Sound Based Tool Condition Monitoring for Precision Milling

The wear state of a milling tool is a critical factor influencing the quality of finished workpieces in milling processes. A worn-down tool, especially for long manufacturing processes, can create high costs through wasted machine time and materials. Tool condition monitoring (TCM) systems aim to optimise tool usage. Still, many existing solutions require complex and expensive setups or are intrusive to the mechanical structure and control of the process.
In this thesis, a tool condition monitoring system is introduced based on the capture and analysis of a mechanical manufacturing process which can easily be installed on an industrial milling machine due to its contactless nature. Compared to conventional methods of acoustic data analysis, the captured microphone recordings used to establish the metrics were processed in relation to the current tool angle. For this purpose, the recorded data was split into smaller segments corresponding to specific processing steps. With the application of synchronous demodulation, these angle-dependent acoustic emissions were depicted as characteristic signal patterns ("flower patterns") for each edge of a given tool. As tool damage and wear elicit a change in the emitted machine sounds over time, the "flower patterns" deform as well, especially when the quality of the individual edges declines. Through the comparison of these changes over the lifespan of a tool, metrics were defined and analysed which can provide reliable statements about the condition changes of a milling tool during the manufacturing process.