摘要： 提出一种基于人工免疫模型的故障诊断方法。根据免疫系统机理构建模型框架，模拟T细胞和B细胞功用，分别设计模型中的T模块和B模块。T模块采用实向量阴性选择算法生成异常检测器，完成系统的异常状态检测；B模块响应系统实际状态，运用聚类原理动态进化，形成告警信息反馈至T模块。2个模块相互作用，共同实现系统状态的在线实时检测。应用结果表明，该模型具有正确性和有 效性。

关键词: 人工免疫模型, 故障诊断, 阴性选择, 聚类

Abstract: This paper presents a fault diagnosis method using artificial immune model. Using immunological principles, the frame of model is created. The T-module and B-module of the model are designed by simulating functions of T-cells and B-cells. The T-module discriminates anomaly states from the states of the system. The B-module responds the actual states of system and dynamically evolves using clustering principle. The alert information is fed back to the T-module. The interaction of these two modules made the system to achieve on-line detection. This model is successfully applied to rudder fault diagnosis of a certain pilotless. Results confirm the feasibility and effectiveness of this method.

Key words: artificial immune model, fault diagnosis, negative selection, clustering

中图分类号: 

• TN791