改进局部保持投影算法及其在故障诊断中的应用  

Improved Locality Preserving Projections Algorithm and Its Application in Fault Diagnosis

作　　者：赵娜[1] 张忠杰 ZHAO Na;ZHANG Zhongjie(Department of Mechanical and Electronic Engineering,Jincheng Institute of Technology,Jincheng Shanxi 048026,China;Department of Mechanical and Electronic Engineering,Shanxi Institute of Technology,Yangquan Shanxi 045000,China)

机构地区：[1]晋城职业技术学院机械与电子工程系,山西晋城048026 [2]山西工程技术学院机械电子工程系,山西阳泉045000

出　　处：《机械设计与研究》2021年第5期215-218,223,共5页Machine Design And Research

基　　金：中国职业教育学会职业教育教学改革课题(1910350)。

摘　　要：针对局部保持投影(LPP)的邻域参数设置主要依靠经验这一缺点,对LPP邻域参数的设置进行了改进设计,提出了改进局部保持投影算法(ILPP)。相比于LPP,ILPP中引入了核密度估计,使得邻域参数会根据数据集的样本密度而自适应的变化,可以更好地对高维数据集的流形结构进行保持。将ILPP用于齿轮典型故障的诊断中,实验结果表明,ILPP可以对不同故障类型数据进行有效区分,实现了故障诊断精度的提高。Aiming at the problem that the setting of neighborhood parameter for locality preserving projections(LPP)is mainly based on experience,an improved design of neighborhood parameter for LPP is presented and improved locality preserving projections(ILPP)is proposed in this paper.Compared with LPP,kernel density estimation is introduced in ILPP.Thus neighborhood parameter will change adaptively according to the density of the dataset and better data structure can be obtained.The experiment results of ILPP used in gear fault diagnosis show that the different fault type datasets can be distinguished effectively,and better diagnosis accuracy can be obtained.

关 键 词：局部保持映射 邻域参数 核密度估计 故障诊断 齿轮 

分 类 号：TH132[机械工程—机械制造及自动化]