基于特征参数监测的电力电子系统主电路故障统一预测方法研究

The prognostics and health management(PHM)of airborne equipment plays an important role in ensuring the security of flight and improving the ratio of combat readiness. The widely use of electronics equipment in aircraft is now making the PHM technology for power electronics devices become more important. It is the main circuit devices that are proved to have high failture rate in power electronics equiment.Thus, research on fault prognostics of the main circuit is conducted in this project.The failture modes of the main circuit devices as well as their characteristic parameters recession law are firstly found out. The fault characteristic of the whole system and the way it is affected by the main circuit devices structural and parametric faults are further obtained by using time and frequecy domain analysis to the mathematical models of buck and boost circuits that contain parasitc parameters as an example. A simple and reliable method to extract the main circuit fault characteristics is also proposed in this project. Unified fault prognostics and diagnostics method of main circuit in power electronics system is then obtained. The characteristic of main circuit devices' structural faults and parametric faults can be exreacted and predicted using minimum testing points.The research results obtained in this project lay the foundation on realizing PHM of the power electronics equipment.

故障预测及健康管理（PHM）对于保障飞行器安全可靠和提高战备完好率具有重要作用。随着电力电子装置在飞机中的应用越来越广泛，急需研究电力电子装置的PHM技术。故障预测是PHM的重要基础，本项目针对电力电子装置中故障率较高的主电路器件的故障预测方法展开研究。在研究主电路器件故障模式及特征参数衰退规律基础上，以BUCK和BOOST电路为对象，利用数值仿真方法建立含寄生参数的系统数学模型，从时域和频域全面探寻主电路器件结构性故障和参数性故障对系统特性的影响规律及故障特征；研究并提出一种简单可靠的主电路故障特征及统一提取方法；以此为基础，研究常用电力电子主电路的统一故障预测及诊断方法。研究目标是获得一种通用且易实现的电力电子主电路器件故障特征统一提取方法，利用统一的最少检测点和硬件，实现主电路器件结构性故障和参数性故障特征的统一提取和预测，为实现电力电子装置的机内故障预测和健康管理奠定基础。

故障预测及健康管理（PHM）对于保障飞行器安全可靠和提高战备完好率具有重要作用。随着电力电子装置在飞机中的应用越来越广泛，急需研究电力电子装置的PHM技术。故障特征提取是PHM的重要基础，本项目针对电力电子装置中故障率较高的主电路器件的故障特征提取方法展开研究。在研究主电路器件故障模式及特征参数衰退规律基础上，分别从结构性故障和参数性故障的角度展开研究。针对开关管结构性故障，以三相逆变和三相整流电路为研究对象，提出了两种开路故障特征提取方法以及四种故障分类方法，对于实现电力电子装置的机内故障在线诊断以及在线定位具有重要意义。针对开关管参数性故障，以BUCK和BOOST电路为研究对象，分别提出了以导通电阻、阈值电压、振铃频率、饱和电阻为故障特征的故障特征提取方法，对于开关管的故障预测和健康管理具有重要意义。针对滤波电容参数性故障，以BUCK和BOOST电路为研究对象，分别提出了以损耗角与等效串联电阻为故障特征的故障特征提取方法，对于电容的故障预测和健康管理具有重要意义。针对滤波电容以及滤波电感的参数性故障，以BUCK和BOOST电路为研究对象，基于电路混杂模型，分别提出了参数辨识的优化方法，有效提高了电容与电感的参数辨识精度；针对开关管、滤波电容、滤波电感参数性故障，以BUCK电路为研究对象，提出了一种故障特征统一提取方法。根据提取的特征参数，对预测算法进行了初步的研究。为实现电力电子装置整体的故障预测和健康管理奠定基础。