1

Part 1 / Chapter 1: Administration

N T

ES
TIO

TIN
INSPEC
SUPPLEMENT

G
Eng. Yousuf Imran - CFPS
MA
IN T E N A N CE

CONDUCTING FIRE PUMP

FLOW TESTS
By Bruce H. Clarke

Editor’s Note: In many buildings, the fire pump is a critical element for delivering a sufficient
amount of water to a sprinkler or standpipe system. It is imperative that, where a pump is installed,
it is properly maintained and that the functional performance of the pump is confirmed through
the required flow tests in Chapter 8. This supplement looks at the process for conducting these tests,
and more importantly, how the information ascertained during the testing should be interpreted to
confirm pump functionality.

NFPA 25 does not provide a specific procedure or set of steps Tools/Equipment

that must be followed to conduct a flow test in accordance
The test equipment required should be in good working con-
with Chapter 8. Chapter 8 simply provides a requirement that
dition and of such quality as to not be questioned should a
the test must be conducted. The following sections provide
flow test fail. All gauges used in the test (i.e., those for the
an approach for conducting these flow tests. It is important
suction and discharge sides of the pump as well as the pi-
to keep in mind that the approaches shown below are not
tot tube) should have recent calibration within the past 6
mandated by NFPA 25 and might not be appropriate for all
to 12 months. Gauges should be calibrated at least annually
installations. The types and models of system components
by means of a dead-weight or calibrated tester throughout
installed vary greatly from installation to installation; there-
the range of operation. Calibration sheets should be kept for
fore, specific procedures outlined in the manufacturer’s litera-
each gauge and correction factors affixed to the back of each
ture should also be reviewed.
gauge before a test series is begun. The flow characteristics
of the test nozzles should be known, as well. While listed play
Sample Testing Procedure pipes have standard flow performance curves, some propri-
etary flow devices, such as hose monster test nozzles, have
Purpose
equipment specific flow charts. If flow meters are used, they
The following is a description of the testing procedures nec- require pre-test calibration. Typical tools and equipment used
essary to conduct the annual test required in 8.3.3.1. This test in the annual pump flow testing are as follows:
allows the owner/inspector to determine the fire pump’s
1. A handheld tachometer1 is required to measure the op-
capability based on its current operating condition. The owner/
erating speed of the fire pump and verify it on the en-
inspector can then compare the current data to the certified
gine.
shop test curve to quantify the pump degradation since its
2. A pitot tube together with a test-pressure gauge2,
installation.
suitable for the pressures to be expected, is required.

653

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 653 01/11/19 6:44 PM

 654 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests

[Usually, a 60 psi (4 bar) gauge is satisfactory.] Test- protective equipment (PPE) are typically required. If PPE
quality gauges should be used in accordance with ASME is not handy, then the start of pump testing can often be
B40.100, Pressure Gauges and Gauge Attachments, having delayed while the PPE is being located.
an accuracy of ±1 percent. The use of quality test gauges
produces results that are considered reasonably accurate
within the scope of the testing procedure. Care should
be taken to protect the gauges from rough handling.
While pitot tubes for fire protection flow testing are not
listed, the gauges used should be listed and should be
acceptable for water or water/air systems and not only
air systems.
A pitot-tube-and-gauge assembly, such as the one
shown in Exhibit S1.1, is indispensable in conducting
flow tests from hydrants and nozzles. The small open-
ing at the end of the tube — not over 1⁄16 in. (1.6 mm)
in diameter — is inserted in the center of the stream, in
a direct line with the flow, at a distance in front of the
opening equal to one-half the opening diameter. Veloc-
ity pressure is registered on the gauge attached to the
tube. Some available test nozzles, including an integral
pitot tube, are pitotless flow measuring features; they EXHIBIT S1.1 Typical Pitot Tube Assembly.
come with an attached test gauge that allows for the
direct measurement of flows without the use of a sepa-
Positive Displacement Pump Flow Tests. The pump flow
rate pitot tube.
for positive displacement pumps is completed to meet the
3. Test-pressure gauges suitable for measurement of
specified rated performance criteria. Typically, one perfor-
expected suction and discharge pressures on the fire
mance point is required to establish positive displacement
pump are required. Note that the gauge used on the
pump performance. Many positive displacement pumps also
suction side of the fire pump should be a compound
have special applications, such as a component in foam or
pressure. The vacuum gauge should have a pressure
water mist systems. For these types of pumps, manufactur-
range of at least twice the static pressure available from
ers’ data should be referenced. The pump flow test for foam
the suction supply. The gauge used on the discharge side
positive displacement pumps typically is completed using a
of the fire pump should have a range of at least twice the
flowmeter or orifice plate installed in a test loop back to the
rated working pressure of the fire pump.
foam concentrate tank or the inlet side of a water pump. The
4. A sufficient number of test nozzles are required,
flowmeter reading or discharge pressure should be recorded
including hoses for connection to test headers where
and should be in accordance with the pump manufacturer’s
needed [typically 50 ft (15.2 m) lengths of lined 2½ in.
flow performance data. If orifice plates are used, the orifice
(65 mm) hose might be used for this purpose], to per-
size and corresponding discharge pressure to be maintained
mit the measurement of a flow rate up to 150 percent
on the upstream side of the orifice plate should be made
of the rated capacity of the fire pump. Test nozzles can
available to the authority having jurisdiction (AHJ). Flow rates
include UL play pipes or other approved test nozzles.
should be as specified while operating at the system design
The flow characteristics of the test nozzles must be
pressure.
known.
In addition to manufacturers’ recommendations, systems
5. The various wrenches and other specific tools neces-
with positive displacement pumps should reference ANSI/
sary to allow for required adjustments during the test
Hydraulic Institute 3.6, Rotary Pump Tests, and Chapter 12,
and for the installation of required test equipment are
Chapter 13, and Annex C of NFPA 750, Standard on Water Mist
required.
Fire Protection Systems. ANSI/Hydraulic Institute 3.6 applies to
6. For electric pump tests, a multimeter and/or clamp-on
industrial rotary positive displacement pumps; it also includes
ammeter is required for taking voltage and amperage
detailed procedures on the setup and methods for conduct-
readings.
ing hydrostatic test and performance tests of such pumps.
7. A timing device is required to measure the duration of
the fire pump run (on acceptance tests). Steam-Driven Turbine Fire Pumps. While Chapter 13 of
8. A form for recording test data is required. NFPA 20, Standard for the Installation of Stationary Pumps for
9. While not required for the actual pump test data Fire Protection, is dedicated to these types of pumps, they
­collections, approved hearing protection and personal are rarely installed today due to maintenance and reliability

2020 NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 654 01/11/19 6:44 PM

 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests 655

c­ oncerns both on the driver and steam source. A steam tur- discover this during an annual test rather than an actual
bine driver for a fire pump should always be kept “warmed fire. Where the pump churn pressure will significantly
up” to permit instant operation at full rated speed. The auto- exceed the rated pressure for the system, the fire pump
matic starting of the turbine should not be dependent on any discharge control valve might be closed prior to the op-
manual valve operation or period of low-speed operation, eration of the fire pump to limit the exposure of the con-
and this should be verified as part of the test. Steam turbines nected systems. Consideration might be given to closure
are provided with governors to maintain a predetermined of the fire pump discharge control valve prior to conduct-
speed, with some adjustment for higher or lower speeds. ing the test. This could limit the exposure of the connected
Desired speeds below this range can be obtained by adjust- systems to the resultant pressure surge during the start-up
ing the main throttle valve. of the fire pump. Another method to avoid potential water
hammer effects is to partially open a test header outlet at
Vertical Turbine and Horizontal Split Case Pump Flow
start-up, and then close it for the churn pressure reading.
Tests. As required by 8.3.3.1, pump flow and pressure, as well
7. Replace suction and discharge pressure gauges with the
as drive shaft speed (rpm), is measured at a minimum of three
appropriate calibrated test gauges. (For a vertical turbine
points: churn (no flow), rated (100 percent flow), and peak/
fire pump, there will only be a discharge gauge.)
overload (150 percent flow).
8. Install flow test equipment on the test header as re-
quired, to allow for the maximum flow rate. This might
Procedure Steps
include the use of hoses connected to the test header
and extending them to flow measurement devices, such
Pre-Flow Activities/Planning.
as listed play pipes or other test nozzles.
1. If the site has a fire alarm system, notify the alarm moni- 9. Confirm that all test equipment, such as play pipes,
toring company and place the fire alarm system on test. hoses, play pipe test stands, and so forth, are adequate-
2. Notify all facility representatives that testing is going to ly secured or protected against movement. Manually
be conducted. In some cases, the amount of water flow- securing play pipes (i.e., by hand holding) is not accept-
ing on the ground might require extra maintenance able and could cause serious personal damage. Swivel
personnel. Or, suction pressure requirements could af- elbows, if installed on test headers, must be locked in
fect site production equipment (e.g., site cooling tow- place with set screws to avoid spinning during testing,
ers could trip off line due to low supply pressure, and so which could cause serious personal damage. In some
forth). cases, the connection of flow test equipment — such
3. Notify the local fire department and/or any other AHJ as play pipes — can be made directly to the test header.
that might want to attend, such as property insurance The use of a flow diffuser can aid in preventing move-
representatives. ment, as well as breaking the directed stream.
4. Complete the required inspection activities in Section 10. For testing using hose-connected flow-measuring devic-
8.2 to ensure pump system readiness for full flow test- es or UL play pipes directly connected to the test header,
ing. This maintenance should include checking battery ensure that the individual hose connection valves on
electrolyte levels and specific gravity, inspecting cable the test header are in the closed position. Then, slowly
conditions, and checking for any corrosion on pipes and open the main test header control valve pressurizing the
other system equipment. test header. For other flow measuring devices directly
5. Review the fire pump assembly nameplates, noting the connected to the test header, ensure that the main test
following: header control valve is closed. Then, fully open a single
a. Voltage rating (electric pumps) hose connection valve, readying the arrangement for
b. Rated speed the first flow point, with the flow being controlled by the
c. Rated capacity (flow) main test header control valve. Any time the use of an ad-
d. Churn pressure ditional hose connection valve is required with this latter
e. Rated pressure arrangement, the additional hose connection valves are
f. Overload pressure to be fully opened and the flow again controlled by the
6. Ensure all valves are in the proper position. Normally, position of the main test header control valve rather than
pump discharge valves that supply site fire systems or fire the individual hose connection valves.
system loops should be kept fully open throughout both 11. Check the surrounding area for potential uncontrolled
weekly testing and the annual pump test. These are closed water flow that could cause damage — plugged drain
sometimes due to concern that the “high pump pressure” pipes, local construction with open pits, berms subject to
or “pressure fluctuations” might damage the fire system hydraulic erosion, and so forth. Check any relief valve or
where pipe integrity is a concern. However, if fire mains cooling water discharge outlets as well as the discharge
are subject to damage due to pump pressures, it is best to points for the fire pump flow to ensure that there are no

NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems 2020

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 655 01/11/19 6:44 PM

 656 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests

obvious conditions that would prevent water from being safety purposes, the adjustment of the packing gland
discharged safely or cause direct damage in the immedi- should be made while the pump is not running. Care
ate vicinity. Where the discharge of water is to an area should be exercised to ensure the glands are not tight-
subject to potential freezing conditions, the facility rep- ened to the point of breaking.
resentative should be advised of the potential for icing 3. Monitor the fire pump operation for any unusual noise,
conditions. vibration, or other signs of malfunction.
12. Check the fire pump packing glands for a slight discharge 4. For electric-driven pumps, verify that the operation of
(slow drip) of water, adjusting the packing gland nuts as the circulation (casing) relief valve has a steady stream of
needed to achieve approximately 1 drip per second. If water to ensure proper cooling of the pump case.
the packing is completely dry before start-up, it could 5. Check the packing gland box, shaft bearings, and pump
overheat and fail once the pump is turned for the pump casing for overheating approximately every 5 minutes
test. The overheating of bearings could also quickly during the test. The packing gland box and shaft bear-
occur. For safety purposes, the adjustment of the ings might be warm to the touch, but the pump casing
packing gland should be made while the pump is not should remain cool throughout the test.
running. Care should be exercised to ensure the glands 6. If the fire pump is equipped with a main pressure relief
are not tightened to the point of breaking. valve, verify the operation of the pressure relief valve
13. For vertical turbine pumps, make sure the pump is such that outlet pressures do not exceed the pressure
completely primed. rating of the piping downstream of the fire pump. Usual-
14. In cases where a handheld tachometer will be used, ly, this rating is 175 psi (12.1 bar); however, some systems
verify where the point readings will be taken from and might be designed for higher pressures.
that reflective tape is installed as required. Some hand- 7. For electric-driven pumps, the motor voltage and cur-
held tachometers require advance preparation prior to rent on all phases (lines) should be recorded from an ex-
the test for proper measurement, such as the application ternal readout screen. Where no external screen exists,
of a reflective tape on the shaft. Care should be taken energized controller doors should not be opened; other
not to apply reflective tape on a shaft that already has it tests’ results should be used to determine if an issue
applied, because false “double readings” might occur. exists with voltage or current. If a controller door must
Since a reading will be required at each point during the be opened, it should be done so by individuals trained
actual test, care should be taken where the location tape and qualified in electrical hazards and equipped with
is applied, considering the proximity to casing spray, hot the needed safety equipment as outlined in NFPA 70E®,
engine exhaust, accessibility to the sides of a rotating Standard for Electrical Safety in the Workplace®. Also see
shaft, and so forth. 8.3.3.14 and Section 4.9, along with the associated com-
mentary, for additional guidance.
Flow Test Activities/Planning.
8. Record the suction and discharge pressures. Note that
1. For pressure-actuated fire pump controllers, simulate an only the discharge pressure is recorded for vertical tur-
automatic start for the fire pump by creating a pressure bine pumps.
drop in the sensing line to the fire pump controller. This is 9. Record the operating speed (rpm) of the motor.
typically accomplished by slowly opening the drain valve 10. The data recorded in Step 8 and Step 9 should be noted
on the sensing line located near the fire pump controller as the churn test point.
until the fire pump starts automatically. The use of the 11. Initiate the flow of the fire pump by slowly opening either
“start” button on the fire pump controller is not accept- hose control valves on the test header or by slowly open-
able for the purpose of simulating an automatic start. ing the main test header control valve, depending on
Note the start time of the fire pump, and record the start- the arrangement selected. As the valve is opened, take
ing pressure. flow measurements to determine the flow rate, adjusting
  For pressure-actuated fire pump controllers that use each valve position until a combined flow rate equal to
an automatic timer, an automatic opening of a sole- the rated capacity from the nameplate on the fire pump
noid valve in the sensing line to the fire pump controller is achieved. Where the flow rate with the valve fully open
might be used to simulate the automatic start of the fire is inadequate to meet this rated capacity, additional hose
pump. These systems must include a record of the pres- valves must be opened and adjustments made until the
sure drop on the pressure recorder for the controller. total measured flow rate equals the rated capacity. For
  For non-pressure-actuated fire pump controllers, the systems equipped with a main pressure relief valve, this
automatic start can be simulated by other means. should now be verified as closed and not flowing water.
2. Check the fire pump packing gland for a slight discharge If it is flowing water, adjust the pressure relief valve to
(slow drip) of water, adjusting the packing gland nuts as temporarily close.
needed to achieve approximately 1 drip per second. For 12. Repeat Step 8 and Step 9.

2020 NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 656 01/11/19 6:44 PM

 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests 657

13. The data recorded in Step 11 and Step 12 should be not- 29. Upon completion of all testing, notify the fire depart-
ed as the rated (or 100 percent flow) test point. ment and/or the alarm monitoring company (as need-
14. Continue to open hose valves and/or the main test header ed), as well as the facility representatives, that testing is
control valve, as in Step 11, and take flow measurements complete. Reset the fire alarm system as necessary.
to determine the flow rate. Also, adjust the valve(s) posi- 30. Complete the pump test data analysis to determine com-
tion until a flow rate equal to the 150 percent of the rated pliance with 8.3.3.1 and 8.3.7.1.
capacity from the nameplate on the fire pump is achieved,   At each flow test point, the quantity of water dis-
or until the maximum available flow rate is achieved, charging from the fire pump test assembly should be
whichever is lower. Care must be taken to ensure that stabilized as best as possible before taking pitot read-
the suction pressure (at the point of the city connection) ings. When using multiple outlets for pitot readings,
does not fall below an acceptable level. NFPA 20 allows a final readings on all outlets cannot accurately be made
suction pressure of 0 psi (0 bar) for fire pumps connected until all adjustments are completed. Due to hydraulic
to municipal supplies and –3 psi (–0.2 bar) for fire pumps principles, each time one test header outlet is adjusted,
taking suction from a grade level tank. Some local water the flow on others will be affected. That is, opening an
authorities might require a higher maintained pressure additional test header will cause the flow (and related
[typically 20 psi (1.4 bar) where this is a regulation], as the pitot readings) on others flowing to drop to a different
flow rate is increased to 150 percent. number than the reading taken before the additional
15. Repeat Step 8 and Step 9. valve was opened.
16. The data recorded in Step 14 and Step 15 should be not-
Fire Pump Operational Settings. The fire pump system,
ed as the overload (or 150 percent flow) test point.
when started by pressure drop, should be arranged as follows:
17. Shut down the flow by slowly closing the main test head-
er control valve to avoid water hammer. 1. The jockey pump stop point should equal the pump
18. If the run time for the fire pump has not reached 10 min- churn pressure plus the minimum static supply
utes, then allow the fire pump to continue operating pressure.
until a total run time of 10 minutes is reached. Recheck 2. The jockey pump start point should be at least 10 psi
for overheating periodically during the test, and then (0.7 bar) less than the jockey pump stop point.
shut it down manually upon completion of the test. 3. The fire pump start point should be 5 psi (0. 4 bar) less
19. Plot the test data against the manufacturer’s specifica- than the jockey pump start point. Use 10 psi (0.7 bar)
tions before removing test equipment to verify a retest is increments for each additional pump.
not required due to bad data recording and other reasons. 4. Where minimum run timers are provided, the pump will
20. Remove all attached test equipment from the test head- continue to operate after attaining these pressures. The
er, restore the closed position of all hose connection final pressures should not exceed the pressure rating of
valves, and reinstall the hose caps. In locations subject the system.
to freezing, check that the test header properly drains or 5. Where the operating differential of pressure switches
pumps out as necessary. does not permit these settings, the settings should be as
21. Remove the calibrated gauges, and reinstall the origi- close as equipment will permit. The settings should be
nal gauges on the suction and discharge side of the fire established by pressures observed on test gauges.
pump. Any gauges found out of calibration should be re-
placed. Evaluation of Results
22. Inspect and clean any installed intake screens. The text following this Testing Procedure section provides an in-
23. Verify that all of the fire pump supervisory signals (e.g., depth review of interpreting the results of the annual flow test
fire pump running, loss of power, phase reversal) re- and determining if the pump has seen more than a 5 percent
quired by NFPA 20 and NFPA 25 (see 8.3.3.13) have been degradation when compared to the certified shop test curve.
indicated at the fire pump controller as well as being
transmitted to any connected fire alarm panels.
References
24. Verify the main pressure relief valve is properly set and
adjusted. 1. American Society of Mechanical Engineers, Three Park
25. Verify the fire pump start-stop settings (see below). Avenue, New York, NY 10016-5990.
26. If closed, reopen the fire pump discharge control valve, 2. ASME B40.100, Pressure Gauges and Gauge A ­ ttachments,
and conduct a valve status test downstream of the closed 2013.
valve in accordance with 13.3.3.4. 3. National Fire Protection Association, 1 Batterymarch
27. Restore the fire pump to the automatic operating Park, Quincy, MA 02169-7471.
position. 4. NFPA 20, Standard for the Installation of Stationary Pumps
28. Complete a coupling alignment check. for Fire Protection, 2019 edition.

NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems 2020

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 657 01/11/19 6:44 PM

 658 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests

5. NFPA 70E®, Standard for Electrical Safety in the Workplace®,   Per 8.3.7.2.3, pump gross performance not adjusted for
2018 edition. speed must be graphically plotted and evaluated with
6. NFPA 750, Standard on Water Mist Fire Protection Systems, a comparison to owner-provided fire system demands.
2019 edition. And when graphically plotted, the annual gross pump
7. Hydraulic Institute, 6 Campus Drive First Floor North, performance flow should not be less than the system
Parsippany, NJ 07054-4406. requirement demands of all water-based fire protection
8. ANSI/Hydraulic Institute 3.6, Rotary Pump Tests, 2010 e­ dition. supplied by the fire pump.
3. Determine if the fire pump supplies 100 percent of rated
Endnotes flow.
There are many more actions for ITM that are required
1. All test equipment should be calibrated by an approved
as part of an acceptable annual fire pump test, but these
testing laboratory within 12 months prior to the test.
three items listed are the requirements for engineering anal-
2. Test-quality gauges should be used in accordance with
ysis of the actual pump data. Both are required to determine
ASME B40.100, Pressure Gauges and Gauge Attachments,
a “pass/fail” for the test. Other requirements include inspect-
having an accuracy of ±1%. The use of quality test gauges
ing general pump room conditions, including heating and
produces results that are considered reasonably accurate
cooling systems; inspecting and testing control valves and
within the scope of the testing procedure. Care should
controller alarm signals; testing controller start-stop pres-
be taken to protect the gauges from rough handling.
sures; and inspecting and testing numerous mechanical
Gauges should be calibrated at least annually by means
performance features such as master relief valves, circulat-
of a dead-weight or calibrated tester throughout the
ing relief valves, and heat exchanger solenoid valves. These
range of operation before a test series is begun. Calibra-
requirements are covered throughout NFPA 25, primarily in
tion sheets should be kept for each gauge and correction
Chapter 8.
factors affixed to the back of each gauge.
Whenever deficiencies are found during ITM procedures,
The procedure(s) outlined above is not mandated by NFPA 25 the owner (or designee) is responsible for repairs per 4.1.5.
and represents only one approach to conducting this test. Supplemental action after repair might be required per sum-
mary tables in each chapter, including Table 8.6.1 for fire
Fire Pump Flow Test Analysis pump systems. To help ensure a quick repair, it is also required
that any deficiencies and abnormalities observed during in-
The inspection, testing, and maintenance (ITM) tasks outlined spection or testing be reported promptly to the property
in Chapter 8 of NFPA 25 are intended to ensure that a fire owner or designated representative by the test contractor
pump, when installed, will function at or near (within 5 percent per 8.4.1. Additionally, any documented performance issues
of the original specification documentation) the performance should be formally recorded and retained for future reference
criteria the pump was installed for. Confirming functionality per Section 4.3. Thus, deficiencies must be formally reported
is achieved by properly maintaining the equipment and con- and presented to building owners, or those designated as
ducting the necessary tests to perform a q ­ uantifiable analysis responsible for initiating repairs, and not simply verbally
of the pump’s capability based on the data gathered during the communicated.
test. NFPA 25 has two specific data analysis requirements for Data analysis is also required for electric-driven fire
pump flow performance after conducting an annual fire pump pumps regarding recorded current and voltage readings per
test to determine pump acceptance. These are as follows: 8.3.7.2.8 for pump tests to be considered acceptable — but
1. Determine if a fire pump net performance, relative to the this is not part of the flow test analysis.
original design/installation, is acceptable. For data analysis requirement 1, per 8.3.7.2.3, a test must
  Per 8.3.7.2.3, pump net performance adjusted for rated be compared to the original pump performance. The original
speed must be graphically plotted and evaluated with a pump performance data is available typically from a manu-
comparison to the net pressure curve from owner docu- facturer’s pump curve. It can also be recorded from a fire
ments, copies of original manufacturers pump curves, pump design placard or nameplate. Data on these should be
pump nameplate data, or pump retrofit/rebuild docu- the same, as the nameplate data is usually stamped directly
ments. Additionally, per 8.3.7.2.3(3), when graphically from the pump curve data.
plotted the annual net pump performance flow should Paragraph 8.3.7.1.3 of NFPA 25 requires that, where
be no less than 95 percent of the original pump perfor- applicable, the speed-adjusted data be compared to the
mance across the complete flow performance curve (i.e., original pump performance. Net speed-adjusted data can
churn through 150 percent of rated flow). be compared to the owner document pressure curves,
2. Determine if a fire pump gross performance, relative copies of the original manufacturer’s pump curves, pump
to connected fire protection system demands, is nameplate data, or pump retrofit/rebuild documents per
­acceptable. 8.3.7.2.3.

2020 NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 658 01/11/19 6:44 PM

 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests 659

Original factory-new performance data, to use in compari- representing the performance of the pump at the factory that
son, can come from (or be generated from) numerous sources, should be seen once installed. Since these curves are not re-
as now listed. All four acceptable sources now listed should viewed and approved by an AHJ or outside authority at the
contain the exact same data, because all should be generated factory, the term could have been misleading when used with
from the original manufacturer’s pump curve. The term certi- pump curve documents. A certified manufacturer’s curve is the
fied was removed from the requirement because this was never same as an original manufacturer’s pump curve.
an NFPA 25–defined term. In relation to technical documents, On pumps where this information is missing, original
the term typically implies that something is official, is recog- manufacturers can be contacted for this information. Manu-
nized by an authority having jurisdiction, and is based upon a facturers typically keep original pump curve data stored by
review relative to a specific and recognized quality standard. pump serial numbers. If pump serial numbers cannot be
While it could be said that pumps are designed to be listed and l­ocated, many times information can be located by manu-
meet the requirements of NFPA 20, Standard for the Installa- facturers based on the address of the pump installation.
tion of Stationary Pumps for Fire Protection, the certified pump ­Exhibit S1.2 shows the relationship between a manufacturer’s
curves have always only been self-certified by m ­ anufacturers, curve and placard.

C
B D
E
A

A: Serial number: Unique to each pump. Curve and placard must match.
B: RPM: Pump and curve must match.
C: Rated point:
D: Churn pressure
E: 150% rated capacity

EXHIBIT S1.2 Relationship Between Pump Placard and Manufacturer’s Certified Pump Curve. (Courtesy of Liberty Mutual Insurance
Fire Lab)

NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems 2020

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 659 01/11/19 6:44 PM

 660 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests

Before an annual test can be completed, as implied by

the two primary data analysis requirements, some design in-
formation is required. This information should be collected,
reviewed, and analyzed before the annual fire pump test is
scheduled and made available to those responsible for the
analysis. This evaluation includes original pump performance
data system requirement demands of all water-based fire pro-
tection systems supplied by the fire pump. It is also helpful
to have recent and historical copies of past water supply test
results.

Fire Pump Testing Analysis Walkthrough

The following sections of this supplement use a real-world
scenario to walk through a fire pump testing analysis.

EXHIBIT S1.3 Manufacturer’s Original Pump Performance Data

Site Information
Placard. (Courtesy of Liberty Mutual Insurance Fire Lab)
For this example, assume the following data have been col-
lected, reviewed for accuracy, and are available before the
test (see Exhibit S1.3): 2014 Annual Fire Pump Recoded Raw Test Data. The in-
formation shown in Table S1.1 should be recorded directly
Original Pump Performance Data (Determined from One during the pump test. As noted in 8.3.3.10(2), if testing an
of the Means in 8.3.7.2.3). electric-driven fire pump, the voltage and current should also
Driver type: Diesel be recorded for the power supply analysis.
The plotting of hydraulic water data is typically complet-
Rated flow: 1000 gpm
ed using N1.85 graph paper. When plotted, the flow character-
Rated pressure: 85 psi istics of water flow — such as from a fire hydrant — visually
Rated rpm: 1760 rpm appear as a curve. Contrary to common sense where “two
Churn pressures (i.e., maximum net pressure): 97.0 psi points make a line,” for water flow, two points make a curve.
The hydraulic relationship between flow and pressure results
Net pressure at 150 percent: 56.2 psi
in a curve, and the more that is flowed, exponentially more
Building Sprinkler System Placard Design Information friction loss occurs and exponentially less corresponding
(Water-Based Fire Protection Systems Supplied by the pressure is available. Thus, when using standard graph paper
Fire Pump). to plot water supply tests, the results appear more as a “curve”
than a line. This can be complicated for water supply analysis
System 1, Office Area:
because several points must be plotted, and those drawing
Design: 0.10 gpm/ft2 over 1500 ft2 the plot points must be competent in using French curves.
Demand for design (including hoses): 268 gpm at 78 psi By using the N1.85 semi-exponential log paper, the lines on
System 2, Manufacturing Area: the vertical axis remain constant, but the lines on the horizontal
axis increase inwards as the grid moves from left to right, with
Design: 0.28 gpm/ft2 over 2500 ft2 N1.85 taking the curve out of the line. This allows accurate analysis
Demand for design (including hoses): 1254 gpm at 102 psi to be achieved easily and with only two points. Exhibit S1.4 and
System 3, Warehouse Area: Exhibit S1.5 illustrate water supply test data plotted on straight-
line graph paper and N1.85 paper. Exhibit S1.6 and Exhibit S1.7
Design: ESFR sprinkler system
also show the straight-line graph paper and N1.85 paper, this time
Demand for design (including hoses): 1440 gpm at 63 psi with both water supply and pump test data being plotted.
Pump Suction Water Supply Test Data (from City Fire While the benefits of using N1.85 graph paper for water sup-
Hydrant Test or a Hydrant on the Fire Pump Suction Supply). ply analysis do not transfer to fire pump net performance, it
is recommended for fire pump gross performance analysis; it
Static pressure: 45 psi is especially useful if questionable pump test results require a
Residual pressure at flow: 35 psi suction supply review. And, if used for fire pump net analysis,
Flow: 1250 gpm secondary plots are not required for fire pump gross analysis.

2020 NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 660 01/11/19 6:44 PM

 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests 661

TABLE S1.1 Pressure and Flow Data for Fire Pump Testing

Test Pitot S Q Ps Pd Pv Pn QN PN
Test Flow Info Reading (rpm) (gpm) (psi) (psi) (psi) (psi) (gpm) (gpm)

1-Churn No flow n/a 1777 0 45 140

2-Rated 2-1.75 in. play 32, 32 1763 1028 38 121

pipes

3-150% 3-1.75 in. play 32, 32, 32 1731 1542 29 84

Rated pipes

S = pump speed — typically recorded at pump shaft with tachometer

Q = pump discharge flow — combined pitot tube pressure readings converted to flow for given orifice
Ps = pump suction pressure — from reading suction flange gauge at flow
Pd = pump discharge pressure — from reading discharge flange gauge at flow
Pv = pump velocity head — correction calculated based on suction and discharge pipe sizes
Pn = pump net discharge — total pump discharge performance accounting for Pv
QN = pump flow — gross pump flow adjusted for speed
PN = pump pressure — gross pump pressure adjusted for speed

200 EXHIBIT S1.4 Water Supply

1
190 Test Plotted on Straight-
180 Test flow: 1
Line Graph Paper.
115 psi static
170 60 psi at 1200 gpm 1
160
150 1
140
130
120
Pressure (psi)

110
100
90
80
70
60
50
40
30
20
10
0
100 200 300 400 500 600 700 800 900 1000
Flow–X2 (gpm)

NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems 2020

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 661 01/11/19 6:44 PM

 662 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests

EXHIBIT S1.5 Water Supply Test

Plotted on N1.85 Graph Paper. 150
FLOW TEST
140 SUMMARY SHEET
N1.85
130

20 Test flow:
115 psi static
10 60 psi at 1200 gpm

00
Pressure (psi)

90

80
70

60

50
40

30
20

10
0
0 100 150 200 250 300 350 400 450 500 Scale A 550 575
200 300 400 500 600 700 800 900 1000 Scale B 1100 1150
400 600 800 1000 1200 1400 1600 1800 2000 Scale C 2200 2300
800 1200 1600 2000 2400 2800 3200 3600 4000 Scale D 4400 4600
Flow (gpm)

EXHIBIT S1.6 Water Supply and

Pump Test Plotted on Straight-Line 200
1
Graph Paper. 190
180 Test flow: 1
115 psi static
170 60 psi at 1200 gpm 1
160
150 1
140
130
120
Pressure (psi)

110
100
90
80
70
60
50
40
30
20
10
0
100 200 300 400 500 600 700 800 900 1000
Flow–X2 (gpm)

2020 NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 662 01/11/19 6:44 PM

 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests 663

150 EXHIBIT S1.7 Water Supply

FLOW TEST and Pump Test Plotted on
140 SUMMARY SHEET N1.85 Graph Paper.
N1.85
130

20 Test flow:
115 psi static
10 60 psi at 1200 gpm

00
Pressure (psi)

90

80
70

60

50
40

30
20

10
0
0 100 150 200 250 300 350 400 450 500 Scale A 550 575
200 300 400 500 600 700 800 900 1000 Scale B 1100 1150
400 600 800 1000 1200 1400 1600 1800 2000 Scale C 2200 2300
800 1200 1600 2000 2400 2800 3200 3600 4000 Scale D 4400 4600
Flow (gpm)

There are drawbacks to using the N1.85 graph paper, too. E­ xhibit S1.9. If the gross pump discharge has a flow and pres-
A water supply might not exactly follow the set exponential sure that exceeds what each fire system demand needs, the
relationship, such as in cities where municipal pumps are test is considered acceptable. From the example plot shown in
phased on to meet added demands. In these types of cases, Exhibit S1.9, all fire system demands fall under the gross test
no matter which graph paper is used for analysis, systematic data and thus appear acceptably supplied by the fire pump.
and detailed analysis is required. Secondly, N1.85 paper can Figure A.8.3.7.2.3(2)(b) shows another example of a
make reading flow points on the curve somewhat difficult, pump test result plotted and compared with system de-
but this would only typically account for a 1 psi (0.07 bar) to mands. That figure shows that the actual pump discharge
3 psi (0.2 bar) difference at most. can meet system demands. Some things to consider in this
analysis are as follows:
Analysis 1
1. Sprinkler system demands are not always calculated to
The purpose of the first analysis is to look at the overall fire the pump. If the pump is significantly detached from the
pump output to confirm that it will support the necessary sprinkler system, this could be an issue. Many are only
demand. calculated to the base of the riser. Thus, this information
Determine if a fire pump gross performance, relative to must be known.
connected fire protection system demands, is acceptable.   For example, if the System 2 design as listed on the
placard was only calculated to the base of the riser, there
Step 1: Plot gross fire pump performance, which is “Q” at the could easily be another 5 psi (0.34 bar) to 20 psi (1.38 bar)
associated “Pd”. (See Exhibit S1.8.) pressure loss between what was plotted with the gross
Step 2: Plot all provided fire system demands. (See pump discharge and what is available for the fire protec-
Exhibit S1.9.) tion system. In the plot example shown, there is only an
approximate buffer of 3 psi (0.2 bar). After analysis, this
Step 3: Make a determination for the final report if a fire
could mean the pump test is not acceptable and that
pump gross (unadjusted) performance, relative to the
fire system reinforcement or a pump replacement is
original design/installation, is acceptable.
warranted.
This analysis is completed by closely reviewing the data 2. Sprinkler system demands are not always calculated to
in the gross pump test with system demands, as shown in include hose allowances per NFPA 13, Standard for the

NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems 2020

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 663 01/11/19 6:44 PM

 664 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests

Installation of Sprinkler Systems. Also, data shown on riser morning or during the highest demand hours, the results
placards do not always include the sprinkler system de- could have indicated that System 2 only had an accept-
mand with hose allowances included. Thus, this informa- able water supply from the installed fire pump at certain
tion must be known. times of the day.
  For example, if the System 2 design listed on the riser Any significant drop in gross pump output, as shown in
placard was calculated without hose allowance require- Exhibit S1.10, should always be a concern and should be in-
ments, there could be another 100 gpm (379 L/min) to vestigated, since this can also be an indication of a hidden
750 gpm (2835 L/min) required to be added to the sys- closed valve somewhere on the suction supply leading to
tem demand. Also, in the example shown, there is only an the fire pump. This is why it is recommended that historical
approximate 45 gpm (170 L/min) buffer. After analysis, this water supply data, such as from past city fire hydrant tests or
could mean the pump test is not acceptable and that fire hydrant flow tests on the fire pump suction supply, be main-
system reinforcement or a pump replacement is warranted. tained and available during the pump test for comparison.
3. Fire pump water supply fluctuations (Ps) directly affect
fire pump discharge (Pd). As a municipal water supply Analysis 2
pressure drops, so does the fire pump gross perfor-
This analysis of the pump’s capabilities will examine the func-
mance. Municipal supply pressures can fluctuate for vari-
tionality of the pump when compared to the performance
ous reasons. Some cities now cycle or adjust municipal
of the pump during acceptance testing conducted in accor-
pump pressures down to force water conservation in
dance with NFPA 20.
areas that experience drought. Less pressure relates to
less flow. In turn, this creates less household and indus- Determine if a fire pump net performance, relative to the
try consumption. Pressure fluctuations can also occur original design/installation, is acceptable.
on shared water mains where heavy equipment or land-
scape irrigation drops the total system pressure. Step 1: Determine velocity head adjustment relative to net
  For this example, if the pump test were completed pressure (Pn).
in the late afternoon when the water supply from a city The pump test data analysis should include velocity head
was the highest, or if the same test was completed in the correction. Velocity head is created when the discharge pipe

EXHIBIT S1.8 Pump Gross

Test Data Plotted on N1.85 Graph 150
Paper. Test Test
Flow
Pitot
Reading
S Q Ps Pd
140 Info
(rpm) (gpm) (psi) (psi)

130
1-Churn No flow n/a 1770 0 45 140
20 2-1.75°
2-Rated play 32, 32 1763 1028 38 121
10 pipes

00 3-1.75°
3-150% Rated 32, 32, 32 1741 1542 29 84
play
Pressure (psi)

pipes
90

80
70

60

50
40

30
20

10
0
0 100 150 200 250 300 350 400 450 500 Scale A 550 575
200 300 400 500 600 700 800 900 1000 Scale B 1100 1150
400 600 800 1000 1200 1400 1600 1800 2000 Scale C 2200 2300
800 1200 1600 2000 2400 2800 3200 3600 4000 Scale D 4400 4600
Flow (gpm)

2020 NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 664 01/11/19 6:44 PM

 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests 665

150 EXHIBIT S1.9 Pump Gross

System 1-Office area: Test Data with System Demand
140
Demand for design (including hoses): 268 gpm at 78 psi Plotted on N1.85 Graph Paper.
130 System 2-Manufacturing area:
20 Demand for design (including hoses): 1254 gpm at 102 psi

10 System 3-Warehouse area:

Demand for design (including hoses): 1440 gpm at 63 psi
00
Pressure (psi)

90

80
70

60

50
40

30
20

10
0
0 100 150 200 250 300 350 400 450 500 Scale A 550 575
200 300 400 500 600 700 800 900 1000 Scale B 1100 1150
400 600 800 1000 1200 1400 1600 1800 2000 Scale C 2200 2300
800 1200 1600 2000 2400 2800 3200 3600 4000 Scale D 4400 4600
Flow (gpm)

150 EXHIBIT S1.10 Pump

FLOW TEST Gross Test Data Plotted on
140 SUMMARY SHEET N1.85 Graph Paper.
N1.85
130

20

10 Afternoon city water supply

and resulting gross pump test
00
discharge. [Green]
Pressure (psi)

90

80
70

60 Morning city water supply and

resulting gross pump test
50
discharge. [Blue]
40

30
20

10
0
0 100 150 200 250 300 350 400 450 500 Scale A 550 575
200 300 400 500 600 700 800 900 1000 Scale B 1100 1150
400 600 800 1000 1200 1400 1600 1800 2000 Scale C 2200 2300
800 1200 1600 2000 2400 2800 3200 3600 4000 Scale D 4400 4600
Flow (gpm)

NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems 2020

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 665 01/11/19 6:44 PM

 666 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests

size leading to a fire pump differs from the suction pipe size. Pd = pump discharge pressure — from reading discharge
Changes in pipe size change velocity. Since water is an in- flange gauge at flow
compressible fluid, when the pipe size increases or decreases Ps = pump suction pressure — from reading suction flange
without compression as an option, the velocity must change. gauge at flow
For fire pump analysis purposes, the following calcula-
Calculating Vertical Turbine Pump Adjustments. One other
tion can be applied:
adjustment must also be accounted for on vertical turbine
Pn :  (Pd – Ps ) + (Pv ) pumps. Since vertical turbine pumps have impellers that
are typically submerged in a static water source and must
where:
lift the water up to the discharge piping, there are no suc-
Pn = pump net discharge — total pump discharge perfor-
tion gauges, and therefore, there is no Ps. For vertical turbine
mance accounting for Pv
pumps, pressure loss due to the required lift, or Pe, must also
Pd = pump discharge pressure — from reading discharge
be included, and Pn is as follows:
flange gauge at flow
Ps = pump suction pressure — from reading suction flange Pn = Pd + Pe
gauge at flow
Pv = pump velocity head — correction calculated based on where:
suction and discharge pipe sizes Pe = E × 0.433
Pn = pump net discharge — total pump discharge perfor-
To mathematically calculate velocity head, the following mance accounting for Pv
­calculation can be applied: Pd = pump discharge pressure — from reading discharge
flange gauge at flow
Pv = 0.001123 Q2(1/Dd4 – 1/Ds4)
Pe = pump negative pressure (loss) due to elevation head
where: E = distance (ft) between pump discharge gauge and suction
Pv = velocity pressure (psi) water level
Q = measured flow (gpm) In the rare instances where a vertical turbine pump is
Dd = suction pipe diameter measured at gauge attachment ­connected to a pressurized water main, the adjustment meth-
flange (inches) odology discussed for horizontal split case pumps must be
Ds = discharge pipe diameter measured at gauge attachment used with the required elevation adjustments. See Table S1.2.
flange (inches) Step 2: Determine speed-adjusted net flow and pressure
However, while velocity head can be worth noting in (QN and PN ).
some cases, such as for very small pumps with small pipes, A mathematical adjustment must be made for the correc-
this adjustment is negligible and can be ignored for almost all tion of recorded test data to the original pump rated speed (it
standard fire protection pump installations. For this reason, may also include a velocity head adjustment); it must be applied
most contractor pump test sheets do not even have a column when determining flow test performance relative to the original
to record this calculation. pump performance. These corrections are shown in Table S1.2.
For example, for installations where both suction and dis- The three data points on a manufacturer’s pump curve
charge pipes or flanges (measured where gauges are mounted) and placard are generated as part of the manufacturing qual-
are the same, this can always be disregarded, since there is no ity assurance process, such that each pump is tested for flow
velocity head potential. Additionally, most fire pumps have be- and pressure at the speed listed on the manufacturer’s pump
tween 6 in. (150 mm) and 10 in. (255 mm) pipes where, even if curve or placard. At each pump manufacturing site, when a
sizes change, the Pv created is typically negligible — less than manufacturer’s curve is generated, the pump is connected to
2 psi (0.14 bar). For example, a pump with an 8 in. (200 mm) a test driver and the driver is manually adjusted to the same
suction pipe and 10 in. (255 mm) discharge flowing 2000 gpm speed at each test point before flow is recorded. So, on the
(7570 L/min) and a pump flowing 500 gpm (1890 L/min) with a 4 manufacturer’s curve the churn, rated, and 150 percent flow
in. (100 mm) suction and a 6 in. (150 mm) discharge would both points are all taken at the exact same speed, which is listed as
result in less than a 1 psi (0.07 bar) adjustment. For this reason, Pv the pump speed (rpm) at rated flow and pressure. In the same
should be considered to be 0 psi, as is the typical industry prac- way a car does not maintain the same speed along each hill
tice. And therefore, Pn is simply calculated as follows: and valley, a pump and driver that are installed in the field
do not really maintain the same speed across all flow points.
Pn = (Pd – Ps ) In a real installation, for most diesel drivers (and to a
smaller degree, for some electric drivers), the more work
where:
the pump does (i.e., more flow), the more work or load is re-
Pn = 
pump net discharge — total pump discharge quired of the driver. As with most drivers, an increased work
­performance accounting for Pv load ­results in decreased speed. This is similar to what is seen

2020 NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 666 01/11/19 6:44 PM

 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests 667

TABLE S1.2 Fire Pump Output Data with Velocity Head and Speed Corrections
Test Pitot S Q Ps Pd Pv Pn QN PN
Test Flow Info Reading (rpm) (gpm) (psi) (psi) (psi) (psi) (gpm) (gpm)

1-Churn No flow n/a 1777 0 45 140 0 95

2-Rated 2-1.75 in. play 32, 32 1763 1028 38 121 0 83

pipes

3-150% Rated 3-1.75 in. play 32, 32, 32 1731 1542 29 84 0 55

pipes

S = pump speed — typically recorded at pump shaft with tachometer

Q = pump discharge flow — combined pitot tube pressure readings converted to flow for given orifice
Ps = pump suction pressure — from reading suction flange gauge at flow
Pd = pump discharge pressure — from reading discharge flange gauge at flow
Pv = pump velocity head — correction calculated based on suction and discharge pipe sizes
Pn = pump net discharge — total pump discharge performance accounting for Pv
QN = pump flow — gross pump flow adjusted for speed
PN = pump pressure — gross pump pressure adjusted for speed

in automobiles. A vehicle towing a heavy load has a slower Step 3: Make a determination for the final report if a fire pump
speed than the same vehicle towing nothing unless more fuel net (adjusted) performance, relative to the original design/
is supplied via the gas pedal. installation, is acceptable.
For this reason, to make an equal comparison of a fire pump This calculation is completed by comparing the manu-
installed in the field to a test flow to that on a manufacturer’s facturer’s performance curve plot with the adjusted flow and
pump curve, the test data (both flow and pressure) must be pressure plot (three QN and PN points), with judgment rela-
mathematically adjusted to the pump’s rated speed (or rpm). tive to meeting the requirements of 8.3.7.2. If both curves run
Adjusting flow and pressure mathematically for speed within +/- 5 percent of each other, the test is acceptable. This
related to fire pump analysis is typically completed using the is shown in our final plot of Exhibit S1.11.
Affinity Laws as follows: Figure A.8.3.7.2.3(2)(a) shows another example of a net
pump test plotted and compared to an original pump manu-
1. For flow: QN = Q [rated speed (rpm)/measured speed (S)]
facturer’s test curve. Adjusted results of this test closely over-
2. For pressure: PN = Pn [rated speed (rpm)/measured lap (i.e., the pump is performing at or above 95 percent of the
speed (S)]2 original design specifications per the manufacturer’s perfor-
For our case, placard data indicates the pump speed rating is mance curve), which indicate the pump is performing close
1760 rpm. Therefore, the following applies: to the way it did when it was factory new. Also, this plot is a
good indication that the internal parts of the pump are func-
Churn: tioning well and the driver is operating as required for the at-
QN = 0 gpm tached pump.
A final step in any pump test is a written evaluation of
PN = 95 psi × (1760/1777)2 = 93 psi
the conclusions. Per 8.4.1, and 8.4.1.1, the property owner or
Rated: designated representative must be given a written copy of
QN = 1028 gpm × (1760/1763) = 1028 gpm the data analysis (i.e., interpretation of results) with the test
results as part of an annual pump test that includes any noted
PN = 83 × (1760/1763)2 5 83 psi
abnormalities. Any documented performance issues must be
150 percent: recorded and retained for future comparison purposes.
QN = 1542 gpm × (1760/1731) = 1569 gpm This analysis leads to one final question: What must be
done if a fire pump does not pass an annual test? If the fail-
PN = 55 × (1760/1731)2 = 57 psi
ure is in the associated Analysis 1, this must be addressed im-
This data is summarized in Table S1.3. mediately. Without further engineering review, a failed test

NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems 2020

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 667 01/11/19 6:44 PM

 668 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests

TABLE S1.3 Gross Pump Output Data Following Adjustments for Speed
Test Pitot S Q Ps Pd Pv Pn QN PN
Test Flow Info Reading (rpm) (gpm) (psi) (psi) (psi) (psi) (gpm) (gpm)

1-Churn No flow n/a 1777 0 45 140 0 95 0 93

2-Rated 2-1.75 in. play 32, 32 1763 1028 38 121 0 83 1028 83

pipes

3-150% Rated 3-1.75 in. play 32, 32, 32 1731 1542 29 84 0 55 1569 57
pipes

S = Pump speed — typically recorded at pump shaft with tachometer

Q = pump discharge flow — combined pitot tube pressure readings converted to flow for given orifice
Ps = pump suction pressure — from reading suction flange gauge at flow
Pd = pump discharge pressure — from reading discharge flange gauge at flow
Pv = pump velocity head — correction calculated based on suction and discharge pipe sizes
Pn = pump net discharge — total pump discharge performance accounting for Pv
QN = pump flow — gross pump flow adjusted for speed
PN = pump pressure — gross pump pressure adjusted for speed

EXHIBIT S1.11 Pump Plot:

Gross vs. Net Data Plotted on 150
N1.85 Graph Paper. Gross test results Test Test Pitot S Q Ps Pd Pv Pn QN PN
140 Flow Reading (rpm) (gpm) (psi) (psi) (psi) (psi) (gpm) (gpm)
Info
130
1-Churn No flow n/a 1777 0 45 140 0 95 0 93
20 2-1.75°
2-Rated play 32, 32 1763 1028 38 121 0 83 1028 83
10 pipes
3-150% 3-1.75° 32, 32, 32 1731 1542 29 84 0 55 1569 57
00 Rated
play
pipes
Pressure (psi)

90
Net test results:
80
• Net data (red)
70 • Net speed adjusted
data (red dashed)
60 • Mfg. curve (yellow)
50
40

30
20

10
0
0 100 150 200 250 300 350 400 450 500 Scale A 550 575
200 300 400 500 600 700 800 900 1000 Scale B 1100 1150
400 600 800 1000 1200 1400 1600 1800 2000 Scale C 2200 2300
800 1200 1600 2000 2400 2800 3200 3600 4000 Scale D 4400 4600
Flow (gpm)

2020 NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 668 01/11/19 6:44 PM

 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests 669

would indicate that in the event of a fire, there might not be required. Poor performance indicated here is usually the re-
an adequate water supply to acceptably supply the fire sys- sult of a damaged impeller or blocked impeller vane.
tems connected to the fire pump. This translates to the po- The following are some examples of poor net pump test
tential of an uncontrolled fire and, potentially, extended loss curves and possible problems.
of life and property. Exhibit S1.13 shows a failed test because there is more
The solutions are varied and a complete engineering than a 5 percent performance degradation. The test shown in
analysis must be completed, but a logical first step is to verify Exhibit S1.13 is not a failed test per NFPA 25; however, an in-
that all gauges and flow measuring devices are properly cali- vestigation is still warranted. This test, taken as plotted, would
brated and working. This might also involve a pump retest to mean the pump and driver actually improved performance
verify the results. by almost 10 percent from when it was factory new. This is not
If results still show a failed pump test, a possible solution likely. This would most likely be the result of test equipment
would be replacement of the fire pump with a larger unit. not calibrated correctly or poor test data recording. In either
Since the driver size is based on pump size, this will likely case, the test has to be questioned as “too good to be true.”
mean a newer, more powerful driver as well. Additionally, There are numerous reasons for a test, as shown in
since controller components, power supply transformers, and ­Exhibit S1.12. Exhibit S1.10 shows a gross test plot with the
wiring for electric drivers are sized based on motor size, this water supply. Any significant drop in gross pump output per
could also mean that a new transformer, a new controller, and Exhibit S1.10 should always be a concern and should be in-
rewiring are required. vestigated, since this can also be an indication of a hidden
The flow also must be evaluated to see if the issue is closed valve somewhere on the suction supply leading to the
with a water supply source. As previously discussed, a water fire pump. This is why it is recommended that historical water
supplier could be purposely decreasing a municipal water supply data such as from past city fire hydrant tests or hydrant
supply for reasons related to fire protection, such as forced flow tests on a fire pump suction supply be maintained and
water conservation. In these cases, a request can be made to available during the pump test for comparison. Of course,
increase the pump suction water supply. Reduced flow could since Exhibit S1.12 only looks at net performance, this would
also be the result of a valve partially closed downstream of not be detected if only one of the two analyses were com-
the pump. Thus, these are the reasons a full engineering anal- pleted. This could be the result of impeller wear, uncalibrated
ysis must be completed. gauges reading low, or using the wrong flow device pitot
In some cases, the solution is connecting the pump to reading chart. It could also be from a pressure relief valve
another water supply source or parallel source to counter fric- stuck open in a certain position.
tional loss, such as from a recent backflow apparatus instal- Exhibit S1.14 is a failure for the same reason as
lation. In extreme cases, this could also mean adding a fire ­Exhibit S1.12. However, this might not be as big of a concern,
pump suction tank as a new source of site water. And in this because performance is within the requirements at the rated
case, an entirely new fire pump and driver would likely also and 150 percent point. This result could simply be a pressure
be required. relief valve closing as flow increases, as should be the case.
In completing the evaluation for Analysis 2, a failed test For testing, it should be ensured that this valve closes as soon
per 8.3.7.2.3(3) that required further investigation is pump as water flow through the header is started, so as not to affect
flow performance degradation in excess of 5 percent. While the performance curve data.
the causes for failure here could be different from those for Exhibit S1.15 is a failure for the same reason as
Analysis 1, they could also be the same or related. Thus, a full Exhibit S1.12 and Exhibit S1.14. Also, this is much more
engineering analysis must be completed for this, along with critical, as the flow region with the poor performance is
following the same logic and process of issue investigations. typically where most sprinkler system demands would fall
A logical first step would also be to verify that all gauges and on the graph. Reviewing the pump affinity laws, pressure
flow measuring devices were properly calibrated and work- is exponentially affected relative to flow. This can be
ing. A pump retest to verify the results would also be rec- attributed to a partially closed valve on the suction side
ommended. If results are confirmed, maintenance involving of a fire pump or, more likely, an impeller that is damaged
taking apart the pump for visual inspection is then typically or has foreign material lodged in it.

NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems 2020

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 669 01/11/19 6:44 PM

 670 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests

EXHIBIT S1.12 Example of

Failed Test. 150

140 Solid line = mfg. pump curve This shows a failed test:
130 Dashed line = net pump test curve the plotted test is more
than 5% below the
20 manufacturer’s curve.
The test curve is ~10%
10 below the manufacturer’s
curve.
00
Pressure (psi)

90

80
70

60

50
40

30
20

10
0
0 100 150 200 250 300 350 400 450 500 Scale A 550 575
200 300 400 500 600 700 800 900 1000 Scale B 1100 1150
400 600 800 1000 1200 1400 1600 1800 2000 Scale C 2200 2300
800 1200 1600 2000 2400 2800 3200 3600 4000 Scale D 4400 4600
Flow (gpm)

EXHIBIT S1.13 Example of Test

Results That Would Trigger an 150
Investigation.
140 Solid line = mfg. pump curve This shows a failed test:
130 Dashed line = net pump test curve the plotted test is more
than 5% below the
20 manufacturer’s curve.
The test curve is ~10%
10 above the manufacturer’s
curve.
00
Pressure (psi)

90

80
70

60

50
40

30
20

10
0
0 100 150 200 250 300 350 400 450 500 Scale A 550 575
200 300 400 500 600 700 800 900 1000 Scale B 1100 1150
400 600 800 1000 1200 1400 1600 1800 2000 Scale C 2200 2300
800 1200 1600 2000 2400 2800 3200 3600 4000 Scale D 4400 4600
Flow (gpm)

2020 NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 670 01/11/19 6:44 PM

 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests 671

150 EXHIBIT S1.14 Example of

Failed Test Data.
140 Solid line = mfg. pump curve
Dashed line = net pump test curve This shows a failed test:
130 the plotted test is more
than 5% below the
20
manufacturer’s curve
10 at churn.

00
Pressure (psi)

90

80
70

60

50
40

30
20

10
0
0 100 150 200 250 300 350 400 450 500 Scale A 550 575
200 300 400 500 600 700 800 900 1000 Scale B 1100 1150
400 600 800 1000 1200 1400 1600 1800 2000 Scale C 2200 2300
800 1200 1600 2000 2400 2800 3200 3600 4000 Scale D 4400 4600
Flow (gpm)

150 EXHIBIT S1.15 Example of

Failed Test Data.
140 Solid line = mfg. pump curve This shows a failed test:
130 Dashed line = net pump test curve the plotted test is more
than 5% below the
20 manufacturer’s curve
between the rated and
10 150% flow points.
00
Pressure (psi)

90

80
70

60

50
40

30
20

10
0
0 100 150 200 250 300 350 400 450 500 Scale A 550 575
200 300 400 500 600 700 800 900 1000 Scale B 1100 1150
400 600 800 1000 1200 1400 1600 1800 2000 Scale C 2200 2300
800 1200 1600 2000 2400 2800 3200 3600 4000 Scale D 4400 4600
Flow (gpm)

NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems 2020

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 671 01/11/19 6:44 PM

 672 Part 2 / Supplement 1: Conducting Fire Pump Flow Tests

Summary impairment; it can also be a sign of potential issues with the

pump, suction piping, or water supply.
Confirming a fire pump’s ability to provide the required flow
and pressure based on system demand is imperative to main- References Cited
taining the level of life safety anticipated for a given prop-
erty. Properly conducting fire pump flow tests to quantify a National Fire Protection Association, 1 Batterymarch Park,
pump’s functionality is a critical step to making this confirma- Quincy, MA 02169-7471.
tion. In addition to confirming that the pump will supply the NFPA 13, Standard for the Installation of Sprinkler Systems,
necessary pressure and flow, it is important to assess the per- 2019 edition.
formance of the pump against its original performance data. NFPA 20, Standard for the Installation of Stationary Pumps for
Degradation in pump output over time can lead to a system Fire Protection, 2019 edition.

2020 NFPA 25 Handbook: ITM of Water-Based Fire Protection Systems

BK-NFPA-25HB20-190199-Chp24_Supp1.indd 672 01/11/19 6:44 PM