基于遗传小波神经网络的模拟电路故障诊断方法的研究

发布时间：2018-04-21 11:23

本文选题：神经网络 + 故障诊断　；参考：《湖南师范大学》2015年硕士论文

【摘要】：信息处理技术在当今得到了快速发展,电子设备中的电路变得日益复杂,由模拟电路引起的设备故障问题,要得到有效处理却日益棘手。集成电路集成度的变高,元器件本身固有的不稳定性等原因给快速定位故障及处理故障带来更大挑战。面对众多出现的问题,传统故障诊断方法已经不能满足社会发展需求,新的诊断技术迫在眉睫。各国研究者开始尝试新的理论研究,其中神经网络作为智能技术运用于模拟电路故诊断研究得到快速发展,在新的诊断技术方面开辟了新路径,并在一段时间内取得了丰硕的成果。现如今,广大学者开始重视将小波分析,遗传算法等多种理论及其融合理论结合神经网络进行故障诊断的新技术,这为智能化故障诊断技术提供了新的思路。LabVIEW软件作为一款功能强大的图形编程软件,可以提供良好的人工交互界面,已经开始运用于故障诊断技术中,为实现故障诊断的简易化提供了便捷之路。本文以新的诊断技术为背景,将小波分析,遗传算法理论融合到神经网络,结合虚拟仪器(Lab VIEW平台),实现电路故障的可视化诊断。介绍了模拟电路故障诊断的研究背景意义、国内外发展现状、存在问题及分类方法。概述人工神经网络理论,包括其特点、应用以及学习方式。以BP神经网络理论为基础,对小波神经网络结构进行构造及其改进算法进行详细讲解,通过仿真实例进行验证所提算法的正确性,其中包括使用软件ORCAD10.5对待诊断电路进行原始数据提取;利用MATLAB软件平台编程对数据进行多分辨分析,提取故障特征值,构造样本集;基于小波神经网络故障诊断方法的实现:针对神经网络权值问题,利用遗传算法进行优化,改善网络性能,最后通过Lab VIEW软件平台实现编写程序的图形化,搭建神经网络模拟电路故障诊断系统界面,实现诊断过程的可视化,操作简易化。
[Abstract]:With the rapid development of information processing technology, the circuits in electronic devices are becoming more and more complex, but the problems caused by analog circuits are becoming more and more difficult to deal with effectively. The high integration of integrated circuits and the inherent instability of components bring greater challenges to fast fault location and fault handling. In the face of many problems, the traditional fault diagnosis method can not meet the needs of social development, new diagnosis technology is urgent. Researchers all over the world began to try new theoretical research, in which neural network as an intelligent technology used in analog circuits so that the rapid development of diagnostic research, in the new diagnostic technology opened up a new path. And in a period of time has achieved fruitful results. Nowadays, many scholars begin to attach importance to the new technology of fault diagnosis, which combines wavelet analysis, genetic algorithm and fusion theory with neural network. This provides a new idea for intelligent fault diagnosis technology. LabVIEW software, as a powerful graphical programming software, can provide a good interactive interface, and has been used in fault diagnosis technology. It provides a convenient way to realize the simplification of fault diagnosis. In this paper, based on the new diagnosis technology, wavelet analysis and genetic algorithm theory are combined into neural network, and the visual diagnosis of circuit fault is realized by combining virtual instrument with LabLab VIEW platform. This paper introduces the research background significance, development status, existing problems and classification methods of analog circuit fault diagnosis. This paper summarizes the theory of artificial neural network, including its characteristics, applications and learning methods. Based on BP neural network theory, the structure of wavelet neural network structure and its improved algorithm are explained in detail, and the correctness of the proposed algorithm is verified by a simulation example. The software ORCAD10.5 is used to extract the original data from the diagnosis circuit, the multi-resolution analysis of the data is carried out by using the MATLAB software platform, the fault characteristic value is extracted, and the sample set is constructed. The realization of fault diagnosis method based on wavelet neural network: aiming at the weight problem of neural network, the genetic algorithm is used to optimize the network performance and improve the network performance. Finally, the graphical programming is realized through Lab VIEW software platform. The interface of the neural network analog circuit fault diagnosis system is built to realize the visualization of the diagnosis process and the simplicity of operation.
【学位授予单位】：湖南师范大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP183;TN710

【参考文献】