Three parallel algorithms for classical molecular dynamics are presented. The first assigns each processor a fixed subset of atoms; the second assigns each a fixed subset of inter-atomic forces to compute; the third assigns each a fixed spatial region. The algorithms are suitable for molecular dynamics models which can be difficult to parallelize efficiently—those with short-range forces where the neighbors of each atom change rapidly. They can be implemented on any distributed-memory parallel machine which allows for message-passing of data between independently executing processors. The algorithms are tested on a standard Lennard-Jones benchmark problem for system sizes ranging from 500 to 100,000,000 atoms on several parallel supercomputers--the nCUBE 2, Intel iPSC/860 and Paragon, and Cray T3D. Comparing the results to the fastest reported vectorized Cray Y-MP and C90 algorithm shows that the current generation of parallel machines is competitive with conventional vector supercomputers even for small problems. For large problems, the spatial algorithm achieves parallel efficiencies of 90% and a 1840-node Intel Paragon performs up to 165 faster than a single Cray C9O processor. Trade-offs between the three algorithms and guidelines for adapting them to more complex molecular dynamics simulations are also discussed.

Fast parallel algorithms for short-range molecular dynamics

Behaviour of granular soils subjected to internal erosion involves complex coupling between solid–fluid interaction, skeleton deformation and microstructural evolutions. This paper presents a micro–macro investigation on suffusion in idealized gap-graded and well-graded soils using the coupled computational fluid dynamics and discrete element method. The interaction between soil particles and seepage flow is modelled via momentum exchange between two phases. The progressive loss of fine particles subjected to upward seepage flow at various hydraulic gradients is investigated. The fines content, volumetric contraction and void ratio are monitored to identify the changes of macroscopic states of the soil skeleton. In addition, the microstructural evolution is tracked via particle-scale descriptors such as coordination numbers and force chain statistics. Several clogging–unclogging events which are responsible for the sudden changes of fines content and skeleton response are observed during suffusion. A parametric study indicates that the initial fines content and the hydraulic gradient significantly affect the kinetics of suffusion. Microstructural analyses reveal that the removal of fines is accompanied by the reduction in weak contact pairs and particles with low connectivity.

Suffusion-induced deformation and microstructural change of granular soils: a coupled CFD–DEM study

The flow direction is generally different from the gravity direction in geotechnical structures or slopes, the effect of which during suffusion remains unclear. This paper presents a coupled computational fluid dynamics and discrete element method approach to simulate the particle–fluid interaction relevant to this problem. The CFD-DEM approach is first benchmarked by a classic granular system problem, which is then used to investigate the characteristics of suffusion and its impact on the mechanical behavior. Five different angles between gravity and seepage directions for gap-graded soils with two fines contents are examined. Both the macroscopic and microscopic characteristics during suffusion and triaxial loading tests are analyzed. The direction angle is found to play a significant role affecting the erosion process and the mechanical consequence of soils. The results show that the greater the angle is, the harder it is for suffusion to occur and continue.

Investigating the effect of flow direction on suffusion and its impacts on gap-graded granular soils

Coupled particle-fluid simulations of the initiation of suffusion

1 Key Laboratory of Offshore Geotechnics and Material of Zhejiang Province, Hangzhou, China 2 College of Civil Engineering and Architecture, Zhejiang University, Hangzhou, China 3 Department of Civil and Environmental Engineering, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China 4 Department of Civil and Environmental Engineering, The Hong Kong Polytechnic University, Kowloon, Hong Kong, China

A coupled CFD-DEM investigation of suffusion of gap graded soil: Coupling effect of confining pressure and fines content

Internal instability is a form of internal erosion that can occur in embankment dams or flood embankments where the finer fraction of the material is washed out under the action of seepage flow; if undetected this process can progress to cause embankment collapse. Gap-graded materials are particularly susceptible. Skempton and Brogan [1] proposed that a key contributor to instability is the reduced stress transmitted by the finer fraction and that the magnitude of this reduced stress could be inferred from the hydraulic gradients observed at the initiation of particle migration in experiments. Here Skempton and Brogan’s hypothesis is assessed at the particle scale using a discrete element method (DEM) model coupled with computational fluid dynamics (CFD). This contribution discusses validation of the coupled DEM-CFD software prior to describing the simulation of a permeameter experiment. The simulation generated particlescale data at the initiation of instability by considering a gap-graded sample subject to at a hydraulic gradient of 1.0 (upward flow). The results provide insight into the instability mechanism, most notably showing that while the particles that move under seepage flow do indeed transmit relatively small effective stress, a finite proportion of the particles that move transfer relatively large stresses.

/pdf/coupled-dem-cfd-analysis-of-the-initiation-of-internal-54k43sgopu.pdf

Coupled DEM-CFD Analysis of the Initiation of Internal Instability in a Gap-Graded Granular Embankment Filter

Underfilled and gap-graded soils are known to be susceptible to suffusion; a form of internal instability in which the finer fraction of a soil is washed out from the coarser matrix under the action of seepage. This phenomenon poses a risk to embankment dams and flood embankments. The processes and mechanisms operate at the particle scale, and insight can be gained via the particulate discrete element method (DEM). Vir-tual samples can be created using DEM and simulation results can provide information on particle stresses, as well as quantitative information on the fabric of the particulate material. This is important as the amount of stress carried by the finer particles is thought to govern the susceptibility of a given material to suffusion. DEM modelling can also provide information on variation in properties within samples as well as the detailed data needed to quantify the material fabric. DEM models are, however, an idealization of reality and con-strained in particular by the number of particles used and sample preparation method. This study examines key issues relating to the development of virtual samples for use in DEM analysis and also examines the proportion of the applied stress that is carried by the finer particles.

/pdf/using-dem-to-assess-the-influence-of-stress-and-fabric-5gjtyfwzv5.pdf

Kenichi Kawano

Papers

Coupled particle-fluid simulations of the initiation of suffusion

Coupled DEM-CFD Analysis of the Initiation of Internal Instability in a Gap-Graded Granular Embankment Filter

Using DEM to assess the influence of stress and fabric inhomogeneity and anisotropy on susceptibility to suffusion