Degradation Detection of Power Semiconductor Based on Real-time Vds-on Monitoring Techniques

This paper introduces a real-time degradation monitoring method for power semiconductor devices based on the monitoring of the on-state drain-source voltage, which is a representative degradation predictor. A degradation indicator, K-factor, defined as the ratio between the real-time measured drain-source voltage and the nominal value from the datasheet, is proposed to represent the relative variation caused by manufacturing deviation and degradation. K1, which is the initial value of K, shows the calibration process for estimating the inherent manufacturing variation, while a gradual increase I K toward K2 Indicates device degradation. By integrating the voltage error during operation, the proposed algorithm continuously updates K without requiring precise temperature information. Given that the proposed algorithm does not require any offline commissioning process it can be applied to the prognostics of mass-produced power converters. The feasibility of the proposed algorithm is verified with fault samples with different severity stages.