I. Introduction
High-power converter devices not only have obvious advantages in power conversion and power transmission but can also provide high quality and high efficiency power. Therefore, they are widely used in the fields of new energy, rail transportation, and aerospace [1]. The traction converter, as the most important power electronic device in the rail transit system, has shown that about 38% of the traction converter system failures come from the failure of power devices [2]. Among all power devices, the insulated gate bipolar transistor (IGBT) has 42% of the number used because of its characteristics of high current carrying and high switching frequency [3]. As the core device of the traction converter system, the reliability of the IGBT directly affects the reliability of the traction converter system and the rail transit train [4]. Therefore, the reliability study of the IGBT module has received a lot of attention. Lifetime prediction is an important way to improve the operational reliability of IGBT modules. Using the results of the lifetime prediction of the IGBT modules, system maintenance plans can be made on time. It is not only effectively reducing unplanned maintenance but also reduces the system downtime due to converter failure and increases the effective system operation time, which is essential for improving the system reliability.