摘要
将均匀设计与二次响应面相结合 ,发展了一种高效实用的飞机多学科设计的并行子空间优化算法。采用均匀设计给出初始样本点 ,为近似模型提供有效的精确模型信息 ,简单实用的二次响应面作为一种在计算量和精确度上达到合理折衷的近似方法为系统级和学科级的优化模型提供状态变量的近似信息。将该方法用于某通用航空飞机的概念设计 ,考虑气动、重量和性能 3个学科 ,很好地解决了各学科间复杂耦合带来的计算困难 ,证实了该算法的有效性。
We aim to make Concurrent Subspace Optimization (CSSO) more efficient and practical for aircraft conceptual design by combining Uniform Design (UD) with quadratic polynomial Response Surface Method (RSM). In this paper, we explain in much detail how to exploit UD and quadratic polynomial RSM fully to make CSSO more efficient and practical. Here we give only a briefing of our explanation. UD is used to improve the uniformity of initial sample points, and RSM is employed to provide reasonably accurate models approximating the relationships between state variables and design variables. A conceptual design for general-aviation aircraft was performed to test our improved CSSO. This conceptual design problem was divided into three disciplines, including aerodynamics, weight and performance discipline, and the optimization models were set up respectively. Each optimization model was much smaller than the original one and could be performed concurrently. Calculations in our conceptual design of general-aviation aircraft show that the occasional difficulty——failing to converge due to non-uniform initial sample points——when using CSSO simply did not appear; our method unceasingly added new sample points that approached the optimum more closely to make response surface attain higher precision; thus our multidisciplinary integrated optimization method for aircraft conceptual design does appear to be more efficient and practical.
出处
《西北工业大学学报》
EI
CAS
CSCD
北大核心
2005年第1期102-106,共5页
Journal of Northwestern Polytechnical University
基金
国家自然科学基金 (10 3770 15 )资助
关键词
并行子空间优化
响应面
均匀设计
飞机概念设计
multidisciplinary design optimization, aircraft conceptual design, Concurrent Subspace Optimization (CSSO), Response Surface Method (RSM), Uniform Design (UD)
