Development of a greybox model to predict tool wear of PVD coatings for finish turning - GreyProPVD

点击翻译按钮获取中文摘要

Coated carbide tools are frequently used for machining with geometrically defined cutting edges. Depending on process conditions and material combinations, the tools are subject to crater, flank and notch wear generated by abrasion, adhesion, diffusion and tribooxidation. In addition, spalling can occur due to cohesive substrate failure. Despite the significant influence of tool wear on the cutting process, the real application behavior of coated tools cannot be adequately predicted according to the current state of research. While experimental approaches require a high amount of energy and materials resources, purely simulative, FE-based methods are unsuitable due to their high computational effort and time requirements. Both tools are therefore mainly useful for individual investigations that do not consider a wide variation of input variables. However, the extension or combination of simulation-based methods with data-driven black-box approaches offers the possibility to improve the performance and usability of models for tool development. The combined approaches also offer the possibility of integrating changes in tribological and thermophysical tool properties, as well as empirical data, into simulation models of tool and layer development. The mapping of property change is particularly highly relevant for time-resolved wear prediction of PVD coated cutting tools for finishing operations. Main aim of the proposed research project is the configuration of a greybox model to describe the time-resolved continuous tool wear and cohesive substrate failure of PVD coated carbide tools. By describing the wear behavior, an adaptation of different cutting material combinations and process parameters is possible, which leads to reduced tool wear and the avoidance of reject components and thus enables a reduction of resources in material consumption. For this purpose, data based black box models are combined with a simulation-based white box model. The blackbox models are used to derive wear-dependent, time-resolved layer properties and wear constants as well as the prediction of cohesive substrate failure based on the existing layer and substrate residual stresses. The whitebox model is used to calculate contact conditions and thermomechanical loads for time-dependent prediction of local state variables such as normal stress, sliding velocity and tool temperature. The experiments are carried out in continuous cutting conditions. In preparation for the second phase, an extension of the model to interrupted cutting conditions is planned as an intermediate step for milling.