孪生支持向量机在滚动轴承振动故障诊断中的应用

Application of Defect Diagnosis for Rolling Bearing Based on Twin Support Vector Machine

通过对滚动轴承振动信号进行定量分析,从振动故障信号中提取与故障诊断方法有关的故障特征,在传统支持向量机的基础上,研究孪生支持向量机的建模方法,建立基于孪生支持向量机的滚动轴承振动故障诊断模型,并结合粒子群优化算法对故障诊断模型的关键参数进行寻优,从而得到最佳的故障诊断模型。仿真结果表明,将孪生支持向量机建模方法应用于滚动轴承振动故障诊断中,能够取得较好的诊断效果和诊断效率,结合粒子群优化算法进一步提高了故障诊断模型的分类准确率,为滚动轴承的振动故障诊断提供了可行有效的思路。

This method begins with the quantitative analysis of the defect signals received during roll bearing vibrations. Then the defect features are abstracted out of the defect signals through classical defect diagnosis method. And a twin support vector machine is utilized to process these data. In order to obtain a more accurate model, swarm optimization algorithm is applied to optimize the key parame- ters of the model. The result of simulation demonstrates that it is feasible to apply twin support vector machine in defect diagnosis for roll bearings, and the performance and efficiency are superior, which provides a new way for defect diagnosis for roll bearing.