基于神经网络方法的模拟电路故障诊断研究  

Research on analog circuit fault diagnosis based on neural network method

作　　者：周晶晶[1] 程慧华[1] 安明[2] 刘琼俐[1] 

机构地区：[1]武汉军械士官学校指挥控制系 [2]中国人民解放军65583部队保障部

出　　处：《现代电子技术》2015年第23期47-50,共4页Modern Electronics Technique

摘　　要：小波变换是一种时频分析方法,具有多分辨率特性,被誉为数学显微镜;而BP神经网络具有较好的泛化能力,很适合判断电路状态属于哪种故障类型的分类问题。将二者结合起来,采用基于小波神经网络的模拟电路故障诊断方法,应用小波变换对模拟电路幅频响应的采样信号进行故障特征提取,然后利用BP神经网络对各种状态下的特征向量进行分类决策,实现模拟电路的故障诊断。通过对电路进行仿真,证明该方法能够实现故障检测及定位,具有准确率高的特点。Wavelet transform is a time-frequency analysis method, and is honoured as the mathematical microscope be- cause of its multi-resolution characteristics. Since BP neural network has great generalization ability, and is suitable to deter- mine which kind of fault type the circuit status belongs to, the analog circuit fault diagnosis method based on wavelet neural net- work is adopted by combining with the two methods. The wavelet transform is used to extract the fault feature of sampled signal responding to amplitude-frequency in analog circuit, and then the eigenvectors in various status are classified and determined with BP neural network to implement analog circuit fault diagnosis. Simulation of the circuit proves that the method can realize fault detection and location, and has high precision.

关 键 词：小波变换 模拟电路故障诊断 神经网络 特征向量 

分 类 号：TN711-34[电子电信—电路与系统]