CAJ | 학술논문

针对模拟电路的固有复杂性及其传统故障检测方法存在延时大和正确识别率低的问题，提出基于最小二乘支持向量机和Volterra级数的故障诊断方法。采用Volterra级数频域核对电路故障特征进行提取，利用最小二乘支持向量机进行模态分类，最终完成故障诊断。仿真结果表明，该方法与BP神经网络相比提高了系统故障辨识能力与系统故障诊断速度。
침대모의전로적고유복잡성급기전통고장검측방법존재연시대화정학식별솔저적문제，제출기우최소이승지지향량궤화Volterra급수적고장진단방법。채용Volterra급수빈역핵대전로고장특정진행제취，이용최소이승지지향량궤진행모태분류，최종완성고장진단。방진결과표명，해방법여BP신경망락상비제고료계통고장변식능력여계통고장진단속도。
To solve the problem of analog circuit complexity ,long time diagnosis and low correct recognition rate in traditional fault diagnosis ,a new fault diagnosis method was proposed which combined least squares support vector machine (LS-SVM) and Volterra series .The circuit feature extraction was completed by using Volterra frequency-domain core ,LS-SVM was used to do modal classification and finally the fault diagnosis was accomplished .Simulation results show that this method ,compared with traditional BP neural network ,is more efficient and accurate in system fault recognition and diagnosis .