导弹测发控系统故障诊断  

Fault Diagnosis of Test-firing Control System of Missile

作　　者：骆功纯[1] 胡昌华[1] 平振海[1] 张军波[2] 

机构地区：[1]第二炮兵工程学院302教研室,陕西西安710025 [2]第二炮兵装备研究院科研部,北京100085

出　　处：《控制工程》2007年第S3期71-73,共3页Control Engineering of China

摘　　要：针对导弹测发控系统电路原理复杂的特点,采用了一种新的诊断方法-基于粗糙集理论和神经网络的故障诊断。在已知故障模式库的基础上,考虑到各可测点和可观测器件之间的联系与冗余,首先采用粗糙集理论对模式库中的冗余属性进行约简,得到最简样本集,再加入到网络中进行训练,实现故障的诊断定位。与传统的基于神经网络的故障诊断方法相比,该方法对导弹测发控系统故障诊断速度快、定位准确,具有实际应用价值。To the complexity of circuit principle in the missile test-fixing control system,a new method of fault diagnosis based on rough-set theory with neural network is introduced.Considering the relation or redundancy of the accessible nodes and components,on the basis of a known fault pattern base,the simplest sample collection is acquired by redundant attributes' reduction to pattern base with rough-set theory firstly.Then fault diagnosis and location are accomplished by training the collection in neural network.Compared with traditional fault diagno- sis based on neural network,the new method which has the advantages like faster diagnosis and accurate localization is illustrated.

关 键 词：粗糙集 神经网络 导弹测发控系统 故障诊断 

分 类 号：TJ760.6[兵器科学与技术—武器系统与运用工程]