多智能体的分布式智能故障诊断

Fault diagnosis model distributed intelligence based on multi-agent system

中文关键词  多Agent系统(MAS)  智能故障诊断  任务分解  互操作  面向Agent编程(AOP)

英文关键词  multi-agent system (MAS)  intelligence fault diagnosis  task decompose  mutual operation  agent-oriented programming

中文摘要

      基于分布式人工智能的思想 ,将多Agent技术引入复杂故障诊断领域 ,分析了基于MAS的分布式智能故障诊断方法和过程 ;讨论了基于模式聚类的故障求解机制及对诊断问题任务辨识、分解 ;研究了多Agent宏观上的约束和关联 ;设计了应用Agent工作状态的表达机制 ;确定了应用Agent间的工作状态影响关系及多Agent间的交互、协作和通讯 ;构建了多Agent模糊关联模型 ;给出了多Agent诊断系统局部诊断决策与全局诊断决策的集成描述结构 ;建立了一种分布式Agent诊断系统结构及其原型系统 .在某电力企

英文摘要

      Modern malfunction diagnosis is a dynamic,distributed,flexible,real-time and uncertain complicated system.The paper tried to introduce multi-agent technology to the complicated malfunction diagnosis.Based on the distributed artificial intelligence,the paper analyzed MAS based on the distributed intelligent malfunction diagnosis method and process,and discussed pattern-clustering based malfunction solution-finding mechanism and the identification and decomposing of a diagnosis task.It also (studied) the restriction and relationship of multi-agent in the sense of macroscopic,and designed the expression mechanism for applied agent's work status.Else it ascertained the influence and the communication and the interaction among applied agents,constructed a multi-agent fuzzy related model.The paper presented an integrated descriptive structure of local diagnosis decision and global diagnosis decision,established a distributed agent diagnosis system structure and its prototype model system too.When the model was applied to the power enterprise's security monitor system,the result was similar to expert system.The model overcame many deficiencies of precious monitor diagnosis system,enhanced the efficiency of the enterprise security running.Compared with the traditional diagnosis method,the proposed model has the special advantage in the distributed domain.