R5375

IMPLICIT METHOD FOR THE 3D

EULER EQUATIONS
MARTIN KYNCL
JAROSLAV PELANT
VZLU

AEROSPACE RESEARCH AND TEST ESTABLISHMENT

BERANOVY

CH 130, 199 05 PRAGUE, CZECH REPUBLIC

c 2012
R5375
Contents
1 INTRODUCTION 4
2 THE EULER EQUATIONS 4
3 FVM DISCRETIZATION OF THE INVISCID PROBLEM 5
3.1 Inner Face Values Discretization (Riemann Solver) . . . . . . . . . . . . . . . . . . . . 8
3.2 Vijayasundaram Numerical Flux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.3 Boundary Face Values Discretization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
4 THE RIEMANN PROBLEM FOR THE SPLIT EULER EQUATIONS 11
4.1 The Incomplete Initial Value Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5 BOUNDARY CONDITION BY PREFERENCE OF PRESSURE 14
6 BOUNDARY CONDITION BY PREFERENCE OF VELOCITY 15
6.1 B.C. for the Impermeable Wall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
7 BOUNDARY CONDITION BY PREFERENCE OF TOTAL QUANTITIES AND DIREC-
TION OF VELOCITY 16
8 SOFTWARE AND EXAMPLES 19
9 Examples 23
9.1 Laval Nozzle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
9.2 A Body in a Channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
10 CONCLUSIONS 26
List of Figures
3-1 Coordinate transformation for the inner edges in 2D. . . . . . . . . . . . . . . . . . . . 8
3-2 Coordinate transformation for the boundary edge in 2D. . . . . . . . . . . . . . . . . . . 11
4-1 Structure of the solution of the Riemann problem (16),(17),(18) . . . . . . . . . . . . . . 12
4-2 The incomplete Riemann problem for the 1D Euler equations. General form of the solution. 14
5-1 Algorithm for the solution of the problem (16),(17),(31),(33),(32),(34),(35) at the half
line S
B
= {(0, t); t > 0}. The value of the pressure p

is prescribed. Possible situations

are illustrated by the pictures showing
L

HTL

L
region. The sought boundary
state located at the time axis, marked red. . . . . . . . . . . . . . . . . . . . . . . . . . 15
6-1 Algorithm for the solution of the problem (16),(17),(36),(33),(37),(34),(35) at the half
line S
B
= {(0, t); t > 0}. Possible situations are illustrated by the pictures showing
the region
L

HTL

L
with the sought boundary state located at the time axis. . . 16
7-1 Coordinate transformation for the boundary edge
ij
in 2D. The axis x
1
is in the opposite
direction to the prescribed direction of the velocity. . . . . . . . . . . . . . . . . . . . . 17
7-2 Algorithm for the solution of the problem (24)-(29),(40)-(44) at the half line S
B
=
{(0, t); t > 0}. In the case of failure, the problem doesnt have a solution, and the value
for u

is chosen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
8-1 10% cylindrical bump, initial mesh and boundary numbering. . . . . . . . . . . . . . . . 19
8-2 C3DEEI: allowed elements, vertices numbering. . . . . . . . . . . . . . . . . . . . . . . 20
8-3 3D inviscid ow computation (coarse mesh). Mach number and pressure isolines. (C3D) 23
9-1 3D inviscid computation, coarse mesh with 1896 prismatic elements. Pressure and Mach
number isolines, real time of the computation for chosen methods. (C3DEEI) . . . . . . 24
- 2 -
R5375
9-2 3D inviscid computation, coarse mesh with 8208 hexahedral elements. Pressure and
Mach number isolines (C3DEEI). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
9-3 3D inviscid computation, mesh with 109490 prismatic elements. Pressure and Mach
number isolines (C3DEEI). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
9-4 3D inviscid computation, mesh with 118835 prismatic elements. Pressure and Mach
number isolines, real time of the computation for chosen methods. (C3DEEI) . . . . . . 26
- 3 -
R5375
1 INTRODUCTION
The physical theory of the compressible uid motion is based on the principles of conservation laws of
mass, momentum, and energy. The mathematical equations describing these fundamental conservation
laws form a system of partial differential equations. In this article we will work with the Euler equations,
and we focus on the numerical solution of these equations. The correct design of the boundary condi-
tions plays also an important role in the numerical modeling of the processes involved. We choose the
well-known nite volume method to discretize the analytical problem, represented by the system of the
equations in generalized (integral) form. To apply this method we split the area of the interest into the
elements, and we construct a piecewise constant solution in time. The crucial problem of this method
lies in the evaluation of the so-called uxes through the edges/faces of the particular elements. In order
to compute these uxes, various methods can be used. One of the most accurate method (and perhaps
the most accurate method) is based on the solution of the so-called Riemann problem for the 2D/3D split
Euler equations. Unfortunately, the exact solution of this problem cannot be expressed in a closed form,
and has to be computed by an iterative process (to given accuracy). Therefore this method is also one
of the most demanding. Nevertheless, on account of the accuracy of the Riemann solver, we decided
to use the analysis of the exact solution also for the discretization of the uxes through the boundary
edges/faces. The right-hand side initial condition, forming the local Riemann problem, is not known at
the boundary faces. In some cases (far eld boundary) it is wise to choose the right-hand side initial
condition here as the solution of the local Riemann problem with given far eld values, which gives
better results than the solution of the linearized Riemann problem, see [2]. Another boundary condition
based on the exact Riemann problem solution, simulating the impermeable wall on move, was shown in
[20, pages 221-225], where the right-hand side initial condition is constructed in a special way, in order
to obtain the desired solution. We show, that the right-hand side initial condition for the local problem
can be partially replaced by the suitable complementary condition. This idea was introduced by RNDr.
Jaroslav Pelant, CSc. in [17]. We construct own algorithms for the solution of the boundary problem,
and we use it in the numerical examples.
2 THE EULER EQUATIONS
Here we showthe Euler equations in 3D. The systemdescribes the conservation laws of mass, momentum,
and energy. We present the Euler equations in the differential form
w
t
+
3

s=1
f
s
(w)
x
s
= 0 in Q
T
= (0, T). (1)
Here w = w(x, t) is the state vector, x , t denotes the time, Q
T
is called a space-time cylinder, f
s
are the inviscid (Euler) uxes
w = (w
1
, w
2
, . . . , w
N+2
)
T
= (, v
1
, . . . , v
N
, E)
T
,
f
s
(w) = (v
s
, v
s
v
1
+
s1
p, . . . , v
s
v
N
+
sN
p, (E +p) v
s
)
T
with v = (v
1
, . . . , v
N
)
T
denoting the velocity vector, being the density, p the pressure, E the total
energy. The system (1):

t
_
_
_
_

v
1
v
2
v
3
E
_
_
_
_
+

x
1
_
_
_
_
v
1
v
2
1
+ p
v
1
v
2
v
1
v
3
(E + p) v
1
_
_
_
_
+

x
2
_
_
_
_
v
2
v
2
v
1
v
2
2
+ p
v
2
v
3
(E + p) v
2
_
_
_
_
+

xs
_
_
_
_
v
3
v
3
v
1
v
3
v
2
v
2
3
+ p
(E + p) v
3
_
_
_
_
= 0,
The equations (1) represent a system of hyperbolic partial differential equations. We assume that the
partial derivatives exist, such that the differential formulation (1) makes sense. Within the numerical
solution of this problem we usually consider the equations in the more general integral form, which
- 4 -
R5375
allows the discontinuities in the solution. It is necessary to specify the closure conditions. We consider
the thermodynamical relations yielding from the caloric equation of state
p = ( 1)(E v
2
/2). (2)
The system of the Euler equations is rotationally invariant (see e.g. [5, page 108]). We can choose a
new Cartesian coordinate system ( x
1
, x
2
, x
3
) such that the origin is at the point , and the axis x
1
is in
the direction of a chosen unit vector n = (n
1
, n
2
, n
3
), |n| = 1.
This transformation can be written
_
_
x
1
x
2
x
3
_
_
= Q
0
_
_
x
1
x
2
x
3
_
_
+ , with Q
0
=
_
_
n
o
p
_
_
=
_
_
n
1
n
2
n
3
o
1
o
2
o
3
p
1
p
2
p
3
_
_
. (3)
Here Q
0
is the rotation matrix, the elements of the rotation matrix are not all independent. The vectors
n, o, p must have the properties
|n| = |o| = 1, n o = 0, p = n o.
We can choose the vector o = (o
1
, o
2
, o
3
) such that n o = 0, |o| = 1. And then the vector p =
(p
1
, p
2
, p
3
) is determined by p = n o. The axis x
2
points in the direction of the vector o, and the
axis x
3
in the direction of the vector p. The matrix Q
0
is orthogonal, and the vectors are orthogonal unit
vectors. The inverse of the orthogonal matrix Q
0
is Q
1
0
= Q
T
0
. The transformation of the state vector
w yields the state vector q in the new coordinates. It is
q = Q w,
w = Q
1
q,
(4)
with
Q =
_
_
_
_
_
_
1 0 0 0 0
0 n
1
n
2
n
3
0
0 o
1
o
2
o
3
0
0 p
1
p
2
p
3
0
0 0 0 0 1
_
_
_
_
_
_
, Q
1
=
_
_
_
_
_
_
1 0 0 0 0
0 n
1
o
1
p
1
0
0 n
2
o
2
p
2
0
0 n
3
o
3
p
3
0
0 0 0 0 1
_
_
_
_
_
_
.
Using this transformation of the coordinates, the Euler Equations keep the form
q
t
+
3

s=1
f
s
(q)
x
s
= 0.
3 FVM DISCRETIZATION OF THE INVISCID PROBLEM
Here we describe the so-called nite volume discretization of the problem (1) in the domain . By
h
let
us the denote the polyhedral approximation of . The system of the closed polyhedrons with mutually
disjoint interiors D
h
= {D
i
}
iJ
, where J Z
+
= {0, 1, . . .} is an index set and h > 0, will be called a
nite volume mesh. This system D
h
approximates the domain

h
=
_
iJ
D
i
.
The elements D
i
D
h
are called nite volumes. For two neighboring elements D
i
, D
j
we set

ij
= D
i
D
j
=
ji
.
- 5 -
R5375
Here we will work with the so-called regular meshes, i.e. the intersection of two arbitrary (different)
elements is either empty or it consists of a common vertex or a common edge or a common face (in 3D).
The boundary D
i
of each element D
i
is
D
i
=
_

ij

D
i

ij
. (5)
Here the set
D
i
= {
ij
;
ij
D
i
} forms the boundary D
i
. By n
ij
let us denote the unit outer normal
to D
i
on
ij
. Let us construct a partition 0 = t
0
< t
1
< . . . of the time interval [0, T] and denote the
time steps
k
= t
k+1
t
k
. We integrate the system (1) over the set D
i
(t
k
, t
k+1
). With the integral
form of the equations we can study a ow with discontinuities, such as shock waves, too.
_
D
i
_
t
k+1
t
k
w
t
dx dt +
_
t
k+1
t
k
_
D
i
3

s=1
f
s
(w)
x
s
dx dt = 0 (6)
Using the Greens theorem on D
i
it is
_
D
i
3

s=1
f
s
(w)
x
s
dx =
_
D
i
3

s=1
f
s
(w)n
s
dS. (7)
Here n = (n
1
, . . . , n
3
) is the unit outer normal to D
i
. Further we use (5), and we rewrite (6)
_
Di
(w(x, t
k+1
) w(x, t
k
)) dx +
_
t
k+1
t
k

ijD
i
_
ij
3

s=1
f
s
(w)(n
ij
)
s
dS dt = 0 (8)
We dene a nite volume approximate solution of the system studied (1) as a piecewise constant vector-
valued functions w
k
h
, k = 0, 1, . . . , where w
k
h
is constant on each element D
i
, and t
k
is the time instant.
By w
k
i
we denote the value of the approximate solution on D
i
at time t
k
. We approximate the integral
over the element D
i _
D
i
w(x, t
k
) dx |D
i
|w
k
i
. (9)
Further we proceed with the approximation of the uxes. Usually the ux

3
s=1
f
s
(w)(n
ij
)
s
|

ij
is
being approximated by a numerical ux at suitable time instant t
l
3

s=1
f
s
(w)(n
ij
)
s
|

ij
H(w
l
i
, w
l
j
, n
ij
),
with w
l
i
, w
l
j
denoting the approximate solution on the elements adjacent to the edge
ij
at the time
instant t
l
. In the case of a boundary face the vector w
l
j
has to be specied. Here we show the numerical
ux based on the solution of the Riemann problem for the split Euler equations (shown later in Section
4). By w
l

ij
let us denote the state vector w at the center of the edge
ij
at the time instant t
l
, and let
us suppose w
l

ij
is known. Evaluation of these values will be a question of the further analysis, here we
use them to approximate the integrals with the one-point rule
_

ij
3

s=1
f
s
(w(x, t
l
))(n
ij
)
s
dS |
ij
|
3

s=1
f
s
(w
l

ij
)(n
ij
)
s
. (10)
Now it is possible to approximate the system (8) by the following explicit nite volume scheme
|D
i
|(w
k+1
i
w
k
i
) +
k

ij

D
i
|
ij
|
3

s=1
f
s
(w
k

ij
)(n
ij
)
s
= 0. (11)
- 6 -
R5375
With this nite volume formula one computes the values of the approximate solution at the time instant
t
k+1
, using the values from the time instant t
k
, and by evaluating the values w
k

ij
at the faces
ij
. In
order to achieve the stability of the used method, the time step
k
must be restricted by the so-called CFL
condition, see [5]. The crucial problem of this discretization lies with the evaluation of the edge values
w
k

ij
. Or one deals with the problem of nding the face uxes H(w
k
i
, w
k
j
, n
ij
).
Here we will work with the implicit scheme
|D
i
|(w
k+1
i
w
k
i
) +
k

ij

D
i
|
ij
|H(w
k+1
i
, w
k+1
j
, n
ij
) = 0. (12)
The crucial problem of this discretization lies with the evaluation of the face uxes H(w
k+1
i
, w
k+1
j
, n
ij
).
One possibility is to use the linearization via the Taylor expansion of the vector function H(u, v, n).
We assume that the mapping H(u, v, n) is differentiable. Let
DH(u,v,n)
Du
,
DH(u,v,n)
Dv
be the Jacobi
matrix of the mapping H(u, ., .), H(., v, .)
DH
Du
=
_
H
i
u
k
_
5
i,k=1
=
_
_
_
_
_
_
_
H
1
u
1
H
1
u
2
H
1
u
3
H
1
u
4
H
1
u
5
H
2
u
1
H
2
u
2
H
2
u
3
H
2
u
4
H
2
u
5
H
3
u
1
H
3
u
2
H
3
u
3
H
3
u
4
H
3
u
5
H
4
u
1
H
4
u
2
H
4
u
3
H
4
u
4
H
4
u
5
H
5
u
1
H
5
u
2
H
5
u
3
H
5
u
4
H
5
u
5
_
_
_
_
_
_
_
,
DH
Dv
=
_
H
i
v
k
_
5
i,k=1
We will use shorter notation H|
k
ij
= H(w
k
i
, w
k
j
, n
ij
),
DH
Du

k
ij
=
DH
Du
(w
k
i
, w
k
j
, n
ij
),
DH
Dv

k
ij
=
DH
Dv
(w
k
i
, w
k
j
, n
ij
). We approximate the ux H|
k+1
ij
as
H|
k+1
ij
H|
k
ij
+
DH
Du

k
ij
(w
k+1
i
w
k
i
) +
DH
Dv

k
ij
(w
k+1
j
w
k
j
)
Then the scheme (12) can be writen as
_
_
I +

k
|D
i
|

ijD
i
|
ij
|
DH
Du

k
ij
_
_
w
i
+

k
|D
i
|

ij D
i
|
ij
|
DH
Dv

k
ij
w
j
=

k
|D
i
|

ij D
i
|
ij
|H|
k
ij
(13)
Here we used the notation w
i
= w
k+1
i
w
k
i
, i = 1, ..., n with n denoting the number of nite
volumes. The system in a matrix form
Aw = b. (14)
withw = (w
1
, w
2
, ..., w
n
)
T
, b = (

k
|D
1
|

1l

D
1
|
1l
| H|
k
1l
, ...,

k
|Dn|

nl

Dn
|
nl
| H|
k
nl
)
T
,
and A is the block matrix with blocks
a
ij
=
_

_
I +

k
|D
i
|

l
D
i
|
il
|
DH
Du

k
il
, i = j,

k
|D
i
|
|
ij
|
DH
Dv

k
ij
, i = j, j S(i),
0, i = j, j / S(i).
The contributions of the face
ij
(between the elements D
i
, D
j
) into the system Aw = b
_
_
_
_
_
_
_
_
_
. .
I +

k
|D
i
|

il

D
i
|
il
|
DH
Du

k
il

k
|D
i
|
|
ij
|
DH
Dv

k
ij
. .
. .

k
|D
j
|
|
ji
|
DH
Dv

k
ji
. . I +

k
|D
j
|

jl

D
j
|
jl
|
DH
Du

k
jl
. . .
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
.
w
i
.
.
w
j
.
_
_
_
_
_
_
=
_
_
_
_
_
_
_
_
.

k
|D
i
|

il

D
i
|
il
| H|
k
il
.
.

k
|D
j
|

jl

D
j
|
jl
| H|
k
jl
.
_
_
_
_
_
_
_
_
- 7 -
R5375
Further we assume that the ux H is conservative: H(u, v, n) = H(u, v, n), hence
H|
k
ji
= H|
k
ij
,
DH
Du

k
ji
=
DH
Dv

k
ij
,
DH
Dv

k
ji
=
DH
Du

k
ij
.
The construction of the matrix A and the right side b can be made as follows. We set A = I, b = 0. Then
the ux H|
k
ij
and its Jacobians
DH
Du

k
ij
,
DH
Dv

k
ij
are computed for each face
ij
, updating the values in
A, b :
b
i
= b
i
+

k
|D
i
|
|
ij
| H|
k
ij
, b
j
= b
j

k
|D
j
|
|
ij
| H|
k
ij
,
a
ii
= a
ii
+

k
|D
i
|
|
ij
|
DH
Du

k
ij
, a
ij
=

k
|D
i
|
|
ij
|
DH
Dv

k
ij
,
a
jj
= a
jj

k
|D
j
|
|
ij
|
DH
Dv

k
ij
, a
ji
=

k
|D
j
|
|
ij
|
DH
Du

k
ij
.
The main problem of this discretization lies with the evaluation of the face uxes H|
k
ij
and its Jacobian
DH
Du

k
ij
,
DH
Dv

k
ij
.
3.1 Inner Face Values Discretization (Riemann Solver)
One possibility of the face ux evaluation is to approximate the face values w
k

ij
and then compute the
numerical ux
H|
k
ij
=
3

s=1
f
s
(w
k

ij
)(n
ij
)
s
. (15)
To approximate the face values w
k

ij
at time instant t
k
we solve the simplied system (16) in the vicinity
of the face
ij
in time with the initial condition formed by the state vectors w
k
i
and w
k
j
. Using the
rotational invariance of the Euler equations, the system (1) is expressed in a new Cartesian coordinate
system x
1
, x
2
, x
3
with the origin at the center of the gravity of
ij
and with the new axis x
1
in the
direction of n = (n
1
, n
2
, n
3
), given by the face normal n = n
ij
.
x
1
x
2

ij
D
i
D
j
x
1
n
ij
Fig. 3-1: Coordinate transformation for the inner edges in 2D.
Then the derivatives with respect to x
2
, x
3
are neglected and we get the so-called split 3DEuler equations,
see [5, page 138]:
q
t
+
f
1
(q)
x
1
= 0. (16)
The values w
k
i
and w
k
j
adjacent to the face
ij
are known, forming the initial conditions
q( x
1
, 0) = q
L
= Q w
k
i
, x
1
< 0, (17)
q( x
1
, 0) = q
R
= Q w
k
j
, x
1
> 0. (18)
- 8 -
R5375
The transformation matrix Q is dened in (4). In this work we will refer to these initial conditions as
to the left-hand side initial condition (17) and the right-hand side initial condition (18). The problem
(16), (17), (18) has a unique solution in (, ) (0, ), the analysis can be found in [5, page
138], we will show the analysis of this problem later in Section 4. Let q
RS
(q
L
, q
R
, x
1
, t) denote the
solution of this problem at the point ( x
1
, t). We are interested in the solution of this local problem at the
time axis, which is the sought solution in the local coordinates q

ij
= q
RS
(q
L
, q
R
, 0, t). The backward
transformation (4) of the state vector q

ij
into the global coordinates is
w
k

ij
= Q
1
q

ij
= Q
1
q
RS
(Q w
k
i
, Q w
k
j
, 0, t). (19)
The constructed numerical ux can be written as
H(u, v, n) :=
3

s=1
f
s
_
Q
1
q
RS
(Q u, Q v, 0, t)
_
n
s
. (20)
The described process of nding the face values w
k

ij
is independent on the choice of the vectors o, p,
determining the transformation matrix Q dened in (4), see [7].
3.2 Vijayasundaram Numerical Flux
Another possibility of the face ux discretization is to use some of the properties of the solved system
(1). The system is a rst order quasilinear system of partial differential equations, see [4, page 385], and
the ux can be written
3

s=1
f
s
(w)n
s
=
3

s=1
n
s
A
s
(w)w, with A
s
=
Df
s
Dw
.
Using the rotational invariance of the equations of the inviscid ow and the transformation of the
coordinates dened in (4) it is
3

s=1
n
s
A
s
(w)w = Q
1
A
1
(Qw)Qw = Q
1
TT
1
Qw =

T

T
1
w.
Here A
1
(Qw) = TT
1
is diagonizable matrix (due to the hyperbolicity of the problem), shown in [5].
Matrices ,

T = Q
1
T,

T
1
= T
1
Q depend on w as
= diag(u a, u, u, u, u +a),

T =
_
_
_
_
_
_
1 1 1 1 1
U n
1
a U U o
1
a U p
1
a U +n
1
a
V n
2
a V V o
2
a V p
2
a V +n
2
a
W n
3
a W W o
3
a W p
3
a W +n
3
a
h au
|v|
2
2
|v|
2
2
av
|v|
2
2
aw h +au
_
_
_
_
_
_
,

T
1
=
1
a
2
_
_
_
_
_
_
1
2
_
1
2
|v|
2
+ua
_

1
2
(n
1
a +
1
U)
1
2
(n
2
a +
1
V )
1
2
(n
3
a +
1
W)
1
2
a
2
a(v +w)
1
2
|v|
2

1
U + (o
1
+ p
1
)a
1
V + (o
2
+p
2
)a
1
W + (o
3
+p
3
)a
1
va o
1
a o
2
a o
3
a 0
wa p
1
a p
2
a p
3
a 0
1
2
_
1
2
|v|
2
ua
_
1
2
(n
1
a
1
U)
1
2
(n
2
a
1
V )
1
2
(n
3
a
1
W)
1
2
_
_
_
_
_
_
.
Here U = w
2
/w
1
, V = w
3
/w
1
, W = w
4
/w
1
are the components of the velocity in global coordinates,
and u = n
1
U + n
2
V +n
3
W, v = o
1
U + o
2
V +o
3
W, w = p
1
U + p
2
V +p
3
W are the components of
the velocity in local coordinates x
1
, x
2
, x
3
dened in (3). Further a =
_
p/ is the speed of the sound,
h = a
2
/
1
+
|v|
2
2
is the total specic enthalpy, and
1
= 1.
- 9 -
R5375
Now we approximate the ux

3
s=1
f
s
(w
k

ij
)(n
ij
)
s
=

T

T
1
(w
k

ij
)w
k

ij
H|
k
ij
H|
k
ij
= P
+
_
w
k
i
+w
k
j
2
, n
_
w
k
i
+P

_
w
k
i
+w
k
j
2
, n
_
w
k
j
, (21)
with
P
+
=

T
+

T
1
, P

=

T

T
1
,

= diag(

1
,

2
,

3
,

4
,

5
),
+
i
= max(
i
, 0),

i
= min(
i
, 0).
The Vijayasundaram numerical ux can be written as (see [4, page 456] )
H(u, v, n) = P
+
_
u +v
2
, n
_
u +P

_
u +v
2
, n
_
v.
The Jacobi mapping of such a numerical ux at
ij
and time instant t
k
can be approximated by
DH
Du

k
ij
,
DH
Dv

k
ij
with
DH
Du

k
ij
= P
+
_
w
k
i
+w
k
j
2
, n
_
,
DH
Dv

k
ij
= P

_
w
k
i
+w
k
j
2
, n
_
. (22)
Inpractical applications we can choose the directiono =
_
(n
2
/
_
1 n
2
3
, n
1
/
_
1 n
2
3
, 0), n
3
n
2
,
(n
3
/
_
1 n
2
2
, 0, n
1
/
_
1 n
2
2
), n
3
> n
2
,
and then compute p = n o.
3.3 Boundary Face Values Discretization
For the boundary faces we choose the discretization (15), where w
k

ij
must be given or computed
H|
k
ij
=
3

s=1
f
s
(w
k

ij
)(n
ij
)
s
.
Let
ij
be the face of the element D
i
laying at the boundary of the computational area. To approximate
the face values w
k

ij
at time instant t
k
we solve the incomplete simplied system (16) in the vicinity of
the face
ij
in time with the initial condition determined by the state vector w
k
i
. Similarily to Section 3.1
we use the rotational invariance of the Euler equations, and we express the system (1) in a new Cartesian
coordinate system x
1
, x
2
, x
3
with the origin at the center of gravity of
ij
and with the new axis x
1
in the
direction of vector n = (n
1
, n
2
, n
3
). The vector n corresponds to the face normal vector n
ij
pointing
out of the volume D
i
, however for certain boundary conditions it is convenient to dene n in a different
way (still pointing out of the volume D
i
). Then the derivatives with respect to x
2
, x
3
are neglected and
we get the so-called split 3D Euler equations (16):
q
t
+
f
1
(q)
x
1
= 0.
The state vector w
k
i
on the element D
i
adjacent to the face
ij
is known, forming the initial condition
for the local problem (16),(17)
q( x
1
, 0) = q
L
= Q w
k
i
, x
1
< 0.
The transformation matrix Qis dened in (4). We seek the boundary state q

ij
= q(0, t). Once the local
state vector q

ij
is known, we use transformation (4) to evaluate the face state vector w
k

ij
in global
coordinates, it is w
k

ij
= Q
1
q

ij
.
- 10 -
R5375
x
1
x
2

ij
D
i
x
1
n
ij
Fig. 3-2: Coordinate transformation for the boundary edge in 2D.
We suppose that the solution of the problem (16),(17),(23) in
B
is a solution of the problem
(16),(17),(18) in
B
, described later in Section 4, for some (unknown) q
R
satisfying the pressure
positivity condition (30). By adding properly chosen equations into the system (16),(17) it is possible
to reconstruct the boundary state q
B
such that the system (16),(17),(23) has a unique solution in
B
,
see Section 4. We will refer to these added equations as to complementary conditions. The suitable
complementary conditions may give the unique boundary state q
B
, as the solution of the incomplete
system (16),(17) at the half line S
B
= {(0, t); t > 0}. Several choices of the complementary conditions
will be discussed further. Here we introduce the initial-boundary value problem for the 3D split Euler
equations. We consider the system (16) in the set
B
= (, 0) (0, ), equipped with the initial
condition (17), and the boundary condition
q(0, t) = q
B
, t > 0. (23)
The problem of the boundary condition is to choose the boundary state q
B
such a way that the system
(16),(17),(23) is well-posed, i.e. it has a unique solution (entropy weak) in the considered set
B
. We
suppose that the solution of the problem (16),(17),(23) in
B
is a solution of the problem (16),(17),(18) in

B
, described later in Section 4, for some (unknown) q
R
satisfying the pressure positivity condition (30).
By adding properly chosen equations into the system (16),(17) it is possible to reconstruct the boundary
state q
B
such that the system (16),(17),(23) has a unique solution in
B
, see Section 4. We will refer to
these added equations as to complementary conditions. The suitable complementary conditions may
give the unique boundary state q
B
, as the solution of the incomplete system (16),(17) at the half line
S
B
= {(0, t); t > 0}. Several choices of the complementary conditions will be discussed further in the
Section 5 - 7.
4 THE RIEMANN PROBLEM FOR THE SPLIT EULER EQUATIONS
For many numerical methods dealing with the two or three dimensional equations, describing the com-
pressible ow, it is useful to solve the split Riemann problem. Here we will follow in 3D, analogously
it is possible to proceed in 2D. It is the system of the equations (16), equipped with the initial conditions
(17),(18)
q
t
+
f
1
(q)
x
1
= 0, q( x
1
, 0) =
_
q
L
, x
1
< 0,
q
R
, x
1
> 0.
Split means here that we still have 5 equations in 3D, but the dependence on x
2
, x
3
is neglected,
and we deal with the system for one space variable x
1
. The system (16) is considered in the set
Q

= (, ) (0, +). The solution of this simplied system is useful when dealing with the
inviscid uxes within various numerical methods.
The solution of this problem is fundamentally the same as the solution of the Riemann problem for
the 1D Euler equations, see [5, page 138]. In fact, the solution for the pressure, the rst component of
- 11 -
R5375
the velocity, and the density is exactly the same as in one-dimensional case. It is a characteristic feature
of the hyperbolic equations, that there is a possible raise of discontinuities in solutions, even in the case
when the initial conditions are smooth, see [3, page 390]. Here the concept of the classical solution is too
restrictive, and therefore we seek a weak solution of this problem. To distinguish physically admissible
solutions fromnonphysical ones, entropy condition must be introduced, see [3, page 396]. By the solution
of the problem (16),(17),(18) we mean the weak entropy solution of this problem in Q

. The analysis
to the solution of this problem can be found in many books, i.e. [5], [3], [20]. The general theorem
on the solvability of the Riemann problem can be found in [5, page 88]. Here we summarize, that the
problem has a unique solution for certain choice of the initial conditions. This solution can be written
for t > 0 in the similarity form q( x
1
, t) = q( x
1
/t), where q( x
1
/t) : IR IR
3
([5, page 82]). Now
we will show the form of the possible solution q = q( x
1
, t) of the Riemann problem (16),(17),(18). The
solution is piecewise smooth and its general form can be seen in Fig. 4-1, where the system of half lines
is drawn. These half lines dene regions, where q is either constant or given by a smooth function. Let

t
x
1
0

x
1
t
=s
HL
s
TL
=
x
1
t
x
1
t
=u

x
1
t
=s
TR
s
HR
=
x
1
t

HTL

L

R

HTR
Fig. 4-1: Structure of the solution of the Riemann problem (16),(17),(18)
us dene the open sets called wedges (see Fig. 4-1):

L
= {( x
1
, t);
x
1
t
< s
HL
; t > 0},

HTL
= {( x
1
, t); s
HL
<
x
1
t
< s
TL
; t > 0},

L
= {( x
1
, t); s
TL
<
x
1
t
< u

; t > 0},

R
= {( x
1
, t); u

<
x
1
t
< s
TR
; t > 0},

HTR
= {( x
1
, t); s
TR
<
x
1
t
< s
HR
; t > 0},

R
= {( x
1
, t); s
HR
<
x
1
t
; t > 0}.
We will refer to the set
HTL
as to the left wave, and the set
HTR
will be called the right wave. The
solution in
HTL
and in
HTR
is continuous. Let us denote
q|

HTL
= q
HTL
, q|

HTR
= q
HTR
.
The solution in
L
,
L
,
R
,
R
is constant (see e.g. [5, page 128]), we write
q|

L
= q
L
,
q|

L
= q
L
,
q|

R
= q
R
,
q|

R
= q
R
.
It is more convenient to use the vector of primitive variables (, u, v, w, p) rather then the vector of
conservative variables (, u, v, w, E) in solving the Riemann problem. The exact solution of the
Riemann problem has three waves in general, illustrated in Fig. 4-1. The wedges
L
and
L
are
separated by the left wave (either 1-shock wave, or 1-rarefaction wave). There is a contact discontinuity
between the regions
L
and
R
. Wedges
R
and
R
are separated by the right wave (either 3-shock
wave, or 3-rarefaction wave).
For a certain choice of the initial condition the solution is the same on either side of the particular
wave, and the wave diminishes. This situation is covered in the general example with three waves
- 12 -
R5375
in the solution. For example, given the initial condition with q
L
= q
R
, there is no discontinuity in
the solution. General analogy to this example is the solution with three waves separating the regions

L
,
L
,
R
,
R
with the same constant solution. The solution for the primitive variables can be
described as follows:
(, u, v, w, p)|

L
= (
L
, u
L
, v
L
, w
L
, p
L
),
(, u, v, w, p)|

L
= (
L
, u

, v
L
, w
L
, p

),
(, u, v, w, p)|

R
= (
R
, u

, v
R
, w
R
, p

),
(, u, v, w, p)|

R
= (
R
, u
R
, v
R
, w
R
, p
R
).
The folowing relations for these variables hold:
u = u
L
+
_
_
_
_
_
_
_
_
_
(p p
L
)
_
2
(+1)
L
p+
1
+1
p
L
_1
2
, p > p
L
2
1
a
L
_
1
_
p
p
L
_
(1)/2
_
, p p
L
(24)

L
=
_
_
_
_
_
_
_

L
1
+1
p
L
p
+1
p
L
p
+
1
+1
, p > p
L

L
_
p
p
L
_ 1

, p p
L
(25)
s
1
TL
=
_
_
_
u
L
a
L
_
+1
2
p
p
L
+
1
2
, p > p
L
u a
L
_
p
p
L
_
(1)/2
, p p
L
(26)
u = u
R
+
_
_
_
_
_
_
_
_
_
(p p
R
)
_
2
(+1)
R
p+
1
+1
p
R
_1
2
, p > p
R
,

2
1
a
R
_
1
_
p
p
R
_
(1)/2
_
, p p
R
(27)

R
=
_
_
_
_
_
_
_

R
p
p
R
+
1
+1
1
+1
p
p
R
+1
, p > p
R

R
_
p
p
R
_ 1

, p p
R
(28)
s
3
TR
=
_
_
_
u
R
+ a
R
_
+1
2
p
p
R
+
1
2
, p > p
R
u + a
R
_
p
p
R
_
(1)/2
, p p
R
(29)
Here a
L
=
_
p
L
/
L
, a
R
=
_
p
R
/
R
, and denotes the adiabatic constant. Further s
1
TL
denotes
unknown left wave speed, s
3
TR
unknown right wave speed. Note, that the system (24) - (29) is the
system of 6 equations for 6 unknowns p

,u

,
L

R
,s
1
TL
,s
3
TR
. The solution of this system leads to a
nonlinear algebraic equation, and one cannot express the analytical solution of this problem in a closed
form. The problem has a solution only if the pressure positivity condition is satised
u
R
u
L
<
2
1
(a
L
+a
R
). (30)
We will use some of these relations to construct and solve the initial-boundary value problem which will
be the original result of this work.
Remark
Once the pressure p

is known, the solution on the left-hand side of the contact discontinuity depends
only on the left-hand side initial condition (17). And similarily, with p

known, only the right-hand side

initial condition (18) is used to compute the solution on the right-hand side of the contact discontinuity.
4.1 The Incomplete Initial Value Problem
In this part we deal with the incomplete Riemann problem for the Euler equations described by the system
(16) equipped with the left-hand side initial condition (17). This problem is not uniquely solvable in
(, ) (0, ). The short analysis of this problem will help us to understand the construction of the
boundary conditions performed in Sections 5 - 7.
We have shown the unique entropy weak solution of the initial value problem (16),(17), equipped
with the initial condition (18) formed by an arbitrary state q
R
satisfying the pressure positivity condition
(30). Therefore we may seek the solution of the incomplete initial problem (16),(17) in the same general
form, consisting of 4 constant states q
L
, q
L
, q
R
, q
R
separated by 3 waves (see Fig. 4-1). Now the
primitive variables in the right region
R
are unknown.
Here we will point out the number of unknowns and the number of missing equations in particular
regions in order to compose a uniquely solvable system.
The solution in
L
(across 0 wave)
Here the solution
L
, u
L
, v
L
, w
L
, p
L
is known from the initial condition (17).
- 13 -
R5375

x
1
0

L
, u
L
, v
L
, w
L
, p
L
s
1
TL
(17)

R
, u
R
, v
R
, w
R
, p
R
s
3
TR
u

, p

L
v
L
w
L

R
v
R
w
R
Fig. 4-2: The incomplete Riemann problem for the 1D Euler equations. General form of the solution.
The solution in
L

HTL

L
(across 1 wave)
There are three unknowns in the region
L
. It is the density
L
, the pressure p

, and the velocity

u

. Also the speed s

1
TL
of the left wave determining the position of the region
HTL
is not
known. The solution components in
L

HTL

L
region must satisfy the system of equations
(24)-(26). It is a system of three equations for four unknowns. We have to add another equation in
order to get the uniquely solvable system in
L

HTL

L
.
The solution in
L

HTL

L

R
(accross 2 waves)
There are four unknowns determining the solution in
L

HTL

L
region: the density
L
,
the pressure p

, the velocity u

, and the speed of the left wave s

1
TL
. Further the density
R
and
the velocity components v
R
, w
R
in the
R
region are not known. We deal with seven unknowns,
whereas the system of three equations (24),(25),(26) must be satised. We have to add four
properly chosen equations into this system of equations in order to get uniquely solvable system
for 7 unknowns sought in
L

HTL

L

R
region.
The solution in
L

HTL

L

R

HTR

R
(accross 3 waves)
We are interested in 11 unknowns
R
,
L
, p

, u

,
R
, p
R
, u
R
, v
R
, w
R
, s
1
TL
, s
3
TR
. There are 6
equations (24)-(29), that must be satised. We have to add 5 equations in order to get the uniquely
solvable system.
5 BOUNDARY CONDITION BY PREFERENCE OF PRESSURE
In this section we construct the boundary condition, i.e. the state q
B
in (23), preferring the given value
for the pressure p
GIVEN
> 0. This boundary condition corresponds to the real-world problem, when
we deal with the experimentally obtained pressure distribution at the boundary. We add the following
complementary conditions into the system (16),(17)
p

:= p
GIVEN
, (31)
p
R
= p

, (32)

R
:=
GIVEN
, (33)
v
R
:= v
GIVEN
, (34)
w
R
:= w
GIVEN
. (35)
Here p

denotes the pressure in

L

R
part of the solution of the Riemann problem for the split
Euler equations, shown in Section 4. The conditions (31), (32) prescribe the given pressure wherever it
- 14 -
R5375
is possible, solution in
L
is governed by the condition (17). We seek the boundary state q
B
as the
unique solution of the problem (16),(17),(31)-(35) at the half line S
B
= {(0, t); t > 0}. The system
(24)-(29), (31)-(35) is uniquely solvable. The algorithm for the construction of the primitive variables

B
, u
B
, v
B
, w
B
, p
B
at the half-line S
B
is shown in Fig. 5-1. The complete analysis of this problem is
shown in [7]
p > p
L
p p
L
s
1
= u
L

p
L

L
_
+1
2
p
p
L
+
1
2
u = u
L
(p p
L
)
_
2
(+1)
L
p+
1
+1
p
L
_1
2

L
=
L
1
+1
p
L
p
+1
p
L
p
+
1
+1
a
L
=
_

p
L

L
, s
HL
= u
L
a
L
u = u
L
+
2
1
a
L
_
1
_
p
p
L
_
(1)/2
_
s
TL
= u a
L
_
p
p
L
_
(1)/2

L
=
L
_
p
p
L
_
1/
s
1
> 0 s
1
0 s
HL
0 s
HL
< 0
u 0 u < 0 s
TL
0 s
TL
< 0
u 0 u < 0
OUTLET OUTLET INLET OUTLET OUTLET OUTLET INLET
p
B
= p
L
u
B
= u
L
v
B
= v
L
w
B
= w
L

B
=
L
p
B
= p
u
B
= u
v
B
= v
L
w
B
= w
L

B
=
L
p
B
= p
u
B
= u
v
B
= v
R
w
B
= w
R

B
=
R
p
B
= p
L
u
B
= u
L
v
B
= v
L
w
B
= w
L

B
=
L
u
B
=
2

+
1
_
a
L
+

1
2
u
L
_
v
B
=
v
L
w
B
=
w
L
p
B
=
p
L
_
2

+
1
+

1
(

+
1
)
a
L
u
L
_
2

B
=

L
_
2

+
1
+

1
(

+
1
)
a
L
u
L
_
2

1
p
B
= p
u
B
= u
v
B
= v
L
w
B
= w
L

B
=
L
p
B
= p
u
B
= u
v
B
= v
R
w
B
= w
R

B
=
R
Fig. 5-1: Algorithm for the solution of the problem (16),(17),(31),(33),(32),(34),(35) at the half line
S
B
= {(0, t); t > 0}. The value of the pressure p

is prescribed. Possible situations are illustrated

by the pictures showing
L

HTL

L
region. The sought boundary state located at the time axis,
marked red.
6 BOUNDARY CONDITION BY PREFERENCE OF VELOCITY
In this section we construct the boundary condition, i.e. the state q
B
in (23), preferring the given value
for the velocity u
GIVEN
. We add the following complementary conditions into the system (16),(17)
u

:= u
GIVEN
, (36)
u
R
= u

, (37)

R
:=
GIVEN
,
v
R
:= v
GIVEN
,
w
R
:= w
GIVEN
.
Here u

denotes the rst component of the velocity in the region

L

R
.
We seek the boundary state q
B
as the unique solution of the problem (16),(17),(36), (33)-(35) at
the half line S
B
= {(0, t); t > 0}. The system (24)-(29), (36), (33)-(35) is uniquely solvable. The
algorithm for the construction of the primitive variables
B
, u
B
, v
B
, w
B
, p
B
at the half-line S
B
is shown
in Fig. 6-1. The complete analysis of this problem is shown in [7].
- 15 -
R5375
u < u
L
u u
L
D = 4
L
p
L
+
2
L
(
+1
2
)
2
(u
L
u)
2
p =
1
2
_
2p
L
+
1
2

L
(u
L
u)
2
+ (u
L
u)

D
_
s
1
= u
L

p
L

L
_
+1
2
p
p
L
+
1
2

L
=
L
1
+1
p
L
p
+1
p
L
p
+
1
+1
a
L
=
_

p
L

L
, s
HL
= u
L
a
L
p = p
L
_
u+u
L
+
2
1
a
L
2
1
a
L
_
2
1
s
TL
= u a
L
_
p
p
L
_
(1)/2

L
=
L
_
p
p
L
_
1/
s
1
0 s
1
< 0 s
HL
0 s
HL
< 0
u 0 u < 0 s
TL
0 s
TL
< 0
u 0 u < 0
OUTLET OUTLET INLET OUTLET OUTLET OUTLET INLET
p
B
= p
L
u
B
= u
L
v
B
= v
L
w
B
= w
L
:=
L
p
B
= p
u
B
= u
v
B
= v
L
w
B
= w
L

B
=
L
p
B
= p
u
B
= u
v
B
= v
R
w
B
= w
R

B
=
R
p
B
= p
L
u
B
= u
L
v
B
= v
L
w
B
= w
L

B
=
L
p
B
=
p
L
_
2

+
1
+

1
(

+
1
)
a
L
u
L
_
2

1
u
B
=
2

+
1
_
a
L
+

1
2
u
L
_
v
B
=
v
L
w
B
=
w
L

B
=

L
_
2

+
1
+

1
(

+
1
)
a
L
u
L
_
2

1
p
B
= p
u
B
= u
v
B
= v
L
w
B
= w
L

B
=
L
p
B
= p
u
B
= u
v
B
= v
R
w
B
= w
R

B
=
R
Fig. 6-1: Algorithm for the solution of the problem (16),(17),(36),(33),(37),(34),(35) at the half line S
B
=
{(0, t); t > 0}. Possible situations are illustrated by the pictures showing the region
L

HTL

L
with the sought boundary state located at the time axis.
6.1 B.C. for the Impermeable Wall
As a special case of the complementary condition (36) we prescribe zero normal velocity, i.e. we choose
u

:= 0. (38)
Further we use the analysis of the system (16),(17),(36), shown above in 6. We seek the boundary state
q
B
as the unique solution of the problem (16),(17),(38) at the half line S
B
= {(0, t); t > 0}. Using the
analysis in 6 we construct the solution of the system (16),(17),(38) in primitive variables. It is
_
_
_
_

B
u
B
v
B
w
B
p
B
_
_
_
_
=
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_
_

L
1
+1
p
L
p
+1
p
L
p
+
1
+1
0
v
L
w
L
p
_
_
_
_
_
_
_
, with p =
2p
L
+
+1
2

L
u
2
L
+u
L
_
4
L
p
L
+
2
L
(
+1
2
)
2
u
2
L
2
, u
L
> 0,
_
_
_
_
_
_
_
_

L
_
1 +
1
2
u
L
a
L
_ 2
1
0
v
L
w
L
p
L
_
1 +
1
2
u
L
a
L
_ 2
1
_
_
_
_
_
_
_
_
, u
L
0.
(39)
Here
L
, u
L
, v
L
, w
L
, p
L
are prescribed, and a
L
=
_
p
L
/
L
. The solution for the rst component of
the velocity u
B
is equal to the prescribed velocity u

= 0 in this special case.

7 BOUNDARY CONDITIONBYPREFERENCE OF TOTAL QUANTI-
TIES AND DIRECTION OF VELOCITY
Here we will construct the boundary state w
k

ij
, following the procedure and notation described above
in 3.3. The state vector w
k
i
on element D
i
at time instant t
k
is known. Our aim is to prescribe the inlet
- 16 -
R5375
boundary values for the total pressure, total temperature, and the direction of the velocity, if possible.
Let us denote the prescribed values by p
o
,
o
, d. We use the transformation of this problem into a
local coordinate system x
1
, x
2
, x
3
, shown in 3.3, with the axis x
1
being in the opposite direction of the
prescribed velocity direction, given by n = d, |d| = 1. The situation is depicted in Fig. 7-1.
x
1
x
2

ij
D
i
x
1
x
2
n
v
Fig. 7-1: Coordinate transformation for the boundary edge
ij
in 2D. The axis x
1
is in the opposite
direction to the prescribed direction of the velocity.
Then the derivatives with respect to x
2
, x
3
are neglected. Our aim is to evaluate the boundary state
q
B
= q(0, t), such that the system (16),(17),(23) is uniquely solvable.
We add the complementary conditions
p

= p
o
_
1
1
2a
2
0
u
2

_
/(1)
, (40)

R
=
o
_
1
1
2a
2
o
u
2

_
, (41)
u

< 0, (42)
p

= p
R
,
v
R
= 0, (43)
w
R
= 0. (44)
Here
o
> 0, p
o
> 0 are given constants and a
2
o
= R
o
. Further R denotes the characteristic gas
constant, and is the adiabatic constant. The equations (40),(41) are considered for the velocity
u

(
_
2a
2
o
(1)
,
_
2a
2
o
(1)
). The equations (41),(40) come from the idea, that the total pressure p
o
and
the total temperature
o
are known (prescribed) in the region
R
.
The sought solution for q
B
of the system (24)-(29),(40)-(44) was described in [7]. The algorithm for
the construction of the primitive variables
B
, u
B
, v
B
, w
B
, p
B
at the half-line S
B
is shown in Fig. 7-2.
The complete analysis of this problem is shown in [7]. Here we state that the system may not have a
solution for the certain choice of the initial data. In that case the prescribed quantities cannot be preferred
at the boundary. In practical applications we use the boundary condition preffering the pressure (Section
5) with the prescribed pressure p
GIVEN
= p
o
in this unfortunate case. At last, the sought state vector is
evaluated as w
k

ij
= Q
1
q
B
. We use this boundary condition in the case, when the total quantities are
known at the inlet.
- 17 -
R5375
1. compute u

yes
u
L
0

no
F(0) 0
yes
u

= root of (*S)

no
Failed, choose u

= min {0, u
L
+
2
1
a
L
}
u
L
+
2
1
a
L
>
_
2a
2
o
(1)

yes
F(min {0, u
L
+
2
1
a
L
}) < 0

yes

no

no
Failed, choose
u

=
_
2a
2
o
(1)
+

_
2a
2
o
(1)
u
L

yes
u

, use (*R)

no
F(u
L
) < 0
yes
u

, use (*R)

no
u

= root of (*S)
(*S) F(u) = po
_
1
1
2a
2
o
u
2
_
/(1)
p
L

+1
4

L
(u
L
u)
2

(u
L
u)
2
_
4
L
p
L
+
2
L
(
+1
2
)
2
(u
L
u)
2
,
(*R) u =
2
_
u
L
+
2
1
a
L
_

DIS
2
_
1+
2a
2
L
(1)a
2
o
_
po
p
L
_
(1)/
_, DIS = 4
_
po
p
L
_
(1)/ 2a
2
L
(1)a
2
o
_
2a
2
o
(1)
+
4a
2
L
(1)
2
_
po
p
L
_
(1)/

_
u
L
+
2a
L
1
_
2
_
.
2. compute u
B
, p
B
,
B
u
B
= u

, p
B
= p
o
_
1
1
2a
2
0
u
2

_
/(1)
,
B
=
po
Ro
_
1
1
2a
2
o
u
2

_
1/(1)
.
Fig. 7-2: Algorithm for the solution of the problem (24)-(29),(40)-(44) at the half line S
B
=
{(0, t); t > 0}. In the case of failure, the problem doesnt have a solution, and the value for u

is
chosen.
- 18 -
R5375
8 SOFTWARE AND EXAMPLES
The described methods and original boundary conditions were coded in C programming language, and
used in own programs for the solution of the compressible gas ow on unstructured meshes in 3D.
Application of the presented boundary conditions on moving meshes was tested in [12, 6, 13, 18]. In this
work we further used the program ANGENER [1] for the renement of the triangular meshes, TETGEN
(http://tetgen.berlios.de) for the tetrahedral mesh generation, and TECPLOT (http://www.tecplot.com)
for the visualization of the computed results. Here we will shortly describe the developed program, used
to compute the shown examples.
C3DEEI
As a part of this work we attach (attachment [C3DEEI]) the latest version of the developed code for the
solution of the 3D compressible ow, written in C programming language. It is a further development
of the program shown in [7], with the implicit time marching implemented, as described in Section 3. It
can be used for the numerical solution of the 3D Euler Equations on tetrahedral grids. It is based on the
nite volume method. An explicit and implicit time discretization is used. We used the GMRES method
with ILUK preconditioning in order to solve the system (14) in each time step, SPARSEKIT [19] was
used. The software is written for the parallel use, OPENMP directive is used. To approximate the face
values at the inner faces, the program solves local initial-value Riemann problem for the 3D split Euler
equations, shown in Section 4 or the Vijayasundaram numerical ux is used. At the boundary faces, the
initial-boundary modication to this problem is solved, as shown in Section 5 - 7. We will show the use
of this program on the example of channel ow. We chose the channel with 10% cylindrical bump at the
lower wall, as a natural extension of the so-called GAMM channel to 3D. The geometry of the GAMM
channel can be found in [5, page 277]. Here we extend this 2D geometry into a 3D channel with the
thickness 0.2. The initial mesh is shown in Fig. 8-1.
1 2
3
4
5
Fig. 8-1: 10% cylindrical bump, initial mesh and boundary numbering.
The direction of the ow is from left to right. The ow regime in test case is set to the inlet Mach
number M = 0.67. In this regime a shock is generated near the lower wall. The following data were
used in our tests:
initial condition - constant state in the whole domain,

o
= 1.5, v
o
1
= 205.709277, v
o
2
= 0, v
o
3
= 0, p
o
= 101000.
inlet boundary condition - total quantities and velocity direction
The boundary condition from Section 7 was used, with

o
= 255.639, p
o
= 136461, d = (1, 0, 0). These values were obtained using the equations (40)
and (41) for the data in initial condition.
outlet boundary condition - preference of the pressure
The boundary condition from Section 5. with described averaging technique was used, with
p

= 101000.
- 19 -
R5375
impermeable wall boundary condition
The boundary condition in Section 6 was used in the case of the inviscid ow.
The computational results are shown in Fig. 8-3. Anisotropic renement method, shown and coded in
[21], was used with minor changes in the code (renement programs are also attached).
Input and Output Data Files Format
The 3D solution is computed on tetrahedral, prismatic, and hexahedral elements, and all the solver pa-
rameters are given in one input ini le. All comments in the input les start with the sign * and end with
the end of the line.
mesh le format:
The mesh le with tetrahedral, prism, and hexahedral elements is written in ASCII and uses the following
format
n
poin
number of points
n
vol
number of volume elements
n
bfaces
number of boundary faces
x
1
y
1
z
1
coordinates of the 1st point (=vertex 1)

x
n
poin
y
n
poin
z
n
poin
coordinates of the n
poin
th point
P
1
1
P
2
1
P
N1
1
element 1 : index of vertices

P
1
n
vol
P
2
n
vol
P
Nn
vol
n
vol
element n
vol
: index of vertices
B
1
1
B
2
1
B
M1
1
I
1
boundary face 1: index of vertices, followed
by the index of the boundary
B
1
n
bfaces
B
2
n
bfaces
B
Mn
bfaces
n
bfaces
I
n
bfaces
boundary face n
bfaces
: vertices, index
n
pp
number of periodical pairs (points)
Q
1
1
Q
2
1
periodical pair 1 (index of points)

Q
1
npp
Q
2
npp
periodical pair n
pp
The point coordinates are followed by the element vertices and boundary face vertices. The coordinates
of the vertex i are x
i
, y
i
, z
i
. Each volume element D
j
is given by the index of its vertices P
1
j
P
2
j

P
Nj
j
, j = 1...n
vol
, volume element has Nn
vol
vertices, Nn
vol
{4, 6, 8}. The geometric ordering of
the vertices is shown in Fig. 8-2. Each boundary face i = 1...n
bfaces
is given by the indexes of its
vertices B
1
i
, B
2
i
, ...B
Mi
i
, M
i
{3, 4}, and by the index of the boundary condition I
i
{1, ..., NBC}.
The periodical point pairs are given by the vertices indexes Q
1
i
Q
2
i
, i = 1...n
pp
.
1
2
3
4
1
2
3
4
5
6
1
2 3
4
5
6 7
8
Fig. 8-2: C3DEEI: allowed elements, vertices numbering.
ini le format:
The format of the ini le, with the initial solver data, will be shown on the example. There is always
- 20 -
R5375
a keyword followed by the input value. Input and output physical values are given in the International
System of Units (SI). The following keywords are recognized:
EQTYPE
Equations to solve: euler, NS, NS-FVMFEM-split, NS-FVMFEM-nosplit.
TIMEINTEGRATE_PROC
Time integration to use: euler, ec, rk4, implicit_euler, implicit_eulerN.
(Euler, Euler-Cauchy, 4-points Runge-Kutta, Implicit Euler, Implicit Euler - N iterations of the Newton method)
MATRIXSOLVER SOLVER
Matrix solver to use (implicit method)
SPARSEKIT_ILUKGMRES Sparskit GMRES + ILUK preconditioner (default).
SPARSEKIT_ILUKBICGSTAB Sparskit BiCGSTAB + ILUK preconditioner.
CUSP_PBiCGSTAB_CPU CUSP package, BiCGSTAB (TESTING, so far bad preconditioner).
CUSP_PGMRES_CPU CUSP package, GMRES (TESTING, so far bad preconditioner).
MATRIXSOLVER MAXITER N
Matrix solver, maximal number of iterations (50 default)
MATRIXSOLVER RESTART N
Matrix solver, restart after N iterations (applied with GMRES iterative solver, 25 default)
MATRIXSOLVER EPSrelative G
Stopping criterium for the matrix solver, default G=1e-8
EFLUX
Inviscid ux to use: Riemann, Vijayasundaram.
EFLUX_DHDW
Inviscid ux Jacobi matrix estimate: N, VS.
N compute derivatives numerically.
VS via Vijayasundaram ux with xed matrixes.
SPACERECONSTRUCTION_PROC
Space reconstruction used to achieve the higher order method:
none No reconstruction, constant approximations are used (FVM).
FVM0 No reconstruction, constant approximations are used (FVM).
BarthJespersen_P Barth-Jespersen limiter is applied on primitive variables.
BarthJespersen_C Barth-Jespersen limiter is applied on conservative variables.
VanAlbada_P Van Albada limiter is applied on primitive variables (hexaxedral meshes).
VanAlbada_C Van Albada limiter is applied on conservative variables (hexaxedral meshes).
SPACERECONSTRUCTION_START 10000
Computes 10000 iterations with the 1st order method and then it continues with the higher order reconstruction.
SPACERECONSTRUCTION_GRADIENT_RECOMPUTE_STEP 60
Gradients for the reconstruction are recomputed after every 60 iterations.
GRIDNAME X.grid
The name of the base le with the computational mesh. The given string determines also the le-names of the other output and input les. These les are X.grid.sol and
X.grid.sol.bin for the input/output solution le in ASCII and binary format, X.grid.dat for the output TECPLOT le, X.grid.stat for the output computational
statistics, X.grid.conv for the output convergence statistics.
In the case of the MPI run, programworks with the base le-name, altered by prex with a number. This number denotes the id of the splitted domain part (0 .. nparts-1) for example
0_X.grid, 1_X.grid are grid les. Further, the les with volume aliases and send and receive indexes are used with MPI (input les): 0_X.grid.vid,1_X.grid.vid,
0_X.grid.recv,1_X.grid.recv,0_X.grid.send and 1_X.grid.send.
TECPLOT_OUT 1
Output TECPLOT le is created if given 1, no le if 0 is given.
CONV_OUT 0
Output convergence le is created if set to 1, no le if set to 0.
SPLIT_REDMAP_OUT 1
Output temporary le redmap for the renement of the selected elements, with the programmeshrefine.
ANIZO_RPF_OUT 0
Output temporary le anizoRPF for the anisotropic renement point function, which can be used with the program meshrefanizo.
IC ruvwp rho v1 v2 v3 p
Initial condition, here rho v1 v2 v3 p are the given initial values for density, velocity components and the pressure. This initial condition is not used if the solution le
X.grid.sol or X.grid.sol.bin is presented in the execution directory. Example: IC ruvwp 1.29233 0 0 0 101325
CONST KAPA 1.4
Adiabatic gas constant, given as = 1.4.
CONST R 287.04
Ideal gas constant, given as R = 287.04.
CONST RMI_REF 1.716e-05
Viscosity at the
ref
temperature.
CONST T_REF 273.15
The reference temperature
ref
.
CONST S_T 110.4
The Sutherland temperature S.
The Sutherlands law constants, used for the correction of the dynamic viscosity at each face
=
ref
_

ref
_
3/2

ref
+ S
+ S
. (45)
- 21 -
R5375
CONST RKA 0.0211
Ideal gas constant, given as k = 0.0211.
OUT_PHYSICAL_TIMESTEP 10.0003
Periodical output, time-step between each output.
STOP LINFTY 1e-9
Stop the computation when residual drops to the given minimum.
STOP PHYTIME 999
Stop the computation when the given physical time is reached.
STOP ITERATIONS 100000
Stop the computation when the given iteration 100000 is reached. If started from the iteration with the higher number, computations are stopped after 100000 iterations.
USE_NORMED_VALUES 1
All values within the solver are normed to the critical ones. This may enhance the overall accuracy. Input and output les are still written in SI. If USE_NORMED_VALUES 0 is
given, normed values are not used.
CONST ACCURACY 1e-7
Accuracy used within the solution of the edge values problem. It is used by the Riemann solver and for the boundary conditions evaluation.
CONST CFL 20000
Stability time-step restriction, rst step. (CFL 1 for explicit method)
CONST CFLMIN 0.85
Stability time-step restriction, minimal CFL within implicit method.
CONST CFLMAX 100000
Stability time-step restriction, maximal CFL within implicit method.
PRINT_ITERATION 100
Controls the standard output stream, prints values each 100 iterations.
BC
Boundary conditions. Keyword is followed by the number of the boundaries. Then the data for each boundary follows. The id of the boundary is followed by the used boundary
condition with necessary data.
Example (3D GAMM channel):
BC
6
1 inlet fixed_to_T_direction 1.5 205.709277 0 0 101000
2 outlet pavg 101000
3 wallT 273.15
4 wallT 273.15
5 periodic 6
6 periodic 5
3 4 5 6 1 2
This example shows 6 boundaries. The indexes 1-6 correspond to the indexes of the boundaries in the mesh le. At the boundary 1 we prescribe inlet boundary condition conserving
the total pressure, the total temperature, and the direction of the velocity given by the vector, with values computed from the given xed state. Boundaries 3 and 4 represent the
walls with the given temperature T = 273.15. Boundary 2 is an outlet boundary with the pressure p = 101000 in average given. Boundaries 5 and 6 are periodical. The
sequence 3 4 5 6 1 2
prescribes the boundary priority for the points laying on multiple boundaries (boundary 3 has the highest priority).
ini le example (3D GAMM channel):
EQTYPE euler
TIMEINTEGRATE_PROC implicit_euler
CONST CFL 20000
CONST CFLMIN 0.85
CONST CFLMAX 100000
MATRIXSOLVER SOLVER SPARSEKIT_ILUKGMRES
MATRIXSOLVER MAXITER 50
MATRIXSOLVER RESTART 35
MATRIXSOLVER EPSrelative 1e-6
SPACERECONSTRUCTION_PROC FVM0
SPACERECONSTRUCTION_START 5000
SPACERECONSTRUCTION_GRADIENT_RECOMPUTE_STEP 60
GRIDNAME X.grid
TECPLOT_OUT 1
CONV_OUT 0
CONST KAPA 1.4
CONST R 287.04
OUT_PHYSICAL_TIMESTEP 100
STOP LINFTY 1e-8
STOP PHYTIME 50.25
STOP ITERATIONS 50
USE_NORMED_VALUES 1
CONST ACCURACY 1e-6
PRINT_ITERATION 1
IC ruvwp * 1 rho v1 v2 v3 p kappa
1.5 205.709277 0 0 101000
BC
* no boundaries + each boundary condition + boundaries priority for dual mesh
6
1 inlet fixed_to_T_direction 1.5 205.709277 0 0 101000
2 outlet p 101000
3 wall
4 wall
5 periodic 6
6 periodic 5
EOF
- 22 -
R5375
Fig. 8-3: 3D inviscid ow computation (coarse mesh). Mach number and pressure isolines. (C3D)
solution le format:
The solution le stores the nite volume solution in the primitive variables for each element. Solution
le is written in ASCII. Example:
nvol 5
(1) v
1
(1) v
2
(1) v
3
(1) p(1)
. . .
(nvol) v
1
(nvol) v
2
(nvol) v
3
(nvol) p(nvol)
iteration
physical time
Here nvol is the number of volume elements, 5 is the count of the solution components, is adiabatic
constant, symbol (i) denotes the density, v
1
(i), v
2
(i), v
3
(i) are the components of the velocity, and
p(i) is the pressure on element D
i
(i is the element index determined by the the mesh le). Solution
binary le stores the values in the same order as the solution le. It is preferred in the case of the
computation restart.
Program run
The program c3deei is coded with OPENMP directive, and it uses all the systems processors for the
computation by default. We tested the program on the Linux system, the Intel C/C++ compiler was used
for the compilation.
9 Examples
Here we show the computational results of the compressible gas ow, obtained with the own-developed
codes. For other computational results see also [8], [11], [9], [10], [14], [15], [16]. The shown
computations ran on the personal computer with two Intel(R) Xeon(R) CPU E5520 @ 2.27GHz
processors.
9.1 Laval Nozzle
A de Laval nozzle is a convergent-divergent tube used to accelerate subsonic ow to a supersonic speed.
Its purpose is to achieve a homogeneous parallel ow in the test section, without shock waves and other
disturbances. This can be done with careful shaping the cambered sides.
In the converging part, the speed of the ow is increased due to the reducing cross-section area of the
tube. The velocity reaches the speed of sound (M = 1) in the throat of the nozzle and cannot be
increased beyond M = 1. We say that the ow is choked. Then the ow is expanded into a supersonic
speed in the divergent part of the tube.
The specic nozzle we show here is a part of the supersonic wind tunnel at the department of high-speed
aerodynamics in VZLU

. It is a rectangular tube with cambered upper and lower sides. The design of the
- 23 -
R5375
nozzle was chosen to yield the Mach number M = 2.135, the length of the nozzle is 2 meters. At the
inlet, the nozzle is connected to the standard atmosphere, and through the outlet it is connected to the
low-pressure reservoir. To simulate real conditions, we use the inlet boundary condition conserving the
total pressure p
o
= 101325, total temperature T
o
= 273.15 and the uniform direction of the velocity,
perpendicular to the inlet boundary, see Section 7. At the outlet we use the boundary condition
preferring the pressure (Section 5.), with the prescribed value p

= 10490.3808184132. This value is an

adiabatic estimate, obtained with the use of the equation (40) for given p
o
and M. Results were shown
in [16], [7], computed on structured and unstructured meshes, adaptively rened according to the
solution. Here we add more results, computed with the use of the described implicit method on the
three-dimensional unstructured mesh and we show the comparison of the real computational time for
the chosen methods, see Fig. 9-1 - 9-3. Here (VS+VS) means the Vijayasundaram numerical ux +
Jacobian computed using (22), (RS+VS) means the exact Riemann Solver + Jacobian computed using
(22), (RS+N) means the Riemann solver numerical ux + Jacobian evaluated numerically. The speed
advantage of the implicit method is apparent even for the computation on the coarse meshes.
method used CFL iterations real time speedup
1 explicit 0.85 1725 0m28.213s 1.00
2 implicit (RS+N) 8-308 48 0m13.322s 2.12
3 implicit (VS+VS) 8-339 55 0m11.702s 2.41
4 implicit (RS+VS) 8-662 54 0m10.498s 2.69
Fig. 9-1: 3D inviscid computation, coarse mesh with 1896 prismatic elements. Pressure and Mach
number isolines, real time of the computation for chosen methods. (C3DEEI)
- 24 -
R5375
method used nproc CFL iterations real time speedup
1 explicit 1 0.85 30627 33m21.782s 1.00
2 explicit 4 (OMP) 0.85 30627 12m51.273s 2.60
3 explicit 8 (OMP) 0.85 30627 7m42.946s 4.32
4 implicit (RS+N) 1 132-4406 57 2m35.913s 12.84
4 implicit (VS+VS) 1 132-2985 66 2m3.313s 16.23
5 implicit (RS+VS) 1 132-3918 66 1m57.246s 17.11
Fig. 9-2: 3D inviscid computation, coarse mesh with 8208 hexahedral elements. Pressure and Mach
number isolines (C3DEEI).
method used nproc CFL iterations real time speedup
1 explicit 1 0.85 46262 699m57.692s 1.00
2 implicit 1 50 - 6674.01 134 46m47.029s 14.03
Fig. 9-3: 3D inviscid computation, mesh with 109490 prismatic elements. Pressure and Mach number
isolines (C3DEEI).
9.2 A Body in a Channel
Here we present a computational result of the 3D non-stationary inviscid channel ow at Mach number
M = 0.67, the ow is subsonic in this example. A body immersed in the owing uid establishes a
certain wave pattern which evolves in time and eventually exits the channel. The inlet is located left,
outlet right, other boundaries are considered as wall. The simple geometry of 8% double circular arc
DCA08 was chosen as a body.
Initial condition:
= 1.5, v = (205.709277, 0, 0), p = 101000.
Boundary conditions:
inlet (left): BC. Section 7., with p
o
= 136461,
o
= 255.639, direction d = (1, 0, 0),
outlet (right): BC. Section 5., p = 101000,
walls: BC. Section 6., u = 0,
- 25 -
R5375
periodical boundary condition (front and rear boundary).
The computational results and comparison of the real computational time for the chosen methods are
shown in Fig. 9-4. The signicant speedup (236 ) was achieved with the use of the implicit method.
The notation (A25) means that 25 iterations (at most) were allowed in solving (14) with the iterative
GMRES method. The correct choice of the boundary conditions is also important in this case (tested
with the explicit method). It can be shown that the xed boundary conditions give incorrect results near
boundaries, the linearized boundary condition reects the waves into the domain, leading to the
oscillations in the solution. The suggested boundary conditions do not suffer from these drawbacks.
method used CFL iterations real time speedup
1 explicit 0.85 27370 626m44.229s 1.00
2 implicit (A50) 80000-100000 8 4m13.803s 148.16
3 implicit (A50) 80000-200000 7 4m4.848s 153.58
3 implicit (A25) 80000-244141 7 2m38.985s 236.53
Fig. 9-4: 3D inviscid computation, mesh with 118835 prismatic elements. Pressure and Mach number
isolines, real time of the computation for chosen methods. (C3DEEI)
10 CONCLUSIONS
In this paper we worked with the system of equations describing the compressible uid ow in 3D. We
applied the nite volume method for the discretization of these equations, and we have shown the
implicit method, which is considerably faster than the explicit method. The so-called Riemann problem
for the 3D split Euler equations was presented, as the initial-value problem equipped with left-hand side
and right-hand side initial-value condition. The analysis of this problem can be found in many books,
i.e. see [5], [20]. The solution of the Riemann problem led us to the system of nonlinear algebraic
equations, which must be solved with the use of an iterative process. In order to discretize the values at
the boundary, the original initial-value problem was modied, as shown in [17]. The right-hand side
initial-value condition is not known at the boundary faces, and has to be either prescribed or replaced by
suitable conditions. In order to prescribe some variable (for example pressure) by chosen value at the
boundary, the local modied Riemann problem must have a solution, otherwise the value cannot be
used at the boundary. This was the main idea behind the construction of the shown boundary conditions
by preference of certain variable. On the contrary to the solution of the initial-value Riemann problem,
the solution of the modied boundary problems can be written in a closed form. Therefore it is not
computationally expensive to use the constructed boundary conditions in the code.
The algorithms for the solution of the boundary problems were coded and implemented into the
own-developed software for the solution of the inviscid compressible gas ow (the Euler equations).
Numerical examples show superior behaviour of the implicit method.
Acknowledgment
This result originated with the support of Ministry of Industry and Trade of Czech Republic for the
long-term strategic development of the research organisation.
- 26 -
R5375
REFERENCES
[1] V. Dolejs . Angener users guide. Department of Numerical Mathematics,Faculty of Mathematics
and Physics,Charles University, Prague, http://www.karlin.mff.cuni.cz/dolejsi/angen/angen.htm,
2005.
[2] V. Dolejs . Discontinuous Galerkin method for the numerical simulation of unsteady compressible
ow. WSEAS Transactions on Systems,5(5):1083-1090, 2006.
[3] M. Feistauer. Mathematical Methods in Fluid Dynamics. Longman Scientic & Technical,
Harlow, 1993.
[4] M. Feistauer. Numerical methods for compressible ow. In J. Neustupa and P. Penel, editors,
Mathematical Fluid Mechanics. Recent Results and Open Questions. Birkhauser, Basel, 2001.
[5] M. Feistauer, J. Felcman, and I. Straskraba. Mathematical and Computational Methods for
Compressible Flow. Oxford University Press, Oxford, 2003.
[6] M. Kyncl J. Pelant. Applications of the Navier-Stokes equations for viscous turbulent ow on
steady or moving grids with the (k,w) turbulent model. Czech Aerospace Proceedings / Letecky
zpravodaj, 1:26, 2008.
[7] M. Kyncl. Numerical solution of the three-dimensional compressible ow. Masters thesis,
Prague, 2011. Doctoral Thesis.
[8] M. Kyncl and J. Pelant. Applications of the Navier-Stokes equations for 3d viscous laminar ow
for symmetric inlet and outlet parts of turbine engines with the use of various boundary conditions.
Technical report R3998, VZLU

, Beranovych 130, Prague, 2006.

[9] M. Kyncl and J. Pelant. Applications of the Navier-Stokes equations for 2d viscous, compressible
turbulent ow on steady grids with the EARSM turbulent model. Technical report R4300, VZLU

,
Beranovych 130, Prague, 2007.
[10] M. Kyncl and J. Pelant. Applications of the Navier-Stokes equations for 3d viscous, compressible
laminar and turbulent ow in the axissymmetrical channels as parts of turbine engines with
steady or moving walls. Technical report R4301, VZLU

, Beranovych 130, Prague, 2007.

[11] M. Kyncl and J. Pelant. Applications of the Navier-Stokes equations for viscous turbulent ow on
steady or moving grids with the (k,omega) turbulent model. Technical report R4153, VZLU

,
Beranovych 130, Prague, 2007.
[12] M. Kyncl and J. Pelant. Applications of the Navier-Stokes equations for 3d viscous, compressible
laminar and turbulent ow in the channels with steady or moving walls. Technical report R4415,
VZLU

, Beranovych 130, Prague, 2008.

[13] M. Kyncl and J. Pelant. Applications of the Navier-Stokes equations for 3d viscous compressible
turbulent ow with the (k, ) or EARSM turbulent model. Technical report R4607, VZLU

,
Beranovych 130, Prague, 2009.
[14] M. Kyncl and J. Pelant. Numerical simulation of the forced airfoil oscillation with the (k, ) or
EARSM turbulent model. Technical report R4606, VZLU

, Beranovych 130, Prague, 2009.

[15] M. Kyncl and J. Pelant. Applications of the Navier-Stokes equations for 3d viscous compressible
turbulent channel ow with the use of various boundary conditions. Technical report R4916,
VZLU

, Beranovych 130, Prague, 2010.

[16] M. Kyncl, J. Pelant, and J. S

imak. Numerical solution of the transonic compressible gas ow

through Laval nozzles. Technical report R4917, VZLU

, Beranovych 130, Prague, 2010.

- 27 -
R5375
[17] J. Pelant. Arti reports VZLU

, z-65, z-67 to z-73. Prague, 1996-2000.

[18] A. Prachar , M. Kyncl, and J. Pelant. Numerical simulations of the oscillating wing/airfoil
problem. Technical report R4915, VZLU

, Beranovych 130, Prague, 2010.

[19] Youcef Saad. Sparsekit: a basic tool kit for sparse matrix computations. University of Minnesota,
1994.
[20] E. F. Toro. Riemann Solvers and Numerical Methods for Fluid Dynamics. Springer, Berlin, 1997.
[21] M. Zoubek. Adaptive methods for the solution of three dimensional ow. Masters thesis, Charles
University, Prague, 2001. (in Czech).
- 28 -
R5375
VZLU

Aerospace Research and Test Establishment, Beranovych 130, 199 05 Prague

Czech Republic
REPORT DOCUMENTATION PAGE
1. Report No. 2. Report Date 3. Account No. 4. No. of Pages 5. Publication Code
6. Title and Subtitle
7. Authors department (Abbr. No. Name)
8. Autor(s) (Name, Signature) Head of Department (Name, Signature)
9. Responsible Person (Name, Signature) 10. Head of Department (Name, Signature) 11. Technical Director (Name, Signature)
12. Abstract
13. Key Words
R5375 May 2012 IP4401 29 + CD FIR
IMPLICIT METHOD FOR THE 3D EULER EQUATIONS
MOT 4400 Engines
Martin Kyncl
Jaroslav Pelant Jir Fiala
Martin Kyncl
Jarom r Lamka Viktor Kucera
We work with the system of equations describing non-stationary inviscid compressible uid ow in 3D, i.e. the
Euler equations, and we focus on the numerical solution of these equations and on the boundary conditions. In
particular, we choose the implicit nite volume method. Many boundary conditions (i.e. xed, linearized) bring
non-physical errors into the solution, slowing down the convergent process, or even ruining the solution in the
whole domain. We use the analysis of the Riemann problem for the construction of the boundary conditions in
order to match the experimental data. We show, that the unknown one-side initial condition for the local Riemann
problem can be partially replaced by the suitable complementary condition. We suggest such complementary
conditions (by preference of pressure, velocity, total quantities,...) giving physically relevant data. Our method
for the construction of the boundary conditions is robust and its use accelerates the solver convergence to the
desired solution. As a part of this work we present own-developed software for the solution of the 3D Euler
equations, using the implicit method.
compressible gas ow, boundary conditions, nite volume method, implicit method
- 29 -