Collaborative Research Network on New Testing Concepts for Power Electronics

For decades, advances in power electronic applications (e.g., renewable energy, electric vehicles, etc.) have been limited by a significant bottleneck: reliability aspect verification and testing time. At every stage of the development process, new designs and technologies must be tested for months or even years to determine the lifetime. For a given testing method, the testing time can be reduced by increasing the stress level. However, it may risk trigging new failure mechanisms that do not appear in normal operation. Testing under accelerated conditions close to normal operation is time-consuming and usually not considered. Moreover, substantial variations among devices further require additional testing samples. Hence the key research question is how to reduce both the duration of the single test and the number of experiments required in reliability testing. However, conventional methods in this field seem to have remained relatively stagnant. Researchers proposed various approaches to balance the testing time and confidence level case by case, yet the testing bottleneck has not been essentially overcome. To address this challenge and provide new testing concepts, this project is to establish a research network to bridge the necessary expertise across borders and disciplines. Through cross-disciplinary research between the areas of power electronics (AAU) and chemical engineering (MIT), this collaborative network will explore new degradation testing and data analysis concepts for power electronic components.