The 15th IEEE International Conference on Electron Devices and Solid-State Circuits (EDSSC)

502 / 2019-02-28 21:16:01

Data-driven IGBT module life prediction model

insulated gate bipolar transistor; failure mechanism; life prediction model

Plenary track > Power Electronics

With the continuous development of new energy power generation and aerospace, insulated gate bipolar transistors (IGBT) are increasingly being used in such demanding operating environments, and their reliability is more and more important. In this paper, the failure mechanism and life model of IGBT modules are summarized, and the accuracy, limitations and advantages and disadvantages of IGBT life prediction model are compared and analyzed. For these models, the failure process of IGBT cannot be detected in real time, and the continuous development of artificial intelligence in recent years, thus forming a data-driven analytical model. By detecting and extracting the electrical parameters in the corresponding IGBT failure process, the algorithm is used to establish the model, so as to predict the failure electrical parameters and replace the IGBT modules that will be invalid in advance, so as to realize the real-time detection of the reliability of the IGBT module.