<p>In this study, an artificial neural network (ANN) surrogate model is developed to address the computational intensity of the lower-scale discrete-element (DE) representative volume element (RVE) simulation in the hierarchical multiscale off-road mobility model. Accurate prediction of off-road mobility capability is important&nbsp;in making reliable go/no-go decisions in operational planning. The development of high-fidelity off-road mobility simulation, however, poses critical challenges due to the computational complexity stemming from multiple scales involved in the granular terrain dynamics. In consideration of the nature of vehicle-terrain interaction problems, the hierarchical FE-DE multiscale off-road mobility modeling approach was recently proposed by the authors, and it has proven to be effective in predicting vehicle mobility on granular terrain. Despite the massive parallel processing which can be utilized to compute stress responses from many DE RVEs, speedup of the lower-scale DE RVE calculation is required to enable quick prediction of off-road vehicle mobility performance.</p><p>To address this issue, a neural network surrogate model is introduced to the hierarchical multiscale modeling of tire-soil interactions in this presentation. To predict the path-dependent soil failure behavior, the incremental stress and strain relationship of the lower-scale DE RVE is replicated at different states of the stress and strain by exploiting learning capability of artificial neural networks for high-fidelity off-road mobility simulations. By integrating the ANN surrogate RVEs into the multiscale computational framework as a lower-scale model, a new hierarchical FE-ANN multiscale terrain model is developed for high-fidelity off-road vehicle mobility simulations with the scalable parallel computing capability. With the ANN RVEs, the incremental Cauchy stress tensor as well as the tangent moduli can be obtained quickly in each time step. The predictive ability of the new FE-ANN multiscale terrain model is examined for high-fidelity off-road mobility simulations. Furthermore, it is demonstrated that the computational time is substantially reduced by the ANN RVEs due to the fast processing in the lower-scale&nbsp;computation. Details on the formulation, computer implementation, neural network training procedure, as well as the numerical results will be discussed in the presentation at the conference.</p>

Neural Network Surrogated Model for Hierarchical Multiscale Off-Road Mobility Simulation

Paper Proceedings: https://asmedigitalcollection.asme.org/IDETC-CIE/IDETC-CIE2021/volume/85369

ASME 2021 International Design Engineering Technical Conferences & Computers and Information in Engineering Conference (IDETC/CIE2021)

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.