A multiscale simulation approach to grinding ferrous surfaces for process optimization

Abstract

A fundamental optimization of a grinding process usually involves expensive equipment and experimental matrices covering a large parameter space. To aid this often cumbersome procedure, here we present three simulation approaches that are intrinsically related and even use the same software, but consider the grinding process at different levels of detail, thus spanning several length scales. Using a molecular dynamics (MD) model, we subject a nanocrystalline carbon steel work piece to grinding by hard alumina abrasives and study material removal and surface topography. A second, much larger MD model allows us to additionally study the microstructural and stress response of a polycrystalline ferritic work piece with a grain size that qualitatively reproduces macroscopic material behavior. Finally, the material point method is introduced as a way of modeling a machining process at the mesoscale in a mesh-free fashion, which is highly advantageous because it intrinsically treats the large deformations during chip formation correctly without the need for repeated remeshing. We discuss which aspects of the grinding process or the work piece quality may be optimized using the adopted approaches, and we show that although our simulations span almost four orders of magnitude in length, the obtained material removal rates agree well. Thus, the presented mesh-free multiscale approach opens new avenues for simulation-aided optimization of grinding processes.