摘要
针对叶片型面的设计问题,提出一种基于遗传优化算法的多层参数化方法。这种方法类似多层网格法原理,利用Bézier曲线的递推算法进行各层之间的设计变量转化,使得优化迭代过程中,下层群体中得以保存上层的优秀个体。根据遗传算法固有的并行特性构建了局域网并行优化平台,并对基本遗传算法进行了改进,从而大大缩短优化时间、提高优化效率。最后设计了曲线逼近和叶型优化的算例,结果显示多层参数化方法能明显加速收敛,在个体数较少时,效果更为明显。
An adaptive multilevel parameterization algorithm is presented based on genetic optimization algorithm aimed at the design issue on blade profile. According to the principle of multi-grid method, the algorithm uses the recursive algorithm of Bézier curve to transform the design variables and their numbers among up-down levels, saves the prefect individuals from up level to down level in the optimization iterative process. The optimization efficiency is improved by the advancement of simple genetic algorithm, and a local parallel network optimization platform according to the property of genetic algorithm is built, thus leading to less time cost. An example on curve approximation and the other on blade profile optimization are given. Results indicate that the adaptive multilevel parameterization algorithm can accelerate convergence of the iteration, especially, a small number of individuals.
出处
《南京航空航天大学学报》
EI
CAS
CSCD
北大核心
2009年第1期11-15,共5页
Journal of Nanjing University of Aeronautics & Astronautics
基金
空军装备部预研基金资助项目
关键词
气动优化
遗传算法
递推算法
BÉZIER曲线
多层参数化
aerodynamic optimization ； genetic algorithms
recursive algorithm
Bézier curve
multilevel parameterization
