Comparative evaluation of reliability assessment methods of power modules in motor drive inverter

The complete percentile lifetime models are essentially required for a mission profile based reliability evaluation of power modules. However, the information on the complete lifetime models or lifetime data of power modules is rarely fully provided by manufacturers. The Monte Carlo method is a good solution to obtain the time-dependent reliability function when only limited lifetime information such as a certain Bx lifetime is given. However, there is a lack of research on the accuracy of the reliability function obtained by the Monte Carlo analysis. In this paper, a time-dependent reliability function of the IGBT module which is obtained by Monte Carlo method from a certain Bx lifetime is compared with that estimated by the complete percentile lifetime model in the case study of motor drive inverter. The complete percentile lifetime model is developed by the use of power cycling test results of 30 IGBT modules.