Multidimensional position-based monitoring of machining using references of multiple machine tools

Abstract

Tool condition monitoring systems are a prerequisite for autonomous production. However, monitoring small series is challenging as systems usually require references of correct processes for parameterization. As an alternative, these references can be sourced from other machine tools that execute similar processes. Unfortunately, actual processes vary among machine tools, despite machine tools using identical NC-instructions. This work introduces a novel multidimensional position-based online monitoring approach that is designed to overcome process variations among machine tools. This approach utilizes dynamic time warping to match the multidimensional tool paths of machine tools. The results include monitoring of pocket milling on three machine tools using process forces sourced from one, or multiple, other machine tools. A comparison demonstrates that position-based limits for online monitoring, rather than time-based limits, are better suited for transfer among machine tools.