Decision Support by Interpretable Machine Learning in Acoustic Emission Based Cutting Tool Wear Prediction

Abstract

Predictive maintenance is a prominent and active field for applications of machine learning in industry in recent years. The health and wear of equipment directly influences the productivity and quality of the production process. Especially in ultra-precision manufacturing, tool wear has a major impact on the achievable quality while the wear itself cannot be measured directly in-process. In this paper we present a machine learning-based classification of the tool wear in-process using acoustic emission sensors. To increase the interpretability of the process – to open the black box model – we apply a feature importance analysis and use the obtained feature importances to provide augmented data representations to the users. These representations increase the transparency of the model's decision process and assist the users in validating the model's decisions and gain new insight into the phenomenon of tool wear itself.

Authors 7

1. Arno Schmetz Aachen Fraunhofer Institute for Production Technology IPT

   Fraunhofer Institute for Production Technology IPT

   Affiliation as printed

   Fraunhofer-Institute for Production Technology IPT, Aachen, Germany

2. Zengyi Zhen Aachen Fraunhofer Institute for Production Technology IPT

   Fraunhofer Institute for Production Technology IPT

   Affiliation as printed

   Fraunhofer-Institute for Production Technology IPT, Aachen, Germany

3. Daniel Reibert Aachen Fraunhofer Institute for Production Technology IPT

   Fraunhofer Institute for Production Technology IPT

   Affiliation as printed

   Fraunhofer-Institute for Production Technology IPT, Aachen, Germany

4. Sebastian Mayer

   Fraunhofer Institute for Algorithms and Scientific Computing

   Affiliation as printed

   Fraunhofer-Center for Machine Learning and Fraunhofer-Institute for Algorithms and Scientific Computing SCAI, Sankt Augustin, Germany

5. Daniel Zontar Aachen Fraunhofer Institute for Production Technology IPT

   Fraunhofer Institute for Production Technology IPT

   Affiliation as printed

   Fraunhofer-Institute for Production Technology IPT, Aachen, Germany

6. Jochen Garcke

   Fraunhofer Institute for Algorithms and Scientific Computing

   Affiliation as printed

   Fraunhofer-Center for Machine Learning and Fraunhofer-Institute for Algorithms and Scientific Computing SCAI, Sankt Augustin, Germany

7. Christian Brecher Aachen Fraunhofer Institute for Production Technology IPT Laboratory for Machine Tools and Production Engineering (WZL)

   Fraunhofer Institute for Production Technology IPT · RWTH Aachen University

   Affiliation as printed

   Fraunhofer-Institute for Production Technology IPT, Aachen, Germany

   Laboratory for Machine Tools and Production Engineering (WZL), RWTH Aachen University, Germany