摘要
故障特征提取和分类器设计是模拟电路故障诊断的两个重要环节,为了提高模拟电路故障辨识的准确率,提出了提升小波变换与支持向量机相结合的故障诊断方法。根据提升小波变换的原理,提取被测电路单脉冲响应信号的小波系数构成故障特征,建立以支持向量机为分类器的故障诊断系统。该方法对两个滤波器电路的故障诊断取得了满意的效果,在故障模式较多的情况下故障分类的精度达到了99%以上,优于传统的小波方法。
Feature extraction and classifier construction are two important stages for analog circuit fault diagnosis. In order to improve correctness rate of fault identification, an approach based on lifting wavelet transform （LWT） and support vector machine （SVM） is proposed. According to the theory of lifting wavelet transform, the impulse response signal of CUT is sampled and decomposed to form fault features, and then the analog circuit fault diagnosis system is established. The experimental results on two filter circuits show that the presented approach is superior to classical wavelet analysis based methods. The fault classification accuracy can be higher than 99% with respect to large number of fault categories.
出处
《电子测量与仪器学报》
CSCD
2010年第1期17-22,共6页
Journal of Electronic Measurement and Instrumentation
基金
国防基础科研(编号:A1420061264)资助项目
国家自然科学基金(编号:60673011)资助项目
