Consistent process data processing
The "fourth industrial revolution" is increasingly finding its way into many german industrial companies and offers a wide range of optimization options for manufacturing processes with new solutions. In particular, optimization based on plant and process data has increased in recent years due to the spread of data analysis tools based on AI methods. In this context, data from various phases of the plant life cycle, from planning and engineering to operation and maintenance, are relevant for optimization. The data is often stored in a data center without further description, so understanding, allocating, and processing it is hardly possible without the help of production experts. In addition, data often has different structures, formats and quality. Without continuous, automatic processing, the collection, integration and preparation of data for analysis purposes are therefore very time-consuming and typically take up 70% - 80% of the processing effort in analysis projects. This gives rise to two research questions:
- How can plant and process data be structured, enriched with additional contextual information, and stored in such a way that they become easily accessible and usable for subsequent analyses?
- How can data from different sources be integrated as automatically as possible and made available for specific analysis methods?
Through cooperation with industrial companies, data from real production plants and processes will be considered to address the above research questions. Furthermore, methods for contextualization and storage of data as well as for their integration will be developed. Based on this, a concept for end-to-end data handling processes for knowledge discovery will be developed.