Beruflich Dokumente
Kultur Dokumente
D
Boundary condition
PDE must be satisfied in D
Physical Classification
Marching problems involve an open domain (time or
time-like) and initial conditions and/or boundary
conditions.
Example are unsteady heat conduction and boundary
layer (without separation) problems.
x
t
I.C.
B.C.
B.C.
Marching direction
x
y
Physical Classification
Boundary layer equations: (marching in x-direction)
2
2
0
y
u
y
u
v
x
u
u
y
v
x
u
Pure initial value problems:
=
Eq. Berger linear non 0
Eq. Berger linear 0
) ( ) 0 , ( . .
x
u
u
t
u
x
u
a
t
u
x f x u C I
) Eq. wave (
) ( ) 0 , ( & ) ( ) 0 , ( . .
2
2
2
2
2
x
u
a
t
u
x g x u x f x u C I
t
= =
Mathematical Classification
Consider the following general 2
nd
order PDE:
0 = + + + + + + G F E D C B A
y x yy xy xx
The characteristic equation of this PDE is:
0
2
=
|
.
|
\
|
|
.
|
\
|
C
dx
dy
B
dx
dy
A
Whose roots are:
A
AC B B
dx
dy
2
4
2
=
|
.
|
\
|
We can categorize the PDE according to its
discriminant: AC B D 4
2
=
Mathematical Classification...
If 0 > D at ) , (
0 0
y x p , two real characteristics
exist and the equation is Hyperbolic.
If 0 = D at ) , (
0 0
y x p , only one real characteristic
exists and the equation is Parabolic.
If 0 < D at ) , (
0 0
y x p , no real characteristics
exists and the equation is elliptic.
If D changes sign then the PDE is of mixed type.
For example, for 2D potential flow and Transonic
flow, we have:
) 1 ( 4
0 ) 1 (
2
2
M D
M
yy xx
=
= +
Note that a PDE can be of different types in
different regions.
Mathematical Classification...
Wave equation:
) ( ) 0 , ( ), ( ) 0 , ( : . .
,
2
x g x u x f x u C I
u a u
t
xx tt
= =
=
The exact solution is
| |
+
+ + + =
at x
at x
dz z g
a
at x f at x f t x u ) (
2
1
) ( ) (
2
1
) , (
It is seen that
Hyperbolic a D =
2
4
On the t x plane, we can show that two
characteristic lines exist
+ =
+ =
=
2
1
1
c at x
c at x
a dx
dt
Mathematical Classification...
Region 3 is called domain of influence for point P .
Region 2 is called domain (zone) of dependence for
point P .
Region 1, is called domain of silence. This region
does not feel the other two regions.
2
1
1
3
0 0
at x
0 0
at x +
x
t
1/a
-1/a
P
Model Equations
Laplace and Poisson:
= +
) , (
0
y x f
yy xx
Unsteady Heat Conduction:
) (
yy xx t
+ =
First order linear wave equation (Linear Burger Eq.):
0 = +
x t
a
First order nonlinear wave equation (Inviscid nonlinear burger):
0 = +
x t
Viscous Burger Equation:
xx x t
= +
Second order wave equation:
xx tt
a
2
=
System of PDEs
In general, any higher-order PDE can be converted to a
system of first order PDEs.
EXAMPLE: Let us consider the wave equation
xx tt
a
2
= :
0
0
0
2
2
2
2
2
2
=
(
+
(
=
w
v
x a
a
w
v
t
t x
u
a
t
w
x
u
a
x
w
x t
u
x
v
t
v
t
v
x
u
a w
t
u
v
Or equivalently
0
U
A
U
=
x t
Eigen-values of A are a = .
Therefore, we have two distinct real eigen-value.
That is our system is Hyperbolic.
System of PDEs
EXAMPLE: studying the Laplace equation, 0 = +
yy xx
, we
find:
0
0 1
1 0
=
(
+
(
=
v
u
y v
u
x
x
v
y
u
This have two distinct complex eigen-values, i = .
Therefore, the equation is elliptic.
Linear Unsteady System of Equations
Consider
0 = +
y
B
x
A
t
Here, matrices A and B are functions of x t, and y .
is a column vector and is the dependent variable.
is a column vector and is function of x , and y .
The equation is hyperbolic at a point ) , ( t x p if the eigen-values
of A are all real and distinct.
The equation is hyperbolic at a point ) , ( t y p if the eigen-values
of B are all real and distinct.
The equation is parabolic at a point ) , ( t x p if the eigen-values
of A are all real but less than number of equations.
The equation is parabolic at a point ) , ( t y p if the eigen-values
of B are all real but less than number of equations.
Linear Unsteady System of Equations
The equation is elliptic at a point ) , ( t x p if the eigen-values of
A are all complex.
The equation is elliptic at a point ) , ( t y p if the eigen-values of
B are all complex.
The equation is mixed (Hyperbolic/Elliptic) at a point ) , ( t x p if
the eigen-values of A are mixed real and complex.
The equation is mixed (Hyperbolic/Elliptic) at a point ) , ( t y p if
the eigen-values of B are mixed real and complex.
Linear Steady System of Equations
Consider
0 = +
y
B
x
A
Method 1: define
2 3
2 3
4 1
4 1
2
4
b b
a a
b b
a a
R B Q A P
where
PQ R H
+ = = =
=
Then,
Elliptic H
parabolic H
Hyperbolic H
<
=
>
0
0
0
Method 2: re-write the equation as
0 )
( )
( = +
+
j
y
i
x
j B i A
Linear Steady System of Equations
In the characteristic direction n
r
, we can write
y x y x
n B n A n i n B i A T + = + + = ) j
( ) j
(
A wave-like solution will exist if
0 0 = + =
y x
n B n A or T
This gives
0
2
= +
|
|
.
|
\
|
+
|
|
.
|
\
|
P
n
n
R
n
n
Q
y
x
y
x
Or ( PQ R H 4
2
= )
Q
H R
n
n
y
x
2
=
Therefore,
Elliptic H
Parabolic H
Hyperbolic H
<
=
>
0
0
0
Linear Steady System of Equations
EXAMPLE: Steady, inviscid and incompressible flow
0
0
0
=
y
p
y
v
v
x
v
u
x
p
y
u
v
x
u
u
y
v
x
u
In vector form, we have
( ) 1 0 ) (
1 0
0 0
0 1 0
0 0
1 0
0 0 1
0
2 2
= = = + + =
(
(
(
=
(
(
(
=
(
(
(
= = +
y
x
y
x
y x y x
y
U
x
U
n
n
and
v
u
n
n
n n n v n u T
v
v B
u
u A
p
v
u
U with B A
Which means that the equations are of mixed hyperbolic/elliptic
type.
Linear Unsteady System of Equations
EXAMPLE: Unsteady, 1D, inviscid and compressible flow
0
0
1
0
2
=
x
p
u
x
u
a
t
p
x
p
x
u
u
t
u
y
u
x
u
t
In vector form,
(
(
(
=
(
(
(
= =
u u
u
u
p
u with
x t
2
0
/ 1 0
0
0
A U
U
A
U
Whose eigen-values are
a u a u u I + = = , , 0 A
System has 3 real and distinct eigen-values, therefore, it is
hyperbolic.
Second order PDEs
Every second order PDE can be converted to two first order
equations first and then dealt with as before.
EXAMPLE: steady, 2D, viscous, incompressible flow
|
|
.
|
\
|
|
|
.
|
\
|
2
2
2
2
2
2
2
2
Re
1
Re
1
0
y
v
x
v
y
p
y
v
v
x
v
u
y
u
x
u
x
p
y
u
v
x
u
u
y
v
x
u
Set,
b
y
v
x
u
y
u
c
y
v
b
x
v
a
continuum
=
=
Second order PDEs
In vector form, we get
C
y
U
B
x
U
A =
(
(
(
(
(
(
(
(
(
=
(
(
(
(
(
(
(
(
(
=
(
(
(
(
(
(
(
(
=
1 0
Re
1
0 0 0
0
Re
1
0 0 0 0
0 0 1 0 0 0
0 0 0 1 0 0
0 0 0 0 1 0
0 0 0 0 0 1
0 0 0
Re
1
0 0
1 0
Re
1
0 0 0
0 1 0 0 0 0
0 0 1 0 0 0
0 0 0 0 0 1
0 0 0 0 0 0
B A
p
c
b
a
v
u
U
We get,
( )
= +
|
|
.
|
\
|
=
= + =
(
(
(
(
(
(
(
(
(
(
=
0 1
0
0
Re
1
0
Re Re
0 0
Re Re
0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0 0
2 2
2 2 2
x
y
y
y x y
y
y
x
x
y
x
x y
x y
y x
y
n
n
n
n n n T
n
n
n
n
n
n
n n
n n
n n
n
T
Whose eigen-values are imaginary. Therefore, the system is
mixed elliptic/parabolic.