Tutorial 10: Flow Analysis using Differential Methods (II)

[Solutions]
MECH 2210 - Fluid Mechanics (L2)
Spring 23-24

Justin Sun
[email protected]

22 April 2024

Contents
1 Concept Recap 2
1.1 Equations of Motion*** . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.1 Inviscid Flow Assumption - Euler’s Equations of Motion . . . . . . . . . . 2
1.1.2 Euler’s Equations of Motion and the Bernoulli Equation . . . . . . . . . . 3
1.2 Navier-Stokes Equations*** . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Problem Solving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 Exercises 5
2.1 Question 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1.1 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Question 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.1 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3 Question 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3.1 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3.2 Note on our solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

3 Additional Problems 13
3.1 Question 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.1.1 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2 Question 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.2.1 Solution - Part (a) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2.2 Solution - Part (b) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.3 Question 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.3.1 Solution - Part(a): Incompressibility . . . . . . . . . . . . . . . . . . . . . 17
3.3.2 Solution - Part(b): Pressure Gradient - ∇p . . . . . . . . . . . . . . . . . . 18

4 Appendix 21
4.1 Note on the Steady Flow Assumption . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.2 Further on Derivation Techniques for Vector Calculus . . . . . . . . . . . . . . . . 23

1
1 Concept Recap
1.1 Equations of Motion***
The Equations of Motion for a fluid element, derived by considering the forces acting on a fluid
element as shown below:

gives us the following equations:

 
∂σxx ∂τyx ∂τzx ∂u ∂u ∂u ∂u
x: + + + ρgx = ρ +u +v +w
∂x ∂y ∂z ∂t ∂x ∂y ∂z
 
∂τxy ∂σyy ∂τzy ∂v ∂v ∂v ∂v
y: + + + ρgy = ρ +u +v +w
∂y ∂y ∂z ∂t ∂x ∂y ∂z (1)
 
∂τxz ∂τyz ∂σzz ∂w ∂w ∂w ∂w
z: + + + ρgz = ρ +u +v +w
∂y ∂y ∂z ∂t ∂x ∂y ∂z
| {z } | {z }
F in F = ma ma in F = ma

The equations above essentially are Newton’s Second Law, better known as F = ma. The left
side of the equations (the F part of Newton’s Second Law) describes the forces acting on the fluid
element, such as normal stresses (surface force), σ, which is usually our pressure terms, shear
stresses (surface force), τ, and gravitational forces (body force), g. The right side of the equations
(ma part) is the product of the density term and material derivative of the velocity (acceleration),
which describes the motion of the fluid element.

1.1.1 Inviscid Flow Assumption - Euler’s Equations of Motion

The inviscid assumption means that the viscosity is negligible, i.e., viscous forces, or in this case,
the shear stresses, τ are negligible. Therefore, we can set all terms with τ in Eq. (1) to 0, leading

2
to the following set of equations:
 
∂σxx ∂u ∂u ∂u ∂u
x: + ρgx = ρ +u +v +w
∂x ∂t ∂x ∂y ∂z
 
∂σyy ∂v ∂v ∂v ∂v
y: + ρgy = ρ +u +v +w
∂y ∂t ∂x ∂y ∂z
 
∂σzz ∂w ∂w ∂w ∂w
z: + ρgz = ρ +u +v +w
∂z ∂t ∂x ∂y ∂z
While we define the normal stress outwards from the fluid element, we usually define pressure as
inwards (into the fluid element), and since the pressure at a point is the same in all directions,
we must have:
σxx = σyy = σzz = −p
where the negative sign on the pressure term p indicates the opposite direction to the normal
stress. Substituting this into the equations above, we have the following equations:
 
∂p ∂u ∂u ∂u ∂u
x: ρgx − =ρ +u +v +w
∂x ∂t ∂x ∂y ∂z
 
∂p ∂v ∂v ∂v ∂v
y: ρgy − =ρ +u +v +w (2)
∂y ∂t ∂x ∂y ∂z
 
∂p ∂w ∂w ∂w ∂w
z: ρgz − =ρ +u +v +w
∂z ∂t ∂x ∂y ∂z

or equivalently in its vector form:

 
∂V
ρg − ∇p = ρ + (V · ∇)V (3)
∂t

which is named the Euler’s Equations of Motion when we assume inviscid flow.

1.1.2 Euler’s Equations of Motion and the Bernoulli Equation

Note that from the Euler’s Equations of Motion, assuming irrotational flow, i.e., ∇ × V = 0,
we can apply the Bernoulli Equation:
1
p + ρV 2 + γz = c
2
between any two points in a flow field, assuming the other three assumptions:
(a) Steady Flow
(b) Incompressible Flow
(c) Inviscid Flow
still holds. Effectively, we have replaced the along a streamline assumption with a stronger
assumption of irrotational flow, such that we can now apply the Bernoulli Equation between
any two points in a flow field.

3
1.2 Navier-Stokes Equations***
Now suppose we cannot use the inviscid flow assumption, i.e., the fluid is viscous. In that scenario,
our stress terms can be represented as follows:
 
∂u ∂u ∂v
σxx = −p + 2µ τxy = τyx = µ +
∂x ∂y ∂x
 
∂v ∂v ∂w
σyy = −p + 2µ τyz = τzy = µ +
∂x ∂z ∂y
 
∂w ∂w ∂u
σzz = −p + 2µ τzx = τxz = µ +
∂x ∂x ∂z

Substituting the above into our original Equations of Motions Eq. (1), we have the following:

 2
∂ u ∂ 2u ∂ 2u
  
∂u ∂u ∂u ∂u ∂p
x: ρ +u +v +w =− + ρgx + µ + +
∂t ∂x ∂y ∂z ∂x ∂x2 ∂y 2 ∂z 2
 2
∂ v ∂ 2v ∂ 2v
  
∂v ∂v ∂v ∂v ∂p
y: ρ +u +v +w =− + ρgy + µ + + (4)
∂t ∂x ∂y ∂z ∂y ∂x2 ∂y 2 ∂z 2
 2
∂ w ∂ 2w ∂ 2w
  
∂w ∂w ∂w ∂w ∂p
z: ρ +u +v +w = − + ρgz + µ + +
∂t ∂x ∂y ∂z ∂z ∂x2 ∂y 2 ∂z 2

which are the famed Navier-Stokes Equations (technically we also have to include the conti-
nuity equation described in Tutorial 8 notes).

The Navier-Stokes (NS) Equations are the preeminent equations of study throughout the study
of Fluid Mechanics and this branch of research is effectively dedicated to the study of these
equations. They are a set of partial differential equations, where we have been able to find
partial solutions in different numerical domains or under some particular assumptions, but not
a complete, general solution has been found across the whole numerical domain. It is one of
the seven Millennium problems (where solving it scores you a cool USD 1 million) and is more
precisely named the “Navier-Stokes Existence and Smoothness Problem” (see more at
https://www.claymath.org/millennium/navier-stokes-equation/).

1.3 Problem Solving

The above equations are all extremely complex and are effectively impossible to solve manually.
In order to make our lives easier, we again utilise various assumptions to attempt to remove
terms from our partial differential equation, i.e., if our velocity vector has no component in the y
direction, we can remove any terms that include v since their value must be zero.

One key equation to combine with Eq. (5) is the Continuity Equation. It is always good to
include it within your problem solving steps and is nearly always useful!

Once we have effectively simplified our equation, we can then utilise boundary conditions, some-
times hinted in the question and integrate accordingly to find the desired quantity or function. If
your simplified equation is not integrable, there may be further simplifications you need to find.

4
2 Exercises
2.1 Question 1
A layer of water flows steadily down an infinite, inclined plane, as shown below. There is no
velocity in the z direction. The thickness of the layer is uniform and equal to h in the y direction.
Determine, by means of the Navier-Stokes Equations, the relationship between h and the flow
rate per unit width.

2.1.1 Solution
Given there is no velocity in the z-direction, we only need to consider the Navier-Stokes Equations
in the x- and y-directions as follows:
 2
∂ u ∂ 2u ∂ 2u
  
∂u ∂u ∂u ∂u ∂p
x: ρ +u +v +w =− + ρgx + µ + +
∂t ∂x ∂y ∂z ∂x ∂x2 ∂y 2 ∂z 2
 2
∂ v ∂ 2v ∂ 2v
  
∂v ∂v ∂v ∂v ∂p
y: ρ +u +v +w =− + ρgy + µ + +
∂t ∂x ∂y ∂z ∂y ∂x2 ∂y 2 ∂z 2
Notice that on the free surface of the fluid along the x-axis, the gauge pressure will always be
zero, and therefore the pressure gradient along the x-axis on the free surface (at y = h) will be 0,
∂p
i.e., ∂x = 0. - Assumption 1

Furthermore, note that the gravitational acceleration is in the vertically downward direction where
it is not aligned with the y-axis, therefore, decomposing the gravitational acceleration g into the
corresponding x and y directions, we have the following:
gx = g sin θ
gy = g cos θ
where θ is the slope angle. - Assumption 2

From the figure, notice that we only have a velocity component in the x-direction, u. Therefore,
we have v = w = 0. - Assumption 3

The question also states that the flow is steady, i.e., there is no variation in time. - Assumption 4

Lastly, from the continuity equation:

∂u ∂v  ∂w

+ + =0
∂x 
∂y ∂z
| {z }
A.3: v=w=0

5
 ∂u
=0
∂x
Notice that the velocity component is a function of y only, i.e., u = u(y), since it is constant in
the x- and z-direction (for the z-direction, notice that the figure shows a cross-section of the flow
and it repeats infinitely in the z-direction). - Assumption 5.

Therefore, we simplify the Navier-Stokes Equations according to our four assumptions. Since
we have no velocity component in the y-direction, we only need to consider the Navier-Stokes
Equations in the x-direction. We simplify as follows:
   
2
 ∂u
∂u
∂u ∂u

∂p
 ∂u ∂ 2u ∂ 2u

ρ

+u
+ v + w  = −

+ρgx + µ  2 + 2 + 2 

∂t ∂x

∂y
 ∂z ∂x

∂x ∂y ∂z

|{z} |{z} | {z } |{z} |{z} |{z}
A.4 A.5 A.3 A.1 A.5 A.5

∂ 2u
0 = ρ g sin θ +µ 2
| {z } ∂y
A.2
2
∂ u ρg sin θ
2
=−
∂y µ

We now have a 2nd -order differential equation to solve by integrating twice with respect to y as
follows:
Z
∂u ρg sin θ
= − ∂y
∂y µ
ρg sin θ
=− y + c1
µ
Z
ρg sin θ
u= − y + c1
µ
ρg sin θ 2
u=− y + c1 y + c2
2µ
Notice that we can set c1 and c2 as a numerical constant rather than a residual function of x
since we have already assumed that u is only a function of y in A.5. Now let us find c1 and c2 by
applying boundary conditions. Since we have 2 unknowns, we must apply two sets of boundary
conditions.

For our first boundary condition (B.C.1), at y = 0, i.e., on the plane, by the stick boundary
condition, we have u = 0. Therefore, applying the boundary conditions, we get the following:
ρg sin θ 2
u=0=− (0) + c1 (0) + c2
2µ
c2 = 0

Our next boundary condition (B.C.2) is much trickier. Recall our shear stress term where we
have:
du
τ =µ
y dy y

6
At the free surface, i.e., y = h, this shear stress, τ , can be assumed to be zero. Therefore, we
y=h
have the following:
du
τ =µ
y=h dy y=h
 
d ρg sin θ 2
=µ − y + c1 y
dy 2µ y=h
 
ρg sin θ
=µ − y + c1
µ y=h

0 = −ρg sin θh + c1 µ
ρg sin θh
c1 =
µ

which gives us the full function for u = u(y) as follows:

ρg sin θ 2 ρg sin θh
u=− y + y
2µ µ

For the flow rate, q, we can compute it by integrating the velocity over the thickness of the water
layer as follows:
Z h
q= udy
0
ρg sin θ 3 ρg sin θh 2 h
=− y + y
6µ 2µ 0
ρg sin θ 3 ρg sin θ 3
=− h + h
6µ 2µ
ρgh3 sin θ
q=
3µ

2 key assumptions were used here:

1. Gauge pressure at any liquid interface with air (liquid surface) can be treated as 0.

2. Shear stress, τ, at any liquid interface with air (not a wall) can be treated as 0.

7
2.2 Question 2
The x-component velocity, u, is shown below in the figure and has the function:
u(x) = −20(1 − e−x )
Approximate the y-component velocity, v at the point (2, 0.2) in the xy-plane, assuming incom-
pressible flow.

Hint: Use the incompressibility assumption and the continuity equation.

2.2.1 Solution
Utilising the incompressibility assumption, we can simplify our continuity equation into the fol-
lowing:
∂u ∂v ∂w
∇·V = + + =0
∂x ∂y ∂z
∂u ∂v ∂w
+ +  =0
∂x ∂y |{z}
∂z
w=0

Notice that since we are only considering a 2D flow, we have w = 0. Let us now substitute our
known function u(x) into the above continuity equation:
∂u ∂v
+ =0
∂x ∂y
∂v
−20e−x + =0
∂y
∂v
= 20e−x
∂y
Given the above partial differential equation, we can integrate as follows:
Z
v = 20e−x ∂y

v = −20ye−x + c(x)
where c(x) is our integration function residual. To find the unknown function c(x), we apply
particular boundary conditions for v. Notice that from the figure above, the y-component velocity,
v = 0 on the x-axis, i.e., when y = 0. Therefore, we have the following:
v = 0 = −20ye−x + c(x)

8
 0 = −20(0)e−x + c(x)
c(x) = 0

While the above is only valid at y = 0, since we want to find the v at y = 0.2, which is relatively
close to y = 0, we approximate the integration residual function for y = 0.2 using c(x) = 0.
Finally, we find v at (2, 0.2) as follows:

v = −20ye−x
= −20(0.2)e−2
v = −0.54134113294
v = −0.541 (3.s.f.)

9
2.3 Question 3
An incompressible, viscous fluid is placed between horizontal, infinite, parallel plates as shown
below. The two plates move in opposite directions with constant velocities, U1 and U2 , as shown.
The pressure gradient in the x direction is zero, and the only body force is due to the fluid weight.
Use the Navier–Stokes equations to derive an expression for the velocity distribution between the
plates. Assume laminar flow.

2.3.1 Solution
Given it is a laminar flow, we can assume that we only have a velocity component in the x-
direction. Therefore, we only need to analyse the x-component of the Navier-Stokes Equations:
 2
∂ u ∂ 2u ∂ 2u
  
∂u ∂u ∂u ∂u ∂p
x: ρ +u +v +w =− + ρgx + µ + +
∂t ∂x ∂y ∂z ∂x ∂x2 ∂y 2 ∂z 2

We then make our assumptions as follows.

Let us first assume that the flow is steady, i.e., there is no variation in time, and that the flow is
fully developed. - Assumption 1

Note that gravity only acts in the y-direction, therefore we have gx = 0. - Assumption 2

As we have already stated, we assume that there is only an x-component velocity, therefore we
must have v = w = 0. - Assumption 3

Given A.1, let us further assume that given the fully developed flow, there is no pressure gradient
∂p
in the x direction since there is no variation in fluid property in the x-direction, i.e., ∂x = 0. -
Assumption 4

Lastly, from the continuity equation:

∂u ∂v  ∂w

+ + =0
∂x 
∂y ∂z
| {z }
A.3: v=w=0
∂u
=0
∂x

10
Notice that the velocity component is a function of y only, i.e., u = u(y). - Assumption 5.

We then simplify our Navier-Stokes Equation as follows:

   

 ∂u
∂u
∂u ∂u
 ∂p
 ∂ 2u

 ∂ 2u ∂ 2u
ρ 
+u
+ |{z}
v + w
= −
+ρ gx +µ   + +  
∂x2 ∂y 2 |{z}
∂z 2 
  

∂t ∂x ∂y |{z} |{z}

∂z ∂x
  


|{z} 
|{z} |{z} A.3 A.3 |{z} A.2
|{z}
A.3 A.5 A.5 A.4 A.1,A.5 A.5
2
∂ u
0=µ
∂y 2
2
∂ u
0= 2
∂y
since the viscosity µ ̸= 0. We now integrate to find the function of u as follows:
Z Z
u= 0 dy
Z
= c1 dy

u(y) = c1 y + c2

where c1 , c2 are constants. With two unknown constants, we find two boundary conditions.

Our first boundary condition (B.C.1), by the stick boundary condition, must have that the
fluid at the surface boundary must be moving at the same velocity as the surface itself. Therefore,
at the top surface, i.e., y = b, we have u = U1 . Similarly, our second boundary condition (B.C.2),
again by the stick boundary condition, at the bottom surface, i.e., y = 0, we have u = −U2 .
Note the negative sign due to the opposite direction velocities. Therefore, by B.C.1, we have:

u(b) = U1 = c1 (b) + c2

and by B.C.2, we have:

u(0) = −U2 = c2

It is clear that we have c2 = −U2 , and therefore it follows that we have:

U1 = c1 (b) − U2
U1 + U2
c1 =
b
and finally, substituting our expressions for c1 and c2 , we get the velocity distribution as follows:

U1 + U2
u= y − U2
b

11
2.3.2 Note on our solution
Recall the case in Chapter 1 where we assume a linear velocity distribution for a Newtonian fluid
in a parallel plate setting, where the top plate is moving and the bottom plate is fixed, as shown
in the figure below:

Our solution verifies this linear velocity distribution. We can recreate this setting by letting
U2 = 0, resulting in the following velocity distribution:
U1
u= y
b
This previous assumption that we made when computing the shear stresses for Newtonian fluids
can be verified by making the appropriate assumptions with the Navier-Stokes Equations.

12
3 Additional Problems
3.1 Question 4
The velocity component along the x-axis is given as follows:
40
u(x) = 10 −
x2
What is the radius of the cylinder for the figure shown below? Approximate the y-component
velocity, v, at (−3, 0.1) assuming an incompressible flow.

3.1.1 Solution
Note that on the left-most point of the cylinder shown, we must have a stagnation point where
the velocity is zero, which also corresponds to a distance of radius R away from (0, 0) (to the left
side) of the coordinate system. Therefore, we can find the radius R using the velocity function
along the x-axis, where y = 0, as follows:
40
0 = 10 −
(−R)2
40 = 10R2
R=2

Therefore, the velocity at (−3, 0.1) is not zero since it is not at the stagnation point nor is it in
the cylinder. From the incompressible flow assumption, our continuity equation is as follows:
∂u ∂v ∂w
∇·V = + + =0
∂x ∂y ∂z
∂u ∂v ∂w
+ +  =0
∂x ∂y |{z}
∂z
w=0

Notice that since we are only considering a 2D flow, we have w = 0. Let us now substitute our
known function u(x) into the above continuity equation:

∂u ∂v
+ =0
∂x ∂y

13
 80 ∂v
+ =0
x3 ∂y
∂v 80
=− 3
∂y x
Integrating, we find the following:
Z
80
v= − ∂y
x3
80
v=− y + c(x)
x3
where c(x) is our integration function residual. To find the unknown function c(x), we apply
particular boundary conditions for v. Notice that from the figure above, the y-component velocity,
v = 0 on the x-axis, i.e., when y = 0. Therefore, we have the following:
80
v=0=− y + c(x)
x3
80
0 = − 3 (0) + c(x)
x
c(x) = 0

Therefore, we get the final expression for v as follows:

80
v=− y
x3

While the above is only valid at y = 0, since we want to find the v at y = 0.1, which is relatively
close to y = 0, we approximate the integration residual function for y = 0.1 using c(x) = 0.
Finally, we find v at (-3, 0.1) as follows:
80
v=− y
x3
80
=− (0.1)
(−3)3
v = 0.2̇96̇
v = 0.296 (3.s.f.)

14
3.2 Question 5
When water discharges from a tank through an opening at the bottom, a vortex may form with
a curved surface profile as shown below:

Assume that the velocity distribution in the vortex is the same as that for a free vortex. At the
same time the water is being discharged from the tank at point A, it is desired to discharge a
small quantity of water through pipe B. As the discharge through A is increased, the strength of
the vortex, as indicated by its circulation, Γ, is increased.

(a) Find an expression for relating the surface shape, i.e., find a function that relates zs to r as
indicated in the figure below.

(b) Determine the maximum strength that the vortex can have in order that no air is sucked in
at B as indicated in the figure below.

Express your answer in terms of the circulation, Γ. Assume that the fluid level in the tank at a
large distance from the opening at A remains constant and viscous effects are negligible.

15
3.2.1 Solution - Part (a)
It is given that we can model the vortex as a free vortex, hence we can also then apply the
Bernoulli Equation between any two points within the flow field since a free vortex is irrotational.
Selecting points (1) and (2) on the fluid surface as indicated in the figure of Part (a), we can assume
p1 = p2 = 0 and letting z2 − z1 = zs , we apply the Bernoulli Equation as follows:
1 2 1
p1 + ρV + γz1 =p2 + ρV22 + γz2
 2 1 2
1 2 1
ρV = γ(z2 − z1 ) + ρV22
2 1 2
1 2 1 2
ρV = γzs + ρV2
2 1 2
V12 V22
= zs +
2g 2g
Γ
Recall that the stream function of a free vortex is ϕ = 2π θ and the velocity in the polar direction
θ, i.e., the circular velocity within a cross-sectional area of the vortex in an xy-plane, is expressed
as follows for a free vortex:
1 ∂ϕ
Vθ =
r ∂θ
1 Γ
Vθ =
r 2π
Let us then notice the following. At point (1), far away from the origin, we can approximate
our velocity V1 = Vθ ≈ 0 by using a free surface assumption. We then simplify our Bernoulli
expression further and substitute our expression for Vθ as follows:

V12
V22
= zs +

2g


2g
 2
1 1 Γ
0 = zs +
2g r 2π
Γ2
zs = −
8gr2 π 2

3.2.2 Solution - Part (b)

From the figure in Part (b), we notice that air will be sucked in if there is no water covering the
pipe B, and therefore at zs = −0.3m, we must have r ≥ 0.61m. To find the critical (maximum)
vortex strength, Γmax , we simply substitute the critical values for zs and r as follows:

Γ2max
zs = −0.3 = −
8g(0.61)2 π 2
p
Γmax = −(−0.3)(0.61)2 (8g)π 2
= 9.298648523 m2 /s
Γmax = 9.30 m2 /s (3.s.f.)

16
3.3 Question 6
Suppose we have the velocity field, V, as follows:
10x 10y
V= i+ 2 j + 0k
x2 +y 2 x + y2
Is it possible that the flow is incompressible? If so, compute the pressure gradient, ∇p assuming
a frictionless flow with negligible body forces.

3.3.1 Solution - Part(a): Incompressibility

Let us first analyse the continuity equation. If the flow is incompressible, it must adhere to the
incompressible continuity equation as follows:
∂u ∂v ∂w
∇·V = + +  =0
∂x ∂y |{z}
∂z
w=0
∂u ∂v
+ =0
∂x ∂y
We must then differentiate the velocity components. First recall the Quotient Rule for differ-
entiation:
g(x)f ′ (x) − f (x)g ′ (x)
 
∂ f (x)
=
∂x g(x) g(x)2
and we differentiate using the Quotient Rule:
 
∂u ∂ 10x
=
∂x ∂x x2 + y 2
(x2 + y 2 )10 − 10x(2x)
=
(x2 + y 2 )2
10x2 + 10y 2 − 20x2
=
(x2 + y 2 )2
∂u 10y 2 − 10x2
=
∂x (x2 + y 2 )2
and we also have the following:
 
∂v ∂ 10y
=
∂y ∂x x2 + y 2
(x2 + y 2 )10 − 10y(2y)
=
(x2 + y 2 )2
10x2 + 10y 2 − 20y 2
=
(x2 + y 2 )2
∂v 10x2 − 10y 2
=
∂y (x2 + y 2 )2
Substituting these differential expressions back into our continuity equation, we have:
∂u ∂v 10y 2 − 10x2 10x2 − 10y 2
+ = +
∂x ∂y (x2 + y 2 )2 (x2 + y 2 )2

17
 ∂u ∂v
+ =0
∂x ∂y
and therefore the incompressible continuity equation holds, and the given flow is possibly in-
compressible.

(Note that there is a singularity point at (x, y) = (0, 0) where the flow may not be incompressible
since the denominator of the differential fractions cannot be zero.)

3.3.2 Solution - Part(b): Pressure Gradient - ∇p

We are given that the flow is frictionless, and therefore we can assume that the flow is also
inviscid. We can therefore analyse the Euler Equations as follows:
 
∂p ∂u ∂u ∂u ∂u
x: ρgx − =ρ +u +v +w
∂x ∂t ∂x ∂y ∂z
 
∂p ∂v ∂v ∂v ∂v
y: ρgy − =ρ +u +v +w
∂y ∂t ∂x ∂y ∂z
 
∂p ∂w ∂w ∂w ∂w
z: ρgz − =ρ +u +v +w
∂z ∂t ∂x ∂y ∂z

The desired pressure gradient term ∇p is composed as follows:

∂p ∂p ∂p
∇p = i+ j+ k
∂x ∂y ∂z
Let us now make assumptions to simplify our equations above to find the pressure gradient terms.

Firstly, can the velocity components, we have w = 0. - Assumption 1

Next, it is given that we have negligible body forces, i.e., gravitational forces, and we therefore
have gx = gy = gz = 0. - Assumption 2

Lastly, we have a steady flow given the velocity components do not depend on time. - Assumption
3

We simplify as follows:
 
∂p  ∂u
∂u ∂u ∂u 
ρgx − = ρ 
+u +v +w 

|{z} ∂x
∂t ∂x ∂y | {z
∂z}
A.2
|{z}
 A.3 A.1

∂p  ∂v
∂v ∂v ∂v 
ρgy − = ρ 
+u +v +w 

|{z} ∂y
∂t
|{z} ∂x ∂y | {z
∂z}
A.2
A.3 A.1
(((
((((∂w
 
∂p ∂w (( ∂w(((∂w
ρgz − = ρ (((+u +v +w
( ∂z
( ( ((( ∂t ∂x ∂y ∂z
|(( {z }
A.1

18
Therefore, we get the following equations for our pressure gradient components:
 
∂p ∂u ∂u
= −ρ u +v
∂x ∂x ∂y
 
∂p ∂v ∂v
= −ρ u +v
∂y ∂x ∂y

Note that we have already found or were given the following:

10x
u=
x2
+ y2
10y
v= 2
x + y2
∂u 10y 2 − 10x2
=
∂x (x2 + y 2 )2
∂v 10x2 − 10y 2
=
∂y (x2 + y 2 )2

and we find the remaining unknown terms, again by differentiating using the Quotient Rule as
follows:
 
∂u ∂ 10x
=
∂y ∂y x2 + y 2
(x2 + y 2 )(0) − 10x(2y)
=
(x2 + y 2 )2
∂u −20xy
= 2
∂y (x + y 2 )2

and as follows:
 
∂v ∂ 10y
=
∂x ∂y x2 + y 2
(x2 + y 2 )(0) − 10y(2x)
=
(x2 + y 2 )2
∂v −20xy
= 2
∂x (x + y 2 )2

Therefore, we substitute into and simplify our expression for the pressure gradient as follows,
firstly in the x-direction:
 
∂p ∂u ∂u
= −ρ u +v
∂x ∂x ∂y
10x 10y 2 − 10x2
 
10y −20xy
= −ρ + 2
x2 + y 2 (x2 + y 2 )2 x + y 2 (x2 + y 2 )2
100xy 2 − 100x3 −200xy 2
 
= −ρ + 2
(x2 + y 2 )3 (x + y 2 )3
−100xy 2 − 100x3
 
= −ρ
(x2 + y 2 )3

19
 100(xy 2 + x3 )
=ρ
(x2 + y 2 )3
100x(y 2 + x2 )
=ρ
(x2 + y 2 )3
∂p 100x
=ρ 2
∂x (x + y 2 )2

then similarly, in the y-direction:

 
∂p ∂v ∂v
= −ρ u +v
∂y ∂x ∂y
10y 10x2 − 10y 2
 
10x −20xy
= −ρ +
x2 + y 2 (x2 + y 2 )2 x2 + y 2 (x2 + y 2 )2
−200x2 y 100x2 y − 100y 3
 
= −ρ +
(x2 + y 2 )3 (x2 + y 2 )3
−100x y − 100y 3
2
 
= −ρ
(x2 + y 2 )3
100(x2 y + y 3 )
=ρ
(x2 + y 2 )3
100y(x2 + y 2 )
=ρ
(x2 + y 2 )3
∂p 100y
=ρ 2
∂y (x + y 2 )2

Finally, we can express our pressure gradient as follows:

100x 100y
∇p = ρ i + ρ j + 0k
(x2 + y 2 )2 (x2 + y 2 )2

20
4 Appendix
4.1 Note on the Steady Flow Assumption
This section is not required but serves as a note for those who are interested. In the lecture
slides for Chapter 6, Page 11, under “For steady flows”, we have the following relationship that
is utilised (with “Handout” next to it):
1
(V · ∇)V = ∇(V · V) − V × (∇ × V) (5)
2
where the velocity vector V = ui + vj + wk. This section will prove the above mathematical
relationship Eq. (5) rigorously as follows. We first observe the following where × denotes the
vector cross product:

i j k
∂ ∂ ∂
V × (∇ × V) = V ×
∂x ∂y ∂z
u v w
      
∂w ∂v ∂w ∂u ∂v ∂u
= (ui + vj + wk) × − i− − j+ − k
∂y ∂z ∂x ∂z ∂x ∂y
      
∂w ∂v ∂u ∂w ∂v ∂u
= (ui + vj + wk) × − i+ − j+ − k
∂y ∂z ∂z ∂x ∂x ∂y
i j k
u v w
=
! ! !
∂w ∂v ∂u ∂w ∂v ∂u
− − −
∂y ∂z ∂z ∂x ∂x ∂y
         
∂v ∂u ∂u ∂w ∂v ∂u ∂w ∂v
V × (∇ × V) = v − −w − i− u − −w − j
∂x ∂y ∂z ∂x ∂x ∂y ∂y ∂z
| {z } | {z }
i−component j−component
    
∂u ∂w ∂w ∂v
+ u − −v − k
∂z ∂x ∂y ∂z
| {z }
k−component

Notice that we have the following relationship:

∂ 2 ∂u ∂v ∂w
(u + v 2 + w2 ) = 2u + 2v + 2w
∂x ∂x ∂x ∂x
1 ∂ 2 ∂u ∂v ∂w
(u + v 2 + w2 ) = u +v +w (6)
2 ∂x ∂x ∂x ∂x
where we can replace x with y or z and Eq. (6) will still hold. We then simplify the above
i−, j−, k−components denoted above. Firstly, for the i−component, we have the following:
   
∂v ∂u ∂u ∂w ∂v ∂u ∂u ∂w
v − −w − =v −v −w +w
∂x ∂y ∂z ∂x ∂x ∂y ∂z ∂x
   
∂v ∂w ∂u ∂u
= v +w − v +w
∂x ∂x ∂y ∂z

21
    
∂u ∂v ∂w ∂u ∂u ∂u
= u +v +w −u − v +w
∂x ∂x ∂x ∂x ∂y ∂z
   
∂u ∂v ∂w ∂u ∂u ∂u
= u +v +w − u +v +w
∂x ∂x ∂x ∂x ∂y ∂z
1 ∂ 2
i − component = (u + v 2 + w2 ) − (V · ∇)u
2 ∂x
where we utilised Eq. (6). Note that we have:

∂ ∂ ∂
V·∇=u +v +w
∂x ∂y ∂z
where it acts as an operator. Similarly, for the j−component, we have the following:
    
∂v ∂u ∂w ∂v ∂u ∂v ∂w ∂v
− u − −w − =u −u +w −w
∂x ∂y ∂y ∂z ∂y ∂x ∂y ∂x
   
∂u ∂w ∂v ∂v
= u +w − u +w
∂y ∂y ∂x ∂z
   
∂u ∂v ∂w ∂v ∂v ∂v
= u +v +w −v − u +w
∂y ∂y ∂y ∂y ∂x ∂z
   
∂u ∂v ∂w ∂v ∂v ∂v
= u +v +w − u +v +w
∂y ∂y ∂y ∂x ∂y ∂z
1 ∂ 2
j − component = (u + v 2 + w2 ) − (V · ∇)v
2 ∂y
and finally, for the k−component, we have the following:
   
∂u ∂w ∂w ∂v ∂u ∂w ∂w ∂v
u − −v − =u −u +v −v
∂z ∂x ∂y ∂z ∂z ∂x ∂y ∂x
   
∂u ∂v ∂w ∂w
= u +v − u +v
∂z ∂z ∂x ∂y
   
∂u ∂v ∂w ∂w ∂w ∂w
= u +v +w −w − u +v
∂z ∂z ∂z ∂z ∂x ∂y
   
∂u ∂v ∂w ∂w ∂w ∂w
= u +v +w − u +v +w
∂y ∂y ∂y ∂x ∂y ∂z
1 ∂ 2
k − component = (u + v 2 + w2 ) − (V · ∇)w
2 ∂y
Combining the i−, j−, k−components, we have the following:
 
1 ∂ 2 2 2
V × (∇ × V) = (u + v + w ) − (V · ∇)u i
2 ∂x
 
1 ∂ 2 2 2
+ (u + v + w ) − (V · ∇)v j
2 ∂y
 
1 ∂ 2 2 2
+ (u + v + w ) − (V · ∇)w k
2 ∂y
 
1 ∂ ∂ ∂
= i+ j + k · (u2 + v 2 + w2 ) − (V · ∇)V
2 ∂x ∂y ∂z

22
 1
V × (∇ × V) = ∇(V · V) − (V · ∇)V
2

where re-arranging the above gives us the desired Eq. (5).

4.2 Further on Derivation Techniques for Vector Calculus

While the above proof seems extremely cumbersome and is essentially a “brute-force” proof,
a “neater” proof requires more complex concepts such as the Kronecker Delta, denoted as δij ,
and the Alternating Tensor (more commonly known as the Levi-Civita symbol), εijk which are
particularly useful for proofs within vector calculus. The usage of the Einstein notation also helps
simplify the proofs. The desired equation is a specific case of the vector triple product. Other
related identities include:

∇ × (∇ × A) = ∇(∇ · A) − ∇2 A

which is the curl of a curl of a vector field, is also known as the Lagrange triple cross product.
These identities, derived using vector calculus, have numerous useful applications in the study of
physics-related fields such as electromagnetism, gravitation, and of course, fluid dynamics.

23