PCB故障智能诊断与可视化技术研究

发布时间：2018-04-01 00:38

本文选题：印刷电路板　切入点：故障诊断　出处：《上海工程技术大学》2015年硕士论文

【摘要】：随着科学技术的发展,印刷电路板(Printed Circuit Board,PCB)需求量的日益增加,板载电子电路系统也变得越来越复杂,对出厂前故障诊断的要求也越来越高,如何快速、高效地检测PCB板的质量,已经成为生产企业不得不高度重视的问题。然而,传统的人工检测方法存在着劳动强度大、可靠性差(易发生漏检和误检)等缺点,直观、高效的智能检测方法逐步替代人工检测方法成为必然趋势。本文对PCB的智能诊断和可视化技术进行研究,首先介绍了PCB故障诊断技术的背景及其研究意义,比较全面地阐述了PCB的故障诊断技术,包括PCB故障的分类、诊断PCB故障的常用方法等,着重阐述了本文所采用的PCB功能测试法;并针对传统的PCB故障诊断技术过程繁琐、不直观等缺点,提出了基于工控软件Lab VIEW(Laboratory Virtual Instrument Engineering Workbench)和PLC(Programmable Logic Controller,PLC)的PCB故障诊断可视化系统。该系统将采用Lab VIEW和PLC作为主要的开发平台,为了解决PCB故障元件的精确定位问题,本文利用Visual C++6.0封装Visio 2007 Active X Control绘图控件,继而将Active X控件嵌入到Lab VIEW中,通过Lab VIEW友好的人机交互界面定位被检测到的故障元件,以实现系统的可视化;为更好地建立PCB故障智能诊断与可视化系统的模型,本文还引入了隐马尔科夫模型(Hidden Markov Models,HMM),并且对HMM要解决的三个基本问题,即评估问题、解码问题和训练问题,以及针对这三个问题的Foreword-Backward算法、Viterbi算法和Baum-Welch算法进行了详述;为解决HMM初始参数的确立及训练的难题,引入了遗传算法(Genetic Algorithm,GA)和时域上相关的隐马尔科夫模型(Time Correlation HMM,TC-HMM);随后,分别介绍了遗传编码、适应度函数、遗传算子以及时域上相关的隐马尔科夫模型算法的设计;为了能更加清晰地描述PCB故障智能诊断与可视化系统的构建,本文重点阐述了基于遗传算法与TC-HMM相结合的故障诊断方法的实现,最终利用遗传算法改进和优化时域上相关的隐马尔科夫模型的参数初始化训练过程,以建立本文所需的PCB故障智能诊断模型。最后,为验证本文所提出的算法和所建立模型的综合性能,建立了智能诊断模型的实验平台,通过大量实验和测试验证诊断系统的性能。实际结果表明该诊断系统达到了预期设计目的,提高了故障诊断效率,并实现了故障定位的可视化,具有自动化程度高、操作简单、直观等特点,适用于大批量PCB板的故障检测。
[Abstract]:With the development of science and technology, the demand for printed Circuit boards is increasing day by day, and the on-board electronic circuit system is becoming more and more complex.Has become the production enterprise has to attach great importance to the problem.However, the traditional manual detection method has many disadvantages, such as high labor intensity, poor reliability (easy to miss detection and false detection) and so on. Therefore, it is inevitable to replace manual detection method with intuitive and efficient intelligent detection method.In this paper, the intelligent diagnosis and visualization technology of PCB is studied. Firstly, the background and significance of PCB fault diagnosis technology are introduced, and the fault diagnosis technology of PCB, including the classification of PCB fault, is expounded comprehensively.In this paper, the PCB function test method used in this paper is expounded, and the disadvantages of the traditional PCB fault diagnosis technology are discussed, such as tedious, not intuitive, etc.This paper presents a visual PCB fault diagnosis system based on Lab VIEW(Laboratory Virtual Instrument Engineering Workbench and PLC(Programmable Logic Controller.This system will adopt Lab VIEW and PLC as the main development platform. In order to solve the problem of accurate location of PCB fault components, this paper encapsulates the Visio 2007 Active X Control drawing control by Visual C 6.0, and then embeds Active X control into Lab VIEW.In order to realize the visualization of the system by locating the detected fault elements through the friendly man-machine interface of Lab VIEW, the model of the intelligent diagnosis and visualization system for PCB fault is established.In this paper, Hidden Markov models are also introduced, and three basic problems to be solved by HMM, namely, evaluation problem, decoding problem and training problem, as well as the Foreword-Backward algorithm and Baum-Welch algorithm for these three problems are described in detail.In order to solve the problem of establishing and training the initial parameters of HMM, the genetic algorithm (GA) and the time Correlation HMMC-HMMM model in time domain are introduced, and the genetic coding and fitness function are introduced respectively.The design of genetic operator and hidden Markov model algorithm in time domain, in order to describe the construction of intelligent diagnosis and visualization system of PCB fault more clearly,This paper focuses on the implementation of fault diagnosis method based on genetic algorithm and TC-HMM. Finally, genetic algorithm is used to improve and optimize the parameter initialization training process of the related hidden Markov model in time domain.In order to establish the PCB fault intelligent diagnosis model needed in this paper.Finally, in order to verify the proposed algorithm and the comprehensive performance of the established model, an experimental platform of intelligent diagnosis model is established, and the performance of the diagnostic system is verified by a large number of experiments and tests.The practical results show that the system achieves the desired design purpose, improves the efficiency of fault diagnosis, and realizes the visualization of fault location. It has the characteristics of high automation, simple operation, intuitive operation, etc.It is suitable for the fault detection of mass PCB board.
【学位授予单位】：上海工程技术大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN41

【参考文献】