文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > Bivariate spline method for numerical solution of the steady state Navier-Stokes equations
侵权投诉

Bivariate spline method for numerical solution of the steady state Navier-Stokes equations

Abstract. We use the bivariate spline method to solve the steady state NavierStokes equations numerically. The bivariate spline we use in this paper is the space of splines of smoothness r and degree 3r over triangulated quadrangulations. The stream functi

4 for the Navier-Stokes equations. An advantage of using the stream function formulation instead of the traditional velocity-pressure formulation and stream function-vorticity formulation is that the system of nonlinear equations has much smaller size. Indeed, let N, M, and L be the dimension of the C 0 quadratic nite element space, the bivariate spline space to be introduced below, and the linear nite element space, respectively. In the velocity-pressure formulation (cf. Temam'77]), one needs 2N unknown to approximate the velocity vector and L unknowns for the pressure function and thus, the system of nonlinear equations is of size (2N+ L) (2N+ L). In the stream function-vorticity formulation, one needs N unknowns to approximat

e the stream function and N unknowns to approximate the vorticity function and hence, the size of the system is 2N 2N . In the stream function formulation, the size of the system is M M . Since the bivariate splines have a much better approximation power, one may choose M<< N to get the same accuracy. Since the system is nonlinear, we have to use Newton's iterative method and hence, the linearized system has to be solved many times. Thus, the smaller the size of the system, the more robust the uid ow simulation becomes. Since we are going to solve the Stokes and Navier-Stokes equations in the stream function formulation, higher order nite elements are de nitely needed. It is known that setting up a linear system using higher order nite elements takes more time than that using linear or quadratic nite elements. However, we should note that the time for setting up a linear system using the bivariate spline functions is linearly increasing as the number of the triangles in the underlying triangulation while the time for solving a linear system is cubically increasing (direct methods) or quadratically increasing (iterative methods). Since the spline functions usually give a much better approximation than linear and quadratic nite elements, for the same tolerance the spline method requires a smaller number of triangles in a triangulation of the underlying polygonal domain than the linear and quadratic nite element method. Hence, for su ciently small tolerances the total time for computing numerical solutions using the spline method will be less than the linear or quadratic nite element method. Refer to Tables 1a, 1b, 2a and 2b, for the accuracy which the spline method can achieve. Once we compute the stream function, we then solve for the pressure function. It should be noted that as the weak form of the boundary conditions for this problem involve second order derivatives of the stream function, a high order approximation of the stream function is necessary. Refer to Table 4 for the accuracy of the pressure function approximations by the standard linear nite element method. Let us now introduce the bivariate spline space to be used to approximate the stream functions for the Stokes and Navier-Stokes equations. Let} be a quadran-

上一页第4页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email