UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Shear strength of soil-pipeline interfaces under low confining stresses and large displacements Amarasinghe, Ruslan Shanth 2019

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

Item Metadata

Download

Media
24-ubc_2019_september_amarasinghe_ruslan.pdf [ 56.38MB ]
Metadata
JSON: 24-1.0379469.json
JSON-LD: 24-1.0379469-ld.json
RDF/XML (Pretty): 24-1.0379469-rdf.xml
RDF/JSON: 24-1.0379469-rdf.json
Turtle: 24-1.0379469-turtle.txt
N-Triples: 24-1.0379469-rdf-ntriples.txt
Original Record: 24-1.0379469-source.json
Full Text
24-1.0379469-fulltext.txt
Citation
24-1.0379469.ris

Full Text

Shear strength of soil-pipelineinterfaces under low confiningstresses and large displacementsbyRuslan Shanth AmarasingheB.Sc.Engrg. (Hons), The University of Peradeniya, Sri Lanka, 2009M.A.Sc., The University of British Columbia, 2013A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinThe Faculty of Graduate and Postdoctoral Studies(Civil Engineering)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)June 2019© Ruslan Shanth Amarasinghe, 2019The following individuals certify that they have read, and recommend to the Facultyof Graduate and Postdoctoral Studies for acceptance, the thesis entitled:Shear strength of soil-pipeline interfaces under low confiningstresses and large displacementssubmitted by Ruslan Shanth Amarasinghe in partial fulfillment of the require-ments for the degree of Doctor of Philosophy in Civil Engineering.Examining Committee:Prof. Dharma Wijewickreme, Civil EngineeringSupervisorProf. Erik Eberhardt, Geological SciencesSupervisory Committee MemberProf. John Howie, Civil EngineeringSupervisory Committee MemberProf. Donald Anderson, Civil EngineeringUniversity ExaminerProf. Edouard Asselin, Materials EngineeringUniversity ExaminerProf. James Blatz, University of Manitoba, Civil EngineeringExternal ExaminerAdditional Supervisory Committee Members:Prof. Liam Finn, Civil EngineeringSupervisory Committee MemberiiAbstractThe design of onshore and offshore energy pipelines requires that the geotechnicalaspects of the soil-pipe interaction are adequately captured. The soil-pipe interfaceshear strength is an important design parameter that is required for assessing soilloads on pipelines. Laboratory element and physical model testing conducted underrelatively large displacements and soil confining stresses relevant to the soil-pipe in-teraction problem (typically within the range of 3 to 30 kPa) is required to deriveappropriate soil-pipe interaction parameters for use in designs and for the validationand improvement of design guidelines. However, laboratory test methods for assessingsoil-solid interface shear strength under low confining stresses and large displacementsare not well developed, and the availability of laboratory test data conducted underthe above conditions is very limited. A novel macro-scale interface direct shear appa-ratus capable of testing soil specimens of 1 m × 1 m footprint on various solid surfacesunder low confining stresses (2 to 40 kPa) and large displacements (up to 1 m) whileproviding the means to measure pore-water pressure and total normal stress at thesoil-solid interface was developed, and a series of macro-scale interface direct sheartests were conducted to study the effect of soil type, confining stress level, and surfaceroughness of the solid surface on the drained large displacement soil-solid interfacefriction angle using a number of soils and epoxy coated steel surfaces. A novel full-scale axial soil-pipe interaction physical model test apparatus capable of testing pipesof up to 0.45 m (18 inch) in diameter in saturated fine-grained soil was developed anda limited number of tests were conducted using an NPS18 (0.45 m diameter) epoxycoated steel pipe on a fine-grained soil bed. This dissertation presents the detailsof the new apparatus and the tests conducted. The results of the experiments andimportant findings are presented and discussed.iiiLay SummaryPipelines play a major role in the transportation and continuous supply of oil and gasacross the continent. They are also used in deep water in the sea connecting continentstogether. Failure of pipelines should be avoided given the detrimental impact it canhave on the environment and economy. As with any man-made structure, pipelinesare subjected to a variety of natural and operational hazards. Because pipelines areburied, they are exposed to natural hazards such as landslides and earthquakes andit is important to design pipelines to withstand resulting forces. The less one knowsabout how a pipe will interact with the soil around it, the more uncertainty there isand this leads to the unnecessary overuse of resources as the design needs to accountfor the uncertainty. The present research improves our understanding of how differentpipeline materials interact with different soils under different hazard conditions.ivPrefaceThis dissertation contains details of a research program conducted at the Departmentof Civil Engineering, University of British Columbia during the period 2013 to 2018.The research work was undertaken by the author, Ruslan Shanth Amarasinghe, underthe supervision of Professor Dharma Wijewickreme.The work encompassed the design and fabrication of two novel laboratory testapparatus: (i) macro-scale interface direct shear apparatus; and (ii) modified full-scalesoil-pipe interaction apparatus, and carrying out laboratory testing using the abovedevices. The concept designs of the new devices were developed by the author withsuggestions from Professor Wijewickreme. The structural design of the stationaryframe of the macro-scale interface direct shear apparatus was carried out by the authorwith suggestions from Professor Siegfried Stiemer. The fabrication of the structuralmembers of the stationary frame were carried out by a local steel fabrication company,and the assembly of the stationary frame was carried out by the author with theassistance of Mr. Adebayo Adegbola, Mr. YoungHa Hwang, and Ms. April Graves.The design and development of the mobile frame, specimen container, wiper system,linear bearing system, pneumatic load application system, pore-water measurementsystem, total normal force measurement system, linear actuator system, and the dataacquisition and control system of the macro-scale interface direct shear apparatuswas carried out by the author with workshop machining support provided by Mr.Harald Schrempp and Mr. Bill Leung. Discussions held with Mr. Scott Jackson, andguidance provided by Dr. Noboru Yonemitsu with regard to the development of thedata acquisition system is acknowledged.The modified full-scale axial soil-pipe interaction device was developed by modify-ing the UBC ASPIReTM soil chamber. The structural design and implementation ofvPrefacethe mobile walls, stationary walls, test pipe instrumentation, linear actuator system,and data acquisition system were carried out by the author with structural fabrica-tion and assembly assistance provided by Ms. April Graves and Mr. Kellie Liu. Theepoxy coated steel plates used in the macro-scale interface direct shear apparatuswere coated by a local epoxy coating company.The soil specimen preparation procedure for both the macro-scale interface directshear apparatus and the full-scale soil-pipe interaction apparatus was developed bythe author. All laboratory tests conducted using the two new apparatus were carriedout by the author.Interpretation of the test data and discussions were carried out by the author withsuggestions from Professor Dharma Wijewickreme and the supervisory committee.Review of the thesis was carried out by Professor Dharma Wijewickreme and thesupervisory committee. Additional review assistance was provided by Dr. SadanaKarannagoda Gamage.viTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiLay Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvList of Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxxAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxxiv1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Purpose of the Research Program . . . . . . . . . . . . . . . . . 41.2 Scope of the Research Program . . . . . . . . . . . . . . . . . . . 51.3 Organization of the Thesis . . . . . . . . . . . . . . . . . . . . . 62 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82.1 Geotechnical Aspects in the Design of Pipelines Subjected to Rela-tive Deformation . . . . . . . . . . . . . . . . . . . . . . . . . . . 82.1.1 Axial soil-pipe interface shear strength: total stress approach(α method) . . . . . . . . . . . . . . . . . . . . . . . . . 102.1.2 Axial soil-pipe interface shear strength: effective stress ap-proach . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15viiTable of Contents2.2 Internal Friction Angle of Soils at Large Strains and Low Stresses 202.2.1 Friction angle of granular materials . . . . . . . . . . . . 202.2.2 Friction angle of granular materials at low confining stresses 252.2.3 Friction angle of fine-grained soils at low confining stresses 292.3 Friction Angle of Soil-Solid Interfaces at Large Strains and LowStresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342.3.1 Friction angle of sand-polymer interfaces . . . . . . . . . 342.3.2 Friction angle of sand-steel interfaces . . . . . . . . . . . 362.3.3 Friction angle of clay-solid interfaces . . . . . . . . . . . . 382.4 Laboratory Testing of Soil-Solid Interface Friction Angle . . . . . 422.4.1 Direct shear apparatus . . . . . . . . . . . . . . . . . . . 452.4.2 Direct simple shear apparatus . . . . . . . . . . . . . . . 482.4.3 Ring shear apparatus . . . . . . . . . . . . . . . . . . . . 502.4.4 Other devices . . . . . . . . . . . . . . . . . . . . . . . . 512.5 Summary of the Chapter . . . . . . . . . . . . . . . . . . . . . . 532.6 Expected Contributions . . . . . . . . . . . . . . . . . . . . . . . 553 The Macro-Scale Interface Direct Shear Apparatus . . . . . . . . 583.1 Overview of Test Device . . . . . . . . . . . . . . . . . . . . . . . 593.2 Description of Main Components of Test Device . . . . . . . . . 633.2.1 Stationary frame and normal force measurement system . 633.2.2 Mobile frame and actuator assembly . . . . . . . . . . . . 663.2.3 Normal force application system . . . . . . . . . . . . . . 743.2.4 Measurement of pore-water pressure . . . . . . . . . . . . 853.3 Details of Specimen Preparation . . . . . . . . . . . . . . . . . . 913.3.1 Preparation of coarse-grained soil specimens . . . . . . . 933.3.2 Preparation of fine-grained soil specimens . . . . . . . . 943.4 Selection of Shear Displacement Rate . . . . . . . . . . . . . . . 993.5 Interpretation of Data . . . . . . . . . . . . . . . . . . . . . . . . 1013.6 Sources of error and corrective measures . . . . . . . . . . . . . . 103viiiTable of Contents3.6.1 Interface shear stress . . . . . . . . . . . . . . . . . . . . 1043.6.2 Total normal stress . . . . . . . . . . . . . . . . . . . . . 1083.6.3 Pore-water pressure and effective normal stress . . . . . . 1093.6.4 Quality control measures taken to minimize soil specimenvariability . . . . . . . . . . . . . . . . . . . . . . . . . . 1103.7 Summary of the Chapter and Contributions . . . . . . . . . . . . 1114 Experimental Program and Results . . . . . . . . . . . . . . . . . . 1144.1 Materials Tested . . . . . . . . . . . . . . . . . . . . . . . . . . . 1144.1.1 Soil materials . . . . . . . . . . . . . . . . . . . . . . . . 1154.1.2 Solid surfaces . . . . . . . . . . . . . . . . . . . . . . . . 1224.2 Tests Conducted and Observations . . . . . . . . . . . . . . . . . 1294.2.1 Interface shear characteristics of Fraser River sand with dif-ferent solid surfaces . . . . . . . . . . . . . . . . . . . . . 1374.2.2 Interface shear characteristics of Fraser River silt with dif-ferent solid surfaces . . . . . . . . . . . . . . . . . . . . . 1454.2.3 Interface shear characteristics of Kaolinite with different solidsurfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . 1574.2.4 Interface shear characteristics of Strong-Pit FGS with dif-ferent solid surfaces . . . . . . . . . . . . . . . . . . . . . 1684.2.5 Interface shear characteristics of Redstone on the green epoxycoated steel surface . . . . . . . . . . . . . . . . . . . . . 1774.2.6 Interface shear characteristics of Kaosand on the sand-blastedmild steel surface . . . . . . . . . . . . . . . . . . . . . . 1814.2.7 Surface roughness observed at the end of testing . . . . . 1844.2.8 Repeatability test data . . . . . . . . . . . . . . . . . . . 1864.2.9 Displacement rate effects on test data . . . . . . . . . . . 1914.3 Summary of the Chapter, Key Findings, and Contributions . . . 1995 Discussion on Macro-Scale Interface Direct Shear Test Results 202ixTable of Contents5.1 Influence of the Effective Normal Stress on the Soil-Solid InterfaceShear Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2025.1.1 Solid interfaces with Fraser River sand and Fraser River silt 2025.1.2 Kaolinite interfaces . . . . . . . . . . . . . . . . . . . . . 2085.2 Effect of Roughness of the Test Surfaces on the Interface ShearStrength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2145.3 Effect of Rate of Shearing on the Interface Shear Strength . . . . 2225.4 Summary of the Chapter, Key Findings, and Contributions . . . 2246 Full-Scale Axial Soil-Pipe Interaction Testing . . . . . . . . . . . 2306.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2306.2 New Device for Axial Soil Restraint Testing of Buried Pipes in Com-pressible Soils . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2346.2.1 Overview of test device . . . . . . . . . . . . . . . . . . . 2346.2.2 Description of main components of test device . . . . . . 2356.3 Experimental Program . . . . . . . . . . . . . . . . . . . . . . . 2426.3.1 Test pipe . . . . . . . . . . . . . . . . . . . . . . . . . . . 2426.3.2 Preparation of soil-pipe test chamber . . . . . . . . . . . 2426.3.3 Instrumentation . . . . . . . . . . . . . . . . . . . . . . . 2476.3.4 Test procedure . . . . . . . . . . . . . . . . . . . . . . . . 2526.3.5 Test results . . . . . . . . . . . . . . . . . . . . . . . . . 2546.3.6 Interpretation of test data . . . . . . . . . . . . . . . . . 2606.4 Discussion of Soil-Pipe Interaction Test Results . . . . . . . . . . 2676.5 Summary of the Chapter, Key Findings, and Contributions . . . 2747 Summary, Conclusions and Future Work . . . . . . . . . . . . . . 2777.1 Summary of the Key Findings and Contributions . . . . . . . . . 2787.1.1 Development of a novel macro-scale interface direct shearapparatus . . . . . . . . . . . . . . . . . . . . . . . . . . 2787.1.2 Development of a novel full-scale axial soil-pipe interactiontest apparatus . . . . . . . . . . . . . . . . . . . . . . . . 279xTable of Contents7.1.3 Characterization of the drained large-displacement soil-solidinterface shear strength . . . . . . . . . . . . . . . . . . . 2807.1.4 Effect of solid surface roughness on the soil-solid interfaceshear strength . . . . . . . . . . . . . . . . . . . . . . . . 2837.1.5 Full-scale axial soil-pipe interaction testing to study the ef-fect of confining stress level on mobilized axial soil loads onpipelines . . . . . . . . . . . . . . . . . . . . . . . . . . . 2847.2 Limitations and Future Work . . . . . . . . . . . . . . . . . . . . 285Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306A Selection of Optimum Specimen Footprint Size . . . . . . . . . . 307B Surface Roughness and its Measurement . . . . . . . . . . . . . . 309B.1 Characterization of surface roughness profiles . . . . . . . . . . . 309B.1.1 Obtaining the roughness profile . . . . . . . . . . . . . . 312B.2 Surface roughness parameters . . . . . . . . . . . . . . . . . . . . 315xiList of Tables2.1 Interface efficiency factors recommended by the American Lifeline Al-liance (ALA) (ALA, 2001), and the Pipeline Research Council Inter-national (PRCI) (PRCI, 2004, 2009) guidelines . . . . . . . . . . . . 162.2 Summary of previous research on soil-solid interfaces conducted usingthe direct shear apparatus . . . . . . . . . . . . . . . . . . . . . . . 472.3 Summary of previous research on soil-solid interfaces conducted usingthe direct simple shear apparatus . . . . . . . . . . . . . . . . . . . 492.4 Summary of previous research on soil-solid interfaces conducted usingthe ring-shear apparatus . . . . . . . . . . . . . . . . . . . . . . . . 514.1 Composition of the Fraser River silt test-soil (based on X-ray diffrac-tion testing carried out at the UBC Dept. of Earth, Ocean and At-mospheric Sciences). . . . . . . . . . . . . . . . . . . . . . . . . . . 1174.2 Composition of the Kaolinite clay (based on information provided bythe vendor) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1184.3 Composition of the Redstone soil (based on information provided bythe vendor). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1204.4 Properties of fine-grained soils . . . . . . . . . . . . . . . . . . . . . 1224.5 Surface roughness and hardness parameters of the solid test plates . 1294.6 Summary of macro-scale interface direct shear tests. . . . . . . . . . 1344.7 Results of macro-scale interface direct shear tests conducted on FraserRiver sand against the sand-blasted mild-steel surface. . . . . . . . 1394.8 Results of macro-scale interface direct shear tests conducted on FraserRiver sand against the abrasion resistant green epoxy surface. . . . 142xiiList of Tables4.9 Results of macro-scale interface direct shear tests conducted on FraserRiver sand against the grey epoxy surface. . . . . . . . . . . . . . . 1444.10 Results of macro-scale interface direct shear tests conducted on FraserRiver silt against the sand-blasted mild steel surface. . . . . . . . . 1494.11 Results of macro-scale interface direct shear tests conducted on FraserRiver silt against the grey epoxy surface. . . . . . . . . . . . . . . . 1524.12 Results of macro-scale interface direct shear tests conducted on FraserRiver silt against the green epoxy surface. . . . . . . . . . . . . . . 1564.13 Results of macro-scale interface direct shear tests conducted on Kaoli-nite against the sand-blasted mild steel surface. . . . . . . . . . . . 1604.14 Results of macro-scale interface direct shear tests conducted on Kaoli-nite against the grey epoxy surface. . . . . . . . . . . . . . . . . . . 1634.15 Results of macro-scale interface direct shear tests conducted on Kaoli-nite against the green epoxy surface. . . . . . . . . . . . . . . . . . 1654.16 Results of macro-scale interface direct shear tests conducted on Strong-Pit FGS against the sand-blasted mild steel surface. . . . . . . . . . 1704.17 Results of macro-scale interface direct shear tests conducted on Strong-Pit FGS against the grey epoxy coated steel surface. . . . . . . . . . 1744.18 Results of macro-scale interface direct shear tests conducted on Strong-Pit FGS against the green epoxy coated steel surface. . . . . . . . . 1774.19 Results of macro-scale interface direct shear tests conducted on Red-stone against the green epoxy coated steel surface. . . . . . . . . . . 1794.20 Results of macro-scale interface direct shear tests conducted on Kaosandagainst the sand-blasted mild steel surface. . . . . . . . . . . . . . . 1824.21 Results of macro-scale interface direct shear repeatability tests. . . . 1914.22 Results of macro-scale interface direct shear rate effect tests conductedon the grey epoxy surface. . . . . . . . . . . . . . . . . . . . . . . . 199xiiiList of Tables5.1 Interface efficiency factors recommended by the American Lifeline Al-liance (ALA) (ALA, 2001), and the Pipeline Research Council Inter-national (PRCI) (PRCI, 2004, 2009) guidelines . . . . . . . . . . . . 2055.2 Drained large-displacement interface efficiency factors for Fraser Riversand and Fraser River silt obtained at low effective normal stresses. 2055.3 Drained large-displacement interface efficiency factors for Kaoliniteobtained at low effective normal stresses. . . . . . . . . . . . . . . . 2126.1 List of axial soil-pipe interaction tests conducted . . . . . . . . . . . 2526.2 Results of the fill-scale axial soil-pipe interaction tests. . . . . . . . 2666.3 Estimated α values obtained from the full-scale axial soil-pipe inter-action tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267B.1 Typical sampling lengths for roughness of periodic surface profiles(ASME Standard B46.1-2002, 2003) . . . . . . . . . . . . . . . . . . 315B.2 Typical sampling lengths for roughness of non-periodic surface profiles(ASME Standard B46.1-2002, 2003) . . . . . . . . . . . . . . . . . . 315xivList of Figures2.1 Idealized soil-spring models used in analysis of soil-pipe interaction.(a) Continuum model. (b) Soil-spring model. (c) Soil-spring load-displacement response. . . . . . . . . . . . . . . . . . . . . . . . . 92.2 PRCI recommended bounds for the adhesion factor α. Adopted fromPRCI (PRCI, 2009).(C-CORE, 2008; Cappelletto et al., 1998; Honeg-ger, 1999; Paulin et al., 1998; Rizkalla et al., 1996). . . . . . . . . . 112.3 Components of shear resistance of coarse-grained soils (After Rowe(1962)). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212.4 Relationship between peak friction angle and maximum rate of dila-tancy (dv/da)max (After Vaid and Sasitharan (1992)). . . . . . . 222.5 Typical undrained trixial compression test results showing the steadystate and phase transformation lines (After Negussey et al. (1988);Vaid et al. (1990)). . . . . . . . . . . . . . . . . . . . . . . . . . . . 232.6 Direct shear test results of Ottawa sand and Ottawa sand on HDPE(After O’Rourke et al. (1990)). . . . . . . . . . . . . . . . . . . . . 272.7 Thin-specimen direct shear apparatus (After Pedersen et al. (2003)). 302.8 Drained peak shear strength of Kaolinite measured using thin-specimendirect shear apparatus (After Pedersen et al. (2003)). . . . . . . . . 302.9 Drained ring-shear test results on Kamenose soil and Morikawa soil(After Kimura et al. (2014)). . . . . . . . . . . . . . . . . . . . . . 312.10 Drained ring-shear test results on Chenyoulanxi soil and Miaowan3soil (After Kimura et al. (2014)). . . . . . . . . . . . . . . . . . . . 322.11 Drained ring-shear test results on Brown London clay (After Bishopet al. (1971)). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33xvList of Figures2.12 Drained ring-shear test results on Lias clay (After Chandler and John-son (1976)). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332.13 Typical configurations of the modified direct shear apparatus used forinterface shear testing. . . . . . . . . . . . . . . . . . . . . . . . . . 452.14 Typical configuration of the simple shear type apparatus used forsoil-solid interface shear testing (adapted from Uesugi and Kishida(1986b)). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483.1 Photograph of the macro-scale interface direct shear test device. . 613.2 Schematic diagram of the macro-scale interface shear test device andtest setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 623.3 Drawing of stationary frame. (Note that the base plate, test plate,and the mobile frame are not shown). . . . . . . . . . . . . . . . . 643.4 Photograph showing base load cells. . . . . . . . . . . . . . . . . . 653.5 Drawing of stationary frame with the base plate and test plate in-stalled. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 663.6 Drawing of stationary frame with the base plate, test plate, and mobileframe installed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 673.7 Photograph showing the arrangement of pore-water pressure ports onthe test plate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 683.8 End elevation of macro-scale interface direct shear apparatus. . . . 693.9 Photograph showing the north-east horizontal linear bearing. . . . 703.10 Photograph showing mobile frame and specimen container. . . . . 713.11 Photograph showing side load cells. . . . . . . . . . . . . . . . . . 713.12 Photographs showing the geotextile covered rubber wipers of the mo-bile frame. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 723.13 Photograph showing the main load cell and actuator assembly of themacro-scale interface direct shear apparatus. . . . . . . . . . . . . 733.14 Schematic cross-sectional view of the wiper and main load cell assem-bly of the macro-scale interface direct shear apparatus. . . . . . . . 73xviList of Figures3.15 Schematic diagrams showing the normal force application system. (a)Dead weight surcharge loading system, (b) Dead weight surchargeloading system in combination with the pneumatic pressure applica-tion system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 763.16 Photographs showing the load transfer plate of the macro-scale inter-face direct shear apparatus. . . . . . . . . . . . . . . . . . . . . . . 773.17 Photographs showing the pneumatic pressure plate of the macro-scaleinterface direct shear apparatus. . . . . . . . . . . . . . . . . . . . 773.18 Photographs showing the installation sequence of the pneumatic pres-sure application system of the macro-scale interface direct shear ap-paratus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 783.19 Photographs showing the pressure pad matrix device used to deter-mine the spatial distribution of total normal stress. . . . . . . . . . 803.20 Spatial distribution of total normal stress at the interface due to thesurcharge sand layer (contour values are in kPa). . . . . . . . . . . 813.21 Spatial distribution of total normal stress at the interface due to thecombination of surcharge sand layer and the pneumatic pressure ap-plication system (contour values are in kPa). . . . . . . . . . . . . 823.22 Photograph showing the natural latex rubber membrane being pre-pared. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 823.23 Schematic diagram showing specimen container and water tank foot-prints used for total normal stress calculation. . . . . . . . . . . . . 843.24 Photograph of pressure transducer that is used for pore-water pressuremeasurement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 853.25 Details of pressure transducer connection to the test plate. (Amaras-inghe, 2013) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 863.26 Details of pressure transducer saturation procedure. (a) Saturation ofpressure transducer cavity. (b) Attachment of saturated transducerto stanless steel adapter and saturation of adapter. (c) Placement ofporous stone. (Amarasinghe, 2013) . . . . . . . . . . . . . . . . . . 87xviiList of Figures3.27 Schematic diagram showing the initial and final position of soil speci-men with respect to the pore-water pressure transducer ports. . . . 883.28 Time history of pore-water pressure readings for a Kaolinite speci-men on a mild steel test surface obtained during application of sandsurcharge and consolidation. . . . . . . . . . . . . . . . . . . . . . 903.29 Time history of pore-water pressure readings for a Kaolinite specimenon a mild steel test surface obtained during application of pneumaticloading and consolidation. . . . . . . . . . . . . . . . . . . . . . . . 913.30 Photographs showing cleaning and setup of specimen container andtest plate prior to specimen placement. . . . . . . . . . . . . . . . 923.31 Plastic container of 100 L capacity and the electric mixer used in thepreparation of the slurry. . . . . . . . . . . . . . . . . . . . . . . . 943.32 Photograph showing pouring of fine-grained soil slurry into the speci-men container. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 953.33 Photograph showing an as-placed fine-grained soil slurry layer insidethe specimen container. . . . . . . . . . . . . . . . . . . . . . . . . 973.34 Diagram showing the self-weight settled soil specimen and the clearwater surface formed during settlement. . . . . . . . . . . . . . . . 983.35 Typical pore-water pressure dissipation profile during consolidation offine-grained soil specimens under surcharge loading measured on themacro-scale interface direct shear apparatus. . . . . . . . . . . . . 993.36 Variation of the raw unfiltered main load cell reading with the ambienttemperature in the laboratory recorded continuously over the courseof two days. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1053.37 Device friction force of the macro-scale interface direct shear apparatusmeasured over a horizontal displacement range of 80 cm (Specimencontainer was equipped with geotextile-lined wipers, and the test wasconducted over an epoxy coated steel surface). . . . . . . . . . . . 1063.38 Device friction force measured at the end of testing a Kaolinite speci-men on the green epoxy test surface. . . . . . . . . . . . . . . . . . 107xviiiList of Figures3.39 Variation of pore-water pressure at the soil-solid interface measuredduring a macro-scale interface direct shear test conducted using re-constituted fully-saturated Kaolinite clay on a sand-blasted mild steelsurface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1103.40 Photographs showing the sampling of a fine-grained soil specimen atthe end of macro-scale interface direct shear testing for determinationof final average moisture content of the soil specimen. . . . . . . . 1114.1 Grain size distribution of soils tested. . . . . . . . . . . . . . . . . 1164.2 Photographs showing the processing of the silt samples by wet sievingto remove debree. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1164.3 Photogragh (left) and scanning electron images (centre, right) of dryKaolinite (SEM images were obtained using a Phillips XL30 micro-scope at the UBC Department of Earth, Ocean and Atmospheric Sci-ences). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1174.4 Photographs showing the fine-grained soil deposit at the Strong Pitmine in Abbotsford, BC. . . . . . . . . . . . . . . . . . . . . . . . 1184.5 Photograph of Redstone slurry (left) and scanning electron microscopyimages of dry Redstone powder (centre and right) (SEM images wereobtained using a Phillips XL30 microscope at the UBC Departmentof Earth, Ocean and Atmospheric Sciences). . . . . . . . . . . . . . 1194.6 Scanning electron microscopy images of dry Kaosand powder (SEMimages were obtained using a Phillips XL30 microscope at the UBCDepartment of Earth, Ocean and Atmospheric Sciences). . . . . . . 1204.7 Plasticity chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1214.8 Photographs showing roughness measurement of samples of the solidtest plates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1234.9 Laser confocal microscopy images of the solid test surfaces. . . . . 1254.10 Sample surface profiles of the solid test surfaces measured along agauge length of 50D50 (D50 of Kaolinite has been selected). . . . . 126xixList of Figures4.11 Sample surface profiles of the solid test surfaces measured along agauge length of 50D50 (D50 of Fraser River sand has been selected). 1274.12 Roughness size distribution of the solid test surfaces compared withgrain size distributions of test soils). . . . . . . . . . . . . . . . . . 1284.13 Experimentation tree for studying the effects of soil type, interfaceroughness, and effective normal stress level. . . . . . . . . . . . . . 1304.14 Experimentation tree for ascertaining the repeatability of tests con-ducted using the macro-scale interface direct shear apparatus. . . . 1324.15 Experimentation tree for studying the effect of rate of shear displace-ment on the large-displacement interface shear strength at low effectivenormal stresses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1334.16 Macro-scale interface direct shear test results for Fraser River sand onsand-blasted mild steel. . . . . . . . . . . . . . . . . . . . . . . . . 1384.17 Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test ofFraser River sand on sand-blasted mild steel. . . . . . . . . . . . . 1394.18 Macro-scale interface direct shear test results for Fraser River sand ongreen epoxy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1404.19 Abnormal increase in shear stress observed during testing Fraser Riversand against the green epoxy surface at 30 kPa effective normal stress(Test ID: FRS Green 30) caused by the entrapment of sand under-neath the south edge of specimen container. . . . . . . . . . . . . . 1414.20 Macro-scale interface direct shear test results for Fraser River sand ongrey epoxy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1434.21 Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test ofFraser River sand on grey epoxy coated steel. . . . . . . . . . . . . 1444.22 Variation of interface shear stress with shear displacement for FraserRiver silt on sand-blasted mild steel. . . . . . . . . . . . . . . . . . 146xxList of Figures4.23 Variation of pore-water pressure with shear displacement for FraserRiver silt on sand-blasted mild steel. . . . . . . . . . . . . . . . . . 1464.24 Macro-scale interface direct shear test results for Fraser River silt onsand-blasted mild steel. . . . . . . . . . . . . . . . . . . . . . . . . 1474.25 Variation of interface shear stress with effective normal stress for FraserRiver silt on sand-blasted mild steel. . . . . . . . . . . . . . . . . . 1484.26 Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test ofFraser River silt on sand-blasted mild steel targeting σ′n of 3 kPa. . 1494.27 Variation of interface shear stress with shear displacement for FraserRiver silt on grey epoxy. . . . . . . . . . . . . . . . . . . . . . . . . 1504.28 Variation of pore-water pressure with shear displacement for FraserRiver silt on grey epoxy. . . . . . . . . . . . . . . . . . . . . . . . . 1504.29 Macro-scale interface direct shear test results for Fraser River silt ongrey epoxy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1514.30 Variation of interface shear stress with effective normal stress for FraserRiver silt on grey epoxy. . . . . . . . . . . . . . . . . . . . . . . . . 1524.31 Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test ofFraser River silt on grey epoxy coated steel targeting σ′n of 15 kPa. 1534.32 Variation of interface shear stress with shear displacement for FraserRiver silt on green epoxy. . . . . . . . . . . . . . . . . . . . . . . . 1544.33 Variation of pore-water pressure with shear displacement for FraserRiver silt on green epoxy. . . . . . . . . . . . . . . . . . . . . . . . 1544.34 Variation of interface shear stress with effective normal stress for FraserRiver silt on green epoxy. . . . . . . . . . . . . . . . . . . . . . . . 1544.35 Macro-scale interface direct shear test results for Fraser River silt ongreen epoxy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155xxiList of Figures4.36 Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test ofFraser River silt on green epoxy coated steel targeting σ′n of 15 kPa. 1564.37 Variation of interface shear stress with shear displacement for Kaoli-nite on sand-blasted mild steel. . . . . . . . . . . . . . . . . . . . . 1574.38 Variation of pore-water pressure with shear displacement for Kaoliniteon sand-blasted mild steel. . . . . . . . . . . . . . . . . . . . . . . 1574.39 Macro-scale interface direct shear test results for Kaolinite on sand-blasted mild steel. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1584.40 Variation of interface shear stress with effective normal stress for Kaoli-nite on sand-blasted mild steel. . . . . . . . . . . . . . . . . . . . . 1594.41 Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test ofKaolinite on sand-blasted mild steel targeting σ′n of 15 kPa. . . . . 1604.42 Variation of interface shear stress with shear displacement for Kaoli-nite on grey epoxy. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1614.43 Variation of pore-water pressure with shear displacement for Kaoliniteon grey epoxy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1614.44 Macro-scale interface direct shear test results for Kaolinite on greyepoxy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1624.45 Variation of interface shear stress with effective normal stress for Kaoli-nite on grey epoxy. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1634.46 Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test ofKaolinite on grey epoxy coated steel targeting σ′n of 15 kPa. . . . . 1644.47 Variation of interface shear stress with shear displacement for Kaoli-nite on green epoxy. . . . . . . . . . . . . . . . . . . . . . . . . . . 1654.48 Variation of pore-water pressure with shear displacement for Kaoliniteon green epoxy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165xxiiList of Figures4.49 Macro-scale interface direct shear test results for Kaolinite on greenepoxy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1664.50 Variation of interface shear stress with effective normal stress for Kaoli-nite on green epoxy. . . . . . . . . . . . . . . . . . . . . . . . . . . 1674.51 Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test ofKaolinite on green epoxy coated steel targeting σ′n of 15 kPa. . . . 1674.52 Variation of interface shear stress with shear displacement for Strong-Pit FGS on sand-blasted mild steel. . . . . . . . . . . . . . . . . . 1684.53 Variation of pore-water pressure with shear displacement for Strong-Pit FGS on sand-blasted mild steel. . . . . . . . . . . . . . . . . . 1684.54 Macro-scale interface direct shear test results for Strong-Pit FGS onsand-blasted mild steel. . . . . . . . . . . . . . . . . . . . . . . . . 1694.55 Variation of interface shear stress with effective normal stress for Strong-Pit FGS on sand-blasted mild steel. . . . . . . . . . . . . . . . . . 1704.56 Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test ofStrong-Pit FGS on sand-blasted mild steel. . . . . . . . . . . . . . 1714.57 Variation of interface shear stress with shear displacement for Strong-Pit FGS on grey epoxy coated steel. . . . . . . . . . . . . . . . . . 1724.58 Variation of pore-water pressure with shear displacement for Strong-Pit FGS on grey epoxy coated steel. . . . . . . . . . . . . . . . . . 1724.59 Variation of interface shear stress with effective normal stress for Strong-Pit FGS on grey epoxy coated steel. . . . . . . . . . . . . . . . . . 1724.60 Macro-scale interface direct shear test results for Strong-Pit FGS ongrey epoxy coated steel. . . . . . . . . . . . . . . . . . . . . . . . . 1734.61 Variation of interface shear stress with shear displacement for Strong-Pit FGS on green epoxy coated steel. . . . . . . . . . . . . . . . . 1754.62 Variation of pore-water pressure with shear displacement for Strong-Pit FGS on green epoxy coated steel. . . . . . . . . . . . . . . . . 175xxiiiList of Figures4.63 Variation of interface shear stress with effective normal stress for Strong-Pit FGS on green epoxy coated steel. . . . . . . . . . . . . . . . . 1754.64 Macro-scale interface direct shear test results for Strong-Pit FGS ongreen epoxy coated steel. . . . . . . . . . . . . . . . . . . . . . . . 1764.65 Variation of interface shear stress with shear displacement for Red-stone on green epoxy coated steel. . . . . . . . . . . . . . . . . . . 1784.66 Variation of pore-water pressure with shear displacement for Redstoneon green epoxy coated steel. . . . . . . . . . . . . . . . . . . . . . 1784.67 Variation of interface shear stress with effective normal stress for Red-stone on green epoxy coated steel. . . . . . . . . . . . . . . . . . . 1794.68 Macro-scale interface direct shear test results for Redstone on greenepoxy coated steel. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1804.69 Variation of interface shear stress with shear displacement for Kaosandon sand-blasted mild steel. . . . . . . . . . . . . . . . . . . . . . . 1814.70 Variation of pore-water pressure with shear displacement for Kaosandon sand-blasted mild steel. . . . . . . . . . . . . . . . . . . . . . . 1814.71 Variation of interface shear stress with effective normal stress for Kaosandon sand-blasted mild steel. . . . . . . . . . . . . . . . . . . . . . . 1824.72 Macro-scale interface direct shear test results for Kaosand on sand-blasted mild steel. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1834.73 Laser confocal microscopy images of test surfaces before and afterbeing subjected to macro-scale interface direct shear testing. . . . 1844.74 Roughness size distribution curves of the solid test surfaces prior toand after being subject to macro-scale interface direct shear testing. 1854.75 Repeatability test results for Fraser River sand on grey epoxy. . . . 1874.76 Repeatability test results for Fraser River silt on grey epoxy for testsconducted targeting 3.0 kPa effective normal stress. . . . . . . . . 1874.77 Repeatability test results for Fraser River silt on grey epoxy for testsconducted targeting 30.0 kPa effective normal stress. . . . . . . . . 188xxivList of Figures4.78 Repeatability test results for Kaolinite on green epoxy for tests con-ducted targeting 3.0 kPa effective normal stress. . . . . . . . . . . 1884.79 Repeatability test results for Kaolinite on grey epoxy for tests con-ducted targeting 3.0 kPa effective normal stress. . . . . . . . . . . 1894.80 Repeatability test results for Kaolinite on grey epoxy for tests con-ducted targeting 15.0 kPa effective normal stress. . . . . . . . . . . 1894.81 Repeatability test results for Kaolinite on grey epoxy for tests con-ducted targeting 30.0 kPa effective normal stress. . . . . . . . . . . 1904.82 Macro-scale interface direct shear test results for Fraser River silt ongrey epoxy conducted to study rate effects. Target σ′n = 3.0 kPa. . 1924.83 Pore-water pressure dissipation time history of rate effect tests con-ducted for Fraser River silt on grey epoxy targeting σ′n = 3.0 kPa. 1934.84 Macro-scale interface direct shear test results for Fraser River silt ongrey epoxy conducted to study rate effects. Target σ′n = 30.0 kPa. 1944.85 Pore-water pressure dissipation time history of rate effect tests con-ducted for Fraser River silt on grey epoxy targeting σ′n = 30.0 kPa. 1954.86 Macro-scale interface direct shear test results for Kaolinite on greyepoxy conducted to study rate effects. Target σ′n = 3.0 kPa. . . . . 1964.87 Macro-scale interface direct shear test results for Kaolinite on greyepoxy conducted to study rate effects. Target σ′n = 15.0 kPa. . . . 1985.1 Drained large-displacement interface shear strength envelopes for FraserRiver sand. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2035.2 Drained large-displacement interface shear strength envelopes for FraserRiver silt. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2035.3 Variation of the drained large-displacement coefficient of interface fric-tion with effective normal stress level for Fraser River silt on the mildsteel and grey epoxy interfaces. . . . . . . . . . . . . . . . . . . . . 204xxvList of Figures5.4 Large-displacement interface shear strength data obtained from paststudies on various sand materials tested on steel surfaces of differentroughness (Ho et al., 2011; Jardine et al., 1993; Karimian, 2006; Reddyet al., 2000; Rinne, 1989; UESUGI et al., 1989). . . . . . . . . . . . 2075.5 Drained large-displacement interface shear strength envelopes for Kaoli-nite on the three solid surfaces based on macro-scale interface directshear data at low effective normal stresses. . . . . . . . . . . . . . 2095.6 Variation of the large-displacement drained secant interface frictionangle δ with effective normal stress for Kaolinite on the three solidsurfaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2105.7 Variation of the large-displacement drained interface efficiency factorwith effective normal stress obtained from past studies for a numberof fine-grained soils tested on solid surfaces of various roughness (Kuoet al., 2012; Meyer et al., 2015; Najjar et al., 2007a; Pedersen et al.,2003). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2135.8 Roughness size distribution of the solid test surfaces compared withgrain size distributions of test soils). . . . . . . . . . . . . . . . . . 2175.9 Variation of large-displacement drained interface efficiency factor withRa 50D50/D50. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2185.10 Variation of the large-displacement drained interface efficiency factorwith normalized average roughness obtained from past studies for anumber of coarse-grained and fine-grained soils tested on solid surfacesof various roughness (Lemos and Vaughan, 2009; Reddy et al., 2000;Rinne, 1989). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2215.11 Pore-water pressure dissipation time history of rate effect tests con-ducted for Fraser River silt on grey epoxy targeting σ′n = 3.0 kPa. 2235.12 Variation of the drained large-displacement interface friction anglewith rate of shear displacement. . . . . . . . . . . . . . . . . . . . 2246.1 Schematic representations of different soil-pipe interaction modes. . 231xxviList of Figures6.2 Schematic cross-sectional view showing conventional partially buriedaxial soil-pipe interaction test configuration. . . . . . . . . . . . . . 2336.3 Schematic cross-sectional representation of a conventional fully buriedaxial soil-pipe interaction test setup used when utilizing coarse-grainedsoils. (Note: this configuration is also applicable for partially buriedpipe testing) (Huber and Wijewickreme, 2014; Karimian, 2006). . . 2356.4 Schematic cross-sectional view showing the new axial soil-pipe inter-action test configuration. Note: this configuration is applicable forpartially buried pipe testing as well. . . . . . . . . . . . . . . . . . 2366.5 Photograph of the UBC Advanced Soil-Pipe Interaction Research (ASPIReTM)soil chamber before and after incorporating mobile walls. . . . . . 2376.6 Drawing showing perspective view of the axial soil-pipe interaction testconfiguration incorporating mobile walls that provide vertical degreeof freedom to the pipe. . . . . . . . . . . . . . . . . . . . . . . . . 2386.7 Drawing showing plan view of the axial soil-pipe interaction test setup. 2386.8 Photograph showing the interior of the soil chamber highlighting thePVC sheeting of the mobile walls and the seals between the mobilewalls and the stationary walls. . . . . . . . . . . . . . . . . . . . . 2406.9 Photograph of the UBC Advanced Soil-Pipe Interaction Research (ASPIReTM)soil chamber after incorporating mobile walls and installation of testpipe and instrumentation. . . . . . . . . . . . . . . . . . . . . . . . 2436.10 Photograph showing the epoxy coated steel test pipe. . . . . . . . 2446.11 Photographs showing the process of Kaolinite slurry preparation andpumping into the soil chamber. . . . . . . . . . . . . . . . . . . . . 2456.12 Photographs showing the process of wick drain installation prior toconsolidation of the soil bed under surcharge load. . . . . . . . . . 2456.13 Photographs showing the process of application of surcharge load forconsolidation of the soil bed under a normal stress of 3.0 kPa. . . . 2466.14 Photographs showing the 1.0 m thick Kaolinite test soil bed prior toplacement of the test pipe. . . . . . . . . . . . . . . . . . . . . . . 247xxviiList of Figures6.15 Photograph of the pressure sensors being installed on the test pipe. 2486.16 Photograph showing the diaphragm type pressure sensors mounted inthe custom made snesor ports. . . . . . . . . . . . . . . . . . . . . 2496.17 Schematic cross-sectional view of pressure sensors installed on the testpipe wall. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2506.18 Photograph showing the load cell connected to the test pipe and linearactuator rod used to measure the axial pulling force applied to the testpipe. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2516.19 Photograph showing the custom made data acquisition circuit boardused to monitor and record the sensor readings of the full-scale soil-pipe interaction experiments. . . . . . . . . . . . . . . . . . . . . . 2516.20 Photographs showing placement of test pipe. . . . . . . . . . . . . 2536.21 Time history of pressure sensor readings captured during the consoli-dation of the soil bed under the weight of the test pipe. . . . . . . 2546.22 Distribution of pore-water pressure and total normal stresses aroundpipe periphery during consolidation and during axial shearing (Note:level of water table is approximate and inferred based on PT13 data). 2566.23 Full-scale axial soil-pipe interaction test results of grey epoxy coatedtest pipe on normally consolidated kaolinite soil bed (Test-1). . . . 2586.24 Full-scale axial soil-pipe interaction test results of grey epoxy coatedtest pipe on normally consolidated kaolinite soil bed (Test-2). . . . 2596.25 Full-scale axial soil-pipe interaction test results of grey epoxy coatedtest pipe on normally consolidated kaolinite soil bed (Test-3). . . . 2606.26 Theoretical soil-pipe interface contact normal force calculation dia-gram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2626.27 Distribution of theoretical and measured total normal stresses aroundtest pipe during axial shearing of Test-1 (Pressure values indicated arein kPa). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264xxviiiList of Figures6.28 Use of macro-scale interface direct shear test data to determine thecoefficient of interface friction for soil-pipe interaction Tests 1, 2, and3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2696.29 Variation of effective normal stress with axial displacement obtainedfor the three full-scale axial soil-pipe interaction tests. . . . . . . . 2706.30 Comparison of measured and predicted axial forces for the full-scaleaxial soil-pipe interaction tests. . . . . . . . . . . . . . . . . . . . . 2716.31 PRCI recommended bounds for the adhesion factor α. Adopted fromPRCI (PRCI, 2009).(C-CORE, 2008; Cappelletto et al., 1998; Honeg-ger, 1999; Paulin et al., 1998; Rizkalla et al., 1996). . . . . . . . . . 2736.32 Adhesion factors obtained from the full-scale axial soil-pipe interactiontests compared with those recommended by PRCI (C-CORE, 2008;Cappelletto et al., 1998; Honegger, 1999; Paulin et al., 1998; PRCI,2009; Rizkalla et al., 1996). . . . . . . . . . . . . . . . . . . . . . . 274A.1 Estimated device friction ratio for various specimen footprint dimen-sions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308B.1 Schematic Diagram of Surface Characteristics. (ASME Standard B46.1-2002, 2003) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310B.2 Pictorial display of surface texture - Surface Roughness, Waviness andLay. (Bharat, 2001, p. 51) . . . . . . . . . . . . . . . . . . . . . . 311B.3 Examples of Nominal Profiles. (ASME Standard B46.1-2002, 2003) 313B.4 Filtering a Surface Profile. (ASME Standard B46.1-2002, 2003) . . 314B.5 Surface Profile Measurement Lengths. (ASME Standard B46.1-2002,2003) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314B.6 The Rt and Rmax Parameters. (ASME Standard B46.1-2002, 2003) 317B.7 Illustration for the Calculation of Average Surface Roughness Ra.(ASME Standard B46.1-2002, 2003) . . . . . . . . . . . . . . . . . 317B.8 Rt, Rp, and Rv Parameters. (ASME Standard B46.1-2002, 2003) . 317xxixList of Symbolsα Adhesion factor used in soil-pipe interaction analysis using the total stressbased α-approachδ Soil-pipe interface friction angle obtained at large shear displacementη Numerical factor corresponding to the extent and location of the drainageboundaries in a given geotechnical engineering laboratory experimentγ′ Average effective unit weight of soilλc wave length of the high-pass filter used to remove the waviness componentof a given surface profileµ Coefficient of soil-pipe interface frictionφ′ Friction angle of soil obtained at large shear displacementsφ Internal friction angle of soilφµ Inter-particle sliding friction (the intrinsic sliding friction angle of the soil)φcv Critical state friction angle or the constant volume friction angle of the soilφdr Friction angle due to particle rearrangement and damage aloneφg Friction induced by dilation and particle rearrangement (geometrical inter-ference friction angle)φmob Mobilized friction angle of the soilφd Friction angle due to dilation aloneφm Maximum internal friction angle of the soil(σ′n)av Average effective normal stress at the soil-pipe interfaceσnθ Total normal stress acting on the soil-pipe interface at an angle theta fromthe horizontal planeσ0 Atmospheric pressurexxxList of Symbolsσ′n average effective normal stress at the soil-solid interfaceσn Average total normal stress acting on the soil-solid interfaceτim Soil-pipe interface shear stress mobilized during the full-scale axial soil-pipeinteraction testτtotal Total soil-solid interface shear stress obtained from the macro-scale interfacedirect shear apparatus without correcting for the device frictionτi Average soil-solid interface shear stressAwc Planar area of the water surrounding the specimen container in the macro-scale interface direct shear apparatusAc Footprint area of the specimen container of the macro-scale interface directshear apparatusAi Soil-pipe or soil-solid interface contact areaAw Footprint area of the water layer in the water tank of the macro-scale interfacedirect shear apparatusCu Coefficient of uniformity of soilcv Coefficient of consolidation of the soilD Outer diameter of pipeD50 Mean particle diameter of a given soilemax Maximum void ratio of soil determined according to ASTM D4253emin Minimum void ratio of soil determined according to ASTM D4253f Soil pipe interface efficiency factor defined as the ratio of the tangent ofthe large-displacement soil-pipe interface friction angle to the tangent of thelarge-displacement internal friction angle of the soilFBLC Total normal force applied to the base plate of the macro-scale interfacedirect shear apparatusFdf Device friction force measure in the macro-scale interface direct shear appa-ratusFpull Puling force measured by the main load cell of the macro-scale interfacedirect shear apparatusFa Measured axial force in the full-scale axial soil-pipe interaction testxxxiList of SymbolsFi Average interface shear force measured in the macro-scale interface directshear apparatusfs Limit skin friction per unit area at the soil-pipe interfaceFv Total vertical force acting on the soil-pipe interface in the full-scale axialsoil-pipe interaction testGs Specific gravity of soilH Height of soil above the springline of pipeHD Shore-D hardness of the polymers tested in accordance with ASTM-D2240Hs Length drainage path based on one-dimensional consolidation theoryK0 Coefficient of lateral earth pressure at restLg Gauge length over which the surface roughness parameters are evaluatedLi Length of test pipe in contact with the soil bed in the full-scale axial soil-pipeinteraction testLL Liquid limit of soil determined according to ASTM D4318p Mean normal stress as defined in a typical triaxial testPSWH Water pressure of the static water head that is around the soil specimen in thewater tank of the macro-scale interface direct shear apparatus as measuredusing the pressure transducer (PTSH)PI Plasticity index of soil determined according to ASTM D4318PL Plastic limit of soil determined according to ASTM D4318pp11 Pore-water pressure sensor reading measured by pressure transducer PT-11in the full-scale axial soil-pipe interaction testpp12 Pore-water pressure sensor reading measured by pressure transducer PT-12in the full-scale axial soil-pipe interaction testpp13 Pore-water pressure sensor reading measured by pressure transducer PT-13in the full-scale axial soil-pipe interaction testq Deviatoric stress as defined in a typical triaxial testRa 50D50 Ra value of a given surface profile, evaluated over a gauge length correspond-ing to the 50×D50 values of the test soilRmax Maximum roughness parameter of a given surface profilexxxiiList of SymbolsRa Average roughness parameter of a given surface profileRn Normalized roughness parameter defined as the ratio of the maximum rough-ness parameter of a given surface profile to the mean particle diameter of agiven soilSt Soil sensitivity defined as the ratio of the undrained shear strength of undis-turbed soil to the undrained shear strength of remoulded soil at the samewater contentsu Undrained shear strength of undisturbed soiltf Time required for shear failure to occur in a given geotechnical laboratorytestTu Maximum axial shear force at the soil-pipe interface per unit lengthu Average pore-water pressure at the soil-solid interfaceuavg Average pore-water pressure acting on the soil-pipe interface in the full-scaleaxial soil-pipe interaction testUf Degree of pore-water pressure dissipation at failureW ′ Submerged weight of the pipe per unit lengthW Total weight of the test pipe in the full-scale axial soil-pipe interaction testxxxiiiAcknowledgementsI am grateful to my parents for their love.I would like to thank my wife, Anupama, for her love and constant support. Mywork here at UBC would not have been possible without her love and support.My gratitude goes out to my supervisor, Dr. Dharma Wijewickreme, who hasprovided me with support and guidance throughout this endeavour. Be it at nightor day, weekday or weekend, he was always available to consult with whenever Ihad the need to. His encouragement, guidance, and invaluable advice has helpedme immensely, not only to complete this research, but also to develop myself as aprofessional in the community. I am truly indebted and thankful to him for his moralsupport and the time he spared for me.I am grateful to Dr. Yogi Vaid, Dr. John Howie, Dr. Liam Finn, Dr. ErikEberhardt, Dr. Mahdi Taiebat, and Dr. Jonathan Fannin of the Department of CivilEngineering of UBC for their invaluable advice, guidance, and support. A very specialthanks goes out to Dr. Noboru Yonemitsu of the Department of Civil Engineering ofUBC for sharing his experience on and illuminating the beautiful world of electronicsand sensors. Also, I would also like to thank all the faculty members and the staffof the Department of Civil Engineering of UBC for providing an exceptional learningexperience.Special thanks goes out to my friends and colleagues Achala Soysa, Gaziz Siedali-nov, Ainur Siedalinova, Priyesh Verma, Ilaibi Omunguye, Drian Roos, Daniel Barnes,Judy Mei, Sadana Sadana Karannagoda Gamage, Anton Dabeet, Lalinda Weerasekara,Buddhika Samarakoon, Gamini Siriwardana, Lekha Samanmalee, Anajana Punchi-hewa, Thushara Jayasingha, Emalayan Vairavanathan, Ana Valverde, Primal Wije-sekera, and Nuwan Devapriya for their support. Thanks to all my colleagues andxxxivAcknowledgementssenior graduate students of the Geotechnical Group of the Department of Civil En-gineering for their valuable comments and moral support.Thanks are also due to Emilie Lapointe, Xavier Tellier, Della Anggabrata, El-liot Yii, Norman Richardson, Daniel Fortin, Michael Ross Ang, Eric Tam, MehrdadMirhossaini, Ji Lv, Mark Yang, Adebayo Adegbola, Youngha Hwang, April Graves,Kellie Liu, Muin Ahmed and Simon Chen for their contributions during laboratorytesting.Technical assistance of Messrs. Harald Schrempp, Bill Leung, Scott Jackson andJohn Wong of the Department of Civil Engineering Workshop is also deeply appreci-ated.xxxvChapter 1IntroductionEnergy pipelines (onshore and offshore) form an essential part of global lifeline in-frastructure, and they directly contribute to the improvement of society’s standard ofliving. The failure of oil and natural gas pipelines can have long lasting detrimentaleffects on the natural environment and economies. It is therefore important that suchsystems are designed with the intent of minimizing the risk of failure under operatingas well as extreme loading conditions.The design of buried and partially buried pipelines requires special attention withregard to geotechnical engineering aspects. This is because most pipeline systemstraverse vast distances with potential exposure to geotechnical hazards such as slopeinstability, landslides, rockfalls, earthquake-induced ground deformations and lique-faction of soil. Relative permanent ground deformation has been identified as one ofthe key geotechnical hazards that could adversely impact the performance of pipelines.Another commonly encountered problem is the rupture of pipelines due to thermalbuckling. Repeated thermal expansion and contraction cycles are very common dur-ing the operational life of pipelines. During such events, the soil resists the deforma-tion of the pipeline and this leads to a build-up of axial force along the length of thepipeline. The design of pipelines needs to take into account the stresses/strains thatare induced during such events. This essentially involves understanding the com-plex soil-pipe interaction problem that takes place when the pipelines are exposedto geotechnical hazards, while addressing the inherent uncertainties and variabilitiesassociated with soil and environmental conditions.The geotechnical analysis of the soil-pipe interaction problem, treated as a bound-ary value problem, requires proper modeling of the soil behaviour as well as the soil-pipe interface behaviour. Pipelines are usually buried under a soil cover of 0.6 to 1.51Chapter 1. Introductionm, and in offshore deep water projects they are often only partially buried (Brutonet al., 2006; Cathie et al., 2005; Dendani et al., 2008, 2007; Randolph et al., 2011;White and Cathie, 2010). Under these circumstances, the shear displacements at thesoil-pipe interfaces develop under relatively low soil effective stress levels. In addition,the deformations involved in typical soil-pipe interaction problems are associated withlarge strains within the soil as well as at the soil-pipe interface. This requires thatthe numerical models used in the designs are capable of capturing the large-strainsoil-pipe interaction behaviour that would take place under such low confining stresslevels.There is currently very limited experimental data available for the characterizationof the large-strain soil-pipe interface shear strength obtained at low effective stresslevels. This is mainly due to the difficulties encountered in conducting interfaceshear testing under relatively low stress levels. The characteristic response at soil-solid interfaces are obtained using common geotechnical devices such as ring-shearand direct shear apparatus with geotechnical laboratory testing typically conductedat confining stress levels above 50 kPa. At confining stress levels below 50 kPa,such devices tend to produce unreliable data as device compliance effects such asmechanical friction forces of the apparatus become emphasized, particularly whenthe objective is to characterize the interface behavior at effective normal stress levelssuch as 5 kPa encountered under offshore seabed conditions.While conventional ring-shear and direct shear apparatus are still being consideredto study soil-solid interfaces at relatively low stresses despite the aforementionedlimitation, it is prudent to develop devices specifically targeting testing at low stressesin view of producing reliable data. The Tilt-Table apparatus (Najjar et al., 2007b);the Cam-Shear apparatus developed at Cambridge University, UK (Kuo and Bolton,2009); and the Macro-Scale Interface Direct Shear Apparatus (Mark-I) developed atthe University of British Columbia, Vancouver, BC, Canada (Amarasinghe, 2013)are some examples of custom-made devices that are specifically designed for testingsoil-solid interfaces at low confining stresses.Evidence arising from such laboratory testing, accompanied by full-scale and2Chapter 1. Introductionreduced-scale physical model testing, drives the development of design guidelines andcodes of practice for pipeline design. With the recent expansion in offshore pipelinedevelopments, industry funded research projects such as the SAFEBUCK JIP (Brutonet al., 2007, 2008; Carr et al., 2006; White et al., 2011) have led to the updated DNVRecommended Practice F-114 - Pipe-Soil Interaction for Submarine Pipelines docu-ment (DNVGL, 2017), which emphasizes the importance of considering the possiblenonlinearity of the soil-pipe interface shear strength envelope that can be prevalentunder low confining stress conditions, and recommends the use of specialized labo-ratory test devices to characterize the soil-solid interface shear strength under suchconditions. Design guidelines such as the American Lifeline Alliance (ALA, 2001),and Pipeline Research Council International (PRCI, 2009), provide limited guidelineson the use of interface friction parameters under low confining stress conditions.With this background, there is a need for the expansion of the database of labo-ratory soil-pipe interface shear test data to improve design guidelines pertaining tosoil-pipe interaction under relatively low confining stress conditions. The investiga-tion of the variation of the soil-pipe interface shear behaviour with respect to thevariation of effective stresses is of particular importance in this regard.The soil-pipe interaction of pipelines is known to occur at displacement rates lead-ing to partially drained, and at most times close to fully undrained conditions. Thecurrent design codes use the total stress based α-approach when undrained conditionsare expected. However, there is motivation to use the effective stress based approachfor the soil-pipe interaction problem to reduce uncertainties that arise with the useof the α-approach.As mentioned previously, the current database of soil-pipe interface shear tests islimited, and there is a need for expansion in order to develop effective-stress basedframeworks for this problem. There is some evidence to indicate that the internal fric-tion angle of soils, examined at relatively low effective confining stresses, may be de-pendent on the magnitude of the effective confining stress (Bolton, 1986; Chakrabortyand Salgado, 2010; Fannin et al., 2005; Sture et al., 1998). There is limited datashowing that the large-strain soil-solid interface friction may also be dependent on31.1. Purpose of the Research Programthe effective confining stress level when tested at very low effective stresses (Pedersenet al., 2003). However, the current database of such tests is very limited, and moretesting is needed to reach a better understanding of the problem at hand.With this background, it was determined that there is a need to develop a suitablemethod for the experimental investigation of drained shear strength characteristics ofsoil-solid interfaces under the following conditions:1. under relatively low effective normal stresses (typically in the range of 3 kPa to30 kPa); and2. at relatively large shear displacements.This provided the impetus to undertake a novel experimental research programwhich in turn, led to the design, fabrication, and commissioning of a new macro-scaleinterface direct shear apparatus capable of performing large-displacement soil-solidinterface direct shear tests under normal stresses in the range of 3 to 30 kPa. Thedevice permitted conducting laboratory characterization of the behavior of severalsoil-solid interfaces under large-displacements and low effective normal stresses. Thefollowing sections concisely present the purpose and scope of the research work un-dertaken in detail.1.1 Purpose of the Research ProgramThe research program was designed with the intent of obtaining the interface fric-tional characteristics between soils of varying plasticity and various coated and/oruncoated steel surfaces under effective normal stresses ranging from 3 to 30 kPa.Specifically, the effects of: (i) soil type; (ii) magnitude of effective normal stress; and(iii) type of and/or roughness of the solid surface on the drained shear strength ofsoil-solid interfaces at large displacements was investigated. Emphasis was given tocharacterize the shear strength of various soil-solid interfaces at steady drained con-ditions typically achieved at sufficiently large shear displacements. The assessment41.2. Scope of the Research Programof interface behaviour at the soil-solid interfaces at small-strain (displacement levels)was not feasible in the devices and, as such, this aspect was not investigated.Primarily, the research is divided into three components:1. Development of a novel interface direct shear apparatus to address the challengesencountered when testing soil-solid interfaces under relatively low effective nor-mal stresses, and large displacements;2. Characterization of the drained large-displacement interface shear strength atthe soil-solid interfaces - with respect to a number of soils (fine-grained andcoarse-grained) - tested against solid surfaces of various roughness - under verylow effective normal stresses and large shear displacements to improve our un-derstanding of the effects of soil type, effective normal stress, and solid surfaceroughness; and,3. Full-scale laboratory physical model testing of axial soil-pipe interaction underpartially buried conditions.It is expected that this research has the potential to advance the knowledge of andcontribute to the geotechnical engineering profession with respect to the key concernsoutlined earlier.1.2 Scope of the Research ProgramThe specific key tasks undertaken in this research are as follows:1. Design, fabrication, and commissioning of a novel macro-scale interface directshear apparatus capable of performing soil-solid interface direct shear tests un-der the following conditions:ˆ testing of soil-solid interfaces under drained, large-displacement, and rela-tively low normal stress (in the range of 3 to 30 kPa) conditions;ˆ capability to test fine-grained as well as coarse-grained soils against varioussolid surfaces commonly encountered in practice;51.3. Organization of the Thesisˆ capability to introduce a maximum interface shear displacement of 1.0 m;ˆ capability to apply very low shear displacement rates to ensure full-drainageconditions at the soil-solid interface;ˆ provide means of monitoring pore-water pressure at the soil-solid interfaceto allow determination of the interface shear strength in terms of effectivestresses.2. Conduct interface direct shear tests to study the effect of soil type, type ofinterface, and magnitude of effective normal stress on the shear strength ofsoil-solid interfaces.3. Design, fabrication, and commissioning of a novel full-scale axial soil-pipe in-teraction testing setup that would allow for conducting experiments under thefollowing conditions:ˆ capability to test axial shear of NPS18 steel pipes of 3 m length partiallyburied in fine-grained as well as coarse-grained soils;ˆ capability to apply a maximum axial pipe displacement of 0.5 m;ˆ capability to apply very low shear displacement rates to the pipe to ensurefull-drainage conditions within the soil specimen;ˆ provide means of monitoring excess pore-water pressure at the soil-pipeinterface during shearing.4. Conduct full-scale axial soil-pipe interaction testing to study the axial soil-pipe interaction behaviour under partially buried conditions and large axialdisplacements.The following section presents the organization of the thesis.1.3 Organization of the ThesisChapter I presents an introduction to the thesis. An outline of the proposed workcarried out in the research program and the basis for the proposed work is presented.61.3. Organization of the ThesisChapter II presents a literature review covering important geotechnical aspects ofsoil-pipe interaction design and different methods in use for the determination of thesoil-pipe interface shear strength. The chapter is aimed at demonstrating the needfor the development of new laboratory test methods specifically applicable to soil-pipe interaction under relatively low confining stresses. An overview of conventionalinterface shear test devices is presented and the limitations of the devices with re-spect to their use under low confining stress levels and large shear displacements arehighlighted.Chapter III presents details of the newly developed macro-scale interface direct shearapparatus. The principle of operation of the device, function of each component ofthe device, and its important features are described. The chapter is concluded witha detailed description of how each type of soil specimen is prepared for testing.Chapter IV presents details of the macro-scale interface direct shear tests that wereconducted on the new device, and the test data are presented.Chapter V presents a discussion of the test results drawing upon and comparingwith findings from other researchers. Important findings arising from the presentresearch are highlighted.Chapter VI presents the details of the newly developed large-scale axial soil-pipeinteraction test setup, the details of the full-scale axial soil-pipe interaction physicalmodel tests that were conducted, and the test results.Chapter VII summarizes the work undertaken and highlights the important find-ings and contributions arising from the research. Recommendations for future workand improvements are presented.7Chapter 2Literature ReviewThis chapter covers past studies of soil-solid interface shear strength characterizationas related to pipeline design, and highlights key areas that require further investiga-tion. Specifically, this chapter aims to make the case for the requirement to conductfurther research on characterizing soil-solid interface shear strength under three im-portant conditions: (i) drained conditions to allow for characterizing the interfaceshear strength in terms of effective stresses; (ii) at relatively low confining stresses (3to 30 kPa); and (iii) at relatively large shear displacements.First, the current state of practice in the design of onshore and offshore pipelinesis provided, and areas that warrant further research and development are highlighted.Next, a brief outline of the concepts of internal friction angle of coarse-grained andfine-grained soils is presented in order to define the important geotechnical soil shearstrength parameters as this understanding is important before delving into the shearstrength of soil-solid interfaces. The research problem of the present thesis is presentedin reference to the past research work that has already been carried out and furtherwork that is required to improve future practice. The concluding sections of thischapter cover laboratory testing methods of soil-pipe interface friction under loweffective stress levels and large shear displacements.2.1 Geotechnical Aspects in the Design ofPipelines Subjected to Relative DeformationExcept for the simplest of cases, the design of pipelines subjected to soil loadingrequires nonlinear finite element methods of analysis. As per guidelines such as theAmerican Lifeline Alliance (ALA) (ALA, 2001), American Petroleum Institute (API82.1. Geotechnical Aspects in the Design of Pipelines Subjected to Relative DeformationRP2GEO)(API, 2011) and the more recent Pipeline Research Council International(PRCI) (PRCI, 2004, 2009), and the DNVGL Recomended Practice F-114 Pipe-SoilInteraction for Submarine Pipelines (DNVGL, 2017), soil-pipe interaction problemsare widely modeled using pipe elements attached to soil spring elements, where thepipe elements are used to aid in analyzing stresses in the pipeline wall, while the soilresponse is idealized by discrete, nonlinear soil springs representing soil loads actingon the pipeline along three orthogonal axes (axial, vertical, and lateral).SoilPipe𝑥 𝑦𝑧𝑥 𝑦𝑧Spring forceDisplacementTPQPQT(a) (b) (c)Figure 2.1: Idealized soil-spring models used in analysis of soil-pipe interaction. (a)Continuum model. (b) Soil-spring model. (c) Soil-spring load-displacement response.There is much uncertainty related to the estimation of soil spring properties. Thisuncertainty is mostly related to estimation of the soil-pipe interface shear strengthparameters. Except in the case of offshore pipelines, because pipelines are typicallylocated above the water table, the uncertainty in estimating the soil-pipe interfaceshear strength parameters for fine-grained soils is further complicated as the strengthof partially saturated soils is not well defined in geotechnical engineering practice(PRCI, 2004, 2009).It is commonplace to use bi-linear elastic-perfectly plastic models for soil-springs.In such models, the axial shear strength of a given soil-pipe interface is conventionallyestimated based on either the empirical α-method, where the axial shear strength isrelated to the undrained shear strength su of the soil at the soil-pipe interface usingan empirical adhesion factor α, or the effective stress based β-method, where theaxial shear strength is related to the effective interface friction angle. Both methods92.1. Geotechnical Aspects in the Design of Pipelines Subjected to Relative Deformationare equally important in practice, and past studies have primarily focused on: (i)developing α values based on field or laboratory physical model experiments; and(ii) developing correlations of the coefficient of interface friction with effective normalstress and soil-pipe interface properties.In fine-grained soils, depending on the rate of relative soil-pipe displacement, theaxial soil-pipe interaction response may be undrained, drained or partially drained.Therefore, design and assessment is typically based on separate assessments of drainedand undrained axial resistance, with a range bounding both cases (PRCI, 2004, 2009).In the following sections, a discussion of the total stress approach (i.e., α-method)and the effective stress approach (i.e., β-method) recommended in guidelines is pre-sented, and the impetus for the work undertaken in the present research project ishighlighted.2.1.1 Axial soil-pipe interface shear strength: total stressapproach (α method)The total stress approach (α-method) used to evaluate axial soil loads on pipelinesis consistent across many current guidelines such as the American Lifeline Alliance(ALA) (ALA, 2001), the Pipeline Research Council International (PRCI) (PRCI,2004, 2009), and the DNVGL F-114 Recommended Practice (DNVGL, 2017). Esti-mation of soil axial forces has been established based on empirical observations similarto the use of the (α-method) in pile design. In this case, the limit skin friction perunit area fs at the soil-pipe interface is expressed as given by Eq. 2.1:fs = αsu (2.1)where, su is the undrained shear strength of the soil, and α is an empirical adhesionfactor that is determined by laboratory or field testing.The American Lifeline Alliance (ALA) (ALA, 2001), and the Pipeline ResearchCouncil International (PRCI) (PRCI, 2004, 2009) guidelines recommend using Eq. 2.2for estimating the axial soil-pipe shear resistance, while using Fig. 2.2 for selecting102.1. Geotechnical Aspects in the Design of Pipelines Subjected to Relative Deformationthe appropriate α value.Tu = piDαsu (2.2)where, D is the pipe outside diameter,su is the undrained shear strength of the soil backfill,α is the adhesion factor (Fig. 2.2) defined by the lower bound α = 0.7(0.12σ0su)0.8≤ 1and upper bound α = 0.5(0.55σ0su)0.8≤ 1.Figure 2.2: PRCI recommended bounds for the adhesion factor α. Adopted fromPRCI (PRCI, 2009).(C-CORE, 2008; Cappelletto et al., 1998; Honegger, 1999; Paulinet al., 1998; Rizkalla et al., 1996).In the case of offshore pipelines, the α method is extensively used for determiningaxial shear resistance of soil-pipe interfaces when the rate of shear displacement ofthe pipeline provides little time for excess pore-water drainage in fine-grained soils112.1. Geotechnical Aspects in the Design of Pipelines Subjected to Relative Deformation(Randolph and Gourvenec, 2011; White and Cathie, 2010). Appropriate values of theadhesion factor depend on whether the peak or residual axial resistance is required.Laboratory tests are recommended for the specific soil and coating combinationsunder consideration in design. In very soft clays, in the absence of specific data,and subject to the roughness of the interface, the adhesion factor is recommended tobe taken as 1 for the peak resistance and is related to the soil sensitivity St for theremolded strength condition (α = 1/St) (Cathie et al., 2005).When calculating axial and upward vertical soil spring parameters, the soil prop-erties representative of the backfill are required. Soil spring parameters for lateraland bearing springs are generally based on the native soil properties. In practice,when the pipeline is buried in a trench, the backfill material may be carefully cho-sen or designed specifically to account for the geohazard conditions prevalent at thesite (PRCI, 2009). One approach of reducing soil loads on pipelines is the use of ashallow trench with a loose granular backfill. For example, the use of sand as backfillis common in trench designs at earthquake fault crossings (PRCI, 2004). In areaswhere there are no major geohazards the backfill is likely to be the same soil materialthat was excavated from the site. In the case of fine-grained backfill, the soil maybe compacted to design specifications. In difficult terrain conditions such as muskegareas where the soil is primarily composed of highly organic peat, the trenching andinstallation is typically carried out during winter when the ground is frozen. Undersuch conditions, the backfill material may not be compacted. There are often cases,especially at river crossings, where pipelines are installed using horizontal directionaldrilling techniques. In such cases, the in-situ soil properties would be of importancein defining soil spring parameters. It is important to understand that obtaining soilproperties representative of the backfill material may not be practical, and there canbe notable uncertainty and variability in the estimated backfill soil parameters. Theestimated undrained shear strength of the soil would depend on many factors such asrate of shearing and the type of in-situ or laboratory test method used.In practice, a range of in-situ and laboratory tests may be prescribed to char-acterize the soil properties along the route of the pipeline. Due to the availability,122.1. Geotechnical Aspects in the Design of Pipelines Subjected to Relative Deformationpracticality and cost effectiveness, the Standard Penetration Test (SPT) is widelyused for this purpose and is the most common in-situ test in North America. Incritical areas along the pipeline route, such as at the bends of the pipeline or wherespecific geohazards have been identified, a combination of Field Vane Shear Testing(FVT), Cone Penetration Testing (CPT) and laboratory testing of relatively undis-turbed samples may be prescribed.The SPT is generally used to estimate geotechnical properties of coarse-grainedsoils, and it is generally accepted that owing to the nature of the test, estimationof undrained shear strength properties of fine-grained soils using SPT data is notreliable (ASTMD1586/D1586M-18, 2018; CGS, 2006). Many factors such as sam-pling and drilling method, sampler type, rod length and energy efficiency of the SPThammer drop system affect the SPT results (ASTMD1586/D1586M-18, 2018; Howieet al., 2003; Robertson et al., 1983; Robertson, 1986; Schmertmann, 1979; Sy andCampanella, 1991). However, it is not uncommon in practice for the use of the SPTto characterize properties of fine-grained soils based on past local experience andengineering judgment.The FVT is the most common method of determining the in-situ undrained shearstrength. In the case of FVT, a typical approach in practice would be the use of theFVT to obtain the peak and remolded shear strength properties of the native soil(i.e., done in-situ prior to the excavation of the trench), where the remolded strengthis then assumed to be representative of the disturbed backfill material. Based onexperience and engineering judgment, ranges for these parameters would then bespecified for defining soil spring parameters to account for the large uncertainty. Thepipeline design engineer would then use the provided upper and lower bounds of thesoil spring parameters to carryout the pipe stress-strain analyses.The CPT (and more notably CPTu - where pore-water pressure is continuouslymeasured) is becoming more popular as an in-situ testing method, and is useful forestimating the in-situ undrained shear strength (Robertson and Campanella, 1983;Schmertmann, 1978). Interpretation of the undrained shear strength is based onbearing capacity theory. The bearing capacity factor for the CPT is typically obtained132.1. Geotechnical Aspects in the Design of Pipelines Subjected to Relative Deformationfrom empirical correlations. The bearing capacity factor is dependent on the soiltype, and hence, a soil specific calibration of the CPT may be required where pastexperience or literature is not available (Robertson and Campanella, 1983). The mostimportant advantage of the CPT over the SPT is the high repeatability in fine-grainedsoils. The CPT is considered a reliable method of determining soil stratigraphy andsoil behaviour type characterization (Robertson, 2010, 2012).When relatively undisturbed shelby tube samples are available, it is commonpractice to conduct unconfined compression testing to determine the undrained shearstrength. This is especially the case where project budget is limited. Given the highprobability of sample disturbance (Ladd and DeGroot, 2003) even when high qualityshelby tube samples are available, the consolidated undrained triaxial test may bemore appropriate as opposed to the unconfined compression test.The estimation of the undrained shear strength of the backfill and in-situ soilmaterials will inherently be subject to considerable uncertainty. Hence, the use ofa combination of in-situ and laboratory test methods may be required to narrowthe variability of the estimated soil properties. The α method relies heavily on full-scale laboratory physical model test data or field test data for developing appropriateα values for a given design. Given the complex nature of the soil-pipe interactionproblem owing to the multitude of variables involved, full-scale laboratory physicalmodel testing is sought after for producing reliable information. While small-scaletests can be useful for certain applications, they tend to produce errors related toscale effects. Also, full-scale soil-pipe interaction test data is invaluable in calibratingnumerical models that are used for soil-pipe interaction design. The number of suchexperiments that have been incorporated in guidelines is still quite limited, and hencethere is much uncertainty in selecting appropriate design α values. This is especiallytrue with regard to the limited availability of full-scale soil-pipe interaction test datafor fine-grained soils.Given the importance of full-scale laboratory physical modeling for reducing thisuncertainty, the research work undertaken in the present thesis was tailored to producea new method of conducting full-scale axial soil-pipe interaction tests in fine-grained142.1. Geotechnical Aspects in the Design of Pipelines Subjected to Relative Deformationsoils.2.1.2 Axial soil-pipe interface shear strength: effectivestress approachThe effective stress approach (β method) was originally developed for use in estimatingthe skin friction of piles (Burland, 1973), and has been adopted for the estimation ofaxial soil-pipe interface friction based on a fundamental effective stress approach. Inthe case of fully buried pipelines in the at rest condition, the average effective normalstress on the pipe surface is estimated using Eq. 2.3, and based on the assumptionthat failure occurs at the soil-pipe interface, the maximum axial soil load on thepipeline per unit length Tu is estimated using Eq. 2.4.(σ′n)av =(1 +K02)γ′H (2.3)where, K0 is the coefficient of lateral earth pressure at rest,γ′ is the average effective unit weight of the soil, andH is the height of soil above the springline of the pipe.Tu = piD(σ′n)av tan(δ) = piD(1 +K02)γ′H tan(δ) (2.4)where, D is the external diameter of the pipe, andδ is the soil-pipe interface friction angle and is often presented in guidelines in theform of the interface efficiency factor f = δ/φ.In the case of offshore pipelines, when freely draining soils are involved or whenthe shearing rate is known to be slow enough for drained conditions to dominate, theβ method is adopted. In this case the maximum axial shear force at the soil-pipeinterface per unit length is given as:Tu = µW′ = tan(δ)W ′ (2.5)where, W ′ is the submerged weight of the pipe per unit length, and152.1. Geotechnical Aspects in the Design of Pipelines Subjected to Relative Deformationδ is the coefficient of friction for the soil-pipe interface.Equation 2.4 has been widely used and is recommended in guidelines such as ASCE(1984) “Guidelines for the Seismic Design of Oil and Gas Pipelines”, American Life-line Alliance (ALA) ALA (2001) “Guidelines for the Design of Buried Steel Pipe”,and the Pipeline Research Council International (PRCI) guidelines PRCI (2004, 2009)“Guidelines for the Seismic Design and Assessment of Natural Gas and Liquid Hy-drocarbon Pipelines”.The interface efficiency factors recommended by the American Lifeline Alliance(ALA) (ALA, 2001), and the Pipeline Research Council International (PRCI) (PRCI,2004, 2009) guidelines are presented in Table 2.1.Table 2.1: Interface efficiency factors recommended by the American Lifeline Alliance(ALA) (ALA, 2001), and the Pipeline Research Council International (PRCI) (PRCI,2004, 2009) guidelinesPipe coating f = δ/φm*Concrete 1.0Coal tar 0.9Rough steel 0.8Smooth steel 0.7Fusion bonded epoxy 0.6Polyethylene 0.6* Note: φm is defined as the maximum internal friction angle of the soil.In the case of partially buried offshore pipelines, various guidelines are availablefor the selection of an appropriate value for µ but are not yet comprehensive enoughto cover all possible design situations. Three such guidelines are as follows:1. (RP2A-WSD-API, 2000)- µ = tan(φ′ − 5◦)2. (B.S.8010:part3, 1993)- Non-cohesive soil: µ = 0.55− 1.2,- Cohesive soil: µ = 0.3− 1.03. (Bruton et al., 2006) -For fully drained conditions µaxial = tan(δ), where δ is162.1. Geotechnical Aspects in the Design of Pipelines Subjected to Relative Deformationthe angle of friction for the soil-pipe interface.The use of Eq. 2.4 for prediction of axial soil loads on pipelines carries a certaindegree of uncertainty. This uncertainty is primarily related to the variability of theparameters (σ′n)av, and δ. Work done by researchers have focused on full-scale physicalmodel testing of pipes in coarse-grained soils, under loose and dense conditions, toevaluate the robustness of using Eq. 2.4 for pipeline design. In particular, the work byPaulin et al. (1998), and more recently by Karimian (2006) have shown that in densesands, the axial soil loads on pipelines can be over 2 times greater than that estimatedusing Eq. 2.4 as the effect of dilation is not considered. Using centrifuge model testingand numerical simulations, Daiyan et al. (2011) have investigated the effect of axialand lateral soil-pipe interaction coupling effects on the axial soil loads on pipelinesand have shown that coupling effects can be large. Also, Daiyan et al. (2011) haveshown that the weight of the pipeline can affect the axial soil-pipe interaction forcesnotably. These effects are not explicitly accounted for in current guidelines.The ability to determine δ for a given interface to a reasonable degree of accuracyis critical to effectively representing soil loads on a pipeline. In particular, there is aneed to determine the coefficient of friction for soil-pipe interfaces at relatively loweffective normal stresses typically experienced by pipelines. Experimental data on theeffect of normal stress on the value of δ at relatively low effective normal stresses andlarge displacements is scarce. Such data for soil-pipe interaction in fine-grained soilsis further more limited. However, with the recent expansion in the development of oiland gas pipelines in the offshore environment where the seabed conditions are mostlyfine-grained, there has been a number of publications that focus on characterizingthe axial shear strength of soil-pipe interfaces under low confining stresses and largedisplacements in saturated fine-grained soils.Flowlines and pipelines used for the transportation of materials across the seabedat depths exceeding 2 km from surface of the ocean are generally installed by layingthe pipeline on the seabed without burial in a trench. Burying or restraining apipeline at such depths is often not economical (Bruton et al., 2006; Cathie et al.,2005; Dendani et al., 2008, 2007; Randolph et al., 2011; White and Cathie, 2010).172.1. Geotechnical Aspects in the Design of Pipelines Subjected to Relative DeformationBurial of a pipeline may be feasible at relatively shallow depths.An important design consideration in the design of offshore pipelines is the miti-gation of global axial movement of the pipeline on the seabed. An offshore pipelinetypically undergoes several start-up shut-down cycles during its operational life. Withevery start-up, the temperature of the pipeline rises and the pipeline expands. Atshut-down, the temperature drops, and the pipeline contracts. This cyclic loadingproduces a net axial movement of the pipeline due to asymmetry in the frictionalresistance imposed by the seabed or due to some other form of asymmetry such aseffects of seabed slope or relatively higher riser tension at one end. This phenomenonis called ‘pipeline-walking’ (Bruton et al., 2008, 2010; Carr et al., 2006). If not ac-counted for in the design, uncontrolled walking can cause failures at connections.Mitigation of global pipeline walking is commonly achieved by the use of pipelineanchors (Bruton et al., 2010; Perinet et al., 2011).Offshore pipelines are typically operated at high temperatures and pressures. Thehigh temperature causes the pipeline to expand and this leads to the accumulationof axial compressive stresses. Naturally, the tendency of the slender pipeline is torelieve this stress by buckling. Uncontrolled buckling can often adversely affect theintegrity of the pipeline due to excessive stresses in the pipe wall or at end connections(Bruton et al., 2006; Cathie et al., 2005; Dendani et al., 2008; Nielsen et al., 1990;Perinet et al., 2011). Therefore, it is common practice to restrain the pipeline againstsuch movements either by using anchors, rock-dumps, stabilization mattresses or bytrenching (Bruton et al., 2006).However, in deep water applications (at depths greater than 2 km from the seasurface), such measures can present a major financial cost. Generally, a pipeline isa forgiving structure that can tolerate relatively large deformations (Bruton et al.,2006). Therefore, there is currently a move towards designing pipelines such that thestresses are relieved by allowing the pipeline to buckle in a ‘controlled’ manner so thatthe need for restraining measures is minimized (Bruton et al., 2006; Hoor et al., 2014;Randolph et al., 2011). A preferred method is the use of buckle triggering mechanismsthat incorporate sleepers and buoyancy elements to ‘control’ the buckling by forcing182.1. Geotechnical Aspects in the Design of Pipelines Subjected to Relative Deformationthe pipeline to buckle at predetermined locations (Bai et al., 2009; Hoor et al., 2014;Lim et al., 2014; Sun et al., 2012).As developments move to deep water, the use of mitigation measures such asrestraints and buckle initiators can become financially costly and other less expensivetechniques such as snake-lay, where the pipeline laying is carried out in such a way thatartificial out of straightness is introduced with bend radii to help control the bucklinglocations, are sought after(Cooper et al., 2014; Hooper et al., 2004; Preston et al.,1999; Rundsag et al., 2008; Wang et al., 2015). But, given the degree of uncertaintyand variability of geotechnical parameters used in estimating the soil-pipe interactionresponse (Sicilia et al., 2014; Westgate et al., 2016; White et al., 2014), the design ofa pipeline to undergo ‘controlled’ buckling is not a trivial task. Therefore, it is oftenconsidered safer to provide mitigation measures in the form of restraints and buckleinitiators (Hakim et al., 2016; Perinet et al., 2011).Pipeline walking and all other soil-pipe interaction design considerations are pri-marily controlled by the axial friction mobilized at the soil-pipe interface. The soil-pipe interface friction is affected by: (i) the normal stress level at the interface; (ii)the roughness of the pipeline surface; (iii) the soil type; and (iv) drainage conditionsduring movement of the pipeline (Boukpeti and White, 2017; Boylan et al., 2014;Eid et al., 2014; Ganesan et al., 2013; Kuo and Bolton, 2014; Kuo et al., 2012; Naj-jar et al., 2007a, 2003; Taha and Fall, 2014; Wijewickreme et al., 2013). In offshorepipeline applications, the normal stress at the soil-pipe interface is typically in therange of 1 to 10 kPa (Bruton et al., 2009; Najjar et al., 2007a; White et al., 2007).Given that the shear response of a given fine-grained soil and the soil-pipe interfacecan be dependent on the level of effective normal stress at the interface (Najjar et al.,2007a; Pedersen et al., 2003), it is important that laboratory testing is carried out atstress levels similar to those encountered in the field.Laboratory testing for conventional geotechnical engineering works is not typicallycarried out at such low stresses. Also, standard geotechnical laboratory equipment arenot designed to be reliable for testing at such low stress levels. Therefore, researchwork on soil-pipe interaction is typically carried out with the use of custom-made192.2. Internal Friction Angle of Soils at Large Strains and Low Stressesapparatus (Boukpeti and White, 2017; Eid et al., 2014; Ganesan et al., 2013; Kuoand Bolton, 2014; Kuo et al., 2012; Najjar et al., 2007a, 2003; Wijewickreme et al.,2013).As such, with particular reference to soil-pipe interaction research, the databaseof such low-stress soil-solid interface shear strength laboratory tests is very limited.Therefore, there is much uncertainty in the selection of lower and upper bounds fordesign friction factors. There is a need for laboratory test data of soil-solid interfaceshear strength tests conducted at relatively low effective normal stresses and largedisplacements.2.2 Internal Friction Angle of Soils at LargeStrains and Low StressesGiven that the large-strain friction angle of soil-structure interfaces is dependent onthe internal friction angle of the soil, it is important to have some understandingof the internal friction angle of soils when subjected to large strains. As such, abrief discussion of the body of knowledge on this topic is presented in the followingsub-sections.2.2.1 Friction angle of granular materialsThe mobilized shear resistance of a granular material subjected to drained shearingis known to depend on two components of shear resistance: inter-particle sliding fric-tion (the intrinsic sliding friction, φµ); and friction induced by dilation and particlerearrangement (commonly referred to as geometrical interference friction angle φg)(Rowe, 1962; Sadrekarimi and Olson, 2011). Geometrical interference is further di-vided into resistance mobilized due to dilation (φd) and resistance mobilized due toparticle rearrangement and damage (φdr) (Rowe, 1962; Sadrekarimi and Olson, 2011).202.2. Internal Friction Angle of Soils at Large Strains and Low StressesThis relationship is given by Eq. 2.6 and is schematically presented in Fig. 2.3.φmob = φµ + φg = φµ + φd + φdr (2.6)where, φmob is the mobilized friction angle of the soil.Figure 2.3: Components of shear resistance of coarse-grained soils (After Rowe(1962)).During initial stages of shearing, an increase in mobilized angle of friction withincreasing shear strain is observed until a peak value is reached. Further shearingbeyond the peak may either lead to a reduction or no change in the mobilized angle offriction. In either case, with continued shearing, the soil ultimately tends to a constantangle of friction at large strains where deformation occurs indefinitely without achange in void ratio or stress state. This ultimate state is generally referred to as theconstant volume or critical state. The friction angle at this critical state is referredto as the critical state friction angle or the constant volume friction angle (φcv).At the critical state, all contraction, dilation and particle rearrangements cometo a halt and the soil undergoes shearing under constant volume and constant stressconditions (Negussey et al., 1988). Typically, at very low densities and high confin-212.2. Internal Friction Angle of Soils at Large Strains and Low Stressesing stresses the critical state is reached via contraction of the soil volume. However,at most commonly encountered densities and stresses, soils will undergo some initialcontraction followed by dilation until finally reaching the critical state. In this case,the point of maximum contraction is considered to be a transient constant-volumestate (Negussey et al., 1988). It has been shown using drained triaxial compressionand extension tests conducted on Erksak sand that a unique relationship exists be-tween the peak friction angle φ′p and dilation rate at peak (dv/da)max regardlessof the relative density, confining stress at failure, stress path, and mode of loading(Vaid and Sasitharan, 1992). This unique relationship can be clearly observed in thedata presented by Vaid and Sasitharan (1992) in Fig. 2.4. A similar finding has beenreported by Bishop (1971) using conventional triaxial compression tests conductedon undisturbed and remolded soil specimens. It is also notable that as suggested byBishop (1971), and further explored by Vaid and Sasitharan (1992), the extrapolationof the φ′p− (dv/da)max relationship to zero maximum dilatancy rate should providea good estimate of the constant volume friction angle of the soil.𝜑𝑐𝑣′Figure 2.4: Relationship between peak friction angle and maximum rate of dilatancy(dv/da)max (After Vaid and Sasitharan (1992)).222.2. Internal Friction Angle of Soils at Large Strains and Low StressesFigure 2.5: Typical undrained trixial compression test results showing the steady stateand phase transformation lines (After Negussey et al. (1988); Vaid et al. (1990)).When saturated granular materials are subjected to undrained shearing (i.e.,shearing under constant volume), the tendency of the soil to change in volume dueto shearing induces excess pore-water pressures. The initial tendency of the soil tocontract causes excess pore-water pressure to rise until a constant pore-water pressurestate (often referred to as phase transformation) is reached. Further shearing induces232.2. Internal Friction Angle of Soils at Large Strains and Low Stressesa reduction in the excess pore-water pressure due to the tendency of the soil to dilate.The soil undergoes shearing ultimately reaching a steady state of excess pore-waterpressure at large strains (Negussey et al., 1988; Vaid et al., 1990). At the steady statethe soil shearing continues under constant stress and constant pore-water pressureconditions. It has been shown that the friction angle at phase transformation and atsteady state are identical and is unique for a given granular material (Sadrekarimiand Olson, 2011; Vaid and Chern, 1985). There is also evidence indicating that whilethe critical state and the phase transformation friction angles obtained from triaxialtesting are unique for a given soil material, the two friction angle values may notalways be identical (Lade and Ibsen, 1997). Lade and Ibsen (1997) have shown thatwhile some sands can exhibit identical critical state and phase transformation fric-tion angles under triaxial testing, some other sands can exhibit critical state frictionangle values different to the phase transformation friction angle values under triaxialtesting. However, the data presented by Lade and Ibsen (1997) shows considerablescatter on the order of ±5°. Characterization of the critical state friction angle inthe laboratory is not an easy task as the exact void ratio conditions within the shearzone are rarely known.The mobilized friction angle at the critical state is a lower bound of the shear re-sistance of a given granular material. The work by Negussey et al. (1988); Uthayaku-mar and Vaid (1998); Vaid et al. (1990); Verdugo and Ishihara (1996); Wijewickreme(1986) shows that the critical state friction angle is a unique property for a givengranular material and that it is not affected by mode of loading, direction of principalstresses, initial consolidation stress, and initial void ratio.There does not seem to be a consensus on the most appropriate method to beused to estimate the critical state friction angle. Critical state friction angle valuesreported in literature are mostly based on triaxial tests. Laboratory characterizationof the internal friction angle of soils can be carried out using a range of devices.The simple shear test, direct shear test, ring-shear test, unconfined compression test,and triaxial shear test are some of the common tests available in the geotechnicalengineering laboratory. The laboratory test used to characterize the internal friction242.2. Internal Friction Angle of Soils at Large Strains and Low Stressesangle of the soil should be chosen to represent the failure mode expected in the fieldas closely as possible.The definition of soil spring parameters for pipeline design must be representativeof the in-situ or backfill soil properties. Generally, the magnitudes of relative grounddeformations around pipelines during geohazards are relatively large (PRCI, 2004,2009). The design friction angles must be measured in a manner that captures theshearing mode or stress path expected in the field, under the prevalent confining stressconditions. It is also of importance to consider whether the peak or the post-peakfriction angle is relevant for design. For example, considering a pipeline subjected toaxial soil loads due to a landslide, the design may be more conservative if the peakfriction angle is used as opposed to using the post-peak large-strain friction angle asthis would lead to higher soil loads at the critical locations along the pipeline. Incontrast, in the case where a pipeline is subjected to axial thermal expansion, the useof the post-peak large-strain friction angle angle may be more conservative as thiswould result in larger loads at critical locations along the pipeline.In reference to soil loads on pipelines and offshore foundation systems, the soilstresses at the soil-solid interface are relatively low. For example, in offshore pipelinesystems, the normal stresses at the soil-pipe interface can be as low as 3 kPa (Brutonet al., 2009; Najjar et al., 2007a; White et al., 2007). It is hence important to have agood understanding of the shear strength of soils at such low confining stresses. Thenext subsections highlight past research work that has been carried out to study theshear strength of soils at relatively low confining stresses.2.2.2 Friction angle of granular materials at low confiningstressesThere is very limited data available on the critical state shear strength behaviourof coarse grained soils at relatively low confining stress levels. The state in whicha soil mass is continuously deforming at constant volume, constant normal effectivestress, constant shear stress and constant velocity is defined as the critical state of the252.2. Internal Friction Angle of Soils at Large Strains and Low Stressessoil (Poulos, 1981). Estimation of the critical state friction angle in the laboratoryposes many challenges. For example, the true void ratio at the failure plane is notaccurately known, and in most devices the amount of shear strain imposed may notbe sufficient to reach critical state. This is especially true for the triaxial, direct shearand simple shear apparatus. Ring-shear testing may be the most applicable testfor characterizing the critical state friction angle, however, the major shortcomingof the ring-shear apparatus is that the complete stress state is not known and onlythe friction angle and the normal stress at the shear plane is known. The triaxialapparatus seems to be the most relied upon device used for determination of thecritical state friction angle. The advantage of using the triaxial apparatus is that inthe triaxial test the complete stress state of the soil specimen is known. However,the strain levels required to reach the critical state is not achievable in the triaxialapparatus. Also, it is important to consider that conventional devices do not providereliable measurements at relatively low confining stress levels due to errors due todevice friction and compliance effects that become noticeable relative to the measuredparameters at low confining stress levels.In consideration of the above factors, the following discussion aims to capturepast work done in attempting to characterize the critical state friction angle of soils,and it should be understood that in most studies the true critical state may not havebeen attained. When referring to past studies, in most cases the term: “large strainfriction angle” or “friction angle at large displacement” is used. These terms shouldnot be confused with the true critical state friction angle. The friction angle obtainedat large-strain or at large displacement may be close to but not equal to the criticalstate friction angle as the rate of volume change may not be equal to zero at theattained strain levels in most test devices such as the triaxial, simple shear and directshear box Wijewickreme (1986).O’Rourke et al. (1990) present data from a series of conventional direct shear testson Ottawa sand conducted at normal stresses between 20 and 70 kPa. The data showsno dependence of the large strain friction angle on the confining stress. Figure 2.6shows the results obtained by O’Rourke et al. (1990).262.2. Internal Friction Angle of Soils at Large Strains and Low StressesFigure 2.6: Direct shear test results of Ottawa sand and Ottawa sand on HDPE (AfterO’Rourke et al. (1990)).There is limited evidence to show that the friction angle obtained in drainedtriaxial compression at large strains can increase as the confining stress decreases. Anumber of high-precision drained triaxial tests conducted on Ottawa sand in a zero-gravity environment using confining stresses between 0.05 kPa and 1.30 kPa duringthe STS-79 and STS-89 space shuttle missions provide data showing that the frictionangle at large strain increases with decreasing confining stress (Sture et al., 1998).However, as discussed earlier, it should be noted that the triaxial apparatus is notcapable of producing reliable estimation of the critical state at relatively large shearstrains which would be required at low confining stresses.The work by Been et al. (1991) provides data to show that the large strain frictionangle of a coarse grained soil obtained at relatively low stresses can be less than thatobtained at relatively high stresses. The experimental program conducted by Beenet al. (1991) comprised of over 50 triaxial tests on Erksak 330/0.7 sand aimed atstudying the uniqueness of the critical state line in the conventional void ratio andstress spaces. Their data covered tests conducted at confining stresses ranging from10 kPa to 5000 kPa, and at a number of initial void ratios spanning from very lowdensity to high density. Results of both drained and undrained tests are reported,and the results show a decrease in the large strain friction angle with increase in voidratio (i.e., based on the void ratio obtained at the critical state using soil specimens272.2. Internal Friction Angle of Soils at Large Strains and Low Stressesfrozen at the end of testing). Their data shows that at typical large strain voidratios, the large strain friction angle of Erksak sand is approximately 30°, but at veryhigh large strain void ratios obtained in tests conducted at relatively low confiningstresses, the friction angle lies between 12° and 18°. Been et al. (1991) point outthat in the p-q stress space, the data points pertaining to tests conducted at verylow confining stresses are all clustered very close to the origin of the p-q space, andtherefore, plotting data points obtained at very low confining stresses together withthose obtained at typical geotechnical confining stresses results in the data near theorigin of the p-q plot to become insignificant. In this case, the slope of the p-q stresspoints results in a large strain friction angle of 31°.It is of relevance to note that the low stress data presented by Been et al. (1991)shows much scatter and errors can be quite large for triaxial tests conducted at suchlow stresses and large strains. It is therefore difficult to reach a firm conclusionregarding the stress dependency of the large strain friction angle based on the limitedbody of evidence presented by Been et al. (1991).The mode of shear at the soil-pipe interface is likely to be equivalent to thelaboratory plane-strain shearing. As such, the use of the simple shear, direct shear,or ring shear apparatus may be best suited to obtain representative internal frictionangles. However, when using the direct shear or simple shear apparatus, sufficientlylarge shear displacements may not be achieved, and the mobilized internal frictionangles may not be the lowest attainable - this is especially the case at low confiningstress levels.In summary, the concept of critical state friction angle in soils is well establishedand it is accepted that there exists a critical state where a soil exhibits shearing underconstant effective normal stress, constant volume and constant shear velocity underwhich a constant shear stress is measured. The friction angle obtained under theseconditions is termed the critical state friction angle of the soil. However, measurementof the critical state friction angle of soil poses many challenges. Reaching the criticalstate condition requires inducing large shear strains in the soil specimen. the ring-shear apparatus is capable of applying such strain levels but the complete stress state282.2. Internal Friction Angle of Soils at Large Strains and Low Stressesof the soil specimen is not known in the ring-shear test. As such, many researchersrely on the triaxial test apparatus to determine the critical state friction angle of soils.In most cases the triaxial apparatus cannot produce strain levels required to reachcritical state conditions. Therefore different methods have been devised to estimatethe critical state friction angle of soils from triaxial test data. Triaxial test datainvestigating the effect of confining stress level on the critical state friction angle atconfining stress levels applicable to pipeline design is very limited and not sufficientto draw conclusions.2.2.3 Friction angle of fine-grained soils at low confiningstressesClays have been shown to exhibit a curved shear strength envelopes at large strainwhen tested in ring-shear under low confining stress conditions (Mesri and Abdel-Ghaffar, 1993; Terzaghi et al., 1996; Watry and Lade, 2000). Stark and Eid (1994)report results of a large number of ring-shear tests conducted to measure the large-displacement drained internal friction angle of a number of fine-grained soils, andhave shown that the shear strength envelope based on the large strain friction anglesobtained from ring-shear testing of fine-grained soils is curved at low effective normalstresses.Pedersen et al. (2003) report thin-specimen direct shear test data of Kaolinite on anumber of solid surfaces under σ′n values from 0.001 to 2.4 kPa. A tilt-table apparatushas been used to conduct the direct shear tests. Weights were used to apply the normalforce on the soil specimen. The table was tilted in 1° increments while allowing sometime for the dissipation of excess pore-water pressure after each increment. Theeffective stress at the failure plane was calculated based on the assumption that theexcess pore-water pressure had been fully dissipated. The friction angle of their testswas calculated based on the ultimate shear force attained during the tilting of thetable and the normal force acting normal to the table. Figure 2.7 shows a schematicrepresentation of the thin-specimen direct shear apparatus. Figure 2.8 shows the292.2. Internal Friction Angle of Soils at Large Strains and Low Stressesresults of their tests on Kaolinite.Figure 2.7: Thin-specimen direct shear apparatus (After Pedersen et al. (2003)).Figure 2.8: Drained peak shear strength of Kaolinite measured using thin-specimendirect shear apparatus (After Pedersen et al. (2003)).In their tests, the friction angle was defined as the inverse tangent of the ratio of theshear stress at which the sliding of the soil specimen occurred to the effective normalstress on the sliding plane. Their data shows that the peak drained internal shear302.2. Internal Friction Angle of Soils at Large Strains and Low Stressesstrength envelope of Kaolinite obtained from the tilt table apparatus is curved at loweffective stresses. Interpretation of the results of their study poses several challenges.The clay specimens were normally consolidated, however some overconsolidation isunavoidable due to the reduction in normal stress due to the titling of the table.Tilting of the table induces transient excess pore-water pressure changes and thisleads to changes in the stress state within the shear zone, although the incrementin excess pore-water pressure would be small when small tilt angle increments areused. Their data shows that the secant friction angle of Kaolinite obtained usingtheir method increases from approximately 25° at 2 kPa effective normal stress toapproximately 55° at 0.001 kPa effective normal stress.(a) Kamenose soil (b) Morikawa soilFigure 2.9: Drained ring-shear test results on Kamenose soil and Morikawa soil (AfterKimura et al. (2014)).Kimura et al. (2014) have conducted drained ring-shear testing of a number offine-grained soils of various plasticity to study the effect of rate of shear displacementat low displacement rates of 0.01 and 0.5 mm/min on the residual friction angle ofthe soils. Testes were conducted at effective normal stresses in the range of 50 to 400kPa. The authors define the residual friction angle as the ratio of the lowest stablevalue of the shear stress obtained from the ring-shear testing at large displacementsto the effective normal stress applied to the specimen. The displacement rates used312.2. Internal Friction Angle of Soils at Large Strains and Low Stressesare reported to be sufficiently low to produce drained conditions, however no data onthe pore-water pressure within the specimen during shearing is presented. Figures2.9 and 2.10 show the ring-shear test results for some of the soils tested by Kimuraet al. (2014).(a) Chenyoulanxi soil                                                                             (b) Miaowan3 soilFigure 2.10: Drained ring-shear test results on Chenyoulanxi soil and Miaowan3 soil(After Kimura et al. (2014)).The test results of Kimura et al. (2014) indicate that at lower effective normalstresses the secant friction angle of the fine-grained soils obtained from ring-sheartesting at large shear displacements is greater than those measured at higher effectivenormal stresses. Similar observations have been reported by Bishop et al. (1971) andChandler and Johnson (1976). Figures 2.11 and 2.12 show the secant residual frictionangles measured in drained ring shear tests on Brown London clay (Bishop et al.,1971) and Lias clay (Chandler and Johnson, 1976) respectively.322.2. Internal Friction Angle of Soils at Large Strains and Low Stresses0.300.250.300.150.100.05Effective normal stress (kPa)Secant residual friction angle (Deg.)0            40           80          120         160         200         240         28016141210864Figure 2.11: Drained ring-shear test results on Brown London clay (After Bishopet al. (1971)).0.400.300.200.10Effective normal stress (kPa)Secant residual friction angle (Deg.)0 40             80             120           160           200            2402015105Figure 2.12: Drained ring-shear test results on Lias clay (After Chandler and Johnson(1976)).332.3. Friction Angle of Soil-Solid Interfaces at Large Strains and Low StressesGiven this background, it is important to consider the possible effect of confiningstress level on the large-displacement soil-solid interface friction angle of fine-grainedsoils when designing pipelines that will experience relatively low soil confining stresses.The use of friction angles measured at conventional geotechnical stress levels (typi-cally above 50 kPa) instead of those obtained at low stress levels can lead to under-prediction of soil load demand on structures in certain cases. This is especially truefor pipelines.2.3 Friction Angle of Soil-Solid Interfaces atLarge Strains and Low StressesThe work reported in this thesis primarily focuses on the friction angle character-istics of soil-solid interfaces under relatively low effective normal stresses and largedisplacements. Current literature on soil-solid interface shear behaviour under theseconditions is limited. The proceeding sub-sections touch upon some of the work thatis directly related to this field of work.2.3.1 Friction angle of sand-polymer interfacesMost polymer materials used in geotechnical applications tend to be extensible poly-mer sheets. Therefore, testing such materials in the laboratory to determine thesoil-polymer friction angle can be problematic. Testing sand-polymer interfaces com-monly encompass using the direct-shear test; the ring-shear test; or the pullout test(Ingold, 1982; Negussey et al., 1989; O’Rourke et al., 1990). The direct-shear boxis often used with the soil contained in the upper-half of the box and allowing thesoil to rest against the polymer using one of the following methods: (i) the polymerbonded to a rigid substrate and inserted in the lower-half of the box ; (ii) replacingthe lower-half of the box with a large sheet of polymer bonded to a rigid substrate;and (iii) replacing the lower-half of the box with a large sheet of polymer clamped onone end and left free at the other end instead of bonding to a rigid substrate (Ingold,342.3. Friction Angle of Soil-Solid Interfaces at Large Strains and Low Stresses1982). When a sufficiently large interface area is used, the interface direct shear testcan be used to achieve relatively large shear displacements.Based on a series of tests utilizing the direct-shear box, Ingold (1982) reports afew interesting observations with regard to the shear strength characteristics of sand-geotextile interfaces. Whether or not reported friction angles are at the peak or largedisplacement conditions is not reported. The interface friction angles measured inthe direct shear apparatus was found to be less than the internal friction angle of thesand measured in the same apparatus. The interface friction angle was found to behighly dependent on the roughness of the interface (Ingold, 1982).Negussey et al. (1989) report ring-shear test results for angular to round grainedsands tested against an 80-mil smooth high density polyethylene (HDPE) geomem-brane. Negussey et al. (1989) report observing distinct peak interface friction an-gles at small strains, followed by much lower large-displacement friction angles atlarge strains. Both, the peak and large-displacement, interface friction angles werefound to be less than the corresponding internal friction angles of the sands. Large-displacement interface friction angles were found to be as much as 50% lower thanthe internal large-displacement friction angles of the sands. Negussey et al. (1989)report observing damage to the geomembrane in the form of indentations or concen-tric scour tracks post-test. The peak interface friction angle was found to increasewith increasing normal stress, while the large-displacement interface friction anglewas found to be independent of normal stress (Negussey et al., 1989).Rinne (1989) has conducted interface ring-shear tests of Ottawa sand and Targetsand on PVC and HDPE geomembranes to study the effect of angularity of the sand,the roughness of the solid test surface, and the effective normal stress. The large-displacement interface friction angle was found to increase slightly when the effectivenormal stress was increased from 100 to 500 kPa. A general trend of a mild increase ofthe large-displacement interface friction angle with increasing effective normal stresswas observed for both sands on PVC.O’Rourke et al. (1990) report data from a series of sand-polymer interface direct-shear tests conducted at relatively low normal stresses (between 3.5 and 69 kPa) with352.3. Friction Angle of Soil-Solid Interfaces at Large Strains and Low Stressesan emphasis on application to buried pipelines and linings. Ottawa sand was usedin their study. One of the primary aims of these tests was to study the effects ofpolymer hardness on the interface shear mechanism. O’Rourke et al. (1990) reportthe Shore-D hardness (HD) of the polymers tested in accordance with ASTM-D2240(D2240-15, 2010). Two hard epoxy materials with HD = 88 and HD = 69, and anHDPE with HD = 64, after interface direct shear testing with sand, when examinedunder a scanning electron microscope, showed damage to the polymer surfaces in theform of scratch marks. In contrast, a soft epoxy with HD = 44 and a PVC withHD = 34 did not show such scratch marks post-test. Based on these observations,O’Rourke et al. (1990) support that the softer polymers induce more rolling of thesand particles as opposed to harder polymers which induce more sliding.O’Rourke et al. (1990) also report observations on the effects of normal stresson the peak interface friction angle. The observations show that the peak interfaceshear strength envelope is linear within the 3.5 to 69 kPa normal stress range for allthe polymer interfaces tested. O’Rourke et al. (1990) also report observing that anon-plasticized hard PVC with HD = 85 showed a much larger peak interface frictionangle compared to that for a plasticized soft PVC with HD = 35−45. O’Rourke et al.(1990) show that a linear trend exists between the Shore-D hardness of the polymersand the ratio of peak interface friction angle to the peak internal friction angle of thesoil (O’Rourke et al., 1990).A number of sand-polymer interface direct shear tests conducted by Frost et al.(2002) corroborate earlier findings, further adding that the peak interface frictionangle rapidly increases when the hardness of the material is decreased while simul-taneously increasing interface roughness, implying a coupling effect of roughness andhardness.2.3.2 Friction angle of sand-steel interfacesA detailed study of the peak interface friction angle of several sand-steel interfaces hasbeen carried out by Uesugi and Kishida (1986a). They report the use of a simple-sheardevice and a direct-shear device, both customized for testing sand-steel interfaces, to362.3. Friction Angle of Soil-Solid Interfaces at Large Strains and Low Stressesstudy the effects of: (i) interface roughness; (ii) mean particle diameter of sand; (iii)sand type; (iv) coefficient of uniformity; (v) test device type; and (v) normal stress;on the peak interface friction angle. Uesugi and Kishida (1986a) point out that for thesame steel surface, smaller sand particles experience a greater geometric interferencecompared to larger sand particles (i.e., given the same surface roughness of the steelsurface, smaller sand particles experience a greater relative roughness compared tolarger sand particles). Therefore, they propose that the surface roughness can bebetter correlated with the coefficient of interface friction by normalizing it by theparticle size of the soil as given by Eq. 2.7. In Eq. 2.7, the normalized roughness Rnis ratio of the maximum roughness Rmax to the mean particle diameter D50. It is ofrelevance to note that Uesugi and Kishida (1986a) have evaluated Rmax over a gaugelength Lg = D50.Rn =(Rmax(Lg=D50))D50(2.7)Uesugi and Kishida (1986a) show that the coefficient of interface friction can beapproximated to a linear function of normalized roughness with no dependence onD50, bearing in mind that there is some degree of scatter in the data. Uesugi andKishida (1986a) show that the sand-steel interface friction is strongly influenced by:(i) surface roughness of the steel plate; (ii) the mean particle size of the sand; and(iii) the sand type. They also show that there is little influence of: (i) the test type(i.e., direct-shear vs. simple-shear vs ring-shear); (ii) the uniformity coefficient ofthe sand; and (iii) the normal stress level within the range of 98-980 kPa. Whilethe afforementioned findings are well supported by the data set presented by Uesugiand Kishida (1986a), it should be borne in mind that data is limited to three sands.Nevertheless, the evidence is quite convincing.Rinne (1989) has conducted interface ring-shear tests of Ottawa sand and Targetsand on smooth and rough steel to study the effect of angularity of the sand, the rough-ness of the solid test surface, and the effective normal stress. The large-displacementinterface friction angle was found to be insensitive to the effective normal stress, but372.3. Friction Angle of Soil-Solid Interfaces at Large Strains and Low Stresseswas sensitive to the angularity of the sand.Evgin and Fakharian (1997) have conducted a series of sand-steel interface sheartests utilizing a sophisticated multi-axial simple shear device to study the influenceof stress paths. They show that the peak and residual interface friction angles areindependent of stress path and normal stress. Evgin and Fakharian (1997) reportlarge-displacement interface friction angle data for a quartz sand tested on a sand-blasted steel surface using customized constant normal stress, and constant normalstiffness direct simple shear interface test devices under σ′n values in the range of 100to 500 kPa. Their results show that the large-displacement interface friction angle isnot affected by σ′n.It is useful to consider the work done by Jardine et al. (1993) where the resultsof interface direct shear tests of Labenne sand sheared on steel interfaces of differentroughness within a σ′n range of 45 to 70 kPa are presented. The grain size distributionof Labenne sand is approximately similar to that of Fraser River sand, and bothmaterials show the same constant volume internal friction angle of 33° based on K0consolidated undrained triaxial tests. Their results show that the large-displacementinterface friction angle is independent of the initial sand relative density, and is onlymarginally dependent on the normal effective stress level at low effective normal stresslevels.2.3.3 Friction angle of clay-solid interfacesPast work on characterizing the large-displacement drained interface friction angle offine-grained soils is very limited. Data is even more scarce for tests conducted at loweffective normal stresses. Several past studies of significance are highlighted in thetext that follows.Pedersen et al. (2003) report thin-specimen direct shear test data of Kaoliniteon a number of solid surfaces under σ′n values from 0.001 to 2.4 kPa. A tilt-tableapparatus has been used to conduct the interface shear tests. Their data shows thatthe peak drained internal shear strength envelope of Kaolinite as well as its drainedpeak interface shear strength envelopes on a number of solid surfaces are curved at low382.3. Friction Angle of Soil-Solid Interfaces at Large Strains and Low Stresseseffective stresses. Stark and Eid (1994) report results of a large number of ring-sheartests conducted to measure the large-displacement drained internal shear strength ofa number of fine-grained soils, and have shown that the shear strength envelope offine-grained soils is curved at low effective normal stresses.Clark and Meyerhof (1972) report findings of a number of drained interface directshear tests conducted to study the interface shear strength of a medium plastic fine-grained soil on steel surfaces of different roughness as part of an experimental programthat was conducted on characterizing the skin friction of driven piles in fine-grainedsoils. They have used a modified direct shear device of standard size, but withthe bottom plate replaced by a solid block of steel to serve as the test interfacesurface that was instrumented with a pore-water pressure transducer as a means toaid determination of the effective normal stress on the interface. They have conductedthe interface shear tests at σ′n values in the range of 17.6 to 90.9 kPa. Their resultsshowed that for the rough steel interface, the large-displacement drained interfacefriction angle was quite close to the large-displacement drained internal friction angleof the clay measured in direct shear; an interface efficiency factor of 0.96 was observed.For the smooth steel interface, the interface efficiency factor dropped to a value of0.67. Their shear strength envelopes were fairly linear within the σ′n range tested,and no apparent intercept was observed.Littleton (1976) reports data arising from large-displacement drained direct sheartesting of two clay materials on a steel surface. They have used a conventional directshear device of 60 mm size, with the bottom half replaced by the test steel surface.No pore-water pressure measurements were made. However, very low displacementrates have been used to ensure drained shearing in order to aid in determining theeffective normal stress on the interface. The interface shear tests have been conductedat σ′n values in the range of 50.0 to 300 kPa. Interface efficiency factors of 0.80 to0.82 have been observed for the two clay materials. Also, fairly linear shear strengthenvelopes with no apparent intercepts were reported.Tika-Vassilikos (1991) reports drained large-displacement interface ring-shear testresults of London clay on a relatively rough stainless steel surface at σ′n of approx-392.3. Friction Angle of Soil-Solid Interfaces at Large Strains and Low Stressesimately 450 kPa. They report an interface efficiency factor of 0.82. Tsubakiharaand Kishida (1993) present results of drained large-displacement interface shear testsconducted on Kawasaki clay on a number of steel interfaces with different surfaceroughness using a simple-shear and direct-shear type test devices under σ′n of 294kPa. They report interface efficiency factors in the range of 0.72 to 0.9. The interfaceefficiency factors were found to depend heavily on the surface roughness of the steel.Lemos and Vaughan (2009) report results of drained large-displacement ring-sheartesting of a large number of clay-solid interfaces under σ′n values from 100 to 500 kPa.They report interface efficiency factors in the range of 0.60 to 0.85, and the interfaceefficiency factors were found to be heavily dependent on the surface roughness of theinterfaces.Najjar et al. (2007a) report results of drained large-displacement interface sheartests of a number of clay materials on solid surfaces of different roughness using atilt-table apparatus under σ′n from 1.0 to 5.0 kPa. The drained conditions have beenmaintained by controlling the tilt rate of the tilt-table. Large-displacement frictionangles have been obtained by repeated shearing of the same specimen by repeatedtilting of the tilt-table. They report interface efficiency factors in the range of 0.55to 0.90. The interface efficiency factors were shown to depend on the solid surfacematerial type as well as σ′n.It is important to highlight that the literature on the large displacement frictionangle of soils at very low effective normal stresses is limited, and it is acknowledgedthat this is mainly due to experimental challenges involved in conducting tests un-der such low effective normal stresses. Standard geotechnical testing devices aredesigned to be operated at much higher stress levels. Therefore, the study of soilsand soil-structure interfaces at very low effective stresses often requires custom madedevices. The present research program covered in this thesis aims to address some ofthe technical challenges involved in conducting interface shear strength tests underthese challenging conditions by developing a novel macro-scale interface direct shearapparatus.The following section briefly documents the various apparatus conventionally used402.3. Friction Angle of Soil-Solid Interfaces at Large Strains and Low Stressesfor the study of soil-solid interfaces and attempts to show the limitations of each devicein its applicability in pipeline design.412.4. Laboratory Testing of Soil-Solid Interface Friction Angle2.4 Laboratory Testing of Soil-Solid InterfaceFriction AngleCommonly available direct-shear and ring-shear devices do not always serve as thebest option when testing for the soil-solid interface shear strength at a suitable ac-curacy when very low effective normal stresses are used (Fang et al., 2004). This ismainly because of the errors introduced as a result of the relatively small shear areaavailable combined with the friction associated with the mechanical system in thesedevices. Due to low interface shear strengths expected under small confining stresslevels, shear forces measured from small-scale direct shear tests would be relativelysmall and would also be affected by the friction associated with the mechanical com-ponents of the devices (Lehane and Liu, 2013). As such, special modifications arenecessary for accurately and reliably testing for the soil-solid interface shear strengthunder very low confining stresses in such apparatus. For example, the electrical back-ground noise in the data acquisition system of these devices is often at the sameorder of magnitude as the electrical signals generated by the load cells that measureshear loads at very low confining stresses (Likos et al., 2010). Use of precision instru-mentation and data acquisition systems is required in order to ensure that reliablemeasurements are made. In addition, stress non-uniformities are quite common dueto the manner in which the normal stress is applied in most of these devices (Hsiehand Hsieh, 2003; Lings and Dietz, 2004; OSullivan et al., 2004). Moreover, the use ofthe direct-shear device often requires multiple reversals of the shear direction to arriveat the large displacement friction angle, and this does not necessarily represent fieldconditions (Mesri and Huvaj-Sarihan, 2012). Another limitation of these devices isthe difficulty to perform interface shear tests with reliable excess pore-water pressuremeasurements to ensure drained conditions during shearing of fully saturated fine-grained soils. The use of pore-water pressure transducers at the soil-solid interface isrequired in order to measure the pore-water pressure variation at the interface duringshear so that an accurate determination of the effective normal stress can be made(Fox and Stark, 2004). When testing under small effective stress levels, the stress422.4. Laboratory Testing of Soil-Solid Interface Friction Angledistribution at the interface can also be affected by the physical boundaries of thesedevices and this can lead to large errors in determining the interface shear resistance.Hence, the use of conventional apparatus for studying the soil-solid interface shearstrength at large shear strains under very low effective normal stresses may not alwaysbe suitable, unless the devices are modified to minimize such uncertainties.Despite such limitations, traditional testing methods are still quite useful and theyare being used to study soil-solid interface shear strength at relatively low effectivenormal stresses (2.4 kPa to 70 kPa). This is particularly evident in work involv-ing studying geosynthetic clay liners and shallow slope failures where such effectivestresses are expected at the shearing zone. The direct shear device has been used totest for the internal shear strength of geosynthetic clay liners (clay layer between twolayers of geosynthetics) under applied total normal stresses ranging from 2.4 kPa to70 kPa (Gilbert et al., 1996; Zornberg et al., 2005). Because of the difficulties in ascer-taining the pore-water pressure at the soil-solid interface during shearing, the resultsderived from such apparatus are often questionable when fine-grained soils are tested(Bemben and Schulze, 1997). Since the conventional direct shear device providesa limited shear displacement range it often fails to capture the large-displacementinterface shear strength (Gilbert et al., 1996).The ring shear device is also being used for the purpose of characterizing soil-solidinterface shear behaviour in the specialized field of geosynthetics. Testing at totalnormal stresses as low as 50 kPa have been conducted in the past to study the shearstrength of various soil-geosynthetic interfaces using the ring-shear apparatus(Effendi,1995).Most early research has been geared towards characterizing pile shaft friction (i.e.,pile skin friction). Many types of soil-solid interfaces involving common constructionmaterials such as steel, timber and concrete have been widely studied (Bosscher andOrtiz, 1987; Brumund and Leonards, 1973; Fakharian and Evgin, 1997; Potyondy,1961; Reddy et al., 2000). With the expansion in the use of geosynthetics in industryfor geotechnical engineering, the characterization of interface shear strength of varioussoils and geosynthetics has also been given considerable attention (Frost and Han,432.4. Laboratory Testing of Soil-Solid Interface Friction Angle1999; Negussey et al., 1989; O’Rourke et al., 1990; Stark and Poeppel, 1994).It is also important to note that with particular reference to the offshore pipelinedevelopment applications where the soil-pipe interface shear strength at low effectivenormal stress levels is sought after, the current understanding has been mainly derivedfrom data originating from relatively small-scale (less than 0.15 m2) tilt table testswhere both the normal and shear stresses to the interface are applied using gravityas opposed to using a mechanical system (Fang et al., 2004; Liu et al., 1997; Najjaret al., 2007b; Pedersen et al., 2003). The tilt table device is very simple in its designand does provide the means to perform interface shear tests under low confiningstresses. However, the method does not provide for the measurement of excess pore-water pressure during shearing, in turn, making it difficult to determine the effectivenormal stress at the interface accurately. This is further complicated due to thecontinuous changes in applied normal stress during tilting and the lack of control onthe displacement rate of shearing that takes place as the shear surface is tilted.With recent developments in the offshore oil and gas industry, projects involvingthe development of subterranean and submerged pipelines that traverse vast distancesacross continents has also presented new areas of study directly related to soil-solidinterfaces. Often, conventional geotechnical testing devices require modification inorder to test soil-solid interfaces. The research work undertaken in this unique areaof geotechnical engineering has produced an array of novel tools and methods that arebeing used to study the shear strength of soil-solid interfaces (Boukpeti and White,2016; Eid et al., 2014; Ganesan et al., 2013; Najjar et al., 2007a; Wijewickreme et al.,2013). An array of testing methods, including the direct shear test, direct simpleshear test, dual interface shear test, ring shear test, and the tilt table test have beenused in the past to undertake this task (Kishida and Uesugi, 1987; Najjar et al.,2007a; Paikowsky et al., 1995; Rinne, 1989; Uesugi and Kishida, 1986b; Uesugi et al.,1988; Yoshimi and Kishida, 1981).The next sub-sections provide an overview of the tools and methods commonlyused in this field of study.442.4. Laboratory Testing of Soil-Solid Interface Friction Angle2.4.1 Direct shear apparatusThe conventional direct shear box is often used with the soil contained in the upper-half of the box and allowing the soil to rest against the solid test surface using one ofthe following methods (see Fig. 2.13): (i) the solid test surface is bonded to a rigidsubstrate and inserted in the lower-half of the box ; (ii) replacing the lower-half of thebox with a large sheet of the solid test surface bonded to a rigid substrate; and (iii)replacing the lower-half of the box with a large sheet of solid test surface clamped onone end and left free at the other end instead of bonding to a rigid substrate (ASTM-D5321-12, 2012; Ingold, 1982; Lee and Manjunath, 2000). The configurations shownin Figs. 3.15(a) and 3.15(b)) are used for characterizing the small-displacement andlarge-displacement interface shear strength respectively.Loading capUpper box Test soil Test surfaceRigid spacer blockLower box(a)Loading capUpper box Test soil Test surfaceRigid substrate(b)Figure 2.13: Typical configurations of the modified direct shear apparatus used forinterface shear testing.Testing for the interface shear strength using the modified direct shear device isvery much the same as the conventional direct shear test. Typically, the soil specimenis sheared against the solid surface in displacement-controlled mode under a constantnormal stress until the required horizontal displacement is reached or until the limitinghorizontal displacement of the device is reached. The shear displacement of the upperbox, the vertical displacement of the loading cap, and the shear force are measuredcontinuously during shearing. The interface shear strength envelope is obtained froma series of tests assuming a Mohr-Coulomb failure criterion.The direct shear apparatus presents two important advantages over other com-452.4. Laboratory Testing of Soil-Solid Interface Friction Anglemonly available devices: (i) test setup and specimen preparation procedures are rel-atively simple; and (ii) it is often readily available in any geotechnical laboratory.There are commercial devices available for testing relatively large interface areas ofup to 305 by 305 mm2 (Gilbert et al., 1996; Seed and Boulanger, 1991). Consequently,it has been a common choice for characterizing soil-solid interfaces in research andgeotechnical practice. Many applications of the direct shear apparatus in this field ofstudy include testing soil-geomembrane, soil-geotextile, and geomembrane-geotextileinterfaces.The direct shear apparatus presents several limitations however. The original usefor the direct shear device was confined to the characterization of peak shear strengthof soils. Therefore, the conventional shear device only allows for a limited amountof shear displacement. In its original configuration as shown in Fig. 3.15(a), theconventional direct shear device is inadequate for the study of large-displacementinterface shear strength. Multiple reversals of shear is commonly used to producelarge-displacement shear behaviour. However this does not always reproduce actualfield conditions and may not always be reliable. The device is often modified to theconfiguration shown in Fig. 3.15(b) to allow for larger interface shear displacements(Lee and Manjunath, 2000). In addition, end effects, induced by the presence of therigid walls of the soil container, may introduce errors in the test results (OSullivanet al., 2004; Potts et al., 1987). Also, non-uniformity of strains can generate acrossthe soil specimen resulting in undervalued peak shear strengths. A summary of a fewselected interface shear experiments performed using various types of conventionaland modified direct shear devices is presented in Table 2.2. The studies presented inthe table have been selected based on the relevance to offshore pipeline applications.462.4. Laboratory Testing of Soil-Solid Interface Friction AngleTable 2.2: Summary of previous research on soil-solid interfaces conducted using thedirect shear apparatusReference Interface Outline of tests conducted(Potyondy, 1961) Sand/Steel,Sand/Concrete60 mm × 60 mm and 90 mm × 90 mm sheararea used. Monotonic shear under constantnormal load. Investigated the peak interfaceshear strength only. Smooth and rough sur-faces have been tested.(Butterfield and An-drawes, 1972)Sand/Steel Leighton Buzzard - B.S.S. 14-36 sand wastested against polished mild-steel.(Acar et al., 1982) Sand/Steel,Sand/ConcreteMonotonic shear under constant normal load.(Uesugi and Kishida,1986a)Sands/Steel 400 mm × 100 mm shear area used. Fujigawasand and Seto sand tested against low-carbonsteel of various surface roughness. Monotonicshear under constant normal load. Investi-gated the peak interface shear strength andfactors of influence.(Tejchman and Wu,2005)Sand/Steel Karlsruhe sand was tested against steel of var-ious surface roughness. Monotonic shear un-der constant normal load.(Frost et al., 2002) Sand/Steel Modified direct shear apparatus having sheararea of 12.7 cm2 used. Ottawa 20/30 sandand Valdosta blasting sand was tested againsthardened steel and concrete surfaces. Mono-tonic shear under constant normal load.(Dietz and Lings,2006)Sand/Steel Monotonic shear under constant normal load.A winged direct shear apparatus having a 100mm × 100 mm shear area was used. Coarse,medium and fine sands was tested againstmild steel surfaces of various roughness.(Tsubakihara andKishida, 1993)Clay/Steel 100 mm × 100 mm shear area used. Mono-tonic shear under constant normal load or con-stant volume. Kawasaki marine clay testedagainst low-carbon steel surfaces of variousroughness.(Tanaka, 2003) Silt/Steel, Clay/Steel Gap-controlled direct shear apparatus used.Monotonic shear under constant volume. Siltand Kaolinite clay was tested against steel sur-faces of various roughness. A range of sheardisplacement rates have been used.472.4. Laboratory Testing of Soil-Solid Interface Friction Angle2.4.2 Direct simple shear apparatusA very good example of the use of a direct simple shear type apparatus for thecharacterization of soil-solid interface shear strength is presented by (Uesugi andKishida, 1986b). They present a direct simple shear device where the test soil iscontained within a box built from a number of stacked aluminum plates that are freeto displace relative to each other (see Fig. 2.14). The soil specimen rests on a largertest plate that forms the solid interface. A constant normal stress is applied to thesoil, and shearing is achieved by applying a tangential displacement to the solid testplate. With this apparatus, the sliding displacement at the soil-solid contact surfaceas well as the displacement due to the shear deformation of the soil mass can beobtained independently. The larger solid surface ensures that the soil is always incontact with the solid surface resulting in the contact area at the soil-solid interfaceto remain constant during shear. Each aluminum ring is coated with a lubricant tominimize inter-ring friction. A summary of selected interface shear studies performedusing various types of modified direct simple shear apparatus is presented in Table2.3.Displacmenet at interfaceShear deformation of specimenFigure 2.14: Typical configuration of the simple shear type apparatus used for soil-solid interface shear testing (adapted from Uesugi and Kishida (1986b)).482.4. Laboratory Testing of Soil-Solid Interface Friction AngleTable 2.3: Summary of previous research on soil-solid interfaces conducted using thedirect simple shear apparatusReference Interface Outline of tests conducted(Uesugi and Kishida,1986b)Sand/Steel 400 mm × 100 mm shear area used. Mono-tonic shear under constant normal load. Toy-oura sand, Fukushima sand, Fujigawa sand,Glass beads were tested against low-carbonsteel surfaces of various roughness.(Uesugi et al., 1988) Sand/Steel 400 mm × 100 mm shear area used. Mono-tonic shear under constant normal load. Setosand was tested against low-carbon steel sur-faces of various roughness.(UESUGI et al., 1989) Sand/Steel 100 mm × 40 mm shear area used. Monotonicor repeated shear under constant normal load.Toyoura sand, Seto sand, Fujigawa sand wastested against low-carbon steel surfaces of var-ious roughness.(Evgin and Fakharian,1997)Sands/Steel 100 mm × 100 mm shear area used. Mono-tonic shear under constant normal stress orconstant normal stiffness. Dense sand wastested against a rough steel surface. Investi-gated the peak and large displacement inter-face shear strength.(Fakharian and Evgin,1997)Sand/Steel 100 mm × 100 mm shear area used. Cyclicshear under constant normal stress or con-stant normal stiffness. Silica sand was testedagainst a rough steel surface.(Tsubakihara et al.,1993)Clay/Steel 60 mm diameter shear area used. Monotonicshear under constant normal load. Kawasakimarine clay and four types of artificially mixedcohesive soils were tested against low-carbonsteel surfaces of various roughness.(Tsubakihara andKishida, 1993)Clay/Steel 60 mm diameter shear area used. Monotonicshear under constant normal load or constantvolume. Kawasaki marine clay tested againstlow-carbon steel surfaces of various roughness.(Oumarou and Evgin,2005)Sand/Steel 100 mm × 100 mm shear area used. Cyclicshear under constant normal stress. Quartzsand was tested against a rough steel surface.492.4. Laboratory Testing of Soil-Solid Interface Friction AngleAs for limitations, these devices inevitably carry similar drawbacks as the directshear device due to non-uniform distribution of stresses at the interface (Kishida andUesugi, 1987). Large shear displacements representative of conditions at the soil-pipe interface may not be attained in these devices. Also the preparation of the soilspecimen is relatively complicated and the device provides a limited maximum totaldisplacement.2.4.3 Ring shear apparatusThe conventional ring-shear apparatus is also used to study soil-solid interfaces. Thering-shear apparatus provides a number of advantages over the direct shear and directsimple shear devices. Namely, (i) the device allows for unlimited circumferential sheardisplacement at the interface which is preferable in offshore pipeline applications;and (ii) there is no eccentric loading during shear and stress non-uniformity alongthe contact surfaces and the soil specimen are limited. Table 2.4 provides a list ofexperimental investigations on soil-solid interfaces conducted using ring shear devicesrelevant to the work covered in this thesis.The ring-shear apparatus has a few disadvantages: (i) it is relatively complicatedin terms of specimen preparation and test procedure; (ii) there is the possibility ofhaving a displacement gradient across the interface, and as a result, possible variationof shear strain within the soil specimen; and (iii) there is the possibility for the testsurface to change in texture in large-displacement tests, because of shearing of thesoil specimen over the same test surface repeatedly.502.4. Laboratory Testing of Soil-Solid Interface Friction AngleTable 2.4: Summary of previous research on soil-solid interfaces conducted using thering-shear apparatusReference Interface Outline of tests conducted(Yoshimi and Kishida,1981)Sand/Steel 400 mm × 100 mm shear area used. Mono-tonic shear under constant normal load. Toy-oura sand, Tonegawa sand, Niigata sand weretested against low-carbon steel surfaces of var-ious roughness.(Ho et al., 2011) Sand/Steel, Silt/Steel Range of silica sands, and a silica rock floursilt was sheared against roughened steel an-nular interfaces. Monotonic shear under in-cremental constant normal loads.(Negussey et al., 1989) HDPE-geomembrane/sand,HDPE-geomembrane/gravel,HDPE-geomembrane/geotextileMonotonic shear under constant normal load.(Effendi, 1995) VLDPE, PVC, HDPE,Non-woven geotextileswith high-plasticityclay and Ottawa C-109sandMonotonic shear under constant normalstresses within the range: 25 kPa to 300 kPa.(Rinne, 1989) Sand/Steel Monotonic shear under constant normal load.Ottawa sand, Target sand, was tested againststeel surfaces of various roughness.2.4.4 Other devicesThe dual interface direct shear device (Paikowsky et al., 1995), annular interfaceshear device (Brumund and Leonards, 1973) and, more recently, the tilt table device(Najjar et al., 2007b; Pedersen et al., 2003) have been used to characterize the large-displacement friction angle of soil-solid interfaces. These devices are a good exampleof the development of custom test methods focused on studying unique soil-solidinteraction problems.The annular interface shear device is a highly customized device that resemblesa cylindrical triaxial test. The soil specimen is cylindrical in shape, and is enclosed512.4. Laboratory Testing of Soil-Solid Interface Friction Angleby a flexible rubber membrane that is used to apply hydrostatic pressure around thesoil specimen using fluid pressure. A cylindrical solid rod, made of the solid interfacetest material, passes through the soil specimen at the centre along the longitudinalaxis. The test rod is displaced relative to the soil specimen to induce interface shear(see Brumund and Leonards (1973) for details). This device has been primarilyused to study the dynamic interface friction between soils and common constructionmaterials. Unfortunately, it does not seem to be widely used for research or industrialprojects. This may be due to the difficulties (or uncertainty) encountered in theinterpretation of the test data. Specifically, it is generally difficult to measure thenormal stress at the interface, and certain assumptions need to be relied upon toestimate the normal stress. Specimen preparation can also be quite difficult.The dual interface shear device is very similar to the simple shear device describedearlier, but has the unique feature of providing the ability of determining the sheardistribution along the interface through the use of a friction bar (see Paikowsky et al.(1995) for details).The tilt table device provides a relatively simple method of testing for the interfaceshear strength, but it is difficult to control the rate of shear and loading conditionsat the interface since the movement is gravity driven. Further disadvantages are:displacements are not controlled and the post-peak response cannot be measured,test pressures are limited by toppling of the surcharge weights, and non-uniformnormal stresses develop on the interface as the device is inclined. Advantages includethe elimination of internal machine friction, not forcing failure to occur along theinterface, and the ability to perform tests under low normal stresses (Pedersen et al.,2003).While there are many testing devices at our disposal, it should be born in mindthat there are limitations with respect to testing interfaces under stress conditionscommonly encountered in soil-pipe interaction problems. Primarily, careful consider-ation of the reliability of test results is required when testing interfaces under verylow effective normal stresses and large shear strains. Conventional test apparatusare often designed to test soils at stress levels much higher than that encountered in522.5. Summary of the Chapterpipeline applications. Testing under very low confining stress conditions can easilyintroduce reliability and device compliance issues. Therefore, new test methods, orsome form of modification of conventional testing methods, are frequently being usedto overcome such limitations.2.5 Summary of the ChapterThis chapter looked at the past work that has been done to study the internal andsoil-solid interface friction angles with specific emphasis on the effect of confiningstress level. Given that in practice soil-pipe interface shearing occurs at relativelylow confining stress levels (3 to 30 kPa) and the relative soil-pipe shear displacementscan be several times the diameter of the pipeline, the primary focus was to investigateresearch on the large-displacement friction angle of soils and soil-solid interfaces.In summary, the following conclusions can be drawn from the literature reviewed:1. The chapter first looked at the shear strength properties of soil in terms of thepeak and large-displacement internal friction angle as measured in the labo-ratory using common test apparatus. The mobilized internal friction angle ofa given soil was shown to be composed of three components: (i) the intrinsicinter-particle friction angle (φµ); (ii) the friction angle due to dilation (φd); and(iii) the friction angle due to particle reorientation and damage (φdr). Paststudies have shown that at large strains, the mobilized friction angle tends to aunique lower bound value termed “critical state friction angle” (φcv) where in-definite shearing occurs under constant stress and constant volume conditions.It is understood that φcv is a unique material property.2. Studies of the internal friction angle of soils in the soils laboratory have beenextensively carried out using the triaxial, direct shear and ring-shear appara-tus. However, it was shown that at low confining stress levels the required shearstrains to attain critical state conditions would not be possible in these appa-ratus. In addition to the above limitation, the triaxial apparatus is not reliable532.5. Summary of the Chapterfor testing at confining stresses applicable to soil-pipe interaction problems (3to 30 kPa). Only a limited body of data on the large-displacement friction angleobtained from triaxial testing at low confining stress levels is available, and thisdata needs to be treated with the above limitations in mind.3. Ring-shear testing may be the most suitable for characterizing the large-displacementfriction angle of soils with applicability to pipeline design. However, withoutsuitable modification of conventional apparatus, data obtained at low confiningstresses is subject to high variability.4. The current body of research on the effect of confining stress level on the criticalstate friction angle of soils at very low confining stress levels (3 to 30 kPa) isvery limited. More research work is required to investigate the uniqueness ofthe critical state friction angle under conditions applicable to pipeline design.5. Given that the friction angle obtained would be highly dependent on the testconditions applied (e.g., plane strain versus triaxial compression) and the meth-ods used for interpretation of the results, it is of much importance to selectsuitable test methods that would be representative of field conditions. Cur-rent pipeline design guidelines do not provide sufficient guidance as to the bestpractices required for estimating soil friction angle parameters for design.6. Current design guidelines take two approaches for estimating the axial soil-pipeshear resistance: (i) the undrained (α-approach); and (ii) the effective stressbased (β-approach).7. The α-approach heavily relies on empirical correlations of the adhesion factorα and the undrained shear strength su of the soil. Considerable variabilityand uncertainty exists in methods of estimating su. Similarly, the α versussu correlations for pipelines are based on a very limited database of field andlaboratory physical model test results. As such, there is much uncertainty inthe soil-spring parameters derived from the α-approach. There is much needfor further research in this area.542.6. Expected Contributions8. The β-approach relies on estimation of the effective normal stress and the fric-tion angle of the soil and soil-pipe interface at that effective normal stress andat relatively large shear displacement. Current laboratory apparatus used forcharacterizing the friction angle of soils at low confining stresses and large sheardisplacements are not reliable owing to a number of challenges. There is verylittle research that has been carried out to study the uniqueness of the criticalstate friction angle of soils at low confining stresses applicable to pipeline de-sign. More research work is warranted to gain a better understanding of thisarea of interest.9. The measurement of the soil-solid interface friction angle at representative con-fining stresses and large shear displacements has been carried out using a modi-fied form of the direct shear apparatus or the ring-shear apparatus. Custommade apparatus such as the Cam-Shear device and the Tilt Table appara-tus have also been used. Research work investigating the effect of confiningstress level on the large-displacement interface friction angle is very limited.The general consensus seems to be that the large-displacement interface fric-tion angle increases with decreasing confining stress level resulting in a curvedlarge-displacement shear strength envelope. The limited data available in theliterature is subject to much uncertainty as most devices do not provide reli-able data at unconventionally low confining stress levels (3 to 30 kPa). Currentdesign guidelines do not explicitly provide best practices to account for effect ofconfining stress level on the selection of soil spring parameters. More researchwork is required to produce reliable data showing the effect of confining stresslevel on the large-displacement soil-solid interface friction angle.2.6 Expected ContributionsConsidering the knowledge and technology gaps identified in the literature review,the present research work aims to improve the current state of practice through thefollowing contributions:552.6. Expected Contributions1. Development of a novel laboratory shear test apparatus capable of conductingsoil-solid interface shear testing at low confining stresses (3 to 30 kPa) and largeshear displacements.2. Development of a novel full-scale axial soil-pipe interaction physical modelingapparatus for use with fine-grained soils.3. Study the effect of soil type, confining stress level and interface roughness onthe large-displacement soil-solid interface friction angle at confining stress levelsapplicable to pipeline design (3 to 30 kPa).4. Full-scale axial soil-pipe interaction testing of a selected large-diameter epoxycoated pipe in fine-grained soil to study the possible effects of confining stresslevel on the axial soil-pipe shear resistance at large axial displacements.The use of a particular type of device is primarily dictated by whether the resultsobtained from the device would be useful in predicting the actual behaviour of asoil-solid interaction problem encountered in the field. It is often challenging touse conventional devices to capture the soil-solid interface friction under very loweffective normal stresses. Problems are often attributed to the mechanical frictionrelated errors that are emphasized at low stress levels, limited displacement ranges,and difficulty in measuring excess pore-water pressure during shear. Limitationsarising from conventional devices to study the large-displacement interface frictionangle of soil-solid interfaces under very low effective normal stresses (i.e., 3 to 6 kPa)have prompted the development of a new interface shear test device in the researchprogram presented herein.A prudent approach for obtaining more reliable parameters would be to use dataderived from tests conducted on relatively large-scale soil-solid interface configura-tions. Hence the work undertaken in this thesis is an attempt to supplement thebody of apparatus currently being used to study soil-solid interfaces and is expected toprovide the means to measure the corresponding large-displacement interface frictionangle under very low effective normal stresses. In order to overcome such difficulties562.6. Expected Contributionsa large-scale interface direct shear test device was designed and built for conductingthe experiments reported in this thesis. The next chapter provides technical detailsof the newly built device.57Chapter 3The Macro-Scale Interface DirectShear ApparatusAs discussed in the previous chapter, the conventional interface direct shear test isrelatively simple compared to other types of interface testing methodologies, and itis often the preferred method for characterizing the behaviour of soil-solid interfaces.However, as with most conventional geotechnical shear testing devices, it does notalways provide reliable data when testing materials under low confining stress levels.This reliability issue mainly arises because the order-of-magnitude of the shear forcemobilized at the soil-solid interface under small confining stresses is comparable toforces generated due to the device compliance such as the frictional forces of themechanical components.One of the ways to overcome the above issue, is to test a specimen that hasa relatively large interface area, so that the device frictional forces would becomenotably less compared to the mobilized interface shear force.With these considerations in mind, a soil-solid interface shear test device witha large shearing footprint was designed and built by the author. The design andcommissioning of this device (henceforth referred to as the macro-scale interface directshear apparatus (MIDSA)) forms a major part of the scope of work encompassed bythis thesis. The design and fabrication of the new device commenced in May of2013 and it was commissioned in December of 2016. Many technical aspects wereconsidered in the design of the device with an emphasis on providing the means toproduce high quality test data for the characterization of soil-solid interfaces undervery low confining stresses.The newly designed macro-scale interface direct shear apparatus provides the fol-583.1. Overview of Test Devicelowing advantages over conventional interface testing devices:1. Capability of performing interface direct shear tests at very low effective normalstresses - ranging from 3 kPa to 30 kPa with minimal errors associated withdevice compliance effects such as mechanical friction;2. Allows for testing coarse-grained soils as well as fine-grained soils against varioussolid surfaces;3. Provides the means for measurement of excess pore-water pressure as well as thetotal normal stress at the soil-solid interface, which allows for the determinationof the effective normal stress at the interface; and4. Interface pore-water pressure measurements enables determining the large -displacement interface shear strength in effective stress terms even when thetest may be partially drained.Most design aspects (e.g., sizing) of the device were finalized based on mechanics-based, theoretical calculations and information found in the technical literature, whilesome aspects (e.g., approaches for vertical loading) had to be finalized by using trialand error considerations. Therefore, an array of commissioning experiments wasnecessary, to attempt and evaluate different design options and to select the mostsuitable design for the intended purpose. The following sub-sections provide detailsof the apparatus.3.1 Overview of Test DeviceThe basic loading and shear mechanism invoked by the macro-scale interface directshear apparatus is fundamentally similar to that in the conventional interface directshear test. As such, the basic principle of operation of the device is simple andidentical to that of the well-established standard direct shear device (as per ASTMD 3080-04) as outlined below:593.1. Overview of Test Device(i) the test soil specimen is laterally contained within a rigid specimen “container”while the soil specimen is supported on top of a large solid test surface;(ii) a constant normal stress is then applied to the top of the soil specimen; and(iii) after consolidation under the above vertical stress, the specimen is shearedagainst the solid surface to be tested by displacing the container with the soilspecimen at a constant rate of horizontal displacement.In order to address the above considerations, the soil specimen “containment” inthe MIDSA needed to be designed using two primary components: (i) the stationaryframe; and (ii) the mobile frame as detailed in Figs. 3.1 and 3.2. An overview ofthese two components are given below, with additional mechanical details given inthe subsequent sections.The stationary frame is used to support the solid surface to be tested, which caneither be a solid test plate, or any test material of choice coated on or attached to arigid plate, and supports the weight of the test soil specimen. The stationary frameis instrumented with an array of load cells (henceforth referred to as the base loadcells) that allow the measurement of the net vertical force acting on the test surface.Moreover, it serves as a reaction frame for the application of normal stress and sheardisplacement. The mobile frame is used to house the specimen container and it hasthe ability to move horizontally so that shearing could be invoked at the interfacebetween the test soil specimen and the test surface.As shown in Fig. 3.2, the specimen container serves as the rigid lateral contain-ment for the test soil specimen and provides an interface shear area of 1.0 m by 1.0m in plan. A set of thin, linear, and flexible rubber wipers enclosed in non-wovengeotextile sheets are mounted on the bottom perimetric interior of the four side-wallsof the specimen container, so as to prevent the test soil specimen from escaping thecontainment. The device can be used to test coarse-grained as well as fine-grained ma-terials under drained conditions. An array of pressure transducers is used to monitorthe pore-water pressure generation at the soil-solid interface during shearing; thesetransducers are mounted such that their top surface is flush with the test surface.603.1. Overview of Test DeviceMobile frameSpecimen containerTest plate Base plate / water tankActuator assemblyStationary frameMain load cellFigure 3.1: Photograph of the macro-scale interface direct shear test device.As may be noted later, this array of pore-water pressure sensors together with theabove-mentioned base load cells provide an effective means to determine the averageeffective normal stress at the interface.The next section provides detailed information of the various parts of the deviceincluding details related to the application of vertical stress to the test soil. Thebasic steps involved in the preparation of the test soil specimens of coarse-grainedand fine-grained soils are described in a separate subsequent section.613.1. Overview of Test DeviceTest soil specimen (typically 2 cm to 5 cm thick)Solid test plateLayer of sand used for drainage and application of surcharge load Static water bath for keeping test soil specimen submerged.Rigid base plate / shallow water tank 1.0 m4 mRubber/geotextile wiper used for preventing loss of soil from the specimen container.0.3 mNon-woven geotextile layer used for separation of test soil from surcharge/drainage sand layer.Rigid specimen container.Rubber bladder(for applying air pressure).Load cell.Stepper motor/worm and gear.String potentiometer.NorthPore-water pressure transducers. Hinge.Base load cells(to measure total normal force).Stationary frame.Mobile frame + specimen containerPTSHPTSH : Pressure transducer for measureing static water head of the water bath.Figure 3.2: Schematic diagram of the macro-scale interface shear test device and testsetup.623.2. Description of Main Components of Test Device3.2 Description of Main Components of TestDevice3.2.1 Stationary frame and normal force measurementsystemThe rigid stationary frame, constructed of standard structural steel members, servesas the reaction frame for the test device. As shown in Fig. 3.3 the top middle portionof the stationary frame has four support I-beams (beams B1 through B4). These fourI-beams are supported on the array of 12 base load cells (labeled LC1 through LC12),with three load cells supporting each I-beam.As shown in Fig. 3.4, the base load cells are of low profile type and are used tomeasure the vertical force acting on the four support I-beams. A rigid base plate,surrounded by perimeter walls, measuring approximately 1.2 m by 3.0 m in plan, thatwould serve as: (i) a supporting platform for the solid surface to be tested; and (ii) ashallow water reservoir in keeping the test soil specimen submerged in water duringtesting, is mounted on top of the four support I-beams as shown in Fig 3.5. The baseplate together with the four support I-beams form a rigid structure that is supportedon the base load cells. Any vertical force applied to the base plate is measured withthe use of the base load cells. A water pressure transducer (see pressure transducerPTSH in Figs. 3.4 and 3.5) that is attached to the base plate at the north end allowsthe determination of the static water head in the shallow water tank. Because thebase load cells measure the total vertical force acting on the base plate, the weightof the static water head needs to be corrected for when determining the total normalforce acting at the soil-solid interface. This is carefully accomplished using the totalvertical force measured by the base load cells and the water head determined withthe pressure transducer.The specimen of the solid surface to be tested has to be a large plate made ofsteel measuring a maximum of 1.2 m (4 ft) wide, a minimum of 2.4 m (8 ft) long,and typically 6.0 mm (0.25 inch) thick (for compatibility with the dimensions of the633.2. Description of Main Components of Test DeviceB1B4Linear rail (east)Linear rail (west)1.23 m4.00 m1.64 mNorthB2B3LC12LC11LC10LC9LC8LC7LC6LC4LC3LC2Figure 3.3: Drawing of stationary frame. (Note that the base plate, test plate, andthe mobile frame are not shown).device), coated or lined with the solid test material of choice. The solid plate is firmlymounted to the base plate using machine screws. For example, if the interest is toobtain the interface shear strength of a given soil against an epoxy coated steel surface,the epoxy coating is pre-applied to the steel plate and it is then mounted securelyon top of the base plate. This setup, together with the base load cells, provides themeans to determine the total normal force acting at the soil-solid interface.The mobile frame, which supports the weight of the specimen container, is sup-ported on two precision linear rails on the east and west periphery of the stationaryframe (see Figs. 3.5, 3.6 and 3.8). The device is designed in such a way that theweight of the mobile frame and the specimen container is not transferred to the base643.2. Description of Main Components of Test DeviceBase load cellsB1B2B3B4LC5LC6LC7LC8Pressure transducer for measuring static water head  PTSHFigure 3.4: Photograph showing base load cells.load cells. This allows the determination of the total normal force acting on the in-terface arising due to the test soil specimen only. The linear actuator that is usedto apply a constant rate of horizontal displacement to the mobile frame is mountedon the stationary frame at the north end. More details on the actuator assembly isprovided in the following sub-section.The pore-water pressure at the interface is measured using an array of 12 pore-water pressure transducers that are mounted flush with the test surface. Figure 3.7shows the planar spatial distribution of the pore-water pressure transducer apertureson the test plate. This layout of pressure transducers has been carefully selected toensure that an optimal planar area of the soil specimen is monitored for pore-waterpressure during testing. It is of relevance to note that the 12 pore-water pressuretransducers are not to be confused with the pressure transducer PTSH used to de-653.2. Description of Main Components of Test DeviceNorthBase plate / shallow water tankTest plateActuator assembly2.4 m1.2 mPP11PP21PP41PP12PP22PP32PP42PP13PP23PP33PP43PP31Linear rail (east)Linear rail (west)PTSH3.0 mFigure 3.5: Drawing of stationary frame with the base plate and test plate installed.termine the static water head in the water tank. However, the static water headmeasured by PTSH should ideally be the same as the pore-water pressure measuredby the pore-water pressure transducers under static conditions when the test soilspecimen has reached close to 100% consolidation.3.2.2 Mobile frame and actuator assemblyThe mobile frame is of stiff structural steel construction and is supported on thestationary frame via four precision linear bearings (horizontal linear bearings in Figs.3.8 and 3.9). The primary purpose of the mobile frame is to provide support for thespecimen container (see Figs. 3.8 and 3.10).663.2. Description of Main Components of Test DeviceNorthMobile frameTest plateActuator assemblySpecimen containerMain load cellHinge connectionMobile frame load cells (west)1.0 m1.0 m0.3 mLinear rail (east)Linear rail (west)PTSHThreaded rodFigure 3.6: Drawing of stationary frame with the base plate, test plate, and mobileframe installed.The horizontal linear bearings constrain the displacement of the mobile frame tothe north and south directions in the horizontal plane. Hence, any vertical forceapplied onto the mobile frame is transferred to the stationary frame through thehorizontal linear bearings. This allows the mobile frame to act as a reaction frame forvertical forces while at the same time allowing for free displacement in the horizontalplane. The capability of the mobile frame to act as a reaction frame is used forthe normal force application system (further details on this is provided later in thischapter).The specimen container is constructed with 50.0 mm (2 inch) thick solid aluminumslabs joined to form a 1.0 m (Length) × 1.0 m (Width) × 0.3 m (Height) laterally673.2. Description of Main Components of Test DeviceSpecimen container1.0 mPP11 PP12 PP13PP21 PP22 PP23PP31 PP32 PP33PP41 PP42 PP430.1 m0.3 m 0.3 m0.3 m0.2 mTest plateFigure 3.7: Photograph showing the arrangement of pore-water pressure ports on thetest plate.enclosed volume. The specimen container is supported on four load cells (two of theside load cells are shown in Fig. 3.10) that are mounted on the east and west sidesof the mobile frame. The specimen container is laterally constrained by the mobileframe using eight vertical linear bearings, and this allows the specimen containerto be moved in the vertical plane. As such, the weight of the specimen container istransferred to the mobile frame, and the total weight of the mobile frame and specimencontainer is transferred to the stationary frame through the horizontal linear bearings.Therefore, the weight of the specimen container and the mobile frame is not sensed bythe base load cells. The side load cells are used to determine the side-wall friction (i.e.,friction between the test soil specimen and the specimen container) that is mobilizedduring specimen preparation and testing. Fig. 3.11 shows a close-up view of how the683.2. Description of Main Components of Test DeviceMobile frameSpecimen containerVertical linear railSide load cell Side load cellHorizontal linear bearingStationary framePore-water pressure transducersPressure transducer for measuring static head(PTSH)B1 B2 B3 B4 Shallow water tankBase load cells1.7 m EastFigure 3.8: End elevation of macro-scale interface direct shear apparatus.specimen container is supported on the side load cells.As shown in Fig. 3.11 the specimen container is supported on the side load cells viagap adjustment bolts. The gap adjustment bolts allow for the gap between the bottomof the specimen container and the test surface to be adjusted. The gap between thespecimen container and the test surface needs to be set in such a way that the lossof the test soil through the gap is minimized. However, an excessively small gap canresult in portions of the test soil being trapped underneath the specimen containerduring shearing which, in turn, can introduce an unwanted friction force between thespecimen container and the test surface. To address this issue a sufficiently large gap(approximately 1.0 cm) is set between the specimen container and the test surfaceand a flexible wiper is installed on the inside of the specimen container to close thegap. The wiper is a rubber wiper covered by a layer of non-woven geotextile, and it693.2. Description of Main Components of Test DeviceStationary frameHorizontal linear bearingHorizontal linear railMobile frameFigure 3.9: Photograph showing the north-east horizontal linear bearing.remains in contact with the test surface during the course of a test.The selection of the rubber wiper was made by performing a series of tests, with-out a soil specimen in place, to observe the frictional resistance that the wiper rubbingagainst the test surface would produce. It was found that a fairly thin and relativelyflexible wiper made of rubber produced an adequately low frictional resistance. How-ever, the rubber wiper alone was not effective in containing a slurry of fine-grainedsoil. This issue was solved by covering the rubber wiper with a layer of non-woven geo-textile. With all the wipers installed, the frictional force due to the wipers sweepingthe test surface, and the frictional force produced by the horizontal linear bearings,combined, (i.e., total device friction) was found to fall between 0.05 kN (on a smoothepoxy coated steel surface) and 0.15 kN (on a rough sand-blasted mild steel surface).For very low stress testing performed at a normal effective normal stress of approxi-703.2. Description of Main Components of Test DeviceSpecimen containerMobile frameMain load cellTest plateBase plate / water tankHorizontal linear rail (west)Horizontal linear rail (east)Vertical linear railsSide load cells0.3 mRubber / geotextile wiperFigure 3.10: Photograph showing mobile frame and specimen container.Side load cellsMobile frameGap adjustment boltsFigure 3.11: Photograph showing side load cells.713.2. Description of Main Components of Test Device(a) Specimen container (b)(c) (d)Wipers (installed)Test plateWiper coversWipers (installed)Wiper removed from specimen containerGeotextile layerRubber wiper coreFigure 3.12: Photographs showing the geotextile covered rubber wipers of the mobileframe.mately 3 kPa, this range of device friction is between 5% and 15% of the measuredshear force. For tests conducted at normal effective stresses in the range of 10 kPato 30 kPa, the device friction varies between 0.5% to 5% of the measured shear force.It is of relevance to note that for soil-solid interface direct shear testing, the devicefriction is typically measured at the end of each test. At the end of a given test, beforetesting for the device friction, only the test soil specimen enclosed by the specimencontainer is removed, leaving any soil that is trapped between the specimen containerand the test surface untouched. This ensures that any friction developed due to anytrapped soil is also captured in the measurement of device friction. The interfaceshear data obtained during the test is then corrected for the device friction measuredpost test.723.2. Description of Main Components of Test DeviceStepper motorWorm-gear actuatorThreaded rodString potentiometerDouble-hinge connectionMain load cellMobile frameBase plate / water tankFigure 3.13: Photograph showing the main load cell and actuator assembly of themacro-scale interface direct shear apparatus.North (Direction of shear)Test plateVertical linear bearingRubber wiperSpecimen container(Not to scale)Main load cellVertical linear railPin-1 Pin-2Threaded rodTie rod connectionMobile frameFigure 3.14: Schematic cross-sectional view of the wiper and main load cell assemblyof the macro-scale interface direct shear apparatus.The shear force mobilized during interface shear testing is measured using a loadcell that is attached to the mobile frame (main load cell in Fig. 3.13). Shearing of thesoil against the test surface is accomplished by applying a constant rate of horizontal733.2. Description of Main Components of Test Devicedisplacement to the mobile frame using a precision worm-gear linear actuator systemdriven by a stepper motor. The threaded rod of the actuator is connected to themain load cell via a double-hinge pin connection to ensure that no bending momentsare transferred to the load cell during pulling (pulling force is applied towards thenorth direction; see Fig. 3.13). This setup allows for a total shear displacementof 1.0 m at displacement rates ranging between 0.0001 mm/s and 1.0 mm/s. Thedisplacement rate is kept constant over the duration of a given interface shear test.The lower bound displacement rates attainable in this device are used when testingfully saturated fine-grained soils targeting drained conditions.3.2.3 Normal force application systemThe testing for the coefficient of interface friction using the macro-scale interfacedirect shear apparatus is carried out at constant normal effective stresses at theinterface. Also, the normal effective stress is kept relatively uniform across the planararea of the interface. The interface shear test program presented in this thesis coversa target effective normal stress range of 3.0 to 30.0 kPa (details of the test programare provided in a later section of the text). The 1.0 m × 1.0 m planar interface areatranslates this stress range into a normal force range of 3.0 to 30.0 kN. A highlypreferred method of applying a constant normal force is the use of dead weights as asurcharge load on top of the test soil specimen. However, given the relatively largeinterface area used in the macro-scale interface direct shear apparatus, the applicationof this range of normal force poses several complications. The 3.0 kN normal force canbe achieved by using a layer of sand, approximately 0.3 m thick, on top of the test soilspecimen. In this case, the layer of sand fills a gross volume of 1.0 m × 1.0 m × 0.3m resulting in a mass of approximately 360 kg (i.e., 0.3 m3× 1200 kgm−3). However,achieving a normal force of 30.0 kN using a dead weight as surcharge (approximately3000 kg) is not practical. A more practical method of applying larger normal stressesis the use of a pneumatic pressure membrane which, upon application of air pressure,presses against the 1.0 m × 1.0 m soil specimen area to produce the required normalforce. This method requires an air pressure of approximately 30 kPa to produce a743.2. Description of Main Components of Test Devicenormal force of 30 kN. In this case, a reaction frame, capable of transferring the30 kN vertical force to the stationary frame, is required. The macro-scale interfacedirect shear apparatus uses both of the above mentioned methods of normal forceapplication. While normal forces of 3.0 kN and below are achieved using the deadweight surcharge loading system, normal forces above 3.0 kN are achieved using acombination of the dead weight surcharge loading system and the pneumatic pressureapplication system as shown in Fig. 3.15.To produce the dead weight surcharge load, the test soil specimen (which is gen-erally 2.0 to 10.0 cm thick) is first covered with a perforated load transfer plate and alayer of non-woven geotextile. Then, a uniform layer of Fraser River sand of requiredthickness is gradually and systematically air-pluviated on top of the geotextile layer.A commercially available non-woven geotextile having an apparent opening size of0.212 mm (ASTM-D4751) and a permitivity of 1.3 s-1 (ASTM-D4491) is used for thispurpose. The perforated load transfer plate is a 3.0 cm thick perforated plywood platethat is encapsulated in a non-woven geotextile case (see Figs. 3.16-a and 3.16-b). Asshown in Fig. 3.16-a, the edges of the load transfer plate are covered with spongeswhich, once enclosed in the geotextile casing, act as a wiper that produce a close-fitbetween the load transfer plate and the side walls of the specimen container. The loadtransfer plate aids in damping out any planar non-uniformity in the surcharge sandload before it is transferred to the test soil specimen. The load transfer plate andthe layer of non-woven geotextile separate the test soil specimen and the surchargesand layer to ensure that the surcharge sand layer does not contaminate the test soilspecimen, and allows any excess pore-water pressure generated within the test soil todissipate through the surcharge sand layer. Fig. 3.16-c shows the load transfer plateplaced on top of a test soil specimen, and Fig. 3.16-d shows the surcharge sand layerplaced on top of the load transfer plate.As mentioned earlier, the sand surcharge is only sufficient to produce a normalforce of 3.0 kN and less. In order to apply higher normal forces, a pneumatic pressureapplication system is used. The pneumatic pressure plate shown in Fig. 3.17 is a0.5 inch thick solid polycarbonate plate that has a rubber membrane on one side.753.2. Description of Main Components of Test DeviceMobile frameSurcharge load / drainage sand layerNon-woven geotextileTest soil specimenBase plate / water tankTest plateSide load cellStationary frameHorizontal linear bearingSpecimen containerBase load cellsPore-water pressure transducersStatic water layerVertical linear bearingshtshslWiperPerforated load transfer plate(a)Air pressure transducerMobile frame with reaction frame for air pressure plateAir pressure plateAir pressure membrane Surcharge load / drainage sand layerNon-woven geotextileTest soil specimenBase plate / water tankTest plateSide load cellStationary frameHorizontal linear bearingSpecimen containerBase load cellsPore-water pressure transducersStatic water layerVertical linear bearingshtshslWiperPerforated load transfer plate(b)Figure 3.15: Schematic diagrams showing the normal force application system. (a)Dead weight surcharge loading system, (b) Dead weight surcharge loading system incombination with the pneumatic pressure application system763.2. Description of Main Components of Test DevicePerforated plywood plateSponge wiper arrangementLoad transfer plate1.0 m1.0 mSpecimen containerLoad transfer plateSurcharge load sand layerNon-woven geotexille layer(a) (b)(c) (d)Figure 3.16: Photographs showing the load transfer plate of the macro-scale interfacedirect shear apparatus.(a)Solid polycarbonate plate (1/2 inch thick)(b)Nitrile rubber membrane (1/16 inch thick)Clamping bars1.1 m1.1 m1.0 m1.0 mFigure 3.17: Photographs showing the pneumatic pressure plate of the macro-scaleinterface direct shear apparatus.773.2. Description of Main Components of Test DeviceThe rubber membrane is a 1/16 inch thick nitrile rubber sheet having a durometerhardness of 60A as per ASTM D2240-15 (D2240-15, 2010). The rubber membrane isclamped to the pressure plate at the four edges to produce an air-tight seal. Duringpreparation for testing, after applying the surcharge sand load, the specimen is al-lowed to consolidate prior to application of any additional normal forces that may benecessary. In cases where fully saturated fine-grained soils are tested, the increase in(a)Surcharge load sand layer(b)(c) (d)Geotextile covering surcharge load sand layerPneumatic pressure plate installed I-beams of reaction frameFigure 3.18: Photographs showing the installation sequence of the pneumatic pressureapplication system of the macro-scale interface direct shear apparatus.the pore-water pressure at the interface due to the gradual and uniform filling of thesurcharge sand layer is continuously monitored using the array of pore-water pressuretransducers. Once consolidation under the sand surcharge is complete, the surfaceof the sand layer is covered with a layer of non-woven geotextile, and the pressureplate is placed on top of the specimen container. Four structural I-beams placed on783.2. Description of Main Components of Test Devicetop of the pressure plate and tied to the mobile frame act as the reaction frame forthe pneumatic pressure application system. Figure 3.18 shows how the pressure plateis installed together with the reaction frame. Inflating the pressure plate causes therubber membrane to expand and push against the surcharge sand surface. The re-sulting vertical reaction force is transferred to the mobile frame through the reactionframe, and the horizontal linear bearings transfer the reaction force to the stationaryframe. A pressure transducer connected to the air pressure supply line of the pressureplate is used to monitor the air pressure inside the pressure plate. The vertical forceapplied to the surface of the surcharge sand layer gets transferred to the interface andit is measured using the base load cells.Uniformity of total normal stress on the interfaceWhile the total normal force acting on the interface is captured by the base load cells,they do not provide information about the spatial distribution of the total normalstress. It was of particular interest to determine the degree of spatial uniformity ofthe total normal stress that results from the combined loading of the surcharge sandlayer and the pneumatic pressure application system. In order to accomplish thistask, a device capable of measuring the spatial distribution of the total normal stresswithin a 1.0 m × 1.0 m footprint was custom built. This device is comprised of a12.0 mm (0.5 inch) thick steel plate measuring 0.9 m × 1.6 m in plan that serves asa rigid support for an array of 42 load cells that support a 3×7 matrix of pressurepads as shown in Fig. 3.19.The pressure pad matrix device is not a component of the macro-scale interfacedirect shear apparatus, but was merely built to conduct a limited number of experi-ments to determine the spatial distribution of the total normal stress at the interface.The pressure pad matrix was installed underneath the specimen container as shownin Fig. 3.19-d. The pressure pad matrix provides a constant vertical stiffness uni-formly distributed over the entire 1.0 m × 1.0 m footprint which helps to minimizeany effects that may arise due to the vertical deformation of the pressure pads duringloading. Several tests were undertaken to determine the spatial distribution of the793.2. Description of Main Components of Test Device(a) ½ inch thick steel plate (b)0.9 m(c) (d)1.6 mBeam load cell ¼ inch thick steel plate (pressure pad)1.0 m1.0 m0.33 m0.14 mBeam load cells¼ inch thick steel plate (pressure pad)Specimen container3x7 Pressure pad matrixPD11PD12PD13PD14PD15PD16PD17PD21PD22PD23PD24PD25PD26PD27PD31PD32PD33PD34PD35PD36PD371.0 m1.0 mFigure 3.19: Photographs showing the pressure pad matrix device used to determinethe spatial distribution of total normal stress.total normal stress acting on the interface due to the surcharge sand layer and thecombined loading of the surcharge sand layer and the pneumatic pressure applicationsystem. In these tests, Fraser River sand was used as the test soil. Figs. 3.20 and 3.21show the spatial distribution of the total normal stress as captured by the pressurepad matrix for the surcharge sand loading and the combined loading respectively.It is important to understand that the use of a pneumatic pressure applicationsystem has many technical complexities that need to be addressed in order to arrive ata practical system that works to produce reliable stress boundary conditions requiredfor interface direct shear testing. First of all, the stiffness of the membrane resultsin a normal stress distribution that is somewhat higher towards the centre and lessertowards the edges of the interface footprint, the least amount of normal stress being803.2. Description of Main Components of Test DeviceEast-west coordinates of interface plane (m) South-north coordinates of interface plane (m) Figure 3.20: Spatial distribution of total normal stress at the interface due to thesurcharge sand layer (contour values are in kPa).applied at the corners. This issue can be resolved by using a relatively thin andextremely flexible membrane that can easily fill the volume between the pressure plateand the surcharge sand surface. However, producing such a membrane is not a trivialtask. A major time commitment was necessary to determine a suitable method ofproducing such a membrane. After many trial-and-error attempts, a feasible solutionwas reached where a natural latex liquid rubber material was applied to a mold toproduce the required membrane (see Fig. 3.22). However, the latex does not havea good shelf life, and tends to degrade and rupture between tests. Producing a newmembrane using liquid latex took approximately two weeks, which, given typicaltime constraints, made its use impractical. As an alternative, instead of using theliquid latex rubber, a synthetic liquid silicone rubber was later used to produce themembrane. This method was found to work quite well and can be recommended forfuture projects as a membrane that can typically be produced within two to threedays, and this results in a very flexible membrane that has a relatively long shelf life.However, the silicone rubber is relatively expensive, and needs to be used with care813.2. Description of Main Components of Test DeviceEast-west coordinates of interface plane (m) South-north coordinates of interface plane (m) Figure 3.21: Spatial distribution of total normal stress at the interface due to thecombination of surcharge sand layer and the pneumatic pressure application system(contour values are in kPa).Liquid latex applied to mold15.0 cmCured latex membrane being removed from moldFigure 3.22: Photograph showing the natural latex rubber membrane being prepared.to prevent rupture of the membrane. Another, more robust option is the use of athin sheet (typically 1/16 inch or less) of nitrile rubber as the membrane. This wasthe simplest to produce and was less expensive than the silicone rubber membrane.This was used as the final solution for the pressure plate of the macro-scale interface823.2. Description of Main Components of Test Devicedirect shear apparatus.It is also important to understand that the normal force applied by both thesurcharge sand layer and the pneumatic pressure application system is affected bythe side-wall friction mobilized between the surcharge sand layer and the walls ofthe specimen container. This side-wall friction effect also works to produce a non-uniform normal stress distribution on the interface footprint. The side-wall frictioncan be reduced to some extent by using two layers of polyethylene sheets (preferablywith a layer of lubricant in between the sheets) between the surcharge sand and thewalls of the specimen container (Tognon et al., 1999). Given that the side-walls ofthe specimen container are relatively smooth, it is not essential that polyethylenesheeting is used for the side-walls. As mentioned earlier, the macro-scale interfacedirect shear apparatus is equipped with side load cells that are used to monitor theside-wall friction forces when necessary.Nevertheless, it should be borne in mind that the total normal stress applied onthe interface is not perfectly uniform. In spite of this, the total normal stress isassumed to be uniform over the interface and an average total normal stress that actson the interface is used for interpretation of data treating the 1.0 m × 1.0 m interfacefootprint as an elemental area. The average total normal stress is obtained using thetotal normal force measured by the base load cells of the macro-scale interface directshear apparatus and then by dividing the force by the planar area of the interface.Further details on this is provided next.Determination of the average total normal stressAs explained earlier, the base load cells of the macro-scale interface direct shearapparatus measure the total force acting on the base plate. During a typical test theforces acting on the base plate are: (i) the total normal force acting on the interfacefootprint arising from the weight of the saturated test soil, the weight of the sandsurcharge, and the pneumatic loading applied on top of the surcharge sand layer; and(ii) the weight of the water contained in the shallow water tank.In order to obtain the total normal stress acting on the interface footprint, the833.2. Description of Main Components of Test Devicea1a1a2a3Specimen container footprint (area Ac = a12)Footprint of water layer (area Aw = a2.a3)Specimen/interface footprint (area Ai = a2)aaPressure transducer (PTSH) footprintFigure 3.23: Schematic diagram showing specimen container and water tank foot-prints used for total normal stress calculation.total normal force measured by the base load cells must first be corrected for theweight of the water contained in the water tank. Specifically, the vertical force actingon the base plate arising from the water that surrounds the specimen container needsto be subtracted from the total force measured by the base load cells. The planar areaof the water surrounding the specimen container (Awc) can be obtained using Eq. 3.1(see Fig. 3.23 for definitions of the footprint areas). The water pressure (PSWH) ofthe static water head in the water tank is measured using the pressure transducer(PTSH). As such, the total normal stress acting on the interface is obtained usingEq. 3.2. The average effective normal stress is calculated using Eq. 3.3. The methodof measuring pore-water pressure is presented in the next subsection.Awc = Aw − Ac (3.1)σn =FBLC − PSWH .(Aw − Ac)Ai=FBLC − PSWH .AwcAi(3.2)σ′n = σn − u (3.3)843.2. Description of Main Components of Test Devicewhere, σn = average total normal stress acting on the interface,FBLC = total force acting on the base plate as measured by the base load cells,PSWH = static water pressure of the water contained in the water tank as measuredby the pressure transducer (PTSH),Aw, Ac, Ai are planar footprint areas as defined in Fig. 3.23,σ′n = average effective normal stress at the interface, andu = average pore-water pressure at the interface.3.2.4 Measurement of pore-water pressureAs discussed earlier, the macro-scale interface direct shear apparatus is equipped withpore-water pressure transducers which provide measurements for the determinationof the effective normal stress at the interface. This allows for the determination ofthe effective angle of interface friction.1/4NPT to 1/8NPT adapter Pressure transducer (Gauge type, rated for 70 kPa)Figure 3.24: Photograph of pressure transducer that is used for pore-water pressuremeasurement.853.2. Description of Main Components of Test DeviceSoil specimenTest platePressure transducerDe-aired glycerin ordistilled, de-aired waterStainless steel adapterNorth (Direction of shear)Porous stoneData cablePRODUCED BY AN AUTODESK EDUCATIONAL PRODUCTPRODUCED BY AN AUTODESK EDUCATIONAL PRODUCTP RODUCE D BY  AN AUT ODE S K E DUCAT I ONAL  P RODUCTPRODUCED BY AN AUTODESK EDUCATIONAL PRODUCTFigure 3.25: Details of pressure transducer connection to the test plate. (Amaras-inghe, 2013)The pressure transducers used in the device are of Gauge type with a range of0 to 70 kPa (see Fig. 3.24). The sensors are mounted to the test plate through anadapter such that the opening of the adapter is flush with the test surface (see Fig.3.25). This allows for the measurement of pore-water pressure right at the soil-solidinterface level. It is important to make sure that the pressure transducer ports do notaffect the shear behaviour of the soil. For example, having the ports not flush withthe interface can lead to localized strain non-uniformity around the port adverselyaffecting the measurement of pore-water pressure. It is also important that the portsare made as small as possible as the roughness of the port-porous stone surface isdifferent from the interface and this can lead to local strain non-uniformity affectingthe measurement.The pressure transducers and the adapters are well cleaned and fully saturatedwith a de-aired, incompressible liquid medium such as glycerine or water before place-ment of the soil specimen into the mobile frame. Installation of the pressure trans-ducers is carried out by first attaching the adapters to the test plate from underneaththe test plate, and then attaching the saturated pressure transducers to the adapters.863.2. Description of Main Components of Test Device(a) (b) (c)SyringeFlexible rubber tubeDe-aired glycerine or,de-aired distilled waterPRODUCED BY AN AUTODESK EDUCATIONAL PRODUCTPRODUCED BY AN AUTODESK EDUCATIONAL PRODUCTP RODUCE D BY  AN AUT ODE S K E DUCAT I ONAL  P RODUCTPRODUCED BY AN AUTODESK EDUCATIONAL PRODUCTFigure 3.26: Details of pressure transducer saturation procedure. (a) Saturation ofpressure transducer cavity. (b) Attachment of saturated transducer to stanless steeladapter and saturation of adapter. (c) Placement of porous stone. (Amarasinghe,2013)Then each adapter is saturated with the selected liquid medium with the aid of asyringe fitted with a flexible rubber tube.The saturation of the pressure transducer is carried out by first inserting theflexible tube into the cavity of the pressure transducer and then filling the wholecavity with the liquid until the liquid overflowing out of the cavity does not containany air bubbles as shown in Fig. 3.26-a. Care is needed to ensure that no air bubblesare introduced into the cavity during this process. Once the cavity is completely filleda meniscus of the liquid covering the port can be observed as shown in Fig 3.26-a,b.The saturated pressure transducer is then carefully attached to the adapter ensuring873.2. Description of Main Components of Test Devicea tight seal (thread sealant is used to aid in this operation). The adapter cavity isthen carefully filled with the liquid starting from the base of the transducer cavity allthe way to the opening of the adapter at the test surface as shown in Fig. 3.26-b,c.The cavity is filled until no air bubbles are visible in the overflowing liquid. A concavemeniscus is left on top of the adapter opening. The opening is then covered with asaturated porous stone ensuring that no air bubbles are introduced.1.0 m1.0 mPosition of specimen container  at start of shearing.0.8 mPosition of specimen container at end of shearing.Pore-water pressure transducer ports. (4x3 matrix).Test surface.Rows 3 and 4 exposed.Row 1Row 2Row 3Row 4North (direction of shear displacement).Figure 3.27: Schematic diagram showing the initial and final position of soil specimenwith respect to the pore-water pressure transducer ports.883.2. Description of Main Components of Test DeviceThe spatial distribution of the pressure sensors on the test surface has been es-tablished based on ensuring optimum coverage of the soil specimen during shearing.When the specimen container is in the southern end of the test plate, all 12 pore-water pressure sensor ports are enclosed within the specimen container as shown inFig. 3.27. Hence, during the preparation of the test soil, the pore-water pressuresensors allows for the monitoring of the pore-water pressure as well as the spatial dis-tribution of the pore-water pressure within the interface footprint. This is especiallyuseful during the consolidation stage (note that a detailed description of the specimenpreparation procedure is provided in the next section of the text). As the specimenis sheared against the test surface by displacing the specimen container towards thenorth direction, the positions of the pore-water pressure ports with respect to thespecimen container change continuously. As shown in Fig. 3.27, towards the end ofshearing, only two rows of pore-water pressure sensors (namely, rows 1 and 2) remainwithin the footprint of the interface. Hence, it should be understood that only thesetwo rows provide the pore-water pressure at the interface during the entire durationof a given test. However, rows 3 and 4 are still provide the means to check and gainconfidence in the pore-water pressure measurement. As the third and fourth rowsof pore-water pressure sensors exit the footprint of the soil specimen as it is beingsheared, they get exposed to the static water head of the water enclosed in the watertank, and the pressure sensor readings can be checked to see whether they providethe correct water head at that stage.Figure 3.28 shows a time history of pore-water pressure observed during a typicalconsolidation stage of a specimen of kaolinite under a surcharge load sand layer. Inaddition to the pore-water pressure sensor data and the average pore-water pressuredata, included in the figure are the time histories of the static water head PSWH ;the total normal force acting on the base plate FBLC ; the average total normal stressσn; and the average effective normal stress σ′n. The normal stresses are calculatedusing Eqs. 3.2 and 3.3. The application of normal force, using the sand surchargealone, or the combined loading of sand surcharge and the pneumatic loading system,results in the generation of excess pore-water pressure. As the consolidation pro-893.2. Description of Main Components of Test DeviceStress or pressure (kPa) and Force (kN)Time (hour)16 18 20 22 24 260246810 FBLC (kN)PP11 (kPa)PP12 (kPa)PP13 (kPa)PP21 (kPa)PP22 (kPa)PP23 (kPa)PP31 (kPa)PP32 (kPa)PP33 (kPa)PP41 (kPa)PP42 (kPa)PP43 (kPa)Average PP (kPa)Effective normal stress σn' (kPa)PSWH.Awc (kN)Total normal stress σn (kN)PSHW (kPa)FBLCσn σn'PSWH.AwcAverage pore-water pressurePSHWPore-water pressure rise due to placement of sand surcharge.Figure 3.28: Time history of pore-water pressure readings for a Kaolinite specimenon a mild steel test surface obtained during application of sand surcharge and con-solidation.ceeds to completion, the excess pore-water pressure dissipates, and the pore-waterpressure eventually reaches equilibrium with the static water head. The generationof excess pore-water pressure during the gradual placement of the sand surchargelayer can be observed between time stamps of 17 to 17.5 hours. The average excesspore-water pressure thus generated equals the increase in the average total normalstress. Also evident is that the pore-water pressure distribution over the footprint ofthe interface is fairly uniform, with slightly higher values observed towards the centreof the footprint, which is in line with the expected spatial distribution of total normalstress as discussed in the previous subsection. However, as shown in Fig. 3.29, thespatial distribution of pore-water pressure is more pronounced during the applica-tion of higher normal stresses using the pneumatic pressure application system. Asdiscussed earlier, various factors, such as the stiffness of the rubber membrane andthe side-wall friction of the specimen container, could cause a more pronounced spa-tial non-uniformity of normal stress, which can be observed in the excess pore-waterpressure data immediately after applying the pneumatic pressure at time stamp 26.7hours.903.3. Details of Specimen PreparationStress or pressure (kPa) and Force (kN)Time (hour)26 26.5 27 27.5 28 28.5 29 29.5 30 30.5051015202530 FBLC (kN)PP11 (kPa)PP12 (kPa)PP13 (kPa)PP21 (kPa)PP22 (kPa)PP23 (kPa)PP31 (kPa)PP32 (kPa)PP33 (kPa)PP41 (kPa)PP42 (kPa)PP43 (kPa)Average PP (kPa)Effective normal stress σn' (kPa)PSWH.Awc (kN)Total normal stress σn (kN)PSHW (kPa)FBLC σn σn'PSWH.AwcAverage pore-water pressurePSHWPore-water pressure rise due to the pneumatic pressure application.Figure 3.29: Time history of pore-water pressure readings for a Kaolinite specimenon a mild steel test surface obtained during application of pneumatic loading andconsolidation.3.3 Details of Specimen PreparationAs indicated earlier, the macro-scale interface direct shear apparatus can be used totest coarse-grained and fine-grained soils on various solid test surfaces. This sectiondescribes the basic procedure followed in the preparation of each type of test soilspecimen.Prior to preparing the test soil specimen inside the specimen container, it is im-portant that the test plate and the specimen container are thoroughly cleaned. Figure3.30 shows photographs after cleaning and setup of the specimen container prior toplacement of the test soil specimen. The interior walls of the specimen container arecleaned by scrubbing and washing with water to remove any contaminants. The fourwipers of the specimen container are typically removed and are also cleaned beforereinstalling in the specimen container. The wipers are installed after the specimencontainer and the test plate have been cleaned. Cleaning of the test plate is typ-ically done by scrubbing and washing the surface of the test plate with water andsometimes with a degreasing agent, with the aim of removing any materials that havecontaminated the test surface. A test plate is typically reused several times during913.3. Details of Specimen Preparation(a) (b)(c) (d)North wiper in placeClean test plate and specimen containerWiper covers being installedWiper covers installed and specimen container in starting positionGap set as requiredPolyethylene sheeting installedFigure 3.30: Photographs showing cleaning and setup of specimen container and testplate prior to specimen placement.a given test program. Each test plate was reused up to a maximum of 9 times inthe present research program. The need to reuse the test plates primarily arises fromthe cost of the plates and budget constraints. Given that coarse-grained soils caneasily alter the surface texture of a given test plate during testing, all tests involvingcoarse-grained soils were conducted at the very last stage of the testing program (i.e.,testing commenced with the fine-grained soils and progressed towards coarse-grainedsoils). In such cases, the test plate is not removed from the device, but is cleaned wellin between tests to ensure that the there is no residue of the previous test soil presenton the test plate prior to preparing the new test soil specimen. When a test plate isbeing reused for consecutive tests, the pore-water pressure transducers are typicallykept attached to the test plate. During the cleaning operation, it is important to923.3. Details of Specimen Preparationclose the pore-water pressure ports to prevent any contaminants from entering theports.Once the specimen container and the test plate are cleaned, the gap adjustmentbolts are used to set the required clearance between the specimen container and thetest surface. As discussed earlier, the gap between the specimen container and the testplate should not be too small (as this can lead to a notable increase in device friction)or too large (as this can lead to loss of the test soil from the specimen container).Once a suitable gap is set, the wipers are installed to close the gap. A gap between0.5 to 1.0 cm was found to work well in most cases. Once the specimen container andthe test plate are cleaned and setup, the pore-water pressure sensors are saturated.As discussed earlier, the friction between the surcharge load sand layer and thewalls of the specimen container can affect the uniformity of the normal stress distri-bution on the interface. While completely eliminating this effect is not feasible, it ispossible to reduce the side-wall friction by lining the interior walls of the specimencontainer with two layers of polyethylene sheeting, preferably with a layer of greasein between the two sheets (Tognon et al., 1999). Based on many experiments utiliz-ing the side load cells of the device, it was found that this approach can reduce thesidewall friction force by approximately 10%. As discussed earlier, the normal stressuniformity is affected by the stiffness of the pneumatic pressure membrane and wasfound to be important to control.3.3.1 Preparation of coarse-grained soil specimensPreparation of coarse-grained soil specimens is relatively straight forward and involvesuniformly placing the test soil material inside the 1.0 m by 1.0 m footprint of thespecimen container. Prior to placement, the test soil is dried either using an oven orpreferably by natural evaporation under room temperature. Having a dry soil massaids in producing a relatively uniform specimen. The test soil is air-pluviated intothe specimen container in a consistent manner from an approximate height of 0.5 musing a container of 1 litre capacity.933.3. Details of Specimen Preparation3.3.2 Preparation of fine-grained soil specimensThe testing of fine-grained soils in the macro-scale interface direct shear apparatusis carried out under saturated conditions. The determination of the effective normalstress at the soil-solid interface depends upon on the accurate measurement of pore-water pressure generated at the interface during shearing. Therefore, special care istaken to aim for a fine-grained soil specimen that is saturated. In order to producea saturated specimen, the fine-grained soil is first mixed with water to produce arelatively thin slurry (i.e., a slurry with a relatively high water content), and it isthen carefully deposited inside the specimen container.Slurry mixing containerElectric slurry mixerFine-grained soil slurryFigure 3.31: Plastic container of 100 L capacity and the electric mixer used in thepreparation of the slurry.The mass of soil and water needed to produce the slurry are calculated based on thespecimen slurry volume required. Typically, a 2.0 to 5.0 cm thick specimen is sufficient943.3. Details of Specimen Preparationfor interface shear testing. This translates to a slurry volume of approximately 0.03 to0.05 m3. A short vertical drainage path reduces the consolidation and drainage times,and hence, helps to reduce the overall test duration. In a plastic mixing container(see Fig. 3.31) of approximately 100 litre capacity, the fine-grained soil is mixed withwater at the previously calculated mixing ratio. An electric mixer is used for mixingthe material until a homogeneous slurry is produced. Special care is taken to ensurethat a fully saturated slurry is achieved. The slurry is then gradually poured into thespecimen container consistently in small steps using a container of approximately 1litre capacity (see Fig. 3.32). Pouring is carried out from a height of approximately0.2 m above the test surface. Special care is taken to minimize any turbulence in theflow that could result in the entrapment of air pockets within the specimen.Polyethylene sheetingFine-grained soil slurryFigure 3.32: Photograph showing pouring of fine-grained soil slurry into the specimencontainer.The thinner the slurry, the easier it is to remove any air from the slurry. However,an excessive amount of water can lead to segregation. This is especially the case for953.3. Details of Specimen Preparationsoils containing coarse particles (for example, low to medium plastic silts), where seg-regation of the coarser particles can result in a non-uniform specimen. Therefore, theamount of water used to produce the slurry is dependent on the soil type. The slurrydeposition characteristics can be determined by producing slurry samples of differentwater contents and allowing them to settle in separate metal containers. Once settled,the specimens are dried in the oven and are tested for grain-size distribution acrossthe height of the specimen using sieve analysis and hydrometer testing. Especially,the grain-size distribution of the bottom part of each specimen that is in contactwith the base of the container is of interest. Sieve analysis alone is often sufficient asthe goal is to detect segregation of coarse particles to the bottom of the specimen.The bottom part of each specimen should show approximately the same grain-sizedistribution as the original sample to ensure that segregation has not occurred. Thewater content above which segregation occurs is thus determined. This method isrelatively time consuming, and often, a qualitative approach is used to identify theapproximate segregation water content. Generally, a slurry that is sufficiently thinto ensure that mixing and pouring of the slurry can be done without trouble can beproduced by mere observation of the consistency of the slurry during mixing withwater.Once the desired consistency is obtained, the slurry is sampled for water content.New slurry batches for subsequent interface shear tests are prepared at this watercontent. Note that this qualitative approach is not suitable for certain soils. Thesegregation water content of low plastic silts is very difficult to obtain by mere ob-servation of the consistency. Hence, for such soils, the quantitative approach can beused to determine the segregation water content limit. Once the segregation watercontent limit is determined for each soil type, the preparation of the slurry is doneby mixing the soil with water to produce a target water content that is slightly lessthan the segregation limit. The mass of soil and water used is carefully controlled fora given test series, and the prepared slurry is always sampled for water content priorto placement in the device as a quality assurance measure.The pore-water pressure sensor matrix is used to continuously measure the pore-963.3. Details of Specimen PreparationPolyethylene sheetingFine-grained soil slurry (approx. 5 cm thick)NorthFigure 3.33: Photograph showing an as-placed fine-grained soil slurry layer inside thespecimen container.water pressure at the test surface during placement of the slurry. Once the slurry isplaced, the as-placed thickness of the slurry layer is determined by manual height mea-surements. The slurry is then allowed to settle under its own weight. The self-weightsettlement time can vary widely dependent on the type of soil used. For example, fora Fraser River silt slurry layer of 5.0 cm thickness, the self-weight settlement processtakes around 24 hours to complete. For a Kaolinite slurry of 5.0 cm thickness, theprocess can take around two to three days to complete.During the settlement process, the solid material displaces some of the water outof the soil matrix, and at the end of settlement, a layer of clear water can be observedabove the surface of the soil specimen as shown in Fig. 3.34. Once the self-weightsettlement of the slurry is complete, the pore-water pressure recorded from the pore-973.3. Details of Specimen PreparationSelf-weight settled soil specimenTest platePressure transducerClear water surfaceSurface of settled soil specimenThickness of soil specimenHeight to free water surface from test plate surface(Not to scale)PRODUCED BY AN AUTODESK EDUCATIONAL PRODUCTPRODUCED BY AN AUTODESK EDUCATIONAL PRODUCTP RODUCE D BY  AN AUT ODE S K E DUCAT I ONAL  P RODUCTPRODUCED BY AN AUTODESK EDUCATIONAL PRODUCTFigure 3.34: Diagram showing the self-weight settled soil specimen and the clearwater surface formed during settlement.water pressure sensor matrix should be equal to this static water head. This servesas a way to check the proper functioning of the pressure transducers prior to theapplication of the surcharge load.The soil specimen is next covered with the load transfer plate in preparation forapplying the surcharge load. The load transfer plate is subsequently covered witha layer of geotextile, and the desired surcharge load is applied (as described in thesection 3.2.3). The fine-grained soil specimen is then allowed to consolidate under thesurcharge load until the excess pore-water pressure generated during the applicationof the surcharge load is adequately dissipated.Typically, given the low hydraulic conductivity of fine-gained soils, the applicationof the surcharge load is immediately reflected as an equal increase in the pore-waterpressure measured by the pore-water pressure transducer matrix at the soil-solidinterface as shown in Fig. 3.35. It is possible to use the consolidation curve (pore-water pressure dissipation curve) obtained during this process for estimating thecoefficient of consolidation of the test-soil. As discussed earlier, for normal stresses in983.4. Selection of Shear Displacement RateStress or pressure (kPa) and Force (kN)Time (hour)18 20 22 24 26 280246810 FBLC (kN)PP11 (kPa)PP12 (kPa)PP13 (kPa)PP21 (kPa)PP22 (kPa)PP23 (kPa)PP31 (kPa)PP32 (kPa)PP33 (kPa)PP41 (kPa)PP42 (kPa)PP43 (kPa)Average PP (kPa)Effective normal stress σn' (kPa)PSWH.Awc (kN)Total normal stress σn (kN)PSHW (kPa)FBLCσn σn'PSWH.AwcAverage pore-water pressurePSHWPore-water pressure rise due to placement of sand surcharge.Figure 3.35: Typical pore-water pressure dissipation profile during consolidation offine-grained soil specimens under surcharge loading measured on the macro-scale in-terface direct shear apparatus.excess of 3.0 kPa, the pneumatic loading system is used.3.4 Selection of Shear Displacement RateWhen conducting interface direct shear tests with saturated fine-grained soils, asexpected, the rate of shear displacement was found to have a pronounced effect on theexcess pore-water generation during shearing. The macro-scale interface direct shearapparatus provides single drainage through the top of the test soil specimen, and henceprovides the means to conduct fully-drained or close to fully-drained tests dependingon the shear displacement rates used. However, because small strain behaviour isnot of interest in this study, it was considered that it is not critical to have zeroexcess pore-water pressure generation at small strains. To obtain the drained large-displacement interface friction angle, it is necessary to have a state of zero excess pore-water pressure at the interface when large shear displacements are achieved. Hence, asuitable shear displacement rate that leads to a zero excess pore-water pressure stateis reached at that large shear displacement condition needs to be established for eachtest soil.993.4. Selection of Shear Displacement RateThe excess pore-water pressure dissipation characteristics obtained from the pore-water pressure sensor matrix provide data that can be used to calculate the rate ofshear displacement that would promote minimal excess pore-water pressure genera-tion during large shear displacements. In the macro-scale interface direct shear appa-ratus, the test soil specimen is constrained within the four side-walls of the specimencontainer, and is also constrained at the bottom by the test-surface. Free drainage ofexcess pore-water is primarily possible across the top surface of the test soil specimen.These drainage boundary conditions are a close match to Terzaghi’s one-dimensionalconsolidation theory. Using the concepts of one-dimensional consolidation with singledrainage on the excess pore-water pressure dissipation characteristics the coefficientof consolidation cv of the fine-grained soil specimen can be estimated.Determination of the appropriate rate of shear displacement requires an estimateof the time required for the excess pore-water pressure dissipation and amount ofshear deformation required to reach failure as presented by Gibson and Henkel (1954).These two factors depend on the type of soil material and the stress history. Thetheoretical equation proposed by Gibson and Henkel (1954), as given by Eq. 3.4,together with guidelines specified in ASTM-D3080-98 (1998) can then be used tocalculate the time to failure, tf , that would correspond to a given degree of excess-pore pressure dissipation at failure.tf =H2sηcv(1− Uf ) (3.4)where, Hs = thickness of test-soil specimen = length drainage path,cv = coefficient of one-dimensional consolidation of the test-soil specimen,Uf = desired degree of pore-water pressure dissipation at failure,η = numerical factor corresponding to the extent and location of the drainage bound-aries.A value of η equal to 2 is recommended by (Gibson and Henkel, 1954) to representthe prevalent drainage conditions in a typical direct shear test with one sided drainage.1003.5. Interpretation of DataFor macro-scale interface direct shear testing, the time to failure is selected as thetime required to achieve a shear displacement of 400 mm (i.e., approximately halfthe total shear displacement of a typical test); it is considered that this magnitude ofshear displacement is sufficient to mobilize the large-displacement interface frictionangle. Substituting this rationalized value in Eq. 3.4, the value of tf corresponding toan excess pore-water pressure equalization of 85% can be calculated. This, in turn, isused to determine the required displacement rate by dividing the shear displacementat failure by tf . Considering the extensive time duration required for the testing offine-grained soils and that the excess pore-water pressures in the immediate vicinityof the shear zone are monitored during shear displacement, the use of a target pore-water pressure equalization value of 85% was considered fair and reasonable.3.5 Interpretation of DataThe primary measurements obtained from the macro-scale interface direct shear ap-paratus are: (i) the average interface shear force Fi that is applied to the mobile frameduring displacement controlled shearing; (ii) the corresponding shear displacement ofthe mobile frame; (iii) the total normal force applied to the base plate FBLC ; (iv)the static water head of the pool of water inside the water tank as measured by thepressure transducer PTSH PPTSH ; and (v) the average pore-water pressure at theinterface u. These measurements are used to determine the average interface shearstress τi, and the average effective normal stress at the interface σ′n, which allow forthe calculation of the effective angle of interface friction δ. While the mobile frameis displaced at a constant rate, a pulling force (Fpull) is measured by the main loadcell. This pulling force is comprised of: (i) the soil-solid interface shear force Fi,and (ii) the inherent friction of the moving parts of the mobile frame and the frictioncaused by the wipers of the specimen container rubbing against the test surface (here-after, collectively referred to as the device friction (Fdf )). The device friction (Fdf ) istypically measured using the main load cell at the end of conducting a given inter-face shear test by first removing the test soil specimen from the specimen container1013.5. Interpretation of Dataand then applying the same rate of displacement to the mobile frame without a soilspecimen in place. It was found that the inherent device friction was approximately5-10% (or less) of the pulling force measured during a typical macro-scale interfacedirect shear test. The pulling force is presented in Eq. 3.5 using the two constituentforces. Rearranging the terms in Eq. 3.5, the average interface shear stress τi can becalculated as shown in Eq. 3.6.Fpull = τtotalAi = Fdf + τiAi (3.5)τi =Fpull − FdfAi(3.6)As discussed in subsection 3.2.3, the base load cells and PTSH are used to de-termine the total normal stress σn acting on the interface. The average pore-waterpressure at the interface obtained from the pore-water pressure matrix is then usedto determine the effective normal stress σ′n. The total normal stress acting on thetest surface is given by Eq. 3.8 as presented in subsection 3.2.3. The effective normalstress is calculated using Eq. 3.9.Awc = Aw − Ac (3.7)σn =FBLC − PSWH .(Aw − Ac)Ai=FBLC − PSWH .AwcAi(3.8)σ′n = σn − u (3.9)where, σn = average total normal stress acting on the interface,FBLC = total force acting on the base plate as measured by the base load cells,PSWH = static water pressure of the water contained in the water tank as measuredby the pressure transducer (PTSH),Aw, Ac, Ai are planar footprint areas as defined in Fig. 3.23,σ′n = average effective normal stress at the interface, and1023.6. Sources of error and corrective measuresu = average pore-water pressure at the interface.The effective interface friction angle (δ), is then calculated as per Eq. 3.10. Asdiscussed earlier, the pore-water pressure varies continuously as the soil specimen issheared against the test surface during the initial stage of shearing. This excess pore-water pressure slowly dissipates to the static water head at approximately 30-50%of the total shear displacement. During the remaining shear displacement, the pore-water pressure typically remains constant at the static water head level. The large-displacement interface friction angle is obtained at these large shear displacements.δ = arctan(τiσ′n)(3.10)3.6 Sources of error and corrective measuresThe macro-scale interface direct shear apparatus was custom designed and built toproduce high quality large-displacement interface shear strength data for soil-solidinterfaces at relatively low effective normal stresses. The quality and reliability ofthe test data are dependent on the performance of the measurement system, and thequality control measures implemented to reduce experimental variability. The dataacquisition system of the new apparatus was custom made at UBC to produce highresolution data using a 24-bit sigma-delta type analog-to-digital converter (ADC);and the sensors were selected considering the essential resolution, range and linearcalibration curve requirements. As outlined earlier, after the device was developed,a series of commissioning tests were carried out to investigate the repeatability ofthe experimental protocols and to ascertain the degree of uncertainty or variabilityexpected in the test data. Based on the results of the commissioning tests, importantsources of error and variability were identified, and suitable quality control and qual-ity assurance measures have been devised (which include the specimen preparationprocedures described earlier) to minimize the identified errors.This section presents further details of the investigations that were carried out toascertain the important sources of error and variability that can affect the test data1033.6. Sources of error and corrective measuresand the steps that have been taken to minimize these errors. The primary test mea-surements of the device are: (i) the soil-solid interface shear force (for interface shearstress τi determination); (ii) the total normal force acting at the soil-solid interface(for total normal stress σn determination); and (iii) the pore-water pressure at thesoil-solid interface (for determination of the effective normal stress σ′n). The reliabil-ity, repeatability and accuracy of the test data generated by the device are dependenton: (i) the performance of the sensors and data acquisition system; and (ii) the qual-ity control measures taken to reduce variability in the experiments. The followingsubsections discuss the sources of error that are expected in each of these measure-ments, and the quality control measures that have been implemented to minimizethese errors.3.6.1 Interface shear stressThe calculated interface shear stress τi of the macro-scale interface direct shear ap-paratus can be affected by three sources of error: (i) the performance of the mainload cell and data acquisition system; (ii) the friction and other forces correspondingto the mechanical parts of the device; and (iii) variability in the experimentationprocedures and the soil specimen preparation procedures.The measurement resolution of the main load cell with little to no electrical noiseeffects is approximately 0.002 kN. During interface direct shear testing, the dataacquisition setup together with the wiring system used in the device is prone topicking up high frequency electrical noise that is primarily due to the electromagneticinterference caused by the stepper motor coils of the linear actuator. Taking thiselectrical noise into consideration, the raw unfiltered resolution of the main load cellmeasurement drops to 0.03 kN. The electrical noise that is due to the stepper motorcan be filtered out (either during testing, or after testing is complete) by processingthe collected data through a low-pass filter. Filtering the electrical noise improvesthe resolution of the main load cell measurement to approximately 0.004 kN. Thistranslates to an error of approximately ±0.004 kPa in the calculated interface shearstress data.1043.6. Sources of error and corrective measuresIt is of relevance to point out that fluctuations in the ambient temperature can alsoaffect the shear force data, and accounting for temperature effects is important, espe-cially when conducting tests at relatively slow displacement rates where the durationof a given test can span several weeks. Considering the importance of producing highquality test data, the main load cell, as well as all the critical electronic componentsof the data acquisition system were carefully selected to have a very low temperaturecoefficient. By running the data acquisition system continuously over the course ofapproximately 48 hours, and monitoring the main load cell reading with ambient tem-perature, the effect of the change in the ambient temperature in the laboratory on themain load cell readings was determined to be negligible. Figure 3.36 shows the timehistories of the raw unfiltered main load cell reading and the ambient temperature inthe laboratory recorded during the aforementioned test.Main load cell reading (kN)Time (h)5 10 15 20 25 30 35 40-0.00200.0020.004Temperature (°C)2222.52323.524Figure 3.36: Variation of the raw unfiltered main load cell reading with the ambienttemperature in the laboratory recorded continuously over the course of two days.The second source of error and variability in the measured shear force of the ap-paratus is the effect of friction and other forces associated with the mechanical com-ponents of the device. Primarily, the device friction force is due to: (i) the frictionalforce produced by the set of horizontal linear bearings that support the weight of themobile frame; and (ii) the frictional force due to the geotextile-lined rubber wipers1053.6. Sources of error and corrective measuresof the specimen container that are in contact with the solid surface to be tested. Inorder to minimize the device friction force Fdf , precision low-friction horizontal linearbearings were selected; and as described in detail earlier, the specimen container wassupported on the mobile frame to ensure that the weight of the specimen containerwas not transferred to the solid surface to be tested (which also allowed for setting theclearance between the specimen container and the solid test surface to be adjusted asrequired). Based on device friction measurements, taken without the wipers installed,it was found that the inherent device friction (due to the linear bearing friction alone)was approximately 0.008 kN.Device friction force (kN)Horizontal displacement of mobile frame (cm)0 20 40 60 8000.020.040.060.080.10.120.14Figure 3.37: Device friction force of the macro-scale interface direct shear apparatusmeasured over a horizontal displacement range of 80 cm (Specimen container wasequipped with geotextile-lined wipers, and the test was conducted over an epoxycoated steel surface).Shown in Fig. 3.37 is the device friction force that was measured with the wipersinstalled, and tested on top of a smooth clean epoxy coated solid surface, over a hor-izontal displacement range of 80 cm, at a displacement rate of 0.004 mm/s. In thiscase, the device friction was found to vary between 0.06 kN to 0.08 kN. During inter-face shear testing, the soil specimen is able to penetrate into the clearance betweenthe specimen container and the solid test surface, and this can introduce a further in-crease of the device friction force. The magnitude of the device friction that can arisedue to the intrusion of soil into the clearance between the specimen container and thesolid test surface varies depending on the type of soil and solid surface roughness, andit is extremely important that a correction is made to the test data to account for thisfriction force. At the end of a given interface shear test, immediately after removing1063.6. Sources of error and corrective measuresthe soil specimen from the specimen container and before cleaning the specimen con-tainer and test plate, the device friction force needs to be measured again. In thiscase, the device friction force measured, will include the effect of the entrapped soilwithin the clearance between the specimen container and the solid test surface. Thepost-test device friction, as measured from commissioning tests using coarse-grainedand fine-grained soils on different solid surfaces, ranged from 0.15 kN (for relativelysmooth solid test surfaces) to 0.50 kN (on a relatively rough solid surface). The devicefriction force measured at the end of testing a Kaolinite specimen on the green epoxytest surface is illustrated in Fig. 3.38. When testing at low confining stress levels,this force can introduce a notable error to the measured interface shear stress if notaccounted for. For example, when testing at a confining stress level of 3 kPa, the thistranslates to an error of approximately 33% of the measured shear force. Therefore,special quality control measures were put in place to ensure that the device frictionforce Fdf , with the wipers installed, is measured at the end of each macro-scale in-terface direct shear test. The device friction force is then deducted from the shearforce data when calculating the interface shear stress τi. With the above mentionedstringent quality control measures taken to minimize errors and variability, the inter-face shear force measurements of the macro-scale interface direct shear apparatus isreliable to within ±0.004 kN. At a confining stress level of 3 kPa, this translates toan error of approximately 0.3% of the measured interface shear force.Shear stress (kPa)Shear displacement (cm)50 55 60 65 7000.511.52Kao_Green_3Reduction of pulling force with the removal of the soil specimen at end of testing.Device friction force measured after removing the soil specimen at end of testing.Figure 3.38: Device friction force measured at the end of testing a Kaolinite specimenon the green epoxy test surface.1073.6. Sources of error and corrective measures3.6.2 Total normal stressThe base load cell matrix is comprised of 12 load cells, each rated at 2000 kg capacity.The high frequency electrical noise due to the electromagnetic interference causedby the stepper motor coils of the linear actuator is picked up by the measurementsystem. Taking this electrical noise into consideration, the raw unfiltered resolution ofthe main load cell measurement is 0.2 kN. The noise that is due to the stepper motoris filtered out from the base load cell readings to improve the resolution. Filteringthe electrical noise improves the resolution of the base load cell measurement toapproximately 0.01 kN.As described earlier, the total vertical force measured by the matrix of base loadcells is comprised of: (i) the total normal force applied by the test soil specimen onthe solid test surface; and (ii) the vertical force produced by the static water baththat surrounds the specimen container to keep the test soil specimen saturated dur-ing testing. Therefore, the total normal tress σn is calculated by subtracting thevertical force produced by the surrounding water bath from the total vertical forcemeasurement of the base load cell matrix as given by Eq. 3.8. In order to determinethe vertical force produced by the surrounding water bath, the water pressure read-ing of the pressure transducer PTSH is multiplied by the total planar area of thesurrounding water footprint Awc (i.e., the planar footprint of the water tank minusthe exterior footprint of the specimen container). The planar area Awc is calculatedusing the length and width measurements of the footprints of the water tank and thespecimen container. The area Awc has a measurement error of ±0.023 m2 calculatedtaking into account the measurement error of the length measurements which weredone using a standard tape measure. The pressure transducer has a measurementresolution of 0.004 kPa. The error in the calculated planar area coupled with the errorof the pressure sensor measurement produces a total measurement error of ±0.05 kNin the total vertical force produced by the surrounding water head. Therefore, thetotal normal stress acting on the soil-solid interface has a total error of ±0.05 kPa.1083.6. Sources of error and corrective measures3.6.3 Pore-water pressure and effective normal stressThe pore-water pressure sensor matrix is comprised of 12 pore-water pressure trans-ducers, each rated at a measurement range of 75 kPa. The sensor output is internallytemperature compensated and amplified to produce a stable 0 - 5 V output. Thepore-water pressure sensor measurement error is limited to ±0.004 kPa. Combiningthe errors arising from the total normal stress and pore-water pressure measurements,the total error in the effective normal stress is estimated to be ±0.05 kPa.The inclusion of porous stone elements to cover the pore-water pressure trans-ducer openings can affect the pore-water pressure sensor readings during the interfaceshearing process. Localized non-uniform failure mechanisms of the soil specimen inthe vicinity of the porous stone can occur if the porous stone elements are not madeflush and level with the solid test surface. Also, the change in surface roughness thatthe soil specimen experiences as it moves from the solid test surface to the surfaceof the porous stone can cause local non-uniform pore-water pressure measurements.In order to minimize such effects, quality control measures were put in place to en-sure that the porous stones were installed to be flush and level with the solid testsurface. The performance of the pore-water pressure measurement system was testedduring commissioning tests by observing the pore-water pressure readings as a sat-urated fine-grained soil specimen was sheared over a solid test surface. Ideally, atlarge shear displacements, when the pore-water pressure has dissipated to the staticwater head level, there should be no further change in pore-water pressure at thesoil-solid interface with continued shear displacement. In fact, pore-water pressuresensor readings observed during commissioning tests confirmed that the pore-waterpressure measurements showed little to no variation at large shear displacements afterreaching the static water head level.Figure 3.39 shows the pore-water pressure sensor data as recorded by the 12pore-water pressure transducers during a macro-scale interface direct shear test thatwas conducted using a reconstituted, saturated specimen of Kaolinite clay normallyconsolidated under σn of 4 kPa on top of a sand-blasted mild steel surface and sheared1093.6. Sources of error and corrective measuresat a horizontal displacement rate of 0.004 mm/s. The data in Fig. 3.39 shows that atthe onset of shear displacement, there is excess pore-water pressure generated at thesoil-solid interface, which dissipates to the static water head level of approximately0.67 kPa at approximately 20 cm of shear displacement, and remains constant for theremaining 40 cm of shear displacement.Pore-water pressure (kPa)Shear displacement (cm)0 10 20 30 40 50 600123456PP11PP12PP13PP21PP22PP23PP31PP32PP33PP41PP42PP43Figure 3.39: Variation of pore-water pressure at the soil-solid interface measuredduring a macro-scale interface direct shear test conducted using reconstituted fully-saturated Kaolinite clay on a sand-blasted mild steel surface.Based on the above mentioned quality control measures in place, and with the useof the precision base load cell and the pore-water pressure sensor matrix measure-ments, the effective normal stress estimation that is made in the macro-scale interfacedirect shear apparatus is accurate to within ±0.05 kPa and this is considered suffi-cient for studying the interface shear strength parameters under the effective normalstress range of 3.0 to 30.0 kPa.3.6.4 Quality control measures taken to minimize soilspecimen variabilityThe variability in the specimen preparation procedure, and the specimen constitutioncan affect the results of the interface direct shear tests. It is therefore important toadhere to a strict specimen preparation procedure that would minimize variation inspecimen uniformity, thickness, and initial conditions. The specimen preparationprocedures discussed earlier were followed precisely when preparing soil specimens,and parameters such as initial and final (i.e., during soil specimen slurry preparation,1103.7. Summary of the Chapter and Contributionsand at the end of interface shear testing) moisture contents of the soil specimen, theinitial and final thickness of the soil specimen were monitored throughout the testingprogram as a quality assurance process.(a)NorthSampling cups(b) (c)Figure 3.40: Photographs showing the sampling of a fine-grained soil specimen at theend of macro-scale interface direct shear testing for determination of final averagemoisture content of the soil specimen.Figure 3.40 shows photographs taken during sampling of a fine-grained soil speci-men at the end of a macro-scale interface direct shear test (the top surface of the soilspecimen is exposed after removing the surcharge load sand layer). The sampling wascarried out by pushing the sampling cup vertically downward into the soil specimen,and then carefully removing the soil surrounding the cup, and using a spatula to sliceacross the sampling cup opening to extract the cup with the sample. A small ventingorifice was made in the sampling cup to allow air to escape as the cup was pushedinto the soil specimen. The sampling cup was pushed into the soil specimen firmly,until a small amount of the soil specimen extruded out through the orifice. Thesesamples were used to determine the final moisture content of the soil specimen. Fora given test soil type, in order to minimize potential variability of the soil specimenconstitution, the same soil specimen was reconstituted and reused from one test tothe next throughout the test program.3.7 Summary of the Chapter and ContributionsAs identified in the literature review, there are major challenges in using conventionalgeotechnical laboratory test apparatus to characterize the large-displacement soil-1113.7. Summary of the Chapter and Contributionssolid interface shear strength under relatively low confining stress levels applicable topipeline design (within σ′n of 3 to 30 kPa). The present chapter presented details ofthe development of a novel macro-scale interface direct shear apparatus that providesthe following improvements over conventional devices:1. The macro-scale interface direct shear apparatus was developed to have a 1.0 m× 1.0 m interface area to aid in minimizing the device compliance effects suchas that due to device friction, and in maximizing the accuracy in the measuredinterface shear force when testing at low confining stresses applicable to pipelinedesign (σ′n within 3 to 30 kPa). The error introduced by the device friction whenconducting tests at low confining stresses is approximately 0.3% of the measuredshear force. In contrast, this error would be in the order of approximately 100%or more in a conventional interface shear test apparatus when used for testinginterfaces at low confining stresses.2. The new apparatus allows for determining the effective normal stress at thesoil-solid interface during shearing with a resolution of ±0.05 kPa within thetest normal stress range of 3 to 30 kPa. This is achieved by using an array ofbase load cells that are used to estimate the total normal force acting on theinterface, and by using an array of pore-water pressure sensors to measure thepore-water pressure at the soil-solid interface.3. The device is equipped with side-wall load cells to allow for the measurementof side-wall friction force that arises due to the shearing of the soil specimenvertically against side-walls of the specimen container.4. A total horizontal interface shear displacement of 1.0 m is achievable usingthe new apparatus at displacement rates as low as 0.1 µm to aid in obtainingthe drained large-displacement interface friction angle. This displacement is ap-proximately twice the horizontal length of the soil specimen, and approximately20 times the thickness of the soil specimen.In summary, the new apparatus can be treated as a new tool that provides the1123.7. Summary of the Chapter and Contributionsmeans of characterizing soil-solid interfaces for application in pipeline design wheresoil-solid interface testing is required at unconventionally challenging testing condi-tions.113Chapter 4Experimental Program and ResultsThis chapter presents the details of the experiments that were conducted using themacro-scale interface direct shear apparatus to study the friction angle of soil-solidinterfaces at relatively low effective normal stresses and large-displacements. As out-lined at the beginning of this thesis, the primary aim of the laboratory experimentswas to characterize the large-displacement soil-solid interface shear strength of a num-ber of soils against solid surfaces of varying roughness under relatively low effectivenormal stresses. Particularly, it was of interest to study the effects of soil-type (i.e.,coarse-grained versus fine-grained), soil plasticity, surface texture of the solid surface,and the effective normal stress on the large-displacement interface friction angle. Thefollowing sections provide details of the materials used and the experiments carriedout using the macro-scale interface direct shear apparatus. The test data arising fromthe experiments are also presented.4.1 Materials TestedThree soil materials, namely: (i) Fraser River sand; (ii) low-plastic Fraser River silt;and (iii) medium-plastic Kaolinite clay were selected for the research with an empha-sis to cover the spectrum of coarse-grained to fine-grained soils that can be expectedunder real-life conditions. These soils were tested against three solid test surfaces,namely: (i) a sand-blasted mild steel surface; (ii) an abrasion resistant 100%-solidsepoxy surface (green epoxy); and (iii) an abrasion resistant three-coat epoxy sur-face (grey epoxy) having a slightly higher roughness compared to the green epoxy.In addition to the above mentioned soils, a limited number of interface direct sheartests were conducted using three additional soil materials, namely: (i) medium plas-1144.1. Materials Testedtic Strong-Pit FGS; (ii) medium-plastic Redstone clay; and (iii) low-plastic Kaosandclay on some of the above mentioned solid test surfaces targeting relatively low effec-tive normal stresses. Details of the test materials used and the tests conducted areprovided in the following sections.4.1.1 Soil materialsFraser River sandFraser River sand has been extensively used in element testing as well as soil-pipeinteraction testing conducted at UBC over the past 10 years. This material wasobtained through a local supplier in bulk form; and the sand has been dredged fromthe Fraser River in the Lower Mainland of British Columbia, Canada. Fraser Riversand is composed of 40% quartz, quartzite, and chert, 11% feldspar, and 45% unstablerock fragments, and 4% miscellaneous detritus (Garrison et al., 1969). The sand grainsare generally angular to sub-rounded in shape. The sand has an average particle sizeD50 = 0.26 mm, D10 = 0.17 mm, specific gravity Gs = 2.71, and uniformity coefficientCu = 1.6. The maximum and minimum void ratios (emax and emin) for the sanddetermined as per American Society for Testing and Materials Standards ASTM-D4254 (2006) and ASTM-D4253-00 (2006) are 0.94 and 0.62, respectively. Basedon data gathered from past studies conducted at UBC, the constant volume frictionangle φcv of the sand is 33° (Sivathayalan, 2000; Uthayakumar and Vaid, 1998). Thegrain size distribution of the sand together with the fine-grained soils is presented inFig. 4.1.Fraser River siltFraser River silt, a natural non-plastic fine-grained soil, was retrieved from a sitelocated on the south bank of the Fraser River adjacent to the Port Mann bridge inBritish Columbia, Canada. The material was excavated using a commercial excavatorand transported to UBC in 1000-kg bags. This material had a significant organicscontent (mostly pieces of wood and roots) that had to be removed by wet-sieving the1154.1. Materials TestedPercentage finer (%)Particle size (mm)10-310-210-1100020406080100 Fraser River sandStrong-Pit FGSKaoliniteKaosandRedstoneFraser River siltKaoliniteLL = 48%PL =26%PI = 22%Strong-Pit FGSLL = 34%PL = 17%PI = 17%Fraser River siltLL = 21%PL = 18%PI = 3%Fraser River sandKaosandLL = 22%PL = 18%PI = 4%RedstoneLL = 34%PL = 20%PI = 14%Figure 4.1: Grain size distribution of soils tested.material before it could be used for testing. The grain size distribution of the materialis shown in Fig. 4.1. Index tests showed a specific gravity Gs of 2.76. A Liquid limit(LL) of 21% and a plastic limit (PL) of 18% were observed. Ring-shear testing of thesilt at effective normal stresses of 3 and 6 kPa showed a drained friction angle of 33°at relatively large displacements (Eid et al., 2014). The chemical constituents of thematerial were determined based on X-ray diffraction testing carried out at the UBCDept. of Earth, Ocean and Atmospheric Sciences, and are listed in Table 4.1.Figure 4.2: Photographs showing the processing of the silt samples by wet sieving toremove debree.1164.1. Materials TestedTable 4.1: Composition of the Fraser River silt test-soil (based on X-ray diffractiontesting carried out at the UBC Dept. of Earth, Ocean and Atmospheric Sciences).Constituent (%)Actinolite (Ca2(Mg,Fe2+)5Si8O22(OH)2) 6.9Clinochlore ((Mg,Fe2+)5Al(Si3Al)O10(OH)8) 5.7Dolomite-Ankerite (CaMg(CO3)2-Ca(Fe2+,Mg,Mn)(CO3)2) 1.0Illite-Muscovite 2M1 (K0.65Al2.0Al0.65Si3.35O10(OH)2/KAl2AlSi3O10) 3.6K-feldspar (orthoclase) (KAlSi3O8) 6.5Magnetite (Fe3O4) 1.2Plagioclase (albite, albite calcian) (NaClSi3O8-CaAlSi2O8) 39.5Quartz (SiO2) 35.8Kaolinite clayA commercially available Kaolinite clay was selected as a fine-grained soil to repre-sent the clay-size particles. This material is yellowish-white in color. The chemicalconstituents of the material, as per the material datasheet supplied by the vendor,are listed in Table 4.2. Index tests indicated a Gs of 2.66; liquid and plastic limits of48% and 26% respectively, resulting in a PI of 22%. Ring-shear testing of the clayat effective normal stresses of 3, 6, 10, and 25 kPa showed drained friction angles of25.2°, 25.0°, 24.9° and 24.5° respectively (values were obtained at large displacements)(Eid et al., 2014).Figure 4.3: Photogragh (left) and scanning electron images (centre, right) of dryKaolinite (SEM images were obtained using a Phillips XL30 microscope at the UBCDepartment of Earth, Ocean and Atmospheric Sciences).1174.1. Materials TestedTable 4.2: Composition of the Kaolinite clay (based on information provided by thevendor)Constituent (%)Silicon dioxide (SiO2) 45.9Aluminium oxide (Al2O3) 38.7Titanium dioxide (TiO2) 1.70Iron oxide (Fe2O3) 0.30Calcium oxide (CaO) 0.03Magnesium oxide (MgO) 0.10Potassium oxide (K2O) 0.10Sodium oxide (Na2O) 0.08Loss on ignition 13.7Strong-Pit FGSA naturally occurring low-plastic fine-grained soil, grey in color, was retrieved froma deposit located at an open-pit mine (Strong Pit mine) located on King road inAbbotsford, British Columbia, Canada. The material shows a well-graded grain-sizedistribution with Gs = 2.75, LL = 34% and PL = 17%. The Strong-Pit FGS has afines content (smaller than 2 µm particle size) of 25%, which is approximately equalto that found in the Kaolinite clay.Figure 4.4: Photographs showing the fine-grained soil deposit at the Strong Pit minein Abbotsford, BC.1184.1. Materials TestedRedstoneThe Redstone soil is a low-plastic fine-grained soil, dark red in color, that has beenquarried and processed by a company based in Medicine Hat, Alberta, Canada, specif-ically for use in the pottery industry. This material was purchased in bulk form for thepresent research work, and the material was supplied in dry powder form. Redstoneshows a well-graded grain-size distribution closely matching that of the low plasticStrong-Pit FGS presented earlier. Based on index tests carried out at UBC, the spe-cific gravity of the soil was determined to be Gs = 2.75, and a plasticity index of PI= 14% (LL = 34% and PL = 17%) was ascertained. Ring-shear testing of the clayat effective normal stresses of 10, and 25 kPa showed drained friction angles of 26.3°and 24.8° respectively (values were obtained at large displacements). The chemicalconstituents of the material, as per the material datasheet supplied by the vendor,are listed in Table 4.3.Figure 4.5: Photograph of Redstone slurry (left) and scanning electron microscopyimages of dry Redstone powder (centre and right) (SEM images were obtained using aPhillips XL30 microscope at the UBC Department of Earth, Ocean and AtmosphericSciences).KaosandThe Kaosand soil is a low-plastic fine-grained soil, light grey in color, that has beenquarried and processed by a company based in Medicine Hat, Alberta, Canada, specif-ically for use in the pottery industry. This material was purchased in bulk form forthe present research work, and the material was supplied in dry powder form. Based1194.1. Materials TestedTable 4.3: Composition of the Redstone soil (based on information provided by thevendor).Constituent (%)Silicon dioxide (SiO2) 68.5Aluminium oxide (Al2O3) 14.6Iron oxide (Fe2O3) 5.1Potassium oxide (K2O) 2.6Magnesium oxide (MgO) 0.9Titanium dioxide (TiO2) 0.7Calcium oxide (CaO) 0.3Phosphorus pentoxide (P2O5) 0.3Sodium oxide (Na2O) 0.1Loss on ignition 6.5on index tests carried out at UBC, the specific gravity of the soil was determinedto be Gs = 2.75, and a plasticity index of PI = 4% (LL = 22% and PL = 18%)was ascertained. Ring-shear testing of the clay at effective normal stresses of 10, and25 kPa showed drained friction angles of 31.1° and 30.8° respectively (values wereobtained at large displacements).Figure 4.6: Scanning electron microscopy images of dry Kaosand powder (SEM imageswere obtained using a Phillips XL30 microscope at the UBC Department of Earth,Ocean and Atmospheric Sciences).The plasticity charts showing the index properties of the fine-grained soils is pre-sented in Fig. 4.7.The coefficient of consolidation cv of the fine-grained soils was determined using1204.1. Materials Tested+KaoliniteFraser River siltStrong-Pit FGSRedstoneKaosandFigure 4.7: Plasticity chart.the pore-water pressure dissipation data gathered during the consolidation stage of themacro-scale interface direct shear tests. As described in Chapter 3, the fine-grainedsoil specimens were prepared in slurry form, and were of relatively high void ratioat the onset of consolidation under the initial 3 kPa normal stress. In all the testsconducted on Kaolinite, cv was determined to be approximately 1.4 to 1.5 m2/yearat 3 kPa normal stress. At subsequent consolidation stages at higher normal stressesof 15 and 30 kPa, cv was found to be between 2.7 and 3.0 m2/year. Fraser River siltyielded a cv of approximately 2.5 m2/year at 3 kPa normal stress, 7.4 to 8.5 m2/yearat 15 kPa normal stress, and around 22.0 m2/year at 30 kPa normal stress. Table 4.4lists the important properties of the soils tested in this research program.1214.1. Materials TestedTable 4.4: Properties of fine-grained soilsSoil LL(%)Ip(%)D50(µm)CF(%)cv (m2/yr) φ′(deg.)*Fraser Riversand- - 260 - - 33Fraser River silt 21 3 24.8 9.3 2.5 - 22.0 33Kaolinite 48 22 2.8 29.4 1.5 - 3.0 24 - 25Strong-Pit FGS 34 17 10.1 26.0 0.12 -Redstone 34 14 8.0 25.0 1.49 25 - 26Kaosand 22 4 70.0 15.0 2.34 31* Note: For the fine-grained soil, φ′ was obtained using ring-shear testing at largedisplacement under drained conditions and relatively low effective normal stresses.For Fraser River sand, φcv is reported based on past studies conducted at UBC.4.1.2 Solid surfacesA hot-rolled mild-steel plate, of type CS-Type-A (A1011/A1011M12b, 2012), 6.4 mm(0.25 inch) thick, was sand blasted to produce a relatively rough surface texture. Itis important to note that the surface was susceptible to oxidation and it was foundto rapidly spread across the surface within 24 hours after leaving the surface exposedto moisture.Two epoxy coated mild-steel plates were prepared by applying commercially avail-able proprietary epoxy coating materials on mild-steel plates of type CS-Type-A(A1011/A1011M12b, 2012), 6.4 mm (0.25 inch) thick. While the common method-ology of application of epoxy coatings on pipes is through a process called ”FusionBonding”, where the pipe and the coating are heated to high temperatures to fuse thematerial to the pipe surface, this was not a feasible approach for coating the flat testplates that were used for the macro-scale interface direct shear testing. Instead, twoabrasion resistant liquid epoxy coatings, namely, a 100% solids green epoxy, and a1224.1. Materials Testedtwo-part grey epoxy, both typically used in industrial pipeline protection applications,were used to professionally coat the steel plates as per the suppliers specifications bya commercial organization specializing in liquid epoxy coatings. Several mild steelplates were also coated during this process for later use in measuring the coatingproperties. Each coupon was cut across to expose its cross section, and the thicknessof the coating was measured. On average, the coatings were 1.5 mm thick.The three solid surfaces used in this study (i.e., sand-blasted mild steel, greenepoxy, and grey epoxy) were characterized based on their surface texture and hard-ness. The surface texture was quantitatively assessed using two surface metrologytools: (i) an Olympus LEXT OLS3100 laser confocal microscope with built-in surfacemetrology functions; and (ii) Mitutoyo SJ210 type stylus-based mechanical profilome-ter (see Fig. 4.8).(a) (b)(c) (d) StylusSurface profileMitutoyo SJ210 profilometerLaser confocal microscopeGreen epoxy couponSand-blasted steel couponFigure 4.8: Photographs showing roughness measurement of samples of the solid testplates.The Olympus LEXT OLS3100 laser confocal microscope had a maximum mea-1234.1. Materials Testedsurement footprint of 0.25 mm × 0.25 mm, and provided a height resolution of 0.01µm. The laser confocal microscope also provided the means to gather a qualitativeassessment of the surface textures. The stylus profilometer had a height resolution of0.02 µm, and was equipped with a stylus of 2.0 µm tip radius. The size of the stylustip limits the spatial resolution to some extent compared to the capability of the laserconfocal microscope. However, the stylus device provides a much larger linear mea-surement range of up to 17.5 mm, which makes the device suitable for determiningthe macro-scale surface texture of a given surface.Selection of appropriate gauge lengths for evaluation of averageroughnessThe gauge length over which the roughness of the solid surface is evaluated wasselected to be a function of the mean particle size D50 of the test soil of interest.Specifically, the minimum gauge length Lg for the measurement of Ra was selectedas 50 times D50 of the test soil of interest. For example, in the case of Kaolinite, theminimum gauge length was calculated to be approximately 12.5 µm as given by Eq.4.1.Lg−kaolinite = 50×D50−kaolinite = 50× 0.25 µm = 12.5 µm (4.1)In the case of Kaolinite, due to the stylus tip radius of the stylus type profilometerbeing approximately similar in size to the D50 of Kaolinite, this device was determinedto be not suitable for evaluating the Ra of the solid surfaces, and instead, the laserconfocal microscope was used to evaluate Ra. The gauge lengths corresponding to50×D50 of Fraser River silt and Fraser River sand were approximately 1.25 mm and13.0 mm respectively, and in this case, the laser confocal microscope was determinedto be not suitable for measurement of Ra owing to the limited measurement footprint(0.25 µm × 0.25 µm). In this case, the stylus type profilometer was selected forevaluating the average roughness of the solid surfaces.Given the above considerations, the following measurement protocol was used for1244.1. Materials Testedcharacterizing the average roughness of the solid surfaces:(i) For macro-scale interface direct shear tests involving Kaolinite as the test soil,the Ra values of the solid surfaces were evaluated using the laser confocal mi-croscope using a gauge length of 12.5 µm;(ii) For macro-scale interface direct shear tests involving Fraser River silt or FraserRiver sand as the test soil, the Ra values of the solid surfaces were evaluatedusing the stylus type profilometer using a gauge length of 13.0 mm;Shown in Fig. 4.9 are the laser confocal microscopy images of the three solid testsurfaces taken using an objective lens of ×50 magnification and numerical apertureof 0.95 (values recommended for surface roughness characterization by the manufac-turer). Marked on each image is a line indicating approximately 50×D50 = 0.12 mmgauge length (D50 for Kaolinite is approximately 2.5 µm).0.25 mm x 0.25 mm0.12 mm 0.12 mm0.25 mm x 0.25 mm0.12 mm0.25 mm x 0.25 mmGreen epoxy Grey epoxy Sand-blasted mild steelFigure 4.9: Laser confocal microscopy images of the solid test surfaces.Shown in Figs. 4.10 and 4.11 are sample surface profiles measured along selectedgauge lengths. The surface profiles corresponding to Lg of 12.5 µm were obtainedusing the laser confocal microscope, and the surface profiles corresponding to Lg of13.0 mm were obtained using the stylus profilometer.1254.1. Materials TestedZ (um)X (mm)0.02 0.04 0.06 0.08 0.1 0.12-1-0.500.51(a) Green epoxyZ (um)X (mm)0.02 0.04 0.06 0.08 0.1 0.12-2-1012(b) Grey epoxyZ (um)X (mm)0.02 0.04 0.06 0.08 0.1 0.12-20020(c) Sand-blasted steelZ (um)X (mm)0.02 0.04 0.06 0.08 0.1 0.12-30-20-100102030Green epoxy Grey epoxy Sand-blasted steel(d) Three solid test surfaces comparedFigure 4.10: Sample surface profiles of the solid test surfaces measured along a gaugelength of 50D50 (D50 of Kaolinite has been selected).1264.1. Materials TestedZ (um)X (mm)2 4 6 8 10 12-505(a) Green epoxyZ (um)X (mm)2 4 6 8 10 12-15-10-5051015(b) Grey epoxyZ (um)X (mm)2 4 6 8 10 12-40-2002040(c) Sand-blasted steelZ (um)X (mm)2 4 6 8 10 12-40-2002040Green epoxy Grey epoxy Sand-blasted steel(d) Three solid test surfaces comparedFigure 4.11: Sample surface profiles of the solid test surfaces measured along a gaugelength of 50D50 (D50 of Fraser River sand has been selected).For each test surface, nine such surface profile measurements were taken using1274.1. Materials Testedthe to determine the average roughness Ra. Also, for each test surface-soil combina-tion, the surface profile measurements allowed for the determination of the roughnesssize distribution of the test surface, and these data were plotted on a geotechnicalgrain size distribution graph allowing for the convenient comparison of the grain-sizedistribution of a given soil material with the solid surface roughness size distribution.Percent finer (%)Grain size (mm), or Roughness size (mm) 10-610-510-410-310-210-1100020406080100Green epoxyGrey epoxySand-blasted steelKaoliniteRa_50D50 = 7 μm 0.5 μm 0.15 μm(a) Roughness size based on gauge length of 0.12 mm (Lg = 50×D50−Kaolinite)Percent finer (%)Grain size (mm), or Roughness size (mm) 10-610-510-410-310-210-1100020406080100 Green epoxyGrey epoxySand-blasted steelFraser River sandKaosandRedstoneStrong-Pit FGSFraser River siltRa_50D50 = 7 μm 1.8 μm 1.6 μm(b) Roughness size based on gauge length of 13.0 mm (Lg = 50×D50−FRS)Figure 4.12: Roughness size distribution of the solid test surfaces compared withgrain size distributions of test soils).In this case, the roughness size distribution curves (RSD curves) provided an1284.2. Tests Conducted and Observationseffective means of delineating between a relatively rough and a relatively smooth solidtest surface with respect to a selected soil material. The roughness size distributioncurves obtained from the present study are shown in Fig. 4.12.The Ra values of the solid surfaces, evaluated over the selected gauge lengthscorresponding to the 50×D50 values of the test soils, henceforth referred to as Ra 50D50are given in Table 4.5.Table 4.5: Surface roughness and hardness parameters of the solid test platesSurface Ra 50D50(±SD) (µm)Lg = 12.5 µmRa 50D50(±SD) (µm)Lg = 13.0 mmHardness(Shore-D)Green epoxy 0.15 (±0.05) 1.8 (±0.6) 67Grey epoxy 0.50 (±0.25) 1.6 (±0.4) 68Sand-blasted mild steel 7.00 (±3.60) 7.0 (±0.4) -4.2 Tests Conducted and ObservationsThe macro-scale interface direct shear tests conducted during this research were tai-lored to provide data for characterizing the drained large-displacement interface fric-tion angle of soil-solid interfaces, with an emphasis on understanding the effects of:(i) the magnitude of effective normal stress; (ii) the relative roughness of the interface;and (iii) the soil type.As such, these topics of interest were studied by testing the six soil types againstthe three different solid surfaces under effective normal stresses of 3, 15, and 30 kPa(see Fig. 4.13). The three solid surfaces were chosen to represent typical texturesfound on most pipelines. Epoxy-coated surfaces are frequently found on both offshoreand onshore pipelines as a means of corrosion protection (Braestrup et al., 2009; Guoet al., 2005; Revie, 2015).1294.2. Tests Conducted and ObservationsFigure 4.13: Experimentation tree for studying the effects of soil type, interface rough-ness, and effective normal stress level.1304.2. Tests Conducted and ObservationsThe sand-blasted steel surface was chosen to represent the surface of most offshorefoundations and anchoring systems. Characterizing the texture of these surfaces isimportant as the texture and the hardness of the solid surface can affect the soil-solidinterface shear behaviour. The proper characterization of the texture of a surfaceallows development of correlations with interface shear test data that could be usedin engineering designs.It is of relevance to note that the three solid surfaces that were used in the presentstudy, were reused from one test to the next. As described in Chapter 3, each testplate was reused up to a maximum of 18 times in the present research program. Theneed to reuse the test plates primarily arose from the cost of the plates and budgetconstraints. Given that coarse-grained soils can easily alter the surface texture of agiven test plate during testing, all tests involving coarse-grained soils were conductedat the very last stage of the testing program (i.e., testing commenced with the fine-grained soils and coarse-grained soils were tested at the very end of the test program).The changes in surface roughness of the three solid surfaces observed in the macro-scale interface direct shear test program is discussed in Section 4.2.7.It has been shown that soils can exhibit curved shear strength envelopes at rela-tively low effective normal stresses (Mesri and Abdel-Ghaffar, 1993; Stark and Eid,1994; Terzaghi et al., 1996; Watry and Lade, 2000). Given that most soil-pipe inter-action problems entail normal stress conditions that fall within this critical range, itis of interest to study the effect of normal stress on the shear strength characteristicsof soil-solid interfaces. The magnitudes of the effective normal stresses were selectedbased on typical values that are encountered in practical soil-pipe interaction prob-lems. The lower bound of 3 kPa is particularly important when designing partiallyburied pipelines such as those found in offshore projects. The 15 and 30 kPa nor-mal stress levels are of importance for many on-shore soil-pipe interactions problems,where typically pipelines are buried in a trench with a cover of approximately 1 to2 meters; and for offshore foundations. As explained in Chapter 2, there are verylimited data that engineers can rely upon to make judgments on the sensitivity of agiven soil-pipe interface to the confining stress when operating at this lower range of1314.2. Tests Conducted and Observationsstresses.Also, as put forward at the beginning of this thesis, one of the primary goals ofthis research project was to produce high quality, reliable data gathered at relativelylow effective normal stresses. In order to gain a sense of the reliability of the test-ing methodology used to study the effect of confining stress on the interface shearstrength, a series of repeatability tests were also conducted. Details of these testsare provided later in the text. The tests conducted to evaluate the repeatability, areshown in Fig. 4.14.Fraser River sand Grey epoxy coated steel3 kPa3 kPaFraser River silt Grey epoxy coated steel3 kPa3 kPa30 kPaGreen epoxy coated steelKaoliniteGrey epoxy coated steel3 kPa3 kPa15 kPa3 kPa3 kPa30 kPa15 kPa30 kPa30 kPaFigure 4.14: Experimentation tree for ascertaining the repeatability of tests conductedusing the macro-scale interface direct shear apparatus.The rate of shear displacement also plays an important role in characterizing soil-solid interface shear strength. The rate of shear displacement was carefully chosen toensure that a fully drained condition is achieved at large displacements to calculatethe drained large-displacement interface friction angle. A shear displacement rate of1.0 mm/s was used for testing the Fraser River sand. A displacement rate of 0.0125mm/s was used for testing Fraser River silt, and a rate of 0.004 mm/s for testingKaolinite, Strong-Pit FGS, Redstone, and Kaosand was used. The appropriate dis-placement rates for the fine-grained soils were determined based on the consolidation1324.2. Tests Conducted and Observationscharacteristics of the soils (see Chapter 3 for details). While it is of interest to studythe effect of rate of shear on the interface shear strength, given the considerableamount of time required for testing fine-grained soils targeting drained conditions(e.g., typically a test may take two to three weeks to complete depending on the soiltype used), only a limited number of tests were conducted to study the effect of rateof shear. The tests conducted for studying the effect of rate of shear displacementare shown in Fig. 4.15.Fraser River silt Grey epoxy coated steel3 kPa0.05 mm/s0.07 mm/s0.01 mm/s0.14 mm/s0.14 mm/s30 kPa0.01 mm/s0.14 mm/sKaolinite Grey epoxy coated steel3 kPa0.04 mm/s0.002 mm/s15 kPa0.002 mm/s0.04 mm/sFigure 4.15: Experimentation tree for studying the effect of rate of shear displacementon the large-displacement interface shear strength at low effective normal stresses.It was important to ensure that fine-grained soil specimens were prepared in afully-saturated condition (or close to fully-saturated conditions) so that pore-waterpressure measurements done at the interface could be relied upon to determine theeffective normal stress at the soil-solid interface. Hence, reconstituted specimens ofthe fine-grained soils were prepared in slurry form according to the specimen prepa-ration procedures detailed in Chapter 3. Each slurry specimen was prepared at thepredetermined water content that would: (i) prevent segregation of the soil duringsettlement; and (ii) result in a slurry that is fairly thin which helps to produce afully saturated specimen. Each slurry specimen that was poured into the specimencontainer resulted in a thickness of 4 to 6 cm prior to settlement and consolidation.The slurry was sampled to determine the water content from which the void ratio1334.2. Tests Conducted and Observationsprior to consolidation was calculated. Each slurry specimen was allowed to settleunder its own weight prior to the application of the surcharge load, and was thenallowed to consolidate at the selected normal stress level imposed by the surchargeloading. Upon reaching approximately 90% consolidation, the specimen was shearedat a constant rate of shear displacement appropriate for the soil type to produce adrained condition at large displacement. The drained coefficient of interface frictionat large displacement was thus determined based on the macro-scale interface directshear testing. At the end of each test, the surcharge sand layer was carefully removedto expose the soil specimen. This allowed sampling of the specimen to determine thefinal moisture content and the final void ratio. The final thickness of the specimenwas also measured.Table 4.6 lists a summary of the macro-scale interface direct shear tests conductedin the present study. The next sections provide the results gathered from the tests(factual data and observations) conducted during this research project. Key findingswith regard to the effects of soil-type, normal stress, interface roughness etc., arepresented in the next Chapter.Table 4.6: Summary of macro-scale interface direct shear tests.Test IDi Soil Solid Rate (mm/s) σ′n (kPa)FRS Steel 3 FRSand Steel 1.0 4.0FRS Steel 15 FRSand Steel 1.0 15.4FRS Steel 30 FRSand Steel 1.0 33.8FRS Green 3 FRSand Green epoxy 1.0 4.0FRS Green 15 FRSand Green epoxy 1.0 15.4FRS Green 30 FRSand Green epoxy 1.0 30.7FRS Grey 3 FRSand Grey epoxy 1.0 4.0FRS Grey 15 FRSand Grey epoxy 1.0 16.8FRS Grey 30 FRSand Grey epoxy 1.0 35.9Continued on next page1344.2. Tests Conducted and ObservationsTable 4.6 – Continued from previous pageTest ID Soil Solid Rate (mm/s) σ′n (kPa)FRS Grey 3 R FRSand Grey epoxy 1.0 4.0FRSilt Steel 3 FRSilt Steel 0.0125 3.5FRSilt Steel 15 FRSilt Steel 0.0125 15.2FRSilt Steel 30 FRSilt Steel 0.0125 27.9FRSilt Green 3 FRSilt Green epoxy 0.0125 3.4FRSilt Green 15 FRSilt Green epoxy 0.0125 16.1FRSilt Green 30 FRSilt Green epoxy 0.0125 31.1FRSilt Grey 3 FRSilt Grey epoxy 0.0125 3.7FRSilt Grey 15 FRSilt Grey epoxy 0.0125 16.7FRSilt Grey 30 FRSilt Grey epoxy 0.0125 26.2FRSilt Grey 3 R1 FRSilt Grey epoxy 0.0125 3.4FRSilt Grey 3 R2 FRSilt Grey epoxy 0.0125 3.9FRSilt Grey 30 R1 FRSilt Grey epoxy 0.0125 27.4FRSilt Grey 30 R2 FRSilt Grey epoxy 0.0125 25.3FRSilt Grey 3 F0.01 FRSilt Grey epoxy 0.01 3.4FRSilt Grey 3 F0.05 FRSilt Grey epoxy 0.05 3.6FRSilt Grey 3 F0.07 FRSilt Grey epoxy 0.07 3.9FRSilt Grey 3 F0.14 FRSilt Grey epoxy 0.14 3.0FRSilt Grey 3 F0.14 R FRSilt Grey epoxy 0.14 3.0FRSilt Grey 30 F0.01 FRSilt Grey epoxy 0.01 26.6FRSilt Grey 30 F0.14 FRSilt Grey epoxy 0.14 29.0Kao Steel 3 Kaolinite Steel 0.004 3.5Kao Steel 15 Kaolinite Steel 0.004 14.6Kao Steel 30 Kaolinite Steel 0.004 27.0Continued on next page1354.2. Tests Conducted and ObservationsTable 4.6 – Continued from previous pageTest ID Soil Solid Rate (mm/s) σ′n (kPa)Kao Green 3 Kaolinite Green epoxy 0.004 2.6Kao Green 15 Kaolinite Green epoxy 0.004 13.9Kao Green 30 Kaolinite Green epoxy 0.004 24.7Kao Grey 3 Kaolinite Grey epoxy 0.004 3.1Kao Grey 6 Kaolinite Grey epoxy 0.004 5.1Kao Grey 10 Kaolinite Grey epoxy 0.004 8.5Kao Grey 15 Kaolinite Grey epoxy 0.004 16.2Kao Grey 30 Kaolinite Grey epoxy 0.004 36.6Kao Green 3 R1 Kaolinite Green epoxy 0.004 2.1Kao Green 3 R2 Kaolinite Green epoxy 0.004 2.3Kao Grey 3 R1 Kaolinite Green epoxy 0.004 3.1Kao Grey 3 R2 Kaolinite Green epoxy 0.004 3.0Kao Grey 15 R1 Kaolinite Grey epoxy 0.004 16.2Kao Grey 15 R2 Kaolinite Grey epoxy 0.004 12.5Kao Grey 30 R1 Kaolinite Grey epoxy 0.004 37.1Kao Grey 30 R2 Kaolinite Grey epoxy 0.004 36.6Kao Grey 3 F0.002 Kaolinite Grey epoxy 0.002 2.3Kao Grey 3 F0.04 Kaolinite Grey epoxy 0.04 2.6Kao Grey 15 F0.002 Kaolinite Grey epoxy 0.002 16.4Kao Grey 15 F0.04 Kaolinite Grey epoxy 0.04 13.1StrongPitFGS Steel 3 Strong-Pit FGS Steel 0.004 3.2StrongPitFGS Grey 3 Strong-Pit FGS Grey epoxy 0.004 3.3StrongPitFGS Grey 15 Strong-Pit FGS Grey epoxy 0.004 14.6StrongPitFGS Green 3 Strong-Pit FGS Green epoxy 0.004 2.6Continued on next page1364.2. Tests Conducted and ObservationsTable 4.6 – Continued from previous pageTest ID Soil Solid Rate (mm/s) σ′n (kPa)StrongPitFGS Green 15 Strong-Pit FGS Green epoxy 0.004 15.4StrongPitFGS Green 30 Strong-Pit FGS Green epoxy 0.004 25.6Redstone Green 3 Redstone Green epoxy 0.004 2.8Kaosand Steel 3 Kaosand Steel 0.004 3.3Kaosand Steel 15 Kaosand Steel 0.004 14.6i R: Repeatability tests; F: Rate effect tests4.2.1 Interface shear characteristics of Fraser River sandwith different solid surfacesFraser River sand on sand-blasted mild steelThe results obtained from the tests conducted on the sand-blasted mild steel surfaceare shown in Fig. 4.16. The average interface shear stress reached a plateau atapproximately 2 to 4 cm of shear displacement, and remained constant thereafter. Alarge-displacement interface friction angle δ of 28°, which is slightly lower than theconstant volume internal friction angle of the sand (φcv = 33°), was observed for allthree tests conducted at three stress levels between 4 kPa to 35 kPa. No noticeablesensitivity to the effective normal stress was observed.The results of the macro-scale interface direct shear tests conducted on FraserRiver sand against the sand-blasted mild-steel surface are summarized in Table 4.7.Shown in Fig. 4.17 are photographs of the exposed trailing end of the solid testsurface taken at the end of the interface direct shear test conducted targeting σ′n of3 kPa. A sporadic spread of a thin layer of test soil was observed over some parts ofthe solid test surface. This observation was consistent in all three tests.1374.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50051015202530FRSand_Steel_3 | σn' = 4.0 kPaFRSand_Steel_15 | σn' = 15.4 kPaFRSand_Steel_30 | σn' = 33.8 kPa(a) Variation of shear stress with shear displacementCoefficient of interface friction (μi)Shear displacement (cm)0 10 20 30 40 5000.20.40.60.81FRSand_Steel_3 | σ'n = 4.0 kPaFRSand_Steel_15 | σn' = 15.4 kPaFRSand_Steel_30 | σn' = 33.8 kPa(b) Variation of coefficient of interface friction with shear displacement024681012141618200 5 10 15 20 25 30 35 40Shear stress (kPa)Effective normal stress (kPa)(𝛿=28°)0.541.00(c) Effective residual interface shear strength envelopeFigure 4.16: Macro-scale interface direct shear test results for Fraser River sand onsand-blasted mild steel.1384.2. Tests Conducted and ObservationsTable 4.7: Results of macro-scale interface direct shear tests conducted on FraserRiver sand against the sand-blasted mild-steel surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φcv)Ra 50D50/D50FRS Steel 3 4.0 ±0.01 27 ±0.10 0.52 ±0.0020 0.78 0.03FRS Steel 15 15.4 ±0.01 28 ±0.02 0.54 ±0.0004 0.82 0.03FRS Steel 30 33.8 ±0.01 28 ±0.01 0.54 ±0.0002 0.82 0.03Figure 4.17: Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test of Fraser River sandon sand-blasted mild steel.Fraser River sand on green epoxyThe results from the shearing of dry Fraser River sand specimens on the green epoxycoated mild steel surface are presented in Fig. 4.18. Testing the sand on the greenepoxy surface yielded a large-displacement interface friction angle δ of 28° at all threeeffective normal stresses. As such, both the sand-blasted mild steel surface and thegreen epoxy surface produced the same interface friction angle for Fraser River sand.Visual inspection of the epoxy test surface post-test showed signs of mild abrasion.The abrasion was found to be highly random in direction, and no distinct streaks ofscraping oriented in the (north-south) direction of shear were observed.1394.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 60 700510152025FRSand_Green_3 | σn' = 4.0 kPaFRSand_Green_15 | σn' = 15.4 kPaFRSand_Green_30 | σn' = 30.7 kPa(a) Variation of shear stress with shear displacementCoefficient of interface friction (μi)Shear displacement (cm)0 10 20 30 40 50 60 7000.20.40.60.81FRSand_Green_3 | σn' = 4.0 kPaFRSand_Green_15 | σn' = 15.4 kPaFRSand_Green_30 | σn' = 30.7 kPa(b) Variation of coefficient of interface friction with shear displacement0246810121416180 5 10 15 20 25 30 35Shear stress (kPa)Effective normal stress (kPa)(𝛿=27°)0.511.00(c) Effective residual interface shear strength envelopeFigure 4.18: Macro-scale interface direct shear test results for Fraser River sand ongreen epoxy.1404.2. Tests Conducted and ObservationsAs discussed at the onset of this chapter, the roughness of the test surfaces haveto be considered carefully when attempting to draw correlations of the interface shearcharacteristics. It is of relevance to note that in terms of Ra 50D50 values, the sand-blasted mild steel surface was of higher roughness than the green epoxy coated steelsurface. When comparing the roughness size distribution curves of the solid surfaceswith the grain size distribution curve of Fraser River sand (see Fig. 4.12), it can beseen that the roughness size distributions of both, the sand-blasted mild steel andthe green epoxy coated steel, surfaces are much finer than the grain size distributionof Fraser River sand.Entrapment of sand at south edge of specimen container observed after 28 cm of shear displacement.Figure 4.19: Abnormal increase in shear stress observed during testing Fraser Riversand against the green epoxy surface at 30 kPa effective normal stress (Test ID:FRS Green 30) caused by the entrapment of sand underneath the south edge of spec-imen container.It is of relevance to make a comment on the test conducted at the effective normalstress of 30 kPa (Test ID: FRS Green 30). In this test, the interface shear stressshowed a gradual increase after 28 cm of shear displacement. A similar behaviourwas observed when testing Fraser River sand on the grey epoxy surface at 30 kPaeffective normal stress (Test ID: FRS Grey 30) as well. This increase in shear stressappeared to be due to soil being gradually entrapped underneath the trailing edge1414.2. Tests Conducted and Observations(south edge) of the specimen container, which caused the specimen container to liftup from that end and consequently shift some of its weight to the test surface throughthe entrapped sand layer. This was evident in the corresponding increase in the totalnormal force measured by the base load cells of the device, and in the correspondingreduction in the side-wall load cell readings (see Fig. 4.19).The increase in normal stress due to the entrapment of sand underneath the southedge of the specimen container resulted in a corresponding increase in the averageshear stress. Therefore, as can be seen in Fig. 4.18-(b), this had no effect on thecalculated coefficient of interface friction. Subsequent investigation of this problemshowed that making the size of the gap between the test plate and the specimencontainer slightly larger (changing from 5 mm gap to 10 mm gap) eliminated thisissue.The results of macro-scale interface direct shear tests conducted on Fraser Riversand on the green epoxy surface are summarized in Table 4.8.Table 4.8: Results of macro-scale interface direct shear tests conducted on FraserRiver sand against the abrasion resistant green epoxy surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φcv)Ra 50D50/D50FRS Green 3 4.0 ±0.01 27 ±0.10 0.51 ±0.0020 0.78 0.007FRS Green 15 15.4 ±0.01 27 ±0.02 0.51 ±0.0004 0.78 0.007FRS Green 30 30.7 ±0.01 27 ±0.01 0.51 ±0.0002 0.78 0.007Fraser River sand on grey epoxyThe results from the shearing of dry Fraser River sand specimens against the greyepoxy coated mild steel surface are presented in Fig. 4.20. Overall, the sand - greyepoxy interface showed a slightly different behaviour compared to that observed forthe sand - steel and sand - green epoxy interfaces. The large-displacement interfacefriction angle was slightly lower (δ = 26°), and a very mild sensitivity to the effectivenormal stress was observed.1424.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 60 700510152025FRSand_Grey_3 | σn' = 4.0 kPaFRSand_Grey_15 | σn' = 16.8 kPaFRSand_Grey_30 | σn' = 35.9 kPa(a) Variation of shear stress with shear displacementCoefficient of interface friction (μi)Shear displacement (cm)0 10 20 30 40 50 60 7000.20.40.60.81FRSand_Grey_3 | σn' = 4.0 kPaFRSand_Grey_15 | σn' = 16.8 kPaFRSand_Grey_30 | σn' = 35.9 kPa(b) Variation of coefficient of interface friction with shear displacement024681012141618200 5 10 15 20 25 30 35 40Shear stress (kPa)Effective normal stress (kPa)(𝛿 = 26°)0.481.00(c) Effective residual interface shear strength envelopeFigure 4.20: Macro-scale interface direct shear test results for Fraser River sand ongrey epoxy.1434.2. Tests Conducted and ObservationsAs shown in Fig. 4.20-b, the value of µi obtained for the test conducted at theeffective normal stress of 4 kPa was slightly lower than that obtained from testsconducted at effective normal stresses of 15 and 30 kPa. Visual inspection of theepoxy test surface post-test showed signs of mild abrasion. The abrasion was foundto be highly random in direction, and no distinct streaks of scraping oriented in the(north-south) direction of shear were observed.The results of macro-scale interface direct shear tests conducted on Fraser Riversand against the grey epoxy surface are summarized in Table 4.9.Table 4.9: Results of macro-scale interface direct shear tests conducted on FraserRiver sand against the grey epoxy surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φcv)Ra 50D50/D50FRS Grey 3 4.0 ±0.01 23 ±0.10 0.43 ±0.0010 0.65 0.006FRS Grey 15 16.8 ±0.01 25 ±0.02 0.47 ±0.0004 0.72 0.006FRS Grey 30 35.9 ±0.01 26 ±0.01 0.48 ±0.0002 0.75 0.006Figure 4.21: Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test of Fraser River sandon grey epoxy coated steel.Shown in Fig. 4.21 are photographs of the exposed trailing end of the solid testsurface, taken at the end of the interface direct shear test conducted targeting σ′n of 31444.2. Tests Conducted and ObservationskPa. A sporadic spread of a thin layer of test soil was observed to cover some parts ofthe trailing end of the solid test surface. This observation was consistent in all threetests.4.2.2 Interface shear characteristics of Fraser River siltwith different solid surfacesFraser River silt on sand-blasted mild steelFor the tests conducted on Fraser River silt on the sand-blasted mild steel surfacetargeting 3, 15 and 30 kPa effective normal stresses, the variation of the averageinterface shear stress with shear displacement observed during the first 5 cm of sheardisplacement is shown in Fig. 4.22. The variation of pore-water pressure with sheardisplacement for the first 5 cm of shear displacement is shown in Fig. 4.23.During testing at σ′n of 30 kPa, the actuator came to a halt at approximately2.5 cm of shear displacement due to an over-current fault. Rectification of the faultcondition required approximately 30 minutes, and during this time, the soil specimenconsolidated and the pore-water pressure dissipated to the static water head. Imme-diately upon resuming shearing, a negative pore-water pressure transient followed bya rise of pore-water pressure back to the original (i.e., prior to the fault condition)pore-water pressure level occurred within approximately 5 mm of shear displacement.No distinguishable peak in shear stress was observed in any of the tests. At theonset of shearing, the average pore-water pressure at the interface was observed torapidly rise to a peak, and then gradually dissipate until it approached the static waterhead. The rise in pore-water pressure at the beginning of shearing is consistent withthe contractive behaviour typically observed in normally consolidated fine-grainedsoils. The pore-water pressure was observed to reach its peak at around 3 to 4 mmof shear displacement in all three tests, regardless of the total normal stress level.As expected, the magnitude of the peak excess pore-water pressure was observed togenerally increase with increasing applied total normal stress (i.e., peak excess pore-water pressure values of 1.0, 7.0, and 10.5 kPa were observed for the tests conducted1454.2. Tests Conducted and Observationstargeting effective normal stresses of 3, 15, and 30 kPa).Figure 4.22: Variation of interface shear stress with shear displacement for FraserRiver silt on sand-blasted mild steel.Slope ≈ 40 kPa/cmActuator cutoff due to over-current limit.Shear displacement temporarily halted and resumed after approximately 30 minutes.Figure 4.23: Variation of pore-water pressure with shear displacement for Fraser Riversilt on sand-blasted mild steel.The excess pore-water pressure response observed at small displacements should beconsidered with caution when drawing conclusions as the interface direct shear test, inan overall sense, is partially drained at this stage of shearing. The initial shear inducedexcess pore-water pressure is able to continuously dissipate during testing. Therefore,the peak excess pore-water pressure values may not be the maximum possible valuesexpected, for example, as in an ideal undrained test. This is especially true for theFraser River silt material that has a relatively higher permeability compared to aclayey soil. The complete set of test data covering the full range of displacement ispresented in Fig. 4.24.It was also considered of interest to examine the variation of the average interfaceshear stress with the effective normal stress during the progress of the test as shown1464.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 60 70051015202530FRSilt_Steel_3 FRSilt_Steel_15 FRSilt_Steel_30(a) Variation of shear stress with shear displacementu (kPa)Shear displacement (cm)0 10 20 30 40 50 60 7002468101214FRSilt_Steel_3FRSilt_Steel_15FRSilt_Steel_30(b) Variation of average pore-water pressure with shear displacementσ n' (kPa)Shear displacement (cm)0 10 20 30 40 50 60 7001020304050FRSilt_Steel_3 FRSilt_Steel_15 FRSilt_Steel_30(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 60 7000.20.40.60.811.21.4 FRSilt_Steel_3 FRSilt_Steel_15 FRSilt_Steel_30(d) Variation of coefficient of interface friction with shear displacementFigure 4.24: Macro-scale interface direct shear test results for Fraser River silt onsand-blasted mild steel.1474.2. Tests Conducted and Observations0.34 cm2.00 cm70.0 cm0.34 cm3.07 cm70.0 cm0.32 cm4.69 cm70.0 cm(𝛿 = 27° ~ 31°)Figure 4.25: Variation of interface shear stress with effective normal stress for FraserRiver silt on sand-blasted mild steel.in Fig. 4.25. While the curves shown in Fig. 4.25 resemble stress paths, it should bekept in mind that these are not stress paths in the conventional sense. Nevertheless,as shown in Fig. 4.25, all three tests showed a τi-σ′n response at the interface trend-wise similar to the shapes of stress paths observed in undrained shearing of normallyconsolidated soil. It is notable that the initial reduction in effective normal stress(due to generation of excess pore water pressures) and the subsequent recovery to thedrained condition occurred during the first 2 cm of shear displacement. As mentionedearlier, the rise in excess pore-water pressure at the onset of shearing occurs underpartially drained conditions. During dissipation of the excess pore-water pressure,the average interface shear stress continues to increase with the increase in effectivenormal stress along a well defined failure envelope. At large displacements, when theexcess pore-water pressure has fully dissipated to the static water head, the averageinterface shear stress remains constant. The points of shear stress at large displace-ments can be used to interpret and develop a drained large-displacement interfaceshear strength envelope. From the data presented in Fig. 4.25, the drained large-displacement interface friction angle was found to be 29° to 30°; this is approximatelythe same as that determined for Fraser River sand on the sand-blasted mild steel1484.2. Tests Conducted and Observationssurface.The results of macro-scale interface direct shear tests conducted on Fraser Riversilt against the sand-blasted mild steel surface are summarized in Table 4.10.Table 4.10: Results of macro-scale interface direct shear tests conducted on FraserRiver silt against the sand-blasted mild steel surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φ′) Ra 50D50/D50FRSilt Steel 3 3.5 ±0.05 27 ±0.43 0.52 ±0.007 0.78 0.16FRSilt Steel 15 15.2 ±0.05 29 ±0.10 0.55 ±0.002 0.85 0.16FRSilt Steel 30 27.9 ±0.05 31 ±0.06 0.59 ±0.001 0.93 0.16Shown in Fig. 4.26 are photographs of the exposed trailing end of the solid testsurface, taken at the end of the interface shear test conducted targeting σ′n of 3 kPa.A uniform spread of a thin layer of test soil was observed to completely cover thetrailing end of the solid test surface. This observation was consistent in all threetests.Figure 4.26: Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test of Fraser River silt onsand-blasted mild steel targeting σ′n of 3 kPa.1494.2. Tests Conducted and ObservationsFraser River silt on grey epoxyThe shear stress-displacement response of the Fraser River silt on the grey epoxysurface obtained targeting 3, 15 and 30 kPa effective normal stresses, observed duringthe first 5 cm of shear displacement is shown in Fig. 4.27. The variation of pore-waterpressure with shear displacement for the first 5 cm of shear displacement is shown inFig. 4.28. The complete set of test data over the full displacement range is presentedin Fig. 4.29. The variation of the average interface shear stress with the effectivenormal stress during the progress of the tests is presented in Fig. 4.30.Figure 4.27: Variation of interface shear stress with shear displacement for FraserRiver silt on grey epoxy.Slope ≈ 47 kPa/cmFigure 4.28: Variation of pore-water pressure with shear displacement for Fraser Riversilt on grey epoxy.1504.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 60 7005101520FRSilt_Grey_3 FRSilt_Grey_15 FRSilt_Grey_30(a) Variation of shear stress with shear displacementu (kPa)Shear displacement (cm)0 10 20 30 40 50 60 7002468101214FRSilt_Grey_3FRSilt_Grey_15FRSilt_Grey_30(b) Variation of average pore-water pressure with shear displacementσ n' (kPa)Shear displacement (cm)0 10 20 30 40 50 60 700510152025303540FRSilt_Grey_3 FRSilt_Grey_15 FRSilt_Grey_30(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 60 7000.20.40.60.81FRSilt_Grey_3 FRSilt_Grey_15 FRSilt_Grey_30(d) Variation of coefficient of interface friction with shear displacementFigure 4.29: Macro-scale interface direct shear test results for Fraser River silt ongrey epoxy.1514.2. Tests Conducted and ObservationsAll three tests showed a similar τi-σ′n response (see Fig. 4.30). The drained large-displacement interface friction angle was observed to increase from 25° to 27° as theeffective normal stress was increased from 3 to 30 kPa.0.26 cm70.0 cm70.0 cm0.21 cm3.81 cm34.3 cm1.50 cm8.84 cm70.0 cm(𝛿 = 25° ~ 27°)Figure 4.30: Variation of interface shear stress with effective normal stress for FraserRiver silt on grey epoxy.The results of macro-scale interface direct shear tests conducted on Fraser Riversilt against the grey epoxy surface are summarized in Table 4.11. Shown in Fig. 4.31Table 4.11: Results of macro-scale interface direct shear tests conducted on FraserRiver silt against the grey epoxy surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φ′) Ra 50D50/D50FRSilt Grey 3 3.7 ±0.05 25 ±0.38 0.46 ±0.007 0.72 0.04FRSilt Grey 15 16.7 ±0.05 26 ±0.09 0.49 ±0.001 0.75 0.04FRSilt Grey 30 26.2 ±0.05 27 ±0.06 0.52 ±0.001 0.78 0.04are photographs of the exposed trailing end of the solid test surface, taken at the endof the interface shear test conducted targeting σ′n of 15 kPa. The trailing end of thesolid test surface was observed to be relatively clean with only trace amounts of the1524.2. Tests Conducted and Observationssoil material left on the surface of the plate This observation was consistent in allthree tests.Figure 4.31: Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test of Fraser River silt ongrey epoxy coated steel targeting σ′n of 15 kPa.Fraser River silt on green epoxyThe shear stress-displacement response of the Fraser River silt on the green epoxysurface obtained targeting 3, 15 and 30 kPa effective normal stresses, observed duringthe first 5 cm of shear displacement is shown in Fig. 4.32. The variation of pore-water pressure with shear displacement for the same 5 cm is shown in Fig. 4.33.The collected test data over the full displacement range is shown in Fig. 4.35. Thevariation of the average interface shear stress with the effective normal stress duringthe progress of the tests is shown in Fig. 4.34.1534.2. Tests Conducted and ObservationsFigure 4.32: Variation of interface shear stress with shear displacement for FraserRiver silt on green epoxy.Figure 4.33: Variation of pore-water pressure with shear displacement for Fraser Riversilt on green epoxy.0.91 cm 0.29 cm70.0 cm0.31 cm70.0 cm1.15 cm0.18 cm2.67 cm65.0 cm(𝛿 = 27°)Figure 4.34: Variation of interface shear stress with effective normal stress for FraserRiver silt on green epoxy.1544.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 60 700510152025Target effective normal stress = 3 kPaTarget effective normal stress = 15 kPaTarget effective normal stress = 30 kPa(a) Variation of shear stress with shear displacementu (kPa)Shear displacement (cm)0 10 20 30 40 50 60 7002468FRSilt_Green_3FRSilt_Green_15FRSilt_Green_30(b) Variation of average pore-water pressure with shear displacementσ n' (kPa)Shear displacement (cm)0 10 20 30 40 50 60 7001020304050FRSilt_Green_3 FRSilt_Green_15 FRSilt_Green_30(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 60 7000.20.40.60.81FRSilt_Green_3 FRSilt_Green_15 FRSilt_Green_30(d) Variation of coefficient of interface friction with shear displacementFigure 4.35: Macro-scale interface direct shear test results for Fraser River silt ongreen epoxy.1554.2. Tests Conducted and ObservationsIt was interesting to note that a large-displacement interface friction angle of 27°was observed for the Fraser River silt - green epoxy interface at all three effectivenormal stress levels tested. The results of macro-scale interface direct shear testsconducted on Fraser River silt against the green epoxy surface are summarized inTable 4.12.Table 4.12: Results of macro-scale interface direct shear tests conducted on FraserRiver silt against the green epoxy surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φ′) Ra 50D50/D50FRSilt Green 3 3.4 ±0.05 27 ±0.43 0.51 ±0.007 0.78 0.04FRSilt Green 15 16.1 ±0.05 27 ±0.09 0.51 ±0.002 0.78 0.04FRSilt Green 30 31.1 ±0.05 27 ±0.05 0.50 ±0.001 0.78 0.04Shown in Fig. 4.36 are photographs of the exposed trailing end of the solid testsurface, taken at the end of the interface shear test conducted targeting σ′n of 15 kPa.The trailing end of the solid test surface was observed to be relatively clean with onlytrace amounts of the soil material left on the surface of the plate. This observationwas consistent in all three tests.Figure 4.36: Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test of Fraser River silt ongreen epoxy coated steel targeting σ′n of 15 kPa.1564.2. Tests Conducted and Observations4.2.3 Interface shear characteristics of Kaolinite withdifferent solid surfacesKaolinite on sand-blasted mild steelThe shear stress - displacement response of Kaolinite on the sand-blasted mild steelsurface obtained targeting 3, 15 and 30 kPa effective normal stresses, during the first 5cm of shear displacement is shown in Fig. 4.37. The variation of pore-water pressurewith shear displacement for the first 5 cm is shown in Fig. 4.38. The collected testdata over the full displacement range is shown in Fig.4.39. The variation of theaverage interface shear stress with the effective normal stress during the progress ofthe tests is presented in Fig. 4.40.Figure 4.37: Variation of interface shear stress with shear displacement for Kaoliniteon sand-blasted mild steel.Slope ≈ 29 kPa/cmFigure 4.38: Variation of pore-water pressure with shear displacement for Kaoliniteon sand-blasted mild steel.1574.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 6005101520Kao_Steel_3 Kao_Steel_15 Kao_Steel_30(a) Variation of shear stress with shear displacementu (kPa)Shear displacement (cm)0 10 20 30 40 50 600246810Kao_Steel_3Kao_Steel_15Kao_Steel_30(b) Variation of average pore-water pressure with shear displacementσ n' (kPa)Shear displacement (cm)0 10 20 30 40 50 600510152025303540Kao_Steel_3 Kao_Steel_15 Kao_Steel_30(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 6000.20.40.60.81Kao_Steel_3 Kao_Steel_15 Kao_Steel_30(d) Variation of coefficient of interface friction with shear displacementFigure 4.39: Macro-scale interface direct shear test results for Kaolinite on sand-blasted mild steel.1584.2. Tests Conducted and Observations60.0 cm0.34 cm55.0 cm0.40 cm6.93 cm20.0 cm0.75 cm12.0 cm60.0 cm(𝛿 = 23° ~ 27°)Figure 4.40: Variation of interface shear stress with effective normal stress for Kaoli-nite on sand-blasted mild steel.The shear stress response showed no distinguishable peak (followed by a drop)in shear stress. This is also in accord with that expected as the shear behaviour fortypical normally consolidated clays. As that observed in interface shear tests withFraser River sand as well as Fraser River silt, the tests with Kaolinite showed anincrease of shear stress during the first 2 to 5 cm of shear displacement, followed bya plateau that remained constant for the remainder of shearing.The drained large-displacement interface friction angle was found to reduce from27° to 23° as the effective normal stress was increased from 3 kPa to 30 kPa. Theresults of macro-scale interface direct shear tests conducted on Kaolinite against thesand-blasted steel are summarized in Table 4.13.Shown in Fig. 4.41 are photographs of the exposed trailing end of the solid testsurface taken at the end of the interface shear test conducted targeting σ′n of 15 kPa.A uniform spread of a relatively thick (approximately 5 mm thick) layer of test soilwas observed to adhere to (removal of the layer of the soil layer required some effort)and completely cover the trailing end of the solid test surface. This observation wasconsistent in all three tests.1594.2. Tests Conducted and ObservationsTable 4.13: Results of macro-scale interface direct shear tests conducted on Kaoliniteagainst the sand-blasted mild steel surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φ′) Ra 50D50/D50Kao Steel 3 3.5 ±0.05 27 ±0.42 0.50 ±0.007 1.00 2.8Kao Steel 15 14.6 ±0.05 25 ±0.09 0.46 ±0.002 1.00 2.8Kao Steel 30 27.0 ±0.05 23 ±0.04 0.42 ±0.001 0.93 2.8Figure 4.41: Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test of Kaolinite on sand-blasted mild steel targeting σ′n of 15 kPa.Kaolinite on grey epoxyThe shear stress - displacement response of the Kaolinite on the grey epoxy surfaceobtained targeting 3, 15 and 30 kPa effective normal stresses, during the first 5 cm ofshear displacement is shown in Fig. 4.42. The variation of pore-water pressure withshear displacement for the first 5 cm is shown in Fig. 4.43. The collected test dataover the full displacement range is shown in Fig. 4.44. The variation of the averageinterface shear stress with the effective normal stress during the progress of the testsis presented in Fig. 4.45.The drained large-displacement interface friction angles were observed to be equal1604.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 1 2 3 4 502468101214 Kao_Grey_1.5Kao_Grey_2Kao_Grey_3Kao_Grey_10Kao_Grey_15Kao_Grey_30Figure 4.42: Variation of interface shear stress with shear displacement for Kaoliniteon grey epoxy.u (kPa)Shear displacement (cm)0 1 2 3 4 505101520Kao_Grey_1.5Kao_Grey_2Kao_Grey_3Kao_Grey_10Kao_Grey_15Kao_Grey_30Figure 4.43: Variation of pore-water pressure with shear displacement for Kaoliniteon grey epoxy.for the two tests that were conducted at 15 and 30 kPa effective normal stresses.However, the test conducted targeting 3 kPa effective normal stress showed a relativelyhigher drained large-displacement interface friction angle indicating some sensitivityof the coefficient of interface shear strength to the effective normal stress. It wasalso observed that the drained large-displacement interface friction angles betweenKaolinite and grey epoxy were consistently lower compared to those observed for theKaolinite on sand-blasted mild steel interface. The results of the macro-scale interfacedirect shear tests conducted on Kaolinite against the grey epoxy are summarized inTable 4.13.1614.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 6002468101214 Kao_Grey_1.5Kao_Grey_2Kao_Grey_3Kao_Grey_15Kao_Grey_30Kao_Grey_10(a) Variation of shear stress with shear displacementu (kPa)Shear displacement (cm)0 10 20 30 40 50 6005101520Kao_Grey_1.5Kao_Grey_2Kao_Grey_3Kao_Grey_10Kao_Grey_15Kao_Grey_30(b) Variation of average pore-water pressure with shear displacementσ n' (kPa)Shear displacement (cm)0 10 20 30 40 50 600102030405060Kao_Grey_1.5Kao_Grey_2Kao_Grey_3Kao_Grey_10Kao_Grey_15Kao_Grey_30(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 6000.20.40.60.81Kao_Grey_1.5Kao_Grey_2Kao_Grey_3Kao_Grey_10Kao_Grey_15Kao_Grey_30(d) Variation of coefficient of interface friction with shear displacementFigure 4.44: Macro-scale interface direct shear test results for Kaolinite on grey epoxy.1624.2. Tests Conducted and Observations6.05 cm0.59 cm12.1 cm0.26 cm70.0 cm58.0 cm0.44 cm18.4 cm70.0 cm(𝛿 = 15° ~ 25°)Figure 4.45: Variation of interface shear stress with effective normal stress for Kaoli-nite on grey epoxy.Table 4.14: Results of macro-scale interface direct shear tests conducted on Kaoliniteagainst the grey epoxy surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φ′) Ra 50D50/D50Kao Grey 1.5 1.4 ±0.05 25 ±1.22 0.47 ±0.0210 1.00 0.2Kao Grey 2 2.3 ±0.05 22 ±0.51 0.41 ±0.0090 0.86 0.2Kao Grey 3 3.1 ±0.05 19 ±0.34 0.34 ±0.0060 0.74 0.2Kao Grey 10 8.5 ±0.05 17 ±0.10 0.30 ±0.0020 0.66 0.2Kao Grey 15 16.2 ±0.05 16 ±0.05 0.28 ±0.0010 0.62 0.2Kao Grey 30 36.6 ±0.05 16 ±0.02 0.28 ±0.0004 0.62 0.2Shown in Fig. 4.46 are photographs of the exposed trailing end of the solid testsurface, taken at the end of the interface shear test conducted targeting σ′n of 15 kPa.The trailing end of the solid test surface was observed to be relatively clean with onlytrace amounts of the soil material left on the surface of the plate. This observationwas consistent in all three tests.1634.2. Tests Conducted and ObservationsFigure 4.46: Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test of Kaolinite on greyepoxy coated steel targeting σ′n of 15 kPa.Kaolinite on green epoxyThe shear stress - displacement response of the Kaolinite on the green epoxy surfaceobtained targeting 3, 15 and 30 kPa effective normal stresses, during the first 5 cmof shear displacement is shown in Fig. 4.47. The variation of pore-water pressurewith shear displacement for the same displacement range is shown in Fig. 4.48. Thecollected test data over the full displacement range is presented in Fig. 4.49, and thevariation of the average interface shear stress with the effective normal stress duringthe progress of the tests is presented in Fig. 4.50.The drained large-displacement interface friction angles were observed to be equalfor the two tests that were conducted targeting 15 and 30 kPa effective normal stresses.The test targeting 3 kPa effective normal stress showed a distinctively higher drainedlarge-displacement interface friction angle. This indicates some sensitivity of thelarge-displacement interface shear strength to the effective normal stress. It is alsonotable that the drained large-displacement interface friction angles between Kaoliniteand green epoxy were consistently lower compared to those observed for the Kaoliniteon sand-blasted mild steel interface.1644.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 1 2 3 4 50246810Kao_Green_3 Kao_Green_15 Kao_Green_30Figure 4.47: Variation of interface shear stress with shear displacement for Kaoliniteon green epoxy.Slope ≈ 23 kPa/cmFigure 4.48: Variation of pore-water pressure with shear displacement for Kaoliniteon green epoxy.The results of macro-scale interface direct shear tests conducted on Kaoliniteagainst the green epoxy are summarized in Table 4.15.Table 4.15: Results of macro-scale interface direct shear tests conducted on Kaoliniteagainst the green epoxy surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φ′) Ra 50D50/D50Kao Green 3 2.6 ±0.05 18 ±0.37 0.32 ±0.006 0.69 0.07Kao Green 15 13.9 ±0.05 13 ±0.05 0.24 ±0.001 0.50 0.07Kao Green 30 24.7 ±0.05 13 ±0.03 0.24 ±0.001 0.51 0.071654.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 600246810Kao_Green_3 Kao_Green_15 Kao_Green_30(a) Variation of shear stress with shear displacementu (kPa)Shear displacement (cm)0 10 20 30 40 50 600246810Kao_Green_3Kao_Green_15Kao_Green_30(b) Variation of average pore-water pressure with shear displacementσ n' (kPa)Shear displacement (cm)0 10 20 30 40 50 600510152025303540Kao_Green_3 Kao_Green_15 Kao_Green_30(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 6000.20.40.60.81Kao_Green_3 Kao_Green_15 Kao_Green_30(d) Variation of coefficient of interface friction with shear displacementFigure 4.49: Macro-scale interface direct shear test results for Kaolinite on greenepoxy.1664.2. Tests Conducted and Observations60.0 cm0.44 cm3.75 cm0.41 cm10.4 cm58.0 cm0.49 cm12.5 cm60.0 cm(𝛿 = 13° ~ 18°)Figure 4.50: Variation of interface shear stress with effective normal stress for Kaoli-nite on green epoxy.Shown in Fig. 4.51 are photographs of the exposed trailing end of the solid testsurface, taken at the end of the interface shear test conducted targeting σ′n of 15 kPa.The trailing end of the solid test surface was observed to be relatively clean with onlytrace amounts of the soil material left on the surface of the plate. This observationwas consistent in all three tests.Figure 4.51: Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test of Kaolinite on greenepoxy coated steel targeting σ′n of 15 kPa.1674.2. Tests Conducted and Observations4.2.4 Interface shear characteristics of Strong-Pit FGS withdifferent solid surfacesStrong-Pit FGS on sand-blasted mild steelThe shear stress - displacement response of the Strong-Pit FGS on the sand-blastedmild steel surface obtained targeting 3 kPa effective normal stress, during the first 5cm of shear displacement is shown in Fig. 4.52. The variation of pore-water pressurewith shear displacement for the first 5 cm is shown in Fig. 4.53. The collected testdata over the full displacement range is shown in Fig. 4.54. The variation of theaverage interface shear stress with the effective normal stress during the progress ofthe test is presented in Fig. 4.55.Shear stress (kPa)Shear displacement (cm)0 1 2 3 4 500.20.40.60.811.2Figure 4.52: Variation of interface shear stress with shear displacement for Strong-PitFGS on sand-blasted mild steel.u (kPa)Shear displacement (cm)0 1 2 3 4 5012345Figure 4.53: Variation of pore-water pressure with shear displacement for Strong-PitFGS on sand-blasted mild steel.1684.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 6000.511.522.53(a) Variation of shear stress with shear displacementu (kPa)Shear displacement (cm)0 10 20 30 40 50 60012345(b) Variation of average pore-water pressure with shear displacementσ n' (kPa)Shear displacement (cm)0 10 20 30 40 50 60012345(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 6000.20.40.60.811.21.4(d) Variation of coefficient of interface friction with shear displacementFigure 4.54: Macro-scale interface direct shear test results for Strong-Pit FGS onsand-blasted mild steel.1694.2. Tests Conducted and Observations60.0 cm34.1 cm1.14 cm0.17 cm(𝛿 = 29°)Figure 4.55: Variation of interface shear stress with effective normal stress for Strong-Pit FGS on sand-blasted mild steel.Table 4.16: Results of macro-scale interface direct shear tests conducted on Strong-PitFGS against the sand-blasted mild steel surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φ′) Ra 50D50/D50Strong-PitFGS Steel 33.2 ±0.05 29 ±0.52 0.56 ±0.01 - 0.69Reliable data for the internal friction angle of Strong-Pit FGS at large-displacementsand low confining stresses was not available at the time of writing and hence the in-terface efficiency factors are not provided in Table 4.16. Shown in Fig. 4.56 arephotographs of the exposed trailing end of the solid test surface, taken at the end ofthe interface shear test. A thin layer of test soil was adhered to and covered most ofthe trailing end of the solid test surface.1704.2. Tests Conducted and ObservationsFigure 4.56: Photographs showing the exposed trailing end of the solid test platesurface at the end of the macro-scale interface direct shear test of Strong-Pit FGS onsand-blasted mild steel.Strong-Pit FGS on grey epoxy coated steelThe shear stress - displacement response of the Strong-Pit FGS on the grey epoxycoated steel surface obtained targeting 3 and 15 kPa effective normal stresses, duringthe first 5 cm of shear displacement is shown in Fig. 4.57. The variation of pore-water pressure with shear displacement for the first 5 cm is shown in Fig. 4.58. Thecollected test data over the full displacement range is presented in Fig. 4.60, and thevariation of the average interface shear stress with the effective normal stress duringthe progress of the tests is shown in Fig 4.59.The test conducted targeting σ′n of 15 kPa had to be stopped shortly after reach-ing 30 cm of shear displacement due to a rupture of the rubber membrane of thepneumatic loading system.1714.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 1 2 3 4 50123456Strong-Pit FGS_Grey_3 Strong-Pit FGS_Grey_15Figure 4.57: Variation of interface shear stress with shear displacement for Strong-PitFGS on grey epoxy coated steel.u (kPa)Shear displacement (cm)0 1 2 3 4 505101520Strong-Pit FGS_Grey_3Strong-Pit FGS_Grey_15Figure 4.58: Variation of pore-water pressure with shear displacement for Strong-PitFGS on grey epoxy coated steel.30.0 cm1.02 cm14.1 cm0.65 cm60.0 cm (𝛿 = 17° ~ 18°)25.7 cmFigure 4.59: Variation of interface shear stress with effective normal stress for Strong-Pit FGS on grey epoxy coated steel.1724.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 600246810Strong-Pit FGS_Grey_3 Strong-Pit FGS_Grey_15Test stopped due to rupture in pneumatic rubber membrane(a) Variation of shear stress with shear displacementu (kPa)Shear displacement (cm)0 10 20 30 40 50 6005101520Strong-Pit FGS_Grey_3Strong-Pit FGS_Grey_15(b) Variation of average pore-water pressure with shear displacementσ n' (kPa)Shear displacement (cm)0 10 20 30 40 50 60051015202530Strong-Pit FGS_Grey_3 Strong-Pit FGS_Grey_15(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 6000.20.40.60.81Strong-Pit FGS_Grey_3 Strong-Pit FGS_Grey_15(d) Variation of coefficient of interface friction with shear displacementFigure 4.60: Macro-scale interface direct shear test results for Strong-Pit FGS on greyepoxy coated steel.1734.2. Tests Conducted and ObservationsThe results of macro-scale interface direct shear tests conducted on Strong-PitFGS against the grey epoxy coated steel are summarized in Table 4.17. Reliabledata for the internal friction angle of Strong-Pit FGS at large-displacements and lowconfining stresses was not available at the time of writing and hence the interfaceefficiency factors are not provided in Table 4.17.Table 4.17: Results of macro-scale interface direct shear tests conducted on Strong-PitFGS against the grey epoxy coated steel surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φ′) Ra 50D50/D50Strong-PitFGS Grey 33.3 ±0.050 18 ±0.30 0.33 ±0.005 - 0.16Strong-PitFGS Grey 1514.6 ±0.001 17 ±0.06 0.30 ±0.001 - 0.16At the end of testing, the trailing end of the solid test surface was observed to berelatively clean with only trace amounts of the soil material left on the surface of theplate. This observation was consistent in both tests.Strong-Pit FGS on green epoxy coated steelThe shear stress - displacement response of the Strong-Pit FGS on the green epoxycoated steel surface obtained targeting 3 and 15 kPa effective normal stresses, duringthe first 5 cm of shear displacement is shown in Fig. 4.61. The variation of pore-water pressure with shear displacement for the first 5 cm is shown in Fig. 4.62. Thecollected test data over the full displacement range is presented in Fig. 4.64. Thevariation of the average interface shear stress with the effective normal stress duringthe progress of the tests is shown in Fig. 4.63.The drained large-displacement interface friction angles were observed to be ap-proximately the same for the two tests that were conducted targeting 15 and 30 kPaeffective normal stresses.1744.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 1 2 3 4 50246810Strong-Pit FGS_Green_3Strong-Pit FGS_Green_15Strong-Pit FGS_Green_30Figure 4.61: Variation of interface shear stress with shear displacement for Strong-PitFGS on green epoxy coated steel.u (kPa)Shear displacement (cm)0 1 2 3 4 505101520Strong-Pit FGS_Green_3Strong-Pit FGS_Green_15Strong-Pit FGS_Green_30Figure 4.62: Variation of pore-water pressure with shear displacement for Strong-PitFGS on green epoxy coated steel.60.0 cm0.58 cm60.0 cm0.62 cm 31.8 cm2.49 cm0.45 cm60.0 cm(𝛿 = 13° ~ 26°)32.1 cmFigure 4.63: Variation of interface shear stress with effective normal stress for Strong-Pit FGS on green epoxy coated steel.1754.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 600246810Strong-Pit FGS_Green_3Strong-Pit FGS_Green_15Strong-Pit FGS_Green_30(a) Variation of shear stress with shear displacementu (kPa)Shear displacement (cm)0 10 20 30 40 50 60-5051015Strong-Pit FGS_Green_3Strong-Pit FGS_Green_15Strong-Pit FGS_Green_30(b) Variation of average pore-water pressure with shear displacementσ n' (kPa)Shear displacement (cm)0 10 20 30 40 50 6001020304050Strong-Pit FGS_Green_3Strong-Pit FGS_Green_15Strong-Pit FGS_Green_30(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 6000.20.40.60.81Strong-Pit FGS_Green_3Strong-Pit FGS_Green_15Strong-Pit FGS_Green_30(d) Variation of coefficient of interface friction with shear displacementFigure 4.64: Macro-scale interface direct shear test results for Strong-Pit FGS ongreen epoxy coated steel.1764.2. Tests Conducted and ObservationsThe test targeting 3 kPa effective normal stress showed a distinctively higherdrained large-displacement interface friction angle. This indicates some sensitivityof the large-displacement interface shear strength to the effective normal stress. Itis also notable that the drained large-displacement interface friction angles betweenStrong-Pit FGS and green epoxy were consistently lower compared to those observedfor the Strong-Pit FGS on the sand-blasted mild steel interface.The results of macro-scale interface direct shear test conducted on Strong-PitFGS against the green epoxy coated steel are summarized in Table 4.18. Reliabledata for the internal friction angle of Strong-Pit FGS at large-displacements and lowconfining stresses was not available at the time of writing and hence the interfaceefficiency factors are not provided in Table 4.18.Table 4.18: Results of macro-scale interface direct shear tests conducted on Strong-PitFGS against the green epoxy coated steel surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φ′) Ra 50D50/D50Strong-PitFGS Green 32.6 ±0.05 26 ±0.55 0.48 ±0.009 - 0.18Strong-PitFGS Green 1515.4 ±0.05 14 ±0.05 0.25 ±0.001 - 0.18Strong-PitFGS Green 3025.6 ±0.05 13 ±0.03 0.23 ±0.001 - 0.18At the end of each shear test, the trailing end of the solid test surface was observedto be relatively clean with only trace amounts of the soil material left on the surfaceof the plate.4.2.5 Interface shear characteristics of Redstone on thegreen epoxy coated steel surfaceThe shear stress - displacement response of the Redstone on the green epoxy coatedsteel surface obtained targeting 3 kPa effective normal stress, during the first 5 cmof shear displacement is shown in Fig. 4.65. The variation of pore-water pressure1774.2. Tests Conducted and Observationswith shear displacement for the first 5 cm is shown in Fig. 4.66. The collected testdata over the full displacement range is presented in Fig. 4.68. The variation of theaverage interface shear stress with the effective normal stress during the progress ofthe test is presented in Fig. 4.67.Shear stress (kPa)Shear displacement (cm)0 1 2 3 4 500.20.40.60.81Figure 4.65: Variation of interface shear stress with shear displacement for Redstoneon green epoxy coated steel.u (kPa)Shear displacement (cm)0 1 2 3 4 5012345Figure 4.66: Variation of pore-water pressure with shear displacement for Redstoneon green epoxy coated steel.The shear stress response showed a transient peak of 0.6 kPa (followed by a drop toapproximately 0.35 kPa in shear stress) during the first 1.5 cm of shear displacement.Thereafter, the shear stress showed a gradual continuous increase for the remainderof the test. The continuous increase in the shear stress can be explained by lookingat the effective normal stress that was acting on the interface during the test. Thepore-water pressure, after reaching a peak value of 2.2 kPa at approximately 5 mmof shear displacement, showed a continuous gradual decrease for the remainder of the1784.2. Tests Conducted and Observationstest, which resulted in a corresponding transient drop followed by a continuous rise ineffective normal stress. Based on the pore-water pressure dissipation characteristicsobserved during the test, it is clear that the rate of shear displacement of 0.004 mm/swas higher than optimal for reaching a steady drained condition prior to the endof the test. Despite this complication, the pore-water pressure at 60 cm of sheardisplacement was approximately equal to the static water head level, and the τi− σ′ncombination at the last 5 cm of the test was used to calculate the large-displacementinterface friction angle for this interface. During the last 20 cm of shear displacement,the coefficient of interface friction remained constant. The results of the macro-scaleinterface direct shear test conducted on Redstone against the green epoxy coated steelsurface are summarized in Table 4.19.Table 4.19: Results of macro-scale interface direct shear tests conducted on Redstoneagainst the green epoxy coated steel surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φ′) Ra 50D50/D50Redstone Green 3 2.8 ±0.05 12 ±0.23 0.21 ±0.004 0.43 0.220.14 cm1.68 cm40.8 cm60.0 cm(𝛿 = 12°)Figure 4.67: Variation of interface shear stress with effective normal stress for Red-stone on green epoxy coated steel.1794.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 6000.20.40.60.81(a) Variation of shear stress with shear displacementu (kPa)Shear displacement (cm)0 10 20 30 40 50 6000.511.522.53(b) Variation of average pore-water pressure with shear displacementσ n' (kPa)Shear displacement (cm)0 10 20 30 40 50 60012345(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 6000.10.20.30.40.5(d) Variation of coefficient of interface friction with shear displacementFigure 4.68: Macro-scale interface direct shear test results for Redstone on greenepoxy coated steel.1804.2. Tests Conducted and ObservationsDuring the interface shear test, as the mobile frame is displaced horizontally, thetrailing end of the solid test plate gets exposed incrementally allowing for visualinspection of the test surface at the end of the test. The trailing end of the solidtest surface was observed to be relatively clean with only trace amounts of the soilmaterial left on the surface of the plate.4.2.6 Interface shear characteristics of Kaosand on thesand-blasted mild steel surfaceThe shear stress - displacement response of the Kaosand on the sand-blasted mildsteel surface obtained targeting 3 and 15 kPa effective normal stresses, during thefirst 5 cm of shear displacement is shown in Fig. 4.69.Shear stress (kPa)Shear displacement (cm)0 1 2 3 4 50246810Kaosand_Steel_3 Kaosand_Steel_15Figure 4.69: Variation of interface shear stress with shear displacement for Kaosandon sand-blasted mild steel.u (kPa)Shear displacement (cm)0 1 2 3 4 505101520Kaosand_Steel_3Kaosand_Steel_15Figure 4.70: Variation of pore-water pressure with shear displacement for Kaosandon sand-blasted mild steel.1814.2. Tests Conducted and ObservationsThe variation of pore-water pressure with shear displacement for the first 5 cmis shown in Fig. 4.70. The collected test data over the full displacement range ispresented in Fig. 4.72, and the variation of the average interface shear stress with theeffective normal stress during the progress of the tests is presented in Fig. 4.71. Theresults of the macro-scale interface direct shear test conducted on Kaosand againstthe sand-blasted mild steel are summarized in Table 4.20.Table 4.20: Results of macro-scale interface direct shear tests conducted on Kaosandagainst the sand-blasted mild steel surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φ′) Ra 50D50/D50Kaosand Steel 3 3.5 ±0.050 28 ±0.45 0.53 ±0.008 0.88 0.09Kaosand Steel 15 14.0 ±0.001 28 ±0.11 0.53 ±0.002 0.89 0.0960.0 cm18.4 cm21.7 cm0.95 cm1.42 cm60.0 cm(𝛿 = 28°)Figure 4.71: Variation of interface shear stress with effective normal stress for Kaosandon sand-blasted mild steel.1824.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 6002468101214 Kaosand_Steel_3 Kaosand_Steel_15(a) Variation of shear stress with shear displacementu (kPa)Shear displacement (cm)0 10 20 30 40 50 60051015Kaosand_Steel_3Kaosand_Steel_15(b) Variation of average pore-water pressure with shear displacementσ n' (kPa)Shear displacement (cm)0 10 20 30 40 50 600510152025Kaosand_Steel_3 Kaosand_Steel_15(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 6000.20.40.60.811.21.4 Kaosand_Steel_3 Kaosand_Steel_15(d) Variation of coefficient of interface friction with shear displacementFigure 4.72: Macro-scale interface direct shear test results for Kaosand on sand-blasted mild steel.1834.2. Tests Conducted and ObservationsAt the end of each test, the trailing end of the solid test surface was observed tobe relatively clean with only trace amounts of the soil material left on the surface ofthe plate.4.2.7 Surface roughness observed at the end of testingThe initial surface textures of the untested surfaces were subject to change duringinterface shear testing (either due to corrosion or due to mechanical shearing). Mea-surements of the surface roughness of the test plates were carried out at the end ofthe macro-scale interface direct shear testing program. Laser confocal microscopyimages of the three test surfaces used in the present study prior to and after beingsubject to all the macro-scale interface direct shear tests are shown in Fig. 4.73. Theroughness size distribution curves are presented in Fig. 4.74.0.12 mm 0.12 mm0.25 mm x 0.25 mm0.12 mm0.25 mm x 0.25 mm0.25 mm x 0.25 mmGreen epoxy (New) Grey epoxy (New) Sand-blasted mild steel (New)0.25 mm x 0.19 mm 0.25 mm x 0.19 mm 0.25 mm x 0.19 mmGreen epoxy (Used) Grey epoxy (Used) Sand-blasted mild steel (Used)(a) (b) (c)Figure 4.73: Laser confocal microscopy images of test surfaces before and after beingsubjected to macro-scale interface direct shear testing.1844.2. Tests Conducted and ObservationsPercent finer (%)Grain size (mm), or Roughness size (mm) 10-610-510-410-310-210-1100020406080100 Green epoxyGrey epoxySand-blasted steelKaoliniteGreen epoxy (Used)Grey epoxy (Used)Sand-blasted steel (Used)Ra = 0.42 umRa = 2.00 um Ra = 0.65 um(a) Roughness size based on gauge length of 0.12 mm (Lg = 50×D50−Kaolinite)Figure 4.74: Roughness size distribution curves of the solid test surfaces prior to andafter being subject to macro-scale interface direct shear testing.The mild steel surface was particularly difficult to work with given its propensityto undergo oxidation relatively quickly. During a typical macro-scale interface directshear test with a fine-grained soil, where the consolidation stage can typically takemultiple days to complete, the surface of the steel test plate that lies outside theboundaries of the soil specimen undergoes rapid oxidation. It can be clearly seen thatthe initial textures of the solid surfaces were notably different to that of the testedsurface textures. Laser confocal microscopy imaging of the test surfaces conductedpost-test showed that the average roughness of the green epoxy surface had increasedby 180% to Ra 50D50 = 0.42 µm. The average roughness of the grey epoxy surface hadincreased by 30% to Ra 50D50 = 0.65 µm. The average roughness of the sand-blastedsurfaces on the other hand had decreased by 250% to Ra 50D50 = 2.00 µm. Theneed to reuse the test plates primarily arose from the cost of the plates and budgetconstraints. Given that coarse-grained soils can easily alter the surface texture of agiven test plate during testing, all tests involving coarse-grained soils were conductedat the very last stage of the testing program (i.e., testing commenced with the fine-grained soils and coarse-grained soils were tested at the very end of the test program).No measurements of surface roughness were made in between the macro-scale interface1854.2. Tests Conducted and Observationsdirect shear tests. It was assumed that any major changes in the surface roughnessof the surfaces would occur during testing of Fraser River silt and Fraser River sand.4.2.8 Repeatability test dataAs highlighted earlier, the macro-scale interface direct shear apparatus was designedfor conducting soil-solid interface direct shear tests under relatively low effective nor-mal stress conditions to reliably measure the drained large-displacement interfaceshear strength. The large specimen footprint of the device ensures that the measuredinterface shear force is much larger compared to the friction and other mechanicalresistances associated with the mechanical parts of the device; and the inclusion ofpore-water pressure measurements allows for the accurate confirmation of drainedconditions. Also, combined with the base load cell measurements, the device allowsfor the proper estimation of the effective normal stress that is acting at the interfaceduring shear.In order to gain confidence that the new apparatus is performing within designparameters, a series of repeatability tests were conducted on the grey epoxy andthe green epoxy coated solid surfaces. Table 4.21 lists the summary of the testsconducted together with test parameters. Great care was taken to ensure that thespecimen preparation procedures and the testing procedures were repeated properly.For a given soil type, the same soil specimen was reused for conducting each of therepeatability tests.Two tests were conducted using Fraser River sand on the grey epoxy at 3.0 kPaeffective normal stress to check reapeatability. The results of the tests are shownin Fig. 4.75. Testing fine-grained soils typically pose more challenges compared totesting coarse-grained soils. Two sets of repeatability tests, one set targeting 3.0 kPaeffective normal stress and the other targeting 30.0 kPa effective normal stress, wereconducted using Fraser River silt on the grey epoxy surface (see Figs. 4.76 and 4.77).One of the repeatability tests that was conducted targeting 30.0 kPa effective normalstress had to be stopped prematurely due to a rupture in the pneumatic pressuremembrane that occurred at approximately 16 cm of shear displacement. Despite this1864.2. Tests Conducted and Observationsincident, the repeatability data gathered upto 16 cm of shear displacement was laterfound to be within design parameters, and hence, this test was not redone. Very goodrepeatability of testing was observed for the coefficient of interface friction.Shear displacement (cm)Data point index1000 2000 3000 4000 5000 6000020406080FRS_GREY_3 FRS_GREY_3_RShear stress (kPa)Shear displacement (cm)0 20 40 60 8000.511.522.5FRS_GREY_3 FRS_GREY_3_RFigure 4.75: Repeatability test results for Fraser River sand on grey epoxy.μ iShear displacement (cm)0 10 20 30 40 50 60 7000.20.40.60.81FRSilt_Grey_3_R1 FRSilt_Grey_3_R2Stress (kPa)0123456FRSilt_Grey_3_R1 (Shear stress)FRSilt_Grey_3_R2 (Shear stress)FRSilt_Grey_3_R1 (Effective normal stress)FRSilt_Grey_3_R2 (Effective normal stress)Figure 4.76: Repeatability test results for Fraser River silt on grey epoxy for testsconducted targeting 3.0 kPa effective normal stress.1874.2. Tests Conducted and Observationsμ iShear displacement0 5 10 1500.20.40.60.81FRSilt_Grey_3_R1 FRSilt_Grey_3_R2Stress (kPa)01020304050FRSilt_Grey_30_R1 (Shear stress)FRSilt_Grey_30_R2 (Shear stress)FRSilt_Grey_30_R1 (Effective normal stress)FRSilt_Grey_30_R2 (Effective normal stress)Rupture of pnuematic pressure membrane.Figure 4.77: Repeatability test results for Fraser River silt on grey epoxy for testsconducted targeting 30.0 kPa effective normal stress.Shown in Figs. 4.78 to 4.81 are the repeatability test results for Kaolinite on thegrey and green epoxy coated solid surfaces.μ iShear displacement (cm)0 10 20 30 40 50 6000.20.40.60.81Kao_Green_3_R1 Kao_Green_3_R2Stress (kPa)00.511.522.533.54Kao_Green_3_R1 (Shear stress)Kao_Green_3_R2 (Shear stress)Kao_Green_3_R1 (Effective normal stress)Kao_Green_3_R2 (Effective normal stress)Figure 4.78: Repeatability test results for Kaolinite on green epoxy for tests conductedtargeting 3.0 kPa effective normal stress.1884.2. Tests Conducted and Observationsμ iShear displacement (cm)0 10 20 30 40 50 6000.20.40.60.81Kao_Grey_3_R1 Kao_Grey_3_R2Stress (kPa)012345Kao_Grey_3_R1 (Shear stress)Kao_Grey_3_R2 (Shear stress)Kao_Grey_3_R1 (Effective normal stress)Kao_Grey_3_R2 (Effective normal stress)Figure 4.79: Repeatability test results for Kaolinite on grey epoxy for tests conductedtargeting 3.0 kPa effective normal stress.μ iShear displacement (cm)0 10 20 30 40 50 6000.20.40.60.81Kao_Grey_15_R1 Kao_Grey_15_R2Stress (kPa)0510152025Kao_Grey_15_R1 (Shear stress)Kao_Grey_15_R2 (Shear stress)Kao_Grey_15_R1 (Effective normal stress)Kao_Grey_15_R2 (Effective normal stress)Figure 4.80: Repeatability test results for Kaolinite on grey epoxy for tests conductedtargeting 15.0 kPa effective normal stress.1894.2. Tests Conducted and Observationsμ iShear displacement (cm)0 10 20 30 40 50 6000.20.40.60.81Kao_Grey_30_R1 Kao_Grey_30_R2Stress (kPa)01020304050Kao_Grey_30_R1 (Shear stress)Kao_Grey_30_R2 (Shear stress)Kao_Grey_30_R1 (Effective normal stress)Kao_Grey_30_R2 (Effective normal stress)Figure 4.81: Repeatability test results for Kaolinite on grey epoxy for tests conductedtargeting 30.0 kPa effective normal stress.Macro-scale interface direct shear tests involving relatively slow draining fine-grained soils such as Kaolinite typically take two to three weeks to complete dueto the consolidation time and the slow shear displacement rates required to attaindrained conditions. This makes such tests exceptionally sensitive to variations inenvironmental factors such as ambient temperature. Nevertheless, the tests showedvery good repeatability. Also, the results of the test series on Kaolinite showed thatthe device captures the sensitivity of the drained large-displacement interface shearstrength to the effective normal stress level. The repeatability tests data confirmedthat both, fine-grained soils and coarse-grained soils, can be tested within designparameters of the new apparatus with exceptionally good repeatability performance.1904.2. Tests Conducted and ObservationsTable 4.21: Results of macro-scale interface direct shear repeatability tests.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φ′) Ra 50D50/D50FRS Grey 3 4.0 ±0.01 27 ±0.10 0.52 ±0.002 0.78 0.006FRS Grey 3 R 4.0 ±0.01 27 ±0.10 0.52 ±0.002 0.78 0.006FRSilt Grey 3 R1 3.4 ±0.05 25 ±0.43 0.46 ±0.007 0.72 0.06FRSilt Grey 3 R2 3.9 ±0.05 25 ±0.43 0.46 ±0.007 0.72 0.06FRSilt Grey 30 R1 27.4 ±0.05 27 ±0.06 0.51 ±0.001 0.78 0.06FRSilt Grey 30 R2 25.3 ±0.05 27 ±0.06 0.51 ±0.001 0.78 0.06Kao Green 3 R1 2.3 ±0.05 17 ±0.42 0.31 ±0.007 0.65 0.07Kao Green 3 R2 2.1 ±0.05 17 ±0.42 0.31 ±0.007 0.65 0.07Kao Grey 3 R1 3.1 ±0.05 19 ±0.42 0.34 ±0.007 0.73 0.2Kao Grey 3 R2 3.0 ±0.05 19 ±0.42 0.35 ±0.007 0.73 0.2Kao Grey 15 R1 16.2 ±0.05 16 ±0.09 0.28 ±0.002 0.61 0.2Kao Grey 15 R2 12.5 ±0.05 16 ±0.09 0.28 ±0.002 0.61 0.2Kao Grey 30 R1 37.1 ±0.05 16 ±0.04 0.28 ±0.001 0.62 0.2Kao Grey 30 R2 36.6 ±0.05 16 ±0.04 0.28 ±0.001 0.62 0.24.2.9 Displacement rate effects on test dataA limited number of tests were conducted using the fine-grained soils to investigatethe degree of sensitivity of the drained large-displacement interface shear strength onthe shear displacement rate. Fraser River silt and Kaolinite clay were tested on thegrey epoxy coated solid surface at several displacement rates. For a given test soiltype, the initial specimen thicknesses, and the drainage boundary conditions of alldisplacement rate effect tests were kept the same (i.e., single drainage upward intothe surcharge sand layer), and the rate of shearing was the only parameter that waschanged from one test to the other. The results of the rate effect tests are presentedin Figs. 4.82 to 4.87.Fraser River silt was tested on the grey epoxy coated solid surface targeting σ′nof 3.0 kPa at rates of 0.01, 0.05, 0.07, and 0.14 mm/s. One repeatability test wasconducted for Fraser River silt at a displacement rate of 0.14 mm/s targeting σ′n of1914.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 60 7000.511.522.53FRSilt_Grey_3_F0.01FRSilt_Grey_3_F0.05FRSilt_Grey_3_F0.07FRSilt_Grey_3_F0.14FRSilt_Grey_3_F0.14_R(a) Variation of shear stress with shear displacementAvg. pore-water pressure (kPa)Shear displacement (cm)0 10 20 30 40 50 60 70012345FRSilt_Grey_3_F0.01FRSilt_Grey_3_F0.05FRSilt_Grey_3_F0.07FRSilt_Grey_3_F0.14FRSilt_Grey_3_F0.14_R(b) Variation of average pore-water pressure with shear displacementEffective normal stress (kPa)Shear displacement (cm)0 10 20 30 40 50 60 7001234567FRSilt_Grey_3_F0.01FRSilt_Grey_3_F0.05FRSilt_Grey_3_F0.07FRSilt_Grey_3_F0.14FRSilt_Grey_3_F0.14_R(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 60 7000.20.40.60.81FRSilt_Grey_3_F0.01FRSilt_Grey_3_F0.05FRSilt_Grey_3_F0.07FRSIlt_Grey_3_F0.14FRSilt_Grey_3_F0.14_R(d) Variation of coefficient of interface friction with shear displacementFigure 4.82: Macro-scale interface direct shear test results for Fraser River silt ongrey epoxy conducted to study rate effects. Target σ′n = 3.0 kPa. 1924.2. Tests Conducted and Observations3.0 kPa to confirm that the results obtained at relatively high displacement rates arerepeatable. In addition to the tests conducted targeting σ′n of 3.0 kPa, two more testswere conducted targeting σ′n of 30.0 kPa at 0.01 and 0.14 mm/s rates to investigatewhether the applied effective normal stress level has an influence on the sensitivity ofthe large-displacement interface shear strength to the displacement rate.The data for interface direct shear testing of Fraser River silt on the grey epoxycoated solid surface conducted for investigating the effect of displacement rate target-ing σ′n of 3.0 kPa is shown in Fig. 4.82. Shown in Fig. 4.83 are the pore-water pressuretime histories for the same set of tests. The pore-water pressure at the interface wasobserved to rise rapidly at the onset of shearing, and then gradually dissipate towardsthe static water head as shearing continued. The effective normal stress essentiallymirrored the pore-water pressure change. The pore-water pressure dissipation datashown in Figs. 4.82-(b) and 4.83 indicate that the tests conducted at displacementrates of 0.07 and 0.14 mm/s did not have sufficient time for the complete dissipationof the excess pore-water pressure to the static water head.Avg. pore-water pressure (kPa)Time (s)0 5000 10000 15000 20000 25000 30000 3500000.511.522.533.54FRSilt_Grey_3_F0.01FRSilt_Grey_3_F0.05FRSilt_Grey_3_F0.07FRSilt_Grey_3_F0.14FRSilt_Grey_3_F0.14_RFigure 4.83: Pore-water pressure dissipation time history of rate effect tests conductedfor Fraser River silt on grey epoxy targeting σ′n = 3.0 kPa.For the relatively fast displacement rates, the coefficient of interface friction atlarge-displacement was observed to be slightly less than, and was approaching, thevalue observed for the test conducted at the slower displacement rate of 0.01 mm/s.The data for interface direct shear testing of Fraser River silt on the grey epoxycoated solid surface conducted for investigating the effect of displacement rate tar-1934.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 60 7005101520FRSilt_Grey_30_F0.01 FRSilt_Grey_30_F0.14(a) Variation of shear stress with shear displacementAvg. pore-water pressure (kPa)Shear displacement (cm)0 10 20 30 40 50 60 7005101520FRSilt_Grey_30_F0.01 FRSilt_Grey_30_F0.14(b) Variation of average pore-water pressure with shear displacementEffective normal stress (kPa)Shear displacement (cm)0 10 20 30 40 50 60 700510152025303540FRSilt_Grey_30_F0.01 FRSilt_Grey_30_F0.14(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 60 7000.20.40.60.81FRSilt_Grey_30_F0.01 FRSilt_Grey_30_F0.14(d) Variation of coefficient of interface friction with shear displacementFigure 4.84: Macro-scale interface direct shear test results for Fraser River silt ongrey epoxy conducted to study rate effects. Target σ′n = 30.0 kPa. 1944.2. Tests Conducted and Observationsgeting σ′n of 30.0 kPa is shown in Fig. 4.84. The pore-water pressure time historiesobtained for the two tests are shown in Fig. 4.85. Note that for the same initialthickness of the test soil specimen, the resulting final specimen thickness after consol-idation under σ′n of 30.0 kPa is less than that at σ′n of 3.0 kPa. Therefore, for the testsconducted targeting σ′n of 30.0 kPa, the drainage path was much shorter, and thisallowed for much faster dissipation of excess pore-water pressure. Note that, duringthe test conducted at the displacement rate of 0.14 mm/s, the pore-water pressurewas observed to dissipate to almost zero. A gradual recovery of the pore-water pres-sure to the static water head level was observed towards the end of the test. Thecoefficient of interface friction at large displacement was found to be insensitive tothe rate of shear displacement.Avg. pore-water pressure (kPa)Time (s)0 2000 4000 6000 8000 1000005101520FRSilt_Grey_30_F0.01 FRSilt_Grey_30_F0.14Figure 4.85: Pore-water pressure dissipation time history of rate effect tests conductedfor Fraser River silt on grey epoxy targeting σ′n = 30.0 kPa.Based on the interface shear test data for Fraser River silt on the grey epoxycoated solid surface conducted at different shear displacement rates, it is possible todraw a preliminary conclusion that the drained large-displacement interface frictionangle of Fraser River silt is not sensitive to the shear displacement rate at the loweffective normal stress range of 3.0 to 30.0 kPa.A limited number of macro-scale interface direct shear tests were conducted usingKaolinite on the grey epoxy coated solid surface at displacement rates of 0.002 and0.04 mm/s targeting 3.0 and 15.0 kPa effective normal stress levels to investigatethe effect of displacement rate on the large-displacement interface shear strength.1954.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 60 7000.511.52Kao_Grey_3_F0.002 Kao_Grey_3_F0.04(a) Variation of shear stress with shear displacementAvg. pore-water pressure (kPa)Shear displacement (cm)0 10 20 30 40 50 60 7000.511.522.533.54Kao_Grey_3_F0.002 Kao_Grey_3_F0.04(b) Variation of average pore-water pressure with shear displacementEffective normal stress (kPa)Shear displacement (cm)0 10 20 30 40 50 60 70012345Kao_Grey_3_F0.002 Kao_Grey_3_F0.04(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 60 7000.20.40.60.81Kao_Grey_3_F0.002 Kao_Grey_3_F0.04(d) Variation of coefficient of interface friction with shear displacementFigure 4.86: Macro-scale interface direct shear test results for Kaolinite on grey epoxyconducted to study rate effects. Target σ′n = 3.0 kPa.1964.2. Tests Conducted and ObservationsThe test results are shown in Figs. 4.86 and 4.87. During the test conducted atthe faster displacement rate of 0.02 mm/s targeting σ′n of 3.0 kPa, the pore-waterpressure did not have sufficient time to fully dissipate to the static water head level.The test conducted at the faster displacement rate was observed to show a lowercoefficient of interface friction at large displacement compared to that observed forthe test conducted at the slower rate. This seems to be an effect of displacement rate.However, the results are not conclusive because the large-displacement interface shearstrength for the Kaolinite - grey epoxy interface is sensitive to the effective normalstress level at relatively low confining stresses, and the effective normal stresses of thetwo tests were not identical. The macro-scale interface direct shear tests conductedwith Kaolinite on the grey epoxy coated solid surface targeting σ′n of 15.0 kPa atthe displacement rates of 0.002 and 0.04 mm/s, were observed to show the samelarge-displacement coefficient of interface friction.1974.2. Tests Conducted and ObservationsShear stress (kPa)Shear displacement (cm)0 10 20 30 40 50 60 700123456Kao_Grey_15_F0.002 Kao_Grey_15_F0.02(a) Variation of shear stress with shear displacementAvg. pore-water pressure (kPa)Shear displacement (cm)0 10 20 30 40 50 60 70024681012Kao_Grey_15_F0.002 Kao_Grey_15_F0.02(b) Variation of average pore-water pressure with shear displacementEffective normal stress (kPa)Shear displacement (cm)0 10 20 30 40 50 60 70051015202530Kao_Grey_15_F0.002 Kao_Grey_15_F0.02(c) Variation of effective normal stress with shear displacementμ iShear displacement (cm)0 10 20 30 40 50 60 7000.20.40.60.81Kao_Grey_15_F0.002 Kao_Grey_15_F0.02(d) Variation of coefficient of interface friction with shear displacementFigure 4.87: Macro-scale interface direct shear test results for Kaolinite on grey epoxyconducted to study rate effects. Target σ′n = 15.0 kPa.1984.3. Summary of the Chapter, Key Findings, and ContributionsThe results of the displacement rate effect tests are summarized in Table 4.22.Table 4.22: Results of macro-scale interface direct shear rate effect tests conductedon the grey epoxy surface.Test ID σ′n (kPa) δ (deg.) µitan(δ)tan(φ′) Rate(mm/s)FRSilt Grey 3 F0.01 3.4 ±0.05 24 0.45±0.0100.68 0.010FRSilt Grey 3 F0.05 3.6 ±0.05 24 0.44±0.0100.68 0.050FRSilt Grey 3 F0.07 3.9 ±0.05 23 0.42±0.0100.65 0.070FRSilt Grey 3 F0.14 3.0 ±0.05 24 0.44±0.0100.68 0.140FRSilt Grey 3 F0.14 R 3.0 ±0.05 22 0.40±0.0100.62 0.140FRSilt Grey 30 F0.01 26.6 ±0.05 27 0.50±0.0010.78 0.010FRSilt Grey 30 F0.14 29.0 ±0.05 26 0.48±0.0010.75 0.140Kao Grey 3 F0.002 2.3 ±0.05 22 0.41±0.0100.86 0.002Kao Grey 3 F0.04 2.6 ±0.05 17 0.30±0.0100.65 0.040Kao Grey 15 F0.002 16.4 ±0.05 16 0.28±0.0040.61 0.002Kao Grey 15 F0.04 13.1 ±0.05 16 0.28±0.0040.61 0.0404.3 Summary of the Chapter, Key Findings, andContributionsThis chapter presented, in detail, the macro-scale interface direct shear tests con-ducted to study the effect of soil type, confining stress level, and interface roughnesson the drained large-displacement interface friction angle. One coarse-grained soiland five fine-grained soils were tested utilizing two epoxy coated steel plates and one1994.3. Summary of the Chapter, Key Findings, and Contributionssand blasted mild steel plate representing various types of pipeline surfaces of differentroughness and texture. Tests were conducted targeting three effective normal stresslevels of 3, 15, and 30 kPa. Test results were presented and important observationswere highlighted. This chapter also presented the results of a number of repeatabilitytests that were used to gain a better understanding of the degree of variability of themacro-scale interface direct shear tests presented earlier in the chapter. The chapterconcludes with a presentation of results of a limited number of rate effect tests thatwere conducted using the macro-scale interface direct shear apparatus.Based on the work carried out, the following observations and contributions canbe highlighted:1. Based on the literature review presented in Chapter 2, it was identified thatthere is a lack of data available related to the study of the drained large-displacement soil-solid interface friction angle at low confining stress levels ap-plicable to pipeline design. It was highlighted that conventional geotechnicaltest apparatus were not suitable for conducting interface shear testing underthese conditions without modification. The work undertaken in the presentthesis resulted in the development of a novel macro-scale interface direct shearapparatus that addresses many of the unique challenges identified in the lit-erature review. The new apparatus was used to conduct a range of soil-solidinterface direct shear tests producing data that helps to fill the data gap iden-tified in the literature review. By no means does the data completely fill theknowledge gap, but it certainly is sufficient to shed light on the topic underconsideration.2. All tests conducted utilizing the coarse-grained soil on the two of the threesolid surfaces (the green epoxy surface and the sand-blasted mild steel surface)showed linear large-displacement interface shear strength envelopes within the3 to 30 kPa confining stress range. A mild sensitivity of the large-displacementinterface friction angle was observed on the grey epoxy surface where the in-terface friction angle was observed to increase slightly with increasing confining2004.3. Summary of the Chapter, Key Findings, and Contributionsstress level. On all three test surfaces, the large-displacements interface frictionangles were observed to be less than the large-displacement internal frictionangle of the coarse-grained soil with interface efficiency factors in the range of0.65 to 0.82.3. The drained large-displacement interface friction angle of fine-grained soils onthe three solid surfaces were found to be highly dependent on the confiningstress level in the 3 to 30 kPa range. The test data emphasizes the impor-tance of characterizing the soil-pipe interface shear shear strength parametersat confining stress conditions representative of conditions expected in practice.4. When tested on the epoxy coated surfaces, the fine-grained soils were observedto show drained large-displacement interface friction angles smaller than thedrained large-displacement internal friction angles of the soils where interfaceefficiency factors ranged between 0.43 and 0.85. When tested on the sand-blasted mild steel surface, the kaolinite showed drained large-displacement in-terface friction angles equal to the drained large-displacement interface frictionangle of the soil except at the 30 kPa confining stress where a slightly lowerinterface friction angle was observed. These observations show the importanceof considering the effect of surface roughness and texture of the pipeline indetermining soil-spring properties.A detailed discussion of the test results of the present study is presented in thenext chapter.201Chapter 5Discussion on Macro-ScaleInterface Direct Shear Test ResultsThe results obtained from the macro-scale interface direct shear tests provided the op-portunity to add new information to the current body of knowledge, and to comparethe new results with past findings. The next three sections discuss the experimen-tal findings of the present study relative to past work and highlight some findingsthat could contribute to improving current practice and guidelines. In particular,the stress-dependent non-linearity of the interface shear strength envelopes observedbased on data from the macro-scale interface direct shear tests is discussed. Theeffect of surface roughness of the solid surface on the interface shear strength is alsoaddressed.5.1 Influence of the Effective Normal Stress onthe Soil-Solid Interface Shear Strength5.1.1 Solid interfaces with Fraser River sand and FraserRiver siltThe large-displacement drained interface shear strength envelopes obtained for FraserRiver sand as well as Fraser River silt against the three tested solid surfaces are shownin Figs. 5.1 and 5.2. The interface strength envelopes for Fraser River sand arelinear, indicating that the coefficient of interface friction (or interface friction angle(δ) obtained at large-displacements against all the three solid surfaces is not sensitiveto the magnitude of the effective normal stress level within the 3.0 to 30.0 kPa range.2025.1. Influence of the Effective Normal Stress on the Soil-Solid Interface Shear Strength024681012141618200 5 10 15 20 25 30 35 40Shear stress (kPa)Effective normal stress (kPa)FRS_Steel FRS_Green FRS_Grey𝛿 = 28.6°𝜇𝑖 = 0.54𝛿 = 25.6°𝜇𝑖 = 0.48𝛿 = 27.0°𝜇𝑖 = 0.51Figure 5.1: Drained large-displacement interface shear strength envelopes for FraserRiver sand.0246810121416180 5 10 15 20 25 30 35Shear stress (kPa)Effective normal stress (kPa)FRSilt_Steel FRSilt_Green FRSilt_Grey𝐹𝑟𝑎𝑠𝑒𝑟 𝑅𝑖𝑣𝑒𝑟 𝑠𝑖𝑙𝑡𝑅𝑖𝑛𝑔−𝑠ℎ𝑒𝑎𝑟 𝑡𝑒𝑠𝑡 𝑑𝑎𝑡𝑎𝛿 = 28.8°𝜇𝑖 = 0.55𝛿 = 27.5°𝜇𝑖 = 0.52𝛿 = 26.6°𝜇𝑖 = 0.50Figure 5.2: Drained large-displacement interface shear strength envelopes for FraserRiver silt.The large-displacement drained interface shear strength envelope obtained forFraser River silt against the green solid surface, was also found to be not sensitiveto the magnitude of the effective normal stress within the 3.0 to 30.0 kPa range. In2035.1. Influence of the Effective Normal Stress on the Soil-Solid Interface Shear Strengthcontrast, the Fraser River silt - sand-blasted mild steel and the Fraser River silt -grey epoxy interfaces showed some minor sensitivity to the magnitude of the effectivenormal stress. The coefficient of interface friction obtained at large-displacementfailure conditions for both of these interfaces was observed to increase slightly (anincrease of 10%) with the increase of the effective normal stress level from 3.0 kPato 30.0 kPa. This effect can be more clearly seen in Fig. 5.3, where the variation ofthe large-displacement coefficient of interface friction with the effective normal stresslevel is shown for Fraser River silt on the two solid surfaces.0.400.450.500.550.600.650 5 10 15 20 25 30Coefficient of interface frictionEffective normal stress (kPa)FRSilt_Steel FRSilt_Grey FRSilt_Grey Repeat TestsDashed lines indicate the possible variability in the test results.Figure 5.3: Variation of the drained large-displacement coefficient of interface frictionwith effective normal stress level for Fraser River silt on the mild steel and grey epoxyinterfaces.The interface efficiency factors f = tan(δ)/tan(φ′) of the sand and silt on thethree surfaces obtained from these results are presented in Table 5.2. All interfaces,for Fraser River sand as well as Fraser River silt, were observed to show interfaceefficiency factors between 0.7 and 0.9. The interface efficiency factors recommendedby the American Lifeline Alliance (ALA) (ALA, 2001), and the Pipeline ResearchCouncil International (PRCI) (PRCI, 2004, 2009) guidelines are presented in Table2045.1. Influence of the Effective Normal Stress on the Soil-Solid Interface Shear Strength5.1 for comparison.Table 5.1: Interface efficiency factors recommended by the American Lifeline Alliance(ALA) (ALA, 2001), and the Pipeline Research Council International (PRCI) (PRCI,2004, 2009) guidelinesPipe coating f = δ/φm*Concrete 1.0Coal tar 0.9Rough steel 0.8Smooth steel 0.7Fusion bonded epoxy 0.6Polyethylene 0.6* φm is defined as the maximum internal friction angle of the soil.Table 5.2: Drained large-displacement interface efficiency factors for Fraser Riversand and Fraser River silt obtained at low effective normal stresses.Interface σ′n (kPa) δ (deg.) δ/φ′ tan(δ)tan(φ′) µi = tan(δ)FRSand-Steel 3.0 to 35.0 28.6 ±0.10 0.87 0.84 0.54 ±0.002FRSand-Grey 3.0 to 35.0 25.6 ±0.10 0.77 0.74 0.48 ±0.002FRSand-Green 3.0 to 30.0 27.0 ±0.10 0.82 0.78 0.51 ±0.002FRSilt-Steel 3.0 to 30.0 28.8 ±0.43 0.87 0.85 0.55 ±0.007FRSilt-Grey 3.0 to 25.0 27.5 ±0.43 0.82 0.78 0.52 ±0.007FRSilt-Green 3.0 to 30.0 26.6 ±0.43 0.82 0.78 0.50 ±0.007For Fraser River sand and Fraser River silt φ′ = φcv = 33°. The ± errors shown have beencalculated based on low confining stress conditions of σ′n < 5kPa.In the PRCI (PRCI, 2004, 2009) guidelines, there is no explicit guidance givenwith regard to the method to use for estimating the internal friction angle of thesoil applicable to soil-pipe interaction design. In Table 5.1, as given in the PRCIguidelines, the internal friction angle of the soil used in the interface efficiency factorshas been specified as φm “maximum internal friction angle of the soil”. In the ALA2055.1. Influence of the Effective Normal Stress on the Soil-Solid Interface Shear Strength(ALA, 2001) guidelines, the interface efficiency factors have been specified in termsof the “internal friction angle of the soil”. There is no explicit guidance as to whetherthe interface efficiency factors are based on the peak or critical state friction angle.As discussed earlier in the thesis, the peak internal friction angle of a given soilwill be highly dependent on many factors such as the method used to measure andinterpret the friction angle (such as, triaxial versus direct shear versus ring-shear)and initial conditions of the soil specimen. Overconsolidated fine-grained soils ordense coarse-grained soils under relatively low confining stresses will exhibit a peakfriction angle at small shear strains brought about by dilation. The value of the peakfriction angle will highly depend on the initial conditions of the soil specimen andthe stress path imposed. The same soil material tested using different modes of shearunder different initial conditions can show different peak friction angles. Therefore, ifthe design requires the use of the peak friction angle of the soil material in question,special consideration of the ability to replicate the in-situ conditions in the laboratorywill be required. Specifying the interface efficiency factors as a function of the peakfriction angle of the soil is prone to much uncertainty unless the conditions underwhich the interface efficiency factors were developed are well defined. The criticalstate friction angle of a given soil is currently understood to be a unique propertyof the soil material. Hence, expressing the interface efficiency factors in terms of thecritical state friction angle may be a more suitable approach.It is of interest to examine past empirical data on the large-displacement shearstrength of sand-steel interfaces, and shown in Fig. 5.4 is interface efficiency factor(tan(δ)/tan(φ′)) data originating from a number of past studies that have looked atthe effect of confining stress level on the large-displacement interface shear strength ofvarious sand materials on steel surfaces of different roughness. The φ′ values in thesestudies were taken at large displacements. However, in certain cases, especially wherethe direct shear device has been used at low confining stresses, the φ′ values would beaffected by dilation. It is noteworthy that past data on the interface shear strengthof sand-steel interfaces at relatively low effective normal stresses is scarce owing tothe inherent challenges of conducting interface shear testing at low confining stresses2065.1. Influence of the Effective Normal Stress on the Soil-Solid Interface Shear Strengthusing conventional geotechnical testing apparatus (Bruton et al., 2007; Lehane andLiu, 2013; Najjar et al., 2003; White et al., 2012).00.20.40.60.811.20 100 200 300 400 500 600 700 800Interface efficiency factor at large displacementstan(δ) / tan(φ’)Effective normal stress (kPa)Figure 5.4: Large-displacement interface shear strength data obtained from past stud-ies on various sand materials tested on steel surfaces of different roughness (Ho et al.,2011; Jardine et al., 1993; Karimian, 2006; Reddy et al., 2000; Rinne, 1989; UESUGIet al., 1989).In general, the data shown in Fig. 5.4 corresponding to confining stress levels of 50to 750 kPa indicates that for a given sand type, the large-displacement interface shearstrength of the sand-steel interface can be considered to be more or less insensitiveto the effective normal stress level. The macro-scale interface direct shear test data2075.1. Influence of the Effective Normal Stress on the Soil-Solid Interface Shear Strengthindicates that this observation also holds true for low confining stresses of 3 to 30 kPafor Fraser River sand on the sand-blasted mild steel surface.A limited number of studies have shown that the large-displacement interface shearstrength of sand on steel can increase with the increase in effective normal stress level(Rinne, 1989; UESUGI et al., 1989). This increase of the interface shear strengthwith the effective normal stress level has been associated with sand particle crushingat large shear displacements (UESUGI et al., 1989). However, the limited data cur-rently available is not conclusive in this regard, and further research is warranted toinvestigate this effect further.5.1.2 Kaolinite interfacesThe drained interface shear strength envelopes at large-displacement failure condi-tions obtained from tests conducted for Kaolinite on the three solid surfaces areshown in Fig. 5.5-(a). Figure 5.5-(b) shows the same shear strength envelopes withinthe effective normal stress range of 0 to 5 kPa. Shown in Fig. 5.6 is the variationof the derived large-displacement interface friction angles with effective normal stressfor the same data. The Kaolinite-steel interface showed sensitivity to σ′n throughoutthe 3.0 to 30.0 kPa range. It can be seen that the failure envelope shows a slightcurvature indicating the sensitivity of the interface friction coefficient to σ′n.Moreover, the Kaolinite interface shear strength envelope was observed to matchthe drained large-displacement internal friction angle of Kaolinite within the σ′n rangeof 3.0 to 15.0 kPa. The drained large-displacement internal friction angle of Kaoliniteis based on drained ring-shear tests conducted by Eid et al. (2014). The decrease inthe measured large-displacement friction angle with increasing confining stress levelmay be attributed to the effects of dilation and particle rearrangement, where thesoil would show a greater tendency to dilate at lower confining stresses. Whether ornot this stress dependent change in the large-displacement friction angle is due todilation is difficult to ascertain without knowing the volume change state at the shearinterface.2085.1. Influence of the Effective Normal Stress on the Soil-Solid Interface Shear Strength0246810120 5 10 15 20 25 30 35 40Shear stress (kPa)Effective normal stress (kPa)Kao_SteelKao_GreenKao_GreyKaolinite (Ring-shear)(a) Interface shear strength envelopes for Kaolinite on the three solid surfaces00.20.40.60.811.21.41.61.820 1 2 3 4 5Shear stress (kPa)Effective normal stress (kPa)Kao_SteelKao_GreenKao_GreyDashed lines indicate the possible variability in the test results.(b) Extrapolated interfaces shear strength envelopes within the effective normalstress range of 0 to 5 kPa for Kaolinite on the three solid surfacesFigure 5.5: Drained large-displacement interface shear strength envelopes for Kaolin-ite on the three solid surfaces based on macro-scale interface direct shear data at loweffective normal stresses.2095.1. Influence of the Effective Normal Stress on the Soil-Solid Interface Shear Strength0510152025300 5 10 15 20 25 30 35 40Secant drained interface friction angle at large-displacement (deg.)Effective normal stress (kPa)Kao_Grey Kao_Green Kao_Steel𝐾𝑎𝑜𝑙𝑖𝑛𝑖𝑡𝑒𝑅𝑖𝑛𝑔−𝑠ℎ𝑒𝑎𝑟 𝑡𝑒𝑠𝑡 𝑑𝑎𝑡𝑎Dashed lines indicate the possible variability in the test results.Figure 5.6: Variation of the large-displacement drained secant interface friction angleδ with effective normal stress for Kaolinite on the three solid surfaces.The amount of horizontal displacement achieved in the macro-scale interface directshear apparatus is approximately 1200% with respect to the overall thickness of thesoil specimen. Under this large-displacement condition, while it is likely that theshear strain at the soil-solid interface is sufficiently large to produce critical stateconditions, whether or not there is any volume change at the large-displacements isnot certain.It is of further relevance to note that under all three confining stress levels theKaolinite-steel tests had noticeable smearing of the soil on the steel surface. In themacro-scale interface direct shear apparatus the soil specimen is continuously andprogressively exposed to a fresh solid surface at the leading edge of the specimen (i.e.,at the north edge of the specimen). The progressive exposure of the soil specimento a fresh solid surface is likely to induce transient volume change conditions even atlarge-displacements as the shearing zone transitions from soil-soil internal shearingto soil-solid interface shearing. This transient effect would likely not exist in the2105.1. Influence of the Effective Normal Stress on the Soil-Solid Interface Shear Strengthring-shear test. It would be of interest to see whether relatively large scale interfacering-shear tests would produce different observations. The small-scale ring-shear testdata for Kaolinite also shows a reduction in the large-displacement friction angle ofKaolinite with increasing σ′n and the trends observed in both the macro-scale interfacedirect shear and small-scale ring-shear apparatus are remarkably similar. Given thatthe variation of the large-displacement friction angles obtained from both the macro-scale interface direct shear apparatus and the ring-shear apparatus are very similar,it is likely that in both devices at large displacements either the volume changeconditions were similar or there was no volume change and critical state conditionswere in play. The pore-water pressure was observed to settle at the static water headlevel at large displacements. Any dilation induced transient reduction of the pore-water pressure would have sufficient time to dissipate under the displacement ratesused. Whether there was a tendency of the pore-water pressure to drop below thestatic water head during shearing was not possible to determine with certainty as thetests were conducted at very low displacement rates. Nevertheless, as will be shownin Chapter 6, full-scale soil-pipe interaction tests conducted in Kaolinite as part ofthis research project showed that the effect of confining stress level on the interfacefriction angle at large displacements directly affects the measured axial soil resistanceon the pipe.The sensitivity of the large-displacement interface friction angle to σ′n for theKaolinite on the green and grey epoxy surfaces was observed to be much more pro-nounced at σ′n values below 15.0 kPa. There was little to no smearing of the soil on thetwo epoxy surfaces at all three confining stress levels. The drained large-displacementinterface shear strength envelope appears to show a minor inflection just below the3.0 kPa effective normal stress level for the grey epoxy surface as can be seen in Fig.5.5-(b). It is unclear as to what causes this shift to occur, and it is unlikely that thiscould be attributed to experimentation error given the stringent quality control mea-sures taken during the testing program. Nevertheless, further testing is required toinvestigate and substantiate this observation. The results of the macro-scale interfacedirect shear tests of Kaolinite on the three surfaces are summarized in Table 5.3.2115.1. Influence of the Effective Normal Stress on the Soil-Solid Interface Shear StrengthTable 5.3: Drained large-displacement interface efficiency factors for Kaolinite ob-tained at low effective normal stresses.Interface σ′n (kPa) δ (deg.) φ′ (deg.) tan(δ)tan(φ′)Kao-Steel 3.60 ±0.05 26.6* ±0.42 25.2 1.00Kao-Steel 14.6 ±0.05 24.7 ±0.09 24.9 0.99Kao-Steel 27.0 ±0.05 22.8 ±0.04 24.2 0.93Kao-Grey 3.10 ±0.05 20.3 ±0.34 25.2 0.79Kao-Grey 16.2 ±0.05 15.1 ±0.05 24.8 0.58Kao-Grey 25.0 ±0.05 15.1 ±0.02 24.7 0.59Kao-Green 2.45 ±0.05 18.8 ±0.37 26.1 0.69Kao-Green 14.2 ±0.05 14.0 ±0.05 25.2 0.53Kao-Green 24.2 ±0.05 14.0 ±0.03 24.7 0.54* Note: δ > φ′ is likely due to measurement error and is treated as δ = φ′.Past work on characterizing the large-displacement drained interface shear strengthof fine-grained soils at low effective normal stresses is very limited. The variation ofthe large-displacement interface efficiency factor f = tan(δ)/ tan(φ′) with the effec-tive normal stress obtained from the results of the present study are compared withthat of several past studies of significance in Fig. 5.7.The f − σ′n results of the present study are in reasonable agreement with whatis reported in past studies. The large-displacement interface efficiency factor can beseen to be notably sensitive to the effective normal stress level below 10 kPa. Basedon the results of the present study and those of past studies, it seems that in general,the f − σ′n trend lines tend to have a concave shape within the σ′n range of 1 to 30kPa. The exception to this are the Kaolinite - steel (present study) and Kaolinite- acrylic (Pedersen et al., 2003) interfaces which show convex shaped f − σ′n trendlines.2125.1. Influence of the Effective Normal Stress on the Soil-Solid Interface Shear Strength00.20.40.60.811.20 5 10 15 20 25 30 35 40 45Interface efficiency factor ( tan(δ) / tan(φ’) )Effective normal stress (kPa)ccccFigure 5.7: Variation of the large-displacement drained interface efficiency factor witheffective normal stress obtained from past studies for a number of fine-grained soilstested on solid surfaces of various roughness (Kuo et al., 2012; Meyer et al., 2015;Najjar et al., 2007a; Pedersen et al., 2003).The interface efficiency factor can be seen to reduce by approximately 25% as theeffective normal stress is increased from approximately 1 kPa to 15 kPa for almost allthe interfaces shown in Fig. 5.7. The exception to this is the Strong-pit FGS - greenepoxy interface (present study) which showed a reduction in f of approximately 50%as σ′n is increased from 1 kPa to 15 kPa.It is of importance to note that in addition to the effect of effective normal stress,the solid surface material and its roughness also affects the large-displacement in-2135.2. Effect of Roughness of the Test Surfaces on the Interface Shear Strengthterface shear strength. The next section presents a discussion of the findings of thepresent study with reference to the effect of surface roughness of the soil-solid inter-face on the drained large-displacement interface shear strength. The results are alsodiscussed in reference to findings arising from the limited number of past studies onthis topic.5.2 Effect of Roughness of the Test Surfaces onthe Interface Shear StrengthAs discussed previously, the drained large-displacement interface shear strength ofsoil-solid interfaces have been shown to depend on the material and roughness of thesolid surface. Further evidence of this was observed in the present research program.For example, the drained large-displacement interface shear strength of Kaolinite onthe steel surface was observed to be much greater than that observed on the twoepoxy surfaces. Further, differences in the drained large-displacement interface shearstrength was observed between the two epoxy materials as well. Past studies invlovingcharacterizing the shear strength of soil-solid interfaces, such as the work by Clark andMeyerhof (1972); Lemos and Vaughan (2009); Najjar et al. (2007a); Tsubakihara andKishida (1993); Uesugi and Kishida (1986a,b); Yoshimi and Kishida (1981), have useda single roughness parameter (such as: average roughness Ra, maximum roughnessRmax, or normalized roughness Ra/D50) to characterize the surface roughness of solidsurfaces (note: parameters Ra and Rmax are defined in Appendix A).However, there are certain difficulties involved in using a single parameter tocharacterize the surface roughness. One of the salient weaknesses that arises from thisapproach is that the arithmetic mean of the magnitudes of peaks and valleys of twogiven surfaces can yield the same Ra values, even though the two surfaces may possessvery much different profiles. Different researchers have used different gauge lengthsLg in measuring roughness parameters. The gauge length (also known as samplinglength) corresponds to the wave length of the high-pass filter λc that is used to remove2145.2. Effect of Roughness of the Test Surfaces on the Interface Shear Strengththe long wave length component of the surface profile that is typically treated as thewaviness component (note: the concept of gauge length is introduced in AppendixA). This variation also makes it difficult to compare roughness parameters acrossstudies effectively, as they are sensitive to the selected gauge length. For example,Jardine et al. (1993) do not report the gauge length of their roughness measurements.Yoshimi and Kishida (1981) report using a gauge length of 2.5 mm for measuring Rmaxof steel test surfaces, and the basis for selection of the gauge length is not reported.Tsubakihara and Kishida (1993); Uesugi and Kishida (1986a,b) on the other handreport using a gauge length equivalent to the D50 = 0.20mm of the test sand thatthey used in their experiments to measure the roughness of the test steel surfaces. Ina general sense, larger gauge lengths tend to produce larger Ra values, and smallergauge lengths tend to produce smaller Ra values for the same surface profile. Uesugiand Kishida (1986a) have introduced the use of the normalized roughness parameterRn = Rmax/D50, where Rmax is evaluated over the gauge length Lg = D50. Using theapproaches taken by various researchers as background, it was considered of interestto select a gauge length that would be a significant multiplier of the mean particlesize of a given test soil in question.With the above in mind and taking into consideration the capabilities of measure-ment devices and data-interpretation software available to assess roughness, it wasjudged that a minimum gauge length of Lg = 50 × D50 is reasonable for roughnessmeasurements. The average roughness Ra obtained over a gauge length of 50D50 isdenoted by Ra 50D50. In addition, surface profile data was used to determine rough-ness size distribution curves (defined as RSD curves) to characterize the roughnessof the solid test surfaces; the RSD curves depict the distribution of roughness in asimilar fashion to that for particle size distribution curves in geotechnical engineering.The roughness size distribution curves obtained from these measurements arerepeated in Fig. 5.8 for convenience. When considering the roughness size distributioncurves obtained based on a gauge length suitable for interfacing with Fraser Riversand, and Fraser River silt (i.e., Lg = 13.0 mm), the two epoxy coated steel surfaceswere found to show approximately the same average roughness value Ra 50D50 =2155.2. Effect of Roughness of the Test Surfaces on the Interface Shear Strength1.6− 1.8 µm, and the sand-blasted mild steel surface showed an Ra 50D50 value of 7.0µm. From the roughness size distribution curves obtained based on a gauge lengthsuitable for interfacing with Kaolinite (i.e., Lg = 0.12 mm), Ra 50D50 values of 0.15,0.5, and 7.0 µm were obtained for the green epoxy coated steel surface, the grey epoxycoated steel surface, and the sand-blasted mild steel surface respectively.Shown in Fig. 5.9 is the variation of the large-displacement interface efficiencyfactor f = tan(δ)/tan(φ′) with Ra 50D50 obtained based on the results of the macro-scale interface direct shear tests. It is important to note that the interpolation trendlines that connect the data points (dashed lines) have been arbitrarily drawn to fitthe general trend of the limited amount of data available. Nevertheless, the dataprovides some important insights with regard to the effect of surface roughness onthe soil-solid interface shear strength.The data shown in Fig. 5.9 indicates that each soil type has its own uniquef −Ra 50D50/D50 characteristic. In general, it can be seen that for each soil type, thelarge-displacement interface efficiency factor f increases with increasing Ra 50D50/D50.However, the rate of increase of f with Ra 50D50/D50 seems to be dependent on thesoil type. For example, Fraser River sand shows an increase of f from 0.74 to 0.81 asRa 50D50/D50 increases from 0.006 to 0.007 (∆f/∆(Ra 50D50/D50) = 70). In the caseof Fraser River silt, f increased from 0.78 to 0.84 with an increase of Ra 50D50/D50from 0.04 to 0.16, indicating a ∆f/∆(Ra 50D50/D50) rate of 0.5, and in the case ofKaolinite, a ∆f/∆(Ra 50D50/D50) value of 0.74 can be observed when consideringinterfacing with the green and grey epoxy coated steel surfaces.2165.2. Effect of Roughness of the Test Surfaces on the Interface Shear StrengthPercent finer (%)Grain size (mm), or Roughness size (mm) 10-610-510-410-310-210-1100020406080100Green epoxyGrey epoxySand-blasted steelKaoliniteRa_50D50 = 7 μm 0.5 μm 0.15 μm(a) Roughness size based on gauge length of 0.12 mm (Lg = 50×D50−Kaolinite)Percent finer (%)Grain size (mm), or Roughness size (mm) 10-610-510-410-310-210-1100020406080100 Green epoxyGrey epoxySand-blasted steelFraser River sandKaosandRedstoneStrong-Pit FGSFraser River siltRa_50D50 = 7 μm 1.8 μm 1.6 μm(b) Roughness size based on gauge length of 13.0 mm (Lg = 50×D50−FRS)Figure 5.8: Roughness size distribution of the solid test surfaces compared with grainsize distributions of test soils).2175.2. Effect of Roughness of the Test Surfaces on the Interface Shear Strength0.00.10.20.30.40.50.60.70.80.91.00 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18Large-displacement interface efficiency factorf = tan(δ) / tan(φ’)Ra_50D50 / D50FRS FRSilt(a) Fraser River sand and Fraser River silt0.00.10.20.30.40.50.60.70.80.91.00 0.5 1 1.5 2 2.5 3Large-displacement interface efficiency factorf = tan(δ) / tan(φ’)Ra_50D50 / D50Kaolinite FRSand FRSilt(3 kPa)𝜎𝑛′(30 kPa)(15 kPa)(b) Kaolinite compared with Fraser River sand and Fraser River siltFigure 5.9: Variation of large-displacement drained interface efficiency factor withRa 50D50/D50.Within the Ra 50D50/D50 values tested (Ra 50D50/D50 in the range of 0.006 to 0.2),2185.2. Effect of Roughness of the Test Surfaces on the Interface Shear Strengthboth Fraser River sand and Fraser River silt showed large-displacement interfaceefficiency factors (f) close to, but less than, unity. In contrast, Kaolinite was observedto show an f value of unity at Ra 50D50/D50 value of approximately 2.7 when testedat σ′n = 3 kPa. The exact Ra 50D50/D50 at which each soil type would show an fvalue of unity is not known, and more data is needed to establish the correspondingcritical Ra 50D50/D50 values. It is of relevance to note that, in the case of Kaolinite,the variation of f with Ra 50D50/D50 is affected by the effective normal stress levelwithin 3 to 30 kPa range as notable by the downward shift of the f − Ra 50D50/D50trend lines with the effective normal stress level.The large-displacement interface friction angles were found to be greatest for thesand-blasted mild steel surface and lowest for the green epoxy-coated steel surface inthe case of Kaolinite on the three solid surfaces. Looking at the RSD curves of thethree solid surfaces, one can clearly see that the RSD of the sand-blasted mild steelsurface is “coarser” than the GSD of Kaolinite. It is also evident that the range of theRSD of the sand-blasted mild steel surface (approx. 0.02 mm to 0.0001 mm) overlapsthe range of the GSD of Kaolinite (approx. 0.02 mm to 0.001 mm). The smearingthat was observed in tests involving the Kaolinite-steel interface can be explainedbased on the coarser RSD of the steel surface. The coarser roughness features of thesurface would allow the finer portion of the Kaolinite particles to fill such features,ultimately resulting in a soil-soil shear zone as opposed to a soil-solid interface shearzone.It can be seen that both epoxy surfaces show RSD curves much “finer” than theGSD of Kaolinite. When tested with Kaolinite, the green epoxy surface which showsthe finest RSD resulted in the lowest large-displacement interface friction angles.However, the range of the RSD values of the two epoxy surfaces do not overlap withthe GSD of Kaolinite. This indicates that the roughness features of the two epoxysurfaces are much finer than the Kaolinite particle sizes. Despite this being the case,the effect of the surface type on the large-displacement interface friction angle canclearly be observed.It is of relevance to note that the two roughness measurement devices produced2195.2. Effect of Roughness of the Test Surfaces on the Interface Shear Strengthdifferent RSD curves for the same epoxy coated mild steel surfaces. The mechanicalstylus type profilometer is limited in resolution by its stylus tip radius. As can beseen in Fig. 5.8, the stylus type profilometer produced approximately identical RSDcurves for both epoxy surfaces. This shows that selection of a tool with a resolutionand gauge length that is suitable for a given soil type is important in characterizingRSD’s. It is also important to be mindful that the roughness profile obtained us-ing a profilometer is only a two-dimensional representation of the three-dimensionalsolid surface. When reading the RSD curve of a given solid surface, it is implic-itly assumed that the two-dimensional roughness size measured is representative ofa three-dimensional roughness feature. This assumption may not be strictly accu-rate as there may be anisotropy in the roughness features along the two orthogonalhorizontal axes of roughness measurement. In the present research surface roughnesssize distributions based on measurements carried out in the two orthogonal directions(north-south and east-west) did not show a notable degree of anisotropy. Reportingan average RSD based on two sets of orthogonal roughness profiles may be moreappropriate when surfaces showing roughness anisotropy are encountered.A relatively large data set that covers a wide range of roughness values is requiredin order to gain a better understanding of the general trend of the value of f withsurface roughness,. The range of roughness values covered in the present researchprogram is very limited, and further testing to cover a wider range of surface roughnessis warranted. However, further insight into the effect of roughness on f can be gainedby comparing data from past studies. It is of importance to note that the gaugelength used to measure surface roughness is not consistent across studies. Therefore,it is difficult to compare separate data sets in a meaningful manner. Nevertheless,it is possible to gain a sense of the general trend of f with roughness by looking atresults of past studies.Shown in Fig. 5.10 is the variation of the large-displacement drained interfaceefficiency factor with normalized average roughness Ra/D50 obtained from past stud-ies for a number of coarse-grained and fine-grained soils tested on solid surfaces ofvarious roughness. Shown on the same figure is the f − Ra 50D50 data of the present2205.2. Effect of Roughness of the Test Surfaces on the Interface Shear Strength0.00.10.20.30.40.50.60.70.80.91.00.00001 0.0001 0.001 0.01 0.1 1 10Large-displacement interface efficiency factorf = tan(δ) / tan(φ’)Ra or Ra_50D50 / D50Kaolinite (Present study) FRSand (Present study)FRSilt (Present study) Ottawa sand (100 kPa) (Rinne, 1989)Ottawa sand (500 kPa) (Rinne, 1989) Ottawa sand (750 kPa) (Rinne, 1989)Sand1 (50 kPa) (Reddy et al., 2000) Sand1 (100 kPa) (Reddy et al., 2000)Sand1 (150 kPa) (Reddy et al., 2000) Leighton Buzzard sand (50 kPa) (Reddy et al., 2000)Leighton Buzzard sand (100 kPa) (Reddy et al., 2000) Leighton Buzzard sand (150 kPa) (Reddy et al., 2000)Clay (Lemos and Vaughan, 2009) Target sand (100 kPa) (Rinne, 1989)Target sand (500 kPa) (Rinne, 1989) Target sand (750 kPa) (Rinne, 1989)(3 kPa)𝜎𝑛′(30 kPa)(15 kPa)Figure 5.10: Variation of the large-displacement drained interface efficiency factorwith normalized average roughness obtained from past studies for a number of coarse-grained and fine-grained soils tested on solid surfaces of various roughness (Lemos andVaughan, 2009; Reddy et al., 2000; Rinne, 1989).study. In general, it can be seen that the increase in roughness results in an increasein the measured large-displacement interface efficiency factor. Looking at the datafrom Rinne (1989) and Reddy et al. (2000), it can be seen that the f −Ra/D50 trendlines for coarse-grained soils can be affected by the effective normal stress level. Asexpected, it is also evident that the soil type has a notable effect on the interfaceroughness characteristics.It is of relevance to note that the effective normal stresses used in past studies aremuch higher than those used in the present research. Nevertheless, the findings of thepresent study are in general agreement with the above observations when consideringthe effect of surface roughness on the large-displacement interface shear strength.2215.3. Effect of Rate of Shearing on the Interface Shear StrengthHowever, it is evident that there is a need to improve and standardize methods ofcharacterizing the roughness and texture of solid surfaces in a manner suitable forgeotechnical engineering applications. The present research work has provided onesuch approach by relating the minimum gauge length of roughness measurement tothe mean particle size of a given soil, and by introducing the concept of roughnesssize distribution curves to aid in characterizing the surface roughness in a mannerfamiliar to methods in geotechnical engineering. The database of soil-solid interfaceshear strength obtained at large-displacement failure conditions is limited, and thereis much room for improvement. In this respect, the present study has generated highquality data that is of much value to the field of geotechnical engineering.5.3 Effect of Rate of Shearing on the InterfaceShear StrengthThe effect of rate of shear displacement on the soil-solid interface shear character-istics of fine-grained soils is an important aspect and, as described in Chapter 4, alimited number of macro-scale interface direct shear tests were conducted using thefine-grained soils to investigate whether the drained large-displacement interface shearstrength is dependent on the shear displacement rate. However, a detailed investiga-tion of the effect of rate of shear was not carried out in the scope of this thesis, andonly some limited factual observations are presented.All the rate-effect tests were conducted on the grey epoxy coated mild-steel surface.Fraser River silt was tested on the grey epoxy surface targeting σ′n of 3.0 kPa atdisplacement rates of 0.01, 0.05, 0.07, and 0.14 mm/s. Two rate-effect tests wereconducted targeting σ′n of 30.0 kPa at displacement rates of 0.01 and 0.14 mm/s. Alimited number of rate-effect tests were also conducted using Kaolinite on the greyepoxy surface targeting 3.0 and 15.0 kPa effective normal stresses and at 0.002 and0.02 mm/s shear displacement rates.Shown in Fig. 5.11 are the pore-water dissipation time histories of the rate-effect2225.3. Effect of Rate of Shearing on the Interface Shear StrengthAvg. pore-water pressure (kPa)Time (s)0 5000 10000 15000 20000 25000 30000 3500000.511.522.533.54FRSilt_Grey_3_F0.01FRSilt_Grey_3_F0.05FRSilt_Grey_3_F0.07FRSilt_Grey_3_F0.14FRSilt_Grey_3_F0.14_RFigure 5.11: Pore-water pressure dissipation time history of rate effect tests conductedfor Fraser River silt on grey epoxy targeting σ′n = 3.0 kPa.tests conducted using Fraser River silt on the grey epoxy surface. Even at higherrates of shear displacement, the excess pore-water pressure that was generated atthe onset of shearing was observed to dissipate at approximately the same rate ofdissipation regardless of the rate of shearing. This observation was consistent in testsconducted using Fraser River silt as well as Kaolinite. It is of relevance to note thatwhen considering tests conducted at a given effective normal stress level, the specimenthickness was essentially the same in each test, and this ensured that the drainagepath distance (i.e., distance from the solid test plate surface to the top surface of thetest soil specimen) was the same in such tests. Estimation of the effective normalstress at higher rates of shear displacement can become challenging as the pore-waterpressure measurements at the interface become less reliable at higher rates of shearing.The limited number of tests with different rates of shear displacement herein wereconducted with the objective of getting a preliminary assessment of the effect of rateof shearing on the drained large-displacement interface shear strength. The partiallyundrained or any transient response of the interface during the early stages of sheardisplacement was not studied in detail owing to the difficulty in accurately estimatingthe effective normal stress at the interface.Figure 5.12 shows the variation of the drained large-displacement interface frictionangle with rate of shear displacement obtained from the rate-effect tests. The drainedlarge-displacement interface shear strength of Fraser River silt on the grey epoxy2235.4. Summary of the Chapter, Key Findings, and Contributions05101520253035400.000 0.020 0.040 0.060 0.080 0.100 0.120 0.140 0.160Secant drained interface friction angle at large-displacement (deg.)Displacement rate (mm/s)FRSilt-Grey (3 kPa) FRSilt-Grey (30 kPa) Kao-Grey (3 kPa) Kao-Grey (15 kPa)(𝜎𝑛′≈3 𝑘𝑃𝑎)(𝜎𝑛′≈30 𝑘𝑃𝑎)(𝜎𝑛′≈2.8 𝑘𝑃𝑎)(𝜎𝑛′≈2.2 𝑘𝑃𝑎)(𝜎𝑛′≈13 𝑘𝑃𝑎)(𝜎𝑛′≈16 𝑘𝑃𝑎)Figure 5.12: Variation of the drained large-displacement interface friction angle withrate of shear displacement.appears to be insensitive to the rate of shearing within the 0.01 to 0.14 mm/s range. Italso appears that the drained large-displacement interface shear strength of Kaoliniteis insensitive to the rate of shear displacement within the 0.002 to 0.04 mm/s range.Given that only a limited number tests were conducted at only a limited range ofshear displacement rates, the rate-effect test results need to be treated with caution,until further testing can substantiate the findings.5.4 Summary of the Chapter, Key Findings, andContributionsMacro-scale interface direct shear tests were conducted using three soil types (FraserRiver sand, Fraser River silt, and Kaolinite) on three solid surfaces (green epoxycoated mild steel, grey epoxy coated mild steel, and sand-blasted mild steel), at targeteffective normal stresses of 3, 15, and 30 kPa, to characterize the large-displacement2245.4. Summary of the Chapter, Key Findings, and Contributionsinterface shear strength of some soil-solid interfaces commonly encountered in geotech-nical engineering practice. These experiments were used to: (i) expand the databaseof large-displacement drained soil-pipe interface shear strength data obtained at lowconfining stresses; and (ii) study the sensitivity of the drained large-displacementsoil-pipe interface shear strength to the confining stress level, and the solid surfaceroughness. A limited number of tests were conducted to assess the effect of rate ofshearing on the large-displacement drained soil-solid interface shear strength.Some key findings and contributions arising from this work are summarized below:1. The macro-scale interface direct shear tests revealed that for Fraser River sand,the large-displacement drained interface shear strength envelopes are linearwithin the 3 to 30 kPa effective normal stress range on all the three solid surfacesused in the testing program. For Fraser River silt on the three solid surfaces, theinterface shear strength envelopes were found to be slightly nonlinear (concaveupwards) within the same effective normal stress range. The interface efficiencyfactors at large displacement, defined as f = tan(δ)/tan(φcv), for Fraser Riversand as well as Fraser River silt on all three solid surfaces were found to beless than unity. Past work on characterizing soil-solid interface shear strengthof coarse-grained soils under relatively low effective normal stresses is very lim-ited. The available data is in agreement with the results of the present study.2. When interfacing Kaolinite on the two epoxy coated steel surfaces, the large-displacement drained interface shear strength envelopes were found to be non-linear within the 3 to 10 kPa effective normal stress range. The Kaolinite-steelinterface showed sensitivity to the effective normal stress throughout the 3.0to 30.0 kPa range resulting in a slightly curved interface shear strength enve-lope. This observation is in agreement with those of the limited number of paststudies that have been undertaken to characterize the soil-solid interface shearstrength of fine-grained soils at relatively low effective normal stresses.3. The interface efficiency factors recommended in current pipeline design guide-lines such as PRCI and ALA (ALA, 2001; PRCI, 2004, 2009), are expressed2255.4. Summary of the Chapter, Key Findings, and Contributionsbased on the maximum internal friction angle of the soil. The definition of theinterface efficiency factors used in the guidelines is f = δ/φm. Given that theinternal friction angle of soils is highly dependent on many factors includingthe different modes of shearing in different test methods and the stress pathsimposed in each type of test method, it is important that the conditions underwhich the internal friction angle of the soil is measured are more clearly defined.Given that the small-strain peak friction angle of soils would highly depend oninitial conditions, confining stress level and mode of shear, the use of the peakfriction angle of a given soil for defining the interface efficiency factors wouldrequire the listing of all the conditions under which the peak internal frictionangle was obtained. The use of the critical state internal friction angle φcv ofthe soil would be much more appropriate in specifying design interface efficiencyfactors. It is also important to specify whether the interface efficiency factorsare based on small-strain or large-displacement conditions.4. The interface efficiency factors of fine-grained soils recommended in currentpipeline design guidelines such as PRCI and ALA (ALA, 2001; PRCI, 2004,2009), do not take into consideration the possible sensitivity of the interfaceefficiency factors on the confining stress level. The results of the present studyhas shown that the large-displacement interface efficiency factor of fine-grainedsoils on epoxy coated surfaces can increase by a considerable amount when theeffective normal stress drops below 10 kPa. Due to the difficulty in measur-ing the volume change at the shear zone of a test specimen it is not knownfor certain whether the large-displacement interface friction angles measured inthe present study correspond to the condition of critical state. Given the largeshear displacements used, and based on the good agreement of the results withthose obtained from large-displacement ring-shear testing, it is highly proba-bly that the interface friction angles obtained in the present study correspondto the critical state condition where shearing occurs with no volume change.While further research is needed to investigate and fully understand the cause2265.4. Summary of the Chapter, Key Findings, and Contributionsof this increase in the large-displacement interface friction angle with decreasingconfining stress level, the evidence presented in the present work suggests thatthe effect of confining stress level on the large-displacement interface frictionangle needs to be considered in pipeline design. As will be shown in Chapter6, the effect of confining stress level on the large-displacement interface frictionangle was observed to be at play in the full-scale axial soil-pipe interaction testsconducted in the present research.5. A literature review of studies involving characterization of soil-solid interfaceshear strength for geotechnical engineering applications showed that a consis-tent method of selecting a suitable gauge length for roughness measurementshas not yet been established. As a result, different studies have used differ-ent gauge lengths of roughness measurement, and this makes it difficult (if notimpossible) to meaningfully compare the effects of surface roughness on the soil-solid interface shear strength across different studies. It is evident that there isa need to improve and standardize methods of characterizing the roughness andtexture of solid surfaces in a manner suitable for geotechnical engineering appli-cations. The present research work has provided one such approach by relatingthe minimum gauge length of roughness measurement to the mean particle sizeof a given soil, and by introducing the concept of roughness size distributioncurves to aid in characterizing the surface roughness in a manner consistentwith methods in geotechnical engineering.6. The large-displacement interface shear strength of soil-solid interfaces was gen-erally found to increase with increasing Ra 50D50. The rate of increase of thelarge-displacement interface efficiency factor with Ra 50D50 was observed to bedependent on the soil type. There appears to be a critical roughness after whichthe large-displacement drained soil-solid interface shear strength becomes equalto the internal large-displacement drained shear strength of the soil as evi-denced by the interface shear strength characteristics of Kaolinite on the sand-blasted mild steel surface and based on a limited number of results of past2275.4. Summary of the Chapter, Key Findings, and Contributionsstudies reported in the literature (UESUGI et al., 1989). In the macro-scaleinterface direct shear tests conducted using Kaolinite on the sand-blasted mildsteel surface clear evidence of smearing of the soil on the solid surface was ob-served. The results of the present research showed that the position and rangeof the roughness size distribution (RSD) curve of the solid surface relative tothe position and range of the grain size distribution (GSD) of the soil may beuseful in predicting whether the large-displacement drained soil-solid interfaceshear strength would be equal to the internal large-displacement drained shearstrength of the soil. Shearing a fine-grained soil on a solid surface having anRSD that is “coarser” than the GSD of the soil would result in the finer soilparticles filling the coarser roughness features of the solid surface and eventu-ally leading to soil-soil shearing. Further research is needed to clearly identifythe critical RSD parameters for various soil types so that those could be usedin engineering practice as guidelines in selecting appropriate large-displacementinterface efficiency factors.7. The limited number of macro-scale interface direct shear tests conducted at rel-atively low effective normal stresses and at different rates of shear displacementallowed for a preliminary assessment of the effects of rate of shear displace-ment on the interface shear strength. Factual data obtained from these testswere reported, and it appears that within the displacement rates tested, thelarge-displacement drained interface shear strength of Fraser River silt on thegrey epoxy coated solid surface is not affected by the rate of shear displace-ment. The drained large-displacement interface shear strength of Kaolinite onthe grey epoxy coated solid surface appeared to show some sensitivity to therate of shear displacement when tested at effective normal stresses below 10 kPa.Specifically, at effective normal stresses below 10 kPa, the large-displacementinterface friction angle was observed to increase slightly with the increase inrate of shear displacement.8. The database of soil-solid interface shear strength obtained at large-displacement2285.4. Summary of the Chapter, Key Findings, and Contributionsfailure conditions is limited, and there is much room for improvement. In thisrespect, the present study has generated much needed data to help fill gaps inliterature and to help further our understanding of soil-pipe interface frictionunder low confining stresses and large displacements.229Chapter 6Full-Scale Axial Soil-PipeInteraction TestingThis chapter presents details of a new test device that was developed to conductfull-scale axial soil-pipe interaction physical model tests in fine-grained soil beds.As presented at the beginning of this thesis, the main goals of the present researchproject were: (i) development of a reliable method for laboratory testing of soil-solidinterfaces under relatively low effective normal stresses for characterization of thedrained large-displacement interface shear strength; (ii) characterizing the drainedlarge-displacement interface shear strength of a number of soil-solid interfaces com-monly encountered in practice; (iii) development of a reliable method for laboratoryphysical modeling of full-scale axial soil-pipe interaction problems; and (iv) use ofmacro-scale interface direct shear test results to aid in the analysis and prediction ofthe full-scale physical model axial soil-pipe interaction test results. Items (i) and (ii)were covered in Chapters 3 through 5, and the present chapter covers the last two ofthe research goals.6.1 IntroductionDepending on the direction of relative displacement between the soil and the pipeline,soil-pipe interaction testing can be categorized into four different modes: (i) axialsoil-pipe interaction; (ii) lateral soil-pipe interaction; (iii) vertical uplift soil-pipe in-teraction; and (iv) vertical bearing soil-pipe interaction (see Fig. 6.1). As discussed inChapter 2, such experiments are often conducted utilizing some form of a large-scalesoil chamber where the soil bed (or pipe trench) is prepared, the pipe is installed,2306.1. Introduction(c) Vertical uplift(a) Axial interaction(d) Vertical bearing(b) Lateral interactionFigure 6.1: Schematic representations of different soil-pipe interaction modes.and controlled displacements are applied to the pipe to simulate relative movement(Brennodden et al., 1986; Cheuk et al., 2007; ORourke et al., 2016; PRCI, 2004;Wijewickreme et al., 2009).Axial soil-pipe interaction testing of partially buried pipes is important in develop-ing/calibrating/validating models of soil springs and soil-pipe interface stress-strainmodels that can be used for the design of off-shore pipelines. In the case of modelingpartially buried offshore pipelines, tests are often conducted by first placing the testpipe on a soil bed (typically a saturated fine-grained soil bed is used representing thefield conditions prevalent in the offshore environment), and then, upon completion ofthe consolidation of the soil bed under the weight of the test pipe, by applying a con-stant rate of shear displacement to the pipe in the axial direction. The shear stress- displacement response observed in the test is used to develop/calibrate/validate2316.1. Introductionsuitable soil-spring models. In certain instances where any cyclic back and forthmovement of a pipeline is expected, cyclic physical model tests may be conducted tosimulate such movement.The plowing of soil at the leading end of the test pipe causes an additional force(see Fig. 6.2)in addition to the shaft friction that develops on the partial perimeteras the model test pipe is displaced as pointed out in many studies (Bruton et al.,2009; Senthilkumar et al., 2011; White et al., 2011, 2010). Since such plowing of soildoes not occur in a real pipeline, this end-effect error in the experiment needs to beaccounted for to obtain the correct axial soil-pipe shear-displacement response. Thereare two approaches that have often been used to account for this end effect: (i) the useof a relatively long test pipe segment so that the interface shear forces become muchmore pronounced compared to the end-effect forces (White et al., 2011); and (ii) theuse of a three-piece test pipe segment where the ends of the pipe are connected to themiddle segment through load cells so that the end-effect forces can be measured andcorrected for (see Fig. 6.2-(c)) (Bruton et al., 2009; Senthilkumar et al., 2011; Whiteet al., 2010).The first approach is favourable as it is simple and produces boundary conditionsthat are relatively closer to reality. However, in full-scale testing, the use of a longlength of test pipe requires a large test footprint, and requires a large quantity ofsoil, which often makes project costs economically unfeasible. The second approachallows the use of a relatively shorter test pipe and hence a smaller test footprint,but the test pipe needs to be carefully instrumented to allow for the measurement ofend-effect forces. Also, the non-uniform boundary conditions and remolding of thesoil that result due to the plowing action that occurs at the leading end (south endin Fig. 6.2) of the test pipe causes a non-uniformity in the properties of the soil thatthe middle test pipe section is exposed to as displacement continues. There is alsothe possibility of the test pipe “biting” into the soil bed at the leading end duringdisplacement. Studies that employ the second approach often remove soil ahead ofthe leading end of the test pipe to reduce these effects.Given these practical difficulties in obtaining reliable large-scale physical model2326.1. IntroductionSaturated fine-grained test soil bedNorthSoil chamberTest pipeLoad cells for axial and vertical force measurement(a) Test pipe placed on soil bed and the soil allowed to consolidate under the weight of pipeSaturated fine-grained test soil bedSouth (direction of test pipe displacement)Soil chamberTest pipePlowing of soil bed at leading end of test pipeNorth end South endMiddle portion(b) Test pipe displaced axiallySaturated fine-grained test soil bedSouth (direction of test pipe displacement)Soil chamberTest pipePlowing of soil bed at leading end of test pipeNorth end South endMiddle portionLoad cells for measuring end effect forces(c) Instrumented test pipe displaced axiallyFigure 6.2: Schematic cross-sectional view showing conventional partially buried axialsoil-pipe interaction test configuration.2336.2. New Device for Axial Soil Restraint Testing of Buried Pipes in Compressible Soilstest data, the present research work resulted in producing a novel approach of axialsoil-pipe interaction testing that can be used to test pipes of upto 460 mm (NPS18)diameter in saturated fine-grained soils focused on addressing some of the aforemen-tioned concerns. The following sections of the text provide details of the new testmethod. The improvements over previous test methods as well as the limitationsof the new method are presented. Test results of a limited number of preliminaryaxial soil-pipe interaction tests that were conducted using the new method are alsopresented. An analysis of the full-scale axial soil-pipe interaction test results wasconducted in the context of using the macro-scale interface direct shear test resultsfor predicting the axial soil-pipe interaction forces.6.2 New Device for Axial Soil Restraint Testingof Buried Pipes in Compressible Soils6.2.1 Overview of test deviceMany studies that focus on studying axial soil-pipe interaction in coarse-grained soilshave used the test configuration shown in Fig. 6.3. In this case, the test pipe is muchlonger than the soil chamber length, and is inserted into the soil chamber throughtwo circular openings such that the two ends of the pipe protrude out through theopenings. This test configuration is preferred as the ends of the test pipe lie outsidethe test soil boundary. The length of test pipe that is inside the soil chamber maintainsrelatively uniform contact with the test soil. As the test pipe is axially displaced, thesoil-pipe interface shear is mobilized along the length of test pipe that is in contactwith the soil with minimal end effects.This configuration is not directly suitable for use with fine-grained soils; this is be-cause the fine-grained grained soils undergo relatively large consolidation settlementswhen the test pipe is placed on the soil bed. The conventional soil chamber setupdoes not allow the pipe to follow this movement since the pipe gets restrained by therigid side walls of the soil chamber. In order to allow for the pipe to freely displace2346.2. New Device for Axial Soil Restraint Testing of Buried Pipes in Compressible SoilsActuator Load cellNorth rigid soil chamber wallsCoarse-grained test soilTest pipeDirection of  test pipe displacementNorth South rigid soil chamber wallsComplete length of soil-pipe interface is mobilized in shear. End effects minimum.Figure 6.3: Schematic cross-sectional representation of a conventional fully buriedaxial soil-pipe interaction test setup used when utilizing coarse-grained soils. (Note:this configuration is also applicable for partially buried pipe testing) (Huber andWijewickreme, 2014; Karimian, 2006).vertically (to accommodate the consolidation settlements of the soil bed), the newtest method with mobile walls on the south end (leading end of pipe) and north end(trailing end of pipe) of the soil chamber (see Fig. 6.4) was developed by accord-ingly modifying the UBC Advanced Soil-Pipe Interaction Research (ASPIReTM) soilchamber test facility. The details of the new device are presented in Section 6.2.2.6.2.2 Description of main components of test deviceThe UBC Advanced Soil-Pipe Interaction Research (ASPIReTM) soil chamber wasmodified to incorporate a mobile wall system that allows the test pipe to freely movein the vertical plane to accommodate consolidation settlements. A schematic repre-sentation of the new setup is shown in Fig. 6.4. For visual comparison, photographsof the UBC ASPIReTM soil chamber test facility before and after incorporating thenew test configuration are shown in Fig. 6.5. A drawing showing the perspective viewof the new test configuration is presented in Fig. 6.6. The new test configurationis very similar to the axial soil-pipe interaction setup used for coarse-grained soils(Karimian et al., 2006; Karimian, 2006; Wijewickreme et al., 2009) except for thedetails as described below.2356.2. New Device for Axial Soil Restraint Testing of Buried Pipes in Compressible SoilsActuator Load cellSaturated fine-grained test soil bedTest pipe(Note: mobile walls and test pipe are in the raised position, prior to placement of pipe on the soil bed.)North Mobile walls(a) Mobile walls in the raised position prior to placement of test pipe on the soil bedActuator Load cellSaturated fine-grained test soil bed(Note: mobile walls and test pipe have displaced downward due to consolidation of the soil bed.)North Mobile wallsTest pipe(b) Mobile walls displace downward as the soil bed consolidates under the weight of t