基于AVMD与SMA-SVM的滚动轴承故障诊断

Rolling Bearing Fault Diagnosis Based on AVMD and SMA-SVM

现实工程中难以获得大量轴承故障样本,因此大多采用支持向量机进行分类,而传统的智能优化算法优化支持向量机,容易陷入局部最优解,寻优时间长,并且需要人为干预。本文提出了一种自适应变分模态分解(adaptive variational modaldecomposition,AVMD)与黏菌算法(slime mouldalgorithm,SMA),对支持向量机(support vectormachine,SVM)进行智能优化的故障诊断方法,用更合理的惩罚参数与核参数使构建的SMA-SVM模型对小样本数据进行快速准确分类。该方法首先利用AVMD方法对故障信号进行分解,然后计算各IMF分量的样本熵作为特征向量,最后将特征向量输入到所提出的SMA-SVM模型中进行故障识别。并将其与以往传统的优化算法,如遗传算法、粒子群算法的优化支持向量机等故障诊断方式相比较。结果表明,所提出的故障识别方法准确率高,并且缩短了寻优时间,相较于其他方法展现了其优越性,该方法可有效用于轴承的故障诊断。

In view of the fact that it is difficult to obtain a large number of bearing fault samples in reality,support vector machines were mostly used for classification in the past,while the traditional intelligent optimization algorithm for optimizing support vector machines is prone to fall into local optimal solutions,the optimization time is long,and human intervention is required.A fault diagnosis method based on adaptive variable modal decomposition(AVMD)and slime mold algorithm(SMA)to intelligently optimize support vector machine(SVM)is proposed.More reasonable penalty parameters and kernel parameters enable the SMA-SVM model to classify small sample data quickly and accurately.This method firstly decomposes the fault signal by AVMD method,then calculates the sample entropy of each IMF component as the feature vector,and finally inputs the feature vector into the proposed SMA-SVM model for fault recognition.It is compared with the traditional optimization algorithms such as genetic algorithm and particle swarm optimization support vector machine.The results show that the proposed method has high fault identification accuracy and shortens the optimization time.Compared with other methods,it shows advantages and can be effectively used in bearing fault diagnosis.