HTPB推进剂装药工艺研究及力学性能预测

HTPB propellant charging process and prediction of its mechanical properties

关键词:  HTPB推进剂  力学性能  关键工艺  神经网络  性能预测

Key words:  HTPB propellant  Mechanical properties  Key process  Neural network  Performance prediction

摘要:

固体发动机药柱成型过程中，制造工艺对推进剂力学性能有很大影响，通过对130组推进剂装药工艺参数及其力学性能研究，确定了装药过程的工艺参数波动情况；基于多元线性回归分析方法，获得了影响推进剂力学性能的关键工艺参数；以关键工艺参数为输入，建立了BP神经网络并对推进剂的力学性能进行预测。研究结果表明，16个工艺参数对推进剂力学性能有不同程度的影响，其中固化参数、混合时间、混合温度、硫化时间、混合物相对湿度、混合压强等6个关键工艺参数对推进剂常温的力学性能有显著影响；以关键工艺参数及对应推进剂力学性能为依据，建立BP神经网络能够准确地对推进剂的力学性能进行预测，最大抗拉强度的平均误差为4.08%，最大伸长率的平均误差为3.54%。

Abstract:

The manufacturing technology has a great influence on the mechanical properties of solid propellant of solid rocket motor during the molding process. Through the study of the charging process parameters and mechanical properties of 130 pots propellant， the fluctuation of the charging process parameters was determined. Based on the multiple linear regression analysis method， the key process parameters affecting the mechanical properties of the propellant were obtained. Taking the key process parameters as input， a BP neural network is set up and the mechanical properties of the propellant are predicted. The research results show that 16 parameters have different degrees of influence on mechanical properties of propellant， six key process parameters such as the curing parameters， mixing time， mixing temperature， vulcanizing time， relative humidity， mixing pressure and so on have significant effect on the mechanical properties of the propellant temperature. Based on the key process parameters and the corresponding mechanical properties of the propellant， the BP neural network was established to accurately predict the mechanical properties of the propellant. The average error of the maximum tensile strength was 4.08%， and the average error of the maximum elongation was 3.54%.