Multiscale Off-Road Mobility Simulation With Computational Load Balancing for Lower-Scale Discrete-Element Models

Abstract

Scalable parallel computing schemes play an important role in physics-based off-road mobility simulations due to complexities in modeling soil behavior for vehicle-terrain interaction. With the hierarchical multiscale off-road mobility simulation capability, limitations of existing computational deformable terrain models can be eliminated, including the use of phenomenological constitutive assumptions in finite element (FE) approaches as well as high computational intensity of discrete element (DE) models. However, parallel computing algorithms for multiscale simulations need to be carefully developed due to possible unbalanced computational loads occurring in lower-scale RVE simulations, which prevents desirable computational speedup. Therefore, this study aims to develop a scalable hybrid MPI-OpenMP parallel computing framework for hierarchical FE-DE multiscale off-road mobility simulations with a special focus on computational load balancing for the lower-scale DE models.