Preparation of Monitoring Data for Machine Learning (Fill Ges.m.b.H)

Our client is a mechanical engineering company that records data about the operation of its machines at high frequency. The aim of the project is to obtain early indications of wear of machine parts or other anomalies from this data. We cooperate with the RISC Software GmbH, whereby our task is to prepare, smooth and aggregate the data and to obtain feature vectors from them, from which the RISC Software GmbH is to generate the desired indications by means of machine learning. We use a framework developed by us within the MEVSS CD laboratory for data preparation, which can be flexibly parameterized to generate the desired feature vectors. In the course of the work, the framework will be further developed.