UNRESTRICTED

STUDY ON IMPROVING THE

ENERGY EFFICIENCY OF
PUMPS

Contractor: ETSU, AEAT PLC, (United Kingdom)

Participants: CETIM (France), David T. Reeves (United
Kingdom), NESA (Denmark), Technical University
Darmstadt (Germany)
European Commission
February 2001
AEAT-6559/ v 5.1 UNRESTRICTED

ETSU
ETSU is a large UK-based Energy and Environmental consultancy organisation working in
diverse environmental activities including renewable energy, energy efficiency and sustainable
development. ETSU is now part of AEA Technology PLC, and undertakes a broad range of
work both in the UK and Overseas for a variety of private and public sector clients.

The main task of CETIM is to provide mechanical engineering companies with the necessary
technological support to increase its competitiveness. To do this, it is active in the
dissemination of relevant information, undertakes collective research, works for individual
companies, and undertakes work to defend the interests and influence of the French
mechanical engineering industry.

TFA
Darmstadt University of Technology
The chair of Turbomachinery and Fluid Power (TFA) is an educational institution within
the faculty of mechanical engineering of Darmstadt University of Technology
(Germany). The main research activities are aimed at experimental and theoretical
investigations on turbomachinery and fluid powered systems, wehereas the topic
„efficiency potential of centrifugal pumps“ is a major key aspect and the matter of
permanent investigations within the frame of various research projects.

David T. Reeves Pump Engineering Consultant

David Reeves has been a full time independent consultant in pumps and
pumping for fifteen years, advising makers, users, contractors and
consultants on problems, procedures, performance and design. Previously, he
was Director - Fluid Machinery Design for Worthington/Dresser USA, Chief
Hydraulic Engineer for Weir Pumps and worked with three other major pump
companies. He is immediate past Chairman of the Fluid Machinery Committee
of the IMechE.

ii
AEAT-6559/ v 5.1 UNRESTRICTED

NESA is the largest Danish electricity company active in the complex new electricity market,
and is the first Danish electricity distribution company to be a direct participant in the Nordic
electricity exchange, Nord Pool. As a leading electricity company, NESA offers customer
driven products as energy savings pro-ducts and procurement of energy efficient equipment,
and offers energy management to customers.

Acknowledgements

The study group would like to thank the many individuals and organisations who helped with
this study, in particular Guy van Doorslaer, Erich Holzhuter, Steven Schofield of Europump,
and the many other pump manufacturers who took an active interest in the work or who
attended progress meetings. In addition we would like to thank Bob Went of Thames Water
(chairman of UK Pump Centre), Anibal de Almeida of University of Coimbra (Portugal),
Paolo Bertoldi of DG TREN, and the EU/US Enersave Life Cycle Costs project team, all of
whom made particularly valuable contributions to the project.

iii
AEAT-6559/ v 5.1 UNRESTRICTED

Executive Summary
Energy consumption of Pumps

Pumps are the single largest user of electricity in Industry in the European Union, consuming
160 TWhpa of electricity, accounting for 79 Mton CO2. This report recommends policy
options to improve the selection and maintenance of pumps to reduce these emissions. It is
estimated that the better selection of pumps could give cost effective energy savings of 3% or
1.1TWHpa in 2015. This give cumulative saving of 8.5TWhpa by 2015.

The group estimates that further reductions in emissions could be made through improved
system design, and so also supports actions to reduce these system losses. In particular, the
Enersave Life Cycle Costing Guide is commended.

A Procurement scheme for pumps

It is recommended that a User-driven Procurement level scheme is piloted for the ranges of
pumps concentrated on in this study. This will allow users to quickly see the spread of
available efficiencies of pumps to meet their specified duty. It will therefore give a clear basis
on which to assess different pumps offered.

In addition, it is recommended that a “Mean Line” is drawn through this data. This would at
last give specifiers and manufacturers alike a practical guide for assessing the efficiency of
procured pumps.

To make this study in to something achievable with the effort and time available, it
concentrates on single stage centrifugal pumps for clean, cold water pumping duty only, but
the methodology is applicable to other types of pumps as well. `

Other Key Points and Recommendations:

1.) Efforts should continue to be made by manufacturers, Government Agencies and

others to raise the general level of knowledge about pumps and pumping systems.
This is important whichever other policy options are implemented.

2.) It it is recognised that the largest energy savings are to be made through the better
design and control of pump systems. But unfortunately the lack of expertise and
available time to identify best solutions means that promoting the use of more efficient
pumps is very valuable in itself.

It is however also recognised that giving simple guidance on the selection of pumps by
efficiency, does present particular challenges, in particular that:
· Efficiency may fall off fast as operation moves from the Best Efficiency Point
· Pump Efficiency will deteriorate over time

iv
AEAT-6559/ v 5.1 UNRESTRICTED

· Pump operation and hence efficiency is determined by principally two parameters

(head and flow), in addition to other factors such as properties of the fluid, impeller
design, and motor speed selected. This compares unfavourably with labelling schemes
for products such as lights, white goods or electric motors where definition of typical
operation and hence rated efficiency is much simpler.
· Above all, it is crucial that any scheme to aid in the selection of better efficiency
pumps ensures that users are always comparing “like with like.”

It is recognised that simple methods of identifying efficiency are not needed by the
informed buyer, (usually those from larger companies in for example the Water and
Chemicals sectors), where there is anyway a greater awareness of efficiency and
Lifecycle costing.

3.) There is a concern over the effect of tolerances on the published efficiencies of mass-
produced pumps allowed under existing ISO test codes. It is hoped that this will be
remedied. Such changes should not though be allowed to slow down the
implementation of any of the suggested policy options.

4.) Full data on pump characteristics is essential for the informed selection and use of
pumps. It can be difficult to obtain the detailed information required, and so a
database of pumps from different manufacturers would be attractive. However, such a
database would be very large, and so demand large amounts of time and effort to
design and maintain. To check on how useful such a database would really be, it is
suggested that a pilot version with data on a limited range of pumps is included within
EURODEEM and sent out for comment. Also, a close watching brief should be
maintained on the commercially–funded BigMachines.com on-line pump selection and
purchase web-site.

5.) It was clear from discussions with manufacturers that there is not a direct relationship
between price and efficiency. Users will therefore often be in a position to buy a more
efficient pump at little or no price premium.

6.) The energy savings identified are based on manufacturers producing pumps with
efficiency ratings among the current “best in class”. Energy savings beyond these
values are possible using new manufacturing and design techniques. But for the
fastest impact on energy use, the emphasis should be on existing technologies rather
than the possible distraction and longer time to market of newer technologies.

7.) Europump has throughout made constructive comment on the work of the study group,
and given support in many other ways. It is important that this relationship is
developed to future mutual benefit.

8. Support should be given for a proposed simple labelling system for smaller water
circulation pumps. Such types have very similar general characteristics and similar
duties, and a labelling scheme is appropriate.

v
AEAT-6559/ v 5.1 UNRESTRICTED

Contents

1. Policy Options For Improving The Efficiency Of

Pumps 9
1.1 SUMMARY 9
1.2 CONSIDERATIONS IN RANKING PUMPS BY EFFICIENCY 9
1.3 DEVISING A CLASSIFICATION SCHEME 12
1.4 THE RECOMMENDED PROCUREMENT LEVEL APPROACH 13

2. Efficiency Levels Of Pumps 15

2.1 OBJECTIVES 15
2.2 PUMP TYPES AND DESIGN DUTIES 15
2.3 DATA COLLECTION AND ANALYSIS 17
2.4 PROCEDURE FOR ANALYSIS 18
2.5 EFFECT OF SPECIFIC SPEED 19
2.6 EFFICIENCY REFERENCE LEVELS 23
2.7 EFFICIENCY DETERIORATION IN USE 24
2.8 REFERENCES 25

3. Characterisation of Pump Use in the EC 26

3.1 INTRODUCTION 26
3.2 CENTRIFUGAL PUMP USE 26
3.3 METHOD 1 - EU PUMP STOCK BASED ON MOTOR ENERGY
CONSUMPTION 26
3.4 METHOD 2 – ENERGY CONSUMPTION BASED ON ANNUAL SALES
FIGURES 27
3.5 FUTURE PUMPSET USE WITH POTENTIAL SAVINGS 29
3.6 SUMMARY 30

4. User Attitudes to Pump Selection and Ways in

Which They Can Be Changed 31
4.1 ENERGY EFFICIENCY IS NOT A PRIORITY 31
4.2 BARRIERS TO IMPROVING THE EFFICIENCY OF PUMPS 32
4.3 EDUCATIONAL NEEDS 33
4.4 WHAT USERS WANT TO HELP THEM IMPROVE THEIR SELECTION OF
PUMPS 34

5. Design Factors Affecting Pump Efficiency 36

5.1 INTRODUCTION 36
5.2 RESULTS OF THE THEORETICAL INVESTIGATIONS 37
5.3 INFLUENCE OF DIFFERENT VALUES OF SURFACE ROUGHNESS 39

vi
AEAT-6559/ v 5.1 UNRESTRICTED

5.4 THE INFLUENCE OF SMOOTHING SEVERAL PARTS OF PUMPS 39

5.5 INFLUENCE OF DIFFERENT GAP CLEARANCES ON THE INTERNAL
LEAKAGE FLOW RATE 42
5.6 CONCLUSIONS 43

6. Pump Efficiency deterioration 44

6.1 INTRODUCTION 44
6.2 THE COST OF EFFICIENCY DETERIORATION 44
6.3 FACTORS EFFECTING PUMP DETERIORATION 45
6.4 APPROACHES TO MAINTENANCE 46
6.5 MONITORING OF PUMP PERFORMANCE 46
6.6 REPAIR AND RE-CONDITIONING OF PUMPS 47

7. References 48

Appendix 1: EC Guide to Pump Efficiency 51

Appendix 2: Pump energy consumption calculated from

sales data (1998) 60

Appendix 3: Efficiency Characteristics of

Centrifugal Pumps 62

vii
AEAT-6559/ v 5.1 UNRESTRICTED
1. Policy Options For Improving
The Efficiency Of Pumps

1.1 SUMMARY
To improve the efficiency of the installed base of pumps, it is important that the decision
maker(s) are given accurate information to assist in choosing the most efficient pump for their
application. While such information and guidance should be simple, it is crucial that it is not
over-simplified. In particular it is important that products are only compared on a “like for like”
basis.

Through careful consideration of technical and practical commercial issues, it is recommended

that a “procurement level” approach is used in preference to attaching labels to pumps. A draft
brochure for the end user showing the practical implementation of the procurement level
approach is included as Appendix I.

As described in chapter 4, better general education about pumps and pumping systems is
important whichever policy options are selected. All personnel involved in the pump market
would benefit, but in particular users, specifiers and system designers.

1.2 CONSIDERATIONS IN RANKING PUMPS BY EFFICIENCY

The following key points were the basis on which a procurement level approach was seen as
being preferable to an efficiency labelling scheme:

1.2.1 Definition of pump types

The wide range of pumping applications means that there is a very wide range of pump types
manufactured. The BPMA lists 68 distinct types of pumps, although just a small number of
styles account for the bulk of sales and energy use. Within these ranges there will also be
different sub-styles and materials used to take account of the requirements of different users.
Even within the well defined range of pumps considered in this study, there are no standard
pump outputs:dimensions. (There is an ISO code of dimensions and approximate duties for some
types of end-suction pumps (ISO2858-1975 (E), but different manufacturers will offer different
head:flow characteristics within these set dimensions.) Finally, a particular centrifugal pump
from a manufacturers range will be offered with a range of rotational speeds, impeller diameters,
and sometimes, a range of different impellers.

This variety of pump types would make any sort of labelling scheme challenging to design and
apply.

9
AEAT-6559/ v 5.1 UNRESTRICTED

1.2.2 Determination of pump efficiency and operating point

The operating point and hence efficiency of a pump depends on several factors, chiefly NPSH,
properties of the pumped fluid, and the system head and flow. The need to take into account two
parameters (head and flow) in defining the system operating point makes comparing efficiencies
considerably more complicated than with simpler products. (For example, a light bulb only has a
single working point unaffected by external conditions, and a motor performance is determined

D e s ig n D u ty
H ead

Power

E ffic ie n c y
NPSHR

F lo w

P U M P S 4 /2 8
primarily by just load torque.)
Figure (1.1) – Characteristics of a typical centrifugal pump

1.2.3 Operational Efficiency

The rated efficiency of a pump may differ considerably from the actual installed efficiency. This
is for two reasons, the magnitude of which will vary from pump to pump:

The pump will spend much of its time running at considerably below rated duty, where the
efficiency is likely to be much reduced. Users should be strongly encouraged to take more care
in specifying pumps to ensure that they are not over-rated for the actual duty, and also to consider
fitting speed control, multiple pumps in parallel or scheduling to reduce the energy losses from
low flow operation.

The efficiency of the pump will deteriorate over time, and so the actual efficiency of a pump will
not match that of the pump when new. This is a very important issue concerning both the
average efficiency and maintenance costs of a pump.

10
AEAT-6559/ v 5.1 UNRESTRICTED

1.2.4 The effect of Accuracy and tolerances of published

pump data

Whatever schemes to promote the specification of more efficient pumps are implemented,
manufacturers will still have to be trusted to publish data that is truthful. This is though no
different to the current situation regarding the accuracy of published data.

The wide tolerances allowed under current pump test standards severely limits the number of
potential product efficiency bands which it is feasible to have to just two.

1.2.5 Definition of efficiency

It is the view of the study group that existing pump declared efficiencies are misleading, and
accordingly it is hoped that action can be taken to amend ways in which test standards are used
when quoting efficiencies. However, this is not a reason to delay implementation of any
suggested policy actions.

There are several important concerns over quoted efficiencies:

The allowed tolerance on mass-produced pumps by ISO test codes is wide. Therefore it is
strictly hard to advise procurement behaviour on the basis of this standard. To get a much more
useful efficiency figure the purchaser would have to go back to the manufacturer and ask for an
efficiency quote based on the pump being tested to class B. It is however recognised that this
will usually be unrealistically expensive.

Unlike motors, where it is the losses rather than the efficiency that is measured, pump test
efficiencies have a tolerance on the efficiency rather than the losses. The tolerance is based on
the accuracy of the equipment used for the measurement of power, head and flow. This means
that while the tolerance as a percentage of the pump efficiency remains the same across the
whole efficiency range of pumps, the allowed tolerance expressed as a percentage of the losses
on more efficient pumps is wider than on less efficient pumps. Furthermore, it is precisely this
higher efficiency part of the market where efficiency is considered an important purchasing
parameter. Similarly, less efficient pumps have a tighter tolerance. This results from the relative
accuracy which is constant for the measurement of power, head and flow.

It is not unusual to prepare special samples (eg special polishing) for efficiency testing.

EUROPUMP has suggested that quoted catalogue efficiencies of some manufacturers may be
different from actual shipped product.

11
AEAT-6559/ v 5.1 UNRESTRICTED

1.3 DEVISING A CLASSIFICATION SCHEME

1.3.1 Key considerations

Despite the very real issues summarised above which relate to the need for careful interpretation
of a simple pump efficiency value, there is clear user demand for some simple way of indicating
in a fairly simple way which pump would use the least energy in a particular application. This is
particularly the case for the smaller commodity type pumps considered in this study.

But for specialist and engineered pumps where more effort is anyway spent in the selection of
pumps, such simple indicative guidance on efficiency is not so important. Similarly, well-
educated users are anyway proficient in the proper selection of pumps, and so further information
is not so important for such people. However, the complexities of the decision making routes
when purchasing pumps means that a clear indicator of pump efficiency would be of great
benefit.

Considerations made clear that it is only practical to have a single efficiency dividing line. Two
options for the use of this line have been considered:

Labelling of pumps
Publishing a “procurement level” allowing buyers to compare the efficiency of a pump they are
considering with others available.

In considering which option to recommend, the group also took account of the additional
considerations relating to attaching a label to a particular pump.

Pump efficiency varies with speed (eg a pump optimised for a 2-pole motor may operate less
efficiently when connected to a 4-pole motor).
Pump efficiency varies with impeller diameter, with it being the norm for pumps to be offered
with a variety of diameters. (Efficiency will also vary with impeller fitted, since a number of
impellers may be fitted within the same casing).
Pump efficiency can vary with different materials and when offered to different specifications,
(eg API specifies large wear-ring clearances).
There would be confusion on pumpsets if both the motor and pump had labels with different
efficiency ratings.

A database of pumps would be useful to help users to select the most efficient pump for their
particular application. As with all databases of equipment, care would need to be taken that users
were always looking at pumps appropriate for their particular application, and that the
information was correct and up to date.

The pumps extension to the Eurodeem database of motors will include several tools to assist
users in pump selection:
A simple guide for users on selecting the right type of pump for different applications.
A database allowing users to compare the efficiency of different pumps for a specified duty. This
database tool could automatically calculate a procurement level efficiency for the duty, showing
clearly which pumps exceed this level.

12
AEAT-6559/ v 5.1 UNRESTRICTED
Closely related to this will be a European version of the US PSAT (Pump System Assessment
Tool) software.

Figure 1.2 - EURODEEM motor selection software

A particular concern about the growth of databases and e-commerce at the expense of more
traditional literature and contact with the supplier is that buyers will be more likely to buy on the
basis of the few parameters they have selected (eg price, delivery, perhaps efficiency). This
could lead to less time being spent on consideration of other key criteria for a particular
application.

1.4 THE RECOMMENDED PROCUREMENT LEVEL APPROACH

1.4.1 Why the procurement level approach?

As a result of the previous considerations, the study group recommends that the procurement
level approach is used to promote the better selection of pumps.

This would in practice be plots of pump best-efficiencies and a “mean-line”. This approach
differs essentially from that of labeling by assessing the pump efficiency based on its actual

13
AEAT-6559/ v 5.1 UNRESTRICTED
application and duty. This is important because since pumps are not made in standard ratings, in
practice a purchaser will be comparing pumps operating at different load points. It also
overcomes the issues of different impeller diameters and types, and of different motor speeds. A
procurement level scheme also has the benefit of being completely separate from manufacturers,
since no actual pump listing is needed.

The published “procurement level” would therefore give the user a clear guide as to how the
efficiency of the pump they are considering compares with the best-efficiencies of others of the
same type. This can help them in choosing a pump, and be the basis of a minimum efficiency
procurement specification. Manufacturers might also wish to claim that some of their pumps,
subject to certain product options and operating criteria, can achieve the level of the published
‘mean line’. Manufacturers will also be given the opportunity of comparing the best-efficiencies
of their pumps with other makes and may decide to make design improvements.

1.4.2 Might pump labelling still be valid?

The study group recommends that in some commodity type markets, in particular the smaller
domestic/light commercial water circulation pump market, where the in-line pump and motor are
always sold as a standardised integrated unit, some form of labelling might be possible.

1.4.3 Where to draw the efficiency line?

A ‘mean line’ of best-efficiencies based on data collected from many different manufacturers
will be plotted, for guidance only. The rated duty will rarely coincide with the pump best-
efficiency point. The rated flow will usually be below or above the best-efficiency flow. The
rated head will usually require a reduced diameter impeller. A survey has suggested that less
than one in five quotes will fall above the ‘mean line’ of best-efficiency points.

14
AEAT-6559/ v 5.1 UNRESTRICTED

2. Efficiency Levels Of Pumps

2.1 OBJECTIVES

To obtain and analyse data on the performance of pumps currently in manufacture, in a form
which can be used as a basis for formulating proposals for rating pumps by efficiency level. Also
to consider why pump efficiency deteriorates in use.

2.2 PUMP TYPES AND DESIGN DUTIES

An examination of the market shows that about 26% of purchases by value are single stage
pumps. These are mainly of the end suction type, which are either close coupled with the
impeller mounted on an extension of the motor shaft or long coupled with their own bearings,
and the double entry axially split case type.

Regarding the operating ranges, we have looked at the areas of maximum purchase value. This
is the basis for the chosen sizes, flows, heads and powers. The 50 Hz speeds of two pole and
four pole motors only are selected. The resulting data is shown in Table 2.1. The corresponding
shaft power range is quite wide, from 1.2 to 680 kW.

2-pole 4-pole

Approx. Approx.
Flow Head Flow Head
Outlet Outlet
Pump Size (m3/h) (m) (m3/h) (m)
Size (mm) Size (mm)
Type Limit

Double Max 150 400 150 300 1500 140

Entry
Split Min 80 130 50 100 150 15
Case

End Max 100 200 130 150 400 50

Suction
own Min 50 30 20 50 30 10
bearings

End Max 100 200 130 100 100 50

Suction
close Min 32 15 20 50 30 10
coupled
Table 2.1. Pump flow and head limits selected

15
AEAT-6559/ v 5.1 UNRESTRICTED

1000

Duty Coverage - 2 pole

DESC
Double Entry Split Case
200 kW
96 kW
Head (m)

ESOB
100 End Suction Own Bearings
ESOB
DESC
ESCC
23 kW End Suction Close Coupled

ESCC

1.4 kW 2.5 kW

10
10 100 1000 10000

Flow (m3/h)

Fig 2.1. Range coverage at best-efficiency duties (2 pole)

1000

Duty Coverage - 4 pole

ESCC ESOB DESC

End Suction Close Coupled End Suction Own Bearings Double Entry Split Case

680 kW
Head (m)

100

20 kW
DESC
67 kW

ESOB

ESCC
7.4 kW
1.2 kW
10
10 100 1000 10000
3
Flow (m /h)

Fig 2.2 Range coverage at best-efficiency duties (4 pole)

Figs 2.1 & 2.2 fully define the areas of interest. These indicate the exact ranges of best-
efficiency flows and heads which we will consider. The two pole coverage (Fig 2.1) shows the

16
AEAT-6559/ v 5.1 UNRESTRICTED
simple rectangular boxes of head and flow as defined by Table 2.1. However, at four pole speed
(Fig 2.2) rectangular boxes would have produced some abnormally low or high specific speeds at
those corners of the boxes which are not directly defined by Table 2.1. It was therefore necessary
to change the rectangles to trapezia to avoid those duties where pumps in parallel or series (or
even multistage) would be much more efficient.

2.3 DATA COLLECTION AND ANALYSIS

Data has been collected from pump makers in the form of hydraulic performance curves showing
head, flow, efficiency and NPSH at maximum impeller diameter. The NPSH figure is needed
since we will ignore any pump requiring more than 10m NPSH as ‘abnormal’. This data is
readily available and a large amount has been collected. No data will be published in a form that
makes it identifiable with any particular maker.

The simplest pump efficiency to analyse is that at best-efficiency flow and full diameter impeller.
However, this fails to take into account two relevant facts:

On average, pumps operate at flows below their best-efficiency flow.

Pumps usually operate over a range of flows. Thus, an efficiency curve which has a high
peak value but falls off rapidly away from the peak is not ideal.

We therefore considered a ‘weighted efficiency’ of the form:

Weighted efficiency = (60%Qx1)+(80%Qx2)+(100%Qx2)+(120%Qx1)

6
(where 60%Q means efficiency at 60% of best-efficiency flow)

This ‘weighted efficiency’ would be the better measure for labelling pumps. However, as the
project progressed it became evident that the labelling of pumps (as done with electric motors) is
not a practical proposition for many reasons, as explained in Chapter 1.

We therefore decided to use a ‘procurement level’ approach. By this method, anyone selecting a
pump for their principal duty can compare the quoted efficiency at that duty with the best
efficiency of other pumps in the market. They can then see immediately whether a more efficient
pump is likely to be available and should therefore be sought.

It is clear that the ‘procurement level’ scheme cannot use a weighted efficiency and must show
best efficiencies only. This is simply because the person selecting the pump will normally be
concentrating on their primary duty point and will not be interested in calculating a weighted
efficiency from the above formula. This is particularly the case since, in most instances, the
formula is unlikely to match their specific application. In other words, it would not be possible
to compare like with like.

17
AEAT-6559/ v 5.1 UNRESTRICTED
2.4 PROCEDURE FOR ANALYSIS

There are many ways of plotting curves of pump efficiency against the mutual variables of head,
flow and speed. One of the most traditional is to plot best efficiency against specific speed (Ns).

Ns = N x Q1/2 / H3/4 (where N = speed, Q = flow, H = head) ……………….. (i)

However, this is normally done for a discrete number of flows, since large pumps are more
efficient than small pumps. This means that analysis by plotting efficiency points for many
pumps on a family of curves of this type is impractical, since most of the flow rates for points to
be plotted will be different from those chosen for the individual curves.

The method we have chosen is to plot best efficiency against flow for each type of pump and
running speed, as shown typically in Fig 2.3. This is also the basis of the method used by the
Hydraulic Institute of the USA (Ref 1) amongst others.

85

80
Pump Efficiency + C (%)

75

Pump Ns (rev/min, m3/h, m)

0 1000 2000 3000 4000
0
70
2
4
Correction = C

6
8
10
65
12
14
16

60
20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200

Pump Flow (m3/h)

Fig 2.3. Typical plot of best efficiency against flow

(For illustration only, not to be used for selection)

Note that in this chapter, specific speed is calculated using the following units:

Speed - rev/min
Flow - m3/h
Head – m

18
AEAT-6559/ v 5.1 UNRESTRICTED
2.5 EFFECT OF SPECIFIC SPEED

The disadvantage of Fig 2.3 is that head variation must be allowed for, since at a given flow
pumps with high or low heads will have lower efficiencies than pumps with ‘optimum’ heads,
even though the pumps are of equal ‘quality’.

This variation of head is accommodated by making a correction for specific speed. In Fig 2.3,
best efficiency is plotted against flow for several pumps. A typical specific speed correction
curve is also shown. Using this curve, at a given flow a pump of 2500 specific speed would have
its efficiency point plotted directly, whereas a pump of 800 specific speed would have 8 points
added to its efficiency before plotting. In this way, efficiency scatter should be a reflection of
pump ‘quality’ only.

Anderson (Ref 2) analysed hundreds of pumps and deduced a specific speed correction factor:

C = 29 [ log10 (2650 / Ns)]2

This has been used for the ‘typical’ correction curve shown in Fig 2.3.

Other authors have indicated other correction factors. Ten such curves are shown in Fig 15 of
Ref 3. The correction factors tend to vary with the particular test results analysed. Because of
the method of analysis, different corrections can also be derived for different flows. The
Hydraulic Institute in Ref 1 chose to use Anderson’s correction factor.

There is no universally agreed correction for specific speed. Fig 2.4 shows a comparison of
specific speed corrections at 180 m3/h and 2900 rev/min using:

Anderson and Hydraulic Institute (Refs 2 & 1)

Average maximum attainable efficiency from other authors (Ref 3)
Maximum practically attainable efficiency (Europump, Ref 3)
Theoretically attainable efficiency (Darmstadt, Ref 4)

These show appreciable differences at low specific speed. We decided that there was insufficient
agreement in the data to use different corrections for different flows, and therefore to use a single
correction curve for the limited range of flows covered in our analysis. To decide which
correction to use, we plotted two curves of efficiency against flow for end suction own bearings
pumps at 2900 rev/min. For one curve we used the specific speed correction of Anderson and for
the second we used the correction derived from the maximum practically attainable efficiency
curve from Europump, Ref 3. By identifying the points of low specific speed (high correction
factor) it was evident that the correction curve giving the minimum scatter of pump efficiency
would be an intermediate curve. By trial and error we arrived at this ‘optimum’ correction curve,
which has also been plotted on Fig 2.4, designated ‘Chosen for SAVE’. The formula for this
‘optimum’ correction factor is:

C = 48 [ log10 (2650 / Ns)]2 ………………………………… (ii)

19
AEAT-6559/ v 5.1 UNRESTRICTED
There is no strict theoretical justification for this approach but we believe that it serves our
purposes satisfactorily. It is possible that more detailed research may suggest varying correction
factors for different types of pumps.

Pump Ns

0 500 1000 1500 2000 2500 3000 3500 4000

0

2
Theoretically
attainable

6
Correction = C

Maximum by
other authors
8

10 Anderson
and
Hydraulic Chosen
Chosen
Institute for
for SAVE
SAVE
12

Maximum
practically
14
attainable

16

Fig 2.4. Specific speed corrections at 180 m3/h and 2900 rev/min

85 50

80 45

75 40

70 35
'Lower Efficiency'
Efficiency correction = C
Pump Efficiency + C (%)

65 30

Catalogue 'mean'
60 25
of best-efficiency
points
130m head
55 110m 20

90m

50 15
70m 20m

45 50m 10

40m
40 5
30m

20m
35 0
30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
3
Pump Flow (m /h)

Fig 2.5. Efficiencies of End Suction Own Bearings pumps at 2900 rev/min

20
AEAT-6559/ v 5.1 UNRESTRICTED
90 50

85 45

80 40

75 35

'Lower Efficiency'

Efficiency correction = C
Pump Efficiency + C (%)

70 30

65 25

Catalogue 'mean'
of best-efficiency
60 20
50m head
points

40m
55 15

30m
50 10
15m 10m
20m
45 5
15m

10m
40 0
20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400
3
Pump Flow (m /h)

Fig 2.6. Efficiencies of End Suction Own Bearings pumps at 1450 rev/min
85 60

80 55

75 50

70 'Lower Efficiency' 45

65 Catalogue 'mean' 40

Efficiency correction = C
Pump Efficiency + C (%)

of best-efficiency
60 points 35

55 30

50 25
130m head

45 110m 20

90m
20m
40 15
70m

35 10
50m

40m
30 5
30m
20m
25 0
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
3
Pump Flow (m /h)

Fig 2.7. Efficiencies of End Suction Close Coupled pumps at 2900 rev/min

21
AEAT-6559/ v 5.1 UNRESTRICTED
90 50

85 45

80 40

75 35

Efficiency correction = C
'Lower Efficiency'
Pump Efficiency + C (%)

70 30
Catalogue 'mean'
of best-efficiency
65
points 25
50m head

60 40m 20

55 30m 15

25m

50 10
20m

15m
45 5

10m

40 0
30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
3
Pump Flow (m /h)

Fig 2.8. Efficiencies of End Suction Close Coupled pumps at 1450 rev/min
90 50

85 45

80 40

75 35
'Lower Efficiency'

Efficiency correction = C
Pump Efficiency + C (%)

70 30
Catalogue 'mean'
of best-efficiency
points
65 25

60 20

55 15
150m head

130m
50 10
110m

90m

45 5
70m

50m

40 0
130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400

Pump Flow (m3/h)

Fig 2.9. Efficiencies of Double Entry Axially Split Case pumps at 2900 rev/min

22
AEAT-6559/ v 5.1 UNRESTRICTED
95 50

90 45

85 40

80 35
'Lower Efficiency'
Catalogue 'mean'

Efficiency correction = C
Pump Efficiency + C (%)

75 of best-efficiency 30
points

70 25

65 20
140m head
15m

120m 20m
60 15
40m
100m
30m
55 80m
10

60m
50 5
50m
40m
30m
20m
45 0
100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500

Pump Flow (m3/h)

Fig 2.10. Efficiencies of Double Entry Axially Split Case pumps at 1450 rev/min

2.6 EFFICIENCY REFERENCE LEVELS

Six ’procurement level’ plots have been produced showing pump best efficiency against flow
(Figs 2.5 to 2.10). These are based on pump selection curves published by many different
manufacturers. The reliability of manufacturers’ curves has sometimes been questioned.
However, since these are the curves used by the manufacturers when quoting pump efficiencies,
anyone selecting a pump is justified in comparing the selected/quoted efficiency with the
‘procurement level’ plots.

Each plot shows a ‘mean’ efficiency line with a ‘lower efficiency’ area below the ‘mean’. This
provides a guide to the user of the plot and a convenient way of defining a level of efficiency.
However, the plots could readily be used for their intended purpose without a ‘mean’ line.

Since each plot refers to a stated speed, it is possible to dispense with the ‘optimum’ specific
speed correction curve by relating the efficiency correction at a given flow to the pump head,
using formulae (i) and (ii). In this way, it is possible for less experienced people to make use of
the plots without needing to understand the concept of specific speed.

In Fig 2.5, the scatter of efficiencies lies roughly between six points above and ten points below
the ‘mean’. It is important that this extent of efficiency variation is brought to the notice of all
involved, to emphasise the scope for choosing pumps of better efficiencies.

23
AEAT-6559/ v 5.1 UNRESTRICTED
For guidance on using Figs 2.5 to 2.10 see Appendix 1.

Theoretically
95 attainable

90

Maximum
Pump Efficiency at Optimum Specific Speed

85 practically
attainable

80
(%)

75 Catalogue
'mean'

Hydraulic
Institute
70 'Large
Anderson Pumps'

Hydraulic
65
Institute
'ANSI/API'

60

55
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
3
Pump Flow (m /h)

Fig 2.11. Comparison of best-efficiency levels from different sources

Fig 2.11 provides best-efficiency comparisons between the ‘Catalogue ‘mean’’ from Fig 2.5,
‘Maximum practically attainable’ (Ref 3), ‘Theoretically attainable’ (Ref 4), ‘Hydraulic Institute’
(Ref 1) and ‘Anderson’ (Ref 2). All the curves relate to optimum specific speed. They show
reasonable agreement in curve shape except that the ‘Theoretically attainable’ curve shows a
relatively high efficiency at low flows. This relative increase in efficiency would clearly involve
a high additional cost which may be hard to justify in a small pump.

2.7 EFFICIENCY DETERIORATION IN USE

There is little point in encouraging the use of higher efficiency pumps if their efficiencies fall off
rapidly in service. A small pump on clean water supply duty has been known to lose over 20%
of its new efficiency in less than two years of operation (Ref 5). Although this is an extreme
case, it has been shown that larger clean water pumps can lose, on average, around 5% of their
new efficiency in the first five years of operation (Ref 6).

Part of this loss in efficiency is due to wear at the impeller/casing wear rings (see section 5.5 for
more details on this). The rate of wear will tend to be much greater in pumps which operate at
flows away from the best-efficiency flow due to increased shaft deflection.
Most of the loss in efficiency will normally be due to a build-up of corrosion products in cast

24
AEAT-6559/ v 5.1 UNRESTRICTED
iron casings. This occurs with practically all pumps working with clean cool water. The rate of
corrosion is affected by the catalytic effects of bacteria which are usually present even in treated
drinking water.

Methods of maintaining pump efficiencies at high levels are covered in Chapter 6.

2.8 REFERENCES

1. ‘Efficiency prediction method for centrifugal pumps’, Hydraulic Institute, 1994.

2. Anderson, H.H., ‘Centrifugal pumps and allied machinery’, Elsevier Advanced

Technology, 1994.

3. ‘Attainable efficiencies of volute casing pumps’, Europump Guide No. 2, 1999.

4. Stoffel, B. and Lauer, J., Summary of the final report on the research project for VDMA
‘Theoretically attainable efficiency of centrifugal pumps’, Technical University of
Darmstadt, 1994.

5. Reeves, D.T., ‘Deterioration of a centrifugal pump in service’, Allen Engineering

Review, November 1960.

6. Fleming, J., ‘Identification and implementation of effective pumping system energy cost
savings’, Pumping Cost Savings in the Water Supply Industry Seminar, IMechE, May
1989.

25
AEAT-6559/ v 5.1 UNRESTRICTED
3. Characterisation of Pump Use
in the EC

3.1 INTRODUCTION
This chapter considers current and future pump use within the EU. The purpose is to give agreed
“best estimates” of current energy consumption and potential for energy savings. Against this the
cost-effectiveness of future energy saving policy options can be measured. There are four main
sub-sections in this section, including:
· Centrifugal pump use
· EU pump stock based on motor energy consumption
· Energy consumption based on annual sales figures
· Potential savings in future pump energy use

3.2 CENTRIFUGAL PUMP USE

The centrifugal pumps considered in this section of the report constitute just part of the total
centrifugal pump market, and include;
· End suction close coupled centrifugal pumps
· End suction own bearing centrifugal pumps
· Double entry split case centrifugal pumps

Two main sources of data were used to characterise the EU energy consumption of pumps;
· Pump stock based on motor energy consumption, using data from SAVE II study1 on EU
motor energy consumption. This study is based on a comprehensive study of motor energy
use in European industry, and so gives a very good estimate of overall pump energy
consumption.
· Annual Sales Figures, using EU pump production/sales data1. This data gives a much more
detailed breakdown of pump use , by type, by county.

3.3 METHOD 1 - EU PUMP STOCK BASED ON MOTOR ENERGY

CONSUMPTION
The total electricity consumption for motors is estimated to be 797 TWh (SAVE II)1 in the whole
of the EU. This is based on the assumption that motors account for 69% of the total electricity
consumption of Industry and 38% of total electricity consumption in the tertiary sector.

Figure (3.1) shows the split of total motor energy consumption, showing that pumps consume
approximately 20% of total motor energy consumption2, equivalent to 160 TWh pa.

1
SAVE II, Improving the penetration of energy-efficient motors and drives, 2000.
1
Data presented by national associations, public statistical offices and supplied to the study group by europump.
2
This is based on 1996 data presented in the SAVE II1 report.

26
AEAT-6559/ v 5.1 UNRESTRICTED

Pumps
20%

Other
30%

Fans
Conveyors 18%
4%
Cool Compressors
11% Air compressors
17%

Figure (3.1) – Split of motor energy consumption, by application.

3.4 METHOD 2 – ENERGY CONSUMPTION BASED ON ANNUAL

SALES FIGURES
This data was supplied to the study group by Europump and was derived from data presented by
the national associations, public statistical offices and EUROSTAT. It includes 1998 figures for
EU pump production, imports and exports, allowing the total purchase value by country to be
calculated. This data was then used to derive an approximate annual energy use of these pumps,
using the following methodology. (The table detailing these calculations is included as Appendix
2.)
45,000

40,000
Average Annual Energy Usage (GWhpa)

35,000

30,000

25,000 End Suction Close

Coupled (GWhpa)
20,000
End Suction Own
15,000 Bearing (GWhpa)

10,000 Double Entry Split

Case (GWhpa)*
5,000

-5,000
D UK F B DK FL I NL AU SP
End Suction Close Coupled (GWhpa) 6,000 15,700 16,100 7,200 700 500 14,000 4,000 1,600 7,000
End Suction Own Bearing (GWhpa) 38,200 7,600 23,100 2,800 7,600 2,400 27,500 2,500 3,600 10,900
Double Entry Split Case (GWhpa)* -1,200 7,300 700 -700 300 - 1,200 1,000 100 4,800

Country

Figure (3.2)*–Average annual energy usage of the three centrifugal pump types considered, by
country

*
The negative values shown in this figure are as shown in the statistics, although it is unclear what this indicates.

27
AEAT-6559/ v 5.1 UNRESTRICTED
Source: National associations, public statistical offices and EUROSTAT, supplied to the study group by EUROPUMP.

Assuming that all pumps are sold as a pumpsets, the study group estimated an average pumpset
cost of 200Euro/kW. This produced a value for the annual sales (kW) of pumps.

20 years was suggested as an estimate lifetime for the types of pumpset considered. This
produced the total installed pumpset capacity (GW).

It was estimated that the average pumpset runs for 4,000hpa1, (although it should be noted that
this figure is an average for all motors), and that the motors driving these pumps run at an
average of 65% of their rated electrical power. The calculated energy usage (GWhpa) for the
three types of centrifugal pump investigated in this study are shown in figure (3.4). This
calculation does not take into account any motor losses.

COST (MILLION Euro)

TYPE COUNTRY Belgium Austria Czech France Germany Spain UK TOTAL
Republic
Hand Pumps - - - 7.12 16.25 - 7.54 30.91
Reciprocating Pumps 4.14 15.18 1.01 51.05 266.86 0.82 48.78 387.84
Dosing & proportioning - - - 38.02 104.51 - 16.41
Inline Piston 4.14 - 0.03 4.7 65.05 - 10.64
Other Reciprocating - - 0.98 8.33 97.3 - 21.73
Rotary Pumps 2.57 5.5 2.453 60.29 191.25 3.18 135.7 400.95
Gear pumps 2.51 - 0.64 27.57 26.76 - 14.04
Vane pumps 0.06 - 0.003 3.48 - - 4.73
Screw pumps - - 1.81 3.18 71.76 - 2.66
Other rotary pumps - - - 26.06 92.73 - 114.3
Centrifugal Pumps 88.50 74.49 15.6 560.48 1076.82 186.23 451.2 2453.31
Submersible single stage - - 0.18 43.17 165.71 37.34 30.6
Submersible multistage 42.63 - 0.08 33.93 82.07 65.34
Glandless circulators - - 0.93 182.99 159.36 39.92
Centrifugal < 15mm outlet - - 0.18 6.51 5.18 -
Channel/side channel - - - 9.69 67.39 1.79 -
Single St./S. entry monobloc - - 1.37 53.17 67.14 22.96 72.14
Single stage, single entry 26.59 - 4.37 81.65 250.01 39.87 72.88
single stage > 1 entry - - 0.23 62.11 7.82 18.36 39.47
centrifugal multistage 3.68 - 5.42 1.82 121.32 41.93 62.83
other centrifugal - - 0.64 37.57 92.22 9.02 48.79
other centrifugal multistage 15.6 - 2.2 47.87 58.6 14.96 19.22
Other Pumps - - - - 121.37 12.96 - 134.33

Table (3.1) – Pump production data for different types of pump

Source: PRODCOM, supplied by EUROPUMP, (1998 figures)

Table (3.1) is a summary of data produced by PRODCOM supplied by EUROPUMP, giving a

breakdown of pump production, split by type and country. Figure 3.2 uses the data to highlight
the proportion of output accounted for by the types considered in this study.

28
AEAT-6559/ v 5.1 UNRESTRICTED

End suction close coupled

9%

Other pumps
4%
Rotary Pumps
End suction ow n bearing
12%
15%

Reciprocating Pumps
11%
Double entry split case
2%

Other Centrifugal Pumps

47%

Figure (3.3) - Split of sales of all pump in the EU, by type, by size.

3.5 FUTURE PUMPSET USE WITH POTENTIAL SAVINGS

For future pumpset energy use, a time horizon of 2015 was looked at.

Using the data shown in chapter 2, on average, energy savings of c3% can be made by users
selecting higher rather than lower efficiency pumps.

It is expected that it will not be until the year 2020 that all less efficient pumps will be replaced,
due to their estimated 20 year lifetimes. Using these assumptions, calculations were undertaken
showing the possible future energy saving, table (3.2).

This shows that savings of 1.1 TWhpa could be made in 2015 from the use of the higher
efficiency pumps identified. This would account to a cumulative saving of 8.5 TWh by 2015.

29
AEAT-6559/ v 5.1 UNRESTRICTED

Year business as uptake of Saving with 3% Energy consumption

usual (TWhpa) efficient increase in pump with uptake of
pumps efficiency efficient pumps
(TWhpa) (TWhpa)
2000 41.5 0% 0.0 41.5
2001 42.0 5% 0.1 41.9
2002 42.5 10% 0.1 42.3
2003 43.0 15% 0.2 42.8
2004 43.5 20% 0.3 43.2
2005 44.0 25% 0.3 43.7
2006 44.5 30% 0.4 44.1
2007 45.1 35% 0.5 44.6
2008 45.6 40% 0.5 45.1
2009 46.2 45% 0.6 45.5
2010 46.7 50% 0.7 46.0
2011 47.3 55% 0.8 46.5
2012 47.9 60% 0.9 47.0
2013 48.4 65% 0.9 47.5
2014 49.0 70% 1.0 48.0
2015 49.6 75% 1.1 48.5
CUMULATIVE
TOTAL 726.7 8.5 718.3

Table (3.2) – Projected pump energy use and potential savings

3.6 SUMMARY
· Pumps in the EU consume 160 TWhpa of electricity. This represents 20% of energy
consumed by electric motors. This is equivalent to 14% of the total electricity consumed in
industry and commerce.
· Centrifugal pumps in the EU use 117 TWhpa of electricity. This represents 73% of all pump
energy consumption.
· The centrifugal pumps on which the calculations in this chapter are based account for 35%
(by capacity) of all centrifugal pumps. This is equivalent to 26% of all pumps.
· From the uptake in the EU of the higher efficiency pumps identified, (chapter 2), an energy
saving of 1.1 TWhpa could be made in 2015. This would give a cumulative saving of
8.35TWh by 2015. If this was inclusive of all pumps the cumulative energy saving value is
expected to at least double.
· The two methods used to obtain estimates of energy consumption by pumps involved several
assumptions and were necessarily imprecise, however the results give reasonably similar
answers. Therefore both results can be used with a reasonable degree of confidence.

30
AEAT-6559/ v 5.1 UNRESTRICTED
4. User Attitudes to Pump
Selection and Ways in Which
They Can Be Changed

4.1 ENERGY EFFICIENCY IS NOT A PRIORITY

A survey of Users and specifiers showed the following general order of priorities when
considering the purchase of new pumps, (figure (4.1)):

(1) KEY CRITERIA

· Immediate concerns
(2) SECONDARY · Price
CRITERIA · Deliveries
· Reliability · Known name
· Maintenance (3) SOMETIMES
CONSIDERED
· Energy costs

Figure (4.1) – Priorities when purchasing a new pump

But in terms of overall costs of ownership, the order of importance of factors in determining
Life-Cycle Costs is actually the opposite of the priorities listed:

Purchase
5%
Maintenance
10%

Energy
85%

Figure (4.2) - Example of Lifecycle Costs of Ownership of a pump

31
AEAT-6559/ v 5.1 UNRESTRICTED

4.2 BARRIERS TO IMPROVING THE EFFICIENCY OF PUMPS

There are lots of barriers to the reduction in energy consumption of pumps, here divided loosely
into three categories. Contrary to initial impressions, often it is the organisational and
educational factors that are the main barrier to improved efficiency, rather than technical or
economic factors.

4.2.1 Organisational Behaviour

· Many pumps are sold to intermediaries such as Distributors and OEM suppliers, who are
driven by customers to supply at lowest cost.
· Within pump users’ organisations, the department responsible for buying the pump is
different from that paying the running costs.
· De-manning means that there is less effort available for energy saving considerations. This is
compounded on many sites by a low level of skills.
· There is often little visibility of the costs of pumping.
· Fast turnaround on the repair of pumps means that often non-specialists undertake repairs in a
poor way.

4.2.2 Technical/Educational
· There are trade-offs between efficiency and maintenance effort /safety requirements.
· Uncertainty over system characteristics, and allowances for future plant expansion,
frequently mean that pumps are considerably over-sized for the duty.
· Promoting improved efficiency of pumps should not be divorced from system efficiency
considerations.
· Pumps are often poorly maintained, and are not given attention until they start to cause
problems or stop working altogether.
· There is considerable difficulty in measuring even the most basic of parameters, for instance
flow.
· Inability to correctly choose pumps.
· Scepticism over manufacturers efficiency data.
· Pumps with low duty (ie low running hours and/or low load) are unlikely to justify the greater
costs of being more efficient.
· It is easier to replace a pump with the same type as before rather than re-assessing the
requirement.
· Misapplication can quickly de-grade the efficiency of a “good” pump.

4.2.3 Economic
· For most users, energy saving is treated as if it is less important than either first cost, ease of
maintenance or reliability.
· Payback on a more energy efficient solution is not adequate.

32
AEAT-6559/ v 5.1 UNRESTRICTED
· Large sites may either buy or generate low cost electricity which makes payback times
unattractive.
· Total energy costs as a proportion of total costs may be too low to warrant any significant
effort.
· Any energy savings made can be difficult to assess.
· Acceptance tests to verify performance are only economic on larger pumps.
· Lack of capital for purchasing new or more expensive pumps.

4.3 EDUCATIONAL NEEDS

Which ever strategies are adopted, education is an essential supporting element. Financial
incentives are a good short term measure for encouraging a change in behaviour, but are unlikely
to be cost effective if run for a long time. Legislation can enforce market changes, but must be
carefully thought through if it is to be workable.

From the study it is clear that all markets participants would benefit from some education – eg.
Pump maintenance engineers, Plant designers, Non-technical management, Pump suppliers. The
following key suggestions on best approaches to promoting energy efficiency came out of the
discussions with such personnel:

Operation .
Overall - Duty System- Equip- -
Control- Detail Balan- Mainte-
Planning Cycle lay out ment ing design cing nence

Specification Designing the pump system Installing and operation

Figure (4.3) – Stages in the specification, design and operation of pumps

· Energy saving is currently only an important issue for a small number of users, and so in
many cases it will be better to promote general pump selection and maintenance best practice,
which will often lead to energy saving as an “incidental” extra.

· The importance of correct maintenance, to both maintain pump reliability and to reduce
energy costs, should be promoted.

· The importance of whole life costing (LCC) as the best way to assess different pump and
system options. The Enersave guidance will be useful in helping to achieve this.

· Similarly, in order to make messages interesting and relevant, and to help ensure that they
lead to actual action, information should be given on specific types of applications with quite
detailed best practice solutions.

· We should not divorce the issue of the pump from system issues. However, users are in a
better position to choose a better pump than to make other system changes that are likely to
require extra skills and time.

33
AEAT-6559/ v 5.1 UNRESTRICTED
During the study the group reviewed many examples from the wide range of materials available,
including the following:
· Literature – Guides, case studies.
· Training courses
· Videos
· Software Tools
Ø pump system modelling software (eg Pump-Flo),
Ø pump selection software (eg EURODEEM),
Ø pump system efficiency tools (ie US P-SAT).
It was observed that there are actually many useful pump related materials around in existence
that deserve a much wider audience, but that attention to selecting the best routes for
dissemination is now the “weak link” for ensuring wide readership and hence changes in
behaviour.

Main dissemination routes include Seminars, energy agency databases, equipment suppliers
databases, Articles / Press releases. To further encourage the wider use of these materials, the
study group suggests a web-based resource book listing all useful materials available within the
EU.

4.3.1 On site advice

Free or subsidised on site advice would give sites a quick idea of which pumps, if any, should be
looked at in more detail to make energy savings. More detailed advice could also be offered on
selected pumping systems with high energy savings potential.

4.3.2 Encouraging third party involvement

“Contracting out” part or all of the purchase, installation and maintenance of pumps to a
qualified third party can be an excellent way of ensuring better practice in companies without the
time or expertise to devote to pumps. This sort of approach naturally encourages a longer time
frame when considering different options. Sharing energy and/or cost savings with third parties
can really help encourage longer term and more imaginative measures.

4.3.3 Research and Development

Several manufacturers commented that there is a need for more funding for R&D into the design
of more efficient pumps. However, the study group considers that the potential from
encouraging the use of the best pumps already available is much greater than that from
designing even more efficient pumps. Accordingly, it is recommended that funds are first
devoted to the promotion of best practice in the selection and use of pumps, with further funds
for R&D following later.

4.4 WHAT USERS WANT TO HELP THEM IMPROVE THEIR

SELECTION OF PUMPS

34
AEAT-6559/ v 5.1 UNRESTRICTED
Pump users/specifiers were also asked for their thoughts on the usefulness of other possible tools
to help their decision making process. Overall, the less well informed the respondent, the
simpler the information that they want.

Simulation software. Mainly used by specialist consultants. Other respondents thought that
they would find it hard to know enough about their system to be able to make good use of this
software.

Pump selection Database. It was thought that specialists would find this useful, but that less
informed users weren’t sure how useful they would find it. A demonstration programme would
be useful in order to get better feedback on this.

Procurement Level. All users thought that this could be useful, but in particular the less
informed users. Again, a short leaflet to better explain this scheme is necessary to get better
feedback.

Since the better informed users are already aware of the importance of energy efficiency, and will
usually take account of it during pump selection, it is the view of the study group that these Users
need little further support. Indeed, there is already a well-established commercial market in
simulation software, and internet pump procurement organisations, and so there much less need
for further Governmental effort in these areas. Instead, effort should be better focussed on
supplying tools to improve the behaviour of the less informed users, where there is much more
scope for improvement.

4.5 SUMMARY

Despite energy costs in most cases dominating the overall life cycle costs of ownership of a
pump, there are a wide range of barriers to energy efficiency being given greater importance
when making purchasing decisions. There are accordingly a range of measures that can be
applied to help overcome these barriers. Of particular importance is the need for better education
on pumping issues, without which any other initiatives will struggle to make a significant and
ongoing impact. Simple tools to aid in the selection of pumps, such as the procurement level
scheme illustrated in the brochure in Appendix1, will greatly assist the often relatively
uneducated user in making a much better decision.

35
AEAT-6559/ v 5.1 UNRESTRICTED
5. Design Factors Affecting Pump
Efficiency

5.1 INTRODUCTION
Due to the fact that the majority of the pump manufacturers within the EU have reached a level
of know how which enables them to carry out hydraulically correct designed centrifugal pumps,
the value of the practically attainable overall pump efficiency h of these machines is mainly
influenced by factors such as surface roughness of parts which are in contact with the flow as
well as the internal leakage flows through the sealing gaps. Especially the surface roughness of
hydraulic parts strongly depends on the manufacturing techniques used. Further on the surface
quality is a property which can get worse during life time of a pump and thereby causes energy
losses during pump operation.

To quantify the effects of these above mentioned factors the following investigations on single-
stage centrifugal pumps were carried out at the chair for Turbomachinery and Fluid Power at
Darmstadt University of Technology:
· The influence of different values of surface roughness
· The influence of smoothing several parts of pumps
· The influence of different gap clearances on the internal leakage flow rate

The specific speeds (as defined in figure 5.2) of the considered pumps covered the range from ns
= 10 min-1 up to ns = 100 min-1 (corresponding to values from 520 min-1 up to 5200 min-1 in US-
units) and represents the typical field of application of standard centrifugal pumps. By the aid of
the similarity laws it is possible to transfer the results obtained for one pump size to another
(respectively from one speed of rotation to another).

To carry out the investigations a special software tool was used, which was developed within the
scope of a former research project named “Attainable Efficiencies of Volute Casing Pumps”
sponsored by the Research Fund of the German Pump Manufacturer Association.

The main capability of this program is to estimate the maximum theoretically attainable
efficiency hmax,th of volute casing pumps. We explicitly want to point out that this software tool is
no CFD code. To determine the friction losses for the parts shown in fig. 1 respectively the
leakage flow rates through the sealing gaps the program uses differential equations as well as
simplified mathematically loss approaches. All calculations are carried out on the base of a
hydraulic design process considering common industrial design standards in respect to the
geometrical settings.

36
AEAT-6559/ v 5.1 UNRESTRICTED
Volute
Considered Losses

Diffusor
- hydraulic losses:
Inner surfaces
of impeller friction losses
Sealing gaps deceleration losses
Outer surfaces wake losses
of impeller mixing losses (Volute)
Bearings & sealing

- volumetric losses (gap flow)

Inlet
- mechanical losses

Fig. 5.1 Loss-causing components of a centrifugal pump

In order to evaluate the efficiency values estimated by the computer program additional
experimental investigations were carried out at a centrifugal pump with a specific speed of
ns = 12 min-1. This special test pump, designed according to usual industrial standards was
equipped with very narrow sealing gaps (radial gap clearance equal to 0.1 mm) as well as
hydraulic smooth surfaces. For this optimized pump the value of the inner efficiency hi was
measured at a special high precision test rig and compared to the theoretical value obtained by
the computer program. The comparison of both values (for pump operation at the point of best
efficiency) results in a very good agreement.

5.2 RESULTS OF THE THEORETICAL INVESTIGATIONS

For the purpose within this SAVE study the above mentioned software tool was partly modified
respectively extended in its capabilities. All the following figures which demonstrate the
influence of the parameters surface roughness as well as gap clearance show efficiency values h
respectively differences of efficiency values Dh in per cent points that were plotted versus the
value of specific speed ns as defined in fig. 5.2 (where n is the speed of rotation, Q the rate of
flow and H the pump head).

37
AEAT-6559/ v 5.1 UNRESTRICTED
90
h Rate of Flow
[%] 85 Specific speed:
80 1
Q 2

75
ns = n × 3
4
3
70
Flow [m /h] m3
Q = 18 ns [min ], n [min -1 ], Q [
-1
], H [m]
Q = 36
s
65
Q = 72
Q = 180
60 Q = 360 n s , U S - un its 51 . 64 ns
Q = 720
55 Q = 1800

50
10 100 nS
-1
[min ]
In the following:
n = 1450 min -1
Q = 180 m3 h

Fig. 5.2. The influence of rate of flow

As shown in fig. 5.2 every specific speed value corresponds to a typical impeller geometry,
which means that low specific speeds characterize more radial extended impellers while higher
specific speeds correspond to mixed flow respectively more axial types of impellers. Fig. 5.2
exemplarily shows the dependence of the overall efficiency h on the rate of flow respectively
pump size at constant speed of rotation (n = 1450 min-1). It can be stated that an increase of the
rate of flow leads to higher values of the overall pump efficiency, which is the effect of an
increasing Reynolds number Re. The above figure also shows that the efficiency values for very
low specific speeds are definitely smaller than for higher ones, which is due to the geometric as
well as hydraulic attributes of such types of pumps.

In respect to a better comparability all results of the investigations shown in the following
diagrams were generated for operating conditions characterized by a flow rate of 180 m3/h and a
speed of rotation n = 1450 min-1.
mechanical losses
volumetric losses
100
h
[%] 95
impeller friction losses
90
vain losses
85

80
effective power
75
mechanical
70
gap
65
outer impeller
60
vains
55
10 100 nS
-1 n = 1450 min -1
[min ]
Q = 180 m3 h

Fig. 5.3: Partial losses within a centrifugal pump

38
AEAT-6559/ v 5.1 UNRESTRICTED

Depending on the value of specific speed which directly corresponds to the shape of the impeller
different influences on the losses caused by several pump components can be mentioned. For
pumps with lower specific speeds volumetric losses as well as losses due to disk friction at the
back and front shroud of the impeller are very significant. This also applies to the losses within in
the volute casing. For higher specific speeds the influence of blade friction losses within the
impeller dominates and mainly determines the level of the overall efficiency h. According to
former investigations based on statistically evaluated data it is known, that the largest potential
regarding an improvement of efficiency does exist at low specific speeds.

5.3 INFLUENCE OF DIFFERENT VALUES OF SURFACE

ROUGHNESS

95
h
[%] 90

85

80 Surface roughness kS
all smooth
75
0.024 mm
70 0.1 mm

65 0.2 mm
0.4 mm
60

55

50
10 100 nS
-1
[min ]
Fig. 5.4: The influence of surface roughness

Fig. 5.4. demonstrates the general influence of different values of surface roughness whereas all
inner surfaces of the pump show identical conditions.
It is remarkable that the gain of efficiency due to smoothing the inner surfaces of a centrifugal
pump is estimated more than 5 per cent points compared to pumps showing top quality sand-
cast-rough surfaces (surface roughness ks » 0.024 mm). Compared to pumps showing a very low
surface quality (e.g. due to low quality of manufacturing, corrosion or incrustation which can
result in a value for the surface roughness up to ks = 0.4 mm) a theoretical efficiency
improvement of even more than 20 per cent points could be estimated for pumps of very low
specific speed (ns = 10 min-1).

5.4 THE INFLUENCE OF SMOOTHING SEVERAL PARTS OF

PUMPS
Since smoothing a whole pump is a very cost intensive manufacturing process (especially for
small and medium sized pumps produced by a normal sand cast method) the influence of

39
AEAT-6559/ v 5.1 UNRESTRICTED
smoothing only several parts of the pump (i.e. volute, casing, outer surface of impeller, inner
surface of impeller) was investigated theoretically by the aid of the described software.

The result of this parameter study shows, that also in case of partial smoothing the maximum
efficiency improvement is to be expected for pumps with low specific speeds and can reach
values of roughly 6.5 per cent points (e.q. in case of smoothing the outer surfaces of a radial
impeller showing a origin surface roughness ks = 0.2 mm). With respect to an overall surface
treatment of the impeller the investigations showed also, that a smoothing of the inner surfaces is
primarily favourable for pumps with higher specific speeds (ns > 30 min-1), where the hydraulic
losses were mainly quantified by the flow velocity within the impeller.

The Influence of only partly smoothing the outer surface of the impeller

Due to the fact that especially for radial pump impellers (ns < 30 min-1) smoothing the outer
surfaces of the impeller front and back shroud (by turning) is a very efficient and less costly
procedure to reduce the losses, i.e. improvement of pump efficiency, the effect of smoothing the
impeller front and back shroud only partly was also investigated by an appropriate theoretical
parameter study.

As a result of this study it could be estimated, that smoothing only 40 % of the outer surface of
the back and front shroud (starting the turning process at the impeller outlet diameter D2) an
efficiency improvement of roughly 5.5 per cent points still can be estimated. This value decreases
to 3.5 per cent points in the case of smoothing only 20 % of the outer surface of the back and
front shroud. Du to this fact there is no need to smooth the impeller at smaller diameters where
turning gets more difficult because of the more complicated impeller contour.

As a validation of the above mentioned effect for the test pump (ns = 12 min-1) available at the
chair of Turbomachinery and Fluid Power at Darmstadt University of Technology an
improvement of efficiency of about 2 per cent points could be measured by smoothing 50 per
cent of the outer surface of the impeller back and front shroud (whereas the original surface
roughness ks before smoothing showed a very low (good) value of roughly 0.03 mm).

Fig. 5.5 summarises the theoretical estimation results obtained by the several parameter studies.
The labelled efficiency values roughly quantify the maximum gain of efficiency (in per cent
points) that can be expected by smoothing the wetted surfaces of a centrifugal pump showing a
surfaces roughness equal to a ks value of 0.2 mm. Depending on the specific speed of a pump the
efficiency values can be significantly less.

40
 whole pump: ... < 18.5%
AEAT-6559/ v 5.1

whole impeller volute: ... < 5% casing: ... < 4.5%

UNRESTRICTED

outer surface inner surface: ... < 5%

20% 40% 100%

of outer surface: ... < 3.5% of outer surface: ... < 5.5% of outer surface: ... < 6.5%

41
Fig. 5.5 : Maximum improvement of efficiency for several smoothing steps (estimated by theoretical calculations for medium
size pump of 180 m3/h))
AEAT-6559/ v 5.1 UNRESTRICTED

5.5 INFLUENCE OF DIFFERENT GAP CLEARANCES ON THE

INTERNAL LEAKAGE FLOW RATE

Volumetric losses are mainly caused by the existence of a suction-sided sealing gap which serves
as a throttle in order to reduce the secondary flow from the impeller out- to inlet as well as a an
additional pressure-sided sealing gap which usually belongs to the axial thrust balancing system
of a single-stage centrifugal pump. This internal leakage flows strongly depend on the clearance
of the sealing gaps. Fig. 5.6 shows the change in efficiency due to a variation of the gap clearance
(the change in efficiency refers to a smooth gap with a radial gap clearance of 0.6 mm).

Gap Clearance

6
Dh
[%] 5

2 Gap Clearance
0.1 mm
1
0.3 mm
0
0.5 mm
-1 0.7 mm
0.9 mm
-2

-3

-4
10 100 nS
-1
[min ]
Fig. 5.6: The influence of secondary flow through the sealing gaps

The above diagrammed estimation results demonstrate that a reduction of gap clearance for
instance from 0.6 mm to 0.3 mm can improve the pump efficiency about 3 per cent points. A
possible additional treatment to reduce the internal leakage flows is to furnish one or both gap
surfaces with circumferential notches (s. Fig. 5.7).

smooth notched

Fig. 5.7: Different types of sealing gaps

42
AEAT-6559/ v 5.1 UNRESTRICTED
The theoretically carried out parameter study showed, that notching gaps of the type usually used
in standard centrifugal pumps (cylindrical gaps with a relatively short gap length), only leads to a
slight improvement of the pump efficiency.

5.6 CONCLUSIONS

The presented results show, that for single-stage standard centrifugal pumps within a range of
specific speed ns = 10 min-1 up to ns = 100 min-1 (corresponding to values from 520 min-1 up to
5200 min-1 in US-Units), the highest potential for an efficiency improvement can generally be
found in the region of low specific speeds. All efforts aimed at an improvement of the surface
quality of several parts of the pump which are in contact with the flow cause a gain of efficiency.
With regard to the manufacturing costs which result from such additional surface treatments the
smoothing of the outer front and back shroud of the impeller can be proposed as a cost-efficient
procedure to improve the efficiency. Furthermore it could be shown that it is recommendable to
reduce the clearance of the sealing gaps to the smallest possible value in order to increase the
volumetric efficiency.

It should also to be mentioned, that the conditions of the surfaces as well as the sealing gaps
within a centrifugal pump normally depend on the time of operation, which means that there is a
strong necessity to check these parameters at reasonable intervals during the life time of a pump.

43
AEAT-6559/ v 5.1 UNRESTRICTED

6. Pump Efficiency deterioration

6.1 INTRODUCTION
Unlike much other industrial or commercial equipment, the efficiency of most pumps will
deteriorate, even in normal use. While this subject is strictly beyond the scope of this study, the
importance of this on the efficiency of the installed base of pumps in the EU means that it is
appropriate for the key issues to be summarised here for reference. Without maintenance, a
centrifugal cold water pump can eventually lose about 10% - 15% of its original efficiency.

Best maintenance practice both reduces the deterioration in efficiency, and improves the
reliability of pumps. The Life Cycle Costing approach to the design and maintenance of systems
shows clearly the benefits of proper attention to maintenance.

Although in many systems the lifetime energy costs will exceed those of maintenance, since
reliable operation of a pumping system is often of critical concern to the site where it is used,
effective maintenance has a higher priority than energy saving. Talking to companies about
maintenance usually gets a much better response than talking about energy saving, and so the
close link between maintenance and energy saving is important.

Improperly applied, a pump with a high “new” efficiency can be worn to total failure within just
a few hours. A thorough knowledge of the duty is therefore needed to reduce wear on the pump.
In fact, most pump problems are due to poor system design, a high percentage of which can relate
to the suction side.

6.2 THE COST OF EFFICIENCY DETERIORATION

There is no one agreed figure on how much efficiency a pump will lose if not maintained, but
UK data on many centrifugal pumps being used on cold water shows the following general
points:
· Most of the loss occurs in the first few years of life.
· After about ten years the loss starts to level out.
· The overall drop in efficiency for an unmaintained pump can be around 10 – 15%.
· An unmaintained pump can fail catastrophically after about 20 year’s service.
· The introduction of solids, increased temperature or speed will deteriorate the pump quicker.

These figures show that the effect of efficiency deterioration is of a very similar magnitude to the
spread of available efficiencies for new pumps. Figures 6.1 and 6.2 show the effect of
deterioration on pump performance.

44
AEAT-6559/ v 5.1 UNRESTRICTED

Figure 6.1: Effect of deterioration on pump characteristics

10 –
15%

Figure 6.2 The effect of maintenance on pump efficiency and lifetime

6.3 FACTORS EFFECTING PUMP DETERIORATION

Correct selection of pumps. The right type and capacity of pump must be selected, and the
correct material specified.
Correct mounting. The pump must be properly grouted to its bed, and the motor/pump shafts
correctly aligned.
Suction side piping. The inlet flow must be smooth, with no vortices. Narrow piping creates
low pressure, and hence makes pumps more susceptible to cavitation. (As a general rule, try to
keep at least 10 x diameter of pipe as a straight run into the pump.)
Entrained air. As little as 2% entrained air into a centrifugal pump can dramatically effect the
pump performance and reliablity.
Dry running. This causes heat build up and eventually seizure.

45
AEAT-6559/ v 5.1 UNRESTRICTED
Operation away from BEP. At the pumps best efficiency point the fluid enters and then exits
the pump efficiently. As operation moves further from BEP, the fluid recirculates within the
pump, causing out of balance forces, which will increase wear on seals, wear rings and bearings
through greater shaft deflection.
Contaminated bearing grease. Contaminated grease will cause faster bearing erosion.
Contaminated oil bath bearings. For these bearings, oil should be replaced at regular intervals.
(Tests have shown that 20ppm of water in oil can reduce bearing life from 24,000 hours to 2,200
hours.)

While failure of seals or bearings may be immediately apparent, the deterioration in pump
hydraulics is unlikely to be obvious. This can have many sources; chemical erosion, physical
abrasion, deposition. The use of glass or resin coatings can help to increase and maintain a
good hydraulic efficiency over a long period of time, and for larger pumps many users specify
these coatings as standard. Improvements in efficiency of 2-3% are typical. This is a practice
that should be encouraged.

6.4 APPROACHES TO MAINTENANCE

Maintenance is a time consuming and costly activity, but is essential to minimise Life Cycle
Costs and to minimise unplanned downtime. The approach to maintenance will be determined
by several key considerations:
· Criticality of the process
· Consequences of failure
· Severity of service
· Sensitivity of pump design to operating conditions

Deciding when to do maintenance on a pump is a balance between cost and system performance.
There are essentially three options:
· Do nothing until something goes wrong. Not generally recommended, and the maintenance
cost “savings” from this approach may not be as large as sites might hope. However, in some
less critical applications this might be appropriate.
· Periodic strip-down. This is good, but relies on a good knowledge of a particular pump and
its applications to know what time between maintenance is appropriate.
· Condition-based maintenance. Regular or on-line monitoring of the pump condition can give
accurate indication of when attention is needed. This is arguably the best method, but
demands investment in time/monitoring equipment to give the data needed.

6.5 MONITORING OF PUMP PERFORMANCE

At the very simplest, a periodic walk around a site, listening to and feeling for unusual vibration
from pumps, will give a clue as to which are starting to wear. At a more advanced level,
vibration can be monitored and recorded using standard vibration sensors, and alignment either
with conventional techniques or newer laser alignment.

On-line pump monitors are available with varying degrees of complexity, including measurement
of temperature, pressure, motor current etc, which can sound alarms to show abnormal operation.
Alternatively, and at lower cost, critical parameters can me measured with hand instruments, and

46
AEAT-6559/ v 5.1 UNRESTRICTED
performance over time recorded. If a pump operates at a steady duty, then simply measuring and
noting increases in current can indicate how efficiency is deteriorating.

In situ measurement of pump efficiency is costly, but where it is economic it is often done using
the thermodynamic method of pump efficiency determination. This method is already used in
the UK and elsewhere by pump users with large energy bills, in particular water companies. For
best accuracy, this technique does need to be done by trained personnel, and so it is usually done
by consultants rather than company personnel.

Regular maintenance should include adjustment of seals and lubrication of bearings as a

minimum, and adjustment of wearing ring positions where this is possible.

6.6 REPAIR AND RE-CONDITIONING OF PUMPS

Pump manufacturers prefer pumps to be returned to either them or an accredited agent for repair
or re-conditioning. This helps them control the quality of repair and hence customer satisfaction.
But particularly with smaller lower cost pumps, for speed of repair they will usually be repaired
by a local repairer, who may only have a very limited knowledge of proper pump repair.
Training of these organisations represents an important way to improve the efficiency of the
installed base of pumps. Training courses on this subject already exist, (Hydraulic Institute in the
US, and Association of Electrical and Mechanical Trades in the UK).

Whoever does the actual pump repair, it is important that they create the original clearances used
in a new pump, and only use correct parts, as pirate parts may not be exact replacements and so
can give poor service.

47
AEAT-6559/ v 5.1 UNRESTRICTED

7. References

A Guide to LCC Analysis for Pumping Systems. 2001 Hydraulic Institute & Europump

Energy Efficient Actions for Electric Motors, 1996. SAVE, for EC DGXVII.

Study of the technical and cost/benefit analysis of energy efficient improvements in industrial
three-phase induction motors, November 1999. SAVE, for EC DGXVII.

Proceedings EEMODs ’99 Energy Efficiency in Motors and Driven Systems, January 2000.
SAVE, for EC DGXVII. Published by Springer.

Improving the penetration of energy efficient motors & drives, 1999. SAVE, for EC DGXVII.

Proceedings International Workshop Accelerate innovation and market transformation of energy

efficient products, February 1999. EC/IEA/DETR.

United States Industrial Electric motor Systems Market Opportunities Assessment, December
1998. US Motor Challenge / Office of Industrial Technology.

Efficient Electric motor systems for industry - Report on Roundtable Discussions of Market
Problems and Ways to Overcome them, November 1993. US DOE Office of Energy Demand
Policy and Office of Industrial Technologies

Energy-Efficient Motor Systems, 1991. Nadel, Shepard, Greenburg, Katz, de Almeida. ACEEE,

Sulzer Centrifugal Pump Handbook, Sulzer Pumps, 2nd Edition,1998. Elsevier, Oxford, UK.
ISBN 1 85617 346 1.

Centrifugal Pumps, Karassik & McGuire, 2nd Edition, 1998. Chapman & Hall, New York.
ISBN 0-412-06391-3.

Pump Users Handbook, R.Rayner, 4th Edition, 1995. Elsevier, Oxford, UK. ISBN 1 85617 216
3.

Centrifugal Pumps and Allied Machinery, H.Anderson, 4th Edition, 1994. Elsevier, Oxford, UK.
ISBN 1 85617 231 7.

European Pumps and Pumping, 1994. CHW Roles & Associates Ltd, Sunbury on Thames, UK.
ISBN 0 907485 07 3.

Attainable Efficiencies of volute casing pumps, Europump, 1999. Elsevier, Oxford, UK. ISBN
1 85617 357 7.

48
AEAT-6559/ v 5.1 UNRESTRICTED
Centrifugal Pumps – The state of the art and new opportunities, IMechE, 2000.Professional
Engineering Publishing Ltd, London, UK. ISBN 1 86058 283 4.

Energy saving in the design and operation of pumps, IMechE,1996. Profesional Engineering
Publishing Ltd, London, UK. ISBN 1 86958 015 7.

49
AEAT-6559/ v 5.1 UNRESTRICTED

Appendices

CONTENTS

Appendix 1 EC guide to pump efficiency

Appendix 2 Pump energy consumption calculated from
Eurostat data (1998)
Appendix 3 Efficiency characteristics of centrifugal pumps

50
V 5.1 AEAT-6559-3 UNRESTRICTED

8. Appendix 1: EC Guide to
Pump Efficiency

NB. There have been further consultations with EUROPUMP since this Guide was produced.
This has resulted in a booklet entitled ‘European Guide to Pump Efficiency for Single Stage
Centrifugal Pumps’. This builds on the ‘EC Guide to Pump Efficiency’ by providing
additional information relating to average selections for efficiency, enabling the plots to show
areas of ‘Optimum Efficiency Selections’, ‘Efficiency Selections’ and ‘Lower Efficiency
Selections’. This booklet is available from:

http://energyefficiency.jrc.cec.eu.int

or hard copy from:

Mike Birks
Future Energy Solutions (FES)
B154
Harwell
Didcot
Oxfordshire
OX11 OQJ
United Kingdom
Tel: +44 1235 433468
Fax: +44 1235 433990
Email: mike.birks@aeat.co.uk

51
V 5.1 AEAT-6559-3 UNRESTRICTED
Purpose of the Guide
This guide is for anyone wishing to buy pumps and save money. The aim is to help you choose a pump
of good efficiency. This will reduce your energy costs. In some cases the saving from just one
additional point of efficiency can pay for your pump. We hope this Guide will also reduce the energy
used across the EU, benefiting everyone through a better environment. For the origin of the Guide see
Annex A.

Pump types and duties

This Guide should only be used for single stage centrifugal pumps handling clean water at up to 95°C.
Pump types covered are:

End suction pumps having their own bearings.

Close coupled end suction pumps, with the impeller on the motor shaft.
Double entry pumps having an axially split casing.

The applicable ranges of flow and head are shown on Figs 1 and 2 for pumps running at nominally
2900 and 1450 rev/min.

1000

DESC
Double Entry Split Casing
200 kW
96 kW
Head (m)

ESOB
100 End Suction Own Bearings
ESOB
DESC
ESCC
23 kW End Suction Close Coupled
ESCC

1.4 kW 2.5 kW

10
10 100 1000 10000
3
Flow (m /h)

Fig 1. Range coverage at best-efficiency duties at 2900 rev/min

52
V 5.1 AEAT-6559-3 UNRESTRICTED

1000

ESCC ESOB DESC

End Suction Close Coupled End Suction Own Bearings Double Entry Split Casing

680 kW
Head (m)

100

20 kW
DESC
67 kW

ESOB

ESCC
7.4 kW
1.2 kW
10
10 100 1000 10000
3
Flow (m /h)

Fig 2. Range coverage at best-efficiency duties at 1450 rev/min

How to use the efficiency plots

Performance curves published by manufacturers have been used to produce six plots (Figs 3 to 8).
Each point plotted shows the performance at the best-efficiency duty at maximum impeller diameter,
after correcting for pump head (see Annex B). A line is included, for guidance only, showing the
‘Catalogue ‘mean’ of best-efficiency points’. To use the plots:

Decide which type of pump you want.

Choose the flow and head at which you would like maximum efficiency.
Check that the chosen flow and head are within the ranges of Figs 1 or 2.
Get efficiency quotes from manufacturers (or use published information).
Enter your chosen flow and head on the plot that suits the quotes (Figs 3 to 8).
Read correction factor ‘C’ on the right-hand axis.
Add ‘C’ to the efficiency that has been quoted.
Plot (quoted pump efficiency + correction ‘C’) at your chosen flow.
If the point lies in the ‘Lower Efficiency’ area, seek higher efficiency quotes.

Important note: Although a quoted low efficiency may be due to poor pump quality, it is more likely to
be due to your chosen duty not coinciding with that pump’s best-efficiency point. Your flow may be
below or above the optimum for that pump. Your head will probably require a reduced diameter
impeller. A survey has suggested that you are unlikely to receive more than one quote in five above the
‘mean’ best-efficiency line.

Worked example:

Chosen pump type? End suction with own bearings.

53
V 5.1 AEAT-6559-3 UNRESTRICTED
Chosen duty for maximum efficiency? 80 m3/h at 110 m.
Quoted pump performance? 60% efficiency at 2900 rev/min.
(Check materials, suction performance, etc, are satisfactory)
Is chosen duty within ranges covered? From Fig 1, yes.
From Fig 3: ‘C’ = 14.
Plot on Fig 3: ‘Pump Efficiency + C’ = 60 + 14 = 74%.
Is this satisfactory? Fig 3 suggests that an additional 4 to 8 points is possible.
Action: Seek further quotes.

Loss of pump efficiency with time

A pump of high efficiency is of little value if the efficiency falls rapidly with time. You can minimise
this risk. Choose materials carefully, particularly for wear rings. Avoid high and low flow operation
in relation to your chosen duty. Ask for cast iron casings to be protectively coated if the water is known
to cause serious roughening due to corrosion.

Life cycle costs

Pump efficiency is just one factor which influences the cost of owning and running a pump. Attention
to the design of the system in which the pump operates and the method of pump control can achieve
very large savings. For more details see Ref 2.

85 50

80 45

75 40

70 35
'Lower Efficiency'

Efficiency correction = C
Pump Efficiency + C (%)

65 30

Catalogue 'mean'
60 25
of best-efficiency
points
130m head
55 110m 20

90m

50 15
70 20m

45 50m 10

40m
40 5
30

20m
35 0
30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
3
Pump Flow (m /h)

Fig 3. Efficiencies of end suction pumps with their own bearings at 2900 rev/min

54
V 5.1 AEAT-6559-3 UNRESTRICTED
90 50

85 45

80 40

75 35

'Lower Efficiency'

Efficiency correction = C
Pump Efficiency + C (%)

70 30

65 25

Catalogue 'mean'
of best-efficiency
60 20
50m head
points

40m
55 15

30m
50 10
15 10m
20
45 5
15

10
40 0
20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 400
3
Pump Flow (m /h)

85 60

80 55

75 50

70 'Lower Efficiency' 45

65 Catalogue 'mean' 40 Efficiency correction = C

Pump Efficiency + C (%)

of best-efficiency
60 points 35

55 30

50 25
130m head

45 110m 20

90m
20m
40 15
70m

35 10
50m

40m
30 5
30m
20m
25 0
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
3
Pump Flow (m /h)

Fig 4. Efficiencies of end suction pumps with their own bearings at 1450 rev/min
Fig 5. Efficiencies of close coupled end suction pumps at 2900 rev/min

55
V 5.1 AEAT-6559-3 UNRESTRICTED
90 50

85 45

80 40

75 35

Efficiency correction = C
Pump Efficiency + C (%)

'Lower Efficiency'
70 30
Catalogue 'mean'
of best-efficiency
65 points 25
50m head

60 40m 20

55 30m 15

25m
50 10
20m

15m
45 5

10m

40 0
30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
3
Pump Flow (m /h)

90 50

85 45

80 40

75 35
'Lower Efficiency'

Efficiency correction = C
Pump Efficiency + C (%)

70 30
Catalogue 'mean'
of best-efficiency
points
65 25

60 20

55 15
150m head

130m
50 10
110m

90

45 5
70

50
m
40 0
130 140 150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400
3
Pump Flow (m /h)

Fig 6. Efficiencies of close coupled end suction pumps at 1450 rev/min

Fig 7. Efficiencies of double entry axially split casing pumps at 2900 rev/min

56
V 5.1 AEAT-6559-3 UNRESTRICTED
95 50

90 45

85 40

80 35
'Lower Efficiency'

Efficiency correction = C
Catalogue 'mean'
Pump Efficiency + C (%)

75 of best-efficiency 30
points

70 25

65 20
140m head
15m

120m 20m
60 15
40m
100m
30m
55 80m
10

60m
50 5
50m
40
30
20
45 0
100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500
3
Pump Flow (m /h)

Fig 8. Efficiencies of double entry axially split casing pumps at 1450 rev/min
References

1. ‘Improving the energy efficiency of pumps’, Report produced for the European Commission (DG
Energy – SAVE), 2000.

2. ‘A guide to life cycle cost analysis for pumping systems’, Europump and Hydraulic Institute, 2000.

3. ‘Efficiency prediction method for centrifugal pumps’, Hydraulic Institute, 1994.

4. Anderson, H.H., ‘Centrifugal pumps and allied machinery’, Elsevier Advanced Technology, 1994.

5. ‘Attainable efficiencies of volute casing pumps’, Europump Guide No. 2, 1999.

6. Stoffel, B. and Lauer, J., Summary of the final report on the research project for VDMA
‘Theoretically attainable efficiency of centrifugal pumps’, Technical University of Darmstadt, 1994.

ANNEX A - Origin of this Guide.

This Guide and Ref 1 were produced by a Project Team comprising:

ETSU (UK) Project Leader

CETIM (France)
David T Reeves (UK)
NESA (Denmark)
University of Darmstadt (Germany)

57
V 5.1 AEAT-6559-3 UNRESTRICTED
The work was funded principally by the European Commission with additional funding provided by the
Project Team.

ANNEX B - Correction for pump head

Correction ‘C’, as shown on Figs 3 to 8, is based on pump flow, head and speed. It is actually a
correction for pump Specific Speed. Pumps of relatively low head (high Specific Speed) and high head
(low Specific Speed) lose efficiency due to unavoidable secondary hydraulic losses. Ref 3 shows a
curve to allow efficiency correction for Specific Speed. This is actually the correction proposed by
Anderson in Ref 4. Alternative curves to correct for Specific Speed can be derived from Ref 5 and Ref
6.

None of the these curves appear to provide an ‘optimum’ correction for the published manufacturers’
data analysed for this Guide. We have therefore chosen to use an intermediate Specific Speed
correction curve which produces the minimum scatter of the points plotted. There is no strict
theoretical justification for this approach but we believe that it serves well for the purpose for which
this Guide is intended. (Further information can be found in Ref 1.)

ANNEX C - Comparison with other sources

For practical reasons it has only been possible to source a limited amount of data to produce the plots
in this Guide. To assess how meaningful the results are, it is useful to compare them with other
sources.

The best efficiencies at optimum Specific Speed for end suction pumps having their own bearings at
2900 rev/min from Fig 3 are replotted below in Fig 9 and marked ‘Catalogue ‘mean’’. Additional
curves (all at optimum Specific Speed) are derived from:

Ref 3 (Fig 1A), marked ‘Hydraulic Institute’.

Ref 4 (Fig 7.3), marked ‘Anderson’.
Ref 5 (Figs 7 to 10), marked ‘Maximum practically attainable’.
Ref 6 (using Figs 7 to 10 of Ref 5), marked ‘Theoretically attainable’

From Fig 9 we deduce that the Catalogue ‘mean’ curve is suitable for the guidance purposes intended.
The ‘Hydraulic Institute ‘ANSI/API’’ curve is low, particularly at low flows. This is probably mainly
due to the use of relatively large wear ring clearances, as required for pumps in special materials or to
meet the American Petroleum Institute (API) Standard.

58
V 5.1 AEAT-6559-3 UNRESTRICTED

Theoretically
95 attainable

90

Maximum
Pump Efficiency at Optimum Specific Speed

85 practically
attainable

80
(%)

75 Catalogue
'mean'

Hydraulic
Institute
70 'Large
Anderson Pumps'

Hydraulic
65 Institute
'ANSI/API'

60

55
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
3
Pump Flow (m /h)

Fig 9. Comparison of efficiencies of end suction pumps from various sources

59
V 5.1 AEAT-6559-3 UNRESTRICTED

9. Appendix 2: Pump energy

consumption calculated
from sales data (1998)

60
 V 5.1 AEAT-6559-3 UNRESTRICTED

Source: National associations, public statistical offices and Eurostat, supplied to the study group by Europump
Total Total Total Total Total Av. Annual Life Installed Operating Load Pump energy Total pump
Country Type Production Imports Exports Market Market Price Sales Capacity Hours consumption energy consumption*
(Million DM) (Million DM) (Million DM) (Million DM) (million €) €/KW (kW) (Years) (GW) (Hrs/a) (%) (GWhpa) (GWhpa)
Germany End Suction Close Coupled 132 28 115 45 23.0 200 115,042 20 2.30 4000 65% 5,982
End Suction Own Bearing 500 23 236 287 146.7 200 733,715 20 14.67 4000 65% 38,153
Double Entry Split Case 15 4 28 -9 -4.6 200 -23,008 20 -0.46 4000 65% -1,196 42,900
UK End Suction Close Coupled 141 10 33 118 60.3 200 301,667 20 6.03 4000 65% 15,687
End Suction Own Bearing 142 15 100 57 29.1 200 145,720 20 2.91 4000 65% 7,577
Double Entry Split Case 77 2 24 55 28.1 200 140,607 20 2.81 4000 65% 7,312 30,600
France End Suction Close Coupled 104 24 7 121 61.9 200 309,336 20 6.19 4000 65% 16,085
End Suction Own Bearing 159 34 19 174 89.0 200 444,831 20 8.90 4000 65% 23,131
Double Entry Split Case 4 1 0 5 2.6 200 12,782 20 0.26 4000 65% 665 39,900
Belgium End Suction Close Coupled 52 7 5 54 27.6 200 138,051 20 2.76 4000 65% 7,179
End Suction Own Bearing 7 19 5 21 10.7 200 53,686 20 1.07 4000 65% 2,792
Double Entry Split Case 0 1 6 -5 -2.6 200 -12,782 20 -0.26 4000 65% -665 9,300
Denmark End Suction Close Coupled 1 6 2 5 2.6 200 12,782 20 0.26 4000 65% 665
End Suction Own Bearing 44 14 1 57 29.1 200 145,720 20 2.91 4000 65% 7,577
Double Entry Split Case 2.5 0 0 2.5 1.3 200 6,391 20 0.13 4000 65% 332 8,600
Finland End Suction Close Coupled 3 2 1 4 2.0 200 10,226 20 0.20 4000 65% 532
End Suction Own Bearing 13 6 1 18 9.2 200 46,017 20 0.92 4000 65% 2,393
Double Entry Split Case 0 0 0 0 0.0 200 0 20 0.00 4000 65% 0 2,900
Italy End Suction Close Coupled 124 12 31 105 53.7 200 268,432 20 5.37 4000 65% 13,958
End Suction Own Bearing 187 28 8 207 105.8 200 529,195 20 10.58 4000 65% 27,518
Double Entry Split Case 8 1 0 9 4.6 200 23,008 20 0.46 4000 65% 1,196 42,700
Netherlands End Suction Close Coupled 19 12 1 30 15.3 200 76,695 20 1.53 4000 65% 3,988
End Suction Own Bearing 10 21 12 19 9.7 200 48,573 20 0.97 4000 65% 2,526
Double Entry Split Case 7.5 0 0 7.5 3.8 200 19,174 20 0.38 4000 65% 997 7,500
Austria End Suction Close Coupled 6 8 2 12 6.1 200 30,678 20 0.61 4000 65% 1,595
End Suction Own Bearing 19 10 2 27 13.8 200 69,025 20 1.38 4000 65% 3,589
Double Entry Split Case 1 0 0 1 0.5 200 2,556 20 0.05 4000 65% 133 5,300
Spain End Suction Close Coupled 46 12 5 53 27.1 200 135,494 20 2.71 4000 65% 7,046
End Suction Own Bearing 79 10 7 82 41.9 200 209,633 20 4.19 4000 65% 10,901
Double Entry Split Case 36 0 0 36 18.4 200 92,034 20 1.84 4000 65% 4,786 22,700
TOTALS 817.1 4,085,285 81.71 212,435 212,400

* Only centrifugal pumps covered in this study

61
V 5.1 AEAT-6559-3 UNRESTRICTED

10. Appendix 3:
Efficiency
Characteristics of
Centrifugal Pumps

This section gives a quick overview of the fundamentals of choosing and using a pump for
best efficiency. This is not an exhaustive guide, but is designed to be just sufficient for non-
technical personnel to get a better understanding of the technical background to this work.
Other publications (such as the Enersave guide) are recommended as giving much fuller
explanations. The following applies to most types of centrifugal pumps only.

Part load characteristics of pumps

Pumps are always defined by the basic Pump characteristics below (figure A3.1). They show
the relationship between head, power and efficiency against flow. It is important to see just
how "peaky" the efficiency might be, showing that running at a duty (head and/or flow) below
rated duty is likely to lead to a significant reduction in pump efficiency. The Best Efficiency
Point (BEP) of a pump is ideally at the rated duty point.

Figure (A3.1) – Centrifugal pump characteristics

Running at reduced flow, or indeed above rated flow, will accelerate pump wear and might
give operational problems, (figure A3.2).

62
V 5.1 AEAT-6559-3 UNRESTRICTED

Figure (A3.2) – Onset of adverse effects when operating a pump away from its peak
efficiency flow

In practice pumps are usually found to be over-rated for the duty, since either the demand
varies, and/or the system designer has been prudent and over-sized the system. This problem
is generally over-come by throttling the flow with a valve. But deliberately restricting the
system flow is far inferior to better matching of the pump to the actual system requirements.
This is because it often causes additional wear, (figure A3.2), and reduces system efficiency,
(figure A3.3).

Figure (A3.3) – Illustration of the effect on efficiency of throttling a pump

One pump - different duties

63
V 5.1 AEAT-6559-3 UNRESTRICTED

Reducing the diameter of the impeller will make an existing pump run more efficiently at a
lower flow without throttling, (figure A3.4). Manufacturers commonly offer the same pump
housing (volute) with a range of impeller diameters because of this. Manufacturers may also
offer different styles of impeller to cope with different duties.

Figure (A3.4) – Pump characteristics showing various impeller diameters

Often a manufacturer will offer the same pump with different motor options to allow the one
pump to be used over a much wider range of duties. For instance, changing from the most
common 4-pole motor to a faster 2-pole motor will enable the same pump to deliver twice as
much peak flow and 4 times the head. (The effect of running a pump with 4 and 2 pole
motors is the same as what happens when running at 50% and 100% speeds as shown in
figure A3.5).

Variable Speed Drives allow a pump to operate efficiently over a wide range of speeds and
hence duties, and so are very good for saving energy, (figure A3.5). They are particularly
useful in systems where there is a wide variation in demanded flow.

64
V 5.1 AEAT-6559-3 UNRESTRICTED

Figure (A3.5) – Effect of speed reduction on pump characteristics

What this all means is that the same basic pump can serve different duties depending on both
the diameter of impeller fitted and the speed of the motor chosen. The power consumption
will vary with these parameters and with other factors such as viscosity. It is therefore not
useful, and indeed would be very misleading, to characterise pumpsets by power rating. The
size of motor required therefore needs to be determined for each application.

Selecting a pump
When selecting a pump, a manufacturer will use "tombstone" curves, which show their ranges
of pumps to cover a range of duties (figure (A3.6). Ideally, the duty you want will be roughly
20% below the maximum flow shown on the tombstone, which corresponds to the BEP of the
selected pump (each tombstone is built up from the individual pump curves in figure A3.4).
But for economic reasons they have to restrict the number of pumps that they offer. This
means that even a manufacturer of particularly efficient pumps may lose out, when quoting
efficiencies in competition with less efficient pumps whose BEP just happens to be nearer the
requested performance. The worked example in the box makes this clearer.

65
V 5.1 AEAT-6559-3 UNRESTRICTED

Figure (A3.6) – “Tombstone” curves for the selection of pumps by duty.

In addition, pumps wear over time (figure A3.7), but their efficiency can be maintained by
refurbishment figure (A3.8).

Figure (A3.7) – Effect of wear on pump characteristics

66
V 5.1 AEAT-6559-3 UNRESTRICTED

Figure (A3.8) – Average wear trends for maintained and unmaintained pumps

Worked example
This worked example shows the key issues that arise when selecting a pump on the basis of
efficiency.

67
V 5.1 AEAT-6559-3 UNRESTRICTED

BEP (B)

BEP (A) = 86%

= 82% 83%
84 80%

Desired (Quoted) 82 Pump B

Duty Point
80%

Actual (Installed) 77%

Duty Point Pump A
Head

Pump Efficiency Pump A Pump B

Flow
Peak (BEP) 82% 86%
At desired (quoted) duty point 81.5% 80%
At actual (installed) duty point 79% <77%

A User requests quotes for a pump at a particular Desired duty. Manufacturers A and B offer
the pumps shown, which are the best that they can offer from the ranges that they have.
(These two pumps are excerpts from the two manufacturers’ ranges of the type shown in
figure A3.6).

There are two important points:

While pump B has a higher BEP, at the desired duty, pump A actually has a higher efficiency
than pump B.

Over-specifying the duty means that at the actual installed duty, the efficiency of the pump
will be considerably less than quoted. (In this particular case it would be better to use a
reduced diameter impeller, or perhaps a quite different pump to either of those quoted for.)

Key points
· Pump efficiency can decrease significantly when the pump is operating away from the
designed Best Efficiency Point. Over-specifying the duty when specifying a pump will
therefore mean much increased energy costs.
· Pumps are not made to standard duties, which makes comparing efficiencies less simple
than with products that are made to standard duties, (such as motors).
· A manufacturer of pumps with a high design (BEP) efficiency may often lose out to
another manufacturers of a less efficient (BEP) pump, depending on where the actual duty
point requested lies within the performance curves of the pumps.
· The same pump will usually be offered with different impellers to give good performance
at lower duties.
· The same pump will often be offered with different speed motors to allow it to cover a
much wider range of duties.

68
V 5.1 AEAT-6559-3 UNRESTRICTED

· Pump performance and efficiency deteriorates over time.

· Power consumption of a particular pump, and hence rating of motor, is determined by the
application.

69