复合固体推进剂燃烧性能模拟计算的神经网络方法

推进技术 ›› 1996, Vol. 17 ›› Issue (4): 72-76.

复合固体推进剂燃烧性能模拟计算的神经网络方法

邓鹏图,田德余,庄逢辰

国防科技大学材料科学与应用化学系;国防科技大学材料科学与应用化学系;国防科技大学材料科学与应用化学系

发布日期:2021-08-15

A NEURAL NETWORK FOR MODELING CALCULATIONS FOR COMPOSITE PROPELLANTS

Dept. of Material Science and Applied Chemistry, National Univ.of Defense Technology, Changsha, 410073;Dept. of Material Science and Applied Chemistry, National Univ.of Defense Technology, Changsha, 410073;Dept. of Material Science and Applied Chemistry, National Univ.of Defense Technology, Changsha, 410073

Published:2021-08-15

摘要/Abstract

摘要： 在总结已有燃烧模型的基础上，重点考虑压强、氧化剂的重均粒径、氧化剂的质量浓度三种主要影响因素，提出了一种基于误差反传（ＢＰ）神经网络的复合固体推进剂燃烧性能模拟计算方法，计算结果和实验值吻合较好，这为推进剂配方的计算机辅助设计提供了一种新方法。

关键词: 神经元机;复合推进剂;推进剂燃速;计算化学

Abstract: he traditional modeling of composite propellant combustion, which strives to representthe combustion process in mathematical terms, is restricted by the knowledge of the combustionmechanism. In this paper, a new scheme has been proposed to calculate the combustion of compositepropellants without considering the specific combustion processes by applying the back-propagation(BP) neural network, which has the capability to "learn" system characteristics through nonlinearmapping. The computed results are in good agreement with the experimental data, which shows thatthe scheme offers a new way for CAD of propellant formulation.

Key words: Neural machine;Composite propellant;Propellant burning rate;Computational chemistry

邓鹏图,田德余,庄逢辰. 复合固体推进剂燃烧性能模拟计算的神经网络方法[J]. 推进技术, 1996, 17(4): 72-76.