基于故障树的故障诊断专家系统软件平台设计

发布时间：2018-06-04 00:18

本文选题：专家系统 + 综合数据库　；参考：《南京理工大学》2017年硕士论文

【摘要】：现代化工业具有设备大型化、生产连续化、高度自动化和高度经济化的特点,而一旦这些设备出现故障,就可能带来巨大的经济损失,甚至灾难性事故。故障诊断技术能够预防、减少或消除故障的发生,这对于提高系统的安全性、可靠性和有效性具有重要意义。因此,开发智能化、实用化的故障诊断系统十分必要,而目前大多数故障诊断软件都较为专用,缺乏较为通用的故障诊断软件,本文在分析了多种故障诊断方法的基础上,确定了专家系统为主要研究对象,设计了集传统专家系统、模糊专家系统和故障树诊断方式于一体的故障诊断通用软件平台。本文的主要研究内容与成果如下:1、设计了综合数据库,将不同表现方式和不同来源的信息映射成征兆事件,并设计了具体映射方式。2、通过对不同类型专家系统知识表现形式及诊断原理的分析,设计了专家系统知识库,将传统产生式规则改变为INPUT-OUTPUT结构规则,能够表述传统规则、模糊规则等不同形式的知识,增强了规则表示能力,INPUT-OUTPUT结构规则还包含推理控制信息进行冲突消解,提高推理效率。3、设计了能够组织和管理多故障树的故障树软件模块,降低了对大型复杂系统建立单一故障树的困难度,不同故障树之间通过设立优先级表明故障树的诊断顺序。4、设计并实现了故障树的图形化建树和自动化显示功能,能够简单高效的建立、维护和分析故障树。5、将所设计的专家系统与故障树相结合,设计了故障诊断软件平台的推理机制,最终完成了故障诊断软件平台的设计和编写。专家系统作为一种智能化诊断方式,在故障诊断领域应用广泛并且具有很高的诊断效率和准确度。通过对综合数据库、规则库、故障树的设计,增强了系统的知识表示能力使系统具有更强的适用性,利用设计的知识库分别设计了规则推理方式和故障树推理方式,并将两种推理方式相结合形成了故障诊断软件平台的推理机制。最后通过对发射车电气系统故障诊断项目和基于三比值法的电力变压器故障诊断方式的应用,验证了所设计的故障诊断软件平台的通用性和有效性。
[Abstract]:Modern industry has the characteristics of large-scale equipment, continuous production, high automation and high economic, but once these equipment failure, it may bring huge economic losses, even catastrophic accidents. Fault diagnosis technology can prevent, reduce or eliminate the occurrence of fault, which is of great significance to improve the safety, reliability and effectiveness of the system. Therefore, it is necessary to develop an intelligent and practical fault diagnosis system. At present, most of the fault diagnosis software are special and lack of common fault diagnosis software. The expert system is chosen as the main research object, and a general software platform for fault diagnosis is designed, which integrates traditional expert system, fuzzy expert system and fault tree diagnosis. The main research contents and results of this paper are as follows: 1. A comprehensive database is designed to map information from different representations and sources into omen events. By analyzing the knowledge representation and diagnosis principle of different types of expert systems, the knowledge base of expert system is designed. The traditional production rules are changed into INPUT-OUTPUT structure rules, and the traditional rules can be expressed. Different forms of knowledge, such as fuzzy rules, enhance the ability of rule representation. The rules also include reasoning control information to resolve conflicts, improve reasoning efficiency. 3. The software module of fault tree which can organize and manage multi-fault trees is designed. The difficulty of establishing a single fault tree for large and complex systems is reduced. The diagnosis sequence of fault tree is indicated by setting priority among different fault trees, and the graphical tree building and automatic display function of fault tree are designed and realized. The reasoning mechanism of fault diagnosis software platform is designed by combining the designed expert system with fault tree. Finally, the fault diagnosis software platform is designed and compiled. As an intelligent diagnosis method, expert system is widely used in fault diagnosis field and has high diagnosis efficiency and accuracy. Through the design of comprehensive database, rule base and fault tree, the knowledge representation ability of the system is enhanced to make the system more applicable, and the rule reasoning mode and fault tree reasoning mode are designed by using the designed knowledge base, respectively. The reasoning mechanism of fault diagnosis software platform is formed by combining the two reasoning methods. Finally, through the application of fault diagnosis project of electric system of launch vehicle and fault diagnosis mode of power transformer based on three-ratio method, the generality and validity of the designed fault diagnosis software platform are verified.
【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP182;TP277

【相似文献】