Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Dilatancy characterization of sands using the resistivity cone penetration test Kokan, Matthew J. 1992

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

Item Metadata


831-ubc_1992_fall_kokan_matthew.pdf [ 2.56MB ]
JSON: 831-1.0050465.json
JSON-LD: 831-1.0050465-ld.json
RDF/XML (Pretty): 831-1.0050465-rdf.xml
RDF/JSON: 831-1.0050465-rdf.json
Turtle: 831-1.0050465-turtle.txt
N-Triples: 831-1.0050465-rdf-ntriples.txt
Original Record: 831-1.0050465-source.json
Full Text

Full Text

DILATANCY CHARACTERIZATION OF SANDS USING THERESISTIVITY CONE PENETRATION TESTBYMATTHEW J. KOKANB.A.Sc., The University of British ColumbiaA THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE FACULTY OF APPLIED SCIENCEDEPARTMENT OF CIVIL ENGINEERINGWe accept this as conformingto th ‘‘‘ - — —THE UNIVERSITY OF BRITISH COLUMBIAAPRIL 1992Øcopyright Matthew J. Kokan, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of___________________________The University of British ColumbiaVancouver, CanadaDate ‘y /99_—DE-6 (2188)iiABSTRACTThe determination of in situ dilatancy in sands has been adifficult and elusive objective in site investigations.Difficulty with sampling, as well as problems with extrapolatinglaboratory results to field performance have caused geotechnicalengineers to place increased reliance on in situ tests. One suchtest which is gaining acceptance in the geotechnical professionis the cone penetration test (CPT). As with other in situ tests,the CPT can be used to predict dilatancy behaviour of sandsbased on empirical correlations. As with most empirical methods,correlations are often site specific and sensitive to variablesthat are not readily measurable. A new, geophysically based,technique has been developed to determine dilatancycharacteristics of sands in situ.The resistivity cone penetration test (RCPT) employs astandard 10 sq cm piezo cone, paired with a module whichmeasures soil resisitivity at different electrode spacings. Theresistivity is recorded in a semi—continuous manner along withregular CPT data. The resistivity measurements at differentelectrode spacings can be used to infer sand densities atdifferent distances from the penetrating probe. By comparing theresistivity close to the probe with the resistivity further awayfrom the probe, it is possible to observe the shear inducedvolume change caused by penetration of the probe. This approachis analytical and does not require water sampling, nor is itiiisite specific.Data are presented to illustrate the effectiveness of thistechnique in determining in situ dilatancy of sands. Thetechnique is compared to existing empirically based approachesfor prediction of dilatancy. Finally, possible futureapplications of the RCPT are discussed.ivTABLE OF CONTENTSABSTRACTTABLE OF CONTENTSLIST OF TABLESLIST OF FIGURESLIST OF SYMBOLSACKNOWLEDGEMENTSiiivviviixxi1. INTRODUCTION 12. PROCEDURES AND EQUIPMENT2.1 Research Vehicle2.2 Test Equipment2.3 The U.B.C. Resistivity Module2.4 Data Collection and Reduction333473. MEASUREMENT OF SOIL RESISTIVITY3.1 Introduction3.2 Resistivity measurement3.3 Calibration of the Resistivity Module94. FACTORS AFFECTING RESISTIVITY MEASUREMENTIN SOILS 184.1 Introduction 184.2 Groundwater Effects on MeasuredResistivity4.3 Soil Effects on Measured ResistivityEffect of FinesEffect of Organic MaterialsResistivity Measurement in Sands4.4 Sununary 325. USE OF THE RCPT TO DETERMINE SAND BEHAVIOUR .. 345.1 Premise of Research5.2 Constraints on the Analysis5.3 Theoretical Basis5.4 Dilation Parameter6. FIELD INVESTIGATION PROGRAM AND RESULTS . .6.1 Description of Field Investigation6.2 Arthur Laing Bridge Site6.3 Miller Road Site6.4 Alex Fraser Bridge Site6.5 U.B.C. Pile Research Site6.6 Kwantlen College Site3438404953535560656772910114. PREDICTION OF DILATANCY BEHAVIOUROF SAND FROM RCPT DATA . . . . 777.1 Introduction 777.2 Arthur Laing Bridge Site 787.3 Miller Road Site 817.3 Alex Fraser Bridge Site 837.5 U.B.C. Pile Research Site 877.6 Kwantlen College Site 877.7 Sununary 908. PRACTICAL USE OF THE DILATION PARAMETER . 928.1 Introduction 928.2 CPT-D Correlations 938.3 SPT-D Correlations 998.4 Relative Density - D Correlations 1018.5 Densification Control 1018.5.1 Alex Fraser Bridge Site . . . 1038.5.2 Kwantlen College Site 1038.6 Effect of Fines on Soil Behaviour In Situ 1069. CONCLUSIONS AND RECOMMENDATIONS . . . 11010. POSSIBLE APPLICATIONS 113REFERENCES . 115APPENDIX A 118APPENDIX B 125viLIST OF TABLESTABLE PAGE7.1 Summary of SPT results from Laing Bridge site . . 81viiLIST OF FIGURESFIGURE TITLE PAGE2.1 UBC Resistivity Cone Penetrometer 53.1 The effect of dissolved solids and temperatureon specific conductance (after Carr, 1982) . . .. 123.2 Schematic diagram of calibration setup 143.3 Variation in calibration factor with excitationlevel and measured electrode potential 164.1 A comparison between bulk resistivityand pore fluid resistivity measured at LaingBridge site 234.2 A comparison between formation factors calculatedfrom three different electrode spacings at LaingBridge site 244.3 A comparison of resistivity (77.5 mm spacing)and friction ratio at the UBC pileresearch site 264.4 Observed relationship between apparentformation factor and cone bearing normalizedwith respect to horizontal effective stress(after Weemees, 1990) 294.5 A comparison of resistivity (77.5 mm)spacing)and cone bearing at Laing Bridge site 304.6 A comparison between resistivity vs conebearing for largest and smallest electrodespacings from Laing Bridge site where porefluid resistivities are constant 315.1 Volumetric strain vs relative radial distancefrom penetrometer axis (after Chong,1988) .. 455.2 Measured dilatancy behaviour of a densesand during a pressuremeter test, using aresistivity method (after Windle & Wroth,1975). 465.3 Flownet from electrical analog model showingdistribution of current with distance fromthe resistivity module 486.1 Map of Fraser Delta showing locations ofresearch sites 54viii6.2 Arthur Laing Bridge research site 566.3 RCPT sounding from Arthur Laing Bridge site . . . 576.4 A comparison between the smallest electrodespacing, the largest electrode spacing andpore fluid resistivities measured at theArthur Laing Bridge site 586.5 Miller Road research site 616.6 RCPT sounding from Miller Road site 626.7 A comparison of bulk resistivity and conebearing in the vadose zone at the MillerRoad site 646.8 Alex Fraser Bridge research site 666.9 RCPT sounding from Alex Fraser Bridge site . 686.10 UBC pile research site 696.11 RCPT sounding from UBC pile research site . . 716.12 Kwantlen College research site 736.13 RCPT sounding from Kwantlen College researchsite (not treated with DC) 746.14 RCPT sounding from Kwantlen College researchsite (treated with DC) 757.1 A comparison between dilation parameter andnormalized cone bearing with depth forLaing Bridge site 797.2 A comparison between dilation parameter andnormalized cone bearing with depth for MillerRoad site 827.3 A comparison between dilation parameter andnormalized cone bearing with depth for AlexFraser Bridge site (south side) 847.4 A comparison between dilation parameter andnormalized cone bearing with depth for AlexFraser Bridge site (north side) 857.5 A comparison between dilation parameter andnormalized cone bearing with depth for theUBC pile research site 88ix7.6 A comparison between dilation parameter andnormalized cone bearing with depth for theKwantlen College site 898.1 Normalized cone bearing vs dilation parameterfor Arthur Laing Bridge site (from 4-14 m) . 948.2 Normalized cone bearing vs dilation parameterfor Miller Road site (from 8-18 m) 958.3 Normalized cone bearing vs dilation parameterfor Alex Fraser Bridge site (from 4—14 m) . . . 968.4 Normalized cone bearing vs dilation parameterfor Arthur Laing Bridge, Miller Road andAlex Fraser Bridge sites . 988.5 Laing Bridge SPT (N1)60 - dilation parametercorrelation from LAING11.CPT, using measured(N1)60 and (N1)60 from CPTINT v4.0.22(4—14 m) 1008.6 Relative density vs dilation parameter forLaing Bridge site, interpreted from CPTINTv4.0.22 (from 4 to 14 metres) 1028.7 Comparison between improved and unimproved soilat Alex Fraser Bridge site (south side) usingnormalized cone bearing and dilation parameteras improvement indexes 1048.8 Comparison between improved and unimproved soilat Kwantlen College site using normalized conebearing and dilation parameteras improvement indexes 1058.9 Comparison between dilation parameter, frictionratio and normalized cone bearing in thedensified region at Kwantlen College site . . . 108xLIST OF SYMBOLSR = resistanceV = voltageI = currentp = resistivityPb = bulk resistivitypf = pore fluid resistivityA = cross—sectional area1 = path length of flowK = electrode calibration factorF = apparent formation factor• = porositym = soil shape factorD = dilation parameterQ = cone bearing= normalized cone bearing= vertical effective stressxiACKNOWLEDGEMENTSI would like to extend my gratitude to my thesis advisor, Dr.R.G. Campanella, for providing me with the opportunity to carry outthis research and lending his valuable insight.I would also like to thank David J. Woeller of ConeTecInvestigations Ltd. and technicians Scott Jackson, Art Brookes andHarald Schrempp for their assistance.Additional thanks are also extended to fellow graduatestudents Jaiwei Wang, Renato pinto de Cunha and especially ThandavaMurthy for their help in carrying out my field work.Finally, I would like to thank Marnie and my parents for theirsupport and understanding throughout the last several months.I wish to acknowledge the financial support of the NaturalSciences and Engineering Research Council, Canada, in the form ofa graduate research assistants grant.11. INTRODUCTIONThe electronic cone is rapidly becoming the tool of choicefor geotechnical engineers who have gained a wide experience inin situ testing. The cone penetration test (CPT) is a fast andaccurate way of logging stratigraphy as well as measuringimportant geotechnical parameters. Digital data acquisitiontechnology allows near continuous recording of sleeve friction,cone bearing and pore pressure. Since its inception, electroniccone testing technology has been constantly evolving, with thedevelopment of new cone designs, as well as new tests to measurean increasing number of in situ properties. Recently aresistivity module was developed at U.B.C (Campanella andWeemees, 1990) which permits in situ resistivity to be measuredalong with the standard cone parameters. In situ resistivity isrelated to several factors, one of which is soil porosity. Atechnique has been developed that uses resistivity to measurechanges in porosity of sands caused by penetration inducedshear. The change in porosity is intimately related to thedilatancy behaviour of sand in situ. This method does notrequire water or soil sampling and at this point appears not tobe site specific. The purpose of this report is not to providean argument in favour of large scale deployment of a test, butrather to introduce a new technique and to provide somerationale for its use. Data will be presented to illustrate thetechnique as well as to show how it can be used parallel toother tests for evaluation of in situ behaviour and field2performance. This method shows particular promise since it isanalytically based and therefore does not require labcorrections, energy measurement or any type of normalizationprocedures.32. PROCEDURES AND EQUIPMENT2.1 RESEARCH VEHICLEThe in situ geotechnical group at U.B.C. is fortunate tohave an excellent support vehicle from which cone penetrationtesting is conducted. The vehicle is a modified Ford LN600 withreinforced chassis and a power take—off used to operate thetruck’s hydraulics and on board electrical equipment. Its set upis adjustable to accommodate different types of tests includingmechanical cone, electrical cone, fixed piston sampling,pressuremeter and other in situ tests. The twin piston hydraulicram has a total pushing capacity of 40,000 lbs. Because of thedead weight of the truck, the safe pushing capacity is about18,000 lbs without installation of reaction anchors or weights.The deepest penetration accomplished to date, without reactionanchors or ballast weight, is 80 iu. For more information on theU.B.C. research vehicle please refer to Campanella & Robertson(1981)2.2 TEST EQUIPMENTThe resistivity cone penetration test is a modification ofthe standard piezo cone test. Originally the test was performedwith a U.B.C. built 15 sq cm cone, around which the resistivitymodule was designed. Since then, the module has been paired withstandard 10 sq cm cones. Two different cones were used, a4commercially available one built by Hogentogler and a U.B.C.designed and built one. The cones are capable of measuring conebearing (Qc), local sleeve friction (Fs), pore pressure (U),temperature (T) and inclination (I) simultaneously. The U.B.C.cone has the ability to measure pore pressure at two differentlocations at the same time. In addition the cones are equippedwith geophones or accelerometers to record shear wave signalsfor determination of dynamic soil properties. The ability tomeasure the resistance to current flow in soils, has been one ofthe more recent developments in penetration technology at U.B.C.The five electrode array resistivity module is located behindthe piezo cone, as shown in figure 2.1. Analog cone andresistivity signals are sent up to the data logging system.2.3 THE U.B.C. RESISTIVITY MODULEThe resistivity cone shown in figure 2.1, is the latestmodification of this instrument by the U.B.C. in situ group. Itis an adaptation of the cone described by Campanella and Weemees(1990), with the resistivity module located behind a standardpiezocone measuring bearing, sleeve friction and pore pressure.There are five electrodes allowing resistivity measurements tooccur across four channels. The electrode separation varies from9.5 mm to 77.5 mm providing different depths of lateralpenetration. The concept is that the closer spacings willprovide resistivity measurements from soil immediately adjacentto the module, whereas the wider spacings will provide readings10cm2PIEZOCONERESISTIVITYMODULEU-’120QJ<E1II.DIMENSIONSINmm4536•CURRENTELECTRODEELECTRONICSFORRESISTIVITYPLASTICINSULATIONPROBEFigure2.1UBCResistivityConePenetrometer6from soil further away from the module. In addition, the smallerspacings allow for detection of finer layers. The module hasbeen modified since originally developed, with the addition ofone electrode, giving an additional two channels across whichresistivity can be measured. The inner three electrodes areberyllium—copper while the outer two are brass. The choice ofelectrode materials is important since the electrodes transmitand receive current. They should have stable and linearconductive properties at varying current levels. The two outerbrass electrodes are the current supply electrodes. They arealso used as measurement electrodes. Of the materials tested byWeemees (1990), brass offered the best compromise betweenconduction and wear characteristics.Electrode conduction is an important consideration in the designof a resistivity module, as is circuitry, since it is necessaryto have stable conduction over varying current densities. If thelatter is not achieved then the soil resistivity cannot beaccurately predicted from the measured voltage. Ultimately theperformance of the electrode and hence the module is related tothe thickness, width and composition of the electrodes. Thewidth of the current electrodes are 5 mm while the three innerelectrodes are 2.5 mm. The three inner electrodes are passivemeasurement electrodes. They too must have the characteristicsmentioned above for the supply electrodes, though typically thenon—linearity problems are not as severe as in the currentsupply electrodes. The required input current depends on how7conductive the soil is, so that lower resistivity soils requirea higher input current. It is necessary to maintain the signalat the measurement electrodes between certain limits to ensurethat an accurate voltage can be measured. The supply electrodesuse an AC source set at 1000 Hz, which can be varied inamplitude to vary the current supplied to the currentelectrodes. The resistivity module itself is powered using aconstant voltage power supply, separate from, the cone powersupply, to ensure a minimization of noise. For more informationon the design and specific electronics of the module the readershould refer to Weemees, (1990). All of the signal conditioningcircuitry for the piezo cone as well as the resistivity moduleis contained within the probes.2.4 DATA COLLECTION AND REDUCTIONData is logged using the U.B.C. in-situ testing digitaldata acquisition system. This system uses a 10 bit A/D convertorand is capable of simultaneously recording eight channels. Datacan be downloaded from the data acquisition system onto floppydisk for analysis. Before the resistivity data can be used itmust be reduced into a usable form. The resistivity data is atpresent recorded as voltages by the data acquisition system soit must be converted to resistivities. Furthermore, because thecentre of the resistivity module is located 0.8 m behind thetip, the depth of the resistivity measurements must be correctedso that they correspond to the same depth as the other CPT8measurements. A simple Fortran program was developed for thedata reduction, a listing of which can be found in appendix A.93. MEASUREMENT OF SOIL RESISTIVITY3.1 INTRODUCTIONGeophysical techniques are being increasingly used inpenetration testing technology for collection of geotechnicaldata. These methods are often superior to other methods in thatthey are fast, relatively non—invasive and do not requireextensive lab testing of soils. Compared to surface resistivitymethods, downhole or penetration type tests are more specificand require less interpretation. For the U.B.C. ResistivityModule, the electrode spacing is small (between 9.5 nun and 77.5nun) so that averaging occurs over a very small distance. Thevariable spacing allows for measurement across different sizedintervals providing different levels of focusing into theformation wall, and different levels of averaging in depth. Thegreater the spacing, the greater the penetration of the electricfield into the formation and the greater the path length of thefield. The combination resistivity module and electric cone isa displacement device which makes it ideal for soil resistivitymeasurement since the electrode contact with the soil isassured. The lateral stresses on the penetration devices arehigh enough and the soils are abrasive enough that theelectrodes are not able to transfer any appreciable soil or porefluid in the penetration process. This eliminates thepossibility of carry over effects. Because the resistivitymodule is located behind a piezo cone that measures cone10bearing, sleeve friction and pore pressures, there is excellentgeotechnical and stratigraphic information available from thesite in addition to the resistivity measurements.3.2 RESISTIVITY MEASUREMENTResistivity of the soil is not directly measured, rather itis inferred from the electrical resistance (R) of the soil.Voltage (V) is recorded across a receiver electrode pair givena constant supplied current (I). The resistance of the soil,from Ohm’s Law, is then:R=V/I (1)The resistance is not a fundamental soil property, butrather depends on the path length (1) and the cross-sectionalarea (A) of the effective resistive unit. Resistivity is afundamental soil property which is related to resistance by thefollowing relationship:p = (A/l)*R (2)In the case of the resistivity cone, with the assumptionsthat the soil acts as a homogeneous isotropic media, theelectrodes act as perfect conductors and the resistivity modulecircuitry acts as a perfect current supply source, then theratio A/l is constant. This ratio can be determined bycalibration of the module. The first assumption that the soil11acts as a homogeneous isotropic media is required for theanalysis since we cannot consider directional properties of thesoils in this test. Furthermore it is likely to be metreasonably well for the interval of resistivity measurement thatis being made. The latter assumptions become significantly inerror at particular measurement levels. This will be discussedbriefly in the following sections.3.3 CALIBRATION OF THE RESISTIVITY MODULEThe resistance of the soil as measured at the electrodes ofthe resistivity module is related to the bulk resistivity of thesoil by a factor that is dependant on the size and separation ofthe electrodes. This factor, referred to as K by Campanella andWeemees (1990), can be determined by performing a calibrationtest. The calibration test uses an ionic solution of KC1 torelate the resistance of the solution, as measured with theresistivity module, to the conductance of the solution, asmeasured using a portable conductivity meter (Omega CDH-3).Different concentrations of KC1 were used to establish therelationship between resistivity and resistance, referred to asK (where K = A/l). Conductivity, the inverse of resistivity, islinearly related to the concentration of dissolved salt, whenthe salts are composed of monovalent ions. However conductivityis also linearly related to temperature, as can be seen infigure 3.1. Because of these two factors, it is important thatthe solution that is being used for calibration is both well12500001/)0I400000tJ30004I.0200000ONwiooo.0C3000C-)C,)0! 2000IjJ0z.4I—C-)0z0C-)0(000L.0w0-TEMPERATURE (DEGREES CELSIUS)• Variation in the conductivity of an aqueous solution of KCI (10 molar concentration, very low) at several temperatures. Adapted from Hem (1959:98).Figure 3.1: The Effect of Dissolved Solids and Temperature on Specific ConductanceI I I I I500 000 1500 2000 2500 3000 3500DISSOLVED SOLIDS (UG/L)Variation in the conductivity of water samples from the Gila River at Bylas,Arizona, as the concentration of conductive colloidal particles changes.Adapted from Hem (1959:100).0 10 20 30 40(after Carr, 1982)‘3mixed and isothermal throughout the calibration procedure. Thelatter was accomplished by using a Laudia RX2O temperature bath.The temperature bath is equipped with an external pump, so thatcommunication can occur between the temperature bath itself andanother external vessel. The bath was connected in series withthe calibration tank as shown in figure 3.2. This arrangementallowed for accurate calibrations. A temperature of 10 degreescelsius was chosen for calibration since it is in the range ofseasonal ground temperatures in the lower mainland of Vancouver.It was recognized by the writer during early calibrationsthat the calibration factor K was not constant as previouslydescribed. In fact the calibration factor varied quitesignificantly with measured voltage, particularly at high andlow voltage levels. It also varied slightly with the level ofsupplied excitation voltage. Since the calibration factor issupposed to be constant (1/A) it became apparent that the moduleitself was having some influence on the measured results. Theobserved strong non—linearities at the high and low measuredvoltages can be explained by the following:1. The circuitry for the excitation current was designed insuch a way that grounding can take place to the cone body.The degree to which grounding takes place depends on thelevel of excitation voltage. Therefore at high excitationvoltages, required for measurement in low bulk resistivitysoils, a greater proportion of the total voltage is lost14Temperature ControlledBathCalibrationTankResistivity ConePenetrometerFigure 3.2: Schematic Diagram of Calibration Setup.15through grounding effects.2. At higher excitation current capacitive effects caused bythe AC current can become noticeable.3. In low bulk resistivity media the total impedance providedby the soil is within the same range as the impedance ofthe circuit so that the impedance of the circuit becomesnoticeable. Low measured soil resistivities correspond tolow measured voltages.The observed non—linearities were most noticeable for thelarge electrode spacings (current electrodes). Figure 3.3 showshow the calibration factors vary with excitation level andvoltage measured at the electrodes. This figure shows that thevalues of the calibration factors begin to increase considerablyat an electrode potential of about 1.5 volts. At about 0.5volts, the measured calibration factor starts to become muchlarger with any further reduction in electrode potential. Alsothe calibration factor appears to vary for different excitationlevels. This can in part be explained by points 1 and 2 in thepreceding discussion. The voltage dependence was accounted forin the data reduction program used by allowing the value of thecalibration factor to vary with voltage. A combination ofconstants and linear and power functions were used to describethe relationship between calibration factor and measuredelectrode potential. For more details on the calibration factors160.70 -9.5 mm ELECTRODE CALIBRATIONCONSTANT TEMP = 10 DEGREES CELSIUS8o A Q 0- 4o0o 00 08 0 0LLz - 00-HEXCITATION LEVELCO 00000 0.84 V RMS—J 2.72 V RMS..... 8.73 V RMS(-)0.50 — I I I I I I I I I I I I I I I I I I0.00 2.00 4.00 6.00 8.00ELECTRODE POTENTIAL (VRMs)15.0-- 77,5 mm ELECTRODE CALIBRATION- CONSTANT TEMP = 10 DEGREES CELSIUSC $I— •oC..)<1-(Th ILLj_ ‘-‘•-- 0z9 0 00 0F— 8 0 0•8 0 0Cc) 0EXCITATION LEVEL00000 0.84 V RMS(_) 888A 2.72 V RMSI.... 8.73 V RMS1 1 .0 I I I I I I I I I I I I I 1 1 I I I I0.00 2.00 4.00 6.00 8.00ELECTRODE POTENTIAL (VRMS)Figure 3.3 Variation in calibration factor with excitation level andmeasured &ectrode potentialused, the reader is directed to appendix A.17184.0 FACTORS AFFECTING RESISTIVITY MEASUREMENTS IN SOILS4.1. INTRODUCTIONThe electrical resistivity measured across the electrodesis the bulk resistivity of the soil and is referred to as theapparent resistivity. It is a combination of the resistivity dueto the soil skeleton itself as well as the resistivity of thepore fluid. In saturated soils the latter effect dominates. Thebulk resistivity of the soil is related to the resistivity ofthe pore fluid by the formation factor (F), which was originallydefined by Archie (1942) as follows:F= PbIPf (3)where Pb = bulk soil resistivitypf = resistivity of the pore fluidThe formation factor is in turn related to soil porosity byArchie’s equation:F = (4)where = soil porositym = constant, related to the shape ofsoil particles19This particular relationship holds for saturated porousmedia where the constituent particles are non—conductiverelative to the pore fluid. Normally F measured in the field istermed apparent formation factor, since there may be somesurface conduction taking place as well as other intergranularpore water effects that contribute to the measured bulkresistivity. The intrinsic porosity can be determined only ifthe soil matrix behaves as a perfect nonconductor.When looking at a bulk resistivity profile one sees theeffects of both the soil and the groundwater. While it isusually not possible to consider one variable of interest (soilor groundwater) exclusively, it is often possible to account forthe other and remove it in a somewhat qualitative way. Forexample, to focus on lithologic contributions it is oftenreasonable to make some assumptions with respect to groundwaterresistivity and hence exclude its effects from the analysis.Similarly it may be possible to exclude the effects oflithology. To accurately analyze the effects of either soil orgroundwater it is necessary to draw samples of the groundwater.Because of the constraint of permeability this will likely onlybe appropriate in granular soils. By applying a few interpretiveprinciples as well as some basic laws of groundwater flow,qualitative observations of variation in lithology andgroundwater chemistry can be made.20The sampling technique employed by the writer is the BATgroundwater sampling system (Torstennson, 1984). This systemuses a vacuum evacuated sample tube that draws a sample about 30ml in size in through a double ended syringe from a sealedsampling chamber. The technique employed to draw representativewater samples is as follows:1. Install the sampling tip to the desired test depth2. Wait 2 - 5 minutes for water to enter the sampling tip3. Draw the first sample and discard it (purge).4. Draw the next sample for testing.5. Repeat step 4 until sufficient volume has beencollected.6. Measure the resistivity of the sample.Usually 30 ml is sufficient to obtain a resistivity estimatewith the Omega portable conductivity meter. The sampling tipused was specially machined to have an outside diameter slightlygreater than the outside diameter of the cone and resistivitymodule. This ensures good contact between the sampler and theformation walls and minimizes the effort to install the samplingtip.4.2 GROUNDWATER EFFECTS ON MEASURED RESISTIVITYUnderstanding the evolution of groundwater permits one tomake estimates of how the resistivity of the groundwater can be21expected to change with depth in a soil profile. In theunsaturated or vadose zone the resistivity is very high becauseof the insulative effect of air and resistivity is inverselyrelated to the moisture content. In this zone it is extremelydifficult to qualitatively disassociate the effects ofgroundwater saturation and groundwater chemistry. This fact incombination with the fact that the majority of the test sites inthe Fraser Delta have water tables only a few metres below theground surface table have encouraged the writer to give only abrief consideration of partially saturated conditions.As a general rule deeper, older groundwater sources tend tobe higher in total dissolved solids (TDS) than younger,shallower ones (Freeze & Cherry, 1979). Generally, groundwatertends to evolve chemically towards the composition of sea wateras it flows further along it’s flow path. When surface waterinfiltrates the ground it continues to pick up dissolvedconstituents as it travels deeper into the soil. This fatehowever may be tempered by the lack of dissolvable constituentsor augmented by the sudden appearance of increasedconcentrations of dissolved constituents. In the case of theFraser Delta, upland freshwater recharge at depths as well asinfiltration of brackish water in coastal areas can producechanges in TDS that are greater than those normally expected.For example, of those sites tested in the western delta,proximity to the river usually translated into abrupt TDSchanges below 10 iti of depth. An example of this is the Arthur22Laing Bridge site on Sea Island. This effect is usually veryobvious, and is marked by a rapid smooth decline in resistivitywithin 1-2 metres in depth. Figure 4.1 shows a bulk resistivityprofile from the Laing Bridge site where a large change in bulkresistivity is noticeable at about 9 metres. The pore fluidsamples taken at 1 metre intervals demonstrate the stronginfluence of TDS (or salinity) change on the bulk resistivity.4.3 SOIL EFFECTS ON MEASURED RESISTIVITYSoil structure and chemistry play a lesser though importantrole in determining bulk resistivity. Changes in stratigraphyare often marked by changes in resistivity. Because themagnitude of these changes are often much smaller than thosecaused by groundwater chemistry changes, they tend to be moresubtle. In situations where groundwater resistivities are verylow, the structural variations in the soil and thus thestratigraphic variations become the dominant factor affectingbulk resistivity (Carr, 1982). Even at higher groundwaterresistivities qualitative changes can be detected. The apparentformation factor is intrinsically related to the soil type, andcan be considered constant for a given soil at constant voidratio. Figure 4.2 shows formation factor plotted with depth forthe Arthur Laing Bridge site. The general trend is for formationfactor to increase with depth. This is expected as usually soildensity increases with depth. This trend is evident for allelectrode spacings, although it is most pronounced for the230RESiSTIVITY (ohm—rn)510H0LU1520Figure 4.1: A comparison between bulk resistivity and pore fluidresistivity measured at Laing Bridge site.249,5 mm ELECTRODE SPACING26 mm ELECTRODE SPACING77.5 mm ELECTRODE SPACINGFigure 4.2: A comparison between formation factors calculated fromthree different electrode spacings at Laing 5ridge site.FORMATION FACTOR0 1 2 3 4I I I I I I i I I I I Ib‘I...5LOOSE TO MEDIUMLOOSE SAND05.10115-20F0LUMEDIUM DENSE TODENSE SANDFINER SILTY SANDMEDIUM DENSECLAYEY SiLT25largest spacing (77.5 mm). Changes in formation factor are dueto changes in soil characteristics, which can be a result ofeither change in density, or change in soil composition. Theformation factor is dependent on the physical structure of thesoil particles as well as any chemical contributions the soilmight have. As with hydraulic conductivity, electrolyticconductivity is controlled more by the mean pore size than bythe soil porosity. As a result coarser granular soils willexhibit lower resistivities than finer soils.4.3.1 EFFECT OF FINESThe above statements with respect to porosity vsresistivity relations cannot be extrapolated to plastic soils.These soil types contain significant amounts of conductiveconstituents such as clay particles, Similarly soils of glacialorigin often contain significant amounts of clay particles,causing their electrical properties to be difficult to predicton the basis of porosity alone. In general fine grained orplastic soils will produce a decrease in resistivity which isdemonstrated by the friction ratio resistivity relationshipshown in figure 4.3. Friction ratio (indicative of an increasein fines) is inversely related to resistivity. On figure 4.3this is noted by the fact that the highs in friction ratiogenerally line up with the lows in resistivity. The greater thepercentage of plastic fines present, the greater the percentageof clay minerals and hence the greater the surface conduction26RESISTIVITY (ohm—rn)0 10 20 30 40 50I I )_•__I I I I I i_.._I........_i__i I-..--:.-:-.-i--=-..‘S-7‘S.-F---- FRICTION RATIO (%)RESISTIVITY (ohm—rn)-‘-7—--•-I I I I I I I I I I1 2 3 4 5FRICTION RATIO (%)Figure 4.3: A Comparison of resistivity (77.5 mm spacing) andfriction ratio at the U.B.C. Pile Research site.---=——=— — —-.--------------------.——-‘.-,..-_____-S. — —EU1O12141618—2O-2224-0‘¼--- -‘C127effect. An increase in surface conduction corresponds to adecrease in measured resistivity.4.3.2 EFFECT OF ORGANIC MATERIALSThe presence of any organic material in soils may alter thesoils resistivity by increasing the osmotic potential andreducing the free energy of the water (Carr, 1982). Ions thatare released into solution during the decomposition process tendto sequester water particles and remain attracted to the organicmaterial. This effect produces osmotic suction that tends tomaintain a high water content in these soils. Also organicmatter tends to promote soil aggregation which increases thesoils’ porosity and its moisture holding capacity (Carr, 1982).Thus one could expect that soils having high organic contentswill tend to be higher in water content and higher inconcentration of dissolved ionic constituents. Both of thesefactors will contribute to a lower resistivity compared to othersoil types.4.3.3 RESISTIVITY MEASUREMENT IN SANDSSince sand particles themselves are non—conductive, theflow of current through saturated sand must occur byelectrolytic conduction. Hence the bulk resistivity measured insaturated sands is largely a function of the pore waterconductivity and the amount of conductive pore space. The amount28of conductive pore space is influenced by a number of soilrelated factors such as porosity, tortuosity of pore space aswell as grain size and shape factors (Urish 1981). It wasrecognized by Campanella & Weemees (1990) that there was arelationship between cone bearing, normalized for the effect ofchanges in effective overburden stress, and formation factorwhich suggested that the measured bulk resistivity must berelated in some way to density and therefore to porosity. Figure4.4 shows the relationship obtained from the two electrodespacings that were used. It was observed that for the largeelectrode spacing a relationship existed between normalized conebearing and formation factor. Figure 4.5 shows a portion of aresistivity profile from the Laing Bridge site presented overtop of the cone bearing results. Referring back to figure 4.1 itis clear that below 10 in the groundwater resistivity isconstant. However figure 4.5 indicates that there are somenoticeable changes in the bulk resistivity profile measured withlarge electrode spacing which are paralleled by changes in conebearing. These changes are not related to changes in plasticfines content, as the soil is a clean sand. Given that the porefluid resistivity is constant it is possible to look at bulkresistivity changes as a function of changes in cone bearing.Figure 4.6 shows resistivity as a function of cone bearing forthe 9.5 nun and the 77.5 mm electrode spacings. A relationship isobserved for the large electrode spacing but not the smallelectrode spacing. This is consistent with observations byCampanella and Weemees in figure 4.4.290z00IzLJa.a04.03.93.837.3.63.53,,5McDONALD FARM SAND (6 to 14.5m)?.LLU OUTER ELECTRODESQ9O INNER ELECTRODES100CONE SEARING / HORIZONTAL EFFECTIVE STRESS (Ko=O.55)Figure 4.4: Observed relationship between apparent formationfactor and cone bearing normalized with respect tohorizontal effective stress (after Weemees, 1990).RESISTIVITY10CONE BEARING (bar)Figure 4.5: A Comparison of Resistivity (77.5 mm spacing) andCone Bearing at Laing Bridge site.306(ohm—rn)0 5 15 20 25-I.—-I—CONE BEARING (bor)RESISTIVITY (ohm—rn)8/-——‘————10 —0w12114—16 liii75 100 125 15010-2-77.5mmELECTRODESPACING-9.5mmELECTRODESPACINGci-DOXJ•ciom-9am-oociO00•DOOIIDUIJIDODIXIDI-•cmEIcocaE00000•cm000CI8-ooCcitiDm03•CI000-0000-Dci0.aacia>--0ciI—-0E11ci009]—.0ci:icoo>0000cia>N_8RtB0F-O00-aaDCl)-a-a1Li0LUaOa60-0-a00a-50___________________________111111111111111iiiiL0iii111111IIiiiiioioo2000100200CONEBEARING(bar)CONEBEARING(bar)Figure4.6:AComparisonbetweenresistivityandconebearingforthelargestandsmallestelectrodespacingsfromLaingBridgesiteinsand,whereporefluidresistivitiesareconstant(10-14iu.)324.4 SUMMARYArchie’s equation relating formation factor to soilporosity is mainly valid in sandy but not clay soils. Clay soilscontain polarized clay minerals which themselves can exhibitconsiderable conductivity. In this case the formation factormust be corrected for surface conduction effects if porosity isto be inferred. In reviewing the literature, it became apparentthat while application of Archie’s formula was widespread, therewere bounding conditions on its validity. Some researchers havesuggested that while Archie’s equation may be valid for brinefilled rocks in oil field interpretation, it does not hold forfreshwater sands (Campbell & Lehr, 1973). In contrast, studiesby Jackson et al (1978) showed that for varying marine sands thevalue of m is between 1.39 and 1.58. This range of values for inindicates that F is largely related to soil porosity, arguing infavour of the validity of Archie’s equation. Urish (1981)contends that the porosity can be related directly to F in bothmodels and field tests for clay free fresh water glacial outwashdeposits. However the formation factors that are measured aresensitive to the resistivity of the pore water. Thus theformation factor as described by the Archie equation is not aunique formation parameter. It is believed that the effects ofsurface conduction become more important at lower pore fluidresistivities (Urish 1981), therefore Archie’s equation appearsto be sensitive to changes in pore fluid resistivites.33Most of the sands in the Fraser Delta that were examinedhad bulk resistivites that ranged between 5 and 50 ohm-rn. Giventhis range of measured resistivites, it is prudent to cautiouslyassume the validity of Archie’s equation to deltaic sanddeposits considered herein.345. USE OF THE RCPT TO DETERMINE SMD BEHAVIOURBecause of the dependence of measured bulk resistivity onporosity it has long been believed that there should be some wayof exploiting this dependency to measure density. Delft (1982)built what they called a density logging cone, which usedresistivity measurements from reconstituted density tests aswell as field tests to directly estimate in situ sand density.This approach was not always successful. While sand behaviour iswell understood from laboratory testing, current in situtechniques cannot provide a purely analytical basis for lookingat soil behaviour. Because of the difficulty in obtainingrepresentative samples from granular soils, they cannot betested under field conditions. Hence in situ behaviour of sandshas been a subject of some confusion and even speculation forgeotechnical engineers. While there is a reasonableunderstanding of soil behaviour in the laboratory, the use ofoften crude index tests as gauges of field behaviour continues.5.1 PREMISE OF RESEARCHIt is understood by previous researchers that formationresistivity or bulk resistivity is related to pore fluidresistivity, soil type, porosity and tortuosity. Furthermorecone bearing can be related to formation factor, which can bemathematically related to porosity in clean well rounded sands.Therefore it was believed that formation factor could be used as35a quantitative means of determining sand density.Dense sands show higher bulk resistivity values than loosesands in a pore fluid of a fixed resistivity due to theirsmaller pore space per unit volume. Because the resistivitymodule is a displacement device, soil around the probe issheared during penetration of the probe. The degree ofdisturbance decreases with increasing radial distance away fromthe probe. For the RCPT, one would expect that an electrodespacing capable of allowing the current path to pass into theless disturbed formation would be necessary to obtain a measureof relatively undisturbed or “far field” bulk resistivity. Thisrequires as large a spacing as practically feasible. Smallelectrode spacings would in general measure resistivities nearthe instrument, in the “near field”, where the soil has beensheared considerably by penetration. In sand one would expectthat the density in this near field zone of severe remouldingand shearing would be relatively constant, and close to criticaldensity. The critical density is the density at which shearstrain occurs at constant volume (Lee and Seed, 1967). Henceonly the formation factor calculated using the large electrodespacing should be related to density and cone bearing. This isdemonstrated quite clearly in figure 4.3. If pore fluidresistivities were to remain constant, then bulk resistivitywould vary the same as formation factor. Bulk resistivity wouldbe expected to vary with cone bearing only for the largeelectrode spacing. This was observed in figure 4.5 and36demonstrated in figure 4.6.The formation factor approach to studying density in sandsis limited by the fact that it requires water sampling and porefluid resistivity determination. Getting representative watersamples for formation factor calculations is by no means atrivial exercise. Sampling can be very time consuming sinceformation permeabilities limit the flow of fluid into thesampler. A single sample may take up to a half hour to obtain.The resultant formation factors are only representative ofdiscrete points in the soil profile so that the data setcollected may be incomplete for a particular analysis.Furthermore there are errors associated with the measurementitself. Handling the sample and bringing it into a differentenvironment can change the temperature and hence the resistivityof the sample. There are also errors associated with theequipment used for conductivity measurement. While in principleformation factor determination can be used for estimation ofabsolute density by Archie’s equation, it has proven to be quitedifficult in practice. Extensive site specific correlations arerequired for this to be achieved. In the Netherlands thisapproach was used by Delft (1982) with very limited success.Another approach to the understanding of in situ density insand has been developed by the writer. The remainder of thischapter will describe the methodology used. Rather than tryingto look at the problem of measuring absolute density and37comparing it to reference densities of sands, the writer hasundertaken to consider only the density or volume change thatoccurs during shear. One way of interpreting the volume changecharacteristics of the sand during penetration induced shear isto compare the formation factors between the inner and outerelectrodes. However this has no particular advantage overcomparing the resistivity values from the inner and outerelectrodes themselves. The latter approach has the advantagethat water samples are not necessary. By comparing near fieldresistivity to far field resistivity, it is possible to observethe relative differences in sand porosity or density between thenear field and the far field. Knowing the differences allows oneto determine the in situ volume change behaviour of sands. Ifthe assumption is made that the disturbed sand near the probe isat or near critical state, then one has an in situ referencedensity and corresponding bulk resistivity. Because denser sandswill have higher bulk resistivities than looser sands,comparison of near and far field bulk resistivity shouldindicate whether the soil has become looser or denser near theprobe. Therefore sands can be identified as contractive ordilative based on their volume change behaviour duringpenetration induced shear.It is important to reaffirm the assumptions that allowthese statements to be made. Firstly we must assume that themedium which the electrodes are contacting is laterallyhomogeneous and to some extent vertically homogeneous. This38applies to both soil and groundwater. If these assumptions areinvalid there may be problems in comparing the relative valuesof bulk resistivity for the larger and smaller electrodespacings. While lateral homogeneity is generally a reasonableassumption, vertical may not be. The smaller the electrodespacing, the higher the chance that very small stratigraphicvariations will influence the bulk resistivity. While smallvariations can be important in the near field, they may beundetectable in the far field where averaging is taking place byvirtue of the relatively long current path. Therefore some caremust be taken to ensure that the differences due to scale can beaccounted for.5.2 CONSTRAINTS ON THE JNALYSISAs indicated above there is a relationship between porosity(density) and resistivity, however the relationship is nonunique. It is applicable when resistivity is measured in soilswith a non—conductive soil skeleton, or soils that obey Archie’sLaw. Soils that have fines can have a significant proportion ofthe applied current conducted by surface conduction rather thanelectrolytic conduction. As a result, the measured resistivitycan be effected not only by variation in porosity, but also byvariation in fines content. Without any way of accounting forchanges in resistivity due to the addition of fines, it would bedifficult to account for changes due to density alone.Furthermore penetration must be drained so that volume changes39are permitted to occur. If penetration is undrained, volumechanges will be opposed by pore pressure development.The initial design of the U.B.C. resistivity module was topermit it to operate within a large range of resistivity valueswith as high a degree of accuracy as possible. Because oftenthere is a compromise in the design of measurement circuitrybetween range and resolution, a primary design for such a modulerequires that both are optimized. The range and resolutionchosen for operation was one that could accommodate mostsedimentary geological conditions that could contain eitherconductive or insulative contaminants. The module was designedto respond to contrasts in resistivity either as a result ofconductivity increases or decreases. During development theobjectives of the research were such that there wasn’t a needfor the resistivity response to be highly accurate as long as itwas observable.The accuracy of the bulk resistivity measurements madedepend primarily on three factors: the resolution of the A/Dconvertor used in the data acquisition, the absolute value ofthe measured voltage and the accuracy of the calibrationfactors. While the A/D resolution was satisfactory and themeasured voltages could be controlled by adjusting the suppliedcurrent, it was found that the calibration constants describedby Weemees (1990), which were used to convert the measuredvoltages to resistivity, were not truly constant. As discussed40previously, defining the true calibration values over the rangeof measured voltages was a difficult exercise. After a precisioncalibration it was decided that the calibration factors could bedescribed by a relatively simple mathematical expression thatfitted the calibration data. For three of the spacings used thecalibration factor could be approximated as piece—wise linearover particular voltage range. A polynomial function betterdescribed the current electrode (77.5 mm spacing) calibrationfactor. The need for such precise measurement arose in thecourse of the authors research, when it was recognized that thedifferences in resistivity due to slight differences in porositycan be very small. As a result changes in resistivity could beobscured by the inaccuracy of the data.Due to the constraints imposed by soil type and resistivitymodule design, it was decided that the research objectivesshould be constrained such that only clean sands, having bulkresistivity values that fall between the accurately measurablerange are analyzed. It should be noted that while mainly cleansand behaviour will be discussed, the method described could inprinciple be applied to any soil type to show volume changebehaviour.5.3 THEORETICAL BASISDilatancy refers to the volume change that sand undergoeswhen it is sheared. The volume change characteristics of sands41have been understood since 1885, when Reynolds, who not knowingthe significance of this observation, showed experimentally thatdense sands dilate when sheared. It was Cassagrande who in 1936showed that both the peak friction angle and the volume changebehaviour of sands depended on the initial void ratio or densityof the sand. Naturally occurring sands can have substantialdifferences in their dilatancy characteristics. Variations inmean grain size, grading, mineralogy and in situ stress can allinfluence the dilatancy behaviour. These observations have beenrepeatedly demonstrated under various laboratory conditions.Steady state concepts are the most recent framework in whichdilatancy behaviour has been discussed. The basis for this typeof analysis is that sand has a steady state line which exists invoid ratio — effective confining stress space (Vaid et al,1990). The steady state line can be determined from triaxialtests in the laboratory. The position of a sand element withrespect to the steady state line determines its behaviour whensheared. It is well understood that increasing void ratio ordecreasing density promotes volume decrease or contractionduring shear. Similarly decreasing void ratio or increasingdensity promotes volume increase or dilation during shear.Furthermore increasing confining stress promotes volume decreaseor contractive behaviour. In practical engineering problemswhich involve shallow depths, density or void ratio is usuallythe more important variable. Hence in situ density has becomethe crucial parameter in determining field behaviour. Theposition of the steady state line described is different for42sands of different composition. Moreover even for a given sand,such as Ottawa Sand (ASTM C-log), the position of the steadystate line is also stress path dependant (Vaid et al, 1990).Therefore it appears that steady state concepts must be appliedin the context of the specific soil tested and the direction ofloading.Evidence for volume change behaviour in situ has been moredifficult to demonstrate than in the laboratory. Various indextests, both static and dynamic, have been used to establishempirical relationships between some type of measured fieldparameter and dilative or contractive behaviour given fixedloading conditions. While steady state concepts have beendiscussed mainly in the context of laboratory tests, recentlythey have been extended to in situ tests (Been et al, 1987,Sladen and Hewitt, 1989). In their state parameter approach,Been et al relate CPT cone bearing to the state parameter whichis used to define the contractive dilative boundary or steadystate line. The state parameter approach is an attempt to relatea field index value (CPT cone bearing) to a model based onbehaviour in triaxial tests (Been et al, 1987). The stateparameter is defined as the difference between the void ratio ofthe sand at its initial state (before testing) and the voidratio at steady state conditions at the same mean effectivestress. CPT cone bearing is related to the state parameter asdetermined in laboratory tests using chamber test results. Thestate parameter concept is attractive since it seems to be able43to relate soil behaviour determined in the laboratory to a veryreliable in situ test. It is also attractive since it relatesthe volume change characteristics of sand to a single parameterwhich incorporates both the effects of void ratio and meannormal stress level. The inherent shortcoming of this approachlies not in the theory or methodology adopted by the authors,but rather in the implementation of their approach. The methodrequires the independent assessment of the steady state line andin situ horizontal stress. While the estimation of in situhorizontal stress can usually be done, determination of thesteady state line requires extensive laboratory testing to beconducted on representative samples. Sladen and Hewitt proposea contractive dilative boundary for sand, based on cone bearing,from back analyses on hydraulic fill structures in the BeaufortSea (1989). The authors examined CPT data from various fillstructures and were able to separate data sets from structuresthat failed from those that did not. The boundary separatingthese data sets was inferred to be the contractive dilativeboundary. Because both these methods rely on empiricalrelationships that were extracted from very specific data bases,the potential exists for misinterpretation of new data.The methodology proposed herein by the writer differs fromthe above in a fundamental manner. The volume change associatedwith shearing is observed directly. Therefore there is no needfor laboratory or field correlations to predict dilatancy insitu. Dilatancy can be estimated by looking at the volume change44that has occurred during shear at various distances away fromthe penetration probe. The level of shear induced by thepenetration of a probe, such as the cone used in the CPT,decreases as the radial distance away from the probe isincreased. Chong (1988) describes the volumetric strain as afunction of radial distance to be a critically dampedcompression wave. Figure 5.1 shows the relationship derived byChong, who did his work in a chamber, using thermistors tomeasure density changes radially around the probe. The datapresented is for medium dense and dense sand. The difference inthe amount of dilation, indicated by volumetric strain, betweenthe medium dense and dense samples is substantial. Thepenetration of the electric field from the resistivity module isproportional to the spacing, so that by varying the spacing onecan expect to vary the level of volumetric strain that ismeasured. Resistivity methods have been used before to measurevolume changes in granular soils. Windle and Wroth (1975) usedresistivity to determine the volume changes that occur during apressuremeter test. They used a specially equipped pressuremeterthat had electrodes placed on the outer membrane. The tool wasdeployed into a pre-cut hole and then back filled with sand. Aconstant current AC source was used as with the U.B.C.resistivity module. The authors were able to measure not onlythe dilation and contraction during expansion of thepressuremeter membrane, but also the dependence of measuredstrain on electrode spacing. Their results are presented infigure 5.2. Curve A represents expansion with an electrode45}EEjR:EECotnprsio -:r’Ez:::z[‘IIJ 0 fxpt volu._j— Way. !unction 1’ A • 0.11 eL. c00t2’ £d’—ooe, x,,”17p.173) kglrn , Stag. 4— a Ept yak..—— Wav lunction 4.J1.0.0608— 8c00’ • d —0.0125 , *— p.l5lekgIrn3,Stog.4I ii i 11 I Iii I__ II.12 4 • 8 10 1214 18 182022I.Iativ. radial distonc.Figure 5.1: Volumetric Strain vs Relative Radial Distance fromPenetrometer Axis. (after Chong, 1988)+ 0.03— 0I46vol change (%)812 16fO//0PLOTS OF VOIDS RATIO VERSUS STRAIN FOR TWO VALUES OF XFOR THE DENSE TESTFigure 5.2: Measured Dilatancy Behaviour of a Dense Sand Duringa Pressuremeter Test, Using A Resistivity Method(after Windle & Wroth, 1975)8400 4PLOTS OF VOLU€ CHANGE VERSUS STRAIN FOR TWO VALUES OF xFOR THE DENSE TESTA.04O.30 4 8 12 16For Electrode Spacings X: A = 1 inchB = 1/2 inch47spacing of 1 inch, while curve B was for a 1/2 inch spacing. Theauthors suggest that the level of “bulk” current penetrationinto the formation walls is about equal to 1 to 1.5 times theelectrode spacing. The term bulk current is used since thecurrent distribution tends to be quite broad, though the bulk ofit passes between certain limits. When the U.B.C. resistivitymodule was modelled the findings were concurrent with those ofWindle and Wroth. Figure 5.3 shows the current distributionresults from an electrical analog model for the currentelectrodes of the U.B.C. resistivity probe. The lines thatintersect the plane of the module (with electrodes) at rightangles are equipotential lines or lines of equivalent voltage.The lines that are perpendicular to these are the flow lines.The quantity of current flow (dl) between any two flow lines isalways the same (d11 = d12 = d13 .. etc.). The sum of all thecurrent discharges (dl’s) must always be equal to the totalcurrent supplied. Because there is no fixed boundary controllingflow as radial distance from the probe increases, theoreticallythe applied field extends infinitely. In reality the fieldbecomes extremely weak as distance increases beyond the outerflow line bounding the current discharge d14, at about half theelectrode spacing. Hence significant flow does not occur atlarger distances. Soil heterogeneities and stress induces volumechanges will cause the true conductivity of the soil to be nonuniform around the probe. This non uniformity will cause adistortion of the field in the direction of higher conductivity.The modelling was done with a DC source so that 1000 Hz AC48Figure 5.3: Flownet from Electrical Analog Model ShowingDistribution of Current with Distance From theResistivity Module.49effects, if any, cannot be observed. Nevertheless the modelprovides a useful gauge of current distribution in the soil. Thesmaller electrode spacings will see a smaller voltage dropcompared with the larger spacings and will cause discharge ofcurrent flow that is closer to the probe. Similarly largerspacings will cause discharge of current that is further awayfrom the probe.5.4 DILATION PARAMETERIn order to have some way of quantifying the dilatancybehaviour it is necessary to establish a parameter that isrelated to the amount of dilation or contraction that isobserved during penetration induced shearing. For this reasonthe parameter D, the dilation parameter, was chosen. The valueof D is defined by the writer as follows:D= Pbo95IPbv5 (5)where: 0 = dilation parameterp = bulk resistivity measured at 9.5mmelectrode spacingPbv5 = bulk resistivity measured at 77.5mmelectrode spacingThe significance of the dilation parameter and its relationshipto the formation porosities can be seen in the following:501. define the formation factors for the two differentelectrode spacingsF095= P1,o95/Pwo95 = (O95) formation factor @ 9.5mmF5 Pb775/Pw775 = m(n formation factor @ 77.5mm2. express the dilation parameter D in terms of thegroundwater resistivities and formation porositiesD = ( ) /(mm)Pwv5)3. but since the formation characteristics and pore fluidresistivity can be considered constant at a givendepth point the equation can be simplified to thefollowing:D=95)/(775) (6)Hence the dilation parameter reduces to the ratio ofthe porosities.The dilation parameter was defined in this manner since ofthe four electrode spacings, the 9.5mm and the 77.5mm spacingsrepresented the smallest and the largest available on the moduleat the time of the writers research. This allows for comparisonof the largest variation in soil disturbance. At the early51stages in the development of the technique it was necessary tomaintain the highest contrast possible, since there wereinaccuracies in measurement resulting from the limitedresolution of the module. Also fine tuning of calibrations hadto be done to ensure the field observations remained consistent.While it is possible to use a different set of measurementelectrodes to calculate a different D, the variation in responseof the electrodes in the soil becomes smaller. The ratio of theresistivities is the signal being measured. As the signalbecomes smaller the signal to noise ratio increaseproportionally. Because measurements at two different electrodesare required, there are two sources of noise. When adding twomeasured responses with random noise the signal tends to beamplified with respect to the noise. Because we are dividing oneresponse by the other the random noise can be amplified. The useof a quotient term as the dilation parameter was chosen since itimplicitly normalizes with respect to the bulk resistivityvalue. In this way variation in the pore water conductivity doesnot have an effect on the value of D, provided the assumption oflateral homogeneity in groundwater is met. Hence there is noneed for pore fluid sampling. By increasing the spacing it maybe possible to increase the penetration of the electric field tothe point where truly undisturbed formation resistivities can bemeasured. Since the large and small electrode spacingresistivity measurements have different effective path lengths,they have different vertical as well as horizontal resolutions.The smaller electrode spacing can respond to thinner soil layers52than can the larger electrode spacing. This can produceperceptible differences in very stratified soil deposits such asin the Fraser River Delta. In order to overcome this, the Dvalues calculated were treated with a seven point smoothingfunction. Seven point smoothing was chosen since the ratio ofthe larger to smaller electrode spacing was about seven. Thesmoothing also helps to remove some of the high frequency noisethat is observed.Upon examination of the in situ methods available it isclear that a need exists for developing some type of analyticaltechnique for understanding dilatancy behaviour of sands. Whilethe pressuremeter shows great promise as the definitiveanalytical tool for in situ testing, it cannot match the RCPTtest in terms of speed and effectiveness. Furthermore thepressuremeter requires interpretation of a stress strain curveto extract comparable dilatancy information. In order todemonstrate the usefulness of the test, a field program wascarried out at several research sites around the lower mainland.536. FIELD INVESTIGATION PROGRAM AND RESULTS6.1. DESCRIPTION OF FIELD INVESTIGATIONIn selection of field test sites there were severalimportant criteria that had to be satisfied. Firstly they had tobe amenable to cone testing. Because of Vancouver’s proximity tothe Fraser Delta, the U.B.C. in situ geotechnical group hasaccess to several good research sites in the lower Fraser Delta.These sites offer a good variation of soil types. The lowerFraser valley is especially rich in fluvial sands, deposited indifferent environments by a prograding Fraser River. These sandsshow some variation in density, grain size and grading dependingon the specific depositional history of the site. Of mostconcern to designing geotechnical engineers are the loosepossibly liquefiable sands and non plastic silts that tend to befound between five and fifteen metres of depth.The present Fraser Delta is a consequence of the latestphase of deposition from post glacial ice ablation and FraserRiver scour. The Fraser Delta consists of a complex sequence ofglacial, galciofluvial, glaciomarine and fluvial sediments thatextend up to several hundred metres in depth. The most recentsediments are the fluvial sediments, resulting from continuouscycles of scour and deposition caused by flooding and channelrelocation. The locations of the U.B.C. research sites that werevisited can be seen in figure 6.1.I-i.CD H(DP)CD C) rtCD CD‘1 CD H 3 C,) 0 0 C) 0 0 H)U-’556.2 ARTHUR LAING BRIDGE SITEThe site is adjacent to the southern off ramp of the ArthurLaing Bridge at the Vancouver International airport, as shown infigure 6.2. It is at the north eastern limit of Sea Island nextto the North Arm of the Fraser River and in relatively closeproximity to the present delta front of the Fraser River.The RCPT profiles from one of the two closely spacedsoundings carried out are presented in figure 6.3. The top 4metres of soil consist of organic rich silts and loose siltysands as indicated by the high friction ratios. Below this is asequence of clean sand extending to about 14 metres. This sandshows a general trend towards increasing density with depth.There is a major change in cone bearing at seven metres, wherea large increase is observed. Both the friction ratio and thedynamic pore pressure response confirm the fact that the sandbetween 4 and 14 metres is clean. Below 14 metres the sandbecomes quite silty in layers. The resistivity profile indicatesa moderate level of total dissolved solids in the groundwater inthe upper 10 metres of the deposit. The approximate limit ofpotable water is 12 ohm—in which corresponds to about 500 mg/ltotal dissolved solids. A sharp change in the bulk resistivityis noted at 10 metres, at which point the resistivity of thegroundwater halves. The fact that the groundwater is controllingthe bulk resistivity response is evidenced by the parallelsbetween groundwater and bulk resistivities seen in figure 6.4.5‘ISCA L EAPP !‘)<‘IooO 4_; 44,t4 L.vFFTLf7H Aew-dzJ7-Figure 6.2: Arthur Laing Bridge Re5earch SiteUrt5U,a)tn‘VC(15C4J4E0--44JU0’•—I‘Vt,CCordc_) -cn57CIn -LU8 -LI,rI-a-1Akzwe1,,‘.0a)C0’(SJ5’.eW) H1d30585E10Li1520Figure 6.4: A Comparison Between the Smallest Electrode Spacing,the Largest Electrode Spacing and Pore FluidResistivities Measured at Laing Bridge Site.0RESISflVITY (ohm—rn)2059The latter figure shows both the large electrode (77.5mm) andthe small electrode (9.5mm) spacing results plotted on top ofone another, with the groundwater resistivity samples plotted asdiscrete points. At the scale provided it is apparent that theresistivity values for the two spacings plot very close to oneanother and exhibit the same basic trend. Both measurements arecontrolled by the groundwater resistivity. On closer examinationhowever slight differences between the two electrode spacingsare visible, especially in the region where the groundwaterresistivity flattens out (10 to 17 rn). The implications of thesedifferences is the basis of the proposed method for dilatancycharacterization. Looking back to figure 6.3, after the lowersilty sand at about 19 metres, there is a clayey silt unitcontaining some thin silty sand lenses. The resistivity measuredin this unit is of particular interest. A bulk value of 25 - 30ohm—rn, compared to about 6 — 8 ohm—rn in the sands above. Thisresponse is atypical of a sand to clayey silt transition. Aspreviously pointed out, surface conduction in the clay mineralphase of the soil usually results in a decrease in bulkresistivity. The equilibrium piezornetric level in the silty clayhas not been well defined with depth, so while an uplandsrecharge of freshwater is a possible explanation, this cannot besubstantiated.606.3 MILLER ROAD SITEThis site is approximately 400 meters south of Laing Bridgesite, also on the eastern limit of Sea Island. Figure 6.5 showsa site plan with the sounding locations as well as a side viewof the fill. It differs from the above site in that it was aproposed routing for the Sea Island to Lulu Island connectorbridge and has a sand fill up to 6.5 m above grade on the site.This fill has been in place for about 20 years according toMinistry of Highways sources. The fill has acted as a preloadcausing settlements in the native soils beneath it. Taking thesand unit weight at 19.5 kN/m3, the surcharge load caused by thesand pile 6.5 metres thick would be approximately 125 kPa. Theeffect of this on the behaviour of granular soils beneath it isbasically twofold. Firstly there will be some compaction inthese soils, especially those that are compressible and close tothe original ground surface. Densities of these soils willtherefore increase. Secondly the vertical effective stresses andthus confining stresses at depth will increase. The effect ofincreased confining stress can be a significant factor incontrolling soil behaviour at depth.Figure 6.6 shows one of the RCPT profile from this site.The sand fill in the upper 6.5 in of the profile is typified byvariable cone bearing arising from the placement technique.Since the sand berm was brought up to final height by applyingseveral smaller ‘lifts’ the density is expected to be variable,Figure 6.5:SLC,T ION A-A’NIL Ev ‘T3ALL.WQQ61—--.AItMiLLR OAOA’Miller Road Research Site624-JC)C’,ci)zLiiciíaLI)lflLi)ía0a)—404-i4-I0’“-I0CfIC-)cD4-I“-Iai-LULUoClI—U, ILUII-.,IzH1d30(SJ.ew)63with the top of the lift in general receiving greater compactionenergy and therefore showing greater density. Directly under thesand pile are the overbank silts and silty sands characteristicof this part of Sea Island. These deposits are underlain by asequence of sands with silt layers. The resistivity profiledemonstrates the present water table level to be at about 6.5metres below the top of the fill, at about the same level as thefirst native soils are present. As indicated these soils aresilty with some clay, causing the resistivity profile toapproach the 10 ohm—rn level. Once into the cleaner sands thebulk resistivity appears to be on average 40 ohm-rn through untilabout 18 metres where the resistivity drops. This drop as in theLaing Bridge site, is controlled by the changing pore fluidconductivity. The bulk resistivities are in the same range asthe Laing Bridge site as expected since the distance between thesites is less than 400 m.Since most of the Fraser Delta is only several metres abovetidal groundwater levels, with most of the crustal soils beingoverbank silts and clays, it is rare that a thick sequence ofsand is found above the water table. Because of this the writerhas not considered partially saturated conditions in theresearch findings presented so far. This site presents a uniqueopportunity to look at relatively uniform dredged river sand atvarious placement densities. Figure 6.7 shows the resistivityfrom 0 to 5 m along side the cone bearing. It is quite evidentthat a relationship exists between resistivity and cone bearing.RES)STIVITY (ohm—rn)0 500 1000 1500 20000—I I I I I I I I I I I I I I I Il_I I I I I1—2-LU -3-4—CUNh. ARING (bar)RESISTIVITY (ohm—rn)__,I I I I I I I I I I I I I I50 100 150 200CONE EARINC (bar)Comparison of Bulk Resistivity and Cone Bearingin the Vadose Zone at Miller Road Site.64/——‘———1—450— 4250Figure 6.7: A65The relationship is an inverse one, with high cone bearingcorresponding to relatively low values of bulk resistivity. Thisis consistent with what is expected. Denser sands will havesmaller void ratio and smaller effective pore size compared withlooser sands. Because of this denser sands have an increasedcapacity to retain water by capillary suction. Therefore theywill, have higher moisture contents in the unsaturated zone thanlooser sands. Higher moisture contents are evidenced in figure6.7 as lows in the resistivity profile.6.4 ALEX FRASER BRIDGE SITEThis U.B.C. research site is located on Annacis Island inNew Westminster B.C. Figure 6.8 shows a site plan with thelocation of the test soundings. The site includes the north pierof the Alex Fraser Bridge as well as the northern approach spansupport piers. The Alex Fraser Bridge was completed in 1986.During the preconstruction and the construction phases of theproject an extensive site investigation was conducted. As aresult it is perhaps one of the most tested sites in the lowermainland, and most studied from a liquefaction point of view.There has been extensive vibro—replacement densification doneunder the main pier as well as under the approach span piers.Annacis Island is located upstream from the previous sites asshown on figure 6.1. It is a smaller channel island, similar toother islands upstream from the mouth of the delta. It mostlikely originated as a sand bar and with continuous deposition66OiO 5 IoOPPox. ScM-EFigure 6.8:)RASI. CPT/NFRA54. cp-rASZ. aprAlex Fraser Bridge Research Site67and seasonal flooding attained its present form.Looking at figure 6.9 one can see that the site ispresently covered with sand fill to a depth of 3 metres. Beneaththe sand fill there is approximately 1 metre of silty overbankdeposits which are underlain by loose sand and silt layers to adepth of 7 metres. Beneath 7 metres the sand is more or lessclean, though occasional thin silt lenses can be found. The siltlenses are marked by sharp decreases in bulk resistivity. Thesands are characteristically loose near the surface throughoutthe site, hence vibro—densification was recommended by thegeotechnical consultants for the main span portion of theproject (Bazett & McCammon, 1986). The bulk resistivity at thesite shows a relatively low level of total dissolved solids inthe pore fluid. The sand up to a depth of 15 m is fairlyuniform. Below this depth it gets quite stratified compared toother sites. The stratigraphic variability is reflected in theresistivity response. A total of four resistivity soundingswhere conducted at this site.6.5 U.B.C. PILE RESEARCH SITEThis site is located at the eastern end of Lulu Island inRichmond, as seen in figure 6.1. Much of eastern Lulu Island iscovered with peats and other organic rich sediments. Thesesediments accumulate in slow moving organic rich waters. Figure6.10 shows the site plan and test locations at the site.Ua)I)a)a)xci)ECE-a)Ua)‘-ICziI-. LlJuJ _J68IU,-.a a:1c’’Li,C’)I-.LJw‘3_, -‘I-11d30(sJe’.eW)692 Om=1.4.— FRASER RIVER 4.0 / pc.2.cPrfl4V’’SAND FILL(2m DEEP)UBCPRS0APPROX IMATESCALEDIKE ROADFigure 6.10: U.B.C. Pile Research Site70This site differs from most of the previous sites in thatthe native surf icial deposits are organic rich to 15 metres indepth. Figure 6.11 shows the RCPT profile from the pile researchsite. After the initial 2 metres of fill, indicated by the highcone bearing at the beginning of the sounding, there is a longsequence of soft organic materials. Between two and three metresthe high friction ratio and low resistivity combined with thelack of pore pressure generation suggest the presence of afibrous peat. From 3 to 12 metres the friction ratio and dynamicpore pressure indicate a silty soil which exhibits a plasticbehaviour. Previous work at the site confirm this soil to be anorganic silt (Davies, 1987). Between 12 and 15 metres the soiltype appears to change again. This change is marked by a suddendrop in resistivity, and an increase in dynamic pore pressureand friction ratio. These resistivity and pore pressure changessuggest an increase in clay mineral content through thissequence. Below the organic soils is a sand unit with varyingsilt content. The sand is very stratified and continues until28.5 metres. Below the sand are interbedded sands and clayeysilts. The bulk resistivity profile in the top twenty metres isquite variable. Most of this variability attributable to changesin soil type. The resistivity low between 2 and 3 metresdemonstrates the likely presence of fibrous peat. From 3 to 12metres the organic silt shows variability which probably is dueto different amounts of clay mineral phase. From Twelve tofifteen metres the resistivity decreases markedly. The sandsshow characteristically higher resistivites than do the fineDEPTH(rrTQtrs)wa.iii-I6zC)emCHCD0ct0CH0C0CDCDCDDiHC)z.r1CD-Vcl-ir’iC-irtr’,)—a:n;.‘—0C,•‘IIIIIIIIIIIIIIIIIIIIIII•I&IL72grained soils. At about 22 metres the resistivity begins todecrease rapidly, indicating the conductivity of the pore fluidis increasing.6.6 KWMTLEN COLLEGE SITEThe location of the new Kwantlen College site, presentlyunder construction can be seen on figure 6.1. The site is innorth Richmond, near the Landsdowne Shopping Centre. The sitewas treated with dynamic compaction in a region encompassing theproposed building footprint, since the surf icial soils weredeemed unsuitable in their present form. Two RCPT soundings wereconducted at the site, one at a location that was densified bydynamic compaction and one that was not. The locations of thesoundings and a site plan are contained in figure 6.12.Figure 6.13 shows a profile from an untreated portion ofthe site. The site is highly variable and this profile is notnecessarily indicative of what may be present at anotherlocation on the site. Covering most of the site is a sand andgravel fill to the depth of 0.5 to 1 metres. Beneath this is asequence of clayey silts followed by sand to 18 metres. The sandis silty in locations as indicated by the friction ratio anddynamic pore pressure. Figure 6.14 shows a profile about 11meters from the above sounding that was treated by dynamiccompaction. Before the soil was treated the surficial silts andclayey silts were excavated and replaced with sand fill. The41CW,4NT2. cPr — —€wMi13..cPT1:1000I ED6E Of D.C. I___I A--i-7 3L4JLDIN6 FooT PPIAJI---”- -- -------- --------I1’LANOSI)OWNE ROADFigure 6.12: Kwantlen College Research SiteCONEBEARINGSLEEVEFRICTIONFRICTIONRATIOPOFEPESS1ERESISTIVITYINTERPRETEDUt(bar)ft(bar)Rf(X)U(metres)(oh.—.)STRATIGRAPIfY020002.505—1003007500‘‘0.0••a____________isuiFBEHIHDIll’CLAYEISflTI*SAND10IC101010I— 0 w QI20202C20I2C‘.‘30•••3()II•I3013030Figure6.13:RCPTSoundingfromKwantlenCollegeResearchSite(NotTreatedwithD.C.)(Kwant2.cpt)750‘V.r4cC)4J..-4 ._4oa)4J‘—Ia)UUrI I I I I‘.4c,,r....6‘.4t._ Till I •l I I I Ia-4a)4.).1-ICf2cirJ•ja)a)a)kilta-4h(U) I-IJ.d3076upper 4 meters show an unusually dense surficial sand unit. Thesand becomes silty after 3 metres in depth and then grades intoa looser silty sand between 4.5 and 5.5 metres. From 5.5 metresto 18 metres there is a relatively clean sand unit with somesilty layers. The soils beneath 4 metres are native howevertheir densities have been increased considerably by thecompaction process. The resistivity profiles from the twosoundings are similar through the sequence of native soils. Inthe sands the bulk resistivity is generally higher where thecone bearing is higher and resistivity tends to decrease withincreased friction ratio. Between 4 and 8 metres in thedensified soil, the normal relationship between cone bearing andresistivity is not observed. High cone bearings are accompaniedby relatively low resistivites. Decreasing void ratio orincreasing density normally causes resistivity to increasebecause the amount of conductive pore space is reduced. Thisdifference in trend can be explained by the presence ofsignificant fines, as evidenced by the increase in frictionratio and dilative dynamic pore pressure. The dynamic porepressure recorded behind the tip in this region drops as low as—9 metres, bordering on cavitation pore pressure. This largenegative pore pressure is unable to recover during the 60seconds pauses in penetration that rods are being added. Thissuggests that penetration through the soil is undrained.Negative dynamic pore pressures are generally seen in dilativesands, and non—plastic silts. The densified soils between 3 and7 metres are silty sands.777. PREDICTION OF DILATANCY BEHAVIOUR OF SAND PROM RCPT DATA7.1 INTRODUCTIONThe sites examined demonstrate the type of variability thatcan be expected, in terms of pore fluid characteristics as wellas soil characteristics. The dilatancy behaviour of sands iseffected by mineralogical and grain size factors in addition toplacement density and confining pressure. Significant variationsin mean grain size, grading and mineralogy have been observed inthe Fraser Delta (Armstrong, 1984). It is expected that thefactors effecting dilatancy behaviour also effect measured conebearing, however it is not clear whether they are effected inthe same manner. This topic will be addressed as the data ispresented.The data presented in this chapter will show the dilationparameter (D) plotted with depth together with normalized conebearing. The normalization applied is that proposed by Sladenand Hewitt (1989), which is based on back analysis of failuresof hydraulic fill structures. The Sladen and Hewittnormalization is as follows:Qc= Qc/(a’)°65 (7)This normalization format was adopted since it appeared tobe superior in terms of correcting for overburden effects on78measured cone bearing as well as appearing contiguous with SPTbased results from Seed (Robertson & Woeller, 1992). Verticaleffective stresses were estimated using an assumed soil density.Normalization is mainly an attempt to account for changes inobserved behaviour caused by increasing effective overburdenstress. For this analysis normalization is seen as a good way ofstandardizing the data presented while at the same time allowingcomparison of the resistivity method of sand characterization toa CPT based method. Sladen and Hewitts’ criterion for thedilative contractive boundary is:Qc = 70 barSands that are less than 70 bar are considered loose orcontractive, while sands that are greater than 70 bar areconsidered dense or dilative. Clearly this criterion cannot berigerously applied to any type of sand. Dilatancy in sands iseffected by factors such as mineralogy, age and OCR. The Sladenand Hewitt criterion will be checked to see if it is valid forFraser Delta sands.7.2 ARTHUR LAING BRIDGE SITETwo soundings were carried out at this site, spaced about1 metre apart. They were located about 2 metres from a mudrotary hole where SPT’s were done at five foot intervals. Hencegrain size data in the sands was available. Figure 7.1 shows theEr 10F—0LU020NORMALIZED• - - -- DILATION PARAMETERNORMALIZED CONE BEARING79Figure 7.1: A Comparison Between Dilation Parameter and NormalizedCone Bearing with Depth for Laing Bridge Site.NORMALIZED CONE BEARING (bar) CONE BEARING (bar)70151.0DILATION PARAMETER, DD= 9095/P775DILATION PARAMETER, 080dilation parameter and normalized cone bearing computed to 20metres for the two RCPT’s done at the site. The normalized conebearing is shown as an indicator of the in situ relative density(Robertson and Campanella, 1989). The sequence of sand between4 and 14 metres is a good one for the analysis since it is cleanand shows considerable variability in density. There is anoticeable relationship between normalized cone bearing anddilation parameter. The increases in cone bearing are insequence with decreases in dilation parameter and hence increasein inferred dilation. In the upper part of the zone from 4 to 14m, the sands are loose to medium loose as indicated bynormalized cone bearing being in the range of 40 to 60 bar. Thedilation parameter recorded in this region was greater thanunity on average, indicating that contraction was taking place.Contrastingly in the lower sands, which appear to be quite denseindicated by the normalized cone bearings greater than 70 bar,D is less than unity. This indicates that dilation is occurringin the sands.As mentioned earlier while cone bearing is related to density, other factors can also influence cone bearing. One ofthese factors is grain size. The grain size analyses from theSPT’s are summarized in table 7.1. For more detail on theseanalyses the reader is referred to appendix B. The analysesindicate a very uniform medium sand is found between 4 and 14metres depth. A first approximation of horizontal effectivestresses can be made assuming a constant K0 of about 0.5. If I<81and grain size are both invariant with depth, a goodrelationship between normalized cone bearing and dilationparameter is expected. Figure 7.1 show this to be the case.TABLE 7.1: SUMMARY OF SPT RESULTS FROM LAING BRIDGE SITEDEPTH (m) D50 (irnu) (N1)1.7 0.03 N/A3.2 0.08 54.7 0.35 116.2 0.45 177.8 0.30 159.3 0.28 910.8 0.28 1512.3 0.275 1513.9 0.30 2215.4 0.25 1216.9 0.26 2518.4 0.18 2720.0 0.030 621.5 0.023 223.0 0.053 27.3 MILLER ROAD SITEThe two soundings carried out at this site were spacedabout 25 in apart. They were done at two different fill heights,along a gentle grade slope as seen on figure 6.3. Similarresults can be observed in both of the soundings. Figure 7.282-“ DILAflON PARAMETERNORMALIZED CONE BEARINGm I I I1.5Figure 7.2: A Comparison Between Dilation Parameter and NormalizedCone Bearing with Depth for MiNer Road Site.NORMALIZED CONE BEARING (bor)0NORMALIZED CONE BEARING70I I I I I I__l_I___(bor)14010121416C4/LRES2CPTE=aw182022D Po95’P7750.5 1.0DILATION PARAMETER, 083shows dilation parameter plotted along side normalized conebearing from 8 to 22 metres. This section was chosen since itincludes all of the clean sands that are present beneath thewater table. The trends observed in the previous site are againevident in the clean sands. Most of the sands appear to have Dless than unity, indicating that they are dilative. Oneobservation that can be made is that the response of D tochanges in cone bearing appears to be attenuated somewhat wherethe sands appear to be more highly stratified. The stratigraphicvariability of the site precludes comparison of the twosoundings on the basis of confining stresses. Therefore while itis understood that changes in overburden stresses are expectedto influence the dilatancy of sands in situ, this point couldnot be demonstrated from the data collected at the site.7.4 ALEX FRASER BRIDGE SITEFour test soundings were carried out at different locationsaround the site. Two of the soundings were beside support pierN2, while the other two were at the north end of the site faraway from the piers. Figure 7.3 shows the computed normalizedcone bearing and dilation parameters for the two tests done nearthe pier where some densification had been done. Figure 7.4shows the same for the tests done at the north end of the sitewhere no densification had been done. Both figures show similartrends. In the upper looser sands there is some contraction,indicated by a greater than unity value of dilation parameter.842I2Lii0202530NORMALIZED CONE 5EARINO (bar) NORMALIZED CONE BEARING (bar)D=• -- -- DILATION PARAMETERNORMALIZED CONE BEARINGFigure 7,3: A Comparison Between Dilation Parameter and Normalized ConeBearing with Depth for Alex Fraser Bridge Site (South Side).0 70 140 0 70 1401015DILATION PARAMETER, 0I0zIDIL)00IL)-J0z0-o0zcowz0C-)0IL)-Jtr:0za,Ca)0.-DNt0L.00(1)-D0ca00’-D0zc!LfID0U)(w)Hld3a0U)C)(‘U(‘UC)86Lower down there is mainly minor dilation, though contraction isquite prominent in the interval 25 to 30 metres. Therelationship between normalized tip resistance and dilationparameter while evident, is not nearly as good as for theprevious two sites. It appears that most of the sand lies on theless than unity or dilative side of the figure. This site isquite different stratigraphically from the previous sites. Theamount of layering observed in the sands is much greater,however the range of densities is smaller. The magnitude of thechanges in D are often not in proportion to the magnitude of thechanges in normalized cone bearing. Furthermore some of thechanges in D appear not to be in sequence with changes innormalized cone bearing. This may not be totally unexpected.Since the resistance to penetration is influenced by soilseveral cone diameters ahead of the tip it is possible that conebearing values may not be a truly representative estimate of thesoils density at the depth of the cone tip. This is particularytrue when stiffer soil overly softer soils. In this case conebearing may not be able reach its full value in the stiff soil,since the soft soil beneath it will begin deforming before thecone reaches it (Robertson and Campanella, 1989). Furthermorethe compressibilty of the soil effects it’s dilatancy behaviour.More compressible sands will develop lower cone bearings thanless compressible sands at the same relative density. Theimplications of these points are that normalized cone bearingmay not be a good an index of fundamental soil behaviour orfield performance for natural deposits.877.5 UBC PILE RESEARCH SITEFigure 7.5 shows a comparison of dilation parameter andnormalized cone bearing with depth at the pile research site.This site does not have much clean sand for analysis, howeversome interesting observations with regards to generalizedapplication of the resistivity dilatancy analysis can be made.From 3 to 15 metres in depth the normalized cone bearing isrelatively unchanged. In the previous chapter it was observedthat dynamic pore pressure and friction ratio responsessuggested there were some major changes in soil response topenetration at 3 metres and 12 metres in depth. Both thesedepths show large changes in D. These observations suggest thatthis approach to analyzing soil behaviour could be extended toother soil types besides clean sands.7.6 KWANTLEN COLLEGE SITEAs mentioned previously test soundings were done in dynamiccompaction treated and untreated locations at the site. Theresults from these tests can be seen in figure 7.6. In bothsoundings there is good agreement between increases innormalized cone bearing and dilation parameter in the sands,except for between 3 and 7 metres in depth. The profile on theleft of figure 7.6, labelled KWANT2.CPT is from the sounding inthe densified region of the site. From 3 to 7 metres dilationparameter does not seem to show the expected variation with=LUNORMALIZED CONE SEARiNG bcr)88D= Po95!P775- DILATION PARAMETER—NORMALIZED CONE BEARINGFigure 7.5: A Comparison eetween Dilation Parameter and Normalized ConeBearing with Depth for the UBC Pile Research Site,NORMALIZED CONE BEARING (bor)70DILATiON PARAMETER, D1.0DILATION PARAMETER, DELU1214NORMALIZED CONE eEARING (bar) NORMALIZED70CONE BEARING (bar)140 21089D — / - - - - DILATION PARAMETER—P095 P775— NORMALIZED CONE BEARINGFigure 7.6: A Comparison etween DHation Parameter and NormalizedCone Bearing with Depth for Kwantten College Site.0 70 140 20I I I I I I I010DILATION PARAMETER, D1.0DILATION PARAMETER, 090normalized cone bearing seen in sands. Instead of dilating,which is expected for sand with a normalized cone bearing valueof 200, the material behaves as borderline dilative contractive.This is the zone in which the friction ratio and dynamic porepressure in figure 6.14 indicate the presence of considerablefines. Hence fines could be influencing the behaviour of thesand. Because some technical problems were encountered with theequipment, the two RCPT’s done at Kwantlen College are slightlyincomplete in terms of the data collected. The smaller (9.5 mm)electrode spacing was periodically grounding which causedreadings to be missed at certain depths in both soundings. As aresult the dilation parameter was not calculated for the entiredepth as is normally the case.7.7 SUMMARYFrom the data presented, it appears that a trend ofdecreasing dilation parameter with increasing normalized conebearing can be demonstrated in clean sands. The relationship ismost pronounced when the sands are relatively uniform, such asat Arthur Laing Bridge site and Miller Road site. In highlylayered sands, such as Alex Fraser Bridge site, observed changesin dilation parameter do not seem to be as large, in magnitude,as the observed changes in normalized cone bearing. This isexpected since the resistivity measurement from the largeelectrode spacing averages over a relatively large distance. Incontrast the value of cone bearing is an instantaneous91“snapshot” of the resistance to penetration at a specific depthpoint. Thus resistivity cannot respond to a thin soil layer thatcone bearing can. While dilation parameter is used herein as ameasure of dilatancy of sand, it shows promise as a parameterfor soil classification, as demonstrated from the UBC PileResearch site data.928. PRACTICAL USE OF THE DILATION PARAMETER8.1. INTRODUCTIONThe bulk of the data relating to field performance ofgranular soils comes from the standard penetration test (SPT).The main driving force for continued use of this test is thefact that compared to other in situ tests, there is a very largedata base available, most notably from areas where large seismicevents have been relatively common. While this argument infavour of the SPT might seem a bit self-perpetuating, in essenceit is the only deterministic option available using anempirically based approach. Two of the performance criteria thathave been well established using SPT blow count data,liquefaction and ground settlement, are a function of volumechange behaviour or compressibility. Both liquefaction andsettlement are characteristic of strain softening or contractivesands (Lee and Seed, 1967). Laboratory studies have shown thatthe term contractive needs to be applied cautiously sincefactors such as confining stress, silt content and direction ofloading have an immense impact on the behaviour of sands(Keurbis & Vaid, 1989). From an “engineering in the FraserDelta” point of view the concern is normally for sands to adepth of 10 metres. Because it is believed that the dilationparameter D is related to the dilatancy characteristics of cleansands in situ, a comparison with existing field performancecriterion is both relevant and useful. Wherever possible, data93collected will be compared with other test data to determine ifcorrelations are possible with existing methods.8.2 CPT-D CORRELATIONSBecause the test used to measure D is a modified CPT, it isnatural to use this test as a first basis of comparison. Thedata presented so far have shown that there is some basis formaking a correlation between Qc or normalized Qc and D. Thiscorrelation will be explored by analyzing data from thedifferent sites and attempting to make some generalizations thatare not site specific. Figures 8.1 through 8.3 show normalizedcone bearing vs dilation parameter, averaged at 25 cm intervals,for some of the tests done. The tests selected were ones whererelatively uniform sands were present at less than 20 m belowthe ground surface. Above 20 in there is usually substantialvariability in the density of the sands penetrated. Below 20 inmost sands are quite dense, and while they may be appropriate toconsider, they were left out so as not to bias the analysistowards sands of higher density. Also, for settlement andliquefaction analyses, most of the time the soils of concern areshallow. The results of figure 8.1 from Laing Bridge site showan extremely good correlation between normalized cone bearingand dilation parameter. Figure 8.2, showing similar data for theMiller Road site also shows a good trend but the slope appearssomewhat steeper than the previous site. Figure 8.3 shows theresults from Alex Fraser Bridge site. Unlike the previous sites9400000 LANG11.CPT150- •....LANG12,CPT0C0• .0I00z • II• •001000ILii- 0zo ..C(_)• ICw - Io •50- S00oz -I0— i I I I I I I I I I• 0.7 0.8 0.9 1.0 1.1 1.2 1.3DILATION PARAMETER; DFigure 8.1: Normalized Cone Bearing vs Dilation Parameter forArthur Laing Bridge Site (from 4—14 m).95- 00000 MILRES1.CPT1 50 ••••I MILRES2.CPT•0-00Z -•ê-Q•0.100-• •LU000 •zooSoc0LiJ 05Q 0 0o 0z 00 8I I I I I cc I I II II I II I0.7 0.8 0.9 1.0 1.1 1.2 1.3DILATION PARAMETER, DFigure 8.2: Normalized Cone Bearing vs Dilation Parameter forMiller Road Site (from 8—18 m).96(3z100LUz0C-)0LU-J0z1 50DD FRAS1.CPT‘I’ll FRAS4,CPTooooo FRAS2.CPT•.‘.. FRAS4.CPT00 C.. 00500I. C•.••0C00 I•• C CI:CCC00C.0.9DILATION1. 1.1PARAMETER, D1 .2 1Figure 8.3: Normalized Cone Bearing vs Dilation Parameter forAlex Fraser Bridge Site (from 4—14 m).97there doesn’t appear to be as clear a correlation between thetwo parameters. The spread in data is likely attributable tofactors affecting cone bearing but not dilatancy. Some of thesefactors include grain size, grading and are discussed in furtherdetail in Robertson & Campanella (1989). In addition there canbe factors that effect measured dilatancy to a larger degreethan cone bearing. One of these observed in the data so far isthe presence of fines. Fines cause a decrease in observeddilation, which in terms of the figures described above, wouldcause the data points to shift towards the right hand side ofthe figure.Figure 8.4 shows all of the preceding data presented atonce. From this figure it seems that the lower bound of the datacan be defined by a straight line. The upper bound of the datacannot be defined as clearly. There are many data points thatescape the major trend of the data, which is towards decreasingdilation parameter with increased normalized cone bearing. It isbelieved that the presence of fines might be responsible atleast in part for these data points.Looking at the Miller Road site data on figure 8.4, itseems clear that on balance there is less dilation for a givenvalue of normalized cone resistance than for Laing Bridge site.It is believed that confining stress may be at least in partresponsible for this difference. Between 5 and 6.5 meters ofsand fill is present at the locations of the two RCPT’s98L0zwz0C)0Ui-J0z00000000 LAING BRIDGE•‘••e MILLER ROADALEX FRASER BRIDGE•.“ ALEX FRASER BRIDGE00 .0000: °° • • ••••0I1 501 005000.\ 0aC00a . aIa000IDILATION PARAMETER, DFigure 8.4: Normalized Cone Bearing vs DilationArthur Loing Bridge, Mifler Road, AlexParameter forFraser Bridge Sites.99performed at Miller Road site. This corresponds to an increasein vertical effective stress between 100 and 125 kPa. Anincrease in vertical effective stress results in increasedconfining stress which has been demonstrated to decreaseobserved dilative behaviour in lab tests on sands.Sladen and Hewitt (1989) suggest a value of Qc = 70 bar forthe boundary between contractive and dilative behaviour. Giventhe data presented in figure 8.4, which shows D = 1 having amean value of 55 bar (normalized), Sladen and Hewitts value maybe overly conservative when applied to Fraser Delta sands. Theirdata is from back analyses on failures in hydraulically placedsands, hence it may not be realistic to apply this criterion toother sands with different steady state parameters.8.3 SPT-D CORRELATIONSSPT data were available for only one of the sites examined.Since most field performance data available are based on thistest it is particularly appropriate for comparison. Arthur LaingBridge site had SPT’s done about 5 metres away from an RCPT, asshown in figure 6.2. Figure 8.5 shows SPT values plotted asa function of D from a nearby RCPT. Also plotted are thecomputed by the computer program CPTINT v4.0.22. The measureddata shows a good fit with the computed data. This figuredemonstrated that a relationship can also be proposed betweenSPT N1( values and the Dilation parameter. With the limited data100XXXXX FROM SPTaooao FROM CP11NT (LAING1 1 .CPT)- 0- 0- 0 0- *0• 020- 0 0 0• D0 0• © 0*o0 0 01 0*0*0*00 00 0 00.7 0.8 0,9 1.0 1.1 1.2 1.3DILATION PARAMETER (D)Figure 8.5: Laing Bridge SPT (N1)- dilation parameter correlationfrom LAING11.CPT, using measured (N1) and (N1)from CPTINT v4.5.22 (4- 14 m).101available it seems unreasonable to draw further conclusions fromthe results.8.4 RELATIVE DENSITY- D CORRELATIONSAlthough relative density (Dr) is a difficult parameter toevaluate, it continues to be used as a guide in design. Recentcalibration chamber testing, has provided correlations betweencone resistance (Qc) and relative density for several referencesands (Robertson and Cainpanella, 1989). One of the referencesands, Ticino Sand, is a moderate compressibility sand that issimilar in characteristics to sand found in the Fraser Delta(Campanella, 1991). Using the computer program CPTINT v4.O.22 itwas possible to estimate Dr for sands tested. Figure 8.6 showsrelative density as a function of D for the two soundings doneat Laing Bridge site. The results suggest a good relationshipbetween D and Dr.8.5 DENSIFICATION CONTROLOne of the areas in which the CPT is becoming thepreeminent test is compaction or densification control.Generally contractors are required to densify to a particularcontrol specification, to a specified depth. This specificationis generally cone bearing. It is believed that the dilationparameter could be used as an independent method of assessingground improvement for any method of densification.1021 00:ooocoLAjNGjj.CPT..... LAING12.CPT-o9_o60- (‘•.(J -z •LU•LU 0 0 •00Afl._tLI0LU- 00020-0—•— I I I I I I I I I0.7 0.8 0,9 1.0 1.1 1.2 1.3DILATION PARAMETER, DFigure 8.6: Relative Density vs Dilation Parameter for Loing Bridgesite, interpreted from CPTINT v4.5.22 (from 4 to 14 metres).1038.5.1 ALEX PRASER BRIDGE SITEAs mentioned in chapter 6 densification was carried out atthe site at the locations of the pier structures. Figure 8.7shows normalized cone bearing and dilation parameter for twoRCPT’s carried out at different distances from the pier N2, asshown in figure 6.8. Densification was recommended to 15 in, andwas carried out using the vibro replacement technique. SoundingFRAS2.CPT was done 15 m east of pier N2, while FRAS3.CPT wasdone about 30 m east of pier N2. FRAS2.CPT was just outside thedensification footprint that was specified by the consultants(Golder Assoc.), while FRAS3.CPT was over 15 in away from thedensification work. The change in normalized cone bearing isnoticeable as shown in figure 8.7, with in most cases anincrease in normalized cone bearing observed in the soundingcloser to pier N2. Looking at the dilation parameter, a decreasein dilation parameter is observed concurrent with increases innormalized cone bearing. This would suggest that the densifica—tion has improved the site characteristics by causing the sandto undergo positive changes in density.8.5.2 KWANTLEN COLLEGE SITEAt this site the technique employed was dynamic compaction.Before compaction began the surf icial silts and clayey siltswere excavated and replaced with sand fill. Figure 8.8 shows acomparison between improved and unimproved soil at KwantlenzCw8O101214DILATION PARAMETER, DFRAS2CPT (IMPROVED)FRAS3CPT (UNIMPROVED)104Figure 8. 7: Comparison Between Improved and Unimproved Soil at AlexFraser Bridge Site (South Side), Using Normalized Cone Bearingand Dilation Parameter as Improvement Indexes.NORMALIZED CONE EARINC (bar)70= p095/p77105zI—a1014DILATION PARAMETER, DKWANT2.CPT (IMPROVED)---- KWANT3.CPT (UNIMPROVED)Figure 8 . 8: Comparison Between Improved and Unimproved Soil at KwantlenCollege, Using Normalized Cone Bearing and Dilation Parameteras Improvement Indexes.NORMALIZED CONE SEARING012D Po95’P775106College. Missing dilation parameter data and different soiltypes obscures some of the effects, however it appears that inthe sand beneath the zone of replacement, very littleimprovement is measured in terms of D, although significantincreases in Qc are observed. In the zone of replacement,between 2 and 4 metres, there is strong dilation in the improvedsoil profile. From 4 to 7 metres in depth, which is native soilin both soundings, the improvement is not particularlyrecognizable in D, however plastic fines are present in thesand which are believed to be controlling the volume changebehaviour. This will be discussed further in the next section.The fact that D remains unchanged in the treated soil from 7 to10 metres is unexpected. A possible explanation for this is thatwhile dynamic compaction increases density of granular soils, italso increases the in situ stresses (especially horizontal) inthe soil. Therefore both increase in density and increase in insitu stresses would contribute to a higher Qc. While increasingdensity promotes dilative behaviour, increasing confining stressdecreases dilative behaviour. It is possible that the effectsare acting in equal proportions, so as to cancel each other out.8.6 EFFECT OP FINES ON SOIL BEKAVIOUR IN SITUThe effect of fines content on soil behaviour and thereforeliquefaction resistance has long been a subject of interest.While the presence of fines is known to cause a decrease inpenetration resistance, their effect on soil behaviour remains107a contentious issue. Recent studies on the effect of finescontent on undrained strength of sands seem to contradict oneanother. Keurbis and Vaid (1989) suggest that for non-plasticfines, a sand skeleton can accommodate up to 20 % fines withoutany reduction in monotonic or cyclic strength. They contend thatthe behaviour of the sand will be governed by the void ratio ofthe sand skeleton and that the silt will not contribute at allto the undrained behaviour. On the other hand Troncoso (1990)found that silt content decreases cyclic strength in tailingssands.Figure 8.9 shows friction ratio plotted with dilationparameter, alongside normalized cone bearing for the treatedsounding at the Kwantlen College site (KWANT2.CPT). This figureclearly illustrates the change from dilative to borderlinedilative/contractive behaviour accompanying the appearance offines in the soil. Furthermore there appears to be somerelationship between the magnitude of friction ratio andmagnitude of D (tendency towards contraction). This relationshipdemonstrates the influence of plastic fines on the behaviour ofthe sand skeleton during penetration induced shear. If the sandis not able to drain during penetration, as the dynamic porepressure in figure 6.14 seems to suggest, then the volume changeassociated with dilative behaviour cannot occur until cavitationpressure is achieved. The data from this site suggests that iffines control the drainage of soil during penetration then theyalso control the ability to measure volume change with theEI—U-iDILA11ON PARAMETERFRICTION RATIOFigure 8. 9: A ComparisonNormalizedCollege site.108FRICTION RATIO ()0.5 1.0NORMALIZED CONE BEARING (bar)0 70 140 210I I I I I I iiDILATION PARAMETER, DD= p0/p5Between DilationCone Bearing inParameter, Friction Ratio andthe densified region at Kwantlen109approach described herein. In figure 6.14, the increase infriction ratio from 3 to 5 metres is accompanied by a decreasein pore pressure and a decrease in resistivity. Theseobservations are consistent with the decrease in dilationparameter observed in figure 8.9, suggesting that an increase inclay fraction occurs with increase in fines. This exampledemonstrates that for dilatancy behaviour to be recognized usingthe resistivity technique, penetration must be occurring underdrained conditions. For sandy soils that are not fully drainedduring penetration, the volume change behaviour can be observedqualitatively by measuring dynamic pore pressure behind the conetip. Dilation during penetration induced shear is marked by adecrease in dynamic pore pressure (beneath hydrostatic) behindthe tip.1109. CONCLUSIONS AND RECOMMENDATIONSThe data presented illustrate a useful application ofgeophysical data to the study of practical problems ingeotechnics. The measurement of relative porosity change duringpenetration induced shear is a concept that is both novel andimportant for understanding dilatancy behaviour of sands insitu. By comparing the resistivity measurements at differentelectrode spacing it is possible to consider the effect ofdifferent levels of strain on porosity, since strain level inthe soil decreases as the distance from the probe increases.It has been demonstrated that an analytically based methodcan be used to qualitatively observe dilatancy in sands.Comparison of the results of this method with anotherempirically based approach (Sladen and Hewitt, 1989),demonstrates that care must be taken when applying empiricallybased criteria to naturally deposited soils. Naturally occurringsands can have highly variable steady state lines in void ratioconfining stress space, and hence it can be very difficult topredict their dilatancy behaviour based on a parameter such ascone bearing. Furthermore, finer materials that can beaccommodated within a sand skeleton can control its behaviourduring shearing.In order to improve the resistivity module operations aswell as better define the operational constraints of the11]technique outlined, the following recommendations are proposed:1. Redesign the module with a greater range of spacingsso that a greater depth of penetration into theformation wall can be achieved. Compare ratios oflarger spacings to confirm that the conclusions hereinare substantiated.2. Confirm the observation of Chong (1988) relating tothe change in density around a penetrating probe andextend the analysis to look at a greater range ofplacement densities. Chongs’ technique usingthermistors to measure density changes seems to be thebest way to accomplish the above.3. Perform additional field testing at DC and Vibro sitesto confirm the observation presented. In particularbefore and after densification testing should be doneto confirm the effect of dynamic compaction on theparameter D.4. Examine the effects of confining stress on D in situ.A good location for this would be the Miller Roadsite, except rather than performing tests at largespacings along the ramp, perform one test at the topof the fill and then another close beside the f ill.This should reduce the soil variability observed and112allow direct observation of changes in D withconfining stress.5. Correlate D to other in situ parameter such asdilation angle measured by the SBPT or Kd measuredwith the DMT.6. Perform lab tests on samples in drained and undrainedloading by placing electrodes into the base and cap ofa triaxial cell. By using an AC excitation at 1000 Hzas in the module, behaviour during cyclic loadingcould be examined as well. It may also be appropriateto adapt a simple shear device to measure resistivityin the zone of shearing during the test.11310. POSSIBLE APPLICATIONSThere are several practical applications where the RCPT canbe successfully employed to increase the level of confidence fordesign specifications. Densification is a particularly goodapplication as has been demonstrated in the precedingdiscussion. Measurement of D can be used as an additionalspecification for densification. Soil improvement specificationscan be modified to include a required D value for densification.This may or may not be appropriate for dynamic compaction forreasons discussed in the preceeding sections. Because of theuniformity of sands used in construction and the resistivityrange of the pore fluid, the RCPT would be useful in settingdesign specifications for man made sand islands such as thoseconstructed in the Beaufort Sea for drilling platforms. From theresults presented it is reconuuended that the resistivitydilatancy method be applied to analyses where drainedpenetration is assured. If penetration is undrained dilativecontractive behaviour may be better predicted from dynamic porepressure assessment, as mentioned in section 8.5.It is the writers belief that when perfected, the dilationparameter technique for resistivity can be used as a practicaladdition to a field investigation currently using the CPT orseismic CPT as its principal investigative tool. It should alsoencourage those who are currently using other in situ tests togain similar information to consider use of the RCPT. While atpresent the test may not be advantageous for prediction of field114response, the fact that it is analytical in nature suggests itcould become a more precise method in the future, once all ofits limitations are fully understood.115REFERENCESArchie, G.E., 1942, The electrical resistivity log as an aid indetermining some reservoir characteristics. Trans. Am.Inst. Mm. Eng., Vol. 146, PP. 54-62.Armstrong, J.E., 1984. Environmental and engineeringapplications of the surf icial geology of the FraserLowland, British Columbia. Geologic Survey of Canada, paper83—23.Bazett, D.J. and N.R. McCanunon, 1986. Foundations of the Annaciscable-stayed bridge. Canadian Geotechnical Journal, Vol.23,pp. 458—471.Been, K., Jefferies, M.G., Crooks, J.H.A. & Rothenburg, L.,1987. The cone penetration test in sand: part II, generalinference of state. Geotechnique, Vol. 37, No.3, pp. 285-299.Campanella, R.G., 1991. Personal communications.Campanella, R.G. and I. Weemes, 1990. Development and use of anelectrical resistivity cone for groundwater contaminationstudies. Canadian Geotechnical Journal, Vol.27, pp. 557-567.Campanella, R.G. and P.K.Robertson, May 1981. Applied ConeResearch. Soil Mechanics Series No.46, Department of CivilEngineering, The University of British Columbia.Campbell, M.D. & J.H. Lehr, 1973. Water Well Technology, McGraw-Hill, New York. pg.l84.Carr, C., 1982. Handbook on Soil Resistivity Surveying. Centerfor American Archeology Research Series, Vol.2, pp. 47-101.Chong, M.K., 1988. Density changes of sand on cone penetrationresistance. Penetration Testing 1988, ISOPT-1. ED. DeRuiter, Vol.2 pp. 707-714.Davies, M., 1987. Predicting axially and laterally loaded pilebehaviour using in situ testing methods. M.A.Sc. thesis,Department of Civil Engineering, The University of BritishColumbia, Vancouver B.C.Delft Soil Mechanics Laboratory, 1982. Results of densitymeasurements in situ in sand at the Holmen site in Dranimen,Norway. Report: BO—262520/23 SE—690276/2Freeze, R.A. and J.A. Cherry, 1979. Groundwater. Prentice-Hall,Inc., Englewood Cliffs,N.J. pp. 279—286.116Jackson, P.D., D. Taylor—Smith and P.N. Stanford, 1978. Resistivity-porosity-particle shape relationships for marinesands. Geophysics, Vol. 43, No. 6, pp. 1250—1268.Kuerbis, R. and Y.P. Vaid, 1989. Undrained behaviour of cleanand silty sands. In Proceedings of Discussion Session onInfluence of Local Conditions on Seismic Response, XIIInternational Conf. on Soil Mechanics and Foundation Eng.,Rio de Janeiro, pp. 91—100.Lee, K.L. and H.B. Seed, 1967. Drained strength characterizationof sands. Journal of the Soil Mechanics and FoundationDivision Proceedings of the ASCE, Vol. 93, No. SM6, pp.117—141.Robertson, P.K. and R.G. Campanella, R.G., November 1989.Guidelines for geotechnical design Using CPT and CPTU. SoilMechanics Series No.120, Department of Civil Engineering,The University of British Columbia.Robertson, P.K. and D.J. Woeller, 1992. Seismic cone penetrationtesting for evaluating liquefaction potential. CanadianGeotechnical Journal, accepted for publication.Sladen, J.A. and Hewitt, K.J., 1989. Influence of placementmethod on the in situ density of hydraulic sand fills.Canadian Geotechnical Journal, Vol. 22, pp. 564—578.Torstennson, B.A., 1984. A new system for groundwatermonitoring. Ground Water Monitoring Reveiws, Vol.4, No.4,pp. 131—138.Troncoso, J.H., 1990. Failure risks of abandoned tailings dams.In Proceedings of the International Symposium on Safety andRehabilitation of Tailings Dams, ICOLD. Sydney, Australia.May 1990.Urish, D.W., 1981. Electrical resistivity-conductivityrelationships in glacial outwash aquifers. Water ResourcesResearch, vol. 17, No. 5, pp. 1401—1408.Vaid, Y.P., Chung, E.F.K. and Keurbis, R.H., 1990. Stress pathand steady state. Canadian Geotechnical Journal, Vol. 27,No. 1.Weemees, l.A., 1990. Development of an Electrical ResistivityCone for Groundwater Contamination Studies. M.A.Sc thesis,Department of Civil Enginnering, The University of BritishColumbia., Vancouver B.C.117Windle, D. and C.P. Wroth, 1975. Electrical resistivity methodfor determining volume changes that occur during apressuremeter test. In Proceedings of the Conference on InSitu Measurement of Soil Properties, June 1-4 1975., NorthCarolina State University. Raleigh, North Carolina.Specialty Conference of the Geotechnical EngineeringDivision, ASCE. Vol. 1, pp. 497—510.‘‘aAPPENDIX A119PROGRAM REDUCECC THE FOLLOWING PROGRAM CONVERTS VOLTAGES AT THE ELECTRODES OF THEC RESISITIVIriy CONE TO RESISTIVITY VALUESC = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = —====—— = = = = = = = = = = = = = = = = = = =REAL KOSS, K165, K260,K775, VINPUT,TIPBL,FSBL,LPPBL,UPPBL, BLVO95REAL BLV.L65,8LV260, BLV775,DEPTH,BRG,FRICT,UPP,LPP,V095,V165REAL V260,V775,DEPTHR,FRR,VONE,VTWQ,VTHREE,VFQUR,DC1,DC2,D03REAL UEPTHA, UIL,DEPTHS,WT,SIGVP,SLGV,UWIJ,UWS,QC1,CON1,CON2REAL I3COR, PCZUR, COR13CHA1CTE’i fuSSCHARACTER.i LE RE) LEIMI’EI;ER SK1IA, SKIPL,PAN,LIQPRiNT , ‘YOU ARE USING MA’I’THEW KOKANS PROGRAMPRINT X, FOR REDUCING RESISITIVITY CONE DATA’PRINT , COLLECTED USING THE USC RCPT’PRINT ,C ENTER THE NAME OF THE DATA FILE TO BE READPRINT , ‘ENTER THE NAME OF THE DATA FILE TO BE REDUCED’READ ‘(All)’, DFILEC ENTER THE NAME OF THE NEW DATA FILE TO BE CREATEDPRINT *, ‘ENTER THE NAME OF THE NEW DATA FILE’READ ‘(All)’, NFILEC ENTER CORRECTiONS TO BE MADE RD BASELINESPRINT ‘, ‘WHAT WAS THE TIP CORRECTION ?‘READ , SCORPRINT *, ‘WHAT WAS THE SLEEVE CORRECTION ?‘READ ‘, FCORC INITIALIZE (‘liii AMOUNT OF HEADER ‘10 UK SKI PPIIUSKIPA = 12SKIPH IC INI’l’IALIZE CALISRATION CONSTANTSK260 = 0.0153C INiTIALIZE THE VALUE OF CORRECTED DEPTHDEPTHA = 0.025C QUERY THE USER FOR NEED TO CHANGE SUPPLY VOLTAGEPRiNT , ‘WAS THE INPUT VOLTAGE CHANGED DOWNHOLE ? (Y=l/N=O)READ *, PANC READ IN THE INPUT VOLTAGE SUPPLIED BY THE SIGNAL GENERATOR120IF (PAN .EQ. 1) THENPRINT *,PRINT *, ‘THIS PROGRAM ALLOWS ONE TO VARY THE INPUT VOLTAGEPRINT , UP TO FOUR TIMES DURING THE DATA REDUCTION’PRINT *,PRINT , WHAT WAS THE INPUT VOLTAGE AT THE START (RXS) ?‘READ *, VONEPRINT ‘, AT WHAT DEPTH WAS THE VOLTAGE CHANGED 7’READ * DC1PRINT *, ‘WHAT WAS THE INPUT VOLTAGE 7’READ *, VTWOPRINT *, ‘AT WHAT DEPTH WAS THE VOLTAGE CHANGED AGAIN 7’READ , DC2-RINC , ‘WHAT WAS ‘I’HE INPUT VOLTAGE 7’READ , VTHREEPRINT ‘, ‘AT WkiI DEPTH WAS IRE VOLTAGE CHANGED AGAIN ?‘READ , DC3PRINT *, ‘WHAT WAS THE INPUT VOLTAGE ?‘READ , VFOURELSEPRINT *,PRINT , ‘ WHAT WAS THE INPUT VULTAGE 7READ ‘, INPUTEND IFC QUERY USER ABOUT LIQUEFACTION CRITERION OUTPUTPRINT *,PRINT , ‘ DO YOU WISH TO 0UTPU’r THE LIQUEFACTIONPRINT , CRITERION (Y=i/N=O) ?READ , LIGIF (LIQ EQ. 1) THENPRINT , WHAT IS THE DEPTH OF THE WATER TABLE 7’READ , WI’PRINT , ‘INPUT UNSATURATED UNIT WEIGHTREAD *, UWUPRINT , ‘INPU’I SATURATED UNIT WEIGH’!’READ , UWSEND IFC OPEN ‘iRE FILESOPEN (,FILE = DFILE,STATUS = ‘OLD’)OPEN (,FILE = ‘RHO’)OPEN )8,FILE = ‘BARS’)OPEN (9,FILE = ‘SLEEVE’)C FiND AND READ IN THE BASELINE DATA12100 55 1 = 1, SKIPAREAD 5, (A0) ) Mi55 CONjREAD (5,’(819 4)) TIPBL,FSBLLPPBLUPPBL+BLV26O,HLV.l.,SDO 65 J 1, SKIPSREAD (5, (ADO) ) MISS65 CON’fIt4uEC ENTER HEADERS IN’jo NEW FILEC WRI (I°’(A3t),A11)’)<REDUCED UBC CPT DATA FILE>:‘,NFILEC WRITE (10,’ (A50) ) DEPTH/BRG/FRICT/pp/RO9CC READ IN DATA5 READ (SF1o.5,9F94) END 85)+ VI 65, VU 95, V 260, V175C CALCULATE ABS)LUTE VOLTSV095= V095 + ULVO9SV165= V165+ BLV165V260 V260+ NLV250V775= V775+ BLV775C CORRECT CONE BASELINESFRCIT = FRICT + FCORBRG = NRC + BCORC COMPUTE THE DESIRED CALIBRATION CONSTANT FOR VmeasIF ((V095 .LE.5.1) .AND. (V095 .GE. 1.4)) THENKOUS 0.01J2-ELSE IF (VU95 .GT.6.2) THN/KU5 = (U.b1ELSE IF (V095 .LT. 1.1)* V095) / 47.5END IFIF ((V165 .LE. 3.95) .AND. (V165 .GE. 1.15)) THENK165 = 0.01655 THENELSE IF(V165.GT.3.9v165) / 47.5K165THENELSE IF (V165 .LT. 1.1V165) / 47.5END IF122K775 = (13.32 - 2.217 * V775 + 1.666 * V775 ** 2 — 0.541 *+ V7’/5 ** 3 + 0.074 * V775 * 4— 0.0033 * V775 ** 5) / 950C SELECT INPUT VOLTAGE TO BE USED IN CALCULATIONiF (PAN .EQ. I) THENI F ( DEP’lH.EQ . U . U ) THENVINPUT = VONEELSE IF (DEPTH .EQ. DC1) THENVINPUT = VTWOELSE iF cDEPTH . EQ. DC2) THENVINPUT = VTHREEELSE IF (DEPTH .EQ. DC3) THENVINPUT VFOUREND IFEND IFC CONVERT VOLTAGES TO RESISTIVITYR095 = V095 / (KU95 VINPUT)R165 = ViEb / (K165 VINPUT * 0.88)R260 = V2b0 / (K260 * VINPUT * 0.93)R775 = V’17S / (K775 * VINPUT)IF (R775 .GE. 32) THENCOR8 = 1.04ELSECOR8 = U — 0.222091 + 0.371674 * R775 — 0.0468098 * R7752 + 0.00331759 * R775 ** 3— 1.48359E—4 * R775 ** 4 + 4.40+784E—6 * R775 ** 5 — 8.79212E—8 * R775 ** 6 + 1.15194E—9 *+R775 ** 7— 9.41016E—12 R775 ** 8 + 4.31424E—14 * R775 ** 9+— 8.43562E—17 * R775 ** 10END IFR775 = R775 * COR8C CuNVERI RED 111 V ITY DATA TO DEPTH AND WRITE TO FILEIF (DEPTHA .011. 0.80) THENDEPTHR = DEPTHA - 0.80WRITE (6,(F8.3,4F9.3)’) DEPTHR,R095,R165.R260,R775END IFC CONVERT THE SLEEVE VALUE FOR CONEPLOTIF (UEPTHA .GE. 0.10) THENDEPTHS = DEPTHA - 0.10123WRI’iE (9, (F.3,F9.3)’ ) DEPTHS,FRICTEMIl iFC CORRECt PORE PRESSURE (15 SQ CM TO 10 SQ CM)LPP LPP 2C COMPUTE NORMALIZED VALUES FOR LIQUEFACTION CRITERIONIF (LIQ .IIQ. 1) THENIF (DEPTHA .GT. WT) ‘rHENSIGVP = WT * UWU + (DEPTHA — WT) * UWSELSES I GVP = DEPTHA UWUEND IFSLGV USPIHA IJWUQOl = URO SCJRT(i0I.3 / SIGVP)CONI (LkG (1.5 SIGV / 101.3) / 2) / (SIGVP / 101.3)CON2 Uk / (SIGVP /1U1.J) 0.65END IFC WRITE TO FILEIF )LIQ .EQ. 1) THENWRITE (8,(Fi.3,6F9.3)) DEPTHA,BRG,LPP,UPP,QC1,CON1,+CON 2ELSEWRITE (8, (F8.3,3F9.3) ) DEPTHA,BRG,LPP,UPPEND IFDEPTKA = DEPTHA + 0.025GO TO 755 CLOSE (5)CLOSE (6)CLOSE (8)CLOSE (9)C MERGE THE THREE HOLDING FILES INTO A DATA FILE FORMATTEDC FOR CONEPLOT124ODEN ( 6, 61 LE - RHOUPON (1,I1LE OARSUPON’,, P110 SLEEVEUPON (.LU,EILE NEILE)C READ IN THE DA1’A FROM HOLDING FILES90 ROAD (6,(F.3,4F9.3),END = 95) DEPTHR,R095,R165,R260,R775IF (LIQ .EQ. 1) THENREAD (0,(F8..3,6F9,3)’) DEPTHA,BRG,LPP,UPP,QC1,CON1,CON2ELSEREAD U,(IN.3,3F9.3)’) DEPTHA,BRG,LPP,UPPEND IFREAD (9, ‘(FN.3,09.3)’ ) DEPTHS,FR1CTC CALCULATE FRICTiON RATIOC FRR = FR I CI * IOU / BRGC CALCULATE DILATION CONSTANTDII = (1(095 / P175)C WRITE THE DATA TO THE NEW FILEIF ILIO .Ej. 1)IHENWRITE (10, ‘(F8.3,8F9.3)’) DEPTHA,BRG,FRICT,LPP,R095,R260,R’175,DIL,CON2ELSEWRiTE (iO,’(l’ 8.3,8F9.3)) DEPTHA,BRG,FRICT,LPR,RD 9 5, Ri 6 5, R 2 60 ,R 77 5ON. (VGu TO 9U65 CLOSE )b)CLOSE (8)CLOSE (9)CLOSE (10)125APPENDIX B126c)N>: j1i1 Ii’LI; Nii•iii_______zQI.aC0I’,0r1-C0C0JJ1 “1L 1’ HIJi ‘L IL_..J ULi.W “I” .u IL .i......... ...J... £.LL.II If ••TTI•11] I 111(111. ‘H fl i huH:. II1’hI Efl Ii H:..1 LC‘Uz—C’zaCzCC’zaC..CLa‘IIIi11{IijHi1i IIII iI1JrrFF-E I—ri . .. . F 1i:.I;__••••••_•_. II,I!huI.II!,uIi1::l,;..IIr—a’ .‘U=. .-..j1. .‘LI ‘II HI H 1111111 iii;i:h ‘III:’ I. ..L . . ...•44: ••i•V•’1• .:.....,.‘.;. f..I.....J.U... L.j.. Izci—‘- 0r0. r-.q-f jI..0CCC0.L.U. .‘...0 0 0 C 0 0a000w0L0.L Ie /0,1D—g:PC!%TGPA5SNG0a.—a*...00000000000rTjlT:HIHi,i1iI:jIJ_______11iTh1TUi‘t!iTrn00aaEI:EO]’.±L!W:‘;L:i:1HJ;f:rIT11I,‘;Iil;,Ii:l’it1’i‘_JJ!‘LILLL—JLI’IWh—‘IiIIif-J-J1’Cmcn3:C>.zI,)0I0H0CC-,0fllC7”I,0>‘7.7”7”-J>...LhUHThDb’¶—---E‘‘-H:.—-:..:,;,.S•....:......i__.L......._...7I’‘IL,hullaL.fLIj_i!:iIt!==—---:tf___:___i‘iiiiiiiiiiT17•I’ilililillII!HHIHL’rIhiflijJIjjj*:JL’i_Ldhz’_LJILiiijjJjijii71HWJI! :‘“!‘f:,I_!_i_...IIII‘(III5’!iC7!.j,.t.,lI!IIlI7I11II—:;ir;‘1;;litfftff1IHit!hl!hl!itlI!1.H1I1i[tiIJ08.nC.p7.’a0CC..zL.LT128••I!00ea• IV0• I0I.I Ti I I?• S0•:10I:: I,10‘I000iiII..C..Si V0a001FCTL‘•.0••‘:;..•••,i—I •r:.1!.;j’ihII!hi:j:I IIIl IiI’IUNCb0aai•; V. VI •‘,. .• I•,• V..ii0•I•u:;‘ijji :.CII0a00NISSVd DD’3‘I,wCAIN SIZE AN1YSISr129a00Ii aI,!,VIV‘:10t000I,II liii11C- •V V—J.L!.J.U. .LCr..[Li.I!:1VI:C._______T•:• VVc/IC..••f•..:‘11‘—1-I’ tI:1I.-)Uz0CjII:Iiii!i.0VI:IijjflT1 fli01!Iii:•VVV V V•V VVVCI’‘I0C09U)VNtSVd V)3dSISA1’fNYfl’sNI’yDF—__PECE’T.GPASSING000•4——aS.j•.:.L.:.±+..*.;[..Jrrr-:.;...‘•flT1k;‘itrh00CJN-I’JIji:—’—4x.-J000S0NmU,0>N115fri‘riIiI‘II’91i]ii1inn;iiwi ijIfIfffI:.jiii..p.1.41•..4.“..4r‘.I:Ii,.II’,,•:1:0n0‘,‘I,C)n0>‘-4zmLA-4I.0CzaLfl••iI.;.‘..••.;...•......I.:.;:.,.IJ...L...:I1JIIJF!Irjiii!1if1r11ifiil1111’J11!f1Ti!1J..I;.:1111‘I-•-Dii.;....II!;iti.10C‘•-‘00Nr--4uII--J0J.aC.0--,.:!J’1’1rI.i’’.IIi;;I1!UliI,!IIi’’’I1IJillIIIII11111IilEETfI,9llI.:‘iIji..___.:_.‘‘‘‘..........,.....L.L..ii111.[!.L.1:.,:.:;4. ..T.TTTT;,:,...TTT?]ttfljtIt’.iijIT.:.tTT.•IiII:I;ui!iuuIjiil;i.iIjjJ,’.:___j___..i:iliiii:LIIil.!..Li_:Ii.!j..1!iJJ_fjJjIl[!i.i‘1l1..1.LiiIIill]!IIIhIjjjji‘I,p:Ji[F05a0zNm0000tIiiiijj)JJJiIJ;:-II.’’II,IIoT1’!“°N‘e’ca;p1———PERCENTAGEPASSINGI.a0a—*.aa800000000000‘fliFFfi!F[1iiififiiI1fi[HFFlLNIn,ImU,3:Cz“IRI,0mI,C0PMCzaV.’3-/C)zC——I-J6J5-;0CC)zCS,r;i1i,iT]NIfii1f’t9IIFIiifiliitiffit:•UwLTh!iliwutirurir’j1 °•rTT•.,I....Irr1.1 LiIi.i.’.I’•i•_jjJII_L1I1lIII$IIi‘I,_IIj••j•_j,f1IIIIIIIji!Ijiij:IiIIIJ!I’1liiiIi!Ii1IiIIItjl}iiiI._!_j1I::ij._.J’L’UiiiijijjiifhjH;I1IIJIWHWi.II!hii’’IIIIII1111111II:°IJ.u•L_.____.-7i:9;, J_T.‘L!LLLLII.!.!LIL.I,jjLI’’ff111Ijn};IflHliif!1WI’rrrwFir...Lijr‘..IJ.jr1ff11—Iqffrit‘LthMP3TfT!.; -.IIHIfth’90iTfl[iji1Iii1]9Triirn‘IlILiJI.FfjiJIIIfIjI“aLijji1fi.[jiji1jjhi‘L1jj’iijiiI :.F!mr’jj”1hhI!ltI‘tliiQq.06.0I‘‘‘ IIII:1:LIilET1Uad,‘UNad,zC1324Iiii.Li,wNtSSVd 33d133I, IV VTgI! II i:0a0IIi:ICIl0.i I.:IiJ.iI II IIIC—I00--i-rcU.,NzC:i‘I•I•.II!Vrir:11.1_iIII.: 00‘lhjVV —V.0•VIJWIllV‘!‘:iII.C!Tim.i •1VV. ii 0DNISvd Dl’3dr&134--— I—siz 4NL1Y$ISV1VC-r(c.)‘UzC‘25--VNISd v3dU)w1Ce. (j135NISYd LN33dI I j2i07L I,7 0a0020•IoIC-00LII!llII!1I!ll’1I,,u!I1I,[I,I,Ill;IiII,l1j 1VI 1 jlji’I [‘ IHH:ffffflfff±fr :th;, L ;ifHIJ1IiIJtjjIiJjII!J’IIj’I 1 IiLWj’fl1 fj I“ -ill-i W.LI_J Th’ L1 U. JilLLii4 .4 4 ni 4JjJ 4jJJ4 ij4 W JW H .W lUg L1I LJWL IL.W ULL_ Li.!’ ‘L: ..f4r. 14.. Lj__J1 didiliIf1 1!T Ub riLLL LH4Ht44J..:ii .Li..L.i!! fFfi . .I.[.J...fL... J...L J. .1.oil II— (It’jj_jj..!.i_i_.i.!’ i’I. I•.., .. liii 1I. .r1 I 1141 i_iwLW ‘ .ii.:_i.L r!1•.,‘: f’ II ‘II:,.,,,..... ri , IiI’hi IILhhII H ii ‘I I I Ii riii rTT r- I FriiTflrjj’ , LJ F , II 1 i..i . •,I’tii .1 ‘Ii II I . iTT ‘ T7 •-:# , ‘ .p,,+ L.; w ‘ i __‘ j;1 1.IIii1I..j. .‘1i1•:LLi u1fj1j iii1i “UHI I’ fjj[II IIJUIII! [UI ii IIii ‘I‘-IbiwzV.a40V—C,‘Ua40V• biaOWa0z04C’,0CLUU,U,zC,0wIV-4-I‘I.—C1rrII.ii •f(i I1!1t ij1lllJIii III, ,i:ilh,I 11111 II ,iIhI,II ‘,:I’ ‘HIHj 114:1 If—••11 —‘.l•..._________ii Ii.. ‘I fTC0 0 0 0 0 0 e 0 00 • a .. S S — —C0S0GRAIN SIZE ANALYSIS136S 0r%. C-‘II00r.0S0• ,oCC.)4z0II • I.1£U,wNISSd q33dDIe;137H-:————————1Ii00000000C0• .nCia0aa0i1Ii’i IjIjiI :Ii. — I —‘. fTi7r I II — r hI. IJI 44_jJ ‘4.LL’ ._!L..L. ..JLth1 LJL.i.L’ L L:i’1-___.i1_ J —.._.L .._L± — -. :±--I iTTTIiir‘ ffiU’I: •,Jjj}:..::I’II III!,IIIiLIiIIjI:IIIIIII0iio r •.—.,. —-+t .... ..;.tUhi_——7r,1 IIj,C I I‘:L!_f ‘tfiiJiii;’ I:J —I•FTT_:. I!1.ji . . 1I I :;.iI•1 .i .Ui”LLLiiLJJLL JW: J, fl 1, Jjj HL :J.:i’ I 1.11 .L;_ —j.ii::Jii‘rc4.C1N.(0l1-a,—H.-C•u’t\J6w .aC•-Q0I)z0usa0UH-....Li.I.lfI.!.l.J.4iJJ JJ4.j4ij..LL !i1!’N:::‘ 4fT-r—i: .t:z:::.--::1 T:ONISSVd 3d00aGAIN SIZE ANYSISJV138%,iiii th!i! !I!0000w>00zI1tiIlj1ILL’ Li0!iI!iIL (II0LiiIIIit::,.40aC0C.:I; lJi !li!!-:;4.i41±iVC.SC;I’’!.Wif4:IIjLL,i!.Li. ILI.LLLUI,a..••p•.:•‘:• :ViC01 Vii•r:’’.lI,Ii..;:: 1:1I:;’!0VV0CVVI’iIllIi::I IVL I!1 thV•V. IVV.1Ci,.:I‘:if VIV V IV IIII0SC00L1wDNISSVdr.L’)L3d_,I,#i’1Ti#I#H1ffl’I•tl•j•I•LitifI1________!h././i.11111.1111I1i!LLL__1:11i.iJijjILi[1liiiilIi11111IiI1111111[Ti‘iiiifIlIiiiIli;lIIIIiI;;!!!I!:!.i!I1(l(tuIIiuh::ITTT1Hi’iI11111iigII1iiIIiiIi!IIII!,!!,:1:/IiioII:g..—-r.ii;I,1::,iHrifH;iirLLSISATh’Nv3ZISNiYO00PE,CENrAGEPASSING0‘d••11j1ffIiif(titi‘HItiI(tjiIfIfiI:•i1IiJ•i]]ufi1111iririif1fffItiIfI 1•I•I•1I1Ifi•I•I•IfffliIlIffiTIIIilOhiT1iiThZifliii;HIIIIII1’Hr——----—---_HIIITT;;:j—cn—1lii_iliiii;r’—p-___0L1HItE!_L:_I••____i—-H.L_..5.0/_•—z‘•.N——rn.,—.:_lI__‘Hi:1—r—’I.-.1iL1:1.—!-IC.I.’II%IIIHIIHIH!jflliIiIi(H!iI!1li111_.!I/IflI/IIII—I.1,I_1ij.1jliii1111111;uii’.).L.J.J....rIJIIiIiIIliiiIIff11ftIiftfrth‘liftifri‘iI//I/fib11111111LiW1:1JiiLLi.III-i1JjjJiIfIiII!It!ii/1IIIiiIiIiIiIILLWI!IIftUji.I_1..iili;I.LiiIIIiLi.LIL1///1llfl//U////IIIif/‘i’’/jT’,/”II1/t,iifl/i/ii.;1/!/////f!IiiI/!f!:y6Ir140IL0iJ>C WC“ITiI:I.1-: I TVIjf.i,’i’I‘‘A0Ti1Ifi‘llLLUWiIiIII. fl:i!llIV. 3’0LWlIjj00I.,E1IIIIII0‘1,3‘, :1V.000%00T .1 1ITTTff‘Jjii’U;iIII I ,1Tdfl1C.Siii.i;C-1iHIjhC—0I IV , ‘VI:[;‘:1::.:C‘VI.,.C—ijf!ij1I:IiIi.‘!‘U3z0I II]IiII1l!i1..4I.;:.I Ti•.:‘-V.’.VVVVV•VVVVViIi;lVI: jill—0c/i—TIrlliIHJd.1-.—I ‘IC0V.V.,... ,VVV0SCIii3,.:1IT‘‘IVC— 4.....,. 40a0.1OS.’cd v.3dci,000.0I--C_.zzz J””’ Noi fIDg


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items