基于支持向量机和邻域粗糙集的旋转机械振动故障诊断

Vibration Fault Diagnosis of Rotating Machinery Based on SVM and Neighborhood Rough Set

针对旋转机械振动故障诊断,提出了一种基于邻域粗糙集模型的改进支持向量机的诊断方法。该方法对选取的特征向量进行约简,并将约简结果作为支持向量机的输入特征向量来诊断旋转机械的工作状态。实验结果表明:与属性约简前的支持向量机诊断模型相比,该方法能够在提高分类精度的前提下缩短系统的诊断时间,取得了良好的效果。

Model of improved SVM based on neighborhood rough set is proposed for the vibration fault diagnosis of rotating machinery. The method reduces the selecting feature vectors. And the reduction result is used for SVM input feature vectors to determine the working status of rotating machinery. The experimental result shows that： compared with SVM diagnosis model of non-reduction, this method can improve the classification accuracy as well as shorten the diagnosis time, and achieves good results.