THE EFFECT OF FOAM ADDITIVES ON THE FIRE SUPPRESSION

EFFICIENCY OF WATER MIST

Andrew K. Kim, Bogdan Z. Dlugogorski, and Jack R. Mawhinney

National Fire Laboratory

Institute for Research in Construction
National Research Council Canada
Ottawa, Canada KIA OR6

ABSTRACT

Water mist has been shown to be an effective fire suppressant. In this paper, the
effectiveness of using a small quantity of additive in water mist is investigated. The additives
used in the study were film-forming and foam-forming agents. Tests were conducted using crib
fires, and heptane and diesel pool fires.

Tests were conducted in a 3.5 m by 3.1 m test enclosure, with walls constructed of
perforated steel, which provided unrestricted ventilation. The benefit of adding a small quantity
of foam agent to the water mist was observed in the suppression of pool fires. Water droplets,
containing solutions of foaming agents, tended to expand slightly in the spray, eventually
forming a thin layer of foam on the surface of burning heptane and diesel pool fires. The
establishment of such a layer provided a means for efficient extinction of pool fires.

INTRODUCTION

Water mist systems have been demonstrated to be effective fire suppression systems.
Flame cooling and oxygen displacement by steam are considered to be the dominant
mechanisms of extinguishment by water spray and it is known that fine watcr mists are more
efficient than coarser sprinkler sprays at absorbing heat. The development of a fine water mist
suppression system, however, was delayed because of its relatively high cost. The recent
development of high performance nozzles and the withdrawal of halon for environmental
reasons has brought about renewed interest in water mist fire suppression systems as potential
alternatives to halon.

Recently, the National Fire Laboratory (NFL) has been working on a series of research
projects which attempts to combine an understanding of the dynamics of extinguishment using
water mist, compartment factors, and mist generation with the practical aspects of applying the
technology. This includes a project to develop a fire suppression system using water mist to
replace existing halon systems in machinery spaces on ships and ongoing research to develop
early detectiodsuppresion of fires in telecommunication and other electronic facilities using
water mist suppression systems. In addition, as part of a project with National Defence Canada
to investigate the effectiveness of Compressed-Air-Foam (CAF) systems in a fixed piping
arrangement to suppress Class A and Class B fires, a series of tests was conducted to determine
the effectiveness of fine water mist for open space fires using single fluid swirl-type nozzles. In

347
the test series, the effectiveness of using small amounts of foam-forming additives in the fine
water mist was also investigated. Tests were conducted using wood crib fires, and heptane and
diesel pool fires. This paper summarizes the results of the experiments that compared the
relative importance of different suppression mechanisms of the fine water mist system as well as
the effectiveness of using additives in the water mist.

TEST SET-UP AND PROCEDURES

All tests were conducted using the NFL's calorimeter facility. The calorimeter facility
includes a 2.4 m by 3 m canopy hood which is connected to a 13 m long, 0.56 rn diameter
exhaust duct. The exhaust duct contains a Pitot-static probe, thermocouples and gas sampling
ports to measure the gas flow rate, temperatures, CO, CO2 and 0 2 concentrations, as well as a
smoke meter to measure the smoke production rate. The heat release rate of the fire during the
test was determined using the oxygen consumption method. In addition, the concentration of
unburned hydrocarbons and the amount of water vapour present in the exhaust gases were also
measured.

Fire tests were conducted in a 3.5 m by 3.1 m and 3.3 m high test enclosure, with walls
constructed of perforated steel to break up the convective air currents without limiting the
ventilation rate. The enclosure was instrumented with thermocouples and heat flux meters and
was placed under the collection hood of the calorimeter facility to measure the changes in fire
behaviour during the test. Figure 1 shows the details of the instrumentation and test set-up. A
thermocouple tree containing 6 thermocouples at 0.3 m vertical intervals was placed above the
centre of the fuel. Three heat flux meters were placed around the enclosure. One was located
directly above the fuel, facing downward, at 2.4 m above the floor. Two were located 1.7 m
away from the centreline of the fuel, one at 1.7 m and another at 2.4 m above the floor, both
viewing the fire. Fuels were placed either on the floor or on a 0.7 m high platform at the centre
of the test enclosure. The enclosure, with its easy access to modify the experimental set-up and
complete visibility of fire behaviour during the test, allowed a systematic study of the various
parameters of these suppression systems.

A 0.9 rn diameter pan with a lip height of 100 mm was used for the heptane and diesel
pool fire tests. Tests were also conducted using 0.6 m by 0.6 m and 0.3 m high wood cribs made
of 40 mm by 40 mm pine sticks. For the crib and the diesel pool fire tests, the fires burned for
approximately 2 min before activation of the suppression system, to allow the fire to reach a
fully developed stage. For the heptane pool fires, a 1-min pre-burn was allowed since the
heptane pool fire reached steady burning conditions in a shorter time than the other fuels.

NOZZLES

Two nozzle types were used in the tests, the 7G-5 (Spraying Systems Company Model
314 7G5)' , and a standard pendent sprinkler. The 7'3-5 nozzle is a swirl type pressure nozzle
which relies on hydraulic pressure to force water through small diameter orifices at a high
* Certain commercial products are identified i n this paper to adequately specify the experimental procedure. In no
case does such identification imply recommendations or endorsement by the National Research Council, nor docs it
imply that the product or material identified is the hest available for the purpose.

348
velocity. The spray angle of the 7G-5 nozzle is 150'. The performance of the nozzle was
evaluated for the open (unenclosed) fire case, with and without a foam-forming agent. Tests
were conducted with one or two 7G-5 nozzles. The single 7G-5 nozzle was mounted right above
the fuel at the ceiling of the enclosure (3 m above the floor). When two 7G-5 nozzles were used,
the nozzles were mounted 2 m apart at the ceiling of the enclosure, equal distances from the fuel.

The drop-size distribution of the water sprays from the 7G-5 nozzle was measured using
a Greenfield Instruments Model 700A Spray Drop Size Analyzer' . Details of the drop size
measurements are given in Reference [I]. The volumetric mean diameter (D,o.s) at the centre of
the spray, measured 0.7 m from the nozzle, was 100 microns at a pressure of 550 kPa (5.43 bar).
The largest diameter for 90% of the spray volume (DVo.y)at the centre of spray for the 7G-5
nozzle was 300 microns. The standard sprinkler had a DvO.sof 440 microns and D,o.yof 1000
microns measured at the centre of the spray at a distance of 1 m from the sprinkler at a pressure
of 180 kPa(1.77 bar).

The spray flux density from each nozzle configuration was obtained by measuring the
rate at which the spray fell on a collecting surface. A total of 169 collecting cups of 0.1 m
diameter were laid on the floor at a grid spacing of 0.18 m, covering the whole area of the nozzle
spray. Spray flux densities on the floor from a single and twin 7G-5 nozzles, located 3 m above
the floor, were measured. The spray flux density from the 7G-5 nozzle varied from spot to spot
on the floor. As shown in Figure 2, the spray flux density in the coverage area for the twin
nozzle case, ranged from 3 to 18 L/min/d. The figure also shows that the spray flux density,
under non-fire conditions, at the spot where the fuel was normally located in the fire tests, is
approximately 7 L/min/m2. For a single 7G-5 case, that value was approximately 18 L/min/m2.

ADDITIVES

Small amounts of a foam-forming Class A concentrate and a film-forming Class B

concentrate were used in the tests to investigate the effectiveness of using these in the water mist
system. The Class A foam concentrate used in the test was a Silvex* solution. This type of
foam is primarily used on fires involving Class A or wood fires. It is made from hydrocarbon-
based surfactants and lacks filming properties, however, it possesses excellent wetting
properties. In the test series, a small amount of Class A concentrate, equivalent to 0.3% of the
water flow rate, was injected into the suppression system.

The Class B foam concentrate used in the test was an Aqueous Film Forming Foam
(AFFF) concentrate. This type of foam is normally used on flammable liquid fires. It is made
of fluorocarbon-based surfactants and has strong film-forming characteristics. The amounts of
AFFF concentrate used in the tests were 1% and 3% of the water flow rate.

*Certain commercial products are identified in this papcr lo adequately specify the experimental procedure. In no
case does such identification imply recommendations or endorsement by the National Rcsearch Council, nor does it
imply that the product or material identified is the hest available for the purpose.

349
RESULTS AND DISCUSSIONS

Twenty tests were conducted. Three dry tests, one with a crib fire and heptane and diesel
pool fires, were conducted without any suppression system for comparison purposes. In the
remaining tests, the heptane, diesel and crib fires were suppressed with water mist, with and
without additives, and with sprinklers. Table 1 shows a summary of the test results.

HEPTANE FIRES

Single nozzle: The test results with a single 7G-5 nozzle, located directly above the fuel,
indicated that the effectiveness of the water mist in extinguishing a fire depends on the type of
fuel and the momentum of the mist. With a 0.9 m diameter heptane pool fire on the floor, the
water mist from a single 7G-5 nozzle, located 3 m abeve the fire, reduced the flame size
substantially, however, it failed to extinguish the fire until 4 min into the test. When the heptane
fuel was raised above the floor by 0.7 m, the water mist from the 7G-5 nozzle, located 2.3 m
above the fire, produced an initial flare-up and made the flame very turbulent. The mist spray
then quickly reduced the flame size, however, and pushed the flame to one side of the pan,
eventually extinguishing it at 30 s. The difference in the performance in these tests could be
from the difference in mist momentum when it hits the flame. Water mist loses its momentum
as it travels through hot fire gases which are flowing in counter direction with a strong buoyant
force. When a nozzle is closer to the fuel surface, the mist, with its higher momentum, appears
to be able to better penetrate the fire plume and push the water vapour created in the outer
regions of the flame onto the fuel surface.

Figure 3 compares the effect of the water mist, with and without foam additives, on
heptane pool fires. The Figure shows the heat release rates (HRR) of heptane fires, without
suppression, with water mist suppression and with water mist and additive suppression. The
heptane pool fire, which could not be extinguished by water mist from a single 7G-5 located 3 m
above the fire, was extinguished when a 0.3% Class A foam solution was added to the water
mist. The water mist spray with the additive was able to reduce the flames. With continued
application onto the fuel surface, the fire was extinguished in 1 min 10 s. When the heptane was
re-ignited, and water mist with additive was applied again, a similar phenomenon occurred; that
is, initial knockdown of the fire and extinguishment at 57 s. The foam additive in the water mkt
produces a thin foam solution layer on the fuel surface. This thin foam layer reduces the thermal
feedback from the flame to the fuel surface and this reduction in the heat flux to the fuel surface
reduces the vaporization rate of the liquid fuel and contributes to the extinguishment. Figure 3
clearly shows that the water mist reduces the HRR but does not extinguish the fire. The HRR
plot for water mist and additive shows a quick knockdown and extinguishment within 1 min.

Adding 1% of Class B foam to the water mist extinguished the heptane pool fire but
required more time than with 0.3% Class A foam. Adding 3% Class B foam to the mist seemed
to work better than the case with 0.3% Class A foam, by producing a thicker and more stable
foam solution layer on the fuel surface. However, the extra cost involved, due to the need for
larger quantities of foam concentrate, needs to be considered.

350

-___
 Two nozzles: When tests were conducted with a heptane pool fire placed in the middle
of two 7G-5 nozzles, which were located 2 m apart and 3 m above the fuel, the water mist
controlled but did not extinguish the flame. As discussed previously, the two nozzle system
produced a lower spray flux density at the centre of the space where the fire was located,
compared to a single nozzle located directly above the fire. Also, the spray momentum at the
centre was lower compared to the single nozzle case. During the fire test, it was observed that
the water mist was not penetrating the plume very well and little mist was reaching the fuel
surface.

In the test with 0.3% Class A foam concentrate added to the water mist, it was observed
that there was not much difference in fire behaviour during suppression compared to the test
with water mist only. The fire was controlled but not extinguished. A probable explanation is
that the water mist did not have sufficient momentum to penetrate the fire plume. The water
mist cooled the fire by evaporation, thus achieving control. However, since the mist with the
Class A foam additives could not reach the fuel surface to form a barrier between the fuel and
the flame zone, fuel vaporization was not reduced and the fire continued to burn.

When two standard pendent sprinklers were used for suppression of the heptane pool fire,
fire control was not achieved and the fire size remained almost the same for the duration of the
test. The primary advantage of water mist over coarser sprays is the enhancement of the rate at
which the spray extracts heat from the hot gases and flame. These current test results clearly
show the efficiency of cooling using water mist compared to sprinklers.

DIESEL FIRES

Sinele nozzle: For a diesel pool fire, a single 7G-5 nozzle, located 3 m above the fire,
produced a violent outburst of flame with the initial application of water mist, followed quickly
by knockdown of the flame and extinguishment in less than 10 s. It was observed that the initial
flare-up was much more prominent in the diesel fire tests than in the heptane fire tests. This
momentary intensification of a fire with the application of water spray has been reported by
other researchers [2,3].

Two nozzles: The diesel pool fire was extinguished with two 7G-5 nozzles located 3 m
above the fuel surface, whereas the same nozzles could not extinguish a heptane fire. Cooling of
flames by water mist was sufficient to extinguish the diesel fire because of the high flashpoint of
the diesel fuel (60'C). Water mist spray produced an initial flare-up in the diesel fire, but
quickly reduced the fire size and extinguished the fire in 1 min 10 s. When 0.3% Class A foam
concentrate was added to the water mist, extinguishment of the diesel fire was achieved in less
than 30 s. When sprinklers were used on a diesel fire, the fire size was reduced very slowly and
after 3 min 45 s, the fire was extinguished.

For liquid fuel fires, the evaporation of the mist cools the flame which, in turn, reduces
the radiant heat flux to the surface of the fuel, resulting in a reduction in the evolution of
flammable vapours [4]. The combination of reduced flame temperature and reduced evolution
of vapours results in a reduced burning rate and, in some cases, complete extinction. Fires in
liquid fuels with flashpoints above normal ambient temperature, such as diesel, can be

351
extinguished relatively easily by flame cooling. The present test series clearly showed this
effect.

WOOD CRIB FIRES

Figure 4 shows the heat release rate of the crib fire without suppression. The Figure
shows that the crib fire requires approximately 2 min from the time of ignition for the fire to
involve the entire crib. After the initial 2 min of this development stage, the crib burned steadily
for approximately 4 min with a heat release rate of 400 kW. Since the suppression system was
activated at 2 min from the time of ignition, the suppression system was considered to be
effective if it extinguished the fire within 4 min from the suppression activation; that is, before
the crib fire started to decay.

&$e. nozzle; When a single 7G-5 nozzle was located directly above the crib, the water
mist spray quickly knocked down the fire and extinguished it in less than 30 s. Adding 0.3%
Class A foam concentrate to the water mist did not change the performance of the water mist in
the single 7G-5 nozzle crib fire test.

Two nozzles: In the crib fire tests, the pine wood crib was placed on a platform 2.3 m
below two 7G-5 nozzles, spaced 2 m apart. The approximate spray flux density at the crib
location was 7 Umidm2, measured under non-fire conditions. Water mist reduced the fire. size
substantially, with the flame on the surface of the crib extinguished at 4 min 25 s. There were
still, however, some flames inside the crib, and complete extinction of the crib fire was not
achieved during the 5 min test.

When 0.3% Class A foam concentrate was added to the water mist, there was very little
difference in the performance of the water mist suppression system. This is shown in Figure 4,
where the heat release rates of the crib fires are shown for tests, with and without the additive, as
well as the heat release rate of the crib fire during un-suppressed test, are shown. These plots
show almost identical heat release rate values for the water mist tests, with and without an
additive.

The difference in the results between the single and two 7G-5 nozzle tests is probably
due to the fact that the single 7G-5 nozzle provides higher spray flux density (18 L/min/m2 vs
7 L/min/m2) and momentum in the central region (directly below the nozzle) than the two 7G-5
nozzle configuration.

For comparison, a wood crib fire test was conducted using two sprinklers instead of mist
nozzles. Water spray from the two sprinklers substantially reduced the fire size and pushed the
fire to the mid-portion of the crib, with a flame height of 0.3 m to 0.6 m above the crib. The
flame on the surface of the crib was extinguished at 3 min, but there was still some flames inside
the crib, which were finally extinguished at 4 min 45 s. Even though the sprinklers extinguished
the crib fire, the flow rate of the two sprinklers was more than twice the water flow rate of the
two 7G-5 nozzles (190 L h i n compared to 70 Umin).
 NRCC's study [ 5 ] on Compressed-Air-Foam systems showed that, for wood crib fires, a
thick foam blanket on the crib surface is needed to stop the burning. The present NRCC
experiments, with 0.3% Class A foam concentrate in the water mist, indicate that the surfactant
quality of Class A foam, at these low concentrations, did not have much effect on suppressing
the crib fires.

CONCLUSION

The current NRC test series indicates that, in an open fire case with unrestricted
ventilation and with the water mists used, cooling of the flame seems to be the dominant
mechanism rather than oxygen displacement. The latter mechanism is more effective in an
enclosed fire. Adding a small quantity of Class A or B foam concentrates to the water mist,
significantly improved the performance of the water mist system in suppressing liquid fuel pool
fires. A thin layer of foam solution on the pool surface reduced the amount of radiant heat
energy that was absorbed by the fuel.

In crib fire tests, the addition of a small amount of foam additive to the water mist did
not significantly change the performance of the suppression system. For liquid fuel fires, the
foam concentrates in the water mist produce a thin foam layer on the liquid fuel surface, which
reduces the vaporization rate of the fuel. For wood crib fires, a thin foam layer on the crib
surface is not sufficient to improve suppression. Other studies have shown that, a thick foam
blanket on the crib surface is needed to stop the burning [ 5 ] .

REFERENCES

I. Mawhinney, J.R., "Characteristics of Water Mist for Fire Suppression in Enclosures" in

Proceedings of the Halon Alternatives Technical Working Conference, Albuquerque, New
Mexico, USA, 1993.
2. Kokkala, M.A., "Extinction of Liquid Pool Fires with Sprinklers and Water Sprays", Valtion
Teknillinen Tutkimuskeskus, Statens Teniska Forskningscentral (Technical Research Centre
of Finland), Espoo, 1989.
3. Jones, A. and Thomas, (3.0.."The Action of Water Sprays on Fires and Explosions",
Transactions of the Institution of Chemical Engineers, 71: Part B, 41-49, 1993.
4. Mawhinney, J.R., Dlugogorski, B.Z. and Kim, A.K., "A Closer Look at the Fire
Extinguishing Properties of Water Mist" 4th IAFSS Conference, Ottawa, Canada, June,
1994.
5. Dlugogorski, B.Z., Kim, A.K. and Crampton, G.P., "Performance of Class A Compressed-
Air-Foam from a Fixed System'' to be presented at International Wildland Fire Foam
Symposium and Workshop, Thunder Bay, Ontario, Canada, May 3-5, 1994.

ACKNOWLEDGEMENTS

The authors appreciate the assistance of George Crampton, Robert Onno and Michael Ryan
in the construction of the test apparatus and carrying out experiments.

353
 Fuel NozzleNozzle Nozzle Additive Fire Extinguishment
type number height (m) reduced ? time (min: s )
Heptane None 0 N/A None No No
Heptane 7G-5 1 3 None Yes No
Heptane 7G-5 1 2.3 None Yes 0 : 30
Heptane 7G-5 1 3 0.3% A Yes 1 : 10
Heptane 7G-5 1 3 1% B Yes 2 : 00
Heptane 7G-5 1 3 3% B Yes 0 : 40
Heptane 7G-5 2 3 None Yes No
Heptane 7G-5 2 3 0.3% A Yes No
Heptane Sprinkler 2 3 None No No
Diesel None 0 N/A None No No
Diesel 7G-5 1 3 None Yes 0 : 10
Diesel 7G-5 2 3 None Yes 1 : 10
Diesel 7G-5 2 3 0.3% A Yes 0 : 30
Diesel Sprinkler 2 3 None Yes 3 : 45
Crib None 0 N/A None No No
Crib 7G-5 I 2.3 None Yes 0 : 30
Crib 7G-5 1 2.3 0.3% A Yes 0 : 30
Crib 7G-5 2 2.3 None Yes No
Crib 7G-5 2 2.3 0.3% A Yes No
Crib Sprinkler 2 2.3 None Yes No
 7G-5 nozzle
.
heat flux
meter
0 [
Thermocouple

-
E heat flu)
Heat flux meter
p!
(1
't; meters
I
0

I
. o

fuel 0 Fuel

3.5 rn _____)

Side View

3.5 m

t
E
r
6
0.5 m heat meter

I A

Plan View

Figure 1. Schematic diagram of the test set-up

355
 1
0
0
v
0
In
0 0
5) -
0
0
0
In
l-

...
 0 0
0 0 0 0
(D N 0 0
7 7 a d 0

-
.-
c
E
v

e..
i=
N
N
0

- E
8
M
C
.e
m
m
5
.-
3
m
p!
L;
8
C
Y
Cd
a
0
................................................................................................................ ., c
*8
e:
e:
3:
c.i
8
&
I
,M
J l t 3 4 g I m ~ ~ ~ 1 ~ ~ ~ , 1
iz , ,
 0 0 0 0
0 0 0 0
OD (0 d N 0
OD
a0

b b

(D
(D

In
In
A

.-
c
E
v

!i=i
Q d

0 0

N
N

- 0
0
0
0
0
0
0
0
0
.-

m (D d N