Improved Inverse Solutions for On-Line Machine Tool Monitoring

Abstract

The identification of tool/chip interface temperatures from remote sensor measurements is a steady inverse heat transfer problem that arises in online machine tool monitoring. In a previous paper we developed a set of inverse approaches, vector projection inverse methods, specifically for this problem. These methods rely on two types of sensor measurements: temperatures and heat fluxes. However, because of the extreme ill-conditioning of the tool configuration we studied previously, only a very limited amount of information could be obtained using any of the inverse approaches examined. In an effort to understand the impact of physical parameters on the conditioning of the problem we examined two modifications to the simulated cutting tool: we increased the thermal conductivity of the tool insert, and we reduced the thickness of the tool insert. Inverse solutions were computed on both configurations with all methods for two temperature profiles and various noise levels. The estimated tool/chip interface temperature for the high conductivity tool showed no improvement compared to the original configuration, since the temperature profiles on the sensor surface were unchanged. However, for the thinner tool, the estimated temperatures were substantially more accurate than with the original configuration. With this thinner tool configuration, an optimal set of four sensors could be used to estimate these temperatures at the tool/chip interface to within a few degrees, even with noisy sensor data.