模拟电路故障诊断的新方法

New method of analog circuit fault diagnosis

本文提出了用多小波神经网络诊断模拟电路故障的新方法。根据多小波的多分辨率分析思想,构造了一种多小波神经网络,其激励函数为具有紧支撑集、对称性和正交性的多尺度函数和多小波函数。用多小波神经网络和单小波神经网络对实例电路进行故障诊断,仿真结果表明,与单小波神经网络相比,多小波神经网络不仅学习收敛速度快,而且能够更准确地对模拟电路进行故障诊断。

A new method of analog circuit fault diagnosis based on multiwavelet neural network is presented in this paper. According to the muhiresolution analysis thought in muhiwavelet space, an orthonormal multiwavelet neural network model for learning is obtained. The activation functions of the model are compactly supported, symmetric and orthonormal multiscaling functions and multiwavelet functions. The multiwavelet neural network and wavelet neural network are used in the fault diagnosis of an example circuit respectively. Simulation results indicate that multiwavelet neural network converges more rapidly than wavelet neural network and can carry out fault diagnosis of analog circuit more accurately.