基于DE-LSSVM的滚动轴承故障诊断

Rolling Bearing Fault Diagnosis Based On DE-LSSVM

摘要:

针对滚动轴承故障诊断难以获得大量样本的问题以及LS-SVM 模型参数选择方法易陷入局部最优的缺点，提出了一种集合经验模态分解能量熵和差分进化算法（DE）优化最小二乘支持向量机相结合的轴承故障诊断方法。首先原始振动信号采用EEMD分解得到一组固有模态函数（IMF），从有效本征模态函数IMF分量中提取的能量特征作为输入建立支持向量机，通过计算不同振动信号的能量熵值大小来判断轴承的故障损伤程度。为了提高模型的诊断精度，采用差分进化算法对LS-SVM的结构参数进行优化，并与LS-SVM和PSO-LSSVM模型相比较。结果表明，DE-LSSVM 模型的故障分类准确性得到了提高，可以有效应用于滚动轴承故障诊断中。

Abstract:

Aiming at the problem of the samples is difficult to get in rolling bearing fault diagnosis, combining the drawbacks of the LS-SVM model is easy to fall into local optimum, a new bearing fault diagnosis method based on differential evolution and least squares support vector machine is proposed. Firstly, extracting the different fault condition of the bearing vibration signal,Sthen the method of EEM D is used to analyze the bearing vibration signal, get the IMF component, and calculate the correlation dimension and Energy entropy of IMF. These data are used to train the DE-LS SVM model. Using differential evolution algorithm to optimize the structure parameters of LSSVM, improve the diagnostic accuracy of the model. The results show that, compared with non-optimized SVM and PSO-LSSVM, the DE-LSSVM model’s fault classification accuracy and efficiency of diagnosis have been improved, it can be applied to the fault diagnosis of rolling bearings.