BSI 599 P.TEsting

BRITISH STANDARD BS 599:1966
Incorporating
Amendment Nos. 1, 2
and 3
Methods of testing
Pumps
UDC 614.894.24
BS 599:1966
Co-operating organizations
The Mechanical Engineering Industry Standards Committee, under whose

supervision this British Standard was prepared consists of representatives
from the following Government departments and scientific and Industrial
organizations:
Associated Offices’ Technical Committee Institution of Heating and Ventilating

Association of Consulting Engineers Engineers*
Association of Mining Electrical and Institution of Mechanical Engineers*
Mechanical Engineers Institution of Mechanical Engineers
Board of Trade (Automobile Division)
British Chemical Plant Manufacturers’ Institution of Production Engineers
Association* Locomotive and Allied Manufacturers’
British Compressed Air Society Association of Great Britain
British Electrical and Allied Manufacturers’ London Transport Board
Association* Machine Tool Trades Association
British Gear Manufacturers’ Association Ministry of Defence, Army Department
British Internal Combustion Engine Ministry of Defence, Navy Department
Manufacturers’ Association Ministry of Labour
British Iron and Steel Federation (H.M. Factory Inspectorate)
British Mechanical Engineering Federation* Ministry of Power
British Railways Board* Ministry of Public Building and Works
Crown Agents for Oversea Governments and Ministry of Technology — National
Administrations Engineering Laboratory*
Electricity Council, The Generating Board Ministry of Transport
and the Area Boards in England and Wales* National Coal Board*
Engineering Equipment Users’ Association National Physical Laboratory
Gas Council (Ministry of Technology)
Institute of Marine Engineers* Radio Industry Council
Institute of Civil Engineers Royal Institute of British Architects
Institution of Gas Engineers
The Government departments and scientific and industrial organizations

marked with an asterisk in the above list, together with the following, were
directly represented on the Committee entrusted with the preparation of this
This British Standard, having standard:
been approved by the
Mechanical Engineering
Industry Standards Committee Association of Hydraulic Equipment Institution of Electrical Engineers
and endorsed by the Chairman Manufacturers Institution of Mining Engineers
of the Engineering Divisional
British Hydromechanics Research Institution of Water Engineers
Council, was published under
the authority of the General Association Metropolitan Water Board
Council on British Pump Manufacturers Association
22 February 1966
© BSI 04-1999
First published, 1935 Amendments issued since publication
First revision, May, 1939
Second revision, February, 1966 Amd. No. Date Comments
1920 March 1976

The following BSI references
relate to the work on this
standard: 2359 September
Committee reference MEE/29 1977
and MEE/29/4
Draft for comment AB(MEE)2965 6145 January 1991 Indicated by a sideline in the margin
ISBN 0 580 32626 8

BS 599:1966
Contents
Page
Co-operating organizations Inside front cover
Foreword iii
Section 1. General
1 Scope 1
2 Classes of measurement 1
Section 2. Symbols and terminology
3 Symbols 1
4 Terminology 2
Section 3. Test procedure and standard installation conditions
5 Hydraulic pressure test 5
6 Duration and range of performance test 5
7 Control of head 5
8 Test arrangements 5
a. Standard test conditions 5
b. Pumping installation under standard conditions 8
c. Pumping installation under submerged conditions 8
d. Borehole and deep-well pumps 8
e. Self-priming pumps 8
f. Pumping installation under non-standard conditions 8
Section 4. Procedure for measurement of head, discharge,
speed and power input
9 General 15
Sub-Section A. Measurement of head
10 Pressure tappings and instrument connecting lines 15
a. Inlet head 15
b. Outlet head 15
c. Special layouts 15
11 Pump total head 16
12 Instruments for pressure measurement 16
Sub-Section B. Measurement of discharge
13 General 18
14 Weighing tank method 18
15 Volumetric tank method 19
16 Uncalibrated orifice plates, venturi tubes and nozzles 19
17 Uncalibrated notches, weirs and flumes 20
18 Dilution method of flow measurement — constant rate injection 20
19 Velocity area method of flow measurement 20
Sub-Section C. Measurement of speed and pump power input
20 Measurement of speed 21
21 Measurement of pump power input 21
a. Measurement of torque 21
b. Electric power measurements 21
c. Types and grades of measuring devices 21
Sub-Section D. Overall efficiency of a pumping unit
22 Overall efficiency 22
Sub-Section E. Fluctuations and variations in readings
and limits of error
23 Permissible fluctuations in readings and use of damping 22
© BSI 04-1999 i
BS 599:1966
Page
24 Limits of error 23
Section 5. Cavitation testing
25 General 25
26 Measurement of pressure 25
27 Measurement of temperature and vapour pressure 25
28 Measurement of discharge 25
29 Pump power input 25
30 Interpretation of results 25
Section 6. Tests with vaporizing liquids
31 General 26
32 Method of test 26
Section 7. Tests with viscous liquids
33 General 26
34 Measurement of pressure 26
36 Measurement of temperature 26
37 Measurement of viscosity 27
38 Permissible variation in viscosity 27
Section 8. Testing of pumps for liquids containing suspended solids
39 General 27
40 Capability of dealing with solids 27
41 Measurement of duty and efficiency 27
43 Measurement of weight density 28
Appendix A The estimation and analysis of limits of error 29
Appendix B Effect of installations and conditions 37
Appendix C Model testing and tests at modified speeds 38
Appendix D Values of some U.K. units in terms of SI units 40
Figure 1 — Illustration defining the positional datum for
different types of impeller pumps 4
Figure 2 — Flow normalizer 6
Figure 3 — Arrangements of single tappings and ring
connections for measurement of head using manometers 9
Figure 4 — Arrangement of single tappings and ring
connections for measurement of head using Bourdon gauges 13
Figure 5 — Device for measuring pressure-plus-velocity-head
and location of measuring points 17
Figure 6 — Baffle tank and arrangement for varying inlet pressure 18
Table 1 — Types and grades of measuring devices 22
Table 2 — Maximum permissible fluctuations in readings
for Class A, B and C measurements 23
Table 3 — Permissible damped fluctuations for Class A, B
and C measurements 23
Table 4 — Maximum permissible limits of uncertainty 24
Table 5 — Limits of variation between repeated measurements
of the same quantity 24
Table 6 — Permissible variations in kinematic viscosity 27
ii © BSI 04-1999
BS 599:1966
Foreword
This standard makes reference to the following British Standards:

BS 89, Electrical indicating instruments.
BS 269, Rules for methods of declaring efficiency of electrical machinery
(excluding traction motors).
BS 353, Methods of testing water turbine efficiency.
BS 1042, Methods for the measurement of fluid flow in pipes.
BS 1042-1, Orifice plates, nozzles and venturi tubes.
BS 1042-2, Pitot tubes1).
BS 1780, Bourdon tube pressure and vacuum gauges.
BS 3435, Measurement of electrical power and energy in acceptance testing.
BS 3680, Methods of measurement of liquid flow in open channels.
BS 3680-2A, Dilution methods. Constant rate injection.
BS 3680-3, Velocity area methods.
BS 3680-4A, Thin plate weirs and venturi flumes.
This British Standard which has been prepared under the authority of the
Mechanical Engineering Industry Standards Committee forms one of a series
providing methods of test for pumps. It supersedes BS 599, BS 1939 and also:
BS 722:1937, Borehole and well pump tests.
BS 723:1937, Sewage pump tests.
BS 724:1937, Vaporizing liquid pump tests.
Standard arrangements and procedures are described for testing a pump
individually without reference to its final installation conditions, or the effect
upon it of any associated fittings. These will be the usual conditions in which a
pump is tested at the maker’s works.
Pump performance may be affected by conditions of the final site installation and
procedures are described for carrying out “Standard tests” on certain types of
installation of which an overall performance is required. The conditions in which
pumps are finally installed, however, often do not permit reliable test
measurements, and recommendations are made concerning the procedure to be
adopted where the layout precludes tests in conformity with the standard or
where installation tests are required which cover plant ancillary to the pump
itself.
Since the publication of this British Standard in 1939, much experimental
evidence has been obtained, not only on the flow conditions at the inlet and the
outlet of pumps, but also on the precautions necessary when using various
methods of measuring flow and head. It has now been established that high
standards of accuracy cannot be obtained by using some of the simple and
convenient methods of measurement hitherto recommended.
The standard lays down different methods of measuring pressure, discharge and
power, and these have been placed in three Classes, A, B and C, in descending
order of accuracy, and the degree of accuracy to be expected in each class is
specified in Section 4.
It is expected that the methods of measurement used for testing the majority of
pumps will be those classified as Class C. The methods of measurement classified
as Class B may be employed where a higher degree of accuracy is required.
Class A methods of measurement will necessitate laboratory conditions, and may
be employed exceptionally on tests where the attainment of high accuracy is held
to justify the greatly increased time and cost expended; such tests might for
instance, be undertaken on prototypes or models of very large pumps which could
not themselves conveniently be tested.
1) In the course of preparation.
© BSI 04-1999 iii

BS 599:1966
Tests may be made in which the methods of measurement do not all fall into one
class and the method of arriving at the combined limit of error when calculating
the pump efficiency is described in Appendix A.
Notes on the estimation of limits of error, and the effect of disturbed flow on head
measurements are given in Appendix A and Appendix B.
Acknowledgement is made to British Hydromechanics Research Association
Reports numbers SP 625 and PR 685 for the material for Figure 2a and
Figure 2b.
NOTE 1 Where metric equivalents are stated the figures in British units are to be regarded as the
standard. The metric conversions are approximate. More accurate conversions should be based on the
tables in BS 350, “Conversion factors and tables”.
NOTE 2 In place of the customary, but incorrect, use of the pound and kilogramme as units of force,
the units called pound-force (abbreviation lbf) and kilogramme-force (abbreviation kgf) have been
used in this standard. These are forces which when acting on a body of mass one pound, or kilogramme
respectively, give it an acceleration equal to that of standard gravity.
NOTE 3 Except for the hydraulic pressure test (see Clause 5) the requirements for Class A, B and C
tests are given in BS 5316, “Acceptance test for centrifugal, mixed flow and axial pumps”,
Part 1, “Class C tests”, Part 2, “Class B tests” and Part 3 “precision class tests”.
A British Standard does not purport to include all the necessary provisions of a
contract. Users of British Standards are responsible for their correct application.
Compliance with a British Standard does not of itself confer immunity
from legal obligations.
Summary of pages
This document comprises a front cover, an inside front cover, pages i to iv,
pages 1 to 40 and a back cover.
This standard has been updated (see copyright date) and may have had
amendments incorporated. This will be indicated in the amendment table on
the inside front cover.
iv © BSI 04-1999
BS 599:1966
Section 1. General
1 Scope
This British Standard specifies the methods to be employed in the determination of the performance and
efficiency of pumps handling fluids which behave as homogeneous liquids.
The formulae quoted are based on the assumption that the liquid handled is clean water at a temperature
of 62 °F (16.6 °C) with a weight density of 62.35 lb/ft3 (998.75 kg/m3).
It also includes details of the special precautions to be taken where the tests are to be carried out using
liquids which are hot, are viscous, are liable to evolve vapour or gases, vary in density or contain solids in
suspension.
The standard does not specify requirements for the testing of oil hydraulic positive displacement (OHPD)
pumps.
2 Classes of measurement
To facilitate estimation of the accuracy with which a test is conducted, methods of measurement are
specified in three classes, A, B and C, in descending order of accuracy. The procedures applicable for
measurement in each class are laid down in the following sections, and the choice of method may be made
in accordance with the requirements of the test.
A 95 per cent degree of confidence is used throughout this standard. The estimated limits of error
corresponding to this degree of confidence in derived quantities (for example efficiency) can be calculated
in accordance with the provisions for Appendix A.
The limits of uncertainty laid down in this standard are those applicable to measurements taken and to
the quantities calculated therefrom; they do not apply to discrepancies between specified and actual
performance.
Section 2. Symbols and terminology

3 Symbols
Discharge (volume rate of flow) Q
Discharge (ft3/secor m3/sec Qc
Discharge (gals/min) Qg
Head H
Total inlet head Hi
Total outlet head Ho
Pump total head (Ho – Hi)
Net positive suction head NPSH
Pressure head hp
Inlet pressure head hpi
Outlet pressure head hpo
Positional head z
Velocity head hv
Velocity V
Inlet velocity Vi
Outlet velocity Vo
Gravitational acceleration g
Vapour pressure head hpv
Horsepower hp
Pump power input Pi
Pump power output (water horsepower) Po
© BSI 04-1999 1
BS 599:1966
Gallons gal
Cubic feet ft3
Cubic metres m3
Weight density w
Kinematic viscosity v
Pump efficiency ½
Overall efficiency ½o
Driving unit efficiency ½D
4 Terminology
The positional datum. (See Figure 1.) For centrifugal, mixed flow or axial flow pumps, the positional
datum shall be taken as the elevation of the highest point of the entrance edges of the first stage impeller
blades, whatever the disposition of the shaft.
For reciprocating pumps the positional datum shall be taken as the evaluation of the horizontal plane
through the centre of gravity of the volumes swept out by the rams, pistons or corresponding parts.
Head. This is the excess energy content of the liquid per unit weight above that of still liquid at absolute
zero pressure at the positional datum, expressed in units of length.
It comprises three components:
i) Pressure head. The pressure energy per unit weight of the liquid, it is equal to:
 
absolute pressure  absolute pressure 
-------------------------------------------------------------------------------------- ×  1 – ------------------------------------------------------------------------
weight density at atmospheric  2 × bulk modulus of fluid
pressure and test temperature  at test temperature 
In many cases the second term in brackets may be neglected (i.e. the fluid may be considered
incompressible). For the purpose of this standard, this may be done when the value of the expression
absolute pressure × 100
 --------------------------------------------------------------------------------------------------------------------------------------------- 
 2 × bulk modulus of fluid at test temperature
is one quarter or less of the permissible percentage for the appropriate class of pressure measurement.
NOTE Where the test temperature exceeds the boiling point of the liquid at atmospheric pressure, which is the case with many
boiler feed pumps, the pressure head may be taken as the absolute pressure multiplied by the mean specific volume. The mean
specific volume should be the arithmetic mean of the actual specific volume at the pump inlet and the specific volume
corresponding to pump inlet entropy and pump outlet pressure. For water these values can be obtained from steam tables.
ii) Positional head Positional head at any point in a liquid is the height of that point above the positional
datum.
iii) Velocity head. The kinetic energy per unit weight of the liquid. It is defined by the expression:
2
V
h v = -------
2g
Total inlet head. (Hi) The algebraic sum of the pressure head, the positional head, and the velocity head at
the point of measurement on the inlet side of the pump.
Total outlet head. (Ho) The algebraic sum of the pressure head, the positional head and the velocity head
at the point of measurement on the outlet side of the pump.
Pump total head. The head imparted to the liquid by the pump. It is the excess of the outlet head over the
inlet head, (Ho – Hi).
Net positive suction pressure head, (NPSH) more accurately termed the Net positive inlet head, is the inlet
head less the vapour pressure of the liquid expressed in feet of the liquid pumped.
Power. This is expressed in units of horsepower:
One horsepower is equivalent to 33 000 ft lb/min (745.700 watts).
2 © BSI 04-1999
BS 599:1966
Pump power input. The power absorbed by the pump, as measured at the pump input shaft.
Volume. This is expressed in either U.K. gallons or cubic feet or cubic metres.
1 U.K. gallon = 0.160 544 ft3 (0.004 546 092 m3).
Discharge (volume rate of flow). This is the volume rate of flow and shall be expressed in either gallons
per minute (gal/min) or cubic feet per second (ft3/s) or cubic metres per second (m3/s).
Weight density. The weight per unit volume of a liquid expressed in pounds per cubic foot (lb/ft3) or
kilogrammes per cubic metre (kg/m3).
Pump power output (water horsepower). This is defined by the expression:
Normal test condition. For the purposes of this standard a representative operating condition shall be
selected, which shall be known as the normal test condition.
Pump efficiency. The ratio of the pump power output to the pump power input.
Overall efficiency. The ratio of the pump power output to the rate of input of energy to the driving unit.
Driving unit efficiency. The ratio of the output power of the driving unit to the rate of input energy to the
driving unit.
© BSI 04-1999 3
4
BS 599:1966
Definitions
The positional datum for centrifugal, mixed flow or axial flow pumps shall be taken as the elevation of highest point of entrance edges of the first
stage impeller-blades, whatever the disposition of the shaft.
For reciprocating pumps the positional datum shall be taken as that of the horizontal plane through the centre of gravity of the volumes
For other pumps the nearest equivalent elevation to the above shall be taken.
© BSI 04-1999
Figure 1 — Illustration defining the positional datum for different types of impeller pumps
BS 599:1966
Section 3. Test procedure and test installation conditions

5 Hydraulic pressure test
All parts of the pump subject to pressure shall be submitted to a hydraulic test pressure not less
than 1.5 × the maximum pressure which can occur within the pump, for a minimum period of five minutes,
or such maximum pressure plus 1 000 lbf/in2 (70.3070 kgf/cm2) (6.894 76 × 106N/m2), whichever is the
less.
6 Duration and range of performance test

To verify the mechanical conditions of the pump it shall be run continuously for not less than one hour. The
duration of the test shall be sufficient to attain consistent results within this period.
Each test run shall begin only when steady test conditions have been established.
All measurements relating to a single test condition shall be taken simultaneously.
The time of all readings must be recorded and where multiple readings are taken to reduce random errors
(see Clause 24) they shall be taken at equal intervals of time.
As far as possible results should be computed as the test proceeds so that any errors and inconsistency in
reading can be detected and investigated during the test.
Where it is necessary to determine performance over a range of operating conditions a sufficient number
of observations shall be taken to establish the performance within the limits of error stated in Section 4.
7 Control of head
The head required for testing may be obtained by throttling in the outlet pipe. If it is necessary to reduce
the inlet pressure by throttling in the inlet pipe reference should be made to Section 5, Clause 27.
8 Test arrangements
a. Standard test conditions. Ideally a pump should be tested under the conditions in which it is intended
to be used. A standard test arrangement should be used if site conditions permit, or are unknown, or do not
need to be considered. This arrangement is described below, together with other permissible arrangements.
i) Inlet connections.
© BSI 04-1999 5
BS 599:1966
Figure 2a.
Figure 2a and Figure 2b are taken from BHRA reports numbers SP 625 and PR 685.
See Foreword.
Figure 2 — Flow normalizer
6 © BSI 04-1999
BS 599:1966
Figure 2 — Flow normalizer

A flow normalizer of diameter larger than that of the pump inlet may be used with a straight taper pipe at
the outlet end, but the specified length of parallel pipe must still be included.
Where the inlet velocity head does not exceed 5 per cent of the pump total head the inlet conditions may be
as specified for Class B measurements.
© BSI 04-1999 7
BS 599:1966
The nominal bore of the straight parallel pipe shall be the same as the bore of the pump inlet to which it is
connected.
ii) Outlet connections.
The nominal bore of the straight parallel pipe shall be the same as the bore of the pump outlet to which it
is connected.
b. Pumping installation under standard conditions. Standard tests can be carried out on a
combination of a pump and associated fittings of the final site installation or an exact reproduction thereof
or fittings introduced for testing purposes and taken as forming part of the pump itself, provided that
connections on the inlet and outlet sides of the whole combination can be made in accordance with
Subclause 8 a. Measurements shall then be taken in accordance with Subclauses 10 a i) and 10 b i).
c. Pumping installation under submerged conditions. Where a pump, or a combination of a pump and
its fittings, is tested or installed in conditions where standard connections on either inlet or outlet as
described in Subclause 8 a cannot be made owing to inaccessibility or submergence, measurements shall
be taken in accordance with Subclauses 10 a iii) and 10 b iii).
d. Borehole and deep-well pumps. Borehole and deep-well pumps cannot usually be tested with their
complete lengths of delivery main, and consequently the loss of head in the portions omitted, and the power
absorbed by any shafting therein, cannot be measured. A thrust bearing, if fitted, would also be more
lightly loaded than when eventually installed.
It is preferable, therefore, that the specified normal test condition (see Clause 4) of a pump of this type
should be based upon the conditions under which measurements can actually be made. The measurements
of head shall then be taken in accordance with Subclause 8 a and 8 b. Additional losses of head and power
caused by the remainder of the installation can then be assessed and the overall performance confirmed by
the test on site, if required.
e. Self-priming pumps. When a pump is specified to be self-priming it shall be tested at the specified
elevation above the still surface of the liquid and drawing through an inlet pipe of the specified volume and
disposition both vertical and horizontal.
When the pump is specified to deal with a mixture of water and air, it shall be tested with a controlled inlet
of air, through a calibrated device for air measurement.
When the pump is to be tested at a position where the local barometric pressure differs from that specified
or where the liquid is other than cold water, steps shall be taken to ensure that the NPSH conforms with
that under the specified condition.
f. Pumping installation under non-standard conditions. Where it is impracticable to test a pump in
accordance with the standard test conditions specified in Subclauses 8a and 8b, or where alternatively it is
required to ascertain the performance of a pump, not under standard conditions, but under the conditions
in which it will work when installed, or where it is desired to test the performance of a pumping installation
including items other than the pump itself, then the following decisions shall be made and recorded with
the test results:
i) The precise range of the items to be included in the installation to be tested.
NOTE Where auxiliary plant is involved, whether the consumption of each individual item is or is not to be debited to the
installation.
ii) The exact positions in the proposed layout at which inlet head and outlet head (or their difference) are
to be measured.
iii) The point in the circuit, the flow past which is to be credited to the installation.
iv) The nature of the supply of energy (e.g. mechanical, electric, fuel, steam, hydraulic, pneumatic).
v) The position, or positions, at which the supply of energy is to be measured.
vi) The method as laid down in the foregoing specification which is to be used for hydraulic
measurements; the method as laid down in other relevant British Standards to be used for the
measurement of power, fuel, etc., both quantitative and, where necessary, qualitative; the method of
measurement to be used where no British Standard exists.
vii) The extent to which the conditions pertaining during the tests may vary from the specified
conditions.
viii) Where lubricating oil or other consumable stores are to be taken into consideration, the methods of
measurement to be used for these.
ix) The proposed date(s), usually referred to the date of installation, of the tests, their duration and the
frequency with which readings are to be taken.
8 © BSI 04-1999
© BSI 04-1999
BS 599:1966
Figure 3a — Ring connections for measurement of head using liquid pumped in manometers
9
10
BS 599:1966
© BSI 04-1999
Figure 3b — Ring connections for measurement of head using heavy liquid in manometers
© BSI 04-1999
BS 599:1966
Figure 3c — Single tappings for measurement of head using heavy liquid in manometers
11
12
BS 599:1966
Figure 3d — Ring connections for measurement of head using heavy liquid in differential manometer
© BSI 04-1999
© BSI 04-1999
BS 599:1966
Figure 4a — Ring connections for measurement of head using Bourdon gauges on inlet and outlet
13
14
BS 599:1966
© BSI 04-1999
Figure 4b — Single tappings for measurement of head using Bourdon gauges on inlet and outlet
BS 599:1966
Section 4. Procedure for measurement of head, discharge, speed and

electrical power input
9 General
This section specifies the standard methods of measurement and instruments to be used for the
determination of the inlet head, outlet head, discharge, pump total head, speed and pump power input.
These methods are divided into three Classes A, B and C in descending order of accuracy, according to the
degree of precision to be attained.
The limits within which it may be expected, with a probability of 20 to 1, that a given reading will lie are
known as the 95 per cent confidence limits. The limits corresponding to each method are given in Section 4,
Sub-section C, together with the method of combination to determine the 95 per cent confidence limits for
derived quantities, for example, efficiency. A short explanation of the meaning and derivation of these
limits will be found in Appendix A.
Sub-Section A. Measurement of head

10 Pressure tappings and instrument connecting lines
Static pressure tappings shall conform with the requirements shown in Figure 3 and Figure 4, and shall
be free from burrs and irregularities and flush with the inside of the bore dimension.
The bore of the pressure tapping shall not be greater than 0.10 of the nominal bore of the pipe or less
than 1/16 in (1.587 mm) in diameter. The length of a pressure tapping hole shall not be less than twice its
diameter.
The bore of the pipe containing the tappings shall be clean, smooth and resistant to chemical reaction with
the liquid being pumped. Any coating such as paint applied to the bore must be intact. If the pipe is welded
longitudinally, tapping holes shall be displaced as far as possible from the weld consistent with their equal
spacing around the pipe circumference.
Care shall be taken to ensure that pipes directly connected to the pipe containing the measuring station
are of equal bore to and concentric with that pipe.
Instrument connecting lines must be free from leaks, of uniform bore and at least equal to the bore of the
pressure tapping, but not less than ( in (3.175 mm) in diameter. Valves, cocks and other fittings used on
the lines must be so constructed as to cause no restriction in the bore when fully open.
a. Inlet head. i) Installations in accordance with Subclause 8 a i).
ii) Installations in accordance with Subclause 8 b. Where a pump is tested in combination with fittings
forming part of the site or test installation the provisions of Subclause 10 a i) above shall be applied to the
inlet flange of the fittings and not to the inlet flange of the pump. This procedure debits against the pump
all energy losses caused by fittings on the inlet side.
iii) Installations in accordance with Subclause 8 c. The inlet head shall be as defined in Clause 4 and is
equal to the positional head of the still surface of the volume of liquid in which the pump is tested or from
which it draws, plus the pressure head equivalent to the absolute pressure on that surface. This
assumption debits against the pump all energy losses caused by fittings on the inlet side.
b. Outlet head. i) Installation in accordance with Subclause 8 a ii).
ii) Installations in accordance with Subclause 8 b. Where a pump is tested in combination with fittings
forming part of the site or test installation the provisions of Subclause 10 a i) above shall be applied to the
outlet flange of the fittings and not to the outlet flange of the pump. This procedure debits against the pump
all energy losses caused by fittings on the outlet side.
iii) Installations in accordance with Subclause 8 c. The outlet head shall be as defined in Clause 4 and is
equal to the positional head of the still surface of the volume of liquid into which the pump delivers plus
the pressure head equivalent on that surface. This assumption debits against the pump all energy losses
caused by fittings on the outlet side.
c. Special layouts. Provided that the installation conditions of Section 3 can be complied with,
measurements of delivery head of submerged pumps may be taken in accordance with Subclause 8 a.
© BSI 04-1999 15
BS 599:1966
Where, however, the outlet flange of a pump of this type is defined, for convenience, as being situated at
some distance from the pump proper, so that it is preceded by a length of outlet pipe, and by a bend or bends
forming an integral and invariable part of the installation, the outlet head measurement shall also be in
accordance with Subclause 8 b.
11 Pump total head

The pump total head is by definition the excess of the outlet head over the inlet head, and may accordingly
be determined from measurements of these two quantities. Alternatively, a device measuring differential
pressure may be used to obtain directly the difference between outlet and inlet pressure head. In this case,
for Class A, where the velocity head does not exceed 5 per cent of the pump total head, at normal test
conditions, and for Class B and Class C measurements, the pump total head is to be calculated as follows:
In this formula the velocity head is calculated as though there were uniform flow in both inlet and outlet
pipes.
12 Instruments for pressure measurement

General. In all cases provision must be made for purging the connecting lines between the pressure
tappings and any type of pressure measuring device. Before each reading it must be ensured that these
connecting lines are filled with liquid at a known temperature, or in the case of dry tube manometers that
they are clear of liquid.
Pressure measurements should preferably be made with liquid column manometers, and must be so made
for Classes A and B when the pump total head at the normal test condition is less than 50 ft (15.24 m).
16 © BSI 04-1999
BS 599:1966
Figure 5 — Device for measuring pressure-plus-velocity-head and location

of measuring points
© BSI 04-1999 17
BS 599:1966
Figure 6 — Baffle tank and arrangement for varying inlet pressure
Sub-Section B. Measurement of discharge

13 General
Any device or method which by calibration has been demonstrated to be capable of measuring discharge
with errors not exceeding the limits appropriate to a particular class of measurement (i.e., A, B or C), under
the conditions with which it will be used during the test, may be used for measurements to that class of
accuracy.
In the following clauses requirements are specified for methods which may be considered primary
standards, such as the use of weighing tanks or volumetric tanks, together with the requirements to be
fulfilled when using uncalibrated devices, or other methods.
14 Weighing tank method

This method is capable of measuring the mean value only of the discharge during the period concerned. It
is subject to the errors of the weighing and of the time-measuring apparatus employed, those due to the
non-instantaneity of diversion of the flow into and from the weighing tank and those involved in the
determination of specific gravity.
18 © BSI 04-1999
BS 599:1966
The time during which liquid is collected shall be not less than 500 times the sum of the times taken to
divert the flow into and from the measuring tank. Note that in the case where two tanks are used
alternately during the measuring period, the flow being diverted from one to the other, only the times of
initial diversion into and final diversion from the measuring system are to be taken into account, not the
times of the intermediate diversions from tank to tank.
15 Volumetric tank method

This method, as in the case of the Weighing tank method, is only capable of measuring the mean value of
the discharge during the period.
It is subject to the errors of mensuration and of the time-measuring device employed. Where possible,
initial calibration should be carried out by weighing liquid of known density into the measuring tank, and
in every case a leakage test of the tank must be carried out and correction made for leakage if necessary.
The tank must be adequately sheltered so that the level and the level-measuring devices etc. are not
disturbed by wind or rain.
Where a flow diverter is employed the method is subject to error due to the non-instantaneity of diversion
mentioned in Clause 14.
The method is in general subject to errors in measurement of levels which are not stationary and which
may be non-uniform. In such a case the levels must be simultaneously measured within stilling-tubes, at
not fewer than four widely separated positions within the tank, usually by hook-gauges, designed so that
the point is vertically below the point of suspension.
The time of collection shall not be less than 500 times the time of diversion as specified in Clause 14. The
change in level which is measured shall be not less than three feet.
16 Uncalibrated orifice plates, venturi tubes and nozzles

The measurement of discharge can be carried out using a square-edged orifice plate, or a venturi tube or a
nozzle, fitted in a straight run of pipeline which must be of circular cross-section and full of water on both
sides of the device, the two tapping points of the device being connected to a differential manometer.
Information on the characteristics of uncalibrated devices of these types are given in BS 10422), to which
reference should be made.
For the purposes of this British Standard the area ratio m of an uncalibrated device shall not exceed 0.36
in the case of an orifice plate or venturi tube, nor 0.25 in the case of a nozzle.
When liquid column manometers are used, water from the tapping points enters the manometer in the
same way as shown diagrammatically in Figure 3, for differential pump head measurements. The water
column tube can either be open ended (for low pressure only) or closed as shown, and the air connecting the
two limbs compressed to the amount required to permit the differential head to be read on the scale.
Alternatively mercury, or a second liquid heavier than water, is used in a suitable manometer. The
difference in level of the heavy liquid columns multiplied by the “effective” specific gravity gives the
equivalent head of water. The “effective” specific gravity is the difference between the specific gravity of the
heavy liquid or compressed gas and that of the water. For mercury this difference is 12.56 at normal
temperature. It is essential that there is no break in continuity (e.g. by air pockets through failure to vent
properly) of the water between the pipeline and the reading surface in the manometer.
The smallest scale division on any instrument used for observing the differential pressure across an orifice
plate, venturi tube or nozzle shall be not less than 0.05 in (1.27 mm) in length and shall represent a
pressure not greater than 1 per cent of the differential pressure corresponding to the flow at the Normal
Test Condition for Class A measurements, 2.5 per cent for Class B measurements and 5 per cent for
Class C measurements.
Where instruments other than those of the liquid column type are employed for measurement of the
differential pressure across an orifice plate, venturi tube, or nozzle, evidence shall be provided of
calibration against a dead weight or liquid column standard immediately before and after the test for
Class A and B measurements.
2) BS 1042, “Methods for the measurement of fluid flow in pipes”, Part 1, “Orifice plates, nozzles and venturi tubes”.
© BSI 04-1999 19
BS 599:1966
Care must be taken in the installation to ensure that cavitation shall not occur in the measuring device. In
the case of orifice or venturi flow measurements this is ensured only if the absolute pressure of the low
pressure tapping is higher than the vapour pressure of the liquid at the temperature in the pipeline. When
cavitation does occur, e.g., in the throat of a venturi tube, it is usually audible.
The coefficients quoted for uncalibrated orifices, venturi tubes or nozzles in BS 10423) are based on the
conditions that the flow is free from pulsations and surges, and has a normal velocity distribution
undisturbed by the close proximity of bends, elbows, valves and other fittings or by the pump itself. Their
values may be seriously altered if this is not the case. The necessary straight lengths of pipe required
upstream of the device for various types of fittings, or combination thereof, may be found by reference to
BS 10423).
Consequently any valve used for controlling the flow shall preferably be placed downstream of the orifice,
venturi or nozzle, and either adequate straight lengths of pipe must be provided upstream of the device, or
a flow normalizer installed in accordance with the provisions of Clause 8 must be used. In some cases it
may be necessary to reduce pressure by a valve close to the pump outlet flange. In this case it is essential
to provide either sufficient straight length of pipe (see BS 10423)) or a flow normalizing device
(see Clause 8) between the control valve and the measuring device. Care must be taken to ensure that the
indications of the device are not affected by air coming out of solution at the control valve. The presence of
air can usually be detected by operating the air-vents on the measuring device.
17 Uncalibrated notches, weirs and flumes

Specifications for the construction and installation of notches, weirs and flumes are given in
BS 3680-4A4), to which reference should be made.
For the purposes of this British Standard the smallest scale division of any instruments used for observing
head shall be not greater than that corresponding to 1.5 per cent of the flow for Class B and 3 per cent of
the flow for Class C measurements.
18 Dilution method of flow measurement — constant rate injection

This is a method for determining discharge by measuring the dilution of an injected solution, which in
certain circumstances is suitable for site testing. This method is equally applicable to the measurement of
flow in open channels and closed conduits, and when used the procedure adopted shall satisfy the
requirements specified in BS 3680-2A5).
19 Velocity area method of flow measurement

This is a method of determining discharge by integration of velocities measured at a number of selected
points in a plane normal to the mean direction of flow.
The devices commonly used in the application of this method are:
a) Current meters — for measurement of flow in open channels (see BS 3680-36)).
b) Current meters — for measurement of flow in closed conduits (see BS 3537)).
c) Pitot tubes (see BS 1042-28)).
Use of the above mentioned methods shall in all cases be carried out in accordance with the requirements
of the appropriate British Standard.
4) BS 3680, “Methods of measurement of liquid flow in open channels”, Part 4A, “Thin plate weirs and venturi flumes”.
5) BS 3680, “Methods of measurement of liquid flow in open channels”, Part 2A, “Dilution methods. Constant rate injection”.
6) BS 3680, “Methods of measurement of liquid flow in open channels”, Part 3, “Velocity area methods”.
7) BS 353, “Methods of testing water turbine efficiency”.
8) BS 1042, “Methods for the measurement of fluid flow in pipes”, Part 2, “Pitot tubes” (in course of preparation).
20 © BSI 04-1999
BS 599:1966
Sub-Section C. Measurement of speed and pump power input

20 Measurement of speed
21 Measurement of pump power input

The pump power input shall be derived from measurement of speed and torque, or determined from
measurements of the electrical power input to an electric motor of known efficiency, directly coupled to the
pump.
a. Measurement of torque.
b. Electric power measurements. Where the electrical power input to an electric motor coupled
directly to the pump is used as a means of determining the pump power input, the following condition
shall be observed:
i) The electric motor shall be operated at between 40 per cent and 100 per cent of its rated power.
ii) The power factor during the tests shall be between 0.7 and unity.
iv) The motor efficiency shall for Class A and Class B tests have been measured either by an
input-output test, or by back-to-back tests on similar machines.
The electrical instruments used during measurements of motor efficiency shall conform with the
requirements specified in Table 1, and conditions in Subclause 21 b iii) shall be observed.
The electrical power input to the driving motor shall be measured by the two wattmeter method in the
case of a.c. motors or by one wattmeter in the case of d.c. motors.
c. Types and grades of measuring devices.
i) Indicating instruments. The type and grade of indicating instrument used for measuring
electrical power shall be in accordance with BS 34359) and equal to or better than, those shown in
Table 1, the grade of accuracy being that given in BS 8910).
ii) Integrating meters. Single-element and double-element integrating meters for a.c. shall be equal
to or better than those shown in Table 1.
iii) Timing devices. Time measurements shall be obtained from timing devices having an error not
exceeding one second per hour.
iv) Instrument transformers. Voltage and current transformers shall be equal to or better than
those shown in Table 1, the class of accuracy of the instrument transformer being that given in
BS 8111).
9) BS 3435, “Measurement of electrical power and energy in acceptance testing”.

10) BS 89, “Electrical indicating instruments”.
11) BS 81, “Instrument transformers”.
© BSI 04-1999 21
BS 599:1966
Table 1 — Types and grades of measuring devices

Class of measurement
Measuring devices
Class A Class B Class C
D.C. voltmeter and
D.C. ammeter
Type Permanent magnet Permanent magnet Permanent magnet
moving coil moving coil moving coil
Grade Speciala Precision Industrial portable with a
scale length of 5 in or more
Shunt
Grade Speciala Precision Industrial
Wattmeter
Type Single element Single element Single element
electrodynamic electrodynamic electrodynamic
Grade Speciala Precision Industrial portable with a
scale length of 5 in or more
Induction integrating
meters Specialb BS 37-8c BS 37-8c
Transformer
Voltage Class AL Class BL Class B
Current Class AL Class AM Class BM
Frequency indicator Short-range Short-range Short-range
NOTE A potentiometer with standard resistors and voltage dividing box may be used instead of indicating instruments for
d.c. current and voltage measurement.
a
Class A requires higher accuracy grade instruments (and shunts) than at present covered by BS 89:1954, “Electrical indicating
instruments” (see the Foreword to that specification).
b
BS 37, “Electricity meters”, Part 10, “Alternating current, high precision kilowatt-hour meters for efficiency testing of electrical
plant”.
c
BS 37, “Electricity meters”, Part 8, “Alternating current precision kilowatt-hour meters for testing purposes”.
Sub-Section D. Overall efficiency of a pumping unit

22 To determine the overall efficiency of a pumping unit only the input and output energy are measured.
In this test the proportion of losses between driving agent and pump is not established, or any losses
associated with intermediate machinery such as gear box or variable speed device. The test measurements
of speed and torque are not required.
Sub-Section E. Fluctuations and variations in readings and limits of error

23 Permissible fluctuations in readings and use of damping
Short term oscillations about a mean value, which occur during the time that a single observation is being
made, are referred to as “fluctuations”; those changes in value which take place between one reading and
the next are termed “variations”.
The following table gives the maximum permissible fluctuations in readings for the various classes of
measurement. In the case of pressure, all valves or cocks in the instrument connecting lines must be fully
open when establishing the range of fluctuation.
Where an instrument is used for instantaneous measurement, the fluctuations of the instrument shall not
exceed the stated percentage of the mean value.
22 © BSI 04-1999
BS 599:1966
Where the construction or operation of a pump is such that pressure fluctuations of greater magnitude than
those given in Table 2 are present, pressures may, in certain cases, be measured by means of an instrument
capable of providing integration over at least one complete cycle of fluctuation. The calibration of such an
instrument must comply with the provisions of Clause 12 for the appropriate class of measurement.
Table 2 — Maximum permissible fluctuations in readings for Class A, B and C measurements
Permissible fluctuation — percentage of meana reading at normal test condition
Reading
Inlet pressure head

Outlet pressure head
Differential pressure head
between outlet and inlet
(where measured directly) ± 2.0 ± 4.0 ± 10.0
Differential head across
venturi meter of orifice
plate (where used)
Discharge ± 1.0 ± 2.0 ± 4.5
Speed ± 0.25 ± 0.75 ± 2.25
Torque
Power input to driving unit ± 0.50 ± 1.50 ± 4.5
a
In the case of inlet pressure head and outlet pressure head the permissible percentage fluctuation is to be calculated on the total
head.
a. Damping. Providing that the fluctuations do not exceed the limits given in the above table, damping
may be introduced in the instrument connecting lines to the pressure measuring instruments (including
those measuring differential head across a venturi meter or orifice plate where used) to reduce fluctuations
to not less than the values given in Table 3.
Table 3 — Permissible damped fluctuations for Class A, B and C measurements
Permissible damped fluctuation — percentage of meana reading at normal test condition
± 0.5 ± 1.0 ± 3.0

a
In the case of inlet pressure head and outlet pressure head the permissible percentage fluctuation is to be
calculated on the total head.
Any damping devices used must conform to the following conditions:

i) The resistance to flow of the device must be the same for flow in either direction through it.
ii) If damping is used on instruments measuring differential pressure, identical devices set to produce
identical resistance to flow must be used in each of the two connecting lines to the instrument.
24 Limits of error
Provided the conditions laid down in this British Standard regarding installation and instrumentation
have been met and the experimental procedure given in this Clause is followed, then the limits of error in
a test will be less than the values given in Table 4 for the appropriate class of each measurement.
© BSI 04-1999 23
BS 599:1966
Table 4 — Maximum permissible limits of uncertainty

(based on 95 per cent confidence limits of overall combined error as defined in Appendix A)
Quantity Class A Class B Class C
per cent per cent per cent
Rate of flow
Pressure head
Speed ± 0.5 ± 1.5 ± 3.0
Torque
Electrical power input
Overall efficiency (computed from discharge, head and
electrical power input) ± 0.7 ± 2.0 ± 4.0
Pump efficiency ± 0.8 ± 2.3 ± 4.5
The experimental procedure requires that repeated sets of observations of the measured quantities shall
be made at the Normal Test Condition, only speed and temperature being allowed to be controlled. Throttle
valve settings, water levels, glands, balance water etc., must be left completely unaltered. The differences
between these repeated readings of the same quantities will be a measure of the unsteadiness of the test
conditions, which are at least partly influenced by the pump under test as well as the installation.
A minimum of three sets of observations shall be taken, and the value of each separate measurement and
of the efficiency derived from the measurements in each set shall be recorded. The percentage difference
between the largest and smallest values of each quantity shall not be greater than that given in Table 5. It
will be noted that a wider tolerance is permitted if the number of readings is increased up to the maximum
requirements of nine readings.
Table 5 — Limits of variation between repeated measurements of the same quantity
(based on 95 per cent confidence limits as defined in Appendix A)
Number of sets of observations Maximum permissible percentage difference between largest
and smallest readings of each quantity
Rate of flow 
Pressure head  3 0.25 0.7 1.4

Speed  5 0.5 1.5 3.0
Torque 
 7 0.7 2.0 4.0
Electrical power input  9 0.9 2.5 5.0

Efficiency 
The arithmetic mean of all the readings for each quantity is to be taken as the actual value for the purposes
of the test.
No reading or selection of readings in the set of observations may be rejected because it lies outside the
limits given in Table 5. Instead, the cause must be ascertained, the condition rectified and a new complete
set of observations made, i.e., all the readings in the original set must be rejected.
if the excessive variation is not due to some error of procedure or instrumentation, and therefore cannot be
rectified, there are two alternatives.
i) The measurements of the quantity must be taken to have limits of error associated with the lower class
for which the variation can be tolerated, as given in Table 4 and Table 5.
ii) The limits of error may be calculated by statistical analysis as given in Appendix A.
It is unlikely that variations outside the requirements of Table 5 will occur for the majority of pump tests
carried out under well controlled conditions in a laboratory or at a manufacturer’s works. Worked examples
to illustrate the procedure are given in Appendix A (Clause A6, Examples 1 and 2).
24 © BSI 04-1999
BS 599:1966
Section 5. Cavitation testing

25 General
The performance of a pump may be affected by the formation of vapour filled cavities where local pressures
fall below the vapour pressures of the liquid being pumped. This phenomenon, known as cavitation,
depends on the NPSH at which the pump is operating and on the design of the pump. The limits within
which a pump may be operated depend greatly on the condition of the liquid being pumped and on the
resistance of the materials of construction to physical damage which may be caused by cavitation.
The object of cavitation testing is to determine the performance of the pump at various values of NPSH.
Some means of varying the NPSH is therefore necessary.
In all other respects the test installation shall conform with the provisions of Section 3, for the appropriate
class of measurement except that where Class C methods of pressure measurement are used the inlet
pressure may be varied by means of a valve followed by a baffle tank as shown in Figure 6, and having the
dimensions specified in that figure.
Where Class A and B methods of pressure measurement are used, methods of inlet pressure variation
which do not involve the restriction of the flow in the inlet pipe must be employed.
26 Measurement of pressure
The provisions of Section 4, Clause 10, shall apply, but special precautions are necessary to prevent errors
due to air or vapour bubbles in the gauge connecting lines and gauges attached to the inlet pipe. The
connecting lines should be as short as possible and provision for purging from an auxiliary supply or from
the pressure side of the pump should be made.
27 Measurement of temperature and vapour pressure

Where conditions are such that the inlet pressure approaches the vapour pressure of the liquid being
pumped, careful measurements of the liquid temperature must be made, since under these conditions small
variations in temperature may cause large changes in vapour pressure and hence in the value of the NPSH.
The temperature measuring device shall be situated not less than five inlet pipe diameters upstream of the
pressure measuring station with its sensitive element not less than ! of an inlet pipe diameter from the
pipe wall. The smallest division on the temperature scales shall correspond to a change in vapour pressure
representing not more than 4 per cent of the NPSH for Class C measurements, 2 per cent for Class B
measurements and 1 per cent for Class A measurements. Alternatively, suitable arrangements may be
made for direct measurement of the vapour pressure of the liquid in the inlet pipe.
28 Measurement of discharge
The flow shall be measured in accordance with the provisions of Section 4, Sub-section B.
Care must be taken in the installation to ensure that cavitation shall not occur in the flow meter. In the
case of orifice or venturi flow measurements, this is ensured only if the absolute pressure of the low
pressure tapping is higher than the vapour pressure of the liquid at the temperature in the pipe line. When
cavitation does occur, e.g. in the throat of a venturi tube, it is usually audible.
29 Pump power input

The pump power input shall be determined in accordance with the provisions of Section 4,
Sub-section C.
30 Interpretation of results
Where the highest degree of accuracy is required, tests must be carried out using liquid at the actual
conditions for which information is required.
Under other circumstances, calculations of performance at conditions other than those obtaining during
the test may be made, but it is not possible to specify limits of error for such calculated results. In particular
it should be noted that thermodynamic effects may cause deviation in actual performance from
performance calculated using affinity rules.
© BSI 04-1999 25
BS 599:1966
Section 6. Tests with vaporizing liquids

31 General
Tests on pumps intended for use with vaporizing liquids (other than complete cavitation tests) are made to
establish that the pump will perform satisfactorily at specified inlet pressures and temperature.
32 Method of test
Pumps intended for use with vaporizing liquids shall be tested in accordance with the provisions of
Section 5, Cavitation testing.
Section 7. Tests with viscous liquids

33 General
The performance of a pump may vary substantially with the viscosity of the liquid being pumped, and it is
not possible to specify any general rule whereby performance with a liquid of one viscosity can be used to
predict performance with a liquid of different viscosity.
If therefore it is desired to apply the provisions of this standard and the appropriate limits of error to
determine the performance of pumps intended to deal with liquids having a kinematic viscosity greater
than that of clean cold water, such tests must be conducted with the actual liquid under the conditions at
which it will be used.
34 Measurement of pressure
The provisions of Section 4, Clause 10 shall apply. Gauge connecting lines should be as short as possible,
and care must be taken to ensure that sufficient time is allowed for gauge readings to reach their final
value.
The provisions of Section 4, Sub-section B, shall generally apply, but for liquids with a viscosity
exceeding 2 centistokes, the procedures specified in BS 104212) to obtain the working coefficient for the
orifice or venturi tube, providing that the Reynolds number is not less than 20 000, shall be adopted.
For lower Reynolds numbers, BS 104212) describes special devices designed to give more constant
coefficients, and it is recommended that one of these should be adopted.
The following provisions apply to other forms of flow measurements:
a) no form of weir may be used.
b) if calibrated flow meters are used the calibration must be carried out with liquid at the conditions at
which the pump test is conducted.
c) if volumetric tanks are used, care must be taken to ensure that the liquid in the tank is free from air
bubbles and surface foam before readings are taken. In cases where it is not possible to observe this
precaution, volumetric tanks should not be used.
36 Measurement of temperature
Because of the marked variation of viscosity with temperature exhibited by most liquids it is essential that
temperature should be measured during the pump test, unless arrangements are made to measure
viscosity directly. For this purpose a temperature measuring device shall be provided with its sensitive
element not less than five or more than ten inlet pipe diameters upstream from the pump inlet flange, and
not less than ¼ of an inlet pipe diameter from the pipe wall. Reliable information on the viscosity and
density of the liquid over a range of temperature including that at which the pump test is conducted must
be available.
Where a venturi meter or orifice plate is used for flow measurement, similar provision for temperature
measurement must be made not less than five or more than ten pipe diameters upstream of the meter.
26 © BSI 04-1999
BS 599:1966
37 Measurement of viscosity
As an alternative to temperature measurement as specified in Clause 36 above, or where reliable
information on the viscosity and density variations with temperature of the liquid pumped is not available,
provision may be made to measure the kinematic viscosity of samples taken continuously during the pump
test.
Sampling points shall be provided not less than three nor more than five inlet pipe diameters upstream of
the pump inlet flange, and not less than three or more than five pipe diameters upstream of a venturi meter
or orifice plate used for flow measurement.
38 Permissible variations in viscosity

The following table gives the maximum permissible variations in kinematic viscosity during the course of
the pump test for the various classes of power or torque measurements:
Table 6 — Permissible variation in kinematic viscosity
Class of pump power input or Permissible variation in
torque measurement kinematic viscosity from that at
specified conditions
A ± 5 per cent
B ± 12 per cent
C ± 20 per cent
Section 8. Testing of pumps designed to deal with fluids containing solids in

suspension
39 General
Pumps designed to deal with fluids containing solids in suspension may, if desired, be tested with cold
water by the methods already described In addition, however, testing may be required usually on site, to
establish:
a. that the action of the pump is not objectionably impeded by the solids.
b. that the duty and efficiency when dealing with the solid-containing fluid are satisfactory.
NOTE This British Standard covers only fluids whose state is sufficiently stable not to change significantly (e.g. by settlement or
coagulation) during the pumping and measuring processes. Thus it is applicable to the pumping of normal sewage, but not, for
instance, to gravel-pumping.
40 Capability of dealing with solids

This must be determined by a prolonged run, usually on site, covering the whole range of duty required.
The duty (measured if necessary with the help of the devices described in Clause 41) must not, after a
settling period of one hour, deteriorate during such a run. Moreover, inspection of the internal parts of the
pump after the run must disclose that they are reasonably free from impediment or blockage and also that
no damage, nor wear in excess of that envisaged in the contract, shall have occurred.
41 Measurement of duty and efficiency

Measurement of Speed, Torque and Power-input shall be exactly as described for the case of pumps dealing
with fluids unencumbered by solids.
Measurement of head. In order to prevent any solids from entering the gauge-tubes, a counterflow of clean
liquid or of air must continually emerge at their points of connection. The amount of this flow must be
limited so that the velocity-head in the gauge tube never exceeds 1 per cent of the head, or of the differential
head, being measured. A source of clean liquid, or of filtered air, at a suitable pressure, regulated usually
by passage through a small orifice, must be provided, and provision also must be made for periodic
inspection, and if necessary cleaning, of the orifice.
Alternatively, head may be measured by the use of a suitably located flaccid diaphragm immersed in the
fluid, and containing clean liquid or air. Care must be taken to ensure that the diaphragm is truly flaccid —
i.e., that it has not, through leakage or other cause, become deformed, nor become buried in, or encrusted
with, deposit so that its flaccidity is impaired.
© BSI 04-1999 27
BS 599:1966
Where the specific gravity of the liquid in the gauge-pipes and/or manometers differs from that of the main
flow, correction shall be made exactly, as described in Clause 10.
This must be carried out by a method sensibly unaffected by the presence of the solids. Allowable methods
include the use of:
a. Weighing tank (corrected if necessary for weight density).
b. Volumetric tank.
c. Venturi tubes, venturi flumes, nozzles, other differential pressure devices listed for use with clean
water. In all cases the differential pressure head must be measured by the use of a counter-flow of clean
liquid or air, corrected if necessary for weight density, as described for measurement of pressure head.
d. Magnetic meters, subject to the limitations laid down for clear water tests.
e. Weirs and notches, as listed for use with clear water. Levels must be measured by means of a stilling
chamber, its connection to the main channel being kept clear by a slow counter-flow of clear liquid, and
correction to be made if necessary for specific weight.
Adequate means must be adopted to keep the edges of weirs, weir-plates and notches clear of obstructing
solids.
43 Measurement of weight density

This must be done by weighing a measured representative volume of the solid-encumbered liquid. if
necessary a number of samples must be measured and weighed.
28 © BSI 04-1999
BS 599:1966
Appendix A The estimation and analysis of limits of error

A1 General
The degree of uncertainty in any measurement (for example the efficiency of a pump) can properly be found
only by a special examination of all the sources and contributory causes of error in the test arrangement
and equipment being used and fluctuations and variations of the phenomenon being measured. It is usually
not practicable to make an exhaustive study of this sort for every measurement taken, and so assessments
of reliability of measurements will be more or less exact depending on the amount of evidence available and
the analysis made. What is required is a conclusion that the value indicated by the measurements is
unlikely to have differed from the true value at the time of the test by more than a stated amount.
Both systematic and random errors or uncertainties must be taken into account, and it is important to
remember that observations of unsteady phenomena, such as those encountered in pump testing, are not
necessarily exactly repeatable. For this reason, the limits of error cannot be arbitrarily predicted before the
tests have taken place.
Satisfactory estimates can be made, however, where a standard installation and calibrated instruments
are to be used to carry out the pump tests. The basis of such estimates will be the prior analysis of several
tests carried out in the installation with similar pumps to establish which Class limits are applicable for
the various measurements. This analysis should follow the procedure laid down in Clauses A2 – A5 of this
Appendix. Thereafter it will not be necessary to analyse fully each test carried out, provided that the overall
accuracy is rechecked completely at approximately yearly intervals and the limits of variation given
in Table 5 of Clause 24 are not exceeded for the required Class of measurement.
The user of the standard need not therefore estimate on the basis of the maximum limits of Table 4 of
Clause 24 once he has established the accuracy of his particular test arrangement.
A2 Analysis of errors
This standard is mainly concerned with the measurement of the basic quantities, pump head, discharge,
torque and shaft speed (or input power) at what is termed the Normal Test Condition.
The concept of a Normal Test Condition implies that during a test these quantities will remain sensibly
constant until the circuit resistance or speed is deliberately altered in order to reduce the uncertainty of
the values used to define that condition. To obtain the pump characteristics, pump head, input power and
efficiency are plotted against discharge, using results from a number of different test conditions. The limits
of error will change for each of these conditions, and repeated readings will be taken only for those points
for which it is considered important to calculate limits of error, or where a single reading appears to differ
from the trend of readings taken at other conditions.
This is the normal type of pump test, and Clause A4 of this Appendix deals with the method used to
calculate the limits of uncertainty of the results of each test condition studied in such circumstances.
Although it is assumed that the reader is familiar with the general terminology and concepts of statistics,
the principles particularly applicable to analysis of pump test results are given in Clause A3.
A3 Random and systematic errors
Repetition of a set of measurements using the same equipment can reduce the uncertainty of the random
error contribution, but has no effect on the systematic error. The latter can be reduced only if equipment of
a higher standard of accuracy is used, or if the same equipment can be calibrated to a higher standard. The
systematic error does not contribute in any way to the scatter of repeated observations.
a. Random errors. If the scatter of repeated observations of the same quantity about a mean value occurs
in a truly random manner, then a sufficiently large number will tend to be grouped in a pattern known
as the normal or Gaussian distribution of error. Other patterns are possible, but in the absence of
contrary evidence, random variation with the consequent normal distribution will be assumed.
If a set of n repeated observations of the same quantity is considered in which the normal distribution
of errors occurs, then the arithmetic mean, M, of the numerical values O1, O2, O3...On, is given by:
(A1)
© BSI 04-1999 29
BS 599:1966
and the standard deviation of these observations is:
(A2)
As n increases, the value of M tends to approach the true mean value of the instrument readings within its
fundamental limits of precision. For example, if a pressure 50 lbf/in2 (3.51 kgf/cm2) is to be measured on a
gauge with 1 lbf/in2 (0.070 kgf/cm2) graduations, the true mean value can be established to within
approximately ± 0.5 lbf/in2 (0.0351 kgf/cm2), but not within say ± 0.05 lbf/in2 (0.003 51 kgf/cm2), for which
an instrument of greater precision would have to be used.
If n were sufficiently large and the variations were normally distributed, it would be found that 95 per cent
of the readings would be grouped about the mean value within certain limits, the magnitude of which would
be related to the standard deviation Sn as given in Equation (A2). It could therefore be stated that the odds
were 19 to 1 that any single one of these observations lay within these limits, which are called the 95 per
cent confidence limits of error or uncertainty. Obviously 99 per cent of the readings would lie within wider
limits, which could be similarly expressed in terms of Sn, but for practical pump testing 95 per cent limits
are generally accepted as providing sufficient certainty, and are used throughout this standard.
The arithmetic mean of a small number of consecutive measured values of the same quantity will, in
general, differ by a greater amount from the true mean value than would the arithmetic mean of a larger
number of consecutive readings. The 95 per cent confidence limits attached to the mean of n readings are:
where x is a function of n as given in Table A1.

Table A1
n 3 5 7 9 30 Z
x 4.3 2.8 2.5 2.3 2 1.96
NOTE The values of x for n = Z associated with 99 per cent
and 99.5 per cent confidence limits are 2.5 and 3.4 respectively.
The procedure, therefore, to determine the 95 per cent confidence limits of uncertainty due to the scatter
of experimental results about the mean is as follows.
1) Determine the mean value of the readings, using Equation (A1).
2) Calculate the standard deviation from the value of M so obtained.
3) Multiply this value of Sn by the appropriate value of
The answer so obtained will indicate within what limits 95 out of 100 mean values obtained under identical
conditions might be expected to lie with respect to the true mean indicated value of the instrument being
used.
Table 4 in Clause 24 is based on maximum limits of random error of all quantities, measured and derived,
as given in Table A2.
Table A2
± 0.35 ± 1.0 ± 2.0
30 © BSI 04-1999
BS 599:1966
b. Systematic errors. The mean indicated value of an instrument reading may differ from the true mean
absolute value of the quantity being measured because of inherent and constructional limitations of the
instrument, and because of the limitations of its calibration. This source of uncertainty is called
systematic error. To confine this error within limits appropriate to the various classes of measurement,
conditions are laid down throughout this standard relating to scale lengths, divisions, calibration, and
installation. The maximum permissible limits of systematic error in the measurement of each basic
quantity are given in Table A3 and are the same as that given in Table A2 for random error.
Table A3
± 0.35 ± 1.0 ± 2.0
Derived quantities, such as pump power input calculated from separate measurements of torque and speed,
and pump efficiency, calculated from all the measured quantities, will have limits of systematic error which
can be calculated from those of the component measured quantities, as explained in Subclause A3 c.
Some instruments are less well suited for accurate measurement than others, although any instrument
where systematic error can be proved to be within the required limits will obviously be applicable for the
appropriate Class of measurement (see Clause A6, Example 3).
c. Combination of errors. The limits of systematic error in a derived quantity may be calculated by
combining in square law propagation the systematic errors of the linear component quantities as, for
instance,
Limit of percentage systematic error for pump efficiency =
(A5)
This combination of the separate errors of torque and speed measurement is the reason why the maximum
permissible limits for input power if based thereon are higher than those for torque and speed separately.
1⁄ 2
 2 2
For example: ±  0.5 + 0.5  = ± 0.7
 
© BSI 04-1999 31
BS 599:1966
The limits of random error for derived quantities can be calculated in this way only in respect of that part
of the error which results from inaccuracies of reading. That part of the random error in the derived
quantity which results from the departures of the quantities, when measured, from their mean values
cannot be so calculated except where (as happens only rarely in practice) the components are mutually
independent. The reason may be clearly understood by studying conditions in the region of pump best
efficiency. Each of the component quantities may be fluctuating in a random manner about its mean value.
Since, however, the quantities are dependent, the effect of those fluctuations may result in movements
along the lines of the pump characteristics. Since the efficiency curve is at zero slope at its maximum, quite
large changes of the component quantities may in consequence not be reflected in similar changes of
efficiency. Efficiency, or any other derived quantity, must therefore be calculated for each separate set of
readings, and the standard deviation of repeated values for the efficiency must be worked out in exactly the
same way as described in A3 a for the individual component quantities in order to estimate the limits of
random error. These limits may then be added to those of systematic error, using the square-law
propagation method, as, for instance,
(A6)
A4 Analysis of error at the normal test condition

The experimental procedure for taking repeated measurements for the Normal Test Condition is given in
Clause 24. If the maximum percentage difference between any two readings in a set is greater than that
given for nine consecutive readings in Clause 24, the test should be treated as one in which test conditions
are varying, and outside the scope of this standard. If, however, the conditions of Clause 24 are fulfilled,
the limits of error will be less than the maximum permissible limits given.
Closer limits will be obtained if the systematic and random errors are calculated according to Clause A3 of
this appendix, and combined as given to obtain overall limits of error for both basic and derived quantities.
A worked example of this procedure is given in Clause A6 (Example 2).
If n readings of a quantity are taken, and the percentage difference between the largest and smallest is
found to be 2d, the standard deviation cannot be less than:
2 - (A7)
S n (minimum) = ± d ------------
n–1
and cannot be more than:
n+1 (A8)
S n (maximum) = + d ------------- for an odd number of readings.
n
32 © BSI 04-1999
BS 599:1966
If a full analysis were carried out, therefore, the standard deviation would be found to lie somewhere
between the extreme values given by Equations (A7) and (A8). The maximum permissible percentage
differences given in Clause 24 have been based on Equation (A8) and therefore may be expected to give
limits of error less than those calculated from the worst possible distribution. The reduction obtained by
the complete analysis is not, however, large for small numbers of readings, and the extra work involved
would be justified only in special cases.
A5 Estimation of limits of error
The systematic error in the instruments being used will be determined from the results of their calibration.
The random error will be estimated either from the maximum limits as given in Clause A3, or from such
limits as have been shown to be associated with the particular test installation being used, as discussed in
Clause A1.
An example of estimation of limits of error is given in Clause A6 (Example 4).
A6 Worked examples
Example 1. Analysis of pump test results carried out under controlled conditions to ensure compliance with
Table 5 of Clause 24.
Example 2. More detailed analysis of above results to calculate closer limits of error.
Example 3. Calculation of systematic errors in measuring equipment.
Example 4. Combination of systematic and random errors.
Table A4 — Typical pump test readings
Observation Speed Rate of flow Pump head feet of Input power Efficiency
number rev/min gal/min water horsepower per cent
1 1447.2 1051.6 61.46 23.59 83.00
2 1447.3 1047.9 61.73 23.59 83.07
3 1447.3 1049.0 61.51 23.55 83.03
4 1447.2 1050.4 61.41 23.54 83.06
5 1447.1 1046.5 61.62 23.53 83.06
6 1447.1 1054.3 61.34 23.65 82.87
7 1447.2 1052.4 61.28 23.56 82.93
8 1447.2 1052.7 61.43 23.63 82.96
9 1447.1 1049.8 61.55 23.58 83.03
The above Table gives a typical set of pump test observations on which the ensuing examples will be based.
Example 1. Analysis of pump test results carried out under controlled conditions to ensure
compliance with Table 4 of Clause 24
Assume the above results had been obtained under the conditions laid down in Clause 24 for repeated sets
of observations, the instrumentation all being such that Class A requirements were satisfied throughout.
First, the minimum three readings (Observations 1, 2 and 3) would be taken. The maximum percentage
difference between largest and smallest readings of each quantity is then calculated as follows
Speed. negligible (less than 0.01 per cent).

Rate of flow. 3.7 × 100
1051.6 – 1047.9 = 3.7 i.e . = 0.35 per cent.
1051.6
Pump head. 0.27 × 100
61.73 – 61.46 = 0.27 i.e . = 0.44 per cent.
61.73
Input power. 0.04 × 100
23.59 – 23.55 = 0.04 i.e . = 0.17 per cent.
23.59
© BSI 04-1999 33
BS 599:1966
Referring now to Table 4, Clause 24 it will be seen that flow and head measurements satisfy Class B
requirements (maximum permissible percentage difference 0.7 per cent) but not Class A (0.25 per cent).
Speed and power satisfy Class A. If Class A standard is required throughout, at least two more sets of
observations will be required. Provided no wider scatter is found, Class A limits for the five
observations (0.5 per cent) will be satisfied.
Taking then the first five observations in Table A4 maximum differences are as follows:
Speed. negligible.
Rate of flow. 1051.6 – 1046.5 = 5.1 i.e. 0.48 per cent.
Pump head. 61.73 – 61.41 = 0.32 i.e. 0.52 per cent.
Input power. 23.59 – 23.53 = 0.06 i.e. 0.25 per cent.
Pump head only now lies outside the Class A limits, because of the low reading obtained in Observation 4.
Taking the next two readings, and repeating the calculations for Observations 1 to 7,
Speed. negligible.
Head and flow are now just outside the limits. Two more observations may be taken to attempt to establish
Class A requirements. Repeating calculations for Observations 1 to 9,
Speed. negligible.
which is the same as for the first seven calculations, since all readings in Observations 8 and 9 lay within
the limits of scatter of the previous seven observations.
The maximum permissible percentage difference for nine readings (Class A) is 0.9 per cent. This is
satisfied. It only remains now to take the mean value of each quantity measured:
Speed. 1 (1447.2 + 1447.3 + 1447.3 + 1447.2 + 1447.1 + 1447.1 + 1447.2 + 1447.2 + 1447.1) = 1447.2
9
Rate of flow. 1 (1051.6 + 1047.9 + 1049.0 + 1050.4 + 1046.5 + 1054.3 + 1052.4 + 1052.7 + 1049.8) = 1050.5
9
Pump head. 1 (61.46 + 61.73 + 61.51 + 61.41 + 61.62 + 61.34 + 61.28 + 61.43 + 61.55) = 61.48
9
Power. 1 (23.59 + 23.59 + 23.55 + 23.54 + 23.53 + 23.65 + 23.56 + 23.63 + 23.58) = 23.58
9
Efficiency. 1 (83.00 + 83.07 + 83.03 + 83.06 + 83.06 + 82.87 + 82.93 + 82.96 + 83.03) = 83.00
9
34 © BSI 04-1999
BS 599:1966
Example 2. Detailed analysis of above results to calculate closer limits of error

A more detailed analysis based on Example 1 can be made, using the pump head readings only as an
example.
First three readings.
1
Mean = --- (61.46 + 61.73 + 61.51) [From (A1)] i.e.
3
M = 61.57
Expressed as percentages, the three readings are:
O1 = 99.82
O2 = 100.26
O3 = 99.90
and M = 100.00
Hence Sn 1⁄
 ( 100 – 99.82 )2 + ( 100 – 100.26 ) 2 + ( 100 – 99.90 )2  2
=  -------------------------------------------------------------------------------------------------------------------------------------- 
 3–1 
1 ⁄2
 0.182 + 0.262 + 0.102 
=  ---------------------------------------------------------  (From A2)
 2 
= 0.235
Therefore 95 per cent confidence limits attached to the mean of the three readings are
xS n 4.3 × 0.235 (Value of x from Table A1)

= -----------------------------
----------
n 3
= 0.58 per cent
Similar calculations may be carried out in turn for 5, 7 and 9 readings.

The last gives:
M = 61.48
Sn = 0.137 per cent
95 per cent confidence limits xS n 2.3 × 0.137
= ± ----------- = -----------------------------
n 9
= ± 0.10
This is obviously well within the required ± 0.35 (Table A2).

Example 3. Calculation of systematic errors in measuring equipment
A single example can be taken — that of an orifice plate calibrated in situ against a volumetric tank.
At the mean rate of flow given in the basic example (Example 1), 1050.50 gal/min, the height of mercury
column is 25.63 inches (scale is assumed to be marked in tenths of inch, with vernier on ft3/s to read to
hundredths of an inch).
In the calibration, the meter coefficient was found to be substantially constant from 640 gal/min upwards.
Seven values of the coefficient were obtained in this range as follows.
k = 207.7, 207.9, 208.0, 208.2, 207.5, 208.5, 207.4
© BSI 04-1999 35
BS 599:1966
where kÆh is the flow in gal/min when h is the differential measured in inches on the mercury column. The
mean value of k is 207.9. The deviations of each of the measured values of k are therefore
Bk = – 0.20, zero, + 0.1, + 0.3, –0.4, + 0.6, –0.5
Squaring,
(Bk)2 = 0.040, zero, 0.01, 0.09, 0.16, 0.36, 0.25
Adding
C(Bk)2 = 0.91
Dividing by one less than the number of readings (h – 1) = 7 – 1 = 6
2
Then ( S n ) = 0.91
6
and 0.91 (Equation A2)

Sn = -----------
6
= 0.39
Expressed as a percentage,
0.39 × 100
S n = ---------------------------- = 0.19,
207.9
and the 95 per cent confidence limits attached to the mean are
xS n (Table A1)
2.5 × 0.19
± = ± --------------------------
n 7
= ± 0.18 per cent
The volumetric tank against which the meter was calibrated would itself have been calibrated. For
instance, carefully measured weights of water may have been added in increments, and the corresponding
levels recorded. All the factors which could contribute to possible sources of error would have been noted
e.g., timing accuracy, leakage and evaporation losses, temperature effects etc. and combined in square law
propagation. If the result was say, ± 0.12 per cent 95 per cent confidence limits of error in the measurement
of flow by the volumetric tank, the corresponding limits for the value of k above would be
1⁄
2 2
{ 0.18 + 0.12 } 2 = ± 0.22 per cent
This would now be carried forward as a systematic error in the subsequent use of the flowmeter.
Thus, if this meter had been used for the flow measurements tabulated, the 95 per cent confidence limits
of overall error attached to the mean of the nine readings could be calculated as follows.
Example 4. Combination of systematic and random errors
The mean rate of flow for the nine readings was 1050.5. Putting this equal to 100 per cent, the percentage
deviations of the other readings were
ÖQ per cent 0.06, 0.25, 0.14, 0.01, 0.38, 0.36, 0.18, 0.21, 0.07
(ÖQ)2 0.0036, 0.0625, 0.0196, 0.0001, 0.1444, 0.1296, 0.0324, 0.0441, 0.0049
Adding, C(ÖQ)2 = 0.4412
0.4412
Sn = = 0.235
9 – 1.0
95 per cent confidence limits of random error of the mean flow,
xS n 2.3 × 0.235
Ö R = ---------- =
n 9
= ± 0.18 per cent
36 © BSI 04-1999
BS 599:1966
In Example 2, the systematic error was estimated as

Ös = ± 0.22 per cent,
therefore the 95 per cent confidence limits of overall error are
2 2 1⁄ 2
ÖO = { ( B S ) + ( BR ) } [Equation (A3)]
2 2 1⁄
2
Ö O = { 0.22 + 0.18 }
= ± 0.28 per cent
Appendix B Effect of installation conditions

B1 Effect on pump performance
Subclause 8f of Test arrangements — draws attention to the effect of installation conditions on the
performance of a pump. Where it is the performance under installed conditions which is to be tested,
measures to ensure normalisation of the flow conditions would be neither necessary nor permissible.
B2 Effect on measurement of head
Departure from the normal regime of flow at the measuring station can cause inaccuracy in head
measurement in three main ways:
i) By causing static pressure variations which give rise to variation in pressure around the circumference
of the pipe.
ii) By causing mal-distributions of velocity which result in the true velocity-head differing substantially
from that calculated as though there were uniform velocity in the pipe.
iii) Where swirl occurs, by causing static pressure-variations across the pipe, though the pressure may
be uniform around the circumference.
The first type of error is to be dealt with by adherence to the provisions of Subclause 8 a ii) for checking
pressures measured at individual tappings.
The second is allowed for in the limits of error specified for Class B Methods of Measurement, and
eliminated by the traversing specified for Class A Methods.
In the third case, the presence of swirl may be detected by preliminary traverses with a yaw-detecting
instrument, by the use of the device shown in Figure 5, or by the following procedure:
In addition to the static pressure tappings at the specified measuring position, are provided a similar set
of tappings at a position where swirl is known to be absent (e.g. the sump or reservoir where such are
involved, otherwise a position sufficiently remote from the pump) which is designated the Reference
position. The fitting of valves or other variable restrictions between the two sets of tappings must not be
altered during the test. The difference between the heads as measured at the measuring position is then
plotted to a base of (discharge)2. Deviation of the resulting graph from a straight line passing through the
point representing zero pressure-difference at zero flow will indicate that swirl is present at the measuring
position.
Although absence of such deviation does not necessarily indicate zero swirl, in practice the errors arising
from an assumption of zero swirl where deviation is absent would not, at any rate for working conditions
approximating the designed duty, exceed those allowable in Class B or Class C tests.
Where swirl occurs this usually happens to a significant extent only at rates of flow outside ± 50 per cent
of the designed flow. In such cases the corrections to be applied to the reference-station heads for such flows
are those given by producing the straight line obtained as described above. The range covered by the
straight line must be not less than 25 per cent of the normal operating flow and the line must be defined
by not less than five observations.
The precautions to be taken to avoid errors in the measurement of discharge are described in Section 4,
Sub-section 4B.
© BSI 04-1999 37
BS 599:1966
Appendix C Model testing and tests at modified speeds

C1 General
Limitations on plant size or power available may make it difficult to conduct tests on certain pumps. In this
case, the desired information may often be obtained from tests on a scale model. Further, there are some
occasions where a full size machine cannot be run at the intended service speed, or where there are
unavoidable variations in speed during the course of a test.
C2 Testing Conditions
If the testing conditions of an impeller pump, or of a scale model of the pump, be so adjusted that the flow-
pattern within the pump, or within the scale model, either of which may be run at various speeds, is
geometrically similar to the flow pattern within the prototype under any initial set of conditions (design
conditions) then fundamental dynamical analysis shows that:
N 3
Q 1 = -------1 ⋅ Ì ⋅ Q 2
N2
N 2 3
H 1 =  -------1 ⋅ Ì ⋅ H 2
N 
2
where Q1 and H1 are the discharge and head under design conditions.
Q2 and H2 are the discharge and head under testing conditions.
N1 and N2 are the design and testing rotational speeds respectively.
Ì is the ratio dimensions of prototype/dimensions of model.
Necessary conditions that the design and testing flow patterns may be similar are:
a. That the relationship of dynamic and viscous forces shall be maintained. This is nearly the case
provided that, v1 and v2 being the kinematic viscosities of the fluid under design and testing conditions
respectively, the ratio of the design and testing Reynolds numbers does not depart too greatly from unity.
b. That no excessive variation occurs as between design and test conditions, in the proportion of the
boundary layer effects to those influences which govern the whole flow. This condition is usually satisfied
for clean cold water, provided the hydraulic mean depth (ratio cross-section area/perimeter) of any
cross-section of any flow passage is not less than:
In the case of Class C tests 1/ in (6.35 mm)

4
In the case of Class A and B 1 in (25.4 mm)
The model tests and modified-speed tests referred to above are not applicable to pumps handling viscous
or non-homogeneous fluids, nor to pumps operating under cavitating conditions.
With clean cold water the variation in v is unlikely to be significant. In that case, the limits of error
prescribed in this standard will remain valid for modified-speed tests providing
N2
4 /5 < < 5 /4
N1
Provided that the above limitations are observed, the performance at the normal test condition of a pump
may be determined by calculation from tests at modified speed.
Tests at modified speed or on a model in which the above limitations are not observed may, of course, be
carried out and the results obtained within specified limits of error. These limits of error may not, however,
be applied to calculations of performance at full speed or full size.
38 © BSI 04-1999
BS 599:1966
C3 Efficiency
Experience suggests that the efficiency of a full size pump is usually greater than that of a smaller model,
but insufficient information as yet exists on the comparative performance for a general “step-up” formula
to be recommended.
© BSI 04-1999
BS 599:1966
Appendix D Values of some U.K. units in terms of SI units

NOTE For unit symbols used in this table, see BS 1991-1.
Length Density
1 yd 0.9144 m 1 lb/in3 2.767 99 × 104 kg/m3
1 ft 0.3048 m 1 lb/ft3 16.0185 kg/m3
1 in 0.0254 m 1 lb/UKgal 99.7764 kg/m3
1 mile 1609.344 m
Force
1 pdl 0.138 255 N
1 lbf 4.448 22 N
Area
1 in2 6.4516 × 10–4 m2
1 ft2 0.092 903 0 m2 Pressure
1 yd2 0.836 127 m2 1 lbf/in2 6894.76 N/m2
2 6 2
1 mile 2.589 99 × 10 m
Energy (work, heat)
1 ft pdl 0.042 140 1 J
Volume 1 ft lbf 1.355 82 J
1 in3 1.638 71 × 10–5 m3 1 cala 4.1868 J
1 ft3 0.028 316 8 m3 1 Btua 1055.06 J
1 UKgal 0.004 546 092 m3
Power
1 hp 745.700 W
Velocity
1 ft/s 0.3048 m/s
1 mile/h 0.447 04 m/s Temperature
1 Rankine 5/9 of Kelvin unit
unit
Mass (= 1 Fahren- (= 5/9 of Celsius unit) heit unit)
1 lb 0.453 592 37 kg
Numbers printed in bold type are exact.
a
As defined by the 5th International Conference on Properties of Steam, London, 1956, and used by ISO.
40 © BSI 04-1999
blank
BS 599:1966
BSI — British Standards Institution

BSI is the independent national body responsible for preparing
British Standards. It presents the UK view on standards in Europe and at the
international level. It is incorporated by Royal Charter.
Revisions
British Standards are updated by amendment or revision. Users of

British Standards should make sure that they possess the latest amendments or
editions.
It is the constant aim of BSI to improve the quality of our products and services.
We would be grateful if anyone finding an inaccuracy or ambiguity while using
this British Standard would inform the Secretary of the technical committee
responsible, the identity of which can be found on the inside front cover.
Tel: 020 8996 9000. Fax: 020 8996 7400.
BSI offers members an individual updating service called PLUS which ensures
that subscribers automatically receive the latest editions of standards.
Buying standards
Orders for all BSI, international and foreign standards publications should be
addressed to Customer Services. Tel: 020 8996 9001. Fax: 020 8996 7001.
In response to orders for international standards, it is BSI policy to supply the

BSI implementation of those that have been published as British Standards,
unless otherwise requested.
Information on standards
BSI provides a wide range of information on national, European and

international standards through its Library and its Technical Help to Exporters
Service. Various BSI electronic information services are also available which give
details on all its products and services. Contact the Information Centre.
Tel: 020 8996 7111. Fax: 020 8996 7048.
Subscribing members of BSI are kept up to date with standards developments

and receive substantial discounts on the purchase price of standards. For details
of these and other benefits contact Membership Administration.
Tel: 020 8996 7002. Fax: 020 8996 7001.
Copyright
Copyright subsists in all BSI publications. BSI also holds the copyright, in the
UK, of the publications of the internationalstandardization bodies. Except as
permitted under the Copyright, Designs and Patents Act 1988 no extract may be
reproduced, stored in a retrieval system or transmitted in any form or by any
means – electronic, photocopying, recording or otherwise – without prior written
permission from BSI.
This does not preclude the free use, in the course of implementing the standard,
of necessary details such as symbols, and size, type or grade designations. If these
details are to be used for any other purpose than implementation then the prior
written permission of BSI must be obtained.
If permission is granted, the terms may include royalty payments or a licensing

agreement. Details and advice can be obtained from the Copyright Manager.
BSI Tel: 020 8996 7070.
389 Chiswick High Road
London
W4 4AL

BSI 599 P.TEsting

Uploaded by

Copyright:

Available Formats

BSI 599 P.TEsting

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

BSI 599 P.TEsting

Uploaded by

Copyright:

Available Formats

BRITISH STANDARD BS 599:1966

The Mechanical Engineering Industry Standards Committee, under whose

Associated Offices’ Technical Committee Institution of Heating and Ventilating

The Government departments and scientific and industrial organizations

1920 March 1976

ISBN 0 580 32626 8

This standard makes reference to the following British Standards:

1) In the course of preparation.

© BSI 04-1999 iii

Section 2. Symbols and terminology

Section 3. Test procedure and test installation conditions

6 Duration and range of performance test

Figure 2 — Flow normalizer

Section 4. Procedure for measurement of head, discharge, speed and

Sub-Section A. Measurement of head

11 Pump total head

12 Instruments for pressure measurement

Figure 5 — Device for measuring pressure-plus-velocity-head and location

Figure 6 — Baffle tank and arrangement for varying inlet pressure

Sub-Section B. Measurement of discharge

14 Weighing tank method

15 Volumetric tank method

16 Uncalibrated orifice plates, venturi tubes and nozzles

17 Uncalibrated notches, weirs and flumes

18 Dilution method of flow measurement — constant rate injection

19 Velocity area method of flow measurement

Sub-Section C. Measurement of speed and pump power input

21 Measurement of pump power input

9) BS 3435, “Measurement of electrical power and energy in acceptance testing”.

Table 1 — Types and grades of measuring devices

Sub-Section D. Overall efficiency of a pumping unit

Sub-Section E. Fluctuations and variations in readings and limits of error

Inlet pressure head

Class A Class B Class C

± 0.5 ± 1.0 ± 3.0

Any damping devices used must conform to the following conditions:

Table 4 — Maximum permissible limits of uncertainty

Class A Class B Class C

Section 5. Cavitation testing

27 Measurement of temperature and vapour pressure

29 Pump power input

Section 6. Tests with vaporizing liquids

Section 7. Tests with viscous liquids

38 Permissible variations in viscosity

Section 8. Testing of pumps designed to deal with fluids containing solids in

40 Capability of dealing with solids

41 Measurement of duty and efficiency

43 Measurement of weight density

Appendix A The estimation and analysis of limits of error

and the standard deviation of these observations is:

where x is a function of n as given in Table A1.

3) Multiply this value of Sn by the appropriate value of

± 0.35 ± 1.0 ± 2.0

± 0.35 ± 1.0 ± 2.0

A4 Analysis of error at the normal test condition

and cannot be more than:

Speed. negligible (less than 0.01 per cent).

Example 2. Detailed analysis of above results to calculate closer limits of error

Expressed as percentages, the three readings are:

xS n 4.3 × 0.235 (Value of x from Table A1)