一种基于小波分析和神经网络的模拟电路故障诊断的方法研究  

ON WAVELET ANALYSIS AND NEURAL NETWORKS-BASED ANALOG CIRCUIT FAULT DIAGNOSIS

作　　者：杨立儒[1] 李红波[1] 郑红岩[1] 曾繁景[1] 李铁峰[1] 

机构地区：[1]信息工程大学信息工程学院,河南郑州450002

出　　处：《计算机应用与软件》2011年第3期223-226,共4页Computer Applications and Software

摘　　要：提出了一种基于小波分析和神经网络的电路故障诊断方法。首先用PSPICE(Simulation Program with Integrated Circuit Emphasis,即集成电路编程仿真技术)电路仿真软件对电路进行仿真;然后对电路的输出节点电压信号进行小波分解,提取各频段的能量作为故障样本;最后利用神经网络的并行处理结构和非线性映射能力,实现故障诊断。仿真实验结果表明该方法对容差模拟电路故障定位具有较高的准确率,为模拟电路故障诊断技术开辟了一条道路,为模拟电路故障诊断技术开辟了一条道路。A new method of fault diagnosis of analog circuit based on wavelet analysis and neural network is presented in this paper.The method uses PSpice as the tool to simulate the circuit first,then uses wavelet to decompose voltage signal of the output node of the circuit,picks up the energy of each band as the fault sample,at last uses parallel processing structure of neural network and nonlinear mapping capability to realise the fault diagnosis.Simulation result shows that the proposed method has high accuracy in diagnosing the faults of tolerance analog circuits and blazes a trail for fault diagnosis technique of analog circuit.

关 键 词：PSPICE 小波分析 神经网络 故障诊断 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]