0% found this document useful (0 votes)
225 views10 pages

SIMPLE Algorithm in CFD

This chapter discusses solution algorithms for pressure-velocity coupling in steady computational fluid dynamics simulations. It introduces the challenges of nonlinearities in the momentum equations and the pressure-velocity linkage. The SIMPLE algorithm is presented as a semi-implicit method for pressure-linked equations that iteratively solves for velocity and pressure corrections until the fields converge. Variants like SIMPLER, SIMPLEC and PISO are also mentioned. The key principle of SIMPLE is that it adjusts pressure in cells based on mass imbalances, driving flow from high to low pressure to satisfy continuity.

Uploaded by

Hsan Haddar
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
0% found this document useful (0 votes)
225 views10 pages

SIMPLE Algorithm in CFD

This chapter discusses solution algorithms for pressure-velocity coupling in steady computational fluid dynamics simulations. It introduces the challenges of nonlinearities in the momentum equations and the pressure-velocity linkage. The SIMPLE algorithm is presented as a semi-implicit method for pressure-linked equations that iteratively solves for velocity and pressure corrections until the fields converge. Variants like SIMPLER, SIMPLEC and PISO are also mentioned. The key principle of SIMPLE is that it adjusts pressure in cells based on mass imbalances, driving flow from high to low pressure to satisfy continuity.

Uploaded by

Hsan Haddar
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/ 10

Contents

 Chap>1 Introduction

 Chap>2 Conservation laws of fluid motion and boundary conditions
Computational Fluid Dynamics  Chap>3 Turbulence and its modeling ‐‐ skipped

 Chap>4 The finite volume method for diffusion problems

 Chap>5 The finite volume method for convection‐diffusion problems

 Chap>6 Solution algorithms for pressure‐velocity coupling in steady flows
Dr, Yasser Ben Salah  Chap>7 Solution of discretized equations

 Chap>8 The finite volume method for unsteady flows
National Engineering School 
Monastir

1 2

Contents
 Introduction
 The staggered grid
 The momentum equations
 The SIMPLE algorithm
CHAPTER 6: SOLUTION ALGORITHMS FOR    Assembly of a complete method
PRESSURE‐VELOCITY COUPLING IN STEADY    The SIMPLER algorithm
FLOWS  The SIMPLEC algorithm
 The PISO algorithm
 General comments on SIMPLE, SIMPLER, SIMPLEC and PISO
 Worked examples of the SIMPLE algorithm
 Summary

3 3 4
Introduction Introduction
 The convection of a scalar variable φ depends on the magnitude and direction of the    Two (2) new problems
local velocity field.  Non‐linearities
 How to find flow field?  The convective terms of the momentum equations contain non‐linear quantities (ρu2).
 In the previous chapter, we assumed that the velocity field was ”somehow ”known.  All three equations are intricately* coupled because every velocity component appears in each  
momentum equation and in the continuity equation.
 In general the velocity field is, however, not known and emerges as part of the overall  
solution process along with all other flow variables.  The pressure‐velocity linkage
 The most complex issue to resolve is the role played by the pressure. It appears in both momentum  
equations, but there is evidently no (transport or other) equation for the pressure.
 Governing equations (2D steady‐state N‐S eqs.)  If the flow is compressible, 
 continuity eq  used as transport eq for density
 Energy eq is the transport eq for TT
 Pressure may then be obtained by using eq of state, p = p(ρ,T)
 If the flow is incompressible the density is constant and hence by definition not linked to the  
pressure.
 In this case coupling between pressure and velocity introduces a constraint in the solution of 
the  flow field
– If the correct pressure field is applied in the momentum equations the resulting velocity field should satisfy  
continuity.

 Both problems can be resolved by adopting an iterative solution strategy such as SIMPLE
 Patankar and Spalding
5 6
*  (FR) 
intricately de façon complexe !  SIMPLE : Semi‐Implicit Method for Pressure Linked Equations

() Introduction
Principle behind SIMPLE  SIMPLE algorithm
• The principle behind SIMPLE is quite simple!
 The convective fluxes per unit mass F through cell faces are evaluated from so‐called guessed  
• It is based on the argument that fluid flows from regions with high pressure to low pressure.
– Start with an initial pressure field. velocity components.
– Look at a cell.  A guessed pressure field is used to solve the momentum equations
– If continuity is not satisfied because there is more mass flowing into that cell than out of the cell, the 
 A pressure correction equation, deduced from the continuity equation, is solved to obtain a  
pressure in that cell compared to the neighboring cells must be too low.
– Thus the pressure in that cell must be increased relative to the neighboring cells. pressure correction field, which is in turn used to update the velocity and pressure fields.
– The reverse is true for cells where more mass flows out than in.  To start the iteration process we use initial guesses for the velocity and pressure fields.
– Repeat this process iteratively for all cells.  The process is iterated until convergence of the velocity and pressure fields.
• The trick is in finding a good equation for the pressure correction as a function of mass imbalance. These equations 
will not be discussed here but can be readily found in the literature.

Improvements on SIMPLE
• SIMPLE is the default algorithm in most commercial finite volume codes.
• Improved versions are:
– SIMPLER (SIMPLE Revised).
– SIMPLEC (SIMPLE Consistent).
– PISO (Pressure Implicit with Splitting of Operators).
• All these algorithms can speed up convergence because they allow for the use of larger underrelaxation factors 
than SIMPLE.
• All of these will eventually converge to the same solution. The differences are in speed and stability.
• Which algorithm is fastest depends on the flow and there is no single algorithm that is always faster than the other 
ones.
7 8
Contents The staggered grid
 Introduction  Where to store the velocities?
 The staggered grid If you stagger, you walk very unsteadily, for    If the velocities and the pressures are both defined at the nodes of an ordinary CV a  
 The momentum equations example because you are ill or drunk. highly non‐uniform pressure field can act like a uniform field in the discretized  
CV Faces values
 The SIMPLE algorithm momentum equations.
 Checker board problem
 Assembly of a complete method
 The pressure at the central node (P)
 The SIMPLER algorithm
 does not appear
 The SIMPLEC algorithm  dp/dx, dp/dy
 The PISO algorithm  Zero at all the nodal points  
 General comments on SIMPLE, SIMPLER, SIMPLEC and PISO even with the oscillations
 Zero momentum source
 Worked examples of the SIMPLE algorithm
 Summary
This behaviour is obviously Checker board pressure field  
non‐physical.

 Solutions
 Use a staggered grid
 Rhie‐Chow interpolation
9 10

() The staggered grid
The idea is :  Harlow and Welch (1965)
 Scalar variables
 Pressure
to evaluate scalar  variables, such as pressure, density, temperature etc., at  ordinary   Stored at 
nodal points  Velocities
 Stored at  and 
 x‐direction: 
but to calculate velocity components on staggered grids centred around the cell faces.  y‐direction: 

 Scalar node
The arrangement for a two‐dimensional flow calculation is shown in Figure 6.2.  (I,J)
 Velocity node
 (i,J)
 (I,j)

11 12
The staggered grid The staggered grid
 Harlow and Welch (1965)  Harlow and Welch (1965)
 Scalar variables  Scalar variables
 Pressure  Pressure
 Stored at   Stored at 
 Velocities  Velocities
 Stored at  and   Stored at  and 
 x‐direction:   x‐direction: 
 y‐direction:   y‐direction: 

 Scalar node  Scalar node


 (I,J)  (I,J) intersection of 2 grid lines
 Velocity node  Velocity node
 (i,J)  (i,J)
 (I,j)  (I,j)
intersection of :
a line defining a cell boundary
and a grid line  mix of subscripts

13 14

The staggered grid Contents


 In the staggered grid arrangement  Introduction
 Pressure gradient  The staggered grid
 At (i,J) and (I,j)  The momentum equations
 The SIMPLE algorithm
 Assembly of a complete method
 The SIMPLER algorithm
 Advantages  The SIMPLEC algorithm
 No checkerboard problem
 The PISO algorithm
 No interpolation to 
calculate  velocities at cell
 General comments on SIMPLE, SIMPLER, SIMPLEC and PISO
faces  Worked examples of the SIMPLE algorithm
 Summary

15 16
The momentum equations The momentum equations
 On staggered grid  Discretized momentum equation
 Scalar grid: I-1, I, I+1, …. , J-1, J, J+1, …. Unbroken grid lines ____  x‐directional momentum eq.
 Velocity grid: i-1, i, i+1, …. , j-1, j, j+1, …. Dashed lines that construct cell faces ------
 Backward staggered grid
1 1
xi  xI  xu y j  y J  y v
2 2

Fig . 6.3

17 18

The momentum equations The momentum equations


 Discretized momentum equation  Discretized momentum equation
 x‐directional momentum eq.  x‐directional momentum eq.

1  I 1,J  I ,J   I ,J  I 1,J  
Fe  u e   ui1,J   ui, J 
2  2   2  

1  I ,J  I 1,J   I 1,J  I 2, J  


Fw  u w   ui,J  ui1, J 
2  2   2  

1
Fn  vn  vnw  une 
2
 1  I 1,J 1  I 1,J vI 1, j 1   I ,J 1  I ,J vI , j1 
2  2   2  

1 Fig . 6.3
Fs  vs  vsw  u se 
2
 1  I 1,J  I 1,J 1   I ,J  I ,J 1 
vI 1,j  

vI , j 
2  2   2   19 20
The momentum equations The momentum equations
 Discretized momentum equation  Discretized momentum equation
 x‐directional momentum eq.  x‐directional momentum eq.

21 22

The momentum equations The momentum equations


 Discretized momentum equation
 Discretized momentum equation
 x‐directional momentum eq.
 x‐directional momentum eq.

23 24
The momentum equations Contents
 Discretized momentum equation  Introduction
 y‐directional momentum eq.  The staggered grid
 The momentum equations
 The SIMPLE algorithm
 Assembly of a complete method
 The SIMPLER algorithm
 The SIMPLEC algorithm
x‐directional momentum eq:
 The PISO algorithm
 General comments on SIMPLE, SIMPLER, SIMPLEC and PISO
 Worked examples of the SIMPLE algorithm
 Summary
 Fe , Fw , Fn , Fs : include the velocity, non‐linear part, use the previous iteration values
 Iterative solution method is required.
 Given a pressure field p, discretized momentum equations of the form can be written for  
each u‐ and v‐control volume and then solved to obtain the velocity fields.
 If the pressure field is correct the resulting velocity field will satisfy continuity.
 As the pressure field is unknown, we need a method for calculating pressure. 25 26

The SIMPLE algorithm The SIMPLE algorithm


 SIMPLE  = Semi‐Implicit Method for Pressure Linked Equations  Semi‐Implicit Method for Pressure Linked Equations
 Patankar and Spalding
 Guess‐and‐correct procedure for the calculation of pressure on the staggered grid  
arrangement
 
ai,J ui,J   anbu nb  p I 1,J  pI ,J Ai,J  bi,J
 Initiation of the SIMPLE calculation
 Guess: p*

 Solve the discretized momentum equations

 Intermediate velocity
– Two systems of equations
– For u* and v*

 Pressure correction, velocity corrections
27 28
The SIMPLE algorithm The SIMPLE algorithm
 Semi‐Implicit Method for Pressure Linked Equations  Semi‐Implicit Method for Pressure Linked Equations

Omission of these terms  
Express u and v using p’
 For simplicity, the following two terms are dropped. Is the main approximation  
of SIMPLE

a '
u
nb nb a '
v
nb nb

'
ai, J ui,J 
 pI' 1,J  p 'I ,J Ai,J 
a I , j vI' ,j  p '
I ,J 1
 p'
I ,J
A I ,j
Express u and v using p’

 Solve the momentum equations with guessed pressure and guessed velocities for non‐linear  
terms.
 The predicted velocity should satisfy continuity equation.

29 30

The SIMPLE algorithm The SIMPLE algorithm


 Semi‐Implicit Method for Pressure Linked Equations  Semi‐Implicit Method for Pressure Linked Equations
 Continuity equation on scalar node  Continuity equation on scalar node

 
 u  A uA  uA  vA  vA 
f
e w n s
uA  
 uAi,J  vAI , j1  vAI ,j  0
i1,J


 uAi1,J  uAi,J  vAI , j1  vAI ,j  0 

Express u and v using p’

Derive the pressure correction eq.
31 32
The SIMPLE algorithm The SIMPLE algorithm
 Semi‐Implicit Method for Pressure Linked Equations  Semi‐Implicit Method for Pressure Linked Equations
 Continuity equation on scalar node  Continuity equation on scalar node
 Rearrange  Cell pressure correction equation

33 34

The SIMPLE algorithm The SIMPLE algorithm


 Semi‐Implicit Method for Pressure Linked Equations  Semi‐Implicit Method for Pressure Linked Equations
 Continuity equation on scalar node  Solve the pressure correction equation
 Collect the cell pressure equations for all cells  Pressure correction p’ can be obtained.
 System pressure correction equation

     
  '   
  p I ,J 1   
     
    
  pI' 1,J   
    
 p I ,J    b I , J  Not the final solution. Why?
  a I ,J 1   a I 1,J  a I 1,J   a I ,J 1 ' '
a I ,J 


 pI' 1, J  
  


'
ai, J ui,J 
 pI' 1,J  p'I ,J Ai,J 
     
  p'I ,J 1    a v '
I , j I ,j  p '
I ,J 1  p 'I ,J A
I ,j
    
      Fe , Fw , Fn , Fs
35 36
The SIMPLE algorithm The SIMPLE algorithm
 Semi‐Implicit Method for Pressure Linked Equations  Semi‐Implicit Method for Pressure Linked Equations
 Under relaxation  Discretize momentum equations from the second iteration

 Pressure under‐relaxation factor
0  p 1

 u‐ and v‐ velocity under relaxation factors

 A correct choice of under‐relaxation factors α is essential for cost‐effective simulations.
 Too large a value of α may lead to oscillatory or even divergent iterative solutions, and a value  
which is too small will cause extremely slow convergence.
 Unfortunately, the optimum values of under‐relaxation factors are flow dependent and must be  
sought on a case‐by‐case basis.
37 38

The SIMPLE algorithm The SIMPLE algorithm


 Semi‐Implicit Method for Pressure Linked Equations
 Pressure correction equation from the second iteration

39 40

You might also like