摘要
针对某指挥仪电路故障诊断专家系统的不足之处 ,用层次结构的多级人工神经网络进行电路的故障诊断 ,用专家知识对学习样本进行筛选 .通过对该电路中差分电路模块的诊断实验表明 ,可以克服专家系统对某些故障无法诊断的不足 。
This paper introduced a new method to diagnose circuits, that is, multiplayer artificial neural network of layer structure and screening out the training samples by the expert knowledge. It overcame the drawback of the circuit fault diagnosis of a certain missile director in the expert system. The experiment results on the difference circuit module in the system showed that the new method can overcome some faults that expert system cannot diagnose and realize the diagnosis on the multiple faults.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2002年第12期23-25,共3页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
航天科工集团某型号项目资助
关键词
指挥仪
电路故障诊断
神经网络
专家系统
director
circuit fault diagnosis
neural network
expert system
