Addressing the criticality of IGBT modules in power electronics and reliability challenges in industrial applications, this study focuses on solving the important problem of solder layer aging failure in IGBT modules. Under defined operating conditions, a specific set of electrical parameters, notably the collector-emitter saturation voltage Vce(sat), serves as a critical indicator for monitoring the degradation of solder layers within IGBT modules. This parameter is indicative of potential solder layer aging failures, which can be identified and studied to predict and mitigate such failures effectively. In order to analyze the aging trend more deeply and accurately, a new nature-inspired meta-heuristic algorithm based on Crested Porcupine Optimizer is introduced. Combining multiple data during the operation of the IGBT module, such as gate voltage, gate current, heatsink temperature, case temperature, collector-emitter current, and Vce(sat), the parameter optimization is carried out by using CPO, and a prediction model based on the optimization of the SVR by CPO is established, which is successfully implemented to predict the accurate prediction of IGBT module solder layer lifetime. The study also validates the effectiveness of the CPO-SVR method on an accelerated aging dataset of IGBT modules provided by NASA Lab. The results of this research contribute to an in-depth understanding and prediction of the aging process of IGBT modules, which in turn improves the reliability and lifetime of the devices.
Read full abstract