故障诊断专家系统中知识库组织和维护技术的研究  

Research of knowledge-base organization and maintenance techniques for a fault diagnosis expert system

作　　者：许亮[1] 李秀喜[1] 周华[1] 钱宇[1] 

机构地区：[1]华南理工大学化工与能源学院,广东广州510640

出　　处：《计算机与应用化学》2006年第11期1124-1128,共5页Computers and Applied Chemistry

基　　金：国家自然科学基金(20376025;20536020)国家杰出青年科学基金(202225620)教育部跨世纪人才基金

摘　　要：在开发化工过程实时故障诊断专家系统的基础上,本文讨论了知识库的组织和维护技术,对知识表示和知识库结构进行了详细介绍。在知识库维护方面,采用面向对象的方法描述领域知识,并实现知识获取和转换；对于知识校验,提出一种基于有向图的知识校验方法。以润滑油生产过程为例,讨论了系统的知识校验过程。本文讨论的技术和策略已应用于润滑油过程实时故障诊断专家系统。实际应用表明,该技术和策略保证了知识库和专家系统的稳定和高效率。Based on an expert system of real-time fault diagnosis of chemical processes developed, techniques of the knowledge-base organization and maintenance are presented in this paper. The knowledge representation and structure of knowledge base are discussed in detail. For the knowledge-base maintenance, an oriented-object method is adopted to describe a domain knowledge, and the knowledge acquisition and transform is implemented. A method is presented for knowledge verification based on a directed graph. The lubricating oil refining process is taken as the industrial implementation ease, the process of knowledge verification is discussed. The techniques and strategies presented in the paper are used to the expert system of real-time fault diagnosis of lubricating oil refining process. It has shown that the techniques and strategies ensure the knowledge base and the expert system to work stably and better efficiently.

关 键 词：专家系统 知识表示 知识库维护 有向图 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]