应用并行子空间设计算法对低雷诺数电动无人机进行了初步设计.电动无人机涉及多门学科,包括气动、重量、推进系统、性能和操稳.首先开发了电动无人机多学科系统分析软件.并行子空间设计算法利用基于人工神经网络的响应面技术作为系统分析的近似模型.电动无人机初步设计中的设计变量共有29个,其中连续设计变量25个,离散设计变量4个.研究表明,并行子空间设计算法能以较少的系统分析次数在可行域内寻找出一个较好的设计点,并在设计过程中具有适应设计要求(约束)可变性的要求,它为飞行器一体化设计提供了一条途径.%The concurrent subspace design (CSD) framework has been used to conduct a preliminary design optimization of an electric powered, unmanned air vehicle (EPUAV) operating at a low Reynolds number. A multi disciplinary system analysis that includes aerodynamics, weights, propulsion, performance and stability and control has been developed for this class of vehicles. The CSD framework employs artificial neural network based response surfaces to provide approximations to the design space. The EPUAV system includes 25 continuous and 4 discrete design variables. The CSD framework was able to identify feasible designs with significant weight reductions relative to any previously considered (i. e. initial database) designs. This was accomplished with a limited number of system analyses. The results also demonstrate the nature of this design framework adaptive to changes in design requirements.
展开▼
机译：应用并行子空间设计算法对低雷诺数电动无人机进行了初步设计.电动无人机涉及多门学科,包括气动、重量、推进系统、性能和操稳.首先开发了电动无人机多学科系统分析软件.并行子空间设计算法利用基于人工神经网络的响应面技术作为系统分析的近似模型.电动无人机初步设计中的设计变量共有29个,其中连续设计变量25个,离散设计变量4个.研究表明,并行子空间设计算法能以较少的系统分析次数在可行域内寻找出一个较好的设计点,并在设计过程中具有适应设计要求(约束)可变性的要求,它为飞行器一体化设计提供了一条途径.%The concurrent subspace design (CSD) framework has been used to conduct a preliminary design optimization of an electric powered, unmanned air vehicle (EPUAV) operating at a low Reynolds number. A multi disciplinary system analysis that includes aerodynamics, weights, propulsion, performance and stability and control has been developed for this class of vehicles. The CSD framework employs artificial neural network based response surfaces to provide approximations to the design space. The EPUAV system includes 25 continuous and 4 discrete design variables. The CSD framework was able to identify feasible designs with significant weight reductions relative to any previously considered (i. e. initial database) designs. This was accomplished with a limited number of system analyses. The results also demonstrate the nature of this design framework adaptive to changes in design requirements.
展开▼
