Improving the efficiency of electrical systems through innovative fault prediction technologies

Innovative fault prediction technologies to improve the efficiency of electrical systems will be developed by R-PODID, a new research project coordinated by the University of Bologna and supported by the Chips Joint Undertaking, the European initiative dedicated to microelectronics and semiconductors. The aim is to develop an automated fault prediction solution, which does not require access to energy-intensive cloud systems, for electric drives, power modules and power devices that can be seamlessly integrated into power converters.

"With R-PODID, we want to work together to integrate artificial intelligence technologies directly into power systems in order to reduce energy consumption, thereby promoting a future in which electrical drives and power devices can operate with greater efficiency, reliability and safety," explains Marco Crescentini, professor at the 'Guglielmo Marconi' Department of Electrical, Electronic and Information Engineering at the University of Bologna and member of the 'Ercole De Castro' Research Centre for Electronic Systems (ARCES), which is coordinating the project. "The innovations that will result from this project can be applied in a wide range of industrial sectors, such as conveyor belts in assembly lines, industrial lighting, automotive traction inverters and heavy vehicle test benches.

The researchers will first work to develop a methodology to generate artificial intelligence models for fault prediction from sparse training data sets, typical in industrial and reliability applications.

The ability to obtain accurate and reliable data is a key element for artificial intelligence-based systems, so the European researchers in the R-PODID project will also investigate innovative sensors and methods for measuring physical parameters to obtain data with a high information content. Demonstrations will be developed using innovative semiconductor technologies, such as silicon carbide and gallium nitride, which complement the models and sensors developed without compromising the performance of the system in terms of energy efficiency. All these elements will enable the development of both 24-hour fault prediction technologies for power converters and electric drives, and reliability prediction sensors in power modules that can ensure proactive maintenance strategies.

Initiated by a consortium of 33 partners and co-funded by the European Commission with more than €7 million out of a total budget of almost €24 million, R-PODID is the first project to be coordinated by the University of Bologna within the Chips JU. This result is also thanks to the ARCES Research Centre’s extensive experience in international microelectronics projects and that reinforces the University's role as a leader in semiconductor research and innovation.