Kriging模型在翼型反设计中的应用研究  

Kriging-based airfoil inverse design

作　　者：刘俊[1] 宋文萍[1] 韩忠华[1] 许建华[1] 樊艳红[1] 

机构地区：[1]西北工业大学翼型叶栅空气动力学国防科技重点实验室,陕西西安710072

出　　处：《空气动力学学报》2014年第4期518-526,共9页Acta Aerodynamica Sinica

基　　金：国家自然科学基金(10902088;11272265);航空科学基金(2011ZA53008)

摘　　要：目前,基于Kriging代理模型的优化方法在气动优化设计中得到了广泛的应用,然而却并未能应用于气动反设计中。本文将翼型气动反设计问题转化为优化问题,利用基于Kriging模型的优化方法进行了翼型的单目标、多目标反设计,验证了Kriging模型在翼型单目标、多目标反设计中的适应性,并研究了模型加点准则对设计结果的影响;此外,将本文方法与多项式响应面模型优化方法、基于Adjoint的优化方法进行了比较,结果表明,该方法明显优于基于多项式响应面的优化方法,且与Adjoint方法相比,设计效率相当且具有更好的通用性。最后将本文方法应用于接近工程实际的在已有翼型基础上修改压力分布进行反设计,验证了本文方法的可行性。Up to date,the Kriging-based optimization methods are widely used in the aerodynamic opti-mization design,however,it hasn’t been successfully applied to the aerodynamic inverse design problems. In this research,the Kriging-based optimization method is applied to the single and multi-point airfoil inverse design,and its feasibility to inverse design problems is validated;besides,the choice of infill criteria is inves-tigated.Moreover,this method is compared to the polynomial response surface based method （P-RSM）and the Adjoint-based method,the results show that the Kriging-based method outperforms the P-RSM,and has comparable efficiency with Adjoint-based method but better flexibility and easier to use.Finally,the current method is applied to engineering inverse airfoil design problems and its feasibility is validated.

关 键 词：KRIGING模型 气动优化设计 翼型反设计 多目标反设计 Navier—Stockes方程 

分 类 号：V211.3[航空宇航科学与技术—航空宇航推进理论与工程]