Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Quality Improvement of Milling Processes Using Machine Learning-Algorithms

: Frye, Maik; Schmitt, Robert H.

International Measurement Confederation -IMEKO-, Budapest; International Measurement Confederation -IMEKO-, Technical Committee Technical Diagnostics -TC 10-:
16th IMEKO TC10 Conference "Testing, Diagnostics & Inspection as a Comprehensive Value Chain for Quality & Safety" : Berlin, Germany, September 3-4, 2019
Budapest: IMEKO, 2019
ISBN: 978-92-990084-1-6
Conference "Testing, Diagnostics & Inspection as a Comprehensive Value Chain for Quality & Safety" <16, 2019, Berlin>
European Commission EC
H2020; 739592; EPIC
Centre of Excellence in Production Informatics and Control
Conference Paper
Fraunhofer IPT ()
quality improvement; Predictive Process Control; machine learning; artificial intelligence; data preprocessing; Artificial Neural Networks; random forest; gradient boosting

The increasing digitalization and industrial efforts towards artificial intelligence foster the use of Machine Learning (ML)-algorithms in the production environment. Within production, different application areas and use-cases arise for the usage of ML. In this paper, we focus on the implementation of ML-algorithms for a milling process where critical process conditions are predicted. Based on the predicted process conditions, the machining parameters can be adjusted in advance to avoid critical conditions of the process. The avoidance of critical process conditions increases the quality of the products, since quality characteristics such as surface roughness or dimensional deviations can be influenced. To ensure the transferability of the results to other applications, we follow a methodical approach. The results of the ML-models are discussed critically and further steps are derived in order to use ML-models successfully in the future.