摘要
针对目前IGBT功率模块内部键合线故障需要通过复杂辅助电路才能实现识别、监测的难题,本文提出一种根据IGBT模块实时开关波形对键合线故障甚至是并联IGBT芯片故障进行评估分析的新方法。从IGBT模块结构特性入手,分析得出键合线及并联IGBT芯片的故障特性将在模块栅极关断电压的变化上得到体现,然后对未塑封的IGBT模块进行了人为的键合线挑断试验得到了预期的分析处理数据。考虑到IGBT模块从投运至失效过程中的数据信息量极为巨大,需要对大量数据进行有效压缩并挖掘出其中的有用信息。本文采用小波奇异熵理论对栅极电压信号进行了故障特征提取,并比较了不同故障类型对特征量变化的影响。研究成果为实现IGBT模块的健康状态评估提供了重要依据。
To deal with the problem that the implementation of recognition and monitoring oI bond wire failure inside IGBT module needs complex auxiliary circuit. A new assessment method of bond wire and paralleled IGBT chip failure by use of the real-time switching waveforms of IGBT module is presented in this paper. The work starts from the structural characteristics of IGBT module, and draws the conclusion that both bond wire failure and paralleled IGBT chip failure would change the gate turn-off voltage waveform of IGBT module. The expected data for analyzing and processing are derived through the test of cutting the bond wires of an open sample manually on purpose. Taking into account the data quantity of IGBT module from putting into operation to failure is very large, and it is need to compress the data and excavate the useful information contained in it. Fault feature of gate voltage signal is extracted by use of the wavelet singular entropy theory, and the fault characters of different fault types are compared. The research results provide an important basis for realizing the health assessment of IGBT module.
出处
《电工技术学报》
EI
CSCD
北大核心
2013年第6期165-171,共7页
Transactions of China Electrotechnical Society
基金
国家自然科学基金(51077137
51137006)
科技部国际科技合作(2010DFA72250)
输配电装备及系流安全与新技术国家重点实验室(2007DA10512711101)资助项目
