Vous êtes sur la page 1sur 9

One Dimensional Unsteady Euler Equation:

The unsteady 1-D Euler Equations can be written in the general conservation form-
0
U F
t x

+

-------- (1)
where,

U Dependent variables Vector

u
e



,
and, F Flux Vector

2
( )
u
p u
e p u

_

+


+
,
Here,
e
Energy per unit volume

2
1
2
u +

2
1
2
v
c T u +

2
1
1 2
R
T u


2
1
,
1 2
p
or e u

------(A)
So, we can write,
2
1
1 2
p
e p u

+ +

------(B)
We can also write these equations separately as Continuity, Momentum and Energy
Equation as,
Continuity Equation:

( ) ( )
0
u
t x

+


0
u
u
t x x


+ +

---------(2)
Momentum Equation:
Page 1 of 9

2
( ) ( )
0
u p u
t x
+
+



2
2 0
u p u
u u u
t t x x x



+ + + +



0
u u u p
u u u
t x x t x x


_
+ + + + +


,


1
0
u u p
u
t x x

+ +

---------(3)
Energy Equation:

( ) {( ) }
0
e e p u
t x
+
+



2 3
1 1
0
1 2 1 2
p pu
u u
t x



1 1
+ + +
1 1

] ]


2 3 2
1 1 1 3
0
1 2 1 1 2 2
p u u p u
u u p u u u
t t t x x x x



_ _ _
+ + + + + +


, , ,



2
1 1 1
0
2 1
u u u p p p u
u u u u u p
t x x t x x t x x



_ _ _ _
+ + + + + + + +


, , , ,


0
p p u
u p
t x x


+ +

---------(4)
Thus, equations, (2), (3) and (4) can be written in terms of the primitive variable
vector, V, given by,
V u
p
_



,
Thus, the 1DEE can also be written as,

0
V V
A
t x

+

-----------(5)
where,

0
1
0
0
u
A u
p u



,
-----(6)
Page 2 of 9
i.e.
0
1
0 0
0
u
u u u
t x
p p
p u

_
_ _




+





, ,
,
--------(7)
1DEE in conservation form: 0
U F
t x

+

-------- (1)
1DEE in non-conservation form: 0
U U
A
t x

+

---------(8)
where,
F
A
U

Flux Jacobian
It will be shown later that,

1
A N AN

, so that both Aand

A
have identical eigenvalues.
Expressing F in terms of the dependent variables
We have,
1
2
3
F
F F
F
_



,

2
( )
u
p u
e p u

_

+


+
,
and,
1
2
3
U
U U
U
_



,


u
e



,
So,
1 1
F u U
------------(9)
Now,
2
2
F p u +
( )
2 2
1
1
2
e u u
_
+

,
[p is substituted from (A)]
( )
2 2
2 2
3
1 1
1
1
2
U U
U
U U

_
+

,
( )
2
2
2 3
1
3
, 1
2
U
or F U
U

_
+ +

,
--------(10)
Page 3 of 9
And,
3
( ) F e p u +

pu eu +

( )
2
2 3 2 2
3
1 1 1
1
1
2
U U U U
U
U U U

_ _
+

, ,


3
2 3 2
3 2
1 1
1
,
2
U U U
or F
U U
_


,
-------(11)
So, the flux vector can be expressed in terms of the dependent variables as,

2
2
2
3
1
3
2 3 2
2
1 1
3
( 1)
2
1
2
U
U
F U
U
U U U
U U


_



_
+

,


_



,
,
--------(12)
The Flux Jacobian
So,
1 1 1
1 2 3
1 2 3 2 2 2
1 2 3 1 2 3
3 3 3
1 2 3
( , , )
( , , )
F F F
U U U
F F F F F F F
A
U U U U U U U
F F F
U U U
_














,
--------(13)
A is called Flux Jacobian, whose elements are given by,

1
1
0
F
U


1
2
1
F
U


1
3
0
F
U


2
2 2 2
2
1 1
3 3
2 2
F U
u
U U
_ _

, ,

2 2
2 1
(3 ) (3 )
F U
u
U U


( )
2
3
( 1) 1
F
U

( )
3
3 3 2 3 2
2 3
1 1 1
( 1) 1
F U U U eu
u
U U U


2 2
3 3 2
2
2 1 1
3( 1) 3( 1)
2 2 2
F U U e u
U U U


3 2
3 1
F U
u
U U

Page 4 of 9
Thus,

2
3 2
0 1 0
3
(3 ) ( 1)
2
3
( 1) ( 1)
2
A u u
eu e
u u u





_


_


,


+

,
----------(14)
Also,
2
1
2
p
H c T u +

2
1
,
2
v
or H c T u + ---------(15)
Now, ( )
3
1
eu
u

+
( )
2 3
1
1
2
v
u c T u u
_
+ +

,
( )
3 3
1
2
v
u c T u u

+
( )
3
2 3 3
1
1
2 2 2
v
u
u c T u u u


_
+ + +

,

( )
3 3
( 1)
1
2
Hu u u

+

3
( 1)
2
u Hu


Also, ( )
2
3
1
2
e
u



( )
2 2
1 3
1
2 2
v
c T u u
_
+

,
( )
2 2 2 2
1 1 3
1
2 2 2 2
v
c T u u u u


_
+ +

,

( )
2 2
1 3
1
2 2
H u u

_
+

,
( )
2
1 H u
Thus, we can write,
2
3 2
0 1 0
3
(3 ) ( 1)
2
1
( 1)
2
A u u
u Hu H u u


_



_


,


_



, ,
-------(16)

Page 5 of 9
The right eigenvectors are given by,

1
1
r u a
H ua
_



,

2
2
1
1
2
r u
u
_




,
And,
3
1
1 r u
H ua
_

+


+
,
It can be shown that F AU ; homogeneity property. This property is made use by
Steger and Warming to construct their Flux Vector Splitting Scheme.
Similarity of the matrices Aand

A
Now, 1DEE in primitive variables is,

0
V V
A
t x

+

-----(5)\
Now,
2
1
1 2
dU d u
p
u

_





+

,

2
1 2
d
du ud
dp u
udu d

_


+


+ +

,
Page 6 of 9

2
1 0 0
0
1
2 1
d
u du
dp
u
u

_

_






,

,
or, dU NdV ------(17)
where,
2
1 0 0
0
1
2 1
N u
u
u

,
-------(18)
Thus,
U V
N
t t


--------(19)
And,
U V
N
x x


--------(20)
Now, multiplying both sides of equation (5) by N

0
V V
N N A
t x

+

0
V V
N N A
t x

+

( )
1
0
U V
N AN N
t x

( )
1
0
U U
N AN
t x


+

--------(21)
Comparing equation (21) with equation (8), we see that,

1
A N AN

---------(22)
Thus,

( )
( ) A A
----------(23)
Proof:
If x is an eigenvector of

A
, then,

Ax x

N Ax Nx
Let,
1
x N y

So,

1 1
N AN y NN y


Ay y
Thus, is an eigenvalue of A also, with corresponding eigenvector as,
y Nx
Page 7 of 9
To find the eigenvalues of A
As

A
is simpler, so we shall find the eigenvalues of

A
, since

( )
( ) A A
.
Characteristic equation is,

| | 0 A I

0 0
1
det 0 0
0
u
u
p u

,

( ) ( )
2
0
p
u u

1

1
]
( ) ( )
2
2
0 u u a
1

]

, , , , u or u a or u a +
The system,
0
t x
U AU +
is called hyperbolic, if A has real eigenvalues and
distinct eigenvectors. The system is hyperbolic if A is diagonizable with real eigenvalues.
So,
( ) , A u u a t
Right eigenvectors of A:
. .
i i
Ar r
Left eigenvectors of A:
. .
i i
l A l
Matrix of the right eigenvectors of A: R = [r
1
r
2
r
3
]
Matrix of the left eigenvectors of A:
1
2
3
l
L l
l
_



,
A can be written in Jordan Canonical Form as,
1 1
A R L R R L L


,
So,
1 1
, , L R and R L


And, Diagonal matrix with the eigenvalues of A as the diagonal elements.

1
2
3
0 0
0 0
0 0



,
Euler equation is written as,
Page 8 of 9
0
U U
A
t x

+


1
0
U U
R R
t x


+


1 1
0
U U
R R
t x


+

If we freeze the Jacobian (locally), then,


0
U U
t x

+

,
which represents the wave equation (decoupled):


0
i i
i
U U
t x


+

, where,
1, 2, 3 i
----------(24)
Thus from equation (24) we can infer that the eigenvalues,
i

, of the flux Jacobian


A correspond to wave speeds.

i
U is invariant along direction,
i
dx
dt
.

i
U is called Riemann invariant.
It is seen that for
| | u a <
, there are waves present, which move in opposite
directions. As the numerical scheme must upwind difference the waves, so in such cases,
one-sided or central differencing for the flux fails.
Page 9 of 9

Vous aimerez peut-être aussi