基于Kriging参数优化的液氧系统预冷充填仿真  

Simulation of pre-cooling filling process of liquid oxygen system based on parameter optimization of Kriging

作　　者：张晓光 任孝文 高玉闪 邢理想 ZHANG Xiaoguang;REN Xiaowen;GAO Yushan;XING Lixiang(Xi’an Aerospace Propulsion Institute,China Aerospace Science and Technology Corporation,Xi’an 710100,China)

机构地区：[1]中国航天科技集团有限公司西安航天动力研究所,西安710100

出　　处：《航空动力学报》2024年第12期369-376,共8页Journal of Aerospace Power

摘　　要：针对惰性气体吹除作用下的低温液体火箭发动机液氧系统预冷充填过程,建立了一种两组分两相充填模型。由于包含7个控制方程,模型参数复杂度提升,直接开展参数优化存在优化效率低、参数矩阵非线性强的制约因素。基于Kriging代理模型采用粒子群优化算法,高效寻得两组分两相充填模型在多因素耦合影响下的全局最优参数组合,有效提升了模型的仿真精度。对某型发动机液氧系统预冷充填的仿真研究表明:仿真结果与试验数据的相对误差仅2.75%;喷注器氧容腔内的压力爬升曲线呈现台阶上升特征,且随入口压力的增大液氧充填速率加快。For the pre-cooling filling process of liquid oxygen system of cryogenic liquid rocket engines under the action of inert gas blowing,a two-component two-phase filling model was established.Due to the inclusion of 7 control equations,the complexity of the model parameters increased,and there were constraints on the low optimization efficiency and strong nonlinearity of the parameter matrix for direct parameter optimization.Based on the Kriging surrogate model,the particle swarm optimization algorithm was used to efficiently find the global optimal parameter combination of the two-component twophase filling model under the influence of multi-factor coupling,which effectively improved the simulation accuracy of the model.The simulation study on the pre-cooling and filling of the liquid oxygen system of a certain type of engine showed that the relative error between the simulation results and the test data was only 2.75%.The pressure climb curve in the liquid oxygen dome of injector showed the characteristics of step rise,and the liquid oxygen filling rate accelerated with the increase of inlet pressure.

关 键 词：Kriging代理模型 预冷充填 液氧 粒子群优化 两相 

分 类 号：V434.1[航空宇航科学与技术—航空宇航推进理论与工程]