Data-driven Prognostic for IGBT Failure Precursor Parameter Based on Informer

The reliability of insulated gate bipolar transistor (IGBT) will affect the safety and stability of the power electronic systems directly. Therefore, the research on the failure precursor parameter prediction of IGBT has attracted wide attention. This paper uses the Informer model to predict the failure precursor parameter of IGBT, and the predicted value of this parameter can be used to estimate the remaining useful life of IGBT. The turn-off peak voltage of collector-emitter is selected as the failure precursor parameter by analyzing the failure mechanism of IGBT. Three evaluation indicators are used to estimate the prediction precision of the Informer model, the LSTM model and the Transformer model. The results show that the prediction accuracy of the Informer model is slightly higher than the other two models, but its model training speed is much faster than other models. The Informer model can meet the IGBT failure precursor parameter prediction application well due to the large amount of data generates by long-term operation.