首页 | 本学科首页   官方微博 | 高级检索  
     


Two dimensional shape optimization using partial control and finite element method for compressible flows
Authors:Shuji Nakajima  Mutsuto Kawahara
Affiliation:1. College of Engineering, Mathematics and Physical Sciences, University of Exeter, Exeter, UK;2. Department of Mathematics, Imperial College, London, UK;1. CWI – Centrum Wiskunde & Informatica, Amsterdam, the Netherlands;2. Faculty of Aerospace Engineering, Delft University of Technology, Delft, the Netherlands;3. DIAM, Delft University of Technology, Delft, the Netherlands;1. Sección de Matemática, Sede de Occidente, Universidad de Costa Rica, San Ramón de Alajuela, Costa Rica;3. Mathematical Institute, University of Oxford, A. Wiles Building, Woodstock Road, Oxford OX2 6GG, United Kingdom;1. School of Mathematical Sciences, Tianjin Normal University, Tianjin 300384, PR China;2. Department of Mathematics & Statistics, Utah State University, Logan, UT 84322, USA;3. School of Materials Science and Engineering, Nankai University, Tianjin 300350, PR China;4. Department of Mathematics, University of South Carolina, Columbia, SC 29028, USA;5. Beijing Computational Science Research Center, Beijing 100193, PR China
Abstract:The objective of this study is to determine the two dimensional shape of a body located in a compressible viscous flow, where the applied fluid force is minimized. The formulation to obtain the optimal shape is based on an optimal control theory. An optimal state is defined as a state, in which the performance function defined as the integration of the square sum of the applied fluid forces is minimized due to a reduction in the applied fluid forces. Compressible Navier–Stokes equations are treated as constraint equations. In other words, the body is considered to have a shape that minimizes the fluid forces under the constraint of the Navier–Stokes equations. The gradient of the performance function is computed using the adjoint variables. A weighted gradient method is used as the minimization algorithm. The volume of the body is assumed to be the same as that of the initial body. In the case of the algorithm used in this study, both the creation of a structured mesh around the surface of the body and the smoothing procedure are employed for the computation of gradient. In this study, a remeshing technique based on the structured mesh around the body changing its configuration in the iteration cycle is employed. For the correction to keep the volume constant, the surface coordinates are moved along the radial direction. For the discretization of both the state and adjoint equations, the efficient bubble function interpolation presented previously by the authors [18] is employed. The algorithm, which is known as the partial control algorithm, is applied to the numerical procedure to determine the movement of the coordinates. In the case of the gradient method, in order to avoid the convergence of the final shape to the local minimum shape, the new algorithm, which is called the partial control algorithm, is presented in this study. In numerical studies, the shape determination of a body in a uniform flow field is carried out in 2D domains. The initial shape of the body is assumed to be an elliptical cylinder. The shape is modified by minimizing the applied fluid forces. Finally, the desired shape of a body, whose performance function is reduced and converged to a constant value, is obtained. By carrying out a procedure that involves the use of the partial control algorithm, the desired shape of a body, whose performance function is reduced further, is obtained. Stable shape determination of a body in a compressible viscous flow is carried out by using the presented method. It is indicated that the optimal shape can be obtained by using the partial control algorithm.
Keywords:
本文献已被 ScienceDirect 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号