节点文献

基于小波神经网络的模拟电路故障诊断的研究

Analog Circuit Fault Diagnosis Based on Wavelet Neural Network

【作者】 王勇

【导师】 李春明;

【作者基本信息】 内蒙古工业大学 , 控制理论与控制工程, 2006, 硕士

【摘要】 模拟电路故障诊断一直是一个十分必要且有意义的课题。本文对容差模拟电路故障诊断算法的小波神经网络实现作了讨论和研究,分别在直流电路和交流电路下对该算法进行了仿真,取得了较好的诊断效果。小波神经网络由于把神经网络的自学习特性和小波的局部特性结合起来,具有自适应分辨性和良好的容错能力,因此在故障诊断领域得到了广泛的应用。本文首先介绍了模拟电路故障诊断的理论,建立了集小波分析与神经网络于一体的紧致型小波神经网络。根据元件存在容差时,电路输出响应在一定区间内变化的特性,结合模拟电路故障诊断即是对故障特征进行分类的特点,给出了利用小波神经网络进行模拟电路故障诊断的算法,同时论述了利用小波神经网络实现诊断系统的过程。最后利用MATLAB软件分别设计了基于BP网络和小波神经网络的故障诊断程序,并对两者进行对比仿真实验,证明了小波神经网络更适合处理情况复杂的模式分类问题。本文虽取得了一些成果,但距离工程实际应用仍有一定差距,还需要在后续的研究中进一步加以改进。

【Abstract】 The theory of analog circuit fault diagnosis is very important and significative. In this paper, an algorithm using wavelet neural network for fault diagnosis in analog circuit with tolerance is proposed, and is simulated on the direct circuit and the alternating circuit and receives better result.By reason of the combination of self-learning characteristic of neural network and local characteristic of wavelet analysis, wavelet neural network has self-adapting and favorable fault tolerant ability, so it has been applied in fault diagnosis field far and wide. The theory of analog circuit fault diagnosis is introduced in this paper firstly, and the tight wavelet neural network is constituted taking the nonlinear Morlet wavelet radices as the stimulant function. Based on the characteristics that the circuit’s responses vary in certain interval when the components have tolerances, combined with the features that the analog circuit fault diagnosis is a question of the fault characteristic classification in essence, an algorithm is presented in this paper in which the wavelet neural network is used for fault diagnosis in analog circuit. At the same time, the process of the design of a diagnostic system using wavelet neural network is discussed.At last, corresponding programs are designed for training the BP neural network and wavelet neural network by MATLAB, and simulation analysis proved that wavelet neural network is adaptive in complex pattern classifying.In this paper, some results are obtained, but there is also some difference from the practical application, and it still should be studied in the future.

  • 【分类号】TP183;TN710
  • 【被引频次】14
  • 【下载频次】397
节点文献中: 

本文链接的文献网络图示:

本文的引文网络