Knowledge integration for fault severity estimation
|Coordinator||Luleå tekniska universitet - Institutionen för system- och rymdteknik|
|Funding from Vinnova||SEK 4 594 197|
|Project duration||September 2019 - September 2022|
|Venture||Strategic innovation programme for process industrial IT and automation – PiiA|
|Call||Digitization of industrial value chains|
Purpose and goal
Develop an AI solution for digitizing knowledge of machine damage and maintenance needs in paper machines and demonstrate an efficient user interface. By automating routine tasks, maintenance engineers should be able to focus on complex tasks and preventative maintenance. The development of AI-based tools for digitization of domain knowledge, data analysis and decision support is also expected to contribute to the development of scalable condition monitoring and decision support innovations, for example in domains like remote monitoring and equipment performance services.
Expected results and effects
The project´s result goal is to reduce unplanned stops by 25% and loss of production by 4 MSEK per paper machine and year, and to enable 10% remanufacturing of large rolling element bearings through early fault identification. This will be made possible by the development of AI-based tools for analysis of measurement data and documents in the condition monitoring systems and improvement of user interfaces, which is expected to strengthen the companies development and innovation capacity and the business relationship between the system supplier and the paper industry.
Planned approach and implementation
The project is a collaboration between process industries (Smurfit Kappa Piteå and SCA Munksund), a systems supplier (SKF) and researchers in machine learning (LTU and RISE). Measurement data and documentation from modern condition monitoring systems installed in the paper machines will be analyzed and supplemented with in-depth investigations of bearing faults. Machine learning methods and tools for data and document analysis will be integrated to automatically identify faults and maintenance needs in paper machines, and a user interface will be demonstrated.