颗粒流理论模拟砂土的力学性质
2024-04-13
来源:年旅网
维普资讯 http://www.cqvip.com 第14卷第3期 2002年8月 岩土工程师 Vo1.14 No.3 Aug.,2002 Geotecnmical Engineer Simulation of Sand Properties by Particle Flow Code ZHOU JIAN&CHI Yong (Department of geotechnical engineering,Tongji University,Shanghai。P.R.C。200092) Abstract:Based on the theory of Particle Flow Code(PFC),the biaxial test and shear—band evolution of sand are simulated in the paper.The macro stress—strain relation of sand is reproduced by PFC mode1. The variation rules of macro properties of the sand sample with meso—parameters,such as particle size and friction coefficient of particle,are studied in the paper.Comparing the results of sand shear—band in lab test with PFC model,and studying the formation and extension process of the shear—band as meso parameters,such as particle size,stiffness and coefficient,are varied.It is shown that the mechanism of sand shear—band can be simulated perfectly by PFC mode1. Keywords:Particle Flow Code;sand;stress—strain relation;shear—band 颗粒流理论模拟砂土的力学性质 周健池永 摘 要:文中采用颗粒流理论模拟了砂土双轴试验和砂土剪切带的形成与发展机理。通过颗粒流数值 模拟试验再现了砂土应力~应变关系曲线,同时研究了砂土细观参数变化时其宏观力学性质的反应。 对比研究了室内试验与颗粒流试验砂土剪切带的形成与发展机理,同时研究了当细观参数如颗粒大小、 刚度、摩擦系数变化时,砂土剪切带的变化规律。 关键词:颗粒流 砂土 应力一应变关系 剪切带 dient of displacement.Zbib&Aifantis(1989)analyzed 1 Introduction the evolution of the deformation within the shear— band of metals.Their theory was supported by ex— The shear—band can be observed both in the shear test of sample and in the process of foundation slip. perimental observations which are more abundant in metallurgy,than in soil mechanics.They intended to examine the structure of shear—band in soil but did not proceed due to lack of experimental data.Bardet& The formation of shear—band is induced by localization of distortion fields in sand,so the study on the mecha. nism of shear—band in sand should be started from me. somechanical leve1. Proubet(1 9 9 1)investigated the structure of persistent shear—・band in granular materials by numerically simu・・ lating an idealized assembly of two— dimensional parti・・ cles.The displacement,volumetric strain,void ratio, rotation of the particles,are examined within the shear—band.Although the numerical method can t re— place lab test to simulate shear—band,it can be taken as a selected numerical method to simulate the struc. The former study results show that the thickness of shear—band in sand depend on the particle size of sand.Roscoe(1 9 70)and Muhlhaus&Vardoulakis (1987)reported that the width of shear—band is about 8~10 times the mean grain diameter.Hill(1962), Mandel(1963),Rudnicki&Rice(1975)analysed the emergence and inclination of shear—band as a bifurca. tion problem in continuum mechanics.However,they didn't consider the thickness of shear—band since their constitutive equations depended only on the first gra— ture of shear—band.Cundall(1989),(1990),and (1991)modeled shear—band formation by FLAC.One aspect that is not treated well by FLAC is the thick— ness of a shear band.In reality,the thickness of a 维普资讯 http://www.cqvip.com 2 岩土工程师 — 第3期 band is determined by internal features of the material,such as grain sizeThese features are not built in— .course of a simulation.The calculation cycle is illus. trated in Figure 1. to FLAC constitutive modelsAlthough the overaII .physics of band formation is modeled correctlv bv FLAC,band thickness and band spacing are grid—de— pendent. Furthermore,if the strain—softening mode1 一Update panicle+waI posit OnS一and set of contacts 1 Law ofMot Oil Force—Displaceme (applied to each particle) (applied to each c is used with a weakening material,the load/displace. ment relation generated by FLAC for a simu1ated test is strongly grid—dependent.This difficulty is not a major concern in PFC,because the program is used to model behavior at the particle IeveI. The theory of particle flow is introduced and the biaxial test and the formation of shear—band in sand are simulated in the paper,and the relation of meso structure and macro mechanical response of sand is obtained. 2 Theory Background 2.1 Assumptions PFC2D provides a particle—flow model containing the following assumptions. (1)The particles are trea ted as rigid bodies. (2)The contacts occur over a vanishingly small area(i.e.,at a point). (3)Behavior at the contacts uses a soft—contact approach wherein the rigid particles are allowed to o— verlap one another at contact points. (4)The magnitude of the overlap is related to the contact force via the force displacement law,and all overlaps are small in relation to particle sizes. (5)Bonds can exist at contacts between particles. (6)A1l particles are circular;however,the clump logic supports the creation of super—particles of arbi. trary shape.Each clump consists of a set of overlap.. ping particles that act as a rigid body with a deform. able boundary. 2.2 Theory of PFC The calculation cycle in PFC2D requires the re. peated application of the law of motion to each parti— cle,a force—displacement law to each contact.and a constant updating of wall positions.Contacts,which may exist between two bails or between a ball and a wal1,are formed and broken automatically during the relative motion resuhanl tbi’ce J—moment constitutive law Contact force Ftgure 1 Calculation cycle in PFC2D 2.2.1 Force—Displacement Law The force—displacement law relates the relative displacement between two entities at a contact to the contact force acting on the entities.There are two types of contact in PFC model that is ball—ball and ball—walI contacts. The contact force vector F can be resolved into normal and shear components with respect to the contact plane as Fi—F7+F; (1) where F and F denote the normal and shear component vectors,respectively. The normal contact force vector is calcuIated bv F7一K U n (2) where K is the normal stiffness at the contact. The value of K”is determined by the current contact— stiffness mode1. The shear contact force is computed in an incre— mental fashion.When the contact is formed,the to— tal shear contact force is initialized to zero.Each sub— sequent relative shear—displacement increment results in an increment of elastic shear force that is added to the current value. The shear component of the contact displace— ment—Increment vector occurring over a timestep of △t is calculated by △U;一 △ (3) and is used to calculate the shear elastic force—incre— ment vector △F;一一K △U; (4) Where K is the shear stiffness[force/displace— ment]at the contact.The shear stiffness is a tangent modulus. The new shear contact force is found by sum— 维普资讯 http://www.cqvip.com 第3期 岩土工程师 ming the old shear force vector existing at the start of the timestep with the shear elastic forceincrement vector F;一F +△F; (5) The values of normal and shear contact force de— termined by Equations.(2)and(5)are adj usted to satisfy the contact constitutive relations. 2.2.2 Law of Motion The motion of a single rigid particle is deter— mined by the resultant force and moment vectors act— ing upon it and can be described in terms of the trans— lational motion of a point in the particle and the rota— tional motion of the particle. The equations of motion can be expressed as two vector equations,one of which relates the resultant force to the translational motion and the other of which relates the resultant moment to the rotationa1 motion.The equation of motion can be written in the vector form as F 一;97( 一g。)translational motion (6) Ml—Hl rotational motion (7 Where F is the resultant force,the sum of a11 externally applied forces acting on the particle: is the total mass of the particle;and gf the body force acceleration vector(e.g.,gravity loading);M is the resultant moment acting on the particle;and H is the angular momentum of the particle. 3 PFC Simulation of Sand Property 3.1 Stress—strain Simulation Based on the biaxia1 test results of Fumio Tat— suoka etc.(1986),the stress—strain property of sand is simulated from mesomechanical leve1.The initial sample dimensions are ho(height)一10.5cm,wo (width)一4era and Zo(1ength)=8cm.The ratio ho/ w0—2.6 was employed to minimize the effects of end friction on the results.Plane strain compression tests were performed at an axia1 strain rate 0.25 /min. For al1 the tests,fresh Toyoura Sand was used which has a mean grain size of 0.16mm,a uniformity coeffi— cient of 1.46,a specific gravity of 2.64 and an angu— lar to sub—angular shape.The stress—strain property will be modeled by PFC model firstly,and adjusting the meso—parameters of the PFC model to ana1vze the eitects of meso—parameters variation on the macro properties of the sand. The PFC specimen—genesis procedure is com— prised of the two steps.Firstly,a rectangular speci— men consisting of arbitrarily placed particles confined by four frictionless walls is generated,and then the particle parameters are set to the particle assembly. The size distribution of the particles satisfies a uni— form distribution with specified values of minimum and maximum radii(Figure 2).The control parame— ters are shown in Table 1.The comparison of nu— merical results and the biaxial test results is shown in Figure 3. Table 1 Sand parameters in PFC sample ),L (a) (a)Wall(b)Wall and particles(before expansion)(c)Speci— men after particle expansion(d)Specimen after equilibrium(e) Specimen after loading Figure 2 Simulation of biaxial lab test of sand specimens by PFC2D 5 4 董3耋2 胃 1 0 0 2 4 6 8 Ⅸ s 面( Figure 3 Comparison between PFC model and lab test result 维普资讯 http://www.cqvip.com d- ∞譬∞趸 4 岩土工程师 念 ∞∞。耵s鼍ad趸第3期 & O 6 0 O 0 O 2.0 O 2 4 6 8 10 axial strain(%) Figure 4 Stress-strain curves of various friction coefficient The Dartic1es in PFC2D is different from real sand sample,and the porosity is not the real porosi— ty,so the higher particle friction coefficient is needed to reproduced the soften property of dense sand. 3.2 Effects of Mesoparameters on Macroscop— 舍dI三一∞∞2苗 p(B ic Properties 3.2.1 Particle Radii and Packing In order to examine the effects of particle size and Dacking on the macroscopic moduli and peak strength,the particle radii of PFC sample is varied from R 一0.36mm to R 一0.8mm.Observing the variation range of peak strength and moduli when the particle radii are changed.The peak strength versus different confining stress in different particle radii is shown in figure 5. 0 0.2 0.4 0.6 0.8 1 confining stress(MPa) Figure 5 Relation of peak strength and confining stress Figure 5 shows that the responses ot peak strength a11 fall within a reasonable band that results from the heterogeneity of packing in the PFC2D mod— e1s.One could eliminate the effects on the peak strength distribution by adj usting the particle radii and packing. Figure 6,which is stress—strain relation in dif— ferent Darticle size,shows that the initial modulus in— crea ses with decreasing particle size. 4 3 2 2.0 1.5 1.0 0.5 0.0 0 2 4 6 8 10 axial strai ̄(%) Figure 6 Stress strain relation of sand in different particle radii 3.2.2 Relation of Particle Friction Coefficient and Friction Angle For an unbonded PFC model,the peak strength is controlled by particle friction coefficient and the e— lastic modulus is controlled by ball—ball contact mod— ulus and hal1 stiffness ratio. As an unbonded material is loaded,particle rear— rangement arising from sliding at contacts produces a non—linear stress—strain response in which the modu一 1us continua11y decreases until a peak axial load is reached.The peak load will vary as a function ot con— finement and will increase with increasing values of Dartic1e friction coefficient.The friction angle of sand is linearly changed with the variation of particle fric— tion coefficient(see figure 7). 40 E 三 哉35I 一 .—————————一 蜀 I l量30 董25 l ————~-—————_| —————— 20 『 。.一30 40 50 6o 70 8o frictioncoefifcient Figure 7 Relation of particle fricnon coefficient and friction angle 3.3 Shear Band Evolution The emergence and evolution of shear band in sand is simu1ated by tracking the displacement field and volumetric strain curves in biaxial test of PFC mode1. When axial stra【in is about 0.32 ,small spikes are detected in the stress—strain curves before the main peak.They correspond to the beginning of rap— id structura1 changes in sand which ultimately form a failure mechanism with shear bands.The displace一 l 维普资讯 http://www.cqvip.com 一H(=巴毳2 里 oo —第3期 岩土工程师 0000000000000000000 ment fields of sand in different axial strain are shown in Figure 8.Figure 9 shows the variation of confining stress as a function of axial strain.The confining stress is practically constant when axial strain is less than 0.32 .However,it starts to vary when axial strain exceeds 0.32 ,implying that the stress be— comes slightly heterogeneous within the sample as a result of strain localization.Figure 10 shows the vol— umetric strain in PFC mode1.When axial strain is a- bout 0.37 ,the sand sample starts to dilate. (a)占=0.32% (b)占=2.8% (c)占=10% Figure 8 Displacement fields of sand in different axial strain 43 5 4 3 3 4 3l 429 427 425 423 42l 4l 9 4l 7 4l 5 4l 3 4l l 409 407 405 403 40l 3 9 9 Figure 9 Axial strain vs.confining stress curves .1.60 .1.40 .1-20 .1.00 .0.80 .0.60 .0.40 .0.20 0 0.00 0 0.20 v:。1。+ 。 t’ ,q 、 Figure 10 Variation of volumetric strain vs.axial strain Figure 8 shows the emergence and evolution of shear bands in sand. It was found that the results of PFC simu1ation are similar to the lab test results by Fumio Tatsuoka etc.(1986).Namely,it was ob— served that a shear band or shear bands starts to de— velop before the peak stress condition and that even before the clear formation of shear bands the deform— ation within a sample is not uniform with the degree of non—uniformity being larger at lower confining stress.In contrast to this,the degree of the concen— tration of deformation to a shear band or shear bands was larger at a larger confining stress,and the aver— age width of shear band was smaller at a larger confi— ning stress.(See Figure 11).It is also observed from Figure 1 1 that larger failure area was formed at lower confining stress and linear failure characteristic was formed at larger confining stress. 0.05 0.1 0.5 4.0 Figure 1 1 Displacement fields of PFC model in different confining stress(kgf/cm ) 3.4 Mesoparameters Effects on Shear Band 3.4.1 Particle Size The displacement fields of different particle radii at the end of test were recorded as Figure 12.It is shown that the particle size has obviously effects on the width and inclination of persistent shear band. The inclination angle and the width of shear bands increase with the increasing of the particle radii.The larger failure area or shear bands was formed at smal— ler Darticle radii and linear failure characteristic was observed at larger particle size. (a)r=0.36mm (b)r=O.4mm (c)r--0.8mm Figure 1 2 Displacement fields of sand in different particle radii 3.4.2 Particle Contact stiffness Observing the variation of shear band character— istic in sand with different particle contact stiffness when the other parameters of the sample were con一 维普资讯 http://www.cqvip.com 岩土工程师 第3期 stant.The intersection of shear band at the upside of the sample when the stiffness is lower,and the inter— section fall with the increasing of stiffness and form shear bands.The direction and inclination of shear band also vary with the changing of particle contact stiffness(See Figure 13). Ec=0.3GPa Ec=O.63GPa Ec:1GPa Flgure l3 Displacement fields o{sand in different contact stiffness 3.4.3 Particle friction coefficient Figure 14 shows the displacement fields distribu— tion of different friction coefficient.It is shown that the width of persistent shear band decrease or more centralize with the increasing of the particle friction coefficient,and the inclination,direction and 1ocation of shear band also vary with the changing of friction coefficient. (a)l=20 曩 (b)I=40 (c)t 60 (d), 80 Figure 14 Shear band of sand in different friction coefficient 4 Conclusions The stress—strain relation and the characteristic of persistent shear band in sand has been analysed by PFC.The responses of peak strength all fall within a reasonable band that results from the heterogeneity of packing in the PFC2D models.0ne could eliminate the effects on the peak strength distribution by adj us— ting the particle radii and packing.The initial modu— lus increases with decreasing particle size.The incli— nation angle and the width of shear bands increase with the increasing of the particle radii.It is also ob— served that larger failure area was formed at 1ower confining stress and linear failure characteristic was formed at larger confining stress for sand.The de— gree of the concentration of deformation to a shear band or shear bands was larger at a larger confining stress,and the average width of shear band was smaller at a larger confining stress. The intersection of shear band at the upside of the sample when the particle contact stiffness is 1ow— er,and fal1 with the increasing of stiffness.The fric— tion angle of sand is linearly changed with the varia— tion of particle friction coefficient.The strain of sam— pie when shear band emergence increase with the in— creasing of friction coefficient. In order to reproduce the peak strength of real sample by PFC2D,the value of particle friction coef— ficient adopted in PFC model of the paper is very large.It is desired to obtain more perfect results by improving the contact bond constitutive type of the mode】. AcknowledgEments The financial support of the National Science Foundation(grants 50178054)and the Key Subject Construction Proj ect of Shanghai municipality are ac— knowledged. References 1.Bardet,J.P.&Proubet,J.1991.A numerical investiga— tion of structure of persistent shear bands in granular media. Geotechnique,4l(4):599~6l3 2.Chi Yong.Mesomechanical Research on the Engineering Mechanical Property of soil——Simulating Stress——strain rela—- tion and shear band by PFC:Doctor.Shanghai:Tongji Uni— versity,2002 3.Cundall,P.A.{ Strack,O.D.L.1979.A discrete numer— ical model for granular assemblies.Geotechnique,29:47~ 65 4.Cundall,P.A.1989.Numerical Experiments on Localiza— tion in Frictional Materia1.Ingenieur—Archiv,59:148~ 159 5.Cundall,P.A.1990.Numerical Modeling of Jointed and Faulted Rock,in Mechanics of Joint— ed and Faulted Rock.Rotterdam:A.A.Balke— ma.1l~18 (下转第4O页) 维普资讯 http://www.cqvip.com 40 岩土工程师 第3期 ②三个测井井底的位移量均为零或接近于零,表 明位移的深度影响范围约为一倍桩长。 ③三个测井测得位移最大值(位移极大值Sma )分 ②K3测井的位移极值为Kl测井的约2倍,表明 由于防护措施的实施,土体位移减少了1/2。 别为l3.85mm、38.55mm和27.94mm,出现部位(极 大值深度H 分别在H=l3.5m、l1.0m和l1.Om处, 表明S 一般出现在测井深度(也即桩长L)的1/3左 右处(即H ≈1/3L),而K 测井处由于受防护孔影 响,极大值深度H 下移至桩长的约1/2处(即H ≈ 1/2L)。 7 结 论 通过采取以上防护措施,对主厂房的防护取得成 功,整个沉桩施工过程中,主厂房设备及仪表运行正 常,通过本工程可以得出以下有益的结论: (1)采用取土卸压孔减小被保护对象处的土体位 (2)土体水平位移与时间(即位移与压桩量)的关 系 移量,结合采取信息化施工,作为沉桩挤土的防护措 施,是可行和有效的。 ①以K2井H=H 深度处为例,当压桩量完成l/ 3时,该处位移达到最终值的34%,当压桩量完成2/3 (2)监测资料表明,由于防护措施的实施,土体位 移减小了一半。事实上,如果增加取土孔直径或减小 取土孔间距,可以更进一步减小挤土效应。 (3)挤土位移的深度影响范围约为一倍桩长,且中 时,该处位移达到最终值的69%。表明位移量与压桩 量基本是成正比关系的。 ②当压桩全部结束3d时,K2、K3测井H 处的位 移仍分别增加0.16mm和0.63mm,表明土体位移具 有滞后性。压桩结束3d后的位移已很小,表明土体巳 基本稳定。 (3)K1、K2、K3三测井位移的比较 部位移较浅部及深部大。 (4)挤土位移量与压桩量基本成正比关系。 (5)挤土位移具有一定的滞后性。 参考文献 [1]《桩基工程手册》编写委员会.桩基工程手册.北京:中国 建筑工业出版社,1995 ①K2测井的5 分别为K1、K3测井的2.78和 1.38倍,表明中部由于挤土叠加效应显著,土体位移 较两侧大。 [2]杨英华主编.土力学.北京:地质出版社。1987 (上接第6页) nique,l987,37:27l~283. 1O.Roscoe,K.H.The influence of strains in soi1 me— chanics.Geotechnique,1970,2O:l29~l7O, 6.Cundal1,P.A.199 1.Shear Band Initiation and E— volution in Frictional Materials,in Mechanics Computing in 1 9 90s and Beyond(Proceedings of 11.Rudnicki,J.W.&Rice,J.R.Conditions for the 1ocalization odf deformation in pressure—sensitive dilatant materia1.J.Mech.Phys.Solids。1975,23:371~ 394 the Conference,Columbus,Ohio,May 1991), Vo1.2:Structura1 and Materia1 Mechanics.New York:ASME,1279~1289 7.Fumio Tatsuoka et a1.1986.Strength and Deform— 12.Zbib。H.M.&Aifantis。E.C.A gradient dependent flow theory of plasticity:application to metal and soil instabili- ties.App1.Mech.Rev.。1989,42(11):295~304 ation Characteristics of Sand in Plane Strain Corn— pression at Extremely Low Pressures.Soil and 13.Zhou Jian et a1.Simulation of biaxial test on sand by particle Foundation,l986,26(1):65~84 flow code.Chinese Journal of Geotechnical Engineering。 2000。22(6):701~704 8.Hil1。R.Acceleration waves in solids.J.Mech. Phys.Solids,10:1~l6. 9.Muhlhaus。H.B.&Vardoulakis,I.The thickness of shear bands in granular materials.Geotech一 1 4.Zhou Jian et a1.The Method of Particle Flow and PFC2D Code.Rock and Soil Mechanics。2000。21(3):271~274