高等土力学高等土力学 (3).pdf
Development and Application of a Large-Scale Static andDynamic True Triaxial Apparatus for GravelAbstract:To investigate the mechanical characteristics of gravel under general stress conditions,a large-scale static and dynamic truetriaxial apparatus was developed at Tsinghua University.The largest specimen was 200?200?400 mm,and the applied pressure canmeet the research requirements of large-scale geotechnical structures.A set of verification experiments with cyclic loadings of differentdirections and frequencies was performed to verify the accuracy of the loading system.Then,a static true triaxial experiment with con-stant mean principal stress,p,and generalized shear stress,q,was carried out to study the deformation characteristics of gravel underchanging Lodes angle,u.The results show that changing Lodes angle,u,causes plastic deformations in all principal stress directions.In addition,a series of experiments with both axial and lateral cyclic loadings was carried out to study the deformation characteristics of agravel specimen under different lateral dynamic conditions.Resultsfrom the experiments indicate that lateral dynamic loads significantlychange the residual and reversible strains in all three principal directions.DOI:10.1061/(ASCE)GM.1943-5622.0001096.2018American Society of Civil Engineers.Author keywords:Large-scale;Truetriaxialapparatus;Gravel;Staticanddynamic;Ovalspecimencap.IntroductionGravel is widely used in earth-rockfill dams,railway and highwaysubgrades,bridge piers,weak foundation treatments,high backfills,and so forth,because it has low compressibility,high shear strength,andstrongantiliquefactioncapacityunderseismicload(Voznesenskyet al.2013).To date,the static characteristics of gravel have beenextensively investigated using conventional triaxial experiments(Lade et al.1996;Suwal and Kuwano 2013).However,in practice,thegravelisinatruetriaxialstressstate,andthedirectionofthemajorprincipal stress alternates.In complex stress states,the mechanicalbehavior of gravel is quite different from the results of conventionaltriaxial experiments,showing inherent and stress stateinduced ani-sotropy.These features are difficult to measure accurately in conven-tionaltriaxialexperiments.Moreover,many large structures made of gravel,such as highearth-rockfill dams,are located in earthquake-prone areas.Projectsdamaged by an earthquake lead to severe consequences.At present,the dynamic characteristics of gravel are generally studied usingconventional dynamic triaxial experiments.The cyclic loading isusually applied in one direction and the initial stress state is underconventional triaxial conditions,which is different from actualdynamic loads;hence,the obtained experimental results cannotthoroughly represent the dynamic behavior of gravel.Therefore,astudy of static and dynamic characteristics of gravel under complexstress conditions,especially under true triaxial stress states,has sig-nificanttheoreticalandpracticalvalue.As one of the most important instruments used in geotechnicalexperiments,the true triaxial apparatus(TTA)has been signifi-cantly advanced since Kjellman(1936)developed the first one.With the development of the soil constitutive theory,it is urgentthat the more complex stress paths are realized to reveal the stress-strain relations in soil more systematically and comprehensively;this call for study greatly promoted the development of the TTAandotherexperimenttechnologies.The TTA can be broadly divided into three categories accord-ing to different boundary conditions.The first type is the rigidboundary apparatus,in which the principal stresses are applied byrigid plates(Ibsen and Praastrup 2002;Matsuoka et al.2002;Ismail et al.2005).One of the most representative apparatus isthe Cambridge TTA(Pearce 1971;Airey and Wood 1988).Thesecond type is the flexible boundary apparatus by which the prin-cipal stresses are inflicted by flexible bladders(Sture and Desai1979;Yamada and Ishihara 1979;Sivakugan et al.1988;Reddyet al.1992;Mandeville and Penumadu 2004;Choi et al.2008;Voznesensky et al.2013;Hoyos and Macari 2001).Ko and Scott(1967)developed a stress-controlled TTA in 1967,which is themodel of many TTAs of the flexible boundary type used today.The third type is the mixed boundary apparatus,which is the com-bination of the rigid boundary type and the flexible boundarytype;that is,the principal stresses are applied by the combinationof rigid plates and flexible surfaces.The mixed boundary TTA1Ph.D.Candidate,State Key Laboratory of Hydroscience andEngineering Dept.of Hydraulic Engineering,Tsinghua Univ.,Beijing100084,China.E-mail:2Professor,State Key Laboratory of Hydroscience and EngineeringDept.of Hydraulic Engineering,Tsinghua Univ.,Beijing 100084,China.E-mail:3Postdoctoral Research Associate,Dept.of Civil and EnvironmentalEngineering,Univ.of Massachusetts Amherst,Amherst,MA 01003(corre-spondingauthor).E-mail:4Professor,State Key Laboratory of Hydroscience and Engineeringand Dept.of Hydraulic Engineering,Tsinghua Univ.,Beijing 100084,China.E-mail:5Senior Engineer,State Key Laboratory of Hydroscience and EngineeringDept.of Hydraulic Engineering,Tsinghua Univ.,Beijing 100084,China.E-mail:6Senior Engineer,State Key Laboratory of Hydroscience andEngineering Dept.of Hydraulic Engineering,Tsinghua Univ.,Beijing100084,China.E-mail:Note.This manuscript was submitted on May 4,2017;approved onSeptember 26,2017;published online on January 9,2018.Discussion pe-riod open until June 9,2018;separate discussions must be submitted forindividual papers.This paper is part of the International Journal ofGeomechanics,ASCE,ISSN 1532-3641.ASCE04018004-1Int.J.Geomech.Int.J.Geomech.,2018,18(3):04018004 Downloaded from ascelibrary.org by Yongkang Wu on 01/09/18.Copyright ASCE.For personal use only;all rights reserved.was first invented by Green(1971),and the apparatus designed byLade and Duncan(1973)and Lade(1978)is also very famousbecause of the advantages it has for reducing friction betweentwo adjoining boundaries and controlling confining pressure eas-ily in the chamber.The mixed boundary apparatus is the one mostwidely used(Lo et al.1994;Lade and Abelev 2003;Shapiro andYamamuro 2003;Alshibli and Williams 2005;AnhDan et al.2006;Leo et al.2008;Yin et al.2010).Although significant progress has been made on TTAs,most oftheapparatusesarestilllimitedinsize,loadingcapacity,ordynamicperformance,which means they cannot meet the requirements ofthe static and dynamic true triaxial experiment research of gravel ata high-stress state and other complex conditions.To meet theserequirements,theTTAsshouldaddressthefollowingconditions:1.To test gravel of different particle size gradations,largerspecimen sizes are needed.However,the specimen size ofmost available apparatuses ranges between 44?44?44and 102?102?102 mm.Large specimens,necessary forcoarse-grained materials such as gravel,are rarely tested in athree-dimensional(3D)stress space because of the size limi-tations of the apparatus.2.The height-width ratio of the specimen should be considered.The specimen of some large-scale TTAs designed for gravel arecubic,such as 241?241?241 mm(Choi et al.2008),whichmay affect the formationof shearbands during the experiment.3.A large confining pressure is needed.The maximum confiningpressure applied by most of the apparatuses is less than1.0 MPa,which cannot meet the stress-state demand of highearth-rockfill dams and other large projects.4.The ability to apply dynamic stresses is required.Most of theapparatuses can only perform static experiments or conven-tional dynamic experiments with cyclic loading in the axialdirection.However,the dynamic mechanical characteristicsunder a true triaxial state are also essential.5.It is well worth making the control method more convenientand making measurements easier and more accurate,especiallyfor the pore-water pressure and the strain in the direction of theminor principal stress.It is also very important to keep thespecimen symmetrical during deformation in experiments.To meet the conditions outlined in the previous paragraph,alarge-scalemixedboundarytypestaticand dynamictruetriaxial ap-paratus was developed at Tsinghua University and is referred to astheTHU-SDTTA.Thispaperfirstintroducesthedesign,characteris-tics,and experimental procedure of the THU-SDTTA.Then theresults from a series ofexperiments with cyclicloadings at differentdirections and frequencies are presented to verify the applicabilityand reliability of the THU-SDTTA.Finally,the static and dynamicmechanical characteristics of gravel under the conditions of a truetriaxial stress state are presented on the basis of a true triaxial staticexperiment of complex stress paths and a series of experimentsunderdifferentlateraldynamicconditions.Design ofthe THU-SDTTAThe configuration of the THU-SDTTA is shown in Fig.1.The ap-paratus consists of a hydraulic pressure system,loading frames,ameasurement system,a control system,and auxiliary equipment.The hydraulic pressure system includes hydraulic rams,water chill-ers,and alternating current(AC)servomotors.The loading framesinclude a vertical loading frame,a horizontal loading frame,and apressure chamber.The measurement system includes load cells,LVDTs,pore-pressure transducers,and a volumetric-change mea-surement device.The control system includes servovalves,digitalservocontroller,and a computer.The auxiliary equipment includesaspecimencompactionmold,acrane,andsoforth.The vertical load and the lateral load in one horizontal directionwere applied separately bytwopairsof rigidloading plates.Each ofthe two directions can be set as the major principal stress directionor the medium principal stress direction.The minor principal stressisappliedbywaterpressureinthechamber,andthemaximumstaticconfining pressure can reach 20 MPa.To test gravel of differentparticle-size gradations,specimens of different sizes were used.The maximum specimen size was 200 mm in width(x-direction),200 mm in length(y-direction),and 400 mm in height(z-direction).With a height-width ratio of 2:1,the influence of loading plates onthe failure surface can be reduced(Kramer andSivaneswaran 1989;Shapiro and Yamamuro 2003;Lenart et al.2014).The THU-SDTTA is one of the worlds largest static and dynamic TTAs,andit can also be used for experiments of other materials,such as clay,sand,concrete,softrock,andthelike.Hydraulic Pressure SystemThe lateral and vertical loads and the confining pressure areapplied by separate hydraulic rams and AC servomotors.Therange of static stress in the lateral and vertical directions isapproximately 050 MPa,and the dynamic stress range is approx-imately 025 MPa.The frequency and the displacement rangesare approximately 03.0 Hz and approximately 0100 mm,respectively.The static confining pressure range is approximately020 MPa,and the maximum pressure of dynamic loading is5.0 MPa.To prevent the hydraulic fluid from overheating duringlong experiments under high pressure,water chillers were in-stalled.The hydraulic pressure system is connected to the loadingsystem via hydraulic pipes.Loading FramesThe framework of the loading system is shown in Fig.2.The verti-cal and lateral stresses are applied through the vertical and the hori-zontal loading frames,respectively.The confining pressure isFig.1.ConfigurationoftheTHU-SDTTA ASCE04018004-2Int.J.Geomech.Int.J.Geomech.,2018,18(3):04018004 Downloaded from ascelibrary.org by Yongkang Wu on 01/09/18.Copyright ASCE.For personal use only;all rights reserved.applied by pressurized water in the steel chamber.Both the top andthe bottom of the vertical loading frame were equipped with actua-tors and load pistons to ensure the horizontal symmetrical plane isundisturbed while a specimen deforms.The horizontal loadingframe was equipped with a self-balancing system and steel pulleysthat can roll smoothly on steel rails to ensure that the frame movesalong a horizontal line symmetrically so that the specimen candeform symmetrically in this direction.Therefore,the loading sys-tem ensures that the three symmetrical planes of the specimen areundisturbedduringdeformation.Fig.2.LoadingframesintheTHU-SDTTA:(a)photograph;(b)schematicdiagram ASCE04018004-3Int.J.Geomech.Int.J.Geomech.,2018,18(3):04018004 Downloaded from ascelibrary.org by Yongkang Wu on 01/09/18.Copyright ASCE.For personal use only;all rights reserved.The vertical and the lateral rigid sliding plates connect to eachother by sliders and grooves to allow relative movement in the con-finingpressurechamber(Leoetal.2008)andtoreducedisturbancesand frictions between the plates.The grooves and slides weresmoothed by polishing and painted with a certain thickness ofgrease.The stable and small friction coefficients of the loadingframes were measured before the instrument is used.In strain-controlled experiments,the friction needs to be deducted from thestresses,while in stress-controlled experiments,the friction needstobecalculatedinadvanceandaddedtotheappliedstresses.The THU-SDTTA is a mixed-boundary-type TTA.The verticaland the lateral loads in one horizontal direction are applied by twopairs of rigid loading plates.The stresses of the specimen appliedby the rigid loading plates(caps)may not be ideally uniform,andthe strains that result from displacement of the plates(caps)alsomay not be ideally uniform.These kinds of limitations exist in alldevices that userigid plates(caps)toapplyloads anddisplacements(Yinetal.2010).Measurement SystemOn each loading surface,one LVDT is equipped.The load cells areequipped at both ends of the vertical loading frame,while only oneend of the horizontal loading frame is equipped Fig.2(b).In mostof the mixed-boundary-type TTAs,the strain in the minor principalstress direction is converted from the volume change.The responseof the volume change is slow under the action of a high-frequencydynamic load,and the measurement contains significant errors.Toovercome this disadvantage,a pair of water submersible LVDTs,with the thread outside,were installed in the middle of the cylindri-cal pressure chamber at the direction of the minor principal stress(Fig.3).The volume monitoring was also used as an additionalcheck.The LVDTs were fixed on the frames or the confining pres-sure chamber.The apparatus was set on a concrete foundation withathicknessof1m.Theframesaremadeofthicksteel,andthethick-ness of the confining pressure chamber is 50 mm.Hence,thedeforming or displacing of the mounting fixtures for the LVDTscanbeneglected.Control SystemWhen the frequency of the dynamic load is high and the deforma-tion is significant,the stability of the load is difficult to maintain,and the loads in different directions are difficult to coordinate whenmultidirection dynamic loads are applied.To solve these difficul-ties,servovalves and accumulators were used in each source ofpressure(Santucci de Magistris et al.1999).Inaddition,a varietyofcommon load combinations were programmed in advance to beautomaticallycontrolledbycomputers.Auxi