基于RBF神经网络的模拟电路故障诊断的研究

Fault Diagnosis of Analog Circuit Based on RBF Neural Network

对于模拟电路故障诊断问题,由于存在着容差以及非线性等问题,传统的诊断方法计算复杂,实用性比较差。神经网络作为现代故障诊断的代表方法,越来越受到人们的重视。常用的BP神经网络由于其收敛速度慢,结构难以确定等不足,在诊断过程中往往不能达到满意的效果。为了解决这一问题,论文提出了基于径向基函数(RBF)神经网络的模拟电路故障诊断方法。通过建立电路仿真模型,使用小波分解能量作为特征向量,利用RBF神经网络实现模拟电路的故障诊断。实验证明,使用这一方法,可以有效实现模拟电路的故障诊断。

For fault diagnosis of analog circuit problems,due to the existence of tolerance and non-linear problems,traditional diagnostic methods are complex and have relatively poor practicality.As a representative of the modern fault diagnosis method,more attention has been paid on neural network.Commonly used BP neural network has slow convergence,and is difficult to determine the structure and other issues,so ideal goal in the diagnosis process cannot be achieved.In order to solve this problem,this paper presents a method on fault diagnosis of analog circuit based on RBF neural network.Through the establishment of circuit simulation model,using wavelet energy as feature vectors,RBF neural network is used to achieve fault diagnosis of analog circuit.Experimental results show that this method can effectively implement fault diagnosis of analog circuit.