Lamarckian Particle Filter for IGBT Remaining Useful Life Prediction in Smart Grids

Insulated gate bipolar transistors (IGBTs) find extensive application in diverse power electronic converter systems. Accurate prediction of the remaining useful life (RUL) is of great significance for the prognostics and health management of IGBTs. This paper introduces a novel method for forecasting the RUL of IGBTs utilizing the Lamarckian particle filter (LPF). The technique employs a Lamarckian inheritance mechanism to bolster lightweight particles through the support of their heavier counterparts. This strategy efficiently tackles the issue of particle impoverishment and markedly diminishes the variance in predictive and estimative outcomes. The efficacy of the suggested RUL forecasting technique is substantiated through experiments, leveraging degradation datasets procured from the NASA Ames Research Center. The results show that the LPF method offers improved RUL prediction and estimation for IGBTs.