三阶灰色神经网络模型的建立及其在导弹故障预报中的应用

Construction of GNNM(3,1) and its application in missile fault prediction

为了解决导弹武器故障数据量小,无法利用现有的方法对其进行故障预报的问题,在分析灰色预测模型与BP网络算法的基础上,对现有的一阶灰色神经网络模型做了进一步优化和完善,并在此基础上建立了基于时间响应函数的三阶灰色神经网络GNNM(3,1)模型,推导出了模型的理论算法。将此算法应用于导弹惯性器件故障预报中,通过实例证明了GNNM(3,1)模型对导弹武器故障预报的有效性。

The data for missile faults is limited in quantity, thus it is difficult to realize missile fault prediction with existed methods. Based on the analysis to grey prediction model and BP net work algorithm, the first - order grey neural network model（GNNM （1,1）） is optimized and improved. Then the third- order GNNM （GNNM（3,1） ） based on time response function is constructed, and the theoretical algorithm for the model is deduced out. The algorithm has been used in fault prediction of the inertial device of missile. Practical examples proved that GNNM（3,1 ） is effective in fault prediction for missiles.