基于MEA-BP算法的IGBT结温预测模型  

Prediction Model of IGBT Junction Temperature Based on MEA-BP Algorithm

作　　者：孟昭亮[1] 吕亚茹 高勇[1] 杨媛[2] 吴磊 Meng Zhaoliang;LüYaru;Gao Yong;Yang Yuan;Wu Lei(School of Electronics Information,Xi'an Polytechnic University,Xi'an 710048,China;School of Automation and Information Engineering,Xi'an University of Technology,Xi'an 710048,China;CRRC Xi'an Yongdian Electronic Co.,Ltd.,Xi'an 710016,China)

机构地区：[1]西安工程大学电子信息学院,西安710048 [2]西安理工大学自动化与信息工程学院,西安710048 [3]西安中车永电电气有限公司,西安710016

出　　处：《半导体技术》2020年第11期856-862,共7页Semiconductor Technology

基　　金：国家自然科学基金资助项目(51477138);高校人才服务企业工程项目(2017080CG/RC043(XALG009))。

摘　　要：针对IGBT芯片被封装在模块内部,芯片结温无法直接测量的问题,提出了基于思维进化算法(MEA)优化的反向传播(BP)(MEA-BP)神经网络算法的IGBT结温预测算法模型。首先,利用温敏电参数(TSEP)法搭建IGBT模块饱和压降实验平台;然后,从实验数据中提取338组饱和压降与集电极电流数据作为TSEP,表征其与IGBT模块结温的关系;最后,利用MEA-BP神经网络算法将提取出的电气参数建立结温预测模型,对结温进行预测。实验结果表明,MEA-BP神经网络算法的结温预测值平均绝对百分比误差在集电极电流小于临界电流时为0.114,在大于临界电流时为0.062,比遗传算法(GA)优化的BP(GA-BP)神经网络算法以及经典BP神经网络算法能更准确预测IGBT结温。To solve the problem that the IGBT chip is encapsulated in the module,so that the chip junction temperature cannot be measured directly,an IGBT junction temperature prediction algorithm model based on the mind evolutionary algorithm(MEA)optimized back-propagation(BP)(MEA-BP)neural network algorithm was proposed.Firstly,the saturation voltage drop experiment platform of the IGBT module was built by using the temperature sensitive electrical parameters(TSEPs)method.Then,338 groups of saturation voltage drop and collector current data were extracted from the experimental data as TSEPs to characterize their relationship with the junction temperature of the IGBT module.Finally,the junction temperature prediction model was established with the extracted electrical parameters by using the MEA-BP neural network algorithm to predict the junction temperature.The test results show that the average absolute percent error of the junction temperature of the MEA-BP neural network algorithm is 0.114 when the collector current is less than the critical current and 0.062 when it is greater than the critical current.Compared with the genetic algorithm(GA)optimized BP(GA-BP)neural network algorithm and classical BP neural network algorithm,the MEA-BP neural network algorithm can predict the IGBT junction temperature more accurately.

关 键 词：IGBT 反向传播(BP)神经网络 温敏电参数(TESP) 结温预测模型 MEA-BP算法 

分 类 号：TN322.8[电子电信—物理电子学]