【作者】 王艳；

【导师】 彭良玉；

【作者基本信息】 湖南师范大学 ， 电路与系统， 2015， 硕士

【摘要】 模拟电路故障诊断技术的研究开始于1960年,目前已在国内外取得了大量有效的科研成果,逐步形成了完善的系统理论,在电路理论中占据非常重要的地位。同时,随着电子工业的飞速发展,电器设备的集成度越来越高,日趋模块化和功能化。但是由于模拟电路自身存在的非线性、连续性、元器件参数容差等特性使得模拟电路故障诊断的难度非常大。采用传统的模拟电路故障诊断方法已难以满足实际工程应用的需求,所以亟需探求新的现代化模拟电路故障诊断技术。诸如神经网络、小波分析、模糊理论、遗传算法等人工智能技术的出现和发展,形成了这一领域新的研究方向。针对模拟电路故障诊断的模糊性和不确定性等问题,采用人工智能新技术的现代模拟电路故障诊断方法为常规方法所不能解决的各类问题带来了新的解决思路。本文系统地分析了几类传统的模拟电路故障诊断方法以及基于智能理论的现代模拟电路故障诊断方法。在此基础上,研究了将BP神经网络、小波分析、小波包分析等理论应用于模拟电路故障诊断中的方法,并引入改进粒子群优化算法优化神经网络的连接权值,达到加快网络收敛速度和提高诊断正确率的目的,进一步提升网络的性能。主要工作有:(1)阐述了模拟电路故障诊断课题的背景意义及当前国内外的发展状况,总结了传统的故障诊断技术以及近年来发展较快的智能故障诊断技术;(2)系统的研究了神经网络、小波分析、小波包分析等理论知识,探索了将这几种技术应用于模拟电路故障诊断中的方法,并选取待测电路进行了仿真分析,用实例证明了该方法的有效性与可行性;(3)对模拟电路故障诊断中最为关键的技术——特征向量的提取进行了详尽的分析与研究。应用小波多分辨分析和小波包分析等技术提取故障特征,并进一步探索将两种方法提取的故障特征向量融合成新的特征向量,作为故障诊断的故障集。通过对待测电路做实例研究的诊断结果表明了此方法的优异性;(4)将粒子群算法引入基于小波神经网络的模拟电路故障诊断中,利用改进粒子群优化算法对小波神经网络的连接权值进行适当的优化,加快了神经网络的收敛速度,并且训练后的网络具有较好的鲁棒性。更多还原

【Abstract】 Fault diagnosis technology of analog circuits was first proposed in 1960. Currently, it has made great academic progresses in the world, and gradually formed a perfect system theory, taken one of the most important roles on circuit theory.At the same time, with the rapid development of the electronics industry, the size and structure of electronic equipment gradually become functionated and modularized. But because of the inherent characteristics in analog circuits such as the non-linearity, continuity and tolerance of component parameters which makes the analog circuits fault diagnosis very difficult. The traditional fault diagnosis technology of analog circuits is difficult to achieve expected effects in practical applications. So with the emergence and development of the artificial intelligence technologies such as neural network, wavelet theory and fuzzy theory has become a new orientation in this filed. In order to solve the problem of the fuzzy and uncertainty of the fault diagnosis of analog circuits that cannot be solved by conventional methods, the modern analog circuit fault diagnosis which using the artificial intelligence brought the new ideas.The paper systematically analyzes the several kinds of traditional and modern diagnostic techniques.And makes intensively studies on the application of fault diagnosis based on the theories of neural network,wavelet analysis, wavelet packet analysis and particle swarmalgorithm. Moreover, to speed up the network convergence speed and the diagnostic accuracy and further enhance the performance of the network, the improved particle swarm algorithms are used to optimize the connection structure of neural network. The main researches of this paper is as follows:(1)Explaining the background significance of analog circuit fault diagnosis technology. Summarizing the traditional fault diagnosis methods and modern diagnosis methods based on artificial intelligence technologies.(2)Explaining the theories of neural network, wavelet transform, wavelet packet transform, and illustrate the superiority of method based on those technologies applied in fault diagnosis with elected circuits.(3)As the key step for fault diagnosis of analog circuits,stress on discusses the extraction of fault feature vector. The methods, wavelet extraction and wavelet packet extraction bounds of this two technologies are explored to extract the fault feature vectors. The superiority of applications on those methods are illustrated with elected circuits.(4)Use PSO algorithm to optimize the wavelet neural network and improve its structure. The result show this improved method can greatly improve performance of the wavelet neural network.更多还原