Beruflich Dokumente
Kultur Dokumente
Harsha
Asst. Prof., MIED, IIT Roorkee
COMPUTATIONAL METHODS IN FLUID FLOW
AND HEAT TRANSFER
Vector Gradient:
Divergence:
Directional Derivative:
Laws
CONSERVATION OF MASS
Consider a volume fixed in space (Figure 1). The
rate of increase of mass inside it is the volume
integral
CONSERVATION OF MOMENTUM
Consider the motion of the infinitesimal fluid
element shown in Fig. 2. Newtons law requires that
the net force on the element must equal mass times
the acceleration of the element. The sum of the
surface forces in the x1 direction equals
which simplifies to
NAVIERSTOKES EQUATIONS
NAVIER
The constitute equation for a Newtonian fluid is
given as
(6)
(8)
Where 2ui is the Laplacian of ui
Euler equation
(10)
Continuity equation
kinematic
viscosity
(constant)
density
(constant)
pressure
external force
(such as
gravity)
Acceleration term:
change of velocity
with time
Advection term:
force exerted on a
particle of fluid by the
other particles of fluid
surrounding it
Pressure term:
Fluid flows in the
direction of
largest change
in pressure
change
body
in
= advection + diffusion + pressure + force
velocity
with time
y - component:
z - component:
y
z
2a
Fluid flow direction
Fixed plate
y - component:
z - component:
y
z
2a
Fluid flow direction
Fixed plate
u = function of (t, x, y, z)
v = function of (t, x, y, z)
z direction:
w = function of (t, x, y, z)
(1)
u = function of (y)
v = function of (y)
z direction:
w=0
(2)
v
=0
y
u v w
+
+
=0
x y z
v = constant
v=0
or
Fixed plate
v=0
2a
Fluid flow direction
Fixed plate
w=0
y - component:
z - component:
+ 2 + g x = 0
x
y
p
y - component:
+ g y = 0
y
z - component:
p
+ g z = 0
z
y - component:
z - component:
p
2u
+ 2 =0
x
y
(4)
=0
y
p is not a function of y
=0
z
p is not a function of z
p is a function
of x only
2 u 1 p
=
2
y
x
(5)
p
It means p =
x
= constant
2 u 1 p p
=
=
2
y
x
(6)
p
where p =
is the constant pressure gradient in the x-direction
x
Since u is only a function of y, the partial derivative becomes an
ordinary derivative.
Therefore, (5) becomes
d 2u
p
=
2
dy
(7)
p y 2
u=
+ C1 y + C 2
2
(8)
u=0
u=0
at
at
y=a
y = a
no-slip
boundary
condition
y=a
Fixed plate
y
x
2a
Fluid flow direction
y = a
Fixed plate
C1 = 0
and
p a 2
C2 =
2
p 2
(
u=
a y 2 ) Parabolic
velocity profile
2
(8)
u =U
u=0
at
at
y=a
y = a
no-slip
boundary
condition
y=a
Moving plate
at velocity U
2a
Fluid flow direction
y = a
Fixed plate
and
U p a 2
C2 = +
2
2
u=
p 2
U
(
(a + y )
a y2 )+
2
2a
y=a
Moving plate
at velocity U
y
x
2a
Fluid flow direction
y = a
Fixed plate
Therefore, we get
u=
U
(a + y ) Linear velocity profile
2a
Point to remember:
- Pressure gradient gives parabolic profile
- Moving wall gives linear profile
d 2u
g
=
sin
2
dy
x direction :
y direction :
z direction :
u = function (t , x, y , z )
v = function (t , x, y , z )
w = function (t , x , y , z )
(1)
Steady, fully developed flow and therefore no change in time and in the flow
direction. Channel is not bounded in the z-direction and therefore nothing
happens in the z-direction.
x direction :
u = function ( y )
y direction :
z direction :
v = function ( y )
w=0
(2)
v
=0
y
u v w
+
+
=0
x y z
v = constant
or
v=0
v
z
v=0
w=0
y - component:
z - component:
+ 2 + g x = 0
x
y
p
y - component:
+ g y = 0
y
z - component:
p
+ g z = 0
z
y - component:
z - component:
2u
= gx
2
y
d 2u
g
=
sin
2
dy
p
= g y = g cos
y
=0
z
p is not a function of z
(4)
What was
asked to be
derived in
Exercise 1
d 2u
g
=
sin
2
dy
done
d 2u
g
=
sin
2
dy
(4)
z
At y = 0, u = 0
At y = d, du = 0
dy
Applying BC 2, we get
du g
=
sin y + A
dy
A =
sin d
(5)
(6)
du g
=
sin (d y )
dy
(7)
z
d
y2
+ B
sin dy
Integrating equation (7), we get u =
2
Applying BC 1, we get
B=0
(8)
(9)
y2
Combining equations (8) and (9), we get u =
sin dy
2
(10)
d 2u
g
=
sin
2
dy
done
done
d 2u
g
=
sin
2
dy
(4)
z
BC 1:
At y = 0, u = 0
BC 2:
At y = d, du = 0
dy
u=0
du g
=
sin y + A
dy
g
y2
sin
+ Ay + B
Integrating equation (12), we get u =
2
d
sin
2
(12)
(13)
g
dy y 2
u =
sin
2
2
(14)
z
d
Q = u dy
0
g
dy y 2
Q =
sin dy
2
2
0
d
g
dy
g
d 3 d 3 gd 3
y
=
Q =
sin
sin =
sin
6 0
6 12
4
4
2
(15)
z
d
y2
u =
sin dy
2
gd 3
Q=
sin
3
(10)
(11)
g
dy y 2
u =
sin
2
2
gd 3
Q=
sin
12
(14)
(15)
3Q
d =
(1 / 3 )
u=
1 exp(Vy / )
U
1 exp(Vh / )
V
U
h
x direction :
u = function ( y )
y direction :
z direction :
v = function ( y )
w=0
(1)
u v w
+
+
=0
x y z
v
=0
y
v = constant or v = 0
v =V
U
h
w=0
y - component:
z - component:
2u
u
p
V
+ 2
=
y
y
x
p
= g
y
p
=0
z
d 2 u V du
=
2
dy
dy
(3)
V
U
h
d 2 u V du
du
V
=
=
where
=
dy 2 dy
dy
(3)
At y = 0, u = 0
BC 2:
At y = h, u = U
du
= exp(y + A)
dy
u=
(4)
(5)
exp(y + A) + B
0=
U=
exp( A) + B
exp(h + A) + B
(6)
(7)
exp( A) =
U
exp(h ) 1
B=
exp( A) =
U
exp(h ) 1
u=
U
U
1 exp(y )
exp(y )
=
U
exp(h ) 1
exp(h ) 1 1 exp(h )
u=
1 exp(Vy / )
U
1 exp(Vh / )
(8)
V
U
h
u=
1 exp(Vy / )
U
1 exp(Vh / )
(8)
(i) For small V, expand exp(Vy/) and exp(Vh/) using Taylor series as follows:
(Vy / ) 2 (Vy / ) 2
+
+ K
1 1 + (Vy / ) +
3!
2!
U
u=
(Vh / ) 2 (Vh / ) 2
1 1 + (Vh / ) +
+
+ K
2!
3!
For small V, we can ignore the terms with power. We then get
Vy /
y
u=
U= U
Vh /
h
V
U
u=
1 exp(Vy / )
U
1 exp(Vh / )
(8)
For very large Vh/, exp(Vh/) goes to infinity. Therefore. Divide equation (8) by
exp(Vh/). We then get
u=
exp(Vy / ) exp( Vh / )
u=
U
( 1)
u = U exp[ V (h y ) / ]
U
h
Thanks