基于CEEMDAN-SVM的液压泵故障诊断方法研究  

Fault Diagnosis of Hydraulic Pump Based on CEEMDAN-SVM

作　　者：李锋 林阳阳 赵辉 晁苏全 王浩 

机构地区：[1]第二炮兵工程大学,陕西西安710025

出　　处：《液压与气动》2016年第1期125-129,共5页Chinese Hydraulics & Pneumatics

摘　　要：由于液压泵故障振动信号微弱和不平稳的特性,造成特征向量提取和故障诊断困难。针对这些问题,提出一种CEEMDAN与支持向量机结合的故障诊断方法。将传感器测得的液压泵故障振动信号进行CEEMDAN分解,得到多个固有模态函数(IMF),并计算其样本熵作为支持向量机的输入特征向量,以诊断液压泵的故障类型。液压泵故障诊断实验结果证明了该方法的有效性和优越性。The characteristics that hydraulic pump fault vibration signal is weak and unstable, bring difficulty to vector extraction and fault diagnosis. To solve these problems, a method is put forward for fault diagnosis based on the combination of CEEMDAN and support vector machine. by CEEMDAN to obtain the intrinsic mode functions （IMF） as the input vectors of support vector diagnosis experiment demonstrate the Decompose the fault vibration signal of hydraulic pump , and calculate the sample entropy of them, which work machine in diagnosis of hydraulic pump. The results of hydraulic pump fault effectiveness and superiority of the method.

关 键 词：CEEMDAN 样本熵 支持向量机 

分 类 号：TH137[机械工程—机械制造及自动化] TH322