A hybrid experimental and simulation approach to evaluate the calibration of tool wear rate models in machining

An important part of the tool wear prediction process is the calibration of empirical models which relate the process variables to the tool wear rate. In this paper, finite element simulations together with limited number of experiments were used to calibrate the empirical wear rate models. The cutting tool investigated was uncoated tungsten carbide which was used in orthogonal machining of AISI 1045. Four different wear rate models were evaluated during calibration from two aspects; the least amount of quantitative assumptions involved and independency from the proportions of dominant wear mechanisms. Following on from here, investigation was also carried out to determine the minimum number of experiments for calibrating the empirical models such that the predicted wear rates agree well with experimental results in specific range of cutting parameters. Results showed that Usui’s wear rate model was the most robust as there was no restriction in selecting process parameter magnitude during calibration. It was also possible to calibrate Usui’s model based on two cutting tests only. However, the number of experiments used during calibration should be increased when workpiece material is sensitive to either strain hardening or strain rate.