飞机变压整流器健康管理算法研究  

Health Management Algorithm of Aircraft Transformer Rectifier Unit

作　　者：李一卓 高朝晖[1] 郗展 王爽 李小宁 梁源颀 LI Yi-zhuo;GAO Zhao-hui;XI Zhan;WANG Shuang;LI Xiao-ning;LIANG Yuan-qi(School of Automation,Northwestern Polytechnical University,Xi'an 710129,China;AVIC Luoyang Institute of Electro-Optical Equipment,Luoyang 471000,China)

机构地区：[1]西北工业大学自动化学院,陕西西安710129 [2]航空工业洛阳电光设备研究所,河南洛阳471000

出　　处：《测控技术》2022年第7期29-36,56,共9页Measurement & Control Technology

基　　金：航空科学基金(20200046053001)。

摘　　要：变压整流器是飞机电源系统中的重要设备,可以使飞机供电系统不再配备直流发电机,变压整流器在飞机电源系统中应用非常广泛,将健康管理技术应用于飞机变压整流器对其故障进行提前预警,具有重要现实意义。在对12脉波变压整流器的原理进行分析后,得出变压整流器的绕组绝缘老化和电容老化是导致其系统老化的主要因素,通过变压整流器的模型,借助仿真数据对比分析,确定了与电容和绝缘绕组相关的系统外部老化特征参数;并通过高斯混合模型、马氏距离对数据进行分类,再通过训练BP神经网络实现对变压整流器健康状态的评估。The transformer rectifier unit(TRU) is an important equipment in aircraft power supply system, which can make the aircraft power supply system no longer equipped with DC generator.TRU is widely used in aircraft power supply system.It is of great significance to apply health management technology to aircraft TRU for early warning of its faults.After analyzing the principle of 12 pulse TRU,it is concluded that the aging of winding insulation and capacitance of TRU are the main factors leading to the system aging.Through the model of transformer rectifier and the comparative analysis of simulation data, the external aging characteristic parameters of the system related to capacitance and insulation winding are determined.The data are classified by Gaussian mixture model and Mahalanobis distance, and then the health state of transformer rectifier is evaluated by training BP neural network.

关 键 词：变压整流器 健康状态评估 高斯混合模型 BP神经网络 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]