Fully decoupled monolithic projection method for natural convection problems

Xiaomin Pan, Kyoungyoun Kim, Chang Hoon Lee, Jung-il Choi

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

To solve time-dependent natural convection problems, we propose a fully decoupled monolithic projection method. The proposed method applies the Crank–Nicolson scheme in time and the second-order central finite difference in space. To obtain a non-iterative monolithic method from the fully discretized nonlinear system, we first adopt linearizations of the nonlinear convection terms and the general buoyancy term with incurring second-order errors in time. Approximate block lower-upper decompositions, along with an approximate factorization technique, are additionally employed to a global linearly coupled system, which leads to several decoupled subsystems, i.e., a fully decoupled monolithic procedure. We establish global error estimates to verify the second-order temporal accuracy of the proposed method for velocity, pressure, and temperature in terms of a discrete l2-norm. Moreover, according to the energy evolution, the proposed method is proved to be stable if the time step is less than or equal to a constant. In addition, we provide numerical simulations of two-dimensional Rayleigh–Bénard convection and periodic forced flow. The results demonstrate that the proposed method significantly mitigates the time step limitation, reduces the computational cost because only one Poisson equation is required to be solved, and preserves the second-order temporal accuracy for velocity, pressure, and temperature. Finally, the proposed method reasonably predicts a three-dimensional Rayleigh–Bénard convection for different Rayleigh numbers.

Original languageEnglish
Pages (from-to)582-606
Number of pages25
JournalJournal of Computational Physics
Volume334
DOIs
Publication statusPublished - 2017 Apr 1

Fingerprint

Natural convection
free convection
projection
convection
Poisson equation
Buoyancy
Factorization
Linearization
Nonlinear systems
Decomposition
Rayleigh number
linearization
Temperature
nonlinear systems
buoyancy
factorization
norms
Computer simulation
Convection
Costs

All Science Journal Classification (ASJC) codes

  • Numerical Analysis
  • Modelling and Simulation
  • Physics and Astronomy (miscellaneous)
  • Physics and Astronomy(all)
  • Computer Science Applications
  • Computational Mathematics
  • Applied Mathematics

Cite this

@article{34a8ecafc2ce42df92a33216614d87b4,
title = "Fully decoupled monolithic projection method for natural convection problems",
abstract = "To solve time-dependent natural convection problems, we propose a fully decoupled monolithic projection method. The proposed method applies the Crank–Nicolson scheme in time and the second-order central finite difference in space. To obtain a non-iterative monolithic method from the fully discretized nonlinear system, we first adopt linearizations of the nonlinear convection terms and the general buoyancy term with incurring second-order errors in time. Approximate block lower-upper decompositions, along with an approximate factorization technique, are additionally employed to a global linearly coupled system, which leads to several decoupled subsystems, i.e., a fully decoupled monolithic procedure. We establish global error estimates to verify the second-order temporal accuracy of the proposed method for velocity, pressure, and temperature in terms of a discrete l2-norm. Moreover, according to the energy evolution, the proposed method is proved to be stable if the time step is less than or equal to a constant. In addition, we provide numerical simulations of two-dimensional Rayleigh–B{\'e}nard convection and periodic forced flow. The results demonstrate that the proposed method significantly mitigates the time step limitation, reduces the computational cost because only one Poisson equation is required to be solved, and preserves the second-order temporal accuracy for velocity, pressure, and temperature. Finally, the proposed method reasonably predicts a three-dimensional Rayleigh–B{\'e}nard convection for different Rayleigh numbers.",
author = "Xiaomin Pan and Kyoungyoun Kim and Lee, {Chang Hoon} and Jung-il Choi",
year = "2017",
month = "4",
day = "1",
doi = "10.1016/j.jcp.2017.01.022",
language = "English",
volume = "334",
pages = "582--606",
journal = "Journal of Computational Physics",
issn = "0021-9991",
publisher = "Academic Press Inc.",

}

Fully decoupled monolithic projection method for natural convection problems. / Pan, Xiaomin; Kim, Kyoungyoun; Lee, Chang Hoon; Choi, Jung-il.

In: Journal of Computational Physics, Vol. 334, 01.04.2017, p. 582-606.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Fully decoupled monolithic projection method for natural convection problems

AU - Pan, Xiaomin

AU - Kim, Kyoungyoun

AU - Lee, Chang Hoon

AU - Choi, Jung-il

PY - 2017/4/1

Y1 - 2017/4/1

N2 - To solve time-dependent natural convection problems, we propose a fully decoupled monolithic projection method. The proposed method applies the Crank–Nicolson scheme in time and the second-order central finite difference in space. To obtain a non-iterative monolithic method from the fully discretized nonlinear system, we first adopt linearizations of the nonlinear convection terms and the general buoyancy term with incurring second-order errors in time. Approximate block lower-upper decompositions, along with an approximate factorization technique, are additionally employed to a global linearly coupled system, which leads to several decoupled subsystems, i.e., a fully decoupled monolithic procedure. We establish global error estimates to verify the second-order temporal accuracy of the proposed method for velocity, pressure, and temperature in terms of a discrete l2-norm. Moreover, according to the energy evolution, the proposed method is proved to be stable if the time step is less than or equal to a constant. In addition, we provide numerical simulations of two-dimensional Rayleigh–Bénard convection and periodic forced flow. The results demonstrate that the proposed method significantly mitigates the time step limitation, reduces the computational cost because only one Poisson equation is required to be solved, and preserves the second-order temporal accuracy for velocity, pressure, and temperature. Finally, the proposed method reasonably predicts a three-dimensional Rayleigh–Bénard convection for different Rayleigh numbers.

AB - To solve time-dependent natural convection problems, we propose a fully decoupled monolithic projection method. The proposed method applies the Crank–Nicolson scheme in time and the second-order central finite difference in space. To obtain a non-iterative monolithic method from the fully discretized nonlinear system, we first adopt linearizations of the nonlinear convection terms and the general buoyancy term with incurring second-order errors in time. Approximate block lower-upper decompositions, along with an approximate factorization technique, are additionally employed to a global linearly coupled system, which leads to several decoupled subsystems, i.e., a fully decoupled monolithic procedure. We establish global error estimates to verify the second-order temporal accuracy of the proposed method for velocity, pressure, and temperature in terms of a discrete l2-norm. Moreover, according to the energy evolution, the proposed method is proved to be stable if the time step is less than or equal to a constant. In addition, we provide numerical simulations of two-dimensional Rayleigh–Bénard convection and periodic forced flow. The results demonstrate that the proposed method significantly mitigates the time step limitation, reduces the computational cost because only one Poisson equation is required to be solved, and preserves the second-order temporal accuracy for velocity, pressure, and temperature. Finally, the proposed method reasonably predicts a three-dimensional Rayleigh–Bénard convection for different Rayleigh numbers.

UR - http://www.scopus.com/inward/record.url?scp=85009958857&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85009958857&partnerID=8YFLogxK

U2 - 10.1016/j.jcp.2017.01.022

DO - 10.1016/j.jcp.2017.01.022

M3 - Article

VL - 334

SP - 582

EP - 606

JO - Journal of Computational Physics

JF - Journal of Computational Physics

SN - 0021-9991

ER -