Вы находитесь на странице: 1из 16

1

Flux Numerical
Methods
2
Flux Basics
The finite-volume formulation of the conservation equations resulted in the
equation



where was the flux of the flow across the control surface resulting from
the approximation of the surface integral. For a finite-volume cell, the flux
was expressed as


Where



It was assumed that the flux was uniform over the cell face.
F P
dt
Q d

=
F

=
=
nf
f
f
F F
1

( ) | |( )
f f f
dS n Q g v F

= D

3
Fluxes on a Hexahedral Cell
A hexahedral cell contains 6 quadrilateral faces, thus





where again,

=
=
=
6
1

nf
f
f
F F
( ) | |( )
f f f
dS n Q g v F

= D

Face f
( f is an index for the face )
f
F

( )
f
dS n
Area normal
vector for
face f

4
Numerical Flux on a Cell Face
The numerical flux on a cell face is




The normal area vector is usually easily defined for a quadrilateral or
triangular cell face. The focus of the rest of this discussion is on numerical
methods for computing



at a cell face.
We first will assume that is a known velocity for the cell face.
( ) | |( )
f f f
dS n Q g v F

= D

( ) | |
f
Q g v D

( )
f
dS n
g

5
Numerical Flux on a Cell Face
We first consider that we have the states of
the flow on the left and right of the cell
face, Q
L
and Q
R
. Our objective is to find the
cell face flux.



One can define



A consistency condition for the numerical
flux is that if Q
L
= Q
R
, then
f
F

( )
f
dS n
R
Q
L
Q
( )
R L f f
Q Q F F ,

=
( )
L f L
Q F F

= ( )
R f R
Q F F

=
R L
F F

=
6
Central Difference Method
A central difference method for computing the flux is simply



The central-difference method works okay for elliptic components of the flux
because there is no preferred direction for the propagation of information.

A simple central difference is often unstable, especially in the presence of
strong gradients. One solution is to add some second-order and fourth-order
dissipation (artificial viscosity) to the flux.



Methods for computing D
(2)
and D
(4)
vary, but generally use second and
fourth-order differences with switches to handle variations in Q.
( )
R L f
F F F

2
1

+ =
( )
) 4 ( ) 2 (

2
1

D D F F F
R L f
+ + + =
7
Use of Central Difference Method
For the Navier-Stokes equations, the viscous shear stress and heat flux terms
in the viscous component D
V
of the non-convective component D are elliptic
and those flux components can be computed with the central difference
method.

Similarly, the fluxes of the turbulence and chemistry equations can be
computed using the central difference method.

The convective portion of the flux and the pressure term in the inviscid
component D
I
of the Navier-Stokes equation have a hyperbolic character.
This wave nature can be put to use to compute the flux using upwind methods.
8
Upwind Methods
We expressed the non-convective portion of the flux of the Navier-Stokes
equation as


This results in the cell-face flux being expressed as


or



We will now focus on computing the inviscid flux using upwind methods.
The focus will be on the use of Roes Upwind Flux-Difference Splitting
Method.
V
NS
I
NS NS
D D D + =
( ) | |
V
NS
I
NS NS f
Q g v F D D =

V I
f f f
F F F

+ =
I
f
F

9
Roe Upwind Flux-Difference Method
The Roe upwind flux-difference method computes the inviscid flux as:



where is the flux difference computed as,
( )
f
I
R
I
L
I
f
F F F F

2
1

2
1

A + =
+
A + A = A F F F

m m
m
m
w r F o

=
+ +
= A
5
1
) (

m m
m
m
w r F o

=

= A
5
1
) (

A
Roes method is the default flux method in WIND.
10
Roe Upwind Flux-Difference Method
The are the eigenvalues that represent the speed of the waves. The (+)
indicate positive eigenvalues and the (-) indicates negative eigenvalues.
The are the right eigenvectors that represent the direction of propagation
of the waves.
The are the Riemann invariants and represent the strength of the wave,



m

m
r

m
w o
2
1
c
p
w
o
o o =
u n w n w o o o
3 1 2
=
w n v n u n
c
p
w o o o

o
o
3 2 1 4
+ + + =
v n u n w o o o
1 2 3
=
w n v n u n
c
p
w o o o

o
o
3 2 1 5
+ + =
11
Roe Upwind Flux-Difference Method
The differentials are computed as







Flow properties at the face are computed using Roe-averaging





Similar for computing v, w, and h
t
.


L R
o =
L R
p p p = o
L R
u u u = o
L R
v v v = o
L R
w w w = o
L R
=
2
2 / 1 2 / 1
2 / 1 2 / 1
R L
R R L L
u u
u


+
+
=
12
Higher-Order Projection
The choice of values of Q
L
and Q
R
have several option:
1) Use the values of the finite-volume cells to the left and right of the
face. This is a zero-order evaluation and will result in a spatially first-order
flux.
2) Use an extrapolation of neighboring finite-volume cells to form a first-
order evaluation of Q at the face. This will result in a spatially second-order
flux.

) (
2
1
1
+ =
i i i L
Q Q Q Q
Q
i
Q
i+2
Q
i+1
Q
i-1
Q
R
Q
L
f
F

) (
2
1
1 2 1 + + +
=
i i i R
Q Q Q Q
13
Variation Limiting
The simple extrapolation formulas assume a smooth variation of Q; however,
discontinuities in Q are possible (i.e. shocks). Need some mechanism to
sense such discontinuities and limit the variation of Q in these extrapolation
formulas. Modify the extrapolations by introducing a limiter |,







This gets into the topic of TVD (Total Variational Diminishing) flux limiting
methods, which we will not get into here. The essential role of the limiter is
to make | 0 in the presence of large variations, which make the flux
spatially first-order.
) (
2
1
1
+ =
i i i L
Q Q Q Q |
) (
2
1
1 2 1 + + +
=
i i i R
Q Q Q Q |
14
Examples of Limiters
The possible functions (and theory) for limiters is varied. A couple examples
include:

Superbee:


Chakravarthy:

Where r is some ratio of the flow properties and indicates the amount of
variation in the solution. An example is



The | is a compression parameter 1 s | s 2, where a value toward 1 makes
the limiter more dissipative.
{} | | | | , min , 0 max ) ( r r=
( ) () | | 2 , min , 1 , 2 min , 0 max ) ( r r r= |
i
i
Q
Q
r
o
o
1 +
=
15
Flux Vector Splitting
An alternative to flux-difference splitting is flux-vector splitting that
considers that the inviscid flux can be linearly separated


van Leers flux-vector splitting has the general form of
+
+ = F F F
I
f

( )
( )
( )
( ) ( ) | | ( ) | |
(
(
(
(
(
(

+
+
+
+
=

1 2 2 1 2
2
2
2
1

2
2
2 2
3
2
1

c v v q
c v n w
c v n v
c v n u
f F
n n
n
n
n
mass
( )
4
1
2

n
mass
M c
f

c
v
M
n
n
=
n v v
n
=

16
Other RHS Methods
Other methods for the right-hand-side (RHS) that will not be discussed:
Methods available for 3
rd
to 5
th
order spatial accuracy.
Roes method is modified to allow non-uniform grids.
Roes method as used in the OVERFLOW code is available.
Coakley method is available
HLLE method is available (similar to Roes method)

Next