基于MAS的分布式智能故障诊断模型与关键技术研究

现代故障诊断已是一个动态的、分布的、柔性的、实时的和不确定的复杂系统，将多Agent技术引入复杂故障诊断领域，是求解复杂过程的故障诊断问题的一种新的尝试。文中讨论基于multi-agent system(MAS)的分布式智能故障诊断方法和过程，设计一种Agent诊断系统结构及其原型系统，对诊断问题任务辨识、分解、各Agent的内部诊断机制、多Agent间的交互、协作、关联模型以及诊断决策问题等进行深入研究。在一电力企业安全监控系统的应用中．该模型能快速、准确地进行故障成因分析，并给出合理的、建设性的决策意见。与传统的诊断方法相比体现了Agent技术的特有优势。

RESEARCH ON DIAGNOSIS MODEL DISTRIBUTED INTELLIGENCE AND KEY TECHNIQUE BASED ON MAS

Fault diagnosis, a complicated system, is virtually dynamic, distributed, flexible, unfixed and useful introducing multi-agent system to the filed of complicated fault diagnosis is a new attempt to solve its process problems. The paper discusses the methods and process of the distributed intelligence fault diagnosis based on multi-agent system(MAS), To design an agent systematical system or prototype system. In order to do task decompose, distribution of the problem of fault diagnosis. To proceed on further research on inner diagnosis mechanism, multi-agent mutation, cooperation, correlation model and decisive problems of diagnosis, etc., the emphasis on the agent diagnosis representation, and the framework model, and the function & definition of separate task agent and strategies for the concrete analysis. When applied on monitor and controlling system of electric enterprise, the model can offer us a quick and accurate analysis and also give us reasonable, constructive and decisive advice. Compared wish the traditional diagnosis, the model embodies distinguished advantages of multi-agent system. Multi-agent represents efficiencies on the distributed software development.