基于相量分析与支持向量机的交流电路故障诊断  

AC Circuit Fault Diagnosis Based on Phasor Analysis and Support Vector Machines

作　　者：黄亮[1] 侯建军[1] 刘颖[1] 宋伟[1] 李赵红[1] 

机构地区：[1]北京交通大学电子信息工程学院,北京100044

出　　处：《北京交通大学学报》2008年第5期30-33,共4页JOURNAL OF BEIJING JIAOTONG UNIVERSITY

基　　金：国家"863"计划资助项目(2007AA01Z241-2);北京交通大学科技基金资助项目(2006XM002)

摘　　要：针对工作在交流状态下的模拟电路,提出采用相量分析和支持向量机(SVM)相结合的方法,实现交流电路的故障诊断.首先构建交流电路的数学模型,编写相量分析程序,得到输出信号的幅度和相位;对电路的所有状态进行运算,得到故障样本和测试样本.然后用故障样本对SVM进行训练;最后用训练好的SVM对测试样本进行模式识别.通过一个交流电路故障诊断的实例表明,该方法具有算法简单、泛化能力强等优点,可以快速准确地判断出故障元件.To diagnose the fault of AC circuits including several dynamic components, a innovative method based on phasor analysis and support vector machines is presented. Firstly, the circuit program is compiled in MATLAB software, and phasor expressions of output signals are obtained. Secondly, the fault samples are sent into SVM to train the SVM. Thirdly, the test samples are classified by trained SVM. Finally, a fault diagnosis example of AC circuit is provided. The result shows that this method has the advantages of simple algorithm, high efficiency, great capability in generalization and classification.

关 键 词：支持向量机 故障诊断 相量分析 交流电路 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP206.3[自动化与计算机技术—控制科学与工程]