期刊文献+

基于长短期记忆神经网络参数辨识的IGBT模块老化判据

IGBT module aging criterion based on long short term memory neural network parameter identification
下载PDF
导出
摘要 铝键合线脱落或者断裂是造成IGBT内部失效的主要原因,由于应用者无法观测到模块内部,从而无法正确认识模块的失效情况,安全措施不及时,导致系统发生更严重故障。研究了考虑杂散参数的多芯片封装高压大功率IGBT模块的工作特性,以某品牌三芯片并联结构IGBT模块的内部封装特性作为研究基础,分析了模块工作状态随栅极电容与寄生电感差异变化关系。重点研究了模块内部芯片间铝键合线的寄生参数,构建了基于长短期记忆神经网络的杂散参数辨识方法,提出了一种可快速精确诊断IGBT内部缺陷与失效率的方案。仿真和实验验证了该方案的正确性。 The fall off or fracture of the aluminum bonding wire is the main reason for the internal failure of the IGBT.Since the application cannot observe the inside of the module,the failure of the module cannot be correctly understood,and the lack of timely safety measures leads to more serious failure of the system.In this paper,the working characteristics of multi-chip packaged high-voltage and high-power IGBT modules considering stray parameters are studied.Based on the internal packaging characteristics of a three-chip parallel structure IGBT module of a certain brand,the relationship between the working state of the module and the difference between gate capacitance and parasitic inductance is analyzed.The parasitic parameters of the aluminum bonding wire between chips in the module are mainly studied,and a stray parameter identification method based on long-term and short-term memory neural network is constructed and a scheme for quickly and accurately diagnosing the internal defects and failure rate of the IGBT is proposed.The correctness of the scheme is verified by simulation and experiment.
作者 金声超 罗玮 暨力 黄晓宏 吴文宝 JIN Shengchao;LUO Wei;JI Li;HUANG Xiaohong;WU Wenbao(Zhixin Energy Technology Co.,Ltd.,Wuhan 430206,China)
出处 《电气应用》 2024年第11期96-103,共8页 Electrotechnical Application
基金 IGBT功能测试技术研究及装置开发(JEPCC-KYXM-2022-034)。
关键词 IGBT缺陷/失效 铝键合线 长短期记忆神经网络(LSTM) 参数辨识 IGBT defect and failure aluminium bonding wires long short term memory parameter identification
  • 相关文献

参考文献19

二级参考文献270

共引文献140

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部