My research topic is “Reliable Predictive Maintenance for Real-Time decision support”, and the research plan is to implement predictive maintenance strategies on real industrial process under the PRUNUS project (Predictive Maintenance from a system perspective).
Industrial Maintenance helps to keep running the process, which consists of different machines, tools, and components. There are different ways, but the two dominants are Corrective and Preventive maintenance. In corrective way, activities perform when unexpected failure occur in operational process and need immediate action to start the process. Preventive maintenance is commonly use, in which planned activities perform to keep the process functional, but it cost high and reduce machines availability due to periodic maintenance. Predictive maintenance emerges as a possible solution, it aims to predict maintenance need by continuous analysis of process. It can help to minimize maintenance cost, prevent unexpected failure, and extend machine availability.
Gathering of data and digital twins
The research is based on Big Data gathered from industrial processes. Development of Digital Twin models and new methods together with signal analysis and process knowledge. These will take place offline and first verified against historical data to ensure the realistic reflection of industrial machines and tools. When the twin model / method is successful, it will run parallel with the real-time process. Later, the developed method will expend further for global optimization of system of system in terms of maintenance.