基于LMD和主分量分析的齿轮损伤识别方法  

Fault Detection Method for Gear Condition Based on LMD Method and Principle Component Analysis

作　　者：杨斌[1] 程军圣[1] 

机构地区：[1]湖南大学汽车车身先进设计制造国家重点实验室,长沙410082

出　　处：《振动．测试与诊断》2013年第5期809-813,912,共5页Journal of Vibration,Measurement & Diagnosis

基　　金：国家自然科学基金资助项目(51075131;51175158);湖南省自然科学基金资助项目(11jj2026);中央高校基本科研业务费专项基金资助项目;湖南大学汽车车身先进设计制造国家重点实验室自主课题资助项目(61075002)

摘　　要：针对齿轮故障振动信号的非平稳特征,提出了基于局部均值分解(local mean decomposition,简称LMD)和主分量分析的齿轮损伤识别方法。首先,对齿轮振动信号进行局部均值分解,将其分解成为若干个PF分量(product function,简称PF);然后,选取包含主要损伤信息的PF分量。从PF分量中提取能量和时域统计量等特征参数,组合成初始特征参数向量矩阵,并进一步对初始特征参数向量矩阵进行主分量分析,得到齿轮振动信号的主特征分量,建立距离判别函数,从而对齿轮工作状态进行识别。实验数据分析结果表明,本方法能有效地识别齿轮损伤类型。According to the non-stationary characteristics of gear fault vibration signals,a fault detection method for gear condition base on LMD method and principle component analysis is proposed.The LMD method is used to deal with the non-stationary signals,which can decompose the vibration signals into a finite number of product functions(PFs).Then the PFs containing the main fault information are selected, and the energy and the time domain statistics feature parameters can be extracted from the PFs,from which the initial feature vector matrixes can be formed.By applying the principle component analysis technique to the initial feature vector matrixes,the principle components are obtained.The euclidean distance is used for classification of gear working condition.The analysis results show that the proposed method can effectively identify the gear condition and fault pattern.

关 键 词：局部均值分解 齿轮 损伤识别 主分量分析 距离判据 

分 类 号：TH113.1[机械工程—机械设计及理论]