基于神经网络的数模混合电路故障诊断模型设计

神经网络是智能故障诊断中运用最为广泛的诊断方法之一,能应对复杂数字电路和模拟电路故障诊断;基于VC++的神经网络诊断系统具有运算速度快、输入输出界面丰富、易于实现在线故障诊断等特点,而Matlab提供强大的神经网络工具箱,在网络训练方面有优势;利用参数文件将二者优势结合起来,可实现功能完善的数模混合电路智能故障诊断系统;诊断实例表明,使用该方法实现数模混合电路故障诊断系统,具有高建模速度和高诊断精度的特点。

Design of Mixed-Signal Circuits Fault Diagnosis Model Based on Artificial Neural Networks

Artificial Neural Networks(ANN)was used widely in intelligent fault diagnosis system and can deal with complex logical cir- cuit and analog circuit.An ANN fault diagnosis system based on VC runs fast,has abundant user interface and can realize online fault diagnosis.Matlab supplies powerful ANN toolbox and has advantage in training networks.A perfect intelligent fault diagnosis system for mixed-signal circuits will be built by combining VC and Matlab with parameter files.The example showed that mixed-signal circuits fault diagnosis systems with this method can be built rapidly and diagnose faults accurately.