Abstract

A modular computational framework for the identification of friction parameters in metal machining applications is presented. Numerical simulation of such processes using mesh-based techniques (usually) necessitates the re-meshing procedure, which is time-consuming and hard to parallelize. Therefore, the present framework synthesizes the advantages of mesh-free methods with GPU parallel computing, offering an efficient tool for the optimization procedures in thermomechanical modeling of cutting problems. The proposed approach employs an inverse method to determine the unknown coefficients of a temperature-dependent friction model in high-speed metal cutting. We found good agreement between the numerical results and experimental data through a quantitative-qualitative comparison.