Research On Thermal Coupling Network Model Of Power Module And Optimization Of Package Parameters

The rapid development of power devices makes them widely used in aerospace,wind power,electric vehicles and other fields.However,the increasingly high-power density and harsh operating environment bring great challenges to the reliability of power devices.The losses of the power module during operation will cause a cyclical temperature in the module,which will lead to alternating thermal stress within the module.Then each layer of material is squeezed or stretched,causing module failure during long-term operation.Therefore,it is necessary to estimate the junction temperature of the module in the running state.So that,on the one hand,measures can be taken to make the module work in a safe temperature range.On the other hand,the junction temperature of the module can be optimized.As the power density of the module increases,the distances between the chips are getting closer.The temperature of the chips will influence each other when the module is working,which will cause further increases of the chips junction temperature.That is,the junction temperature of the chips includes not only the self-heat generated by its own loss but also the thermal influence from other chips.However,the traditional thermal network model does not take into account the thermal effects between the chips.Using the traditional thermal network model to predict the junction temperature of the chips will result in a lower temperature prediction and a larger error of the junction temperature.So,a new thermal network needs to be established to accurately obtain the junction temperature of the power module.Therefore,the heat conductions in the module are analyzed firstly.The thermal cross coupling mechanism is also analyzed based on the lateral conduction of the internal heat in the module.Then a cross-coupling thermal impedance model is established based on the temperature of other chips when a single chip works.Also,the thermal resistances and thermal capacities in the module are obtained through finite element simulation.In addition,the average junction temperatures are compared which are obtained by thermal cross-coupling model,traditional thermal network model and finite element model.Secondly,based on the proposed thermal cross-coupling network model,an electro-thermal coupling model for active load circuits is constructed to predict the chip temperature under different operating conditions.At the same time,an active load test platform is built to verify the model.Further,an electro-thermal coupling model for a three-phase motor is constructed based on the traditional thermal model and the cross-coupling thermal network model respectively to simulate the module temperatures of the converter when the automobile is running in the U.S.urban dynamometer driving schedule(UDDS),which is used to study the influence of the coupling model on the module lifetime.Finally,the thermal resistance of the power module and the thermal stress during operation has been analyzed to optimize the module package parameters by the finite element simulation software.Comparing the junction temperature of the pre-and post-optimization modules under the UDDS,it can be seen that the optimized module obtains a smaller die junction temperature,which can improve the reliability of the module to some extent.