基于Agent与故障树的故障诊断专家系统的研究

发布时间：2018-01-21 03:08

本文关键词： Agent 故障树二叉故障树故障诊断专家系统　出处：《大连交通大学》2012年硕士论文　论文类型：学位论文

【摘要】：故障诊断专家系统是故障诊断技术的一个发展方向,其系统的设计、采用的技术及知识库、推理机的设计等问题成为故障诊断专家系统中要解决的首要问题。结合省自然科学基金项目“客运专线列车故障诊断专家系统关键技术研究”(项目编号：20102014),针对电力机车的电气系统的故障的特点,研究了开发故障诊断专家系统需要用到的相关技术,分析了故障诊断专家系统的架构设计及开发方法。根据电力机车电气系统的故障特征,以当前故障诊断技术的现状为背景,以故障诊断专家系统要解决的首要问题为目标,结合人工智能技术、计算机技术、故障诊断技术等领域的知识,设计一个通用性高、可靠性好的电力机车电气系统的故障诊断专家系统。设计了作为系统组成单元的各个Agent的功能及任务；将传统的专家系统进行改造,使其Agent化,并具备Agent的特点；重点阐述了在本系统中是如何依据故障树模型来进行知识的获取；将以往常规的规则存储方式改为故障现象、故障原因、诊断结论分别进行存储的三级存储机制；在故障树模型的基础上进行改造,结合数据结构的知识,提出二叉故障树模型,设计并实现了故障诊断专家系统的推理机制。将该故障诊断专家系统应用于电力机车的电气部分中,以电气系统中某些故障树为例,详细阐述了基于Agent与故障树的故障专家系统的设计细节与完整的系统建立过程,并对系统仿真实验的结果进行分析,结果证明将Agent与故障树应用于故障诊断领域能够有效地提高诊断任务的工作效率。系统具有良好的扩展性,使得系统能够轻松升级以适应新的需求；系统较高的通用性使得系统无需进行太多改动就可以应用于其他领域及设备的故障诊断中。
[Abstract]:Fault diagnosis expert system is a developing direction of fault diagnosis technology. The design of fault diagnosis expert system, the adopted technology and knowledge base, and the design of inference machine have become the most important problems to be solved in fault diagnosis expert system. Combined with the project of Provincial Natural Science Foundation, "Research on key Technology of Train Fault diagnosis expert system for passenger dedicated Line" (item No.: 20102014), aiming at the characteristics of electric system fault of electric locomotive. In this paper, the related technologies used in the development of fault diagnosis expert system are studied, and the frame design and development method of the fault diagnosis expert system are analyzed, according to the fault characteristics of electric system of electric locomotive. Based on the current situation of fault diagnosis technology and the primary problem of fault diagnosis expert system, this paper combines the knowledge of artificial intelligence, computer technology, fault diagnosis technology and so on. A fault diagnosis expert system for electric system of electric locomotive with high generality and good reliability is designed, and the functions and tasks of each Agent are designed as a component unit of the system. The traditional expert system is reformed to make it Agent and has the characteristics of Agent. How to acquire knowledge according to fault tree model in this system is expounded emphatically. The conventional rule storage method is changed to the three-level storage mechanism of fault phenomenon, fault reason and diagnosis conclusion respectively. Based on the fault tree model and the knowledge of data structure, the binary fault tree model is proposed, and the reasoning mechanism of fault diagnosis expert system is designed and implemented. The fault diagnosis expert system is applied to the electric part of electric locomotive, and some fault trees in the electric system are taken as an example. The design details of fault expert system based on Agent and fault tree and the complete process of system establishment are described in detail, and the results of system simulation experiment are analyzed. The results show that applying Agent and fault tree to fault diagnosis can effectively improve the work efficiency of the diagnosis task. The system has good expansibility and the system can be easily upgraded to meet the new requirements. Because of its high versatility, the system can be applied to fault diagnosis in other fields and equipments without making too many changes.
【学位授予单位】：大连交通大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH165.3;TP182

【引证文献】