基于BP神经网络的空空导弹贮存可靠性预测模型研究

针对空空导弹在长期的贮存过程中环境因素对其可靠性的影响,提出了一种基于BP神经网络的建模方法。该方法首先对空空导弹进行分析,确定神经网络的结构;然后对环境信息进行加权融合,形成输入样本,再将检测时刻可靠度的点估计作为输出样本,训练神经网络;当误差达到要求后,根据在导弹贮存期间所采集的环境信息预测导弹当前的可靠度。

Research of Storage Reliability Prediction Model Based on BP Neural Network for Air-to-Air Missile

Aiming at the influence of environment to reliability during the long-term storage of air-to-air missile,a method of construct model based on BP neural network is presented.This method analyzes the air-to-air missile and confirms the structure of its neural network firstly,then trains the neural network with input sample and output sample.The input sample is formed by fusing the environment information with weight and the output sample is the estimate of reliability when the missiles are checked.When the error of training results meets the requirement,the missiles'reliability can be predict with the environment information during storage.