基于Multi-Agent技术的大型机电设备故障诊断系统

通过对大型机电设备的故障特点的分析,以及对Multi-Agent系统架构的描述,提出了一种基于Multi-Agent技术的大型机电设备故障诊断结构模型.此MAS系统结构采用了层次模型设计,符合故障诊断的推理方法.文章着重阐述了各类Agent的功能以及它们之间的协作方式.该模型分为监测层、数据处理层和管理层等三个层次,其中监测层用于数据采集和简单故障诊断,数据处理层用于复杂故障诊断和人机交互,管理层用于管理整个MAS系统和协调各Agent的合作关系.各Agent之间通信以KQML语言为基础,并扩展了一组专用的动作表达式以满足故障诊断的需要;文章还描述了复杂故障诊断任务中各Agent之间的协作方式.

A Multi-Agent Architecture of Fault Diagnosis for Large-Scale Electromechanical System

Based on the study of the fault diagnosis process for large-scale electromechanical systems and the multi-agent system,this paper proposes a multi-agent architecture(a multi-agent system,MAS) for the diagnostic process.The architecture adopts a hierarchy model which includes a monitoring layer,an interface and synthesis layer and a management layer.The monitoring agents answer for data acquisition and simple fault diagnosis.The interface agents answer for human-machine interaction.The synthesis agents answer for complex fault diagnosis.The management agent answers for managing the whole MAS and coordinating the work mode among agents.The communication between each agent is accomplished by using the KQML language.Especially,a set of KQML expressions are defined for fault diagnosis process.The collaboration mode between each agent is also proposed for the complex fault diagnosis.