基于局部边缘判别投影的发动机故障诊断方法  

Fault Diagnosis Method of Engines Based on Locality Margin Discriminant Projection

作　　者：梁华 吕丽平[2] 王成勇 LIANG Hua;LYU Liping;WANG Chengyong(School New Energy Vehicle,Nanning Vocational College of Technology,Nanning 530008,China;School of Information Engineering,Zhengzhou Shengda College of Economics Trade and Management,Zhengzhou 451191,China;Fifth Military Represent Office of Chongqing District,Chongqing 404000,China)

机构地区：[1]南宁职业技术学院新能源汽车学院,南宁530008 [2]郑州升达经贸管理学院信息工程学院,郑州451191 [3]驻重庆地区第五军代表室,重庆404000

出　　处：《噪声与振动控制》2023年第3期90-94,109,共6页Noise and Vibration Control

基　　金：国家自然科学基金资助项目(61272527);河南省科技厅自然科学资助项目(152102210261);广西职业教育教学改革研究重点项目(GXGZJG2022A054)。

摘　　要：在线性判别分析(Linear Discriminant Analysis,LDA)的基础上,局部边缘判别投影(Locality Margin Discriminant Projection,LMDP)重新定义类间散布矩阵和类内散布矩阵,使得数据样本中异类样本在低维空间中的距离更远、同类样本在低维空间中的距离更近,增强数据样本的可区分度。为更好提取发动机的故障特征,实现发动机故障有效诊断,以LMDP为核心,结合特征提取方法和模式识别方法,给出基于LMDP的发动机故障诊断流程。发动机故障诊断结果表明,LMDP可实现发动机不同故障类型的有效区分,显著提升后续的诊断精度,具有一定的优势。Based on linear discriminant analysis(LDA)and locality margin discriminant projection(LMDP),the interclass dispersion matrix and the intra-class dispersion matrix are redefined,which makes the distance between heterogeneous samples in the low-dimensional space longer and the distance between homogeneous samples in the lowdimensional space closer,and enhances the discriminability of data samples.In order to extract the fault feature of engine better and improve fault diagnosis accuracy,combined with feature extraction method and pattern recognition method,an engine fault diagnosis process of LMDP is proposed.The engine fault diagnosis results show that LMDP can effectively distinguish different fault types of the engine and significantly improve the diagnosis accuracy.

关 键 词：故障诊断 线性判别分析 局部边缘判别投影 散布矩阵 发动机 

分 类 号：TK42[动力工程及工程热物理—动力机械及工程]