南京大学学报(自然科学版) ›› 2011, Vol. 47 ›› Issue (3): 276–280.

• • 上一篇    下一篇

 格子行走方法求解一维对流扩散方程*

 蒋建国, 吴吉春**   

  • 出版日期:2015-04-10 发布日期:2015-04-10
  • 作者简介: ( 南京大学地球科学与工程学院水科学系, 南京, 210093)
  • 基金资助:
      国家自然科学基金 ( 40725010, 40672160)

 Lattice walk method for the convection-diffusion equation

 Jiang Jian Guo, Wu J i Chun
  

  • Online:2015-04-10 Published:2015-04-10
  • About author:( Department of Hydrosciences, School of Earth Sciences and Engineering,
    Nanjing University, Nanjing, 210093, China)

摘要:  本文提出了全新的格子行走方法求解对流扩散方程. 对流扩散方程广泛应用于水中污染物的运移模拟. 然而在对流占优的情况下, 常用的求解对流扩散方程的有限差分方法和有限元法会出现严重
的数值弥散和数值振荡等不良数值现象. 本文提出的格子行走方法, 综合了随机行走方法( random walkmethod) 和格点玻尔兹曼方法(lattice Boltzmann method)的优点, 有效克服了对流占优条件下的数值弥
散现象. 在给定格子长度和时间步长的情况下, 格子行走方法所能计算的流速范围远远超过了有限差分方法的适用范围. 格子行走方法构建了相邻时刻格点上溶质分布的转移矩阵(transition matrix) , 使得格
点上的离散分布与高斯分布的 0- 2 阶矩相等. 通过转移矩阵, 可以直接计算溶质分布随时间的变化过程. 本文用一维瞬时点源模型和一维连续点源模型检验了格子行走方法的适用性. 计算结果显示格子行
走方法得到的结果与解析解吻合得非常好, 而运用有限差分方法却出现严重的数值弥散.

Abstract:  T his paper presents a novel lattice walk method to simulate the convection diffusion equation. The convection-diffusion equation is widely used to describe solute transport. However, conventional methods, e. g. the
finite difference method or the finite element method, to solve this equation may result in numerical dispersion and oscillation, especially when convection velocity is large.
The lattice walk method combines the virtues of random walk method and lattice Boltzmann method (LBM), to simulate convection-diffusion process. The positions of solute particles are limited on lattice points and the jumping
probability of particles on regular lattices is determined by Gaussian distribution of random walk. We construct the transition matrix by the equality of the moments ( the zeroth, first and second order) of discrete distribution on
lattices and Gaussian distribution of random walk. By iteration, we can calculate the evolution of the concentration of solute. Then one dimensional instantaneous point source model and constant concentration point source model are used
to test this novel method. The results of one dimensional instantaneous point source model are shown. When convection dominates the convection-diffusion process, the result given by lattice walk method agrees well with the
analytical solution while the solution of finite difference method suffers from serious numerical dispersion. For the diffusion dominated example, the concentration obtained from lattice walk method is consistent with that from finite
difference method. T herefore we can see the most important character of lattice walk method that it can simulate solute transport at convection velocities far beyond the range of conventional methods.
The results of constant concentration point source model are shown. Our results of lattice walk method agree well with analytical solutions and have little dispersion in the front of the concentration curve. However, the
concentration obtained by finite difference method suffers rather serious numerical dispersion even if Pe number is not too large. T herefore from the results of this model we can see that lattice ?walk method can exclude numerical
dispersion, which has puzzled finite difference method and finite element method for a long time.

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!