基于聚类分类算法的IGBT健康状态分类研究  

Research on IGBT state classification based on cluster and classification algorithm

作　　者：王志远 孙鹏菊[1] 王海波[1] 杨舒萌 WANG Zhi-yuan;SUN Peng-ju;WANG Hai-bo;YANG Shu-meng(State Key Laboratory of Power Transmission Equipment and System Security and New Technology(Chongqing University),Chongqing 400044,China)

机构地区：[1]输配电装备及系统安全与新技术国家重点实验室(重庆大学),重庆400044

出　　处：《电工电能新技术》2021年第11期1-8,共8页Advanced Technology of Electrical Engineering and Energy

基　　金：国家自然科学基金重点项目(51137006)、国家自然科学基金项目(51577020)。

摘　　要：绝缘栅双极型晶体管(IGBT)是功率变流器中最常用,也是故障率最高的元器件,因此其健康状态分类评估极为重要。文中基于聚类分类算法建立了IGBT模块的健康状态分类评估模型。首先根据聚类分类原理,简述了IGBT模块状态分类模型建立的步骤。然后以饱和压降和短路电流作为老化特征量,分析了老化过程特征量的变化趋势。最后搭建了IGBT模块健康状态分类评估模型,基于已有的数据集对模型进行了检测,模型分析结果与测试结果基本一致,验证了健康状态分类评估模型的准确性。The insulated gate bipolar transistor(IGBT)is the most commonly used component in power converter,whitch has the highest failure rate.Therefore,the states assessment of module health is very important.In this paper,the state classification model of IGBT is established based on cluster and classification algorithm.Firstly,according to the principle of cluster and classification algorithm,the steps of establishing IGBT module state classification model are described.Then,the change trend of aging characteristics is analyzed by taking saturation voltage and short-circuit current as aging characteristics.Finally,the IGBT module state classification model is built,and the model is tested based on the existing data set.The results of model are basically consistent with the test results,which verifies the accuracy of the state classification model.

关 键 词：绝缘栅双极型晶体管 K-MEANS聚类 SVM分类 状态分类评估模型 

分 类 号：TN323[电子电信—物理电子学]