基于粗糙集约简的飞机发电机故障诊断决策研究  

Research of aircraft generator fault diagnostic decision based on attribute reduction in variable precision rough sets

作　　者：荆涛[1] 王家林[1] 石旭东[1] 陈厚合[2] 

机构地区：[1]中国民航大学电子信息与自动化学院,天津300300 [2]东北电力大学电气工程学院,吉林吉林132012

出　　处：《计算机应用研究》2017年第4期1101-1104,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(51377161);中央高校科研基金资助项目(3122013D002)

摘　　要：针对现有飞机发电机故障诊断流程繁琐、诊断结果精确度低的问题,从数据挖掘角度出发,引用变精度粗糙集约简算法处理现有诊断流程系统,为了在小样本条件下更好地提取有效信息,利用对象集定义的二元关系和依赖空间给出了变精度粗糙糙集的β下近似协调集的判定定理。采用了下近似属性约简的飞机发电机故障诊断决策,根据现有诊断决策构造原始诊断决策表,并在小样本条件下按照约简规则约简,结合专家经验构造决策约简表,通过粗糙集中右边界域的准确度和覆盖度验证了该方法的有效性和普适性。Aiming at the complicated process and the low accuracy of diagnostic results in existing aircraft generators fault diagnosis, from the view of data mining,in order to better extract valid information from small samples, the paper introduced a variable precision rough sets to dispose the diagnostic information system. It used the binary relation and dependence space to get judg- ment theorems for judging 13 lower approximation consistent sets. It used aircraft generators fault diagnostic decision based on 13 lower approximation attribute reduction algorithm, constructed the original diagnosis decision table according to the existing diagnosis deeision, and redueed it by the reduction rule in the small samples. Combined with the experience, the decision table was constructed, and the validity of the method was verified. The proposed decision method can be applied to other regions.

关 键 词：飞机整体驱动发电机 数据挖掘 变精度粗糙集 诊断决策 准确度 覆盖度 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]