基于马氏距离的航空装备故障预测研究  

Study on Aerial Equipment Failure Forecast Based on Mahalanobis Distance

作　　者：辛龙[1] 周越文[1] 孔庆春[1] 杨召[1] 

机构地区：[1]空军工程大学自动测试系统实验室,西安710038

出　　处：《计算机测量与控制》2014年第7期2052-2054,2058,共4页Computer Measurement &Control

摘　　要：故障状态类型的判别是航空装备故障预测系统的核心环节,它直接影响到故障预测的准确性;针对航空装备的故障状态类型判别问题,提出一种基于马氏距离的故障预测方法;首先介绍了马氏距离,其次建立了状态数据库矩阵及状态判别模型,并给出了基于马氏距离的故障预测流程;最后将该方法用于某型飞机火控系统的故障预测中,使得在线和离线的平均故障预测准确度分别达到98.48%和97.77%,表明马氏距离在航空装备的故障预测中有较好的应用和推广价值。Distinguishing failure states is a core technology in aerial equipment failure forecast system, and it influences the accuracy directly. To solve the problem of failure distinction on aerial equipments, a mathod of failure forecast based on Mahalanobis distance is brought out. Firstly, Mahalanobis distance was introduced. Then, the state database and distinguishing model were founded, and the failure forecast process was given. In the end, the method was applied to the failure forecast of a certain airplane fire control system. The result shows that the average precision of on-line forecast is 98.48%, while the one of off-line forecast is 97. 77%. It can be apparently indicated that Mahalanobis distance has a high value in application and it deserves to be spread widely, especially in the area of aerial equipment failure forecast.

关 键 词：马氏距离 航空装备 故障预测 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]