Title: Meshless single grain cutting simulations on the GPU

Authors: Matthias Röthlin; Hagen Klippel; Mohamadreza Afrasiabi; Konrad Wegener

Addresses: Institute of Machine Tools and Manufacturing (IWF), Department of Mechanical Engineering, ETH Zürich, Switzerland ' Institute of Machine Tools and Manufacturing (IWF), Department of Mechanical Engineering, ETH Zürich, Switzerland ' Institute of Structural Mechanics (IBK), Department of Civil Engineering, ETH Zürich, Switzerland ' Institute of Machine Tools and Manufacturing (IWF), Department of Mechanical Engineering, ETH Zürich, Switzerland

Abstract: Physically sound prediction of grinding results necessarily starts with the single grain interaction with the material, which is especially important for engineered grinding tools (EGT). The deformations encountered in single grain simulations are severe, especially due to the large negative rake angles. Conventional software tools using FEM suffer from the induced extreme distortions of the grid, necessitating costly remeshing operations. Meshless methods are not limited in the amount of deformation they can reproduce and thus are a promising alternative, which also has the potential for extreme parallelisation on the graphics co-processor (GPU). With the drastically increased computational speed unprecedented simulation resolutions can be achieved at reasonable computation times allowing parameter studies on different friction and constitutive models with first order complete meshless formulations. In the interest of reproducibility, the software tool developed is made open source and can be obtained from https://github.com/mroethli/iwf_mfree_gpu_3d.

Keywords: meshless methods; GPGPU computing; simulation; metal cutting; grinding; engineered grinding tools; smoothed particle hydrodynamics; particle methods; CUDA; open source.

DOI: 10.1504/IJMMS.2019.103488

International Journal of Mechatronics and Manufacturing Systems, 2019 Vol.12 No.3/4, pp.272 - 297

Received: 06 Apr 2019
Accepted: 14 Jun 2019

Published online: 30 Oct 2019 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article