采用弹道修正技术的红外干扰弹性能优化  

Performance optimization of infrared interference decoy based on trajectory correction technique

作　　者：邢炳楠 杜忠华[1] 杜成鑫 XING Bingnan;DU Zhonghua;DU Chengxin(Key Laboratory of Intelligent Ammunition Technology, School of Mechanical Engineering, Nanjing University of Science and Technology, Nanjing 210094, China)

机构地区：[1]南京理工大学机械工程学院智能弹药技术国防重点学科实验室,江苏南京210094

出　　处：《国防科技大学学报》2022年第2期141-149,共9页Journal of National University of Defense Technology

基　　金：国家自然科学基金资助项目(1772160,11372142);江苏省科研创新计划资助项目(SJKY19_0352)。

摘　　要：针对弹道修正弹的高维非线性特性导致的性能优化难题,改变概念设计阶段传统的串行设计方式,提出了一种基于实验设计(Design Of Experiments,DOE)和响应面(Response Surface Methodology,RSM)的智能优化算法,定义基本的弹丸结构模型以及相关的设计参数。在DOE的基础上,将设计变量映射到性能准则,生成Stochastic Kriging响应曲面,并训练神经网络识别不稳定设计。通过多种群遗传算法确定最优弹丸设计,改变其代价函数,可以产生反映性能权衡的最优设计Pareto前沿。仿真结果表明,对于基于弹道修正方案的红外干扰弹的技术改造,所提算法在其概念设计阶段可快速、精确地得到性能要求下的最优设计配置,为伴飞任务的实现提供了保障。In order to solve the performance optimization problem caused by the high-dimensional nonlinear characteristics of the trajactory correction decoy and improve the traditional serial design method in the conceptual design stage,an intelligent optimization algorithm based on DOE(design of experiments)and RSM(response surface methodology)was proposed and the basic projectile structure model and related design parameters were defined.Based on the DOE,the design variables were mapped to the performance criteria to generate the stochastic Kriging response surface,and the neural network was trained to identify the unstable design.Multi population genetic algorithm was used to determine the optimal projectile design.By changing its cost function,the Pareto frontier reflecting the performance tradeoff could be generated.Simulation results show that as for the technical transformation of infrared interference decoy based on trajectory correction,the proposed algorithm can obtain the optimal design configuration quickly and accurately in the conceptual design stage,which can guarantee the following accompanying flight mission.

关 键 词：红外干扰弹 弹道修正 实验设计 Stochastic Kriging 神经网络 多种群遗传算法 

分 类 号：TP43[自动化与计算机技术]