Abstract: Compared with the traditional power spectrum, the vector power spectrum can fuse the power information from different channels, and reflect more comprehensive information. The support vector machine (SVM) is a powerful tool for solving the small-sample fault recognition. Here, combining vector power spectrum and SVM, a new fault diagnosis approach is proposed, this approach is that vector power spectrum is used as eigenvectors, SVM as a classifier. The proposed approach has been successfully applied to the fault diagnosis of rotating machinery. In addition, The proposed approach is compared with the recognition approach based on he vector power spectrum and radial basis function network (RBFN). The experiment result shows that the proposed approach is very effective, especially in the small samples.

Key words: Information fusion, Vector power spectrum, Support vector machine (SVM), Fault diagnosis

摘要： 与传统功率谱相比，矢功率谱融合了多通道的能量信息，反映的信息更全面。而支持向量机为解决机械故障诊断中有限的故障样本识别问题提供了一种有力的工具。基于此，结合矢功率谱和支持向量机，提出了一种故障诊断的新方法。该方法是以矢功率谱作为特征输入到支持向量机的多故障分类器进行故障识别，并应用到旋转机械故障诊断中。同时，该方法还与基于矢功率谱的径向基函数网络识别结果进行了比较，实验结果表明，该方法是有效的，尤其在小样本情况下，SVM识别效果明显优于径向基函数网络。

关键词: 信息融合, 矢功率谱, 支持向量机, 故障诊断