Research On Junction Temperature Estimation In IGBT Modules Based On Multi-model

In recent years,high-power power electronics technology has been widely used,which supports the development of new energy generation,aerospace,rail transit,smart grid,electric vehicles and other emerging areas.With the continuous improvement of power capacity,the reliability requirement of the system is getting higher and higher.Insulated gate bipolar transistors(IGBT)modules are the core components of power electronic systems.The junction temperature is an important parameter for reliability evaluation and has received extensive attention from industry and academia.From the point of view of engineering application and taking electric vehicle as the application background,this paper studies the method of IGBT module junction temperature estimation based on multi-model,puts forward an on-line junction temperature estimation method with engineering significance.The accuracy of junction temperature estimation of the proposed scheme is verified by building a motor driver test platform and a finite element simulation model.Finally,the on-line junction temperature estimation of electric vehicle motor driver is realized.Firstly,this paper introduces several common methods of junction temperature estimation,summarizes the advantages and disadvantages of each method,and analyses the applicability of several methods for engineering applications.Based on the analysis of the applicability of several common junction temperature estimation methods,the scheme for on-line junction temperature estimation is locked according to the characteristics of IGBT module used in electric vehicle motor driver.In IGBT module,a negative temperature coefficient thermistor NTC is usually built in to obtain the temperature inside the module.In many converter products,the temperature value read by NTC in IGBT module is used as temperature protection,but it does not directly reflect the IGBT junction temperature.Based on this protection mechanism,this paper explores the relationship between junction temperature and NTC temperature under specific working conditions,builds a data model between two temperatures through a large number of experiments under multiple heat dissipation conditions.The data model is applied to online conditions to verify the feasibility and accuracy of the method of junctiontemperature estimation.Secondly,by comparing and analyzing several common IGBT module junction temperature estimation methods,it is found that the junction temperature estimation method based on thermal impedance model is more suitable for engineering applications.In this paper,the method of obtaining Foster thermal impedance model and the necessary loss calculation method are introduced.Based on the relationship between NTC temperature and junction temperature,two methods of IGBT module junction temperature estimation based on Multi-model are integrated.Kalman filter in modern control theory is used to combine the two models to obtain the optimal junction temperature estimation.Finally,the differences between the multi-model based junction temperature estimation and the finite element simulation results are compared through experiments and finite element simulation analysis,which verifies the feasibility and accuracy of the multi-model based junction temperature estimation method.The application of on-line junction temperature estimation is extended,and an overtemperature protection strategy of IGBT module in electric vehicle motor driver is proposed,which provides a basis for safe and reliable operation of electric vehicle.