基于粒子群算法的旋转机械故障诊断

Rotating Machinery Fault Diagnosis Based on Particle Swarm Optimization

为有效解决旋转机械故障诊断需人工干预以及诊断结果不准确等问题,提出基于粒子群算法的旋转机械故障诊断方法。首先,利用振动传感器采集故障信号,通过小波阈值消噪算法完成故障信号去噪处理,降噪后的振动信号进行SPWVD时频图图像纹理特征参量提取,将提取的各种参量汇总,组成故障特征集。然后,利用SVM及其学习法搭建旋转机械故障诊断分类器,把故障特征集录入到分类器中,通过粒子群算法对分类器中的参数进行优化处理。最后,得到输出的智能识别结果,完成旋转机械故障的诊断。实验结果表明,所提方法得到了准确率更高的旋转机械故障诊断结果。

In order to effectively solve the problems of manual intervention and inaccurate diagnosis resulting in rotating machinery fault diagnosis,a rotating machinery fault diagnosis method based on particle swarm optimization algorithm is proposed.Firstly,the vibration sensor is used to collect the fault signal,and the wavelet threshold denoising algorithm is used to denoise the fault signal.After noise reduction,the texture feature parameters of SPWVD time-frequency image are extracted from the vibration signal,and the extracted parameters are summarized to form a fault feature set.Secondly,by using SVM and its learning method,a rotating machine fault diagnosis classifier is built,the fault feature set is input into the classifier,and the parameters in the classifier are optimized by particle swarm optimization algorithm.Finally,the output intelligent recognition results are obtained to complete the fault diagnosis of rotating machinery.The experimental results show that the proposed method has higher accuracy in rotating machinery fault diagnosis.