Interoperable Information Flow as Enabler for Efficient Predictive Maintenance

Industry 4.0 enables the modernisation of machines and opens up the digitalisation of processes in the manufacturing industry. As a result, these machines are ready for predictive maintenance as part of Industry 4.0 services. The benefit of predictive maintenance is that it can significantly extend the life of machines. The integration of predictive maintenance into existing production environments faces challenges in terms of data understanding and data preparation for machines and legacy systems. Current AI frameworks lack adequate support for the ongoing task of data integration. In this context, adequate support means that the data analyst does not need to know the technical background of the pilot’s data sources in terms of data formats and schemas. It should be possible to perform data analyses without knowing the characteristics of the pilot’s specific data sources. The aim is to achieve a seamless integration of data as information for predictive maintenance. For this purpose, the developed data-sharing infrastructure enables automatic data acquisition and data integration for AI frameworks using interoperability methods. The evaluation, based on two pilot projects, shows that the step of data understanding and data preparation for predictive maintenance is simplified and that the solution is applicable for new pilot projects.