Efficient Simulation of Metal Cutting Processes

Today’s manufacturing techniques demand precise modeling of the metal cutting process. Such numerical modeling has proved to be a fundamental component as it can capture some of the process variables that are not feasible to measure experimentally in metal cutting. These models can also survey the parameter space more comprehensively, providing a basis for the optimization of cutting conditions, tool geometries, and other parameters. However, a recurring question arises: what is an efficient way to do so? One efficient way to do so is to use meshfree methods. Introduced in 1977, the Smoothed Particle Hydrodynamics (SPH) is the most popular meshfree method, as evidenced by a myriad of interesting applications in various fields. Technically, the main advantage of this meshfree type of discretization in metal cutting simulations is its ability to naturally handle large deformations with no theoretical limit and without the caveat of mesh distortion. Literature has it that most of the meshfree metal cutting simulations are performed with particles of uniform spacing, leading to a single-resolution configuration. Needless to say, a finer discretization for the small areas of the domain is desired. In fact, there is a surprisingly low number of accounts that adopt a multi-resolution approach in meshfree metal cutting simulations.

In a nutshell, a review of previous studies reveals that meshfree simulation of metal cutting processes is not really mature when comparing it to other fields of application. Therefore, to partially fill this void, researchers from the Institute of Machine Tools & Manufacturing (IWF) in Switzerland: Mr. Mohamadreza (Mamzi) Afrasiabi, Dr. Matthias Roethlin, Mr. Hagen Klippel, Professor Konrad Wegener, proposed a new study in which they sought to accelerate and increase the size of simulations for an orthogonal metal cutting process by incorporating dynamic refinement into the spatial discretization of the PDEs. Their work is currently published in the research journal, International Journal of Mechanical Sciences.

In their approach, a stabilized meshfree scheme was formulated in the updated Lagrangian framework for the thermomechanical simulation of metal cutting. To further optimize the runtime, a dynamic refinement algorithm via particle splitting was adopted from the state of the art. The research team then implemented a modified Johnson-Cook constitutive model to include the strain softening phenomenon resulting from damage in the machining of titanium alloys. All in all, using the refinement algorithm, the metal cutting process was simulated with four models and cross compared to assess the workability and efficacy of the proposed developments.

The authors presented the reliability of their approach by comparing the results against available FEM/experimental data. Results also showed that the refinement algorithm allowed for high resolution simulations in a more reasonable amount of time, saving almost 70% of the computational effort. As a result of their work, 1 mm of (fairly) high resolution metal cutting could be simulated in about 3 h ending with roughly 12,000 particles.

In summary, a dynamic refinement approach was implemented to leverage the available in-house code, as a groundwork towards optimal adaptivity in meshfree metal cutting simulations. The simulated cutting force and chip morphology were seen to agree with available FEM and experimental data. In fact, almost 70% of the computational cost was saved by employing the dynamic refinement approach. In an interview, with Advances in Engineering, Mamzi Afrasiabi, first author, highlighted that one main contribution of their work was the effective tailoring of the spatial refinement in the updated Lagrangian frame, tuned for an orthogonal metal cutting problem.