This action might not be possible to undo. Are you sure you want to continue?
CFD Inside – Finite Volume Method
Applied CFD
Santiago Laín Beatove
Adapted from material of A. Bakker “Applied CFD” & An Introduction to CFD
of S. Laín
2
Overview of numerical methods
• Many CFD techniques exist.
• The most common in commercially available CFD programs are:
– The finite volume method has the broadest applicability (~80%).
– Finite element (~15%).
• Here we will focus on the finite volume method.
• There are certainly many other approaches (5%), including:
– Finite difference.
– Finite element.
– Spectral methods.
– Boundary element.
– Vorticity based methods.
– Lattice gas/lattice Boltzmann.
– And more!
3
Finite difference method (FDM)
• Historically, the oldest of the three.
• Techniques published as early as 1910 by L. F. Richardson.
• Seminal paper by Courant, Fredrichson and Lewy (1928) derived
stability criteria for explicit time stepping.
• First ever numerical solution: flow over a circular cylinder by
Thom (1933).
• Scientific American article by Harlow and Fromm (1965) clearly
and publicly expresses the idea of “computer experiments” for the
first time and CFD is born!!
• Advantage: easy to implement.
• Disadvantages: restricted to simple grids and does not conserve
momentum, energy, and mass on coarse grids.
4
• The domain is discretized into a series of grid points.
• A “structured” (ijk) mesh is required.
• The governing equations (in differential form) are discretized
(converted to algebraic form).
• First and second derivatives are approximated by truncated
Taylor series expansions.
• The resulting set of linear algebraic equations is solved either
iteratively or simultaneously.
i
j
i
j
Finite difference: basic methodology
5
coextrusion
metal insert
contours of velocity magnitude
• Earliest use was by Courant (1943) for solving a torsion problem.
• Clough (1960) gave the method its name.
• Method was refined greatly in the 60’s and 70’s, mostly for
analyzing structural mechanics problem.
• FEM analysis of fluid flow was developed in the mid to late 70’s.
• Advantages: highest accuracy on coarse grids. Excellent for
diffusion dominated problems (viscous flow) and viscous, free
surface problems.
• Disadvantages: slow for large problems
and not well suited for turbulent flow.
Finite element method (FEM)
6
• First welldocumented use was by Evans and Harlow (1957) at Los
Alamos and Gentry, Martin and Daley (1966).
• Was attractive because while variables may not be continuously
differentiable across shocks and other discontinuities mass, momentum
and energy are always conserved.
• FVM enjoys an advantage in memory use and speed for very large
problems, higher speed flows, turbulent flows, and source term
dominated flows (like combustion).
• Late 70’s, early 80’s saw development of bodyfitted grids. By early 90’s,
unstructured grid methods had appeared.
• Advantages: basic FV control volume balance does not limit cell shape;
mass, momentum, energy conserved even on coarse grids; efficient,
iterative solvers well developed.
• Disadvantages: false diffusion when simple numerics are used.
Finite volume method (FVM)
7
• Divide the domain into control volumes.
• Integrate the differential equation over the control volume and
apply the divergence theorem.
• To evaluate derivative terms, values at the control volume faces
are needed: have to make an assumption about how the value
varies.
• Result is a set of linear algebraic equations: one for each control
volume.
• Solve iteratively or simultaneously.
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Finite volume: basic methodology
8
Control volume
Computational node
Boundary node
Cells and nodes
• Using finite volume method, the solution domain is subdivided
into a finite number of small control volumes (cells) by a grid.
• The grid defines the boundaries of the control volumes while the
computational node lies at the center of the control volume.
• The advantage of FVM is that the integral conservation is
satisfied exactly over the control volume.
9
• The net flux through the control volume boundary is the sum of
integrals over the four control volume faces (six in 3D). The
control volumes do not overlap.
• The value of the integrand is not available at the control volume
faces and is determined by interpolation.
Typical control volume
P
E
W
N
S
SW SE
NE NW
j,y,v
i,x,u
n
e
s
w
∆x
∆y
δx
w
δx
e
δy
n
δy
s
10
Discretization example
• To illustrate how the conservation equations used in CFD can be
discretized we will look at an example involving the transport of a
chemical species in a flow field.
• The species transport equation (constant density, incompressible
flow) is given by:
• Here c is the concentration of the chemical species and D is the
diffusion coefficient. S is a source term.
• We will discretize this equation (convert
it to a solveable algebraic form) for the
simple flow field shown on the right,
assuming steady state conditions.
c
P
c
E
c
N
c
S
c
W
( )
i
i i i
c c
u c S
t x x x
 
∂ ∂ ∂ ∂
+ = +

∂ ∂ ∂ ∂
\ ¹
D
11
Discretization example  continued
• The balance over the control volume is given by:
• This contains values at the faces, which need to be determined
from interpolation from the values at the cell centers.
j,y,v
i,x,u
c
P
c
E
c
N
c
S
c
W
A
n
,c
n
A
e
,c
e
A
s
,c
s
A
w
,c
w
e e e w w w n n n s s s
e w n s p
e w
n s
A u c A u c A v c A v c
dc dc dc dc
A A A A S
dx dx dy dy
− + − =
− + − + D D D D
, , , : areas of the faces
, , , : concentrations at the faces
, , , : concentrations at the cell centers
, , , , , , , : velocities at the faces
, , , , , , , : velocities at the
w n e s
w n e s
W N E S
w n e s w n e s
W N E S W N E S
A A A A
c c c c
c c c c
u u u u v v v v
u u u u v v v v
Notation
cell centers
: source in cell P
: diffusion coefficient
P
S
D
12
Discretization example  continued
• The simplest way to determine the values at the faces is by using
first order upwind differencing. Here, let’s assume that the value
at the face is equal to the value in the center of the cell upstream
of the face. Using that method results in:
• This equation can then be rearranged to provide an expression
for the concentration at the center of cell P as a function of the
concentrations in the surrounding cells, the flow field, and the
grid.
( ) /
( ) / ( ) / ( ) /
e P P w W W n P P s S S e E P e
w P W w n N P n s P S s P
A u c A u c A v c A v c A c c x
A c c x A c c y A c c y S
δ
δ δ δ
− + − = −
− − + − − − +
D
D D D
13
Discretization example  continued
• Rearranging the previous equation results in:
• This equation can now be simplified to:
• Here nb refers to the neighboring cells. The coefficients a
nb
and b will be
different for every cell in the domain at every iteration. The species
concentration field can be calculated by recalculating c
P
from this
equation iteratively for all cells in the domain.
( / / / / ) ( / )
( / )
( / )
( / )
P n P e P w w n p e e s s W w W w w
N n n
E e e
S s S s s
P
c A v A u A x A y A x A y c A u A x
c A y
c A x
c A v A y
S
δ δ δ δ δ
δ
δ
δ
+ + + + + = + +
+
+
+ +
D D D D D
D
D
D
P P W W N N E E S S
nb nb
nb
a c a c a c a c a c b
a c b
= + + + +
= +
∑
14
Finite Volume Method for Diffusion Problems
15
Finite Volume Method for Diffusion Problems
16
Finite Volume Method for Diffusion Problems
17
Finite Volume Method for Diffusion Problems
18
Finite Volume Method for Diffusion Problems
19
Finite Volume Method for Diffusion Problems
20
Finite Volume Method for Diffusion Problems
Tridiagonal matrix
21
Finite Volume Method for Diffusion Problems
22
Finite Volume Method for Diffusion Problems
23
Finite Volume Method for Diffusion Problems
Boundary conditions
0 = +
(
¸
(
¸

¹

\

−
−

¹

\

−
x qA
x
T T
A k
x
T T
A k
W P
w
P E
e
δ
δ δ
24
Finite Volume Method for Diffusion Problems
Boundary conditions
25
Finite Volume Method for Diffusion Problems
Boundary conditions
26
Finite Volume Method for Diffusion Problems
Boundary conditions
27
Finite Volume Method for Diffusion Problems
28
Finite Volume Method for Diffusion Problems
29
Finite Volume Method for Diffusion Problems
30
Finite Volume Method for Diffusion Problems
Exercise!
31
Finite Volume Method for Diffusion Problems
32
Finite Volume Method for Diffusion Problems
Exercise!
33
Finite Volume Method for ConvectionDiffusion
Problems
34
Finite Volume Method for ConvectionDiffusion
Problems
35
Finite Volume Method for ConvectionDiffusion
Problems
36
Finite Volume Method for ConvectionDiffusion
Problems
→Just for now!
37
Finite Volume Method for ConvectionDiffusion
Problems
P
e
E
φ(x)
φ
P
φ
e
φ
E
● We determine the value of φ at the
face by linear interpolation between the
cell centered values.
38
Finite Volume Method for ConvectionDiffusion
Problems
39
Finite Volume Method for ConvectionDiffusion
Problems
40
Finite Volume Method for ConvectionDiffusion
Problems
● Other properties are also considered: Consistency, stability,
convergence and accuracy
41
Finite Volume Method for ConvectionDiffusion
Problems
42
Finite Volume Method for ConvectionDiffusion
Problems
Discrete approximation
The discrete approximation is consistent with the differential equation iff
43
Finite Volume Method for ConvectionDiffusion
Problems
44
Finite Volume Method for ConvectionDiffusion
Problems
45
Finite Volume Method for ConvectionDiffusion
Problems
46
Finite Volume Method for ConvectionDiffusion
Problems
47
Finite Volume Method for ConvectionDiffusion
Problems
48
Finite Volume Method for ConvectionDiffusion
Problems
49
Finite Volume Method for ConvectionDiffusion
Problems
50
Finite Volume Method for ConvectionDiffusion
Problems
→Property of fluid flow
51
Finite Volume Method for ConvectionDiffusion
Problems
52
Finite Volume Method for ConvectionDiffusion
Problems
1 −
Γ = Γ
i w i e
Guarantees that diffusive fluxes are equal
from both sides of the CV
53
Finite Volume Method for ConvectionDiffusion
Problems
54
Finite Volume Method for ConvectionDiffusion
Problems
55
Finite Volume Method for ConvectionDiffusion
Problems
56
Finite Volume Method for ConvectionDiffusion
Problems
57
Finite Volume Method for ConvectionDiffusion
Problems
58
Finite Volume Method for ConvectionDiffusion
Problems
59
Finite Volume Method for ConvectionDiffusion
Problems
60
Finite Volume Method for ConvectionDiffusion
Problems
Secondorder upwind scheme
• We determine the value of φ from the
cell values in the two cells upstream of
the face.
• This is more accurate than the first
order upwind scheme, but in regions
with strong gradients it can result in face
values that are outside of the range of
cell values. It is then necessary to apply
limiters to the predicted face values.
P
e
E
φ(x)
φ
P
φ
e
φ
E
W
φ
W
Flow direction
interpolated
value
61
Finite Volume Method for ConvectionDiffusion
Problems
Hot fluid
Cold fluid
T = 100ºC
T = 0ºC
Diffusion set to zero
k=0
62
Finite Volume Method for ConvectionDiffusion
Problems
8 x 8
64 x 64
Firstorder Upwind Secondorder Upwind
Grid refinement coupled with a higherorder interpolation
scheme will minimize the false diffusion
63
Finite Volume Method for ConvectionDiffusion
Problems
64
Finite Volume Method for ConvectionDiffusion
Problems
65
Finite Volume Method for ConvectionDiffusion
Problems
66
Finite Volume Method for ConvectionDiffusion
Problems
67
Finite Volume Method for ConvectionDiffusion
Problems
P
e
E
φ(x)
φ
P
φ
e
φ
E
x
L
This is based on the analytical solution
of the onedimensional convection
diffusion equation.
68
Finite Volume Method for ConvectionDiffusion
Problems
P
e
E
φ(x)
φ
P
φ
e
φ
E
W
φ
W
Flow direction
● A quadratic curve is fitted
through two upstream nodes and
one downstream node.
● This is a very accurate scheme,
but in regions with strong
gradients, overshoots and
undershoots can occur. This can
lead to stability problems in the
calculation.
69
Finite Volume Method for ConvectionDiffusion
Problems
Diagrama de variable normalizada
En 1D con el flujo de W a E se define la variable normalizada de P a E
W E
W
φ φ
φ φ
φ
−
−
=
ˆ
Esquemas lineales
( )
( ) QUICK
UDS
CDS
P e
P e
P e
φ φ
φ φ
φ φ
ˆ
2 1
8
3
ˆ
ˆ ˆ
ˆ
1
2
1
ˆ
+ =
=
+ =
Diagrama de variable normalizada
(Gaskell & Lau, 1988)
P e
vs φ φ
ˆ
.
ˆ
La condición suficiente para que un es
quema sea de segundo orden es que
pase por el punto Q (0.5, 0.75)
(Leonard, 1988)
70
Finite Volume Method for ConvectionDiffusion
Problems
Diagrama de variable normalizada
Criterio de acotación convectivo (Gaskell & Lau, 1988) (Acotación)
Un esquema acotado ha de tener una representación que satisfaga:
● continua
● f (0) = 0
● f (1) = 1
●
●
( )
P
f φ
ˆ
P
φ
ˆ
∀
( ) ( ) 1
ˆ
0
ˆ ˆ
1
ˆ
< < > <
P P P P
f f φ φ φ φ
( ) 1
ˆ
0
ˆ ˆ ˆ
> < =
P P P P
f φ φ φ φ
El único esquema lineal que las
satisface es UDS. Existen esque
mas no lineales que cumplen las
condiciones
71
Note on accuracy of numerical schemes
• Each of the previously discussed numerical schemes assumes some
shape of the function φ. These functions can be approximated by Taylor
series polynomials:
• The first order upwind scheme only uses the constant and ignores the
first derivative and consecutive terms. This scheme is therefore
considered first order accurate.
• For high Peclet numbers the power law scheme reduces to the first
order upwind scheme, so it is also considered first order accurate.
• The central differencing scheme and second order upwind scheme do
include the first order derivative, but ignore the second order derivative.
These schemes are therefore considered second order accurate.
QUICK does take the second order derivative into account, but ignores
the third order derivative. This is then considered third order accurate.
... ) (
!
) (
.... ) (
! 2
) ( ' '
) (
! 1
) ( '
) ( ) (
2
+ − + + − + − + =
n
P e
P
n
P e
P
P e
P
P e
x x
n
x
x x
x
x x
x
x x
φ φ φ
φ φ
72
Finite Volume Method for ConvectionDiffusion
Problems
73
Finite Volume Method for ConvectionDiffusion
Problems
74
Finite Volume Method for ConvectionDiffusion
Problems
Uniform grid with 10 CV’s
Oscillations
False diffusion
5 =
∆x
Pe
75
Finite Volume Method for ConvectionDiffusion
Problems
Uniform grid with 40 CV’s
25 . 1 =
∆x
Pe
False diffusion
76
Finite Volume Method for ConvectionDiffusion
Problems
Nonniform grid with 10 CV’s
7 . 0 , 31 . 0 , 0125 . 0
max min
= = ∆ = ∆
e
r x x
False diffusion
77
Finite Volume Method for ConvectionDiffusion
Problems
Overview of numerical methods
• Many CFD techniques exist. • The most common in commercially available CFD programs are:
– The finite volume method has the broadest applicability (~80%). – Finite element (~15%).
• Here we will focus on the finite volume method. • There are certainly many other approaches (5%), including:
– – – – – – – Finite difference. Finite element. Spectral methods. Boundary element. Vorticity based methods. Lattice gas/lattice Boltzmann. And more!
2
Finite difference method (FDM)
• Historically, the oldest of the three. • Techniques published as early as 1910 by L. F. Richardson. • Seminal paper by Courant, Fredrichson and Lewy (1928) derived stability criteria for explicit time stepping. • First ever numerical solution: flow over a circular cylinder by Thom (1933). • Scientific American article by Harlow and Fromm (1965) clearly and publicly expresses the idea of “computer experiments” for the first time and CFD is born!! • Advantage: easy to implement. • Disadvantages: restricted to simple grids and does not conserve momentum, energy, and mass on coarse grids.
3
• A “structured” (ijk) mesh is required. • The resulting set of linear algebraic equations is solved either iteratively or simultaneously.Finite difference: basic methodology • The domain is discretized into a series of grid points. 4 . • First and second derivatives are approximated by truncated Taylor series expansions. i j i j • The governing equations (in differential form) are discretized (converted to algebraic form).
• Advantages: highest accuracy on coarse grids. mostly for analyzing structural mechanics problem. • FEM analysis of fluid flow was developed in the mid. • Method was refined greatly in the 60’s and 70’s. • Disadvantages: slow for large problems and not well suited for turbulent flow. Excellent for diffusion dominated problems (viscous flow) and viscous.to late 70’s. coextrusion metal insert contours of velocity magnitude 5 . free surface problems.Finite element method (FEM) • Earliest use was by Courant (1943) for solving a torsion problem. • Clough (1960) gave the method its name.
• Advantages: basic FV control volume balance does not limit cell shape. turbulent flows. efficient. • Disadvantages: false diffusion when simple numerics are used.Finite volume method (FVM) • First welldocumented use was by Evans and Harlow (1957) at Los Alamos and Gentry. early 80’s saw development of bodyfitted grids. unstructured grid methods had appeared. energy conserved even on coarse grids. iterative solvers well developed. momentum and energy are always conserved. higher speed flows. Martin and Daley (1966). • Late 70’s. momentum. 6 . and source term dominated flows (like combustion). mass. By early 90’s. • FVM enjoys an advantage in memory use and speed for very large problems. • Was attractive because while variables may not be continuously differentiable across shocks and other discontinuities mass.
• Solve iteratively or simultaneously.Finite volume: basic methodology • Divide the domain into control volumes. values at the control volume faces are needed: have to make an assumption about how the value varies. • Result is a set of linear algebraic equations: one for each control volume. • • • • • • • • • • • • • • • • • • • • • • Integrate the differential equation over the control volume and apply the divergence theorem. • To evaluate derivative terms. 7 .
Cells and nodes • Using finite volume method. Boundary node Control volume Computational node 8 . the solution domain is subdivided into a finite number of small control volumes (cells) by a grid. • The grid defines the boundaries of the control volumes while the computational node lies at the center of the control volume. • The advantage of FVM is that the integral conservation is satisfied exactly over the control volume.
v i.y.Typical control volume • The net flux through the control volume boundary is the sum of integrals over the four control volume faces (six in 3D).u w s SW S SE P e NE δyn δys E ∆y δxw δxe 9 . ∆x NW N n W j. The control volumes do not overlap.x. • The value of the integrand is not available at the control volume faces and is determined by interpolation.
cS 10 . incompressible flow) is given by: ∂c ∂ ∂ ∂c + (ui c) = D +S ∂t ∂xi ∂xi ∂xi • Here c is the concentration of the chemical species and D is the diffusion coefficient. assuming steady state conditions. S is a source term. • The species transport equation (constant density. cN • We will discretize this equation (convert it to a solveable algebraic form) for the cE cW cP simple flow field shown on the right.Discretization example • To illustrate how the conservation equations used in CFD can be discretized we will look at an example involving the transport of a chemical species in a flow field.
v i.cs j. vN .cw As. cn . which need to be determined from interpolation from the values at the cell centers. cN An. vS : velocities at the cell centers S P : source in cell P D: diffusion coefficient 11 . cN . un . c S : concentrations at the cell centers uw . vW .ce Notation Aw .Discretization example . uS . u E .continued • The balance over the control volume is given by: Aeue ce − Awuw cw + An vn cn − As vs cs = DAe dc dc dc dc − DAw + DAn − DAs + Sp dx e dx w dy n dy s • This contains values at the faces. ce .x. vs : velocities at the faces uW . vn . vw . An . Ae . ve . cE . u N . cs : concentrations at the faces cW .y. us . vE . ue . As : areas of the faces cE cW cw .cn cP Aw.u cS Ae.
Discretization example . Using that method results in: AeuP cP − AwuW cW + An vP cP − As vS cS = DAe (cE − cP ) / δ xe − DAw (cP − cW ) / δ xw + DAn (cN − cP ) / δ yn − DAs (cP − cS ) / δ ys + S P • This equation can then be rearranged to provide an expression for the concentration at the center of cell P as a function of the concentrations in the surrounding cells. the flow field. and the grid.continued • The simplest way to determine the values at the faces is by using first order upwind differencing. let’s assume that the value at the face is equal to the value in the center of the cell upstream of the face. Here. 12 .
Discretization example . The coefficients anb and b will be different for every cell in the domain at every iteration. The species concentration field can be calculated by recalculating cP from this equation iteratively for all cells in the domain. 13 .continued • Rearranging the previous equation results in: cP ( An vP + AeuP + DAw / δ xw + DAn / δ y p + DAe / δ xe + DAs / δ ys ) = cW ( AwuW + DAw / δ xw ) + cN (DAn / δ yn ) + cE (DAe / δ xe ) + cS ( As vS + DAs / δ ys ) + SP • This equation can now be simplified to: a P c P = aW cW + a N c N + a E c E + a S c S + b = ∑a nb nb nb c +b • Here nb refers to the neighboring cells.
Finite Volume Method for Diffusion Problems 14 .
Finite Volume Method for Diffusion Problems 15 .
Finite Volume Method for Diffusion Problems 16 .
Finite Volume Method for Diffusion Problems 17 .
Finite Volume Method for Diffusion Problems 18 .
Finite Volume Method for Diffusion Problems 19 .
Finite Volume Method for Diffusion Problems Tridiagonal matrix 20 .
Finite Volume Method for Diffusion Problems 21 .
Finite Volume Method for Diffusion Problems 22 .
Finite Volume Method for Diffusion Problems Boundary conditions TE − TP T −T ke A − k w A P W δx δx + qAδx = 0 23 .
Finite Volume Method for Diffusion Problems Boundary conditions 24 .
Finite Volume Method for Diffusion Problems Boundary conditions 25 .
Finite Volume Method for Diffusion Problems Boundary conditions 26 .
Finite Volume Method for Diffusion Problems 27 .
Finite Volume Method for Diffusion Problems 28 .
Finite Volume Method for Diffusion Problems 29 .
Finite Volume Method for Diffusion Problems Exercise! 30 .
Finite Volume Method for Diffusion Problems 31 .
Finite Volume Method for Diffusion Problems Exercise! 32 .
Finite Volume Method for ConvectionDiffusion Problems 33 .
Finite Volume Method for ConvectionDiffusion Problems 34 .
Finite Volume Method for ConvectionDiffusion Problems 35 .
Finite Volume Method for ConvectionDiffusion Problems → Just for now! 36 .
P e 37 .Finite Volume Method for ConvectionDiffusion Problems φ(x) φP φe φE E ● We determine the value of φ at the face by linear interpolation between the cell centered values.
Finite Volume Method for ConvectionDiffusion Problems 38 .
Finite Volume Method for ConvectionDiffusion Problems 39 .
Finite Volume Method for ConvectionDiffusion Problems ● Other properties are also considered: Consistency. convergence and accuracy 40 . stability.
Finite Volume Method for ConvectionDiffusion Problems
41
Finite Volume Method for ConvectionDiffusion Problems
Discrete approximation The discrete approximation is consistent with the differential equation iff
42
Finite Volume Method for ConvectionDiffusion Problems
43
Finite Volume Method for ConvectionDiffusion Problems 44 .
Finite Volume Method for ConvectionDiffusion Problems 45 .
Finite Volume Method for ConvectionDiffusion Problems 46 .
Finite Volume Method for ConvectionDiffusion Problems 47 .
Finite Volume Method for ConvectionDiffusion Problems 48 .
Finite Volume Method for ConvectionDiffusion Problems 49 .
Finite Volume Method for ConvectionDiffusion Problems → Property of fluid flow 50 .
Finite Volume Method for ConvectionDiffusion Problems 51 .
Finite Volume Method for ConvectionDiffusion Problems Γe i = Γw i −1 Guarantees that diffusive fluxes are equal from both sides of the CV 52 .
Finite Volume Method for ConvectionDiffusion Problems 53 .
Finite Volume Method for ConvectionDiffusion Problems 54 .
Finite Volume Method for ConvectionDiffusion Problems 55 .
Finite Volume Method for ConvectionDiffusion Problems 56 .
Finite Volume Method for ConvectionDiffusion Problems 57 .
Finite Volume Method for ConvectionDiffusion Problems 58 .
Finite Volume Method for ConvectionDiffusion Problems 59 .
Finite Volume Method for ConvectionDiffusion Problems Secondorder upwind scheme • We determine the value of φ from the cell values in the two cells upstream of the face. It is then necessary to apply limiters to the predicted face values. φW φ(x) φP interpolated value • φe W P e φE E Flow direction 60 . but in regions with strong gradients it can result in face values that are outside of the range of cell values. This is more accurate than the first order upwind scheme.
Finite Volume Method for ConvectionDiffusion Problems Hot fluid T = 100ºC Diffusion set to zero k=0 Cold fluid T = 0ºC 61 .
Finite Volume Method for ConvectionDiffusion Problems Firstorder Upwind Secondorder Upwind 8x8 64 x 64 Grid refinement coupled with a higherorder interpolation scheme will minimize the false diffusion 62 .
Finite Volume Method for ConvectionDiffusion Problems 63 .
Finite Volume Method for ConvectionDiffusion Problems 64 .
Finite Volume Method for ConvectionDiffusion Problems 65 .
Finite Volume Method for ConvectionDiffusion Problems 66 .
Finite Volume Method for ConvectionDiffusion Problems φ(x) φP This is based on the analytical solution of the onedimensional convectiondiffusion equation. φe P e E φE x L 67 .
overshoots and undershoots can occur. φW φP φ(x) φe W P e φE E Flow direction 68 . This can lead to stability problems in the calculation. but in regions with strong gradients.Finite Volume Method for ConvectionDiffusion Problems ● A quadratic curve is fitted through two upstream nodes and one downstream node. ● This is a very accurate scheme.
φ P La condición suficiente para que un esquema sea de segundo orden es que pase por el punto Q (0. 1988) ˆ φe = ( ˆ vs.Finite Volume Method for ConvectionDiffusion Problems Diagrama de variable normalizada En 1D con el flujo de W a E se define la variable normalizada de P a E ˆ = φ − φW φ φ E − φW Esquemas lineales 1 ˆ 1 + φP 2 ˆ ˆ φ =φ ˆ φe = e P ( ) ) CDS UDS QUICK 3 ˆ 1 + 2φ P 8 ˆ Diagrama de variable normalizada φ e (Gaskell & Lau. 1988) 69 .75) (Leonard.5. 0.
Finite Volume Method for ConvectionDiffusion Problems Diagrama de variable normalizada Criterio de acotación convectivo (Gaskell & Lau. 1988) (Acotación) Un esquema acotado ha de tener una representación que satisfaga: ˆ ˆ ● f φ P continua ∀φ P ● f (0) = 0 ● f (1) = 1 ˆ ˆ ˆ ˆ ● f φP < 1 f φP > φP 0 < φP < 1 ˆ ˆ ˆ ˆ ● f φP = φP φP < 0 φP > 1 ( ) ( ) ( ) ( ) El único esquema lineal que las satisface es UDS. Existen esquemas no lineales que cumplen las condiciones 70 .
These schemes are therefore considered second order accurate.Note on accuracy of numerical schemes • Each of the previously discussed numerical schemes assumes some shape of the function φ. 71 . These functions can be approximated by Taylor series polynomials: φ ( xe ) = φ ( x P ) + φ ' ( xP ) 1! ( xe − x P ) + φ ' ' ( xP ) 2! ( xe − x P ) + .. • The first order upwind scheme only uses the constant and ignores the first derivative and consecutive terms. but ignore the second order derivative.. • The central differencing scheme and second order upwind scheme do include the first order derivative. This scheme is therefore considered first order accurate. so it is also considered first order accurate. This is then considered third order accurate... • For high Peclet numbers the power law scheme reduces to the first order upwind scheme. but ignores the third order derivative. + 2 φ n ( xP ) n! ( xe − x P ) n + .. QUICK does take the second order derivative into account.
Finite Volume Method for ConvectionDiffusion Problems 72 .
Finite Volume Method for ConvectionDiffusion Problems 73 .
Finite Volume Method for ConvectionDiffusion Problems Uniform grid with 10 CV’s Oscillations False diffusion Pe∆x = 5 74 .
25 75 .Finite Volume Method for ConvectionDiffusion Problems Uniform grid with 40 CV’s False diffusion Pe∆x = 1.
re = 0.Finite Volume Method for ConvectionDiffusion Problems Nonniform grid with 10 CV’s ∆xmin = 0.31.0125.7 False diffusion 76 . ∆xmax = 0.
Finite Volume Method for ConvectionDiffusion Problems 77 .
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.