Have you ever wondered how it’s possible that you can obtain credits for participating in a lecture called ‘Industrial Machine Technology’ without ever touching the cold blemished steel of a real industrial machine? Yes, so do we!
The tutorial lecture ‚Software Engineering for Machine Learning Applications in Manufacturing’ goes into its third round and will now incorporate a practical project using real milling machines!
The use of data driven technologies in mechanical engineering does not only require profound knowledge about programming and state-of-the-art machine learning approaches, the main challenge behind it, is the conflation of both, knowledge from the domain of software engineering and the domain of industrial machinery. The new practical project will help to blur the line between these two major aspects.
The aluminum processing milling machines are retrofitted with high frequency acceleration sensors that will be used to sample the machines vibration characteristics for worn and sharp machine tools. Following the CRISP-ML development standard procedure, the students will learn to read out the machines NC-data, link it to the time series data from the retrofitted sensors and use these data to train a model making it capable to predict the current tool condition. The goal is to develop a condition monitoring dashboard application that will determine and display the current tool condition during production.