Remaining useful life estimation for thermally aged power insulated gate bipolar transistors based on a modified maximum likelihood estimator

Power electronics converters (PCVs) such as inverters and rectifiers are crucial parts in industrial applications and energy conversion and are commonly used in several power conversion systems. Previous works have only focused on the fault diagnosis of PCVs; however, recently more consideration has been paid on predicting failures. Thus, reinforcing the reliability of power semiconductor devices is crucial for extending the lifetime of the PCVs-based electrical power systems. This paper presents a novel prognostic approach for estimating the remaining useful life (RUL) of the power insulated gate bipolar transistors (IGBTs) used in PCVs. An IGBT-accelerated aging database set under thermal overstress is utilized to evaluate the proposed approach. A data-driven method based on the feature extraction process and a modified maximum likelihood estimator (MLE) technique to RUL estimation of IGBT have been developed and applied. This method is validated to effectively predict the PVC failures based on an accelerated aging experimental data set.