基于特征优化与改进KNN的航空发动机故障诊断  

An Aero engine Gas Path Fault Diagnosis Method Based on Feature Optimization and Improved KNN

作　　者：周寒 莫李平 刘渊 王奕首[1] 卿新林[1] ZHOU Han;MO Li-ping;LIU Yuan;WANG Yi-shou;QING Xin-lin(Xiamen University,Xiamen 361000,China;Hunan Aviation Powerplant Research Institute,AECC,Zhuzhou 412000,China)

机构地区：[1]厦门大学,福建厦门361000 [2]中国航发湖南动力机械研究所,湖南株洲412000

出　　处：《航空计算技术》2023年第2期45-49,共5页Aeronautical Computing Technique

基　　金：国家自然科学基金项目资助(51975494)。

摘　　要：为有效解决航空发动机气路故障诊断难题,建立了基于特征优化与改进KNN的航空发动机气路故障诊断模型。利用特征优化算法对发动机故障特征进行处理,包括特征增维与近邻成分分析算法;将特征优化后的特征输入改进KNN算法,建立基于特征优化与改进KNN算法的故障诊断模型;为验证所建立故障诊断模型的准确性,在四台CFM56-7FB发动机数据上进行实验验证,结果表明:基于特征优化与改进KNN算法的故障诊断模型的准确率可达98%以上,能够达到智能诊断的目的。In order to solve the problem of aero engine gas path fault diagnosis effectively,a model based on feature optimization and improved KNN algorithm is proposed.Firstly,feature optimization algorithm was used to process engine fault features,including feature expand and neighbourhood components analysis.Secondly,the optimized features were input into improved KNN algorithm to establish a fault diagnosis model based on feature optimization and improved KNN algorithm.In order to verify the accuracy of the established fault diagnosis model,the established fault diagnosis model was verified by experiments on the data of four CFM567FB engines.The results show that the accuracy of the fault diagnosis model based on feature optimization and improved KNN algorithm can reach more than 98%,which can achieve the purpose of intelligent diagnosis.

关 键 词：航空发动机 故障诊断 特征优化算法 改进KNN算法 

分 类 号：V231[航空宇航科学与技术—航空宇航推进理论与工程]