摘要
介绍了如何利用神经网络对复杂导弹系统可靠性进行预测。对利用导弹地面测试数据和可靠性数据预测未来导弹可靠性问题进行了描述。给出了建模思路和方法框架。探讨了基于因子分析的可靠性数据特征选择方法、循环神经网络(RNN)、长短时记忆网络(LSTM)、门循环网络(GRU)以及输出数据融合方法。在实例中利用该方法对其飞行可靠性进行了预测,验证了方法的可行性与有效性。该方法可以充分利用复杂系统地面测试数据、可靠性数据中隐含的信息以及多种神经网络的优点,具有一定的通用性与实用性。
This paper introduces how to use neural network to predict the reliability of complex missile systems.First,using missile ground test data and reliability data to predict future missile reliability issues is described.Then,modeling ideas and the method framework are given.Then,the reliability data feature selection method based on factor analysis,cyclic neural network(RNN),long and short time memory network(LSTM),gate loop network(GRU),and output data fusion method are discussed.Finally,the flight reliability is predicted by the method proposed in this paper,and the feasibility and effectiveness of the method are verified.The method can make full use of the ground test data of the complex system,the information hidden in the reliability data,and the advantages of various neural networks,and has certain universality and practicability.
作者
苏续军
吕学志
方丹
陈晓
SU Xu-jun;LYU Xue-zhi;FANG Dan;CHEN Xiao(Shijiazhuang Campus of Army Engineering University,Shijiazhuang 050003,China;Unit 32179 of PLA,Beijing 100012,China)
出处
《火力与指挥控制》
CSCD
北大核心
2020年第5期170-177,共8页
Fire Control & Command Control
基金
国家自然科学基金(U1435218)
中国博士后科学基金(2017M613360)
军内装备科研基金资助项目(012016012600B12507)。
