首次将反向传输(BP)神经网络模型和经遗传算法(GA)优化后的BP(GA-BP)网络模型应用于研究多种工艺助剂和不同测试温度对复合固体推进剂力学性能的影响。结果表明:所建模型对复合固体推进剂力学性能均具有一定的预测能力,通过对未参与建模的5组数据的预测计算,GA-BP模型的抗拉强度预测的平均相对误差由BP模型的15.66%降低至12.86%,断裂伸长率由BP模型的11.24%降低至3.9%,预测精度得到了提高。
For the first time,the back propagation neural network(BP neural network) model and BP network model optimized by genetic algorithm (GA) were applied to study the effects of various processing aids and different test temperatures on the mechanical properties of composite solid propellant.The results showed that the model had a certain ability to predict the mechanical performance of composite solid propellant.Through the prediction of five groups of data that did not participate in the modeling calculation,it was found that the average relative error of tensile strength of prediction of BP model before and after the optimization of GA decreased from 15.66% to 12.86%,their elongation at break decreased from 11.24% to 3.9%,the prediction precision was improved.
[1] 庞维强,张教强,刘何,等.醇胺类键合剂在丁羟推进剂中的应用进展[J].化学推进剂与高分子材料,2005,3(6):10-13.
[2] Ozkaya B,Demir A,Bilgili M S.Neural network prediction model for the methane fraction in biogas from field-scale landfill bioreactors[J].Environmental Modeling & Software,2007,22(6):815-822.
[3] 刘佩进,吕翔,何国强.基于人工神经网络的燃速相关性研究[J].推进技术,2004,25(2):156-158.
[4] 于莹莹,陈燕,李桃迎.改进的遗传算法求解旅行商问题[J].控制与决策,2014,29(8):1483-1488.
[5] Cheng H,Chao F.Simulation of fluid-solid interaction on water ditching of an airplane by ALE method[J].Science Direct Journal of Hydrodynamics,2011,23(5):637-642.
[6] 巩敦卫,朱美强,郭西进,等.一种新的基于混沌变异解决早熟收敛的遗传算法[J].控制与决策,2003,18(6):686-689.
[7] 彭基伟,吕文华,行鸿彦,等.基于改进GA-BP神经网络的湿度传感器的温度补偿[J].仪器仪表学报,2013,34(1):153-160.
[8] 刘奕君,赵强,郝文利.基于遗传算法优化BP神经网络的瓦斯浓度预测研究[J].矿业安全与环保,2015,42(2):56-60.
PDF
(1429KB)
