基于小波神经网络的模拟电路故障诊断

Analogue circuit faults diagnosis based on wavelet neural network

摘要:

介绍了模拟电路故障诊断的神经网络方法及小波神经网络,以一带通滤波器为例,提出了一种基于输出灵敏度分析,利用多频测试生成故障特征向量训练小波神经网络进行故障诊断的方法。仿真结果表明小波神经网络作为故障分类器具有收敛速度快,诊断准确等特点。

Abstract:

The paper introduces the method of analogue circuit faults diagnosis based on neural network and the theory of wavelet neural network, offers a method which used the output sensitivity analysis and multi-frequency faults feature vectors extraction. The simulation result shows that the wavelet neural network has the properties of fast convergence and accurate diagnosis as a faults classifier.