Sie sind auf Seite 1von 83

Electric Submersible Pumps Mohamed Dewidar 2013

Chapter 2


1
Centrifugal Pumps
Table of Content
Section Content Page
1 Definition 3
2 Theory 4
2.1 Centrifugal force
2.2 Peripheral velocity and head
2.3 Specific gravity
3 General concept 10
4 Energy 10
5 Energy equations for an ideal flow 12
6 Horsepower 12
7 The basic equations for centrifugal pumps 14
8 Characteristics of ideal pump and degree of
reaction 18
9 Impeller with finite number of vanes 21
10 Hydraulic losses in pump and plotting
characteristic curve 24
11 Pump efficiency 26
12 Similarity formulas 28
13 Specific speed and its relation to impeller
geometry 31
13.1 Definition
13.2 Specific speed basics
13.3 Specific speed derivation
14 Net positive suction head & cavitation 38
15 NPSH specific speed 41
16 System curve 42
17 Electric Submersible Pumps 44
17.1 Pump stages
17.2 Types of impellers
17.3 Impeller constructions
18 Recommended operating range 50
18.1 Impeller thrust
18.2 Impeller thrust washers
18.3 Shaft thrust
19 Pump configurations 61
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


2
20 Fluid viscosity effect on centrifugal pumps 64
21 Affinity laws 79
22 Friction loss in pipes 81
23 ESP design examples 83

















Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


3
Centrifugal Pumps
2.1. Definition
By definition, a centrifugal pump is a machine that
imparts energy to a fluid. This energy helps a liquid to
flow, rise to a higher level, or both.
The centrifugal pump is an extremely simple machine. It is
a member of a family known as rotary machines and consists
of two basic parts:
1. The rotary element or impeller fig (2.1) and
2. The stationary element or diffuser fig (2.2).



Fig (2.1) Impeller
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


4



Fig (2.2) Diffuser
The centrifugal pumps function is as simple as its design.
It is filled with liquid and the impeller is rotated.
Rotation imparts energy to the liquid causing it to exit
the impellers vanes at a greater velocity than when it
entered. This outward flow reduces the pressure at the
impeller eye, allowing more liquid to enter. The liquid that
exits the impeller is collected in the diffuser where its
velocity is converted to pressure before it leaves the
pumps discharge.
2.2. Theory
2.2.1. Centrifugal Force
A classic example of the action of centrifugal force
is shown in fig (2.3) below. Here, we see a ball of water
swinging in a circle. The swinging ball generates a
centrifugal force that holds the water in the ball. Now,
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


5
if a hole is bored in the ball, water will be thrown out.
The distance the stream carries (tangent to the circle) and the
volume that flows out (per unit time) depends upon the
velocity (in ft/sec) of the rotating ball. The faster the
ball rotates the greater the centrifugal force and
therefore the greater the volume of water discharged and the
distance it carries.

Fig (2.3)
The description above could be considered that of a simple
centrifugal pump. It demonstrates that the flow and head
(pressure) developed by a centrifugal pump depends upon the
rotational speed and, more precisely, the peripheral
velocity of its impeller (ball).
In the above figure, the string is in tension and this
means it pulls in both directions. The force pulling the
ball towards the middle is the centripetal force and the
force pulling to outward is the centrifugal force.
Consider the velocity vector before and after point P has
revolved a small angle d, fig (2.4).
The magnitude of v
1
and v
2
are equal so lets denote it simply
as v. The direction changes over a small period of time dt
by d radians.
ds is almost the length of an arc of radius r. If the angle
is small, the length of an arc is radius x angle, so it
follows that,



Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


6













Fig (2.4)
Example
Calculate the centrifugal force acting on a small mass of 0.5 kg rotating at
1500 rev/minute on a radius of 300 mm.
Solution
= 2N/60 = 2 x x 1500/60 = 157 rad/s
Centrifugal acc. =
2
R = (157)
2
x 0.3 = 7395 m/s2.
Centrifugal force = Mass x acc. = 0.5 x 7395 = 3697 N
2.2.2. Peripheral Velocity & Head
Gravity is one of the more important forces that a
centrifugal pump must overcome. You will find that the
relationship between final velocity, due to gravity, and
initial velocity, due to impeller speed, is a very useful
one.
If a stone is dropped from the top of a building it's
velocity will increase at a rate of 32.2 feet per second
for each second that it falls. This increase in velocity
is known as acceleration due to gravity. Therefore if we
ignore the effect of air resistance on the falling stone, we
can predict the velocity at which it will strike the
ground based upon its initial height and the effect of
acceleration due to gravity.
d
V
1

V
2
r
s
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


7
The equation that describes the relationship of velocity,
height, and gravity as it applies to a falling body is:
v
2
= 2gh
Where:
v = The velocity of the body in ft/sec
g = The acceleration due to gravity @ 32.2 ft/sec/sec (or
ft/sec2)
h = The distance through which the body falls
For example if a stone is dropped from a building 100 feet high:
v
2
= 2 x 32.2 ft/sec2 x 100 ft
v
2
= 6440 ft2/sec2
v = 80.3 ft/sec
The stone, therefore, will strike the ground at a velocity of 80.3 feet per
second.
This same equation allows us to determine the initial
velocity required to throw the stone to a height of 100
feet. This is the case because the final velocity of a
falling body happens to be equal to the initial velocity
required to launch it to height from which it fell. In the
example above, the initial velocity required to throw the
stone to a height of 100 feet is 80.3 feet per second, the
same as its final velocity.
The same equation applies when pumping water with a
centrifugal pump. The velocity of the water as it leaves
the impeller determines the head developed. In other words
the water is thrown to a certain height. To reach this
height it must start with the same velocity it would attain
if it fell from that height.
If we rearrange the falling body equation we get:
h = v
2
/2g
Now we can determine the height to which a body (or water)
will rise given a particular initial velocity. For example,
at 10 Ft per Sec:
h = 10
2
ft
2
/sec
2
/ (2 x 32.2) ft/sec
2

h = 100 ft
2
/sec
2
/ 64.4 ft/sec
2
= 1.55 ft
If you were to try this with several different initial
velocities, you would find out that there is an interesting
relationship between the height achieved by a body and its
initial velocity. This relationship is one of the
Fundamental laws of centrifugal pumps and we will review
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


8
it in detail a little later. As a finale to this section,
let's apply what we have learned to a practical application.
Example:
For an 1800 RPM pump, find the impeller diameter necessary to develop a
head of 200 feet.
First we must find the initial velocity required to develop a head of 100 feet:
v
2
= 2 gh
v
2
= 2 x 32.2 ft/sec2 x 200 ft
v
2
= 12880 ft2/sec2
v = 113 ft/sec
We also need to know the number of rotations the impeller undergoes
each second:
1800 RPM / 60 sec = 30 RPS
Now we can compute the number of feet a point on the impellers rim
travels in a single rotation:
113 ft/sec / 30 rotations/sec = 3.77 ft/rotation
Since feet traveled per rotation are the same as the circumference of the
impeller we can compute the diameter as follows:
Diameter = Circumference /
Diameter = 3.77 ft / 3.1416
Diameter = 1.2 ft or 14.4 inch
Therefore an impeller of approximately 14.4" turning at 1800 RPM will
produce a head of 200 Feet.
2.2.3. Specific Gravity
Why head is usually expressed in feet?
The Specific Gravity of a substance is the ratio of the
weight of a given volume of the substance to that of an
equal volume of water at standard temperature and pressure
(STP).
Assuming the viscosity of a liquid is similar to that of
water the following statements will always be true
regardless of the specific gravity:
1. A Centrifugal pump will always develop the same head in
feet regardless of a liquids specific gravity.
2. Pressure will increase or decrease in direct proportion
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


9
to a liquids specific gravity.
3. Brake HP required will vary directly with a liquids
specific gravity.
The Fig (2.5) below illustrates the relationship between
pressure (in psi) and head (in ft) for three liquids of
differing specific gravity.


Fig (2.5)
We can see that the level in each of the three tanks is 100
feet. The resulting pressure at the bottom of each varies
substantially as a result of the varying specific gravity.
If, on the other hand we keep pressure constant as measured
at the bottom of each tank, the fluid levels will vary
similarly.
A centrifugal pump can also develop 100' of head when
pumping water, brine, and kerosene. The resulting
pressures, however, will vary just as those seen in the
above Figure. If that same pump requires 10 HP when pumping
water, it will require 12 HP when pumping brine and only 8
HP when pumping kerosene.
The preceding discussion of Specific Gravity illustrates
why centrifugal pump head (or pressure) is expressed in
feet. Since pump specialists work with many liquids of
varying specific gravity, head in feet is the most
convenient system of designating head.
When selecting a pump, always remember that factory tests
and curves are based on water at STP. If you are working
with other liquids always correct the HP required for the
specific gravity of the liquid being pumped.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


10
2.3. General Concept
As previously mentioned the pump is a machine which
imparts head to a fluid.
From the physical aspect, the work of a pump consists in
transforming the mechanical energy of a motor (drive) into
fluid energy. i.e., imparting power to a flow of fluid
passing through it. The energy imparted to the fluid in the
pump enables the former to overcome hydraulic resistances
and rise to a geometric elevation.
The key idea is that the energy created by the centrifugal
force is kinetic energy. The amount of energy given to the
liquid is proportional to the velocity at the edge or vane
tip of the impeller. The faster the impeller revolves or the
bigger the impeller is, then the higher will be the velocity
of the liquid at the vane tip and the greater the energy
imparted to the liquid.
The pump diffuser catches the liquid and slows it down. In
the discharge nozzle, the liquid further decelerates and its
velocity is converted to pressure according to Bernoullis
principle.
Let's outline the steps a liquid encounters as it moves from
suction to discharge.
1. Rotation of the impeller and the shape of the vane
entrances forces liquid to move from the eye of the
impeller into its vanes.
2. During rotation the curved shape of the vanes causes
liquid to continue to flow towards the vane exits.
3. This flow causes a partial vacuum at the eye which allows
atmospheric or some other, outside pressure to force more
liquid into the eye thus regenerating the entire process.
4. As liquid flows through the vanes, it gains velocity and
reaches its maximum velocity just as it exits the vanes.
5. Upon exiting the vanes, liquid enters the diffuser where
most of its kinetic energy of motion is transformed into
pressure energy.
2.4. Energy
Liquid can possess three forms of hydraulic energy:
Potential energy due to elevation.
Kinetic energy due to velocity, and,
Pressure energy due to weight or force.
In physics we say that energy can neither be created nor
destroyed it can only change its state or form. Therefore
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


11
these three forms of energy must be able to live in harmony
and their quantities, at a given point in time, will follow
a simple law known as The Conservation of Energy. Now, for
our purposes, we can pretty much eliminate potential energy
because there is little or no elevation change from a pump's
suction to its discharge. But as we saw above, a liquid's
velocity undergoes a large change as it moves through the
impeller and diffuser. Since velocity is, in fact, energy it
must be replaced by some other form as it decreases. That
other form of energy is pressure and it arises of it own
accord as the diffuser volume increases and velocity
decreases.
The fig (2.6) below shows a pipe with water flowing from
left to right at Q gpm. As water reaches the center of the
pipe it encounters a section that has a reduced diameter but
a short distance away the pipe returns to its original
diameter. Notice the three pressure gauges the one on the
left points to 12 o'clock while the one in the center is at
10 o'clock.













Fig (2.6)
The gauge on the right displays just a little less than the
one on the far left (due to friction loss). In other words
pressure drops as water enters the constricted area of the
pipe but it returns to nearly its original pressure as it
exits the constricted area.
What is happening here? Well, if flow is to remain constant
(Q gpm) the velocity of the water must increase as it
travels through the constricted area of the pipe. We see
this in nature when a slow moving river enters and exits a
narrow gorge. And, it confirms our statement about energy as
one form (velocity) increases, another form (pressure) must
decrease and vice versa. Bernoulli's theorem states that
during steady flow the energy at any point in a conduit is
the sum of the velocity energy, pressure energy, and the
potential energy due to elevation. It also says the sum will










Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


12
remain constant if there are no losses. In our example
above, the small loss in pressure seen in the right hand
portion of the pipe is due to increased friction in the
narrow section. Depending upon the circumstances, pressure
loss due to friction may be a result of the generation and
dissipation of heat or it could be due to a small increase
in velocity due to a change in laminar flow. As friction
increases laminar flow becomes less symmetrical which,
essentially, reduces the diameter of the conduit.
2.5. Energy equation for an ideal flow
The pumping head is the sum of increase in pressure
head (static head) and the increase in specific kinetic
energy (dynamic head).


The second term (dynamic head) is usually smaller than the
first term (static head) and if the intake and discharge
pipes have the same diameter (d
1
=d
2
, whence v
1
=v
2
) and
1
=

2
, it is zero and

The rate of discharge of a pump is also called its delivery
or capacity, denoted Q.
2.6. Horsepower
The power of a pump is defined as the energy imparted
by the pump to the fluid flow per second.
Horsepower is the horse which could lift 150 pounds a height
of 220 feet in 1 minute.




Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


13














Q = m
3
/sec
H = m
= kg/m
3

Q = m
3
/sec
H = m
= Specific Gravity

Q = GPM
H = Ft
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


14
= Specific Gravity

Like any other driven machine, a pump consumes more power
than its given off. The ratio of the actual power developed
by the pump (water horsepower) to the power supplies by the
pump (shaft horsepower) given the efficiency of the pump:


HP = Water HP
BHP = Brake HP (Shaft HP)

Hence, the shaft horsepower


and, taking into account Eq. (2.1).

This is the equation used in choosing pump drives.
Total or overall, pump efficiency takes into account three
types of energy losses:
- Hydraulic losses due to friction and turbulence.
- Volume losses due to leakage through internal passages.
- Mechanical losses due to mechanical friction in bearings,
packing, etc.
2.7. The basic equations for centrifugal pumps
A centrifugal pump operates in the following manner;
the principal working unit is a vaned rotor, called
impeller which is made to revolve at high speed. It
accelerates the incoming fluid, increasing both pressure and
absolute velocity of the latter, driving it to the diffuser.
By interaction between the vanes and fluid, the mechanical
energy of the drive is transformed into the energy of flow.
In the diffuser the kinetic energy of the fluid is partly
converted to pressure energy.

The impeller of a centrifugal pump, fig (2.7), consists of
two disks like walls, called shrouds, one of which is
mounted on the shaft. The other shroud, coupled the former
by the vanes, has a hole in the centre, called eye. The
vanes are curved, cylindrical or have complex surfaces. The
fluid enters the impeller along the axis of rotation through
the eye, flows radially outward between the vanes, and is
discharged around the entire circumference into the
diffuser.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


15

Fig (2.7)


Fig (2.8) Flow pattern through centrifugal pump impeller

The motion of the fluid through the passages between the
vanes can be regarded as consisting of two motions fig
(2.8):
- Motion of transport (rotation of the impeller).
- Motion relative to the impeller.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


16
Hence, the absolute velocity C of the fluid can be found as
the vector sum of the peripheral velocity U and the
relative velocity W.
Taking a fluid particle sliding along a vane, one can
construct a velocity parallelogram for the entrance of the
particle to and discharge form, the vane, assuming the
relative velocity W to be tangent to the vane and the
peripheral velocity U tangent to the corresponding circle.
A similar velocity parallelogram can be constructed for any
point on the vane. The subscript 1 refers to the entrance
section, and subscript 2 to exit section of the vane.
The angle between the vectors of the peripheral and absolute
velocities is , and the angle between the tangent to the
vane and the tangent to the circumference is , with the
corresponding subscripts.
In general case the angle changes with the pump
performance, i.e., the speed of rotation n of the impeller
(the velocity U) and the discharge Q (the velocity W).
Angle determines the inclination of a vane at every point
and consequently, does not depend on pump performance.
In order to develop the basic equation of centrifugal pump
theory, we shall accept the following two assumptions:
1. The impeller consists of an infinite number of uniform
vanes of zero thickness (Z = , =0) This means that we
assume such flow in the passages between the vanes in
which the geometry of all stream tubes in the relative
motion is identical and corresponds exactly to the vane
geometry, and the velocities depend only on the radius
and are the same on a circle of a given radius. This
possible in the case when each differential stream tube
is guided by its vane.
2. The pump efficiency is unity (=1), i.e., there is no
energy losses in the pump and shaft horsepower is
converted completely into water horsepower.
This is possible in the case of an ideal fluid, no
leakage in the pump and no mechanical friction in packing
and bearings.
Thus, to facilitate our theoretical investigation of a
centrifugal pump we have substantially idealized its
performance. We shall call such a centrifugal pump, in which
Z = and = 1, an ideal pump. After considering the
theory of the idealized pump we shall, naturally, proceed to
deal with real pumps.
Let us develop two equations:
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


17
1. The power equation, it means that the power supplied to
the impeller shaft is equal to the energy imparted every
second to the fluid in the pump,



) 5 . 2 ( =
Ht
Q T e
2. The equation of momentum, it means that the torque acting
on the pump shaft is equal to increase in the angular
momentum of the fluid in the impeller per second.
Denoting by r
1
the radius of the cylindrical surface on
which the entrance edge of the vanes are located, and by
r
2
the peripheral radius of the impeller, we have,



From Esq. (2.5) and (2.6) the head developed by an idealized
pump is,

This is the basic equation not only for centrifugal pumps
but also for all rotodynamic machines, such as fans,
compressors, and turbines.
Attention should be paid to the fact that the head developed
by an idealized centrifugal pump in terms of column of
pumped liquid does not depend on type of liquid (i.e. on its
specific weight).
As a rule, a liquid entering the impeller has no whirl
component. In the vane passages it moves in radial
direction. This means that the vector C
1
is pointed along
the radius and the angle
1
= 90
O
.
Consequently, the second term in Eq. (2.7) vanishes and the
equation takes the form,

Where U
2
= r
2
= peripheral velocity at vane exit;
C
2u
= projection of absolute velocity on direction of
peripheral velocity, i.e. the tangential component of
absolute velocity C
2
.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


18
Equation (2.8) shows that for centrifugal pump to deliver
high head the peripheral velocity must be greater and,
secondary, the vector C
2u
must be large enough i.e.
sufficient whirl should be imparted to the fluid. The former
is achieved by increasing the speed of rotation and impeller
diameter, the later is attained by providing a sufficient
number of vanes of suitable size and shape.
2.8. Characteristics of ideal pump and degree of
reaction
Equation (2.8) is inconvenient for calculation as does
not contain the rate of discharge Q. Therefore let us
rewrite it to express the head H
t
as a function of
discharge Q and impeller radius.
From the velocity triangle for the impeller exit, fig (2.9),


C
2m
= projection of absolute exit velocity on radius i.e.
the radial component of vector C
2
.
The rate of discharge through the impeller can be expressed
in terms of radial component C
2m
and impeller radius as
follows:

(Fig 2.9) Velocity triangle at impeller exit

Where
b2
= width of vane slot at exit, hence,


Substitution of this expression into equation (2.9) yields

Substituting the obtained expression (2.11) for the
tangential velocity component U
2u
in Eq. (2.8) we obtain
another form of the basic ideal pump equation:

Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


19
This equation can be used to plot the theoretical
characteristics of an idealized centrifugal pump, i.e.
curve of the head generated by the pump as a function of
the discharge for constant speed of rotation. It is from
equation (1.12) that the characteristic curve of such a
pump is a straight line the inclination of which depends on
the value of the vane angle
2
. The following are three
possible cases:

1. Angle |
2
<90
O
, in this case cot|
2
is positive and the head
Ht
decrease with the charge increasing.
2. Angle |
2
=90
O
, cot|
2
=0, and head
Ht
does not depends on
the discharge and equal to U
2
2
/g
3. Angle |
2
>90
O
, cot|
2
is negative and the head
Ht

increases with discharge.
These three theoretical pump characteristics are shown in
Fig (2.10).


Fig (2.10) Ideal centrifugal pump characteristics
The corresponding vane shapes and velocity parallelograms
for the same value of U
2
and C
2m
are presented in Fig (11.a,
b, and c).

Fig (2.11) Vane design and velocity parallelograms
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


20
It thus follows that the optimum head is produced by a
forward curved vane, when
2
>90
O
and the head is highest. In
practical, however, the efficiency of such pump is low, and
the performance of backward curved vanes at
2
<90
O
is found
to be preferable.
Backward curved vanes are in fact more commonly used, the
vane angle usually being about 30
O
. Radial vanes (
2
=90
O
)
are also employed, but the result is lower efficiency.
In order to understand why pump efficiency falls with the
angle
2
increasing we must examine the component of the
head H
t
and the way in which the relation between them
changes with
2
.
The head H
t
, or what the same thing, the total increase in
the specific energy of a fluid in an impeller, comprises
the increase in specific energy of pressure and specific
kinetic energy, i.e.


Or introducing another notation,

Expressing the velocities v
1
and v
2
in terms of their radial
and tangential components, we have,

Assuming the intake and exit areas of the impeller to be
approximately equal, we can consider that C
1m
= C
2m
.
Furthermore, as pointed out before, there is usually no
pre-rotation of the fluid at the impeller intake, and C
1u
=
0. Consequently, instead of the foregoing we have,

Taking this expression into account, we can now find from
Eq. (2.13) the so-called degree of reaction of the pump,
which is the ratio of the head imparted to the fluid by the
pressure increase to the total head:

Using Eq. (2.8), the latter expression can be written as
follows,

Whence finally, after substituting for C
2u
from Eq. (2.9),
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


21

It will be observed from this expression that the greater
C
2u
/U
2
and the smaller angle|
2
, the greater the portion of
the head
Ht
that produced by pressure increase, i.e. the
higher the degree of reaction of the pump. With angle |
2

increasing the portion of the head
Ht
representing the
increase in the kinetic energy becomes grater. The kinetic
energy, in turn, is associated with higher exit velocity of
the fluid from the impeller, which results in considerable
energy losses and lower pump efficiency. That is why it is
not expedient to use vanes with large values of|
2
, i.e.
forward-bent vanes.
It follows from Eq. (2.15) that for radial vanes ( |
2
=90
O
)
the degree of reaction is and at |
2
<90
O
it is more than
but less than unity.
2.9. Impeller with finite number of vanes
So far we have been investigating the performance of an
idealized centrifugal pump with an infinite number of vanes
and unity efficiency.
In going over to real pumps we shall begin with eliminating
the first assumption, retaining the second for the time
being.
Real pumps usually have from six to twelve vanes. The
relative flow through the vane passages is not so laminar
as assumed before and the velocity distribution is not
uniform. On the leading surface of the vane, denoted by
plus sign in fig (1.11a), the pressure is higher and the
velocity is lower, and the velocity distribution in the
passage is approximately as shown.
The velocity distribution can be regarded as the resultant
of two flows: one with uniform velocity distribution, as
was the case when = z (fig 2.11b), and the rotational flow
inside the passage in the opposite direction of the
rotation of the impeller (Fig 2.11c). In pure form the
rotational flow is present when the discharge through the
impeller is zero (Q=0).
In view of the non-uniformity of the distribution of the
relative and absolute velocities in the vane passages when
the number of vanes is finite, the mean velocity for a
circle of given radius is introduced. Our interest is the
mean value of the tangential component of the absolute exit
velocity C
2u
which determines the head developed by the
pump. This component is smaller for a finite number of
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


22
vanes than for an infinite number because the less the
number of vanes the less the whirl imparted to the fluid by
the impeller. In the absence of vanes (z=0) the whirl is
zero, i.e. C
2u
= 0, and the fluid (in the ideal case) issues
from the impeller in a radial direction.
A reduction of the velocity C
2u
in passing over to a finite
number of vanes is also accounted for by the prerotation
mentioned before. This relative motion gives rise to an
additional absolute velocity C
2u
at the outer periphery of
the impeller, Fig 2.12c, which is directed opposite to C
2u

and, hence, is subtracted from the latter.

Fig (2.12) Flow pattern through vane passage
Owing to this velocity triangle at the impeller exit
changes, In Fig (2.13), the solid lines give the velocity
vectors for finite number. The construction was made for
identical values of U
2
and C
2m
, i.e. for identical
rotational speeds and rates of discharge. The primed values
are for the case of a finite number of vanes.
A reduction of the tangential component C
2u
in transition to
a finite number of vanes results in a drop in the pumping
head. The head that would have been generated by a pump if
there were no head losses inside the pump is called
theoretical or ideal head denoted by
Htz
. From Eq. (2.8), we
have

We shall call the ratio of
Htz
to
Ht
the vane number
coefficient ():

So, the head in equation is,

Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


23

Fig (2.13) Change of velocity triangle in going over to
finite number of vanes
The problem now is to determine the numerical value of .
Obviously, the coefficient depends first and foremost on
the number of vanes z, through it is also affected by the
length of the vanes, which depends on the ratio (r
1
/r
2
) and
on the angle of inclination of the vanes|
2
.
Theoretically, investigations reveal that does not depend
on the operation condition of a pump, i.e. on
H
Q
pump
, or n.
It is wholly determined by impeller geometry and is
constant for a given impeller.
Without going into the theory of the effect of the number
of the vanes on the head, here is the conclusion of this
theory as represented by a formula for :

Where

= (0.55 to 0.65)+0.6 sin|
2


Here, for example, is the value of for |
2
=30
O
and r
1
/r
2
=
0.5

z 4 6 8 10 12 16 24
0.624 0.714 0.768 0.806 0.834 0.870 0.908

Thus at z-->, -->1

As the ratio between
Htz
and
Ht
is constant for a given
pump, the theoretical characteristic curve for a finite
number of vanes, like the characteristic curve of an
idealized pump with a uniform speed of rotation
(n=constant) is a straight line. At
O
90
2
= | , it parallel to
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


24
characteristic curve of an idealized pump, and at
O
90
2
< | it
intersects the later on the axis of abscissas as
Htz
=0 and
Ht
=0 at the same discharge
|
t
2
2 2 2
cot
2
u b r
Q =
2.10. Hydraulic losses in pump and plotting
characteristic curve
As stated earlier,
Htz
is the head that would have
been developed by a pump if there were no head losses
inside it. The actual head
Hpump
is less than the
theoretical head by the total losses inside the pump:

Where E
hpump
= total head losses in the pump (at intake, in
impeller and diffuser).
The ratio of the actual head to the theoretical head for a
finite number of vanes is called hydraulic efficiency,
denoted by q
h
thus,

Hydraulic efficiency is always higher than total efficiency
as it takes into account only head losses inside the pump.
It follows from Esq. (2.18) and (2.21) that,

H H H t
h
tz
h
pump
= = q q

Where
Ht
is given by Esq. (2.7) and (2.12)
The hydraulic losses inside the pump hpump
are conveniently
treated as the sum of the following two components.
1. Ordinary hydraulic losses i.e. losses due to friction
and partly to eddy formation inside the pump. As
turbulent flow is the common regime in a centrifugal
pump, this type of head losses increases approximately
as the square of the discharge and can be expresses by
the equation:


Where
k1
is constant depending on hydraulic efficiency and
the dimensions of the pump.
2. Shock losses at impeller and diffuser entrance. if the
relative velocity W
1
of the fluid at the entrance to a
vane passage is tangent to the vane, the fluid is
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


25
entering the impeller smoothly, without shock or eddy
formation. Shock losses in this case are nil. But this
possible only at some definite rated or normal discharge
Q
O
and corresponding radial entrance velocity (C
1m
)
0
,
(Fig 2.14).

Fig (2.14) velocity parallelogram at impeller entrance
If the actual discharge Q is more or less than the rated
discharge Q
O
and the radial entrance velocity C
1m
is more or
less than (C
1m
)
0,
the relative velocity W makes an angle
with the tangent to the vane and the fluid past the vane at
some positive or negative angle of approach. The effect is
that of the fluid impinging on the vane, with eddies forming
on the opposite side. Thus, energy is degraded in the impact
and eddy formation. The velocity parallelograms for the same
peripheral velocities corresponding to these non-rated
operating conditions are shown by broken lines in Fig
(2.14).
One of the parallelograms corresponds to the inequality Q >
Q
0
the other, to Q < Q
0
.
Shock losses can be assumed to vary as the square of the
difference between the actual discharge and the discharge
when they are zero, i.e.,

Shock losses at the diffuser entrance are of the same
nature as it the impeller intake, the minimum being at
about the same rate of discharge Q
0
and included in the
quantity h
2
.
The total loss of head inside the pump is the sum of the
two losses considered, i.e.,


Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


26
The characteristic curves of a pump at uniform rotation
speed (n = constant) are plotted as following:
First draw for H as a function of Q at n = constant the
theoretical characteristic curves for z = and for a finite
number of vanes Z.
These are straight lines, Fig (2.15). Below the Q axis,
plot curves for the two components h
1
and h
2
of the head
losses in the pump. Summing the coordinates of the two
curves gives the curve h
pump
as a function of the discharge.
Now, in accordance with Eq.(2.20) subtract h
pump
from H
tz

which gives the curve H
pump
= f(Q), i.e. the actual
characteristic of the pump for a constant speed.
The curve H
pump
= f(Q) in Fig (2.15) typical of a centrifugal
pump. The maximum value of head H
pump
is commonly obtained
neither at zero discharge nor at Q = Q
0
but at some
intermediate value of Q.
Plotting the characteristic curves by the method described
is not very accurate in view of the difficulty of
determining the coefficients k
1
and k
2
in Esq. (2.22) and
(2.23). Therefore the characteristics are commonly plotted
by direct experiment, i.e., in testing a pump.

Fig (2.15)
2.11. Pump Efficiency
The energy losses in a pump taken into account in
rating the overall efficiency q are:
1. Hydraulic losses, examined in the previous section and
determined by hydraulic efficiency [Eq.(2.21)]:

H
h H
H
H
tz
pump tz
tz
pump
h

= = q
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


27
2. Volumetric losses, due to leakage through clearance
spaces between impeller and diffuser. The impeller drives
the fluid from the suction pipe to discharge pipe, but
because of the pressure drop it produces some of the
fluid leaks back (Fig.2.16)

Fig (2.16) Leakage in pump impeller
The actual discharge of a pump at outlet is Q, and then
the discharge through the impeller is equal to,

Where q = internal leakage
Volumetric energy losses are evaluated by so called
volumetric efficiency

3. Mechanical losses, which include energy degradation due
to friction in packing and bearings as well as surface
friction of the fluid on the impeller. Denoting loss of
power due to friction by HP
m
and total shaft horsepower
by HP, the mechanical efficiency of a pump is:

The numerator of this expression represents the so called
hydraulic horsepower and can be expressed by the formula:

Now let us write the expression of the overall efficiency
of a pump as a ratio of the water horsepower to the shaft
horsepower:

Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


28
And multiply the numerator of this expression by HP
h
and the
denominator by the same quantity, using Eq. (2.28):

After re-arranging:

I.e the overall efficiency of a pump is equal to product of
its hydraulic, volumetric, and mechanical efficiencies.
2.12. Similarity Formulas
Let us investigate similar operating conditions of
homologous centrifugal pump. Hydrodynamic similarity is
provided by geometric, kinematic, and dynamic similarity.
As applied to centrifugal pumps, kinematic similarity means
similarity of the velocity triangles constructed for any
corresponding points of the impeller. Dynamic similarity is
insured by equality of the Reynolds' numbers of the flows
through the pumps in question.
When operating conditions of centrifugal pumps are similar
a proportionality between heads generated and lost, and
between actual delivery and leakage, is observed. It can
therefore be assumed that homologous pumps have the same
hydraulic and volumetric efficiencies. The mechanical
efficiencies of homologous pumps will generally vary, but
the total efficiency can nonetheless be assumed equal
without much error.
Let us consider similar operating conditions of two
homologous centrifugal pumps. The values referring to the
first pump are denoted by the additional subscript I and to
the second, by the subscript II, fig (2.17).

Fig (2.17) Similarity of centrifugal pump
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


29
Taking into account that the peripheral velocities of the
impellers are proportional to the number of rpm times the
respective impeller diameter D, the condition for kinematic
similarity at the impeller exit can be written as follows:

( n; U = r; so, U nD)
From Eq. (2.10),

And from geometric similarity
( )
( )
b
b
D
D
II
I
II
I
2
2
=
We can write from Eq. (2.30)

This means that the rate of discharge of homologous pumps
under similar operating condition conditions is proportional
to the rpm and cube of diameter.
From Eq. (2.8), the theoretical heads for an infinite number
of vanes are proportional to the product of the peripheral
and tangential velocities, while the vane-number coefficient
is the same for homologous impeller. Consequently,


Taking into account (2.30),

The actual head generated by the pump is,

H
H H tz
h
pump
= =q
But, as (
h
)
I
=(
h
)
II
instead of Eq. (2.32) we can write,

I.e., the actual heads developed by the homologous pumps
under similar operating conditions are proportional to the
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


30
square of the product of the rpm times the impeller
diameter.
From the expression of the water horsepower of the pump, Eq.
(2.2), and develop equations (2.31) and (2.32), we can
write the relationship between the power generated by
homologous pumps under similar operation conditions:

If we wish to consider similar operating conditions of the
same pump at different rotational speed n
1
and n
2
, the Eqs
(2.31), (2.32), and (2.33) become simpler, as D and are
the same all through and take the form,




Subscripts 1 and 2 are denoting the different rpm values.
Esq. (2.34) and (2.35) are used to compute pump
characteristics for different relative speeds. If the
relationship between H and Q at n
1
= constant is given, a
similar curve for n
2
= constant be obtained by computing the
abscissas of the points of the former curve (the rate of
discharge) proportionally to the rpm ratio, and the
ordinates (the head), proportionally to square of that ratio
(Fig 2.18)

Fig (2.18) Pump characteristics for different relative
speeds
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


31
Thus it is possible to calculate and plot the
characteristics of a pump for any desired rotative speed and
to draw a series of characteristic curves for the pump at
different values of n, fig (2.18).
The points A1, A2, A3, A4, on these curves joined by
coordinate relationship given in (2.34) and (2.35) represent
similar operating conditions. The other rows of points (B1,
B2, B3, B4), (C1, C2, C3, C4)...Etc, give a second, third,
etc, rows of similar operating conditions.
It is easy to develop the equations of the curves joining
the points of similar operating conditions. According to
Esq. (2.34) and (2.35) for any row of points we can write,

Q
H
Q
H
Q
H
2
3
3
2
2
2
2
1
1
= = =Const
1


Hence, for a row of similar operation conditions we have,
H =Const
1
Q
2

For another row,
H =Const
2
Q
2

Consequently, the points representing similar operating
conditions in an H-Q coordinate system are located on
parabola of second power through the point of origin shown
in Fig (2.18).
2.13. Specific Speed and its relation to impeller
geometry
There is a multitude of pump designs that are
available for any given task. Pump designers have needed a
way to compare the efficiency of their designs across a
large range of pump model and types. Pump users also would
like to know what efficiency can be expected from a
particular pump design. For that purpose pump have been
tested and compared using a number or criteria called the
specific speed (N
S
) which helps to do these comparisons. The
efficiency of pumps with the same specific speed can be
compared providing the user or the designer a starting point
for comparison or as a benchmark for improving the design
and increase the efficiency.
2.13.1. Definition
Specific speed is defined as the speed of an
imaginary pump geometrically similar in every respect to the
actual pump and capable of delivering unit quantity against
a unit head (gpm and ft in US (English) units or m
3
/s and m
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


32
in SI units). Mathematically, the specific speed for a pump
is given by,

Specific speed has dimensions of L
3/4
/T
3/2
. The specific speed
of all pumps, similar in shape, is the same regardless of
size.
Specific speed of a pump can be expressed in the non-
dimensional form as,

Where,
N
s
*
= the specific speed in non-dimensional form
n = impeller speed (rpm)
Q = the discharge rate m
3
/s or gpm
H = the head in m or ft
g = gravity acceleration (9.81 m/s
2
or 32.2 ft /s
2
)

Pumps are traditionally divided into 3 types, radial flow,
mixed flow and axial flow. There is a continuous change from
the radial flow impeller, which develops pressure
principally from the action of centrifugal force, to the
axial flow impeller, which develops most of its head by the
propelling or lifting action of the vanes on the liquid.
2.13.2. Specific Speed basics
The specific speed is largely related to the
impeller discharge angle, relative to the inlet. Pumps in
which the discharge of the impeller is directly radial to
the suction; that is, where the flow transitions rapidly
from one plane to the other, have a low specific speed and
are called radial flow impellers. The N will be in the
neighborhood of from 500 to 1700 (in US units). These pumps
will usually exhibit a low flow to head ratio. At the
other end of the scale, the fluid will be discharged from
the impeller along the same axis as it enters. These
impellers (or propellers) have high specific speeds,
generally above 9000 (in US units), and are referred to as
axial flow impellers (or propellers) and have a high flow
to head ratio.
Between these two extremes fall:
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


33
Mixed Flow impellers, with an N
S
range of from around
4000 to 9000 (in US units) begin to transition away from
the suction axis, but discharge between the axial and
radial angles, and generally exhibit a high flow to
moderate head ratio.
Francis Vaned Impellers, between mixed and radial flow,
with NS values from around 1700 to 4000. Francis vaned
impellers are frequently discussed in the industry, and
are simply impellers which have vanes curvature such that
the transition from the inlet axis to radial axis is
completed more gradually (sometimes is considered as high
speed radial flow).
The specific speed determines the general shape or class of
the impeller. As the specific speed increases, the ratio of
the impeller outlet diameter, D
2
, to the inlet or eye
diameter, D
1
, decreases. This ratio becomes 1.0 for a true
axial flow impeller.
Radial flow impellers develop head principally through
centrifugal force. Pumps of higher specific speeds develop
head partly by centrifugal force and partly by axial force.
A higher specific speed indicates a pump design with head
generation more by axial forces and less by centrifugal
forces. An axial flow or propeller pump with high specific
speed generates it's head exclusively through axial forces.
Radial impellers are generally low flow high head designs
whereas axial flow impellers are high flow low head designs.
2.13.3. Specific Speed (N
s
)derivation
The similarity formula obtained in the previous
section can be used to develop a useful practical factor for
calculating and designing centrifugal pumps which is
commonly known as specific speed.
From Eq. (2.31)

Substituting into Eq. (2.32) yields

Rearranging and raising to the power of we obtain

Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


34
This expression is valid no only two homologous pumps I and
II but for any number of homologous pumps operating under
similar conditions.
Suppose that among these homologous pumps we have a standard
pump which delivers a capacity of one GPM at a head of one
foot, so equation 2.37 becomes,


Where:
N
s
is the specific speed
n is rpm
Q is the pump delivery
H is the pump head
The physical meaning of the quantity N
s
is the rpm of a
standard pump homologous with a given pump and generating
under similar operating conditions, a head H
s
= 1 ft (in US
units)or 1 m (in SI units)at a rate of discharge of Q
s
= 1
gpm (in US units) or 1 m
3
/sec (in SI units).
The hydraulic and volumetric efficiencies of the two pump
are the naturally same.
Centrifugal pumps may be classified according to the
specific speed (in SI units) as follows:
1. Radial flow (low speed) N
s
~<= 20 D
2
/D
1
= 2.2-3.5
2. Radial flow (normal speed) N
s
~> 20-40 D
2
/D
1
= 2.2-1.8
3. Radial flow (high speed) N
s
~> 40-70 D
2
/D
1
= 1.8-1.3
4. Mixed flow N
s
~> 70-160 D
2
/D
1
= 1.3-1.1
5. Axial flow, or propeller N
s
~> 160 D
2
/D
1
= 1
The impeller shapes corresponding to these five types are
presented schematically in fig (2.19)

Fig (2.19)
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


35
The specific speed in U.S units is,

Where:
n in rpm
Q in gpm
H in ft
Centrifugal pumps may be classified according to the
specific speed (in US units) as follows:
1. Radial flow impeller (low speed): N
s
~<= 500-1700
2. Francis flow impeller (Radial high speed):
N
s
~> 1700-4000
3. Mixed flow impeller: N
s
~> 4000-8000
4. Axial flow impeller, or propeller: N
s
~> 8000
Specific speed identifies the approximate acceptable ratio
of the impeller eye diameter (D1) to the impeller maximum
diameter (D2) in designing a good impeller.
Ns: 500 to 4000; D
1
/D
2
< 0.5 Radial flow pump
Ns: 4000 to 8000; D
1
/D
2
> 0.5 Mixed flow pump
Ns: 8000 to 12000; D
1
/D
2
= 1 Axial flow pump
These figures for Ns and D1/D2 ratio are not restrictive,
rather, there is a big amount of overlap in the figures as
pump designers push the envelope of operating range of the
different types of pumps.
Then, the impeller shapes corresponding to these five types
are presented schematically in fig (2.19a)

Fig (2.19a)
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


36
The impeller shapes corresponding to the types are presented
schematically in fig (2.19b) for both SI and Us Units

Fig (2.19b)
Example:
It is required to have a pump has a capacity of 1500 US
gal/min (0.0944 m
3
/s) at 100 ft (30.5 m) of head and is
rotating at 1760 rev/min, what type of impeller has to be
used?
Solution
In US (English) units (gpm, and ft)
N
s
= 1760 x (1500)
0.5
/ 100
0.75
= 2156 radial flow, fig
(2.20)
In SI (Metric) units (m
3
/s and m),
N
s
= 1760 x (0.0944)
0.5
/ 30.5
0.75
= 42 radial flow, (fig
(2.19)
Note
Dividing the US units by 51.64 will yield the SI equivalent
specific speed.
N
s
(SI units) = 5156/51.64 = 41.7 =~ 42
Many pump types have been tested and their efficiency
measured and plotted in Fig (2.20).
Notice that larger pumps are inherently more efficient.
Efficiency drops rapidly at specific speeds of 1000 or less.
Example:
A pump has a capacity of 500 US gal/min at 97 ft of head and
is rotating at 1750 rev/min, calculate:
a. Specific speed. b. Define type of impeller
c. Expected efficiency
Solution
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


37
a. N
s
= 1750 * 500
0.5
/ 97
0.75
= 1266
b. From fig (2.19) it is radial flow impeller pump
c. From fig (2.20 and 2.20') the expected efficiency is ~75-
78%

Fig (2.20)

Fig (2.20')
The specific speed determines the general shape or class of
the impeller as depicted in Fig. (2.21). As the specific
speed increases, the ratio of the impeller outlet diameter,
D
2
, to the inlet or eye diameter, D
1
, decreases. This ratio
becomes 1.0 for a true axial flow impeller.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


38

High N
s
Low N
s

Fig (2.21)
Under certain conditions the specific speed N
s
characterizes
the ability of a pump to develop head and ensure a certain
delivery.
The higher specific speed the less head (for a given Q and
n) and the greater the capacity (for a given H and n).
Specific speed depends on impeller design. Pumps with low
specific speed have impellers with small relative width
b
2
/D
2
but a high value of D
1
/D
2
i.e. long vanes, which is
necessary to obtain a higher head. Flow through such an
impeller is in a plane perpendicular to the axis of
rotation.
With N
s
increasing the ratio D
2
/D
1
(as well as D
2
/D
0
)
decreases, i.e. the vanes are shorter and the relative width
of the impeller b
2
/D
2
is greater. Furthermore, the flow
through the impeller departs from the plane of rotation and
becomes increasingly three dimensional.
In the limit, at maximum value of N
s
, the flow is along the
axis of rotation and the impeller is of the axial flow type.
2.14. Net Positive Suction Head (NPSH) and
Cavitations
The Hydraulic Institute defines NPSH as the total
suction head in feet absolute, determined at the suction
nozzle and corrected to datum, less the vapor pressure of
the liquid in feet absolute. Simply stated, it is an
analysis of energy conditions on the suction side of a pump
to determine if the liquid will vaporize at the lowest
pressure point in the pump.
The pressure which a liquid exerts on its surroundings is
dependent upon its temperature. This pressure, called vapor
pressure, is a unique characteristic of every fluid and
increases with increasing temperature. When the vapor
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


39
pressure within the fluid reaches the pressure of the
surrounding medium, the fluid begins to vaporize or boil.
The temperature at which this vaporization occurs will
decrease the pressure of the surrounding medium decreases.
A liquid increases greatly in volume when it vaporizes. One
cubic foot of water at room temperature becomes 1700 cu. ft.
of vapor at the same temperature.
It is obvious from the above that if we are to pump a fluid
effectively, we must keep it in liquid form. NPSH is simply
a measure of the amount of suction head present to prevent
this excess vaporization at the lowest pressure point in the
pump.
NPSH required is a function of the pump design. As the
liquid passes from the pump suction to the eye of the
impeller, the velocity increases and the pressure decreases.
There are also pressure losses due to shock and turbulence
as the liquid strikes the impeller.
The centrifugal force of the impeller vanes further
increases the velocity and decreases the pressure of the
liquid. The NPSH required is the positive head in feet
absolute required at the pump suction to overcome these
pressure drops in the pump and maintain enough of the liquid
above its vapor pressure to limit the head loss, due to the
blockage of the cavitation vapor bubble, to 3 percent. The
3% head drop criteria for NPSH required is used worldwide
and is based on the ease of determining the exact head drop
off point. Most standard low suction energy pumps can
operate with little or no margin above the NPSH required,
without seriously affecting the service life of the pump.
The NPSH required varies with speed and capacity within any
particular pump. Pump manufacturers curves normally provide
this information.
NPSH Available is a function of the system in which the pump
operates. It is the excess pressure of the liquid in feet
absolute over its vapor pressure as it arrives at the pump
suction. Fig (2.22) shows four typical suction systems with
the NPSH Available formulas applicable to each.
It is important to correct for the specific gravity of the
liquid and to convert all terms to units of feet absolute
in using the formulas.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


40

Fig (2.22)
P
B
= Barometric pressure, in feet absolute.
V
P
= Vapor pressure of the liquid at maximum pumping
temperature, in feet absolute.
p = Pressure on surface of liquid in closed suction tank,
in feet absolute.
H = Static suction lift in feet (positive or negative).
H
f
= Friction loss in feet in suction pipe at required
capacity
Cavitation is a term used to describe the phenomenon, which
occurs in a pump when there is insufficient NPSH Available.
The pressure of the liquid is reduced to a value equal to or
below its vapor pressure and small vapor bubbles or pockets
begin to form. As these vapor bubbles move along the
impeller vanes to a higher pressure area, they rapidly
collapse.
The collapse or implosion is so rapid that it may be heard
as a rumbling noise, as if you were pumping gravel. In high
suction energy pumps, the collapses are generally high
enough to cause minute pockets of fatigue failure on the
impeller vane surfaces. This action may be progressive, and
under severe (very high suction energy) conditions can cause
serious pitting damage to the impeller.
The accompanying noise is the easiest way to recognize
cavitation.
Besides possible impeller damage, excessive cavitation
results in reduced capacity due to the vapor present in the
pump. Also, the head may be reduced and/or be unstable and
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


41
the power consumption may be erratic. Vibration and
mechanical damage such as bearing failure can also occur as
a result of operating in excessive cavitation, with high and
very high suction energy pumps.
The way to prevent the undesirable effects of cavitation in
standard low suction energy pumps is to insure that the NPSH
Available in the system is greater than the NPSH required
(NPSHR)by the pump. High suction energy pumps require an
additional NPSH margin, above the NPSH Required. Hydraulic
Institute Standard (ANSI/HI 9.6.1) suggests NPSH margin
ratios of from 1.2 to 2.5 times the NPSH required, for high
and very high suction energy pumps, when operating in the
allowable operating range.
2.15. NPSH Specific Speed
In designing a pumping system, it is essential to
provide adequate NPSH available for proper pump operation.
Insufficient NPSH available may seriously restrict pump
selection, or even force an expensive system redesign. On
the other hand, providing excessive NPSH available may
needlessly increase system cost.
Suction specific speed may provide help in this situation.
Suction specific speed (N
ss
) is defined as:

Where:
n = Pump speed RPM
gpm = Pump flow at best efficiency point at impeller inlet
NPSH
R
= Pump NPSH required at best efficiency point.
Experience has shown that 9000 is a reasonable value of
suction specific speed.
Pumps with a minimum suction specific speed of 9000 are
readily available, and are not normally subject to severe
operating restrictions.
Example:
Pump flows 2,000 gpm; head 600 ft at 3550 rpm. What NPSH
will be required?
9000= 3550 x (2000)
0.5
/ NPSH
R
3/4

NPSH
R

3/4
= 17.7
NPSH
R
= 46 ft
A related problem is in selecting a new pump, especially at
higher flow, for an existing system. Suction specific speed
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


42
will highlight applications where NPSH
A
may restrict pump
selection.
Example:
Existing system: Flow 2000 GPM; head 600 ft.: NPSH
A
30 ft.
What is the maximum speed at which a pump can be run without
exceeding NPSH available?
9000= n x (2000)
0.5
/ 30
3/4

n = 2580 rpm
Running a pump at this speed would require a gear and at
this speed, the pump might not develop the required head. At
a minimum, existing NPSH
A
is constraining pump selection.
2.16. System curve
For a specified impeller diameter and speed, a
centrifugal pump has a fixed and predictable performance
curve. The point where the pump operates on its curve is
dependent upon the characteristics of the system in which it
is operating, commonly called the System Head Curve...or,
the relationship between flow and hydraulic losses* in a
system. This representation is in a graphic form and, since
friction losses vary as a square of the flow rate, the
system curve is parabolic in shape.
By plotting the system head curve and pump curve together,
it can be determined:
1. Where the pump will operate on its curve.
2. What changes will occur if the system head curve or the
pump performance curve changes.
+ No static head All friction
As the levels in the suction and discharge are the same, Fig
(2.23), there is no static head and, therefore, the system
curve starts at zero flow and zero head and its shape is
determined solely from pipeline losses. The point of
operation is at the intersection of the system head curve
and the pump curve. The flow rate may be reduced by
throttling valve.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


43

Fig (2.23) No Static Head - All Friction
+ Positive static head
The parabolic shape of the system curve is again determined
by the friction losses through the system. But in this case
there is a positive static head involved. This static head
does not affect the shape of the system curve or its
steepness, but it does dictate the head of the system
curve at zero flow rate.
The operating point is at the intersection of the system
curve and pump curve. Again, the flow rate can be reduced by
throttling the discharge valve, Fig (2.24).

Fig (2.24) Positive Suction Head
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


44
2.17. Electric submersible pumps
2.17.1. Pump Stages
Electric submersible pumps (ESP) are multi-staged
centrifugal pumps. Each stage consists of a rotating
impeller and stationary diffuser.
The pressure energy change is accomplished as the liquid
being pumped around the impeller, and as impeller rotates,
the rotating motion of the impeller imparts a rotating
motion to the liquid. Actually, there are two components to
the motion imparted to the liquid by the impeller. One
motion is in a radial direction outward from the center of
the impeller. The motion is caused by centrifugal force. The
other motion moves in direction tangential to outside
diameter of the impeller. The resultant of these two
components is the actual direction of flow (as mentioned
earlier in this chapter).
So, the impellers function is to transfer energy by
rotation to the liquid passing through it, thus raising the
kinetic energy of the fluid.
The diffusers function is to change the high velocity
energy into relatively low velocity energy, while inversing
the pressure energy.
The diffuser section then converts the fluid kinetic energy
to potential energy, increasing the fluids potential energy
as it passes through the stage.

Fig (2.25) Impeller








Fig (2.26) Pump Stage
Upthrust washer
Downthrust washer
Eye washer
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


45

Fig (2.27) Fluid path through the vanes
2.17.2. Types of Impellers
As previously mentioned, there are mainly three
types of impellers, the ones using in ESP are, the radial
flow impeller and the mixed flow impeller.
The difference between these two types of designs is
described by the pump impeller vane angles and the size and
shape of the internal flow passages.
Radial flow design
A radial flow impeller has, as previously mentioned,
specific speed of 500-4000 (US units) vane angels at close
to 45 degree, and therefore, is usually found in pumps
designed for lower flow rates and high head.

Fig (2.28) Radial flow impeller
Mixed flow design
A mixed flow impeller has, as previously mentioned, specific
speed of 4000-8000 (US units) vane angels at close to 45
degree, and therefore, is usually found in pumps designed
for higher flow rates.

Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


46

Fig (2.29) Mixed flow impeller
For more details see item 2.13 of Specific Speed and its
relation to impeller geometry
2.17.3. Impeller constructions
ESP pumps come mainly in two basic varieties:

2.17.3.1. Floater Construction.
Each impeller is free to move up and down the
shaft, so it is said to "float" on the shaft.



Fig (2.30) Radial flow impeller
Why use a floater pump?
Let's look at a floating impeller in detail.
Since a floating impeller is free to move up and down the
shaft, the only thing to stop it is either the upper or
lower diffuser. "Thrust washers" are provided at all mating
surfaces between the impeller and diffuser to absorb any
thrust generated.

There is a small amount of
free play in the coupling such
that the pump shaft can fall
down to where the impellers
ride directly on the lower
diffusers or on the downthrust
washers if available.
Impeller is in full
down position
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


47
Thrust
Washers

Fig (2.31) Washers in impeller
The green area shows, fig (2.32), the "up thrust" washer
between the impeller and upper diffuser.

Force
Upthrust is
absorbed
here


Fig (2.32) up thrust force

The green area shows the "down thrust" washers between the
impeller and lower diffuser, Fig (2.33).

Downthrust
is absorbed
here
Force

Fig (2.33) down thrust force
We lose efficiency in the down thrust position because of
the fluid's ability to re-circulate from the high pressure
to low pressure eye area. In addition to loss in efficiency,
this can promote erosion in the diffuser in abrasive fluids
fig (2.34).
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


48


Fig (2.34) Fluid leakage
"Floater" Pump is normally used because:
Each stage handles its down thrust; a very large number of
stages can be put in a pump without having to worry about
protector bearing capacity.
Very good with mild abrasives since they prevent material
from getting into the radial bearing area.
Easier field assembly - no shimming required.

Fig (2.35) Floater Pump


Motor
Protector
Pump
Impeller
Thrust
Is there
any Thrust
seen here?
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


49
2.17.3.2. Compression Construction.
In a compression pump, all the impellers are
rigidly fixed to the shaft so that if an impeller wants to
move up or down, it will take the shaft with it.
The impeller is normally sitting down on its lower diffuser
during assembly due to gravity. Because of this, the pump
shaft is "raised" with shims in the coupling so that the
impeller is not allowed to touch the diffuser after final
assembly. This allows all thrust developed in the pump
shaft to be transferred to the protector shaft directly.
Every impeller is fixed to the shaft rigidly so that it
cannot move without the shaft moving. All the impellers are
"compressed" together to make one rigid body.


Fig (2.36) Pump shimming Ser.675 or larger


Fig (2.37) Pump shimming Ser.562 or smaller
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


50
Note:
Consult Field Service Manual to determine the appropriate
shimming recommended for every type of pump and stage.
Why Use "Compression" Pumps?
Some stages generate too much thrust to be handled by a
thrust washer in the stage.
Some fluids (e.g., liquid propane) do not have enough
lubricity to properly lubricate a thrust washer.
If abrasives or corrosives are present, it may be
beneficial to handle the thrust in an area lubricated by
motor oil rather than well fluid.
Occasionally in very gassy wells, the flow volume changes
so drastically within the pump that parts of a floater
pump could be in very severe thrust while others are not
so a compression pump could be one alternative.
Since all the thrust is handled in the protector, as long
as the protector has a great enough capacity, the pump
operating range can be extended over a much wider area
without any increased wear or reduced life.

Fig (2.38) Compression impellers
2.18. Recommended Operating Range (ROR)
What exactly is the ROR and why is it important?
ROR is the discharge range which the pump must be worked
within it.
All Thrust
Carried
here
Protector
Thrust
Bearing
Motor
Thrust
Bearing
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


51

Most people look at pump operating range as defining
"thrust" limits where the stage goes into down thrust on the
left side of the range and it goes into up thrust on the
right side of the range. If the stage is within the range,
it is thought to be balanced with no net thrust in either
direction.
Graphically it would look like this:
This is almost always wrong.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


52

Before we worry too much about how much thrust we have, we
need to know what impeller thrust is?
2.18.1. Impeller Thrust
Under normal operating conditions, fluid re-
circulates on top and underneath the impeller.

Fig (2.39) Operation @ BEP
Accordingly, the fluid creates pressure on the upper and
lower impeller shrouds.

Fig (2.40) Pressure acts on the upper and lower shrouds.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


53
Since the cross sectional area on the upper shroud is
larger, the net force of the pressure is down.

Fig (2.41) Pressure acts on the upper > lower shrouds.
This causes the impeller to be moved down. This is a
positive downward force termed Downthrust

Fig (2.42) impeller moved down
At some point where the volume of fluid going up into the
pump, i.e. increase the momentum due to increase the fluid
velocity (momentum = mv kg.m/s), will lift the impeller up,
overcoming the Downthrust pressure. This causes the
impeller to be moved up. The downward force is now reversed
(negative); it is termed Upthrust

Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


54
Fig (2.43) impeller moved up due to upthrust
This creates a larger cross sectional area on the bottom
shroud which increases the pressure underneath, this causes
the impeller to be held up increases the Upthrust.

In normal operation, the pump is designed to operate in
slight to moderate Downthrust.

Fig (2.44) the pump, under normal operating condition
Based on the previous explanation, it is concluded that,
there are three forces acting on the impeller. The sum of
these three forces is the total thrust, they are:
Gravity acting on the buoyed mass of the impeller, its
direction is always downward.
The net force resulting from the differential pressure in
the stage its direction is either downward or zero (zero
occurs at wide open flow - no pressure)
The force from the momentum of the fluid coming into the
stage its direction is either upward or zero (zero occurs
at shut-in or no flow condition)
Gravity: An Impeller has a mass which is acted on by
gravity which pulls it downward toward the Earth.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


55

Fig (2.45)
Pressure: The pressure times the area equals force. There
are both a downward force and an upward force.
The downward force is always larger except when the pump
generates no pressure (wide open flow).


Fig (2.46)
Momentum: The fluid entering the bottom of the impeller is
forced to change direction. This change in momentum exerts
an upward force on the impeller except when there is no flow
(shut-in).





Direction of Fluid Flow
Fig (2.47)
F = mg
Where g is the acceleration due to gravity
F
An impeller adds pressure to the fluid
so that the pressure on the top side is
greater than the pressure on the
bottom
side.
Low Pressure
High Pressure
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


56
Pressure: The downward arrows represent the larger force
due to difference in surface areas.

Fig (2.48)
In general, larger diameter impellers will have a higher
down thrust than smaller impellers for the same flow rate.
Why?
Because they have a larger surface area on which the
pressure difference can operate. They also have more mass.
Is it possible to affect the down thrust caused by pressure
in any way?

What if we could reduce the pressure on top of the impeller?
If we could lower the pressure on the top of the impeller,
this would reduce the thrust.
By using a "balance ring" between the impeller and diffuser
and drilling "balance holes" in the upper impeller skirt,
we can re-circulate lower pressure fluid over the majority
of the upper surface.



Fig (2.49)
+ =
The net difference between the two
forces is the down thrust due to
pressure.
Low Pressure Fluid
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


57
Conclusion: Under normal operating conditions, fluid
recirculation on the top and bottom side of the impeller
cause forces to be applied on the upper and lower impeller
shrouds. When the recirculation forces are greater on the
upper shroud, the impeller is moved down which is termed
downthrust. When the recirculation forces are greater on the
lower shroud, the impeller is moved up which is termed
upthrust. The magnitude recirculation forces depends upon
the flow rate going thru the impeller vs. its head -
capacity, i.e., its operating range. Downthrust increases as
the flow through the stage decreases (or on the left-hand
side of the pump curve). Upthrust increases as the flow
through the stage increases (or on the right-hand side of
the pump curve).

Fig (2.50) Recommended Operating Range
In ROR, most ES pumps are designed to operate in mild to
moderate downthrust.
2.18.2. Impeller Thrust Washers
Washers are fitted to the impeller to prevent wear;
there are three types of washers, upthrust, Downthrust, and
eye washers.

Fig (2.51)
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


58


2.18.3. Shaft thrust
We said that the impellers each handle their own
thrust so why do we have to worry about shaft thrust?
The total thrust is made up of two components:
1. The impeller thrust, already discussed in the previous
section, and
2. The shaft thrust.
Shaft Thrust: Compression Pump
In the compression pump, we could not separate the shaft
thrust from the impeller thrust since they were coupled
together.
Shaft Thrust: Floater Pump
In a floater, since the impeller can move on the shaft it
makes sense that the shaft can also move within the
impeller.
Forces acting on shaft:
Pressure acting on top of the shaft
Pump shaft mass
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


59
So will the pump discharge pressure times the shaft cross
sectional area give us the shaft thrust?
Force = pressure * area
Note: The force, due to the weight of the shaft, is not
usually significant so we will ignore it for now.
Force = Pressure * Area, or
Thrust = Discharge Pressure x Shaft Cross Section Area


A protector serves to equalize the pressure between the
well fluid and the inside of the motor. Therefore the
motor pressure should be roughly equal to the pump intake
pressure.
This is not exactly true since the motor pressure will be
slightly higher due to the weight of a few feet of motor
oil but this increase will be insignificant compared to the
large pressure differential created by the pump.

It can be shown that all pressures cancel out except for
the pressure on the top of the shaft. It can also be shown
that, regardless of the various diameters, couplings, etc.
that the net shaft force can be calculated from the
following:
Shaft Thrust: Floater Pump
Pump Shaft
Protector Shaft
Motor Shaft
Protector Thrust
Runner
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


60
Force = (P
d
-P
i
) *A
s

From this equation, it can be seen that the larger the
shaft diameter, the greater the force for a given pressure
difference.
For two pumps doing the same work, the one with the larger
shaft diameter will generate more shaft thrust.
Reviewing our example of the GN2100 which produced 1850
feet of head, this 1850 feet is a true pump differential so
we can plug it directly into the equation. Or can we?
Let: Specific Gravity = 1.0, and convert to psi:
Pressure = 800
31 . 2
0 . 1 1850
=
x
psi [Pd-Pi]
Force = (Pd-Pi) *A
s
= 800 psi * A
s

Where do we get the area?
**From the catalog.
For a GN2100, the shaft area is 0.6013 sq. in.
Force = 800 psi *0.6013 = 481 lbs
Remember
We neglected the weight of the shaft, and this is usually
safe to do unless you are very close to fully loading a
protector bearing.
Larger diameter pumps have larger diameter shafts so, for
the same amount of work, they have higher shaft thrust.
On the other hand, larger diameter pumps can and will
generally use larger diameter protectors, and larger
protectors can handle much higher thrust loads.
But we must consider either the shaft thrust of a floater
or the total thrust of a compression pump when selecting
a protector, as we will see later.
Operating outside the range requires a larger pump and
motor to be purchased as well as more electricity to
operate it.
One very real concern with establishing operating ranges
is the API RPS2 on pump testing:
+ +/- 5% head and flow
+ Less than +8% on HP within the operating range
+ +/- 10% on efficiency at BEP.

Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


61

2.19. Pump configurations
ESP pumps come in several different configurations.
Most pumps (especially the smaller diameter ones) come as
"center tandems" (or CT type).
Other types are "upper tandems" (UT), "lower tandems" (LT)
and "single" (S) pumps, fig (2.52).
The actual pump stages are no different regardless of what
"type" it is. The difference in the pumps depends on what
it looks like on the ends.

Fig (2.52)
A "single" pump has an
intake and discharge
head intrinsic to the
pump itself. No other
pumps can be attached to
it.
An "upper tandem" pump
has a discharge head
but no intake section.
It can be placed on
top of
another pump or an
intake section.
A "center tandem" pump
has no intake or is
charge inherent and must
have these provided in
some fashion either with
another pump or an
intake section and/or
bolt-on head.
A "lower tandem" pump
has an integral intake
but no discharge head.
It can go below another
pump or else be
completed with a bolt-on
discharge head.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


62
Single Pump
A "single" pump has an intake and discharge
head intrinsic to the pump itself. No other pumps can be
attached to it, fig (2.53).


Fig (2.53)
Upper tandem
The upper tandem has either another pump below it or else
an intake section to complete the assembly, fig (2.54).

Fig (2.54)
Lower Tandem
The lower tandem has either another pump above it or else
a bolt-on discharge to complete the assembly.
Lower tandems are especially common in the larger
diameter, higher flow rate pumps.

Discharge Head
Intake
Pump
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


63
This helps to reduce entrance losses associated with
higher flow rates and also, in some cases, allows a gas
separator to be built directly into the intake where a
standard add-on separator could not handle the fluid
throughout, fig (2.55).

Fig (2.55)
Center Tandem
Center tandem pumps offer the most flexibility. If the
required number of stages for the well cannot fit into a
single section, more sections can be added until the
stage requirement is met.
Since a CT pump can be either a single (with a bolt-on
discharge and intake added) or a part of a larger pump,
inventory requirements are greatly reduced, fig (2.56).


Fig (2.56)
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


64
2.20. Fluid Viscosity Effect on Centrifugal Pumps
2.20.1. Viscosity
The viscosity of a fluid is that property which
tends to resist a shearing force. It can be thought of as
the internal friction resulting when one layer of fluid is
made to move in relation to another layer.
Consider the model shown in fig (2.57), which was used by
Newton in first defining viscosity. It shows two parallel
planes of fluid of area A separated by a distance dx and
moving in the same direction at different velocities V
1
and
V
2
.

Fig (2.57)
The velocity distribution will be linear over the distance
dx, and experiments show that the velocity gradient, dv/dx ,
is directly proportional to the force per unit area, F/A .
F = x dv/dx
Where:
is constant for a given liquid and is called its
viscosity.
The velocity gradient, dv/dx , describes the shearing
experienced
by the intermediate layers as they move with respect to each
other.
Therefore, it can be called the "rate of shear", S. Also,
the force per unit area, F/A , can be simplified and called
the "shear force" or "shear stress," t. With these
simplified terms, viscosity can be defined as follows:
t = x S
Viscosity = = t/S = shear stress / rate of shear
Newton made the assumption that all materials have, at a
given temperature, a viscosity that is independent of the
rate of shear. In other words, a force twice as large would
be required to move a liquid twice as fast. Fluids which
behave this way are called Newtonian fluids. There are, of
course, fluids which do not behave this way, in other words
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


65
their viscosity is dependent on the rate of shear. These are
known as Non-Newtonian fluids.
There are two related measures of fluid viscosity:
1. Dynamic (or absolute) viscosity.
The absolute viscosity of a liquid is defined as the
resistance to flow and shear under the forces of internal
friction as described above. This internal friction is
caused by the resistance of the liquids molecules moving
relative to each other. The larger the molecules, the higher
the internal resistance and consequently, the higher the
viscosity. The commonly units of absolute viscosity is
described below.
In the SI (international system)system the dynamic viscosity
units are N.s/m
2
(Newton.Second/meter
2
), Pa.s (Pascal.sec)
or kg/m.s where,
1 Pa.s = 1 N.s/m
2
= 1 kg/m.s (N = kg.m/s
2
)
The dynamic viscosity is also often expressed in the
metric CGS (centimeter-gram-second) system as,
g/cm.s, dyne.s/cm2 or poise (p)
where,
1 poise (p) = dyne.s/cm
2
= g/cm.s
Pa.s = kg/m.s = 1000g/100cm.s = 10 g/cm.s = 10 poise
Or, N = kg.m/s
2
= 1000 g.100 cm/s
2
=10
5
g.cm/s
2
= 10
5
dyne
where dyne = g.cm/s
2

Pa.s = N.s/m
2
= 10
5
dyne.s/10
4
cm
2
= 10 dyne.s/cm
2
= 10
poise
Pa.s = 10 Poise, then poise = 0.1 Pa.s
For practical use the Poise is to large and it is usual
divided by 100 into the smaller unit called the
centiPoise (cp) where, 1 p = 100 cp
2. Kinematic Viscosity
Kinematic viscosity is a measure of a liquids resistance
to flow and shear under the forces of gravity.
It is the ratio of absolute or dynamic viscosity to density.
Kinematic viscosity can be obtained by dividing the absolute
viscosity of a fluid with its mass density
= /
Where:
= kinematic viscosity
= absolute or dynamic viscosity
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


66
= density
In the SI system the theoretical unit is m
2
/s or commonly
used Stoke (St) where 1 St = 10
-4
m2/s or 1 m
2
/s = 10,000
Stoke
Since the Stoke is an unpractical large unit, it is usual
divided by 100 to give the unit called Centistokes (cs)
where
1 St = 100 cs, or 1 cs = 10
-6
m
2
/s, m
2
/s = 1,000,000 cs
Seconds Saybolt Universal (SSU) is used also to measure the
kinematic viscosity. The efflux time is Seconds Saybolt
Universal (SSU) required for 60 milliliters of a petroleum
product to flow through the calibrated orifice of a Saybolt
Universal viscometer, under carefully controlled
temperature.
1cs = 0.220 x SSU 180/SSU
3.20.2. Types of liquid
There are two basic types of liquid that can be
differentiated on basis of their viscosities behavior they
are:
1. Newtonian
These are fluids where viscosity is constant and
independent of shear rate, and where the shear rate is
linearly proportional to shear stress. Examples are water
and oil
2. Non- Newtonian
These are liquids where the shear rate-shear stress
relationship is nonlinear. They can be divided into four
categories.
Bingham Plastic fluids are those in which there is no
flow until a threshold shear stress is reached. Beyond
this point, viscosity decreases with increasing shear
rate. Most slurries have this property.
Dilatant fluids are those of which viscosity increases
with increasing shear rate. Example are candy mixture,
clay slurries and quicksand.
Pseudo-plastic fluids are similarly to Bingham plastic
fluids except there is no definite yield stress. Many
emulsions fall into this category.
Thixotropic fluids are those which viscosity decreases
to minimum level as their shear rate increases. Their
viscosity at any particular shear rate may vary,
depending on the previous condition of the fluid.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


67
Examples are asphalt, point, molasses, and drilling
mud.
3.20.3. Effect of viscosity on centrifugal pumps (using
charts)
When sizing a pump for a new application or
evaluating the performance of an existing pump, it is often
necessary to account for the effect of the pumping fluid's
viscosity. We are all aware that the head-capacity curve
presented in the pump vendor catalogs are prepared using
water as pumped fluid. These curves are adequate for use
when actual fluid that we are interested in pumping has a
viscosity that less than or equal to that of water. However,
in case of crude oils this is not the case.
Heavy crude oils can have viscosities high enough to
increase the friction drag on pump's impeller significantly.
The additional horsepower required to overcome this drag
reduces the pump efficiency. There are several analytical
and empirical approaches available to estimate the magnitude
of this effect.
Before beginning, it is vital to emphasize the importance of
having an accurate viscosity number on which to base our
estimates.
The viscosity of most liquids is strongly influence by
temperature. This relationship is most often shown by
plotting two points on semi-log grid and connecting them
with straight line. The relationship is of the form:

Where:
= the absolute viscosity of the fluid
A and B = constants
T = Absolute temperature of the fluid
Plotting this relationship requires knowledge of two data
points and using them effectively requires some adjustment
as to the normal operating temperature that might be
expected during other off design condition such as start up.
The effect of pressure on viscosity of most fluids is small.
For example an increase of pressure of 33 bars (~480 psi)
for mineral oil is equivalent to temperature drop of 1
O
C.
The performance of centrifugal pumps is affected when
handling viscous liquids. A marked increase in brake
horsepower, a reduction in head, and some reduction in
capacity occur with moderate and high viscosities.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


68
Fig (2.58) and Fig (2.59) provide a means of determining the
performance of a conventional centrifugal pump handling a
viscous liquid when its performance on water is known. It
can also be used as an aid in selecting a pump for a given
application.
Limitations on Use of Viscous Liquid Performance
Correction Chart
Reference is made to Fig (2.58), this chart is to be used
only within the scales shown (100-10000 gpm). Do not
extrapolate.
Fig (2.59) scale is 10-100 gpm
Use only for pumps of conventional hydraulic design, in the
normal operating range, with open or closed impellers. Do
not use for mixed flow or axial flow pumps or for pumps of
special hydraulic design for either viscous or non-uniform
liquids.
Use only on Newtonian (uniform) liquids. Gels, slurries,
paper stock and other non-uniform liquids may produce widely
varying results, depending on the particular characteristics
of the liquids.
Symbols and Definitions Used in Determination of Pump
Performance When Handling Viscous Liquids.
These symbols and definitions are:
Q
vis
= Capacity when pumping Viscous liquid, gpm
H
vis
= Head when pumping Viscous liquid, feet
E
vis
= Efficiency when pumping Viscous liquid, per cent
BHP
vis
= Brake Horsepower required by the pump for the
viscous liquid.
Q
W
= Capacity when pumping water., gpm
H
W
= Head when pumping water, feet
= Specific Gravity
C
Q
= Capacity correction factor
C
H
= Head correction factor
C
E
= Efficiency correction factor
1.0 Q
w
= Water Capacity at BEP (Best Efficiency Point).
1.2 Q
w
= Water Capacity at 1.2 * 1.0 Q
W.

0.8 Q
w
= Water Capacity at 0.8 * 1.0 Q
W

0.6 Q
w
= Water Capacity at 0.6 * 1.0 Q
W


Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


69

Fig (2.58) Viscosity Corrections for Capacities => 100 GPM
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


70

Fig (2.59) Viscosity Corrections for Capacities =< 100 GPM
The following equations are used for determining the viscous
performance when the water performance of the pump is known:
Q
vis
= C
Q
X Q
w

Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


71
H
vis
= C
H
x H
w

E
vis
= C
E
x E
w


C
Q
, C
H
and C
E
are determined from Fig (2.58) or fig (2.59)
which is based on the water performance.
Instructions for Preliminary Selection of a Pump for a
Given Head-Capacity-Viscosity Condition
Given the desired capacity and head of the viscous liquid to
be pumped and the viscosity and specific gravity at the
pumping temperature, above figures (2.58 or 2.59) can be
used to find approximate equivalent capacity and head when
pumping water.
The following equations are used for approximating the water
performance when the desired viscous capacity and head are
given and the values of C
Q
and C
H
must be estimated from fig
(2.58) or fig (2.59) using Q
vis
and H
vis
, as:
Q
W
(approx.) = Q
vis
/C
Q

H
W
(approx.) = H
vis
/C
H

Enter the chart Fig (2.58) or fig (2.59) at the bottom with
the desired viscous capacity, (Q
vis
) and proceed upward to
the desired viscous head (H
vis
) in feet of liquid. For
multistage pumps, use head per stage. Proceed horizontally
(either left or right) to the fluid viscosity, and then go
upward to the correction curves. Divide the viscous capacity
(Q
vis
) by the capacity correction factor (C
Q
) to get the
approximate equivalent water capacity (Q
w
approximately).
Divide the viscous head (H
vis
) by the head correction factor
(C
H
) from the curve marked "1.0 x Q
w
" to get the approximate
equivalent water head (H
w
approximately). Using this new
equivalent water head capacity point, select a pump in the
usual manner.
The viscous efficiency and the viscous brake horsepower may
then be calculated.
However, the procedure has sufficient accuracy for most pump
selection purposes. Where the corrections are appreciable,
it is desirable to check the selection by the method
described below.
Example
Select a pump to deliver 750 gpm at 100 feet total head of a
liquid having a viscosity of 1000 SSU (220 ct) and a
specific gravity of 0.90 at the pumping temperature.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


72
Enter the chart Fig (2.xx)with 750 gpm, go up to 100 feet
head, over to 1000 SSU, and then up to the correction
factors:
C
Q
= 0.95
C
H
= 0.92 (for 1.0 Q
w
)
C
E
= 0.635
Q
w
= 750/0.95 = 790 gpm
H
w
= 100/0.92 = 109 feet head
Select a pump for a water capacity of 790 gpm at 109 feet
head. The selection should be at or close to the maximum
efficiency point for water performance. If the pump selected
has efficiency on water of 81 per cent at 790 gpm, then the
efficiency for the viscous liquid will be as follows:
E
vis
= 0.635 x 81% = 51.5 per cent
The brake horsepower for pumping the viscous liquid will be:
BHP
vis
= 750 x 100 x 0.90/3960 x 0.515 = 33.1 hp
For performance curves of the pump selected, correct the
water performance as shown below.
Instructions for Determining Pump Performance on a Viscous
Liquid When Performance on Water is Known
Given the complete performance characteristics of a pump
handling water, determine the performance when pumping a
liquid of a specified viscosity.
From the efficiency curve, locate the water capacity (1.0 x
Q
w
) at which maximum efficiency is obtained.
From this capacity, determine the capacities (0.6 x Q
w
),
(0.8 x Q
w
) and (1.2 x Q
w
).
Enter the chart at the bottom with the capacity at best
efficiency (1.0 x Qw), go upward to the head developed (in
one stage) (H
w
) at this capacity, then horizontally (either
left or right) to the desired viscosity, and then proceed
upward to the various correction curves.
Read the values of (C
E
) and (C
Q
), and of (C
H
) for all four
capacities.
Multiply each head by its corresponding head correction
factor to obtain the corrected heads. Multiply each
efficiency value by (C
E
) to obtain the corrected efficiency
values which apply at the corresponding corrected
capacities.
Plot corrected head and corrected efficiency against
corrected capacity. Draw smooth curves through these points.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


73
The head at shut-off can be taken as approximately the same
as that for water.
Calculate the viscous brake horsepower (BHP
vis
) from the
formula given above.
Plot these points and draw a smooth curve through them which
should be similar to and approximately parallel to the brake
horsepower (BHP) curve for water.
Example
Select a pump to deliver 750 gpm at 100 feet total head of
water. Calculate the parameters if the fluid viscosity is
1000 SSU (220 ct) and a specific gravity of 0.90 at the
pumping temperature.
Enter the chart Fig (2.xx)with 750 gpm, go up to 100 feet
head, over to 1000 SSU, and then up to the correction
factors:
C
Q
= 0.95
C
H
= 0.92 (for 1.0 Q
w
)
C
E
= 0.635
Q
vis
= 750 x 0.95 = 712.5 gpm
H
vis
= 100 x 0.92 = 92 feet head
E
vis
= 0.635 x 81% = 51.5 per cent
The brake horsepower for pumping the viscous liquid will be:
BHP
vis
= 712.5 x 92 x 0.90/3960 x 0.515 = 28.9 hp
CALCULATIONS
Q
w
(Q
w0.6Q
, Q
w0.8Q
, Q
w1.0Q
, Q
w1.2Q
) 450 600 750 900
H
W
114 108 100 86
E
w
72.50% 80% 81% 79.5%
BHPw 16.1 18.4 20.8 22.1
Sp.Gr 0.9 0.9 0.9 0.9
VISCOSITY
SSU 1,000 1,000 1,000 1,000
Cs 220 220 220 220
C
Q
0.95 0.95 0.95 0.95
C
H
0.96 0.94 0.92 0.89
C
E
0.635 0.635 0.635 0.635
Q
vis
427 570 712.5 855
H
vis
109.5 101.5 92 76.5
E
vis
46% 50.8% 52.5% 50.5%
BHP
w
23.1 25.9 28.9 29.4
These calculations can be plotted on a performance curve as
follows:
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


74

Fig (2.60) Performance curve
3.20.4. Effect of viscosity on centrifugal pumps (using
calculations)
It was convenient to define a term known as pseudo-capacity:

Where:
= viscosity of the fluid in cs
H = Head in ft at BEP
Q = Capacity in gpm at BEP
Pseudo-capacity is used with the following polynomial
coefficients to determine viscosity correction terms that
are very closed to those given in fig (2.xx) and fig (2.xx).
These polynomials have been checked throughout the entire
range of the figures and appear to give answers within 1% of
those found using figures.
The polynomial used is of form:

Where:
C
x
is the correction factor that must be applied to the term
in question
D
xn
are polynomial coefficients listed in the following
table
P is the pseudo-capacity term defined above
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


75
Correction
Factor
Polynomial Coefficients
D
x1
D
x2
D
x3
D
x4
D
x5
D
x6

C
h
1.0522 -3.5120E-02 -9.0394E-04 2.2218E-04 -1.1986E-05 1.9895E-07
C
Q
0.9873 9.0190E-03 -1.6233E-03 7.7233E-05 -2.0528E-06 2.1009E-08
C
H0.6
1.0103 -4.6061E-03 2.4091E-04 -1.6912E-05 3.2459E-07 -1.6611E-09
C
H0.8
1.0167 -8.3641E-03 5.1288E-04 -2.9941E-05 6.1644E-07 -4.0487E-09
C
H1.0
1.0045 -2.6640E-03 -6.8292E-04 4.9706E-05 -1.6522E-06 1.9172E-08
C
H1.2
1.0175 -7.8654E-03 -5.6018E-04 5.4967E-05 -1.9035E-06 2.1615E-08
For example, we had looking for a pump that could deliver
100 ft of head at capacity 750 gpm, 1000 SSU (220 cs)
viscosity we would proceed as following:
Q
vis
= C
Q
X Q
w

H
vis
= C
H
x H
w

E
vis
= C
E
x E
w


Or
Q
w
= C
Q
/ Q
vis

H
w
= C
H
/ H
vis

E
vis
= C
E
x E
w



P = 1.95*(220)
0.5
*[0.04739*(100)
0.25746
*(750)
0.5
]-
0.5
= 14.03386
C
Q
= 0.9873+0.009019*14.03386-0.0016233*14.03386^2
+0.000077233*14.03386^3-0.0000020528*14.03386^4+
0.000000021009*14.03386^5 = 0.939443 (= 0.95 from the
chart)
H
1.0Q
= 1.0045-0.002664*14.03386-0.00068292*14.03386^2+
0.000049706*14.03386^3-0.0000016522*14.03386^4+
0.000000019172*14.03386^5 = 0.9163 (0.92 from the
chart)
C = 1.0522-0.03512*14.03386-0.00090394*14.03386^2
+0.00022218*14.03386^3-0.000011986*14.03386^4
+0.00000019895*14.03386^5 = 0.638772 (0.635 from the
chart)
Q
w
= 750 / 0.939443 = 798 gpm (790 gpm based on chart)
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


76
H
w
= 100 / 0.9163 = 109 ft (109 ft based on chart)
E
q
= 81 x 0.638772 = 51.7 % (51.5% based on chart)

BHP
vis
= 750 x 100 x 0.9/3960 x 0.517 = 33 hp (33.1 hp based
on chart)
Correction Factor Comparison
C
q
C
Q
C
H0.6
C
H0.8
C
H1.0
C
H1.2

From Chart 0.635 0.95 0.96 0.94 0.92 0.89
From Polynomial 0.639 0.939 0.958 0.939 0.916 0.887
2.18.5. Simplified way to calculate the effect of viscosity
on centrifugal pumps when performance on water is
known
Many years ago a generalized procedure was developed
by the Hydraulic Institute (HI) for correcting centrifugal
pump performance when handling viscous fluids. Their
procedure was universally accepted by pump manufactures and
industry users. HI has recently issued a revised procedure
for predicting pump performance based on new research and
field test data. The revised procedure also corrects the
shortcomings in the earlier method and has added
mathematical formulas for use with computers.
It is widely accepted that viscous fluids affect the
performance of centrifugal pumps. Since the performance of
most centrifugal pumps is determined from water, a procedure
is needed to correct the performance curves when pumping
viscous fluids. Fig (2.61) illustrates a typical pump
performance curve based on water and how it is affected when
pumping viscous fluids. In many applications the difference
in water and viscous performance are significant.
The previous HI procedure was based on a graphical
representation that provided correction factors at four
operating points: 60, 80, 100 and 120 percent of the best
efficiency point (BEP). Using these correction factors at
the four operating points, the user was able to prepare a
corrected pump performance chart of flow versus total head
(Q-H) and flow versus efficiency (Q-), when handling
viscous fluids. The procedure had limitations but proved to
be accurate under a wide range of conditions.
The recently published HI procedure has many of the same
features as the older method. However, many improvements
were made to improve the accuracy and correct problems in
the earlier procedure. Some of the noteworthy changes made
in the latest version are presented below.
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


77

Fig (2.61) Sample performance chart
1. HI previously had two separate graphs for determining
pump viscous effects, one for pump flow rates less than
100 gpm and another for flow rates over 100 gpm. The two
graphs gave conflicting results if the pump operated at
100 gpm. The new procedure has only one procedure
regardless of the flow rate.
2. HI now uses a new basis for determining the correction
factors C
H
, C
Q
, and C

. The new method is based on a


performance factor, called Parameter B, and includes
terms for viscosity, speed, flow rate and total head.

B = Performance factor
K = 16.5 for SI units
= 26.5 for US units
V
vis
= Kinematic viscosity in cs
H
BEP.w
= Water head per stage at BEP in m (ft)
Q
BEP.w
= Water flow rate at BEP in m3/h (gpm)
N = Pump shaft speed in rpm
3. The latest HI procedure allows the user to determine
viscous performance from graphs or equations. With the
previous method only graphical methods were used to
calculate correction factors. Fig (2.62) and (2.63)
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


78
illustrate the revised graphical method and gives the
flow rate correction factor (C
Q
), head correction factor
(C
H
) and efficiency correction factor (C

), based on the
performance Parameter B given above.

Fig (2.62) correction factors for C
Q
and C
H


Fig (2.63) correction factors for C


4. The following equations are used with the correction
factors to determine the corrected pump performance for
viscous fluids
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


79

5. The previous method required the user to find the pump
BEP before calculating corrections factors at 60, 80, 100
and 120 percent of BEP. The revised procedure uses a
similar approach of finding the pump BEP but corrections
factors can be determined, with caution, at any flow
rate, not necessarily at 60, 80, 100 and 120 percent of
BEP.
The following are some factors derived from the latest
addition of the HI procedure for correcting pump
performance. They may be helpful when considering when to
correct pump water-performance curves for viscous fluids.
- The performance factor, Parameter B, is a good indicator
of how viscosity influences pump performance. Parameter B
takes into consideration fluid viscosity, speed, total head
and flow rate at BEP.
- Whenever the performance Parameter B is one or less the
fluid has no affect on the head curve, Q-H. Fig (2.62) shows
that the head and flow correction factors are less than
three percent when Parameter B equals three and about ten
percent when Parameter B equals seven.
- Viscous fluids have a more dramatic effect on pump power.
Fig (2.63) shows that the efficiency decreases by almost 12
percent when Parameter B equals three and over 30 percent
when Parameter B equals seven.
2.21. Affinity laws
The affinity laws are derived from a dimensionless
analysis of three important parameters that describe pump
performance: flow, total head and power (refer to item
2.12). The analysis is based on the reduced impeller being
geometrically similar and operated at dynamically similar
conditions or equal specific speed. If that is the case
then the affinity laws can be used to predict the
performance of the pump at different diameters for the same
speed or different speed for the same diameter.
The affinity laws were developed using the law of
similitude which provide 3 basic relationships.

Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


80
Flow vs. diameter and speed

Total Head vs. diameter and speed

Power vs. diameter and speed

Where:
Subscripts 1 and 2 denote the value before and after the
change.
HP is the power
n is the speed
D is the impeller diameter
H is the total head.
If the speed is fixed the affinity laws become:



If the diameter is fixed the affinity laws become:



Accordingly, the Affinity Laws for centrifugal pumps
describe the impact of changes in speed or impeller
diameter on pump flow, head, and HP. The useful tool in
predicting pump performance changes when speed is changed,
such as variable speed drive.
Example:
A pump is running with following parameters at 50 Hz,
Q = 500 bpd, H = 4000 ft, BHP = 16
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


81
What will be the new parameters if the frequency changed to
be 60 Hz?
Solution:
Q
2
= Q
1
x (n
2
/n
1
) = 500 x 60/50 = 600 bpd
H
2
= H
1
x (n
2
/n
1
)
2
= 4000 x (60/50)
2
= 5760 ft
BHP
2
= BHP
1
x (n
2
/n
1
)
3
= 16 x (60/50)
3
= 28 hp
2.22. Friction Loss in pipes (tubing)
Tubing friction is the loss of energy which occurs in
tubing with flowing fluid. At the wall, the fluid velocity
is zero whereas it attains a maximum value at the tubing
center. Due to these velocity differences across the
tubing, see fig (2.64), the fluid molecules rub against
each other. This transforms kinetic energy to heat energy
which can be considered as lost.

Fig (2.64) Velocity profile in pipes
To maintain a flow in the tubing, an amount of energy
corresponding to the energy which is lost must constantly
be added. Energy is supplied by static pressure difference
from inlet to outlet. It is said that it is the pressure
difference which drives the fluid through the tubing.
The loss in the tubing depends on the fluid velocity, the
inside diameter of the pipe, length, viscosity of the fluid
and inner surface roughness of the pipe (tubing).
The head loss is calculated by the following equation:



Where:
P = pressure loss in psi
Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


82
h
l
= Head loss in ft
= fluid density in slug/cuft
v = fluid velocity in ft/sec
g = gravity constant (32.2)
D = pipe inside diameter in ft
f = Friction coefficient (factor)
The friction coefficient is not constant but depends on
whether the flow is laminar or turbulent. This is described
by the Reynolds number, R
e
:

Where:
= Fluid dynamic viscosity in lb
f
.s/ft
2
= Fluid kinematic viscosity in ft
2
/S
If Re < 2300 this means the flow is Laminar
If Re > 4000 this means the flow is turbulent
If 2300 < Re < 4000 this means the flow is transition
Laminar flow only occurs at relatively low velocities and
describes a calm, well-ordered flow without eddies. The
friction coefficient for laminar flow is independent of the
surface roughness and is only a function of the Reynolds
number. The following applies for pipes with circular
cross-section:

Turbulent flow is an unstable flow with strong mixing. Due
to eddy motion most pipe flows are in practice turbulent.
The friction coefficient for turbulent flow depends on the
Reynolds number and the pipe roughness.
If the flow is turbulent then the friction parameter f is
calculated based on the following equation:

Where:
= Roughness of the pipe
If the flow is transition then the friction parameter f is
calculated based on the following equation:

Electric Submersible Pumps Mohamed Dewidar 2013
Chapter 2


83
f in transition equation is the friction factor calculated
in turbulent flow.
Example
Tubing size: 2-7/8 ID: 2.441
Tubing length: 6000 Q: 2000 bfpd
S.G: 0.95 : 10 cp
: 0.0068
Solution
D = 2.441/12 = 0.20342 ft
Q = 2000 * 0.00004874 = 0.09748 cuft/sec
[bpd = 0.00004874 cuft/sec]
v = Q/A = 0.09748 (/4)* (0.20342)
2
= 3 ft/sec
= 0.95 * 62.4 = 59.28 lb/cuft
= 59.28 32.16 = 1.811 slug/cuft
= 10 * 0.0000209 = 0.000209 lbf.s/ft
2

[1 cp = 0.0000209 lbf.s/ft
2
]
R
n
= DV = 0.20342 * 3 * 1.811 0.000209 = 5,289
R
n
> 4000 ==> Turbulent flow
f =[1.14+ 2*log
10
(D )]
-2
=[1.14+2*log
10
(0.20342 0.0068)]
-
2
= 0.05975
h
L
= f * (LD) * (V
2
2g) =
0.05975*(6000/0.20342)*(3
2
/(2*32.16)) = 246 ft
DP = * g * h
L
= 1.811*32.16*246/144 = 99.5 psi
2.23. ESP Design Examples
See chapter 9

Das könnte Ihnen auch gefallen