Local particle refinement in terramechanical simulations

TL;DR

This work introduces a layer-based local particle refinement strategy for DEM in terramechanics, balancing high surface resolution with coarser subsurface particles to reduce computational cost. By modeling the vertical particle size profile with $d(z) = d_{min} + γ z$ up to $z_{max}$ and discretizing into layers $d_n = r d_{n-1}$, the authors demonstrate through triaxial tests and a coupled pressure-sinkage and shear-displacement experiment that bulk soil properties are preserved within a few percent while achieving substantial reductions in particle count and compute time. Across 36 refined beds, normalized errors stay between $3.4\%$ and $11\%$ relative to high-resolution references, whereas particle counts drop by $2.3$–$25\×$ and run times by $3.1$–$43\×$, with the most aggressive refinements offering larger efficiency gains at modest accuracy costs. The findings support the practical utility of targeted refinement for large-scale DEM terramechanics simulations and point toward adaptive or radial refinement strategies for future work.

Abstract

The discrete element method (DEM) is a powerful tool for simulating granular soils, but its high computational demand often results in extended simulation times. While the effect of particle size has been extensively studied, the potential benefits of spatially scaling particle sizes are less explored. We systematically investigate a local particle refinement method's impact on reducing computational effort while maintaining accuracy. We first conduct triaxial tests to verify that bulk mechanical properties are preserved under local particle refinement. Then, we perform pressure-sinkage and shear-displacement tests, comparing our method to control simulations with homogeneous particle size. We evaluate $36$ different DEM beds with varying aggressiveness in particle refinement. Our results show that this approach, depending on refinement aggressiveness, can significantly reduce particle count by $2.3$ to $25$ times and simulation times by $3.1$ to $43$ times, with normalized errors ranging from $3.4$\% to $11$\% compared to high-resolution reference simulations. The approach maintains a high resolution at the soil surface, where interaction is high, while allowing larger particles below the surface. The results demonstrate that substantial computational savings can be achieved without significantly compromising simulation accuracy. This method can enhance the efficiency of DEM simulations in terramechanics applications.

Figures (11)

• Figure 1: Illustration of the idea of local particle refinement for accelerating vehicle-terrain simulations. Fine particles are used at the terrain surface where they interface with vehicle's tractive devices. Image courtesy of Peter Norlindh at Algoryx Simulation.
• Figure 2: Illustration of a soil bed with vertical particle refinement profile $d(z)$ discretized in layers indexed $n$ with particle size $d_n = r d_{n-1}$ and layer thickness $\eta d_n$. Below $z_\text{max}$, the particles have size $d_\text{max}$.
• Figure 3: The container and the plate with their dimensions. Particles are coloured by their diameter.
• Figure 4: Track sinkage in simulations using particle beds with different scaling aggressiveness. The red curve is the reference case for uniform bed $d = 8.5$ mm.
• Figure 5: Coefficient of resistance, $F_T/F_N$, with the identified region marked. In the three later zones, a chosen peak is marked in each region. The red curve is the reference case for uniform bed $d = 8.5$ mm.