The field of predictive maintenance has been progressing in the last years with the decreasing cost of sensors and the increasing availability of condition monitoring data from components and systems. Despite the substantial progress made in this field in the last years, there are still a lot of challenges to be solved, particularly in transferring new scientific research on features, algorithms and models to real-world industrial applications properly interfacing the different disciplines involved.
There is a gap between the scientific advances on algorithm development and their applicability in the real-world due to the limited clarification of assumptions and prerequisites in each method and a lack of a clear guidance of how and when to use the specific algorithms.
In a modern industrial asset, data-driven predictive maintenance approaches comprise merging of different data sources and balancing system availability, safety and economical risk against maintenance costs. Due to the novelty and diversity of the data-driven predictive maintenance field, the applied methods are very broad and diverse, and clear guidelines and recommendations on the suitability of the different approaches are missing.
The mission of the group "Data-driven Predictive Maintenance of Industrial Assets” is to bring together experts in the field of condition-based and predictive maintenance in order to -Define the most promising gap where predictive maintenance has not yet been leveraged due to missing data analytics processes.