欢迎访问《兵工学报》官方网站,今天是 分享到:

兵工学报 ›› 2023, Vol. 44 ›› Issue (10): 2944-2953.doi: 10.12382/bgxb.2022.0493

• • 上一篇    下一篇

基于支持向量回归模型的弹用冲压发动机性能预测及优化

张宁, 史金光*(), 王中原, 赵新新   

  1. 南京理工大学能源与动力工程学院, 江苏 南京 210094
  • 收稿日期:2022-06-07 上线日期:2023-10-30
  • 通讯作者:

Performance Prediction and Optimization of Ramjet for Projectiles Using Support Vector Regression Model

ZHANG Ning, SHI Jinguang*(), WANG Zhongyuan, ZHAO Xinxin   

  1. School of Energy and Power Engineering, Nanjing University of Science and Technology, Nanjing 210094, Jiangsu, China
  • Received:2022-06-07 Online:2023-10-30

摘要:

为提高弹用固体燃料冲压发动机的工作性能和优化设计效率,提出一种与弹内有限空间适配的发动机性能预测及优化方法。采用带转捩的剪切应力输运和涡概念耗散方程,建立其内弹道计算模型,获得流场结构与性能参数;在此基础上,基于支持向量回归方法构建了发动机性能参数的预测模型,并结合改进非支配排序遗传算法对发动机结构进行优化。研究结果表明:新建立的内弹道计算模型,可较好地模拟发动机内燃烧与流动过程;构建的预测模型可靠性较高,与高可信度模型相比,最大相对误差小于3%;发动机结构优化后,燃烧室缩短了13.88%,补燃室增长了13.50%,燃烧效率、推力和比冲分别提高12.02%、24.22%、20.28%。

关键词: 固体燃料冲压发动机, 代理模型, 多目标优化模型, 性能预测

Abstract:

To improve the working performance of solid fuel ramjet for projectiles and shorten the optimization period, the transitionshear stress transfer and vortex concept dissipation equations are used to establish an internal ballistic calculation model, and the flow field and performance parameters are obtained. Then, based on the support vector regression method,a prediction model of the performance parameters is built, and the ramjet structure is optimized using the NSGA-Ⅱ algorithm. The results show that the internal ballistic calculation model can simulate the combustion and flow process in the ramjet well. At the same time, the constructed prediction model has high reliability, and the maximum relative error is less than 3% compared with that of the high-confidence model. After the ramjet is optimized, the combustion chamber is shortened by 13.88%, the aft mixing chamber is increased by 13.50%, and the combustion efficiency, thrust and specific impulse are increased by 12.02%,24.22% and 20.28%, respectively.

Key words: solid fuel ramjet, surrogate model, multi-objective optimization model, performance prediction

中图分类号: