基于小波网络的模拟电路故障诊断  

Wavelet network -based fault diagnosis approach for analog circuits

作　　者：谭杰[1] 尹新[1] 谢胜曙[1] 何怡刚[1] 

机构地区：[1]湖南大学电气与信息工程学院,410082

出　　处：《微计算机信息》2008年第16期165-166,199,共3页Control & Automation

基　　金：国家自然科学基金(50677014);高校博士点基金(20060532002);湖南省自然科学基金(06JJ2024);教育部新世纪优秀人才支持计划(NCET-04-0767);国家863计划资助

摘　　要：根据实际应用模拟电路中所存在噪声的问题,提出利用相位差来进行故障诊断的新方法。针对噪声存在于各个频段,利用不同频段的带通滤波器对诊断电路进行滤波,得到各种频率的信号,在正常模式和故障模式下分别对这些信号进行相位差和幅值差的特征提取,建立故障字典;然后利用小波神经网络具有非线性映射的优势,对各种状态下的故障特征进行分类决策,从而实现故障诊断。此方法通过实例证明不但诊断准确,而且很切合实际。According to the existed noise of analog circuit in practical use, a new method is presented which uses the phase difference to carry out fault diagnosis. Because the noise is in each frequency band, we make use of band-pass filter in different frequency to filter the diagnosed circuits to get the signal of different frequency. In the normal and fault condition, the character of the signal based on different phase and different magnitude are extracted. Then establish the diagnosis dictionary. Finally, the fault character can be classified by making use of the wavelet neural network that is good at non-linear mapping. Thus, the fault diagnosis is achieved .The simulation and testing results show that the proposed approach for fault diagnosis of analog circuits is effective and practical.

关 键 词：故障诊断 相位差 小波神经网络 故障特征 滤波器 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TN710[自动化与计算机技术—控制科学与工程]