灰狼算法优化VMD与SVM的滚动轴承故障诊断

Fault Diagnosis of Rolling Bearing Based on Grey Wolf Algorithm Optimizing VMD and SVM

针对VMD模态分量K和惩罚因子α选取不当导致分解度较低以及SVM超参数设置较差导致故障诊断准确率低下的问题,提出一种基于GWO-VMD和GWO-SVM的滚动轴承故障诊断的方法。首先利用GWO对VMD进行寻优得到一系列固有模态分量(IMFs),分析并提取最优模态分量子集的中心频率以及能量构建故障诊断特征集,将提取的特征集输入SVM中进行分类,采用GWO对SVM惩罚因子C和核函数g进行优化。实验结果表明,GWO有较大的寻优能量,通过VMD分解后得到高质量分量,并且GWO应用于SVM时仅需1.028 s便能达到最佳适应度值,且分类准确率可达99.78%。

In order to solve the problems of low resolution caused by improper selection of VMD modal component K and penalty factorαand low accuracy of fault diagnosis caused by poor setting of SVM super parameters,a rolling bearing fault diagnosis method based on GWO-VMD and GWO-SVM is proposed.Firstly,GWO is used to optimize VMD to obtain a series of intrinsic mode components(IMFs).The center frequency and energy of the optimal mode component subset are analyzed and extracted to construct fault diagnosis feature set.The extracted feature set is input into SVM for classification.GWO is used to optimize SVM penalty factor C and kernel function g.The experimental results show that GWO has a large optimization energy,and it can get high-quality components through VMD decomposition.When GWO is applied to SVM,it only takes 1.028 s to achieve the best fitness value,and the classification accuracy can reach 99.78%.