QPSO-WT和QPSO-SVM在滚动轴承故障诊断中的应用

为了解决小波降噪软阈值选择非最优以及SVM算法中惩罚参数、核函数参数的设置问题,将小波变换、支持向量机分别与量子行为粒子群优化算法QPSO(quantum-behaved particle swarm optimization,)相结合,利用QPSO优化小波阈值以及优化SVM输入参数,进行全局寻优,并将之应用到滚动轴承故障识别中。实验中,QPSO-WT滤波后信号具有更高的信噪比和更低的MSE,QPS0-SVM对10种不同状态的轴承进行故障诊断,对于多分类的情况该方法的识别精确度达到了87.67%,与SVM和RBF神经网络对比,从而进一步证明了该方法的有效性,说明该方法能够满足实际工况下的故障诊断要求。

The Applications of QPSO-WT and QPSO-SVM in Fault Diagnosis of Rolling Bearing

For the problems of the wavelet threshold is not global optimal solution and punishment parameter and kernel function parameter setting problem in SVM algorithm, improved filtering algorithm and recognition algorithm based on wavelet threshold and SVM and quantum-behaved particle swarm optimization(QPSO)are proposed to improve above questions, and then applying this method to extract features in rolling bearing fault diagnosis. In experiments, QPSO-WT is better than traditional wavelet threshold in filtering, ten bearings with different conditions were diagnosed by QPSO-SVM, getting the result that Accuracy is as high as 87.67%, and Comparing with SVM and RBF neural network further confirmed the effectivity of this method.