EMG 2301: FLUID MECHANICS II

Name: Obadiah K. Maswai

Office: ELB 204
Contact: 0727411231
Email: [email protected]

Notes meant for undergraduate students in the Departments of

Mechanical Engineering/Agricultural & Biosystems
Engineering in the Jomo Kenyatta University of Agriculture and
Technology
PREAMBLE
Course Outline

Types of fluid flows. Reynolds number. Momentum equation: applications of lin-

ear and angular momentum equations. Jet propulsion. Steady flow between solid
boundaries; applications in dashpots and slider bearings. Steady flow in pipes. Un-
steady flows in closed pipelines; water hammer; surge tanks; shafts; surge control.
Power transmission through pipelines. Pipe networks. Flow in open channels; the
optimum cross-section of a channel; varying flow. Dimensional analysis: Theorem;
dimensionless groups; physical significance of dimensionless groups; similarity laws.

Prescribed text books

(i) Douglas, J.F., Gasiorek J.M. & Swaffield J.A., (2001), Fluid Mechanics, Pren-
tice Hall, 4th Ed.

(ii) Munson B.R., Young D.F. & Okiishi T.H. (1998) Fundamentals of Fluid Me-
chanics, John Wiley and Sons, 3rd Ed.

(iii) Roberson J.A., Crowe C.T. & Elger D.F. (1999) Engineering Fluid Mechanics,
John Wiley and Sons, 9th Ed.

(iv) Bansal R.K. (1992) Fluid Mechanics and Hydraulic Machines, R.K. Laxmi
Publications, 4th Ed.

(v) Massey B.S. (1994) Mechanics of Fluids, Van Nostrand Reinhold (UK) Co.
Ltd, 5th Ed.

ii
CHAPTER 1

Introduction

This unit deals with the analysis of fluid in motion - fluid dynamics. The motion of
fluids can be predicted in the same way as the motion of solids are predicted using
the fundamental laws of physics together with the physical properties of the fluid.

1.1 Types of Fluid Flow

Definitions
Uniform flow: The flow velocity has the same magnitude and direction at every
point in the fluid.
Non-uniform: If at a given instant, the velocity is not the same at every point
the flow is non-uniform. (In practice, by this definition, every fluid that flows near
a solid boundary will be non-uniform - as the fluid at the boundary must take the
speed of the boundary, usually zero. However if the size and shape of the of the
cross-section of the stream of fluid is constant the flow is considered uniform.)
Steady: A steady flow is one in which the conditions (velocity, pressure and cross-
section) at a point do not change with time.
Unsteady: If at any point in the fluid, the conditions change with time, the flow is
described as unsteady. (In practise there is always slight variations in velocity and
pressure, but if the average values are constant, the flow is considered steady.

Combining the above we can classify any flow in to one of four type:

(i) Steady uniform flow. Conditions do not change with position in the stream
or with time. An example is the flow of water in a pipe of constant diameter
at constant velocity.

(ii) Steady non-uniform flow. Conditions change from point to point in the

1
 stream but do not change with time. An example is flow in a tapering pipe
with constant velocity at the inlet - velocity will change as you move along the
length of the pipe toward the exit.

(iii) Unsteady uniform flow. At a given instant in time the conditions at every
point are the same, but will change with time. An example is a pipe of
constant diameter connected to a pump pumping at a constant rate which is
then switched off.

(iv) Unsteady non-uniform flow. Every condition of the flow may change from
point to point and with time at every point. For example waves in a channel.

1.2 Laminar and Turbulent Flow

The flow of real fluids exhibits viscous effect, that is they tend to stick to solid
surfaces and have stresses within their body. The shear stress resulting from viscosity
may be included in the equations of motion for small element of fluid, but the
resulting partial differential equations (the Navier-Stokes equations) have no known
solutions. For studying the behavior of real fluids engineers must therefore resort
to experiments. One such experiment is the use of Osborne Reynold apparatus
(named after Osborne Reynold (1842-1912)) to demonstrate different flows. A fine
nozzle was connected to a small reservoir of liquid dye to discharge a coloured
filament into the inlet of a glass tube. The flow rate in the glass pipe was controlled
by a valve. By observing the behaviour of the stream of dye, the flow was classified
according to the velocity of flow into:

• Laminar flow

This kind of flow occur at very low velocities. The particles of fluid are moving
in straight lines.

• Transition flow

The flow that occur in the transition from laminar to turbulent flow.

2
 • Turbulent flow

This kind of flow occur at high velocities. The fluid particles move in a disor-
derly manner. A single particle follows an erratic path involving movements
in three dimensions.

1.3 Criterion of Flow

As laminar and turbulent flow are widely different in their nature and effect, it is
of prime importance to know the conditions under which each may be expected to
exist, and the laws governing them. The primary parameter correlating the viscous
behaviour of all Newtonian fluid is the dimensionless group Reynold Number, Re,
which express the ratio of inertia forces to viscous forces, given as

ρul
Re = (1.1)
µ

where ρ is the density, l is characteristic length, u is velocity and µ is dynamic

viscosity.
It is evident from Eq. 1.1 that a high value of ρ, l and u, or a small value of µ, gives
a high Re. Conversely, a low value of Re is brought about by high viscosity, or low
density, low velocity or small size. A high value of Re indicates that inertia forces
dominate the flow while viscous forces play only a small part. On the other hand, a
low value of Re indicates that viscous forces dominate the flow while inertia forces
play a small part.

The flow can be characterizes using Reynold number as:

Laminar flow: Re < 2000
Transitional flow: 2000 < Re < 4000
Turbulent flow: Re > 4000

3
1.4 Momentum Equations

When a fluid flows over/on a solid body, it exert force on it. For example: force
exerted on a solid surface by a jet of fluid impinging on it, aerodynamic forces (lift
and drag) on aircraft wing, the force on a pipe-bend by fluid flowing within it, etc.
These forces are hydrodynamic forces (due to a moving fluid) and are associated
with a change in momentum.

The magnitude of the hydrodynamic forces on the body due to a moving fluid is
determined by Newton’s second law of motion ”the net force acting on a body in any
direction is equal to the rate of increase of momentum of the body in that direction”;

F = ma (1.2)

The law usually need to be expressed in a form particularly suited to steady flow of
a fluid.

1.5 Linear momentum equations for steady flows

Consider a steady flow which is non-uniform flowing in a control volume (stream

tube) where A is cross sectional area, u is velocity and ρ is density and subscript 1
and 2 represent conditions at entry and exit respectively. In a short interval δt, a
volume of the fluid moves from the inlet a distance uδt.
The mass of fluid entering the control volume in time δt is

M ass entering control volume = volume × density = ρ1 A1 u1 δt (1.3)

Hence the momentum entering the control volume is given by

M omentum entering control volume = mass × velocity = ρ1 A1 u1 u1 δt (1.4)

Similarly, the momentum leaving the stream tube is given by

M omentum leaving control volume = ρ2 A2 u2 u2 δt (1.5)

The force exerted by the fluid is calculated using Newton’s 2nd law;

4
 Force = Rate of change of momentum

ρ2 A2 u2 u2 δt − ρ1 A1 u1 u1 δt
F = (1.6)
δt
For a steady flow, continuity requires that

Q = A1 u1 = A2 u2 (1.7)

and for a constant density, ρ1 = ρ2 = ρ, then Eq. 1.6 reduces to

F = Qρ(u2 − u1 ) (1.8)

Note: F is the total force (FT ) that result from the force acting on the surface of
the body, FR , pressure force, Fp , and body forces, FB , i.e.

FT = FR + Fp + FB (1.9)

We are interested in the force acting on the surface of the body, FR .

1.6 Angular momentum equations for steady flows

Consider a steady flow which is non-uniform flowing in a control volume (stream

tube). The inlet velocity vector, u1 , makes an angle θ1 with the x-axis, while at the
outlet, velocity u2 makes an angle θ2 to the x-axis. Therefore, the forces are resolved
in the directions of the co-ordinate axes. Force in x direction, Fx is given as,

Fx = Rate of change of momentum in x-direction

Fx = ṁ(u2x − u1x ) (1.10)

Fx = Qρ(u2 cos θ2 − u1 cos θ1 ) (1.11)

Force in y direction, Fy is given as

Fy = Rate of change of momentum in y-direction

5
 Fy = ṁ(u2y − u1y ) (1.12)

Fy = Qρ(u2 sin θ2 − u1 sin θ1 ) (1.13)

Similar to linear momentum equations, Fx and Fy are the total force given by Eq
1.9 for x and y direction respectively. The resultant force, Fresultant is given as,
q
Fresultant = Fx2 + Fy2 (1.14)

The angle, α at which this force act is given as,

 
−1 Fy
α = T an (1.15)
Fx
Note: The force exerted by the fluid on the solid body touching the control volume
is opposite to F. So the reaction force, R is given as

R = −FR (1.16)

1.7 Application of Momentum Equations

Typical application of momentum equations are:

(i) Force due to the flow of fluid round a pipe bend.

(ii) Force on a nozzle at the outlet of a pipe.

(iii) Impact of a jet on a plane surface.

(iv) Force due to flow round a curve vane.

1.7.1 Force due to the flow of fluid round a pipe bend

Consider a horizontal pipe with a constant cross section, which bend through an
angle of θ. Since the fluid in the pipe is changing direction, a force will act on the
bend. Therefore, determination of this is necessary in order to determine the force
in which the support must withstand.

6
Analysis

(i) Calculation of total force

Force in x-direction is

FT x = Qρ(u2x − u1x ) = Qρ(u2 cos θ − u1 ) (1.17)

Force in y-direction is

FT y = Qρ(u2y − u1y ) = Qρ(u2 sin θ − 0) = Qρu2 sin θ (1.18)

(ii) Calculation of pressure force

Fp = Pressure force at 1 - Pressure force at 2

Pressure force in x-direction is

Fpx = p1 A1 cos 0 − p2 A2 cos θ = p1 A1 − p2 A2 cos θ (1.19)

Pressure force in y-direction is

Fpy = 0 − p2 A2 sin θ = −p2 A2 sin θ (1.20)

(iii) Calculation of body forces. Body forces due to gravity are negligible, i.e.
FB =0.

(iv) Determination of force acting on the surface of the body, FR . It is calculated

using Eq. 1.9

FT = FR + Fp + FB (1.21)

⇒ FR = FT − Fp

FRx = FT x − Fpx = Qρ(u2 cos θ − u1 ) − p1 A1 + p2 A2 cos θ (1.22)

7
 FRy = FT y − Fpy = Qρu2 sin θ + p2 A2 sin θ (1.23)

The resultant force on the fluid is given by

q
2 2
FR = FRx + FRy (1.24)

The angle, α at which this force act is given as

 
−1 FRy
α = T an (1.25)
FRx

The force on the bend R is the same magnitude but opposite in direction.
Note: The flow is confined within a pipe, hence the static pressure vary from point
to point and forces due to the differences of static pressure must be taken into
account.
Example 1
A 600 reducing pipe bend tapers from 800 mm diameter at inlet to 400 mm diameter
at outlet. The pressure at inlet and outlet is 160 kPa and 146.5 kPa respectively
and the rate of flow of water through the bend is 0.675 m3 /s. Neglecting friction,
calculate the net resultant force exerted by the water on the bend.

1.7.2 Force on a nozzle at the outlet of a pipe

Because the fluid is contracted at the nozzle, forces are induced in the nozzle.

Analysis

(i) Calculation of total force ( x-direction only)

FT x = Qρ(u2 − u1 ) (1.26)

Since Q = A1 u1 = A2 u2 (continuity equation), then Eq. 1.26 reduces to

 
2 1 1
FT x = Q ρ − (1.27)
A2 A1

8
 (ii) Calculation of pressure force (x-direction only)

Fpx = p1 A1 − p2 A2 (1.28)

Assignment 1: show that,

Q2 ρ
 
1 1
p1 = 2
− 2 (1.29)
2 A2 A1

(iii) Calculation of body forces. Body forces due to gravity are negligible, i.e.
FB =0.

(iv) Determination of force acting on the surface of the body, FR . It is calculated

using Eq. 1.9

⇒ FRx = FT x − Fpx
Q2 ρA1
   
2 1 1 1 1
FRx = Q ρ − − 2
− 2 (1.30)
A2 A1 2 A2 A1

Example 2
The the rate of flow of water through the nozzle of a fire fighting equipment is 0.25
m3 /s. Neglecting friction, calculate the net resultant force that the fireman must
withstand. Assume that the fireman is holding the nozzle horizontally.

1.7.3 Impact of a jet on a plane surface

(a) A jet hitting a flat stationary plate, i.e. at angle of 900

Analysis

(i) Calculation of total force

FT = Qρ(u2x − u1x ) = −Qρu1x (1.31)

At outlet the jets are parallel to the plate with no component in the x-direction,
i.e, u2x = 0. The system is symmetrical, hence the forces in y-direction is zero,
Fy =0.

9
 (ii) Pressure force is zero (the pressure at both the inlet and outlets are atmo-
spheric).

(iii) Body forces is zero (acting perpendicular).

(iv) Determination of force acting on the surface of the body, FR . It is calculated

using Eq. 1.9
⇒ FRx = FT x − Fpx − FBx

FRx = FT x = −Qρu1x (1.32)

Force on the plane, R, is the same magnitude but in the opposite direction, i.e.
R = −FRx .
Example 3
A water jet of velocity u impinges normal to a flat plate which moves to the right
at a velocity uc . Find the force required to keep the plate moving at a constant
velocity if the jet density is 1000 Kg/m3 , jet area is 3 cm2 , u and uc are 20 m/s and
15 m/s respectively. Neglect the gravitation force.

(b) A jet hitting an inclined stationary plate

The reaction force normal to the plate is to be calculated, hence we choose an axis
system such that the x-direction is normal to the plane. The first step is to calculate
velocities u1 , u2 , and u3 using Bernoulli’s equation,

p1 u21 p2 u22 p3 u23

+ + z1 = + + z2 = + + z3 (1.33)
ρg 2g ρg 2g ρg 2g

All pressures are atmospheric p1 = p2 = p3 = 0, and the height differences are

negligible, i.e. z1 = z2 = z3 , hence u1 =u2 = u3 . Continuity requires that,

Q1 = Q2 + Q3

A1 u1 = A2 u1 + A3 u1

A1 = A2 + A3 (1.34)

10
Analysis

(i) Calculation of total force

FT = ρ ((Q2 u2x + Q3 u3x ) − Q1 u1x ) = −Q1 ρu1x = −Q1 ρu1 cos θ (1.35)

At outlet the jets are parallel to the plate with no component in the x-direction,
i.e., u2x = u3x = 0. The system is symmetrical, hence the forces in y-direction
are zero, Fy =0.

(ii) Pressure force is zero (the pressure at both the inlet and outlets are atmo-
spheric).

(iii) Body force is zero (acting perpendicular).

(iv) Determination of force acting on the surface of the body, FR . It is calculated

using Eq. 1.9
⇒ FRx = FT x − Fpx − FBx

FRx = FT x = −Q1 ρu1 cos θ (1.36)

Example 4
A water jet, moving at a flow rate of 0.35 m3 /s hit an inclined plate (300 to the
horizontal). If the diameter of the jet is 150 mm, calculate the force exerted on the
plate.

1.7.4 Force due to flow round a curve vane

The forces due to fluid flowing over a curved vane is similar to that of fluid flowing
on a pipe-bend, but the pressures are atmospheric p1 = p2 = 0 and both the cross-
section and velocities remain constant.

11
Analysis

(i) Calculation of total force

Force in x-direction is

FT x = Qρ(u2x − u1x ) = Qρ(u2 cos θ − u1 ) (1.37)

But Au1 = Au2 , hence Eq. 1.37 can be written as,

Q2 ρ
FT x = (cos θ − 1) (1.38)
A
Force in y-direction is,

FT y = Qρ(u2y − u1y ) = Qρ(u2 sin θ − 0) = Qρu2 sin θ (1.39)

(ii) Pressure force is zero (the pressure at both the inlet and outlet are atmo-
spheric).

(iii) Body force is zero (acting perpendicular).

(iv) Determination of force acting on the surface of the body, FR . It is calculated

using Eq. 1.9
⇒ FR = FT − Fp
FRx = FT x and FRy = FT y (1.40)

The resultant force on the fluid is given by

q
2 2
FR = FRx + FRy (1.41)

The angle, α at which this force act is given as,

 
−1 FRy
α = T an (1.42)
FRx

12
Example 5
A jet of water flows smoothly on to a stationary curved vane which turns it through
600 . The initial jet is 50 mm in diameter, and the velocity, which is uniform is 36
m/s. As a result of friction, the velocity of water leaving the surface is 30 m/s.
Neglecting gravity effects, calculate the net resultant force on the vane.

13
CHAPTER 2

Laminar Flow Between Boundaries

2.1 Introduction

Laminar flows occurs at velocities low enough for forces due to viscosity to predomi-
nant over inertia forces and thus, if any individual particle attempts to stray from its
prescribed path, viscosity firmly restrain it. Viscous stresses arises whenever there
is relative movement between adjacent particles of fluid. For a Newtonian fluid, it
is given by,
∂u
τ =µ (2.1)
∂y
∂u
where τ is the shear stress, µ is dynamic viscosity, and ∂y
is the rate at which the
velocity u increases with coordinate y perpendicular to velocity. The following cases
of laminar flow between solid boundaries are to be considered:

(i) Steady laminar flow in circular pipes (The Hagen-Poiseuille law).

(ii) Steady laminar flow between parallel plates.

(iii) Simple cylindrical dash-pot.

(iv) Slider bearing.

2.2 Steady laminar flow in circular pipes (The Hagen-Poiseuille law)

The first documented results about the laminar flow in pipes was the work of German
engineer G. H. L Hagen (1797-1884) which was published in 1839 (flow of water
through small brass tubes). Around the same time, similar work was done by a
French physician J. L. M. Poiseuille (1799-1869) which was published in 1840 (study
blood through veins).
Consider a short length δx of pipe with constant internal radius R inclined at an
angle θ. Let us take a cylinder (small element of fluid) of radius r, moving from left

14
to right with velocity u inside a slightly larger cylinder of radius r + δr, moving in
the same direction with velocity u + δu. In our analysis we assume that the part of
the pipe considered is far enough from the inlet for conditions to have settled.
Net force due to pressure in direction of flow = πr2 (p − (p + δp))
Force due to weight of fluid in direction of flow = −πr2 ρgδx cos θ
Force due to mean shear stress, τ0 in direction of flow = 2πrτ0 δx
The net force on the fluid, F, in the direction of flow is,

F = πr2 (p − (p + δp)) − πr2 ρgδx cos θ + 2πrτ0 δx (2.2)

For steady flow which is fully developed, i.e., the velocity distribution over the cross-
section does not change with distance x, there is no increase in momentum in this
direction and therefore, net force in this direction is zero. Hence,

F = πr2 (p − (p + δp)) − πr2 ρgδx cos θ + 2πrτ0 δx = 0 (2.3)

which can be rearranged as,

2πrτ0 δx = πr2 (δp + ρgδz) (2.4)

where δx cos θ = δz. For fluid at constant density, p∗ = p + ρgz is the piezometric
pressure. Therefore Eq. 2.4 can be reduced to,
r δp∗
τ0 = (2.5)
2 δx
Then in the limit, δx −→ 0
r dp∗
τ0 = (2.6)
2 dx
We also know that for a Newtonian fluid,
∂u
τ =µ (2.7)
∂r
Therefore combining Eqs. 2.6 and 2.7 and knowing that velocity only varies with
radius and not with x, i.e. partial changes to total differential,
du r dp∗
µ = (2.8)
dr 2 dx

15
which can be rearranged as,
r dp∗
du = dr (2.9)
2µ dx
Integrating both sides gives,
r2 dp∗
u= +A (2.10)
4µ dx
At the pipe walls, i.e at r = R, u = 0 (no slip conditions). If we apply this boundary
condition to Eq. 2.10, we get,

R2 dp∗
0= +A (2.11)
4µ dx

R2 dp∗
⇒ A=− (2.12)
4µ dx
Substituting for A in Eq. 2.10, gives the velocity at any given point as,

1 dp∗
 
u=− (R2 − r2 ) (2.13)
4µ dx

Equation 2.13 is known as the Hagen-Poiseuille equation. It shows that the

velocity is maximum at the centre, i.e. at r = 0. The velocity distribution (velocity
profile) over the cross section can be represented graphically by plotting u against
r.

Discharge Through the Pipe

Consider an elemental cylinder (small element of fluid) of radius r and thickness δr.
The discharge in the elemental cylinder, δQ, is given by,

δQ = AU = 2πrδru (2.14)

Substituting for the u using Eq. 2.13 gives,

2π dp∗
 
δQ = − (rR2 − r3 )δr (2.15)
4µ dx

Integrating both sides for the whole radius gives,

Z r
2π dp∗
Z  
δQ = − (rR2 − r3 )δr (2.16)
0 4µ dx

16
which on solving result in,

dp∗ r 2 R2 r 4
 
π
Q=− ( − ) (2.17)
2µ dx 2 4

Hence, the discharge at any radius r is given by Eq. 2.17. The discharge through
the entire section, i.e. r = R is given by,

πR4 dp∗
 
Q=− (2.18)
8µ dx

The mean velocity ū of the fluid may be given as,

Discharge for entire section

ū = (2.19)
Cross − section area
4 dp∗
− πR


R2 dp∗
 
8µ dx
ū = =− (2.20)
πR2 8µ dx
Maximum velocity umax (at r = 0 ) is,

R2 dp∗
 
umax =− (2.21)
4µ dx

Therefore in a pipe ū = (1/2)umax .

2.3 Steady laminar flow between parallel plates

Consider the flow between two parallel infinitely long (to avoid end effects) plates.
As the flow is laminar, there is no movement of fluid in any direction perpendicular
to the flow, and hence pressure varies only in the direction of flow. The flow is
caused by the difference in piezometric pressure between the two ends.

Consider an elemental fluid particle. If the thickness of the element is δz, then:
Net force due to pressure in direction of flow = (p∗ − (p∗ + δp∗ ))δyδz
Force due to shear stress in direction of flow = ((τ + δτ ) − τ )δxδz
Momentum equation requires that,

(p∗ − (p∗ + δp∗ ))δyδz + ((τ + δτ ) − τ )δxδz = ma (2.22)

17
But for steady, fully developed flow there is no acceleration and so the RHS of Eq.
2.22 is zero, i.e.,

(p∗ − (p∗ + δp∗ ))δyδz + ((τ + δτ ) − τ )δxδz = 0, (2.23)

which reduces to,

− δp∗ δy + δτ δx = 0 (2.24)

Dividing through by δxδy and proceeding to the limit δy ⇒ 0 and δx ⇒ 0, we get,

∂p∗ ∂τ
= (2.25)
∂x ∂y

Substituting for shear stress for a Newtonian fluid using Eq. 2.1, then Eq. 2.25 can
be written as,
∂p∗
 
∂ ∂u
= µ (2.26)
∂x ∂y ∂y
p∗ varies only in the x direction and u varies only in y direction, therefore, the
partial derivative can be replaced with total derivatives. Then, integration of Eq.
2.26 wrt y, yields,
dp∗ du
y = µ + A, (2.27)
dx dy
which can be rearranged as,

dp∗
µdu = ydy − Ady (2.28)
dx

An integration of this equation on both sides, gives,

dp∗ y 2
µu = − Ay + B (2.29)
dx 2

Boundary conditions are,

at y = 0, u = 0 ,
at y = c, u = 0
Applying the first B.C. to Eq. 2.29 gives B = 0. And applying 2nd B.C. gives,

dp∗ c
A= (2.30)
dx 2

18
Inserting the values of A and B in Eq. 2.29 gives the velocity at any point y as,

1 dp∗
 
u= (y 2 − cy) (2.31)
2µ dx

The velocity distribution (velocity profile) over the cross section can be represented
graphically by plotting u against y.

Assignment 2: Find the discharge at any point y and show that the total discharge
for a fixed breadth b for flow between two stationary plates with a distance separating
them as c is,
−bc3 dp∗
 
Q= (2.32)
12µ dx

2.4 Simple cylindrical dash-pot

A dash-pot is a device for damping vibrations of machines, or rapid reciprocating

motions, using a fluid with high viscosity. A piston P, connected to the mechanism
whose movement is to be restrained, moves in a concentric cylinder C, the diameter of
which is only slightly greater than that of the piston. The cylinder contains a viscous
oil. If the piston is caused to move downward, oil is displaced from underneath
through the small annular clearance between the piston and the wall of the cylinder.
For an upward movement of the piston, the pressure underneath reduces, and the
resulting difference of piezometric pressure between the top and bottom of the piston
causes the flow of oil from top to bottom.
To analyse the forces in a dash-pot, we use the analysis of flow between a fixed plate
and a moving plate, i.e. one of the surface is the piston, the other one is the inner
wall of the cylinder.
Let us consider a downward movement of the piston with velocity u. Note that this
velocity is negative to the direction of flow.
Therefore, boundary conditions are:
at y = 0, u = 0 ,
at y = c, u = −u

19
Applying the first B.C. condition to Eq. 2.29 gives B = 0. And applying 2nd B.C.
gives,
dp∗ c2
− µu = − Ac (2.33)
dx 2
µu dp∗ c
⇒A= + (2.34)
c dx 2
Inserting the values of A and B in Eq. 2.29 gives the velocity at any point y as,
 ∗ 2  ∗  
dp y y dp c µu
u= − + , (2.35)
dx 2µ µ dx 2 c
which can be rearranged as,
dp∗
 
1 yu
u= (y 2 − yc) − (2.36)
2µ dx c
The equation shows that flow may occur even without a difference in piezometric
∗

pressure provided that one boundary is moving. In such a case dp dx
= 0 and
Q = (1/2)πDuc (This situation is known as Coutte flow after M. F. A Coutte
(1858-1945)).
To calculate the discharge between the cylinder and the piston, we consider an
elemental fluid,
Q = Au = πDudy (2.37)

Substituting for u using Eq. 2.36 and integrating for the whole length c, we get,
Z Z c   ∗ 
1 dp 2 yu
dQ = πD (y − yc) − dy, (2.38)
0 2µ dx c
which can be solved to get,
  ∗ 3 
dp c cu
Q = πD − − (2.39)
dx 12µ 2
This must be equal to the rate at which oil is being displaced by the piston, Qp ,
πD2
Qp = up (2.40)
4
dp∗


If the end effects are neglected, − dx
may be considered constant and may be
written as ∆P/L, where ∆P is the difference of piezometric pressure from the bot-
tom to top of the piston.

20
Therefore equating Eqs. 2.39 and 2.40 (note that u = up ), gives
 3
πD2
  
∆P c cu
πD − − = u (2.41)
L 12µ 2 4

∆P c3
 
D
⇒u +c =− (2.42)
2 6µL
Force balance in a dash-pot

Force due to pressure, Fp is given as,

πD2
 
Fp = ∆p (2.43)
4

The upward shear force is usually negligible compared to other forces. Hence bal-
ances of forces give,
πD2
 
∆p −F −W =0 (2.44)
4

2.5 Fundamentals of the Theory of Hydrodynamic Lubrication

The flow of lubricant in various types of bearing can be assumed to be laminar.

Although high velocities may be involved, the thickness of the film of lubricant is so
small that the Reynolds number is usually below its critical value for that system.
The primary function of the lubricant is to separate the bearing surfaces, and as
long as the lubrication is effective, there is no direct contact between the surfaces.
If the film of lubrication is to keep the bearing surfaces apart, it must be capable
of sustaining a load. One way of achieving this is to supply the fluid lubricant
to the space between the surfaces at sufficiently high pressure from some external
sources. This provides hydrostatic lubrication. Hydrodynamic lubrication
on the other hand, is achieved when high pressures is produced in the lubricant as
a result of the shape and relative motion of the bearing surfaces.

Slider Bearing

Consider a slider bearing with a very small angle. To give steady conditions, the
slipper is considered fixed and the bearing plate moves. The bearing plate moves

21
with a velocity u. This motion is equivalent to the motion between two parallel
surfaces with a pressure gradient. The only difference is that they are inclined to
each other. Therefore it is governed by Eq. 2.29.
The boundary conditions are:
at y = 0, u = U
at y = h, u = 0
Applying the first B.C. to Eq. 2.29 gives B = µU . And applying 2nd B.C. gives,

dp∗ h2
0= − Ah + µU (2.45)
dx 2
dp∗ h µU
⇒A= + (2.46)
dx 2 h
Inserting the values of A and B in Eq. 2.29 gives the velocity at any point y as,
 ∗ 2  
dp y h  y
u= 1− +U 1− (2.47)
dx 2µ y h

For every section continuity requires that

Z h
Q= udy (2.48)
0

h
dp∗ y2
Z     
h y 
Q= 1− +U 1− dy (2.49)
0 dx 2µ y h
which can be solved to get,

h3 dp∗
 
Uh
Q= − (2.50)
2 12µ dx

Hence,  
∗ 6µU 12µQ
dp = − dx (2.51)
h2 h3
Integrating both sides yields,
Z x Z x
∗ dx dx
p = 6µU − 12µQ +A (2.52)
0 h2 0 h3

At x = 0, p∗ = p0 ⇒ A = p0 , then,
Z x Z x
∗ dx dx
p (x) = 6µU 2
− 12µQ 3
+ p0 (2.53)
0 h (x) 0 h (x)

22
At x = L, p∗ = p0 , therefore,
Z L Z L 
U dx dx
Q= 2 3
(2.54)
2 0 h (x) 0 h (x)

But h(x) is given as h(x) = tan α(a − x), and for small α, i.e as α → 0 tan α → α.
Therefore,
h(x) = α(a − x) (2.55)
Z L Z L 
U dx dx
Q= (2.56)
2 0 (α(a − x))2 0 (α(a − x))
3

which can be solved to get

U αa(a − L)
(2.57)
2a − L
The pressure distribution in the bearing in x direction, can be obtained by substi-
tuting the value of h(x) in Eq. 2.53, which on solving gives

6µU (x(L − x))

p∗ (x) = p0 + (2.58)
α2 (2a
− L)(a − x)2

Since, under the slipper, x is always less than L and L is less than 2a, the term
2a − L is always positive; p∗ is thus greater than p0 and the bearing can sustain a
load.
If the distance a approaches infinity, i.e, if the surface become parallel, then p∗ = p0
for all values of x.
The load which a bearing can support is determined by the total net thrust, which
the fluid exerts on the unit width of either bearing surface. The net thrust per unit
width, T, is obtained from the integral
Z L Z L
6µU x(L − x)
T = (p − p0 )dx = 2 dx, (2.59)
0 α (2a − L) 0 (a − x)2

which can be solved to get,

   
6µU a 2L
T = 2 ln − (2.60)
α a−L 2a − L

23
Exercise

(i) Two parallel plates are 4 mm apart and a steady viscous flow of oil is taking
place between them. If the pressure drop is 12 kN/m2 per meter length of
plates and the dynamic viscosity for oil is 5 × 10−2 N s/m2 , find the local
velocity at 1 mm from the lower plate, flow per meter width (total discharge),
and the maximum shear stress.

(ii) A dash-pot consist of a cylinder 7 cm diameter in which it slides a piston 8

cm long having a radial clearance of 1 mm. The cylinder is filled with an oil
of viscosity 0.1 N s/m2 . Calculate the velocity of the piston when acted upon
by a load having a mass of 18 kg.

(iii) The radial clearance between a plunger and the walls of a cylinder is 0.075
mm, the length of the plunger is 250 mm and its diameter is 100 mm. There is
a difference in pressure of water on the two ends of the plunger of 207 kN/m2
and the viscosity of the water is 1.31 × 10−3 kg/m-s. Treating the flow as if
it is occurring between parallel flat plates, estimate the rate of leakage.

(iv) A dash-pot consist of a piston 143.5 mm diameter working concentrically in a

cylinder of 143.6 mm bore. The cylinder contains oil of viscosity 0.08 kg/m-s.
Calculate the force which must be applied to the piston to give it a velocity
0f 0.003 m/s, when 250 mm of piston is in the cylinder.

(v) Oil of viscosity 0.048 kg/m-s flow through an 18 mm diameter pipe with a
mean velocity of 0.3 m/s. Calculate the pressure drop which occurs over a
length of 45 m of pipe. Calculate also the velocity at a distance of 3 mm from
the walls of the pipe.

(vi) A plane bearing plate is traversed by a very wide, 150 mm long, inclined slipper
moving at 1.5 m/s. The clearance between slipper and the bearing plate is
0.075 mm at the toe and 0.025 mm at the heel. If the load to be sustained by

24
the bearing is 500 kN per meter width, determine the viscosity of the lubricant
required.

25
CHAPTER 3

Turbulent Flow in Pipes

The flow in pipes is usually turbulent. Therefore, the energy is dissipated through
various losses, the main one being frictional losses. The minor losses in pipes occur
at abrupt enlargement (exit), at abrupt contraction, in bends, and in pipe fittings.

3.1 Frictional Losses

The dissipation of energy by fluid friction result in a fall of piezometric head in the
direction of flow. Frictional losses, hf , are given by Darcy’s formula:

∆p 4f l ū2
hf = = . (3.1)
ρg d 2g

It can also be expressed as

f l ū2
hf = . (3.2)
m 2g
where f is the friction factor, ū is the mean velocity, d is the pipe diameter, g is
gravitational force, and m is the hydraulic mean depth/hydraulic radius given by
the ratio of cross-sectional area of flow, A to the perimeter in contact with fluid, P.
For a circular section m = A/P = (πd2 /4)/πd = d/4. This equation is named after
French engineer Henri Darcy (1803). The friction factor depends on the relative
roughness of the pipe surface and the Reynold number of flow.
To relate the value of f to the mean shear stress τ¯w at the wall of the pipe, we
consider a short length, dx of pipe. The head lost to friction is

f dx ū2
dhf = (3.3)
m 2g
2gm dhf
f= (3.4)
ū2 dx
For a steady, fully developed flow (see Eq. 2.4-2.6), the wall shear stress is

A dp∗ dp∗
τ̄w = =m (3.5)
P dx dx

26
The drop in piezometric pressure, p∗ is

− dp∗ = ρgdhf (3.6)

Therefore,
dhf 1 dp∗ τ̄w
=− =− (3.7)
dx ρg dx ρgm
Substituting in Eq. 3.4 gives
 
2gm τ̄w |τ̄w |
f =− = (3.8)
ū2 ρgm (1/2)ρū2
It is usual to consider only the magnitude of τ̄w .

3.1.1 Variation of Frictional Factor

Frictional factor is a function of Reynold number and the relative roughness k/d.
Experimental data on friction in pipes may be represented diagrammatically in what
is frequently referred to as Stanton or Moody diagrams (Named after the works
of E. Stanton (1865-1931), J. Nikuradse (1894-1979) and L. F. Moody (1880-1953)).
A number of emperical formulae have been put forward to describe the value of f.
Blasius, showed that for turbulent flows in smooth pipe:

f = 0.079(Re)−1/4 for 3000 < Re < 105 (3.9)

Moody himself produced the following approximate formulae for f :

" 1/3 #
k 106

f = 0.001375 1 + 20000 + (3.10)
d Re

It gives values within ±5% for Reynold numbers between 4000 and 107 and values
of k/d up to 0.01.

Exercise

(i) Determine the head lost to friction when water flows through 300 m long of
150 mm diameter galvanized steel pipe at 50 litres/s. For galvanized steel
k = 0.15 mm and kinematic viscosity of flow, ν = 1.14 × 10−6 m2 /s.

27
 (ii) Calculate the steady rate at which oil will flow through a cast-iron pipe 100
mm diameter and 120 m long under a head difference of 5 m. For cast-iron
pipe k = 0.25 mm and kinematic viscosity flow, ν = 1 × 10−5 m2 /s.

(iii) Determine the size of galvanized steel pipe needed to carry water a distance
of 180 m and 85 litres/s with a head loss of 9 m.

(iv) Water flows through a smooth galvanized pipe at a mean velocity of 2 m/s. If
the pipe diameter is 300 mm, calculate the pressure gradient per unit length
necessary. If instead of water, oil of density 800 kg/m3 and viscosity 0.027
kg/ms was used, what would be the pressure gradient per unit length?

3.2 Minor Losses

3.2.1 Loss at Abrupt Enlargement (Sudden expansion)

Fluid entering from the smaller pipe is unable to follow the abrupt deviation of
the boundary, consequently turbulent eddies form in the corners and result in the
dissipation of energy as heat.
Applying Bernoulli’s equation between point 1 and 2, we get
p1 u2 p2 u2
+ 1 = + 2 + he (3.11)
ρg 2g ρg 2g
where he is the energy loss at the expansion.
Net force acting toward right is

p1 A1 + p0 (A2 − A1 ) − p2 A2 (3.12)

p0 is the mean pressure of eddying fluid. Experimentally, it has been found that
p0 ' p1 . Therefore Eq. 3.12 can be written as

A2 (p1 − p2 ) (3.13)

The conservation of momentum requires that this force equal to the rate of increase
in momentum in the same direction, i.e

A2 (p1 − p2 ) = ρQ(u2 − u1 )

28
 ρQ
p1 − p 2 = (u2 − u1 ) = ρu2 (u2 − u1 ) (3.14)
A2
From Eq. 3.11
p1 − p2 u21 − u22
he = + (3.15)
ρg 2g
and using Eq. 3.14 for p1 − p2 , we have
u2 (u2 − u1 ) u21 − u22
he = + (3.16)
g 2g
which can be solved to get
1
he = (u1 − u2 )2 (3.17)
2g
From continuity A1 u1 = A2 u2 , therefore,
2 2
u21 u22 A2
 
A1
he = 1− = −1 (3.18)
2g A2 2g A1
 2
A1
The term 1 − A 2
is a constant, k and Eq. 3.18 can be written as

u21
he = k (3.19)
2g
 2
A1 u21
If A2  A1 , then 1 − A2
' 12 , i.e k = 1. Hence, he = 2g
. This happens at the
submerged outlet of a pipe discharging into a large reservoir e.g to atmosphere or a
tank. In such circumstances, the head loss is usually termed the exit loss for the
pipe.

3.2.2 Loss Due to Sudden Contraction

Although an abrupt contraction is geometrically the reverse of an abrupt enlarge-

ment, it is not possible to apply the momentum equation to a control volume between
section 1 and 2. This is because, just upstream of the junction, the curvature of the
streamlines and the acceleration of the fluid cause the pressure at the face to vary
in an unknown way. Therefore, the loss at a sudden contraction is given as

Loss due to sudden contraction = loss up to vena contracta + loss due to

sudden expansion beyond vena contracta

29
Loss up to vena contracta is negligible. Loss due to sudden expansion beyond vena
contracta, hc is obtained using Eq. 3.18;
2 2
u22 A2 u22 1
 
hc = −1 = −1 , (3.20)
2g Ac 2g Cc
 2
1
where Cc is the coefficient of contraction given as Cc = Ac /A2 . Cc
− 1 is a
u2
constant, k and therefore, the loss can be expressed as hc = k 2g2 .

3.2.3 Power Transmission Through Pipes

Pipes are employed while transmitting hydraulic power in conjunction with hydraulic
machines.

From conservation of energy;

4f l ū2
H2 = H1 − hf = H1 − , (3.21)
d 2g

where H1 and H2 are total energy head in point 1 and 2 respectively.

But ū = 4Q/πd2 , therefore

4f l 16Q2
H2 = H1 −
d π 2 d4 2g
Q2 f l
⇒ H2 = H1 − 0.33 5 (3.22)
d
The power at the exit, P is given as

P = ρgH2 Q

Q3 f l
 
⇒ P = ρg H1 Q − 0.33 5 (3.23)
d
It is important to establish the conditions under which a pipeline will transmit
maximum power for a given size, and also to determine for a given condition, the
smallest pipe size to transmit maximum power.
To maximize power at the exit, then

Q2 f l
 
dP
= 0 = ρg H1 − 0.99 5 (3.24)
dQ d

30
i.e
0 = ρg(H1 − 3hf )
H1
⇒ hf = (3.25)
3
Therefore, to get maximum power at the exit, we must ensure that the head lost
due to friction is a third of the total head available at entry.
By definition,
power output
Efficiency, η = × 100 (3.26)
power input
Hence for the maximum power,
H2 H1 − H1 /3
η= × 100 = × 100 = 66.7% (3.27)
H1 H1
If a nozzle is fitted at the exit, discharging to the atmosphere at a velocity ūn , then
H2 = ū2n /2g, and
ū2n Q2 f l ū2 fl
H1 = + 0.33 5 = n + 0.20 ū2 (3.28)
2g d 2g d
Continuity requires that ūd2 = ūn d2n ,
 4
2 dn
⇒ ū = ū2n
d
Therefore,  4
ū2n f l dn
H1 = + 0.20 ū2n (3.29)
2g d d
" #
H 1
⇒ ū2n =
4
0.05 + 0.20 fdl ddn
The power transmitted available at the exit, can be written as a function of the
nozzle geometry for a given head H1 as
" #3/2
ū2n πd2n
  
ρπ 2 H1
P = ρgH2 Q = ρg un = dn (3.30)
2g 4 8 dn 4
0.05 + 0.20 fdl


d

Thus for a given available head H1 and pipe, the power can be maximized wrt nozzle
diameter dn by setting ∂P/∂dn = 0 in Eq. 3.30. It can be shown that the power
will be maximum when,
1/4
d5

dn = (3.31)
8f l

31
3.2.4 Pipes in Series

Pipes are said to be in series when pipes of different materials and diameters are
connected end to end.

For pipes in series;

Total loss in energy = Loss due to friction in each pipe +

minor losses in each pipe

and for steady flow, continuity requires that

Q1 = Q2 = ... = Qn (3.32)

3.2.5 Pipes in Parallel

For parallel connection, the head causing flow is the same in each pipe. Bernoulli’s
equation is applied to each pipe independently and continuity requires that,

QT = Q1 + Q2 ...Qn (3.33)

3.3 Pipe Networks

The most common pipe networks are water distribution systems for municipalities.
These systems have one or more sources and numerous loads.

Flow can be through numerous path. Whichever is the path, the requirement is that
the value of pressure at J is unique. It is impossible to have two values of pressure
for the same point. Thus

hJ = hA − hf 1 = hB − hf 2 = hC − hf 3 = hD − hf 4 . (3.34)

Generally,
4fi li ūi
hf i = for i = 1, ...n. (3.35)
di 2g

32
This result in n equations for n junctions. Continuity requires that the net mass
outflow from a junction is equal to zero. Thus

Q1 + Q2 + Q3 − Q4 = 0 (3.36)

Therefore, we have n + 1 equations with n + 1 unknowns, i.e Q1 , Q2 , ..., Qn , and hJ .

Hence the equation is closed and can be solved.
In general we can achieve these equations if we are able to satisfy the following two
conditions:

(i) Continuity: The flow into a junction of network must be equal to the flow out
of the junction and must be satisfied for all junctions.

(ii) The head loss between any two junctions in a closed loop must be the same
regardless of the path in series of the pipes taken to get from one junction
to the other. This means that the algebraic sum of the head losses around a
given loop must be equal to zero. The positive and negative head loss being
given by the sense of flow wrt to loop, e.g clockwise or anticlockwise direction.

X
hl,clockwise = hl,AB + hl,BC
X
hl,antclockwise = hl,AD + hl,DC

Therefore
hl,AB + hl,BC = hl,AD + hl,DC (3.37)

These equations are written for all junctions and loops in the network. Therefore,
it can be seen that it is cumbersome to simultaneously solve all these equations.
Iteration is normally used where initial guess is made.

33
Exercise

(i) A system of three interconnected reservoirs A, B and C by pipes 1, 2 and 3 to

form a junction J has the following data:
pipe 1; 0.8 m diameter, 2000 m long, f = 0.005,
pipe 2; 0.5 m diameter, 500 m long, f = 0.00625,
pipe 3; 0.6 m diameter, 1000 m long, f = 0.0075.
The elevations of water surfaces in reservoirs A, B and C are 30 m, 20 m, and
5 m, respectively. Determine the discharge through each pipe.

(ii) In a hydroelectric plant the head available is 450 m of water. 25 cm penstock

pipe with friction factor of 0.0035 is used. Determine the maximum power
that can be developed. The length of the pipe line is 3600 m. Also determine
the power transmitted for u = 4.5 m/s and u = 3 m/s.

(iii) Two pipes of 0.35 m and 0.25 m diameter and length 2000 m and 1500 m with
f values 0.00525 and 0.0045, respectively connected in series carry water from
a reservoir to a supply system, the head available being 8 m. Determine the
flow quantity neglecting minor losses. If a valve is fitted at the change of the
section so that the loss coefficient k for the valve is 3.1, determine the flow
quantity without neglecting minor losses. Take inlet and exit loss coefficients
as 0.5 and 1 respectively.

(iv) Two reservoirs are connected by parallel pipes with the following details:
pipe 1; 0.2 m diameter, 800 m long, f = 0.0055,
pipe 2; 0.3 m diameter, 1200 m long, f = 0.005,
pipe 3; 0.4 m diameter, 900 m long, f = 0.00475.
i) If the head between the reservoirs is 15 m of water, determine the total flow
rate.
ii) If the total flow rate is 0.66 m3 /s, determine the individual flow and the
head.

34
Homework
A system of four interconnected reservoirs A, B, C and D by pipes 1, 2, 3 and 4 to
form a junction J has the following data:
pipe 1; 0.625 m diameter, 2500 m long, f=0.0065,
pipe 2; 0.875 m diameter, 1200 m long, f=0.0042,
pipe 3; 0.45 m diameter, 1500 m long, f=0.008.
pipe 4; 0.32 m diameter, 1750 m long, f=0.006.
The elevations of water surfaces in reservoirs A, B, C and D are 30 m, 25 m, 20 m
and 15 m, respectively. Determine the discharge through each pipe.

3.4 Unsteady Flow

In this section we consider unsteady-state phenomena applied to some important

engineering applications, such as unsteady flow in closed conduit, water hammer,
surge tanks and shafts.

3.4.1 Unsteady Flow in Closed Pipeline System

If the flow in a long pipe is shut, then the fluid upstream will adjust itself to this
stoppage only if there is a hydraulic communication between the point of closure
and the upstream. This communication takes place through a pressure wave that
travels at sonic speed through the fluid. Depending on the magnitude of pressure in
the wave, it could have destructive effects on the pipeline. Once the flow conditions
have been changed,the fluid undergoes an unsteady flow, even though for a short
time before it attains new equilibrium.
Rigid column theory is used to analyse unsteady flow as discussed in the next section.
Rigid Column Theory
The theory is valid on the following assumptions:

(i) The fluid is incompressible. The fluid become incompressible if the rate at
which velocity changes is much lower than the rate at which the pressure wave

35
 travels.

(ii) The velocity at which the pressure wave travels can be considered to be in-
finitely large.

Consider a pipeline that drains fluid from an open reservoir and discharges to the
atmosphere. When the valve at the end of a pipe line of length L as shown in Figure
above is instantaneously opened, there is a time lapse before the flow reaches steady
state. When the valve first opens, the whole of head H is used for accelerating the
flow. As the velocity increases, however, the head used for acceleration decreases
owing to the fluid friction loss hf , and discharge energy hd . Consequently, the
effective head available to accelerate the liquid in the pipe becomes ρg(H − hf − hd ).
So the equation of motion (momentum equation) of the liquid in the pipe is as
follows, putting A as the sectional area of the pipe,

dv̄
ρg(H − hf − hd )A = ρAL (3.38)
dt

But
4f l v̄ 2 v̄ 2 v̄ 2
hf = =k , and hd =
d 2g 2g 2g
Therefore, substituting the values of hf and hd in Eq. 3.38, results in

v̄ 2 L dv̄
H − (k + 1) = (3.39)
2g g dt

Assume that velocity v̄ becomes v̄0 (terminal velocity) in the steady state (dv̄/dt =
0). Then
(k + 1)v̄02 = 2gH (3.40)
2gH
⇒k= −1
v̄0
Substituting the value of k into Eq. 3.39,

v̄ 2
 
L dv̄
H 1− 2 = (3.41)
v̄0 g dt

36
 v̄02
 
L
dt = dv̄
Hg v̄02 − v̄ 2
or  
Lv̄0 v̄0 + v̄
t= ln (3.42)
2Hg v̄0 − v̄

t = ∞ if v̄ = v̄0 , i.e to attain v̄ = v̄0 , we require infinite time. To avoid this anomaly,
we evaluate the time taken to bring v̄ = 0.99v̄0 , which works out to

Lv̄0
t = 2.646 (3.43)
gH

3.4.2 Surge Tanks and Shafts

When the generator load requirements vary, the speed of the turbine is adjusted
and the control valves are adjusted accordingly to give the appropriate flow rate.
If the generator is not under any load, then the speed of the turbine must go to
zero and hence the control valves completely shut the flow. Now the pressure wave
propagates upstream and gives rise to the pressure in the penstock. This may cause
a very severe damage to the system. A surge tank is introduced to divert flow
from the penstock as shown in the figure. The surge tank should be as close to the
turbine as possible.

3.5 Water Hammer

Whenever a valve is rapidly closed in a pipe, a pressure wave is created upstream of

the valve and travels up the pipe at the speed of sound. This pressure wave is called
a water hammer and is sometime so great that it can cause pipe failure. Therefore,
in design proper understanding of this water hammer phenomena is necessary.

Example
Water is to be discharge from a tank through a piping system of 609.6 m long
into the atmosphere under a constant head of 6.1 m. Valves, fitting and pipe in the

37
 P
system have a combined resistance coefficient, k = 62. After the valve is suddenly
opened to permit flow, how much time will it take to attain 90% of the final velocity.

38
CHAPTER 4

Open Channel Flow

4.1 Introduction

Flow in rivers, irrigation canals, drainage ditches and aqueducts are some examples
for open channel flow. These flows occur with a free surface and the pressure over
the surface is atmospheric. While in closed conduits the flow is sustained by pres-
sure difference, the driving force in open channel flow is due to gravity, and is
proportional to the bed slope. The depth of flow is not restrained and this makes
the analysis more complex. As most of the flow are large in scale and as viscosity
of water is lower, Reynolds number are high. Hence the flow is generally turbulent.
Froude number is the important parameter in the general study of open channel
flow which is free surface flow. The balance of gravity forces and surface friction
forces controls the flow. Changes in channel cross-section and changes in the slope
cause changes and readjustments in the flow depth which may or not propagate
upstream.

Definitions
Hydraulic radius Rh - It is the ratio of flow area, A, to the wetted perimeter, P.
Hydraulic depth- It is the ratio of flow area to top width. This represents the
average depth of the section.
An open channel flow is termed laminar if Re < 500, turbulent if Re > 1250 and
transitional if 500 < Re < 1250
Froude number, Fr - defines as

u
Fr = √ (4.1)
gy

Flows with Fr < 1 are referred to as subcritical, Fr > 1 are referred to supercritical,
and Fr = 1 are referred to as critical.

39
4.2 Classification of Open Channel Flows

Open channel flows are generally classified according to the rate of change of free
surface depth:

(i) Uniform Flow - the depth and velocity remain constant along the length of
flow. For such flow the slope and area should be uniform. The slope of the free
surface is governed by the way in which the slope and the area change. The
value of Froude number characterizes the nature of the flow in such situations.

(ii) Gradually varying flow - the depth changes gradually, due to area or slope
changes.

(iii) Rapidly varying flow - the slope change rapidly or suddenly, due to sudden
changes of area or slope.

4.2.1 Uniform Flow

Also known as flow at normal depth. This is the simplest and common type of flow
that occurs when conditions are steady and slope is not steep. It is also non
accelerating flow.

For steady uniform incompressible flow the height of the water level and area are
constant, hence the velocity is constant.
Continuity Equation - ρ1 A1 u1 = ρ2 A2 u2 and for incompressible flow (ρ1 = ρ2 )
and constant area (A1 = A2 ), then u1 = u2 , i.e constant velocity.
Momentum conservation - The velocity is constant (no acceleration), hence the
momentum flux through the control surface is zero. As the pressure distribution is
hydrostatic the net pressure force on the control volume is zero. Only body force
due to gravity and friction forces/ on the wetted surfaces, act on the control volume.

− Ff + mg sin θ = 0 (4.2)

40
 Ff = mg sin θ = mgSb (4.3)

where mg sin θ is the component of gravity force parallel to the flow, Ff is the
friction force on the wetted surface, and Sb is bed slope, which is defined as the ratio
of change in elevation over a length with the length, i.e Sb = sin θ.
Energy conservation - Applying Bernoulli’s equation between point 1 and 2, we
have
p1 u2 p2 u2
+ 1 + z1 = + 2 + z2 + hl (4.4)
ρg 2g ρg 2g
where hl is the head loss due to friction between point 1 and 2.
Since pressure is essentially hydrostatic at any cross-section

p1 p2
= y1 = y2 (4.5)
ρg ρg

Therefore Eq. 4.4 can be rewritten as

u21 u2
y1 + + z1 = y2 + 2 + z2 + hl (4.6)
2g 2g

For steady flow, u1 = u2 , y1 = y2 and for a length L, then

hl = z1 − z2 = LSb (4.7)

The head loss due to friction in steady flow between two sections equals the change
in elevation of the bed.

Specific Energy
Specific energy, E, is defined as the energy of the fluid referred to the bottom of the
channel as the datum.

Thus, the specific energy at any section is given by

u2
E= +y (4.8)
2g

41
Equation 4.8 may be written in terms of volumetric flow rate by substituting from
continuity the value of u:
(Q/A)2
E= +y (4.9)
2g
For a rectangular channel whose width is b, A = by, and if q is defines as flow rate
per unit width, then Eq. 4.9 can be written as
(qb/by)2 (q/y)2
E= +y = +y (4.10)
2g 2g
The specific energy for uniform flow is constant. The energy grade line, hydraulic
grade line and the channel bed are all parallel.

Chezy’s Equation for Discharge

For the steady uniform flow control volume shown above with length L,
Gravitational force = ρALg sin θ
Frictional force = τw P L, where τw is wall shear stress. Hence, substituting this
values into Eq. 4.3, gives
τw P L = ρALg sin θ (4.11)

or
τw = ρgRh Sb (4.12)

where Rh = A/P . For large values of Reynolds number the friction factor is inde-
pendent of Reynolds number and wall shear stress is proportional to the dynamic
pressure u2 /2g and is independent of viscosity. Hence τw = Kρu2 /2g where K is a
constant of proportionality. Substituting in Eq. 4.12, gives
u2
Kρ = ρgRh Sb (4.13)
2g
which result into
2gg
u2 = Rh Sb (4.14)
K
Equation 4.14 can be written as
p
u2 = C Rh Sb (4.15)

42
 2gg
where C = K
is a constant normally referred to as Chezy’s coefficient, which
is determined experimentally. Eq. 4.15 is known as Chezy’s Equation. The
discharge can be calculated as
p
Q = ρAV = ρAC Rh Sb (4.16)

The dependence of u on hydraulic radius, Rh , was modified by Manning through

experimental correlation to give what is commonly known as the Manning’s equation
given as
2/3 1/2
Rh Sb
u= (4.17)
n
n is known as Manning resistance coefficient and is dependent on surface materi-
als. Its values are obtained experimentally depending on the surfaces of the wetted
perimeter. Typical values of n are given in the table below;

Wetted perimeter n
Natural rivers 0.035
Flood plains, pastures 0.035
Excavated gravelly earth channel 0.022
Grass lined channels 0.010
Concrete lined channels 0.012
Brick worked lined channels 0.015
Rubble masonsry channels 0.025
Brass 0.011
Steel 0.002

The Optimum Cross-section of a Channel

The cost excavation of a channel can be reduced if the cross-section is minimized and
yet convey a given discharge at a given slope. Conversely, for a given cross-section
the discharge should be maximum for the section to be economical.
Since Q = Au, then from Manning’s formula;
1 2/3 1/2
Q = A Rh Sb (4.18)
n

43
or  2/3
1 A 1/2 A5/3
Q=A Sb ∝ 2/3 (4.19)
n P P
For a given Q, n, and Sb , A is minimum when P is minimum.
For a given A, n, and Sb , Q is maximum when P is minimum. Thus to either
maximize Q or minimize A, then we need to minimize P. P is minimum when
dP
=0 (4.20)
dy

(a) Rectangular section

A = by and P = b + 2y
Thus for a given value of A
A
P = + 2y (4.21)
y
Hence
dP A
=− 2 +2=0 (4.22)
dy y
Therefore,
A = 2y 2 = by ⇒ b = 2y (4.23)

This is equivalent to inscribing a semi-circle tangential to the beds and sides.

Since
A by 2y.y y
Rh = = = = (4.24)
P b + 2y 2y + 2y 2
Then the most economical rectangular section is when Rh = y/2, i.e half the depth
of flow. (b) Trapezoidal section
P = b + 2(y 2 + x2 )1/2 and A = by + yx. But x = y tan θ
Therefore

P = b + 2(y 2 + y 2 tan2 θ)1/2

A = by + y 2 tan θ ⇒ b = (A − y 2 tan θ)/y (4.25)

For a given value of A,

A
P = − y tan θ + 2y(1 + tan2 θ)1/2 (4.26)
y

44
and therefore
dP A
= − 2 − tan θ + 2(1 + tan2 θ)1/2 = 0 (4.27)
dy y
Substituting for A from Eq. 4.25 yields;

b + 2y tan θ = 2y(1 + tan2 θ)1/2 (4.28)

But b + 2y tan θ is top width and y(1 + tan2 θ)1/2 is sloping side.
Therefore,

Top width = 2 X sloping side

Or from Eq. 4.28

b
+ y tan θ = y(1 + tan2 θ)1/2 (4.29)
2
If we draw CD ⊥ AB
y CD
cos θ = AB
= AC
but AC = AB ⇒ y = CD = Radius of a semi-circle center C.
The hydraulic radius Rh is
A by + y 2 tan θ (b + y tan θ)y (b + y tan θ)y y
Rh = = 2 1/2
= = = (4.30)
P b + 2y(1 + tan θ) b + (b + 2y tan θ) 2(b + y tan θ) 2
In order to determine the optimum side slope (optimum value of angle θ) for a given
section of area A and depth y.
For analysis, we let tan θ = a , so that Eq. 4.26 becomes
A
P = − ya + 2y(1 + a2 )1/2 (4.31)
y
For optimum values,
dP 1
= −y + 2y × 2a(1 + a2 )−1/2 = 0 (4.32)
da 2
2a
⇒1=
(1 + a2 )1/2
p
Hence, a = 1/3, i.e θ = 300

45
(c) Triangular section
A = y 2 tan θ and P = 2y sec θ
Thus for a given value of A
1/2 1/2 1/2
A sec2 θ A(1 + tan2 θ)
  
A
P = 2 sec θ =2 =2 = 2[A(tan θ+cot θ)]1/2
tan θ tan θ tan θ
(4.33)
Therefore,
d(P 2 ) d
= 4[A(tan θ + cot θ)] = 0 (4.34)
dθ dθ
⇒ sec2 θ − cosec2 θ = 0 (4.35)

Hence,
cos2 θ = sin2 θ ⇒ θ = 450

The optimum half angle is 450 .

 1/2
A
y= = A1/2 (4.36)
tan 450

A y2
Rh = = = 0.353y (4.37)
P 2y sec 450

4.2.2 Gradually varied and Rapidly varied flows

The variation of the flow depth could be as a result of the changing bottom slope,
changing cross sectional area, and obstruction across a portion of the channel.
A gradually varying flow is when dy/dx <<< 1.
A rapidly varing flow when dy/dx ' 1.
Hydraulic Jump
Under certain conditions in an open channel flow, it is possible that the fluid depth
will change very rapidly over a short length without any change in the channel
configuration. Such changes may be approximated as discontinuity with dy/dx = ∞.
This discontinuity is called a hydraulic jump.
Consider an hydraulic jump that occur in a rectangular channel. Assume a one-
dimensional flow and consider only net changes and not the complex flow within the

46
hydraulic jump. For the short distance within the hydraulic jump, assume that the
shear forces are negligible.
Momentum equation between point 1 and 2 is

F1 − F2 = ρQ(u2 − u1 ) = ρu1 y1 b(u2 − u1 ) (4.38)

Pressure forces are hydrostatic and therefore

y1 ρgby12
F1 = ρgA1 ȳ = ρgby1 = (4.39)
2 2
Similarly, F2 = (ρgby22 )/2. Substituting this into Eq. 4.38 and re-arranging yields;
y12 y22 u1 y1
− = (u2 − u1 ) (4.40)
2 2 g
Continuity equation requires that

y 1 u1 b = y 2 u2 b (4.41)

and application of Bernoulli’s Eq. between 1 and 2 gives

u21 u2
y1 + = y2 + 2 + hl (4.42)
2g 2g
Substituting for u2 in Eq. 4.40 using 4.41 gives
y12 y22 u21 y1 y1
 
− = −1 (4.43)
2 2 g y2
which can be simplified to
 2  
y2 y2
+ − 2F r12 = 0 (4.44)
y1 y1
Equation 4.44 is a quadratic equation in y2 /y1 whose solution is
   
y2 1
q
2
= −1 ± 1 + 8F r1 (4.45)
y1 2
p
1 + 8F r12 > 1 (always) and since y2 /y1 must always be positive, then the negative
solution of Eq. 4.45 is not possible. And hence
   
y2 1
q
= −1 + 1 + 8F r12 (4.46)
y1 2

47
y2 /y1 can be plotted against Fr using Eq. 4.45.

Using Eqs. 4.41 and 4.42, the head loss across the jump can be expressed as
 2 !
u21 y1
hl = (y1 − y2 ) + 1− (4.47)
2g y2

Dividing through by y1 yields

"  2 #
hl y2 F r12 y1
=1− + 1− (4.48)
y1 y1 2 y2

Equation 4.48 indicates that the head loss will be negative if F r1 < 1 and since
negative head loss violate the second law of thermodynamics (viscous effects dissi-
pate energy; they can not create energy). It is therefore impossible to produce a
hydraulic jump if F r1 < 1. A flow must therefore be supercritical to be able to
produce a hydraulic jump. In open channel flow, hydraulic jump occur in dam spill
ways, sluice gate and weirs. It is useful in dissipating kinetic energy and thus pre-
vent the scouring of the channel bottom by agitation. The importance of hydraulic
jump analysis is the prediction of occurrence, size and location of the jump.

Exercise

(i) Water flows in a rectangular channel at depth of 30 cm and with a velocity of

16 m/s. If a downstream sill forces a hydraulic jump, what will be the depth
and velocity downstream of the jump? What head loss is produced by the
jump?

(ii) Determine the flow rate of water through a rectangular channel 3 m wide with
a flow depth of 1 m. The bed slope is 1 in 2500. Find also the wall shear
stress. Assume Manning’s constant = 0.11.

(iii) Determine the slope required for a flow of 1500 litres of water per second for
a brass pipe of 2 m diameter flowing half full.

48
(iv) Find the most economical dimensions of an irrigation channel of rectangular
section from the following data:
Discharge = 15 m3 /s
Bed fall = 0.3 m per km
Manning’s constant, n=0.015

(v) If the section in Q4 was a triangle, find the most economical dimensions.

(vi) Find the most economical dimensions of an irrigation channel of trapezoidal

section from the following data:
Discharge = 20 m3 /s
Bed fall = 0.5 m per km
Manning’s constant, n=0.025

49
CHAPTER 5

Dimensional Analysis and Similarity Laws

Dimensional analysis is a useful tool in reducing the number of experiment, which

could be performed to understand a particular situation (important in planning
experiments). The number of parameters can be reduced generally to three by
grouping relevant variables to form dimensionless parameters.

5.1 Methods of Determination of Dimensionless Group

(i) Intuitive method - this method relies on basic understanding of the phe-
nomenon and then identifying competing quantities like types of forces or
lengths etc. and obtaining ratios of similar quantities.

(ii) Rayleigh method - A functional power relation is assumed between the param-
eters and then the values of indices are solved for to obtain the grouping.

(iii) Buckingham Pi Theorem - The application of this theorem provides a fairly

easy method to identify dimensionless parameters (numbers). This method is
illustrated in the next section.

5.2 Buckingham Pi Theorem

If a relation among n parameters exist in the form

f (q1 , q2 , ..., qn ) = 0,

then the n parameters can be group into n-m independent dimensionless ratios or
pi (π) parameters expressed in the form

g(π1 , π2 , ..., πn−m ) = 0,

or
π1 = g(π2 , π3 , ..., πn−m ),

where m is the number of dimensions required to specify the dimensions of all the
parameters; q1 , q2 , ..., qn .

50
 Table 5.1: Units and Dimensions of Variables
Variable Unit (SI) Dimension (MLTθ)
Mass kg M
Length m L
Time s T
Force N ML/T2
0
Temperature C or K θ
Area m2 L2
Volume m3 L3
Volume flow rate m3 /s L3 /T
Mass flow rate kg/s M/T
Velocity m/s L/T
Angular velocity Rad/s 1/T
Force N ML/T2
Pressure, stress, N/m2 M/LT2
Power W, J/s ML2 /T3
Density kg/m3 M/L3
Dynamic viscosity kg/ms, Ns/m2 M/LT
Kinematic viscosity m2 /s L2 /T

5.2.1 Determination of π Groups

The following are steps used in determination of π groups:

(i) List all the parameters that influence the phenomenon concerned. There are
normally three types of parameters in fluid flow; fluid properties, flow geome-
try, and flow parameters.

(ii) Select a set of primary dimensions. The most commonly used sets are MLT
(mass, length, and time) system, FLT (force, length and time) system, and
MLTθ.

51
(iii) List the dimensions of all parameters in terms of the chosen set.

(iv) Select from the list of parameters a set of repeating parameters equal to
the number of primary dimensions.

(v) Set up a dimensional equation with the repeating set and one of the remaining
parameters, in turn to obtain n-m such equations, to determine n-m number
of π terms. The equation is in the form

π = qm+1 q1a .q2b .q3c , ..., qm

d
,

(vi) Check whether π terms obtained are dimensionless.

Exercise

(i) Pressure drop, ∆p per unit length in flow in a circular pipe is depended on:
flow velocity, u, diameter of the pipe, D, density of fluid, ρ, and dynamic
viscosity, µ. Determine the π groups.

(ii) Drag force on a sphere is affected by flow velocity, u, diameter, D, density of

fluid, ρ, and dynamic viscosity, µ. Determine the π groups.

(iii) The force impacted by water on a turbine is affected by diameter, D, velocity,

u, density, ρ, viscosity, µ and rotational speed, N. Determine the π groups.

5.3 Similitude and Model Testing

In the engineering point of view model can be defined as the representation of

physical system that may be used to predict the behavior of the system in the
desired aspect. The system whose behavior is to be predicted by the model is
called the prototype. Models resemble the prototype but are generally smaller
in size, cheaper to build and test, and may be operated with different fluids, at
different pressures, velocities etc. The model should be similar to the prototype in
the following manner:

52
 (i) Geometric similarity - all the similar linear dimensions of the model and pro-
totype should have the same ratio.

(ii) Dynamic similarity - the ratio of forces on fluid elements at corresponding

points in the model and prototype should be the same.

(iii) Kinematic similarity - the streamline patterns for the model and prototype
should be the same. This is achieve when dynamic and geometric similarity
exist.

5.3.1 Conditions for Similarity Between Models And Prototypes

Dimensional analysis provides a good basis for laying down the conditions for simi-
larity. The PI theorem shows that the performance of any system (prototype) can
be described by a functional relationship of the form given by

π1p = f (π2p , π3p , ..., πnp )

The PI terms include all the parameters influencing the system and are generally
ratios of forces, lengths, energy etc. If a model is to be similar to the prototype and
also function similarly as the prototype, then the PI terms for the model should also
have the same value as that of the prototype or the same functional relationship as
the prototype, i.e
π1m = f (π2m , π3m , ..., πnm )

For such a condition to be satisfied, the model should be constructed and operated
such that simultaneously

π1m = π1p , π2m = π3p , ..., πnm = πnp (5.1)

Equation 5.1 provides the model design conditions. It is also called similarity
requirements or modelling laws.

53
5.3.2 Types of Model Studies

Model testing can be broadly classified on the basis of the general nature of flow
into four types. These are

(i) Flow through closed conduits - Flow through pipes, valves, fittings and
measuring devices are dealt under this category. The conduits are generally
circular, but there may be changes along the flow direction. As the wall shear
is an important force (viscous), Reynolds number is the most important
parameter. Compressibility effect is negligible at low mach numbers. (M<0.3).

(ii) Flow around immersed bodies - Aircraft, Submarine, cars and trucks and
recently buildings are examples for this type of study. Models are usually
tested in wind tunnels. As viscous forces over the surface and inertia
forces on fluid elements are involved in this case also, Reynolds number
of the model and prototype should be equal. When the flow speed increases
beyond Mach number 0.3 compressibility effect on similarity should be con-
sidered.

(iii) Flow with free surface - Flow in canals, rivers as well as flow around ships
come under this category. In these cases gravity and inertia forces are
found to be governing the situation and hence Froude number becomes the
main similarity parameter.

(iv) Flow through turbomachinery - Pumps as well as turbines are included

in the general term turbomachines. Similarity of this type is beyond the
scope of this course

Dimensionless groups

ρuD
(i) Reynold number, Re = µ

54
 u
(ii) Froude number, F r = (gl)0.5

u
(iii) Mach number, M = c

P
(iv) Euler number, Eu = ρu2

ρu2 l
(v) Weber number, W e = σ

Exercise

(i) A model of 1/8 geometric scale of a valve is to be designed. The diameter of

the prototype is 64 cm and it should control flow rates up to 1m3 /s. Determine
the flow required for model testing. The valve is to be used with brine in a
cooling system at -100 C. The kinematic viscosity of brine at the saturated
condition is 6.956 × 10−6 m2 /s. For model testing water at 300 C is used.
Kinematic viscosity is 0.8315 ×10−6 m2 /s

(ii) A 1/6 scale model of a submarine is tested in a wind tunnel using air of density
28.5 kg/m3 and viscosity 18.39 10−6 kg/ms at a speed of 36.6 m/s. Calculate
the corresponding speed and drag of the prototype when submerged in sea
water with density 1025 kg/m3 and viscosity 1.637 × 10−3 kg/ms

(iii) 1/30 geometric scale model is to be used to simulate the flow characteristic in
a canal to be used in an irrigation scheme. The canal is 25 m and has a flow
rate of 200 m3 /s. Determine the flow rate required to test the model. Also
determine the time scale for the model. Viscous and surface tension effects
may be neglected.

55