【摘要】 针对汽轮发电机组故障诊断问题的复杂性、实时性要求,提出了主要由监测与诊断Agent、诊断方法Agent、管理Agent、黑板Agent等组成的分布式多智能体系统(MAS)。利用功能分解方法,实现了多级流模型(MFM)和分布式MAS协作的故障诊断方法,其基本思想是:当现场发生异常时,监测与诊断Agent之间协同,依靠汽轮发电机组MFM模型,通过系统中能量、物质或信息流关系建立故障搜索路径,实现故障的快速定位;根据已定位的故障节点(设备)特征,在管理Agent的协调下,诊断方法Agent和监测与诊断Agent协同组成针对性很强的MAS,完成故障的进一步精确诊断。更多还原

【Abstract】 A distributed MAS(Multi-Agent System) composed of the monitoring and diagnosis Agent,diagnostic method Agent,management Agent,blackboard Agent and so on is proposed for the fault diagnosis of steam turbine generator unit.The functional analysis method is used to realize the cooperation of the MFM(Multi-level Flow Models) and the distributed MAS for fault diagnosis.When abnormity occurs,the monitoring and diagnosis Agents cooperate with each other,and based on its MFM,the fault searching path is set up according to the connections among energy,substance and information streams to realize the fast fault locating.According to the features of the faulty node or equipment,the specific MAS is organized by the diagnostic method Agent and the monitoring and diagnosis Agent under the harmonization of the management Agent to realize the precise fault diagnosis.更多还原