基于数据驱动的功率IGBT模块状态监测新方法研究

Because of the direct detection constraints for power module internal faults, it is difficult to monitor and evaluate its conditions as well as guarantee the reliability. Based on the background, this project researches a novel condition monitoring strategy for IGBT module based on data-driven. The mainly innovative research contents include: 1) a precise construction of the semiconductor physics model for IGBT chip and the method of parameters acquisition are proposed considering the different injection of carriers (electron and hole) in IGBT drift region; 2) based on the model, a research about the internal connection between the module early failure and externally electrical parameters is investigated so as to identify the characteristic parameters that can indicate an internal fault. 3) Designing algorithm and proposing the novel condition monitoring strategy for IGBT module based on data-driven, forming a theoretical system and designing scheme. Finally, the simulation and experimental verification is conducted. Research results will have a significant impact on the improvement of condition monitoring theory system and reliability of power converter.

针对功率变换装置中功率模块内部故障不易直接测量，状态难于监测与评估，可靠性难于保证的难题，研究基于数据驱动的功率IGBT模块状态监测新方法。创新性研究内容包括：结合IGBT开关过程中少子注入条件变化，建立精确的芯片物理模型并给出参数提取方法；在模型基础上，研究IGBT模块早期失效与外部可测参数之间的内在关系，揭示其变化规律，确定表征模块内部失效的特征参数；设计算法，提出基于数据驱动的状态监测新方法，形成理论体系及设计方案；最后，进行仿真和实验验证。研究成果对于完善状态监测理论体系，提高功率变换装置可靠性具有重要价值和研究意义。

电力电子装置作为电能变换系统中最为关键的部件，对系统整体性能起着至关重要的影响。如果系统中电力电子装置先兆失效症状出现，且未能快速地有效地完成系统失效过程的预测并进行采取实时的有效的维护措施，则可能造成严重的后果和巨额的经济损失。而电力电子器件作为电力电子装置的核心常常是其失效的主因。目前，电力电子系统大多采用IGBT（Insulated Gate Bipolar Transistor）模块作为功率器件。故对于应用电力电子装置的不同行业及领域的系统来讲，研究IGBT器件的健康状态对提高电力电子系统的整体的可靠性具有重要的意义。.本项目设计了IGBT模块多种电参数在线提取方法，实现了IGBT器件老化过程中电参数变化的实时监控，建立了IGBT器件参数模型；搭建IGBT器件功率循环及参数监测实验平台，获取了器件失效与外部特征参数变化的关系，建立了特征数据库；针对在线测量的导通电阻Vce,on包含测量噪声的问题，提出了采用扩展卡尔曼滤波的方法对在线测量的导通电阻Vce,on进行滤波，并以滤波后的IGBT器件导通电阻Vce,on为特征参数，提出了状态监测新方法，基于此方法的器件寿命预测的最大误差小于20%，有效提高了电力电子系统的可靠性。