基于广义正态分布算法优化支持向量机的电机轴承故障诊断研究

Research on Motor Bearing Fault Diagnosis Based on Generalized Normal Distribution Algorithm Optimization Support Vector Machine

为了提高旋转机械设备故障诊断的准确率,提出一种基于广义正态分布优化(Generalized Normal Distribution Optimization,GNDO)的支持向量机(Support Vector Machine,SVM)算法。在SVM算法的基础上,使GNDO对其进行寻优求解,找到SVM算法的最优参数组,以此推导出GNDO-SVM算法。结果表明,在电机轴承故障诊断中,GNDO-SVM算法相较于SVM算法的分类准确率提高了3.9个百分点,有着更好的有效性和准确性。

In order to improve the accuracy of rotating machinery fault diagnosis,a Support Vector Machine(SVM)algorithm based on Generalized Normal Distribution Optimization(GNDO)is proposed.On the basis of SVM,GNDO is used to optimize the generalized normal distribution optimization.Find the optimal parameter set of SVM algorithm,and then derive the GNDO-SVM algorithm.The results show that compared with the SVM algorithm,the GNDO-SVM algorithm has a higher classification accuracy of 3.9 percentage points,and has better effectiveness and accuracy in motor bearing fault diagnosis.