Sustainable energy saving: A junction temperature numerical calculation method for power insulated gate bipolar transistor module

Abstract

Abstract This study is significant to ensure its safe and reliable operation for promoting the sustainable energy saving. The insulated gate bipolar transistor module was widely applied as the core electronic component in energy-saving and new energy industries. A junction temperature numerical calculation method was proposed to calculate the module operating junction temperature that has closed connection with the reliability and lifetime of module accurately in real time. Various parameters were obtained through tests and surfaces that have connected with the degree of the parameters. Junction temperature polynomial fitting equations under each power cycle were established on the above basis. The corresponding polynomial fitting equation was selected according to the aging degree of the module, and the value of junction temperature was obtained by taking the collector current and the saturation voltage. The polynomial fitting algorithm was compared with the junction temperature prediction algorithm based on particle swarm optimization-support vector machine. This study showed that the absolute error of the junction temperature obtained by polynomial fitting equations is mainly concentrated in the temperature and the maximum absolute error is no more than ±5 Celsius degree, and the average relative error of the junction temperature is about 2%. Therefore, polynomial fitting algorithm is more explicit and accurate to calculate the junction temperature of the module. Because the polynomial fitting method simplifies the method of selecting the junction temperature value with the saturation voltage drop as the temperature calibration criterion, and the process of building database between parameters in the whole aging process is omitted, it realizes the on-line measurement of the junction temperature of the module in the aging process, which has remarkable practical value.