Thanks to visit codestin.com
Credit goes to www.scribd.com

Scribd
Upload
37 views1 page

Cepdata

This short course will cover multigrid methods for solving fluid flow and heat transfer equations. Multigrid methods use several coarse grids with decreasing resolution to efficiently solve large systems of discrete equations. Key components of multigrid include hierarchical grids, point relaxation solvers, interpolation between grids, and V- or W-cycles. The course will describe multigrid theory and implementation, provide code demonstrations, and present convergence results for varying grid sizes applied to heat conduction and incompressible flows using finite difference, finite volume, and meshless methods. It is taught by Professor Pratap Vanka, an expert in computational fluid dynamics and numerical algorithms including multigrid methods.

Uploaded by

manojkarkare
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
37 views1 page

Cepdata

This short course will cover multigrid methods for solving fluid flow and heat transfer equations. Multigrid methods use several coarse grids with decreasing resolution to efficiently solve large systems of discrete equations. Key components of multigrid include hierarchical grids, point relaxation solvers, interpolation between grids, and V- or W-cycles. The course will describe multigrid theory and implementation, provide code demonstrations, and present convergence results for varying grid sizes applied to heat conduction and incompressible flows using finite difference, finite volume, and meshless methods. It is taught by Professor Pratap Vanka, an expert in computational fluid dynamics and numerical algorithms including multigrid methods.

Uploaded by

manojkarkare
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

Multigrid Methods for Fluid Flow and Heat Transfer: A Short Course by Prof. S. P.

Vanka
SPARC Supported Short Course
September 11-15, 2023
The solution of the fluid flow and heat transfer equations presents a significant challenge because of the
coupling and nonlinearity in the governing equations. Further, for accurate solutions, large numbers of
discrete equations need to be satisfied. These discrete equations require many thousands of iterations to
converge to small residuals, especially when fine grids are used. Multigrid methods deploy several coarse
levels and cycle the solution errors between the grids to produce fast iterative convergence. Multilevel
methods are ideal for solving equations governing heat conduction, electromagnetics, incompressible
flows, etc. In a properly implemented multilevel method, the number of iterations to reach a prescribed
level of convergence stays the same for any level of refinement. Multigrid methods consist of four
components. First, several hierarchical grids are defined with decreasing numbers of points (typically by a
factor of 8 in 3D and 4 in 2D) from the previous grid. Second an iterative scheme is selected to solve the
discrete equations on any level. The SOR scheme is the most efficient of the point solvers and is commonly
selected. Other solvers such as ILU with drop tolerance, preconditioned BiCGSTAB and GMRES can also be
used with the multigrid method. The third component of multigrid methods is interpolation which
transfers residuals and corrections between grids. The V- or W-cycle of visiting the grids constitutes the
final component.
This short course will describe the theory, implementation, and results of multigrid methods, as
implemented for heat conduction and incompressible flows. Results will be presented of convergence for
grids of varying number of points. Multigrid methods in the context of finite difference, finite volume and
recently developed meshless methods will be described. The topics covered are:
Lecture 1: Introduction to Multigrid method, conventional iterative solvers, and performance
Lecture 2: Multigrid concept, theory
Lecture 3: Multigrid components
Lecture 4: Model code description and demonstration for heat conduction
Lecture 5: Complex domains: nested refinement, agglomeration coarsening
Lecture 6: Non-overlapping multigrid, algebraic multigrid
Lecture 7: Full Approximation Scheme (FAS) and Full Multigrid (FMG)
Lecture 8: SIMPLE algorithm for fluid flows
Lecture 9: Sample multigrid code for fluid flows
Lecture 10: Results

Pratap Vanka is Professor Emeritus in the Department of Mechanical Science


and Engineering, UIUC. He has pioneered several numerical algorithms
including multigrid methods, Lattice Boltzmann methods, meshless techniques,
GPU computing, and partially parabolic methods. He has taught a graduate level
CFD course at the University of Illinois for 30 years. He is passionate about
developing codes for CFD and heat transfer and has developed more than 25
research level CFD codes since his graduate research at Imperial College. He
worked for his Ph. D. with Professor D. B. Spalding (late), a pioneer in
computational fluid dynamics and computational heat transfer. Pratap Vanka
has published more than 170 papers in journals and reviewed technical
conferences on heat transfer, metal solidification, combustion, and
computational methods. He has received several teaching and research awards.
He is a Life Fellow of ASME, Fellow of APS, Associate Fellow of AIAA, and
recipient of the ASME Freeman Scholar lecture award.

You might also like