Research Experiences for Undergraduates in Computational Methods with Applications in Materials Science
MSU Affiliation
Center for Advanced Vehicular Systems; Center for Computational Sciences
Major
Mechanical Engineering
Research Mentor
Bohumir Jelinek
Creation Date
7-27-2026
Abstract
Wheel-soil interaction simulations are critical for evaluating the performance of off-road vehicles, construction equipment, and planetary exploration systems. The Discrete Element Method (DEM) provides a physics-based approach for modeling these interactions but requires significant computational resources, making efficient use of high-performance computing (HPC) systems essential. This project evaluates the parallel performance of the open- source YADE DEM Wheel-Soil Simulator, a simulation framework developed at Mississippi State University and based on the DEM soil model of Nakanishi et al. (2020). A benchmark simulation representing a fixed 6.3% wheel-slip condition was selected to provide a consistent basis for performance evaluation. Existing benchmark simulations were examined using multiple CPU thread configurations to characterize the current scaling behavior of the simulator. Execution-time data from previous simulations were collected and analyzed to establish baseline performance and identify existing benchmark cases. These results are being used to guide additional scaling studies across multiple computing platforms, including the Morrill and Hercules HPC clusters, with the objective of determining the optimal number of CPU threads for efficient simulation execution. Performance comparisons will be summarized using speed- up curves and execution-time analyses to evaluate parallel efficiency across different hardware architectures. To improve the reproducibility and maintainability of the simulation framework, a GitLab CI/CD workflow is also being implemented to automate simulation testing and validation using containerized YADE environments. The automated pipeline will provide continuous verification of simulation outputs while simplifying future software.
Presentation Date
Summer 7-30-2026
Keywords
soil modeling, discrete element method
Recommended Citation
Londa, Gillian and Jelinek, Bohumir, "Refining and Scaling Particle-Based Soil Models" (2026). Research Experiences for Undergraduates in Computational Methods with Applications in Materials Science. 26.
https://scholarsjunction.msstate.edu/ccs-reu/26
