改进关联规则挖掘算法航空发动机故障检测  

The Fault Detection of Aero-engine Based on the Improved Association Rules Mining

作　　者：芮少辉[1] 张凤鸣[1] 徐显亮[1] 

机构地区：[1]空军工程大学,西安710038

出　　处：《火力与指挥控制》2011年第9期199-202,共4页Fire Control & Command Control

基　　金：国家"863"高技术计划基金资助项目(2007AAJ127)

摘　　要：通过对航空发动机机载数据结构以及几种常用的航空发动机故障检测方法的深入分析,提出采用关联规则挖掘来实现航空发动机故障检测。针对常用的关联规则挖掘算法——Apriori在应对数据量较大的数据库时存在效率瓶颈的问题,对该算法进行了改进。改进后的算法可以不断降低数据库规模和候选项集的数量。通过对航空发动机实际试车数据的挖掘实验,证明了改进算法更加高效简洁。Through deeply researching on the whole structure of the aero-engine data and some popular ways in the aero-engine fault detection,the advice to apply the association rules mining on aero-engine fault detection is raised.However,the classical algorithm of association rules mining——Apriori has the bottleneck of efficiency when dealing with the huge database,so an improved and faster mining algorithm,which can decrease the database scale and reduce the amount of candidate aggregate constantly,is put forward.The improved algorithm is proved to be feasible and effective through the mining in the real data from the aero-engine＇s trial runs.

关 键 词：数据挖掘 关联规则 APRIORI算法 故障检测 

分 类 号：TN956[电子电信—信号与信息处理] O221.6[电子电信—信息与通信工程]