TP III Fluid Mechanics Problem Sheet 1: Unidirectional flows

Problem sheet solutions

Prof. Omar K. Matar, Department of Chemical Engineering,
Imperial College London, ACEX 506, [email protected].

Question 1
Problem statement

W y=H

u = u(y)

P1 P2

y
z
x
y=0

Schematic for flow along a channel

A fluid of density ρ and viscosity µ flows through a channel of height H,

length B, and width W . The pressure at the inlet and outlet of the channel is
P1 and P2 , respectively. Determine the velocity component u as a function of y,
and, subsequently, the flow rate, all in terms of P1 and P2 . Assumptions:

(a) Neglect v and w; assume steady flow; neglect gravity;

(b) Start from the Navier-Stokes equations

Problem solution
• The only non-zero velocity component is u = u(y).

• To determine u(y), look at the x-component of the Navier-Stokes (NS)

equations:
 2
∂ u ∂ 2u ∂ 2u
  
∂u ∂u ∂u ∂w ∂p
ρ +u +v +w = − +µ + + + ρgx (1)
∂t ∂x ∂y ∂z ∂x ∂x2 ∂y 2 ∂z 2

• We now introduce assumptions and approximations.

1
 ∂u
• Steady-state ⇒ ∂t
= 0.
∂2u ∂2u
• u = u(y) so u ∂u
∂x
= w ∂u
∂z
= ∂x2
= ∂z 2
= 0.
• Since v = 0, v ∂u
∂y
= 0.
• Neglect gravity, so ρgx = 0.
• As a result of the above, Eq. (1) becomes
∂p ∂ 2u
0=− + µ 2. (2)
∂x ∂y

• Consider the y-component of the NS:

 2
∂ v ∂ 2v ∂ 2v
  
∂v ∂v ∂v ∂v ∂p
ρ +u +v +w =− +µ + + + ρgy (3)
∂t ∂x ∂y ∂z ∂y ∂x2 ∂y 2 ∂z 2

• But, v = 0 and we neglect gravity, so we are left with

∂p
= 0, (4)
∂y
so p = p(x) only.
• Note from Eq. (2) that since u = u(y), µ∂ 2 u/∂y 2 is independent of x.
Therefore, ∂p/∂x is also independent of x. Since p 6= p(y), it follows that
∂p
= constant. (5)
∂x
• Integration of Eq. (2) w.r.t. y, twice, then gives (note that the partial
derivatives have become total derivatives)
du 1 dp
= y + c1 , (6)
dy µ dx
1 dp y 2
u = + c1 y + c2 . (7)
µ dx 2

• Boundary conditions: no-slip, i.e. u = 0, at y = (0, H). Applying this at

y = 0 gives
0 = 0 + 0 + c2 ⇒ c2 = 0. (8)
Applying the no-slip condition at y = H gives
1 dp H 2
0= + c1 H + c2 . (9)
µ dx 2
So, c1 = −(1/µ)(dp/dx)(H/2), hence
1 dp 1 2 dp P1 − P2 P2 − P 1
u= (y − Hy), where = = , which is negative.
µ dx 2 dx 0−B B
(10)

2
• The flow rate through the channel is
Z A Z H
1 dp 1 2
Q= udA = W (y − Hy)dy
0 0 µ dx 2
W dp H 2 W dp H 3 H 3
Z
= (y − Hy)dy = ( − )
2µ dx 0 2µ dx 3 2
W H 3 dp
= − . (11)
12µ dx

Note that since dp/dx < 0, Q > 0.

3
Question 2
Problem statement

g
x

W (x)

Schematic for film flowing down an incline

Consider the steady flow of a liquid along a flat surface inclined with an
angle α to the horizontal, as shown in the figure above. The density ρ and
viscosity µ of the liquid are constant. We assume that we are sufficiently far
away from the entrance and outlet of the flow that their respective effects can
be ignored. The pressure in the surrounding gas is constant along the gas-liquid
interface (represented by the inclined dotted line in the figure above). Starting
from the Navier-Stokes equations, determine the velocity profile w(x). As one of
the boundary conditions, use that σxz = 0 at the interface.

Problem solution
• With reference to the figure above, at x = h,

p = p0 , a constant, and σxz = 0. (12)

That is, the pressure above the film is constant and the shear stress at the
interface is zero.

• At x = 0, no-slip applies: w = 0.

• Note that we have assumed steady-state, and that w is the only non-zero
component and it depends only on x: w = w(x).

4
• As a result, the z-component of the NS becomes (note the use of the total
derivative for the second derivative of w w.r.t. x)
∂p d2 w
0 = − + µ 2 + ρgz ,
∂z dx
∂p d2 w
0 = − + µ 2 + ρg sin(α). (13)
∂z dx

• The x-component of the NS becomes

∂p
0 = − + ρgx ,
∂x
∂p
0 = − − ρg cos(α). (14)
∂x

• Integration of this equation yields

p(x, z) = −ρgx cos(α) + c1 (z). (15)

• At x = h, p = p0 (see above), so

p(x, z) = ρg(h − x) cos(α) + p0 . (16)

• Substitution of this result in Eq. (13) gives

d2 w
0=µ + ρg sin(α). (17)
dx2

• Integration of this equation w.r.t. x gives

dw
µ = −ρgx sin(α) + c2 , (18)
dx
and one more integration gives
1
µw(x) = − ρgx2 sin(α) + c2 x + c3 , (19)
2
where c2 and c3 are constants.

• c2 and c3 can be eliminated by recalling that σxz = 0 at x = h and w = 0

at x = 0 (see above):

c2 = ρgh sin(α), and c3 = 0. (20)

• Hence w(x) becomes

1 x2
w(x) = ρg sin(α)(hx − ). (21)
µ 2

5
Question 3
Problem statement

Consider the steady flow of a fluid of constant density ρ and viscosity µ in a

“very long”, horizontal tube of length L and radius R; we specify that the tube is
very long because we want to assume that there are no end effects. Require that
w only depends on r. Solve for w(r) by starting from the Navier-Stokes equations.
Note that in the “Fundamental Equations of Fluid Mechanics” document, ur = u,
uθ = v, and uz = w:
 
∂u ∂u 1 ∂u 1 2 ∂u ∂p
ρ +u + v − v +w = ρgr −
∂t ∂r r ∂θ r ∂z ∂r
2 2
   
∂ 1 ∂ 1∂ u 2 ∂v ∂ u
+µ (ru) + 2 2 − 2 + (22)
∂r r ∂r r ∂θ r ∂θ ∂z 2

 
∂v ∂v 1 ∂v 1 ∂v 1 ∂p
ρ +u + v + uv + w = ρgθ −
∂t ∂r r ∂θ r ∂z r ∂θ
2 2
   
∂ 1 ∂ 1∂ v 2 ∂u ∂ v
+µ (rv) + 2 2 + 2 + (23)
∂r r ∂r r ∂θ r ∂θ ∂z 2

 
∂w ∂w 1 ∂w ∂w ∂p
ρ +u + v +w = ρgz −
∂t ∂r r ∂θ ∂z ∂z
2 2
   
1 ∂ ∂w 1∂ w ∂ w
+µ r + 2 2 + (24)
r ∂r ∂r r ∂θ ∂z 2

Problem solution
• Assumptions: steady flow (no time dependence), axisymmetric flow (no
θ dependence), and ‘long’ tube; hence, end effects are negligible, so no
dependence on z. Thus, w = w(r).

• Use the NS in cylindrical coordinates.

6
• Since w = w(r) only, the z-component of the NS becomes (note the use of
total derivatives)  
µ d dw ∂p
r = . (25)
r dr dr ∂z

• Also, since u = 0, the r-component of the NS becomes

∂p
0= , (26)
∂r
so that p = p(z) potentially. However, from Eq. (25), since w does not
depend on z, ∂p/∂z also does not depend on z. Thus,

∂p P 0 − PL PL − P0
= constant = = , which is negative. (27)
∂z 0−L L
so that Eq. (25) becomes
   
d dw 1 PL − P0
r = r = Gr. (28)
dr dr µ L

• Integrate this equation to get,

dw 1 2
r = Gr + c1 , or
dr 2
dw 1 c1
= Gr + , (29)
dr 2 r
where c1 is a constant.

• At r = 0, we assume symmetry so that dw/dr = 0. Hence, c1 = 0, thus

dw/dr = (1/2)Gr.

• Integration of this equation gives

1
w = Gr2 + c2 . (30)
4

• At r = R, we have the no-slip condition: w = 0. Hence, c2 = −(1/4)GR2 .

Thus, w(r) becomes
1
w(r) = G(r2 − R2 )
4  
1 PL − P 0
w(r) = (r2 − R2 ). (31)
4µ L

• Note that w(r) > 0 if P0 > PL .

7
Question 4
Problem statement

Fluid is contained in the annulus between two long concentric cylinders of

radii a1 and a2 rotating about their common axis with angular velocity Ω1 and
Ω2 , as shown in the figure above. This simple flow is called ‘Couette’ flow, and is
used to measure viscosity (see Lectures 4 and 3 of the First and Third Year Fluid
Mechanics notes, respectively). Find the velocity distribution within the annulus.
Assume that the velocity is only in the θ direction; that velocity and pressure are
independent of θ and z; and start from the Navier Stokes equations in cylindrical
coordinates. Note that as in Question 3 above, in the “Fundamental Equations
of Fluid Mechanics” document, ur = u, uθ = v, and uz = w:
 
∂u ∂u 1 ∂u 1 2 ∂u ∂p
ρ +u + v − v +w = ρgr −
∂t ∂r r ∂θ r ∂z ∂r
2 2
   
∂ 1 ∂ 1∂ u 2 ∂v ∂ u
+µ (ru) + 2 2 − 2 + (32)
∂r r ∂r r ∂θ r ∂θ ∂z 2

 
∂v ∂v 1 ∂v 1 ∂v 1 ∂p
ρ +u + v + uv + w = ρgθ −
∂t ∂r r ∂θ r ∂z r ∂θ
2 2
   
∂ 1 ∂ 1∂ v 2 ∂u ∂ v
+µ (rv) + 2 2 + 2 + (33)
∂r r ∂r r ∂θ r ∂θ ∂z 2

 
∂w ∂w 1 ∂w ∂w ∂p
ρ +u + v +w = ρgz −
∂t ∂r r ∂θ ∂z ∂z
2 2
   
1 ∂ ∂w 1∂ w ∂ w
+µ r + 2 2 + (34)
r ∂r ∂r r ∂θ ∂z 2

8
Problem solution
• The assumptions are: steady flow, no θ-dependence (i.e. axisymmetric
flow), no z−dependence. Thus, v is the only non-zero component.

• Consider the θ-component of the NS in cylindrical coordinates (note the

use of total derivatives)
 
d 1 d
0= (rv) . (35)
dr r dr

• Integration of this equation gives

1 d
(rv) = c1 ,
r dr
d
(rv) = c1 r, (36)
dr
where c1 is a constant. Integration of this equation gives
1 2
rv = c1 r + c2 ,
2
1 c2
v = c1 r + . (37)
2 r

• The boundary conditions are:

At r = a1 , v = Ω1 a1 , and at r = a2 , v = Ω2 a2 . (38)

• Application of these conditions gives

1 c2 1 c2
Ω1 a1 = c1 a1 + , Ω2 a2 = c1 a2 + . (39)
2 a1 2 a2

• Simultaneous solution of these equations gives

Ω2 a22 − Ω1 a21
c1 = 1 2 1 2 ,
2
a 2 − a
2 1
1
c2 = Ω1 a21 − a21 c1 . (40)
2