Abstract

In this publication, a novel approach for condition monitoring of power semiconductors is presented. The approach requires no additional sensors and only uses the data that is already available in a power electronic system. The method is based on the analysis of the output control variables of the current controller. Therefore, firstly, the effects of the sensitive parameters that provide information about aging of the power semiconductors on the controller outputs are analysed. This is done by power converter models that allow the variation of sensitive power semiconductor parameters. Based on these results, models are set up, which allow the calculation of sensitive parameters during converter operation - and thus the condition of the power semiconductors. Measurements are performed on the test bench in which the controller outputs are recorded. The chip temperatures of the power semiconductors are measured with a thermal imaging camera in order to evaluate the temperature related effects. Finally, an algorithm is proposed that is able to detect an anomaly of the power semiconductors. The algorithm can even conclude the aging mechanism by detection of a change in the sensitive parameters. Furthermore, detailed data analysis can determine which chip exactly is damaged.