基于Kriging代理模型的舰船桅杆雷达隐身优化  

Radar stealth optimization of warship mast based on Kriging surrogate model

作　　者：杜晓佳[1] 朱显明[1] 丁凡[1] 廖章奇[1] DU Xiaojia;ZHU Xianming;DING Fan;LIAO Zhangqi(China Ship Development and Design Center,Wuhan 430074,China)

机构地区：[1]中国舰船研究设计中心,湖北武汉430074

出　　处：《华中科技大学学报（自然科学版）》2020年第10期109-113,共5页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61273241)。

摘　　要：为降低目标雷达隐身优化计算消耗,将Kriging代理模型与自适应遗传算法(AGA)相结合,构建基于代理模型的雷达波隐身优化快速计算方法.以典型舰船封闭式桅杆为雷达隐身优化对象,以评估角域内雷达散射截面积(RCS)均值为优化目标,在相同初始条件下,分别用AGA和AGA-Kriging方法进行100次重复优化计算.结果表明:AGA-Kriging对电磁散射特性计算次数降低56.81%,并保持良好准确性和稳定性;与初始方案相比,优化后舰船桅杆雷达散射截面积均值降低61%,证明了该方法的良好适用性.To reduce the computational cost of target radar stealth optimization,Kriging surrogate model and adaptive genetic algorithm(AGA)were combined to construct a rapid calculation method for radar wave stealth optimization based on surrogate model.The typical closed mast of warship was taken as the optimization object for radar wave stealth,and the average of radar cross section(RCS)in the domain angle was evaluated as the optimization target.Under the same initial conditions,100 repeated optimization calculations were performed by using AGA and AGA-Kriging methods,respectively.Results show that the AGA-Kriging method can reduce the number of target electromagnetic scattering calculation by 56.81%,while maintains good accuracy and stability.Compared with the original design,the radar cross section of the optimized mast is reduced by 61%,verifing the feasibility of Kriging surrogate model.

关 键 词：舰船桅杆 雷达隐身 代理模型 优化 遗传算法 

分 类 号：U662.2[交通运输工程—船舶及航道工程]