使用流场共轭偏微分方程数值解的气动优化算法研究

A new research field, Numerical Aerodynamic Shape Optimization, bbreviated as NASO, was investigated in this project. NASO bases on Computational Fluid Dynamics and combines it with Nonlinear ptimization Theory and Automatic Grid Generation to optimize the aerodynamic shapes of fluid machines. It can improve the ability of current design system because of the advantages of CFD, and will act as a more and more important role in the future with the development of CFD and Computer Science. In order to make the efficiency of gradient-based system high enough for practical engineering problem, an adjoint method was constructed and the discipline for determining adjoint boundary conditions was presented. The method for numerical solution of adjoint differiential equations was described. A NASO system(soft ware) was established. The results of test-cases demonstrated that the adjoint algorithm can get the gradient of goal function accurately with the computations less than that used by finite difference, especially in the cases whose numbers of design variables are large..

气动造型数值优化作为一个多学科交叉的新兴研究领域，对设计高性能流体机械具有重要意义。该领域研究中出现的共轭算法可快速预测造型改变所引起的气动性能变化。本项目拟开展NS方程的共轭算法研究；并提出了利用共轭信息建立造型气动敏感区的判断算法，由此构造新形式的高效气动造型优化系统；最终得到可快速优化三维粘性流动问题的算法。