基于神经网络的导弹惯性器件故障预报系统研究

Missile Inertia Device Fault Prediction System Based on Neural Network Model

在导弹武器系统当中,及时准确地故障预报对提高导弹的安全性具有极其重要的意义。针对导弹惯性器件故障预报系统的设计要求,考虑到神经网络用于故障预报的优点,在神经网络技术应用于导弹惯性器件的故障预报过程中提出了神经网络的训练算法,把神经网络、预测理论和故障诊断系统有机结合起来建立了一个故障预报系统,利用该系统选取神经网络预测模型对某导弹陀螺随机漂移进行建模和分析,实现了故障预报。实例预测结果证明,给出的神经网络预测模型和训练算法是可行的。

In missile weapon system, exact fault prediction is very important for missile security. According to the design requirements of missile inertia device fault prediction system and the advantages of neural network technique, the application of neural network technique in missile inertia device fault prediction is presented, and the algorithm based on the neural network model in the prediction processes is given. The neural network, theory and fault diagnostic prediction system are combined to form a fault prediction system. As an example, a forecasting for random drift rate of gyro applied for strapdown inertial navigation systems is made. The results show that the forecasting method is effective and feasible for modeling and forecasting for gyro drift.