The disturbed soils, especially fine grain soils, tend to regain their strength over time due to both the consolidation of the excess porewater pressure and thixotropic behavior of soil particles. In this research, the pile installation process and the subsequent consolidation, the thixotropy and load tests for several test piles were modeled using finite element FE model.
A new elastoplastic constitutive model, which was developed based on the disturbed state concept DSC and critical state CS theory, was implemented to describe the clayey soil behavior.
Pile installation was modeled by applying prescribed radial and vertical displacements on the nodes at the soil-pile interface volumetric cavity explanation , followed by vertical deformation to activate the soil-pile interface friction. The soil thixotropic effect was incorporated in the proposed model by applying a time-dependent reduction parameter, which affects both the interface friction and the soil shear strength parameter. The results obtained from the FE numerical simulation included the development of excess porewater pressure during pile installation and its dissipation with time, the increase in effective normal stress at the pile-soil interface, and the setup attributed to both soil consolidation and thixotropy effects at different times after end of driving.
The FE simulation results using the developed model were compared with the measured values obtained from the full-scale instrumented pile load tests to verify the proposed FE model.
The results obtained from verification indicated that simulating soil response using the proposed CSDSC elastoplastic constitutive model and incorporating soil thixotropic behavior in the FE model can accurately predict the pile shaft resistance. Strictly speaking these cal- with heterogeneous soil properties or inclined pile head loads or culations were Eulerian as the mesh was fixed. The Drucker—Prager inclined penetration the axisymmetry analysis should be extended constitutive model was used for the majority of the calculations, as to a full three dimensional analysis.
More recently, similar calculations were made 2. The pile with radius R is already embedded for state model was used. The mesh is fixed. On the lower edge of the do- for the study of cone penetration in a perfectly plastic material main a velocity is prescribed, i. As in the explicit scheme groundwater cannot be The pile material is modelled with an elastic material. The top of incorporated, this resulted in the study of solely undrained Von the pile is fixed, otherwise the pile will not resist the soil, i.
Numerical frameworks capable of large deformation are readily q available. These, however, often lack advanced constitutive models and proper treatment of groundwater flow as is required for the simulation of jacked pile installation. In this study we present re- sults combining large deformations, advanced constitutive model- pile ling and full groundwater coupling. For the current simulations Tochnog [22] was used.
The numer- ical scheme as implemented in Tochnog is based on the method suggested in [23]. In the suggested operator-split method first L the normal Lagrangian calculation is performed for the non-advec- tive terms and subsequently the advective terms of the Eulerian description are calculated.
In the Tochnog implementation all path dependent terms like material velocity, stress and state variables are advected through the mesh using the SUPG method [24]. H Stress wave effects are excluded in the analyses. The soil is rep- resented by a hypoplastic model [25]. This implies a restriction to jacked displacement piles in sandy soils, as no other soils are con- sidered. The constitutive model restricts the analysis to a single loading—unloading cycle. Modelling pile installation soil In the first approach the pile is fixed and the soil moves along the pile.
This is similar to [26,19]. In this approach the entire pile installation phase is actually a non-stationary flow of soil, as not inflow of material with prescribed velocity all material has passed through the entire domain.
At the end of pile installation a stationary phase is reached and the calculated W stresses and strains are numerically correct. In the non-stationary phase it can be questioned whether this is a proper modelling of Fig. Boundary conditions of the fixed pile approach. The con- 3. Constitutive model vective terms, the quantities flowing through the domain, that are prescribed on the lower boundary are the initial stress, strain, For both approaches the soil will be modelled using a hypoplas- density, and pore water pressures.
On the side boundaries the ticity soil model. Hypoplastic models follow the assumption of [27] material velocity normal to the boundary is set to zero. In order that a cohesionless granular material can be governed by the prop- to ensure convergence of the calculations a force q, in equilibrium erties of the grain skeleton or the contact forces between the grains with the initial stress field, is applied to the top boundary.
This and the density of the material, if effects like cementation and force boundary condition is preferred to a closed boundary, as ap- structural anisotropy are not considered. In the hypoplasticity model this reduction is the current void ratio e, the Cauchy granulate stress tensor Ts and introduced by reducing the deviatoric part of the stress with a lin- the stretching tensor of the granular skeleton D.
This stretching ten- T ear factor after each calculation step. As the strength and stiffness response of the soil due to the stress opposed to elastoplastic models in hypoplasticity the soil behaviour dependency of the hypoplasticity model. As a result Ts will be 0. A critical state, where soil density boundary conditions and material properties to all nodes located does not change with increasing shear deformation is therefore on the line or within the influence radius of this line.
By positioning incorporated in the constitutive model [28]. All is created. The pile penetration is simulated by moving this geom- three are pressure dependent properties. These values have a refer- etry line downward.
All nodes which fall within the influence area ence counterpart at a low reference pressure ed0, ei0, ec0 that can will be mapped to the geometry line. As in this case the geometry be determined in the laboratory and is sufficient to determine the line prescribes material properties as well as a vertical velocity, all characteristic void ratios at all stress levels, using two additional nodes of the pile will move with the same velocity. For each stress state the void ratio will stay between the most dense The moving pile approach allows for a more realistic simulation ed and most loose state ei.
The version of hypoplasticity used in this paper is as proposed by Wolffersdorff [25]. His version of hypoplas- of the jacking process of a displacement pile as the pile is initially located above the soil surface and stepwise penetrates the soil.
Model parameters were derived by Elmi Anaraki Also, the initial conditions in the soil are more easily obtained as there is not a pile pre-embedded in the soil. This time a stress level, [30] for Baskarp sand, following the methods described in [31]. Groundwater flow Both approaches are modelled in a completely coupled frame- geometry line work, whereby the influence of pore water pressure and dissipa- influence radius of tion is taken into account.
The effective stress state changes due to the resulting pore pressure change and therefore, the strength and stiffness of the soil will change. Essentially, consolida- tion occurs at a rate determined by the time scale of the pore water dissipation and the excess pore pressure generation, resulting from soil deformation and the time needed to reach the equilibrium state again by inflow or outflow of pore water. In soils with high permeability the rate of deformation needs to be high before node in soil non-negligible excess pore pressures can develop.
However, the v not prescribed relative rate of pile installation may well be high enough for such Fig. Boundary conditions of the moving pile approach. The change in hydraulic head is h.
The gradient in vm links material deformation resulting from the constitutive equation to the groundwater flow equation. This re- sults in a fully coupled analysis including pore pressure changes resulting from void ratio changes. Mesh used for the centrifuge test simulation, this corresponds to initial and final situation of the fixed pile approach and the final situation of the moving pile gravitational acceleration g and the position relative to the refer- simulation.
Simulation of centrifuge tests Both FEA models described before will be used to simulate three pile installation tests performed in the geotechnical centrifuge of Deltares, as presented in [3]. The test involved installation of an instrumented pile in Baskarp sand at different initial densities. All tests results have been scaled to prototype scale for compari- son.
Material parameters for the hypoplasticity model are given in Table 1. Completely saturated conditions are modelled, with the phreatic surface at the top of the model.
The value used for j is rather high, which is due to the scaled model conditions. An axi-symmetric mesh with quadrilateral elements is used for the calculations, as shown in Fig. The mesh is refined near the pile. This mesh corresponds to the initial and final state of the fixed pile approach and to the final state of the moving pile approach.
An overview of the geometry and boundary conditions is shown in Fig. The mesh radius R and height H of the domain are respectively 5 m and 10 m. Geometry and boundary conditions for the two modelling approaches of the turbed shear strength was found. In both approaches the horizontal velocity on the left and right boundary are set at 0.
In the fixed pile approach the initial stress, strain, density, pore water pressure and pore water velocity are Table 2 prescribed on the lower boundary, as well as a vertical material Overview of the calculations.
On the other hand in the moving listed. The numerical approach, the corresponding model test, the pile approach only the initial porosity is set, while the stress is ob- initial porosity n0 not to confuse with exponent n in the constitu- tained from an initial gravity calculation.
In this case also the ver- tive model , the initial stress state r0, the surcharge load q and the tical velocity is set to zero on the lower boundary. Oscillations in the stress results of the moving pile approach; oscillation length is approximately one element height.
Comparison of calculated and measured stress and density response during pile installation for the fixed pile and moving pile approach. Results to the stress derived from the measured force on the pile base, cor- rected for the stationary pore pressures.
This assumes drained con- In order to compare the numerical results with the centrifuge ditions are valid, which is reasonable in view of the calculated pore tests, the evolution of the vertical effective stress below the pile base pressures as presented in Figs.
The calculated as well as the porosity change near the pile shaft are plotted. The effective stress is compared base. The density evolution is presented for a location 1. Fixed pile; calculated horizontal rxx and vertical effective stress ryy distribution after 5 m of pile installation for three different initial densities. For the comparison with the measured results a thou- ment device on the instrumented pile in the centrifuge test. When the results of the moving pile approach are plotted, an The stress results, averaged for the nodes directly below the pile oscillation is clearly present, with magnitude and period closely re- base, for the fixed pile approach show an overprediction of the ini- lated to the element height.
Each time an element is switched from tial stiffness and the final base capacity for the loose and medium soil material to pile material an unloading is apparent, after which dense conditions. Procedia Eng. Hamad, F. Homel, M. Jassim, I. Methods Geomech. Rodger, A. Internal report, R. Sadeghirad, A. Methods Eng. Sulsky, D.
0コメント