论文部分内容阅读
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 2 9个 ,其中连续设计变量 2 5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 application of parallel subspace design algorithm for low Reynolds number electric UAV preliminary design. Electric drone involves many disciplines, including aerodynamics, weight, propulsion system, performance and stability. First developed a multi-disciplinary system analysis software for UAVs. Parallel subspace design algorithm uses response surface technology based on artificial neural network as an approximate model for system analysis. In the preliminary design of electric UAV, there are 29 design variables, of which 25 are continuous design variables and 4 are discrete design variables. The research shows that the parallel subspace design algorithm can find a better design point in the feasible domain with fewer times of system analysis and has the requirements of adapting to the design requirements (constraints) variability in the design process. Design provides a way.