摘要
提出了基于输出电压和电源电流协同分析的模拟电路故障诊断新方法。由于不同故障对输出电压和电源电流谐波频谱的幅度和相位的敏感程度不同,因此使用该方法可以有效地检测到灾难型故障和参数型故障,并且该方法对传统方法无法检测的时延类故障依然有效,在协同分析过程中分别采用欧式距离和马氏距离作为灾难型故障和参数型故障检测的判别依据,并且文中分析了由于输入激励不同、分析谐波次数不同所导致的故障覆盖率问题。与基于输出电压的分析方法相比,基于输出电压和电源电流协同分析的方法具有更高的故障覆盖率。
A new method is proposed for analog circuit fault diagnosis based on the collaborative analysis of output voltage and supply current. Because different faults have different sensitivities to the amplitude and phase of the out- put voltage and supply current harmonic spectrum, catastrophic faults and parametric faults can be effectively detected with this method. And time-delay related faults that can not be detected with traditional methods also can be detected using this collaborative analysis method. The Euclidean distance is adopted as the decision criterion of catastrophic faults and the Mahalanobis distance is adopted as the decision criterion of parametric faults. The fault coverage problems caused by the differences in input excitation or harmonic numbers are also analyzed. The collaborative analysis method based on output voltage and supply current has higher fault coverage compared with the traditional methods based on output voltage.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2013年第8期1872-1878,共7页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金重点项目(60934002)
高等学校博士学科点专项科研基金(20100185110004)资助项目
关键词
模拟电路
灾难型故障
参数型故障
欧式距离
马氏距离
analog circuit
catastrophic fault
parametric fault
Euclidean distance
Mahalanobis distance
