Integration of an Expert System into the IT-Architecture of a Battery Production Plant

Electromobility is one of the biggest leverages to achieve the goal of climate neutrality. The rising demand for electric vehicles leads to a rapid increase in the production capabilities of batteries in Germany. To cope with the rapid increase in production coupled with the high complexity of the different process steps manufacturers need strategies and tools to optimize the ramp-up of production and find a large enough specialized workforce. One way to counteract this is to systematically process and use expert knowledge and make it available on the shop floor by utilizing an expert system software.

The Fraunhofer Research Institution for Battery Cell Production FFB is developing such an expert system that is based on smart learning methods. Some of the key features are user-specific content that is tailored to the worker’s experience and a smart selection of learning content depending on the worker’s location or the machine status.

Integration into the existing IT landscape poses particular challenges during development. A large number of different data sources and communication protocols must be included, which can result in very high implementation costs. Furthermore, the process for integrating new systems and data sources must be automated to ensure the scalability of the system. To deal with these challenges, explicit interfaces to the expert system were developed at the Fraunhofer FFB. Process and plant data are aggregated in an IIoT-platform and transmitted to the expert system via MQTT according to clusters. A meaningful topic structure enables efficient and lightweight data exchange. The digital twin also plays a crucial role. It serves as a single source of truth and provides an overview of the current plant state and the recorded data. It provides a taxonomy that enables the categorization and tagging of all learning content. This makes it easy to find content and assign it to suitable situations.

In summary, an expert system can help to avoid a shortage of skilled workers and accelerate the ramp-up of production. Clearly defined interfaces to the existing IT infrastructure are necessary to connect data from different sources and link it to the content. One concrete solution is the integration of a digital twin and an IIoT-platform via the MQTT communication protocol.