基于EEMD和PSO-SVM的滚动轴承故障诊断

Fault Diagnosis of Rolling Bearing Based on EEMD and PSO-SVM

为实现滚动轴承故障的精确诊断,提出一种基于集成经验模态分解与粒子群算法优化的支持向量机的故障诊断方法。利用EEMD方法分解振动信号,依据经验选取合适的内禀模态函数进行能量值及包络谱特征幅值比等故障特征参量的计算,构建滚动轴承故障特征向量,然后基于少量不同故障部位及故障程度的样本,利用粒子群算法对支持向量机进行参数优化,进而训练样本并建立故障模型,最后对测试样本进行故障诊断,观察该方法的诊断效果。实验表明,该方法可对多种不同故障状态进行诊断,且分类精度高,证明了振动分析与智能算法结合的方法可有效实现滚动轴承的故障诊断。

In order to realize the fault diagnosis for the rolling bearing with accuracy,based on ensemble empirical mode decomposition and support vector machine optimized by particle swarm optimization,a fault diagnosis method is put forward. EEMD method is used to decompose the vibration signal,while a suitable intrinsic mode function is selected on the basis of experience to calculate fault characteristic parameter,including energy value and feature amplitude ratio of envelope spectrum,to build rolling bearing fault feature vector. Then the particle swarm optimization is used to optimize parameters of support vector machine. Based on a small amount of samples with different fault location and fault degree,samples are trained to build fault model. Finally,test samples are conducted on fault diagnosis and the diagnostic effect of this method was observed. The results show that this method can be used to diagnose a variety of fault conditions,and get better classification accuracy. It is proved that the method which combines vibration analysis and intelligent algorithm can effectively realize the fault diagnosis of rolling bearings.