欢迎访问《图学学报》 分享到:

图学学报

• 计算机图形学与虚拟现实 • 上一篇    下一篇

对流-扩散方程数值解的四次 B 样条方法

  

  1. (河海大学理学院,江苏 南京 211100)
  • 出版日期:2020-10-31 发布日期:2020-11-05
  • 作者简介:齐梓萱(1995?),女,河北保定人,硕士研究生。主要研究方向为数值逼近与计算几何。E-mail:qiqizixuanxuan@163.com
  • 基金资助:
    江苏省自然科学基金青年基金项目(BK20160853);河海大学中央高校科研业务费项目(2019B19414)

Quartic b-spline methods for numerical solutions convective-diffusion equations

  1. (College of Science, Hohai University, Nanjing Jiangsu 211100, China)
  • Online:2020-10-31 Published:2020-11-05
  • About author:QI Zi-xuan (1995–), female, master student. Her main research interests cover numerical approximation and computational geometry. E-mail:qiqizixuanxuan@163.com
  • Supported by:
    Jiangsu Natural Science Foundation Youth Fund Project (BK20160853);Hohai University Central University Scientific Research Operating Expense Project (2019B19414)

摘要: 基于四次 B 样条函数,提出一种求解一类对流-扩散方程的四次 B 样条方法。首 先利用光滑余因子协调法,得到有界闭区间上具有均匀节点的一元四次 B 样条基函数表达式。 接着计算在有界闭区间两端点处具有重节点的几种不同情况下的 B 样条基函数表达式,这些样 条基函数具有非负性、单位分解性等良好的性质。然后将一元四次 B 样条函数应用于求解一类 一维对流-扩散方程,其中对于对流-扩散方程的离散过程,对于时间变量的离散采用向前有限 差分,而对于空间变量的离散,引入参数 δ,建立四次样条逼近格式。之后利用四次 B 样条函 数去求解该对流-扩散方程。最后通过具体算例,将四次样条逼近方法与有限差分方法进行比较, 且给出直观的数值误差对比,由此说明样条逼近方法更加简便实用。

关键词: 光滑余因子协调法, 四次 B 样条, 对流-扩散方程, 有限差分法, 微分方程数值解

Abstract: Based on the quartic B spline function, a quartic B spline method for solving a class of convection-diffusion equations was presented. By means of the conformality of smoothing cofactor method, univariate quartic B spline bases were firstly obtained over uniform knots. Secondly, representations of B splines were calculated at the endpoints with multiple knots on the bounded closed interval. In addition, all the quartic B spline bases possessed good properties, such as non-negative property and partition of unity. Thirdly, univariate quartic B spline functions were applied in solving a class of one-dimensional convection-diffusion equations, where discrete in time was realized by forward finite differences and discrete in space was by quartic spline approximation with the parameter δ introduced. Then the convection-diffusion equation was solved by the quartic B spline functions. Finally, according to the numerical example, the comparison was made between the quartic spline approximation method and the finite difference method, and an intuitive comparison of numerical errors was given, which indicates that the former is more convenient and practical than the latter.

Key words: conformality of smoothing cofactor method, quartic B-spline, convection-diffusion equation, finite difference method, numerical solution of differential equation