Core challenges in the operation of technical systems include monitoring, diagnostics and optimization. The process data of machines and plants contain extremely valuable information on wear, causes of errors, and potential for optimization. However, due to the large amounts of and the high complexity of this information, it is not normally found by humans without the aid of technical support, and is ultimately left unused. Methods of machine learning and model-based anomaly detection have the potential for automated extraction of this information, ensuring the value of this information is not lost.
(1) Potential studies for plant operators: The potential of the data that lies in specific machines is checked (prediction of errors, maintenance requirements, optimization of energy efficiency, increase cycle time / throughput / output, quality), and the methods through which the information can be made available is also determined.
(2) Designing of algorithms for analysis systems as well as for specific machines and plants.
(3) Development, demonstration, testing, and benchmarking of large commercial data analysis systems at real production plants is completed in the laboratory 'Machine Learning' as well as SmartFactoryOWL, which is an initiative of the Fraunhofer IOSB-INA and OWL University of Applied Sciences.
This business unit focuses on the development and use of computer-based models to support complex automation systems. This includes:
Targeted support in specification, modeling, implementatoin, demonstration of
- OPC UA ClientServer/PubSub
- OPC UA information models
- Administrative shells and their submodels
- openAAS runtime environment