See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/383204904

48 Fluid Mechanics Problems Problem 1: 1-D Coating Problem

Preprint · August 2024

DOI: 10.13140/RG.2.2.35095.59046

CITATIONS READS
0 58

1 author:

Brian G Higgins
University of California, Davis
145 PUBLICATIONS 2,497 CITATIONS

SEE PROFILE

All content following this page was uploaded by Brian G Higgins on 18 August 2024.

The user has requested enhancement of the downloaded file.

 BGHiggins/UCDavis/Aug16_2024

48 Fluid Mechanics Problems

Teaching/BGHTeaching1/ECH141_FluidMechnaics_2018/NewFluidMechanicsCourse_2023

Problem 1: 1-D Coating Problem

As part of a coating process, a sheet of paper passes between two flat, solid surfaces as shown below.
The sheet is moving with velocity U in the x-direction, and is positioned mid-way between the walls.
The coating fluid is Newtonian with viscosity μ. The length (in the x-direction) L of the walls is large
compared to the dimension H, and the pressure drop from inlet to outlet is ΔP / L. A schematic of the
process is shown in the figure below. You may assume that H/L<<1.

Tasks
(a) Calculate the velocity profile in the region 0 < y < H, assuming that the pressure drop ΔP / L is
known.

(b) To avoid tearing the sheet of paper, it is desired to adjust the pressure drop such that the force
required to move the sheet is approximately zero. Calculate the value of ΔP / L that will result in zero
force being required to move the sheet at a velocity U.

Solution

Part (a)
(a) Since L >> H, we can neglect entrance and exit effects, and take the flow to be unidirectional
(rectilinear). Thus the velocity field is given by
v = vx ex , where vx = vx (y), for 0≤y≤H
2 ECH41MiscProblemsPart2_2024.nb

For the given velocity field the continuity equation for an incompressible fluid is satisfied identically:
∂ vx ∂ vy
∇· v=0  + =0
∂x ∂y

The linear momentum equations with the given velocity field become
∂ 2 vx ∂
0=- +μ , 0= -
∂x 2 ∂y
y

Note: we have used the result that v · ∇ v ≡ 0, when v = vx (y) ex . From the above equations we
can deduce that the pressure field is a function of x only, i.e.,
 =  (x)

Hence the x-momentum equations become

 2 vx
=μ
x  y2

We can integrate this equation to get

1 Δ
vx (y) = - y2 + C1 y + C2
2μ L

The constants of integration can be determined from the no-slip BCs

BC1: vx (0) = U

BC2: vx (H) = 0

Boundary condition BC1 implies that C2 = U. From BC2 we find

U H Δ
C1 = - +
H 2μ L

Thus the velocity profile becomes

H2 Δ y 2 y y
vx (y) = -  + + U 1 - 
2μ L H H H

The first term on the RHS is the component of flow due to the pressure gradient: plane Poiseuille flow.
The second terms is the component of the flow due to the motion of the sheet (viscous drag): plane
Couette flow.

Part b
(b) For no force on the paper sheet , we require that
 vx
τx y = μ = 0 at y = 0
y
 ECH41MiscProblemsPart2_2024.nb 3

From the velocity profile found in Part(a), we have

 vx H2 Δ 2y 1 U
= - + -
y 2μ L H2 H H

At y = 0 the velocity gradient is

 vx H Δ U Δ 2μU
= -  =
y 2μ L H L H2

Hence for
H Δ U
τx y = 0  μ  - =0
2μ L H

This condition is satisfied if

Δ 2μU
=
L H2

Problem 2: Rectilinear Flow

Consider flow between eccentric rotating disks (as illustrated in the figure below.

Both disks rotate with constant angular velocity Ω. The axes of rotation of the disks are separated by a
distance α, and the gap between the disks is β. The spatial coordinates of the fluid particles are given
by
x1 = X1 Cos[Ω t] - X2 Sin[Ω t]

x2 = X1 Sin[Ωt] + X2 Cos[Ω t] + γ X3

x3 = X3

where γ=α/β, and Xi are the material coordinates of the fluid particles in the undeformed state (t=0).
4 ECH41MiscProblemsPart2_2024.nb

Tasks:
(i) Determine the velocity field (in terms of the spatial coordinates xi ) for the motion described above.

(ii) If the temperature field between the rotating disks is given by

x1 x2
θ (x, t) = +
(1 + t) (1 + t)

determine the time rate of change of temperature of a fluid particle in the flow field.

Solution Part (i)

In the problem statement we are given the the spatial coordinates of the motion
x1 = X1 Cos[Ω t] - X2 Sin[Ω t]

x2 = X1 Sin[Ωt] + X2 Cos[Ω t] + γ X3

x3 = X3

Thus the velocity of a fluid particle between the disks is given by

x
=v
t X

Using the spatial coordinates given in the problem we find

v1 = -Ω X1 Sin[Ωt] - Ω X2 Cos[Ω t]

v2 = Ω X1 Cos[Ω t] - Ω X2 Sin[Ω t]

v3 = 0

This is a material description of the problem. To determine the spatial description we need to write
the velocity field in terms of xi .

From inspection of the spatial coordinates we can deduce that

-Ω X1 Sin[Ω t] - Ω X2 Cos[Ω t] = -Ω x2 + Ω γ x3

Ω X1 Cos[Ω t] - Ω X2 Sin[Ω t] = Ω x1

Thus
v1 = -Ω x2 + Ω γ x3 , v2 = Ω x 1 , v 3 = 0

or
v = -Ω (x2 - γ x3 ) e1 + Ω x1 e2
 ECH41MiscProblemsPart2_2024.nb 5

Solution Part (ii)

The time rate of change of temperature of a fluid particle is given by
θ ∂θ
= + v · ∇θ
t X ∂t
Thus we can compute this quantity either in terms of the spatial coordinates of material coordinates.
The former calculation is direct. Recall if we are given the temperature field
x1 x2
θ (x, t) = +
(1 + t) (1 + t)

then
θ x1 x2 (x1 + x2 )
=- - =-
t X (1 + t)2 (1 + t)2 (1 + t)2

∂ ∂ 1 1
∇ θ = e1 +e2 θ= e1 + e2
∂ x1 ∂ x2 (1 + t) (1 + t)

We showed from Part (i) that

v = -Ω (x2 - γ x3 ) e1 + Ω x1 e2

Thus
Ω
v · ∇θ = (x3 γ - x2 + x1 )
(1 + t)

so that
θ (x1 + x2 ) Ω
- + (x3 γ - x2 + x1 )
t X (1 + t)2 1+t

Problem 3 : Analysis of Flow Field

The velocity components for the steady flow of a Newtonian fluid are
A x2
v1 = -ω x2 -
x21 + x22
A x1
v2 = ω x1 +
x21 + x22
v3 = 0

where ω and A are constants.

Tasks :
(i) Is this flow incompressible?
6 ECH41MiscProblemsPart2_2024.nb

(ii) Determine the components of the rate of strain tensor (also called the rate of deformation tensor)
at x1 = 1, x2 = 2.

(iii) What is the force per unit area exerted on a force transducer placed in the flow at the point (1,2).
The direction in which the test surface faces is e1 + 2 e2 .

Solution
(i) For a flow to be called “incompressible”, we require ρ = constant so that
∂ρ ∂ρ
+ ∇ ·ρ v = 0  + ρ ∇ · v + v · ∇ρ = 0
∂t ∂t
Thus if ρ=constant, then
∂ v1 ∂ v2 ∂ v3
∇ ·v = 0  + + =0
∂ x1 ∂ x2 ∂ x3

In this problem we are given the velocity. A simple calculations shows

∂ v1 A x1 x2
=2
∂x 2
x1 2 + x2 2 

∂ v2 2 A x1 x2
=-
∂x 2
x1 2 + x2 2 

∂ v3
=0
∂ x3

Thus
∇ ·v = 0

We are told that the flow is steady but are given no information that ρ is constant. Hence our continu-
ity equation reduces to
v · ∇ρ = 0

This means that if ρ is not constant, the velocity must be orthogonal to the density gradient.

The rate of strain tensor is

1
d =  ∇ v + ∇ v †
2
The components of the velocity gradient ∇ v are
∂ ∂ vj
∇v = ei vj ej = ei ej
∂ xi ∂ xi
 ECH41MiscProblemsPart2_2024.nb 7

or in matrix notation
∂ v1 / ∂ x 1 ∂ v2 / ∂ x 1 ∂ v3 / ∂ x 1
[∇ v] = ∂ v1 / ∂ x 2 ∂ v2 / ∂ x 2 ∂ v3 / ∂ x 2
∂ v1 / ∂ x 3 ∂ v2 / ∂ x 3 ∂ v3 / ∂ x 3

We have already determined ∂ vi / ∂ xi . The off-diagonal terms are

∂ v1 2 A x2 2 A
= - -ω
2
∂ x2 x1 2 + x2 2  x1 2 + x2 2

∂ v2 2 A x1 2 A
=- + +ω
2
∂ x1 x1 2 + x2 2  x1 2 + x2 2

and since the components of ∇ v † are

∂ v1 / ∂ x 1 ∂ v1 / ∂ x 2 ∂ v1 / ∂ x 3
∇ v  =
†
∂ v2 / ∂ x 1 ∂ v2 / ∂ x 2 ∂ v2 / ∂ x 3
∂ v3 / ∂ x 1 ∂ v3 / ∂ x 2 ∂ v3 / ∂ x 3

then
4 A x1 x2 2 A -x1 2 + x2 2 
2 2 2 2
0
x1 + x2  x1 2 + x2 2 
1
[d] = 2 A -x1 2 + x2 2  4 A x1 x2
2 2
- 2
0
x1 2 + x2 2  x1 2 + x2 2 
0 0 0

Evaluating this expression at x1 = 1, x2 = 2 gives

4A 3A
25 25
[d ] = 3A
-4A
25 25

(iii) To determine the force/area acting on the test surface, we need to compute the traction vector
t (n). We are told that the test surface is oriented in the e1 + 2 e2 direction. A unit normal in this
direction is
1
n= ( e1 + 2 e2 )
5

Also
t (n) = T · n = (-p I + τ) · n = -p n + τ · n

For a Newtonian fluid, we have

p 1
t (n) = - (e1 + 2 e2 ) + 2 μ di j ei ej · (e1 + 2 e2 )
5 5

p 2 4
=- (e1 + 2 e2 ) + μ di1 ei + μ di 2 ei
5 5 5
8 ECH41MiscProblemsPart2_2024.nb

Simplifying we get
p 2 2
t (n) = - (e1 + 2 e2 ) + μ d11 e1 + d12 e2
5 5 5

4 4
+ μ d12 e1 + μ d22 e2
5 5

But
4A 3A 4A
d11 = , d12 = d21 = , d22 = -
25 25 25
Hence
p 4Aμ 2Aμ
t (n) = - (e1 + 2 e2 ) + e1 - e2
5 5 5 5 5

- 5p+ 4Aμ 2(5p+Aμ)

= e1 - e2
5 5 5 5

Problem 4 : Drag Force

A biologist has observed a crustacean approximately 1mm in diameter moving at approximately 2.53
cm/s. She would like to know the drag force F resisting the organism motion, but cannot measure
such a small quantity directly. A scale model is therefore built 100 times larger than the crustacean,
and is tested in glycerin. Densities an viscosities for water and glycerin are given below.

Tasks :
(a) Recall that the dimensionless force F / ρ U2 on a submerged object is a function of the Reynolds
number Re only. At what velocity should the model crustacean be tested so that the Reynolds num-
bers for the model and the actual organism are equal?

(b) Assuming the force measured at the velocity from part (a) is 1.3. N, what was the drag force on the
crustacean?
μ ρ
Water : 0.001 Kg / m · s 1000 Kg  m3
Glycerine : 1.5 Kg / m · s 1263 Kg  m3

Solution
◼ Part (a)

First we calculate the Reynolds Number for the crustacean

 ECH41MiscProblemsPart2_2024.nb 9

ρUD 1000 Kg  m3  (0.0253 m / s) (0.001 m)

(e)C = = = 25.3
μ (0.001 Kg / (m · s)

The Reynolds number for the model is

ρUD 1263 Kg  m3  (Um ) (0.1 m)
(e)M = =
μ (1.5 Kg / (m · s)

If the two Reynolds numbers are to be the same, then

(e)M = 25.3 ⟹ Um = 0.3 m / s

Thus the scale model must be tested at 0.3 /m/s

◼ Part (b)

We know the force is related to the Reynolds number as

FC
= f (ReC )
ρw U2C

where f is some unknown function of the Reynolds number. Similarly for the model we have
FM
= f (ReM )
ρG U2M

In the above expressions, ρw and ρG are the densities of water and glycerine, respectively. Now since
ReC = ReM

it follow that
f (ReM ) = f (ReC )

Thus
FC FM
=
ρw U2C ρg U2M

Hence
ρW U2C
FC = FM
ρG U2M 

Evaluating the RHS of the above expression gives

1000 0.0253 2
FC = (1.3 N)
1263 0.3

= 7.32  10-3 N

Note: this force is sufficiently small in magnitude that it will be challenging to determine it
experimentally!
10 ECH41MiscProblemsPart2_2024.nb

Problem 5 : Kinematics Part 1

A 2-dimensional flow field in the x-y plane, bounded by a moving solid plane at y=0 is studied experi-
mentally. Sufficiently far away from the moving solid plane, the velocity of a fluid particle is measured
moving along a horizontal streamline that coincides with the x- axis. The x-component of the two
dimensional incompressible flow field was experimentally found to be describes by the equation
u = x2

Tasks :
Along this streamline determine an expression for
(i) the rate of change of the v component of velocity with respect to y
(ii) the acceleration of the fluid particle along the streamline
(iii) the pressure gradient in the x-direction
(iv) the components of viscous stress tensor τ at a point P (x=3, y=2) on the streamline, assuming that
u = x 2 is a good approximation of the flow even at y=0.

Problem 1 Solution
(i) The continuity equation for this flow is
∂u ∂v
+ =0
∂x ∂y

Thus along the streamline (say at y = ys )where the measurement is made

∂v
= -2 x, at y = ys
∂y

Next we would like to know how v- component of velocity varies with y. This is best accomplished by
integrating the continuity equation from y=0 to y = ys .
ys ∂u ys ∂v
 y +  y = 0
0 ∂x 0 ∂y

We can write this integral as

 ys
 u  y + v (x, ys ) - v (x, 0) = 0
x 0

Noting that v(x,0)=0 (recall at y=0 the flow is bounding by a solid surface moving parallel to the x-axis.
Thus we have
 ys
 u  y = -v (x, ys )
x 0

At this point we do not know how u(x,y) varies with y. All we know is that at y = ys , u( x, ys ) = x 2 . We
will suppose therefore that u is only a function of x. Clearly, this would be reasonable if the speed of
 ECH41MiscProblemsPart2_2024.nb 11

the “plate” at y=0 is a function of x. If we make this assumption then

u (x, y) = x2 for all y > 0

At the point P on the streamline we have

u
ys = -v (x, ys )
x
which gives for any y
v (x, y) = -2 x y

Thus
∂v
= -2 x, v (x, y) = -2 x y
∂y

Note that the value of v(x,y) is required for later calculations

(ii) The acceleration a = a x e x + a y e y of the fluid particle along the streamline is

∂u ∂u ∂v ∂v
ax = ρ u +v , ay = ρ u +v
∂x ∂y ∂x ∂y

for the given velocity field we have

ax = ρ x2 (2 x) + (-2 xy) (0) = 2 ρ x3

ay = ρ x3 (-2 y) + (-2 x y) (-2 x) = ρ -2 x3 y - 4 x2 y

= -2 ρ x2 y (x + 1)

Hence the acceleration along the streamline

a = 2 ρ x3 ex - 2 ρ x2 y (x + 1) ey

(iii) From the x-momentum equation we have

∂p ∂2 u ∂2 u
= -ax + μ +
∂x ∂ x2 ∂ y2

= -2 ρ x3 + μ (2) + μ (0)

= -2 ρ x3 - μ

(iv) The components of the viscous stress tensor are

2μ ∂u
μ  ∂u + ∂v
 4 μ x -2 μ y
∂x ∂y ∂x
[τ] = = 
μ  ∂u + ∂v
 2μ ∂v -2 μ y -4 μx
∂y ∂x ∂y

Evaluating at x=3 y=2 gives

12 ECH41MiscProblemsPart2_2024.nb

12 μ -4 μ
[τ] =  
-4 μ -12 μ

Problem 6 : Kinematics Part 2:

Consider the following flow field:

x1 2 x2 3 x3
v1 = , v2 = , v3 =
1+ t 1+ t 1+ t

Tasks
(i) Determine the acceleration of a fluid particle using a spatial description of the flow field.

(ii) Determine a material description of the flow field (i.e. the velocity is expressed in terms of
X1 , X2 , X3 ). What is the acceleration of the fluid particle in terms of material coordinates?

(iii) Show that (i) and (ii) are equivalent.

Problem Solution
(i) the acceleration of a fluid particle is given by
v Dv ∂v
a= ≡ = + v · ∇v
t  Dt ∂t
Thus for the spatial description
∂ v1 x1 ∂ v2 2 x2 ∂ v3 3 x3
=- , =- , =-
∂t (1 + t)2 ∂t (1 + t)2 ∂t (1 + t)2

and the components of v · ∇ v are

∂ v1 ∂ v1 ∂ v1 ∂ v1 x1
(v · ∇ v ) · e1 = v1 + v2 + v3 = v1 =
∂ x1 ∂ x2 ∂ x3 ∂ x1 (1 + t)2

∂ v2 ∂ v2 ∂ v2 ∂ v2 4 x2
(v · ∇ v ) · e2 = v1 + v2 + v3 = v2 =
∂ x1 ∂ x2 ∂ x3 ∂ x2 (1 + t)2

∂ v3 ∂ v3 ∂ v3 ∂ v3 9 x3
(v · ∇ v ) · e3 = v1 + v2 + v3 = v3 =
∂ x1 ∂ x2 ∂ x3 ∂ x3 (1 + t)2

Thus the spatial description of the acceleration is

∂ v1 ∂ v1
a1 = + v1 = 0,
∂t ∂ x1
 ECH41MiscProblemsPart2_2024.nb 13

∂ v2 ∂ v2 2 x2
a2 = + v2 = ,
∂t ∂ x2 (1 + t)2

∂ v3 ∂ v3 6 x3
a3 = + v3 = ,
∂t ∂ x3 (1 + t)2

(ii) To determine the material description we need the particle paths. Thus
 x1 x1
v1 = =  x1 = 1 (1 + t)
t 1+ t

 x2 2 x2
v2 = =  x2 = 2 (1 + t)2
t (1 + t)

 x3 3 x3
v3 = =  x3 = 3 (1 + t)3
t (1 + t)

The acceleration is then

2 x1
a1 = =0
 t2 

2 x2
a2 = = 2 2
 t2 

2 x3
a3 = = 6 3 (1 + t)
 t2 

(iii) Since we have shown that x2 = 2 (1 + t)2 , we can eliminate 2 in a2 to get

2 x2 2 x2
a2 = = 2 2 =
 t2  (1 + t)2

and for a3 we have

2 x3 6 x3
a3 = = 6 3 (1 + t) =
2
t  (1 + t)2

These are the spatial components of the velocity which are the same as found in (i)

Problem 7 : Kinematics Part 3:

Problem Statement
If v = z i + 2 x j - 4 x2 z k is a velocity field, determine the total mass flow rate M due to
v across a cylinder x2 + y2 = 25 lying in the first octant between the planes z = 0 and z = 3. That
14 ECH41MiscProblemsPart2_2024.nb

is, determine

M =  v · n S
S

where n is the unit normal to the surface S: Hint: If f (x, y, z) = 0, describes a surface, then its
unit normal is n = ∇ f / 〚∇ f〛, where 〚∇f〛 is the magnitude of the vector ∇ f. The projected area
theorem is useful.

Problem Solution
The unit normal to the surface f (x, y, z) = 0 is n = ∇ f / 〚∇ f〛. For our surface
f (x, y) = x2 + y2 - 25, the unit normal is
1
n = 2 x i + 2 y j 4 x2 + 4 y2 = (x i + y j)
5
The flux across the curved surface is :
1
v·n = (x z + 2 x y)
5
Thus
1
M =  v · n S =  (x z + 2 x y)  S
S S 5
To evaluate this integral we make use of the projection area theorem:
y
 x  z = 〚n · j〛  S   x  z = S
5
Hence
5
 v · n S =  (v · n)  x  z
S Sxz y

Thus the Mass Flow rate is given by

3 5 xz+2xy 3 5 xz
M=  x z =   + 2 x x z
0 0 y 0 0 25 - x2

But
t
 t = - a2 - t2
2 2
a -t

Thus the Mas flow rate across the curved surface is

3
M =  -z 25 - x2 + x2  - -z 25 - x2 + x2   z
0 x=5 x=0
 ECH41MiscProblemsPart2_2024.nb 15

3 195
M =  (5 z + 25)  z = = 97.5
0 2
Note that the velocity has a component in the k coordinate direction. Thus there will also be a
flux across the top and bottom surface at z = 0 and z = 3. We can use the projection theorem again
to find the flux across these surfaces. In this case n = ± k so that
v · n = ∓ x2 z, 〚n · k〛  S = dx  y
5 25-y2 625 π z 625 π z
Mxy = ∓   x2 z  x  y = -
0 0 16 z=0 16 z=3

Mxy = -368. × 155

Problem 8 : Kinematics Part 4:

Problem Statement
If v = 12 y i + 2 x j - 3 k is a velocity field, determine the total flow rate M due to v across the
plane x + 2 y + 8 z = 12 for x ≥ 0, y ≥ 0, z ≥ 0.

Problem Solution:
The unit normal to the surface f (x, y, z) = 0, is n = ∇ f / 〚∇ f〛. For our problem
f = x + 2 y + 8 z - 12. Thus ∇ f = i + 2 j + 8 k and 〚∇ f〛 = ∇ f · ∇ f = 69 so that the unit
normal is
1
n= (i + 2 j + 8 k)
69

The mass flow rate M across the surface with unit normal n is given by
x y
M =  v · n S =  v · n
S  〚n · k〛

where we have used the result from the projected area theorem
8
〚n · k〛  S =  x  y, where 〚n · k〛 =
69

and  is the projected area of the surface onto the x-y plane. Thus the limits of integration are
0 ≤ y ≤ 6 - x / 2, and 0 ≤ x ≤ 12. From our definitions we have
12 y + 4 x - 24
v·n =
69

Evaluating the integral for the mass flow rate gives

16 ECH41MiscProblemsPart2_2024.nb

12 6-x/2 3 x
M=  y+ - 3  x  y = 72
0 0 2 2

Problem 9 : Kinematics Part 5:

Problem Statement
Given the following simple plane flow
v1 = x1 / (1 + t), v2 = x2 , v3 = 0

Determine the streamlines at time t, the particle paths, and the streak-lines through xo . Plot the
streamlines in the x2 - x1 plane. Take xo = {1, 2}.

Problem Solution:
The path lines are found by integrating
 x1  x2  x3
v1 ≡ = x1 / (1 + t), v2 ≡ = x2 , v 3 ≡ =0
t t t
The result is
x1 = X1 (1 + t), x2 = X2 et , x3 = X3

We can use the first equation to get an explicit expression for t

t = x1 / X2 - 1

Substituting into the equation for x2 gives

x2 = X2 e-(1-x1 /X1 )

To plot the particle paths x1 - x2 plane, we select values for X1 , X2 and then compute the resulting
path. These calculations can be readily done in Mathematica using ParametricPlot . Here is a way
to plot the particle paths. First we define an array of potential {X1 , X2 } positions for our path lines:

I n [ ] : = InitialPositions = Transpose[{Range[1, 5], Range[1, 5]}]

O u t [ ] =

{{1, 1}, {2, 2}, {3, 3}, {4, 4}, {5, 5}}

Next we define a set of replacement rules that assign the variables X1 , Y1 to these positions:

I n [ ] : = InitialConditions = Map[Thread[{X1 , X2 }  #] &, InitialPositions]

O u t [ ] =

{{X1  1, X2  1}, {X1  2, X2  2}, {X1  3, X2  3}, {X1  4, X2  4}, {X1  5, X2  5}}

We would like also to mark the initial particles on the plot. This can be readily done using a graphic
primitive called Point. The following statement creates the graphic primitives for each of the
particles
 ECH41MiscProblemsPart2_2024.nb 17

I n [ ] : = InitialParticles = Map[Point[#] &, Transpose[{Range[1, 5], Range[1, 5]}]]

O u t [ ] =

{Point[{1, 1}], Point[{2, 2}], Point[{3, 3}], Point[{4, 4}], Point[{5, 5}]}

With those preliminaries out of the way, we now use ParametricPlot to graph our path lines. The
graphic primitives are added to the plot using the Epilog option.

I n [ ] : = ParticlePathLines =
ParametricPlot[Evaluate[{X1 (1 + t), X2 Exp[t]} /. InitialConditions],
{ t, 0, 2}, PlotStyle  Thick, PlotRange  All, AxesOrigin  {0, 0},
Epilog  {RGBColor[1, 0, 0], PointSize[0.03], InitialParticles},
Frame  True, FrameLabel  {Style["x1 ", 16], Style["x2 ", 16]}, AspectRatio  1,
PlotLegends  {"X0 =1,Y0 =1", "X0 =2,Y0 =2", "X0 =3,Y0 =3", "X0 =4,Y0 =4", "X0 =5,Y0 =5"}]
O u t [ ] =

30

X0 =1,Y0 =1
20
X0 =2,Y0 =2
x2

X0 =3,Y0 =3
X0 =4,Y0 =4
10 X0 =5,Y0 =5

0 5 10 15

x1

Consider next the streamlines. We can use the definition of the streamline: the velocity is everywhere
tangent to the streamline. Mathematically this is represented as
 xs
=v
α
where α is a parameter for the streamline (e.g arc length), and xs is the position vector for the STREAM-
LINE, at any point in time. Note: it is in general not the position vector for the particle path. For the
given velocity field we have
 x1  x2  x3
v1 ≡ = x1 / (1 + t), v2 ≡ = x2 , v 3 ≡ =0
α α α
Integrating while holding t fixed gives
18 ECH41MiscProblemsPart2_2024.nb

x1 = x1 (0) e α /(1+t) , x2 = x2 (0) eα , x3 = x3 (0)

where xi (0) are the positions in the flow field defined by α = 0. If we eliminate α we get
x2 1/(1+t)
x1 = x1 (0)
x2 (0)

We can readily visualize the stream lines using the program we developed for the particle paths. Let us
take a snapshot of the streamlines at t=2, the last time step used in the particle path calculation. The
location of the streamline is defined at the same positions we used to define the particles at time t=0.

I n [ ] : = Streamlines =
ParametricPlot[Evaluate[{X1 Exp[α / (1 + 2)], X2 Exp[α]} /. InitialConditions],
{ α, 0, 2}, PlotStyle  Thick, PlotRange  All, AxesOrigin  {0, 0},
Epilog  {RGBColor[1, 0, 0], PointSize[0.03], InitialParticles},
Frame  True, FrameLabel  {Style["x1 ", 16], Style["x2 ", 16]}, AspectRatio  1,
PlotLegends  {"X0 =1,Y0 =1", "X0 =2,Y0 =2", "X0 =3,Y0 =3", "X0 =4,Y0 =4", "X0 =5,Y0 =5"}]
O u t [ ] =

30

X0 =1,Y0 =1
20
X0 =2,Y0 =2
x2

X0 =3,Y0 =3
X0 =4,Y0 =4
10 X0 =5,Y0 =5

0 2 4 6 8 10

x1

If we combine the plots we readily see that the path-lines and streamlines do not coincide

I n [ ] : = Streamlines2 =
ParametricPlot[Evaluate[{X1 Exp[α / (1 + 2)], X2 Exp[α]} /. InitialConditions],
{ α, 0, 2}, PlotStyle  Thick, PlotRange  All, AxesOrigin  {0, 0},
Epilog  {RGBColor[1, 0, 0], PointSize[0.03], InitialParticles},
Frame  True, FrameLabel  {Style["x1 ", 16], Style["x2 ", 16]}, AspectRatio  1];
 ECH41MiscProblemsPart2_2024.nb 19

I n [ ] : = Show[Streamlines2, ParticlePathLines]
O u t [ ] =

30

X0 =1,Y0 =1
20
X0 =2,Y0 =2
x2

X0 =3,Y0 =3
X0 =4,Y0 =4
10 X0 =5,Y0 =5

0 5 10 15

x1

In the above plot each streamline and particle path line originates from the same point at time t=0.
Note though the trajectory of the streamline and the particle path line do not coincide.

For the streamline (at t=2) that passes through x1 (0) = 5, x2 (0) = 5, it is possible to find the
coordinates for a particle when α=2. We need to solve
x1 = x1 (0) e α /(1+t) , x2 = x2 (0) eα , with α = 2, t = 2

The coordinate values are.

x1 = 9.73867, x2 = 36.9453

Now let us use the particle path line equations to determine the trajectory of a particle that at t = 2,
is located at x1 = 9.74 , x2 = 36.95. Recall a particle path is defined by its initial location at
t = 0. Thus we need to solve for X1 , X2 Using the equations for the particle path lines we have
X1 = x1 / (1 + t) X2 = x2 e-t

Evaluating these expressions at t = 2, x1 = 9.74, x2 = 36.95 gives

X1 = 3.24622 X2 = 5.

Here is the trajectory of that particle

20 ECH41MiscProblemsPart2_2024.nb

I n [ ] : = aPathLine = ParametricPlot[Evaluate[{X1 (1 + t), X2 Exp[t]} /. {X1  3.24622, X2  5}],

{ t, 0, 2}, PlotStyle  Red, PlotRange  All, AxesOrigin  {0, 0},
Epilog  {RGBColor[1, 0, 0], PointSize[0.03], Point[{3.25, 5}]},
Frame  True, FrameLabel  {Style["x1 ", 16], Style["x2 ", 16]}, AspectRatio  1]
O u t [ ] =

30

20
x2

10

0 2 4 6 8 10

x1

If we superimpose this trajectory onto the previous plots we see that it becomes tangent to the stream-
line at t = 2, at location x1 = 9.7, x2 = 36.9
 ECH41MiscProblemsPart2_2024.nb 21

I n [ ] : = Show[aPathLine, Streamlines2, ParticlePathLines, PlotRange  {{0, 10}, {0, 35}}]

O u t [ ] =

35

30

25

20
X0 =1,Y0 =1
X0 =2,Y0 =2
x2

15 X0 =3,Y0 =3
X0 =4,Y0 =4
10
X0 =5,Y0 =5

0 2 4 6 8 10

x1

The above plot also shows that streamlines and path lines do not in general coincide when you have a
time dependent flow.

Consider next the computation of the streak line. We know that the path lines are given by
x1 = X1 (1 + t), x2 = X2 et , x3 = X3

The streak line is defined as the line on which all particles lie at some instant in time that passed
  
through a common point. Suppose the common point is x1 , x2 , x3 . Now for t < τ, the particles
  
that pass through x1 , x2 , x3 are
  
X1 = x1 / (1 + τ), X2 = x2 e-τ , X3 = x3 .

Hence the equation for the streak line is


x1  
x1 = (1 + t), x2 = x2 et-τ , x3 = x3
(1 + τ)

We can use Mathematica to plot the streak line when t = 2, and for 0 < τ < 2. The x - y point at
which we tag particles is (5, 5)
22 ECH41MiscProblemsPart2_2024.nb

I n [ ] : = aStreakLine = ParametricPlot[Evaluate[{3 X1 / (1 + τ), X2 Exp[2 - τ]} /. {X1  5, X2  5}],

{τ, 0, 2}, PlotStyle  {Thickness[0.005], RGBColor[0, 1, 0]}, PlotRange  All,
AxesOrigin  {0, 0}, Epilog  {RGBColor[1, 0, 0], PointSize[0.03], Point[{5, 5}]},
Frame  True, FrameLabel  {Style["x1 ", 16], Style["x2 ", 16]}, AspectRatio  1]
O u t [ ] =

30

20
x2

10

0 5 10 15

x1

In the following plot we combine the streak lines with the path lines and streamlines calculated
previously.
 ECH41MiscProblemsPart2_2024.nb 23

I n [ ] : = Show[aStreakLine, aPathLine, Streamlines2,

ParticlePathLines, PlotRange  {{0, 10}, {0, 35}}]
O u t [ ] =

35

30

25

20
X0 =1,Y0 =1
X0 =2,Y0 =2
x2

15 X0 =3,Y0 =3
X0 =4,Y0 =4
10
X0 =5,Y0 =5

0 2 4 6 8 10

x1

Problem 10 : Integral Analysis of Plunging Cylinder in tube

An object with a circular cross-section and a rounded end is being pushed into a cylinder with velocity
U as shown in the Figure below. The cylinder radius is R1 and the radius of the object is R2 . As the
object moves downward in the -z direction, water must flow up because the bottom end of the
cylinder is closed.

Tasks
Calculate the net flow Q of water up at the cross-section AA’.
24 ECH41MiscProblemsPart2_2024.nb

Solution
We will illustrate several methods to solve this problem.

◼ Method 1:

Consider at any instant in time a FIXED control volume that encompasses all the fluid below the plane
AA’.
A macroscopic balance over the fluid in the control volume gives

 ρ  V = - ρ v · n  A
 V A

Since the control volume V is fixed, and ρ is constant, we have

0 = - ρ v · n  A -  ρ v · ez  A
AR AAn

where AR is the area of the rounded end of the object, and AAn is the area of the annulus at AA ' with
unit normal n = ez . Note that v · n = 0 along the vertical side of the rod, along the tube side wall,
and tube bottom for all time. But the flow through the annulus is

Q= ρ v · ez  A = -  ρ v · n  A
AAn AR

At any instant in time the velocity of the rounded end of the rod is
v = -U ez

Thus

 ρ v · n  A = U ρ  ez · n  A
AR AR

so that
 ECH41MiscProblemsPart2_2024.nb 25

Q = U ρ  ez · n  A = ρ U π R22
AR

Note : we have used the projected area theorem to write

2
 ez · n  A = π R2
AR

◼ Method 2:

We pick a control volume as before except it moves with the motion of the rod. Thus the macroscopic
mass balance is

 ρ  V = - ρ ( v - w) · n  A
 V (t) A (t)

Along the vertical surface of the rod the no-penetration condition requires that (v - w) · n = 0. At
the rounded tip the control volume moves with velocity
w = -U ez

and since the mass flux at the rounded tip of the rod must be zero, the only non-zero term on the RHS
of the macroscopic mass balance is at A A’. Thus
V
ρ = -  ρ v · n  A = -Q
t An

The final task is to determine V

t
. Clearly the time rate of change of the control volume V(t) is related
to how fast the rod sweeps out volume
V
= v · n  A = -U  (ez · n )  A
t Arod Arod

The integral on the RHS is simply the projection of the surface area of the rod onto the xy-plane. Thus
V
= -U π R2 2  Q = U π R22
t
We can also determine V/t by the following method. Since
L2
V (t) = π R21 L2 - π  R (z)2  z
L1 (t)

where L2 is the length of the tube at the plane A A’ ( with the origin of the coordinate system located at
the bottom of the tube). Note that L2 (t) is the location of the rod tip at time t, and R (z) is some
arbitrary
profile for the rod. Now using Leibniz’s rule
V  L2  L2 -L1 (t)
= -π  R (z)  z = -π  R (ξ)2  ξ
t t L1 (t) t 0

 L1  L1
=π R22 = -U π R22 as U=-
t t
26 ECH41MiscProblemsPart2_2024.nb

Problem 11 : Kinematics Part 6 :

Problem Statement
In a certain region the flow, the velocity components are
v1 = -A x31 + x1 x22  e-kt

v2 = A x21 x2 - x32  e-kt

v3 = 0

where A and k are given constants, xi are the spatial coordinates and t is time.

Tasks:
(i) If the intensity of illumination of a fluid particle at (x1 , x2 , x3 ) at time t is
I = Be-3 t  x21 + x22 + x23 

where B is a known constant, determine the rate of change of illumination experienced by the fluid
particle which is at point (1, 2, -2) at time t = 1

Problem Solution
We are required to find how the intensity of a fluid particle  changes
I DI ∂I
≡ = + v · ∇I
t  Dt ∂t
where  = {1, 2, -2} at t = 1. The components on the RHS are
∂I
= -3 Be-3 t  r2 , where r2 = x21 + x22 + x23 
∂t
and
∂I ∂ 1 ∂ 1 ∂r
(∇ I) · e1 = = Be-3 t = Be-3 t
∂ x1 x1 r2 ∂r r2 ∂ x1

Now
∂ 1 1 ∂r 1 2 x1 x1
=- , and = =
∂r r2 r ∂ x1 2 x21 + x22 + x23 r

so that
x1 x2 x3
∇ I = Be-3 t e1 + e2 + e3 
r r r
 ECH41MiscProblemsPart2_2024.nb 27

Putting everything together we get

I 3 x1 x2 x3
= Be-3 t - + v1 + v2 + v3
t  r2 r r r

Substituting in for the velocity components v1 , v2 , v3 we get

I 3 1 1
= Be-3 t - - A x41 + x21 x22  e-kt + A x21 x22 - x2 x32  e-kt 
t 2 r r
 r
and then evaluating at  = {1, 2, -2} , t = 0 (recall  = x at t = 0) we get

I 3 A (1 + 4) (4 + 16) 1 25
= B - + +A  = B - + A
t  9 3 3 3 3

Problem 12 : Kinematics Part 7 :

Tasks
Find the trajectory for a fluid particle in the following flow field:
x1 x2 x3
v1 = , v2 = , v3 =
1 + a1 t 1 + a2 t 1 + a3 t

where a1 = a2 = 1, a3 = 0

Problem Solution
We start with the definition for the velocity of a fluid particle :
x
=v
t 

From the problem statement then

 xi
= xi / (1 + ai t), i = 1, 2, 3
t
We can integrate this equation to find the particle path line as a function of time:
 xi t
=
xi (1 + ai t)

1 1
  xi =  t
xi (1 + ai t)
28 ECH41MiscProblemsPart2_2024.nb

1
Ln (xi ) = Ln (1 + ai t) + Ln (c)
ai

Thus using the fact that xi (0) = i , we get

xi = i (1 + ai t)1/ai

From the problem statement we are given

a1 = 1, a2 = 1 / 2 , a3 = 0

Hence the parametric form for the path line is

t 1/2
x1 = 1 (1 + t), x2 = 2 1 + , x3 =  3
2
Using the first equation, we can solve for t:
t = (x1 / 1 ) - 1

Substituting this result into the expression for x2 gives the particle path in the x1 - x2 plane,
parametrized by 1 , 2
x2 = 2 (1 + (x1 / 2 1 ) - 1 / 2)1/2

Here is a plot of two trajectories ( 1 = 2 = 1 Red curve, and 1 = 1, 2 = 2, Blue curve)

 ECH41MiscProblemsPart2_2024.nb 29

I n [ ] : = res = Map[ParametricPlot[Evaluate[Re[{x, y} /. {X1  #〚1〛, X2  #〚2〛}]],

{t, -1, 5}, PlotStyle  {Thick, #〚3〛}, Axes  False,
Frame  True, FrameLabel  {Style["x", 16], Style["y", 16]},
Epilog  {PointSize[0.03], Point[{{1, 1}, {1, 2}}]},
PlotRange  All, AspectRatio  1] &, {{1, 1, Red}, {1, 2, Blue}}];
Show[{res〚1〛, res〚2〛}]
O u t [ ] =

3.5

3.0

2.5
y

2.0

1.5

1.0

0 1 2 3 4 5 6

In the above plot the particle trajectory for x < 1, corresponds to negative time.

Problem 13 : Rectilinear flow (Example 1)

Flow in a screw pump is sometimes modeled by using a shallow cavity like that shown below, which is
much longer than it is high (i.e., L >> B). The flat surface at y = B moves in the positive x-direction
with velocity U. Fluid near the plate at y = B is moved by viscous drag, but the presence of vertical
walls (past which no fluid flows) causes a recirculation to occur at the ends of the cavity. The flow
near these vertical walls is quite complicated, but in the middle of the cavity it is laminar and unidirec-
tional, as shown. The fluid in the cavity is Newtonian and incompressible.

Tasks
(a) Derive an expression for the velocity profile v = u (y) ex in the region of unidirectional, laminar
flow. Your answer may contain an unspecified pressure gradient dP / dx.

(b) Show by a macroscopic mass balance that the net volumetric flow across any vertical cross sec-
tion of the channel must be zero.
30 ECH41MiscProblemsPart2_2024.nb

(c) Use your results from parts (a) and (b) to derive an expression for the pressure gradient dP / dx.

(d) Calculate the force required to move the plate at a velocity U, given that the plate length is L and
its width (in the z direction) is W. Neglect effects due to the complicated flow at the channel ends. A
schematic of the flow is shown below:

Hint: The total force acting on the moving plate with surface area  in the e x direction is

Fx =  t (n) · ex  A, where t (n) = T · n

A

where n is the unit normal to the moving plate: n = e y , and T = Txx e

T = Txx ex ex + Txy ex ey

Problem 7 Solution
Problem 14 : Rectilinear flow (Example 2)

A cylindrical wire with radius R1 is being pulled through a horizontal cylindrical shell with radius R2 ,
where R2 > R1 . The wire (shown in orange in figure below) is centered within the shell, its velocity is U
in the z-direction, and there is no pressure drop over the shell length L. The fluid in the annulus
R1 < r < R2 (shown in blue in the figure below) is a Newtonian liquid with viscosity μ. You may assume
the fluid flow in the annulus is rectilinear, i.e., v = vz (r) ez . A schematic of the flow geometry is shown
below:

Note: the rectilinear flow v = vz (r ) ez in cylindrical coordinates satisfies the continuity equation identi-
cally, and the equations of motion for steady rectilinear flow in cylindrical coordinates with v = vz (r) ez
reduce to:
 ECH41MiscProblemsPart2_2024.nb 31

∂p 1   vz
- +μ r =0
∂z r r r

Tasks
For this problem you may assume the only driving force for rectilinear flow in the annulus is the
motion of the cylindrical wire, i.e., there is no pressure gradient in the flow. With the above assump-
tions calculate

a) the velocity profile vz (r) in the annulus.

b) the force in the z-direction required to move the wire with velocity U.

Hint for part (b): The local force on the moving wire can be determined from the traction vector (
which is the force per unit area): t (n) ≡ T · n, where n is the unit normal to the surface.

Then the total force acting on the wire with surface area  in the ez direction is

Fx =  t (n) · ez  A
A

Recall the stress tensor T is given by

T = -p I + τ

The components of τ in cylindrical coordinates noting that v = v (r ) ez are

∂ vz ∂ vz
τ=μ ez er + μ er ez
∂r ∂r
which in matrix notation is

0 τrz 0 ∂vz
∂r
[τ] =  =μ
τzr 0 ∂vz
0
∂r

Problem 8 Solution:
Since the problem statement indicates that the only force driving the flow is the motion of the wire,
we can assume that ∂p/∂z=0. Thus vz (r) must satisfy
1   vz
μ r =0
r r r
we can integrate this ODE to get
vz (r) = C1 Loge (r) + C2 note : loge (r) ≡ Ln (r)

The constants of integration can be found from the boundary conditions. At the surface of the moving
wire, the no slip condition gives
32 ECH41MiscProblemsPart2_2024.nb

BC1: vz (R1 ) = U

while at the wall of the tube, we have

BC2 : vz (R2 ) = 0

BC2 gives
C2 = -C1 Ln (R2 )

while BC1 gives

U = C1 Ln (R1 ) + C2

= C1 Ln (R1 ) - C1 Ln (R2 )

= C1 Ln (R1 / R2 )

Thus
C1 = U Ln (R2 / R1 ), C2 = -U Ln (R2 / R1 ) Ln (R2 )

and the velocity profile is

vz (r) = U Ln (R2 / R1 ) {Ln (r) - Ln (R2 )}

(b) The force on the moving wire can be determined from the traction vector (recall the traction vector
is just the local force per unit area):

t (n) = T · n

where n is the outward directed unit normal at the surface of the moving wire. For our problem
n = er . Now the component of the force/area acting along the surface of the moving wire is

fz ≡ t (n) · ez = (T · er ) · ez = ez · T · er

But
T = -p I + τ

so that
ez · T · er = -p (er · ez ) + ez · τ · er

∂ vz
= ez · τ · er = τrz = μ
∂r
 vz 1
= μ = μ U Ln (R2 / R1 )
r r
We can now compute the force acting on the moving rod:
 vz 1
fz = μ = μ U Ln (R2 / R1 )
r R1 R1
 ECH41MiscProblemsPart2_2024.nb 33

The total force over the wire with length L is

L 2π
Fz =   fz R1  θ  z = μ L U {Ln (R2 / R1 )}
0 0

Problem 15 : Transient rectilinear flow in a pipe

A pipe of length L is filled with water and is connected to a reservoir at z = 0 such that the pressure
there is held fixed at P0 . A valve located at z = L is closed for t < 0. At T = 0+ , the valve is opened
inducing a transient flow in the pipe.

Tasks
(i) If at z = L the pressure PL < P0 is constant, determine an expression for the transient volumetric
flow in the pipe. Hint: It is helpful to use dimensionless variables

(ii) Plot your result as a function of time in suitable dimensionless variables.

Problem 8 Solution

See the following notebook: TransientFlow in a Pipe.nb

Problem Statement in Dimensionless Variables

We consider rectilinear flow in a pipe. The velocity field is then given by
v = u (r, t) ez

When this velocity profile is substituted into the equations of motion we obtain
∂u ∂ 1 ∂ ∂u
=- +ν r
∂t ∂z r ∂r ∂r

∂ ∂
0=- , 0=-
∂r ∂θ
It follows from integrating the last two equations that the pressure field as the following
dependencies
 =  (z, t)

In the problem statement we are told that the pressure drop is held fixed. Thus =(z) only. The
mathematical statement for our problem is then
∂u  ν ∂ ∂u
=- + r 0 < r < R, t > 0
∂t z r ∂r ∂r
34 ECH41MiscProblemsPart2_2024.nb

IC : u (r, 0) = 0

BC1 : u (R, t) = 0

BC2 : u (0, t) < ∞

To solve for the velocity profile it will be convenient to introduce dimensionless variables. Let Uo be
some measure of the magnitude of the velocity (average speed). Let Δ = P0 - PL be the applied
pressure difference that drives the flow. We will use the following dimensionless variables
U = u / Uo , τ = t ν  R2 , ξ = r / R, P =  / Δ, ζ = z / L

In terms of these variables the PDE becomes

∂U Δ R2 P 1 ∂ ∂U
=- + ξ , 0 < ξ < 1, τ > 0
∂τ ν LUo ζ ξ ∂ξ ∂ξ

IC : U (ξ, 0) = 0
BC1 : U (1, τ) = 0
BC2 : U (0, τ) < ∞

Analysis: Steady State Solution

Since the PDE is inhomogeneous, it will turn out to be convenient to solve for the auxiliary variable
W(ξ,τ) defined as
W (ξ, τ) = U (ξ, τ) - Us (ξ)

where Us (ξ) is the steady state solution. The steady state solution satisfies
Δ R2 P 1   Us
0=- + ξ , 0<ξ<1
ν LUo ζ ξ ξ ξ

BC1 : Us (1) = 0
BC2 : Us (0) <∞

Integrating twice and applying the BCs gives

G
Us (ξ) = - 1 - ξ2 ,
4
where we have defined
Δ R2 P
G≡
ν LUo ζ

We can relate G to the volumetric flow rate Q:

R
Q =  2 π r us (r)  r
0
 ECH41MiscProblemsPart2_2024.nb 35

or in terms of dimensionless variables

2 π R2 Uo G 1
2
Q=-  1 - ξ  ξ  ξ
4 0

Integrating we find
π R2 Uo G
Q=-
8

But given the definition for G we can write

Δ R2  P
G=
ν L Uo  ζ

Substituting the above expression for G into the equation for the flow rate Q we get
P 8QνL
=-
ζ Δ π R4

This equation can be integrated with the BCs

P (0) = P0 / Δ,
P (1) = PL / Δ

The result is
Δ 8Qν
=
L π R4

Analysis: Transient Solution

Consider next the transient equation for W(ξ,τ)

∂W 1 ∂ ∂W
= ξ , 0 < ξ < 1, τ > 0
∂τ ξ ∂ξ ∂ξ

IC : W (ξ, 0) = -Us (ξ)

BC1 : W (1, τ) = 0
BC2 : W (0, τ) < ∞

This problem can be readily solved by separation of variables. If we let W (ξ, τ) = H (τ) ϕ (ξ), we
find on substitution that
H (τ) ∼ Exp -λ2 τ

and ϕ (ξ) satisfies the following eigenvalue problem:

36 ECH41MiscProblemsPart2_2024.nb

 ϕ
ξ + λ2 ξϕ = 0
ξ ξ

BC1 : ϕ (1) = 0
BC2 : ϕ (0) <∞

A candidate solution to the PDE is the zero order Bessel function of the first kind: J0 (ξ).

For non trivial solutions we require

J0 (λn ) = 0 n = 1, 2, 3, …

and the solution is

ϕn (ξ) = Cn J0 (λn ξ)

where J0 (λ ξ) is the zero order Bessel function of the first Kind. The general solution is
∞
W (ξ, τ) =  An Exp -λ2n τ ϕn (ξ)
n=1

The initial condition determines the coefficients An :

∞
-Us (ξ) =  An ϕn (ξ)
n=1

The coefficients An can be found using the orthogonality properties of the Bessel functions
1
 ϕn ϕm ξ  ξ = 0 if n ≠ m
0

Thus we find
1
 Us ξ ϕn  ξ
0
An = -
1
2
 ϕn ξ  ξ
0

where
G
Us (ξ) = - 1 - ξ2 
4
In order to compute the coefficients we need to determine G. Earlier we showed that the flow rate is
given by
π R2 Uo G
Q=-
8
Thus if we take Uo = Q / π R2 , it follows that
G = -8.
 ECH41MiscProblemsPart2_2024.nb 37

The dimensionless flow rate during the start-up process is

 1
Q (τ) =  2 π U (ξ, t) ξ  ξ
0

where
∞
U (ξ, τ) = Us (ξ) +  An Exp -λ2n τ ϕn (ξ)
n=1

Analysis: Mathematica Computations Part 1

In this section we illustrate how to write a Mathematica program to determine the transient flow rate
(or any other property of the solution). First we define a few functions

I n [ ] : = G = -8;
G
Us [ξ_] := - 1 - ξ2 
4
ϕ[ξ_, n_] := BesselJ[0, λn ξ]

To determine the eigenvalues we must solve J0 (λ) = 0. We can use FindRoot or we can use the
available function called BesselZeros

The roots are then readily determined. Here is a list of the first 30 roots

I n [ ] : = roots = N[BesselJZero[0, Range[30]]]

O u t [ ] =

{2.40483, 5.52008, 8.65373, 11.7915, 14.9309, 18.0711, 21.2116,

24.3525, 27.4935, 30.6346, 33.7758, 36.9171, 40.0584, 43.1998,
46.3412, 49.4826, 52.6241, 55.7655, 58.907, 62.0485, 65.19, 68.3315,
71.473, 74.6145, 77.756, 80.8976, 84.0391, 87.1806, 90.3222, 93.4637}

For later computations it will be useful to define the following function, which picks out the nth root
from the list of roots

I n [ ] : = λn_ := roots〚n〛

Here is an example what this function does

I n [ ] : = λ6
O u t [ ] =

18.0711

Next we need to define a function that determines the coefficients An :

1
N Us [ξ] ϕ[ξ, n] ξ  ξ
0
I n [ ] : = An_ := An = -
1
N ϕ[ξ, n]2 ξ  ξ
0
38 ECH41MiscProblemsPart2_2024.nb

In the above we make use of the trick An_ := An =...This forces Mathematica to store any calculated
values and thereby avoids computing the coefficients repeatedly. The first 30 coefficients are shown
below

I n [ ] : = Table[An , {n, 1, 30}]

O u t [ ] =

{-2.21604, 0.279555, -0.0909529, 0.0419818, -0.0232725, 0.0144424, -0.00967574,

0.00685136, -0.00505906, 0.00386029, -0.00302442, 0.00242154, -0.00197437,
0.00163479, -0.00137167, 0.00116423, -0.000998181, 0.000863489, -0.000752931,
0.000661219, -0.000584415, 0.000519544, -0.000464323, 0.000416982,
-0.000376132, 0.000340673, -0.000309722, 0.00028257, -0.000258637, 0.000237448}

Analysis: Mathematica Computations Part 2

The following function gives the transient velocity profile defined by a partial sum . The parameter
nMax determines the number of terms in the partial sum.
nMax
I n [ ] : = U[ξ_, τ_, nMax_] := Us [ξ] +  An Exp-λ2n τ ϕ[ξ, n]
n=1

The transient flow rate is given by (using nMax=20)

 1
I n [ ] : = Q[τ_] =  2 π U[ξ, τ, 20] ξ  ξ
0
O u t [ ] =

-6.78228 × 10-6 -3850.01 τ - 8.34897 × 10-6 -3470.03 τ -

0.0000103953 -3109.79 τ - 0.0000131088 -2769.29 τ - 0.0000167683 -2448.53 τ -
0.0000217987 -2147.51 τ - 0.0000288651 -1866.22 τ - 0.0000390413 -1604.68 τ -
0.000054124 -1362.87 τ - 0.000077246 -1140.81 τ - 0.000114143 -938.479 τ -
0.000175947 -755.891 τ - 0.000285843 -593.043 τ - 0.000496596 -449.934 τ -
0.000942681 -326.563 τ - 0.0020228 -222.932 τ - 0.00520018 -139.04 τ -
0.0179261 -74.887 τ - 0.108273 -30.4713 τ - 3.00584 -5.78319 τ + π

Here is a plot of the transient flow rate

 ECH41MiscProblemsPart2_2024.nb 39


I n [ ] : = PlotQ[τ], {τ, 0, 1.5}, PlotRange  {0, 3.5}, PlotStyle  Thick,

Frame  True, FrameLabel  Style["τ", 16], Style"Q(τ)", 16
O u t [ ] =
3.5

3.0

2.5

2.0
 (τ)
Q

1.5

1.0

0.5

0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

Finally we can display the velocity profile across the tube at select times

I n [ ] : = Quiet[Plot[{U[ξ, .02, 30], U[ξ, .1, 30], U[ξ, .2, 30], U[ξ, .5, 30], U[ξ, 1, 30]},
{ξ, -1, 1}, PlotRange  {0, 2}, Frame  True,
FrameLabel  {Style["ξ", 16], Style["U(τ,ξ)"]},
PlotLegends  {"τ=0.02", "τ=0.1", "τ=0.2", "τ=0.5", "τ=1"}]]
O u t [ ] =
2.0

1.5

τ=0.02
τ=0.1
U(τ,ξ)

1.0
τ=0.2
τ=0.5
0.5
τ=1

0.0
-1.0 -0.5 0.0 0.5 1.0

Problem 16 : Rectilinear flow between two flat plates

Calculate the stress profile, velocity profile and volumetric flow rate (per unit width) for a fluid flowing
between two flat plates separated by a distance H, as shown below. The pressure drop over the
length L is Δp . Do the calculation for a Newtonian fluid with viscosity μ. Show below is a schematic of
the flow field:
40 ECH41MiscProblemsPart2_2024.nb

Problem Solution:
The Navier-Stokes equations for an incompressible fluid are
∇ ·v = 0

∂v 
ρ + e3 · ∇ v = -∇ p + μ ∇2 v + ρ g
∂t
If we ignore entrance and exit effects, the flow in the channel may be assumed to be rectilinear (or
unidirectional). Thus the velocity field can be expressed as
v = vz (t, x) ez

For steady flow, and translational symmetry (no variation in the y-direction) the N/S equations reduce
to
∂ vz
=0
∂z

0 = -∇ p + μ ∇2 vz ex + ρ g

In the absence of gravitational effects, the components of the N/S equations reduce to{
∂ vz
Continuity : =0 (i)
∂z

∂p 2 vz
z - component of momentum : =μ (ii)
∂z  x2

∂p
x - component of momentum : =0 (iii)
∂x

Note that the continuity equation is satisfied identically as vz = vz (x). Equation (iii) advises us that
p = p(z). Hence the RHS of Eq. (ii) must be a constant. Integrating Eq. (ii) from z = 0 to z = L gives:
PL ∂p L 2 vz 2 vz
 z =  μ  z ⟹ (PL - P0 ) = L μ
p0 ∂z 0  x2  x2
 ECH41MiscProblemsPart2_2024.nb 41

or
Δp 2 vz
- = μ , where Δp = P0 - PL
L  x2
The above ODE can be integrated directly to give:
Δp x2
vz (x) = - + C1 x + C2
L 2
To determine the constants C1 and C2 we make use of the no-slip boundary conditions at the walls of
the channel:
H
BC1 : vz = 0 at x = ,
2
H
BC2 : vz = 0 at x = -
2
 vz
BC3 : = 0 , at x = 0
x
Note BC3 is a statement that the flow is symmetric about x = 0. Hence the symmetry condition can
be used in place of BC2 (Note: the application of symmetry conditions often simplifies the algebra.)
The application of the above the BCs gives the following result for the velocity field vz (x):
Δp 2
H2
vz (x) = - x -
2μL 4

Thus the velocity profile is a quadratic function of x, symmetric about x=0.

The volumetric flow rate is given by

H H
W
2 2
Q =   H vz (x)  x  y = 2 W  vz (x)  x
0 - 0
2

Integrating we obtain the following expression for the volumetric flow rate:
 Q Δp
Q≡ = H3
W 12 μ L

In a Newtonian fluid the viscous stress tensor is given by

1
τ=2μD , where D = ∇ v + ∇ v† 
2
Recall that
∂ vj
∇v = ei ej
∂ xi

∂ v1 ∂ v1 ∂ v2 ∂ v2
= e1 e1 + e2 e1 + e1 e2 + e2 e2
∂ x1 ∂ x2 ∂ x1 ∂ x2
42 ECH41MiscProblemsPart2_2024.nb

We can write the components of ∇v in matrix form, as illustrated below:

∂v1 ∂v2 ∂v1 ∂v1
∂x1 ∂x1 ∂x1 ∂x2
[∇ v] = ∂v1 ∂v2
, [∇ v] = †
∂v2 ∂v2
∂x2 ∂x2 ∂x1 ∂x2

Thus
[ τ] = μ [∇ v] + [∇ v]† 

2 ∂v1
 ∂v2 + ∂v1

∂x1 ∂x1 ∂x2
=μ
 ∂v2 + ∂v1
 2 ∂v2
∂x1 ∂x2 ∂x2

Now let subscript 1 denote the component in the ez direction and subscript 2 denote the component
in the ex direction. And since vz = vz (x) we have
 vz
0
x
[τ]=μ  vz
0
x

In tensor notation the viscous stress tensor is

 vz  vz
τ =μ ex ez + ez ex
x x
The normal components of the stress tensor are
ex · τ · e x = 0
ez · τ · e z = 0

while the shear stress is

 vz Δp
ex · τ · ez = τxz = τzx = μ = - x
x μL

But recall that

Δp H2
vz (x) = - x2 -
2μL 4

Thus
 vz Δp Δp
μ = - 2x= - x
x 2μL μL

Hence
Δp
τzx = - x
μL

Thus the stress profile is a linear function of x.

 ECH41MiscProblemsPart2_2024.nb 43

Problem 17 : Turbulent flow in a cylindrical tube

At a Reynolds number of 110, 000, the turbulent velocity profile in a cylindrical tube with radius R is
given by
Vz = Vz,max (1 - r / R)1/7

Tasks
Determine the average velocity 〈Vz 〉 for this profile.

Problem Solution:
The definition of the average velocity is
Q
〈vz 〉 ≡
A
where Q is the volumetric flow rate in the pipe, and A is the cross-sectional area of the pipe. Mathemati-
cally we can represent the volumetric flow rate as
R R
Q =  vz 2 π r  r = 2 π Vz,max  (1 - r / R)1/7 r  r
0 0

Consider the integral:

R
1/7
 (1 - r / R) r r
0

We can make the substitution

r
1 -  ≡ p ⟹ r = -R p + R,  r = -R 
R
Thus we can write the integral
R 0
1/7
 (1 - r / R) r  r = -R  p1/7 (R - R p)  p
0 1

Integrating gives
R
1/7
1 0 1 0
 (1 - r / R) r  r = -R  p1+1/7  + R2  p1+8/7 
0 1/7 + 1 1 1 + 8/7 1

7 7 49 R2
= R2 - R2 =
8 15 120
We can check our answer using Mathematica :
44 ECH41MiscProblemsPart2_2024.nb

R
1/7
I n [ ] : =  (1 - r / R) r r
0
O u t [ ] =
49 R2
120

The average velocity is then

49 R2 49
〈vz 〉 = Vz, max  π R2 = vz, max
120 120 π

Problem 18 : Computation of fluid viscosity

A pressure drop of 200 Pa is recorded over a 1 m length of 10 mm diameter (ID) tubing. The flow rate
Q is 60 mm3  s, the density ρ is 1000 kg  m3 , and the fluid is Newtonian.

Tasks:
Use the above data to compute the viscosity of the fluid (Note: be sure to check that the flow is
laminar).

Problem Solution :
Recall that the Reynolds Number for flow in a pipe can be expressed as:
ρ 〈v〉 R
e =
μ

Let us assume the flow is laminar. Then the flow rate is determined by
4
R4 Δp 5  10-3  200
Q=π =π = 60  10-9
8μ L 8μ 1

Solving the above expression for μ gives:

4
5  10-3  200 25 π
μ=π = = 0.818 Pa s
8 60  10-9  1 96

We must now check to see that our assumption that the flow is laminar flow was correct.
ρ 〈v〉 R 1000 60  10-9
e = = 5  10-3  = 0.004669
0.818 2
μ π  5  10-3 

Based on the above calculation, we can conclude that the flow is laminar!

Problem 19 : Computation of two dimensional flows and their

properties
 ECH41MiscProblemsPart2_2024.nb 45

Consider the following linear, two-dimensional flows given below:

(i) vx = Gy, vy = 0, vz = 0 (a "Couette" flow)
(ii) vx = (G / 2) y, vy = (G / 2) x, vz = 0 (a purely straining flow)
(iii) vx = (G / 2) y, vy = -(G / 2) x, vz = 0 (a purely rotational flow)

Note that addition of flows (ii) and (iii) yields flow (i).

TasksL
For the flows (i)-(iii), calculate
a) the velocity gradient, rate-of-strain and vorticity tensors;
b) the viscous stress tensor (assuming the fluid is Newtonian with viscosity μ);
c) the streamlines. Plot the streamlines for these three steady flows.

Problem Solution:
(a) The velocity gradient is given by
∂ vj
∇v = eiej
∂ xi

The rate of strain tensor is given by

1
D= ∇ v + ∇ v † 
2
And the vorticity tensor is
1
Ω= ∇ v - ∇ v† 
2
The components of the velocity gradients for the 3 flows are

G
0 0 0 0 0
2
(i) [∇ v ] = G 0 0 , (ii) [∇ v] = G
0 0
2
0 0 0
0 0 0

0 -G 0
2
(iii) [∇ v ] = G
0 0
2
0 0 0

The components of the rate of strain tensors for the 3 flows are:
46 ECH41MiscProblemsPart2_2024.nb

1 0 G 0 1 0G0
(i) [D] = G 0 0 , (ii) [D] = G00
2 0 0 0 2 000

0 0 0
(iii) [D] = 0 0 0
0 0 0

The components of the vorticity tensor for the 3 flows are

1 0 -G 0 000
(i) [Ω ] = G 0 0 , (ii) [Ω ] = 000
2 0 0 0 000

1 0 -G 0
(iii) [Ω ] = G 0 0
2 0 0 0

(b) The viscous stress tensor is given by

τ=2μD

Thus the components of the viscous stress tensor for the three flows are
0 G 0 0G0
(i) [τ ] = μ G 0 0 , (ii) [τ ] = μ G 0 0 , (iii) [τ ] = 0
0 0 0 000

(c) For the steady flows the stream lines are given by
 xs
=v
α
where α is a parameter for the streamline (e.g., arc length), and xs is the position vector for the
STREAMLINE, at any point in time. Recall that the velocity is everywhere tangent to the streamline.

For flow 1 we have:

 x1  x2
= G x2 , =0
α α
Solving the above set of equations gives:
x2 = x2 (0), x1 = Gx2 (0) t + x1 (0)

where xi (0) is the position along the streamline when α = 0. Thus if we are given values for G and
xi (0) we can plot the streamline. Here is the parametric plot of the streamlines in the
x1 - x2 plane. We pick the following initial conditions
 ECH41MiscProblemsPart2_2024.nb 47

I n [ ] : = initCond = {{X1  0, X2  .25}, {X1  0, X2  .5}, {X1  0, X2  0.75}, {X1  0, X2  1},

{X1  0, X2  1.25}, {X1  0, X2  1.5}, {X1  0, X2  1.75}, {X1  0, X2  2}};
G = 0.5;

I n [ ] : = ParametricPlot[Evaluate[{G X2 t + X1 , X2 } /. initCond], {t, 0, 10}, PlotStyle  Blue,

Frame  True, FrameLabel  {Style["x", 16], Style["y", 16]}, AspectRatio  0.6]
O u t [ ] =

2.0

1.5
y

1.0

0.5

0 2 4 6 8 10

The plot of streamlines shows the familiar characteristic of plane Couette flow, i.e. a profile that is
linear in y.

For flow (ii) we have

 x1 G  x2 G
= x2 , = x1
α 2 α 2
To solve this system of equations we can write the equations as:
G
x
 0 x1
 1= 2
 
t x 2
G
0 x2
2

This linear system of ODE equations can be readily solved by seeking solutions of the form
xi ∼ ξi Exp (λ α). Substituting this expression into the system of equations results in the following
eigenvalue problem
G
-λ ξ1
2
G
 =0
-λ ξ2
2

For a non-trivial solution we require

G
-λ
2 2
G2
Det G
=0 ⟹ λ - =0
-λ 4
2
G
Thus the eigenvalues are λ1 = 2
, λ2 = - G2 . The corresponding eigenvectors are:
48 ECH41MiscProblemsPart2_2024.nb

1 -1
ξ (1) =  , ξ (2) =  
1 1

Thus the general solution is:

G G
x = C1 ξ(1) Exp α + C2 ξ (2) Exp- α
2 2
At α = 0, we have x1 (0) = 1 , and x2 (0) = 2 . We can use these initial conditions to evaluate the
constants C1 and C2 . The result is:
(1 + 1 ) (2 - 1 )
C1 = and C2 =
2 2
Hence the equations for the streamline are
(1 + 1 ) G (2 - 1 ) G
x1 = Exp α - Exp- α
2 2 2 2

(1 + 1 ) G (2 - 1 ) G
x2 = Exp α + Exp- α
2 2 2 2
We next illustrate how to plot the stream lines using Mathematica's ParametricPlot. First, we con-
struct a set of rules for the initial values {1 , 2 }

In[1]:= initialCond =
{{1  0, 2  0.25}, {1  0, 2  0.5}, {1  0, 2  0.75}, {1  0, 2  1},
{1  0, 2  1.25}, {1  0, 2  1.5}, {1  0, 2  1.75}, {1  0, 2  2}};

The equations for the streamlines are

(1 + 2 ) G (2 - 1 ) G
In[2]:= Eqn1 = Exp α - Exp- α;
2 2 2 2
(1 + 2 ) G (2 - 1 ) G
Eqn2 = Exp α + Exp- α;
2 2 2 2

We use ParametricPlot to display the streamlines for 0 ≤ α ≤ 1

 ECH41MiscProblemsPart2_2024.nb 49

In[4]:= ParametricPlot[Evaluate[{Eqn1, Eqn2} /. G  2 /. initialCond], {α, 0, 1},

PlotStyle  Thick, Frame  True, FrameLabel  {Style["x", 16], Style["y", 16]},
AspectRatio  0.5, PlotLegends  {"χ2 =0", "χ2 =0.25",
"χ2 =0.25", "χ2 =0.5", "χ2 =0.75", "χ2 =1.0", "χ2 =1.25", "χ2 =1.5",
"χ2 =1.75", "χ2 =2.0"}]

3.0
χ2 =0
2.5
χ2 =0.25
2.0
χ2 =0.25
y

1.5
χ2 =0.5
Out[4]=
1.0
χ2 =0.75
0.5
χ2 =1.0
0.0
χ2 =1.25
0.0 0.5 1.0 1.5 2.0
χ2 =1.5
x

For flow (iii) we have

 x1 G  x2 G
= x2 , =- x1
α 2 α 2
To solve this system of equations we can write the equations as
G
 x 0 x1
 1= 2
 
t x 2 - G
0 x2
2

We can use the same procedure as we did for flow (ii), except now the eigenvalues are imaginary:
λ1 = -  G2 , λ2 =  G2 . The corresponding eigenvectors are
 -
ξ (1) =  , ξ(2) =  
1 1

Instead of doing the calculations by hand we will use Mathematica's DSolve function to get a solution
to the ODEs directly. Here is the result:

In[5]:= ODEsol = First

G G
DSolvex '[α]  y[α], y '[α]  - x[α], x[0]  1 , y[0]  2 , {x[α], y[α]}, α
2 2
Gα Gα Gα Gα
Out[5]= x[α]  Cos  1 + Sin  2 , y[α]  -Sin  1 + Cos  2 
2 2 2 2

We use the same values for the initial conditions to plot the streamlines for G=2. The plot is shown
below
50 ECH41MiscProblemsPart2_2024.nb

In[6]:= ParametricPlot[Evaluate[{x[α], y[α]} /. ODEsol /. G  2 /. initialCond],

{α, 0, 10}, AspectRatio  1, PlotStyle  Thick,
FrameLabel  {Style["x", 16], Style["y", 16]}, Frame  True, PlotLegends  {"χ2 =0",
"χ2 =0.25", "χ2 =0.25", "χ2 =0.5", "χ2 =0.75", "χ2 =1.0", "χ2 =1.25", "χ2 =1.5",
"χ2 =1.75", "χ2 =2.0"}]

1
χ2 =0
χ2 =0.25
χ2 =0.25
0
y

χ2 =0.5
Out[6]=
χ2 =0.75
χ2 =1.0
-1
χ2 =1.25
χ2 =1.5

-2

-2 -1 0 1 2

Thus we have seen that plane Couette flow is a combination of a a purely straining flow and a rota-
tional flow. The linear combination of these two flows results in a flow with streamlines parallel to the
x-axis .

Problem 20 : Non-Newtonian tube flow:

Unlike Newtonian fluids, the viscosity η of non-Newtonian fluids is a function of the rate of deforma-
tion. In tube flow, this relation is expressed as:
 vz
η=η
r
where the magnitude of the velocity gradient  vz /  r is known as the "shear rate." Profes-
sor/Inventor R.L. Powell is convinced that he can measure this viscosity function simply by measuring
velocity profiles and pressure drops in tube flow. If that is the case then such a measurement could be
made "on-line", with minimal modification of a process stream, an application that has an obvious
appeal. In this problem we derive the underlying theory for such a device. The following tasks are
listed below:
 ECH41MiscProblemsPart2_2024.nb 51

Tasks:
a) Consider fully developed, Poiseuille flow in a cylindrical tube with radius R. Using the macroscopic
momentum balance with a fixed, cylindrical control volume V having a length L, show that the wall
shear stress is related to the pressure drop by
R ΔP
τr z = - at r = R
2 L
b) Use the stress equations of fluid motion to calculate the stress profile τr z (r). You may assume
that vz = vz (r), and τr z = τr z (r). The other components of the velocity and stress are taken to
be zero.

c) Professor Powell's device is used to measure the velocity profile of tomato paste flowing through a
tube with radius R = 2 cm. From the measurements one is able to show that the velocity field can be
expressed by the following equation:
r 5
vz (r) = (50 cm / s)1 -   
R
Use the following constitutive relation
 vz  vz
τr z = η
r r
to calculate the function η over the range of shear rates that are present in the tube in the region
0.1 < r / R < 1.0. The pressure drop is such that Δp  2RL  = 50 Pa

The nomenclature
 vz
η
r
states that the viscosity η is in general a function of the shear rate vz
r

Problem Solution:
(a) Consider a fixed control volume in a region of the pipe where the flow is steady and rectilinear.


n
r
z
 z
-n  z
n

z=0 z=L

In our analysis we will use the macroscopic momentum balance, which is given by:
52 ECH41MiscProblemsPart2_2024.nb

∂
 (ρ v)  V +  ∇ · (ρ v v)  V =  ρ g  V +  ∇ · T  V
 ∂t   

The above equation can be simplified by applying the divergence theorem to the second and fourth
integrals. The result is:
∂
 (ρ v)  V +  ρ (v v · n)  A =  ρ g  V +  T · n  A
 ∂t   

We apply the above macroscopic momentum balance to the control volume (shown as a dashed line
in the above figure):

 ρ v (v · n )  A =  ρ g  V +  n · T  A
Ae V As

The quantity Ae is the area of the control volume through which fluid enters/exits, and As is the total
surface area of the control volume, which includes the surface area of the tube wall and Ae . The above
integral is a vector equation, and for our purposes we take the ez component, which is aligned in the
direction of flow:

 ρ vz (v · n)  A =  n · T · ez  A
Ae As

Note that we have assumed that g · e z ≡ 0 .

Now vz (r) is the same at z = 0 and z = L (a consequence of the rectilinear flow assumption). But
the outward directed normal n changes sign at the entrance and exit (see Figure above). Thus when
we integrate over Ae at z = 0 and Ae at z = L, the terms cancel and our momentum balance reduces
to

0 =  n · T · ez  A
As

On the surface wall of the tube n = er so that

n · T · ez = e r · T · ez ≡ τrz

At the entrance and exits we have n = ± e z so that

n · T · ez = ± e z · T · ez ≡ ∓ P ± τzz

For rectilinear flow τzz = 0 . (Note: this is not a general result for any non-Newtonian fluid, but will
suffice for a Generalized Newtonian fluid that exhibits power-law behavior)

Our z- component of momentum becomes

 τrz  A +  P A -  P A = 0
As (R) Ae (0) Ae (L)

For a rectilinear flow assumption, τrz (r = R)is a constant. Likewise P is constant at a given cross-
section of the pipe. Integrating the above equation gives
2 π RLτrz (R) = {P (L) - P (0)} π R2
 ECH41MiscProblemsPart2_2024.nb 53

We can rearrange the above expression to obtain:

R ΔP
τrz (R) = - , where ΔP = P (0) - P (L)
2 L
(b) The z-component of momentum (assuming rectilinear flow) is given by:
 1 
0=– + (r τrz )
z r r
Integrating the above equation with respect to r gives
r  C
τrz = +
2 z r
For τrz to be finite at r = 0, we require C = 0. Thus we have
r  r ΔP
τrz = =-
2 z 2 L
Note that this is the same result as found in Part (a).

(c) Since we are given the velocity profile

r 5
vz (r) = 50 1 -   
R
it is a simple matter to determine the rate of shear:
 vz  250 r4 250 r 4
≡γ=- =-  
r R5 R R
We are told that the shear stress is related to the shear rate by the following constitutive equation:
 vz  vz
τr z = η
r r
This equation can be rearranged to give
τrz r Δp 1   vz r ΔP
η=  =-  , where γ ≡ , and τrz = -
γ 2 L γ r 2 L

Furthermore we are told that the pressure drop is such that

R
Δp = 50 Pa
2L
Substituting into our previous result gives
r 1
η = -  (50) 
R γ
 
Thus if we are given γ, then a plot of η versus γ defines the relation for η. Note for the model used
here the viscosity would be infinite at the center of the pipe ξ = 0, where the shear rate is zero. Since

we know the velocity gradient, we can compute γ for different locations in the tube
54 ECH41MiscProblemsPart2_2024.nb

 250 r 4 250 4
γ=-   =- ξ
R R R
In the problem statement we are not given R. Thus consider a pipe with radius R = 0.1 m. Then the
viscosity η for 0.1 < ξ < 1 is
250
In[7]:= Plot- ξ4 /. R  0.1, {ξ, 0.1, 1},
R

FrameLabel  {Style["ξ", 16], Style["γ", 16]}, PlotStyle  Thick, Frame  True

-500

-1000
γ


Out[7]= -1500

-2000

-2500
0.2 0.4 0.6 0.8 1.0

And for our viscosity we have

1 R
η=
5 ξ3

Note that the viscosity of the fluid becomes infinite at the pipe center, a consequence of the rheologi-
cal model selected for the viscosity! A more realistic model would avoid this problem.
R
In[8]:= Plot /. R  0.1, {ξ, 0.1, 1}, FrameLabel  {Style["ξ", 16], Style["η", 16]},
5 ξ3
PlotStyle  Thick, Frame  True, PlotRange  All

20

15

10
η

Out[8]=

0
0.2 0.4 0.6 0.8 1.0

ξ
 ECH41MiscProblemsPart2_2024.nb 55

Problem 21 : Flow in a crude oil pipeline:

Crude oil in a refinery must be pumped uphill through a pipe with radius R and length L. The pipe is at
an angle θ relative to the ground (i.e., θ = 0 corresponds to a horizontal pipe).

Tasks
Calculate the pressure drop Δp (NOT the dynamic pressure!) needed to move the fluid through the
pipe at a volumetric flow rate Q. The fluid is Newtonian, with viscosity μ and density ρ.

Problem Solution:
We will assume that the pipe is sufficiently long so that over its length L the flow is rectilinear, that is
the fluid velocity in the pipe can be expressed as v = vz (r) ez . We will also assume that the flow is
steady and has azimuthal flow symmetry. If we choose a cylindrical coordinate system aligned along
the axis of the pipe then the equations of motion reduce to
∂p 1   vz
0=- + ρ gz + μ r
∂z r r r

∂p
0=- + ρ gr
∂r

1 ∂p
0=- + ρ gϕ
r ∂θ
where gr , gϕ , and gz are the components of the gravitational body force in the er , eϕ , and ez directions.
For our problem:
g = -g Cos[90 - θ] ez + g Sin[90 - θ] eξ

= -g Sin[θ] ez + g Cos[θ] eξ

The unit vector eξ is related to eϕ and er by the following relation.

eξ = -Cos[ϕ] er + Sin[ϕ] eϕ

This follows by noting that in the plane of the pipe cross-section, the base vectors of a local Cartesian
coordinate system (i , j) are related to the er and eϕ by
i = Cos[ϕ] er - Sin[ϕ] eϕ , j = -Sin[ϕ] er + Cos[ϕ] eϕ

and we can identify that

eξ = -i.

Consider first the r-momentum equation

56 ECH41MiscProblemsPart2_2024.nb

∂p
0=- - ρ g Cos[θ]  Cos[ϕ]
∂r
Integrating this equation we get
p (r, z, ϕ) = -ρ g Cos[θ]  Cos[ϕ] r + f (z, ϕ)

where f (z, ϕ) is an arbitrary function of integration. Consider next the ϕ -momentum equation.
1 ∂p
0=- + ρ g Cos[θ]  Sin[ϕ]
r ∂θ
Integrating this equation we get
p (r, z, ϕ) = -ρ g Cos[θ]  Cos[ϕ] r + g (z, r)

It then follows that

f (z, ϕ) = g (z, r) = F (z)

and
p (r, z, ϕ) = -ρ g Cos[θ]  Sin[ϕ] r + F (z)

If we differentiate the expression for p (r, z, ϕ) with respect to z we get

∂p F
=
∂z z
Substituting for in the z-component of momentum gives
∂p
∂z
F 1   vz
0=- ρ - g Sin[θ] + μ r
z r r r
which can be regrouped as
1   vz
G (z) = μ r
r r r
where we have used
F
G (z) ≡ + ρg Sin[θ]
z
Since the LHS is a function of z and the RHS is a function of r, we can conclude that
1   vz G (z)
r = = constant
r r r μ

Integrating we get
r2
vz (r) = G + C1 Ln[r] + C2
4μ

The constants of integration can be found by applying the no-slip condition at r = R , and that the
velocity must be finite at r = 0. Thus we have:
 ECH41MiscProblemsPart2_2024.nb 57

BC1 : vz (R) = 0

BC2 : vz (0) <∞

Applying these BCs the constants C1 and C2 take on the following values:
R2 G
C1 = 0, C2 = -
4μ

Hence the velocity field is given by:

G
vz (r) = r2 - R2 
4μ

The flow rate Q is found by integrating the velocity profile over the cross-sectional area of the pipe.
R G R
2 2
Q = 2 π  vz (r) r  r = 2 π  r - R  r  r
0 4μ 0
4
πGr
Q=-
8μ

Our final task is to determine G (z),which recall is related to F (z). It follows from the definition
of G (z) that
F 8μQ
=- + ρ g Sin[θ]
z π R4
Integrating gives
8μQ
F (z) = - + ρ g Sin[θ] z + C1
π R4
At z = 0, let the average pressure be P0 . Thus
1 2π R
P0 = 〈p (r, z, ϕ)〉z=0 =    p (r, z, ϕ) r  r  ϕ
π R2 0 o z=0

Substituting p (r, z, ϕ) = -ρ g Cos[θ] Sin[ϕ] r + F (z) into the above expression and
integrating gives
P0 = F (0)

A similar calculation at z = L gives

F (L) = PL

Applying these results we obtain the following expression for F(z):

8μQ
F (z) = - + ρ g Sin[θ] z + C1
π R4
And since P0 = F (0), we find
C1 = P0
58 ECH41MiscProblemsPart2_2024.nb

Thus the pressure drop required to pump the oil at a flow rate Q over a length L is:
8μQ
P0 - PL ≡ ΔP = - ρ g Sin[θ] L
π R4

Problem 22 : Rectilinear Flow in a pipe with an elliptical cross-

section

Tasks:
Determine the pressure drop-flow rate relation for rectilinear flow in a pipe with an elliptical cross-
section (major axis is a, minor axis is b ). Hint: look for solutions of the form
x22 x23
v1 (x2 , x3 ) = α + -1
a2 b2

where the velocity field is given by:

v = v1 (x2 , x3 ) e1

Show that your answer reduces to the well-known expression for a pipe with a circular cross-section
when a = b.

Problem Solution:
for rectilinear flow , the v1 component must satisfy

μ ∇2 v1 = ≡G
 x1

where
∂2 v1 ∂2 v1
∇2 v1 ≡ +
∂ x22 ∂ x23

We will look for solutions of the form

x22 x23
v1 = α + -1
a2 b2

This form for the velocity satisfies the no-slip condition on the wall of the tube which is described by
x22 x23
F (x2 , x3 ) = + -1=0
a2 b2
where a and b are the major and minor axes of the ellipse. To determine the constant α, we substi-
tute the above expression into the momentum equation . The result is
2 2
μα + =G
a2 b2
 ECH41MiscProblemsPart2_2024.nb 59

Rearranging gives
G a2 b2
α=
2μ a2 + b2

so that
G a2 b2 x22 x23
v1 = + -1
2μ a2 + b2 a2 b2

The flow rate Q is determined by integrating the velocity profile over the cross-section of the tube.
Because of symmetry we can perform the integration in the first quadrant: Thus
b ξ
Q =   v1  x2  x3 , where ξ = a 1 - x23  b2
0 0

This integration can be readily done in Mathematica (let v = v1 , x = x2 , y = x3 ). First, we define

the variable v to be the velocity profile

G a2 b2 x2 y2
I n [ ] : = v= + -1 ;
2 μ a2 + b2 a2 b2

The first integration gives

ξ y2
I n [ ] : = u =  v  x /. ξ  a 1- // Simplify
0 b2
O u t [ ] =
y2
a3 G -b2 + y2  1-
b2

3 a2 + b2  μ

The second integration gives

b
I n [ ] : = Q =  u y
0
O u t [ ] =
a3 b3 G π
-
16 a2 + b2  μ

In our case the Sign[b] is 1. (Mathematica does not assume that b is real). Thus the flow rate is

I n [ ] : = 4 Q /. Sign[b]  1
O u t [ ] =
a3 b3 G π
-
4 a2 + b2  μ

Recall that G ≡ 
x1
. Integrating over the length L, and defining Δ =  (0) -  (L), we can solve for
Q
60 ECH41MiscProblemsPart2_2024.nb

a3 b4 π Δ
Q=
4 b2 a2 + b2  μ L

Now for a pipe with circular cross section we let a  b:

b4 π Δ
Q=
8Lμ

This is the standard formula for Poiseuille flow in a pipe with radius b

Problem 23 : Computing the viscosity in pipe flow

A pressure drop of 200 Pa is recorded over a 1 m length of 10 mm diameter (ID) tubing. The flow rate
Q is 60 mm3  s, the density ρ is 1000 Kg  m3 , and the fluid is Newtonian.

Tasks
Compute the viscosity (be sure to check that the flow is laminar).

Solution: Problem 8j:

The Reynolds Number is
ρ 〈v〉 R
e =
μ

Let us assume the flow is laminar. Then the flow rate is determined by
4
R4 Δp 5  10-3  200
Q=π =π = 60  10-9
8μ L 8μ 1

Solving for μ gives

4
5  10-3  200 25 π
μ=π = = 0.818 Pa s
8 60  10  -9 1 96

Let us now check to see that our assumption of laminar flow was correct.
ρ 〈v〉 R 1000 60  10-9
= 5  10-3  = 0.004669
0.818 2
μ π  5  10-3 

Thus we can conclude that the flow is laminar

Problem 24 : Analysis of 2-D Flow Field

A 2-dimensional flow field in the x-y plane, bounded by a moving solid plane at y=0 is studied experi-
mentally. Sufficiently far away from the moving solid plane, the velocity of a fluid particle is measured
 ECH41MiscProblemsPart2_2024.nb 61

moving along a horizontal streamline that coincides with the x- axis. The x-component of the two
dimensional incompressible flow field was experimentally found to be describes by the equation
u = x2

Tasks
Along this streamline determine an expression for
(a) the rate of change of the v component of velocity with respect to y
(b) the acceleration of the fluid particle along the streamline
(c) the pressure gradient in the x-direction
(d) the components of viscous stress tensor τ at a point P (x=3, y=2) on the streamline, assuming that
u = x 2 is a good approximation of the flow even at y=0.

Problem 9 Solution
(i) The continuity equation for this flow is
∂u ∂v
+ =0
∂x ∂y

Thus along the streamline (say at y = ys )where the measurement is made

∂v
= -2 x, at y = ys
∂y

Next we would like to know how v- component of velocity varies with y. This is best accomplished by
integrating the continuity equation from y=0 to y = ys .
ys ∂u ys ∂v
 y +  y = 0
0 ∂x 0 ∂y

We can write this integral as

 ys
 u  y + v (x, ys ) - v (x, 0) = 0
x 0

Noting that v(x,0)=0 (recall at y=0 the flow is bounding by a solid surface moving parallel to the x-axis.
Thus we have
 ys
 u  y = -v (x, ys )
x 0

At this point we do not know how u(x,y) varies with y. All we know is that at y = ys , u( x, ys ) = x 2 . We
will suppose therefore that u is only a function of x. Clearly, this would be reasonable if the speed of
the “plate” at y=0 is a function of x. If we make this assumption then
u (x, y) = x2 for all y > 0

At the point P on the streamline we have

62 ECH41MiscProblemsPart2_2024.nb

u
ys = -v (x, ys )
x
which gives for any y
v (x, y) = -2 x y

Thus
∂v
= -2 x, v (x, y) = -2 x y
∂y

Note that the value of v(x,y) is required for later calculations

(ii) The acceleration a = a x e x + a y e y of the fluid particle along the streamline is

∂u ∂u ∂v ∂v
ax = ρ u +v , ay = ρ u +v
∂x ∂y ∂x ∂y

for the given velocity field we have

ax = ρ x2 (2 x) + (-2 xy) (0) = 2 ρ x3

ay = ρ x3 (-2 y) + (-2 x y) (-2 x) = ρ -2 x3 y - 4 x2 y

= -2 ρ x2 y (x + 1)

Hence the acceleration along the streamline

a = 2 ρ x3 ex - 2 ρ x2 y (x + 1) ey

(iii) From the x-momentum equation we have

∂p ∂2 u ∂2 u
= -ax + μ +
∂x ∂ x2 ∂ y2

= -2 ρ x3 + μ (2) + μ (0)

= -2 ρ x3 - μ

(iv) The components of the viscous stress tensor are

2μ ∂u
μ  ∂u + ∂v
 4 μ x -2 μ y
∂x ∂y ∂x
[τ] = = 
μ  ∂u + ∂v
 2μ ∂v -2 μ y -4 μx
∂y ∂x ∂y

Evaluating at x=3 y=2 gives

12 μ -4 μ
[τ] =  
-4 μ -12 μ
 ECH41MiscProblemsPart2_2024.nb 63

Problem 25 : Computing Flux across a Specified Plane

In this problem you are required to determine the flux across a specified plane. Let the velocity field
be given by
v = 12 y i + 2 x j - 3 k

Tasks
Then determine the total flux J due to v across the plane
x + 2 y + 8 z, for x ≥ 0, z ≥ 0

Problem 9b Solution:
The unit normal to any surface f(x,y,z)=0 is given by
n  ∇f / ∇f

For our problem we are given the equation for the surface f(x,y,z)
f (x, y, z) = x + 2 y + 8 z - 12

Given the equation for the surface we can compute the unit normal to that surface. We proceed as
follows:

First we compute the gradient of the surface

∇ f = i + 2 j + 8 k, ∇ = ∇f · ∇f = 69

so that the unit normal is

1
n= (i + 2 j + 8 k)
69

The flux J across the surface with unit normal n is given by

x y
J =  v · n S =  v · n
S Γ n·k

where we have used the result from the projected area theorem
8
n·k  s =  x  y, where n·k =
69

and Γ is the projected area of the surface onto the x-y plane. Thus the limits of integration are
0 ≤ y ≤ 6 - x / 2, and 0 ≤ x ≤ 12

From our definitions we have

64 ECH41MiscProblemsPart2_2024.nb

12 y + 4 x - 24
v·n =
69

Evaluating the integral for the flux gives

x y 12 6-x/2 12 y + 4 x - 24 x y
J =  v·n = 
Γ n·k 0 0 69 8 69

12 6-x/2 3 x
=  y+ - 3  x  y = 72
0 0 2 2

Problem 26 : Magnetic Strength of a Continuum

A magnetic field strength of a continuum is given by

Exp[-A t]
λ=
r
where
r2 = x21 + x22 + x23 , and A is a constant

Tasks
If the velocity components of the continuum are
v1 (x1 , x2 , x3 , t) = B x1 x3 t,
v2 (x1 , x2 , x3 , t) = B x22 t2 ,
v3 (x1 , x2 , x3 , t) = B x3 x2

determine the rate of change of the magnetic intensity for the particle at position P(2,-1,2) when t=1

Problem Solution
Let λ be the intensity of a fluid particle X. Then the intensity of the fluid particle is given by
λ ∂λ
= + v · ∇λ
t X ∂t
It follows from the given magnetic strength that
∂λ A
=- Exp[-At]
∂τ r
To compute ∇λ , we will note the following
∂ 1 ∂ 1 ∂r 2 ∂r
= =-
∂ x1 2 ∂r 2 ∂ x1 3 ∂ x1
r r r

but
 ECH41MiscProblemsPart2_2024.nb 65

r2  x21 + x22 + x23 ⟹ r = ± x21 + x22 + x23

Thus
∂r 1 2 x1 x1 x1
= = =
∂ x1 2 x21 + x22 + x23 x21 + x22 + x23 r

Now
∂ 1 1 ∂r 1 x1 x1
- =-  =-
∂ x1 r r2 ∂ x1 r2 r r3

Thus we can conclude for any component xi we have

∂ 1 xi
=-
∂ xi r r3

Consider next the quantity ∇λ, which in component form is

∂λ ∂λ ∂λ
∇ λ = e1 + e2 + e3
∂ x1 ∂ x2 ∂ x3

Let us first evaluate ∂ λ / ∂ x1 :

∂ 1 ∂ 1 x1
Exp[-A t] = Exp[-A t] = -Exp[-A t]
∂ x1 r ∂ x1 r r3

Thus
∂λ ∂λ ∂λ
v · ∇ λ = v1 + v2 + v3
∂ x1 ∂ x2 ∂ x3

=
x1 x2 x3
B x1 x3 t -Exp[-A t] + B x22 t -Exp[-A t] + B x2 x3 t -Exp[-A t]
3 3
r r r3

Exp[-At]
=- B x21 x3 t + B x32 t2 + B x23 x2 
3
r
Hence
λ A B x21 x3 t + B x32 t2 + Bx23 x2 
=- Exp[-A t] - Exp[-A t]
t X r r3
Now evaluating the RHS of the above expression at P(2,-1,2) and t=1, we get
λ (3 A + B)
= -Exp[-A]
t X 9

Problem 27 : Kinematics of Unsteady Extensional Flow

A researcher from Spooks Inc. claims to have developed a new flow apparatus that is able to produce
66 ECH41MiscProblemsPart2_2024.nb

a general unsteady extensional flow described by

v = a F (t) {x1 e1 +  (x2 e2 + x3 e3 )}

You are instructed by your supervisor to verify the claim by carrying out a set of flow visualization
experiments using water as a test fluid. You are to take
1
f (t) = , a=1
1+t

Tasks
(i) If the flow field must satisfy ∇ · v = 0, what is the value of ?

(ii) Determine the trajectory of a dyed fluid particle that at time t = 0 was at x1 = x2 = x3 = 2. Sketch
the trajectory.

Problem Solution
(i) Let us write the equation of continuity in component form
∂ v1 ∂ v2 ∂ v3
∇· v= + + =0
∂ x1 ∂ x2 ∂ x3

From the problem statement we have

v1 = af (t) x1 , v2 = af (t)  x2 , v3 = af (t)  x3

Thus
1
∇ · v = af (t) {1 + 2 S} = 0   = -
2
(ii) The ODEs that describe the particle path line are
 x1 x1  x2 x2  x3 x3
= v1 = , = v2 = - , = v3 = -
t (1 + t) t 2 (1 + t) t 2 (1 + t)

Integrating we get
x1 = X1 (1 + t), x2 = X2  1 + t , x 3 = X3  1+t

Now if at time t = 0, x1 = x2 = x3 = 2, then

x1 = 2 (1 + t), x2 = 2  1 + t , x3 = 2  1+t

We can use Mathematica's ParametricPlot3D to display the particle path

 ECH41MiscProblemsPart2_2024.nb 67

I n [ ] : = ParametricPlot3D2 (1 + t), 2  1 + t , 2 1 + t , {t, 0, 10},

AspectRatio  1, AxesLabel  {Style["x", 16], Style["y", 16], Style["z", 16]}

O u t [ ] =

Problem 28 : Kinematics of 3-D flow field

Consider the following flow field:

x1 2 x2 3 x3
v1 = , v2 = , v3 =
1+ t 1+ t 1+ t

Tasks
(i) Determine the acceleration of a fluid particle using a spatial description of the flow field.
(ii) Determine a material description of the flow field (i.e. the velocity is expressed in terms of
X1 , X2 , X3 ). What is the acceleration of the fluid particle in terms of material coordinates?
(iii) Show that (i) and (ii) are equivalent.

Solution: Part (i)

(i) the acceleration of a fluid particle is given by
v Dv ∂v
a= ≡ = + v · ∇v
t  Dt ∂t
68 ECH41MiscProblemsPart2_2024.nb

Thus for the spatial description

∂ v1 x1 ∂ v2 2 x2 ∂ v3 3 x3
=- , =- , =-
∂t (1 + t)2 ∂t (1 + t)2 ∂t (1 + t)2

and the components of v · ∇ v are

∂ v1 ∂ v1 ∂ v1 ∂ v1 x1
(v · ∇ v ) · e1 = v1 + v2 + v3 = v1 =
∂ x1 ∂ x2 ∂ x3 ∂ x1 (1 + t)2

∂ v2 ∂ v2 ∂ v2 ∂ v2 4 x2
(v · ∇ v ) · e2 = v1 + v2 + v3 = v2 =
∂ x1 ∂ x2 ∂ x3 ∂ x2 (1 + t)2

∂ v3 ∂ v3 ∂ v3 ∂ v3 9 x3
(v · ∇ v ) · e3 = v1 + v2 + v3 = v3 =
∂ x1 ∂ x2 ∂ x3 ∂ x3 (1 + t)2

Thus the spatial description of the acceleration is

∂ v1 ∂ v1
a1 = + v1 = 0,
∂t ∂ x1

∂ v2 ∂ v2 2 x2
a2 = + v2 = ,
∂t ∂ x2 (1 + t)2

∂ v3 ∂ v3 6 x3
a3 = + v3 = ,
∂t ∂ x3 (1 + t)2

Solution: Part (ii)

To determine the material description we need the particle paths. Thus
 x1 x1
v1 = =  x1 = 1 (1 + t)
t 1+ t

 x2 2 x2
v2 = =  x2 = 2 (1 + t)2
t (1 + t)

 x3 3 x3
v3 = =  x3 = 3 (1 + t)3
t (1 + t)

The acceleration is then

2 x1
a1 = =0
 t2 
 ECH41MiscProblemsPart2_2024.nb 69

2 x2
a2 = = 2 2
 t2 

2 x3
a3 = = 6 3 (1 + t)
 t2 

Solution: Part (iii)

Since we have shown that x2 = 2 (1 + t)2 , we can eliminate 2 in a2 to get
2 x2 2 x2
a2 = = 2 2 =
2
t  (1 + t)2

and for a3 we have

2 x3 6 x3
a3 = = 6 3 (1 + t) =
2
t  (1 + t)2

These are the spatial components of the velocity which are the same as found in (i)

Problem 29 : Venturi Channel: Bernoulli Analysis

Air flows through a Venturi channel of rectangular cross section as shown in the figure below. The
width of the rectangular channel is 0.06 m and the height at the exit is 0.04 m. The air may be taken to
incompressible and viscous effects are negligible.

Tasks
(i) Determine the flow rate when water is drawn up 0.10 m in a small tube attached to the static
pressure tap at the throat where the channel height is 0.02 m.
(ii) Determine the channel height h2 at section (2) where, for the same flow rate as in part (i), the water
is drawn up 0.05 m.
(iii) Determine the pressure needed at section (1) to produce this flow.
70 ECH41MiscProblemsPart2_2024.nb

Flow Parameters: ρw = 1000 kg  m3 , ρair = 1.2 kg  m3 , g= 9.8 m  s2 , Pa = kg  m · s2 

Problem Solution:
(i) Let us designate the cross sectional area of the duct at the throat as
At = ht W

Thus the flow rate Q at the throat is

Q = At Vt , where Vt is the air speed at the throat.

Now using Bernoulli’s equation we have

1 1
ρ V2t + Pt = ρ V2exit + Pexit
2 2
Now
Pexit = P0 , Pt = P0 - ρw g Ht , where Ht = 0.1 m

Hence Bernoulli’s equation becomes

2 ρw g
V2t - V2exit = (Pexit - Pt ) = 2 Ht
ρ ρ

But since the flow rate is constant, we have

At
Q = At Vt = Aexit Vexit ⟹ Vexit = Vt
Aexit

Thus Bernoulli’s equation becomes

At 2 ρw
V2t 1 - =2 Ht
Aexit ρ

Solving for Vt gives

2 ρw
g Ht 2 ρw
g Ht
ρ ρ
Vt = =
At 2 ht 2
1-  1- 
Aexit hexit

Using the parameter values given we get

1000
2× × 9.8 m  s2 (0.1 m)
1.2
Vt = = 46.6667 m / s
2
1 -  0.02 m 
0.04 m

Hence the flow rate is

Q = W ht Vt = (0.06 m) (0.02 m) (46.6667 m / s)

= 0.056 m3  s
 ECH41MiscProblemsPart2_2024.nb 71

(ii) Now the air speed at station 2 is

ρ ρ
V22 + P2 = V2t + Pt
2 2
Thus
2 2 ρw g 2 ρw g
V22 = V2t + (Pt - P2 ) = V2t - (Ht - H2 ) = V2t - (0.05 m)
ρair ρair ρair

Solving for V2 gives

2 ρw g
V2 = V2t - (0.05 m)
ρair

2 × 1000 × 9.8 m  s2
= (46.6667 m / s)2 - (0.05 m)
1.2

= 36.8932 m / s

Then noting that the flow rate at station ( 2) is

Q = A2 V2 = Wh2 V2

Solving for h2 gives

Q 1
h2 = Q / (W V2 ) =
W V2t -
2 ρw g
(0.05 m)
ρ

0.056 m3  s 1
=
0.06 m 36.8932 m / s

= 0.0252982 m

(iii) Now applying Bernoulli’s equation between station 1 and the throat gives
ρ ρ
V21 + P1 = V2t + Pt
2 2
Solving for P1 gives
ρ ρ
P1 = V2t - V21  + Pt = V2t - V21  + P0 - ρw g Ht
2 2
Thus the pressure drop (relative to atmospheric pressure P0 ) required to drive the flow is
ρ
P1 - P0 = V2t - V21  - ρw g Ht
2
72 ECH41MiscProblemsPart2_2024.nb

1.2 kg  m3
 (46.6667 m / s)2 - (36.8932 m / s)2  - 1000 kg  m3 9.8 m  s2 0.05 m
2

490. kg
= 490.002 kg  m s2  -
m s2

=0

Thus the pressure at P1 is the same as P0 which also follows from a Bernoulli calculation between
station 1 and the exit!
ρ ρ
V21 + P1 = V2exit + P0
2 2
and since the cross section of the channel at station 1 is the same as at the exit, the velocities at the
the two stations must be the same (flow rate constant)
v1 = vexit

Hence
P1 = Pexit

This calculation shows that if fluid arrives at station 1 with velocity v1 , and there are no viscous losses,
then as air flows through the venturi, one has a exchange of kinetic energy and pressure energy along
the venturi with no overall pressure drop across the venturi.

Problem 30 : Energy Analysis of a Centrifugal Pump

A schematic of our pump is shown in the following figure. The fluid enters the vertical tube of radius r1
and exits the casing with radius r2. The impeller is also shown in the figure. Note that its vanes are
oriented radially. Thus the fluid velocity leaving the pump is tangential to the vanes and is denoted as
v2 . The tangential velocity of the vanes at r2 is u2 . Hence the relative velocity of the fluid leaving the
pump at r2 is vrel . As shown in the figure the fluid entering the pump is along the axis of rotation.
 ECH41MiscProblemsPart2_2024.nb 73

Task:
Derive an expression that shows the the relationship between pressure drop and the volumetric
throughput of the pump.

Problem Solution:
The transfer of energy to the fluid passing through the centrifugal pump can be accounted for through
the conservation of angular momentum. For steady flow the macroscopic balance of angular momen-
tum reduces to
 
 Maxis =  (r  V) mout - (r  V) min

This is known as Euler's formula for centrifugal pumps and simply states that the sum of moments
about the axis of rotation is equal to the net change in angular momentum of the fluid entering and
leaving the control volume. Let τ be the torque about the axis and for our system the balance of
angular momentum becomes
 
τ =  (r  V) mout - (r  V) min

The angular momentum of the entering fluid is min r1 vw1 where vw1 is the tangential velocity of the

fluid at r1 . Likewise the angular momentum of the fluid leaving the pump is mout r2 vw2 . By conserva-
  
tion of mass min = mout = m. Thus our balance of angular momentum becomes

τ = m (r2 vw2 - r1 vw1 )

If the rotor is spinning at angular velocity ω, then the power (work done per unit time) is
74 ECH41MiscProblemsPart2_2024.nb

P= τω J s-1 or Watts

Recall that the tangential velocity of the rotor at r1 and r2 are

u1 = ω r1 , u2 = ω r2

Hence the power transmitted to the fluid is

 
P = m (r2 ω vw2 - r1 ω vw1 ) = m (u2 vw2 - u1 vw1 )

Then the rate of work done per unit mass of fluid is

P
E =  = (u2 vw2 - u1 vw1 )
m
This is also equal to g H, where H is the Euler head. Thus
Δp
E=gH= = (u2 vw2 - u1 vw1 )
g

Thus by computing the RHS of of the above equation we can determine the pressure head developed
by the pump. In our case the fluid enters the pump in the axial direction so there is no contribution to
angular momentum as vw1 = 0 and we are left with
Δp
E=gH= = u2 vw2
g

Since vw2 is radially outward (tangential to the blades of the impeller) we can write
 
Q Q
vw2 = =
A 2 πr2 h

where h is the height of the impeller. Hence if we know the power imparted to the impeller and the
angular velocity of the impeller, we can determine the pressure drop across the pump and hence
determine the volumetric throughput of the pump.

Problem 31: Analysis of a Underwater Detonation

Suppose at t = 0 an explosive charge initially of radius a0 is detonated under water. Subsequently,

the explosive gaseous debris from the charge creates a cavity of radius a (t) and the pressure of the
debris is uniform for 0 ≤ r ≤ a (t) . There is no mass transferred across the cavity interface.

Tasks :
(i) Show that for a spherically expanding cavity, the velocity of the water given by v = m (t)  r2 er
satisfies the equation of continuity for an incompressible liquid.

(ii) Choose a fixed control volume in the liquid with radius R >> a(t). Write down a macroscopic
mass balance for this control volume and then deduce what m (t) is.

(iii) Redo task (ii) but instead of a fixed control volume choose one that moves with some arbitrary
 ECH41MiscProblemsPart2_2024.nb 75

velocity w

Problem Solution:
(i) In spherical coordinates the continuity equation for an "incompressible fluid" is

1 ∂ 1 ∂ 1 ∂
∇ ·v = r2 vr  + (vθ Sin (θ)) + (vϕ )
r2 ∂ r r Sin (θ) ∂ r rSin (θ) ∂ ϕ

where the velocity when expressed in terms of spherical base vectors is

v = vr er + vθ eθ + vϕ eϕ

Since we are given

vr = m (t)  r2 , vϕ = 0, vθ = 0

it follows that
1 ∂
∇ ·v = {m (t)} = 0
r2 ∂ r
(ii) We take our fixed control volume to have radius r=ℛ >>a(t) for all t of interest. The macro-
scopic mass balance for this control volume is

 ρ  V = -  n · jrel  A
t V A

Now the volume V is made up of two parts the gas bubble VG and the liquid VL so that

 ρG  V +  ρL  V = -  e r · ρ v  A
t VG VL A

where we have used the fact that at a stationary surface

jrel = ρ (v - w) = ρ v, as w = 0

Taking the densities to be constant we can integrate the volume integrals and surface integral to get
 4 4
 π a 3 ρG + π ℛ3 - a3  ρL  = -4 π ℛ2 ρL vr
t 3 3
But we are told that no mass is transferred across the bubble interface. Hence a control volume
located at r = a (t) gives

 ρG  V = 0
t VG

which implies
 4
 π a 3 ρG  = 0
t 3
Using this result in our previous macroscopic balance gives
76 ECH41MiscProblemsPart2_2024.nb

-4 π a2 a
t
ρL = -ρL vr 4 π ℛ2

Solving for vr gives

a 2 a
vr =   at r = ℛ
r t
from which it follows that
a
m (t) = a2
t
(iii) In this calculation we take as our control volume a spherical volume with radius r = R (t) such
that surface velocity of the control volume is w = R e . Applying a macroscopic mass balance to our
t r
moving control volume, we get

 ρG  V +  ρL  V = -  er · ρ ( v - w)  A
t VG VL A

Evaluating the integrals as before we find

 4 4 R
 π a 3 ρG + π R3 - a3  ρL  = -4 π R2 ρL vr -
t 3 3 t
Differentiating, we get
R a R
4 ρL πR2 - 4 ρL πa2 = -4 πR2 ρL vr + 4 ρL πR2
t t t
or
a 2 a
vr =  
r t
Note at the surface of our control volume r = R

Problem 32 : Stress Field Analysis: Example 1

The stress tensor at a point P in a fluid body is given by

T = e 1 e 2 + 2 e 1 e 3 + e 2 e 1 + α e 2 e 2 + e 2 e 3 + 2 e3 e 1 + e 3 e2

Tasks :
(i) Is the fluid at P in equilibrium? Indicate your reasoning.

(ii) Determine α so that the traction vector t(n) on some plane with unit normal n passing through
the point P will be zero, i.e., t(n) = 0 . Give the unit normal for this traction free plane.
Hint: select n = n1 e1 + n2 e2 + n3 e3 , ( where n·n =1) and then determine the values of ni
so that t(n) = 0.
 ECH41MiscProblemsPart2_2024.nb 77

Problem 11 Solution:
(ii) for the fluid to be in equilibrium we require that there be no relative motion. So for hydrostatic
equilibrium, the stress tensor must be isotropic
T = -pI

In this case the stress tensor at the point P has the following structure
1 1 2
[T] = 1 α 1
2 1 0

Clearly this stress tensor is not isotropic, that is

-p 0 0
[T] ≠ 0 -p 0
0 0 -p

Thus the fluid is not in equilibrium.

(ii) We are required to find n such that t(n) = 0. Let

n = n1 e1 + n2 e2 + n3 e3

Then
t(n) · n = 0

But
t(n) = T · n = n1 e2 + 2 n1 e3 + n2 e1 + σ n2 e2 + n2 e3 + 2 n3 e1 + n3 e2

= (n2 + 2 n3 ) e1 + (n1 + σ n2 + n3 ) e2 + (2 n1 + n2 ) e2

For this vector to be zero, we require all the components to be zero

n2 + 2 n3 = 0, n1 + σ n2 + n3 = 0, 2 n1 + n2 = 0

We solve the first and third equations to get

n2 n2
n3 = - , n1 = -
2 2
Substituting into the second equation we find
σ n2 - n2 = 0  σ=1

Hence our unit normal is (taking n2 = 1 : note we can select any value)
1 1
n= - e1 + e2 - e3  1/4 + 1 + 1/4
2 2
Simplifying gives the unit normal to the traction free plane.
78 ECH41MiscProblemsPart2_2024.nb

1
n= (-e1 + 2 e2 - e3 )
6

Problem 33 : Stress Field Analysis: Example 2

The state of stress at a point P in a fluid is given by

20 10 -10
[T] = 10 30 0
-10 0 50

Tasks:
Determine the normal stress and the shearing stress on a surface intersecting the point P and parallel
to the plane 2x+y-3z=9 . Hint: recall that the force per unit area acting at a point in the fluid on a plane
with unit normal n is
t(n) = T · n, where T = Tij ei ej

Problem Solution
The unit normal to the plane F(x,y,z) = 0 is
∇F 2 i +j -3 k
n= =
∇ F 14

Thus the traction (stress) vector at the point P is

1 20 10 -10 2
t(n) = T · n ⟹ 10 30 0 1
14 -10 0 50 -3

2 2 2
= 40 , 25 , -85 
7 7 7

Note that the magnitude of the stress vector is

t(n)  = t(n) .t(n)

2 2 2 2 2 2
= 40 , 25 , -85 .40 , 25 , -85 
7 7 7 7 7 7
= 30 3

Now the normal component of the traction vector is

2 2 2
t(n) · n = 40 , 25 , -85 .{2, 1, -3}  14
7 7 7
 ECH41MiscProblemsPart2_2024.nb 79

360
=
7
To compute the magnitude of the shear stress, note that the stress vector can be represented as
t(n) = TN n + TS eS

Thus the magnitude of the shear stress is

360 2 30 3
Ts = t(n) 2 - T2N = 2700 - = = 7.42307
7 7

Problem 34 : Stress Field Analysis: Example 3

In a continuum the stress field is given by T with components

x1 2 x2 1 - x2 2  x1 0
T= 2 3
1 - x2  x1 x2 - 3 x2   3 0
0 0 2 x3 2

Tasks:
Determine the body force distribution if the equations of statics are to be satisfied throughout the
continuum.
Is the continuum a fluid?

Problem Solution:
For static equilibrium our continuum body must satisfy
∇ ·T + ρ b = 0

In component form we have

∂
ei · (Tj k ej ek ) + ρ bl el
∂ xi

which gives
∂ Tj k
ek + ρ bl el = 0
∂ xj

Next we consider the components of this equation in the orthogonal directions.

The e1 component is:
∂ Tj1
+ ρ b1 = 0
∂ xj

Now evaluating the derivative in the above equation using the stress tensor given in the problem
statement we find
80 ECH41MiscProblemsPart2_2024.nb

2 x1 x2 - 2 x2 x1 + ρ b1 = 0  b1 = 0

The e2 component is
∂ Tj 2
+ ρ b2 = 0
∂ xj

This gives
1 - x22  + x22 - 1 + ρb2 = 0  b2 = 0

The e3 component is
∂ Tj 3
+ ρ b3 = 0
∂ xj

which gives
4
4 x3 + ρ b 3 = 0  b 3 = - x3
ρ

Hence the body force necessary for static equilibrium is

4
b=- x3 e3
ρ

Comment: Note the stress tensor has non-zero off-diagonal terms. This means there is a shear stress
component. Since a fluid cannot support a shear stress in static equilibrium, our continuum body is
not a fluid. It could be a solid though!

Problem 35 : Stress Field Analysis: Example 4

A test surface in a moving fluid is oriented in three directions in succession, and the traction forces are
measured.
The three directions in which test surface was oriented are:
(i) (e1 + e2 )  2

(ii) (e1 - e2 )  2

(iii) e3

The measured tractions on the test surface F  L2  are

(i) (e1 + e2 ) μ - 2

(ii) (e1 - e2 ) -μ - 2

(iii) e3
 ECH41MiscProblemsPart2_2024.nb 81

where the ei ' s are unit base vectors for a rectangular Cartesian coordinate system.

Tasks:
Determine the components of the stress tensor at the measuring point.

Problem Solution:
In this problem we are given the traction values on three mutually orthogonal planes:

t (e1 ) = (e1 + e2 ) μ - 2


t (e2 ) = (e1 - e2 ) -μ - 2


t (e3 ) = e3

where ei are the unit normals to the test planes given by

e1 = (e1 + e2 )  2


e2 = (e1 - e2 )  2


e3 = e 3

From our definition of the traction vector (viz. t (n) = n · T) we have

 
t (e1 ) = e1 · T = (e1 + e2 ) μ - 2

But in terms of the ei basis we can write

  
e1 · T = e1 · Ti j ei ej = e1 · ei  Ti j ej

and using the fact that e1 = (e1 + e2 )  2 we have
 1  1 
e1 · e 1 = , e1 · e2 = , e1 · e3 = 0
2 2

Thus
 1 1 
e1 · T = T 1 j ej + T2 j ej = t (e1 ) = (e1 + e2 ) μ - 2
2 2

The components of this vector equation are

1 1
T11 + T21 = μ - 2 (i)
2 2
82 ECH41MiscProblemsPart2_2024.nb

1 1
T12 + T22 = μ - 2 (ii)
2 2

1 1
T13 + T23 = 0 (iii)
2 2

Using the same procedure we can show that

 1 1 
e2 · T = T1 j ej - T2 j ej = t (e2 ) = (e1 - e2 ) -μ - 2
2 2

The components of this vector equation are

1 1
T11 - T21 = -μ - 2 (iv)
2 2

1 1
T12 - T22 = μ + 2 (v)
2 2

1 1
T13 - T23 = 0 (vi)
2 2

Finally from the last test plane

 
e3 · T = T3 j ej = t (e3 ) = e3

and the components of this vector equation are

T31 = 0 (vii)

T32 = 0 (viii)

T33 = 1 (ix)

Thus we have 9 equations to determine the nine components of the stress tensor. Solving we find
-2 μ 2 0
Ti j = μ 2 -2 0
0 0 1

Problem 36 : Stress, Velocity and Volumetric Flow Analysis of flow

between two flat plates

The Navier-Stokes equations for an incompressible fluid are

∇ ·v = 0
 ECH41MiscProblemsPart2_2024.nb 83

∂v
ρ + v · ∇ v = -∇ p + μ ∇2 v + ρ g
∂t
If we ignore entrance and exit effects, the flow in the channel may be assumed to be rectilinear ( or
unidirectional). Thus we may assume the velocity is given by
v = vz (t, x) ez

For steady flow with translational symmetry (no variation in the y-direction) the Navier-Stokes equa-
tions reduce to
∂ vz
=0
∂z

0 = -∇ p + μ ∇2 vz ex + ρ g

If we ignore gravitational effects, the components of the N/S equations simplify to give
∂ vz
Eq.(i) =0
∂z

∂p ∂2 vz
Eq. (ii) =μ
∂z ∂ x2

∂p
Eq. (iii) =0
∂x
Note that the continuity equation is satisfied identically as vz = vz (x). Equation (iii) advises us that p =
p(z). Hence the RHS of Eq. (iii) must be a constant. Integrating Eq.(ii) from z = 0 to z = L gives
Pl ∂p L ∂2 vz
 z =  μ z
P0 ∂z 0 ∂ x2

∂2 vz
(PL - P0 ) = L μ
∂ x2

Tasks:
(i) Determine an expression for the velocity field

(ii) Determine the volumetric flow rate per unit between the two plates.

(iii) Derive an expression for the viscous stress tensor τ

(iv) Compute the normal components of the viscous tenor τ

(v) Compute the shear components of the viscous stress tensor

84 ECH41MiscProblemsPart2_2024.nb

Problem Solution:
Rearranging the above results gives
ΔP ∂2 vz
- =μ , where ΔP = P0 - Pl
L ∂ x2
This ODE can be integrating directly to give
ΔP x2
vz (x) = - + C1 x + C2
L 2
To determine the constants C1 and C2 we make use of the no-slip boundary conditions applied at the
wall of the channel.
BC1 : vz = 0, at x = H / 2

BC2 : vz = 0, at x = -H / 2

 vz
BC3 : = 0, at x = 0
x
Note that BC3 is a statement that the flow is symmetric about x=0. We are at liberty to use BC3 instead
of BC2. Keep in mind symmetry conditions often simplify the algebra. Applying BC1 and BC2, or BC1
and BC3 we find
ΔP H2
vz (x) = - x2 -
2μL 4

The volumetric flow is given by

W H/2 H/2
Q=  vz (x)  x  y = 2 W  vz (x)  x
0 -H/2 0

Integrating gives
 Q ΔP
Q= = H3
W 12 μ L

To compute the stress profile, we note that for a Newtonian fluid the viscous stress tensor is given by
1
τ = 2 μ D, where D= ∇ v + ∇ vT 
2
Recall that
H/2 ∂ vi
 vz (x)  x = ei ej
-H/2 ∂ xj

∂ v1 ∂ v1 ∂ v2 ∂ v2
= e1 e1 + e2 e1 + e1 e2 + e2 e2
∂ x1 ∂ x2 ∂ x1 ∂ x2
 ECH41MiscProblemsPart2_2024.nb 85

We can write the components in matrix form

∂v1 ∂v2 ∂v1 ∂v1
∂x1 ∂x1 T ∂x1 ∂x2
[∇ v] = ∂v1 ∂v2
, [∇ v] = ∂v2 ∂v2
∂x2 ∂x2 ∂x1 ∂x2

Thus the viscous stress tensor τ becomes

[τ] = μ [∇ v] + [∇ v]T 

2 ∂v1
 ∂v2 + ∂v1

∂x1 ∂x1 ∂x2
[τ] = μ
 ∂v2 + ∂v1
 2 ∂v2
∂x1 ∂x2 ∂x2

Now let subscript “1” denote the component in the ez direction and subscript “2” denote the compo-
nent in the e x direction. And since vz = vz ( x ) we have
∂ vz
0
∂x
[τ] = μ ∂vz
0
∂x

In tensor notation the viscous stress tensor is

∂ vz ∂ vz  vz  vz
τ=μ ex ez + ez ex = μ ex ez + ez ex
∂x ∂x x x
The normal components of the stress tensor are
ex · τ · ex = 0 = ez · τ · ez = 0

while the shear stress is given by

 vz Δp
ex · τ · ez = τxz = τzx = μ =- x
x μL

Problem 37 : Analysis of 2-D Flow field

A 2-dimensional flow field in the x-y plane, bounded by a moving solid plane at y=0 is studied experi-
mentally. Sufficiently far away from the moving solid plane, the velocity of a fluid particle is measured
moving along a horizontal streamline that coincides with the x- axis. The x-component of the two
dimensional incompressible flow field was experimentally found to be describes by the equation
u = x2

Tasks:
Along this streamline determine an expression for
(a) the rate of change of the v component of velocity with respect to y
(b) the acceleration of the fluid particle along the streamline
(c) the pressure gradient in the x-direction
(d) the components of viscous stress tensor τ at a point P (x = 3, y = 2) on the streamline, assuming that
86 ECH41MiscProblemsPart2_2024.nb

u = x2 is a good approximation of the flow even at y = 0.

Problem solution
(i) The continuity equation for this flow is
∂u ∂v
+ =0
∂x ∂y

Thus along the streamline (say at y = ys )where the measurement is made

∂v
= -2 x, at y = ys
∂y

Next we would like to know how v- component of velocity varies with y. This is best accomplished by
integrating the continuity equation from y=0 to y = ys .
ys ∂u ys ∂v
 y +  y = 0
0 ∂x 0 ∂y

We can write this integral as

 ys
 u  y + v (x, ys ) - v (x, 0) = 0
x 0

Noting that v(x,0)=0 (recall at y=0 the flow is bounding by a solid surface moving parallel to the x-axis.
Thus we have
 ys
 u  y = -v (x, ys )
x 0

At this point we do not know how u(x,y) varies with y. All we know is that at y = ys , u(x, ys ) = x2 . We will
suppose therefore that u is only a function of x. Clearly, this would be reasonable if the speed of the
“plate” at y=0 is a function of x. If we make this assumption then
u (x, y) = x2 for all y > 0

At the point P on the streamline we have

u
ys = -v (x, ys )
x
which gives for any y
v (x, y) = -2 x y

Thus
∂v
= -2 x, v (x, y) = -2 x y
∂y

Note that the value of v(x,y) is required for later calculations

(ii) The acceleration a = a x e x + a y e y of the fluid particle along the streamline is

 ECH41MiscProblemsPart2_2024.nb 87

∂u ∂u ∂v ∂v
ax = ρ u +v , ay = ρ u +v
∂x ∂y ∂x ∂y

for the given velocity field we have

ax = ρ x2 (2 x) + (-2 xy) (0) = 2 ρ x3

ay = ρ x3 (-2 y) + (-2 x y) (-2 x) = ρ -2 x3 y - 4 x2 y

= -2 ρ x2 y (x + 1)

Hence the acceleration along the streamline

a = 2 ρ x3 ex - 2 ρ x2 y (x + 1) ey

(iii) From the x-momentum equation we have

∂p ∂2 u ∂2 u
= -ax + μ +
∂x ∂ x2 ∂ y2

= -2 ρ x3 + μ (2) + μ (0)

= -2 ρ x3 - μ

(iv) The components of the viscous stress tensor are

2μ ∂u
μ  ∂u + ∂v
 4 μ x -2 μ y
∂x ∂y ∂x
[τ] = = 
μ  ∂u + ∂v
 2μ ∂v -2 μ y -4 μx
∂y ∂x ∂y

Evaluating at x=3 y=2 gives

12 μ -4 μ
[τ] =  
-4 μ -12 μ

Problem 38 : Kinematics of Simple Plane Flow

Problem Statement
Given the following simple plane flow
v1 = x1 / (1 + t), v2 = x2 , v3 = 0

Determine the streamlines at time t, the particle paths, and the streak-lines through xo . Plot the
streamlines in the x2 - x1 plane. Take xo = {1, 2}.

Problem Solution:
The path lines are found by integrating
88 ECH41MiscProblemsPart2_2024.nb

 x1  x2  x3
v1 ≡ = x1 / (1 + t), v2 ≡ = x2 , v 3 ≡ =0
t t t
The result is
x1 = X1 (1 + t), x2 = X2 et , x3 = X3

We can use the first equation to get an explicit expression for t

t = x1 / X2 - 1

Substituting into the equation for x2 gives

x2 = X2 e-(1-x1 /X1 )

To plot the particle paths x1 - x2 plane, we select values for X1 , X2 and then compute the resulting
path. These calculations can be readily done in Mathematica using ParametricPlot . Here is a way
to plot the particle paths. First we define an array of potential {X1 , X2 } positions for our path lines:

I n [ ] : = InitialPositions = Transpose[{Range[1, 5], Range[1, 5]}]

O u t [ ] =

{{1, 1}, {2, 2}, {3, 3}, {4, 4}, {5, 5}}

Next we define a set of replacement rules that assign the variables X1 , Y1 to these positions:

I n [ ] : = InitialConditions = Map[Thread[{X1 , X2 }  #] &, InitialPositions]

O u t [ ] =

{{X1  1, X2  1}, {X1  2, X2  2}, {X1  3, X2  3}, {X1  4, X2  4}, {X1  5, X2  5}}

We would like also to mark the initial particles on the plot. This can be readily done using a graphic
primitive called Point. The following statement creates the graphic primitives for each of the
particles

I n [ ] : = InitialParticles = Map[Point[#] &, Transpose[{Range[1, 5], Range[1, 5]}]]

O u t [ ] =

{Point[{1, 1}], Point[{2, 2}], Point[{3, 3}], Point[{4, 4}], Point[{5, 5}]}

With those preliminaries out of the way, we now use ParametricPlot to graph our path lines. The
graphic primitives are added to the plot using the Epilog option.
 ECH41MiscProblemsPart2_2024.nb 89

I n [ ] : = ParticlePathLines =
ParametricPlot[Evaluate[{X1 (1 + t), X2 Exp[t]} /. InitialConditions],
{ t, 0, 2}, PlotStyle  Thick, PlotRange  All, AxesOrigin  {0, 0},
Epilog  {RGBColor[1, 0, 0], PointSize[0.03], InitialParticles},
Frame  True, FrameLabel  {Style["x1 ", 16], Style["x2 ", 16]}, AspectRatio  1,
PlotLegends  {"X0 =1,Y0 =1", "X0 =2,Y0 =2", "X0 =3,Y0 =3", "X0 =4,Y0 =4", "X0 =5,Y0 =5"}]
O u t [ ] =

30

X0 =1,Y0 =1
20
X0 =2,Y0 =2
x2

X0 =3,Y0 =3
X0 =4,Y0 =4
10 X0 =5,Y0 =5

0 5 10 15

x1

Consider next the streamlines. We can use the definition of the streamline: the velocity is everywhere
tangent to the streamline. Mathematically this is represented as
 xs
=v
α
where α is a parameter for the streamline (e.g arc length), and xs is the position vector for the STREAM-
LINE, at any point in time. Note: it is in general not the position vector for the particle path. For the
given velocity field we have
 x1  x2  x3
v1 ≡ = x1 / (1 + t), v2 ≡ = x2 , v 3 ≡ =0
α α α
Integrating while holding t fixed gives
x1 = x1 (0) e α /(1+t) , x2 = x2 (0) eα , x3 = x3 (0)

where xi (0) are the positions in the flow field defined by α = 0. If we eliminate α we get
x2 1/(1+t)
x1 = x1 (0)
x2 (0)

We can readily visualize the stream lines using the program we developed for the particle paths. Let us
90 ECH41MiscProblemsPart2_2024.nb

take a snapshot of the streamlines at t=2, the last time step used in the particle path calculation. The
location of the streamline is defined at the same positions we used to define the particles at time t=0.

I n [ ] : = Streamlines =
ParametricPlot[Evaluate[{X1 Exp[α / (1 + 2)], X2 Exp[α]} /. InitialConditions],
{ α, 0, 2}, PlotStyle  Thick, PlotRange  All, AxesOrigin  {0, 0},
Epilog  {RGBColor[1, 0, 0], PointSize[0.03], InitialParticles},
Frame  True, FrameLabel  {Style["x1 ", 16], Style["x2 ", 16]}, AspectRatio  1,
PlotLegends  {"X0 =1,Y0 =1", "X0 =2,Y0 =2", "X0 =3,Y0 =3", "X0 =4,Y0 =4", "X0 =5,Y0 =5"}]
O u t [ ] =

30

X0 =1,Y0 =1
20
X0 =2,Y0 =2
x2

X0 =3,Y0 =3
X0 =4,Y0 =4
10 X0 =5,Y0 =5

0 2 4 6 8 10

x1

If we combine the plots we readily see that the path-lines and streamlines do not coincide

I n [ ] : = Streamlines2 =
ParametricPlot[Evaluate[{X1 Exp[α / (1 + 2)], X2 Exp[α]} /. InitialConditions],
{ α, 0, 2}, PlotStyle  Thick, PlotRange  All, AxesOrigin  {0, 0},
Epilog  {RGBColor[1, 0, 0], PointSize[0.03], InitialParticles},
Frame  True, FrameLabel  {Style["x1 ", 16], Style["x2 ", 16]}, AspectRatio  1];
 ECH41MiscProblemsPart2_2024.nb 91

I n [ ] : = Show[Streamlines2, ParticlePathLines]
O u t [ ] =

30

X0 =1,Y0 =1
20
X0 =2,Y0 =2
x2

X0 =3,Y0 =3
X0 =4,Y0 =4
10 X0 =5,Y0 =5

0 5 10 15

x1

In the above plot each streamline and particle path line originates from the same point at time t=0.
Note though the trajectory of the streamline and the particle path line do not coincide.

For the streamline (at t=2) that passes through x1 (0) = 5, x2 (0) = 5, it is possible to find the
coordinates for a particle when α=2. We need to solve
x1 = x1 (0) e α /(1+t) , x2 = x2 (0) eα , with α = 2, t = 2

The coordinate values are.

x1 = 9.73867, x2 = 36.9453

Now let us use the particle path line equations to determine the trajectory of a particle that at t = 2,
is located at x1 = 9.74 , x2 = 36.95. Recall a particle path is defined by its initial location at
t = 0. Thus we need to solve for X1 , X2 Using the equations for the particle path lines we have
X1 = x1 / (1 + t) X2 = x2 e-t

Evaluating these expressions at t = 2, x1 = 9.74, x2 = 36.95 gives

X1 = 3.24622 X2 = 5.

Here is the trajectory of that particle

92 ECH41MiscProblemsPart2_2024.nb

I n [ ] : = aPathLine = ParametricPlot[Evaluate[{X1 (1 + t), X2 Exp[t]} /. {X1  3.24622, X2  5}],

{ t, 0, 2}, PlotStyle  Red, PlotRange  All, AxesOrigin  {0, 0},
Epilog  {RGBColor[1, 0, 0], PointSize[0.03], Point[{3.25, 5}]},
Frame  True, FrameLabel  {Style["x1 ", 16], Style["x2 ", 16]}, AspectRatio  1]
O u t [ ] =

30

20
x2

10

0 2 4 6 8 10

x1

If we superimpose this trajectory onto the previous plots we see that it becomes tangent to the stream-
line at t = 2, at location x1 = 9.7, x2 = 36.9
 ECH41MiscProblemsPart2_2024.nb 93

I n [ ] : = Show[aPathLine, Streamlines2, ParticlePathLines, PlotRange  {{0, 10}, {0, 35}}]

O u t [ ] =

35

30

25

20
X0 =1,Y0 =1
X0 =2,Y0 =2
x2

15 X0 =3,Y0 =3
X0 =4,Y0 =4
10
X0 =5,Y0 =5

0 2 4 6 8 10

x1

The above plot also shows that streamlines and path lines do not in general coincide when you have a
time dependent flow.

Consider next the computation of the streak line. We know that the path lines are given by
x1 = X1 (1 + t), x2 = X2 et , x3 = X3

The streak line is defined as the line on which all particles lie at some instant in time that passed
  
through a common point. Suppose the common point is x1 , x2 , x3 . Now for t < τ, the particles
  
that pass through x1 , x2 , x3 are
  
X1 = x1 / (1 + τ), X2 = x2 e-τ , X3 = x3 .

Hence the equation for the streak line is


x1  
x1 = (1 + t), x2 = x2 et-τ , x3 = x3
(1 + τ)

We can use Mathematica to plot the streak line when t = 2, and for 0 < τ < 2. The x - y point at
which we tag particles is (5, 5)
94 ECH41MiscProblemsPart2_2024.nb

I n [ ] : = aStreakLine = ParametricPlot[Evaluate[{3 X1 / (1 + τ), X2 Exp[2 - τ]} /. {X1  5, X2  5}],

{τ, 0, 2}, PlotStyle  {Thickness[0.005], RGBColor[0, 1, 0]}, PlotRange  All,
AxesOrigin  {0, 0}, Epilog  {RGBColor[1, 0, 0], PointSize[0.03], Point[{5, 5}]},
Frame  True, FrameLabel  {Style["x1 ", 16], Style["x2 ", 16]}, AspectRatio  1]
O u t [ ] =

30

20
x2

10

0 5 10 15

x1

In the following plot we combine the streak lines with the path lines and streamlines calculated
previously.
 ECH41MiscProblemsPart2_2024.nb 95

I n [ ] : = Show[aStreakLine, aPathLine, Streamlines2,

ParticlePathLines, PlotRange  {{0, 10}, {0, 35}}]
O u t [ ] =

35

30

25

20
X0 =1,Y0 =1
X0 =2,Y0 =2
x2

15 X0 =3,Y0 =3
X0 =4,Y0 =4
10
X0 =5,Y0 =5

0 2 4 6 8 10

x1

Problem 39 : Analysis of three Linear 2-D flows

Consider the three linear, two-dimensional flows given below:

(i) vx = G y, vy = 0, vz = 0 ( a Couette flow )
(ii) vx = (G / 2) y, vy = (G / 2) x, vz = 0 (a purely straining flow)
(iii) vx = (G / 2) y, vy = -(G / 2) x, vz = 0 (a purely roational flow)

Note that the addition of flows (ii) and (iii) yields flow (i). For flows (i), (ii) and (iii) calculate:

(a) the velocity gradient, rate-of-strain and vorticity tensors;

(b) the viscous stress tensor (assuming the fluid is Newtonian with viscosity μ);
(c) the streamlines and then plot the streamlines for these three flows.

Solution

Part(a):
The velocity gradient is given by
∂ vj
∇v = ej ei
∂ xi
96 ECH41MiscProblemsPart2_2024.nb

The rate of strain tensor us given by

1
D= ∇ v + ∇ vT 
2
and the vorticity tensor is
1
Ω= ∇ v - ∇ vT 
2
The components of the velocity gradients for the three flows are
G
0 0 0 0 0 0 -G 0
2 2
(i) [∇ v] = G 0 0 , (ii) [∇ v] = G
0 0 , (iii) [ ∇ v] = G
0 0
2 2
0 0 0
0 0 0 0 0 0

The components of the rate of strain tensors for the three flows are
0 G 0 1 0G0 000
(i) [D] = G 0 0 , (ii) [D] = G 0 0 , (iii) [ D] = 000
0 0 0 2 000 000

The components of the vorticity tensors for the three flows are

1 0 -G 0 000 1 0 -G 0
(i) [Ω] = G 0 0 , (ii) [Ω] = 0 0 0 , (iii) [ Ω] = G 0 0
2 0 0 0 000 2 0 0 0

Part(b):
The viscous stress tensor is given by
τ=2μD

Thus the components of the viscous stress tensor for the three flows are
0 G 0 0G0 000
(i) [τ] = μ G 0 0 , (ii) [τ] = μ G 0 0 , (iii) [ τ] = 000
0 0 0 000 000

Part(c):
For steady flows the streamlines are given by
 xs
=v
α
where α is a parameter for the streamline ( e.g. arc length), and xs is the position vector for the stream-
line, at any point in time. Note: the velocity is everywhere tangent to the streamline.

◼ Flow (i)

For flow (i) we have:

 ECH41MiscProblemsPart2_2024.nb 97

 x1  x2
= G x2 , =0
α α
Solving we find
x2 (0), x1 = G x2 (0) t + x1 (0)

where x1 (0) is the position along the streamline when α = 0. Thus if we are given values for G and
xi (0) we can plot the streamline. Here is a parametric plot of the streamline in the x1 - x2 plane. We pick
the following initial conditions
x1 (0) = 0, x2 (0) = 0.25

I n [ ] : = initCond = {{X1  0, X2  0.25}, {X1  0, X2  0.5},

{X1  0, X2  0.75}, {X1  0, X2  1}, {X1  0, X2  1.25},
{X1  0, X2  1.5}, {X1  0, X2  1.75}, {X1  0, X2  2}};
G = 0.5;

I n [ ] : = ParametricPlot[Evaluate[{G X2 t + X1 , X2 } /. initCond], {t, 0, 10},

PlotStyle  Thick, Frame  True, FrameLabel  {Style["X1 ", 16], Style["X2 ", 16]},
PlotRange  {{0, 10}, {0, 2.5}}, AspectRatio  0.75, PlotLegends 
Placed[{"X1 =0, X2 =0.25", "X1 =0, X2 =0.5", "X1 =0, X2 =0.75", "X1 =0, X2 =1.0",
"X1 =0, X2 =1.25", "X1 =0, X2 =1.5", "X1 =0, X2 =1.75", "X1 =0, X2 =2"}, Right]]
O u t [ ] =
2.5

2.0
X1 =0, X2 =0.25
X1 =0, X2 =0.5
1.5
X1 =0, X2 =0.75
X2

X1 =0, X2 =1.0
1.0
X1 =0, X2 =1.25
X1 =0, X2 =1.5
0.5
X1 =0, X2 =1.75
X1 =0, X2 =2
0.0
0 2 4 6 8 10

X1

Time Dependent Rectilinear FlowsThe plot of the streamlines shows the familiar characteristic of
plane Couette flow

◼ Flow (ii)

For flow (ii) we have

 x1 G  x2
= x2 , = G2 x1
α 2 α
98 ECH41MiscProblemsPart2_2024.nb

To solve this system of equations, we can write the equations as

G
 x1 0 x1
2
 =  
t x2 G
0 x2
2

This linear system of ODEs can be readily solved by seeking solutions of the form
xi ≈ ξi Exp[λ α]

Substituting this expression into this system of equations results in the following eigenvalue problem:
G
-λ
2 ξ1 0
 = 
G
-λ ξ2 0
2

For a non-trivial solution we require that

G
-λ G2
Det G
2
 = 0, ⟹ λ2 - =0
-λ 4
2

Thus the eigenvalues are

G G
λ1 = , λ2 = -
2 2
The corresponding eigenvectors are
1 -1
ξ(1) =  , ξ(2) =  
1 1

Thus the general solution is

G G
x = C1 ξ(1) Exp α + C2 ξ(2) Exp- α
2 2
Now at α = 0, x1 (0) = X1 , and x2 (0) = X2 . Using these initial conditions, we can evaluate the constants C1
and C2 :
(X1 + X2 ) (X2 - X1 )
C1 = , C2 =
2 2
Hence the equations for the streamlines are
(X1 + X2 ) G (X1 - X2 ) G
x1 = Exp α - Exp- α
2 2 2 2

(X1 + X2 ) G (X1 - X2 ) G
x2 = Exp α + Exp- α
2 2 2 2
Next we illustrate how to plot the streamlines using Mathematica’s ParametricPlot function:

I n [ ] : = initialCond = {{X1  0, X2  0.25}, {X1  0, X2  0.5},

{X1  0, X2  0.75}, {X1  0, X2  1}, {X1  0, X2  1.25},
{X1  0, X2  1.5}, {X1  0, X2  1.75}, {X1  0, X2  2}};
 ECH41MiscProblemsPart2_2024.nb 99

The equations for the streamlines are

(X1 + X2 ) G (X2 - X1 ) G
I n [ ] : = Eqn1 = Exp α - Exp- α
2 2 2 2

(X1 + X2 ) G (X2 - X1 ) G
Eqn2 = Exp α + Exp- α
2 2 2 2
O u t [ ] =
1 1
- -0.25 α (-X1 + X2 ) + 0.25 α (X1 + X2 )
2 2
O u t [ ] =
1 1
-0.25 α (-X1 + X2 ) + 0.25 α (X1 + X2 )
2 2

We use ParametricPlot to display the streamlines for 0 ≤ α ≤ 1

I n [ ] : = ParametricPlot[Evaluate[{Eqn1, Eqn2} /. G  2 /. initialCond],

{α, 0, 4}, AspectRatio  1,
PlotStyle  Thick, Frame  True, FrameLabel  {Style["X1 ", 16], Style["X2 ", 16]},
PlotRange  {{0, 2}, {0, 3}}, AspectRatio  0.75, PlotLegends 
Placed[{"X1 =0, X2 =0.25", "X1 =0, X2 =0.5", "X1 =0, X2 =0.75", "X1 =0, X2 =1.0",
"X1 =0, X2 =1.25", "X1 =0, X2 =1.5", "X1 =0, X2 =1.75", "X1 =0, X2 =2"}, Right]]
O u t [ ] =
3.0

2.5

X1 =0, X2 =0.25
2.0
X1 =0, X2 =0.5
X1 =0, X2 =0.75
X2

1.5
X1 =0, X2 =1.0
X1 =0, X2 =1.25
1.0 X1 =0, X2 =1.5
X1 =0, X2 =1.75

0.5 X1 =0, X2 =2

0.0
0.0 0.5 1.0 1.5 2.0

X1

◼ Flow (iii)

For flow (iii) we have

100 ECH41MiscProblemsPart2_2024.nb

 x1 G  x2 G
= x2 , =- x1
α 2 α 2
To solve this system of equations, we can write the equations as
G
 x 0 x1
 1= 2
 
t x 2 - G
0 x2
2

We can use the same procedure as we did for flow (ii), except now the eigenvalues are imaginary
G G
λ1 = - , λ2 = 
2 2
The corresponding eigenvectors are
 -
ξ(1) =  , ξ(2) =  
1 1

Instead of doing the calculations by hand, we use Mathematica’s DSolve function to get a solution to
the ODEs directly

I n [ ] : = ODESol = First
G G
DSolvex '[α]  y[α], y '[α]  - x[α], x[0]  X1 , y[0]  X2 , {x[α], y[α]}, α
2 2
O u t [ ] =

{x[α]  1. Cos[0.25 α] X1 + 1. Sin[0.25 α] X2 , y[α]  -1. Sin[0.25 α] X1 + 1. Cos[0.25 α] X2 }

Using the same values for the initial conditions, we can plot the streamlines for G=2:

I n [ ] : = initialCond = {{X1  0, X2  0.25}, {X1  0, X2  0.5},

{X1  0, X2  0.75}, {X1  0, X2  1}, {X1  0, X2  1.25},
{X1  0, X2  1.5}, {X1  0, X2  1.75}, {X1  0, X2  2}};
 ECH41MiscProblemsPart2_2024.nb 101

I n [ ] : = ParametricPlot[Evaluate[{x[α], y[α]} /. ODESol /. G  2 /. initialCond],

{α, 0, 25}, AspectRatio  1, Frame  True, PlotStyle  Thick,
FrameLabel  {Style["X1 ", 16], Style["X2 ", 16]}, PlotLegends 
Placed[{"X1 =0, X2 =0.25", "X1 =0, X2 =0.5", "X1 =0, X2 =0.75", "X1 =0, X2 =1.0",
"X1 =0, X2 =1.25", "X1 =0, X2 =1.5", "X1 =0, X2 =1.75", "X1 =0, X2 =2"}, Right]]
O u t [ ] =

1
X1 =0, X2 =0.25
X1 =0, X2 =0.5
X1 =0, X2 =0.75
X2

0
X1 =0, X2 =1.0
X1 =0, X2 =1.25
X1 =0, X2 =1.5
-1
X1 =0, X2 =1.75
X1 =0, X2 =2

-2

-2 -1 0 1 2

X1

Thus we see that plane Couette flow is a combination of a purely straining flow and a rotational flow.
The linear combination of these two flows results in a flow with streamlines parallel to the x-axis

Problem 40 : Slot Coater

The figure below shows the essential components of the downstream section of a slot coater in the x-y
plane.

The design specifications for this slot coater is to coat a thin liquid film onto a moving web(substrate)
such that the downstream thickness of the liquid film is h0 =1/3 d, where d is the gap in the slot. You
may assume the flow is steady and two dimensional. Further the flow in the upstream slot is rectilin-
ear, v = u(y) e x .
102 ECH41MiscProblemsPart2_2024.nb

Tasks:
(i) Compute the velocity profile in the liquid film far downstream of the slot where the film thickness is
h0 .
(ii) Determine what ΔP/L=P1 - P2 /L, is such that h0 = 1 / 3 d
(iii) What is the volumetric flow rate per unit width of coated film that must be supplied to the slot.
(iv) Calculate the force per unit width required to move the web at a velocity U, over the length of the
slot L. Neglect entrance and exit effects due to the complicated flow at the slot entrance and exit..

Problem Solution
For this problem the flow in the slot and in the liquid film on the moving web is rectilinear. That is
v = u (y) ex

The governing equations for the flow field are then

∂p ∂2 u
0=- +μ + ρgx
∂x ∂ y2

∂p
0=- + ρgy
∂y

Since gravity is acting in the y direction only we have

gx = 0, gy = -ρ g

(i) Far downstream along the moving web, it follows that ∂p/∂x=0, and the boundary conditions on the
flow are
No Slip at y = 0 : u (0) = Uw
∂u
No Shear at y = h0 : =0
∂ y y=h0

The x- component of momentum is then

∂2 u
μ =0
∂ y2
BC1 : u = Uw , at y = 0

∂u
BC2 : = 0, at y = h0
∂y

Integrating the momentum equation gives

 ECH41MiscProblemsPart2_2024.nb 103

u (y) = C1 y + C2

Applying the BCs gives

u (y) = Uw

We have plug flow in the liquid film

(ii) In the slot the flow is also rectilinear and the governing equations are
∂p ∂2 u
0=- +μ
∂x ∂ y2

∂p
0=- -ρg
∂y

Integrating the last equation gives

p (x, y) = -ρgy + f (x)

Thus
∂p f
- =- ≡ G (x)
∂x x
which shows that ∂p/∂x is at most a function of x. Integrating the x-component of momentum gives
1 y2
u (y) = - G (x) + C1 y + C2
μ 2

The BCs are

BC1 : u (0) = Uw

BC2 : u (d) = 0

Solving for the constants C1 and C2 gives

1 y2 dy y
u (y) = - G (x) - + Uw 1 -
2μ 2 2 d

Now the flow rate per unit width in the slot is

d 1 d3 d3 d
Q =  u y = - G (x) - + Uw d -
0 2μ 6 4 2

G (x) d3 d
= + Uw
μ 24 2

Now far down stream the volumetric flow rate per unit width is
Q = Uw h0
104 ECH41MiscProblemsPart2_2024.nb

If we require h0 = d / 3, then
d G (x) d3 d
Uw = + Uw
3 μ 12 2

Simplifying gives
4 μ Uw ∂p
G (x) = - =-
d2 ∂x
P2 - P1 4 μ Uw P1 - P2 4 μ Uw
= ⟹ =-
L d2 L d2
Note this means that P2 > P1 .

(ii) The volumetric flow rate supplied to the slot must be

d
Q = Uw h0 = Uw
3
(iv) In the slot the shear stress at the moving web is
∂u -Uw dG d μ Uw
τxy = μ =μ + =G -
∂y y=0 d 4μ 4 d

Thus the force per unit width of web is

4 μ Uw d μ Uw 2 μ Uw
F=- - =-
2 4 d d
d

Problem 41 : Stress Tensor Analysis

The state of stress at a point P in a fluid is given by

20 10 -10
[T] = 10 30 0
-10 0 50

Tasks:
Determine the normal stress and the shearing stress on a surface intersecting the point P and parallel
to the plane 2x+y-3z=9 . Hint: recall that the force per unit area acting at a point in the fluid on a plane
with unit normal n is
t(n) = T · n, where T = Tij ei ej

Problem Solution
The unit normal to the plane F(x,y,z)=0 is
∇F 2 i +j -3 k
n= =
∇ F 14
 ECH41MiscProblemsPart2_2024.nb 105

Thus the traction (stress) vector at the point P is

1 20 10 -10 2 2 2 2
t(n) = T · n ⟹ 10 30 0 1 = 40 , 25 , -85 
14 -10 0 50 -3 7 7 7

Note that the magnitude of the stress vector is

t(n)  = t(n) .t(n)

2 2 2 2 2 2
= 40 , 25 , -85 .40 , 25 , -85 
7 7 7 7 7 7
= 30 3

Now the normal component of the traction vector is

2 2 2
t(n) · n = 40 , 25 , -85 .{2, 1, -3}  14
7 7 7
360
=
7
To compute the magnitude of the shear stress, note that the stress vector can be represented as
t(n) = TN n + TS eS

Thus the magnitude of the shear stress is

360 2 30 3
Ts = t(n) 2 - T2N = 2700 - = = 7.42307
7 7

Problem 42: Boundary Layer Analysis 1

An incompressible viscous fluid with density ρ flows pass a solid flat plate which has width b. The flow
initially has a uniform velocity U∞ before contacting the plate. The velocity profile at location x is
estimated to have a parabolic shape
2y y 2
u (x, y) = U∞  - 
δ (x) δ (x)

for y ≤ δ(x) and u = U∞ for y ≥ δ(x). The quantity δ(x) is the boundary layer thickness.

Tasks
(a) Write down the continuity equation and determine the upstream height from the plate h of a
streamline which has a height δ at the downstream location. Express your answer in terms of δ.

(b) Determine the force the fluid exerts on the plate over the distance x. Express your answer in terms
of ρ, U∞ , b, and δ . You may assume that the pressure imposed on the boundary layer is everywhere is
atmospheric pressure. A schematic of the flow geometry is shown below:
106 ECH41MiscProblemsPart2_2024.nb

Solution: Part(a)
We select a control volume that encompasses the flow in the boundary layer from x=0 and x = xbl as
shown in green in the figure below

Let the width of the plate be b. Then material balance over the control volume is given by
h δ
 U∞ b  y =  u (y) b  y
0 0

Recall we are told that the velocity profile at xbl is given by

2y y 2
u (x, y) = U∞  - 
δbl δbl

Evaluating the integral gives

δ 2y y 2
(U∞ b h) =  b U∞  -  y
0 δbl δbl

2 δbl
= b U∞
3
Then solving for h we find
 ECH41MiscProblemsPart2_2024.nb 107

2
h= δbl
3

Solution: Part(b)
To assess the force acting on the plate we consider the same control volume as in Part (a) and write
down the steady macroscopic momentum balance. The results is

 (ρ v) (v · n)  A =  ρ g  V +  tn  A
e  s

Here
tn = T · n

is the traction force acting on the surface of the plate. If we ignore the effects of gravity, then we see
that the total force acting on the plate is equal to

2
h
2
δ 2y y 2 2
 (ρ v) (v · n)  A = -  ρ U∞ b  y +  ρ U∞  -  b y
e 0 0 δbl δbl

δ 2y y 2 2
= -ρ U2∞ bh + ρ b U2∞   -  y
0 δbl δbl

δ 4 y2 y4 4 y3
= -ρ U2∞ bh + ρ b U2∞  + - y
0 δ2bl δ4bl δ3bl

8
= -ρ U2∞ b h + ρ b U2∞ δbl
15

2 8
= -ρ U2∞ b δbl + ρ b U2∞ δbl
3 15

2
=- b ρ U2∞ δbl
15
This is the force exerted by the fluid on the plate. Thus the force on the plate is
2
F= ρ U2∞ b δbl
15

Problem 43: Boundary Layer Analysis 2

Part(a):
Consider the flow of a Newtonian fluid in an arbitrary control volume (t), with surface area (t)
108 ECH41MiscProblemsPart2_2024.nb

moving with velocity w. The point equations that describe the fluid motion are continuity and linear
momentum:
∂ρ
+ ∇ ·ρ v = 0
∂t

∂v
ρ + ρ v · ∇v = ∇ · T + ρ g
∂t
where ρ is the fluid density, T is the stress tensor, and g is the gravitational body force. Start from Eq.
(a) and (B) and derive the macroscopic mass and momentum balances for the motion of a the fluid in
the control volume (t). Be sure to state all the vector identities that you use, and show all your
mathematical steps in your derivation. You need not prove the generalized Reynolds Transport
Theorem. Express your answer as

 ρ V +  ρ (v -w) · n  A = 0
t  (t)  (t)


 ρ v V +  ρ v (v - w) · n  A =  tn  A +  ρ g V
t  (t)  (t)  (t)  (t)

Part (b):
Consider a plate of width b and length L placed parallel to a steady uniform flow v = Uo e x ( see Figure
below). The pressure is assumed to be uniform and so it has no net force on the plate. The viscous
stresses acting on the the plate give rise to a finite drag force on the plate. Use the macroscopic
integral balances from Part (a) to find the drag D on the one side of the plate in terms of the flow
properties ρ, U0 , and δ and the plate dimensions L and b. Ignore the effects of gravity. Express your
answer in the form
δ
D = ρb  u (U0 - u)  y
0 x=L

For the solution see

ECH141 Lecture6c_BoundaryLayerFlow_August2018

Problem 44: Derivation of the Macroscopic Mass and Momentum

Balances

Problem Statement
Consider the flow of a fluid in an arbitrary control volume V (t), with surface area A (t) moving with
velocity w. The point equations that describe the fluid motion are continuity and linear momentum:
 ECH41MiscProblemsPart2_2024.nb 109

∂ρ
+ ∇ ·ρ v = 0
∂t

∂v
ρ + ρ v ·∇ v = ∇ ·T + ρ g
∂t

where ρ is the fluid density, T is the stress tensor, and g is the gravitational body force.

Task:
Start with the above equations and derive the macroscopic mass and linear momentum balances for
the motion of the fluid in the control volume  (t). Be sure to state all vector identities that you use,
and show all your mathematical steps in your derivation. You need not prove the generalized
Reynolds Transport Theorem or divergence theorem. Express your answer as


 ρ V +  ρ (v - w) · n  A = 0
t V (t) A (t)


 ρ v V +  ρ v (v - w) · n  A =  t (n)  A +  ρ g V
t V (t) A (t) A (t) V (t)

From this expression deduce a word statement for the axioms of conservation of mass and balance of
linear momentum.

Problem Solution:
In this solution we will make use of the generalized Reynolds Transport Theorem (GRTT)
 ∂
  V =  V +  ∇ · ( w )  V
t V (t) V (t) ∂t V (t)

where  may be a scalar, vector, or tensor. The last term on the right hand side can be further manipu-
lated using the divergence theorem.

 ∇ · ( w )  V =   w · n A
V (t) A (t)

Thus
 ∂
  V =  V +   w · n A
t V (t) V (t) ∂t A (t)

Now consider first the continuity equation. Integrating over the arbitrary moving control volume
V (t).
110 ECH41MiscProblemsPart2_2024.nb

∂ρ
  + ∇ · ρ v  V = 0
V (t) ∂t

We use the GRTT to take the time derivative outside the integral

 ρ V -  ∇ · (ρ w)  V +  ∇ · (ρ v)  V = 0
t V (t) V (t) V (t)

This can be expressed as


 ρ V +  ∇ · ρ ( v - w)  V = 0
t V (t) V (t)

And using the divergence theorem we get


 ρ V +  ρ ( v - w) · n  A = 0
t V (t) A (t)

Consider next the linear momentum equation

∂v
ρ + ρ v · ∇v = ∇ · T + ρ g
∂t
Integrating over an arbitrary moving control volume V (t) with surface velocity w gives
∂v
 ρ  V + ρ v · ∇v V =  ∇ · T  V + ρ g V
V (t) ∂t V (t) V (t) V (t)

Next we make use of the following vector identities

∂v ∂ ∂ρ
ρ = (ρ v) - v
∂t ∂t t

∇ ( · ρ vv) = ρ v · ∇ v + v ∇ · ρ v

Introducing these definitions into the integrated form of the linear momentum equation gives
∂ ∂ρ
  (ρ v) - v + ∇ ( · ρ v v) - v ∇ · ρ v  V =  ∇ · T  V + ρ g V
 (t) ∂t t  (t)  (t)

But from continuity equation we have

∂ρ
= -∇ · ρ v
∂t
Hence our integrated momentum equation simplifies to
∂
  (ρ v) + ∇ ( · ρ v v)   V =  ∇ · T  V + ρ g V
 (t) ∂t  (t)  (t)

Finally we use the Generalized Reynolds Transport theorem on the first term, and the divergence
theorem on the second and third terms to get

 ρ v V +  ρ v (v - w) · n  A =  t (n)  A +  ρ g V
t  (t)  (t)  (t)  (t)
 ECH41MiscProblemsPart2_2024.nb 111

where we have used the result

 ∇ · T V =  n · T A =  t (n)  A
 (t)  (t)  (t)

where the surface velocity of the control volume is w. For a moving control volume (CV) we can state
the axiom for conservation of mass as
Time rate of change of mass in the CV =
net influx of mass through the surface of the CV

The axiom for balance of linear momentum is

Time rate of change of linear momentum in the CV =
net influx of linear momentum through the surface of the CV
+the net surface forces acting on teh surface of the CV
+ the net body forces acting on the CV

Problem 45: Flow between Eccentric Rotating Disks

Problem Statement
Consider flow between eccentric rotating disks. Both disks rotate with constant angular velocity Ω.
The axes of rotation of the disks are separated by a distance α, and the gap between the disks is β. The
spatial coordinates of fluid particles in such a flow are given by
x1 = X1 Cos (Ω t) - X2 sin (Ω t)
x2 = X1 sin (Ω t) + X2 cos (Ω t) + γ X3
x3 = X3

where γ=α/β, and Xi are the material coordinates of the fluid particles in the undeformed state (i.e. at
t=0). A schematic of the flow geometry is shown below.

Task (i):
(i) Determine the velocity field (in terms of the spatial coordinates xi ) for the motion described above.
112 ECH41MiscProblemsPart2_2024.nb

(ii) If the temperature field between the rotating disks is given by

x1 x2
θ (x, t) = +
(1 + t) (1 + t)

determine the time rate of change of temperature of a fluid particle in the flow field.

Solution

Part(i)
We are given the spatial coordinates of the motion:
x1 = X1 Cos (Ω t) - X2 sin (Ω t)
x2 = X1 sin (Ω t) + X2 cos (Ω t) + γ X3
x3 = X3

Now the velocity of a fluid particle is given by

x
=v
t X

We can evaluate the velocity components from the spatial coordinates:

 x1
Hence ≡ v1 = -X1 Ω Sin (Ω t) - Ω X2 Cos (Ω t)
t X

 x2
≡ v2 = Ω X1 Cos (Ω t) - Ω X2 Sin (Ω t)
t X

 x3
≡ v3 = 0
t X

This is a material description of the problem. To determine the spatial description we need to write
the velocity field in terms of xi . From inspection of the spatial coordinates we can deduce that
-Ω X1 Sin (Ω t) - Ω X2 Cos (Ω t)
= -Ωx2 + Ω γ x3 Ω X1 Cos (Ω t) - Ω X2 Sin (Ω t)
= Ω x1

Thus
v1 = -Ωx2 + Ω γx3 , v2 = Ωx1 , v3 = 0

or
v = -Ω (x2 - γx3 ) e1 + Ω x1 e2
 ECH41MiscProblemsPart2_2024.nb 113

Part(ii)
The time rate of change of temperature of a fluid particle is given by
θ ∂θ
≡ + v · ∇θ
t X ∂t
Thus we can compute this quantity either in terms of spatial coordinates or material coordinates. The
former calculation is more direct. Since we are given
θ (x, t) = x1 / (1 + t) + x2 / (1 + t)

then
∂θ x1 x2 (x1 + x2 )
=- - =-
∂t x (1 + t)2 (1 + t)2 (1 + t)2

∂ ∂ 1 1
∇ θ = e1 +e2 θ= e1 + e2
∂ x1 ∂ x2 (1 + t) (1 + t)

We show from Part (i) that

v = -Ω (x2 - γx3 ) e1 + Ω x1 e2

Hence we can compute

Ω
v · ∇θ = (x3 γ - x2 + x1 )
(1 + t)

The time rate of change of temperature of a fluid particle is given by

θ (x1 + x2 ) Ω
=- + (x3 γ - x2 + x1 )
t X (1 + t)2 (1 + t)

Problem 46: Flux Across a Specified Plane

Problem Statement
In this problem you are required to determine the flux across a specified plane. Let the velocity field
be given by
v = 12 y i + 2 x j - 3 k

Then determine the total flux J due to v across the plane

x + 2 y + 8 z, for x ≥ 0, z ≥ 0
114 ECH41MiscProblemsPart2_2024.nb

Solution:
The unit normal to any surface f(x,y,z)=0 is given by
n  ∇f / ∇f

For our problem we are given the equation for the surface f(x,y,z)
f (x, y, z) = x + 2 y + 8 z - 12

Given the equation for the surface we can compute the unit normal to that surface. We proceed as
follows:

First we compute the gradient of the surface

∇ f = i + 2 j + 8 k, ∇ = ∇f · ∇f = 69

so that the unit normal is

1
n= (i + 2 j + 8 k)
69

The flux J across the surface with unit normal n is given by

x y
J =  v · n S =  v · n
S Γ n·k

where we have used the result from the projected area theorem
8
n·k  s =  x  y, where n·k =
69

and Γ is the projected area of the surface onto the x-y plane. Thus the limits of integration are
0 ≤ y ≤ 6 - x / 2, and 0 ≤ x ≤ 12

From our definitions we have

12 y + 4 x - 24
v·n =
69

Evaluating the integral for the flux gives

x y 12 6-x/2 12 y + 4 x - 24 x y
J =  v·n = 
Γ n·k 0 0 69 8 69

12 6-x/2 3 x
=  y+ - 3  x  y = 72
0 0 2 2
 ECH41MiscProblemsPart2_2024.nb 115

Problem 47: Fluid flows between two porous, parallel flat plates

Problem Statement
An incompressible fluid flows between two porous, parallel flat plates as shown in the Figure below.
An identical fluid is injected at a constant speed V through the bottom plate and simultaneously
extracted from the upper plate at the same velocity. There is no gravity force in x and y directions
g x = g y =0). Assume the flow to be steady, fully-developed, 2D, and the pressure gradient in the x-
direction to be a constant ,i.e., ∂p/∂x =constant.

A schematic of the flow geometry is shown below:

Note: The solution to this problem is based on the following paper:

Hafeez, H and Ndikilar, C. E., Flow of a viscous fluid between two parallel porous plates with bottom
injection and top suction, C. E., Progress in Physics, 10, pp 49-51, (2014).

Tasks:
(a) Write the continuity equation and show that the y-component of velocity is constant at v = 𝑉.

(b) Simplify the x-momentum equation and find the appropriate differential equation for the x- veloc-
ity component u.

(c) To solve the inhomogeneous differential equation, first solve for the homogeneous part and then
find a particular solution.

(d) Apply boundary conditions and find C1 and C2

(e) Plot the velocity profile in the channel as a function of the Reynolds number for the flow:
e= V h/ν for the flow.
116 ECH41MiscProblemsPart2_2024.nb

Solution: Part(a)
The continuity equation of a steady incompressible fluid in 2-D is given by
∂u ∂v
∇ ·v = 0 ⟹ + = 0, where v = u ex + v ey
∂x ∂y

The key assumption in this problem is that the v- component of flow is a constant value. That is
v = V = constant

The assumption that the flow is fully developed, means that the velocity does not change in the flow
direction . Hence from the continuity equation we can conclude that the component of flow in the
flow direction does not change. Hence
∂u
=0 ⟹ u (y) only !
∂x
Thus a velocity field in our porous channel is given by
v (x, t) = u (y) ex + V ey

Note that at the walls of the channel

u (0) = u (h) = 0, v (0) = V, v (h) = V

Solution: Part(b)
As noted in Part (a) the velocity field is given by
v (x, y) = u (y) ex + V ey

Further we are told that the density of the fluid is constant hence linear momentum equation reduce
to a single equation
u ∂p 2 u
x - component : ρV =- +μ , -h < y < h
y ∂x  y2

BC1 : u (h) = 0

BC2 : u (-h) = 0

THus we need to solve this problem to find the the velocity u(y). It is worth noting that the y- compo-
nent of momentum reduces to
∂p
y - component : 0= ⟹ p = p (x) only
∂y
 ECH41MiscProblemsPart2_2024.nb 117

Solution: Part (c)

It is convenient to write our equation in the following form
2 u V u 1 ∂p
- =- Eq. (A)
 y2 ν y μ ∂x

This is a linear inhomogeneous equation for u(y). Let us examine the homogeneous part first
 2 uh V  uh
- = 0,
 y2 ν y

This is an equation with constant coefficients , so we can look for solutions of the form
uh (y) ∼ Exp[p y]

This p must satisfy the following indicial equation:

V
p2 - p=0
ν
Hence the roots are
V
p1 = 0, p2 =
ν
and the solution is of the form
uh (y) = A Exp[p1 y] + B Exp[p2 y]

where A and B are constants. Thus the solution is

uh (y) = A + B Exp[Vy / ν]

To find a particular integral we let

uPI (y) = ay2 + by + c

It follows then
u 2 u
= 2 a y + b, =2a
y  y2

Substituting these quantities into our ODE we get

Vb 2aV 1 p
2a- - y=
ν ν μ x

Note that the RHS of the above equation does not depend on y, thus we can conclude
ν 1 p
a = 0, b=-
V μ x

Thus our particular integral is

118 ECH41MiscProblemsPart2_2024.nb

ν 1 p
uPI (y) = - y+C
V μ x

Thus our general solution is

u (y) = uH (y) + uPI (y)

ν 1 p
= D + B Exp[V y / ν] - y, where D = A + C
V μ x

Solution Part (d):

We showed in the previous part that the general solution is given by
ν 1 p
u (y) = D + B Exp[V y / ν] - y
V μ x

Our task now is to determine the coefficients D and B

Now this equation must satisfy the no-slip BC at y=± h, i.e, u(h)=u(-h)=0. Working through the
algebra we find
ν h p
Exp V h  ν h p
ν h p
V μ x ν V μ x
D=- + , B=
Sinh V h  V μ x Sinh V h 
ν ν

Thus our velocity profile becomes

ν h p
Exp V h  ν h p
ν h p
ν y p
V μ x ν V μ x
u (y) = - + + -
Sinh V h  V μ x Sinh V h  V μ x
ν ν

It is convenient to define the wall Reynolds number e:

Vh
e =
ν
and then noting we can write
e V h ν
= , ⟹ =
h ν e V
Thus our velocity profile can be expressed as
ν h p
Exp V h  ν h p
ν h p
ν y p
V μ x ν V μ x
u (y) = - + + -
Sinh V h  V μ x Sinh V h  V μ x
ν ν

h2 1  p y Exp[e] - Exp[e y / h]
=-  -1+ 
e μ  x h Sinh[e]

The final working solution is

 ECH41MiscProblemsPart2_2024.nb 119

u (y) 2 y Exp[e] - Exp[e y / h]

=  -1+ 
umax e h Sinh[e]

where
h2 p
umax = -
2μ x

where umax is the center line velocity for impervious Poiseuille flow.

For very small e ( or small vertical velocity), then the last term in the brackets [**] can be expanded
in a power series, where Sinh[e]≈e.

WE will use Mathematica’s Series function

Exp[e] - Exp[e y / h]
I n [ ] : = Series , {e, 0, 3}
Sinh[e]
O u t [ ] =

y 1 y2 h2 y - y3  e2 -h4 + 2 h2 y2 - y4  e3

1- + - e + + + O[e]4
h 2 2 h2 6 h3 24 h4

Thus we get
u (y)
=
umax
2 y y 1 y2 h2 y - y3  e2 -h4 + 2 h2 y2 - y4  e3
 - 1 + 1 - + - e + + + O[e]4
e h h 2 2 h2 6 h3 24 h4

y2
= 1- +  (e)
h2

which shows that the Poiseuille solution is recovered.

In the case of very large e numbers, we can write

u (y) 2 y Exp[e] - Exp[e y / h]
=  -1+ 
umax e h Sinh[e]

2 y Exp[e] - Exp[e y / h]
=  -1+2 
e h Exp[e] - Exp[-e]

2 y 1 - Exp[-e (1 - y / h)]
=  -1+2 
e h 1 - Exp[-2 e]

2 y
=  - 1 + 2
e h
120 ECH41MiscProblemsPart2_2024.nb

2 y
= 1 + 
e h
which represents a straight line distribution. Of course in reality the no- slip condition at y=h must be
satisfied. Thus the velocity must drop to zero as the upper wall is approached.

Here is a plot of the velocity profile:

2 Exp[Re] - Exp[Re ξ]
I n [ ] : = velocity1[Re_, ξ_] := ξ-1+
Re Sinh[Re]

I n [ ] : = Plot[{velocity1[.01, ξ], velocity1[1, ξ], velocity1[3, ξ],

velocity1[5, ξ], velocity1[10, ξ], velocity1[20, ξ]}, {ξ, -1, 1},
Frame  True, FrameLabel  {Style["ξ= y/h", 16], Style["U/Umx ", 16]},
PlotLegends  {"Re=0.01", "Re=1", "Re=3", "Re=5", "e=10" , "e=20"} ]
O u t [ ] =

1.0

0.8
Re=0.01
0.6 Re=1
U/Umx

Re=3
0.4
Re=5
0.2 e=10
e=20
0.0
-1.0 -0.5 0.0 0.5 1.0

ξ= y/h

Note for large e, the velocity profile approaches a straight line with a sudden drop-off near the wall
as e becomes large, so that the no-slip BC is satisfied.

Problem 48: Analysis of a Centrifugal Pump

A schematic of our pump is shown in the following figure. The fluid enters the vertical tube of radius r1
and exits the casing with radius r2. The impeller is also shown in the figure. Note that its vanes are
oriented radially. Thus the fluid velocity leaving the pump is tangential to the vanes and is denoted as
v2 . The tangential velocity of the vanes at r2 is u2 . Hence the relative velocity of the fluid leaving the
pump at r2 is vrel . As shown in the figure the fluid entering the pump is along the axis of rotation.
 ECH41MiscProblemsPart2_2024.nb 121

Solution
The transfer of energy to the fluid passing through the centrifugal pump can be accounted for through
the conservation of angular momentum. For steady flow the macroscopic balance of angular momen-
tum reduces to
 
 Maxis =  (r  V) mout - (r  V) min

This is known as Euler's formula for centrifugal pumps and simply states that the sum of moments
about the axis of rotation is equal to the net change in angular momentum of the fluid entering and
leaving the control volume. Let τ be the torque about the axis and for our system the balance of
angular momentum becomes
 
τ =  (r  V) mout - (r  V) min

The angular momentum of the entering fluid is min r1 vw1 where vw1 is the tangential velocity of the

fluid ar r1 . Likewise the angular momentum of the fluid leaving the pump is mout r2 vw2 . By conserva-
  
tion of mass min = mout = m. Thus our balance of angular momentum becomes

τ = m (r2 vw2 - r1 vw1 )

If the rotor is spinning at angular velocity ω, then the power (work done per unit time) is
P= τω J s-1 or Watts

Recall that the tangential velocity of the rotor at r1 and r2 are

u1 = ω r1 , u2 = ω r2

Hence the power transmitted to the fluid is

View publication stats

122 ECH41MiscProblemsPart2_2024.nb

 
P = m (r2 ω vw2 - r1 ω vw1 ) = m (u2 vw2 - u1 vw1 )

Then the rate of work done per unit mass of fluid is:
P
E =  = (u2 vw2 - u1 vw1 )
m
This is also equal to g H, where H is the Euler head. Thus,
Δp
E=gH= = (u2 vw2 - u1 vw1 )
g

Hence by computing the RHS of of the above equation we can determine the pressure head developed
by the pump. In our case the fluid enters the pump in the axial direction so there is no contribution to
angular momentum as vw1 = 0 and we are left with
Δp
E=gH= = u2 vw2
g

Since vw2 is radially outward (tangential to the blades of the impeller) we can write
 
Q Q
vw2 = =
A 2 πr2 h

where h is the height of the impeller. Hence if we know the power imparted to the impeller and the
angular velocity of the impeller, we can determine the pressure drop across the pump and hence
determine the volumetric throughput of the pump.