基于人工免疫算法的Multi-Agent故障监测诊断系统  

The Multi-Agent Fault Detecting and Diagnosis System Using Real-Value Negative Selection

作　　者：武兵[1] 熊诗波[1] 林健[2] 

机构地区：[1]太原理工大学机械电子工程研究所,山西太原030024 [2]太原理工大学网络中心,山西太原030024

出　　处：《太原理工大学学报》2008年第3期257-260,共4页Journal of Taiyuan University of Technology

基　　金：国家自然科学基金重点资助项目(50335030)

摘　　要：根据复杂机电设备的分布式、自动化、集成化高的特点,将Multi-Agent技术应用于复杂机电设备的故障监测,提出了一个基于层次模型的监测系统构架。各监测Agent中采用实值反向选择免疫算法,对设备运行中的异常情况进行识别。应用该系统对某选煤厂电机系统进行在线监测,提高了系统监测的智能化、自适应能力,并具有良好的可扩展性。The complex machine systems have some distinguishing characteristics as distributed, automatically and high-level integrated. The features of Multi-Agent systems （MASs） have been proved to fit those characteristics. And the artificial immune systems （AISs） have similar features. This paper proposed a AIS-MAS fault detecting system for the main high-power motor which works in a coal preparation factory. The system has a hierarchical structure including fault detecting agent, collaborating agent and interface agent. In each agent, it can be regarded as a independent immune system. The real-value negative selection algorithm is employed for the abnormal detection. As a result, the fault detecting system is proved that it have the features of intelligentization, adaptive ability and better extensibility.

关 键 词：故障监测 MULTI-AGENT系统 人工免疫系统 实值反向选择算法 

分 类 号：TP266[自动化与计算机技术—检测技术与自动化装置]