Research on Fault Diagnosis Based on SVM

摘要：

　　In order to improve the fault diagnosis precision of electron system,a method based on wavelet packet transform and SVM was proposed.Fault diagnosis method based on SVM was researched on in this paper because of the complexity of electron system,difficulty of fault diagnosis method and special advantages of SVM.Wavelet packet transform is used to extract fault features from the signal of the circuit output voltage.The specific feature extraction method is introduced.Improved QPSO algorithm was proposed to improve the training speed and class precision of SVM.At last the method mentioned above was applied to a circuit.The result showed that this method was very good.

关键词： SVM fault diagnosis QPSO improved QPSO multi-class SVM

作者: Li Wanling Wang Zhensheng Deng Daquan Fang Jun

作者单位: Ordnance Technology Research Institute Shijiazhuang,China Equipment Department Troop 95829 Xiaogan,China

会议类型: 国际会议

会议名称: 2013 International Conference on Energy Research and Power Engineering(2013能源研究与电力工程国际会议)(ERPE2013)

会议地点: 郑州

会议语种:英文

页码: 1283-1286

在线出版日期: 2013-05-24（万方平台首次上网日期，不代表论文的发表时间）