During each switching operation, strong thermal loads occur in power electronics, which over time can damage a module to such an extent that it fails. For this reason, special CMOS sensor chips from Hahn-Schickard are to be integrated into power-modules for the first time in the project. These sensor chips shall continuously monitor several signals that allow to detect the causes of damage during operation. The evaluation of the monitored signals by means of artificial intelligence (Embedded AI/Deep Learning) allows to detect damage patterns and material fatigue at an early stage (Predictive Maintenance). Together with a suitable simulation model (Digital Twin), the operating parameters of the modules are then to be optimised in such a way that damage is minimised (leading to an increase in the lifetime and a cost reduction). The practical testing of this approach will be carried out in close cooperation with Schweizer Electronic AG and other project partners.
The project is part of the ECSEL Innovation Action iREL40 with 79 partners from 14 countries. The project brings together world-leading reliability experts and European manufacturing know-how to create a sustainable pan-European reliability community.