基于KPCA和PSOSVM的异步电机故障诊断

发布时间：2018-05-04 04:27

本文选题：核主元分析 + 支持向量机　；参考：《振动.测试与诊断》2014年04期

【摘要】：针对异步电机故障振动信号具有较强的非线性特征,而传统的线性分析方法易造成振动信号非线性成分的丢失这一情况,提出一种核主元分析和粒子群支持向量机相结合的异步电机故障诊断方法。利用核函数实现输入空间到高维特征空间的非线性映射以及对映射数据的主元分析,得到原始样本的非线性主元,实现特征提取和数据压缩,将获得的核主元特征通过支持向量机进行模式识别。采用距离比值法和粒子群算法分别对核主元分析和支持向量机的参数进行双重优化选择。实验结果表明,该方法能有效提取故障信号的非线性特征,具有较强的非线性模式识别能力,相比主元分析和支持向量机方法,分类效果更好,实时性更强,可快速有效实现异步电机故障诊断。
[Abstract]:The fault vibration signal of asynchronous motor has strong nonlinear characteristics, but the traditional linear analysis method is easy to cause the loss of the nonlinear component of the vibration signal. A fault diagnosis method for asynchronous motor based on kernel principal component analysis (KPCA) and particle swarm support vector machine (PSO) is proposed. Using kernel function to realize nonlinear mapping from input space to high dimensional feature space and principal component analysis of mapping data, the nonlinear principal components of original samples are obtained, and feature extraction and data compression are realized. The kernel principal feature is recognized by support vector machine (SVM). The distance ratio method and particle swarm optimization algorithm are used to select the parameters of kernel principal component analysis and support vector machine respectively. The experimental results show that this method can extract the nonlinear features of fault signals effectively and has a strong nonlinear pattern recognition ability. Compared with principal component analysis and support vector machine, the classification effect is better and the real time is better. The fault diagnosis of asynchronous motor can be realized quickly and effectively.
【作者单位】：湖南科技大学机械设备健康维护省重点实验室;
【基金】：国家高技术研究发展计划(“八六三”计划)资助项目(2013AA041105) 国家自然科学基金资助项目(51105138) 湖南省教育厅资助项目(11A034) 湖南省科技计划资助项目(2012GK3100) 湖南省高校科技创新团队支持计划资助项目
【分类号】：TM343

【相似文献】