Journal of Systems Engineering and Electronics ›› 2010, Vol. 32 ›› Issue (1): 205-210.

• 可靠性 • 上一篇    下一篇

基于历史数据的测试任务约简和故障诊断

方甲永1,肖明清1, 王磊1, 李斌1,2   

  1. (1. 空军工程大学工程学院自动测试系统试验室, 陕西 西安 710038;2. 空军驻江苏地区军事代表室, 江苏 南京 210016)
  • 出版日期:2010-01-23 发布日期:2010-01-03

Test tasks reduction and fault diagnosis based on historical data

FANG Jia-yong1,XIAO Ming-qing1,WANG Lei1,LI Bin1,2   

  1. (1. ATS Lab., Engineering Coll., Air Force Engineering Univ., Xi’an 710038, China;2. The Military Representatives of Air Force Equipment Dept. in Jiangsu, Nanjing 210016, China)
  • Online:2010-01-23 Published:2010-01-03

摘要:

针对部队航电组件测试任务繁重、故障定位率低的问题,提出一种利用历史数据来简化测试任务、提高故障定位率的方法。利用粗糙集信息系统理论,建立了航电组件故障信息系统模型;基于测试任务辨识函数和诊断允许误差对测试任务进行约简;基于最短测试时间选出最优测试任务集;利用贝叶斯最大后验概率进行故障诊断推理,将诊断问题归结为不等式约束极值问题;用0-1规划隐数算法求得最优解。最后以某型飞机惯导部件为例验证了方法的快速有效性。

Abstract:

Aiming at the problem of complex test tasks and low fault location accuracy to avionics, a method based on historical data is proposed to reduce the test tasks and improve the fault location accuracy. A model of the avionics information system is built by using the rough set theory; the test tasks are reduced by using the novel discernibility function and minimum diagnosis error; the optimized test task set with minimum test time is chosen; the Bayes maximal posterior probability is introduced into fault diagnosis reasoning and the problem is concluded as an extremum problem with inequality constraints; the optimized result is got by the 0-1 programming implicit enumeration. The application to the inertial navigation component proves that the method is fast and effective.