论文部分内容阅读
基于支持向量机(SVM)的故障诊断方法是当前主要的模拟电路故障诊断方法之一,但由传统的二分类SVM组成的故障分类器对新故障模式缺乏处理能力。针对该问题,提出了结合单类支持向量机(OCSVM)和SVM的故障诊断方法。该方法采用OCSVM对故障数据进行检测和初步分类,采用SVM来提高分类性能;最后,采用脉宽调制电路进行故障诊断实验,实验结果说明了所提出的故障诊断方法的有效性。
The fault diagnosis method based on Support Vector Machine (SVM) is one of the main methods of fault diagnosis for analog circuits at present. However, the fault classifier consisting of the traditional two-class SVM lacks the processing capability for the new fault modes. To solve this problem, a fault diagnosis method combining single support vector machine (SVM) and SVM is proposed. The method uses OCSVM to detect and classify the fault data, and uses SVM to improve the classification performance. Finally, the pulse width modulation circuit is used to diagnose the fault. The experimental results show the effectiveness of the proposed fault diagnosis method.