Abstract: The reliability assessments of traditional insulated gate bipolar transistor (IGBT) devices often rely on overvoltage and overcurrent testing or power cycling and thermal testing methods. While these methods require stringent data acquisition conditions and fail to establish a link between accelerated experimental data and actual service life. A method of correlating the two by a Coffin-Manson-Arrhenius extended index mode is proposed, analyzing the relationship between IGBT device reliability and its internal junction temperature, calculating power loss and building the electric thermal equivalence model. Then, more accurate simulation of IGBT electric heating behavior in actual work. Ultimately, based on simulation data, a Coffin-Manson-Arrhenius extended index model is established, circumventing the issue of disconnection between accelerated experimental results and actual situations encountered in traditional methods. Simulation results indicate that this study is accurate and effective for the life assessment of IGBT devices, offering a novel strategy for reliability assessment of IGBT devices, which contributes to enhancing the reliability and stability of IGBT devices.

    Key words: insulated gate bipolar transistor; junction temperature detection; electric thermal equivalence model; power loss; life prediction