Optimizing Maintenance Strategies Using Predictive Maintenance and Intelligent Data Analysis
If, in an industrial production plant, a machine fault leads to downtime production slows down and enormous costs can result. Modern control systems cannot assess their status in a way that would enable them to derive relevant information for the maintenance engineer. Machine diagnosis and forecasting aims to remedy this shortcoming. The objective is to predict the moment by when maintenance measures must be taken in order to prevent potential damage or downtime. Using methods of data mining and machine learning (e.g. deep learning, which at SCCH is not only used here, but also in object recognition and predicting the behaviour of people in safety zones, for instance) fault forecasting models are generated to find an “early warning point”. The key here is the combination of expertise and data-based fault forecasting models. Partners on the “Smart Maintenance” project are BMW Motors, BRP-Rotax, Messfeld and Montanuniversität Leoben. Partners on the “inFADIA” COMET project are Fronius International, ENGEL Austria and Rubble Master HMH.