摘要
针对现有电力电子电路故障诊断方法存在的不足,研究了采用高阶谱分析和支持向量机(support vector machine,简称SVM)的电力电子电路故障诊断和定位方法。首先利用高阶谱中的双谱技术分析、处理和提取电路状态的故障信息特征;然后设计和采用多类层次支持向量机分类器作为故障模式的训练和识别器,其中,分类器的结构利用模糊C-均值算法(fuzzyC-means,简称FCM)进行了优化;最后采用一个实际的Buck功率电路进行了建模、仿真和验证。结果表明,采用该方法对电力电子电路故障的诊断和定位率可达99%以上,达到了较为理想的诊断精度。
Aiming at drawbacks of current methods for power electronic circuit fault diagnosis, this paper proposed a method of power electronic circuit fault diagnosis based on higher-order spectrum(HOS) analysis and support vector machines(SVMs). Firstly, the faulty circuit information feature was analyzed, processed and extracted by bi-spectrum analysis, and then the fault modes were trained and recognized by a hierarchical multi-class SVMs, whose structure is optimized by fuzzy C-means(FCM) algorithm. Finally, a common Buck power circuit was modeled, simulated and diagnosed to test the proposed method. The results show that the fault detection and location accuracy is up to 99%, which is ideal for fault diagnosis.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2007年第10期62-66,共5页
Proceedings of the CSEE
基金
国家自然科学基金项目(60374008
60501022)
~~航空科学基金项目(04I52068)
关键词
电力电子电路
故障诊断
高阶谱
支持向量机
模糊C-均值
power electronic circuit
fault diagnosis
higher-order spectrum
support vector machines
fuzzy C-means