THE SEWAGE

PUMPING
HANDBOOK
 Foreword

Foreword

The use of submersible pumps in sewage and and construction is described in Section 2. Pump
drainage pumping applications has increased performance is dealt with in Section 3, offering
greatly in the last decades since they entered the methods for the calculation of pump performance
market. The introduction of heavy-duty submers- in various installations. Factors affecting pump
ible pumps with motor power ratings exceeding selection are also discussed. Section 4 offers infor-
500 kW has also made them available for central mation on pump testing. Basic design of pumping
municipal pumping duties. The good service stations is discussed in Section 5, offering design
record and high quality standard attained by information for both large and small applications.
these pumps has all but excluded the use of con- Continuous regulation of submersible pump oper-
ventional pumps in municipal service. ation by frequency control is described in Section
6. The concept of whole-life cost for pumps and
By the same token, the special characteristics of pumping installations is presented in Section 7.
submersible pumps have also required the devel- Matters relating to pump commissioning are pre-
opment of new knowledge on their implementa- sented in Section 8, whereas pump operation and
tion, such as the design of pumping stations. This servicing is described in Section 9. Section 10 deals
work has been advanced by both pump manu- with pumping station control and monitoring.
facturers and municipal engineers. Appendix A offers information on the hydraulic
characteristics of common pipe components for
The intention of this book is to bring the newest pipeline loss calculations. Appendix B presents a
information on both submersible pumps and method for the determination of sewage pump-
pumping stations to the use of all concerned ing station capacity and pump starting frequency.
professional people in a concise form. The book is
divided into Sections according to the related One objective of the book has been to make the
topics. contents easy to read and comprehend. The pre-
sentation is therefore enhanced with a large num-
Basic pump theory is described in Section 1, pro- ber of illustrations, providing examples of and
viding a reference background for the assessment complementary information on the matter at
of pump performance. Submersible pump design hand.

3
 Table of Contents

Table of Contents 3.3.2

3.3.3
Local Losses ....................................................43
Rising Main Characteristic Curve .............43
3.4 Rising Main Size ............................................44
1 Pump Theory ............................................. 7 3.4.1 Economy .........................................................44
1.1 The Head Equation .........................................7 3.4.2 Free Passage for Solids ................................ 45
1.1.1 Flow with Losses or Addition of Energy ....7 3.4.3 Avoiding Settling of Solids and Sludge ... 45
1.1.2 Fluid Flowing from a Container .................. 8 3.4.4.Water Hammer .............................................. 45
1.2 The Basic Pump Equation ............................. 8 3.4.5 Avoiding Water Hammer ........................... 47
1.3 Pump Curve and Losses ...............................10 3.5 Pump Duty Point ..........................................48
1.3.1 The Effect of Finite Number of Vanes ......10 3.5.1 Single Pump Operation ...............................48
1.3.2 Friction Losses Hf ..........................................10 3.5.2 Parallel Operation, Identical Pumps ........48
1.3.3 Discontinuity Losses Hs ...............................10 3.5.3 Parallel Operation, Different Pumps ........48
1.3.4 Leakage Losses Hv .........................................10 3.5.4 Serial Operation ........................................... 49
1.3.5 Other Losses ....................................................11 3.5.5 True Duty Point ............................................ 49
1.4. Cavitation and NPSH .....................................11 3.6 Sludge Pumping ........................................... 49
1.4.1 Definition of NPSH .......................................12 3.7 Complex Rising Mains ................................ 49
1.4.2 Reference Plane .............................................12 3.7.1 What Goes on in a Complex
1.4.3 Required NPSH ...............................................12 Rising Main? ................................................. 49
1.4.4 Available NPSH ..............................................14 3.7.2 Determination of Head ................................51
1.4.5 NPSH Safety Margin ..................................... 15 3.7.3 Pipe Size and Flow Velocity .........................51
1.4.6 Damming up of Suction Wells................... 15 3.7.4 Choice of Pump ..............................................51
3.7.5 Confirming Measurements ........................51
2 Pump Construction ................................. 16 3.8 Duty Point Evaluation for
2.1 General ............................................................16 Parallel Pumping Stations .......................... 52
2.2 Pump ................................................................18
2.2.1 Impellers ..........................................................18 4 Testing of Pumps .....................................54
2.3 Motors ............................................................. 27 4.1 Testing Arrangements ................................. 54
2.3.1 General ........................................................... 27 4.1.1 Production Testing ....................................... 54
2.3.2 Explosion-proof Motors .............................. 27 4.1.2 Field Testing, Duty Point .............................56
2.3.3 Motor Cooling ............................................... 27 4.2 Acceptance Tests .......................................... 57
2.3.4 Motor Tightness ...........................................29 4.2.1 Testing Standards ......................................... 57
2.3.5 Motor Bearings .............................................. 31
2.3.6 Motor Protection Devices .......................... 32 5 Pumping Stations ....................................59
2.4. Pump Connection ........................................ 34 5.1 Pumping Station Basic Design ..................59
2.5 Construction Materials, 5.1.1 Wet Well Volume and Surface Area ........59
Corrosion and Wear ..................................... 36 5.1.2 Pumping Station Inlet Pipe ....................... 60
2.5.1 Corrosion Resistance ................................... 36 5.1.3 Wet Well Floor Shape ................................. 60
2.5.2 Wear Resistance ........................................... 37 5.1.4 Stop Levels ...................................................... 61
2.5.3 Abrasive Liquids ............................................ 37 5.1.5 Start Levels .....................................................62
5.1.6 Suction Pipe Dimension and Design .......62
3 Pump Performance .................................. 38 5.1.7 Pumping Station Internal Pipework ........63
3.1 Pump Head .................................................... 38 5.1.8 Flushing Devices ...........................................63
3.1.1 Submersible Pumps ..................................... 38 5.1.9 Odour Problems in Pumping Stations .... 64
3.1.2 Dry-installed Pumps ....................................39 5.1.10 Pumping Station Design Examples ......... 64
3.2 Pump Performance Curves ........................39 5.1.11 Dry-installed Pump Positions ....................67
3.2.1 H Curve ...........................................................39 5.2 Package Pumping Stations ........................ 68
3.2.2 Efficiency Curves ..........................................40 5.2.1 Out-of-doors Pumping Stations .............. 68
3.2.3 Power Curves .................................................40 5.2.2 Indoor Pumping Stations ...........................70
3.2.4 NPSH Curve ....................................................40 5.3 Pumping Stations with
3.3 Pipe Losses and Rising Main Column-installed Pumps ............................70
Characteristic Curves ...................................41 5.4 Pumping Station Dimension Selection ... 72
3.3.1 Friction Losses ................................................41 5.4.1 Regular Sewage Pumping Stations .......... 72

5
 Table of Contents

5.4.2 Stormwater Pumping Stations ..................72 10.2.1 Wet Well Water Level Sensors ..................89
5.4.3 Combined Sewage Pumping Stations 10.2.2 Current Sensor ............................................. 90
and Retention Basins ...................................73 10.2.3 kWh Meter .................................................... 90
5.5 Pump Selection ..............................................74 10.2.4 Phase Failure Relay ...................................... 90
5.5.1 Pump Selection Based on Pump Curves ..74 10.2.5 SARI 2 Monitoring Device .......................... 90
5.5.2 Observing Pump Efficiency .........................74 10.2.6 ASM 3 Alarm Status Module ......................91
5.5.3 Number of Pumps .........................................75 10.3 Pump Control Units ......................................91
5.6 Special Considerations ................................76 10.3.1. Control Features ............................................91
5.6.1 Pump Vibrations ............................................76 10.3.2 Condition Monitoring Features ................92
5.6.2 Pump Noise ....................................................77 10.3.3 Parameters and Signals ..............................92
10.3.4 Data Logging and Analysis .........................93
6 Frequency-controlled Sewage Pumps ........... 78 10.3.5 User Interface ................................................93
6.1 General ........................................................... 78 10.4 Remote Control and Monitoring System 93
6.1.1 Pump Motor Selection ................................ 78 10.4.1 Different Levels for Remote Control ........93
6.1.2 Maximum Frequency .................................. 78 10.4.2 Software and Hardware .............................94
6.1.3 Minimum Frequency and Minimum 10.4.3 Data Transmission ....................................... 95
Performance.................................................. 79 10.4.4 Alarm Transfer ............................................... 95
6.1.4 Pump Frequency Curves ............................. 79 10.4.5 System Integration ......................................96
6.1.5 Pump Clogging .............................................80 10.5 Internet & WAP Based Remote Control
6.1.6 EMC Cable Requirement ............................80 and Monitoring .............................................96
6.1.7 Bearing Currents ..........................................80
6.1.8 High Tension ..................................................81 Symbols .............................................................98
6.1.9 Explosion-proof Motors ..............................81
6.1.10 Guaranteed Values .......................................81 APPENDIX A ...................................................... 100
6.1.11 Tests with Frequency Controller
(String Tests) ...................................................81 APPENDIX B....................................................... 108
6.1.12 Collaboration with the Pump
Manufacturer .................................................81

7 Pump Whole-life Cost Evaluation .......... 82

7.1 General ........................................................... 82
7.2 Calculation Period ........................................ 82
7.3 Investment Costs ......................................... 82
7.4 Energy Costs .................................................. 83
7.4.1 Efficiency Over Time .................................... 83
7.4.2 Energy Usage Calculations ........................ 84
7.5 Maintenance Costs ...................................... 84
7.6 Cooperation With Pump Suppliers .......... 85
7.7 Life Cycle Cost Publication ......................... 85

8 Commissioning ....................................... 86

9 Operation and Service ............................ 87

9.1 Safety .............................................................. 87

10 Pumping Station Control and

Condition Monitoring ............................. 88
10.1 Local Control Methods ............................... 88
10.1.1 Manual Control Units ................................. 88
10.1.2 Relay-based Control Units ......................... 88
10.1.3 Programmable Logic Controllers .............. 88
10.2 Sensors for Pump Control and
Condition Monitoring .................................89

6
 Pump Theory 1

1 Pump Theory are consequently called static head, pressure head

and kinetic head, respectively.

This section is a primer of fluid pumping theory The equation is essential for fluid mechanics and
and provides the reader with the theoretical back- can be used to account for many hydrodynamic
ground knowledge essential for deeper under- phenomena, such as the decrease in pressure that
standing of the pumping process. accompanies a reduction in a flow cross section
area. In this case the fluid velocity increases, and
1.1 The Head Equation for the total head to remain constant and assum-
ing the potential head remains unchanged, the
Figure 1 shows part of continuous fluid flow in a pressure term or static head, must decrease.
duct. Between the two observation sections 1 and
2 no energy is transferred to or from the fluid and Fig. 1
the flow is assumed to be frictionless. Thus the
total energy of the fluid relative to a horizontal v1
reference plane T at the two sections must be p1
equal. The total energy comprises components for
potential energy, pressure energy and kinetic 1
energy, and for a fluid particle with a mass m the Q
energy at the observation sections is as follows: v2
h1
p2
Section 1 2 2
h2
Potential mgh 1 mgh 2
Energy T
Section showing flow of liquid through two obser-
Pressure p p
Energy mg ------1 mg ------2 vation cross sections. T is a reference plane for the po-
ρg ρg tential heads h1 and h2, p1 and p2 are the prevailing
pressures and v1 and v2 the fluid velocities at sections
Kinetic 1 2 1 2 1 and 2.
--- mv 1 --- mv 2
Energy 2 2

1.1.1 Flow with Losses or Addition of

where ρ is the fluid density and g the acceleration Energy
of gravity. If there are losses in the flow between section 1
and section 2 in Figure 1, the head equation 1 can
For a flow without losses the total energy in sec- be written
tion 1 and 2 will be equal, thus
2 2
p v p v
p 1 2 p 1 2 h 1 + ------1 + -----1- = h 2 + ------2 + -----2- + H r (2)
mgh 1 + mg ------1 + --- mv 1 = mgh 2 + mg ------2 + --- mv 2 . ρg 2g ρg 2g
ρg 2 ρg 2

where Hr is the head loss.

Dividing both sides of the equation with the term
mg it is obtained
If energy is added to the flow by placing a pump
2 2 between section 1 and section 2 in Figure 1, the
p v p v equation 2 can be written
h 1 + ------1 + -----1- = h 2 + ------2 + -----2- (1)
ρg 2g ρg 2g
2 2
p v p v
h 1 + ------1 + -----1- + H = h 2 + ------2 + -----2- + H r (3)
This equation is called Bernoulli's equation after ρg 2g ρg 2g
the engineer who first derived it. The terms of the
equation are expressed as heads, and the terms
where H is the pump total head.

7
 1 Pump Theory

1.1.2 Fluid Flowing from a Container To accommodate for losses present, a flow coeffi-
An example of the application of the Bernoulli cient µ is added to equation 6, whence
equation is the calculation of the flow rate of a
fluid flowing freely from an open container. q 1 = µA 2 2gh
(7)
Figure 2 shows an open container with an outlet
orifice near the bottom. For practical purposes the The flow coefficient µ is dependent on the shape
area A1 is assumed much larger than the orifice of the orifice, and can be obtained from text
area A2, and the atmospheric pressure p1 in the books on the subject. If the fluid level in the con-
tainer is allowed to recede, the level height h will
container is equal to that outside the orifice, p2.
change, which will have to be accommodated for
in calculations.
Fig. 2
A1 p1 1.2 The Basic Pump Equation
The basic pump equation is used to calculate and
design geometrical shapes and dimensions of
v1 centrifugal pumps. The basic pump equation is
h also used to deduce the pump Q/H curve.

A pump impeller vane and its associated velocity

v2 vectors are shown in Figure 3.
T
v = absolute fluid velocity
A2 p2 = p2 w = velocity relative to the vane
u = perimeter velocity
Section of a fluid container with an outlet orifice near vu = tangential component of absolute velocity
the bottom. A1 and A2 are the cross section areas of the
surface and the outlet orifice, h the height difference vm = radial component of absolute velocity
between surface and orifice centre line, v1 surface re-
cession velocity and v2 liquid outlet velocity through The relative velocity is parallel to the vane at any
the orifice. Ambient pressure is constant. given point.

Also v u1 = v 1 ⋅ cos α 1 and v u2 = v 2 ⋅ cos α 2

Choosing the centre line of the orifice as the refe-
rence plane T, the term h2 is equal to zero and h1
equal to h. Because A1 is much larger than A2, the Assuming the flow to be without losses and the
2 number of vanes infinite (∞), the familiar basic
v1 equation of pump theory can be derived using the
kinetic head -----
2g
- can be assumed as zero. Thus
laws of mechanics. This relationship is known as
the head equation 1 can be written the Euler equation and is expressed as:

2 1
v H t∞ = ---(u 2 v u2 – u 1 v u1)
h = -----2- (4) g . (8)
2g

where the index t refers to a flow without losses

whence and ∞ refers to the assumption of infinite number
of vanes ensuring complete fluid direction.
v2 = 2gh (5)
In an actual pump neither of these assumptions
For volume flow without losses is obtained can be satisfied, as friction losses are always
present, and the finite number of vanes will not
q 1 = A 2 2gh direct the flow entirely in the direction of the
(6) vane.

8
 Pump Theory 1

The reduction in head caused by losses in the flow It can be shown that both ηh and k are less than
is taken into account by the hydraulic efficiency ηh, unity. They will not be discussed in further detail
and the reduction due to the deviation of the flow here.
from ideal angle β2 is accounted for by a vane
coefficient k. With these modifications, the Euler Centrifugal pumps are normally designed with
equation for an actual pump reads as follows: α1 = 90 °, whence vu1 = 0.

η Thus the basic pump equation is simplified to

H = -----h-(ku 2 v u2 – u 1 v u1) (9)
g
u 2 v u2
H = kη h --------------
- (10)
g

Fig. 3

u2
v2

vm
2
2

vu
2

w1

d2
vm d1
1
vu
1
v1 1

u1

Pump impeller vane with the velocity triangles at leading and trailing edges. Fluid absolute velocity v, relative velocity
w, vane perimeter velocity u, liquid absolute velocity tangential component vu and radial component vm.

9
 1 Pump Theory

1.3 Pump Curve and Losses The effects of the discontinuity losses are shown
in Figure 4.
The ideal head obtained from the Euler equation
is independent of the volume rate of flow Q. If the Fig. 4
Q/Ht∞ curve is plotted, Ht∞ is indicated by a
straight line. The real Q/H curve is derived from H Reduction of flow, Q
caused by leakage losses, Hv
this by subtracting the effects of the finite num-
ber of vanes and various other losses that occur
Effect of finite number of vanes Ht
within the pump. Please refer to Figure 4.
Friction losses Hr

1.3.1 The Effect of Finite Number of Vanes HN Discontinuity losses Hs

As noted earlier, a finite number of vanes
decreases the head by the vane factor k. Taking
this into account, the theoretical head Ht is
obtained. It can be written: QN Q

H t = kH t∞ (11) True pump Q/H curve (H) reduction from theoretical

pump head Ht∞ .
Ht is not perfectly linear because the vane
coefficient is slightly dependent on the volume
rate of flow Q. The head reduction from Ht∞ to Ht Fig. 5
is not caused by flow losses, but by deviation of w1'
losses
the fluid from the ideal flow angles due to the v1'
finite number of vanes. Q > QN

Q = QN
1.3.2 Friction Losses Hf Q < QN w1

Friction losses occur as the fluid flows through the

w1" v1
passages of the impeller and the pump casing.
They increase approximately with the square of v"1
the flow rate Q.
1
u1
1.3.3 Discontinuity Losses Hs
Discontinuity losses are generated in the follow- Vane leading edge relative velocities (w) and losses
at various flow rates. Minimum losses occur at the
ing areas:
pump nominal flow when the fluid angle of attack
• At the vane leading edge where the fluid hits
is equal to the vane leading edge angle β1.
the vane tip. The loss is smallest at the pump
design point, where the fluid contacts the vane
at the vane angle β1. The losses increase with
increasing deviation of the contact angle from 1.3.4 Leakage Losses Hv
the vane angle β1, see Figure 5. Leakage losses occur at the clearance between
• At the vane trailing edge losses occur due to impeller and pump casing. Even if the clearance is
eddies shed by the vane. These increase ap- kept as small as possible, a small backflow passes
proximately with the square of the flow rate. from the high pressure area at the impeller rim to
• In the pump casing at flow rates other than the low pressure area of the impeller eye. Thus
the design value, when the flow velocity at the the flow through the impeller is slightly larger
casing differs from that at the impeller perime- than the flow out of the pump casing, and the
ter. The effect is shown in Figure 6. The velocity pump head is met with a reduced flow, the differ-
differences create turbulence leading to losses, ence being the leakage loss Hv. The effect of the
growing with increasing difference of actual leakage loss is shown in Figure 4. As the pump
flow from design flow. wears out, this loss will increase.

10
 Pump Theory 1

1.3.5 Other Losses 1.4. Cavitation and NPSH

There are further losses in a centrifugal pump, not
affecting the Q/H curve, but that will increase the Cavitation is caused by the formation and collapse
motor shaft power requirement. These include: of vapour bubbles in a liquid. Vapour bubbles
• friction losses at the impeller outside surfaces form when the local static pressure in a flowing
• shaft seal friction losses liquid decreases to or below the liquid vapour
• bearing friction losses pressure at ambient temperature. When the bub-
ble, or void, moves with the flow to an area with a
For submersible pumps, the last two items are higher pressure, it will rapidly collapse. The implo-
included in the motor losses. sion causes a transitory, extremely high local
shock wave in the fluid. If the implosion takes
place near a surface, the pressure shock will, if
Fig. 6
Velocity in casing occurring repeatedly, eventually erode the surface
material.

The cavitation phenomenon will typically occur in

Absolute velocity after impeller (vu) centrifugal pumps at a location close to the impel-
ler vane leading edge, see Figure 7. Cavitation may
also lower the pump Q/H curve and efficiency. A
cavitating pump emits a typical rattling noise, like
Resulting losses
sand being pumped through the pump. No pump
Q<QN Q=QN Q>QN
material will completely withstand cavitation, so
Effect of the difference of velocities in the pump cas- care should be exercised if the pump operating
ing and at the impeller perimeter. Pump casing di- conditions present a risk of cavitation.
mensions are designed to accommodate nominal
flow at the perimeter speed, leading to losses at other
Wear marks from cavitation typically occur locally
flow rates.
and consist of deep pittings with sharp edges. The
pittings can be several millimetres deep, please
refer to Figure 8.
Fig. 7 Imploding Normally pump curves published for submersible
vapour bubbles
pumps are drawn so that a pump in normal sub-
Vapour bubbles
merged installation will not cavitate as long as
the duty point is on the allowed section of the Q/
H curve.

Fig. 8

( Q > QN )

Pumped fluid hitting the vane leading edge at an

angle different than the vane angle. Eddies and low
pressure zones will form on the other side of the
vane. If the pressure falls below the vapour pressure,
vapour bubbles will form. Moving entrained in the
flow to an area with higher pressure, they will even-
tually implode. The consequent high pressure im-
pact may lead to pitting and erosion of the adjacent
structure. Typical cavitation pitting in impeller.

11
 1 Pump Theory

Fig. 9
pmin
g HORIZONTAL PUMP VERTICAL PUMP
Hrt

h
Minimum Pressure
NPSH v02
required 2g
pb
g
hA
Reference Plane

ht
pb

Dimensions and reference pressures for NPSH calculations.

If the submersible pump is installed dry with a Figure 10 shows the principle of static liquid pres-
suction pipe, the installation must be checked for sure distribution in inlet pipe, pump and pressure
cavitation. In these cases the concept of NPSH is pipe of a dry pump installation.
used.
1.4.2 Reference Plane
1.4.1 Definition of NPSH The reference plane is the plane on which the
NPSH is the acronym for Net Positive Suction NPSH calculations are performed. It is the hori-
Head. The following pressure heads are used for zontal plane through the centre point of the circle
the calculation of NPSH: described by the tip of the vane leading edge. For
horizontal pumps the reference plane coincides
ht = inlet geodetic head with the shaft centre line. For vertical pumps the
hA = height difference between reference location of the reference plane is stated by the
plane and tip of vane leading edge. pump manufacturer.
Hrt = flow losses in inlet pipe
2 1.4.3 Required NPSH
v
-----0- = pressure drop caused by inlet velocity The required NPSH is obtained from the following
2g
equation:
∆h = local pressure drop at vane leading edge
pb = ambient pressure at liquid level 2
v
pmin = minimum static pressure in pump NPSH required = h A + -----0- + ∆h (12)
2g
pv = liquid vapour pressure at prevailing
temperature This is also called the pump NPSH value. It can be
presented as a function of flow as shown in Figure
The pressure heads are shown in Figure 9. 11. It is independent of temperature and type of
liquid being pumped. The pump manufacturer is
In order to avoid cavitation, the minimum static required to state NPSH as a numeric value or
pressure in the pump (pmin) must be larger than curve.
the liquid vapour pressure, or
Any pump will actually have different NPSH-val-
p min > p v ues, depending of definition of occurrence, as can

12
 Pump Theory 1

Fig. 10

Liquid static pressure

pb
ht

ht

pb
g

Lowest pressure in pump

Vapour pressure

pmin
Absolute 0 pressure
g pv
g

Pressure variation in a dry pump installation. Distribution of static liquid pressure in inlet pipe, pump and pressure
pipe.

be seen in Figure 12. According to the testing stan-

Fig. 11
dards used by pump manufacturers, the NPSHr is
defined as the situation where pump head is
decreased by 3% due to cavitation. This value is
defined as NPSH3. NPSH required

Light cavitation can be harmless to the pump if (m)

the vapour bubbles do not implode near the
pump structural parts, such as the impeller vane.

The difference between the various NPSH values

is greater in pumps with impellers with few
vanes. Thus single-vane impellers have the great-
est differences in NPSH values with the difference
being caused by the NPSH3 curve dropping, and Q
QN
the tests thus giving too favourable readings.
Therefore an NPSHr curve based on the 3% rule of
Typical variation of required NPSH with pump
the standard is a poor base for a cavitation risk
flow rate.
assessment in pumps with few vanes. The NPSHr

13
 1 Pump Theory

curve published by a pump manufacturer should

Fig. 13
in principle guarantee that no damages will occur
in the pump if the pump is operated above it. This Temp ( C) Head (m)
is especially the case for wastewater pumps, 100 pv
which have a low number of impeller vanes. The 10
g
problem is that there is no accurate way of testing
and establishing such an NPSH value. 9

8
1.4.4 Available NPSH
90 7
The available NPSH indicates the pressure availa-
ble for the pump suction under the prevailing 6
conditions. This may be called the pumping sta- 5
80
tion NPSH.
4

p pv 70 3
NPSH available = ------b – H rt – h t – ------
ρg ρg (13)
60 2
The term ht is positive when the reference plane is 50
above the liquid surface and negative if below it. 1
40
Available NPSH is determined by the pumping 0,5
30
station designer.
20
10
Fig. 12
Vapour pressure for water as a function of temper-
ature.
NPSH NPSHF (Cavitation free)

NPSHonset of noise

NPSHonset of material loss

NPSH0 (0% Head drop)

Fig. 14
NPSH3 (3% Head drop)

Barometric
pressure
m H 2O

Different NPSH curves.

Figure 13 shows vapour pressure for water as a

function of water temperature. Altitude km

Figure 14 shows atmospheric pressure as function Atmospheric pressure as function of elevation above
of elevation above sea level. sea level.

14
 Pump Theory 1

1.4.5 NPSH Safety Margin

NPSHavailable ≥ NPSHrequired + Safety margin

The NPSH margin must be sufficient to allow for

variations in a situation where the real conditions
may differ from those theoretically calculated.
The suction pipe flow losses may be inaccurately
estimated and actual pump operation point may
differ from the theoretical because of variations in
the Q/H curve and inexact pressure pipe resis-
tance calculations. Harmful cavitation may occur
earlier than expected, or at greater NPSH values
than NPSH3 (Figure 12). Manufacturing technical
variations of the shape of the vane leading edge
may affect cavitation behaviour. The required
NPSH may also be affected by the inlet pipe
shape.

For horizontally installed pumps with straight

suction pipes, a safety margin of 1 to 1,5 m is suit-
able.

For vertically installed pumps the safety margin

should be set at 2 to 2,5 m, provided that a reduc-
ing bend is used before the pump inlet. Bend cen-
treline curvature radius should be at least D1 + 100
mm, where D1 is the diameter of the larger open-
ing.

The matter of NPSH, safety margins and measur-

ing methods for NPSH are discussed in details in
the EUROPUMP publication “NPSH FOR ROTODY-
NAMIC PUMPS,REFERENCE GUIDE ”(1997).

1.4.6 Damming up of Suction Wells

In practical installations situations may arise,
where the liquid level on the suction side rises
and the pump head decreases so that the pump
duty point moves to a sector where NPSHr > 10 m.
No cavitation will occur, however, since NPSHavail-
able will also rise and still be larger than NPSHre-
quied. Typical installations where this situation will
arise are dry dock drainage pumping, sewer block-
age situations and drainage pumping with vary-
ing suction liquid levels.

15
 2 Pump Construction

2 Pump Construction Many submersible pumps can also be installed

dry like conventional pumps. This type of installa-
tion ensures uninterrupted operation of the
This section describes the construction of modern installation in case of flooding of the dry well.
electric submersible pumps. Various designs and
the main parts of the pumps are discussed as well Submersible pumps are available for a number of
as topics concerning pump operation and mainte- applications with different requirements, and dif-
nance. The study is limited to pumps for munici- ferent designs for various special applications
pal sewage, drainage and raw water. have been devised.

2.1 General A submersible pump comprises a waterproof

motor and matching pump components. The
The submersible pump is a unit combining a pump components include the impeller, the pump
pump and an electrical motor to an enclosed unit, casing and the required connection parts for dif-
suitable for submersible installation in a wet well ferent installation alternatives. These include a
holding the liquid to be pumped. The submersible guide shoe for submersible installation onto a
pump may be connected to the pressure piping matching connecting baseplate, a stand for porta-
with a special baseplate connection at the bottom
of the wet well for ease of installation and
removal, or it can be installed connected with a Fig. 16
flexible hose or other arrangements with riser
pipes. Power to the unit is fed through one or
more flexible cables, supplied with the pump in
lengths suitable for the installation.

Fig. 15

Section of a GRUNDFOS 17 kW submersible pump

showing details of motor and pump. The pump is fit-
Section of a GRUNDFOS 2,4 kW submersible pump ted with a guide shoe for use with a submerged base-
showing details of motor and pump. The pump is fit- plate in the wet well, facilitating easy pump
ted with a guide shoe for use with a submerged base- installation and removal. The pump casing is adjust-
plate in the wet well, facilitating easy pump able with set screws for maintenance of impeller suc-
installation and removal. tion clearance.

16
 Pump Construction 2

ble pumps and the necessary connection flanges, Fig. 17

stand for dry-installed pumps and seat ring for
column-installed pumps.

The motor is a dry squirrel-cage electric motor

matching a range of pump parts for various
duties. Motor and pump have a common shaft
with the bearings and shaft seals housed in the
motor. The motor also includes watertight cable
inlets and a handle for lifting the unit.

Figure 15 shows a modern small submersible sew-

age pump and Figure 16 a medium-sized submers-
ible sewage pump. Submersible sewage pumps
are available with motors rated from under 1 up to
500 kW for duties ranging from light portable use Section and outline of a 160 kW submersible pump.
to large city sewerage system main pumping The pump is intended for horizontal dry installation
installations. A submersible pump for dry installa- and connects with integral flange joints to both suc-
tion is shown in Figure 17. tion and pressure pipework. The submersible design
permits flooding of the installation without risk of
damage to the pump.

Fig. 18

S
S
S

a b c
closed semi-open open

Different impeller designs. The closed impeller has integral shrouds on both sides of the vanes, whereas the semi-
open impeller incorporates only one shroud on the back side. An open impeller consists of only a hub and vanes, re-
lying on close clearances (s) to the pump casing for its function.

17
 2 Pump Construction

2.2 Pump Fig. 19

The pump comprises the impeller and the pump

Closed Impeller
casing as well as ancillary devices and fittings.
Open Impeller
2.2.1 Impellers
Submersible pumps are fitted with different
impeller designs depending on intended use. The
various impellers can be classified as Operating time
• impellers for sewage pumps
• impellers for macerating pumps Test results from a comparison of the effect of wear
• propellers for axial pumps on pump efficiency for different impeller types.

Impellers can also be classified according to con- Impellers for Sewage Pumps
struction as closed, semi-open or open impellers.
In order to avoid pump blockage, or clogging, spe-
These are shown in Figure 18. Semi-open impellers
cial impellers have been developed for sewage
and open impellers rely on the close clearance
pumping. These include single-channel impellers,
between impeller and pump casing (about 0,5
double-channel impellers and vortex impellers.
mm) for their function. The efficiency of these
The design principles of these are shown in Figure
impellers is very sensitive to wear and will
20. For very large sewage pumps, impellers with a
decrease rapidly as the clearance increases. Figure
multitude of vanes may also be used.
19 shows the effect on pump efficiency from wear
on closed and open impellers. The open and semi-
open impellers are also susceptible to impurities Free Passage
becoming jammed between impeller and suction The concept of free passage is of special relevance
plate, slowing down or completely stopping the to sewage pumps. It refers to the ability of the
pump. pump to let solids in the pumped liquid pass

Fig. 20

Winglets

Counterweights

Vortex impeller 1-channel impeller 2-channel impeller

Impeller types for sewage pumps.

18
 Pump Construction 2

through, and thus to the capacity of being non- Fig. 22

clogging. The dimension of the free passage usu-
ally refers to the largest spherical object that may
pass through the impeller and the casing open-
ings. If the free passage is described with two
numbers, it refers to the largest oblong object
that can pass crosswise through the pump.

The ability of a pump to operate without clogging

relates strongly to the free passage, as can be
seen in the diagram in Figure 21. Normally a free
passage of 80 mm will be sufficient for
unscreened sewage in small and medium-sized
pumps. In larger pumps (flow > 100 l/s) the mini-
mum free passage should be at least 100 mm.

Free passage alone does not ensure good proper-

ties against clogging in a pump. Impeller and vane
geometry must also have features that prevent
blockages. Pumps from different manufacturers A GRUNDFOS S-1 single-channel impeller for sew-
have varying qualities in this respect. There are age use. The impeller is semi-axial in design with
cases where a pump clogging problem has been one long continuous vane, ensuring good proper-
solved by changing pump make, even with the ties against clogging. The asymmetric design re-
pumps having equal free passages, same number quires the casting to include counterweight
of vanes and pump speed. The tendency of sew- masses to facilitate static and dynamic balancing
age to choke a pump may vary from one location of the impeller
to another, with “easy” and “difficult” pumping
stations. The design of the sewer line leading up
to the pumping station is important for the func- Single-channel Impellers
tion of the pumps, because they must be able to A single-channel impeller is shown in Figure 22.
handle any agglomeration of solids originating The single vane is designed as long as possible for
there. Real conditions in sewage systems cannot best efficiency within the limits set by the require-
be simulated in laboratories, and the good proper- ment of free passage. The impeller having only
ties of the Grundfos sewage pumps against clog- one route of passage for the pumped liquid
ging are based on long-term practical experience. ensures good inherent characteristics against
clogging. The asymmetric shape requires the
Fig. 21 impeller to include integral counterweights for
balance. Highest attainable efficiency is 70…75%.
Clogging probability

40 60 80 100 120
Free passage (sphere) [mm]

A diagram showing the relation between probability

of clogging and pump free passage. Good safety
against clogging is achieved with 80 mm free pas-
sage.

19
 2 Pump Construction

Fig. 23 Fig. 24

A GRUNDFOS S-2 double-channel impeller. Good A GRUNDFOS SuperVortex impeller. The design in-
properties against clogging are achieved with re- cludes patented vane winglets. The winglets pre-
cessed vane leading edges and a semi-axial design. vent the formation of secondary eddies over the
The symmetric design is inherently in balance. vane edges, greatly improving pump efficiency.

Double-channel Impellers Vortex Impellers

A double-channel impeller is shown in Figure 23. The principle of the vortex impeller is to induce a
The inherent difficulty with double-channel strong vortex in the open pump casing. The
impellers is that long, fibrous impurities may pumping action of the vortex pump is therefore
enter both channels and get caught by the vane indirect, with the impeller being situated outside
leading edges, causing the pump to clog. This sit- the main liquid flow. Vortex impeller pumps have
uation can be alleviated by good vane leading inherently excellent properties against clogging,
edge design, and this can be found only by devel- and the pumps run very smoothly. The use of
opment work under real conditions in difficult small vortex impeller pumps for sewage is
pumping stations. With the right design and a increasing largely due to improved design and
free passage of at least 100 mm, double-channel efficiency in later years. They are also used as sand
impellers can be designed to handle unscreened separation pumps in sewage treatment plants. A
sewage without clogging. Highest attainable effi- vortex impeller is shown in Figure 24. Highest
ciency is 80…85% for double-channel impellers. attainable efficiency is around 50% for vortex
impellers. It is essential to note that the efficiency
Three- and Four-channel Impellers in the flow range 3…15 l/s of vortex pumps is
In very large pumps impellers with three or four roughly equal to that of single-channel pumps.
vanes can be used and still have a free passage of
at least 100 mm and an impeller with good prop- Flow and Head (Q/H) Ranges for Differ-
erties against clogging. Also for these impellers ent Impeller and Submersible Pump
the design of the vane leading edge is decisive. Types
Highest attainable efficiency is 82…86% for these
Figure 25 shows the typical application areas for
impellers.
different sewage pump and impeller types of the
GRUNDFOS range. It can be seen that with
increasing flow and pump size the number of
vanes of the impeller also increases. The diagram
also shows the Q/H area for which submersible
pumps are available for sewerage use. The largest
pump in the Grundfos range has a motor of 520
kW power.

20
 Pump Construction 2

Fig. 25

Vortex 1-channel 2-channel 3-channel 4-channel

Flow and head (Q/H) ranges for different impeller types.

Fig. 26

GRUNDFOS Grinder pump. The macerating unit is made of hardened stainless steel.

21
 2 Pump Construction

Impellers for Macerating Pumps Fig. 27

For installations with very small amounts of sew-
age, macerating pumps have been developed.
Typical applications are pumping stations for sin-
gle homes, small developments or camping areas.
The required flow is very small, sometimes less
than 1 l/s, but the total head may be high because
of long and narrow rising mains. The flow for a
macerating pumps are typically 1…5 l/s, with
heads as high as 50 m.

In macerating pumps the solids are shredded into

small fragments of around 10 mm, which makes it
possible to use rising mains of small dimensions,
usually DN 40…DN 80. For the very small flows
from single pumping stations even smaller piping
will be used, in order to attain a flow velocity of at
least 0,5 m/s.

Macerating pumps may not be used for sewage

with sand content, since the shredding unit is very
susceptible to wear. Where macerating pumps are
considered for larger installations comprising sev-
eral buildings, it is always advisable make a tech-
nical and economic comparison with a solution
based on conventional pumps.

Figure 26 shows a GRUNDFOS macerating pump.

Outside of the impeller is a macerating unit with
sharp cutting elements installed. The macerating
unit is made of hardened stainless steel.
GRUNDFOS propeller pump. Blade angle is
Propellers for Axial Pumps adjustable for best efficiency.
Axial pumps, using the submersible motors from
sewage pumps, have been developed by many
manufacturers. Figure 27 shows a GRUNDFOS The operating range (Q/H area) of the GRUNDFOS
axial-flow pump with adjustable-pitch propeller. propeller pumps is shown in Figure 28. Part of the
The design incorporates trailing fixed vanes that range is also covered by channel-impeller pumps
transform the rotating movement of the water for column installation, which may be a more suit-
into pressure energy, increasing pump efficiency. able choice for many applications. The final choice
Propeller pumps are normally column-installed. between these should be based on desired duty
point and application. The pump manufacturer
Propeller pumps are used for stormwater and should be consulted in the selection process in
flood water pumping, drainage, irrigation and raw difficult projects.
water pumping as well as for effluent pumping in
sewage treatment plants. Propeller pumps are not
suitable for unscreened sewage because of risk of
clogging. Small and medium-sized propeller
pumps are not suitable for sewage treatment
plant internal circulation pumping of e.g. return
sludge, since they may clog and get jammed by
the fibres present in these fluids. Highest attain-
able efficiency for propeller pumps is 75…85%.

22
 Pump Construction 2

Fig. 28
10

0
0 100 200 500 1000 2000 4000 5000

GRUNDFOS propeller pump flow and head (Q/H) range.

For recirculation pumping in sewage treatment

Fig. 29 plants special axial pumps have been developed
as shown in Figure 29. They are intended to oper-
ate at very low heads, only 0,3… 1,0 m, and high
flow rates, up to 2000 l/s. These pumps are
designed non-clogging with back-swept blades,
large (10 mm) clearance between blade tips and
casing, and by omitting the lead vanes. Highest
attainable efficiency for circulation axial pumps is
35…50%. The delivery loss at the exit of the thim-
ble is significant for the head. Additional head can
be attained by using a conical thimble, lowering
losses.

Submersible recirculation pump for sewage treat-

ment plant. The pump is lowered in place along
guide rails.

23
 2 Pump Construction

Fig. 30 Fig. 31

Auxiliary
vanes

GRUNDFOS recirculation pump S

p's
Impeller Auxiliary Vanes
Auxiliary vanes on the outside of the shrouds are ps
an important feature of the impellers of small
sewage pumps. The auxiliary vanes increase the
velocity of the flow of fluid in the space between
the impeller and the pump casing. Figure 31 Pressure distribution
with auxiliary vanes
shows the location of auxiliary vanes on a single-
channel pump impeller.

Auxiliary vanes assist the pump operation by Pressure distribution

without auxiliary vanes
performing the following functions:
• Decrease axial loads on bearings, particularly if
semi-open impellers are being used
• Reduce impeller and casing wear at the suction The effect of the auxiliary vanes on the suction
clearance shroud is a lowered pressure difference p's over the
• Prevent the wedging of fibres in the suction suction clearance. With less back-flow the suction
clearance will last longer and the risk of jamming is
clearance
reduced.
• Prevent fibres and rags from wrapping around
the pump shaft behind the impeller.

The use of auxiliary vanes extending to the

Suction Clearance
shroud perimeter is not possible on large impel-
lers, since at high flow rates they could cause a The clearance between the impeller and the
pressure drop below the vapour pressure of the pump casing should be as small as possible in
liquid, leading to cavitation. Large pumps are, order to reduce leakage losses. The suction clear-
however, less prone to jamming because of high ance is in the order of 0,5...1,0 mm for most cen-
motor torque. Auxiliary vanes are therefore not trifugal pumps. The clearance can be designed
included on the inlet side of large impellers. cylindrical or axial as shown in Figures 32 and 33.

24
 Pump Construction 2

Fig. 32
Pump performance and efficiency over time are
dependent on the suction clearance being kept
within specified limits. The lowering effect of the
suction clearance on pump efficiency and head
can be calculated with the following empirical
equation:

2
∆η ≈ ∆H ≈ K + K – K
(14)

where

2 H
K = 0, 008 ⋅ s ⋅ ----
Q

Q= flow [l/s]
H= head [m]
s= clearance [mm]
∆η and ∆H are proportional.

For semi-open impellers the effect is increased by

the factor 1,5.

Figure 34 shows the results of a test where a

Cylindrical suction clearance. The design is suscepti- pump was operated with varying suction clear-
ble to jamming, since fibres that get wedged in the ance.
space between impeller and casing may accumulate
and drag down the pump. In case of wear, a wear If the suction clearance widens to 2...3 mm for
ring on the suction cover and the impeller need to be impellers without auxiliary vanes and to 4...5 mm
replaced or re-machined. for impellers with auxiliary vanes, it is necessary
to restore the clearance in order to retain pump
performance. If the suction clearance is made
adjustable, this procedure is easily done by service
Fig. 33
technicians in the field, whereas a pump with

Fig. 34

Axial suction clearance. This design is less prone to

jamming, since drag forces will remove wedged ma-
terial towards the pump suction. The clearance can
be made adjustable for ease of maintenance and Effect of different suction clearance dimensions on
wear compensation. pump curve and efficiency.

25
 2 Pump Construction

Fig. 35 Fig. 36

Suction clearance external adjustment system on

dry-installed pumps.

Suction clearance adjustment system with three Impeller Attachment

set screws. The attachment of the impeller onto the shaft
must be both reliable and easy to dismantle.
Removal is necessary for shaft seal maintenance,
and for impeller replacement if the pump is used
fixed impeller suction clearance will have to be for pumping abrasive materials. The impeller may
brought to the shop for overhaul, or worse, have either a cylindrical or tapered fit onto the
scrapped for high costs of spare parts and work. shaft end.

In pumps with adjustable axial suction clearance A shaft joint tapered to the right angle is easy to
performance can always be guaranteed by check- dismantle. The tapered joint is additionally tight-
ing and adjusting the suction clearance during ened with a screw, which makes the joint free of
routine maintenance. Figure 35 shows a submers- play and rigid.
ible pump design, where the suction clearance is
adjusted with the help of three set screws. The joint is keyed for transmission of torque. Solid
impeller mounting is a key component in pump
For dry-installed pumps GRUNDFOS has devel- operational reliability, and great care should
oped a patented design (SmartTrim) allowing the always be exercised when the impeller is disman-
suction clearance to be adjusted and restored tled. It is good practice to always use a torque
without the need of removing the pump or open- wrench when setting the impeller screw. The
ing pipe connections. Adjustment does not affect pump manufacturer issues correct tightening
pipe connections or require re-alignment of these. torque information and possible recommenda-
Figure 36 show the principle. The adjustment is tion of screw lubricant in each case.
done by first closing the clearance and then back-
ing up the adjustment screws 1 mm, after which
the suction cover is tightened against the set
screws with the fastening bolts.

The adjustment margin on the GRUNDFOS pumps

is 10...15 mm, depending on pump size. It is dimen-
sioned to last the lifetime of the impeller.

26
 Pump Construction 2

2.3 Motors can withstand any internal explosive blaze and

prevent it from spreading into the explosive
surroundings. This is referred to as Class D.
2.3.1 General • The motor is designed so that no sparking or
Submersible pump motors are squirrel-cage elec- high temperatures may occur inside the motor.
tric motors wound for regular three-phase or sin- This is referred to as Class E.
gle-phase alternating current supply. Single-
phase pump motors are available only for small An explosion-proof motor is designed and built
pumps (2 kW or less). Motors are available for 50 according to the rules set forth by international
or 60 Hz supply and a number of voltages. The governing bodies (for example, Euronorm 50014
motors are built for submersible operation, Class and 50018). The requirements for class D motors
IP 68 according to IEC. The electrical features of are detailed, involving among others the selection
submersible motors are described in detail later in and gauge of construction materials, casing joint
this book. design and manufacturing tolerances, motor inte-
rior volume utilization as well as strength of the
The submersible pump is a fixed combination of a structure and fasteners. The essential require-
motor and a pump with a common shaft and ment for the joints is that the mating surfaces
bearings. The motor is short-coupled to the pump, have to be longer, as they are supposed to serve as
and some of the pump parts, such as the volute “extinguishing” gaps. Certification and approval
cover may be integral with the motor attachment of a design is always subject to extensive tests,
flange. For best results the pump and motor are where the actual ability to withstand internal
designed together, with one motor frame fitting a explosions, is determined.
range of pump parts for different duties and dif-
ferent operational ranges by the same manufac- Class E explosion-proof motors do not require
turer. Motor and pump sections are selected and extensive structural modifications, but are tested
designed so as to exclude overloading at any duty for internal temperature rise at certain loads. Also
point on the pump curve. internal sparking must be prevented by adequate
gaps between rotating and stationary parts.
Submersible motors are normally air-filled. Small
motors (1,5 kW and less) are also made oil-filled. Usually explosion-proof motors are based on the
The oil used in these is low-viscosity oil used also regular designs of a manufacturer, and form a
in transformers, in order to keep the rotor friction complement to these. The power characteristics
losses a small as possible. The growing losses and are normally not altered, and the pump parts are
lower efficiency prevent manufacturers from common for both. The structural requirements on
making larger oil-filled motors. Oil-filled motors explosion-proof motors make these more expen-
are cheaper than air-filled motors because of sive than regular motors.
smaller number of parts.
2.3.3 Motor Cooling
2.3.2 Explosion-proof Motors Mechanical and electric losses in the motor are
Submersible pumps are available in explosion- converted to heat, which must be dissipated. In a
proof versions for use in environments where the regular submersible pump motor (see Figure 14)
pumped liquid or ambient atmosphere may con- the heat is transferred from the stator casing to
tain explosive gases. This condition may exist, for the liquid through submersion. For cooling pur-
example, at or near petrochemical works but a poses it is normally sufficient if the motor is sub-
space can also be defined as explosive elsewhere, merged to about half of the motor depth. The
if deemed necessary for safety reasons. liquid level may be pumped down all the way for
short periods without risk of overheating the
The principle of explosion-proof motors is their motor.
safety against causing a potentially explosive
atmosphere from igniting. The following two A motor operating in water this way is in fact very
alternative technical solutions are available to fill efficiently cooled, since cooling continues after
the requirement: the motor has stopped. Thus it is possible to allow
• The motor is designed so that the enclosure frequent starts and stops of submersible motors,

27
 2 Pump Construction

which is beneficial for the design of pumping

Fig. 37
installations.

Allowed Water Temperature

Cooling of submersible motors relies on the
pumped liquid, either by submersion or other-
wise. Water temperature is therefore essential.
Usually the motors are rated for +40 °C liquids.
Higher liquid temperatures may be allowed, but
then the pump selection should be referred to the
pump manufacturer. Also the cavitation risk must
be assessed for higher temperatures with an
NPSH analysis, because of higher vapour pressure
of the liquid.

Submersible Motor Cooling in Dry Instal-

lations
Many submersible motors are installed dry for
various reasons. Adequate cooling of these
motors must be ensured, and it can be arranged in
a number of ways:

With a cooling water jacket that encases the

motor or parts thereof. Part of the pumped liquid A dry-installed GRUNDFOS submersible pump with
is diverted through channels from the pump cas- motor cooling jacket. Part of the pumped liquid is
ing into the cooling jacket where it recirculates filtered through a gap of about 0,5 mm and remains
after the casing has filled up. The water enters the recirculating in the cooling jacket, circulated by the
space behind the impeller through a filtering pumping action of the impeller back shroud auxilia-
clearance (about 0,5 mm) and is circulated by the ry vanes. Efficient cooling is provided by heat con-
auxiliary vanes on the back of the shroud around vection from the stator through dissipation into the
the motor stator housing inside the jacket. Excess pumped liquid.
heat is conveyed to the water through forced con-
vection, ensuring efficient cooling. The principle is
shown in Figure 37. The usage of a filtering clear- With thick stator housing walls. This design, suit-
ance and wide enough cooling channels has able for small submersible pumps, employs a
ensured that the system is non-clogging also in thickened stator housing that conveys the heat
practice. A cooling jacket is often optional on from the stator to the pumped liquid. In this con-
small and medium-sized pumps for dry installa- struction the stator housing flange may contact
tions, whereas very large pumps are often the pumped liquid directly or through the oil
equipped with a cooling jacket as standard housing cover flange. The flange can be shaped
regardless of installation method. with a recess or channel for good contact with the
liquid. The stator casing may also be made of alu-
In some cases, where the liquid being pumped is minium in dry-installed pumps to further
unsuitable for circulation in the water jacket, enhance heat dissipation. Figure 38 shows the
external cooling water may be used. In these construction.
cases the pump is modified with external water
connections in the jacket and by plugging the For dry-installed pumps only a cooling water
entrance channels from the pump casing. A safety jacket offers equal or even superior motor cooling
circuit is necessary to protect the pump from to submergence. Other motors may have to be
overheating due to disruption of the external down rated for dry installations, limiting the
cooling water supply. selection of pump components from the match-
ing range.

28
 Pump Construction 2

renewed each time a joint is opened for service to

Fig. 38
ensure tightness.

The electrical cable inlet to the motor must be

reliably watertight. A good design uses compress-
ible rubber grommets that match both cable and
the inlet recess. The grommet is compressed to
prescribed tightness by the shape of the matching
parts when assembled. A cable clamp external to
the sealing carries all outside tensile loads on the
cable, preventing pulling at the seal.

The possibility of water intrusion through the

cable is a reality. If the cable free end is allowed to
be submerged, water may travel by capillary
action between the copper strands of the leads to
the motor. This action is enhanced by the temper-
ature changes of the motor, and water may this
way enter an otherwise undamaged motor. The
condition can arise in new pumps that have been
stored outdoors prior to installation with the
cable free end unprotected.

Most pump manufacturers deliver their pumps

with protecting sleeves on the cable free ends.
A GRUNDFOS submersible pump suitable for dry or Warning labels are attached to warn the storage
submerged installation. The thick-walled lower sec- and installation personnel of the risk of submerg-
tion of the motor serves as a heat conduit to the ing the cable free end.
pumped liquid. The stator casing may be made of
aluminium to further enhance the effect. Securing submersible motor tightness requires
special knowledge and special tooling, and it is
therefore advisable to return the pump to an
authorized shop for repairs. Pump manufactures
With an internal cooling circuit, where a cooling offer training and special tools to their customers.
fluid, e.g. glycol, is circulated by a separate small For owners of large numbers of submersible
impeller on the pump motor shaft. The pump pumps an in-house authorized shop may be war-
incorporates a heat exchanger between pump ranted.
housing and the motor, where the cooling fluid
yields heat into the pumped liquid. System com-
plexity may pose problems.
Shaft Seals
The shaft seal, providing safety against leakage of
the pumped liquid into the motor, is one of the
2.3.4 Motor Tightness most important elements in a submersible pump.
Water intrusion in the motor leads invariably to
damage, or, if detected by safety devices, at least Modern submersible pumps almost exclusively
to pump outage. The chief requirement and use a shaft sealing arrangement with double
design consideration of submersible pump mechanical seals separated by an oil-filled cham-
motors is therefore complete integrity against ber. The arrangement, developed and refined over
leakage. Motor tightness is ensured with good the years, provides adequate protection against
design and continuous quality control including leakage and motor damage in most cases.
tests during manufacturing.
Figure 39 shows a mechanical seal arrangement
All submersible motor joints are machined to fit, used in submersible pumps. There is a lower or
and O-rings are used throughout. The O-rings are primary seal and an upper or secondary seal. The

29
 2 Pump Construction

pumped liquid. The material used today is silicon

Fig. 39
carbide (SiC), which has a hardness around 2000
on the Vickers scale and ranks next to the dia-
mond. The silicon carbide rings can be either solid
or converted. Converted carbide rings are sintered
to SiC to a depth of approximately 1 mm, leaving
the ring interior unchanged. SiC also has very
good resistance against corrosion, and can be
used in all wastewater and dewatering applica-
tions.

If the secondary seal is oil lubricated, a combina-

tion of materials may be used. A stationary ring of
a softer material with good friction properties in
combination with a hard rotating ring provides for
low seal rotation resistance. The oil lubrication
protects the seal against wear. Modern secondary
seals normally have faces of silicon carbide and
A GRUNDFOS double mechanical seal with primary carbon .
and secondary seals.
Modern submersible pumps utilize mechanical
seals custom-designed for the purpose. Good
designs have been developed by most major man-
seals, being separated by an oil bath, operate
ufacturers. A proprietary design combining pri-
under different conditions. This is reflected in
mary and secondary seal is shown in Figure 40.
their construction with different materials. Both
seals comprise two contacting slip rings, one sta-
tionary and one rotating with the shaft. The rings Fig. 40
are pressed against each other by spring force
and, for the primary seal, in addition by the pump
pressure.

Sealing between the slip rings is based on the

extremely smooth and flat contact surfaces of the
slip rings. The surfaces are in such close contact
that no or only a very minute leakage can pass
between them. The flatness and smoothness of
the rings are in the magnitude of 0,0005 mm and
the faces are finished by lapping. The slip rings
seal against the stationary seat or shaft with O-
rings. The material of the O-rings is selected to
withstand high temperature and the corrosive
and dissolving action of the seal oil and the impu-
rities in the pumped liquid.

Notches in the stationary slip rings of the primary

seal secure them in the seat against turning. The
rotating rings are locked similarly with drive pins.
Spring clips or washers keep the stationary rings
in their seats during abnormal pressure situa-
tions.

The material of the primary seal faces is normally

A GRUNDFOS integrated double mechanical seal.
hard, because of the abrasive action of the

30
 Pump Construction 2

All mechanical seals used in submersible pumps 2.3.5 Motor Bearings

must allow rotation in either direction, since
pumps frequently get started in the wrong direc-
Bearing loads
tion or may be turned backwards by back-flowing
water in installations without check valves. The submersible pump bearings carry the com-
bined load of the pump and motor as exerted on
All submersible pumps with double mechanical the common shaft. The following forces act on the
seals have an oil space between the seals. The oil bearings, either radially or axially:
serves the following functions vital to the func- • hydrodynamic radial force
tion of the seals and the pump: • hydrodynamic axial force
• Seal lubrication, especially of the secondary • magnetic radial force
seal • the weight of the rotating parts
• Seal cooling
• Emulsification of possible leakage water, thus The significant forces acting on the bearings are
rendering it less harmful the hydrodynamic forces.
• Seal condition monitoring. By checking the
seal oil during maintenance, the seal condition The hydrodynamic radial force is the resultant of
and rate of leakage can be estimated. the pressure distribution at the impeller perime-
ter in various relative positions to the pump cas-
Overfilling of the seal oil chamber should be ing. The radial force is dependent on a number of
avoided in order for the oil to be able to absorb design factors as well as on the pump operating
leakage water by emulsification and to prevent point.
possible overpressure due to thermal expansion
of the oil. The pump manufacturer provides infor- The hydrodynamic axial force is the resultant of
mation on oil quantity and filling and monitoring the forces induced by the impeller diverting the
methods. flow from axial suction to radial discharge, and
from the pressure difference between suction and
In special applications, where the pumped liquid pressure side of the impeller. The axial force is
contains very fine materials, the primary seal may also strongly related to the pump flow and oper-
open due to material build-up on the slip ring ating point.
faces. In these cases it may be warranted to
arrange for continuous external flushing of the Bearings
seal. These installations are always considered Rolling bearings are used throughout in submers-
separately for each case by manufacturer and cus- ible pump motors. Ball bearings are used for their
tomer. ability to carry both axial and radial loads. In very
large motors a combination of ball bearings and
Mechanical seal life expectance cannot be deter- roller bearings are used because of the large
mined theoretically or even by lab tests. Perfor- forces on the components.
mance over time is also difficult to predict. Seal
life varies greatly from case to case, with service To allow for heat expansion of the shaft and for
records from a few years to over 15 years reported. manufacturing tolerances, the shaft upper bear-
ing is allowed axial movement, whereas the lower
bearing is locked axially.

Bearing selection is governed by international

standards with regard to bearing life. According to
the pump standard ISO 5199, “Bearing rating life
(B10)” should be at least 17500 hours.

Submersible pump bearings are normally lubri-

cated for life at the pump factory, using special
grease suitable for the high operating tempera-
tures allowed in submersible motors.

31
 2 Pump Construction

2.3.6 Motor Protection Devices ing the circuit, making re-start of the pump
Submersible motors are equipped with various possible. Thermal switches in the windings
protection devices for prevention of damages for protect the motor against overheating from
the following reasons: insufficient cooling, and are especially impor-
• overheating tant in pumps depending on submergence for
• water intrusion cooling.
• seal failure • Water intrusion into the sealed motor can be
• bearing failure monitored with a moisture switch that reacts
• winding insulation deterioration to excess moisture. Normally the moisture
switch is connected in series with the thermal
Some protection devices are standard issue switches in a circuit that disconnects the cir-
whereas others may be available as extra equip- cuit breaker coil and stops the motor upon
ment on request only. Large pumps need better opening. Figure 42 shows a moisture switch
protection devices because of the greater eco- that operates when the humidity reaches
nomic values of these pumps. 100%. The moisture switch is non-reversing
and does not reset after tripping. In a common
The protection devices can be divided into inter- circuit with moisture and thermal switches, it
nal devices with sensors inside the motor and can be determined which device has opened,
external devices in the pump motor control panel. since only the thermal switches close again
after cooling. The motor must be opened and
dried out before any attempts to restart it after
Internal Protection Devices the moisture switch has tripped.
The following protective devices are mounted • Water intrusion into the sealed motor past the
inside the motor: shaft seals can be monitored with a leakage
• Thermal switches in stator windings. These are detector sensor in the seal oil chamber. Regu-
normally bimetal miniature switches that lar motor oils used as seal oil in submersible
open at a fixed, preset temperature, please pumps can form an emulsion with up to 30%
refer to Figure 41. Three switches, one in each water content. The leakage detector either
phase, are used in three-phase motors. The reacts on a water content exceeding 30% (con-
switches are connected in series in the control
circuit, which is wired to stop the motor when
opening. The switches reset upon cooling, clos- Fig. 42

Fig. 41

A GRUNDFOS moisture switch. The unit consists of a

number of moisture-sensitive disks stacked onto an
actuating rod, and a micro switch. The hygroscopic
Thermal switch. The unit consists of a miniature bi-
disks expand upon contact with excess moisture,
metal switch that opens according to the switch
pulling the actuating rod. A cam at the rod end trips
temperature rating. The switch is connected in the
the micro switch and breaks the circuit. The unit is
control panel to break the current in case the motor
non-reversing and must be replaced after use.
is overheating.

32
 Pump Construction 2

ductive detectors) or monitors the water con- develops operational difficulties or gets
tent continuously (capacitive detectors). The jammed, when the pump becomes clogged or
latter may be calibrated to trip at any water during loss of phase in the mains supply. Over-
content and used to indirectly observe primary load protection is frequently provided by over-
seal condition by monitoring water intrusion load relays coupled to the motor contactors.
over time (leakage rate). Leakage detectors are These consist of ambient temperature-com-
usually not standard, but available as extra pensated bimetal elements, that trip the cur-
equipment. rent to the contactor coils in case the current
• Water intrusion into the sealed motor through exceeds the set specified value. Overload
capillary flow through the supply cable before relays provide good protection against loss of
pump installation can be prevented by fitting a phase in the supply. The overload relay should
tight protective sleeve over the cable free end be set according to the motor nominal current.
at the factory. The sleeve is not removed until When star delta start is used, the current
the cable is connected at the control panel. through the overload relay is reduced by the
• The condition of the bearings and/or bearing factor 0,58 (1/√3), which must be taken into
grease can be monitored with thermal sensors account when setting the relay. Figure 43
in the bearing bracket. These are installed shows an overload relay.
close to the bearing outer race, and calibrated • The stator winding insulation is monitored by
to register bearing temperature. Thermal sen- an automatic resistance measuring device that
sors are extra equipment. measures the resistance between the phases
and between phases and earth each time the
External Protection Devices pump stops. Alarm levels for resistance can be
set, preventing damages short circuits and
The following protective devices are mounted in
damages to windings.
the motor control panel:
• Short-circuit protection is accomplished by
means of fuses, circuit breakers or electronic
motor protectors. Fuses and circuit breakers
should be dimensioned to withstand the
motor starting current, but the rating must not
exceed that of the supply cable or switchgear.
Where fuses are used, these should be of the
slow type.
• Overload protection is required in a sudden
overload situation, such as when the impeller

Fig. 43

Thermal overload relay. The relay connects to the

motor contactor and breaks the current in case of the
electric load exceeding the set value.

33
 2 Pump Construction

2.4. Pump Connection sturdy and tight connection. The pump is kept in
place by its own weight. Figure 44 shows a sub-
A submersible pump, when installed submerged, mersible pump baseplate and guide rails.
is connected only to the discharge pipe. For fixed
installations a self-connecting baseplate arrange- Figure 45 shows a flexible seal designed in a way
ment is normally used. that the seal action is further enhanced by pump
pressure, ensuring a tight connection at all times.
Submersible Baseplate
Some pump manufacturers offer conversion kits
The concept of the submersible baseplate has for the connection of pumps to older baseplates
been developed over the years for use with sub- or as replacement pumps to some other manufac-
mersible pumps. The arrangement allows for the turer's baseplate. Thus the upgrading or conver-
pump to be lowered into the pump well and sion of existing pumping stations may be done
firmly connect to the discharge pipework without with a minimum of work and costs.
the need of the operating personnel to enter the
well. Likewise the pump can safely be hoisted
from the well for service. The system includes rails
or pipes that guide the pump down onto the Fig. 45
baseplate. A special flange, or guide shoe, on the
pump discharge mates with the joining surfaces
of the baseplate for a firm connection. Well-
designed systems are made to precision and have
machined surfaces and rubber seal rings for a

Fig. 44

Flexible seal between pump pressure flange and

connector. The seal is designed in a way that the
seal action is further enhanced by pump pressure,
ensuring a tight connection at all times.

A GRUNDFOS submersible baseplate. When seated,

the weight of the pump keeps it firmly in place. Pre-
cision-machined connecting surfaces and a rubber
disk seal ensures tightness. Clearance between the
guide shoe and the rails ensures unobstructed hoist-
ing of the pump even in fouled conditions

34
 Pump Construction 2

Fig. 46 Fig. 47

Seat ring

Submersible pump on stand with hose connection.

This installation version is used for temporary or
shifting installations.

Pump column installation. Pump rests on conical

seat ring installed at the bottom of the column
Hose Connection
Figure 46 shows a submersible pump installation
with hose connection. It may be used for tempo-
rary installations or in applications where the Column installation is ideal for submersible pro-
pump is shifted around the wet well for sludge peller pumps, but also for sewage pumps
pumping. intended for large flows and low to moderate
heads. Figure 48 shows the Q/H area on which
column-installed Grundfos pumps are available.
Column Installation For this range column installation is likely to lead
to lower investment costs, but each project
The concept of column installation of submersible
should be assessed individually. Column-installed
pumps has been developed during the past few
pumps have the same efficiency as pumps
years. The pump is lowered into a vertical pipe or
intended for other installation modes, but the
column, where the circular pump casing fits onto
pump curves will differ slightly because of the
a seat ring installed at the lower end of the col-
open pump casing. Column installation is very
umn, please refer to Figure 47. The pump stays in
suitable for return sludge pumping in sewage
place by its own weight and from the pressure
treatment plants. The column pipe may be made
force from the pumping action. The pump casing
of stainless steel or hot dip galvanized steel.
is special-designed for the installation, and is fit-
ted with trailing vanes. The seat ring is conical,
For seawater installations a column made of
ensuring a tight connection between pump and
stainless steel may create a strong galvanic ele-
column. The tight connection and dowels prevent
ment between pump and column, leading to
the pump from spinning loose at start-up
pump corrosion. Especially galvanized pump parts
will rapidly corrode from the galvanic action of
the large cathode area of the column surrounding

35
 2 Pump Construction

Fig. 48

GRUNDFOS column-installed pump flow and head (Q/H) range.

the pump. A lifting chain left in place, for instance, sion. In these applications the natural corrosion
will have to be made of stainless steel. The cast layer, providing the underlying material with nat-
iron pump should be protected by sacrificing ural protection becomes scrubbed away, leading
anodes that are replaced at regular intervals. to rapid corrosion. The use of stainless materials
Painting the column with a paint layer of at least for these vulnerable parts may be warranted.
200 µm thickness prevents the cathode surface
from forming and thus pump corrosion. Corrosion in seawater is dependent on a number
of factors, such as salinity, oxygen content, pollu-
tion and temperature, and the right material
selection must be considered for each case. Sacri-
2.5 Construction Materials, ficial zinc anodes may offer protection against
Corrosion and Wear corrosion in certain cases.

2.5.1 Corrosion Resistance The supply cable sheath material must be able to
withstand oils and other pollutants present in
Cast iron is the main construction material in sub-
sewage. Other rubber parts, such as O-rings are
mersible sewage pumps, with fasteners and hard-
usually made of Nitrile or Neoprene for resistance
ware made of stainless steel. The pump shaft is
against sewage, oil and chemicals.
either made entirely of stainless steel or pro-
tected against contact with the pumped media.
Submersible pumps are also available made
Where the pump or baseplate includes fabricated
entirely of stainless steel for use in highly corro-
steel parts, these are hot-dip galvanized. These
sive liquids, such as process industry effluents.
materials will last for decades in regular sewerage
Stainless steel submersible pumps are 3…4 times
duty.
as expensive as pumps made of regular materials.
In difficult applications the pump manufacturer
In cases, where the pumped liquid contains indus-
may not be able to guarantee the corrosion prop-
trial effluents, the corrosion resistance of cast iron
erties for a specific case, but will cooperate with
may not be sufficient, especially for parts subject
the client to find the right solution for the case.
to fast flow velocities, such as impellers and pump
casings, which will be subjected to erosion corro-

36
 Pump Construction 2

2.5.2 Wear Resistance Fig. 49

The sand content in sewage is on average
between 0,002 and 0,003 % (in volume). The con- Sand content
tent may periodically, e.g. during heavy rainfall Pm [%]
and snow melting be much higher in areas with 20 Pump head H0 [m]
combined sewage and stormwater drain systems. 10
Regular cast iron will last in most applications for 5
10 5
years, but special material may have to be consid- 50 20
1
ered for highly abrasive effluents, such as sewage
0,5
treatment plant sand trap pumping.
0,2
0,1
2.5.3 Abrasive Liquids 1 10 100 1000 10000
Pump service life [h]
Pump performance in an abrasive liquid is
strongly dependent on the content of abrasives in Pump wear rate as a function of sand content and
the liquid. The standard abrasive is common pump head. H0 is pump head at Q=0. Wear rate is
quartz or silicon sand, to which the following can expressed as expected service life of a cast iron im-
be applied directly. peller and is strongly dependent of sand content and
pump head. The graph is based on experiments, and
The sand content is either expressed as volume or can be used generally.
weight content, which are related as follows:

pm ≈ 3·pv (15) Figure 49 is a diagram showing the relations

between the pump wear rate and the sand con-
where pm is weight content and pv is volume con- tent and pump head. High sand contents in the
liquid will have a dramatic effect on pump service
tent in %. Thus pv = 5% equals pm =15%.
life. The effect of the sand content is exacerbated
by high pump head.
With increased sand content the density of the
liquid/sand mixture increases. Since required Pump wear can be minimized using suitable
pump power is directly related to the density of wear-resistant materials and through appropriate
the pumped liquid, required power will have to design. For best results, materials with a hardness
be checked separately in each case to ensure over 500 HB should be used. The difficult
pump performance, whenever liquids with high machineability of hard materials, such as Nihard
sand content are being pumped. For sand trap and some alloyed irons, may require special
pumps in sewage treatment plants, a power impeller and pump casing designs where machin-
reserve of 30% has proven adequate. ing is minimized.

The density of a mixture water and sand can be The use of submersible pumps in abrasive envi-
written ronments must be considered separately on a
case by case basis and using sound engineering
ρ = 1 + 0,007pm (16) judgment.

where pm is expressed in %.
Thus, if pm = 15%, ρ = 1,1 kg/l.

The following factors affect the wear of a pump:

• sand content
• sand quality
• pump material
• pump head
• type of impeller

37
 3 Pump Performance

3 Pump Performance Fig. 50

Pump performance is the result of the interaction

between pump and rising main or pressure pipe-
line. An introduction to pump selection and the
calculation of rising main resistance characteris-
tics are presented.

3.1 Pump Head

3.1.1 Submersible Pumps
In the following the concept of head is applied to
submersible pumps. For practical reasons the
pressure in the pump well, or lower well, is
assumed to be equal to the pressure prevailing in
the receiving, or upper container. Should these
containers be under different pressure, the pres-
sure difference would have to be taken into
account. The difference in atmospheric pressure
can also be disregarded in all practical installa-
tions, since the difference in atmospheric pres-
sure between a receiving container situated, for
instance, 100 m above the pump well is only 0,001
bar or 0,01 m of water.
Head components in submersible pump installations.
Figure 50 shows how the head is defined in a
submersible pump installation. The following
units are used:
H = pump total head (m) The sum of the static head and the dynamic head
Hst= pump static head (m) is the pump total head, thus
Hd = pump dynamic head (m)
H = Hst + Hd (18)
Hgeod= geodetic head (m)
HJ = pipeline losses (m) According to international agreement (Standard
pL = atmospheric pressure in pump well ISO 2548), the total head H according to equation
pU = atmospheric pressures in upper container 18 is used when plotting characteristic curves for
v2 = flow velocity at outlet (m/s) submersible pumps.

g = acceleration of gravity (9,81 m/s2) The total head H is thus available to pump the liq-
uid through the rising main. The pressure or head
If an observation pipe is installed at the pump required to pump a given flow through a pipeline
outlet flange, the pumped liquid will rise in it to a is made up by the geodetic head and the flow
height Hst from the well level. This height repre- losses. Thus can be written:
sents the pump static head. In addition, the liquid
has a velocity v2 at the pump discharge, which H = Hgeod + HJ (19)
can be converted to pressure or dynamic head Hd
with the following equation: The geodetic head Hgeod is the actual physical
difference in height between the liquid levels in
v
2 the pump well and the receiving container. Pipe-
H d = -----2- (17) line flow losses consist of pipe friction losses, local
2g
losses from various fittings in the pipeline

38
 Pump Performance 3

(elbows, valves, etc.) and the outlet loss at the 3.2 Pump Performance Curves
point of discharge.
Centrifugal pump characteristics are normally
Losses due to liquid flow in the well to the pump presented as a set of curves, where the data has
are considered as pump losses in submersible been established through the testing of the
pump installations. If a suction pipe is installed pumps or assessed by the manufacturer for e.g. a
before the pump, it will have to be taken into special impeller diameter. For submersible pumps
account when calculating pipeline losses. the following important information is normally
plotted as curves against the flow rate Q:
3.1.2 Dry-installed Pumps • H head curve
When calculating heads of dry-installed pumps, • η efficiency curve(s)
the situation before the pump will also have to be • P power curves
considered. Figure 51 illustrates the situation.
Figure 52 shows a typical pump performance
In this case it is assumed that the suction well and curve sheet with information important for the
the receiving container are open to the atmo- user.
sphere and that the pressure at the liquid surfaces
is constant. Thus the pump head is the sum of the Fig. 52
geodetic head and the flow losses in the suction
and pressure pipelines. Thus

H = Hgeod + HJt + HJp (20)

where HJt represents flow losses in the suction

pipeline and HJp flow losses in the pressure pipe-
line.

Fig. 51

Typical pump performance curve sheet for submers-

ible pump. The dashed sections of the curves indicate
areas, where pump prolonged use is prohibited. The
reasons for the limitations may be cavitation, vibra-
tions or motor overload.

3.2.1 H Curve
The head or H curve gives the pump total head as
a function of the flow Q. The curve may contain
additional information on pump usage, such as
limits due to cavitation, vibration or motor over-
Pipeline loss components for dry-installed pumps. load.

39
 3 Pump Performance

3.2.2 Efficiency Curves 3.2.3 Power Curves

Pump efficiency η is also a function of the flow The pump required power is also a function of the
rate Q. The efficiency can be indicated as a ratio or flow rate Q. Figure 52 contains both the pump
percentage. For submersible pumps both the power curve and the motor power curve. The
pump efficiency η and the overall efficiency ηgr are motor power is the electric power drawn by the
defined, where ηgr includes motor losses. It is motor and measured at the cable junction box at
the motor. According to international standards
important to distinguish between these defini-
on pump testing the pump power is designated P
tions for efficiency, especially when comparing
and the power absorbed by the motor Pgr. The
pump performance. The losses leading to the
pump efficiency are discussed in Section 1 of this required power can also be calculated using the
book. Thus it can be written: equation

ηgr = ηmot · η (21) ρQgH

P = ---------------- (22)
η
where ηmot is motor efficiency. where
P = power (W)
The efficiency can also be marked on the head ρ = liquid density (kg/m³)
curve, with numbers indicating different effi- Q = volume flow (m³/s)
ciency values. If several head curves for various g = acceleration of gravity (9,81 m/s² )
impeller diameters are plotted in the same graph, H = pump head (m)
these markings can be connected to form η = efficiency
isograms, or operating areas with the same effi-
ciency. The pump performance diagram will then 3.2.4 NPSH Curve
assume its typical appearance as shown in Figure
Since NPSH calculations are performed only for
53.
dry-installed pumps, the NPSH curve is not usually
included on submersible pump data sheets. It will
Fig. 53 be provided by the manufacturer on request if
cavitation is feared in a dry installation, or if
otherwise required by the client.

Results from tests performed with clean water are

applicable as such on normal municipal sewage
and most industrial effluents, since the low solids
content in sewage (less than 0,05%) does not sig-
nificantly affect pump performance.

Pump performance sheet with curves for different

impeller diameters, where the efficiency is indicated
as points directly on the head curves. Impeller selec-
tion is facilitated by connecting the points with the
same efficiency.

40
 Pump Performance 3

3.3 Pipe Losses and Rising Main where

HJp= pipeline loss (m)
Characteristic Curves λ = friction factor
In the following the theory for calculation of flow l = pipeline length (m)
losses in pipelines is presented. Practical calcula- v = flow velocity (m/s)
tions can be made with the help of the detailed g = acceleration of gravity (9,81 m/s²)
instructions with calculation diagrams and nomo- D = pipeline internal diameter (m)
grams presented in Appendix A, or with a com-
puter program. Obtaining the friction factor λ from the diagram
in Figure 54, equation 24 can be solved. Surface
Flow velocities used in sewage pumping are high roughness values (mm) presented in the following
enough to ensure uniform turbulent flow in the table can be used:
piping. Flow losses therefore increase with the
square of the flow velocity. The flow loss of a ris- Pipe material k new k old
ing main is the sum of the friction loss of the pipe-
line constituent parts and the local losses from Plastic 0,01 0,25
the various components and fittings.
Drawn steel 0,05 1,0

3.3.1 Friction Losses Welded steel 0,10 1,0

Pipe friction losses depend on the following fac- Drawn stainless steel 0,05 0,25
tors :
• pipe length Welded stainless steel 0,1 0,25
• pipe internal diameter
• flow velocity Cast iron 0,25 1,0
• pipe wall relative roughness Bituminized cast iron 0,12
• fluid kinematic viscosity.
Asbestos cement 0,025 0,25
A dimensionless relation, Reynold's number is
introduced: Concrete 0,3...2,0

vD (23) The surface of an old pipe material becomes

Re = -------
ν rougher from erosion. Corrosion and sediment
layers forming on the pipe surface may decrease
where the pipe diameter, also leading to higher flow
Re = Reynold's number losses.
v = flow velocity (m/s)
D = pipe internal diameter (m) The effect of pipe diameter change can be calcu-
ν = kinematic viscosity (m²/s) lated with the following relation:

The kinematic viscosity for water is dependent on 5

H' Jp ≈ H Jp ⎛ -----⎞
D'
temperature: ⎝ D⎠ (25)

Thus an increase of pipe diameter from, for

t °C 0 20 40 60 100
instance, 100 mm to 108 mm decreases the flow
ν 10-6 m²/s 1,78 1,00 0,66 0,48 0,30 loss by 30%.

Equation 25 is sufficiently accurate for practical

The equation for pipeline losses can be written: purposes when comparing flow losses in rising
mains of different diameter, particularly since
2
lv accurate surface roughness values are seldom
H Jp = λ ----------- (24)
D2g available.

41
 3 Pump Performance

Fig. 54

TRANSITION ZONE

SMOOTH PIPE

TURBULENT FLOW

RELATIVE SURFACE ROUGHNESS K/d

REYNOLD'S NUMBER Re=

Moody diagram for establishing the friction factor λ. The value of λ is obtained using Reynold's number and the relative
roughness number k/D as parameters, where D is pipe internal diameter in mm and k equivalent surface roughness in
mm. Completely turbulent flow can be assumed in wastewater applications.

Rising main flow losses are frequently calculated equal to the resultant losses of the two true rising
with the help of proprietary computer programs, mains.
also available from some pump manufacturers.
These programs may also suggest some pump The equivalent diameter is calculated with the
selections from the manufacturer’s range to best following equations:
suit the purpose. It is advisable to take a cautious
view on the pump selection suggested by a pro- A. Both rising mains have the same diameter
gram only, and always contact the pump manu-
facturer in dubious cases. D e = 1, 3 ⋅ D (26)

The rising main is sometimes divided into two where D = diameter of the two parallel rising
separate parallel pipelines. They have the same mains
length but may have different diameters or be
made of different materials. The distribution of B. The rising mains have different diameters
flow between the two lines and the ensuing
losses in these lines can be difficult to determine. 2, 65 2, 65 0, 3774 (27)
D = ( D1 + D2 )
Grundfos has developed a method for this, where
the two lines are substituted with a single virtual
rising main. An equivalent diameter is deter- where D1 and D2 are the different diameters of the
mined for this so that the resulting flow losses are parallel rising mains.

42
 Pump Performance 3

The volume rates of flow for the two rising mains where
are calculated wit then following equations: HJn= local loss (m)
v1 = flow velocity 1 (m/s)
A. Both rising mains have the same diameter
v2 = flow velocity 2 (m/s)
Q
g = acceleration of gravity (9,81 m/s²)
Q 1 = ----
2 (28)
If the pipe expansion is designed with a conical
B. The rising mains have different diameters section with an expansion angle of 10°, the loss is
reduced to 40% of the value calculated with equa-
D 1 2, 65
tion 32. This fact is important when expanding
Q 1 = ⎛ ------⎞ ⋅Q the pipe section right after the pump pressure
⎝ D e⎠ (29)
flange, where the flow velocity can be quite high.
Q2 = Q – Q1 By designing the transition with a 10° gradual
(30)
expansion joint, energy can be saved. In a con-
tracting pipe section the losses are much smaller,
The equations above are valid for turbulent flow,
and the conical section can be built much shorter.
which is normal for water pumping. The equa-
tions require that both pipelines have the same
Losses in a section with velocity reduction are
surface roughness.
generally much greater than in section with
increasing velocity.
3.3.2 Local Losses
Changes in pipeline internal diameter and shape, The final component of pipeline loss is the outlet
bends, valves, joints, etc. as included in the rising loss at the end of the rising main. If no expansion
main cause additional losses that comprise both a is provided, the loss equals the velocity head or v²/
friction and turbulence component. The following 2g.
equation is used to calculate the losses:
Loss coefficients for different valves are provided
2 by the manufacturers. Guide values for the most
v
H Jn = ζ ------ (31) common valves used in sewage installations are
2g
presented in Appendix A.
where
HJn= local loss (m) 3.3.3 Rising Main Characteristic Curve
ζ = local resistance factor In sewage installations the pump sump and the
v = flow velocity (m/s) delivery well are open to the atmosphere, and the
g = acceleration of gravity (9,81 m/s²) rising main characteristic curve will contain the
geodetic head and the flow losses only. Figure 55
Local resistance factors for different pipeline ele- shows the general shape of the characteristic
ments and fittings are presented in Appendix A. resistance curve for a pipeline. Since the flow is
The friction loss of these are not included in the turbulent at the flow velocities in consideration, it
local resistance factor, but is calculated as part of can be assumed that the flow loss varies in pro-
the rising main friction loss by including their portion to the square of the flow rate. Thus, if the
length and internal diameter when calculating flow loss at one flow rate is calculated with the
pipeline length. method described above, the other points of the
curve are obtained sufficiently exactly with the
Pipe expansion discontinuity loss can be calcu- following equation:
lated using the Borda equation:
2
H' J = H J ⎛ -----⎞
Q'
2 (33)
( v1 – v2 ) ⎝ Q⎠
H Jn = ----------------------- (32)
2g

43
 3 Pump Performance

Fig. 55 3.4.1 Economy

The economy of an installation is made up by
both procurement costs and operational costs
during its lifetime. A number of installation and
operational costs are directly dependent on rising
main size, and will react to changes in pipeline
size as follows:

With decreased pipeline diameter

• Piping and pipework component procurement
prices will decrease.
• Pumping station procurement cost will
Characteristic resistance curve for a pipeline. Pipe
increase due to increased flow losses with con-
losses (HJ) are plotted against flow rate (Q) and added sequent requirement for larger pumps and
to the geodetic head, which is constant. control equipment. Costs for increased electri-
cal supply systems, such as substations may
increase significantly.
• Operating costs will increase due to higher
energy costs because of pipeline losses.
3.4 Rising Main Size
With different costs having opposite relations to
Rising main size is selected based on the follow- rising main size, an optimal pipeline size may be
ing factors: found. Figure 56 shows the relation. The selection
• economy of an optimal pipeline diameter may be based on
• required internal diameter for the application Figure 57, which shows the optimum flow velocity
• required smallest flow velocity for the applica- for different installations, and is based on several
tion. studies.

Where possible, a more detailed study can, and

Fig. 56
should, be conducted.

Fig. 57

0 20 40 60 80 100 120 140 160 180 200

Guideline values for economically optimal flow veloc-

The relation of key cost components for a pumping ities for submersible pump installations. The figure is
installation as related to rising main size. With costs based on a study of submersible pump installations
having opposite relations to pipeline size and flow using geodetic head, pipeline length, yearly operation
velocity, an optimum can be found. hours and energy cost as parameters.

44
 Pump Performance 3

Pumping station internal piping should be 3.4.3 Avoiding Settling of Solids and
selected so as to minimize component costs with- Sludge
out unduly increasing the flow losses in the sta-
If the flow velocity in a rising main is too low, sand
tion. Figure 58 shows the flow loss in the internal
or sludge may have time to settle, which increases
pipework in a pumping station with two pumps in
the risk of clogging. Settled sludge may harden
duty-standby operation as well as the economical
and form a crust on the pipeline wall that perma-
pipe dimensions, based on several studies.
nently decreases the diameter, leading to
increased flow losses. Larger sludge clots moving
Fig. 58 with the flow may block bends or other fittings in
the rising main.

For municipal sewage a minimum flow velocity of

0,7 m/s is recommended. Where only domestic
sewage is pumped, the minimum flow velocity
may be as low as 0,5 m/s, but if sand is found in
Total the sewage, this lower value is not endorsed. In
installations with varying flow, e.g. where fre-
quency converters are used, the flow velocity may
temporarily be lower.

Where settling is known to occur, flushing out of

the system with all pumps running simulta-
neously at intervals is recommended. The shape
of the pipeline is also important, and sedimenta-
tion is likely to occur in rising mains having a pro-
nounced low, such as pipelines laid underneath
Size recommendation waterways, e.g. a river. In these cases it is recom-
mended to select a higher flow velocity.
Flow loss in the internal pipework in a pumping sta-
tion for each of two submersible pumps in duty-
standby operation as a function of flow. Each indi- 3.4.4. Water Hammer
vidual pipe installation includes a baseplate with Oscillating pressure waves are generated in a liq-
bend, valves, an upper bend and a branch pipe. uid being pumped through a pipeline during
starting and stopping of the pumps. This phe-
nomenon is called water hammering, and, if
severe, may lead to pipeline and equipment dam-
3.4.2 Free Passage for Solids age. The severity of the phenomenon is depen-
For untreated municipal sewage the smallest dent on a number of variables, such as change of
allowable free passage of the rising main is gener- velocity during the reflection cycle, pipe material
ally 100 mm in order to allow passage of solids characteristics as well as liquid characteristics.
without clogging. In pumping stations with small
flows the internal pipework may have a free pas- When the liquid is accelerated or decelerated, a
sage of 80 mm, especially when the pump free transitory pressure wave oscillates back and forth
passage is also 80 mm. until dampened. The oscillating frequency can be
calculated with the following equation:

2L
µ = -----
a (34)

where
µ = reflection cycle duration, during which
the pressure wave oscillates back and forth
once (s)
L = pipeline length (m)
a = pressure wave velocity (m/s)

45
 3 Pump Performance

Pressure wave velocities in clean water in pipes of Since it is difficult to establish the change in flow
different materials can be obtained from the fol- velocity when the pump starts or stops, exact cal-
lowing table: culations of the pressure transient cannot be eas-
ily performed. Only if, for example, a valve is
Pipe material Velocity (m/s) closed within the reflection cycle, and the flow
velocity change ∆v is equal to the flow velocity v,
Steel 900...1300 can the pressure change be accurately calculated.
Because the pressure fluctuates symmetrically,
Cast iron 1000...1200 the pressure may fall below vapour pressure,
Reinforced concrete 1000...1200 causing cavitation with resulting high pressure
transients and noise. Potential locations for these
Plastic 300...500 are pump, valve and pipeline high point. The high
grade vacuum may also cause the pipeline to col-
Sewage and sludge often contain insoluble air or lapse.
gas, which has a significant effect on the pressure
wave velocity, as can be seen from the following In sewage pumping the water hammer pressures
table, where the pressure wave velocity is induced during pump stop are higher than those
expressed as a function of the quantity of insolu- induced at pump starting. In theoretical computa-
ble air in the liquid: tions the objective is to calculate the water retar-
dation immediately after pump stop and the
pressure transient induce at that instance. The
Head = 15 m most uncertain and significant factors to find out
Quantity of insoluble Velocity ratio of are the pump flow resistance and lowest pressure
air as volumetric ratio pressure wave generated in the pump, after the supply current
has been cut off. This information is not readily
0 1,0 available from the pump manufacturers.

10-6 1,0 Another uncertain factor is the air or gas content

in the water or pipeline. The solution here is to
10-5 0,96
analyze for different concentrations in order to
find out the effect of the gas content.
10-4 0,73

0,32 Figure 59 shows the outcome when observing the

10-3
water hammer phenomenon in a twin pump
0,11 installation. The following is noted:
10-2
• Measured reflection cycle duration is 45 sec-
onds. Theoretical calculations for a fluid with-
Dissolved air has no practical effect on pressure out gas or air indicated a duration of only
wave velocity. 12…20 seconds. The difference between the
two values shows that gas is present in the
The pressure transient resulting from a change in water.
the flow velocity within a reflection cycle can be • Immediately after pump stop the pressure in
calculated with the following equation: the pipeline falls to vacuum. Because the
pressure was measured at the discharge
a ⋅ ∆v flange, the pressure inside the pump must
∆h = ± --------------
g (35) have been even lower. It is likely that the pres-
sure inside the pump fell below cavitation
where pressure (-10 m).
∆h = pressure change (m)
a = pressure wave velocity (m/s) Significant for the water hammer phenomenon is
∆v = flow velocity change (m/s) during one that it cannot be heard, since the pressure surge is
reflection cycle fairly slow, but it can be observed with a pressure
g = acceleration of gravity (9,81 m/s²)

46
 Pump Performance 3

Fig. 59 rising main will prevent the pressure from

dropping inside the pump. Dimension of by-
Head [m] pass pipe should be selected one size smaller
than pump pressure flange size.
Pump 2 starts

Pressure surge
Pump 1 starts

Pumps 1 & 2 stop

• Using heavier pipe components that will with-

stand water hammer pressure. Vacuum tran-
Duty point

sients may be more critical to the pipeline and

equipment than pressure surges.

Time [s]

Water hammer sequence pressure measurements

as function of time.

gauge. Noise will be emitted only in case of cavi-

tation or if a valve closes rapidly.

Water hammer is not a common problem in sew-

age installations. Theoretical description of the
problem is difficult because of the large number
of unknown entities.

3.4.5 Avoiding Water Hammer

If water hammer occurs in sewage installations,
the situation can be alleviated with one or several
of the following measures:
• Preventing of simultaneous stopping of two or
more pumps.
• Installing automatic valves with closing times
of 20…30 seconds instead of regular check
valves. Pump stops after valve has closed.
• Stopping pumps slowly with frequency con-
trol.
• Using soft start equipment also for stopping of
the pumps. Complete control of stopping
sequence not always possible.
• Installing automatic air relief valves at points
where negative pressure occurs.
• In cases of cavitation in pump during the stop-
ping cycle, the installation of a by-pass suction
line with check valve from the wet well to the

47
 3 Pump Performance

3.5 Pump Duty Point Fig. 61

3.5.1 Single Pump Operation

By adding the geodetic head (Hgeod) and the pip-
ing loss (HJ), the rising main head is obtained. The
geodetic head is a constant independent of the
flow, whereas the losses increase with approxi-
mately the square of the flow rate Q (see Figure
55). If a pump head curve drawn to the same scale
is overlaid or plotted on the rising main character-
istic curve, the point of operation will be the inter-
section of the curves. At this point the pump head
equals the head required by the rising main. The
pump flow rate Q can then be read directly off the
diagram as illustrated in Figure 60.

Fig. 60 Operating points for two identical pumps operat-

ing singly (B) or in parallel (D). Since the pipeline re-
sistance increases with the flow rate, the combined
output of two pumps (QD) is always less than two
times the output of a single pump. For practical
purposes a single pump can be assumed to have the
operating point C.

Assuming two identical pumps with identical sep-

arate pipework combined by a branch or header
to the rising main operating in parallel, we obtain
the characteristic rising main curve as illustrated
in Figure 61. The duty point for both pumps is
obtained by plotting the sum of two pump head
curves at constant head onto the rising main char-
acteristic curve for two pumps.

Pump point of operation (D) obtained by plotting 3.5.3 Parallel Operation, Different Pumps
pump head curve onto the rising main characteris- When calculating the point of operation for two
tic curve. Total head is the sum of the geodetic head different pumps operating in parallel, different
(Hgeod) and the pipeline loss (HJ). characteristics for the separate pipework up to
the header should be assumed. The following
method for obtaining the points of operation can
be used.
3.5.2 Parallel Operation, Identical Pumps
Parallel operation is the situation where the com- The losses for each pump in their separate pipe-
bined flow of two or more pumps is directed into work before the common header are checked first.
the same rising main. The shape of the character- These can be plotted in the graph as reductions of
istic curve for the rising main will change slightly the heads, reducing the pump curves. The combi-
with the different numbers of pumps operating, nation of these reduced curves at constant head
since each pump has its own discharge line up to gives the combined head curve for the pumps. The
the common point, and the rising main constitu- intersection of this curve and the rising main
tion will therefore vary. characteristic curve is the combined point of oper-
ation. By drawing backwards from this point at

48
 Pump Performance 3

increased suction losses or loss of suction head.

Fig. 62
The designer should design the pumping plant so
that serial connection of pumps can be avoided,
and make certain that there are pumps available
for the intended duty point.

3.5.5 True Duty Point

The true pump duty point will almost always dif-
fer from calculated. The reason for this is the inac-
curacies in all numeric methods for calculation of
rising main losses, as well as the tolerances
allowed in the published pump performance
curves. Furthermore, the properties of the pump
will change with use due to wear, and corrosion or
sedimentation will change the rising main with
Operating points for two different pumps discharg- age. Figure 63 shows the relation between perfor-
ing into a common rising main. Losses in the indi-
mance tolerances. Pump inherent performance
vidual pipework are reduced from the pump head
curves plotted to scale in the graph. The combined
tolerances are discussed in detail in Section 4 of
output curve is obtained using the reduced head this book.
curves, giving the combined operating point D. The
individual operating points are A and B. For the If the duty point is located on the low flow seg-
pumps operating singly, the points of operation ment of the pump Q/H curve, and the rising main
will be C and E, respectively. characteristic curve is steep, the flow tolerance
range can be very large in proportion to designed
duty point. This fact should be taken into account
when selecting the pump.
constant head to the reduced pump curves, the
individual pump operating points can be read at
the original pump curves straight above these
intersections. Likewise, the individual operation
points with the pumps running singly are Fig. 63
obtained by reading from the head curves above
the intersection of the rising main curve and the
reduced pump curves. The method is illustrated in
Figure 62.

3.5.4 Serial Operation

It is possible to connect a number of pumps in
series in order to increase head. The combined
head is obtained by adding the individual heads
at constant flow. The complexity of the arrange-
ment makes it warranted only in rare instances,
and it is nearly always advisable to use a larger
pump from the manufacturer’s range that can do
the job alone.

Submersible pumps can be connected in series

only if the boosting pumps are installed dry, thus
making them different from the lead pump. Tolerance area for the duty point. The true duty point
for an installation may lie within the shaded area
limited by the allowed tolerance zones for pump
Another risk involving pumps connected in series
head curve and rising main characteristic curve.
is the possible failure of the lead pump, which
Pump output Q may vary largely.
may lead to cavitation in the booster pump due to

49
 3 Pump Performance

3.6 Sludge Pumping istic curve is the same as for water. With greater
solids contents the characteristic curve will be
Sludge of varying consistence is frequently being higher, but lack of data on the sludge makes the
pumped by submersible pumps in sewage treat- establishment of a correct curve difficult. Another
ment plant duty. With increasing solids content in practical problem is the fact that the solids con-
the sludge, the rising main flow losses will tent of the liquid in the pump and rising main can
increase while pump performance decreases. momentarily considerably exceed the mean or
When selecting a pump for sludge duty these two design value. In thicker sludges the pump motor
factors must be considered. The effect is illus- cooling may become a problem, depending on
trated in principle in Figure 64. cooling method.

The situation is complicated by the fact that not Generally submersible sewage pumps are suitable
enough is known about the behaviour of sludge in for pumping of treatment plant sludges with a
centrifugal pumps. Treatment plant sludge may solids content of maximum 3%. These sludges
have high gas content, either dissolved or include primary sludge, return sludge and excess
entrained, and this will have a profound effect on sludge, whereas for denser sludges, such as thick-
centrifugal pumps. As a rule, sludge with high sol- ened sludge and digested sludge, positive dis-
ids content also has a high content of gas, which placement pumps are preferred. For these thicker
will lower pump performance significantly. In liquids the pumped volumes are relatively small.
extreme cases the pump will stop pumping when
the separated gas accumulates in the impeller Propeller pumps are not recommended in sewage
eye, preventing it from developing the necessary treatment plant duty because of the risk of clog-
centrifugal force. ging. For return sludge pumping a channel type
pump in vertical column installation is a good
As a precaution when pumping dense sludge, the solution.
pump should be placed as low as possible, to
ensure positive suction head. The use of long suc-
tion pipes should also be avoided, since the pres- 3.7 Complex Rising Mains
sure drop in these is also increased by the solids
content. Long transfer sewer lines frequently have complex
profiles, with low and high turning points. Air or
With a sludge solids content less than 1%, it is usu- gas trapped at the high turning points increases
ally safe to assume that the rising main character- pump head, whereas the low turning point
increases the risk of sedimentation. There are
Fig. 64 cases where a selected pump has proven to be
Pump H curves inadequate, and cases of sedimentation are also
3% known. Exact forecasting of sewer main perfor-
H 2% mance is difficult because of the intermittent
1%
pumping action of the pumps. Water in the main
0% may move as little as 100 metres during one
pumping cycle, and air or gas in the pipeline will
not be removed and the flow will not stabilize in
0% that period.
1%
Rising main curves 2%
3.7.1 What Goes on in a Complex Rising
3%
Main?
In Figure 65 the section YK-VP contains air. As the
Q3% Q2% Q1% Q0% Q
pump starts, the liquid level VP begins to slowly
rise and the air pressure in the section YK-VP
The effect of solids content in sludge on pump head increases and a flow develops from point VP and
curve and rising main characteristic curve. The graph to point PK (v2). As the pump stops, the flow from
shows the principle only, and cannot be used for nu- VP to PK continues for some time, slowly decreas-
meric evaluations. ing. Because the duration of flow from VP to PK is

50
 Pump Performance 3

Fig. 65 Fig. 66

h1
Hgeod
h2 h3 hn

Rising main conditions. Determination of head.

longer than pump running time, the maximum 3.7.3 Pipe Size and Flow Velocity
flow velocity v2 is slower than v1. The low velocity
As noted above, the air or gas collected in the ris-
of v2 and the rising section after point AK may ing main will even the flow velocity in the follow-
increase the risk of sedimentation. The air or gas ing section, causing lower flow velocity in the low
in the section YK-VP prevents the siphoning effect points of the pipeline. This gives reason to choose
from forming, leading to increased geodetic head. a rising main of a dimension small enough to
ensure that the flow velocity does not fall too low.
The exact location of point VP is difficult to esti- Minimum pipe dimension is DN 100, however.
mate accurately. If the amount of air were con-
stant, the location of point VP could be calculated A smaller pipe has also a smaller volume, mean-
as a function of time. In practice the amount of air ing that the water moves a longer distance with
in the pipe will change, and the location of point each pumping cycle, increasing the flow velocities
VP cannot be calculated. If YK is located lower at the low point of the pipeline. From an odour
than PK, the air could in theory be removed with point of view a smaller pipe dimension is better,
an automatic air valve. If point YK is situated since the sewage stays a shorter time in the rising
higher than PK, the air will flow back into the pipe main. A higher flow velocity may also carry out
after the pump has stopped. Automatic air valves some of the air in the pipe. In these cases the
are prone to clog up in sewage. A solution could dimensioning flow velocity (v1) should be at least
be a hand-operated air valve that is opened at cer- 0,8 m/s, in more difficult cases even higher.
tain intervals according to information on air or
gas accumulation gathered over time.
3.7.4 Choice of Pump
In a complex rising main the true head may differ
3.7.2 Determination of Head considerably from calculated. If the calculated
For a rising main with a profile similar to Figure duty point is situated near either end of the pump
66, the pump total head is difficult to estimate Q/H curve allowed section, this pump should not
exactly. An estimation of magnitude can be made, be considered. A pump with a Q/H curve passing
however. The minimum Head (Hmin) is deter- above the calculated duty point should also be
mined with the rising main completely filled and considered, since it offers security of choice and
the maximum head (Hmax) as a situation with all the flow velocity increases.
downward sloping sections air or gas-filled. Thus
3.7.5 Confirming Measurements
Hmin = Hgeod + pipe flow friction losses for total
Since true duty point may differ considerably in
length of rising main cases with a complex rising main, it may be useful
to measure the volume rate of flow a few weeks
Hmax = h1 + h2 + h3 + +hn + pipe flow friction losses after pumping station commissioning, using the
for total length of rising main volumetric method. Comparing measured values
with calculated will show deviations and indicate
Real total head is a value between the maximum true state of the rising main. The measurements
and minimum value. A useful estimation may be can be repeated a few times during the first year
the mean value of Hmax and Hmin. of operation, since gas or air content in the rising
main may change.

51
 3 Pump Performance

Control measurements are necessary after com- Fig. 67

missioning. All rising mains laid out in difficult Common pipeline Discharge
terrains require careful planning and site-specific
Separate pipelines 4
considerations and technical solutions.
3

3.8 Duty Point Evaluation for

Parallel Pumping Stations 1 Hgeod 1

The combined output of two or more pumping

Hgeod 2
stations discharging at different points into the
same common rising main may be determined 2
using a graphical method. The method is
described below.

Figure 67 presents graphically the situation where

two pumping stations operate in parallel dis-
charging into a common rising main. When both
pumping stations are operating, the pump oper-
ating points are governed by the pressure at the
junction point 3, where the outputs of the pump-
Pumping stations in parallel operation. Defini-
ing stations merge in the common rising main. tions and heads.
The total heads for the individual pumping sta-
tions can be separated into components as is
shown in Figure 68. The heads comprise the fol- Fig. 68
lowing components.
H2
H
HJ 3-4 = Pipe loss in the common main
between sections 3 and 4
Hgeod 1= Geodetic head for pumping station 1 HJ2-3
Hgeod 2= Geodetic head for pumping station 2 H1
HJ 1-3 = Pipe loss in separate portion of rising
main between points 1 and 3 HJ1-3
HJ 2-3 = Pipe loss in separate portion of rising
main between points 1 and 3
Hgeod 2
The loss in the common main HJ 3-4 is equal for
both pumping stations. Hgeod 1

The combined output of two pumping stations is

graphically determined by following the steps of
the procedure shown in Figure 69:
1. The geodetic head Hgeod and the pipe loss in HJ3-4
the separate portion HJ are subtracted from
each of the pumping station H curves. The H Station 1 Station 2
curve is taken for one pump or two pumps Head components.
operating, as the case may be. The pipe loss HJ
is also determined accordingly.
3. The reduced H curves 1 and 2 obtained in step 1
2. The head loss curve HJ 3-4 for the common
is plotted onto the head loss curve both com-
main is plotted bined and separately (1+2).

52
 Pump Performance 3

Fig. 69

H H curve of pumps H H curve of pumps

Hgeod 1 Hgeod 2
1
H2 T2
H1 T1
HJ 1-3 HJ 2-3
2

H-curve
in Point 3 H-curve
in Point 3

Q1 Q Q1 Q
Station 1 Station 2

H
1

1 2

B C A
HJ 3-4
C'
B'

Q2 Q'2 Q1 Q'1 QJ1+2 Q

Establishing operating points for pumping stations discharging into a common rising main.

4. The intersection point A between the com- working singly are the intersection points C’ and
bined pumping stations H curve 1+2 and the B’ of the reduced individual head curves 1 and 2
head loss curve HJ 3-4 represents the combined and the head loss curve HJ 3-4 as plotted in step 3
output Q1+2 at the discharge point. above.
5. A horizontal line is plotted through point A,
intersecting the separate head curves 1 and 2 The procedure can be extended for installations
in points C and B respectively. The correspond- with even more pumping stations in a common
ing flow rates at these points, Q1 and Q2, repre- main. Working out the various operating points
becomes, however, an arduous task. Large sewer-
sent the pumping station individual outputs.
age systems comprise collection wells and gravity
6. Plotting the individual outputs Q1 and Q2 onto
sewer sections, breaking the network into sepa-
the individual head curves for each pumping rate pressurised sections that each can be deter-
station, the operating point for each pump is mined exactly. It is therefore unlikely that very
obtained as the intersection points T1 and T2. complex combined calculations will have to be
The operating points for the pumping stations performed.

53
 4 Testing of Pumps

4 Testing of Pumps Testing standards are conventions that are agreed

for use as a gauge for pump performance evalua-
tion. The presentation below offers methods for
Actual pump performance is established or con- their interpretation. It has been kept brief for clar-
firmed through testing. Tests may be conducted ity, and therefore it can and should be used with
for a number of different reasons and at various the appropriate standard text as immediate refer-
locations. In order to reach unambiguous results, ence.
testing standards have been developed and
agreed upon. The standards require controlled
and calibrated testing circumstances and are 4.1 Testing Arrangements
therefore generally not applicable for field or site
testing. 4.1.1 Production Testing
Testing of submersible pumps under controlled
A pump manufacturer tests his pumps at the circumstances requires a testing facility built and
works for both production development purposes calibrated to the standards governing the testing.
and quality control. Corroborated tests may be Testing facilities in a production line should also
required to confirm that pump performance is be designed for efficient handling and connection
according to the terms of purchase or to settle of the pumps so as not to slow down the produc-
performance disputes. tion process. Figure 70 shows the principle of a
submersible pump testing facility. The test rig
Field testing of pumps in actual installations will includes the necessary pipework and instruments
not yield exact data on pump performance, for pressure and flow rate measurement. The test-
because the precision of the testing arrange- ing facility may also include various measure-
ments cannot meet the terms of the testing stan- ments recording devices as well as computing
dards. These tests provide useful information on equipment for the processing and presentation of
pumps and pumping stations, however, and may the measured data. For pump head the total head,
be used for pumping station monitoring, if per- including both the static and dynamic component
formed periodically or if suitably automated. is used.

Fig. 70

Pressure gauge P2 Flow meter

M Control valve

H [m]
Z1.2
Q [l/s]

2D2

D2

Principle of a submersible production pumps testing facility, where water is circulated. All pipework is designed to pro-
vide ideal and known operating conditions for the pressure gauge and flow meter for unambiguous readings.
Obtained data are fed into computer for speedy results and evaluation service.

54
 Testing of Pumps 4

Pump total head was established in equation 18

Fig. 71
and can be written:

2 H
p v
H = ------2 + Z 1, 2 + -----2- + H J (35)
ρg 2g
v22
2g
where
D
v22
p
------2 = pressure gauge reading changed to head 2g
ρg
Z1,2
Z1.2 = pressure gauge height above water level
p2
g
2
v
-----2- = pump dynamic head at pressure measure- QD Q
2g
ment point
Using a pump head curve sheet for operating point
4⋅Q estimation. Measured static head and dynamic
v 2 = -------------- head function against flow rate are plotted. Pump
2
π ⋅ D2 operating point D is obtained graphically.

HJ = Head losses between point of measure and

Fig. 72
pump flange (calculated).
Pressure Gauge
According to the testing standards, the point of
measure of pressure shall be at a distance 2 x D2
from the pump pressure flange. The distance shall
constitute of a straight pipe section. Z1,2

4.1.2 Field Testing, Duty Point

Testing of pumps in actual installations is useful
when information on pump performance with
reasonable accuracy is required, or when pumping Hose
station performance over time is monitored.

The pump flow rate can be accurately estimated D2

with the volumetric method, where the change of
wet well level in a pumping station with known
dimensions is measured against time. If the
incoming flow in the pumping station cannot be
stemmed for the duration of the measurements,
Field testing of a submersible pump. The pressure
the effect of this must be checked separately with
gauge is connected to the pump pressure flange with
the pumps stopped. The pump flow rate can then a flexible hose. The height of the gauge above the
readily be calculated. water level in the well during the event, Z1.2, is re-
corded. Pump outlet diameter D2 is used for calcula-
Pump static pressure is measured with a pressure tion of dynamic head if flow rate is known. Air in the
gauge connected to the submersible pump near flexible hose must be removed after the pump is
the outlet flange. The pump total head can then started.
be calculated using equation 35(HJ = 0). The test
arrangement is shown in Figure 72. able, the pump operating point can be deter-
mined without an estimation of the flow rate. The
If an accurate head curve for the pump is avail- sum of the static head and the gauge height Z1-2 is

55
 4 Testing of Pumps

measured onto the H axis of the pump head curve The essence of the testing standards is to govern
sheet. The function of the dynamic head against a how the tests should be conducted technically
number of flow rates is then plotted onto the and what are the allowed performance toler-
sheet. Pump operating point will be in the inter- ances, unless otherwise agreed. The testing stan-
section of this curve and pump head curve. The dards do not regulate actions to be taken in case a
principle is shown in Figure 71. pump fails to perform according to the tolerances,
or the consequences thereof. The parties should
Site conditions do not fulfil the testing rig require- separately agree on these issues at time of pur-
ments of the testing standards. The results can chase or later.
therefore not be used for pump acceptance tests.
4.2.1 Testing Standards
A pressure gauge connected to the piping of a
The purpose of the testing standards is to define
pumping station can be used for an approximate
in detail how the tests are performed and how the
determination of the pump duty point. The height
test results shall be technically compared with the
of the pressure gauge above the water level in the
guaranteed values. The content of the standards
pump suction sump, the calculated pressure
is mainly as follows:
losses between pump flange and point of mea-
• terms, definitions and symbols
sure as well as the dynamic head (v2/2g) shall be • organizations of tests
added to the reading. For dry-installed pumps the • test arrangements
losses in the suction piping shall be deducted. • measuring uncertainties
These losses are usually minimal. • verification of guarantees

Unless otherwise agreed, the testing standards

4.2 Acceptance Tests stipulate the following guarantee values to be
compared:
Pump acceptance testing is the procedure with
which a pump is confirmed to have the properties Testing Standard ISO 9906 (Grade 1 and 2)
set out in the manufacturer’s sales literature or Pumps over 10 kW power:
contract specifications. Acceptance tests can be • Q/H duty point
routinely carried out by the manufacturer as part • Efficiency η or ηgr
of the manufacturing process or may be executed
in the presence of the customer or his representa-
Pumps less than 10 kW power:
tive.
• Q/H duty point
• Efficiency η
The testing standards contain two major princi-
• Motor power input Pgr (over the range of oper-
ples:
• The pump is tested at a duty point agreed ation)
upon at the time of purchase.
• The pump is tested at any point along the Pumps produced in series with selection made
curve published for the pump. This practice is from typical performance curves (Annex A):
intended for pumps in serial production, and • Q/H duty point
the allowed tolerances are larger than those • Efficiency η
for custom-made pumps. • Pump power input P
• Motor power input Pgr
In the case of pumps in serial production, the
pump manufacturers perform production tests of Testing Standard ISO 2548 (Class C)
their pumps either at a multitude of points along • Q/H duty point
the curve or at three distinct selected points. • Efficiency η or ηgr
These three points are selected at both ends of
Pumps produced in series with selection made
the allowed portion of the pump curve, and at a
from typical performance curves (Annex B):
point at the middle of the curve.
• Q/H duty point
• Motor power input Pgr

56
 Testing of Pumps 4

Testing Standard ISO 3555 (Class B) Fig. 73

• Q/H duty point
• Efficiency η or ηgr H

These standards contain performance values for

+ tH XG
the tolerances of the measured variables. .

- tQ . QG + tQ . QG
HG
If guarantee values are required in specifications - tH . XG
or sales contracts, the following variables accord-
ing to the testing standards are suitable:
• Q/H duty point Measured curve
t
• Efficiency ηgr or η G G

The desired duty point and testing standard to be

used must also be specified.
QG Q
The testing standards do not require testing of
Verification of guarantee on flow rate, head and
the pump NPSHr value unless specifically
efficiency according to ISO 9906.
requested. NPSH tests are difficult and time-con-
suming, and do not provide absolute information
on the possibilities of cavitation, as can be seen in Q/H curve where it is intersected by the straight
Section 1.4.3. The benefits of NPSH tests are thus line passing through the specified duty point QG,
questionable. The testing standard ISO 9906 pro- HG and the zero of the Q, H axes and from where a
vides tolerance factors for the NPSHr value. In the
vertical line intersects the η curve.
testing standards ISO 2548 and ISO 3555 no toler-
ance factors for NPSHr are provided. The guarantee condition on efficiency is within
tolerance if the efficiency value at this point of
Testing Standard ISO 9906 Grade 1 and 2 intersection is higher or at least equal to ηG (1-tη).
The new testing standard ISO 9906 was published
in 2000 and is scheduled to replace the older test- Testing Standards ISO 2548 (Class C) and
ing standards ISO 2548 and ISO 3555. ISO 3555 (Class B)
The ISO 2548 (Class C) corresponds generally to
Grade 1 requires higher accuracy, whereas Grade 2
the standard ISO 9906 Grade 2 and the ISO 3555
allows larger tolerances. Because sewage pumps
(Class B) to the standard ISO 9906 Grade 1. The
usually operate in intermittent duty, Grade 2 is
standard ISO 2548 is suitable for sewage pumps.
suitable for these pumps. Grade 1 is intended for
the testing of fine-tuned process pumps in contin-
In these standards an elliptic graphic method for
uous duty. For the verification of guaranteed val-
the verification of guaranteed values is used. The
ues, a crosshair method is used. The principle is
principle is shown in Figure 74. The efficiency veri-
shown in Figure 73.
fication is performed in the same way as in stan-
dard ISO 9906.
The verification principle as shown in Figure 73
works in the following way:
The verification principle as shown in Figure 74
works in the following way:
A tolerance cross with the horizontal line ±tQ·QG
and the vertical line ±tH·HG is drawn through the An elliptic tolerance zone with semi-axis QGXQ
guarantee point QG, HG. and HGXH is drawn with the guaranteed duty
The guarantee on the head and flow rate has been point QG, HG as centre point.
met if the measured Q/H curve cuts or at least
touches the vertical and/or horizontal line.
The guarantee on the head and flow rate has been
met if the measured Q/H curve cuts or at least
The efficiency shall be derived from the measured
touches the ellipse.

57
 4 Testing of Pumps

The maximum possible true deviation from

Fig. 74
desired volume of flow is also dependent on the
shape of the rising main characteristic curve and
H location of duty point on the pump Q/H curve.
Q
Guaranteed point QG, HG According to the standards ISO 9906, Grade 2 and
ISO 2548, the deviation near the optimal point
HG H may be ± 3 …10%, depending on main curve shape.
HG X H For the standard ISO 9006, Grade 1 and ISO 3555,
the corresponding deviation is ± 2…6%. If the duty
point is in the low volume flow range, and the ris-
Q G XQ ing main characteristic curve is flat, the deviations
Measured curve
above may be much greater.

The normal tolerances for pump efficiency accord-

QG Q
ing to the ISO testing standards are as follows:
• ISO 9906, Grade 2 -5%
Verification of guarantee value on flow rate and • ISO 2548 -5%
head according to ISO 2548 and ISO 3555. • ISO 9906, Grade 1 -3%
• ISO 3555 -2,8%

The efficiency shall be derived from the measured These are proportional values, not percentage
Q/H curve where it is intersected by the straight points.
line passing through the specified duty point QG,
HG and the zero of the Q, H axes and from where a For sewage pumping the ± tolerances of the stan-
vertical line intersects the η curve. dards ISO 9906, Grade 2 and ISO 2548 are quite
acceptable. They are also compatible with normal
The guarantee condition on efficiency is within production variations in manufacturing. More
tolerance if the efficiency value at this point of stringent requirements may cause extra costs in
intersection is higher or at least equal to ηG (1-tη). production and delivery delays. The published
curves of sewage pumps are also based on these
Grundfos has developed an application method of standards, a fact stated on these curves.
the elliptic tolerance zones of the ISO 2548 stan-
dard, making it more convenient to use in Sometimes a condition stipulating that negative
numeric calculations. The method uses the slope tolerances are unacceptable may be raised by cus-
of the tangent to the Q/H curve at the point of tomers. Problems and misunderstandings are a
examination, and makes it possible to numeri- common consequence of this, with difficulties for
cally determine Hmin and Hmax at guaranteed rate both manufacturer and customer. The ISO testing
of flow, so that the ellipse condition is met. standards do not recognize asymmetric tolerance
systems, and the published curves of the manu-
facturers are based on the symmetric tolerance
Other Testing Standards
systems of the ISO standards. If the customer
Many countries have issued national standards finds that a true volume flow lesser than indi-
equivalent to the ISO standards. In the U.S. a cated by the published curve is unacceptable, a
national testing standard issued by Hydraulic better solution than requesting no-negative toler-
Institute is frequently used. This standard differs ances would be to increase the required volume
from the ISO standards with regard to the toler- flow by 3…10% and then select a pump based on
ance system. this figure.

Allowed True Performance Deviations Requiring no-negative tolerances for the pump
The maximum possible deviation from the guar- efficiency does not make much sense, since pump
anteed duty point is composed of inaccuracies of manufacturers would then be forced to lower the
the measurement technology and the allowed published nominal figures. Too low, conservative
tolerances. The testing standards specify the figures would not be representative of the major-
accuracy requirements of the measuring instru- ity of the pumps and lead to misunderstandings
ments and guide values for the allowed tolerances. and confusion.

58
 Pumping Stations 5

5 Pumping Stations In a good design the start and stop levels should
be relatively close to each other for the following
reasons:
The working environment for submersible pumps, • Pump starting frequency becomes high
regardless of size, is the pumping station. Pump- enough to prevent sludge an impurities from
ing station design and construction is decisive for settling onto well floor.
the performance of the pumps, and care and dili- • Pumping station inlet should stay low relative
gence should therefore be exercised whenever to the liquid level in the wet well.
specifying them. The following is a primer of
pumping station design offering hints and advice A guideline maximum value for the effective vol-
for the design engineer and the operator of ume height in small pumping stations is approx. 1
pumping stations. Some aspects of pump opera- m and 2 m in larger pumping stations.
tion and interaction with the pumping station
pipework is also discussed. The effective volume can be substituted with the
wet well surface area using the following equa-
tion:
5.1 Pumping Station Basic Design Q
A W = ------
The decisive factor for pumping station operation 20 (36)
is a good hydrodynamic design. A bad pumping
station design may lead to pump malfunction, where
uneconomical pumping and frequent needs for AW = wet well surface area in m²
pumping station service and cleanout. Q = pumping station total flow rate, l/s

Modern sewage pumping stations are designed For small pumping station flow rates, however,
for pumping of unscreened sewage, and the the surface area will be limited by the physical
design criteria for these differs from those for dimensions of the pumps when submersible
clean water. In the following the design and spe- pumps are used. The surface area will then be
cial requirements of sewage and stormwater
pumping stations are discussed.
Fig. 75
5.1.1 Wet Well Volume and Surface Area
The wet well effective volume should be of correct
size. Too large a volume may lead to accumulation
of sludge in the well, whereas too small a volume
leads to too frequent starting and stopping of the
pumps. The use of modern submersible pumps,
with high allowable starting frequency, leads to
smaller and more efficient pumping station
designs.

The effective sump volume is the volume

between pump start and stop levels, and it can be
determined with the use of nomograms as a func-
tion of allowed starting frequency. A method for
the calculation of the effective sump volume is
presented in Appendix B of this book.

In reality the incoming volume in a pumping sta- B = 1,5 D

tion varies greatly over time, and the mean start- C = 0,8 D
ing frequency will therefore be lower than Recommended pump installation dimensions for
theoretical. submersible pumps.

59
 5 Pumping Stations

larger than obtained with equation 36. Recom-

Fig. 76
mended pump installation dimensions are shown
in Figure 75.

For larger flows the direction of approach towards

the pumps should be square on. If the flow comes
from behind, the submersible baseplates disturb
the flow causing eddies to form. These impede
pump operation, lowering pump performance
and efficiency and increasing the risk of cavitation
and pump vibrations.

5.1.2 Pumping Station Inlet Pipe

Location and size of the pumping station inlet
pipe is important for the function of the pumping
station. Problems encountered in operation of the
pumps are frequently caused by bad inlet pipe
design.

An inlet pipe located too high relative to the liquid

surface or with a high flow velocity may cause
entrainment of air and the formation of eddies in
the water when splashing into the well. Air mixed
into sewage water has a tendency to remain
because of the possibility of air bubbles to adhere
to the solid particles present. A separate calming
Inlet locations to be avoided. Too high an inlet fall
chamber may therefore not alleviate the situation
height may lead to entrained air reaching the pump
at all. inlet directly or along the bench surface, with con-
sequent operational problems in the pumps.
Inlet fall height should always be minimized and
should not exceed 1 m with the water level down,
regardless if the pumping station has a separate Location of the inlet pipe should be as remote as
calming chamber or not. The effect of a high inlet possible from the pump inlets. Figure 76 shows
fall height cannot be effectively alleviated with designs to be avoided.
baffles.
Flow velocity at the inlet should not exceed 1,2 m/
Air entrained in the water has a tendency to s so as to avoid the formation of eddies in the wet
remain inside the pump impeller, where the cen- well.
trifugal force makes it accumulate around the
impeller hub. This may lead to increased power
5.1.3 Wet Well Floor Shape
requirement and lowered performance and effi-
ciency. The risk of cavitation and pump vibrations The shape of the wet well floor is important for
also increases. If the amount of air in the pump is the functioning of a sewage pumping station. A
very high, the pump may cease to function alto- good design prevents bottom sedimentation, but
gether. may also assist in the prevention of scum forma-
tion and the accumulation of flotsam on the sur-
Air is frequently a problem for pumps drawing face. The following principles should be
directly from aeration basins in treatment plants, recognized in a good bottom design:
because of the high content of air. If a pump is
placed in an aeration basin it should be placed as All corners should be benched to a minimum
low as possible, with the suction pipe near the bench angle of 45°, in small pumping station the
bottom. bench angle may be as high as 60°. The angle
may be smaller, if the section is flushed by a flow.

60
 Pumping Stations 5

Bottom area should be minimized and liquid vol- Fig. 78

ume below pump stop level should be kept to a
minimum.
D2
Minimizing the bottom area and the residual vol-
ume, the flow velocities near the inlets will
increase, flushing out possibly settled sludge. A
surface area decreasing with the falling water
G
level leads to less accumulation of surface debris.
0,2 m
D1
5.1.4 Stop Levels L R
Dp
The start and stop levels are specified at the hs 45 Reducing
design stage. They should always be checked for bend
F
function and possibly altered at commissioning in
order to secure good operation.
Recommended installation dimensions for vertical
The stop level should be as low as possible, so that dry-installed submersible pumps. F = 0,5 · D1,
the flow velocity increases toward the end of the v1max = 2,0 m/s, G = Dp, L ≥ D1 + 100 mm, R ≈ L.
working cycle. Limits for the stop level are set by
the required motor cooling submergence or by
the level when air becomes sucked into the pump In pumping stations with two submersible pumps
intake. The latter level cannot always be foreseen, in duty-standby configuration the stop level can
but must be confirmed at trials during pumping normally be set below the motor even if the
station commission. motor is cooled chiefly by submergence, see Fig-
ure 77. The identical pumps are selected so as to
be able to cope with the pumping station flow
Fig. 77 alone, and the risk of the liquid level remaining
for long near the stop level is slight. Submersible
pumps also have protective devices against over-
heating that stop the pump in case of inadequate
cooling conditions.

In pumping stations with a multitude of sub-

mersible pumps running under varying condi-
K tions, the stop level must be set so that the pump
motors always have enough submergence for
adequate cooling. Pumps with cooling water jack-
hs2
ets or other means of heat dissipation indepen-
hs1
E dent of submergence are preferred in such
installations.

The stop level setting for dry-installed pumps is

hs1 = E+a
dependent on the suction pipe inlet height, shape
a = 100-300mm
hs2 = E+k/2 and flow velocity. 200 mm above the suction pipe
inlet is a good rule-of-thumb for this height, and
useful for the designer. The shape of the suction
pipe inlet is important, and good designs are
Recommended stop levels at the design stage.
shown in Figures 78 and 79. For this inlet shape a
hs1 = stop level for two submersible pumps in duty-
standby operation or pumps with cooling indepen- provisional pump stop level height can be calcu-
dent of motor submergence. hs2 = stop level for mul- lated using the following equation:
tipump installations with motors cooled by
submergence. Final stop level settings should be de- h s = 0, 04 Q + 0, 2 (37)
termined during commission trials.

61
 5 Pumping Stations

where Fig. 79
hs = stop level height, m
Q = pump flow rate, l/s D2

In pumping stations with several different stop

levels, such as in frequency-controlled installa-
tions, it is important to program the control 0,2 m
sequence to pump down to lowest stop level at D1
least once per day to clean out the bottom.
hs 45

5.1.5 Start Levels F

If the wet well surface area AW is dimensioned

Eccentric
using equation 36, the first start level in a pump- reducer
ing station with two submersible pumps in duty-
Recommended installation dimensions for horizon-
standby configuration can be set 1 m above the
tal dry-installed submersible pumps. F = 0,5 · D1,
stop level. Where small inflows are encountered, v1max = 2,5 m/s.
the start level may be lower. The second start
level, can be set 0,2...0,3 m above the first.
from case to case. Wider NPSH safety margins
In pumping stations with more than two pumps
may be warranted.
the starting levels should be considered from case
to case. If the pumps have a common stop level, a
Recommended suction pipe inlet designs are
suitable design would be with the first start level 1
shown in Figures 78 and 79. The downward suc-
m above stop level and the following start levels
tion exerts a cleansing flow on the pumping sta-
at 0,3 m intervals from this. If the pump stop lev-
tion floor, and is less prone to suck air from the
els are staggered should the start levels be set at
surface.
or near equal intervals.
In vertical pumps the suction pipe will have to
In pumping stations with dry-installed pumps the
turn 90° to reach the pump suction cover. The
starting levels have to be set above the pump cas-
bend before the pump suction inlet is crucial for
ing in order to ensure that the casings fill up and
the function of the pump, since it causes the flow
the pumps start pumping. For vertical pumps, this
to be irregular. Too sharp a bend may cause impel-
height may be considerable and should be set
ler cavitation, lower pump efficiency and cause
with a margin according to Figure 78.
vibrations. If the pump suction inlet is smaller
than the suction piping, a reducing bend should
Horizontal pumps do not normally require special
be used, minimizing the interference. Figure 78
considerations for the start levels, if the suction
shows suction bend dimension recommenda-
pipe is designed to prevent air pockets from form-
tions.
ing, see Figure 79.
The contraction of the straight inlet pipe to a hori-
5.1.6 Suction Pipe Dimension and Design zontal pump should be eccentric so as to avoid air
Design and dimensioning of the suction pipe is from collecting and possibly blocking the impeller.
important, with bad designs possibly causing
vibrations, lowered pump efficiency and risk of An inlet design with unfavourable flow charac-
cavitation. teristics may cause a pressure drop large enough
to spend the available NPSH and lead to pump
The suction pipe should be dimensioned so that cavitation. The recommended NPSH margin
the flow velocity does not exceed 2,0 m/s for ver- should be observed in installations where the suc-
tical pumps and 2,5 m/s for horizontal pumps. tion pipe geometry gives reason to concern. The
When new, bigger pumps are installed in old concept of cavitation and NPSH and recommen-
pumping stations, these figures may have to be dations for NPSH margins are presented in detail
exceeded. The situation must then be considered in Section 1 of this book.

62
 Pumping Stations 5

5.1.7 Pumping Station Internal Pipework Fig. 80

The internal pressure pipework in a pumping sta-
tion should be selected for a flow velocity of 2...3
m/s. Especially if the sewage contains sand should
the flow velocity be at least 2 m/s, in order for the
sand to be carried with the flow out of the pump.
In frequency-controlled installation this require-
ment may cause problems at low frequencies. Fig-
ure 58 in Section 3 shows size recommendations
and losses typical for piping. The pipework should
have a dimension of at least 100 mm but can be
80 mm in small pumping stations, provided that
the pump free passage is 80 mm.

The use of flexible joints in the internal pipework

is not recommended, since most pipe vibrations
are pressure-induced by the flowing liquid, and
cannot be avoided by the use of flexible joints.
When installing flexible joints the pipe is cut, and
the section will be subject to a separating force
with a magnitude of pump pressure x area. The
pressure near the pump is pulsating at a fre-
quency determined by pump speed and number
of impeller channels, causing the piping and joint
to vibrate. The vibration is more pronounced
when flexible joints are installed. Flexible joints
are also susceptible to damage.
Pressure pipework branch designs. The design
The pressure pipework is normally expanded after should emphasize smooth transition and prevent
the pump, and in order to save energy these tran- sludge in rising main from settling on valves in
sition pieces should have a conical shape with a pump risers when the pumps are stopped.
maximum shank angle α of 10°. Please see Figure
79.

For vertical dry-installed pumps and submersible 5.1.8 Flushing Devices

vortex pumps the check valve should be installed Pumping station flushing devices consist of a
as far away as possible from the pump in order to remote-controlled by-pass valve mounted on the
alleviate possible problems with air in the pump submersible pump before the connection to the
at start-up. pipework. When the valve is opened, the pump
flow is directed back into the wet well, agitating
For horizontally installed large pumps, where the the liquid and causing settled sludge and scum to
shaft bearings include separate bearings for radial disperse. The suspended matter will then be
and axial forces, the check valve must not be pumped out with the liquid when the flushing
placed directly in the vertical pipe from the pump valve is closed.
delivery flange. Possible shocks from a rapidly
closing valve may pound at the pump hard The flushing valve should be of the normally
enough to gradually damage the radial bearings. closed type (e.g. a spring-loaded pneumatic
device) so that in the case of a malfunction the
In multipump installations the pump pressure pumping action will be able to proceed.
pipes should be joined by a branch designed to
prevent settling of solids during pump stoppage In a correctly dimensioned and shaped pumping
into the individual pipes, which may lead to valve station wet well flushing devices are normally not
blockage. Good branch designs are shown in Fig- needed. They serve a function in old, large wet
ure 80.

63
 5 Pumping Stations

wells and in special situations, where the sewage Fig. 81

contains large amounts of e.g. grease. A flushing
device can also be retrofitted without changes to
wet well structures. The flushing devices are
brand specific and detailed information is avail-
able from the pump manufacturer.

5.1.9 Odour Problems in Pumping

Stations
A sewage pumping station may cause odour prob-
lems in its immediate environment. Many factors
affect the situation, such as pumping station
location, sewage quality, situation before the
pumping station and wet well dimensions and
design. If the pumping station is fed by another,
remote pumping station, the sewage transfer
time between the pumping stations may be so
extensive that the sewage turns septic by anaero-
bic action. Septic sewage produces Hydrogen-Sul-
phide (H2S) that, apart from being toxic, also
creates a typical foul odour.

The occurrence of odour problems is practically

impossible to predict. In case of severe problems,
they may be attempted to be corrected by the fol-
lowing measures:
• Lowering start and stop levels, in order to cut
the retention time in the wet well and prevent
sludge from forming.
• Installing a submerged inlet bend in the wet
well, in order to convey the incoming sewage
below the surface, thus preventing aerosols
from forming.
• Installing air filters in the wet well ventilators.
• Dosing odour-preventing chemicals into the
sewer upstream from the pumping station.

5.1.10 Pumping Station Design Examples

Wet well design will depend on pumping station Pumping station design for submersible pumps
size and flow volume. Figures 81…84 show princi- and relatively small flows (Q = 4...50 l/s). The
ples for wet well design for various cases and preferred cross section of small pumping sta-
pumping station sizes. A pumping station with tions is circular, which minimizes liquid surface
submersible pumps for large flows can be area and avoids corners where sludge could ac-
designed according to Figure 83. If the pumps cumulate. Minimum diameter 1,5...2 m to facil-
require it, the stop level can be set at height hs2. itate service workovers.

The flow velocity vD in the expanding section of

the wet well must be high enough to avoid Q
D = --------------------------------
sludge settling. A suitable value for vD is 0,1...0,3 1000 ⋅ V D ⋅ C (38)
m/s when the liquid is at stop level. The dimen-
sion D can be calculated using the relation where
Q = pumping station flow rate, l/s

64
 Pumping Stations 5

vD = flow velocity in expanding section,

0,1...0,3 m/s
D, C = pumping station dimensions, m

Fig. 82

Pumping station design for two submersible pumps and moderate flows (Q = 50...2000 l/s). The elongated wet well
shape is an important feature that places the inlet pipe away from the pumps and prevents the build-up of sludge on
the wet well floor.

65
 5 Pumping Stations

Fig. 83

stop
vmax = 1,2 m/s

Pumping station design for several submersible pumps and large flows. If the pumps are depending on submergence
for cooling, the stop level hs2 is chosen accordingly.

66
 Pumping Stations 5

Fig. 84

stop vmax = 1,2 m/s

Wet well design for a pumping station with multiple dry-installed pumps. Flow velocity across the suction bends vo =
0,3...0,4 m/s with the liquid at stop level. Pump internal distance B can be selected as for submersible pumps, whereas
the distance M should be selected according to inlet fall height, and should ensure an even flow at the suction inlets.

5.1.11 Dry-installed Pump Positions culations for installations where more than one
For dry installation most manufacturers can offer pump is operated simultaneously.
pumps for both vertical and horizontal installa-
tion. Usually a pump in horizontal position offers Large pumps for horizontal installation are fitted
advantages, such as: with slide bars for easy removal of pump motor
• simplified piping with less bends from the pump housing. Please refer to Figure 17
• suction flow to the impeller is even in Section 2.
• lower pump position.

For larger pumps the NPSH safety margin require-

ment may not be met for pumps in vertical posi-
tion, because of pump location and greater
margin requirement, whereas a horizontal pump
will be acceptable. All possible pump duty points
must be considered when performing NPSH cal-

67
 5 Pumping Stations

Fig. 85 5.2 Package Pumping Stations

5.2.1 Out-of-doors Pumping Stations
Package pumping stations are made ready at a
factory for installation on site. The material used
is glass-fibre reinforced plastic (GRP) or, for
smaller pumping stations, Polyethylene (PE), and
the stations are made complete with all internal
pipework and other components in place. Thus
the installation is reduced to the excavation of the
site, laying of a foundation and connecting the
station to the incoming sewer and rising main,
and connection of the control panel to the power
supply and possible telemetry connections.

The buoyancy of the pumping station when

empty requires it to be anchored to a foundation
or concrete slab, which also may be prefabricated
and matched to the pumping station foundation
bolts. The concrete slab mass can be calculated
using the following equation:

M B = 2000 ⋅ V G
(39)

where
MB = concrete mass (kg)
VG = volume of pumping station below water
table (m³)

The pumping station must be vented to prevent

the build-up of toxic or explosive gases. If there is
risk of freezing, the upper part of the pumping
station can be insulated.

Packaged pumping stations are fitted with access

covers that may be made of aluminium or galva-
nized steel and moulded into the structure. The
internal pipework can be either cast iron or thin-
walled stainless steel with fabricated bends and
Wet well-dry well pumping station outlines. Pumps branches. Valves should be cast iron and suitable
can be installed vertically (A) or horizontally (B). Sub- for use in either horizontal or vertical position.
mersible pump construction is protected against ac- Figures 86…88 show typical package pumping sta-
cidental flooding of the dry chamber. A separate tion arrangements.
sump pump is provided in the dry well for drainage
of leakage water. The pumping station control panel
can be installed on the top or in the dry well space
above flood level.

68
 Pumping Stations 5

Fig. 86 Fig. 87

Package pumping station with separate above- Typical package pumping station. The wet well is
ground service building. Wet well collar serve as complete with folding work platforms for valve ac-
foundation for the building. cess and service.

69
 5 Pumping Stations

Fig. 88 5.2.2 Indoor Pumping Stations

Pumping stations for very small capacities can be
installed indoors, for instance in basements of
buildings close to the source of the effluent. These
may be designed as containers with the pumps
integrated or mounted externally. Figure 89
shows a typical arrangement.

5.3 Pumping Stations with

Column-installed Pumps
Pumping stations with column-installed pumps
typically have large pumping capacity, and
especially axial pumps are sensitive to the condi-
tions in the suction chamber. Figure 90 shows
recommended distances between pumps and
between pumps and wall sections. It is very
important that the flow feeding the pumps is
even and that the flow velocity at this point does
not exceed 0,5 m/s.

A package wet well-dry well pumping station. Round Fig. 90

wet well shape adds strength and facilitates manu-
facturing. Dry-installed submersible pumps are safe
against flooding and are easy and clean to maintain.
Intermediate platform offers access to control panel,
mounted above flood level.

Fig. 89

Pumping station for small flows. The pump may be

integrated into the structure of the container and
can easily be removed. Air-tight construction is suit-
able for indoor installation, and the unit can be in- Pump distance and suction flow velocity recom-
stalled near the effluent source. mendations for column-installed pumps.

70
 Pumping Stations 5

Fig. 91

3 3

2 2

1 1

10 50 100 500 1000 5000

Pump immersion depth recommendation according to proposed CEN (draft) standard.

Pump immersion must be sufficient in order for to appear unexpectedly. A vortex can be pre-
suction vortices to be avoided. Figure 91 shows a vented by placing a float on top of it, if possible.
dimensioning recommendation diagram accord-
ing to the proposed new CEN standard (draft). The Pumps are frequently installed in columns in a
appearance of suction vortices is still impossible manner where the water exits straight upwards
to completely predict beforehand. Pump charac- through the column. In these cases the pump
teristics and flow conditions in the suction cham- head and energy use can be changed by the top
ber influence the development of suction vortices. design. A good working design is shown in Figure
The suction chamber shape may induce vortices 92. The pump column is terminated well below

Fig. 92

Top outlet design and head determination for column-installed pumps.

71
 5 Pumping Stations

the weir, allowing the flow to smoothen before 5.4.1 Regular Sewage Pumping Stations
flowing over the weir. In this design the pump Flow rate estimation of residential sewage is nor-
head can be calculated with workable accuracy mally based on population numbers. The flow
using the equations in Figure 92. Losses in the col- rate varies in daily and weekly cycles, the variation
umn can be disregarded in practical examina- being in the range of 0,5...1,5 times average flow.
tions. Industrial effluent must be estimated on a case by
case basis, depending on the plant type in ques-
tion. The amount of leakage water present
5.4 Pumping Station Dimension depends on a number of variables, such as water
table level, local rainfall and soil characteristics
Selection and general condition of the subterranean pipe-
work. It can be estimated as units per pipeline
Pumping station dimensioning is based on the
expected incoming flow, which usually must be length unit, e.g. kilometre or as a ratio related to
estimated without the use of collected data. quantity of the sewage.
Guidance values cannot always be applied, since
flow rates depend on a great number of variables. The possibility of flooding at the pumping station
Figures are available from the sewerage systems with consequent environmental damage must be
designers or, less accessible, from technical litera- taken into account. For this reason sewage pump-
ture on the subject. The possibility of sewerage ing stations have two pumps in duty-standby con-
system future expansions must also be consid- figuration, with each pump capable of handling
ered as reserve capacity or flexibility in pump size peak flow. Thus flooding will not occur in situa-
installation. Sewage pumping station incoming tions when one pump is out of order or being shut
flow is also typically greatly varying with time, down for service. If pumping station capacity is
both in short and long cycle. based on two pumps operating in parallel, a third
pump should be provided as standby. Estate or
Incoming flow estimation always starts with the other private pumping stations may be equipped
analysis of the possible constituent parts. These with a single pump, since the incoming sewage
are normally classified as flow can easily be controlled by restricting the
• residential sewage usage of facilities.
• industrial effluent
• stormwater (rain and melting snow) 5.4.2 Stormwater Pumping Stations
• leakage water Rain water flow rates are considerably larger than
Of these, leakage water is water entering the sew- other stormwater sources, such as melting snow.
erage system from ground water leaks, leaking Dimensioning of the system is based on the larg-
water mains or stormwater inadvertently enter- est anticipated amount that will reach the pump-
ing a separate sewage system through manholes ing station. This amount may not necessarily be
or other entrances, such as worksite excavations. the most severe torrent, for stormwater sewers
are allowed to flood under heavy rain circum-
To correctly dimension a pumping station, the stances because of the relative harmlessness of
type of use must be known. Sewerage systems are rainwater. The design values are also affected by
classified as flood tolerance of the area and the type of urban
• sewers for regular sewage, receiving domestic environment in question. Leakage water addition
and/or industrial effluents only must also be considered.
• stormwater sewers, handling stormwater only
• combined sewers, handling both regular sew- Controlled flooding at the pumping station is
age and stormwater in various proportions. arranged with overflow weirs, discharging in a
suitable direction, such as a ditch or canal.

Stormwater pumping stations do not have the

same required reliability factor as sewage pump-
ing stations, and they can be designed to handle
the maximum flow with all pumps running in
parallel duty.

72
 Pumping Stations 5

Fig. 93

Ejector and pump for retention basins.

5.4.3 Combined Sewage Pumping Sta- consisting of an ejector drawing air from the sur-
tions and Retention Basins face combined with a submersible pump. The
design is shown in Figure 93.
The dimensioning flow rate for combined sewage
pumping stations is the sum of the estimated
The air provided by the ejector makes the mixing
sewage, stormwater and leakage water flow
more effective at low water depths. The required
rates. Reliability requirement is the same as for
pumping power can be estimated at approxi-
sewage pumping stations, making it necessary to
dimension them with at least one pump as mately 70 W per m2 basin bottom area. The ejec-
standby. Combined sewage pumping stations tors should be placed and directed so that the jets
combine in an unfavourable way the properties of flush the solids towards the basin drain.
regular sewage and stormwater pumping sta-
tions, and their use is therefore discouraged.

In conjunction with both stormwater and com-

bined sewage pumping stations, retention basins
may be used for temporary storage of incoming
sewage that exceeds the installed pumping
capacity. As the flow decreases (such as after
heavy rainfall), the basin is emptied by pumping
or through gravity sewer, and normal pumping
station operation can be resumed. Retention
basins may also be used to even out fluctuations
of the incoming sewage flow to a treatment
plant. Essential in a retention basin is to prevent
solids from settling onto the basin floor when
emptying. This can be accomplished by designing
the basin shape “self-cleaning” or by agitating
and mixing the basin content. Special ejector mix-
ers have been developed by pump manufacturers,

73
 5 Pumping Stations

5.5 Pump Selection • Effect of liquid level variations on pump duty

point. Liquid level may vary in suction sump as
well as in discharge reservoir. If the pump duty
5.5.1 Pump Selection Based on Pump points moves into the cavitation area (NPSHr>
Curves 10 m) because of rising suction level and thus
The pumps for a pumping station project are ini- increased suction head, the pump can nor-
tially selected using the methods described in mally be used without consequences, because
Chapter 3 of this book. It is good practice to con- the NPSHA will increase correspondingly. Pump
sider a number of pumps from a manufacturer's cavitation is thus prevented, and only pump
range that have curves passing near the desired power requirement and available motor power
operation point. must be confirmed. It is recommended that
the pump manufacturer is consulted in uncer-
Pumps having curves both above and below the tain cases. It is especially important to check
initial requirement should be included, because all possible level combinations for propeller
other considerations, such as pump efficiency and pumps, since these have very narrow allowed
cost may be economically decisive factors. Q/H bands, because of strongly varying power
curves.
The intermittent character of sewage pumping
station operation allows a wide margin for the
pump selection, giving the designer freedom of 5.5.2 Observing Pump Efficiency
choice beyond a fixed nominal point of operation. With larger pumps, the pump efficiency becomes
Theoretically calculated operating points are increasingly important for the pump selection.
uncertain in any case, since actual head may vary When warranted, whole life costs calculations
due to changing start and stop levels brought should be performed for a number of alternatives.
forth by programmed level control, pump wear Please refer to Section 7. All duty points at differ-
and tolerances in pipeline and pump characteris- ent duty situations should be taken into account.
tic curves. The following three different cases should be
examined separately:
For instance, a pump with a head curve higher A. Two pumps installed in duty/stand-by configu-
than originally desired may offer better overall ration, or all pumps have separate rising
economy, especially if the pipeline characteristic mains.
curve is flat, or the dynamic losses are small com- In these cases the pumps have only one duty
pared to geodetic head. point (if the variations in suction liquid level
are disregarded), and the situation is fairly
The pumping station designer is therefore well easy from an efficiency point of view. Pump
advised to select a pump from the manufacturer's selection should not be based on having a
standard range and to refrain from requesting pump Q/H curve passing near the desired duty,
exactly tuned pumps. The use of standard pumps if the pump best efficiency point falls far away
will also simplify pump spare part service and from it. Another pump with a Q/H curve pass-
later pump replacement, if needed. ing above the desired duty point, but with a far
better efficiency may be found in the same
The pump selection should be checked so as to price range, and be a far better choice.
make sure that the operating point under any cir- B. Several duty pumps with common rising main.
cumstances does not fall outside allowable range In this case the pumps may have several duty
of the pump curve. Operating range restrictions points, depending on number of pumps in
can be imposed for a number of reasons such as operation. Normally the duty point is selected
risk of cavitation or vibration, or overloading. The for the situation when all duty pumps are in
following should be checked: operation at the same time. In order for the
• Single pump duty points in installations with pump efficiency to be as good as possible with
several pumps pumping in parallel into com- less pumps running, the pump should be cho-
mon rising main. Duty points for situations sen so that pump best efficiency point lies to
with one, two and more up to and including all the right of the main duty point, please see
pumps operating in parallel. Figure 94.

74
 Pumping Stations 5

Fig. 94 Fig. 95

max max

Selection point
Selection point

Several duty pumps with common rising

main. Pump ηmax should lie to the right of the selec-
Frequency control operation. Pump ηmax should lie
tion point.
to the left of the selection point.

C. Pumps used with frequency control. even be incorrect. In particular, information on

In order for the efficiency to be acceptable also the following items is important:
at low frequencies, and for the Qmin to be • Will more than one pump use a common rising
small enough, the best efficiency point should main? In this case the rising main characteris-
lie to the left of the main duty point. Please tic curve or the number of duty pumps and
refer to Figure 95. value of geodetic head are needed.
D. Several duty pumps with common rising main • Information on frequency control usage.
and frequency control. • For pumps in column installation information
For this case it is likely that the best choice is a on nature of liquid is needed, in order to ascer-
pump with the best efficiency point coinciding tain the possibility of using axial propeller
as closely as possible with the main duty point. pumps.

Frequently more than one pump, even from the 5.5.3 Number of Pumps
same manufacturer, may be considered for a Sewage pumping station pumps are selected so
desired duty point. One alternative may offer that at least one pump always is on standby. Espe-
lower costs but have lower efficiency than cially in larger pumping stations, the number of
another. The decision between these pumps pumps should be selected so as to optimize pump
should in principle be based on a whole life cost usage and investment cost. The cost of pumping
analysis. This assessment will frequently have to capacity, or pump power in kW, decreases with
be performed by the customer or his consultant, increased pump unit size. On the other hand, the
since the pump manufacturer usually does not requirement of one standby unit will increase the
have all relevant information. The position of the cost of redundancy if few very large units are
buyer may also be significant, since a contractor used. Installation costs are therefore almost con-
may stress purchasing price over operating costs, stant for a given capacity, regardless of number of
whereas the owner will look at total costs. pumps used to meet the requirement, at least
within a reasonable range. Likewise, the energy
Unfortunately, the pump manufacturer fre- cost will remain almost constant, if the pumps
quently has to select and offer pumps with very considered can run near the optimal operating
little or no information on the project, and pump point.
selection may therefore not be optimal, or it may

75
 5 Pumping Stations

Factors affecting pump number selection may 5.6 Special Considerations

also be the requirement of even or continuous
output, which is easier to accomplish with a large
number of pumps. 5.6.1 Pump Vibrations
Most sewage pumps vibrate, at least to some
Unless special requirements are put forth, the extent. Vibrations are caused by residual mechan-
optimal number of pumps for most small to ical unbalance of the rotating parts, pressure pul-
medium size pumping stations is two. sations incited by the impeller vanes and the
hydrodynamic radial force caused by the fluid
In pumping stations with several pumps it is nor- mass rotating with single-vane impellers. Vortex
mally good practice to select identical pumps pumps vibrate much less, since they do not induce
only. In some special cases where the incoming pulsating pressures. For pumps with volute cas-
flow fluctuates randomly and to a great extent, ings the residual imbalance is negligible as com-
e.g. as a consequence of rain storms, it may be pared with the other vibration factors. Improving
sensible to install larger pumps that run only dur- already good balancing procedures by the manu-
ing peak situations. facturer does not have a measurable effect on
vibrations in the pump.
The effect on investment cost by variations in the
number of pumps installed for a given pumping Sewage pump impellers (except for vortex impel-
requirement may vary from one manufacturer to lers) induce higher rates of vibrations than impel-
another, since pump size increments are different lers for clean water, because of the small number
for different manufacturers. Thus increasing the of vanes and large channels. Pump installation
number of pumps may lead to a more favourable method also has an important impact on the
installation for one manufacturer and to a more vibration level. A submersible pump resting on a
expensive one for another. Where a multitude of baseplate stays in place by its own weight only,
pumps are required for the operation of a pump- which increases vibrations as compared to a fixed
ing station, the final number could be left open installation. A vertically installed dry pump may
for the bidding manufacturers to decide within vibrate more than a horizontally installed pump
given limits. Thus a larger number of bidders are because of a different support structure. The suc-
likely to be able to offer competitively. tion bend required may also enhance vibration
levels.

Vibration prediction and calculation information

for sewage pumps is available in the EuroPump
publication “Guide to Forecasting the Vibrations
of Centrifugal Pumps”, 1992 EuroPump. The values
presented in this book are guide lines and valid
when measured at the main bearing closest to
the pump impeller. Any vibration velocity above
10 mm/s (RMS) measured at this point indicates
an abnormal situation in the pump. The reason
can be clogging of pump, operation outside
allowed section of pump Q/H curve, severe cavita-
tion, and high content of air in liquid or damaged
impeller. Possible mechanical imbalance can be
controlled by running the pump out of water,
when the vibration reading should be less than 2
mm/s (RMS).

Vibration frequency for pumps with volute pump

casings equals rotational speed times number of
impeller vanes. If the pump or piping is supported
in such a way that the natural frequency of these

76
 Pumping Stations 5

elements is near the exciting frequency from the noise problems. In severe situations the piping
pump, the system resonance will increase vibra- and dry-installed pump motors can be clad with
tions. In these cases the support structures must sound-proofing insulation.
be stiffened. For a frequency controlled pump the
system may vibrate more at some frequency
because of resonance.

The pressure pulsation induced by the pump

moves ahead in the pressure piping with the liq-
uid flow several metres, causing vibrations in the
pipe wall. Normal piping vibration levels are
below 10 mm/s (RMS). Greater levels may lead to
pipe failure. The reason can be inadequate pipe
support or resonance.

Column-installed pumps have low vibration levels

because of trailing vanes in the casing, effectively
dampening the pressure pulsation. A single-vane
impeller may still cause pressure pulsations
because of the strong hydrodynamic forces
induced.

5.6.2 Pump Noise

Pumping station noise level is affected by the fol-
lowing elements:
• pump vibration noise
• piping vibration noise caused by pressure pul-
sation from the pump or other transmitted
vibrations
• flow in piping. Pipe bends, tee branches and
valves cause disturbances in the flow, emitting
noise
• pumping station acoustic characteristics
• inlet stream in wet well
• pump cavitation.

The sound level in the pumping stations com-

posed by all the above constituents, and sound
information on the pump alone is not very useful,
and cannot be accurately measured on site. Noise
emitted from the piping is usually decisive
because of the large vibrating emission surface.
Correct pump sound level measurement would
require the pump to be located in sound insulated
space, with the piping on the outside. There are
no standards available for allowed sewage pump
sound levels. Sound level measuring procedures
for submerged pump units is difficult to define
because of the practical difficulties involved.

Pumping station noise is not a common problem.

A pumping station built in connection with a resi-
dential or office building may in some cases cause

77
 6 Frequency-controlled Sewage Pumps

6 Frequency-controlled ered when designing the pumping station and

choosing the pumps.
Sewage Pumps
6.1.1 Pump Motor Selection
The supply current modulated by frequency con-
6.1 General trollers is not perfectly sine wave shaped, causing
the motor efficiency to decrease slightly. Nor-
The reasons for using frequency control are in
mally, however, considering pump duty point and
principle as follows:
operation conditions, the pump standard motor
• levelling of flow for process-related technical
can be used, provided that nominal supply fre-
reasons
quency (50 or 60 Hz) is not exceeded. The choice
• energy savings made possible by favourable
of motor should be confirmed with the pump
rising main characteristic curve. manufacturer, since he will have complete infor-
mation on the motor power and heating charac-
Process-related technical reasons for frequency teristics.
control are found in the following applications:
• return sludge pumping applications
• recirculation pumps in nitrogen reduction pro- 6.1.2 Maximum Frequency
cesses In installations with one frequency-controlled
• treatment plant incoming pumping stations. pump and one or more unregulated pumps in par-
allel operation, output control will be irregular at
Alternatively the incoming flow can be regulated the point where a fixed-speed pump is added to
by increasing the number of pumps in the final or taken off duty, unless the frequency-controlled
pumping station. With smarter pump control and pump is allowed to operate at higher than nomi-
increased starting frequency the output can be nal frequency at that point. Typically the required
levelled. Treatment plant basin and channel over frequency is 53 Hz for 50 Hz pumps. The situ-
design can also be used to level the flow in the ation is shown in Figure 96 below.
plant. In other pumping stations the use of fre-
quency control should be considered only if major
energy savings can be expected. Fig. 96
fN fmax fmin
Frequency control brings energy savings only if H
the rising main is long and the geodetic portion of
HJpipe
total head is less than 40%. In installations with a
high geodetic head the energy consumption is H2
bound to rise with frequency control, since the
pump duty point will move to a section of the
pump Q/H curve, where pump efficiency is lower. (fN)
Losses are experienced in the frequency control
unit and it lowers the pump motor efficiency.
Whenever a frequency controller is being consid-
(fmin)
ered for the objective of energy saving only, the
pay-back period of the investment in the control
system should be calculated separately. For this
the variations in flow, pump efficiency at different Q
Q1 Q2
frequencies must be known. The latter is also
dependent on the rising main characteristic curve. Parallel operation of speed-controlled pumps. Only
Also the efficiencies of the frequency controller one pump is speed-regulated at the time. Q1 = Nomi-
and the pump motor at various frequencies must nal output of one pump, Q2 = Regulated output of
be known. one pump at minimum speed. Regulated output of
one pump is extended to Q1 + Q2, preventing repeat-
Frequency control increases risk of pump clog- ed starting and stopping of pumps at full speed. fN =
ging. Should frequency control be selected, a Nominal frequency (50 or 60 Hz), fmax = Maximum
number of conditions and facts should be consid- frequency, fmin = Minimum frequency, η = efficiency.

78
 Frequency-controlled Sewage Pumps 6

If all pumps in a parallel installations are fre- If the rising main characteristic curve is steeply
quency-controlled, the situation with irregular rising, and several pumps may be running simul-
control will not arise, and regulation above nomi- taneously, it may be necessary to define several
nal frequency will not be necessary. minimum pump flow levels depending of number
of pumps in use at the same time. The pumps
If regulation above nominal is needed, this must must then be controlled by a suitable, program-
be stated on all inquiries, in order for the pump mable logic device. A minimum frequency deter-
manufacturer to be able to allow for it in the mined for the maximum number of pumps
pump and motor selection process. In some cases running simultaneously, and then used in all situ-
the use of a standard 60 Hz pump for a 50 Hz ations, with varying minimum performance as
installation with frequency control is favourable, result, depending on pumps in use.
but this solution should be weighed against fol-
lowing drawbacks: 6.1.4 Pump Frequency Curves
• If the frequency controller is out of order and
The pump Q/H curves for different frequencies
the pump can be run at nominal 50 Hz only,
are necessary in order to determine pump perfor-
the output of a 60 Hz pumps falls to 50…80%
mance at various speeds against a given rising
of that of a 50 Hz pump, depending on rising
main characteristic curve. Minimum frequency
main characteristics and friction losses.
must be determined and also pump efficiency at
• The motor is likely to have special windings
various frequencies. Pump curves for different fre-
because of the differences in voltages in stan-
quencies are easily drawn based on affinity rules
dard 60 Hz pumps as compared to standard 50
calculations, but possible limitations on curve
Hz pumps, which will impede future spare
usage can be determined by pump manufacturer
parts service.
alone. It makes sense to request frequency curves
• Pump efficiency may be lower, since some 60
from the manufacturer, with Q, H and η for differ-
Hz pumps are converted from 50 Hz by reduc-
ent frequencies at e.g. 5 Hz intervals.
ing impeller diameter.
Frequency curves can be calculated based on the
6.1.3 Minimum Frequency and Minimum affinity rules using the following equations:
Performance
Allowed minimum frequency for a specific pump f'
Q' = Q ---
is often inquired. A comprehensive answer to this f (40)
question requires information on the installation
and rising main, since the shifting of duty point 2
H' = H ⎛ ---⎞
f'
when reducing operating frequency is dependent ⎝ f⎠ (41)
on pump Q/H curve shape and rising main charac-
teristics. It is recommended that the minimum 3
P' = P ⎛ ---⎞
f'
frequency is worked out from required minimum ⎝ f⎠ (42)
flow with the help of pump Q/H curve and rising
main characteristics. η' = η
(43)
Pumping sewage at too low flow (too low fre-
quency) may lead to excessive pump wear from 2
NPSH R' = NPSH R ⎛ ---⎞
f'
sand or other abrasive matter remaining in the ⎝ f⎠ (44)
pump instead of being pumped out with the liq-
uid. Too small a flow may also lead to clogging of
the pump. If the geodetic portion of pump head is The above equations are valid simultaneously for
high (above 40%), the duty point will move to the the change of a given Q/H duty point and with an
left of the pump Q/H curve, where pump effi- accuracy acceptable for practical considerations.
ciency is lower, and energy costs may increase. As
a general rule, minimum performance can be lim-
ited to 25% of flow at best efficiency at nominal
frequency (Qopt).

79
 6 Frequency-controlled Sewage Pumps

6.1.5 Pump Clogging • Using an automatic valve in lieu of non-return

In frequency control operation the pump clogging valve, and programming it to remain open for
risk increases for the following reasons : some time at certain intervals after the pump
• Liquid level in pumping station stays station- has stopped, allowing the back flow to wash
ary from accurate pumping control, leading to out the pump.
accumulation of debris on the surface and/or
pit bottom. These may be larger than the free When the pump stops suddenly, the amount of
passage of the pump, blocking the impeller at water in the rising main continues to flow, effi-
pump-out. ciently flushing the freely rotating impeller, clear-
• The pumps run continuously for too long peri- ing possible beginning clogging.
ods, preventing the backwash at stopping
from clearing pump from debris accumula- The measures above can be accommodated in the
tions. pumping station planning stage but taken into
• The pumps stop slowly as controlled by the fre- use only if necessary.
quency controller, preventing the backwash
from clearing pump from debris accumula- The use of frequency control for return sludge and
tions. Smooth starting also prevents cleansing recirculation pumps does not increase the risk of
action. clogging. This is also the case for the final pump-
• Low speed in combination with rising main ing station before the sewage treatment plant, if
curves and losses may lead to complicated the sewage screening takes place prior to these
pump internal flow patterns, increasing clog- pumps.
ging susceptibility.
6.1.6 EMC Cable Requirement
Frequency control installations differ from each According to the EU Council Directive on Electro-
other and a comprehensive prediction of pump magnetic Compatibility (EMC), frequency control-
behaviour is impossible. The risk of clogging may ler manufacturers may require that pump motor
be reduced with the following actions: cables are replaced with EMC compatible and
• After pump start-up the frequency is con- approved cables. The requirement of EMC cables
trolled in a manner that the suction well level for submersible pumps may complicate manufac-
falls steadily and reaches pump stop level in an turing and add to costs. EMC cables also compli-
hour under normal conditions, after which cate pump handling, because these cables are less
pump is stopped. supple than regular supply cables. For submers-
• As above, in addition pump is programmed to ible pumps the choice of frequency controller may
run at nominal speed for 30 seconds before be dictated by requirement of EMC cables or not.
stopping at stop level. This increases flow in Use of EMC cables may be avoided with the use of
pump sump and rising main, flushing out pos- emission suppressing filters in the frequency con-
sible debris accumulations. troller.
• Pump is programmed for 1…2 flush-out
sequences per hour by increasing pump speed
6.1.7 Bearing Currents
to nominal for 20 seconds each, after which
pump is stopped without frequency control or In some instances frequency control causes inter-
the frequency is lowered as quickly as possible ference currents through the bearings of large,
to minimum frequency, where it is allowed to air-cooled squirrel-cage motors, causing bearing
remain for approximately 20 seconds, after damage. It is likely that submersible motors are
which normal operation is resumed. less susceptible to these currents, because they
• Pump is programmed to run in the wrong are well grounded by the pipework and submer-
direction for some time before each start. This gence in water, and thus protected. This assump-
will remove any beginning clogging remaining tion is supported by the Grundfos experience that
in the pump from the previous running inter- submersible motors have been free from bearing
val. The frequency when running the pump damages to date. Adding insulation to the bear-
backwards should be lower than nominal in ings would require extensive redesigning of the
order to avoid vibrations, i.e. 30 Hz for a 50 Hz motors and add to costs.
pump.

80
 Frequency-controlled Sewage Pumps 6

6.1.8 High Tension specified pump installation duty speed, and that
For supply voltages above 500 V, frequency con- test bed results are converted to duty point data
trol may cause too high voltage fluctuations for using the affinity law equations.
standard motors. In these cases the motors may
have to be redesigned with special winding insu- 6.1.11 Tests with Frequency Controller
lation and insulated bearings. The use of voltages (String Tests)
higher than 500 V are therefore discouraged in If the overall efficiency of the combination of
combination with frequency control. pump and frequency controller shall be verified,
the situation of the pump manufacturer is diffi-
6.1.9 Explosion-proof Motors cult. This situation requires exact information on
In frequency control the motors may operate at a frequency controller efficiency and pump motor
higher temperature than normally. Thus an explo- efficiency at modulated current, when voltage
sion proof certification of a motor at nominal fre- alternation is different from unmodulated sine
quency may be void for frequency control wave. These data are device-specific and almost
operation. The ex-proof certification of a pump impossible to get accurate information on before-
cable is likely not to be valid for an EMC cable. hand, and they must therefore be postulated. The
test standard also do not specify tolerance values
If an explosion-proof motor is intended for fre- for total efficiencies measured under these condi-
quency control, this must be clearly stated in the tions. The frequency controller must also be made
inquiry documents, in order for the manufacturer available in advance to the pump manufacturer
to be able to correctly assess pump and motor for testing, further complicating things and add-
suitability. Adding frequency control to an exist- ing costs. String tests are of little practical value.
ing installation also warrants contacting the man-
ufacturer for clearance. 6.1.12 Collaboration with the Pump Man-
ufacturer
6.1.10 Guaranteed Values Designing and executing an installation with fre-
The essential requirement of pump performance quency control of pumps is much more compli-
is that pump volume rate of flow matches the cated than simple fixed-speed pump installations.
specified demand and that pump energy costs are The close collaboration between pump manufac-
under control. turer and client is therefore important already at
the planning stage. Guaranteed duty points and
In order to secure total output, the guaranteed testing standard usage should also be agreed
duty point should be according to parallel opera- upon beforehand at contract negotiations when-
tion of pumps. If the rising main characteristic ever possible.
curve is flat (high Hgeod) or each pump has its own
individual rising main, the same duty point is also
suitable for pump efficiency guarantee evalua-
tion. On the other hand, if the rising main charac-
teristic curve is steep or if the geodetic head is
fluctuating, the determination of a rational guar-
antee point for pump efficiency gets difficult. The
guarantee duty point of η may be different from
that of volume rate of flow and head. It makes
sense to separately agree what duty point shall be
used for evaluation of pump efficiency. This point
could be the maximum efficiency point or the
intersection point of pump Q/H curve at nominal
frequency and the rising main characteristic
curve. A duty point at other frequency than nomi-
nal may also be chosen as guarantee point. It
should be noted that, according to the test stan-
dards, test pump speed may differ by ±20% from

81
 7 Pump Whole-life Cost Evaluation

7 Pump Whole-life Cost It should be noted that these methods have fairly
large error margins for energy and maintenance
Evaluation costs because these items are based on forecasts,
such as future pumped volume and wear rate
estimations.
The pump selection process should comprise a
pump life time cost evaluation, including the esti- The decision may also be based on reasons of
mation of all costs of acquiring, operating and principle or commercial grounds. Environmental
maintaining the pumping plant over its fore- aspects may stress energy use and costs. If the
casted life span. The importance of life cycle cost pumps are part of general contract and purchased
evaluations and comparative calculations by a contractor, purchase price alone may be deci-
increases with increasing pumping plant size. For sive.
instance, the energy costs of operating a mid-size
(30 kW) sewage pumps over three years are equiv-
alent to the original pump procurement costs. 7.2 Calculation Period
The useful life time of modern sewage pumps is in
the magnitude of 25 years. A pumping station
may become due for renovation much earlier, e.g.
7.1 General if changes in the neighbourhood lead to increased
pumping needs or zoning measures call for its
Pumping plant whole-life costs are needed for abolition or relocation. Also the unavailability of
project financial and investment feasibility calcu- spare parts may cause early pump obsolescence. A
lations. For instance, in a pumping station renova- suitable period for whole-life economic calcula-
tion project, where old pumps are replaced with tions is therefore 8…10 years from commission.
new ones, the chief investment assessment crite-
rion is life cycle cost evaluation. Correct long-term
calculations will have to take into account pre- 7.3 Investment Costs
dicted energy cost changes, inflation and interest
rates in addition to pump life cycle costs. These Pump purchase prices are obtained from the man-
calculations require financial and project man- ufacturers by inquiry or negotiations. Final price
agement skills in addition to solid pump knowl- may also include other commercial and purchas-
edge. ing costs, such as transportation. Also the effects
of different pump specifications on other acquisi-
Life cycle cost calculations are commonly used for tion costs must be considered. A larger motor
comparison of pumps during purchasing. The may, for instance, require a frequency converter or
alternatives to consider are either different makes mains supply fuse of higher rating, adding invest-
or different models from the same manufacturer. ment costs.
In these comparisons the financial elements nor-
mally have the same proportional magnitude for Figure 97 shows the proportional effect of pump
the various alternatives. As future changes in size on pump cost for pumps of 1500 1/min nomi-
energy costs and maintenance labour costs are nal speed. A pump with lower nominal speed will
difficult to forecast, it makes good sense to sim- generally be more expensive than a pump of the
plify the comparative calculations to comprise life same rating running at higher speed, because of
cycle costs calculations at present-day cost level, larger size. The figure shows that for small pumps
without financial analysis. Thus the analysis can of less than 10 kW rating, purchase price will be
be based on two different approaches: decisive for the life cycle costs.
• The whole-life costs are calculated for the dif-
ferent alternatives at present-day cost level,
and compared.
• A comparison based on the most inexpensive
alternative is performed, where the pay-back
periods for those alternatives with lower oper-
ating and maintenance costs are calculated.

82
 Pump Whole-life Cost Evaluation 7

Fig. 97

200

Price/PN
(%)

100

0
2 10 50 100 500
PN (kW)

Effect of pump size on pump specific cost Price/PN when pump nominal speed is 1500 1/min relative to 10 kW pump.

7.4 Energy Costs factors can be used when performing calcula-

tions:
The energy requirement is correctly calculated
using the efficiency (ηgr), since efficiency is sub- • Closed impeller with adjustable suction clear-
ject to the manufacturer’s guarantee according to ance: -1,5% (ηgr points)
the testing standards, whereas the information • Semi-open impeller with adjustable impeller
on power is not. It is important that the testing clearance: -3,0% (ηgr points)
standard to be used is agreed upon at this stage, • Closed impeller without adjustable clearance:
since different testing standards have different -3,0% (ηgr points)
tolerances for pump efficiency, which can affect • Semi-open impeller without adjustable impel-
the efficiency figures reported by the manufac- ler clearance: -5,0% (ηgr points)
turer. Please refer to Section 4, Testing of Pumps,
for more information. The efficiency reduction factors above suggest
that in practise sewage is pumped with pumps
7.4.1 Efficiency Over Time having significantly lower efficiency than new
With the exception of vortex pumps, the pumping pumps. The higher values for pumps without
efficiency of sewage pumps deteriorates over clearance adjustment possibilities are based on
time because the clearance between impeller and the fact that these pumps run longer intervals
suction cover widens from wear. This change between clearance restorations, because parts
should be taken into account when performing have to be replaced in shop conditions. The effect
energy usage calculations. Based on tests and is greater on pumps with semi-open impellers,
experience, the following efficiency reductions since these pumps wear faster and the efficiency

83
 7 Pump Whole-life Cost Evaluation

is more sensitive to changes in the clearance (see operating hours. If the pump Q/H curve passes
section 2.2.1 Impellers). above the desired duty point the pump output
will be higher and consequently, the operating
7.4.2 Energy Usage Calculations hours fewer. This must be considered when
using the pump operating hour method.
The energy calculation can be performed using
two different methods:
The energy calculation methods are fairly simple
when the pump is operated in a single duty point.
• Appraisal using yearly pumped water volume,
The situation gets more complicated with pumps
first computing specific energy, using the fol-
in parallel duty and if the pump is used with fre-
lowing equation:
quency converter. In parallel duty, calculations
should be performed separately for the different
g⋅H 3
E sp = -------------------------- [ kWh ⁄ m ] duty points, and then by approximating the
η gr ⋅ 3600 (45)
pumped volumes or operating hours accruing in
each of these.
where
With frequency converter the pump has an infi-
H = pump head at duty point [m],
nite number of duty points. A duty point, repre-
sentative of the average pump duty should be
g = 9,81 [m/s2], selected for the calculations in these cases.
Another uncertainty factor when calculating the
ηgr = overall efficiency (pump + motor) at duty energy consumption for frequency-controlled
point [decimal value], pumps is the fact that the overall efficiency of the
system is difficult to accurately determine. For
Liquid density is assumed to be 1000 kg/m3. comparative calculations the pump efficiency
without frequency controller may be used.
The energy consumption is calculated using
specific energy and estimated yearly pumped
volume.
7.5 Maintenance Costs
Normally submersible pumps are recommended
• Appraisal based on operating hours, first com- routine maintenance on a yearly basis. Mainte-
puting power at guaranteed efficiency, using nance includes seal oil control, motor insulation
the following equation: control with resistance meter, suction clearance
check and, if necessary, adjustment, and general
g⋅Q⋅H surface inspection. Most manufacturers recom-
P gr = ------------------------- [ kW ]
η gr ⋅ 1000 (46) mend very similar routines. Distinctions between
pumps from different manufacturers are most
where clear in the possibilities of maintaining and
restoring pump efficiency.
Q = pump volume flow at duty point [l/s],
If a pump is equipped with an adjustable suction
H = pump head at duty point [m], clearance mechanism, the costs for maintaining
pump efficiency does not add to costs, since the
g = 9,81 [m/s2], adjustment can be performed during normal rou-
tine maintenance on site. If, on the other hand,
ηgr = overall efficiency (pump + motor) at duty pump efficiency maintenance requires that spare
point [decimal value], parts be used or the pump brought to shop, the
costs for these measures will have to be taken
Liquid density is assumed to be 1000 kg/m3. into account when the whole-life costs of the
pump are calculated.
The energy consumption is calculated using
the power obtained and estimated yearly

84
 Pump Whole-life Cost Evaluation 7

7.6 Cooperation With Pump Sup-

pliers
Whole-life cost calculations and comparisons are
seldom completely unambiguous, and it is there-
fore reasonable and fair to perform these openly
and in cooperation with the pump vendors, at
least when considering a pump from that sup-
plier. This way possible misunderstandings can be
avoided, and suggestions and alternatives as pro-
posed by the supplier can be taken into account
for best possible selection.

7.7 Life Cycle Cost Publication

The pump manufacturers associations Europump
(Europe) and Hydraulic Institute (USA) have jointly
published a guide for pump life cycle cost (LCC)
evaluation:
PUMP LIFE CYCLE COSTS
A GUIDE TO LCC ANALYSIS
(ISBN 1-880952-58-0)

This publication deals with the complete pump-

ing system from the design stage, existing pump-
ing systems and examples of implemented
improvements.

85
 8 Commissioning

8 Commissioning
During pump commissioning, the following items
should be inspected:
• Check duty point(s) using pressure gauging
and possibly flow metering, using the volu-
metric method, for comparison of these with
projected values and to confirm that true duty
point lies within allowed limits of pump Q/H
curve.
In long rising mains with several high and low
points on the way to the discharge point, the
true situation may take some time to stabilize.
The measurements should therefore be
repeated after some time after commissioning
to confirm duty point.
• Check pump operation for vibrations and
noise. Check for signs of cavitation.
• Compare start and stop levels with projected
values and adjust if necessary. Lowest possible
stop level for dry-installed pumps should be
found by trials, observing suction of air into
inlet pipe.

For submersible pumps combined with large vol-

ume rates of flow, check for surface vortices at
low level. Adjust stop level where necessary.

86
 Operation and Service 9

9 Operation and Service 9.1 Safety

The most important risk factors associated with
Submersible sewage pumps should be subjected the operation of sewage pumps relate to the fol-
to routine inspection and service yearly. The lowing:
scheduled maintenance should be performed on • electricity
site and includes • lifting and handling of pumps
• Oil check and replacement when necessary. • hot surface temperatures of dry-installed
• Suction clearance (impeller/pump housing) pumps
inspection and adjustment, if the clearance • handling of pump parts during service and
has widened to 2 mm or more from wear. For repairs
pumps without adjustment possibility the res- • incidents of fire and explosions in hazardous
toration of the suction clearance and pump environments
performance requires installation of new • health risks from human contact with sewage.
parts.
• Motor insulation resistance metering at con- The following international standards deal with
trol panel. pump and pumping safety issues:
• Checking of lifting chain and lifting eyes and
lugs. • EN 809 (1998)
• General pump inspection and operation con- Pumps and pump units for liquids – common
trol. safety requirements
• prEN 13386 (1999)
The pump operator’s manual provides thorough Liquids pumps – submersible pumps and
information on maintenance. pump units – particular safety requirements
(proposal 2002)
Routine maintenance can be performed by the
owner or by contract service company. Impeller
replacement should be possible on site during
maintenance, if necessary. Shaft seal replacement
and other work on the watertight pump motor
enclosure should always be referred to an autho-
rized workshop.

Spare part availability is not a problem for sub-

mersible pumps made by recognized manufactur-
ers. The pump manufacturing series are long and
parts kept in stock both for assembly of new
pumps and for spare parts service. Pre-stocking of
spare parts is normally not warranted.

87
 10 Pumping Station Control and Condition Monitoring

10 Pumping Station 10.1.2 Relay-based Control Units

In case pump condition monitoring is not
Control and Condition required, automatic relay-based control units for
local control purposes can be used. Relay-based
Monitoring controls are simple units with fixed or adjustable
start and stop levels. They may include sequenc-
All sewage pumping stations, either working indi- ing of multiple pumps, or this may be accom-
vidually or as a part of a sewage network com- plished with additional pump sequencing units.
prised of several pumping stations, should be
reliably controlled in order to provide safe and In case continuous level measurement is used,
efficient operation. Modern electronic control these control units may have freely adjustable
technology offers possibilities to design and build start and stop levels and local level display. In
versatile control and condition monitoring sys- most cases, however, relay-based control units
tems to reduce long-term operation costs and to use preset or manually adjustable level switches,
increase operational reliability. such as float switches.

Unreliable sewage pumping stations represent an Relay-based control units are both easy to use and
ecological as well as economical risk in the form of reliable due to simplicity of design. They are suit-
waste water overflowing into the environment or able for small or secondary pumping stations
basements of buildings. Reliability is therefore the where little or no operational flexibility is
prime concern in design of a control unit for waste required.
water pumping station.
10.1.3 Programmable Logic Controllers
This chapter describes the sensors the reliable Pump control units based on programmable logic
pumping station control is based on, different controllers (PLCs) offer extensive possibilities for
control methods concentrating on the modern pump condition monitoring, data logging and
state-of-the-art control technology and finally a analysis as well as flexible pump control. Design-
network level remote control and monitoring sys- ing of a good pump control unit based on PLCs is
tem and its future possibilities combining Inter- demanding and always requires solid knowledge
net and WAP technology. of the operation and requirements of a sewage
pumping station, in addition to programming
10.1 Local Control Methods skills. Selection of control and measurement sig-
nals, pump and pumping station analysis and
Local control is always needed at site at the pump- choice of level measurement sensors are among
ing station to control the operation of the pumps. the things that has to be considered.
The local control unit can be built to different
technical levels according to requirements of con-
trol features as well as costs.

10.1.1 Manual Control Units

Manual control is the simplest control method. It
consists merely of a switch (normally manual-off-
auto) with the necessary relays and circuit break-
ers to start and stop the pumps. Manual control is
normally never used as the primary pump control,
but is used as a backup control method during
malfunctions of the regular controls and during
pump repair and maintenance work to check the
pump operation. Possibility for manual control
should always exist.

88
 Pumping Station Control and Condition Monitoring 10

10.2 Sensors for Pump Control The function of pressure transmitters is sensitive
to sedimentation, but this can be avoided by cor-
and Condition Monitoring rectly installing the transmitter inside a protective
pipe as shown in Figure 89.
The pump control unit, simple relay-based or ver-
satile PLC, requires different sensors to gather
information from the operation of the pumps as Fig. 89
well as the whole pumping station as shown in
Figure 88. Such sensors provide information on
wet well water level, current consumption of the
pump, condition of the pump primary seal as well
as motor coil insulation and so on.

Fig. 88

Correct installation of the level transmitter is essen-

tial. The completely sealed transmitter is suspended
Multiple sensors provide accurate information for
in the wet well and its piezo-resistive element trans-
the pump control and condition monitoring unit.
mits the level signal to the control unit. The sensor is
used with electronic pump control units and pro-
vides continuous level reading.
10.2.1 Wet Well Water Level Sensors
The basic information required by any automated
pump control system is the water level in the wet Ultrasonic devices are the only choice if the level
well of the pumping station. There are multiple sensors cannot be in contact with the liquid. Mod-
ways to provide this information as there are dif- ern ultrasonic sensors have built-in programma-
ferent types of information available as well. ble functions for various operating conditions and
Depending on the sensor, the current water level ranges. Ultrasonic sensors are also quite expen-
is given as continuous analog signal or on/off sive.
information as the water level passes certain, nor-
mally preset, height positions. Ultrasonic sensors are normally accurate and reli-
able. On the other hand, in waste water applica-
By experience, pressure transmitters offer the tions steam and foam on the liquid surface may
most reliable and economic way for continuously cause level indication errors or complete loss of
measuring the water level in sewage applications. echo, which can lead to interrupted level monitor-
Especially, a piezo-resistive pressure sensor, either ing. Problems arising from such situations can be
embedded in a stainless steel enclosure or inte- avoided with the installation of backup devices
grated in a sealed liquid-filled rubber construc- for the most vital functions, such as a float switch
tion, is excellent for use in waste water. Pressure for high level alarm.
transmitters provide continuous analog current
(0…20 mA or 4…20 mA) or voltage (0…45 mV) sig- Some earlier level sensors were based on capaci-
nal proportional to the water level. tive sensors. This kind of sensor is also installed in

89
 10 Pumping Station Control and Condition Monitoring

the wet well and consists of a rubber or plastic 10.2.3 kWh Meter
bladder connected to a vertical pipe and contain- Pumping stations with a modern electronic con-
ing a reference liquid. A wire is suspended in the troller should always be equipped with a kWh
pipe and connected to a signal transmitter. The meter, which has a potential free pulse output. As
level of the reference liquid in the pipe rises and the meter provides a certain amount of pulses for
falls with the level of the content in the wet well. each kilowatt hour used by the pumping station,
The capacitance of the wire-pipe element the energy consumption can be monitored.
changes accordingly and the signal is transformed
in the transmitter to a suitable signal for the
pump control unit. 10.2.4 Phase Failure Relay
All the three mains phases are connected to the
Capacitive devices are normally reliable but phase failure relay. This device provides an alarm
exposed to malfunctioning due to accumulation signal in case loss of power in the pumping sta-
of sediments on the bladder at the bottom of the tion.
wet well. The device also needs careful installa-
tion and more service than pressure transmitters. 10.2.5 SARI 2 Monitoring Device
The Grundfos SARI 2 is a combined monitoring
Float switches have been used for level control in device for motor insulation resistance and seal oil
waste water pumping applications for many water content. The motor insulation resistance is
years. They provide the simplest means for level measured between one of the mains phases and
control at fixed levels, but do not provide any pos- ground when the pump is stopped and discon-
sibility for continuous level control. In multiple nected from the mains supply. Low insulation
float switch installations there is always a risk for resistance indicates moisture inside the motor,
the control wires getting entangled in each other which could lead to the motor burning and expen-
or the pump cables. Float switches, together with sive repair work.
a relay-based control unit, are used today mainly
in small installations. In case the pump is equipped with a Grundfos
OCT 1 oil condition transmitter, the SARI 2 also
On the other hand, due to their simplicity and reli- continuously monitors the water content in the
ability float switches are even today quite often seal oil chamber. As the primary shaft seal wears
used as a backup or emergency level control sys- in time and water leaks into the oil chamber, the
tem in larger units, too. This usage provides emer- OCT 1 probe indicates the water content in the
gency operation in case of the main level seal oil. This information is routed to the SARI 2
measuring equipment failing. monitoring device, which raises an alarm.

A level bell is another very simple level sensing

device. It consists of an upside-down positioned Fig. 90
plastic cone with air tube between the upper end
of the cone and the controller. As the water level
reaches the cone, air inside the cone and the tube
is compressed causing pressure against a switch
in the controller. As the pressure rises, the switch
contacts and the pump starts. With this kind of
devices the pump is usually stopped after a pre-
set delay.

10.2.2 Current Sensor

The pump input current is monitored by a current
transformer through which one of the three
mains phase cables is routed. Each pump requires
one current transformer for adequate monitoring Grundfos SARI 2 monitoring device. The SARI 2 is
reliability. Current transformers provide an analog mounted on a DIN rail in the control panel. The SARI
signal (0…20 mA or 4…20 mA) proportional to the 2 is intended for both stand-alone alarm use and for
interfacing with a remote monitoring system.
pump input current.

90
 Pumping Station Control and Condition Monitoring 10

10.2.6 ASM 3 Alarm Status Module 10.3 Pump Control Units

All the Grundfos submersible pumps are
equipped with an internal moisture switch as well A modern and versatile pump control unit is
as temperature switches embedded in each based on the use of microprocessors and control
motor coil. These safety devices are connected is software. The unit is likely to be a PLC with a built-
series, and if any one of them trips, the controller in application software for pump control and con-
stops the pump and raises a safety device failure dition monitoring. The user interfaces with the
alarm. With the ASM 3 module these two alarms unit to access necessary control parameters, such
(moisture or overheat) can be separated to pro- as start and stop levels that can easily be checked
vide accurate alarm information. and adjusted. The complete control unit consists
of the electronic controller and an array of auxi-
liary equipment, such as level sensor, current
Fig. 91
transformers, and phase voltage relays, etc., form-
ing an integrated package. Figure 92 shows an
intelligent electronic pump control unit.

10.3.1. Control Features

The main parameter to measure is the water level
in the wet well. A continuous level indicator is
always used in this type of control unit. Several
types of sensors are available, such as a sealed
pressure transformer and ultrasonic devices.

The pump control sequence is normally quite sim-

ple. In a regular duty-standby application, the pre-
Grundfos ASM 3 alarm status module. set operation levels are stop level, start level and
second start level. The duty pump starts when the
water in the wet well reaches the start level, and
Fig. 92 stops when the water has been pumped down to
the stop level. The duty pump is alternated at
each cycle in order to ensure even distribution of
usage and wear between the pumps. The standby
pump starts at the second start level in a situation
where the incoming flow is larger than the capac-
ity of one pump. If more than one standby pump
is installed, these may be started at the same level
simultaneously or at adjustable intervals, or at
different levels.

All the running pumps are stopped simulta-

neously when the level reaches the stop level or at
adjustable intervals. In some multi-pump installa-
The GRUNDFOS PumpManager electronic pump tions all pumps may have different start and stop
control and condition monitoring unit. The unit levels. This, however, makes pump condition
controls all pumping station functions with soft- monitoring calculations more complicated and
ware stored in a PLC circuitry. Operational parame- less reliable.
ters are set using the keypad and the LCD display of
the unit. Pumping station data logged and calcu- In some cases a separate overflow pump with
lated by the unit are readable on the display by en- different characteristics may be installed to han-
tering codes on the keypad or from a remote dle large flows. This pump does not participate in
monitoring system. A scanning function allows the
the sequencing and should be controlled by a sep-
operator to get all important data with a minimum
arate unit independently from the other pumps.
of keystrokes.

91
 10 Pumping Station Control and Condition Monitoring

Flow measurement is possible without a separate below the stop level to allow pump snoring (con-
flow meter. It is done according to the volumetric trolled dry running) at given intervals. This is to
method, where the changes of levels in a wet well prevent sludge accumulation and cake formation
of known dimensions are measured against time. on the wet well surface. In this manner possible
The unit software calculates both incoming and pump blockages can be avoided with savings in
pumped flows with the same accuracy as that of a unscheduled maintenance costs. This has also
magnetic flow meter. The volumetric method is been found, by experience, to be an effective way
also the basis for the measurement of the pump to prevent odour problems.
capacity, which is continuously measured as a
running average of ten latest pump actions. Another possible advantageous operational fea-
ture is to let the pump starting level intentionally
In case of an overflow from a pumping station, it fluctuate around its setting; this is to prevent a
should be possible to estimate the volume accu- sludge rim from forming on the wet well wall at
rately and unambiguously in order for the opera- the start level.
tor to handle possible claims of damage. When
the incoming flow at the time and the duration of The pump control unit is also programmed to
overflow is known, the volume is estimated by the indicate all operation failures in the station, such
unit software for authority reporting. as alarms for high level, low level, pump power
failure, and other alarms based on settings of
Pump motor current measuring is necessary for parameter limits.
protection and condition monitoring purposes.
With adjustable over and under current limits, the 10.3.2 Condition Monitoring Features
unit is set to protect the pump motor in abnormal
The pump control unit performs automatic pump
situations. In case the input current rises above
condition monitoring based on the parameters
the over current limit, where a burnout of the
logged and analyzed. When the rate at which the
motor becomes a risk due to possible pump fail-
water level in the wet well rises and falls during
ure or clogging, the pump is automatically
the pumping cycle is monitored, pumping capac-
stopped. Together with embedded thermal relays
ity of each pump can be calculated. The unit then
or electronic motor protectors in the pump motor
compares these values to the pump nominal
starter this offers a very reliable motor protection.
performance data stored in the memory and
An abnormally low input current indicates that
raises an alarm in case the performance is outside
the pump is not pumping normally, which may be
set tolerance limits.
due to pump impeller wear or failure or gas
entrapment.
The benefit of such a system lies in its capability
to give early warning for slowly developing
Pump running hours and number of starts are
defects that ultimately could lead to sudden and
information needed for scheduling of pump
unexpected pump failure and consequent envi-
maintenance. These are also important informa-
ronmental damage. Also developing pressure pipe
tion in verifying the pumping station operational
work problems can be detected by closely analyz-
design and when determining the correct start
ing pump performance. Another benefit of such a
and stop levels during the commissioning.
system is the monitoring of the operation from
the economical point of view, where maintenance
All the features described above are available in
actions can be planned and executed according to
the GRUNDFOS PumpManager control and condi-
need. This finally leads the pump servicing from
tion monitoring unit, and readable from the inter-
repair-on-failure to preventive and even predictive
face display. This enables the motor control panel
maintenance.
to be simple without separate ammeters, hour
counters and sequencing relays, features that are
all incorporated within the PLC. 10.3.3 Parameters and Signals
The pump control unit needs a number of param-
With the control units entirely controlling the eters in order to operate as required. The parame-
pumps by software, it is simple to embed special ters are entered into the unit based on actual
features within the program. Thus it is possible, pumping station dimensions and units taken
for example, to let the pumping station pump from the plans or measured at site. For calibration

92
 Pumping Station Control and Condition Monitoring 10

either actual dimensions or percentages of refer- 10.3.5 User Interface

ence values can be used. Values to be entered are To access the data and to enter parameters, the
usually various operating levels such as pump operator needs to interface to the pump control
start and stop levels, low and high level alarm as unit. The interface must be at least a small LCD
well as overflow levels, which all correspond to display and a keypad. Using the keypad the user
certain water level in the wet well. Other parame- must be able to enter all necessary parameters
ters usually required are wet well dimensions and and read logged and calculated data. The use of
pump nominal values for the input current and such interface must be simple and logical.
capacity, which are available from the pump data
sheets. Normally some helpful features like an automatic
scanning function, makes the routine checking of
Several signals are necessary for the pump control the data easy and fast. Separate LED signal lights
to operate as planned. These are either digital or are used for indicating alarms and pump running
analog. Digital signals are either input or output status.
signals and indicate an ON or OFF status. Neces-
sary digital input signals are pump running or
standby indication from the circuit breakers as 10.4 Remote Control and
well as potential free contact signals from the Monitoring System
phase voltage relay and energy meter when avail-
able. Digital output signals are needed for start- Waste water pumping stations are designed to
ing and stopping the pumps. incorporate extra capacity in case of too large an
inflow or a pump failure. This reserve volume does
Analog input signals from additional sensors are not, however, prevent overflow in case the fault
used for various continuous measurements. These goes unnoticed for a longer period. Scheduled
signals are, for example, pump motor winding maintenance visits alone cannot prevent all con-
and bearing temperature measurements, pump tingencies possible in pumping stations; there-
sealing house oil condition information, data fore systems for remote control and condition
from an additional flow meter or frequency con- monitoring as well as for alarm relaying have
verter etc. The use of these signals may require an been developed.
additional extension card and special version of
the application software. The visible and audible alarms located outside of
the pumping station used in the earlier years have
10.3.4 Data Logging and Analysis been developed into sophisticated and decentral-
ized remote control systems. These latest systems
The pump control unit must have sufficient mem-
consist of PLC based pump control units control-
ory capacity to log and analyze data over a certain
ling and monitoring the local processes in the
period of time. The unit has to log at least the
pumping stations. These control units also func-
hours run, number of pump starts and incidents
tion as telemetry outstations and are remotely
of abnormal pump motor current. The unit has to
connected to a central computer where special
analyze and calculate flow, pump capacity and
network level administration software is running.
overflow from the logged data. The logged data
can be collected and further analyzed by down-
loading the data at intervals to a portable PC with 10.4.1 Different Levels for Remote Control
suitable software, or continuously by an auto- The modern pump control units enable the
matic remote control system. remote control and condition monitoring system
to be tailored according to the features required
Even in case the pump control unit operates as an by the customer in comparison to the available
outstation of a network level control and monitor- investment funds.
ing system, it must have memory capacity to store
the logged and analyzed data for several days. In case a very simple automatic alarm transfer is
This is due to the fact that vital data must not be preferred, the control unit can be equipped with a
lost even during possible communication break- GSM modem, with which the alarms generated by
down between the outstation and control center. the control unit will be transferred to the GSM
phone of the person on-duty as an SMS message.

93
 10 Pumping Station Control and Condition Monitoring

Fig. 93

The GRUNDFOS solution for the network level remote administration. The local control units at the pumping stations
are connected via normal telephone (PSTN), radio, GSM modem network or any combination of them to the central
control station. Also fixed (leased) line cable pairs can be used, but are nowadays getting rare.

Such a system offers highly increased operational 10.4.2 Software and Hardware
reliability with modest investment costs since The control center consists of a standard PC work-
there is no control center at all. On the other station, a printer for report printing and the spe-
hand, the modern control unit is capable of utiliz- cially designed administration software. The user
ing the whole SMS message by adding the most interface of the software must be mouse-con-
important logged and analyzed information into trolled and menu-driven for flexible and easy use.
the alarm message. Such information could be Depending on system configuration, theoretically
the pump running hours, number of starts, energy an unlimited number pump control units can be
consumption, pumping station inflow and pump controlled and monitored by a single control sta-
capacities, for example. In case such a system cre- tion. Practically the number is limited to about
ates automatic reports and transfers them to the 200 outstations by the time required to gather
person on duty on a weekly basis even without the observation data from the outstations during
any alarm situations, the normal driving around night time.
to the pumping stations can largely be avoided.
The central control station performs remote con-
In case a network level remote control and moni- trol and monitoring, by which real time condi-
toring system is introduced, there are several tions at the pumping stations can be viewed at
ways to build the communication link between any time. This feature largely substitutes for site
the outstations and the control center as visits by the operating personnel. Pumps can be
described in the following sections. started and stopped; levels and other parameters
can be changed, and so forth. The system gathers
all pumping station observations on a daily basis
and stores the data into the databases from

94
 Pumping Station Control and Condition Monitoring 10

which e.g. the pump and flow data for many years companies, to access a pumping station for spe-
backwards can be monitored for further analysis, cial purposes.
if desired. The software also generates numeric
and graphic reports on flows, pump data, alarms Modern GSM telecom technology offers an attrac-
and other parameters. Figure 93 shows a remote tive solution for remote control and monitoring
control arrangement. for far-off outstations with long distances. GSM
often offers the best alternative for retrofit instal-
10.4.3 Data Transmission lations, since PSTN line access installation after-
wards is expensive and availability may be
Although the pump control units operate com-
limited. It is a clear trend that GSM modem con-
pletely independent, the transmission of data is
nections are getting increasingly popular in the
crucial for the remote control systems to work.
future.
The time needed for data transfer can be
decreased, if the pump control unit performs all
data analysis locally and stores the results in its 10.4.4 Alarm Transfer
memory. Only the calculated results, instead of all Alarms raised at an outstation are transferred to
logged data, need to be transferred to the control the control center, where all incoming alarms are
center. This also enables the pump control units stored in the database. The administration soft-
to operate independently without having to be ware running on the control center computer
constantly connected to the control center. includes an automatic categorization of alarms as
well as a calendar of the service personnel on
The results can also be stored at the outstation for duty, according to which it transfers the alarm to
some period of time, usually one week, before it is the right person at the right time (in case the
automatically sent as a package. This is an impor- alarm is categorized to be transferred). Occasion-
tant feature in case there are indefinite break- ally, the control center computer is also equipped
downs in the communication link. with a separate alarm printer, which prints out all
the alarms for later analysis.
Data transmission is always configured to suit the
individual needs. The communication link must The alarms are normally transferred to the GSM
be flexible and normally the public switched tele- phone of the service person on duty as an SMS
phone network, radio modems, GSM modems or (text) message. The message may include, in addi-
any combination of them can be used. Also fixed tion to the alarm text and the station name, more
cable pairs can normally be used, but they have detailed information on the pump status (run-
lately got quite rare due to increasing monthly ning/off/failure), station inflow, pumped volume
fees and uncertain reliability. A modem is needed during the day, other active alarms (which are
in both ends of the communication link to modu- stated as not to be transferred), etc.
late the data for transfer. The choice between the
different transfer methods must be made by the Another way to transfer the alarms is by pager.
customer keeping in mind the building costs, data The control center computer creates the alarm
transmission costs and features both required and report text, contacts the pager operator and sends
offered by each method. the message to be displayed on the pager. Typi-
cally the message contains coded information on
In general, radio modems and fixed cable pair the station identity and type of alarm issued.
(leased lines) are used over short distances and
always in case there is a need for continuous com- If so required, the control center can also transfer
munication such as with control loops between alarms via voice message. A text corresponding to
fresh water reservoir and water intake station. If the alarm and stored vocally on the computer
connected through the public telephone network, hard disk, is retrieved by the software and used for
the pumping station and the central control sta- transmission over telephone to the operator. The
tion can be situated at a practically unlimited dis- control center can be programmed to call differ-
tance from each other. The public telephone ent numbers, until the alarm is acknowledged by
network also makes it possible to authorize third an operator.
parties, such as equipment vendors and service

95
 10 Pumping Station Control and Condition Monitoring

10.4.5 System Integration 10.5 Internet & WAP Based

A wastewater pumping control and monitoring Remote Control and Monitoring
system can be integrated into some other control
system, such as a treatment plant control system, Alarm messages transferred to the service per-
or an integrated water company control system, if sons as SMS messages are purely one-way infor-
combined control is preferred. mation. If the service person would have the
possibility to control the system and change some
Integration does not mean that all systems run in vital parameters from his mobile phone while
the same computer and with the same software. being on the field, the total flexibility by the
It is normally useful to pick the best system for means of a mobile control center can be achieved.
each application and to combine them at a suit-
able level. This could, for example, mean common The latest improvements in remote control and
software for alarm transfer and reporting. To monitoring techniques involve Internet and WAP
make integration possible, the systems should be technology to overcome the limitations of tradi-
designed using standard procedures such as PC tional monitoring systems described above. The
operating systems, standard data transmission Internet/WAP control and monitoring systems
and signal input and output protocols. also enable remote monitoring to be offered as
service to the municipalities. Figure 94 shows an
Internet or WAP based remote control arrange-
ment.

Fig. 94

The GRUNDFOS model for the Internet/WAP based control and monitoring system offers completely mobile control
center from the WAP cellular phone. In addition, the system utilizes Internet for data storage and the customers can
monitor the outstations and create reports from any computer with Internet access. The system also enables the whole
remote control and monitoring to be offered as a contract service.

96
 Pumping Station Control and Condition Monitoring 10

Fig. 95

From the Internet the customers can browse the history data of their own outstations independent on the location. In-
formation can also easily be shared inside the organisation – operation personnel, duty persons, decision-makers as
well as sewage system designers and engineers.

The Internet based control and monitoring system

allows historical data from the outstations to be
viewed and reported from multiple locations, thus
enabling the information to be used wherever
needed. After typing in the user identification the
operation personnel, persons on duty, decision-
makers, sewage system engineers, etc. are able to
browse detailed historical data from the outsta-
tions for years backwards e.g. from their own
office computers. Figure 95 shows the interface
page of the Grundfos Web based control service.

97
 Symbols

Symbols MB concrete mass, kg

MH head measurement uncertainty
MQ flow rate measurement uncertainty
A area m mass
AW wet well surface area NPSH net positive suction head
a pressure wave velocity n rotational speed
D pipe internal diameter nN nominal rotational speed
Fa axial force P pump power input
Fr radial force Pgr motor power input
fH head measurement uncertainty factor p pressure
fmax maximum frequency pb ambient pressure at liquid level
fmin minimum frequency pL atmospheric pressure in pump well
fN nominal frequency pm sand content by weight
fQ flow rate measurement uncertainty pmin minimum static pressure in pump
factor pU atmospheric pressure in receiving well
g acceleration of gravity, 9,81 m/s² pv liquid vapor pressure, sand content
H pump total head (head)
H0 head at zero flow rate by volume
Q volume rate of flow
Hd dynamic head Q0 volume rate of flow at zero head
Hf head friction losses QI volume rate of flow, one pump
HG guaranteed head QII volume rate of flow, two pumps
Hgeod geodetic head Qin incoming flow
HJ head loss in pipeline QN pump nominal flow rate
HJn local head loss q volume flow
HJp head loss in pressure pipeline Re Reynold’s number
HJt head loss in suction pipeline S curve slope
T cycle duration
Hmax maximum allowable head
t pump running time
Hmin minimum allowable head tH head tolerance factor
Hr head loss tQ flow rate tolerance factor
Hrt head losses in inlet pipe Ul line voltage
Hs head discontinuity losses Uph phase voltage
Hst static head u perimeter velocity
Ht theoretical head V volume
Ht∞ ideal head VG pumping station volume below
Hv head leakage losses water table, m3
h height VH effective wet well volume
hA height difference between reference Vh effective wet well volume
plane and tip of vane leading edge v true fluid velocity
hs pump stop level v2 flow velocity at pump outlet
ht inlet geodetic height vm radial component of true velocity
I electric current vu tangential component of true velocity
Il line current w velocity relative to the vane
Iph phase current XH head tolerance factor
k coefficient XQ flow rate tolerance factor
L length Z starting frequency
l length

98
 Symbols

Z1.2 pressure gage height above water level

ZImax maximum pump starting frequency
ZIImax maximum pump starting frequency
Zmax maximum pump starting frequency
β vane edge angle
∆h local pressure drop at vane leading edge
∆h pressure (head) change
∆HL lower allowable head deviation
∆HT allowable head deviation
∆HU upper allowable head deviation
∆QL lower allowable flow rate deviation
∆QT allowable flow rate deviation
∆QU upper allowable flow rate deviation
∆v flow velocity change
ζ local resistance factor
η pump efficiency
ηgr overall efficiency
ηh hydraulic efficiency
ηmot motor efficiency
λ friction factor
µ reflection cycle duration
ν kinematic viscosity
ρ fluid density

99
 Appendix A

Local Resistance Factors

Branches

Q h vh Diverging flows

Qh/Q α = 90° α = 45°

ζh ζs ζh ζs

0,0 0,95 0,04 0,90 0,04

Qs Q
0,2 0,88 -0,08 0,68 -0,06
vs v
0,4 0,89 -0,05 0,50 -0,04
0,6 0,95 0,07 0,38 0,07

0,8 1,10 0,21 0,35 0,20

1,0 1,28 0,35 0,48 0,33

Qh vh Merging flows

Qh/Q α = 90° α = 45°

ζh ζs ζh ζs

0,0 -1,00 0,04 -0,90 0,04

Q Qs 0,2 -0,40 0,17 -0,38 0,17
v vs 0,4 0,08 0,30 0,00 0,19

0,6 0,47 0,41 0,22 0,09

0,8 0,72 0,51 0,37 -0,17

1,0 0,91 0,60 0,37 -0,54

101
 Appendix A

Qh/Q Merging flows

Qh
vh ζh ζs

45 0,0 -0,82 0,06

0,2 -0,30 0,24

Qs Q 0,4 0,17 0,41
vs v
0,6 0,60 0,56

0,8 1,04 0,80

1,0 1,38 1,13

Qh/Q Diverging flows

Qh
vh ζh ζs

45 0,0 0,92 0,06

0,2 0,97 -0,06

Qs Q 0,4 1,12 0,00

vs v 0,6 1,31 0,09

0,8 1,50 0,20

1,0 0,30

102
 Appendix A

Bends

D
D

R R

D
D

R ⁄ D = 1, 5 ;ζ = 0, 4
R ⁄ D = 1, 5 ;ζ = 0, 7

D
R/D 1 2 3 4 6

ζ 0,36 0,19 0,16 0,15 0,21

R R/D 8 10 12 16 20
ζ 0,27 0,32 0,35 0,39 0,41
90

103
 Appendix A

ζ
D α R/D

1 2 4

R 20° 0,07 0,03 0,03

40° 0,13 0,06 0,06

60° 0,20 0,10 0,09

80° 0,27 0,13 0,12

90° 0,32 0,15 0,13

120° 0,39 0,19 0,17

140° 0,46 0,23 0,20

160° 0,52 0,26 0,23

180° 0,60 0,30 0,26

α 20° 40° 50° 70° 80°

ζ 0,03 0,12 0,24 0,54 0,74

α 90° 120° 140° 180°

ζ 1,00 1,86 2,43 3,00

104
 Appendix A

Expansions and Contractions

v1 v2

2
( v1 – v2 )
H Jn = -----------------------
2g

A2 β° k β° k β° k
A1 5 0,13 45 0,93 100 1,06
v1 v2
10 0,17 50 1,05 120 1,05
15 0,26 60 1,12 140 1,04

20 0,41 70 1,13 160 1,02

2
30 0,71 80 1,10
v A1 2
H Jn = ζ -----1- ζ = k ⎛ 1 – ------⎞
2g ⎝ A 2⎠ 40 0,90 90 1,07

H Jn ≈ 0

Friction drag not included

105
 Appendix A

A1

A2 A2/A1 0 0,1 0,2 0,3 0,4

v1 v2 ζ2 0,50 0,46 0,41 0,36 0,30

A2/A1 0,5 0,6 0,7 0,8 0,9

ζ2 0,24 0,18 0,12 0,06 0,02

2
v
H Jn = ζ -----2-
2g

v2 v1
v1<<v2 v2<<v1

2
v 2
H Jn = 0, 5 -----2- v
H Jn = -----1-
2g 2g

106
 Appendix A

Bend Combinations Suction Inlets

ζ = 2 × ζ 90°
ζ = 3, 0

ζ = 0, 2
ζ = 3 × ζ 90°

ζ = 4 × ζ 90° ζ = 0, 05

107
 Appendix A

Valves
ξ-values depend strongly on shape. Factory values should be used when available.

Gate valves without narrowing: ξ = 0,1…0,3

Gate valves with narrowing: ξ = 0,3…1,2

Ball non-return valves ξ ≈ 1,0 (fully open)

Flap non-return valves ξ = 0,5…1,0 (fully open)

ξ-values above are valid for fully open valves. In partly open position, ξ may be 1,5-2 times as high. Depend-
ing on shape and position, a certain minimum flow velocity through the valve is required for it to be regarded
as fully open. Exact information on each valve is available from the manufacturer or supplier.

108
 Appendix B

Pumping Station Starting Substituting with the expression B3 for t in equa-

tion B4:
Frequency and Pumping
Q in T
Capacity V h = Q in T – Q in -----------
Q
- (B5)

Solving equation B5 for T is obtained:

In a pumping station the water volume comprises
the volume below the lowest pump stop level and
Vh Q
the pumpable volume above this level, fluctuat- T = ---------------------------
-
2 (B6)
ing with pump usage and water incoming flow Q in Q – Q in
rate. The starting frequency of the pumps
depends on the available pumpable volume and
Starting frequency is the inverse value of T, hence:
the incoming flow rate.
2
The following different cases are investigated: Q in Q – Q in
Z = ---------------------------
- (B7)
• single pump pumping station Vh Q
• pumping station with two pumps in duty-
standby operation
The starting frequency Z is a function of the ratio
• pumping station with more than two pumps.
Qin/Q and is shown in Figure B1.

Single Pump Station

Fig. B1
Incoming water during one unit of time (cycle)
Zmax
can be expressed as:
Z [%]

V = Q in ⋅ T (B1)

where
Qin = incoming flow rate
T = duration of cycle

The same volume must be removed by the pump Qin/Q

during the cycle, whence
Starting frequency curve Z for a single pump pump-
V = Q⋅t (B2) ing station as a function of the ratio between incom-
ing flow rate Qin and pump capacity Q.
where
Q = pump capacity
t = pump running time Differentiating the equation B7 over Qin is
obtained:
Combining equations B1 and B2 is obtained
dZ Q – 2Q in
------------ = ---------------------
- (B8)
Q in T dQ in Vh Q
t = -----------
- (B3)
Q
Equation B8 equals 0 when Qin = ½Q
When the pump is stopped, the volume between
the start and stop levels Vh fills up during the time Substituting Qin = ½Q to equation B7:
T - t, whence
Q
Z max = --------------
V h = Q in ⋅ ( T – t ) = Q in ⋅ T – Q in ⋅ t (B4) 4 ⋅ Vh (B9)

109
 Appendix B

From this the pumping station capacity Vh is

Fig. B3
obtained:

Q
V h = --------------------- Ql or Qll
4 ⋅ Z max (B10)

The solution to equation B10 is shown graphically

Qin
in Figure B2.
Start level 2
In practice there may be situations where the
Start level 1
incoming flow to a pumping station is very small
and only momentary, for instance in pumping sta- H
tions serving a few households only. In such cases h
Stop level
the selected pump capacity should be selected
much larger, in order to attain high enough a flow
velocity in the rising main to prevent sedimenta- A B
tion. In this situation the Qin/Q ratio remains
small, and the Zmax value is not reached at all or
very seldom only.
Pumping station with two pumps in alternating
duty. The lead pump starts when the water level
rises to start level 1. If the incoming flow exceeds
Fig. B2 the capacity of one pump, the lag pump will start
at start level 2. Pumps alternate between lead and
lag positions with each running cycle.

1000

Vh [m3] Fig. B4

100 T
t Qin /Ql < 1

A A
10 B t
t

1 T
t1 Qin /Ql > 1

A A
0,1 B t
2 10 100 1000 2000 B
Q [l/s] t2
Q = Pump capacity, l/s
Zmax = Maximum starting frequency, 1/h
Vh = Effective wet well volume, m3
Operation time diagram of the duty and standby
pumps in a pumping station for an incoming flow
Diagram for the determination of the effective wet (Qin) both smaller and larger than the capacity of
well volume Vh for a single pump pumping station. one pump (QI).

110
 Appendix B

Two Pumps in Duty-Standby The duty pump alone is able to handle most regu-
Configuration lar incoming flow situations, and the standby
pump will start only if the incoming flow rate
The principle of operating a pumping station with
(Qin) is larger than the capacity of one pump (QI),
two identical pumps is shown in Figure B3. The
pumps assume alternately the positions of duty in which case the water level continues to rise to
(lead) and standby (lag) pump with each running the second start level, starting the standby pump.
cycle. When the water level in the wet well If the combined capacity of two pumps (QII) is
reaches the first start level the duty pump starts. larger than the incoming flow, all pumps stop
The water level is pumped down to the stop level, when the water reaches the stop level.
and the pump stops, allowing the water level to
rise again to the first start level, completing the Figure B4 shows a time diagram of the running
cycle. cycle of two pumps in alternating duty, further
explaining the principle.

Fig. B5

Vh
(m3)

Ql
Qll

Ql (l/s)
Qin/Ql <1 Qin/Ql >1

Qin = Incoming flow rate, l/s

Ql = Pumping capacity when Qin/Ql <1, l/s
Qll = Pumping capacity two pumps when Qin/Ql >1, l/s
Zlmax = Maximum pump starting frequency for Qin<Ql, 1/h
Zlmax = Maximum pump starting frequency for Qin>Ql, 1/h
Vh = Effective wet well volume to start level 1, m3
VH = Effective wet well volume to starting level 2, m3

Nomogram for the determination of the effective wet well volume Vh and the starting frequency Z for a pumping
station with two pumps in duty-standby configuration.

111
 Appendix B

Q in < Q l Fig. B6

Equations B9 and B10 can be used in the situation Zllmax

where the incoming flow is smaller than the
capacity of one pump for the calculation of start-
ing frequency for each pump. With two pumps Z
starting alternately, the expressions are divided Zlmax
by two, whence

Ql
Z lmax = -------------
-
8 ⋅ Vh (B11)
Qin/Ql
Ql Qll
V h = ---------------------
-
8 ⋅ Z lmax (B12) Qin/Ql < 1 Qin/Ql > 1

The solution to equations B11 and B12 are shown Starting frequency curve Z for one pump and two
graphically in Figure B5. pumps in pumping station with two pumps in duty-
standby configuration as function of the ratio be-
Q in > Q l tween incoming flow rate Qin and pump capacity QI.

In the case when the incoming flow is larger than

the capacity of one pump, two additional factors
must be considered. These are the ratio of the The diagram in Figure B7 shows the effect of the
pumping station capacity to the first start level, ratio Vh/VH on Z for constant VH and varying Vh. In
Vh, and the second start level, VH, and the com- this case the ratio QII/QI is 1,6. The conclusion
bined capacity of the pumps QII. The following from Figure B7 is, that ZIImax is reduced and ZImax
equation for the starting frequency can then be increased with lower start level 1.
derived:
The diagram in Figure B8 shows the effect of the
Ql ( VH – Vh ) Q ll V H –1
ratio QII/QI on Z for a constant Vh/VH ratio of 0,8.
Z ll = ------------------------------
2
+ ------------------------------
2
- (B13)
Q in – Q l Q in Q in Q ll – Q in Increasing rising main losses, decreasing QII/QI
also decreases ZIImax.
The expression for ZIImax can be solved by differ-
entiation, but the expression is very complex. A If the pumps are selected so that one pump can
graphic presentation of the solution is presented handle all incoming flows, ZIImax loses signifi-
in Figure B5. cance.

Figure B6 shows the relation between starting fre- Pumping Stations with more than Two
quency and the Qin/QI ratio. The starting fre- Pumps
quency rises sharply at conditions requiring Pumping stations with a multitude of pumps can
parallel duty. The diagram shows a marked peak be divided into the following two design catego-
value ZIImax. ries:
• Stations with common stop level for all pumps
• Stations with different or stepped stop levels
for each pump
The starting cycle of the pumps are normally
alternated between the pumps in order to ensure
even distribution of wear.

112
 Appendix B

Fig. B7 Pumping Station Capacities and Starting

Frequencies
Vh/VH = 1,0
With several pumps installed in a pumping sta-
Vh/VH = 0,4 tion, the starting frequency changes dramatically
with variations in the incoming flow. The starting
Z frequency will vary between zero and peak values,
0,8
0,6 of which there are several.
0,6
0,8 0,4 Great flow fluctuations are typical for sewage
1,0 pumping, and it becomes impossible and also
rather unnecessary to numerically calculate start-
ing frequencies for each pump. With the aid of
design nomograms total pumping capacities and
0 1,0 1,6 average starting frequencies, on which pumping
Qin/Ql station further design in all practical cases can be
based, can be determined.
Starting frequency curves for different Vh/VH ratios
with constant VH and a QII/QI ratio of 1,6. For the different design categories the following
nomograms can be used.

Fig. B8 Common Stop Level

Figure B9 shows a diagram from which VH or Z
can be selected as functions of overall flow rate
Qoverall.
Z Qll/Q=
For both of these categories it is good practice to
divide the total pumping volume (VH) by the start-
2,0
1,8 ing levels at approximately equal intervals if all
1,6
1,4 the pumps are identical. If the pumps have differ-
1,2 ent capacities, the pumping volume may be
divided into intervals proportional to the pump
0 1,0 Qin/Ql capacities. The use of modern electronic level con-
trol equipment facilitates the optimization of the
Starting frequency curves for different QII/QI ratios start levels either manually or automatically.
and a Vh/VH ratio of 0,8.

113
 Appendix B

Fig. B9

VH [m3] Pump Pump Pump Pump

1 2 3 4
Start

V4

Start

V3

Start

V2

Start

V1

Stop

For similar pumps:

V1 V2 V3 etc.

Qoverall [l/s]
Qoverall = Flow rate, l/s
Z = Approx. average starting frequency of pump, 1/h
VH = Pumping volume to top level, m3

Starting frequency nomogram for pumping station with more than two pumps and common stop level.

Stepped Stop Levels Recommended Starting Frequencies

Figure B10 shows a diagram from which VH or Z Pump and control equipment operation and wear
can be selected as functions of overall flow rate is significantly related to the number of starts and
Qoverall. stops over the long period, such as a year, since
very high starting frequencies can be allowed in
the short term. If peak starting frequencies are
used for dimensioning, the occurrence of these
must be investigated. As shown earlier, the peak
starting frequency for one pump ZImax may never
be attained in reality. Likewise, the peak starting
frequency for two pumps in parallel operation
ZIImax is usually much higher (1,5...2 times) than
the ZImax value and only occurs occasionally.

114
 Appendix B

Starting frequencies selection should be checked

against pump and control equipment manufac- Pump power Allowable Z
turers' recommendations. The following guide-
0...5 kW 25 1/h
lines for mean allowable starting frequencies for
submersible pumps may be used: 5...20 kW 20 1/h

20...100 kW 15 1/h
100...400 kW 10 1/h

Fig. B10

VH [m3] Pump Pump Pump Pump

1 2 3 4
Start

V4

Start Stop

V3

Start Stop
VH

V2

Start Stop

V1

Stop

For similar pumps:

V1 V2 V3 etc.

Qoverall [l/s]
Qoverall = Flow rate, l/s
Z = Approx. average starting frequency of pump, 1/h
VH = Pumping volume to top level, m3

Starting frequency nomogram for a pumping station with more than two pumps and stepped stop levels.

115