CN-121995141-A - Power module service life estimation method and device

CN121995141ACN 121995141 ACN121995141 ACN 121995141ACN-121995141-A

Abstract

The invention relates to the technical field of power modules, in particular to a service life estimation method and device of a power module, wherein the method comprises the steps of obtaining electrical working condition parameters and thermal working condition parameters of the power module in an operation period; the method comprises the steps of inputting electrical working condition parameters and thermal working condition parameters into a junction temperature prediction model obtained through pre-training, outputting junction temperature increment predicted values of a plurality of sampling points in an operation period, dividing the operation period into a plurality of continuous time periods, wherein the temperature change trend of all the sampling points in each time period is consistent, the change trend between adjacent time periods is different, calculating the service life of a power module in each time period according to the junction temperature increment, and accumulating the service lives of the time periods to obtain the service life of the power module in the whole operation period. The service life estimation method and device can improve accuracy of the service life estimation result of the power module.

Inventors

CHAI HONGSHENG
DENG JING
ZHAO XIANG
WANG YANAN
WANG ZHUO

Assignees

辰致汽车科技集团有限公司重庆创新研究分公司
辰致汽车科技集团有限公司

Dates

Publication Date: 20260508
Application Date: 20260126

Claims (7)

1. A method for estimating the lifetime of a power module, comprising: acquiring an electrical working condition parameter and a thermal working condition parameter of the power module in an operation period; The junction temperature prediction model is obtained by training based on a historical power module sample data set, and each sample in the sample data set comprises the electrical working condition parameter and the thermal working condition parameter of the power module and the junction temperature increment simulation value or the actual measurement value corresponding to the electrical working condition parameter and the thermal working condition parameter; Dividing the operation period into a plurality of continuous time periods, wherein the temperature change trend of all sampling points in each time period is consistent, and the change trend between adjacent time periods is different; And calculating the service life of the power module in each time period according to the junction temperature increment, and accumulating the service lives of the time periods to obtain the service life of the power module in the whole operation period.
2. The method for estimating a lifetime of a power module according to claim 1, wherein the lifetime of the power module in each period of time is calculated by: in the formula, Is the first The power module lifetime in a single time period, Is the first The sum of junction temperature increment of all sampling points in each time period, And Is constant.
3. The method of claim 1, wherein the electrical operating parameters include operating current, operating voltage, duty cycle, switching frequency, and power factor, and the thermal operating parameters include fluid temperature, fluid flow, and ambient temperature.
4. The method for estimating a lifetime of a power module of claim 1, wherein the junction temperature prediction model is a cyclic neural network model.
5. The method for estimating a lifetime of a power module of claim 4, wherein said recurrent neural network model is pre-trained by: Acquiring a historical power module sample data set generated by three-dimensional thermo-electric coupling simulation or rack actual measurement, wherein the historical power module sample data set comprises time sequence data under a plurality of working condition combinations, and the samples of each time step comprise an electrical working condition parameter, a thermal working condition parameter and a junction temperature increment corresponding to the electrical working condition parameter and the thermal working condition parameter; Dividing a historical power module sample data set into a training set and a verification set according to a preset proportion; And constructing a circulating neural network model, training the circulating neural network model by adopting a training set with the aim of minimizing the mean square error between the predicted junction temperature increment and the real junction temperature increment, and monitoring the training process based on the verification set until the loss function value on the verification set is lower than a preset threshold value.
6. The method for estimating life of a power module according to claim 1, wherein if the junction temperature increment predicted value is zero or more, the characteristic state of the corresponding sampling point is marked as a temperature rising state, otherwise, the characteristic state of the corresponding sampling point is marked as a temperature lowering state, and when the characteristic states of two adjacent sampling points are different, the latter sampling point is determined as a junction temperature change inflection point and is used as a starting point of a new time period.
7. A power module life estimation apparatus, comprising: The data acquisition module is used for acquiring the electrical working condition parameters and the thermal working condition parameters of the power module; the junction temperature prediction module is used for calling a pre-trained cyclic neural network model, inputting the electrical working condition parameters and the thermal working condition parameters into a pre-trained junction temperature prediction model, and outputting junction temperature increment predicted values of a plurality of sampling points in an operation period; The time dividing module is used for judging the temperature change trend of each sampling point according to the junction temperature increment predicted value of each sampling point, dividing the operation period into a plurality of continuous time periods, and enabling the temperature change trend of all the sampling points in each time period to be consistent; and the service life calculation module is used for calculating the service life of the power module in each time period according to the junction temperature increment, and accumulating the service lives of the time periods to obtain the service life of the power module in the whole operation period.

Description

Power module service life estimation method and device Technical Field The invention relates to the technical field of power modules, in particular to a service life estimation method and device of a power module. Background The power module is an integrated power device used in a motor controller in a new energy automobile and mainly used for converting high-voltage direct current into three-phase alternating current. The semiconductor wafer, which is an important functional element in the power module, generates a large amount of heat in the process of alternating current-direct current conversion, and the temperature of the wafer is correspondingly increased. Related researches show that the failure risk of the power module is improved by about 10 percent when the temperature of the wafer is increased by 10 ℃, and the service life of the power module is correspondingly shortened. By monitoring the junction temperature of the power module, a proper life estimation method of the power module can be adopted to effectively obtain the life value of the power module, and the driving safety is improved. Therefore, the power module junction temperature estimation method has very important practical significance. The existing power module life prediction method mainly comprises the steps of estimating the junction temperature of a power module under the current working condition according to a thermal resistance model through the actual running working condition, and obtaining the corresponding life of the power module according to the junction temperature of the current power module and combining a life curve of the power module. Only the junction temperature of the power module corresponding to the working point at the operation time is considered, and the influence of the front working condition on the temperature of the power module is not considered, so that the obtained junction temperature of the power module is insufficient in accuracy, and the service life prediction result is deviated. Disclosure of Invention The invention aims to provide a power module service life estimation method and device, which can improve the accuracy of a power module service life estimation result. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: in a first aspect, the present invention discloses a power module lifetime estimation method, which includes: acquiring an electrical working condition parameter and a thermal working condition parameter of the power module in an operation period; The junction temperature prediction model is obtained by training based on a historical power module sample data set, and each sample in the sample data set comprises the electrical working condition parameter and the thermal working condition parameter of the power module and the junction temperature increment simulation value or the actual measurement value corresponding to the electrical working condition parameter and the thermal working condition parameter; Dividing the operation period into a plurality of continuous time periods, wherein the temperature change trend of all sampling points in each time period is consistent, and the change trend between adjacent time periods is different; And calculating the service life of the power module in each time period according to the junction temperature increment, and accumulating the service lives of the time periods to obtain the service life of the power module in the whole operation period. Further, the calculation formula of the life of the power module in each time period is: in the formula, Is the firstThe power module lifetime in a single time period,Is the firstThe sum of junction temperature increment of all sampling points in each time period,AndIs constant. Further, the electrical operating parameters include operating current, operating voltage, duty cycle, switching frequency, and power factor, and the thermal operating parameters include fluid temperature, fluid flow, and ambient temperature. Further, the junction temperature prediction model is a cyclic neural network model. Further, the recurrent neural network model is pre-trained by: Acquiring a historical power module sample data set generated by three-dimensional thermo-electric coupling simulation or rack actual measurement, wherein the historical power module sample data set comprises time sequence data under a plurality of working condition combinations, and the samples of each time step comprise an electrical working condition parameter, a thermal working condition parameter and a junction temperature increment corresponding to the electrical working condition parameter and the thermal working condition parameter; Dividing a historical power module sample data set into a training set and a verification set according to a preset proportion; And constructing a circulating neural network model, training the circulating neural network model by adopting a training set with