基于全矢色散熵和极限学习机的液压泵故障诊断  

Hydraulic Pump Fault Diagnosis Based on Full Vector Dispersion Entropy and Extreme Learning Machine

作　　者：王廷栋 WANG Tingdong(College of Mechanical and Vehicle Engineering,Taiyuan University of Technology,Taiyuan 030000,China)

机构地区：[1]太原理工大学机械与运载工程学院,山西太原030024

出　　处：《机械设计与研究》2021年第6期112-115,121,共5页Machine Design And Research

摘　　要：为保证提取出的液压泵故障特征能够更好的对其故障进行表征,本文在研究色散熵(DE)的基础上,结合全矢谱理论,提出液压泵故障特征提取的新方法-全矢色散嫡(FVDE)。相比于DE,FVDE中的全矢谱计算实现了同源双通道信号的融合处理,能够更好的突出故障信息,提升故障特征提取的效果。液压泵故障诊断实例表明,FVDE对故障状态的区分度更加明显,极限学习机(ELM)的诊断精度也更高,具有一定的优势。In order to ensure that the extracted hydraulie pump fault features can be more analyzable,a new method of hydraulic pump fault feature extraction,named full vector dispersion entropy(FVDE),is proposed in this paper based on the study of dispersion entropy(DE)and the theory of full vector specrum.Compared with DE,the full vector spectrum calculation in FVDE realized the fusion processing of homologous dual-channel signals,which can belter highlight the fault information and improve the ffeet of fault feature extraction.The experiment results of hydraulie pump show that FVDE can better distinguish the fault state,and extreme learning machine(ELM)also ean provide relative high dignosis accuracy with potential benefits.

关 键 词：色散熵 全矢色散熵 故障诊断 液压泵 

分 类 号：TH17[机械工程—机械制造及自动化]