基于灰色关联排序聚类方法优化故障诊断  

Optimizing fault diagnosis based on gray incidence ranking and clustering method

作　　者：吴雄学[1] 朱从乔[1] 梅正华[1] 

机构地区：[1]海军工程大学船舶与动力学院,湖北武汉430033

出　　处：《舰船科学技术》2006年第4期103-105,共3页Ship Science and Technology

摘　　要：采用灰色关联分析方法,以舰船监控装置部件的固有可靠性、结构、环境和实际运行等4个因素作为评价指标,结合诊断结果,对其故障原因进行排序和聚类,实现决策优化,较好地解决了专家系统只能提供故障原因集而引发的问题。论文还对各评价指标和相关算法进行介绍,并给出了实例。The fault diagnosis expert system doesn＇t give the exact fault reason because of limit of the on-the-spot condition. To solve this problem, this paper sets up a gray incidence ranking and clustering algorithm based on gray incidence analysis. This algorithm is adopted to sort all fault reason corresponding to failure probability. It gives the direction of seleeting fault reason and improves the direction depth of expert system. The valuation indicator includes inherent reliability, structure factor,environment factor and practice factor. The ranking and clustering of fault reason are designed to adjust according to fault frequency. The valuation indicators and theory, of the algorithm are presented;the actual example for application is given.

关 键 词：专家系统 故障诊断 舰船 灰色关联分析 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]