基于EMD和M-FSVM的泵车液压系统故障诊断  

Pump Truck Hydraulic System Fault Diagnosis Based on EMD and M-FSVM

作　　者：王盈[1] 董增寿[1] 

机构地区：[1]太原科技大学电子信息工程学院,山西太原030024

出　　处：《计算机技术与发展》2012年第6期179-181,186,共4页Computer Technology and Development

基　　金：太原市大学生创新创业项目(110148052)

摘　　要：在泵车液压系统的故障诊断技术研究中,如何精确地提取故障特征以及如何实现高精度的分类识别是研究的关键。针对这一问题,文中提出了一种基于经验模态分解(Empirical Mode Decomposition,EMD)算法与模糊多类支持向量机(Multi-class Fuzzy Support Vector Machine,M-FSVM)技术相结合的液压系统故障诊断方法。该方法首先对原始数据信号进行EMD分解,分解成若干个固有本征模态函数之和,再计算EMD能量熵作为M-FSVM的输入数据加以诊断。实验结果表明,该方法能有效地诊断泵车液压系统故障。How to extract fault feature accurately and how to achieve high accuracy of classification is the key in the field of pump truck hydraulic system fault diagnosis technology. In order to solve this problem,put forward a method that based on empirical mode decompo- sition （EMD） and multi-class fuzzy support vector machine （M-FSVM）. Firstly, decompose the original data signal into several intrinsic mode functions using EMD, calculate the energy entropy as the input of M-FSVM. The experimental results show that the method can diagnose the pump truck hydraulic system fault effectively.

关 键 词：泵车液压系统 经验模态分解 模糊多类支持向量机 相关系数 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]