基于高阶谱与支持向量机的电力电子电路故障诊断技术.pdfVIP

27 卷 10 期 中 国 电 机 工 程 学 报 Vol.27 No.10 Apr. 2007 2007 年4 月 Proceedings of the CSEE 2007 Chin.Soc.for Elec.Eng. 文章编号：0258-8013 (2007) 10-0062-05 中图分类号：TM921;TP 18 文献标识码：A 学科分类号：47040 基于高阶谱与支持向量机的电力电子 电路故障诊断技术 江，王友仁，刘 权 ( 南京航空航天大学自动化学院，江苏省 南京市 210016) The Technique of Power Electronic Circuit Fault Diagnosis Based on Higher-order Spectrum Analysis and Support Vector Machines CUI Jiang, WANG You-ren, LIU Quan (College of Automation and Engineering, Nanji ng University of Aeronautics and Astronautics, Nanji ng 210016, Jiangsu Province, China) ABSTRACT: Aiming at drawbacks of current methods for 关键词：电力电子电路；故障诊断；高阶谱；支持向量机； power electronic circuit fault diagnosis, this paper proposed a 模糊C-均值 method of power electronic circuit fault diagnosis based on 0 引言 higher-order spectrum(HOS) analysis and support vector machines(SVMs). Firstly, the faulty circuit information feature 电力电子电路广泛应用于工业、军事、航空航 was analyzed, processed and extracted by bi-spectrum analysis, 天等重要领域，是电源系统和各种动力、控制系统 and then the fault modes were trained and recognized by a 不可或缺的一部分。研究电力电子电路的故障诊断 hierarchical multi-class SVMs, whose structure is optimized by 技术，对于提高系统的可靠性，保证装备的可用性 fuzzy C-means(FCM) algorithm. Finally, a common Buck 具有重大、现实的经济和军事意义。 power circuit was modeled, simulated and diagnosed to test the 目前，电力电子电路故障诊断的方法大体可以 proposed method. The results show that the fault detection and 分为2 种：基于特征提取的方法和基于模式识别的 location accuracy is up to 99%, which is ideal for fault 方法。其中，基于特征提取的方法有：沃尔什变 di