Skip to main navigation Skip to search Skip to main content

A fast method for solving time-dependent nonlinear convection di ffusion problems

  • Qian He
  • , Wenxin Du
  • , Feng Shi*
  • , Jiaping Yu
  • *Corresponding author for this work
  • Harbin Institute of Technology
  • Donghua University

Research output: Contribution to journalArticlepeer-review

Abstract

In this paper, a fast scheme for solving unsteady nonlinear convection diffusion problems is proposed and analyzed. At each step, we firstly isolate a nonlinear convection subproblem and a linear diffusion subproblem from the original problem by utilizing operator splitting. By Taylor expansion, we explicitly transform the nonlinear convection one into a linear problem with artificial inflow boundary conditions associated with the nonlinear flux. Then a multistep technique is provided to relax the possible stability requirement, which is due to the explicit processing of the convection problem. Since the self-adjointness and coerciveness of diffusion subproblems, there are so many preconditioned iterative solvers to get them solved with high efficiency at each time step. When using the finite element method to discretize all the resulting subproblems, the major stiffness matrices are same at each step, that is the reason why the unsteady nonlinear systems can be computed extremely fast with the present method. Finally, in order to validate the effectiveness of the present scheme, several numerical examples including the Burgers type and Buckley-Leverett type equations, are chosen as the numerical study.

Original languageEnglish
Pages (from-to)2165-2182
Number of pages18
JournalElectronic Research Archive
Volume30
DOIs
StatePublished - 2022
Externally publishedYes

Keywords

  • Buckley-Leverett type equation
  • Burgers type equation
  • Nonlinear convection diffusion problems
  • finite element method
  • multistep technique
  • operator splitting

Fingerprint

Dive into the research topics of 'A fast method for solving time-dependent nonlinear convection di ffusion problems'. Together they form a unique fingerprint.

Cite this