumed that the SPT energy ratio was 78% (273 fit-lb), the correlation factor predicted using Equation 4.10 decreases from 2.06 to 1.58. Assuming a similar decrease in the correlation factor predicted using Equation 4.9, the new range of predicted correlation factors becomes 1.39 to 1.58, which is in good agreement with the observed sand correlation factor of 1.5. The use of sand versus gravel correlation factors will be discussed in Section 8.0. Crova et al. (1993) used SPT and ILPT blow counts collected at the same depths to determine the SPT-ILPT correlation factors. They also state that they used the same method to correct the SPT and ILPT blow counts to a standard overburden pressure. Thus, it can be assumed that the applied overburden correction factors do not affect the ratio of SPT to ILPT blow counts. The observed SPT-ILPT correlation factors are 2.0 to 2.6 times the value predicted using the Winterkorn and Fang method. The correlation -51 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 4.0 Proposed SPT-LPT Correlation Method factors predicted using the proposed method are in good agreement with all of the observed correlation factors. The author participated in three N A L P T field programs to gather additional data for calibrating the proposed correlation method. The results of these programs are presented in the next three sections and discussed in Section 8.0. -52 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program 5. KIDD2 N A L P T F I E L D P R O G R A M The author conducted a five day field program at a site called "Kidd2" located on the Fraser River Delta just south of Vancouver, BC. The site is adjacent to an electrical substation and is the property of the British Columbia Hydro and Power Authority (BC Hydro). The intent of the program was to acquire as much data as possible to check and to calibrate the proposed correlation method. To that end, both dynamic and quasi-static penetration tests were performed using SPT and N A L P T equipment. Comparisons of the quasi-static test data and CPTU data are presented and used to check the penetration resistance assumptions of the proposed correlation method in this section. Empirical correlation factors developed using the dynamic SPT and N A L P T data are also presented and compared to the correlation factors predicted using the Winterkorn and Fang and the proposed method in this section. Kidd2 was selected for this investigation for the following reasons: • close proximity to the University of British Columbia (UBC); • extensively characterized during the Canadian Liquefaction Experiment (CANLEX) and during subsequent research by the UBC Civil Engineering and Earth and Ocean Sciences Departments; • anticipated SPT blow count range of 10 to 35 within upper 20 metres, based on C A N L E X results; and, • silt over silty sand stratigraphy (Figure 5.1) allows SPT-NALPT comparison without grain size effects. One piezocone penetration test (CPTU), two SPT and two N A L P T were performed at the separations shown in Figure 5.2. Grain size distribution data from test holes LPT9902 and SPT9904 are shown on Figure 5.3. -53-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program SPT9901 O CPT ^ LPT9903 • LPT9902 O SPT9904 I—h 0 —I—I 1.5m Figure 5.2 Distribution of SPT, N A L P T and C P T U Test Holes at Kidd2. -55-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program 100 90 80 70 60 50 40 30 20 10 0 — > • \ \ \ \\ \ \ \'\ V \ \ \ \ \ \ \ \ \ \ \ — x 'n \\ N J w > \ > 1. 100 10 1 Grain Size (mm) 0.1 0.01 Range from SPT9904 Range from LPT9902 Figure 5.3 Kidd2 Grain Size Distribution Data. -56-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program 5.1 Drilling Method The four test holes were drilled by the mud-rotary method using a Simcoe 5000 rig supplied by Foundex Explorations Inc. A 6.125" OD tricone bit was used to drill to the first test depth and a 4.75" OD tricone bit was used for further advancement of each of the test holes. The level of the bentonite drill mud mixture was kept above the GWT to ensure hole stability. SPT and N A L P T were generally performed at 1.5 m (5') intervals beginning at 4.9 m (16'). Quasi-static (q-s) penetration tests were performed at selected depths between dynamic test depths. 5.2 Quasi-Static Penetration Tests The quasi-static penetration test method and results are presented and discussed below. 5.2.1 Description of Test Method NWJ drill rods were used to lower the SPT or N A L P T split-spoon sampler to the base of the test hole prior to each quasi-static test. Vertical pushing force was applied to the NWJ rods at the surface using a constant flow rate hydraulic ram that was securely mounted on the rig. Additional reaction force was achieved by installing a 3.05 m (10') solid stem auger roughly one metre toward the front of the drill rig from the test hole location. The frame of the rig was securely chained to this anchor and the flow rate of the hydraulic ram adjusted so that the penetration rate would be two cm/s prior to each quasi-static penetration test. Plastic sample retainers were used and sample barrel liners were excluded for all tests. The axial force in the rods was measured using a strain-gauged N W J rod that was developed to measure FF and F V energy during dynamic tests. The instrumented rod was placed in the rod string directly below the pushing head of the hydraulic ram. The strain-gauge output in volts was recorded as a function of time on a strip chart recorder. Axial force was calculated from the strip chart data using a linear calibration factor -57-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program developed at U B C by hydraulically loading the instrumented rod and monitoring the strain gauge output. 5.2.2 Test Results Figure 5.4 is an example of the strip chart output from a quasi-static SPT split-spoon sampler penetration test. Table B - l , Appendix B, summarizes the test depth, the C P T U data recorded at the test depth, the estimated rod weight, the strip chart force recorded at 305 mm (1') sampler penetration and the quasi-static resistance force predicted using Equation 4.1 for each test. It is noted that the average (Rf) values listed in Table B - l are reasonably close to the selected average value of 0.35%. Table B - l also lists the recovery of each test, defined as: Sample Length Recovery = - 100% (5.1) Sampler Penetration Figure 5.5 shows the penetration force recorded at 76 mm (3") intervals for each test. Figure 5.6 presents the CPTU tip resistance (qt) profile and the predicted quasi-static resistance profiles for the SPT and N A L P T split-spoon samplers. The predicted resistance at zero (toe) and 30 cm (12") penetration are plotted on the same graphs. Also plotted on Figure 5.7 are the measured penetration resistance (including rod weight) at 76, 305 and 457 mm (3", 12" and 18") penetration. 5.2.3 Discussion The data in Figure 5.6 show that the measured quasi-static resistance forces at 76 mm (3") penetration are generally close to the predicted toe resistance (i.e. zero penetration) values but the measured values quickly become larger than the predicted values as sampler penetration increases. This difference is more apparent when the measured force is plotted directly against the predicted force, as shown in Figure 5.7. The slope of the -58-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program o • Sampler Penetration Figure 5.4 Sample SPT Quasi-Static Penetration Test Strip Chart Output. -59-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program 12 15 18 21 Penetration (in.) Recovery (%) 24 27 30 Figure 5.5 Summary of SPT and N A L P T Quasi-Static Resistance Versus Penetration. -60-o ui re c g O <8s CM T3 T3 5 w 0) > 2 O .2 c o c c "E — s re o JR Q O |5! 0 5 . 1 § «.r £ i _ t- co O *- o - O c vt Stx. 5 2 = H § 8 •;= P. <0 o c W 3 _ < CO 2 5 H co o c iS w w tr tn i cr h-0_ _ i < 0) «! T3 O JS £ ro — g o o ' j : ro CD CO +-i 1— •*—> CD CD CD c c c 0 0 CD 0. 0_ Q_ - C N co r o T— • A • <=> -, o CM ; O • O -C\l ; ;bar) O • in -cr o • o -o to . o -t o (LU) U}d9Q a a Q SO 3 t ii > ii U im © fa u a at e o cs u ii a » 0M +J CZ) i « S u L. OH •a a CQ -a SI s w CJ 01 V© «/) u >_ S bl) to M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program 0 2 4 6 8 10 12 Predicted Force (kip) Figure 5.7 Comparison of Measured and Predicted Quasi-Static Penetration Resistance Force. -62-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 5.0 Split Spoon Penetration Testing in Gravels Kidd2 NALPT Field Program best-fit relationships shown in Figure 5.7 can be more easily understood i f it is assumed that (Ci) is equal to (C2) and Equation 4.1 is rearranged as follows: F + W*= qc-C-\AE+K-d-Rf- (ID + OD)\ (5.2) in which case the best-fit slopes are equivalent to the constant (C). The values of (C) are 1.2 and 1.46 for the SPT and NALPT, respectively. Thus, it may be appropriate to rewrite Equations 4.9 and 4.10 as: (N6JSPT _ (N/RJSPT.\.2-{ATEl (KERA )NALPT (N/RU)NALPT-\A6-{ATE j S P r - (5.3) ) NALPT and (N60)SPT (ENTHRU) NALPT '1-2'(-<4TE)SFT (A^,)NALPT " (ENTHRU)SPT • 1.46 • (ATE )NALPT This correction will be referred to herein as the quasi-static penetration resistance correction factor. Similar corrections to the correlation calculations for the JLPT and ILPT would require field determination of the (C) value for the particular samplers. 5.3 Dynamic Penetration Tests SPT and N A L P T penetration tests were performed with energy measurement by the hammer velocity (radar), FF and FV methods. 5.3.1 Description of Test Method Input energy for the SPT was provided by a 640 N (144 lb) safety hammer dropped 76.2 cm (30") for a maximum possible energy of 486.5 J (360 ftlb). The accuracy of the scale used to weigh the hammer and the control over the drop height are such that it is reasonable to assume that the maximum possible energy was 475 J (350 ftJb). A 0.61 m (2') A W transducer rod was attached below the N W anvil rod of the safety hammer and the rest of the rod string consisted of 1.52 m (5') A W J rods. The split spoon dimensions - 6 3 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 5.0 Split Spoon Penetration Testing in Gravels Kidd2 NALPT Field Program were within A S T M standards. A plastic sample catcher was used and sample barrel liners were excluded for all tests. An NWJ transducer rod and rod string were used for the SPT performed at 18.6 m (61') in SPT9904. Input energy for the N A L P T was provided by a 1458 N (328 lb) safety hammer dropped 76.2 cm (30") for a maximum possible energy of 1108 J (820 ft-lb). This is a significant variation from the N A L P T "standard" of 1334 N (300 lb) which will require revision of the estimated correlation factor. A 0.61 m (2') NWJ transducer rod was attached to the H W anvil rod of the safety hammer and the rest of the rod string consisted of 3.0 m (10') N W J rods. The split spoon used had an outer diameter of 7.62 cm (3"), an inner diameter of 6.1 cm (2.4") in the open shoe and an inner diameter of 6.35 cm (2.5") in the sample barrel. A plastic sample catcher was used and sample barrel liners were excluded for all tests. Both the SPT and N A L P T hammers were lifted and dropped using the rope and cathead technique with two turns of rope around the cathead. The rope was loosened, not thrown off the cathead, to initiate hammer drops. Weather conditions were dry for most of the field program and it was necessary to wet the section of the rope that was in contact with the cathead to reduce friction. The SPT and N A L P T split spoons were driven 457 mm (18") and 610 mm (24"), respectively, into the soil at the base of the test holes. The number of blows required for each 2.54 cm (1") of penetration were recorded for both test types. Samples were classified in the field, bagged and returned to U B C for storage. 5.3.2 Energy Measurement Hammer kinetic energy was measured for most of the hammer drops using an HPA system belonging to Klohn-Crippen Consultants Ltd. Figure 5.8 shows an example of HPA data gathered during the field program and an annotated, idealized data set. H P A system calibration consists of holding a vibrating tuning fork in front of the antenna. The HPA comes equipped with tuning forks that oscillate at 4.88 m/s and 9.76 m/s (16 ft/s -64-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program .RANDOM. NOISE O o a) ti T a. > S ° ) trt »- o> = = .£ U. v v Zero Velocity Line . HAMMER DROP J RANDOM. I NOISE Peak Acceleration! JVflo 0.2 s Figure 5.8 Sample and Idealized H P A Strip Chart Output. -65-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program and 32 ft/s). The system was calibrated at the beginning and end of each field day but it was never deemed necessary to adjust the calibration factor. Stress wave data were recorded using a Dynamic Energy Monitoring (DEM) system built for research applications by the U B C Civil Engineering Department and Northwood Instruments. The SPT measurement system consisted of a 0.61 m A W transducer rod that was instrumented with four orthogonal strain gauges to measure force and a bolt-on Entran 7270A-6K (6000 g) accelerometer. The top of the transducer rod was placed in the rod strong roughly 1.5 m below the plane of hammer impact. The accelerometer was mounted roughly 25.4 mm (1") below the strain gauges. A linear calibration factor was developed from laboratory static load tests to relate the output voltage from the strain gauges to axial force in the rod. The accelerometer calibration factor was provided by Entran. A n N W J rod was instrumented in the same manner for use during N A L P T . Both force and acceleration data were sampled at 40 kHz, yielding a Nyquist Frequency of 20 kHz (i.e. the system is capable of monitoring signals with frequency content as high as 20 kHz). A total of 4000 data points were recorded on each channel for each hammer blow. The data points were recorded at 0.025 ms intervals, yielding 100 ms of force and acceleration data per hammer blow. Acceleration was digitally integrated to obtain velocity. Figure 5.9 shows the operating screen of the software used to display and store the data gathered by the D E M . The software automatically plots force and velocity following each hammer blow. The FF energy calculated at the input (2L/c) time and the maximum FV energy calculated over the length of the data trace are listed along with the peak force and peak acceleration on the right side of the operating screen after each blow. The user inputs sufficient rod length and impedance prior to the test to calculate FF energy using Equation 3.10 (correction factors K i , K2 and Kc are not applied). FV energy is calculated using Equation 3.12 and is shown plotted with force and velocity as a function of time in Figure 5.9 (this option is not available in real time). -66 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program SPT REVIEW DATA o z •< 10 15 20 TIME (ms) 25 30 35 FILENAME K9904560.DAT DEPTH ( f t ) 56 . OO PEAK FORCE|21 . 3 (K i p s ) 1 : PEAK VEL. 11 0 . 8 ( f t / s ) ' 1 PK. ACCEL. (9) 321 5 FV ENERGY 56 4 () F2 ENERGY |55.3 () Display Controls FV For. V e l . E n e r g y • • • SWEEP EXCLUDED 23] • SITE LOCATION |Kidd2. Near CANLEX DRILLERS/RIG |Foundex, SIMCOE 5000, mud rotary, rope and cathead J DATE |Mar 28, 99] HOLE # KLPT9904 Figure 5.9 Sample DEM Software Operating Screen. -67-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program 5.3.3 Test Results HPA data were not obtained during SPT performed in SPT9901 above 14 m (46") or during the SPT performed in SPT9904 at 14 m (46'). Tables B-2 and B-3, Appendix B, summarize the HPA results from SPT9901 and SPT9904 and LPT9902 and LPT9903, respectively. The average peak velocity and standard deviation are listed for each SPT and NALPT. The average peak fall velocity of the SPT safety hammer was 3.14 m/s (10.3 ft/s), corresponding to an average (ERV) of 65% of the maximum possible SPT energy. The average peak velocity of the N A L P T safety hammer was higher at 3.42 m/s (11.24 ft/s), corresponding to an average (ERV) of 78.5% of the maximum possible energy (1113 J, 820 ft-lb). The calculated standard deviations and observed ranges of recorded velocity suggest that the apparent greater efficiency of the N A L P T hammer used in this study was real and repeatable. Tables B-4 and B-5, Appendix B, summarize the HPA, FF and FV data gathered during the Kidd2 field program. The force and velocity data were visually reviewed and obviously erroneous data removed before the average FF and FV energies listed in Tables B-4 and B-5 were calculated. The most frequently encountered problem was the occurrence of minor accelerometer baseline shifts during the recording period, indicated by velocity traces that were essentially linear with non-zero slopes beyond roughly 40 ms. Rod cross-sectional areas were determined through field micrometer measurements of the inner and outer rod diameters. In most cases, the rod outer diameter met industry standards but inner diameters did not. The SPT rod string consisted of an " A W " transducer rod above " A W J " rods with cross-sectional areas of 7.16 cm 2 and 4.84 cm 2 (1.11 in 2 and 0.75 in 2). The N A L P T rod string consisted of "NWJ" rods with cross-sectional areas of 9.81 cm (1.52 in'). - 6 8 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 5.0 Split Spoon Penetration Testing in Gravels Kidd2 NALPT Field Program The FF energy method requires the input of a single rod impedance (Equation 3.10). Table B-4 summarizes FF energy ratios calculated using both the " A W J " rod and " A W " transducer rod impedances. Raw blow counts recorded during the Kidd2 field program are summarized in Tables B-4 and B-5. SPT blow counts corrected to 60% of the maximum possible SPT energy using the FF (AWJ and A W rod area) and FV energy ratios are summarized in Table B-4. K i and K.2 correction factors were interpolated from the values tabulated in A S T M D 4633-86. The K c correction was not applied. Raw and stress wave energy corrected blow counts are plotted as a function of depth in Figure 5.10a. The data in Table B-5 shows that the N A L P T rod energy ratios are generally around 60%. To minimize errors associated with large energy corrections, the N A L P T blow counts should be corrected to a standard energy of 60%. N A L P T blow counts corrected to 60% of the maximum possible Kidd2 N A L P T energy (60% of 1113 J, 820 ft-lb) using the FF (NWJ rod area) and FV energy ratios are summarized in Table B-5. As the calculated FF energy ratios were almost entirely greater than 100%, only the raw and FV energy corrected blow counts are plotted in Figure 5.10b. 5.4 Discussion of Energy Data The importance of obtaining accurate values of ENTHRU was illustrated in Sections 3.0 and 4.0. The author assessed the reliability of recorded data by checking repeatability, calibration factors (when possible), determining upper bounds for calculated energies and checking force-velocity proportionality. 5.4.1 Data Repeatability The simplest method of assessing the reliability of energy data is to check repeatability. Other than poorly functioning measurement equipment, there are a variety of reasons why dynamic penetration test energy data could lack repeatability including inconsistent hammer drop height, progressive loosening of rod couplings during testing and variable soil conditions. Despite these poor odds, data will be presented herein illustrating that HPA and D E M data recorded during SPT and LPT can be remarkably repeatable. -69-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program N and N 6 10 20 N and N 60 30 40 0 10 20 30 40 50 u -SPT9901 • • • • SPT9904 2 -4 -• 6 -A 4 0 -A & 8 -A cm - A C * 10 - A oAm A » 12 -A CM A Ok • -A a 14 -A 0 - » -A « » 16 - A O * - A OA 18 -A » A O » -A CJA O A 20 -• N O N 6 0 (FV) A N 6 0 (FF, A W J rods) A N 6 0 (FF, A W rods) Figure 5.10a Raw and Energy Corrected SPT Blow Counts Versus Depth . -70-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program N and N c N and N 60 • N O N 6 0 (FV) Figure 5.10b Raw and Energy Corrected N A L P T Blow Counts Versus Depth. - 7 1 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section S.O Kidd2 NALPT Field Program The HPA data recorded during each SPT or LPT at any given depth was qualitatively very repeatable. This observation is supported by the calculated velocity energy ratio standard deviations of three to four percent (Table B-2 and B-3). The author notes, however, that the strip chart output (Figure 5.8) can only be reliably measured to the nearest 0.5 mm, corresponding to a resolution of 0.12 m/s (0.4 ft/s). The velocity energy ratio is proportional to the square of -the velocity and the resulting resolution is significantly worse. For example, SPT hammer fall velocities of 3.17, 3.29 and 3.41 m/s (10.4, 10.8 and 11.2 ft/s) are equivalent to velocity energy ratios of roughly 67, 72 and 79 percent, respectively. A large amount of data is required to overcome the effects of this poor resolution when calculating average values and standard deviations. Though it is likely not possible to accurately quantify the Kidd2 repeatability because of the size of the data sets, the observed qualitative repeatability seems to be confirmed by the highly repeatable D E M data. Figures 5.11 and 5.12 show the D E M data recorded during each of the hammer blows within the 152 mm to 457 mm (6" to 18") sampler penetration range at 9.5 m (31') in SPT9904 and at 18.6 m (61') in SPT9901, respectively. The force data in both figures is highly repeatable until between — and — . The velocity data contain much more c c scatter, presumably because acceleration measurement errors are compounded during integration, but the trend of the data is still very repeatable at both depths. Figure 5.13 compares the average force and velocity data traces recorded at the two test depths. As would be expected, the data are in poor agreement following the — time of the test at c the shallower depth but prior to this time the two force traces are in excellent agreement, illustrating that the force data were not only highly repeatable at individual test depths but also between test depths and test holes. The agreement between the two velocity traces is better than would be expected, based on the appearance of the data in Figures 5.11 and 5.12. This suggests that the scatter observed at each test depth is largely random error that may be eliminated by using the average data traces. -72-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program Figure 5.11 D E M Output Recorded During Hammer Blows Within the 152 mm to 457 mm (6" to 18") Sampler Penetration Range at 9.5 m (31') Depth in SPT9904. - 7 3 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program Figure 5.12 D E M Output Recorded During Hammer Blows Within the 152 mm to 457 mm (6" to 18") Sampler Penetration Range at 18.6 m (61') Depth in SPT9901. -74-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program g. 1*: CD p Q. 1* N o o > Figure 5.13 Comparison of Average Force and Velocity Data Recorded in Two SPT Test Holes at Differing Depths. -75 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 5.0 Split Spoon Penetration Testing in Gravels Kidd2 NALPT Field Program Figures 5.14 and 5.15 show N A L P T D E M data recorded at 9.5 m (31') in Test Hole LPT9903 and at 17.1 m (56') in Test Hole LPT9902. The force data is again highly 6L repeatable until roughly — . The velocity data appears to contain less scatter than the c SPT data but this is probably because a lower number of hammer blows are plotted. Figure 5.16 compares the average force and velocity data traces recorded at the two test 2L depths. The repeatability of the data is good up to the — time of the shallower test but c not as good as observed during the SPT. This may be due to the higher frequency content of the N A L P T data, which is visually apparent in Figures 5.14 through 5.16. The two N A L P T data sets compare reasonably well, however, suggesting that the higher frequencies are still within the recording capabilities of the measurement equipment. 5.4.2 Calibration Factors Having established that the HPA and D E M data are repeatable, the author sought to determine whether the magnitudes of the data were reasonable. The magnitude of any processed data point is entirely dependent on the calibration factor used to convert the voltage output of the measurement equipment to the desired engineering units. The calibration factor of the HPA was easily checked in the field because it was easy to reproduce the test conditions in a very controlled fashion using tuning forks. It would also be relatively simple to check the calibration factor of the force transducer in the field using a high capacity load frame. The force calibration factors were checked in the U B C laboratory at the beginning and end of the Kidd2 investigation and no significant change was observed. Checking the calibration factor of high capacity accelerometers such as those used during the Kidd2 investigation is very difficult because of the difficulty of recreating test conditions in a controlled fashion. For example, typical peak accelerations recorded during the Kidd2 investigation ranged from 5000 to 7000 g. For this reason, calibration factors provided by the accelerometer manufacturer were used during the Kidd2 -76-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program 70 • • • i • • • i • i i •i i i—i—i—i—i i i i i -5 0 5 10 15 20 25 30 35 Time (ms) Figure 5.14 D E M Output Recorded During Hammer Blows Within the 152 mm to 457 mm (6" to 18") Sampler Penetration Range at 9.5 m (31') Depth in LPT9903. -77-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program Figure 5.15 D E M Output Recorded During Hammer Blows Within the 152 mm to 457 mm (6" to 18") Sampler Penetration Range at 17.1 m (56') Depth in LPT9902. -78-M.A.Sc. Thesis, Chris R. Daniel The University of Bri t ish Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program Time (ms) Figure 5.16 Comparison of Average Force and Velocity Data Recorded in Two N A L P T Test Holes at Differing Depths. - 7 9 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program investigation. The use of force-velocity proportionality to check integrated accelerometer data against the more reliable force data will be discussed in Section 5.4.4. 5.4.3 Quality Control Using Upper Bounds The most obvious upper bound that can be used to check HPA and D E M calculated energy ratios is 100%. Although this criterion does not seem very stringent, all of the N A L P T FF energy ratios are greater than 100% and therefore must be in error (Table B-5, Appendix B). The velocity energy ratios calculated from the HPA output provide a more stringent, though less certain upper bound. A number of authors have observed that the actual energy available for penetration of the sampler into the soil may be greater than the velocity energy ratio for very soft soils with low blow counts. The additional hammer energy available for sampler penetration may be calculated using the formula: Additional Potential Energy = m ' S • Sampler Set m-g-30" 0.4 ( 5 ' 5 ) = 100% N The above relationship is illustrated in Figure 5.17. A blow count of three was recorded during the N A L P T at 4.11 m (13.5 ft) in Test Hole LPT9902. The corresponding additional potential energy of 13.3% represents the amount by which the measured rod energy ratio could exceed the velocity energy ratio recorded by the HPA system. Although the effect of the soil-sampler interaction is not felt at the transducer location 2L 2L before — , some portion of the net rod displacement will already have occurred at — c c due to compression of the rods between the measurement point and the sampler. The minor downward translation of the measurement point that occurs between impact and -80-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program • • • • i • i i i i i i i i i i i i i i 0 5 10 15 20 25 30 35 40 45 50 N Figure 5.17 Relationship Between Additional Potential Energy Due to Sampler Set and Blow Count. -81 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program — fails to explain why the N A L P T FF rod energy ratios exceed the velocity energy c ratios by 20% to 60%. Figure 5.18 compares the average SPT and N A L P T FV energy ratios to the "corrected" energy ratio calculated from the sum of the hammer kinetic energy and the additional potential energy calculated using Equation 5.5. Both the SPT and N A L P T F V energy ratios are roughly equivalent to 0.82 times the corrected velocity energy ratios but the scatter in the N A L P T data appears to be high. 5.4.4 Quality Control Using Force-Velocity Proportionality Force-velocity proportionality is perhaps the most commonly used quality control technique for D E M data. The technique is based on the assumptions that any stress wave 2 1 reflections recorded at the transducer location prior to — will be minor and that c Equation 3.6 will continue to be valid despite such reflections. Figure 5.19 compares the average force and velocity data recorded at 18.6 m (61') in SPT9904. The velocity has been multiplied by the impedance of the A W transducer rod to simplify the comparison. If no significant reflections were returned from below the measurement point and the D E M was functioning correctly, one would expect the force 2L to be equal to the velocity multiplied by the rod impedance until — . This is not the case c in Figure 5.19, however, where the velocity is consistently greater than the force in this time range. The author notes that the first major impedance interface below the measurement point is the coupling between the A W transducer rod and the A W J drill rods. The two-way travel time between the measurement point and the interface is 0.2 ms, which is almost instantaneous, compared to the rise time of the velocity. Equations A.5 and A.6 (Appendix A) indicate that the wave reflected at an A W to A W J interface would cause a force-velocity divergence of roughly 40% of the original force magnitude (velocity increases, force decrease). Assuming an initial force magnitude of roughly 16 -82-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program Figure 5.18 Comparison of D E M and "Corrected" H P A Energy Data. - 8 3 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program Figure 5.19 Average SPT Force and Velocity Data Recorded at 18.6 m (61') in SPT9904. -84-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program kip, the anticipated force-velocity divergence is 6.4 kip, which is in good agreement with the early data in Figure 5.19 and raises important questions about the validity of the force-velocity proportionality assumption of the FF method. The FF energy ratios calculated using the A W impedance ranged from 34% to 46% and were considerably lower than the equivalent FV energy ratios, which ranged from 47% to 68%. In contrast, the FF energy ratios calculated using the A W J rod impedance are in fair agreement with the FV values, ranging from 51% to 68%. There is a simple theoretical explanation for this observation. When the downward propagating stress wave encounters the A W / A W J impedance interface, the force and velocity of the transmitted and reflected waves can be estimated using Equations A.3 to A.6 (Appendix A), which are based on requirements of equilibrium and continuity across the interface. Since the axial force and velocity at the interface would be zero prior to stress wave transmission, the net force behind the upward propagating stress wave must equal the force of the transmitted wave to satisfy force equilibrium. Similarly, the net velocity behind the upward propagating reflected wave must equal the velocity of the transmitted wave to satisfy continuity. The force of the transmitted wave is equal to the velocity of the wave multiplied by the impedance of the rod through which the wave is travelling is, in this case, an A W J rod. Therefore, it is not surprising that using the A W J rod impedance to calculate the FF energy appears to give the most reasonable answer because the time for the upward propagating wave is small (0.2 ms) relative to the time period over which the FF energy is calculated. If the A W / A W J interface was the only major impedance interface in the rod string, it would be quite reasonable to use the FF method with the A W J rod area to calculate energy ratios. Figure 5.20 compares the force and velocity data recorded at 17.1 m (56') in LPT9902. The velocity has been multiplied by the impedance of the NWJ transducer rod to simplify the comparison. In this case the drill rods were also NWJ, though the N W J transducer rod was specifically manufactured for this investigation and may have had slightly different dimensions than the NWJ drill rods (it is difficult to accurately measure the inner diameter of " J " series rods because the threads are tapered). Figure 5.20 shows that -85-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program i . . . i . i i i -1 0 1 2 3 4 5 6 7 8 9 Time (ms) Figure 5.20 Average NALPT Force and Velocity Data Recorded at 17.1 m (56') in LPT9902. - 8 6 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program the measured peak force of roughly 65 kip was significantly higher than the equivalent peak product of velocity and NWJ impedance (roughly 40 kip). The author suspects that one of the primary causes of this discrepancy is the delayed reaction of the integrated velocity to major shifts in acceleration. The N A L P T data is a particularly difficult test case because the peak force is roughly three times higher than the SPT peak force but the width of the initial force "spike" is only 0.4 ms, compared to 0.7 ms for the SPT data. As the FF method generated unreasonable N A L P T energy ratios and the quality of the velocity data is lacking, it is not clear which measured energy should be used to correct the measured N A L P T blow counts. The HPA data appear to be consistent but do not account for energy losses during impact and stress wave transmission. It would be appropriate to use the velocity energy ratios i f it could be assumed that the dynamic efficiency (r)d) of the hammer and rod system, defined by Skempton (1986) as: are equivalent for the SPT and NALPT. This assumption would be based on the fact that safety hammers were used for both types of test. Figure 5.18 shows that both the SPT and N A L P T dynamic efficiencies are roughly equal to 0.82, though there is much more scatter in the N A L P T F V energy ratios. It appears that the most reasonable course of action is to use the N A L P T FV energy ratios to correct the blow counts to a standard energy. In this case, the N A L P T blow counts should only be compared to the FV energy corrected SPT blow counts to avoid systematic variations. 5.5 Calibration of Proposed Correlation Method The N A L P T equipment used during the investigation had slightly different properties than those listed for N A L P T in Table 4.1. Thus Equation 4.10 or 5.4 should be used to predict a revised SPT-NALPT correlation factor for the Kidd2 investigation (recall that ERr (5.6) -87-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program Equation 5.4 contains the quasi-static penetration resistance correction factor). The input parameters are: • ( E N T H R U ) N A L P T = 60% of 1109 J (820 ft-lb) = 665 J (492 ft-lb); • (ATE)SPT = 13.55 cm 2 (2.10 in 2); • (ENTHRU )SPT = 60% of 473 J (350 ft-lb) = 284 J (210 ft-lb); and, • (ATE)NALPT = 21.03 cm 2 (3.26 in 2). yielding a predicted correlation factor of 1.50 using Equation 4.10, greater than the value of 1.38 listed in Table 4.5. Equation 5.4 yields a correlation factor of 1.24. The revised N A L P T (R S ) value of 0.93 • 10"5 ft2/lb yields a correlation factor of 0.97, greater than the value of 0.93 listed in Table 4.5. Figure 5.21 compares the measured SPT (N6o) and N A L P T (N(,o), both of which have been corrected to a standard energy using the calculated FV energy ratios. Each data point represents the average SPT (Neo) and N A L P T (N^o) and the error bars represent the range of values recorded at a given depth. The observed correlation factor of 1.4 falls within the range of 1.24 to 1.50 predicted with the proposed correlation method but is 1.44 times the correlation factor predicted using the Winterkorn and Fang method. The correlation factor predicted using Equation 4.10 is slightly unconservative until the 1 2 quasi-static penetration resistance correction factor of —— = 0.82 is applied. For this 1.46 reason, the field determination of these correction factors should be undertaken for the JLPT, ILPT and any other LPT systems prior to predicting correlation factors using the proposed method. A "typical" value of 0.8 could likely be used in the absence of LPT specific correction factors. -88-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 5.0 Kidd2 NALPT Field Program 40 NALPT N60(FV) Figure 5.21 Comparison of F V Energy Corrected SPT and N A L P T Blow Counts Recorded at Kidd2. -89-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program 6. S E W A R D , A L A S K A N A L P T F I E L D P R O G R A M The United States Army Corp of Engineers (USACE) conducted an SPT, N A L P T and DCPT research program near Seward, Alaska in August and September of 1998. The author participated in the field-work and reporting of the project. The site was located on the flood plain shared by the Resurrection River and Mineral Creek. Ross et al. (1969) and McCulloch and Bonilla (1970) describe the effects of the March 27, 1964 magnitude 9.2 earthquake on three highway and three railway bridges which cross the Resurrection River. The two reports present evidence suggesting that the silty, sandy gravel deposits in the flood plain partially liquefied during the earthquake. This was considered unusual because SPT blow counts of 30 to 60 had been recorded in the deposits, which is unusually high for liquefiable deposits. It was postulated that the presence of gravel particles was responsible for the high SPT blow counts. The U S A C E research program was initiated to investigate these "grain size effects". A total of seven test holes or soundings were completed including: • SEWA9801 - SPT hole with energy measurement, located roughly 500m upstream of the main test site; • SEWA9802 - SPT hole with energy measurement; • SEWA9803 - N A L P T hole with energy measurement; • SEWA9804, SEWA9805 and SEWA9807 - DCPT holes; and, • SEWA9806 - N A L P T hole without energy measurement. The approximate relative locations of SEWA9802 through SEWA9807 are shown on Figure 6.1. SEWA9802 and SEWA9803 are the only test holes that can be used to determine an SPT-NALPT correlation factor. Test Hole SEWA9806 may be useful for checking the repeatability of the SEWA9803 blow counts, i f it is assumed that the energy ratios are unchanged between the two holes. -90-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program SEWA9806 0 (NALPT) SEWA9803 (NALPT) O • SEWA9802 (SPT) SEWA9804 SEWA9807 SEWA9805 (DCPT) (DCPT) (DCPT) 1—1—1—1 0 1.5 m Figure 6.1 Distribution of SPT, N A L P T and DCPT Test Holes at Seward, Alaska Main Test Site. - 9 1 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 6.0 Split Spoon Penetration Testing in Gravels Seward, Alaska NALPT Field Program Table C - l (Appendix C) summarizes the results of grain size analyses performed on samples recovered during the investigation. The soils encountered ranged from sandy silts to poorly graded gravel. 6.1 Drill ing Method The seven test holes were drilled by the hollow-stem auger method using a track-mounted Acker Soil Max rig supplied by the USACE. The auger casing had an inner diameter of 10.8 cm (4.25"). The opening at the base of the casing was sealed with a three-winged drag bit during drilling. When the hole had been advanced to the desired test depth, the casing was filled with water and the sand bit was carefully removed. It was usually necessary to continually pour water into the casing due to the high permeability of the Resurrection River deposits. SPT and N A L P T were generally performed at 1.5 m (5') vertical intervals beginning at 4.3 m (14'). 6.2 Dynamic Penetration Tests Dynamic SPT and N A L P T were performed with energy measurement by the FF and F V methods. HPA data were only recorded for six SPT and nine N A L P T hammer blows. 6.2.1 Description of Test Method Input energy for the SPT was provided by a 609 N (137 lb) safety hammer dropped 76.2 cm (30") for a maximum possible energy of 464.1 J (343 ft-lb). As in Section 5.0, it will be assumed that the maximum possible energy was the standard 475 J (350 ft-lb). Other than an " A W " transducer rod, the rod string consisted entirely of " A " rods to a maximum depth of 18.3 m (60'), beyond which " A W " rods were added to the string directly above the sampler. The split spoon dimensions were within A S T M standards. Plastic sample catchers were used and sample barrel liners were excluded during all tests. Input energy for the N A L P T was provided by a 1307 N (294 lb) safety hammer dropped 86.4 cm (34") for a maximum possible energy of 1129 J (833 ft-lb). As in Section 5.0, this is a significant variation from the N A L P T "standard" energy of 1017 J (750 ft-lb), -92-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program which will require revision of the estimated correlation factor. Other than an A W transducer rod, the rod string consisted entirely of 1.52 m (5') long N W J rods. The split spoon used had an outer diameter of 7.62 cm (3"), an inner diameter of 6.10 cm (2.4") in the open shoe and an inner diameter of 6.40 cm (2.52") in the sample barrel. Plastic sample catchers were used and sample barrel liners were excluded during all tests. Both the SPT and N A L P T hammers were lifted and dropped by the rope and cathead technique with two turns of a 2.86 cm (1.125") diameter rope around the cathead. The rope was loosened, not thrown off the cathead, to initiate hammer drops. Weather conditions during the investigation varied from sunny to hard rain. It was necessary to wet the section of the rope that was in contact with the cathead to reduce friction during dry weather. The SPT and N A L P T split spoons were driven 457 mm (18") and 610 mm (24"), respectively, into the soil at the base of the test holes. Blows per 2.54 cm (1") of penetration were recorded for both test types. Samples were classified in the field, bagged and returned to a U S A C E laboratory for gradation testing. 6.2.2 Energy Measurement The Klohn-Crippen Consultants Ltd. HPA was used to monitor a limited number of hammer drops during the Seward research program. Stress wave data were recorded for almost all hammer blows using a Northwood Instruments D E M belonging to BC Hydro. The BC Hydro D E M was very similar to the D E M used during the Kidd2 program, consisting of a 0.61 m (2') A W rod instrumented with four orthogonal pairs of strain gauges to measure force and a specially mounted high capacity accelerometer. The accelerometer was mounted roughly 12.7 mm (0.5') below the strain gauges. Calibration factors for the strain gauges and accelerometer were provided by Northwood Instruments. Both force and acceleration were sampled at 100 kHz, yielding a Nyquist frequency of 50 kHz. 4000 data points were collected on each channel at intervals of 0.01 ms, yielding 40 ms of data per hammer blow. Acceleration was digitally integrated to obtain velocity. The same D E M software used during the Kidd2 program was used during the Seward program. -93-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program 6.2.3 Test Results Table B-2, Appendix C, summarizes the HPA results from SEWA9802 (SPT) and SEWA9803 (NALPT). Peak SPT hammer velocities ranged from 2.56 mis to 2.93 m/s (8.4 ft/s to 9.6 ft/s). Peak N A L P T hammer velocities were higher due to the increased drop height and ranged from 3.04 m/s to 3.29 m/s (10 ft/s to 10.8 ft/s). The SPT and N A L P T velocity energy ratios ranged from 44% to 57% and from 55% to 64%, respectively. Tables C-3 and C-4, Appendix C, summarize the D E M results gathered during testing in SEWA9802 and SEWA9803. The force and velocity data were visually reviewed and erroneous data were identified using the same criteria used during the Kidd2 investigation and removed before the average FF and FV energies listed in Table C-3 were calculated. The SPT rod string consisted of an " A W " transducer rod, " A " and " A W " type rods with cross-sectional areas of 7.16, 7.74 and 7.61 cm 2 (1.11, 1.20 and 1.18 in 2). SPT FF energies calculated using the impedances of the " A " rods and the " A W " transducer rod are listed in Table C-3. The N W J rods in the N A L P T rod string had a cross-sectional area of 9.16 cm 2 (1.42 in 2). N A L P T FF energies calculated using the impedances of the " A W " transducer rod and the "NWJ" rods are listed in Table C-4. Raw blow counts recorded in SEWA9802 and SEWA9803 are summarized in Tables C-3 and C-4. SPT blow counts were corrected to 60% of the maximum SPT energy using the " A " rod and " A W " transducer rod FF energy ratios as well as the FV energy ratios. Similarly, blow counts were corrected to 60% of the maximum possible Seward N A L P T energy (60% of 1126 J or 833 ft-lb) using the " A W " and " N W J " rod FF energy ratios and the F V energy ratios. -94-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program 6.3 Discussion The author had the benefit of experience gained during the Seward research program when designing and conducting the Kidd2 research program. For example, Kidd2 was selected as the location of the program because the silty sand stratigraphy would allow the calculation of correlation factors using data that is free of "grain size effects". In addition, the number and distribution of test holes and several aspects of the energy measurement techniques employed at Kidd2 are based on apparent deficiencies of the Seward results. The effects of these perceived limitations on the Seward data set are discussed below. 6.3.1 Grain Size Analysis Results Performing SPT and N A L P T in gravel deposits is not a good way to develop a reliable SPT-NALPT correlation factor because of grain size effects. Comparison of the grain size distributions of the samples, however, may provide insight into the effect of gravel size particles on the measured blow counts. For example, i f it were shown that the grain size distributions of an SPT and an N A L P T sample obtained at the same depth were identical, it could be assumed that any differences between the two blow counts were strictly due to the different energies and the different areas upon which the soil resistance may act. In contrast, i f the grain size distributions differed beyond the limits of lateral variability, one could assume that blow count variations were due to grain size effects. Figures 6.2 and 6.3 compare the percent gravel and mean grain size (D50), respectively, of each SPT sample to those of the N A L P T samples acquired at the same depths. The best-fit relationships between the SPT and the two N A L P T data sets are essentially identical in both figures, suggesting that the grain size distributions at each depth were more dependent on the size of the sampler than on the lateral variability of the site. The fact that the slopes of the best-fit lines are less than one indicates that the N A L P T samples were generally coarser and contained more gravel. The coarsest gravel particles sampled during N A L P T were between 37.5 mm and 50 mm (1.5" and 2.0"), which would not - 9 5 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program Figure 6.2 Comparison of Percent Gravel in SPT and N A L P T Samples. -96-M.A.Sc. Thesis, Chris R. Daniel The University of Brit ish Columbia Split Spoon Penetration Testing in Gravels Sect ion 6.0 Seward, Alaska NALPT Field Program Figure 6.3 Comparison of Mean Grain Size (D 5 0) of SPT and N A L P T Samples. - 9 7 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program have been sampled by the SPT split spoon because the sampler opening is only 35 mm (1.375"). 6.3.2 Blow Count Repeatability One major limitation of the Seward data set is the lack of multiple SPT and N A L P T blow counts at each depth to allow assessment of lateral variability. The SPT data cannot be compared with confidence due to the large spacing between the two SPT test holes. The two N A L P T holes cannot be compared because SEWA9806 was completed without energy measurement. In order to provide some assessment of site variability, Figure 6.4 compares the uncorrected blow counts from the two N A L P T holes, as summarized in Table 6.1. The plot shows significant scatter between the uncorrected blow counts of the two N A L P T test holes. The slope of the best-fit relationship indicates that slightly higher blow counts were recorded in SEWA9803, for which D E M data was collected. Table 6.1 Comparison of Uncorrected N A L P T Blow Counts. Average Test Dep th m ( f t ) Seward N A L P T B l o w Count (N) S E W A 9 8 0 2 S E W A 9 8 0 6 4.3 30 16 5.8 14 19 7.3 20 12 9.0 29 43 10.3 38 32 11.9 50 53 13.4 32 8 15.0 11 9 16.5 23 21 18.0 44 47 19.5 31 31 21.1 32 27 22.6 30 26 24.1 26 21 -98-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program 60 0 10 20 30 40 50 60 SEWA9803 NALPT Blow Count (N) Figure 6.4 Comparison of Uncorrected N A L P T Blow Counts from SEWA9803 and SEWA9806. -99-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program 6.3.3 Energy Data Quality As discussed in Section 5.4, the principal methods of energy data quality control are the assessment of data repeatability, checking of calibration factors, comparison of results to upper bounds and assessment of force-velocity proportionality. An insufficient amount of data was collected to assess HPA repeatability. Figures 6.5 and 6.6 show SPT D E M data recorded at 18.1 m (59.3') in SEWA9802 and N A L P T D E M data recorded at 19.6 m (64.3') in SEWA9803. Similar to the Kidd2 data, the SPT and N A L P T force velocity are highly repeatable until — and reasonably repeatable c beyond. The SPT and N A L P T velocity data contain more scatter than the force data but the overall trends of the data are still repeatable. The calibration of the HPA system was checked once per test hole and was always determined to be acceptable. The calibration of the D E M force transducer was checked by Northwood Instruments prior to but not immediately after the Seward program. The accelerometer calibration factor was provided by the manufacturer and could not easily be checked independently. In terms of upper bounds, all of the SPT and N A L P T energy ratios were less than 100%. The average SPT rod energy ratios at test depths 13.4 m and 15.0 m (44.1' and 49.2') were low for a safety hammer, ranging from 42% to 47% (Table C-3). These unusually low values are supported, however, by the six SPT velocity energy ratios calculated at those depths, which ranged from 44% to 57%. Table 6.2 compares the average N A L P T rod energies to the available velocity energy ratios. The FF energy ratios calculated using the " A W " transducer rod area are clearly unreasonable compared to the velocity energy ratios. This supports the observation in Section 5.4.4 that it is incorrect to use the area of the transducer rod to calculate FF energy. The NWJ FF and FV energy ratios are slightly greater than and equal to the -100-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program -is q -20 — 10 15 Time (ms) 20 25 30 35 Figure 6.5 D E M Force and Velocity Data Collected During SPT at 18.1 m (59.3') in SEWA9802. -101 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program Time (ms) Figure 6.6 D E M Force and Velocity Data Collected During N A L P T at 19 6 (64.3') in SEWA9803. -102-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program velocity energy ratios, respectively. The blow counts at those depths ranged from 23 to 44, indicating that very little additional energy would have been derived from sampler set (Equation 5.5). The underlying cause of these borderline impossible results is most likely the resolution of the various energy measurement systems. For example, it was shown in Section 5.4.1 that the resolution of the HPA system was in the order of 5%. Table 6.2 Comparison of N A L 3 T Velocity and Rod Energy Ratios*. Test Average Velocity Energy Ratio (%) FF Energy Ratio f (%) FV Energy Ratio (%) Depth m(ff) " A W " Rod Area "NWJ" Rod Area 16.5 (54.1) 61 87.8 64.1 59.7 18.0 (59.2) 58 94.7 69.2 N . A . 19.6 (64.4) 59 88.3 64.5 59.1 N.A. Not Available N A L P T energy quoted as percent of maximum Seward N A L P T energy 1126 J (833 ft-lb). K i and K 2 factors applied to FF energies Figures 6.7 and 6.8 compare the average D E M output recorded during the SPT at 18.1 m (59.3') in SEWA9802 and the N A L P T at 19.6 m (64.3') in SEWA9803, respectively. The velocity has been multiplied by the impedance of the " A W " transducer rod in both figures to ease comparison. The cross-sectional area of the " A " type rods used for the SPT is greater than the area of the " A W " transducer rod. Based on the reasoning presented in Section 5.4.4, one would expect that the force would be greater than the velocity for most of the time period from 2L impact to — . Substituting the " A " and " A W " rod impedances into Equations A.5 and c A.6 (Appendix A) indicates that the total force-velocity deviation should be in the order of 8% of the original force magnitude (force increases, velocity decreases). Assuming an original force of roughly 20 kip, the first deviation should be roughly 1.6 kip. In fact, the force in Figure 6.8 is equal to or slightly less than the velocity for the majority of the -103-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program Figure 6.7 Average SPT Force and Velocity Data Recorded at 18.1 m (59.3') in SEWA9802. -104-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program Figure 6.8 Average N A L P T Force and Velocity Data Recorded at 19.6 m (64.3') in SEWA9803. -105-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program trace. This discrepancy may be due to incorrect assumptions, measurement error or to a problem with the resolution of the instrument. The cross-sectional area of the "NWJ" type rods used for the N A L P T was greater than the area of the " A W " transducer rod. Substituting the "NWJ" and " A W " rod impedances into Equations A.5 and A.6 (Appendix A) indicates that the force and velocity should separate by roughly 24% of the original force magnitude. Assuming an original force of roughly 25 kip, the first separation should be in the order of 6 kip (force increase, velocity decrease). In fact, the force is greater than the velocity by 5 to 20 kips in the early part of the trace. Thus, the predicted and observed deviations are in reasonable agreement. Both the force and velocity in Figure 6.8 contain a high frequency sinusoidal signal with a period of 0.15 to 0.20 ms. The precise distance between the end couplings on the transducer rod is 483 mm (19") and the two way travel time of a reflection within the transducer rod would be 0.19 ms. Thus the high frequency content of the Seward N A L P T D E M data likely results from using a relatively low impedance transducer rod in an otherwise high impedance rod string. 6.4 Correlation Factor The N A L P T equipment used during this investigation had slightly different properties than those listed for N A L P T in Table 4.1 and from those of the Kidd2 N A L P T . Thus Equation 4.10 or 5.4 should be use to predict a revised SPT-NALPT correlation factor for the Seward investigation. Equation 5.4 may be used because the Seward and Kidd2 N A L P T split-spoons had the same dimensions. The input parameters for the formulae are: • ( E N T H R U ) N A L P T = 60% of 1126 J (833 ft-lb) = 676 J (500 ft-lb); • (ATE)SPT = 13.55 cm 2 (2.10 in 2); • (ENTHRU )SPT = 60% of 473 J (350 ft-lb) = 284 J (210 ft-lb); and, -106-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program • (ATE)NALPT = 21.03 cm 2 (3.26 in 2). yielding a predicted correlation factor of 1.53 using Equation 4.10, greater than the value of 1.38 listed in Table 4.5 and the revised factor of 1.50 predicted for the Kidd2 NALPT. Equation 5.4 yields a correlation factor of 1.26. The revised Seward N A L P T (R s) value 5 2 of 0.916 • 10" ft /lb yields a Winterkorn and Fang correlation factor of 0.98, greater than the value of 0.93 listed in Table 4.5 and the revised factor of 0.97 predicted for the Kidd2 NALPT. Figure 6.9 compares the SPT and N A L P T (N6o), both of which were corrected to 60% of the maximum possible energy using the FF energy ratios. The " A " and "NWJ" impedances were used to correct the SPT and NALPT, respectively. Figure 6.10 compares the SPT and N A L P T (N 6 0 ), both of which were energy corrected using the FV energy ratios. Each data point in the two figures represents a single test depth at which both SPT and N A L P T energy corrected blow counts were recorded. The observed correlation factor of both figures is 1.06, which is below the range of 1.26 to 1.53 predicted with the proposed correlation method but is in good agreement with the Winterkorn and Fang prediction of 0.98. Note that the range of recorded correlation factors varied from roughly 0.67 to 1.50 and from 0.74 to 1.73 for the FF and F V energy corrected blow counts, respectively. It will be shown in Section 8.0 that these variations correlate reasonably well with grain size. -107-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program NALPT N 6 0 (FF, NWJ Rods) Figure 6 . 9 Comparison of F F Energy Corrected SPT and N A L P T Blow Counts Recorded at Seward, Alaska Main Test Site. -108-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 6.0 Seward, Alaska NALPT Field Program N A L P T N 6 0 ( F V ) Figure 6.10 Comparison of F V Energy Corrected SPT and N A L P T Blow Counts Recorded at Seward, Alaska Main Test Site. -109-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program 7. K E E N L E Y S I D E D A M N A L P T F I E L D P R O G R A M An N A L P T investigation was conducted on April 19 and 20, 1999 at Keenleyside Dam under the supervision of the author and Dr. John Howie of the University of British Columbia Civil Engineering Department. The program was conducted using equipment and staff mobilized by BC Hydro for an extensive investigation of the dam foundation. Keenleyside Dam is located on the Columbia River near Castlegar, B.C. The dam consists of an earthfill embankment about 430 m long and three gravity structures with a length of about 370 m. The earthfill dam (Figure 7.1) is a zoned fill embankment with an upstream sloping impervious core and a downstream pervious sand and gravel shell. Sand and gravel portions of the earthfill dam below the original river level were constructed by bottom dumping from barges and end dumping from trucks to form an embankment just above river level. Recent modifications to the seismic design parameters for the area, combined with concerns regarding the bottom and end dumping construction methods, prompted B C Hydro to conduct extensive field investigations of the earthfill portions of the dam. One such investigation is described by Lum and Yan (1994). The main purpose of these investigations has been to quantify the potential for liquefaction of the foundation materials including native soils and fill material. There is concern that liquefaction of the dam foundation could initiate a large-scale failure. Figure 7.2 shows the grain size envelope for the sand and gravel fill material. Mean grain sizes ranged from roughly 7 mm to 50 mm and samples typically contained between 60% and 85% gravel. Liquefaction susceptibility is typically quantified using SPT blow counts but the grain size distribution curves suggest that the material is too coarse for the SPT. For this reason, BC Hydro has collected not only SPT but also Becker Penetration Test (BPT) and shear wave velocity data at Keenleyside Dam. Equivalent SPT blow -110-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program Figure 7.1 Plan View of Keenleyside Dam (Lum and Yan, 1994). M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program 24- !?• 1- IV,- V M "10 "20 "40 «60' I00 «?00 00 -90 ' \ \ w 60 -<?A w, - E N V E L O P E F O R ADDOny I U 1 T F I Y 70 -\ 7o; '. O F 0/ T A 60 -fa fa SO -fa ft iO -fa VA fa 30 * fa w fa, 20 - m 10 • '/A 0 -100 0 1' 100 M i l 1 1 | l l 1 1 1 1 1 1 10 111 i i i i r i 111 i i i i i i 1.0 0.1 .( )I ASTM 0«2 GRAIN SIZE - mm B0U10ERS COBBLE GRAVEL SIZE SANO SIZE SILT SIZE | CLAY SIZE SIZE SIZE COARSE | FINE :0ARSE| UEOIUM | FINE FINE GRAINED Figure 7.2 Grain Size Envelope for Keenleyside Dam Sand and Gravel Fill Material (Lum and Yan, 1994). -112-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program counts were estimated from BPT data using the Harder and Seed (1986) and Sy and Campanella (1993) methods. One N A L P T test hole (DH99-20) was completed in the vicinity of test holes DH91-C (BPT) and DH91-D (SPT), as shown on Figure 7.1. 7.1 Drilling Method The test hole was drilled by the mud rotary method using a truck-mounted AP-1000 drill rig. A 12.1 cm (4.75") outer diameter tricone bit was used to drill to each test depth. The level of the drill mud was kept above the groundwater table to improve hole stability. Several mud loss events occurred between 8.0 m and 8.5 m (26' and 28') and it was necessary to install casing to 8.5 m (28'). No additional mud loss occurred below this depth but it was necessary to re-drill the hole from 11.00 m to 20.95 m (36' to 68.7') due to "squeezing" of the drill string. 7.2 Dynamic Penetration Tests Dynamic N A L P T were performed with velocity and rod energy ratio measurements. 7.2.1 Description of Test Meth od The same split-spoon sampler and hammer used for the Kidd2 investigation were used at Keenleyside. The safety hammer was lifted using a winch with a clutch-release. The rod string consisted of 1.52 and 3.05 m (5' and 10') N W rods. N A L P T were performed at 1.52 m (5') vertical spacing between 3.00 m and 14.90 m (10' and 49') and at 18.00 m and 21.00 m (59' and 69'). Blow counts per 2.5 cm (1") penetration and sample descriptions were recorded by BC Hydro personnel (Log included in Appendix D). The Klohn-Crippen Consultants Ltd. HPA system was used to monitor all hammer drops below 7.13 m (23.4'). Stress wave data were recorded for all hammer blows except at 7.13 m and 21.0 m (23.4' and 68.8') using the BC Hydro D E M . Both force and acceleration were sampled at 100 kHz. - 1 1 3 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program 7.2.2 Test Results Table D - l , Appendix D, summarizes the HPA results from DH99-20. Average peak hammer velocities ranged from 3.08 to 3.38 m/s (10.1 to 11.1 ft/s). Standard deviations ranged from 0.15 to 0.21 m/s (0.5 to 0.7 ft/s). Average values and standard deviations of the velocity energy ratios ranged from 63.9% to 76.5% and from 6.2% to 9.0% of the maximum Keenleyside N A L P T energy of 1108 J (820 ft-lb), respectively. Table D-2, Appendix D, summarizes the D E M results gathered during testing in DH99-20. The force and velocity data were visually reviewed and erroneous data were removed before the average FF and FV energies listed in Table D-2 were calculated. The most common problem with the data was minor accelerometer baseline offsets, resulting in uniform non-zero slopes late in the traces. The " A W " transducer rod and " N W " rods had cross-sectional areas of 7.16 cm 2 and 9.16 cm 2 (1.11 in 2 and 1.42 in 2), respectively. N A L P T FF energy ratios calculated using the impedances of the " N W " rods and the " A W " transducer rod are listed in table D-2. K i and K 2 energy correction factors were interpolated from those tabulated in A S T M (1991b) and applied to the FF energy corrected blow counts. Raw N A L P T blow counts recorded in DH99-20 are summarized in Table D-2. Blow counts corrected to 60% of the maximum possible Keenleyside N A L P T energy (60% of 1108 J, 820 ft-lb) using the " N W " and " A W " rod FF energy ratios and the FV energy ratios are also listed. Table D-3 summarizes BPT and SPT data recorded in test holes DH91-3C and DH91-3D. The FV method was used to correct the SPT blow counts to the standard energy. The BPT blow counts have been converted to equivalent energy corrected SPT blow counts using the methods proposed by Harder and Seed (1986) and Sy and Campanella (1993). -114-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program 7.3 Discussion of Energy Data The quality of the energy data may be assessed by checking repeatability, upper bounds and force-velocity proportionality. HPA calibration factors were checked prior to each test and D E M calibration factors were provided by Northwood Instruments. The standard deviation of the HPA velocity energy ratios ranged from 6.2% to 9.0%. This range is higher than the range of 0% to 5.0% recorded for N A L P T during the Kidd2 investigation, indicating that the hammer fall velocity was more variable using the clutch-release winch at Keenleyside than the rope and cathead at Kidd2. This result was anticipated because the driller had considerable difficulty achieving the correct drop height using the clutch-release system and the hammer was often not falling vertically. Figure 7.3 shows N A L P T D E M data recorded during hammer blows recorded at 18.30 m 2L (60') depth in DH99-20. The force and velocity data are highly repeatable prior to — c and the velocity data is uncharacteristically repeatable later in the trace. Figure 7.4 compares the average force and velocity traces recorded during the N A L P T at 18.30 m (60') in DH99-20. The velocity has been multiplied by the impedance of the " A W " transducer rod to ease comparison. It is useful at this point to review the D E M energy data from the Kidd2 and Seward programs: • Kidd2 SPT: 7.16 cm 2 (1.11 in 2) " A W " transducer rod with 4.84 cm 2 (0.75 in2) " A W J " rods. Referring to Figure 5.19, the product of the measured velocity and the impedance of the transducer rod was generally higher than the corresponding force. This was attributed to upward propagating tensile reflections from the " A W " / " A W J " interface. The peak force was roughly 20 kips. The FF energy ratios calculated using the impedance of the " A W J " rods were in good agreement with the F V energy ratios. • Kidd2 N A L P T : 9.16 cm 2 (1.42 in2) "NWJ" transducer rod and rods. Referring to Figure 5.20, the product of the measured velocity and the impedance of the transducer rod was -115-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 7.0 Split Spoon Penetration Testing in Gravels Keenleyside Dam NALPT Field Program Figure 7.3 D E M Force and Velocity Data Collected During N A L P T at 18.3 m (60') in DH99-20. -116-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program Figure 7.4 Average N A L P T Force and Velocity Data Recorded at 18.3 m (60') in DH99-20. M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program generally lower than the corresponding force, particularly within the first force peak. This was attributed to the much higher peak force of roughly 65 kips, which would create a severe loading situation for the accelerometer. A l l calculated FF energy ratios were greater than 100%. The dynamic efficiency calculated using the F V energy ratios was equivalent to the factor calculated using the Kidd2 SPT F V energy ratios data (0.82). • Seward SPT: 7.16 cm 2 (1.11 in2) " A W " transducer rod with 7.74 cm 2 (1.2 in 2) " A " rods. Referring to Figure 6.7, the product of the measured velocity and the impedance of the transducer rod was generally equivalent to the corresponding force. This was attributed to the relatively minor change in area at the " A W " / " A " interface. The peak force was roughly 20 kips. The FF energy ratios calculated using the impedance of the " A " rods were in good agreement with the FV energy ratios. • Seward NALPT: 7.16 cm 2 (1.11 in 2) " A W " transducer rod with 9.81 cm 2 (1.52 in2) " A " rods. Referring to Figure 6.8, the product of the measured velocity and the impedance of the transducer rod was generally lower than the corresponding force. This was attributed to upward propagating compressive reflections from the " A W " / " N W J " interface. The peak force was roughly 30 kips. The FF energy ratios calculated using the impedance of the "NWJ" rods were in good agreement with the F V energy ratios. The theory supporting the use of the rod impedance instead of the transducer rod impedance to calculate FF energy ratios was presented in Section 5.0. The above review indicates that this practice produces FF energy ratios that are in good agreement with the calculated F V energy ratios. The Keenleyside N A L P T D E M data was collected using a 7.16 cm 2 (1.11 in 2) " A W " transducer rod with 9.16 cm 2 (1.42 in2) " N W " rods. Based on the trends observed at Kidd2 and Seward, one would expect that the magnitude of the Keenleyside force measurements would generally be greater than the product of the measured velocity and the impedance of the transducer rod. In contrast, review of Figure 7.4 indicates that the two values are roughly equivalent over the majority of the trace. Accordingly, a review of Table D-2 reveals that the measured F V energies are in fair -118-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program agreement with the FF energy ratios calculated using the " A W " transducer rod impedance and are higher than those calculated using the rod impedance. One explanation for this unexplained trend reversal is that the velocity measurements are in error, possibly due to an incorrect calibration factor (the possible error is likely not a result of limited capacity because the peak force was only 35 kips). The velocity energy ratios calculated using the HPA data can be used to provide an additional assessment of the quality of the D E M data. Figure 7.5 compares the calculated rod energy ratios (Ki and K.2 factors applied to FF energies) to the measured velocity energy ratio corrected using Equation 5.5. The dynamic efficiencies are 0.97, 0.89 and 0.70 for the FV, FF (NW) and FF (AW) energy ratios, respectively. The dynamic efficiency calculated using the FV energy ratio is unusually high and supports the suggestion that the D E M velocity data is in error. As a result, the range of FF energy ratios is likely the only reliable measurement of energy for the Keenleyside data set. 7.4 Correlation Factor SPT-NALPT correlation factors calculated using actual SPT blow counts and equivalent SPT blow counts predicted from BPT data are presented below. The SPT-NALPT correlation factors predicted for the equipment used at Keenleyside Dam were 1.24 and 1.50, with and without the empirical quasi-static resistance correction, respectively. The correlation factor predicted using the Winterkorn and Fang method prediction was 0.97. Figures 7.6 and 7.7 compare the FF (AW) and FF (NW) energy corrected N A L P T (N6o) to the 1991 SPT data, respectively. Each data point compares SPT and N A L P T blow counts recorded at the same depth. The FF (AW) correlation factor (1.07) is considerably lower than the range predicted using the proposed method but is in good agreement with the Winterkorn and Fang prediction. The FF (NW) correlation factor (1.44) is in good agreement with the range predicted using the proposed method but is considerably higher than the Winterkorn and Fang prediction. It should be noted, however, that Figures 7.6 -119-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program 100 n Corrected Velocity Energy Ratio (%) Figure 7.5 Comparison of D E M and "Corrected" HPA Energy Data. -120-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program Figure 7.6 Comparison of F V Energy Corrected SPT and F F (AW) Energy Corrected N A L P T Blow Counts Recorded at Keenleyside -121 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program (^WNALPT Figure 7.7 Comparison of F V Energy Corrected SPT and F F (NW) Energy Corrected N A L P T Blow Counts Recorded at Keenleyside Dam. -122-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program and 7.7 both show a large range of possible correlation factors. Because the results are not very conclusive, it is likely safest to assumed that the observed SPT-NALPT correlation factor lies between 1.07 and 1.44. Figures 7.8 and 7.9 compare the FF (AW) and FF (NW) energy corrected N A L P T (N6o) to Equivalent SPT blow counts predicted using the Harder and Seed (1986) and Sy and Campanella (1993) methods. The slopes of the relationships range from 0.47 to 0.81, much lower than the slopes of the SPT-NALPT relationships. This indicates that the Equivalent SPT blow counts predicted using the BPT data are generally lower than the actual SPT blow counts at the same depths. It is remarkable that the range of average correlation factors in Figures 7.8 and 7.9 does not overlap the range of average correlation factors from Figures 7.6 and 7.7, considering how broad the two ranges are. -123-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program Figure 7 .8 Comparison of Equivalent SPT Blow Counts from B P T Data and FF (AW) Energy Corrected N A L P T Blow Counts Recorded at Keenleyside Dam. -124-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 7.0 Keenleyside Dam NALPT Field Program 4 0 3 5 3 0 2 5 § 2 0 1 5 1 0 0 H • Harder and Seed Method ( N 6 O ) S P T = 0 - 6 4 ( N 6 0 ) N A L P T O Sy and Campanella Method ( N 6 0 ) S P T = 0 .81 ( N 6 0 ) N A L P T o / >r / • y^ O o y o / • ° / • o 1 0 1 5 2 0 ( ^ 6 O ) N A L P T 2 5 3 0 3 5 4 0 Figure 7.9 Comparison of Equivalent SPT Blow Counts from BPT Data and F F (NW) Energy Corrected N A L P T Blow Counts Recorded at Keenleyside Dam. -125-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion 8. DISCUSSION The following sections summarize and discuss the results of the three research programs described herein and address the larger issues of grain size effects and appropriate use of SPT-LPT correlations. 8.1 Standardization of N A L P T Results It is difficult to compare the results of the various N A L P T research programs because the hammer energies were not consistent. There may also have been secondary effects due to the varying shapes of the downward propagating stress waves. It was demonstrated in Section 4.2, however, that the former are much more significant than the latter. Thus, "standardization" of the test results may be achieved using an inverse proportionality relationship to transform the actual results to those that would have been gathered i f identical equipment had been used at the three sites. The relationship between the actual results and the standardized results (indicated by a prime (')) is as follows: 1 6 0 , N A l P T [ ) n a l p t ' 60% • PE ' 60% • PE ' ( 8 - 1 } or NALPT NALPT 77 , (8.2) where (PE) and (PE') are the actual and selected standard maximum potential energies, respectively. Similarly, the relationship between the actual and standardized correlation factors is as follows: -126-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Section 8.0 Split Spoon Penetration Testing in Gravels Discussion ( • ^ 6 0 )sPT _ ( - ^ 6 0 )sPT i^60)NALPT £g )^ NALPT NALPT NALPT or fa 60 )'NALPT (^60 )NALPT ^E Table 8.1 summarizes the results of applying Equation 8.4 to the observed SPT-NALPT correlation factors using a standard energy (PE') of 1014 J (750 ft-lb), equivalent to dropping a 1335 N (300 lb) hammer 0.76 m (30"). Table 8.1 Summary of Observed and Standardized SPT-NALPT Correlation Factors. Research Program NALPT. Potent J ' Maximum lal Energy, (ft-lb) Energy Measurement Method (Rod Type) ( - ^ 6 0 )sPT ( • ^ 6 0 )NALPT Actual, PE Standard, PE ' SPT N A L P T Actual Standard Kidd2 1109 (820) 1014 (750) FV FV 1.4 1.28 Seward, Alaska 1126 (833) 1014 (750) FF (A) FF (NWJ) 1.06 0.95 FV FV 1.06 0.95 Keenleyside Dam 1109 (820) 1014 (750) FV FF (AW) 1.07 0.98 FV FF (NW) 1.44 1.32 8.2 Performance of Proposed Correlation Method The performance of the proposed correlation method was originally summarized in Table 4.5. Table 8.2 presents a revised summary that includes the standardized SPT-NALPT correlation factors. -127-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion Table 8.2 Revised Summary of Observed and Predicted Correlation Factors Test Type Observed Correlation Factors Winterkorn and Fang Equation 4.9 * Equation 4.10 * Format Material Value {N)SP7. (N)LPT C^60 )sPT (NER ) L P R C^60 )sPT i^ER )LPT C^60 )sPT {^60 )LPT Sand 1.28 N A L P T Sand and Gravel 0.95 to 1.32 0.93 1.23 1.13 JLPT {N)SPT Sand 1.5 1.02 1.81 2.06 (N)LPT Gravel 2.0 Sand 1.14 ILPT [N\(60)\SPT Sand and Gravel 0.89 0.44 1.05 0.95 l^l(60) \ L P T Sand and 1.02 Gravel predicted correlation factors. applied to SPT-NALPT The sand correlation factors should be used to check the correlation methods in order to avoid grain size effects. In all cases, the correlation factors predicted using the Winterkorn and Fang method are more conservative than those predicted using the proposed correlation method. The SPT-NALPT and SPT-ILPT correlation factors predicted using the proposed correlation method ranged from 83% to 96% of the observed values, while the predictions of the Winterkorn and Fang method range from 39% to 73%. The JLPT correlation factors predicted using the proposed correlation method range from 121% to 137% of the observed correlation factor, while the prediction of the Winterkorn and Fang method is 68% of the actual value. The SPT-NALPT and SPT-ILPT Winterkorn and Fang correlation factors are overly conservative compared to those of the proposed correlation method. The SPT-JLPT correlation factors predicted using the proposed method are unconservative and in poor agreement with the observed sand correlation factor. The Winterkorn and Fang correlation factor is in equally poor agreement but is, at least, conservative. It is possible that the SPT-JLPT correlation requires an empirical quasi-static resistance correction, -128 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion particularly because the JLPT sampler has a relatively thick wall compared to the N A L P T and ILPT. It should also be noted, however, that there is considerable uncertainty associated with the observed SPT-JLPT correlation factor because neither the velocity nor rod energy ratios were recorded. A review of the available grain size data in Section 8.3 casts additional doubt on the SPT-JLPT correlation factor. 8.3 Grain Size Effects The following quotes from the literature are indicative of the current perception of grain size effects: Ross, Seed and Migliaccio (1969): "... the presence of gravel sizes in a soil can yield high penetration resistance to small sampling tools even if the relative density of the soil mass is low...." U.S. National Research Council (1985): "It is not possible to evaluate the liquefaction susceptibility of [soil containing gravel] using the SPT; the presence of gravel can increase greatly the penetration resistance..." Rollins, Diehl and Weaver (1997): "... the [SPT] penetration resistance may be artificially increased due to the size of the gravel particles and because gravel particles may plug the sampler....' In addition, several researchers recommend the use of "small increment" blow counts, wherein the blows per inch of sampler penetration are recorded and the lowest recorded value is multiplied by 12 to obtain a revised blow count. This procedure implies that SPT blow counts are significantly higher in gravels, presumably due to interaction between very coarse particles and the sampler. While there is considerable evidence indicating that SPT blow counts tend to increase with average particle size in sands (Decourt, 1989), and SPT commonly meet with refusal on very large particles such as boulders, there is very little data in the literature supporting the assertion that SPT blow counts steadily -129-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion increase between coarse sand and boulder size particles. In fact, some researchers who have performed SPT in gravel deposits have noted the opposite effect: Andrus and Youd (1989): "...the low blow count in the loose gravel at both sites suggest a lack of influence of gravel particles and could not have been increased much due to gravel content...." in which case the SPT blow counts may be representative or even lower than justified in coarse-grained soils. Part of the difficulty of assessing "grain size effects", however, is that it is impossible to determine what the SPT blow count in a gravel deposit "should" be. For example, Burland and Burbridge (1985) conclude that: "... an analysis of [settlement records from over 200 sites including 18 gravel sites] indicates that the SPT blow count should be increased by a factor of about 1.25 for the purpose of assessing [gravel and sandy gravel] compressibility..." It is not possible to say whether the recommended increase of blow counts results from a general decrease in compressibility with increasing grain size, a decrease in SPT blow counts with increasing grain size or a combination of the two. For this reason, soil compressibility is not a suitable performance reference for quantifying grain size effects. Comparison of SPT data to that of other in-situ tests is one way of providing an arbitrary performance reference. The ideal in-situ test for comparison would be sensitive to the strength and deformation parameters of the gravel but insensitive to grain size variations. Shear wave velocity measurements are derived from relatively large volumes of soil and thus are likely less sensitive to grain size variations. Comparison of small strain parameters such as shear wave velocity to large strain test results such as SPT blow counts is considered questionable, however, due to the non-linear nature of soil deformation. The principal attraction of LPT data as a performance reference is the similarity of SPT and LPT equipment. As shown in previous sections, the data in SPT-LPT comparison plots are often well represented by a simple linear best fit. -130-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion Unfortunately, observed changes in SPT-LPT correlation factors may be due to grain size effects on the LPT as well as SPT blow counts. It is reasonable to assume that grain size effects on an LPT sampler would be of a similar nature to those acting on an SPT sampler (i.e. both blow counts would be higher or lower than justified). For example, an increase in the observed correlation factor with increasing grain size most likely represents an increase in SPT blow counts, the severity of which may be partially masked by a similar increase in LPT blow counts. The LPT blow counts selected as the arbitrary performance reference should ideally be free of grain size effects. Some measurement of grain size such as the mean grain size or the portion of the grains that are too large to enter the sampler can be used to gauge the severity of the anticipated grain size effects and to determine what size of sampler is required to avoid grain size effects. Table 8.3 summarizes the available grain size information for each of the sites described in the literature and herein. Table 8.3 Summary of Available Grain Size Information Test Type Research Program Source of Sample Material Type D 5 0 (mm) % Coarser than Sampler Inner Diameter SPT LPT N A L P T Kidd2 N.A. Sand 0.14 to 0.33 0 0 Seward, Alaska N A L P T Sand and Gravel 0.9 to 11 Oto 24 N . A . Keenleyside Dam N.A. Sand and Gravel 7 to 50 18 to 57 5 to 45 JLPT Calibration Chamber Grab Samples Sand 0.34 0 0 Gravel* 1.13 to 1.28 0 0 ILPT Messina, Italy ILPT Sand 0.2 to 0.6 0 N . A . Sand and Gravel 1 to 15 Oto 30 N . A . N .A. Not Available * This material is not classified as gravel using the USC or BSC systems. Based on the information in Table 8.3, one would expect that the grain size effect on the SPT blow counts recorded in sand and gravel at the Seward, Alaska, Keenleyside Dam -131 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion and Messina, Italy sites would be more significant than those recorded in gravel during the JLPT calibration chamber tests. In fact, the "gravel" material used to develop the JLPT correlation factor would not be classified as gravel using the Unified Soil Classification (USC) or British Soil Classification (BSC) Systems. Table 8.2 shows that the SPT-NALPT and SPT-ILPT correlation factors decreased with increasing grain size, supporting the hypothesis that SPT blow counts may actually decrease with increasing grain size in some portion of the gravel range of soil particles. It is possible that the increase in observed SPT-JLPT correlation factor with grain size may have been due to the well documented increase in SPT blow counts with increasing average particle size in sands. One major limitation of using correlation factors to quantify grain size effects is the limited accuracy of the correlation factors. For example, the Seward, Alaska and Keenleyside Dam correlation plots display considerable scatter. When performing tests in natural deposits, one must anticipate variations of grain size with depth. Crova et al. (1993) present an interesting plot comparing the observed SPT-ILPT correlation factor to the mean grain size of the sample retrieved (Figure 8.1). The data appears to support a slight decrease in correlation factor with increasing mean grain size, though there is considerable scatter. The U S A C E performed grain size analyses of many of the samples retrieved during the Seward, Alaska research program. Figure 8.2 compares the SPT-NALPT correlation factor recorded at each test depth to the range of mean grain sizes determined using the two N A L P T samples collected at each depth. Similarly, Figure 8.3 compares the correlation factors to the portion of the N A L P T samples that would be too large to enter the SPT sampler. Unlike the data shown in Figure 8.1, there is no clear relationship between correlation factor and mean grain size in Figure 8.2. The data in Figure 8.3 show considerable scatter but the average values of the "oversize" portion of the N A L P T sample correlates reasonably well (r = 0.77) with the observed correlation factors. -132-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion f N H « 0 ) ] S P T 2.0 [ N l ( 6 0 ) l L P r 1.6 1.2 0.8 0.4 "1 < i i i i m | — I I I 11 I I I 1—i | | | in : Hi f 1 S T D Moan J — 1 I I I I I I ! 0.1 1 I I I M I U * I I I M i l l 10 100 dso (mm) Figure 8.1 Comparison of SPT-ILPT Correlation Factors and Mean Grain Size Data from Messina, Italy (Crova et al., 1993). -133-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion 0.1 < co LU o • — i 10 Q) E ro b CD c _c f-Q . CO 100 D 5 0 (mm) Figure 8.2 Comparison of SPT-NALPT Correlation Factors and Mean Grain Size Data from Seward, Alaska Research Program. -134-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion 2.0 0.0 - I — i — . — . — p — , — . — . — . — . — , — , — • — i — • — , — P — , — , — , — , — , — , — , — , — I 0 5 10 15 20 25 Portion of NALPT Sample Coarser than S P T Inner Diameter (%, by weight) Figure 8.3 Comparison of SPT-NALPT Correlation Factors to "Oversized" Portion of N A L P T Samples from Seward, Alaska Research Program. -135 M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion Based on these results, measurement of the oversize portion of the in-situ soil may be a better gauge of grain size effects. No grain size analyses were performed on the samples retrieved at Keenleyside Dam. The samples collected above and below 14.6 m were identified as gravelly sand and sand, respectively. Figure 8.4 indicates which data points on the Keenleyside Dam correlation plot were identified as sand. The fact that the two data points fall on different sides of the best-fit line indicates that there is no systematic variation of the correlation factor with grain size, though the available data is very limited. With regard to the use of "small increment" blow counts, Figures 8.5 and 8.6 show penetration rate data for the SPT and N A L P T samplers, respectively. The gravel curves are no less regular than the sand curves, and it would be very difficult to determine which portions of the curve corresponded to sand versus gravel penetration. For this reason, the use of "small increment" blow counts appears to be unwarranted, at least at the sites investigated herein. 8.4 Use of SPT-LPT Correlations SPT-LPT correlation factors developed in gravel generally differ than those developed in sands, presumably due to grain size effects. As a result, the only reasonable application of gravel correlation factors, other than studying grain size effects as described in Section 8.3, is estimating grain size affected SPT blow counts without damaging SPT equipment on coarse gravel particles. SPT-LPT correlation factors developed in sand are free of grain size effects. The SPT blow counts estimated using sand correlation factors and LPT data from gravel deposits are the blow counts that would have been recorded in a sand deposit i f the LPT blow counts had been recorded in sand. Whether or not it is appropriate to use the estimated SPT blow counts as input for empirical design methods that were originally developed for sands can only be checked by developing soil performance databases. It is quite -136-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion 50 0 5 10 15 20 25 30 35 40 45 50 ( ^ 6 O ) N A L P T Figure 8.4 Keenleyside Dam SPT-NALPT Correlation Data (FV Energy Corrected), Sand Versus Gravel Data. - 1 3 7 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion 100 9 12 Penetration (in.) 15 18 21 Figure 8.5 Dynamic SPT Blows versus Penetration Data from Kidd2 and Seward, Alaska Sites. -138-M.A.Sc. Thesis, Chris R. Daniel The University of Bri t ish Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion 100 0 3 6 9 12 15 18 21 24 27 Penetration (in.) Figure 8.6 Dynamic N A L P T Blows versus Penetration Data from Kidd2 and Keenleyside Dam Sites. - 1 3 9 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 8.0 Discussion likely that existing empirical design methods for sands will have to be modified slightly, just as many sand correlations have been modified for use in non-plastic silts. When sufficient LPT data is available, it will no longer be necessary to use SPT-LPT correlations, as the desired gravel parameters may be estimated directly from LPT blow counts. -140-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 9.0 Conclusion and Recommendations for Future Research 9. C O N C L U S I O N AND R E C O M M E N D A T I O N S F O R F U T U R E R E S E A R C H A method for predicting SPT blow counts from LPT blow counts was presented and checked against an earlier correlation method presented by Winterkorn and Fang (1975) and against actual correlation factors from the literature and from research programs described herein. The method accounts for the surface area upon which soil resistance may act and the actual input energy of the test as opposed to the maximum potential energy of the hammer. In this regard, the proposed correlation method is better aligned with current geotechnical practice than the Winterkorn and Fang method. The method used to predict quasi-static penetration resistance was found to generate values that were too low for both the SPT and N A L P T samplers. A simple empirical correction factor of 0.82 was required to correct this error. Application of the proposed correlation method to samplers other than the N A L P T sampler may require additional quasi-static penetration testing to develop correction factors for other samplers. Accurate measurement of the rod energy ratio is very important for any application of dynamic penetration test results, including the proposed correlation method. The author observed that the correlation factors generated during the three research programs described herein were sensitive to the method of dynamic energy measurement (i.e. FF versus FV), to the assumptions required for the method (e.g. what rod area is used for FF energy calculation) and to the quality of the data. The author employed assessment of data repeatability, checking of calibration factors, upper bounds and force-velocity proportionality as quality control methods. The dynamic energy data were highly repeatable but the force and velocity magnitudes did not consistently match the predictions of simple hand calculations. This suggests that the geometry of the hammer and rods must be considered more explicitly for accurate predictions. In some cases, the calculated energies were clearly in error and could not be used to correct the measured blow counts to a standard energy. SPT blow counts were corrected to 60% of the maximum potential energy (60% of 473 J = 284 J). The average rod energy ratios - 1 4 1 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 9.0 Conclusion and Recommendations for Future Research recorded during N A L P T were generally around 60% of the maximum potential energy of the N A L P T hammer. For this reason, N A L P T blow counts were also corrected to 60% of the maximum potential energy (60% of 1015 J = 609 J). The proposed SPT-LPT correlation method generated predictions that compared favourably with correlation factors from the literature and from the three research programs describe herein. Excluding the SPT-JLPT correlations, the predictions of the proposed correlation method ranged from 83% to 96% of the observed correlation factors while the Winterkorn and Fang predictions ranged from 39% to 73%. The JLPT data is considered unreliable because the data was not corrected to a standard energy. The predicted correlation factors were in better agreement with correlation factors observed in sand than in gravel, presumably because of grain size effects. Excluding the JLPT data, it was generally noted that the observed correlation factors decreased with increasing grain size, suggesting that SPT blow counts were decreasing with increasing grain size in the gravel range of soil particles. This observation is contrary to the current geotechnical perception that SPT blow counts increase with increasing grain size in the gravel range of soil particles. The observed SPT-JLPT correlation factors increased with increasing grain size but the "gravel" material used for the testing would be classified as coarse sand using the USC and BSC systems. Based on the above conclusions, the author suggests the following items for future research: • additional field testing of the proposed correlation method using a range of hammers and samplers specifically selected to provide a broad range of predicted correlation factors; • SPT and LPT calibration chamber testing with energy measurement to investigate grain size effects where the grain size distribution can be carefully controlled; -142-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Section 9.0 Conclusion and Recommendations for Future Research • additional quasi-static penetration tests using a range of split-spoon samplers to modify the proposed correlation method so that empirical quasi-static resistance corrections are not required. This type of investigation may also prove useful for pile-design applications and; • development of an empirical relationship between gravel liquefaction susceptibility chart and LPT blow counts or SPT blow counts predicted using SPT-LPT correlation factors; • development of a reliable method for field-checking dynamic energy monitoring systems; and, • development of simple software for modelling dynamic energy monitoring system data. This tool could be used to investigate uncertainties in the FF and F V energy measurement methods such as which rod area should be used for the FF method, what should the peak force and velocity be and what are the effects of impedance mismatches on calculated energy. -143-Bibliography A S T M . 1991a. Standard method for penetration test and split-barrel sampling of soils (D1586-84). In 1991 Annual Book of A S T M Standards, sect. 4, vol. 04.08. A S T M , Philadelphia, pp. 232-236. A S T M . 1991b. Standard test method for stress wave energy measurement for dynamic penetrometer testing systems (D4633-86). In 1991 Annual Book of A S T M Standards, sect. 4, vol. 04.08. A S T M , Philadelphia, pp. 872-875. Abou-Matar, H . 1990. Evaluation of Dynamic Measurements on the Standard Penetration Test. Thesis presented to the University of Colorado at Boulder in partial fulfillment of the degree of MS. Abou-matar, H. and Goble, G.G., 1997. SPT dynamic analysis and measurements. A S C E Journal of Geotechnical and Geoenvironmental Engineering, 123(10): 921-928. Andrus, R.D. and Youd, T.L., 1989. Penetration tests in liquefiable gravels. Proceedings, 12 t h ICSMFE, Rio De Janeiro, Balkema, Rotterdam, Volume 1, pp. 679-682. Burland, J.B. and Burbridge, M.C. , 1985. Settlement of foundations on sand and gravel. Proceedings, Institution of Civil Engineers, Part 1, December, 78: pp.1325-1381. Clayton, C.R.I. 1990. SPT energy transmission: theory, measurement and significance. Ground Engineering, 23(10): 35-43. Crova, R., Jamiolkowski, M . , Lancellota, R. and Lo Presti, D.C.F., 1993. Geotechnical characterization of gravelly soils at Messina site: selected topics. Predictive Soil Mechanics, Houlsby and Schofield, Ed., Thomas Telford, London, pp. 199-218. Decourt, L, 1989. The standard penetration test, stateOof-the-art report. Proceedings, 12 t h ICSMFE, Rio De Janeiro, Volume 4, pp 2405-2416. De Nicola, A . and Randolph, M.F., 1997. The plugging behaviour of driven and jacked piles in sand. Geotechnique, 47(4): 841-856. Goble, G.G. and Aboumatar, H. , 1992. Determination of wave equation soil constants from the standard penetration test. Proceedings, Fourth International Conference on the Application of Stress-wave Theory to Piles, The Hague, The Netherlands, pp. 99-103. Goble Rausche Likins and Associates, Inc. (GRLWEAP), 1997. G R L W E A P Manual, Version 1997-1, Cleveland, Ohio. -144-Harder, L.F. Jr. and Seed, H.B., 1986. Determination of penetration resistance for coarse-grained soils using the Becker hammer drill. Report No. UCB/EERC-86/06, University of California, Berkeley, USA, 118p. Hatanaka, M . and Uchida, A. , 1996. Empirical correlation between penetration resistance and internal friction angle of sandy soils. Soils and Foundations, 36(4): 1-9. Kaito, T., Sakaguchi, S., Nishigaka, Y. , Miki , K. and Yukami, H. , 1971. Large Penetration Test. Tsuchi-to-Kiso, 19(7): 15-21 (in Japanese). Kovacs, W.D. and Salomone, L .A. 1982. SPT hammer energy measurement. A S C E Journal of the Geotechnical Engineering Division, 108(GT4): 599-620. Lum, K . Y . and Yan, L. , 1994. In-situ measurements of dynamic soil properties and liquefaction resistance of gravelly soils at Keenleyside Dam. Ground Failures Under Seismic Conditions, A S C E Geotech. Special Pub. No. 44, pp. 221-240. McCulloch, D.S. and Bonilla, M.G. , 1970. Effects of the earthquake of March 27, 1964, on the Alaska Railroad. Geological Survey Professional Paper 545-D, United States Government Printing Office, Washington. McLean, F.G., Franklin, A . G . and Dahlstrand, T.K., 1975. Influence of Mechanical Variables on the SPT. Proceedings, Conference on In Situ Measurement of Soil Properties, A S C E , Volume 1, pp. 287-318. Morgano, C M . and Liang, R., 1992. Energy transfer in SPT - Rod length effect. Proceedings, Fourth International Conference on the Application of Stress-wave Theory to Piles, The Hague, The Netherlands, pp. 121-127. Paik, K. and Lee, S., 1993. Behavior of Soil Plugs in Open-Ended Model Piles Driven into Sands. Marine Georesources and Geotechnology, Vol . 11, pp. 353-373. Palacios, A. , 1977. The Theory and Measurement of Energy Transfer During Standard Penetration Test Sampling. Thesis presented to the University of Florida in partial fulfillment of the degree of Doctor of Philosophy. Robertson, P.K., Woeller, D.J. and Addo, K.O. , 1992. Standard penetration test energy using a system based on the personal computer. Canadian Geotechnical Journal, 29: 551-557. Rollins, K . M . , Diehl, N.B. and Weaver, T.J., 1998. Implications of V S -BPT (Ni)60 correlations for liquefaction assessment in gravels. Second International Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics, Seattle Washington, pp. 506-517. -145-Ross, G.A., Seed, H.B. and Migliaccio, R.R, 1969. Bridge foundation behavior in Alaska Earthquake. Journal of the Soil Mechanics and Foundations Division, Proceedings of the American Society of Civil Engineers, 95(SM4): 1007-1036. Schmertmann, J.H., 1979. Statics of SPT. A S C E Journal of the Geotechnical Engineering Division, 105(GT5): 655-670. Schmertmann, J.H. and Palacios, A . 1979. Energy Dynamics of SPT. A S C E Journal of the Geotechnical Engineering Division, 105(GT8): 909-926. Seed, H.B., Tokimatsu, K., Harder, L.F. and Chung, R . M . 1985. Influence of SPT procedures in soil liquefaction resistance evaluations. A S C E Journal of the Geotechnical Engineering Division, 111(GT12): 1425-1445. Skempton, A.W. 1986. Standard penetration test procedures and effects in sands of overburden, relative density, particle size, aging and overconsolidation. Geotechnique, 36(3): 425-447. Suzuki, Y . , Goto, S., Hatanaka, M . and Tokimatsu, K. , 1993. Correlation between strengths and penetration resistances for gravelly soils. Soils and Foundations, 33(1): 92-101. Sy, A . and Campanella, R.G., 1991a. An alternate method of measuring SPT energy. Proceedings, Second International Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics, St. Louis, Mo., pp. 499-505. Sy, A . and Campanella, R.G., 1991b. Wave equation modeling of the SPT. A S C E Geotechnical Engineering Congress, Boulder, Colorado, McLean, Campbell and Harris Ed., A S C E Geotechnical Special Publication. No. 27, Vol . 1, pp. 225-240. Sy, A . and Campanella, R.G., 1993. BPT-SPT correlations with consideration of casing friction. Proceedings, 46 t h Canadian Geotechnical Conference, Saskatoon, Saskatchewan, pp. 401-411. Tanaka, Y. , Kudo, K. , Kokusho, K. and Yoshida, Y. , 1991. Dynamic strength of gravelly soils and its relation to the penetration resistance. Proceedings, Second International Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics, St. Louis, Mo., pp. 399-406. Timoshenko, S. and Young, D.H., 1955. Vibration Problems in Engineering, D. Van Nostrand Company, New York. U.S. National Research Council, 1985. Liquefaction of soils during earthquakes, National Academy Press, Washington, D . C , 240p. -146-Winterkorn, H.F. and Fang, H., 1975. Foundation Engineering Handbook, Van Nostrand Reinhold Company, New York. Yoshida, Y . , Motonori, I. and Kokusho, T., 1988. Empirical formulas of SPT blow-counts for gravelly soils. Proceedings, ISOPT-1 for Penetration Testing, pp. 381-387. - 1 4 7 -A P P E N D I X A STRESS W A V E F O R M U L A E - 1 4 8 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split-Spoon Penetration Testing in Gravels Appendix A Stress Wave Formulae S T R E S S W A V E F O R M U L A E Palacios (1977) presents a detailed review of the use of the one-dimensional wave equation to describe the formation and propagation of stress waves in SPT drill rods. The results of his review include a number of simple, closed form solutions for stress wave magnitudes that are based on equilibrium and continuity requirements. These solutions can be used to assess the quality of stress wave data recorded in the field and are presented herein using the example of a simple hammer striking a simple rod. A l l of the solutions are based on the assumption that there are no energy losses during stress wave formation or propagation. Hammer Impact Figure A . l illustrates several stages in the formation and propagation of stress waves within a simple hammer and anvil system during and following hammer impact. The hammer and anvil are constructed of the same, homogenous material. A transducer element for recording force and velocity versus time is indicated near the bottom of the anvil rod. In Figure A . l a , the hammer is falling with uniform velocity (Vj) and the anvil rod is at rest. There are no axial forces in the hammer or anvil rod. When the hammer strikes the anvil rod, the impact surface of the hammer must decelerate and the impact surface of the anvil rod must accelerate so that the velocities of the two surfaces are equal (continuity requirement). This is a dynamic problem so it is inappropriate to treat the hammer and anvil as rigid bodies. Instead, the surface accelerations are accommodated by elastic strain within the two bodies on either side of the plane of impact. The magnitude of the resulting axial forces must satisfy force equilibrium across the plane of impact and may be calculated as follows: -149-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Appendix A Stress Wave Formulae Reflection at free end Hammer Velocity Anvil Rod (Zero Velocity) Reduced fall velocity Figure A . l Stages in the Formation and Propagation of Stress Waves Within a Simple Hammer and Anvil System. -150-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split-Spoon Penetration Testing in Gravels Appendix A Stress Wave Formulae F = Z R f V , Z R z H J (A.1) V = ( z N z 'H J (A.2) where: ZH = hammer impedance Z R = anvil rod impedance In Figure A . l b , the hammer has struck the anvil rod and equal lengths of material on either side of the plane of impact are strained (recall that the wave propagation velocity, c, is a function of the Young's modulus and mass density but not the cross sectional area). The resulting particle velocity of the material between the plane of impact and the hammer stress wave front is equal to the sum of the original hammer fall velocity (Vj) and the velocity determined using Equation A.2. The particle velocity of the material between the plane of impact and the anvil stress wave front is equal to the force calculated using Equation A . l divided by the anvil impedance. The two velocities are equal. The force in the strained material between the two stress wave fronts can be calculated using Equation A . l and would be positive (compressive) in this case. At the two interfaces between the strained and unstrained material there are large force imbalances that rapidly decelerate and accelerate hammer and anvil particles, respectively, to the translational particle velocity of the strained material. It is common practice to assume that the interfaces are normal to the rod axis (i.e. wave propagation is assumed to be one-dimensional). In Figure A . l c , the upward propagating stress wave arrives at the top of the hammer. The requirement of the resulting "free-end reflection" is that the total force at the top of the hammer is zero, so the stress wave is reflected with equal but opposite force (tension). -151 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split-Spoon Penetration Testing in Gravels Appendix A Stress Wave Formulae The particle velocity due to the new reflected wave is also negative. The new, total particle velocity behind the downward propagating stress wave is equal to the sum of the initial hammer fall velocity (Vj), the negative particle velocity of the original stress wave and the equal negative particle velocity of the reflected wave. This velocity will not be zero unless the hammer and anvil impedance are equal. When the reflected wave reaches the plane of impact (Figure A.le) , the hammer will be momentarily stress free and moving with uniform velocity, similar to the situation in Figure A . l a . The system will then repeat the stress wave propagation sequence with lower magnitude stress waves due to the reduced "impact" velocity of the hammer. Figure A.2 illustrates the shape of the force and velocity traces that would be recorded by the instrumentation package mounted on the hypothetical system in Figure A . l i f the anvil were semi-infinite in length. If the velocity trace were multiplied by the impedance of the anvil rod it would plot directly over the force trace because of force-velocity proportionality. The force and velocity magnitude decrease in steps, the duration of which are equal to twice the length of the hammer (LH) divided by the velocity of the wave front (c). The magnitudes of the force and velocity step decreases (i.e. the "rates of decay") following the peak value are a function of the ratio of hammer to anvil impedance. Impedance Interfaces Stress wave data recorded in the field rarely looks like the data shown in Figure A.2 because actual hammers and rod strings are usually more complex than the simple arrangement shown in Figure A . l (see typical field data presented in Sections 5.0, 6.0 and 7.0). The complexity of a hammer or rod string refers to the number of locations at which the impedance of the material changes due either to a change in cross sectional area, Young's modulus or material density. For example, rod strings typically consist of a series of 1.5 or 3.0 metre (5 or 10 foot) rods. The couplings between rods represent local changes of impedance. When a stress wave encounters an impedance interface, the -152-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Appendix A Stress Wave Formulae 2L|_| / c L|_| = Length of Hammer Figure A.2 Hypothetical Force and Velocity Data from Simple Hammer and Anvil System. -153-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split-Spoon Penetration Testing in Gravels Appendix A Stress Wave Formulae total energy of the incident wave is divided between the stress wave transmitted through the interface and a second, reflected stress wave (Figure A.3). The force and velocity of the transmitted (FT, VT) and reflected (FR, VR) waves are functions of the incident wave force and velocity (Fj, Vj): 2-R FT = \Z I J .3 Z i J (A.3) ( Z A 1 + ^ z 1 J (A.4) (z2 -^--1 u f zA 1 + - ^ - (A.5) V, • vD = 1 - ^ z 1 J z ^ 1 + ^ (A.6) The derivations of Equations A.3 to A.6 are based on force equilibrium and continuity requirements. Both reflected and transmitted waves exhibit force-velocity proportionality, when propagation direction is accounted for. Equations A.3 to A.6 are also valid for incident waves propagating up the rods, as long as the material is propagating from a material with impedance (Zi) into a (Zi) material. The free-end reflection case illustrated in Figure A . l c is a special case of a wave approaching an interface in which (Zi) is equal to zero. The case of a stress wave approaching a fixed -154-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Appendix A Stress Wave Formulae I \ / / \ LL Figure A.3 Transmitted and Reflected Waves Generated at an Impedance Interface. -155-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split-Spoon Penetration Testing in Gravels Appendix A Stress Wave Formulae end (total velocity must equal zero) is another special case represented by (Z2) equal to infinity. It should be noted that the presence of stress wave reflections in the rod string prior to the (2L/c) time theoretically invalidates the required assumptions of the FF energy measurement method. Equation A.3 to A.6 may be used to assess the effect of impedance interfaces on the measured FF energy. -156-APPENDIX B KIDD2 FIELD PROGRAM TEST RESULTS -157-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Appendix B Kidd2 Field Program Test Results Table B-1. Summary of Kidd2 Quasi-Static Penetration Test Data. Test Hole Depth* m(ft) C P T U Data Force*, kN (kip) Recovery (%) 9t bar (ksi) (%) Rod Measured Predicted SPT9901 13.63 (44.7) 76 (1.10) 0.37 1.29 (0.29) 9.34 (2.1) 9.79 (2.2) 92 16.61 (54.5) 119 (1.73) 0.32 1.56 (0.35) 19.1 (4.3) 15.13 (3.4) 94 18.08 (59.3) 156 (2.27) 0.37 1.69 (0.38) 17.4 (3.9) 20.47 (4.6) 89 LPT9902 5.94 (19.5) 25 (0.36) 0.20 0.67 (0.15) 10.7 (2.4) 4.89 (1.1) 79 9.00 (29.5) 82 (1.19) 0.37 0.89 (0.20) 17.4 (3.9) 17.35 (3.9) 29 12.04 (39.5) 113 (1.63) 0.29 1.16 (0.26) 34.7 (7.8) 22.69 (5.1) 67 15.15 (49.7) 96 (139) 0.39 1.42 (0.32) 30.3 (6.8) 20.91 (4.7) 75 18.11 (59.4) 152 (2.21) 0.37 1.69 (0.38) 43.6 (9.8) 32.48 (7.3) 54 LPT9903 14.36 (47.1) 73 (1.06) 0.39 1.29 (0.29) 23.1 (5.2) 15.57 (3.5) 63 16.65 (54.6) 110 (1.60) 0.33 1.56 (0.35) 34.7 (7.8) 22.69 (5.1) 58 18.08 (59.3) 156 (2.27) 0.37 1.69 (0.38) 46.3 (10.4) 33.37 (7.5) 60 SPT9904 7.47 (24.5) 53 (0.77) 0.41 0.76 (0.17) 15.1 (3.4) 7.12 (1.6) 0 11.31 (37.1) 82 (1.18) 0.25 1.16 (0.26) 14.7 (3.3) 9.79 (2.2) 83 18.14 (59.5) 146 (2.11) 0.37 1.69 (0.38) 20.9 (4.7) 19.13 (4.3) 89 * Tip dept h and force at 305 mm (' ft) sampler penetration. Average R f measured over one foot interval above tip depth. -158-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Appendix B Kidd2 Field Program Test Results Table B-2. HPA Data Collected During SPT at Kidd2. Test Hole Starting Test Depth, m(ft) Number of Peak Velocity, m/s (ft/s) Velocity Energy Ratio (%) * Blows Average Std. Dev. Average Std. Dev. 14.02 (46) 23 3.07 (10.1) 0.11 (0.4) 63.0 4.6 9901 15.54 (51) 28 3.03 (10.0) 0.11 (0.3) 61.6 4.3 17.07 (56) 30 3.10 (10.2) 0.09 (0.3) 64.4 3.7 18.59 (61) 40 3.06 (10.0) 0.09 (0.3) 62.5 3.5 4.88 (16) 16 3.12 (10.2) 0.08 (0.3) 65.1 3.6 7.92 (26) 15 3.19 (10.5) 0.05 (0.2) 68.1 2.4 9.51 (31.2) 30 3.16 • (10.4) 0.03 (0.1) 66.9 2.1 9904 12.50 (41) 25 3.21 (10.5) 0.06 (0.2) 68.9 2.7 15.54 (51) 26 3.19 (10.5) 0.04 (0.1) 67.8 1.7 17.07 (56) 31 3.16 (10.4) 0.03 (0.1) 66.8 0.0 18.59 (61) 46 3.19 (10.5) | 0.05 (0.2) 68.1 2.7 SPT energy quoted as percent of maximum standard SPT energy 473 J (350 ft-lb). -159-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Appendix B Kidd2 Field Program Test Results Table B-3. HPA Data Collected During NALPT at Kidd2 Test Starting Test Depth, m(ff) Number of Blows Peak Velocity, m/s (ft/s) Velocity Energy Ratio (%) * Hole Average Std. Dev. Average Std. Dev. 4.11 (13.5) 8 3.34 (10.9) 0.12 (0.4) 74.5 5.0 4.88 (16) 16 3.28 (10.8) 0.10 (0.3) 72.1 4.3 6.40 (21) 16 3.40 (11.2) 0.06 (0.2) 77.5 2.9 7.92 (26) 19 3.41 ( 1 1 . 2 ) 0.00 (0.0) 77.9 0.0 9.45 (31) 27 3.53 ( 1 1 . 6 ) 0.07 (0.2) 83.4 3.2 9902 10.97 (36) 21 3.44 (11.3) 0.05 (0.2) 79.0 2.3 12.50 (41) 23 3.43 (11.3) 0.05 (0.2) 78.8 2.1 14.02 (46) 22 3.47 (11.4) 0.08 (0.2) 80.4 3.5 15.54 (51) 23 3.41 ( 1 1 . 2 ) 0.00 (0.0) 77.9 0.0 17.07 (56) 31 3.40 (11.1) 0.06 (0.2) 77.1 2.6 18.59 (61) 37 3.38 (11.1) 0.06 (0.2) 76.6 2.8 4.88 (16) 10 3.40 (11.2) 0.04 (0.1) 77.3 1.8 6.40 (21) 14 3.43 (11.3) 0.07 (0.2) 78.9 3.3 7.92 (26) 20 3.43 (11.3) 0.07 (0.2) 78.7 3.1 9903 9.51 (31.2) 34 3.41 (11.2) 0.02 (0.1) 77.7 1.0 11.03 (36.2) 21 3.42 (11.2) 0.03 (0.1) 78.3 1.5 12.47 (40.9) 22 3.45 (11.3) 0.09 (0.3) 79.7 4.4 15.58 (51.1) 22 3.45 (11.3) 0.06 (0.2) 79.6 2.7 18.56 (60.9) 24 3.50 (11.5) 0.07 (0.2) 81.8 3.3 NALPT energy quoted as percent of maximum Kidd2 NALPT energy 1108 J (820 ft-lb) -160-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Appendix B Kidd2 Field Program Test Results Table B-4. Summary of SPT Blow Counts and Energy Ratios Recorded at Kidd2. Test Hole Starting Depth, m(ft) N E R v f F F Method F V Method E R r * T K i K 2 N 6 0 * E R r f N 6 0 9901 3.35 (11) 2 N . A . N . A . N . A . 4.88 (16) 6 N . A . 50.9 (34.4) 1.03 1.03 5 (4) 47.2 5 6.40 (21) 7 N . A . 52.0 (35.1) 1.02 1.02 6 (4) 59.2 7 7.92 (26) 13 N . A . 52.2 (35.3) 1.01 1.01 12 (8) 51.2 11 9.45 (31) 33 N . A . 56.5 (38.2) 1.0 1.0 31 (21) 53.6 29 10.97 (36) 22 N . A . 57.9 (39.1) 1.0 1.0 21 (14) 55.4 20 12.47 (40.9) 27 N . A . 54.0 (36.5) 1.0 1.0 24 (16) 52.2 23 14.02 (46) 19 63 56.8 (38.4) 1.0 1.0 18 (12) 51.4 16 15.54 (51) 21 62 58.9 (39.8) 1.0 1.0 21 (14) 51.3 18 17.01 (55.8) 22 65 62.6 (42.3) 1.0 1.0 23 (16) 58.8 22 18.53 (60.8) 30 62 61.8 (41.8) 1.0 1.0 31 (21) 54.7 27 9904 4.97 (16.3) 12 65 51.1 (34.5) 1.03 1.03 11 (7) 52.7 11 7.92 (26) 12 68 61.2 (41.4) 1.01 1.01 12 (8) 55.8 11 9.51 (31.2) 24 67 61.1 (413) 1.0 1.0 24 (17) 58.5 23 12.50 (41) 19 68 63.3 (42.8) 1.0 1.0 20 (14) 63.6 20 14.05 (46.1) 13 N . A . 67.5 (45.6) 1.0 1.0 15 (10) 67.7 15 15.54 (51) 21 68 63.9 (43.2) 1.0 1.0 22 (15) 57 20 17.01 (55.8) 24 67 66.6 (45.0) 1.0 1.0 27 (18) 61.8 25 18.38 (60.3) 36 68 78.7 1.0 1.0 47 j 59 35 N.A. Not Available K i and K 2 factors applied to F F energy corrected blow counts but not to F F energy ratios. Non-bracketed values calculated using "AWJ" rod area 4.84 cm 2 (0.75 in2), bracketed values calculated using area of "AW" transducer rod area 7.16 cm 2 (1.11 in2). "NWJ" rod area 9.16 cm 2 (1.42 in2) used for calculations at 18.38 m (60.3') in SPT9904. SPT energy quoted as percent of standard maximum SPT energy 473 J (350 ft-lb). -161 -M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Appendix B Kidd2 Field Program Test Results Table B-5. Summary of NALPT Blow Counts and Energy Ratios Recorded at Kidd2. Test Hole Starting ER V F FF Method FV Method Depth, m(ft) N ER R * T Ki K 2 N 6 o* E R ; N 6 0 4.11 (13.5) 3 74 83.7 1.06 1.11 - 57.1 3 6.40 (21) 6 78 107.1 1.02 1.03 - 66.8 6 7.86 (25.8) 9 78 101.1 1.0 1.02 - 69.1 10 9.51 (31.2) 15 84 110.5 1.0 1.01 - 72.1 17 9902 11.03 (36.2) 11 79 106.3 1.0 1.0 - 63.4 12 12.50 (41) 11 79 112.1 1.0 1.0 - 81.6 13 14.11 (46.3) 11 81 110.9 1.0 1.0 - 74.5 12 15.61 (51.2) 12 78 119.4 1.0 1.0 - 72.7 14 17.01 (55.8) 16 77 114.6 1.0 1.0 - 66.2 20 18.65 (61.2) 20 77 114.3 1.0 1.0 - 79.8 25 4.94 (16.2) 5 78 113.5 1.03 1.03 - 52.5 4 6.40 (21) 7 79 101.5 1.02 1.02 - 60.3 7 7.96 (26.1) 11 79 96.0 1.01 1.01 - 66.5 12 9903 9.51 (31.2) 16 78 115.4 1.0 1.00 - 68.6 18 11.03 (36.2) 12 78 136.4 1.0 1.0 - 67.7 14 12.47 (40.9) 10 79 114.0 1.0 1.0 - 72.6 12 15.58 (51.1) 11 79 116.7 1.0 1.0 - 63.1 12 18.56 (60.9) 12 82 116.5 1.0 1.0 - 70.0 14 «! and K 2 factors applied to FF energy corrected blow counts but not to FF energy ratios. "NWJ" rod area 9.16 cm 2 (1.42 in2) used for calculation. NALPT energy quoted as percent of maximum Kidd2 NALPT energy 1108 J (820 ft-lb). -162-APPENDIX C SEWARD, ALASKA FIELD PROGRAM TEST RESULTS -163-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Appendix C Split Spoon Penetration Testing in Gravels Seward, Alaska Field Program Test Results Table C-1. Summary of Seward, Alaska Grain Size Analysis Results. Test Depth * m(ft) SEWA9802 (SPT) SEWA9803 (NALPT) SEWA9806 (NALPT) u s e % Gravel f D 5 0 , mm u s e % Gravel f D 5 0, mm u s e % Graved D 5 0, mm 4.3 (14.1) GP 58 8.0 GM 49 4.4 GW 66 11.0 5.8 (19.0) GW-GM 53 5.5 GP-GM 47 3.9 GW 61 7.6 7.3 (24.0) SM 17 0.2 SP-SM 29 0.9 N.A. 9.0 (29.5) N.A. GP-GM 53 5.4 SM 0 0.2 10.4 (34.1) SW-SM 36 2.0 GW-GM 58 7.8 GP-GM 49 4.4 11.9 (39.0) SW-SM 43 3.3 GW-GM 55 6.4 GW-GM 60 8.8 13.5 (44.3) GP-GM 49 4.4 N.A. GM 47 3.4 15.0 (49.2) SM 15 0.1 SP-SM 30 2.8 ML 1 -16.5 (54.1) SP-SM 39 2.4 SP-SM 37 2.4 SM 37 1.7 18.0 (59.1) SW-SM 43 3.3 GW-GM 49 4.5 GM 50 4.8 19.6 (64.3) SW-SM 35 2.0 SP-SM 41 2.8 SW-SM 41 3.0 21.1 (69.2) SM 12 0.6 SM 12 0.5 SM 14 0.5 22.6 (74.1) N.A. N.A. SM 15 0.3 24.1 (79.1) SM 7 0.5 SM 5 0.4 SW-SM 10. 0.5 N.A. Not Available Exact test depth varies between test holes. Unified Soil Classification (USC) Gravel / Sand Boundary = #4 Sieve (4.75 mm). -164-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Appendix C Split Spoon Penetration Testing in Gravels Seward, Alaska Field Program Test Results Table C-2. HPA Data Collected During Seward Program. Test Hole Drop Height cm (in.) Test Depth m(ft) Peak Hammer Velocity of Single Hammer Drop, m/s (ft/s) Velocity Energy Ratio (%) * 2.68 (8.8) 48 13.5 (44.13) 2.80 (9.2) 53 SEWA9802 76.2 2.93 (9.6) 57 (SPT) (30) 2.56 (8.4) 44 15.0 (49.18) 2.68 (8.8) 48 2.80 (9.2) 53 3.17 (10.4) 59 16.5 (54.12) 3.29 (10.8) 64 3.17 (10.4) 59 SEWA9803 (NALPT) 86.4 (34) 3.17 (10.4) 59 18.0 (59.19) 3.17 (10.4) 59 3.04 (10) 55 3.17 (10.4) 59 19.6 (64.35) 3.17 (10.4) 59 3.17 (10.4) 59 SPT energy quoted as percent of standard maximum SPT energy 473 J (350 ft-lb), NALPT energy quoted as percent of maximum Seward NALPT energy 1126 J (833 ft-lb). -165-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Appendix C Split Spoon Penetration Testing in Gravels Seward, Alaska Field Program Test ResuIts Table C-3. Summary of SPT Blow Counts and Energy Ratios Recorded in SEWA9802. Starting Depth, m(ft) N F F Method F V Method E R r * f Ki K 2 N 6 0 * E R / N 6 0 4.3 (14.2) 30 38.8 (41.9) 1.04 1.08 22 (24) 39.4 20 5.8 (19.0) 32 43.7 (47.2) 1.03 1.04 25 (27) N.A. 7.3 (23.9) 10 42.1 (45.5) 1.01 1.02 7 (8) 45.5 8 9.1 (29.9) 23 46.0 (49.7) 1.01 1.01 18 (19) 52.1 20 10.3 (33.9) 61 46.9 (50.7) 1.00 1.00 48 (52) 51.4 52 11.9 (39.0) 54 40.7 (44) 1.00 1.00 37 (40) 44.5 40 13.4 (44.1) 33 41.8 (45.2) 1.00 1.00 23 (25) 46.7 26 15.0 (49.2) 25 43.6 (47.1) 1.00 1.00 18 (19) 46.4 19 16.5 (54.0) 53 39.6 (42.8) 1.00 1.00 35 (38) 41.7 37 18.1 (59.3) 99 45.0 (48.6) 1.00 1.00 74 (80) 48.8 81 19.5 (64.0) 52 N.A. N.A. 21.1 (69.2) 61 46.9 (50.7) 1.00 1.00 48 (52) 48.9 50 22.6 (74.1) 40 46.2 (49.9) 1.00 1.00 31 (34) 48.3 32 24.1 (79.2) 43 43.9 (47.5) 1.00 1.00 31 (34) N.A. N.A. Not Available * HM and K 2 factors applied to F F energy corrected blow counts but not to F F energy ratios. Non-bracketed values calculated using "A" rod area 7.74 cm 2 (1.2 in2), bracketed values calculated using area of "AW" transducer rod 7.16 cm 2 (1.11 in2). + SPT energy quoted as percent of standard maximum SPT energy 473 J (350 ft-lb). -166-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Split Spoon Penetration Testing in Gravels Appendix C Seward, Alaska Field Program Test Results Table C-4. Summary of NALPT Blow Counts and Energy Ratios Recorded in SEWA9803. Starting Depth, m(ft) N FF l\ /lethod FV Method ER r * f Ki K2 N 6 0 * E R / N 6 0 4.4 (14.3) 30 46.1 (63.2) 1.05 1.06 26 (35) 54.1 27 5.9 (19.3) 14 N.A. N A. 7.4 (24.3) 20 52.7 (72.2) 1.02 1.02 18 (25) 56.2 19 8.9 (29.3) 29 55.0 (75.3) 1.01 1.01 27 (37) 54.2 26 10.5 (34.3) 38 55.4 (75.9) 1.00 1.00 35 (48) 52.6 33 11.9 (39.2) 50 65.5 (89.7) 1.00 1.00 55 (75) 61.4 51 13.5 (44.2) 32 59.1 (80.9) 1.00 1.00 32 (43) 54.8 29 15.0 (49.2) 11 63.1 (86.4) 1.00 1.00 12 (16) 57.8 11 16.5 (54.1) 23 64.1 (87.8) 1.00 1.00 25 (34) 59.7 23 18.0 (59.2) 44 69.2 (94.7) 1.00 1.00 51 (69) N. A. 19.6 (64.3) 31 64.5 (88.3) 1.00 1.00 33 (46) 59.1 31 21.1 (69.3) 32 65.6 (89.9) 1.00 1.00 35 (48) 59.1 32 22.6 (74.3) 30 64.6 (88.5) 1.00 1.00 32 (44) 58.1 29 24.0 (78.9) 26 67.5 (92.5) 1.00 1.00 29 (40) 61.2 27 N.A. Not Available ' K, and K2 factors applied to FF energy corrected blow counts but not to FF energy ratios Non-bracketed values calculated using "NWJ" rod area 9.81 cm 2 (1.52 in2), bracketed values calculated using area of "AW" transducer rod 7.16 cm 2 (1.11 in2). NALPT energy quoted as percent of maximum Seward NALPT energy 1126 J (833 ft-lb). -167-APPENDIX D KEENLEYSIDE DAM FIELD PROGRAM TEST RESULTS -168-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Appendix D Split Spoon Penetration Testing in Gravels Keenleyside Dam Field Program Test Results Table D-1. HPA Data Collected in DH99-20 During NALPT at Keenleyside Dam. Starting Test Depth m(ft) Number of Blows Peak Velocity, m/s (ft/s) Velocity Energy Ratio (%) * Average Std. Dev. Average Std. Dev. 7.13 (23.4) 36 10.2 0.6 64.3 7.5 8.93 (29.3) 21 10.1 0.6 63.9 7.9 10.30 (33.8) 45 10.5 0.7 68.5 9.0 11.83 (38.8) 22 10.8 0.5 72.2 7.3 13.38 (43.9) 38 10.7 0.5 71.5 6.2 14.9 (48.9) 28 11.1 0.5 76.5 7.2 17.95 (58.9) 48 10.8 0.5 73.2 6.3 20.97 (68.8) 96 10.6 0.7 70.6 8.8 N.A. - Not Available NALPT energy quoted as percent of maximum Keenleyside NALPT energy 1108 J (820 ft-lb). -169-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Appendix D Split Spoon Penetration Testing in Gravels Keenleyside Dam Field Program Test Results Table D-2. Summary of NALPT Blow Counts and Energy Ratios Recorded in DH99-20. Starting Depth, m(ft) N FF Method FV Method ERR* F Ki K 2 N 6 0 * E R ; N 6 0 2.99 (9.8) 19 32.8 (42.0) 1.38 1.09 16 (20) 62.4 20 4.42 (14.5) 18 33.5 (42.9) 1.16 1.05 12 (15) 64.2 19 5.61 (18.4) 12 31.0 (39.7) 1.08 1.03 7 (22) 56.5 11 7.13 (23.4) 21 N.A. 8.93 (29.3) 8 39.8 (50.9) 1.02 1.01 5 (6) 59.9 8 10.30 (33.8) 20 45.9 (58.8) 1.01 1.00 15 (19) 63.2 21 11.83 (38.8) 16 50.2 (64.3) 1.00 1.00 13 (17) 71.7 19 13.38 (43.9) 26 50.9 (65.1) 1.00 1.00 22 (28) 73.5 32 14.9 (48.9) 9 60.1 (76.9) 1.00 1.00 9 (12) 80.7 12 17.95 (58.9) 24 59.0 (75.5) 1.00 1.00 24 (31) 77.7 31 20.97 (68.8) 45 69.1 (88.3) 1.00 1.00 52 (67) 97.5 73 N.A. Not Available Ki and K 2 factors applied to FF energy corrected blow counts but not to FF energy ratios. Non-bracketed values calculated using "NW" rod area 9.16 cm 2 (1.42 in2), bracketed values calculated using area of "AW" transducer rod 7.16 cm 2 (1.11 in2). NALPT energy quoted as percent of maximum Keenleyside NALPT energy 1108 J (820 ft-lb). -170-M.A.Sc. Thesis, Chris R. Daniel The University of British Columbia Appendix D Split Spoon Penetration Testing in Gravels Keenleyside Dam Field Program Test Results Table D-3. Keenleyside Dam Comparison Data Set. DH91-3c(BPT) DH91-3d (SPT) DH99-20 (NALPT) Depth m(ft) Equivalent (N 6 0 ) S PT Depth m(ft) (N6Cj)sPT Depth m(ft) (N6O)NALPT Harder and Seed Sy and Campanella F2 Method * FV Method 3.0 (10) 8 16 3.1 (10) 23 2.99 (9.8) 16 (20) 20 4.6 (15) 8 13 4.6 (15) 30 4.42 (14.5) 12 (15) 19 5.8 (19) 6 6 6.1 (20) 17 5.61 (18.4) 7 (22) 11 9.1 (30) 6 4 9.0 (29.5) 5 8.93 (29.3) 5 (6) 8 10.4 (34) 8 8 10.7 (35) 15 10.30 (33.8) 15 (19) 21 11.9 (39) 12 15 12.2 (40) 33 11.83 (38.8) 13 (17) 19 13.4 (44) 8 6 13.7 (45) 31 13.38 (43.9) 22 (28) 32 14.9 (49) 13 14 15.2 (50) 23 14.9 (48.9) 9 (12) 12 18.0 (59) 18 24 18.3 (60) 20 17.95 (58.9) 24 (31) 31 21.0 (69) 17 19 21.5 (70.5) 70 20.97 (68.8) 52 (67) 73 Non-bracketed values calculated using "NW" rod area 9.16 cm* (1.42 in*), bracketed values calculated using area of "AW" transducer rod 7.16 cm 2 (1.11 in2). -171 -

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.831.1-0064071/manifest

Comment

Related Items