Using Finite Element Method For Pile-Soil Interface .

Journal of Civil Engineering and Construction Technology Vol. 3(10), pp. 256-272, November 2012Available online at http://www.academicjournals.org/JCECTDOI: 10.5897/JCECT12.024ISSN 2141-2634 2012 Academic JournalsFull Length Research PaperUsing Finite Element method for Pile-Soil Interface(through PLAXIS and ANSYS)Mohammad Mahdi Jalali1*, Syed Hasan Golmaei2, Mohammad Reza Jalali1, AlistairBorthwick1, Mir Khalegh Ziatabar Ahmadi2 and Reza Moradi31University College Cork, Cork, Ireland2Mazandaran University, Sari. Iran3Jahad Daneshgahi of Tarbyat modares university, Tehran, IranAccepted 18 October, 2012This study investigated the pile-soil interaction and its effect on the pile settlement and shear stress inthe interaction zone. The impact of loading on the pile displacement was explored applying MohrCoulomb as well as Hard-soil behavior laws. First, the modeling proposed by the PLAXIS Bulletin (No.16) was duplicated in which Hard-soil behavior law was taken into consideration. Next, the modelingtook place through behavior law. The data were analyzed by PLAXIS. Finally, the results obtained fromthe two models were compared to each other. Moreover the dynamic analysis was performed toinvestigate the interface effect between soil and pile through ANSYS software. The results revealed thatthe variance of velocity at the head or the bottom of the pile does not make a significant difference inthe interface coefficient. Finally, the natural frequencies and the mode shape were determined for thepile.Key words: Pile-soil interaction, PLAXIS, Mohr-coulomb, hard-soil, ANSYS, Finite Element Analysis.INTRODUCTIONWhile piles have been used literally for millennia, manyaspects of piling are not to this day modeled with muchvigor. Pile installation, particularly in the case of drivenpiles, can not yet be simulated accurately.Laterally, loaded piles are designed using a semiempirical technique the p-y method- which is used almostuniversally. The frame work of elasto–plasticity has beenusually used for representing the mechanical behavior ofinterfaces.Relative movements between bodies in contact aredescribed either by a local high velocity gradient or by akinematic discontinuity. In both cases, the developmentof a nonlinear behavior within contact, inducing a veryslow rate of convergence of the global solution except formodels of thin layers.Two major kinds of constitutive equations are used formodeling the soil-structure interface behavior, oftenassociated with the Finite Element method. The first one*Corresponding author. E-mail: mahdijalali76 5@yahoo.com.considers the soil-structure interface as a thin continuum(Desai, 1981; Ghaboussi et al., 1973), thus the thicknessof the interface elements should then be specified. In thesecond approach, the interface zone is replaced by atwo-dimensional continuum (Boulon and Jarzebowski,1991; Gens et al., 1989), subjected to kinematicdiscontinuities and exhibiting tangential as well as normaldisplacement jumps. Most interface models wereemployed in Elastic–plastic form by Boulon, Gens, Desaiand Ghaboussi. This paper aims at investigating pile–soilinteraction through changing pile–soil interface coefficientand studying this phenomenon for the settlement of pile,shear stress and forces in the pile.The numerical modeling is going to be carried out bymeans of the Finite Element method as it allows formodeling complicated nonlinear soil behavior and variousinterface conditions, with different geometries and soilproperties. PLAXIS program will be used, which has aseries of advantages:1. Excess pore pressure: Ability to deal with excess porepressure phenomena. Excess pore pressures are

Jalali et al.computed during plastic calculations in undrained soil.2. Soil-pile interaction: Interfaces can be used to simulateintensely shearing zone in contact with the pile, withvalues of friction angle and adhesion different to thefriction angle and cohesion of the soil.3. Better insight into soil-structure interaction.4. Soil model: It can reproduce advanced constitutive soilmodels for simulation of non-linear behavior.MODEL ELECTION: SOIL, PILE AND INTERFACELinear elastic modelIt is the simplest available stress-strain relationship.According to the Hooke law, it only provides two inputparameters, that is, Young’s modulus E and Poisson’sratio ν. It is not suitable because soil under load behavesstrongly in elastically. However, this will be used to modelthe pile.Mohr-Coulomb modelIt is a perfectly elasto-plastic model of general scope,thus, it has a fixed yield surface. It involves five inputparameters, that is, E and ν for soil elasticity, the frictionangle ϕ and the cohesion c for soil plasticity, and theangle of dilatancy ψ. It is a good first-order model, reliableto provide us with a trustful first insight into the roximation of soil or rock behavior. It is recommendedto use this model for a first analysis of the problemconsidered. For each layer one estimates a constantaverage stiffness. Due to this constant stiffness,computations tend to be relatively fast and one obtains afirst estimate of deformations.Hardening-soil modelIt is an advanced hyperbolic soil model. The maindifference with the Mohr-Coulomb model is the stiffnessapproach. Here, the soil is described much moreaccurately by using three different input stiffneses: triaxial loading stiffness E50, tri-axial unloading stiffness Eurand the oedometer loading stiffness Eoed. Apart from that,it accounts for stress-dependency of the stiffness moduli,all stiffnesses increase with pressure (all three inputsrelate to reference stress, 100 kPa). Hardening SoilModel (HS-model) is formulated in the framework ofclassical theory of plasticity. In this model the total strainis calculated using a stress-dependent stiffness, differentfor both virgin loading and un-/reloading. The plasticstrains are calculated by introducing a multi-surface yield257criterion. Hardening is assumed to be isotropic dependingon both the plastic shear and volume strain. For frictionalhardening a non-associated and for cap hardening anassociated flow rule is assumed.In this paper the Finite Element Analysis is used for amonopole installed in the sand. To this end, the PIAXISand two dimensional Finite Element program (version 8)was run. Tables 1 and 2 show the parameters of the soiland pile in two different Mohr-Coulomb and Hard soilmodels.This study attempts to determine the pile-soil behaviorfor different interface coefficients, thus the modelingresults of the two Mohr–Coulomb and Hard soil arecompared to each other. Figure 1 shows the meshing ofthe Finite Element of the mono pile installed in the sand.Moreover, the permeability of soil is the necessary inputof model for calculations.After developing the meshes of the Finite Elements, theprincipal stresses and the total stress in the soil weresimulated for the measurements. In this study, since thesoil is unsaturated and there is no pore pressure, the totalstress would be equal to the effective stress based onTerzaghi relationship. Figure 2 show the effective stressin the soil.To perform the experiment, the loading of the pile wasdone in five separate, successive phases. The loadingvalues for each phase are shown in Table 3.The deformation of the pile is demonstrated in Figure 3after the loading of the pile in soil was completed.FINITE ELEMENT ANALYSISPile settlement in the soilThe settlement of pile for two different interfacecoefficients of Mohr-Coulomb and Hard soil wereestimated and compared to each other, the results ofwhich are shown in Graphs 1 to 7.In these graphs the dimension of settlement is in termsof meter and that of load for installing of pile is Kpa.According to the reference graphs of Vesic andMeyerhof, at the end of loading, as the final extremity ofthe graph gets closer to the horizontal level and thehorizontal asymptote, the measurement of the pilesettlement becomes more accurate. Graphs 1 to 7indicate the settlement of the pile for the interfacecoefficients of 0.7 to 1 in the two Mohr-Coulomb andHard soil models.The measurements based on the Mohr-CoulombBehavior Law are much more accurate compared tothose performed through Hard soil Behavior Law (Graphs1 to 4), when the appropriate interface is taken intoconsideration. However, the results of Graphs 4 to 7shows that Hard soil Behavior Law is more accurate andprecise compared to Mohr-Coulomb Behavior law, whenthe interface is less effective.

258J. Civ. Eng. Constr. Technol.Table 1. Parameters of pile and soil in Mohr-Coulomb sion(kpa)1Poissonratio0.20.3Young modulus(Mpa)9020000Density(Knm -3)2125ModelMaterialMohr-CoulombLinear elasticSandPileTable 2. parameters of pile and soil in the Hard soil (Mpa)13530000Eoed(Mpa)45Density-3(Knm )2125ModelMaterialHard SoilLinear elasticSandPileFigure 2. The effective stress in the soil.Pile forcesFigure 1. The meshing of the pile and soil.Table 3. Loading values of the pile in each phase.Phase12345Loading intensity (Kpa)10002000300060009000The Finite Element Analysis was also performed for theforces developed in the pile. Furthermore, the interfacecoefficients were measured for these forces. For thispurpose, the forces acting at the lower end or bottom ofthe pile and the pile-soil interface were evaluated. MohrCoulomb and Hard soil Behavior laws were taken intoconsideration to compare the interfaces of the two MohrCoulomb and Hard soil models. The results are shown inGraphs 8 to 14.In these figures, the horizontal axis is the loading effecton the upper end or head of the pile during the six

Jalali et al.nt in bottonof pile (meter)Displacementin botton of pile (meter)Sum-MloadALoad in the head of pile (Kpa)Figure 3. The deformation of the pile.Displacement in bottom of pileGraph 1. The settlement of pile for the interface coefficient of 0.70 (Mc and Hs).Displacement in botton of pile(meter)Displacement in bottom of pile259

260J. Civ. Eng. Constr. Technol.MC 0.75H.S 0.75Sum-MloadALoad in the head of pile (Kpa)Sum-MloadASum-MloadAH.S 0.75Load in the head of pile (Kpa)Load in the head of pile (Kpa)Load in the head of pile (Kpa)MC 0.75tonof pile (meter)Displacementin botton of pile (meter)Displacement in bottom of pileDisplacement in bottom of pileDisplacement in bottom of pileDisplacement in bottom of pileGraph 2. The settlement of pile for the interface coefficient of 0.75 (Mc and Hs).Displacement in botton of pileDisplacement in bottomof pileDisplacementin botton of pile(meter)Displacement in bottom of pileDisplacement in botton of pileDisplacement in bottonof pilein botton of pileDisplacement(meter)Displacement in botton(meter)of pile(meter)Load in the head of pile (Kpa)Sum-MloadASum-MloadALoad in the head of pile (Kpa)Load in the head of pile (Kpa)Sum-MloadALoad in the head of pile (Kpa)loadALoad in the head of pile (Kpa)adA(meter)(meter)Displacement in botton of pileDisplacement in botton of pile(meter)(meter)Displacement in bottom of pileGraph 3. The settlement of pile for the interface coefficient of 0.80 (Mc and Hs).Displacement in bottom of pileMC 0.80MCH.S0.800.80