itsowl-TT-iDEPP: Intelligente Diagnoseplattform zur Erkennung von Prozessanomalien in Produktionslinien

The objective of this project is the realization of an intelligent diagnosis platform for automatic process monitoring of modular production systems. The mentioned platform has the ability to learn models and detect anomalies automatically. In concrete, the normal process behavior of each single production module is learned. The learned models can be used for a reliable detection of process anomalies within the production phase of a plant. Following categories of anomalies can be revealed: (i) discrete signal errors (e.g. Sensor defect), (ii) erroneous continuous signal sequences (e.g. defective process) or (iii) timing errors of the overall system (e.g. wearing of drives). The anomaly detection is done by comparison of the learned model with the actual plant behavior during the current production process. This way, the productivity and reliability of the entire process can be increased and the maintenance cycles as well as the downtime of a plant are reduced to a minimum.