局部均值分解和形态谱的液压泵故障诊断方法  

Research on Hydraulic Pump Fault Diagnosis Method Based on Local Mean Decomposition and Pattern Spectrum

作　　者：孙兆丹 郑直 张何 姜万录[3,4] SUN Zhaodan;ZHENG Zhi;ZHANG He;JIANG Wanlu(Department of Mechanical and Electrical Engineering,Langfang Yanjing Vocational Technical College,Langfang 065000,Hebei,China;College of Mechanical Engineering,North China University of Science and Technology,Tangshan 063210,Hebei,China;Hebei Provincial Key Laboratory of Heavy Machinery Fluid Power Transmission and Control,Yanshan University,Qinhuangdao 066004,Hebei,China;Key Laboratory of Advanced Forging&Stamping Technology and Science,Ministry of Education of China,Yanshan University,Qinhuangdao 066004,Hebei,China)

机构地区：[1]廊坊燕京职业技术学院机电工程系,河北廊坊065200 [2]华北理工大学机械工程学院,河北唐山063210 [3]燕山大学河北省重型机械流体动力传输与控制重点实验室,河北秦皇岛066004 [4]燕山大学先进锻压成形技术与科学教育部重点实验室,河北秦皇岛066004

出　　处：《噪声与振动控制》2020年第2期96-101,共6页Noise and Vibration Control

基　　金：国家自然科学基金资助项目(51875498);河北省省属高等学校基本科研业务费资助项目(JQN2019004)河北省自然科学基金重点资助项目(E2018203339);华北理工大学博士科研启动基金资助项目(0088-28412499);。

摘　　要：针对液压泵故障诊断问题,提出一种基于局部均值分解(LMD)、形态谱和核模糊C均值聚类相结合的方法。首先,用LMD分解液压泵振动信号,得到具有物理意义的若干个模态分量(PFs);其次,选取含有特征信息丰富的3个PFs为数据源,采用基于峰度值、能量和均方差的评价方法,从这3个PFs中提取出各个尺度上的形态谱的3个平均值,将其组成一个向量;最后,采用核模糊C均值聚类方法(KFCM)对不同工况下所有样本进行聚类分析,对液压泵故障进行诊断。此外,将信号采用经验模态方法(EMD)分解、模糊C均值聚类方法(FCM)分析,结果表明LMD和KFCM分别优于EMD和FCM;该方法诊断精度高,是液压泵故障诊断的有效方法。Aiming at the problem of hydraulic pump fault diagnosis,a fusion method based on local mean decomposition(LMD),pattern spectrum and kernel fuzzy C-means clustering(KFCM)is proposed.Firstly,the vibration signal of the hydraulic pump is decomposed by LMD to get several product functions(PFs)which have some physical meanings.Then,the three PFs which have rich feature information are selected as the data source.Using the evaluation method based on kurtosis,power and mean square difference,the three mean values of the pattern spectrums in some scales are extracted from the three PFs,and then the three mean values are combined into a sample.Finally,all the samples of the different working conditions are clustered by the KFCM clustering to diagnose the hydraulic pump faults.In addition,the signals are decomposed by EMD and clustered by FCM.It is concluded that LMD and KFCM are respectively superior to EMD and FCM;the proposed method can efficiently diagnosis the faults with high accuracy and is an effective method for fault diagnosis of hydraulic pumps.

关 键 词：故障诊断 局部均值分解 形态谱 核模糊C均值聚类 液压泵 

分 类 号：TH137[机械工程—机械制造及自动化] TP277[自动化与计算机技术—检测技术与自动化装置]