• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2026, Vol. 62 ›› Issue (7): 15-24.doi: 10.3901/JME.260361

• 特邀专栏:系统工程与数字化 • 上一篇    

扫码分享

基于数据驱动的运载火箭总体回路设计研究

秦曈, 孟庆尧, 魏威, 陈立为, 王紫扬, 李莉   

  1. 北京宇航系统工程研究所 北京 100076
  • 收稿日期:2025-04-17 修回日期:2025-12-23 发布日期:2026-05-25
  • 作者简介:秦曈,男,1983年出生,博士。主要研究方向为运载火箭总体设计、计算流体力学、多学科优化、人工智能辅助设计。E-mail:qinshaomou@sina.com
    魏威(通信作者),男,1986年出生,硕士,研究员。主要研究方向为运载火箭总体设计、数字化转型。E-mail:kunlun109@qq.com

Research on Data-driven General Loop Design of Launch Vehicles

QIN Tong, MENG Qingyao, WEI Wei, CHEN Liwei, WANG Ziyang, LI Li   

  1. Beijing Institute of Astronautical Systems Engineering, Beijing 100076
  • Received:2025-04-17 Revised:2025-12-23 Published:2026-05-25

摘要: 随着运载火箭技术的不断发展,以及高密度、多样化的任务发展需求,对于运载火箭研制和任务分析的方案快速迭代以及设计参数闭环等提出了更高的要求。本文基于MBSE思想,在运载火箭的总体回路设计过程中采用数据驱动的方式实现各专业模型的协同设计,证明了基于模型代替文档的工程可行性。针对总体回路设计中气动特性预示时间长、弹道设计约束多等问题,进一步将试验数据与预示数据相结合,采用人工智能预示和优化算法,大幅提升气动特性模型、弹道计算模型的设计效率,缩短总体回路的设计周期。最终构建了总体协同设计平台,通过调用执行器,实现专业工具模型的调用以及数据在模型间的交互,使得设计数据实现在流程中各专业间的自由传递,又能实时跟踪设计数据的更新与迭代,实现MBSE思想在整个火箭顶层设计回路的“数同源”和“变同知”目标。

关键词: 数据驱动, 总体设计, 基于模型的系统工程, 人工智能

Abstract: With the continuous development of launch vehicle technologies and the growing demand for high-density, diversified mission requirements, higher requirements have been placed on rapid iteration of launch vehicle development and closed-loop design parameter optimization. Based on the principles of Model-Based Systems Engineering (MBSE), this paper adopts a data-driven approach to enable collaborative design of multidisciplinary models during the overall closed-loop design process of launch vehicles, demonstrating the engineering feasibility of model-based processes over document ones. Aiming at addressing issues such as prolonged aerodynamic characteristic prediction time and multiple trajectory design constraints in traditional general loop design, the application of artificial intelligence prediction and optimization algorithms significantly enhances the design efficiency of aerodynamic characteristic models and trajectory calculation models, thereby substantially reducing the design cycle of general loop systems. Through actuator invocation, the platform enables the calling of specialized tools and data interaction between models. This ensures both free flow of design data across disciplines during the process and real-time tracking of design data iterations. These achievements reach the goals of "data homogenization" and "change synchronization" throughout the entire loop of the overall design process.

Key words: data-driven, general design, MBSE, artificial intelligence

中图分类号: