Incremental generation of hierarchical meshes for the thermomechanical simulation of NC-milling processes

摘要

For the production of large structural components, the NC-milling process is often run with minimum quantity lubrication only. Additionally, most of the raw material will be removed by using process parameters that lead to a significant generation of heat. To guarantee error-free products, the thermally induced deformation has to be taken into account. Thus, a simulation of the milling process, which calculates the thermal expansion, can be used to predict deviations and to optimize milling strategies. In this paper a new approach for the simulation of the NC-milling process is presented. It combines a pure geometrical simulation with new techniques of adaptive finite element methods. The key for a fast simulation of linear thermo elasticity is a hierarchical mesh, based on hexahedra with additional geometric primitives on the lowest level of refinement. Using hexahedra permits the application of tensor-product approaches to achieve faster computations, whereas the use of constrained-approximation allows for multi-level hanging nodes in the mesh. In turn, hp-adaptive finite element methods can be applied. The investigation focuses on the basic coupling of the simulation components and the incremental generation of suitable meshes for the finite element analysis and thereby completely avoids any remeshing action. It is based on a hierarchical octree that is extended by an adaptive sub-mesh for each cell that is partially removed during the simulated cutting process. Thus, the mesh is incrementally refined and adapted during the simulation. Based on previous experimental and analytic investigations, a complete simulation cycle will be presented. For the prediction of the heat input, an empirical model has been developed in order to ensure a fast simulation. After the remote finite element analysis, local displacements are applied to the position of the milling tool instantly. In this way very accurate results can be achieved that serve as a basis for further process optimization operations.