Airfoil optimization design based on Gaussian process regression and genetic algorithm
-
摘要: 针对高升阻比风力机翼型前缘曲率半径较大的问题,传统的翼型参数化方法前缘控制能力不足,且基于面元法XFOIL预测精度差的问题,利用增强类函数/形函数转换(CST)参数化方法控制翼型的形状变化、拉丁超立方实验设计、计算流体力学(CFD)流场计算模块、高斯过程回归模型和遗传算法,提出了基于高可信度Reynolds average Navier-Stocks(RANS)和高斯回归模型辅助遗传算法的翼型优化设计方法。结果表明:基于高斯回归模型的翼型优化方法,可以将优化所用CFD计算次数降低一阶,从而大幅度提升优化设计效率。由标准算例超临界翼型RAE2822的降阻设计表明,在百次量级的CFD次数阻力降低43.16%,激波被削弱且升力、力矩和面积严格满足约束。由风力机翼型NACA64618的最大化升阻比优化设计表明,所设计翼型不仅在设计攻角和副设计攻角处升阻比大大增加,在整个小攻角范围内其气动性能都得到了提升,且两个主设计点,无不良阻力的产生。
-
关键词:
- 高斯过程回归(GPR) /
- CST参数化方法 /
- 遗传算法 /
- 贝叶斯优化 /
- 翼型设计
Abstract: In view of the problems that the leading edge curvature of the wind turbine airfoil with high lift-to-drag ratio has large radius,the traditional airfoil parameterization method has insufficient leading edge control ability,and there exists poor prediction accuracy based on the panel method XFOIL,an enhanced class function/shape function transformation (CST) parameterized method was used to control the shape change of airfoil,Latin hypercube experimental design,computational fluid dynamics (CFD) flow field calculation module,Gaussian process regression model and genetic algorithm,based on high-confidence Reynolds average Navier -Stocks (RANS) and Gaussian regression model-assisted genetic algorithm for airfoil optimization design method.The results showed that the airfoil optimization method based on the Gaussian regression model can reduce the number of CFD calculations for optimization by one order,thereby greatly improving the optimization design efficiency.The resistance reduction design of the supercritical airfoil RAE2822 of the standard calculation example showed that the resistance of the CFD frequency in the order of hundreds of times was reduced by 43.16%,the shock wave was weakened and the lift,moment and area could strictly meet the constraints.The maximum lift-to-drag ratio of the wind turbine airfoil NACA64618 showed that the designed airfoil not only greatly increased the lift-to-drag ratio at the design angle of attack and the secondary design angle of attack,but also improved its aerodynamic performance in the entire range of the small angle of attack.And there were two main design points,without bad resistance. -
[1] 安启蒙,黄典贵,吴国庆.基于变分有限元方法的S809翼型的改进设计[J].工程热物理学报,2011,32(5):779-782. [2] 何磊,钱炜祺,刘涛,等.基于深度学习的翼型反设计方法[J].航空动力学报,2020,35(9):1909-1917. [3] RITLOP R,NADARAJAH S.Design of wind turbine profiles via a preconditioned adjoint-based aerodynamic shape optimization[R].AIAA 2009-1547,2009. [4] 白井艳,杨科,李宏利,等.水平轴风力机专用翼型族设计[J].工程热物理学报,2010,31(4):589-592. [5] 陈进,张石强,王旭东,等.基于粗糙度敏感性研究的风力机专用翼型设计[J].空气动力学学报,2011,29(2):142-149. [6] 汪泉,王君,陈进,等.基于多点攻角的风力机翼型优化设计[J].哈尔滨工程大学学报,2016,37(11):1580-1585. [7] BARRETT R,NING A.Comparison of airfoil precomputational analysis methods for optimization of wind turbine blades[J].IEEE Transactions on Sustainable Energy,2016,7(3):1081-1088. [8] 闵新勇.风力机翼型的气动性能模拟与优化设计[D].上海:上海交通大学,2010. [9] 余莉,呼政魁,程涵,等.风力机翼型的多学科设计优化[J].南京航空航天大学学报,2011,43(5):697-700. [10] 琚亚平,张楚华.利用试验设计法建立翼型气动特性的人工神经网络模型[J].航空学报,2010,31(5):893-898. [11] 黄靓,李景银,高远.基于响应面法的风力机翼型气动优化设计[J].流体机械,2011,39(2):21-24. [12] 杨硕,程珩.基于径向基函数模型的风力机翼型多目标优化设计[J].现代制造工程,2016(3):54-58. [13] 王邦祥,陆金桂,王京涛.神经网络在风力机翼型气动性能优化中的应用[J].机械设计与制造,2020,349(3):236-240. [14] 单志辉,刘学军,吕宏强.高斯过程回归在翼型气动性能快速评估中的应用[C]∥ 第4届中国航空学会青年科技论坛文集 ,沈阳:中国航空学会,2010:159-166.[15] 徐亚峰.基于CST参数化方法的飞机翼型快速设计研究[D].南京:南京航空航天大学,2013. [16] 高赫,刘学军,郭晋,等.基于高斯过程回归的连续式风洞马赫数控制[J].空气动力学学报,2019,37(3):480-487. [17] 韩忠华.Kriging模型及代理优化算法研究进展[J].航空学报,2016,37(11):3197-3225. [18] GARDNER J R,KUSNER M J,XU Z E,et al.bayesian optimization with inequality constraints[R].Beijing:International Conference on Machine Learning,2014:937-945. [19] GELBART M A.Constrained Bayesian optimization and applications[D].Boston:Harvard University,2015. [20] RASMUSSEN C E,WILLIAMS C K I.Gaussian processes for machine learning[M].Cambridge,US:Massachusetts Institute of Technology Press,2005. [21] KULGAN B M.Universal parametric geometry representation method[J].Journal of aircraft,2008,45(1):142-158. [22] 李静,高正红,黄江涛,等.基于CST参数化方法气动优化设计研究[J].空气动力学学报,2012,30(4):443-449. [23] 卜月鹏,宋文萍,韩忠华,等.基于CST参数化方法的翼型气动优化设计[J].西北工业大学学报,2013,31(5):829-836. [24] KULFAN B M.Modification of CST airfoil representation methodology[EB/OL].[ 2020-09-01].https:∥brendakulfan.com/docs/CST8.pdf. [25] PIEGL L,TILLER W.The NURBS book[M].Berlin Heidelberg:Springer,1997. [26] 刘俊.基于代理模型的高效气动优化设计方法及应用[D].西安:西北工业大学,2015. [27] GRAMANZINI J R.Adjoint-based airfoil shape optimization in transonic flow[D].Missouri,US:Missouri University of Science and Technology,2015. [28] 陈立立,郭正,侯中喜.自由变形技术在RAE2822翼型优化设计中的应用[J].国防科技大学学报,2018,40(5):45-53. [29] 朱莉.基于神经网络的翼型优化设计方法研究[D].西安:西北工业大学,2007. [30] 苏伟,白俊强.一种代理模型方法及其在气动优化设计中的应用[J].弹箭与制导学报,2008,28(3):199-202. [31] 王超,高正红,张伟,等.自适应设计空间扩展的高效代理模型气动优化设计方法[J].航空学报,2018,39(7):45-63.
点击查看大图
计量
- 文章访问数: 247
- HTML浏览量: 9
- PDF量: 163
- 被引次数: 0