摘要
导弹测试系统电路原理复杂,分析排除故障主要靠人的经验,而人员的流动性又大,不便积累经验。为适应现代化作战需要,增强导弹部队的生存能力,实现及时、快速排除故障。该文根据电路原理及领域专家经验,按导弹地面测试的步骤,逐个分析可能出现故障(指示灯、测试仪表等),建立了导弹测发控系统故障诊断知识库,把BP(back-propagation)神经网络应用于该系统的故障诊断,从而使得诊断、排除故障更加方便、快捷,解决了人员排除故障的困难,又便于知识的扩充与删除,更有利于基层部队排除故障,提高我导弹部队作战能力。
The diagnosis and elimination of faults mainly depend on teehnieian＇s experience. Because of the complexity of circuit principle in the missile test system, the quick flowage of technician is not beneficial for the accumulation of experience. For adapting the need of modern war, strengthening armyg survivability and elimination faults timely, in the paper, we analyze the possible faults （including indicators, test instruments and so on ） one by one according to the circuit principle, domain expert＇s experience and the step of the missile test, then build the knowledge base of fault diagnosis in missile test, launch and control systems and apply BP（ back -propagation）neural network to the systems. It makes the diagnosis and elimination more convenient and quicker and solves the problems of eliminating faults by manpower. This application can advance army＇s battle ability.
出处
《计算机仿真》
CSCD
2005年第7期18-21,共4页
Computer Simulation
关键词
导弹测试
故障诊断
冲经网络
Missile Test
Fault Diagnosis
Neural Network
