基于多种群协同进化免疫多目标优化算法的百叶窗优化研究  

Research on Optimization of Louvered Fin with Hybrid Multi-objective Optimizations Based on EDA and AIS

作　　者：骆清国 赵耀 桂勇 刘红彬 帅刚 LUO Qingguo;ZHAO Yao;GUI Yong;LIU Hongbin;SHUAI Gang(Department of Vehicle Engineering,Army Academy of Armored Forces,Beijing 100072,China;Military Representative Office of the Chinese People s Liberation Army in No.627 Factory,Xiangtan 411100,Hunan,China)

机构地区：[1]陆军装甲兵学院车辆工程系,北京100072 [2]中国人民解放军驻627厂军事代表室,湖南湘潭411100

出　　处：《兵工学报》2019年第4期689-696,共8页Acta Armamentarii

基　　金：国防"十三五"预先研究项目(30105080101)

摘　　要：为降低空气流通阻力、增加空气流量以及获得最大的防护能力,对装甲车辆进气和排气百叶窗结构优化进行研究。建立装甲车辆进气和排气百叶窗数值计算模型,根据冷却系统工作特点搭建冷却风道半实物仿真模型,对比分析优化前的仿真结果和台架测试结果,并给出了优化目标函数表达式。为解决优化效率低、计算量大的问题,确定设计变量并建立设计变量与目标函数之间关系的椭圆基神经网络替代模型。通过分析目标函数随设计变量的变化规律,采用基于多种群协同进化免疫的多目标优化算法对模型进行优化,确定了最终解。利用自组织神经网络对样本点进行数据挖掘,找到了内在规律。研究结果表明,优化后的进气和排气百叶窗在提高防护能力的同时,提高了散热性能。The structure optimization of intake and exhaust louvers of armored vehicle is studied to reduce air flow resistance and increase air flow. A numerical calculation model is established for the intake and exhaust louvers of armored vehicle. According to the operating characteristics of cooling system, a hardware-in-the-loop test-rig of air-cooling duct is set up. The simulated and test results of the duct before and after optimization are compared and analyzed, and the expression of optimization objective function is given. The design variables are determined. In order to solve the problem of low efficiency and large computation, an elliptic basis neural network agent model is established for the relationship between design variables and objective functions. The hybrid multi-objective optimization based on estimation of distribution algorithm (EDA) and artificial immune systems is used to find a final solution by analyzing the variation of objective function with the design variables. The inherent law of sample points is found by using self-organizing neural network for data mining. The research results show that the optimized inlet and exhaust louvers improve the performance of heat dissipation while improving the protection capability.

关 键 词：装甲车辆 动力舱 百叶窗翅片 多目标优化 试验设计 

分 类 号：TJ810.2[兵器科学与技术—武器系统与运用工程]