1 At present, the conventional parameterization methods mainly include B-spline, 6 parametric section (PARSEC), 7 Hick-Henne, 8 class shape transformation (CST), 9 and so on. In the design process of aerodynamic shape, how to reduce the time cost and keep high accuracy has always been one of the difficult problems for designers.įirst of all, the parameterization method has a significant effect on the optimization space and calculation efficiency during the optimization design process. However, it still consumes plenty of computational resources and the numerical calculation results are difficult to be reused. It improves the design efficiency to a certain extent and avoids some financial and personnel restrictions caused by a wind tunnel test. With the improvement of computing ability and the development of numerical simulation technology, the computational fluid dynamics (CFD) method was introduced into the airfoil design process. Such a process takes high time cost and consumes lots of resources. In the past, aerodynamic design for airfoil was based on prior experience and the wind tunnel test. 1 Aerodynamic design for airfoil is a crucial step in aircraft design and has an essential influence on the aerodynamic performance and safety. With the vigorous development of the aerospace industry, higher requirements for aircraft design capability have been proposed. Overall, the DPCNN framework in this paper has outstanding advantages in time cost and prediction accuracy. Finally, the optimization attempt of operating parameters is completed by gradient descent method, which shows good stability and convergence. The prediction absolute errors of physical field of most sample points are lower than 0.002, and the relative errors of aerodynamic performance parameters are lower than 2.5%. When the train size is 0.1, the predicted results can be obtained within 5 ms. It has significant advantages such as good robustness, great convergence, fast computation speed, and high prediction accuracy compared with the conventional machine learning method. The results show that the DPCNN framework can generate substantial perfect airfoil profiles with only three geometric parameters. The airfoil profile parameterization, physical field prediction, and performance prediction are achieved. Aiming at the problems of a long design period and imperfect surrogate modeling in the field of airfoil design optimization, a convolutional neural network framework for airfoil design and performance prediction (DPCNN) is established based on the deep learning method.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |