Coupled smoothed particle hydrodynamics-discrete element method simulations of soil liquefaction and its mitigation using gravel drains
Introduction
Over the past five decades, the U.S. and other seismically active areas have sustained considerable damage resulting from earthquake-induced site liquefaction that was associated with very costly damage to port facilities, bridges, dams, buried pipes, and buildings of all types. The 2011 Tohoku (Japan) earthquake caused an estimated $300 billion in damage. Evidence of wide spread of liquefaction and lateral spreading at unprecedented scale has been observed in many locations and port facilities [1]. Similarly, the 2010–2011 earthquakes that hit the Christchurch city and its surrounding areas (Canterbury Earthquakes, New Zealand) resulted in a devastating damage due to liquefaction. The most common definition of sand liquefaction is that it is a result of water pressure buildup due to squeezing of pore space during rapid earthquake loading, without sufficient time for water to flow through the grains and drain the pressure [e.g., 2]. That is, when the sand is loosely packed, there would be a tendency for the grains to get into a denser configuration during earthquake motion, squeezing pore-water and rapidly increasing the pressure owing to the high bulk modulus of water.
Liquefaction resistance can be improved by: increasing the soil density through compaction, stabilizing the soil skeleton, reducing the degree of saturation possibly by introducing air bubbles into the void space, dissipation of the generated excess pore pressure, and intercepting the propagation of excess pore pressures buildup, among other techniques. Herein, the focus is on gravel drains as one of the widely used liquefaction hazards mitigation method. Sadrekarimi and Ghalandarzadeh [3] argue that the potential benefits of gravel drains include densification of the surrounding granular soil, dissipation of excess pore water pressure, and redistribution of earthquake-induced or pre-existing stress. They also note that the relatively high internal friction resistance of the gravel imparts a significant frictional component to the treated composite, improving both its strength and its deformational behavior.
Computational modeling offers effective means to predict and assess soil behavior in response to earthquakes. In this regard, the coupled (solid-fluid) response of saturated granular soils is commonly modeled using continuum formulations derived based on phenomenological considerations (e.g., the mixture or Biot theories) or homogenization of the micro-mechanical equations of motion [[4], [5], [6]]. Each of these different formulations requires a constitutive model to describe the relationship between effective stresses and strains of the solid phase. For liquefaction problems, constitutive models based on plasticity theory are most commonly used. Several constitutive models have been presented to describe the behavior of saturated granular soils during cyclic loading [e.g, [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19]]. These constitutive relations can be employed in a finite element method or a finite difference formulation to predict the seismic response of a saturated granular deposit. High performance parallel computing simulations have also been presented using continuum-based methods [e.g., 20, 21].
The discrete element method (DEM) provides an alternative effective tool to model granular soils and other geomaterials based on micromechanical idealizations. This method [22] simulates these media as assemblages of interacting discrete particles, and has shown great capability to reproduce the actual behavior of granular soils with simple parameters at the microscale. Numerous attempts have been made at incorporating fluid-particle interaction equations into the discrete element method formulation. One of the popular coupling techniques is to describe the fluid flow by averaged Navier-Stokes equation based on mean multiphase mixture properties and employ well-established semi-empirical equations to calculate the fluid particle interaction forces (e.g. Ref. [23,24]). The fluid equations are discretized over a fixed mesh and solved using a technique known as the finite volume method (FVM). This method has proven to yield satisfactory results in simulating different geotechnical phenomena such as soil liquefaction. It has since gained momentum and was adopted by researchers to model several problems in geomechanics [e.g., [25], [26], [27], [28], [29]]. However, the usage of fixed coarse grid mesh limits its scope and application to the fixed boundary problems.
A more elaborate approach is to model fluid at the pore-scale level to investigate the development of pore pressure due to actual changes in the shape and volume of the pore space caused by particle movements. Zhu et al. [30] developed a pore-scale numerical model using smoothed particle hydrodynamics (SPH) to investigate the flow through porous media. They conducted two-dimensional simulation of flow through periodic arrangements of cylinders and validated the proposed method by comparing the results with those obtained from finite element method. Potapov et al. [31] presented a coupled DEM-SPH method to analyze flows of liquid-solid mixtures. The fluid-particle interaction was obtained by applying no-slip boundary conditions at the solid particles surface. Han and Cundall [32] combined DEM and lattice Boltzmann method (LBM) to simulate flow through porous media at the pore scale. El Shamy and Abdelhamid [33] investigated liquefaction of saturated granular soil deposits by idealizing soil grains through DEM and modeling the pore fluid using LBM. Abdelhamid and El Shamy [34] presented a fully coupled DEM-LBM model to investigate the mechanism of fine particle migration in granular filters. The high accuracy of the pore scale models comes at the price of being computationally expensive, to a degree that makes it impractical to perform numerical simulations with realistic particle sizes using typical desktop computers.
As an alternative to modeling the fluid at the pore scale, SPH could be used to approximate the set of partial differential equations represented by an averaged form of Navier-Stokes equations [35,36] that accounts for the presence of the solid phase and the momentum transfer between the phases. Indeed, SPH is a method that could be generalized to approximate any set of partial differential equations and not necessarily for fluids. For instance, large deformation models of granular materials in a continuum framework have been presented by Chen and Qiu [37], where the saturated soil equations of motions were approximated using SPH. Coupling SPH for the fluid and DEM for the solid phase offers the benefits of overcoming the need for a constitutive model for the solid phase while maintaining the robustness of DEM for large deformation problems and SPH for tracking the fluid motion. Sun et al. [38] presented a Lagrangian-Lagrangian DEM-SPH coupled model for the multiphase flows with free surfaces. They performed dam break and rotational cylindrical tank simulations to showcase the proposed method abilities. Robinson et al. [39] presented a meshless simulation technique based on coupled DEM-SPH algorithm and validated the model by conducting simulations of single particle and constant porosity block sedimentation in a fluid column. Many more examples of coupled DEM-SPH application to various science and engineering problems can be found in the recent literature [[40], [41], [42], [43], [44], [45], [46], [47]].
In this paper, the results of a novel application of SPH-DEM to model soil liquefaction is presented. A key feature of the employed technique is that it does not presume undrained conditions for the granular deposit and allows for spatial fluid movements within the deposit. The responses of loose and dense granular deposits to seismic excitation are first analyzed. As expected, the loose deposit exhibited significant pore pressure development and liquefaction while the dense deposit barely showed any considerable buildup of pore pressure and did not liquefy. A liquefaction mitigation technique through the installation of gravel drains was then introduced to the loose deposit and its effect on mitigating pore pressure buildup was examined.
Section snippets
Coupled SPH-DEM Model
A fully coupled Lagrangian particle-based method is presented herein to analyze the dynamic response of saturated granular deposits subjected to horizontal seismic base excitations. In the SPH scheme, the fluid domain is discretized into a set of individual particles carrying local properties of the fluid such as density and pressure [[48], [49], [50]]. DEM is employed to model the solid particles with proper momentum transfer between the two phases. The presented SPH-DEM technique has several
Validation cases
Two validation tests are presented in this section. Poiseuille flow simulation was performed to validate the SPH-based fluid model and examine the performance of the no-slip, no-penetration boundaries. Furthermore, the accuracy of the coupled SPH-DEM model is demonstrated using particle sedimentation test.
Liquefaction of saturated granular soils
The proposed coupled SPH-DEM approach was used to analyze the response of loose and dense saturated granular deposits as well as modeling gravel drains as a measure to mitigate liquefaction of loose sand deposits. In order to realistically model such boundary value problems, some tools have been utilized to bring the simulations to a manageable size. In this regard, use was made of the high g-level concept commonly used in centrifuge testing to decrease the dimensions of the domain that needed
Conclusions
A three-dimensional fully coupled particle-based model is presented to evaluate the dynamic response and liquefaction of saturated granular deposits and the use of gravel drains as a liquefaction mitigation measure. A microscale idealization of the solid phase is achieved using the discrete element method while the fluid phase is modeled using the smoothed particle hydrodynamics. In this method, the interstitial pore fluid is idealized using averaged Navier-Stokes equations and the
CRediT authorship contribution statement
Usama El Shamy: Conceptualization, Methodology, Investigation, Supervision, Writing-Original draft preparation. Saman Farzi Sizkow: Software, Formal analysis, Data curation, Visualization, Writing-Editing.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgement
This research was partially supported by the US Army Corps of Engineers Engineer Research and Development Center, grant number W9132V-13-C-0004 and the National Science Foundation award number CMMI-1728612. These supports are gratefully acknowledged.
References (72)
- et al.
Modeling of cyclic mobility in saturated cohesionless soils
Int J Plast
(2003) - et al.
Computational modeling of cyclic mobility and post-liquefaction site response
Soil Dynam Earthq Eng
(2002) A simple plasticity theory for frictional cohesionless soils
Soil Dynam Earthq Eng
(1985)- et al.
Modeling deformation banding in dense and loose fluid-saturated sands
Finite Elem Anal Des
(2007) - et al.
Formulation of a sand plasticity plane-strain model for earthquake engineering applications
Soil Dynam Earthq Eng
(2013) - et al.
Calibration and implementation of a sand plasticity plane-strain model for earthquake engineering applications
Soil Dynam Earthq Eng
(2013) - et al.
A unified plasticity model for large post-liquefaction shear deformation of sand
Comput Geotech
(2014) - et al.
Seismic analysis of stone column improved liquefiable ground using a plasticity model for coarse-grained soil
Comput Geotech
(2020) - et al.
Coupled discrete element and finite volume solution of two classical soil mechanics problems
Comput Geotech
(2011) - et al.
Numerical simulation on undrained triaxial behavior of saturated soil by a fluid coupled-DEM model
Eng Geol
(2015)
Understanding the microscopic moisture migration in pore space using DEM simulation
J. Rock Mech. Geotechn. Eng.
Micro-scale modeling of saturated sandy soil behavior subjected to cyclic loading
Soil Dynam Earthq Eng
Liquid–solid flows using smoothed particle hydrodynamics and the discrete element method
Powder Technol
Modeling granular soils liquefaction using coupled lattice Boltzmann method and discrete element method
Soil Dynam Earthq Eng
Three-dimensional simulation of a solid–liquid flow by the DEM–SPH method
J Comput Phys
Fluid–particle flow simulations using two-way-coupled mesoscale SPH–DEM and validation
Int J Multiphas Flow
Numerical analysis of wet plastic particle separation using a coupled DEM-SPH method
Powder Technol
Coupled DEM-SPH simulations of wet continuous screening
Adv Powder Technol
Using SPH one-way coupled to dem to model wet industrial banana screens
Miner Eng
Combined DEM and SPH simulation of overflow ball mill discharge and trommel flow
Miner Eng
Prediction of coupled particle and fluid flows using DEM and SPH
Miner Eng
A coupled SPH-DEM model for micro-scale structural deformations of plant cells during drying
Appl Math Model
A coupled SPH-DEM model for fluid-structure interaction problems with free-surface flow and structural failure
Comput Struct
Simulating free surface flows with SPH
J Comput Phys
Modeling low Reynolds number incompressible flows using SPH
J Comput Phys
SPH without a tensile instability
J Comput Phys
A generalized wall boundary condition for smoothed particle hydrodynamics
J Comput Phys
Observations on the effects of particle rotations on the failure of idealized granular materials
Mech Mater
Experimental micromechanical evaluation of strength of granular materials: effects of particle rolling
Mech Mater
Measured kinematic fields in the biaxial shear of granular materials
Comput Geotech
Comparative study on mesh-based and mesh-less coupled CFD-DEM methods to model particle-laden flow
Powder Technol
The equations of stress in a granular material
Physica A
Diagrams considering well resistance for designing spacing ratio of gravel drains
Soils Found
Effectiveness of vertical drains in mitigation of liquefaction
Soil Dynam Earthq Eng
Preliminary observations of the effects of ground failure and tsunami on the major ports of Ibaraki prefecture
Geotechnical Earthquake Engineering
Cited by (14)
Dynamic response of piles embedded in granular soil to lateral impacts
2024, Soil Dynamics and Earthquake EngineeringThe influence of particle shape on crushing pattern and energy dissipation within a pressurized sand damper
2024, Computers and GeotechnicsSPH-DEM modeling of the seismic response of shallow foundations resting on liquefiable sand
2022, Soil Dynamics and Earthquake EngineeringCitation Excerpt :The main drawback of this technique is the fact that the fluid is assumed to be weakly compressible, which can be compensated for by using a large enough numerical speed of sound that limits the density fluctuations to very small values. The authors previously utilized the above-mentioned SPH-DEM method to analyze geotechnical problems involving soil liquefaction and large deformations, which showed qualitative consistency with published experimental studies [73–75,77]. In this paper, the ability of this coupled scheme to model the seismic response of shallow foundations resting on liquefiable soil is examined.
SPH-DEM simulations of saturated granular soils liquefaction incorporating particles of irregular shape
2021, Computers and GeotechnicsCoupled SPH-DEM simulations of liquefaction-induced flow failure
2021, Soil Dynamics and Earthquake EngineeringCitation Excerpt :DEM is employed to model the solid particles with proper momentum transfer between the two phases. A description of the model components were previously presented by the authors [71] and will be briefly described in the following section for completeness. The responses of level loose and dense level saturated granular deposits overlain by a thin impermeable crust to seismic excitation are first analyzed.
Dynamic mechanical behavior of frozen soil using smoothed particle hydrodynamics
2024, Computational Particle Mechanics