面向数据挖掘的静态电源综合故障诊断研究  

Research of Data Mining Oriented Integrated Error Diagnosis of Solid State Power

作　　者：孙毅刚[1] 曲睿[1] 陈维兴[1] 王慧敏[1] 

机构地区：[1]中国民航大学航空自动化学院,天津300300

出　　处：《计算机测量与控制》2015年第10期3274-3276,3280,共4页Computer Measurement &Control

基　　金：中央高校基本科研业务基金(20001843)

摘　　要：为了对故障诊断提供方向和目标,最终以降低经济损失最大化为目的;在基于桥载设备的安监系统上,对静态电源的故障进行强关联和综合诊断预测,首先将聚类和关联算法结合得到了一次强关系矩阵,然后再经过二次关联加载,得到二次强关系矩阵,最终通过二次强关系矩阵,对故障进行综合诊断;通过数据挖掘软件和在线数据库测试结果表明:二次强关系矩阵能够明显的反映出静态电源故障在实际作业中显著的特点,解决了实际中故障变量和故障模式之间模糊的关系,能够对静态电源的未来状态进行有效的预测。For providing error solution with direction and goal, it gets the purpose that economic losses can be reduced to maximization. Based on safe monitoring system of the bridge load equipment, it can come true that strong relation and integrated diagnostics and forecasting system. Firstly, the first strong relation matrix can be concluded by cluster and association algorithm. Furthermore, the second strong relation matrix can be obtained by loading on the second association. Last, it can be made the integrated diagnose. By the data mining software and the online database testing, the second strong relation matrix can reflect obviously the characteristics on the actual procedure, solve the fuzzy relation in error variable and error model and predict the future state effectively.

关 键 词：静态电源 关联分析 聚类分析 诊断 预测 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]