Molecular dynamics simulation of nano-grinding on single-crystal silicon: A deep learning-based NSGA-II optimization

GC Xu and RX Wang, MATERIALS TODAY COMMUNICATIONS, 46, 112916 (2025).

DOI: 10.1016/j.mtcomm.2025.112916

Nano-grinding is a crucial technique for manufacturing high-precision components from optical brittle materials, such as single-crystal silicon. The grinding conditions significantly influence the quality of the machined surface, yet the interaction mechanisms are complex due to their intricate coupling effects. In this study, molecular dynamics simulation was employed to investigate the effects of abrasive grain radius, grinding depth, grinding temperature, and grinding velocity on surface formation and subsurface damage. To address the coupling effect of the grinding conditions on machined surface characteristics, a deep learning-based NSGA-II optimization framework was developed. This framework integrates a neural network with the NSGA-II algorithm to determine optimal nano-grinding parameters efficiently. The results demonstrate that the proposed framework can efficiently identify an optimal machining parameter set, significantly minimizing elastic recovery and crystal defect depth in approximately 5s. This research not only advances the understanding of the interactive mechanisms between grinding conditions and surface quality but also underscores the innovative application of deep learning methods in the numerical analysis of nano-grinding process, offering a promising approach for optimizing nanoscale machining.

Return to Publications page