Lithium-ion Battery Fire Suppression Using Water

Mist Systems

This is the Published version of the following publication

Ghiji, Matt, Burch, Ian, Suendermann, Brigitta, Gamble, Grant, Novozhilov,

Vassili, Joseph, Paul and Moinuddin, Khalid (2021) Lithium-ion Battery Fire
Suppression Using Water Mist Systems. Frontiers in Heat and Mass Transfer,
17. ISSN 2151-8629

The publisher’s official version can be found at

http://thermalfluidscentral.org/journals/index.php/Heat_Mass_Transfer/article/view/1214
Note that access to this version may require subscription.

Downloaded from VU Research Repository https://vuir.vu.edu.au/43829/

Frontiers in Heat and Mass Transfer (FHMT), 17, 13 (2021) Global Digital Central
DOI: 10.5098/hmt.17.13 ISSN: 2151-8629

Frontiers in Heat and Mass Transfer

Available at www.ThermalFluidsCentral.org

LITHIUM-ION BATTERY FIRE SUPPRESSION USING WATER MIST

SYSTEMS
Matt Ghijia, *, Ian Burchb, Brigitta Suendermannb, Grant Gambleb, Vasily Novozhilova, Paul Josepha, Khalid Moinuddina
a Institute of Sustainable Industries and Liveable Cities (ISILC), Werribee Campus, Victoria University, PO Box 14428, Melbourne, VIC, 8001,
Australia
b Maritime Division, Defence Science & Technology Group, 506 Lorimer St, Fishermans Bend, Melbourne, VIC, 3207, Australia

ABSTRACT
Lithium-ion batteries (LiBs) have superior energy density and lifetime compared to battery technologies such as lead acid. Despite the widespread
application of LiBs in energy storage systems, electronic devices, aerospace and the automotive industry, they present a fire risk. In this study,
experiments were conducted to characterize the thermal behavior of the electrolyte (as the main contributor to LiB fires) using a cone calorimeter;
investigate the interactions of water mist and a Bunsen burner, as a precursor to examining the effectiveness of a water mist suppression system in
extinguishing a LiB fire. In the present work, we have endeavored to systematically study the fire suppression efficacy of water mist by adopting to
some novel approaches. This involved carefully planned laboratory scale explorations that involved a propane gas fueled flame, and subsequently by
using bespoke set up that mimicked fire owing to fuel surge from typical Li-ion cell (18650 cells). In the latter set of fire suppression tests, water
droplets were produced by a fan nozzle and sprayed horizontally toward the jet flame of replica 18560 battery containing only the electrolyte. The
results showed that both fire types (Bunsen burner and LiB) are suppressed rapidly on activation of the water mist fire suppression system for geometries
that enable the water mist direct access to the lift-off zone, between the gas source and base of the flame.

Keywords: Lithium-ion Battery; Thermal Runaway; Fire; Suppression; Water Mist.

lead-acid, nickel-cadmium, zinc bromide and alkaline. However, the

1. INTRODUCTION composition of the electrolyte may be varied to be compatible with the
cathode and anode materials, with additives to improve cycling and
The increased use of renewable energy technologies has put battery lithium ion conductivity. Fire retardants may also be added to terminate
energy storage solutions in the spotlight. Lithium-ion batteries (LiBs) the radical chain reactions of combustion, providing a safer operating
provide outstanding energy density, voltage and lifetime compared to environment (Julien et al. 2016; Ouyang et al. 2019). The separator is a
other battery technologies (Blum and Long Jr 2016). In addition, LiBs porous membrane between the cathode and anode allowing lithium ion
are lightweight and have a low self-discharge rate making them the movement but preventing electrical short-circuiting between the anode
preferred battery technology for electronic handhelds, electric vehicles, and cathode (Julien et al. 2016; Arora and Zhang 2004). The separator
and energy storage systems in airplanes and submarines (Pacala and materials are microporous films or laminates made of polyolefins such
Socolow 2004; Depetro 2016). Significant research has been conducted as polyethylene (PE) and polypropylene (PP).
in the field of materials science to improve the energy density, safety, Despite the widespread application of LiBs, there are major safety
charge/discharge rate and longevity of LiBs. concerns especially in battery-based power storage station, personal
A lithium-ion cell comprises an anode and cathode deposited onto electronic devices, electric vehicles (Ruiz et al. 2018) and airplanes
copper and aluminum current collectors respectively, electrolyte and a (Rodriguez 2013). These battery systems consist of a large number of
separator. Lithium ions are transferred from the anode to the cathode cells where a failure such as thermal runaway in a single cell can
during discharge and from the cathode to the anode during the charging influence neighboring cells and consequently the safety of the entire
process (Tarascon and Armand 2001; Galatro et al. 2020). The most battery system (Lamb et al. 2015). Thermal runaway is a process where
common material used for the anode is graphite due to its high negative an increase in temperature causes internal reactions that are exothermic
potential, and various lithium metal oxides are used for the cathode with and result in further increases in temperature, generally leading to battery
lithium iron phosphate commonly recognized as the best compromise for venting and fire. Pressure relief vents are built into the lid of the battery
energy output and safety. A Solid-Electrolyte-Interface (SEI), a layer casings to control overpressure and avoid explosion by directing venting
permeable to lithium ions but not to the electrolyte, is formed on the gases, and possible jet fire, away from the battery module. The likelihood
anode due to reactions between intercalated lithium ions and the and consequence of LiB fires can be seen over the past two decades
electrolyte during the initial charge. Intercalation is the reversible where more than 300 fires or fire-related incidents with 40 fatalities have
insertion of a molecule or ion into materials with layered structures. The been reported (Depetro 2016).
stability and characteristics of the SEI layer determine LiB thermal The abuse conditions which can initiate thermal runaway are
behavior, safety and lifetime. LiB electrolyte is usually a flammable categorized by: electrical abuse (over-charging/over-discharging) (Ye et
carbonate-based organic solvent, unlike other battery chemistries such as al. 2016); thermal abuse (over-temperature) (Guo et al. 2017);

*
Corresponding author. Email: [email protected]
1
Frontiers in Heat and Mass Transfer (FHMT), 17, 13 (2021) Global Digital Central
DOI: 10.5098/hmt.17.13 ISSN: 2151-8629

mechanical abuse (Ruiz et al. 2018) and internal short circuit (Feng et al. suppressing LiB fire was also reported by Wang et al. (Wang et al. 2016)
2018). Thermal runaway is associated with the occurrence of a variety of while Liu et al (Liu et al. 2018) reported Novec 1230 to also be effective.
reactions involving the SEI decomposition, anode and electrolyte LiB fire suppression can also be achieved by applying large
reactions and electrolyte decomposition (Wang et al. 2012; Ghiji, amounts of water to a battery or by submerging the battery in water (Hill
Edmonds, and Moinuddin 2021). The heat generated from the reactions 2017). Both of these methods extinguish LiB fires and cool the battery,
can propagate to surrounding cells which may result in thermal runaway, inhibiting thermal runaway and preventing re-ignition. Det Norske
cell rupture, gas venting, fire and explosion in the entire battery pack Veritas and Germanischer Lloyd (DNV-GL) (Hill 2017) investigated the
(Feng et al. 2018). Therefore, fire protection measures can be taken at the effectiveness of substances such as encapsulants (F500, FireIce,
cell, module, pack, system and compartment levels (Wilkens et al. 2017). PyroCool), aerosol (Stat-X) and water sprinklers in suppressing a LiB
The fire protection measures range from integrated internal fuses to fire and cooling a battery during thermal runaway. All systems
prevent internal shorting, fire retardants (Feng et al. 2018), power and extinguished the fires but the sprinkler system had enhanced continued
temperature management (Singh et al. 2021), module separation and cooling ability. Egelhaaf et al. (Egelhaaf et al. 2013) also demonstrated
enclosure cooling (Ouyang et al. 2019). Fig. 1 depicts the levels of fire that the addition of surfactant and gelling agents can decrease the amount
protection from cell components to compartment level. of water required for firefighting. Tests conducted by the FAA (Maloney
2014) concluded that water-based suppressants (water, Hartindo AF-31,
Hartindo AF-21, Aqueous A-B-D) are more effective compared with
non-aqueous extinguishants (Hill 2017). The Fire Protection Research
Foundation (Long et al. 2013) through testing demonstrated that water
mist can effectively suppress a fire involving an electric vehicle battery.
It has also been reported that the extinguishing effect of water mist can
be improved by adding 5% F500 solution and 5% anionic nonionic
surfactant to pure water (Luo et al. 2018). The effectiveness of adding
3% aqueous surfactant to water mist on re-ignition of an 18650 type
Lithium Cobalt Oxide battery pack (10 Ah × 4) fire was analyzed by Li
et al. (Li et al. 2015). It was shown that water mist with aqueous
surfactant was more effective in delaying the re-ignition (45 seconds)
compared to ABC powder (10 seconds), and carbon dioxide (8 seconds).
So far, water mist with additives such as surfactants and
encapsulants (Luo et al. 2018) are considered to be promising for
extinguishing and cooling LiBs. However, these additives could be
harmful to the environment (Wang et al. 2019) and thus where possible
water mist alone would be preferred. Water mist is considered due to its
low water usage compared with other water-based suppression systems.
However, further investigation on the thermal behavior of LiBs,
firefighting strategies and suppression mediums are required to establish
a preferred guideline to extinguish LiB fires.
There are typically five mechanisms associated with fire
extinguishment using water mist in a closed compartment (Mawhinney,
Dlugogorski, and Kim 1994): gas-phase cooling; oxygen depletion and
Fig. 1 The levels of fire protection for a LiB system in a compartment flammable vapor dilution; wetting and cooling of the fuel surface;
(Wilkens et al. 2017). radiation attenuation; and kinetic effects. Generally, all mechanisms take
place to some extent during the suppression of LiB fires in enclosures.
The classification of a LiB fire is controversial and can vary due to Conventional water mist systems direct droplets downward. However in
the complexity of battery components, but basically fits into classes A the present work, we have developed a novel technique, where a sheath
(fires in ordinary combustible materials, such as wood, cloth, paper, of fine water droplets was deployed horizontally across the flame front.
rubber, and many plastics), B (fires in flammable liquids, combustible Prior to this set of experiments we also carried out some preliminary
liquids, petroleum greases, tars, oils, oil-based paints, solvents, lacquers, investigations using an in-house-built testing rig consisting of a water
alcohols, and flammable gases), and E (fires that involve energized spray and fires generated by burning propane gas through a Bunsen
electrical equipment) (Kong et al. 2018; Ouyang et al. 2019). Many burner set up. The results showed that fires are suppressed rapidly on
suppression systems, such as halogen-based gases, chemical powder and activation of the mist systems where geometries and configurations of
carbon dioxide can extinguish a LiB fire, but cannot maintain cooling to the effluent mist that were capable of affecting the lift-off zone, between
suppress thermal runaway (Summer 2010). If the LiB is not sufficiently the gas source and base of the flame.
cooled, thermal runaway may continue and the battery or its neighboring
batteries may ignite (Ghiji et al. 2020).
2. MATERIALS AND METHODS
Numerous researchers have studied LiB fires to examine fire
suppression. The best practice and the optimum fire extinguishing A LiB undergoing the process of thermal runaway shows multi-stage
medium for LiBs is yet to be determined (Summer 2010; Ditch and De behavior including cell expansion stage, gas spill and auto-ignition stage,
Vries 2013). It has been reported by the US Federal Aviation jet fire stage, and stable combustion stage (Wang et al. 2017). Depending
Administration (FAA) (Maloney 2014) and US National Fire Protection on the state of charge of the LiB, additional smoking, jet fire, and
Agency (NFPA) (Summer 2010) that halogen-based products can combustion stages can occur. During thermal runaway in a LiB, the
extinguish an open LiB fire, but cannot mitigate the internal battery electrolyte and separator may significantly contribute to heat release in
temperature increase even after the extinguishment of a fire. In studies the presence of oxygen through their ignition and combustion processes.
performed by Rao et al. (Rao et al. 2015), FM200 (heptafluoropropane) Although the electrolyte and separator have low mass in a LiB,
showed superior behavior in suppressing LiB fires compared to carbon approximately 10% and 3% of the cell mass respectively (Ribière et al.
dioxide and powder extinguishants. The effectiveness of FM200 in 2012; Ping et al. 2015), they provide high effective heat of combustion
and account for approximately 80% of the heat release in a LiB fire. As
2
Frontiers in Heat and Mass Transfer (FHMT), 17, 13 (2021) Global Digital Central
DOI: 10.5098/hmt.17.13 ISSN: 2151-8629

shown in Fig. 2, the effective heat of combustion for a commercial During these tests, 5 grams of electrolyte was poured into a stainless
Lithium Manganese Oxide pouch cell was measured as steel cup with a diameter of 52.5 mm. The electrolyte was ignited after
4.03 ± 0.34 KJ g−1 with the electrolyte and separator contributing up to placing the cup under the cone calorimeter hood as shown in Fig. 3.
1.92, and 1.34 KJ g−1, respectively (Ribière et al. 2012).

Fig. 3 Liquid sample fire under the cone Calorimeter.

Fig. 2 Left: the total heat of combustion measured through The electrolyte investigated in this study is the most common
experimentation. Right: heat of combustion of battery components electrolyte used in LiBs, lithium hexafluorophosphate salts (LiPF6)
determined by thermodynamic calculations (Ribière et al. 2012). dissolved in ethylene carbonate (EC) and diethyl carbonate (DEC) (1.0
M LiPF6 in EC/DEC=50/50 v/v) with a density of 1260 kg/m3. The
It is the unpredictability of the electrolyte reaction to heating that is electrolyte was provided by Sigma-Aldrich Ltd Co. The electrolyte has a
the most difficult to define which enhance the complexity of a LiB fire boiling point of 130 °C and a flashpoint of 30 °C. The operating
event. Furthermore, the combustion characteristics of a LiB are conditions and settings of the cone calorimetry are described in Table 1.
influenced by its chemical composition, stored electrical energy and
construction. Table 1 The cone calorimetry operating conditions and settings.
To simplify the fire process and eliminate the effect of other
contributions to the combustion, a representative electrolyte only version Nominal duct flow rate (l/s) 24
of an 18650 (18 mm diameter by 65 mm length) LiB was developed as a Initial mass (g) 5
replica cell. The replica cell concept can also be appreciated by Surface area of the cup (cm²) 21.65
Computational Fluid Dynamics analysis to evaluate and validate their Start and End of test criterion ISO 5660-1:2002
simulation settings and results. A cylindrical steel rod was inserted into Duct diameter (m) 0.114
an empty 18650 cell case to fill the void taken by the battery internals, Ambient temperature (°C) 25
the case was filled with 5 grams of the electrolyte and the safety lid Barometric pressure (Pa) 97000
crimped onto the body. The replica battery contains the same mass of Relative humidity (%) 50
electrolyte as an 18650 LiB. Heat flux (kW/m²) 0
To examine water mist suppression of venting LiB fires, three sets Sampling interval (s) 1
of experiments were undertaken; Orientation Horizontal
1. Characterization of the thermal behavior of the electrolyte, O2 delay time (s) 33
2. Preliminary investigation of the interactions of water mist and CO2 delay time (s) 15
simple (diffusion) flame using a Bunsen burner and various CO delay time (s) 15
water mist nozzle configurations and
3. Examination of the behavior of LiB fires to water mist 2.2 Interaction of water mist with a Bunsen burner flame
suppression.
To investigate the effectiveness of a water mist fire suppression system
2.1 Thermal analysis of the electrolyte using a cone (WMFSS) in extinguishing LiB fires, the suppression system was
calorimeter initially tested using a propane gas Bunsen burner flame. Although the
Bunsen burner presents premix combustion and a LiB shows non-premix
The thermal properties of LiB electrolyte such as the heat release rate combustion, the interactions of water mist droplets with flames are
(HRR), and mass loss are measured using a cone calorimeter (Babrauskas considered similar without the complexity of all the combustion
et al. 2015) and a load cell similar to the tests conducted by Larsson et processes of a LiB. The effects of different flame lengths and distances
al. (Larsson et al. 2018; Larsson 2017; Larsson, Andersson, and from the water mist nozzle to the burner tip on suppression time were
Mellander 2016; Larsson et al. 2014). The HRR, considered the most examined. The flame length is measured using an image processing
important variable in fire hazard assessment (Babrauskas and Peacock technique based on the intensity of flame similar to that presented by
1992), is determined by the oxygen consumption principle, According to Zheng et al.(Zheng et al. 2011). Water was supplied to the nozzles at a
the oxygen consumption principle, each organic liquid, gas or solid pressure of 3.1 bar, generating droplet sizes in the range of 50-100 μm,
releases approximately 13.1 MJ heat per kg of consumed oxygen for according to Spray System Company manufacturer specifications. Two
complete combustion (Ping et al. 2015). The cone calorimeter tests nozzles were used, each configured in a different orientation. The first
conducted in this study are in conformance with ISO 5660 ("Reaction- nozzle was a conventional vertically-orientated brass single-hole hollow-
To-Fire Tests; Heat Release, Smoke Production and Mass Loss Rate"
2015).
3
Frontiers in Heat and Mass Transfer (FHMT), 17, 13 (2021) Global Digital Central
DOI: 10.5098/hmt.17.13 ISSN: 2151-8629

cone nozzle with the orifice diameter of 1 mm and orifice length of airflow into the enclosure. Two LED lights are located at corners of the
5 mm, and flow rate 0.05 l/min, as shown in Fig. 4. hood, illuminating the specimen from two sides. A camera with 11 mm
CMOS sensor is also employed with a capture rate of 240 frames per
second and resolution of 1080 p (1920 ⨉ 1080). The camera was placed
Nozzle in front of the specimen to record the events.

Hollow Cone Nozzle

Spray Pattern

Replica LiB
Fig. 4 Illustration of the vertically-orientated water mist apparatus for the
small-scale fire test on a Bunsen burner flame; the single-hole hollow-
Gas burner
cone conical spray pattern is shown on the left.

The second nozzle was a stainless steel flat-fan nozzle with an Fig. 6 The experimental apparatus developed for the WMFSS tests. The
equivalent orifice diameter of 0.66 mm, and a flow rate of 0.52 l/min image shows the set-up of the horizontal water mist nozzle toward the
producing a spray angle of 110 degrees, oriented horizontally as shown battery and propane gas burner used to heat the replica battery.
in Fig. 5. The rationale behind the horizontal application of water mist is
that it is applied directly to the combustion zone at the base of the flame The fan nozzle was placed at different horizontal and vertical
rather than having to overcome the fire plume momentum. Although this distances from the top of the battery. The configuration of the nozzle
strategy has been investigated computationally for various flame towards the battery and the propane burner can be seen in Fig. 6. The
configurations (typical diffusion flame, stem of a rising fireball, and water mist was activated manually at the onset of fire after venting of the
spread of flame front) (Karpov et al. 2004; Novozhilov 2003; Prasad, Li, electrolyte.
and Kailasanath 1998a; Prasad, Li, and Kailasanath 1998b), to the best
of the authors’ knowledge, it has not been investigated for jet flame 3. RESULTS
venting, as experienced in LiB fires. Similarly, a review of recent
literature revealed no results on experimental work examining horizontal
3.1 Thermal analysis of electrolyte
water mist effects on flame jet venting. Note that the efficacy of the two The combustion properties of the electrolyte were determined using a
WMFSS nozzle orientations cannot be directly compared due to the cone calorimeter. Tests were conducted in triplicate to ensure the
different flow rates and spray dynamics; this comparison is for future repeatability of the measurements. The measured HRR of the electrolyte
works. is shown in Fig. 7. The units of HRR have been converted from kW/m2
to kW, for the comparisons to be made between the experimental and
future numerical modelling of battery electrolyte emissions and fire. As
can be seen from Fig. 7, it takes approximately 200 seconds to burn
5 grams of electrolyte which is similar to the value reported by Ribiere
et al. (Ribière et al. 2012) for a LiB with 0% state of charge. The HRR
increased sharply during the first 20 seconds, with an averaged maximum
of 0.54 kW at approximately 50 seconds after the onset of combustion.
The non-uniform HRR profile is due to the different combustion
properties of the two electrolyte solvents (discussed in section 2.1); the
major contribution to the first peak is the DEC while for the second peak
it is the EC. This is consistent with experiments conducted by Eshetu et
al. (Eshetu et al. 2014).
Fig. 5 A schematic of the horizontally-orientated flat fan water mist 0.7
nozzle. Test 1
0.6 Test 2
2.3 LiB Fire suppression using a horizontal WMFSS Test 3
Heat Release Rate (kW)

0.5 Average
Having gained an understanding of the fire suppression mechanisms
from the Bunsen burner experiments, the horizontal water mist nozzle
0.4
was used to examine the extinguishment of replica LiB fires. The flat
sheet of water mist droplets was injected perpendicularly to the battery, 0.3
in the orientation as shown in Fig. 5. Propane gas burners are frequently
employed by researchers to generate heat to initiate thermal runaway 0.2
(Ruiz et al. 2018; Feng et al. 2018). Thus, the replica battery was heated
by a propane gas burner placed 150 mm below the replica cell, see Fig. 6. 0.1
Heating of the battery leads to exothermic reactions, evaporation of
electrolyte and pressure increase inside the cell which eventually bursts 0.0
the pressure relief vent built into the cap of 18650 cells. To mitigate the 0 50 100 150 200 250
Time (s)
risks associated with LiB fires and the spread of toxic materials, the tests
were conducted under a fume hood, inside an enclosure to allow Fig. 7 Heat release rate (kW) of the electrolyte (1.0 M LiPF6 in
controlled airflow. The enclosure dimensions are 3 m ⨉ 3 m (floor EC/DEC=50/50 v/v) measured using a cone calorimeter 5 grams of
area) ⨉ 3.1 m (height). There are 1.1 m air gaps below each wall to allow electrolyte was poured into a cup and was ignited using a torch.
4
Frontiers in Heat and Mass Transfer (FHMT), 17, 13 (2021) Global Digital Central
DOI: 10.5098/hmt.17.13 ISSN: 2151-8629

The average total heat released (THR) and average effective heat of Additionally, it was observed that the water mist interaction with
combustion (EHoC) of the electrolyte are 74.2 (kJ) and 15.6 (kJ/g) longer flames produced erratic behavior where the flame diminished in
respectively, shown in Table 2. THR is calculated as the area under the size and morphed into unusual geometries above the lift-off zone,
HRR curve in Fig. 7. The EHoC is obtained by dividing the total heat virtually extinguishing, then would restabilize, as water was entrained
released by the total mass loss (Babrauskas and Peacock 1992), also into the combustion at fluctuating rates as a reaction to the flame
presented in Table 2. The residual mass, predominantly LiPF6, is 0.2 g. behavior. This phenomenon may be partially due to the fire enclosure
and will not be explored further here: rather it is the lift-off zone behavior
Table 2. Thermal properties of electrolyte revealed by the cone that is of further interest.
calorimeter. For the horizontal water mist application, extinguishment of Bunsen
burner flames for the flame lengths used in Table 3 was repeated. This
Total Mass Total Heat Average Effective Heat time the fan nozzle distance from the Bunsen burner was adjusted
Loss (g) Released (kJ) of Combustion (kJ/g) vertically and horizontally (refer to Fig. 5). Again, the lift-off zone
Test 1 4.8±0.1 71.4±0.1 15.0±0.2 increased after the activation of the water mist, as a result of water
Test 2 4.7±0.1 77.1±0.1 16.2±0.2 droplets within the combustion zone, and this distance varied depending
Test 3 4.8±0.1 74.1±0.1 15.6±0.2 on the flame length. Rapid flame extinguishment occurred at
Average 4.8±0.2 74.2±0.2 15.6±0.2 approximately 1-3 seconds (i.e. immediately after the water mist reached
the burner) with flame lengths of 80 mm and 140 mm (producing flame
The average effective heat of combustion of electrolyte (1.0 M lift-off zones smaller than 5 mm) for vertical and horizontal distances up
LiPF6 in EC/DEC=50/50 v/v) is measured as 15.6 kJ/g which is in good to 50 mm and 400 mm, respectively. However, the fire extinguishment
agreement with the calculated value of 16.6 kJ/g based on the effective for a flame length of 200 mm, producing a flame lift-off zone of 20 mm,
heat of combustion of EC and DEC measured by Guo et.al (Guo et al. depended on both the horizontal and vertical distances, these are
2019) as 12.15 and 21.06 kJ/g, respectively. presented in Table 4 and Fig. 8.
3.2 Bunsen burner flame extinguishment using vertical and Table 4. Fire extinguishment time (s) using the horizontal WMFSS at
horizontal water mist different nozzle vertical and horizontal distances (mm) from the top of
The effectiveness of a conventional (vertical) water mist system (Fig. 4) the burner with the flame length of 200 mm. NA indicates that
in suppressing Bunsen burner flames was investigated by varying the extinguishment is Not Achieved.
flame length and distance between the water mist nozzle and Bunsen Vertical Horizontal Distance (mm) ±2
burner tip. The different flame lengths are produced by varying the Distance
air/propane gas mixture. 50 100 150 200 250 300 350 400
(mm)
It was observed for all experiments, that a short time after the 10 1±1 1±1 1±1 2±1 2±1 2±1 NA NA
activation of the WMFSS (once a full mist spray had been established),
20 1±1 1±1 1±1 2±1 2±1 NA NA NA
a flame lift-off zone (the distance between the Bunsen burner tip and the
30 NA NA NA 2±1 2±1 2±1 3±1 3±1
base of the flame) developed. This phenomenon is most likely the result
of entrainment of water mist into the combustion zone at the base of the 40 NA NA NA NA 2±1 2±1 3±1 3±1
flame, cooling and diluting the propane/air mixture. The extinguishing 50 NA NA NA NA NA NA NA 3±1
results of these Bunsen burner tests are presented in Table 3 and indicate 60 NA NA NA NA NA NA NA NA
that shorter flame lengths or shorter vertical distances between the burner
and nozzle produced faster extinguishing times. The shorter 80 mm
flames burn closer to the ideal stoichiometric fuel/air ratio; there is less
excess fuel. Also, geometries with the water mist in closer proximity to Nozzle
the fuel source result in higher water flux reaching the fuel source.
Conversely, as the vertical distance or flame length increased,
extinguishment time became longer. Longer flames produced by the
Bunsen burner have a higher fuel mass flow rate which requires larger
quantities of water to absorb the heat.

Table 3. Effect of Bunsen burner flame lengths and distances between

the nozzle and burner tip on the extinguishment time for the conventional
vertical WMFSS. The flame lengths are measured using image post-
processing. NA indicates that extinguishment is Not Achieved.
Vertical Extinguishment Time (s)
Distance Flame Length Flame Length Flame Length Bunsen burner
(mm) = 80 (mm) = 140 (mm) = 200 (mm)
160 1±1 1±1 33±1
190 1±1 1±1 50±1 Fig. 8 Results from Table 4 for the Bunsen burner flame length of 200
220 1±1 6±1 98±1 mm, are superimposed over the horizontal water mist geometry to
250 1±1 23±1 105±1 visualize the extinguishment zones. The image depicts the nozzle at a
280 1±1 78±1 129±1 horizontal distance of 400 mm and a vertical distance of 100 mm from
310 1±1 NA NA the Bunsen burner.
340 1±1 NA NA
At vertical distances greater than 20 mm and horizontal distances of
370 2±1 NA NA
150 mm or less (the green area), the water mist was unable to extinguish
400 3±1 NA NA the flames. At these vertical distances, the main water mist sheet passes

5
Frontiers in Heat and Mass Transfer (FHMT), 17, 13 (2021) Global Digital Central
DOI: 10.5098/hmt.17.13 ISSN: 2151-8629

through the flame, cooling the flame rather than passing through the Table 5. Fire extinguishment time (s) using the horizontal WMFSS at
flame lift-off zone and cooling/diluting the gas mixture. different nozzle vertical and horizontal distances from the top of the
At vertical distances 30 and 40 mm and horizontal distances 200 replica LiB. NA indicates that extinguishment is Not Achieved.
mm and greater (the red area), the mist droplets produce extinguishment.
At these vertical and horizontal distances, the mist sheet curves Vertical Horizontal Distance (mm) ±2
downward into the flame lift-off zone, producing gas cooling and dilution Distance (mm) 100 200 300 400
at the base of the flame resulting in suppression. As the vertical distance 10 1±1 1±1 2±1 2.5±1
is increased, the mist droplets no longer pass through the lift-off zone. It 20 1±1 1±1 2±1 2.5±1
can be concluded that the horizontal WMFSS could suppress the Bunsen 30 1±1 1±1 2±1 2.5±1
burner flame if the water mist passes through the lift-off zone. 40 1±1 1±1 2±1 2.5±1
50 1.5±1 1.5±1 2.5±1 3±1
3.3 LiB fire extinguishment using the horizontal WMFSS 60 2±1 2±1 3±1 3.5±1
The horizontal mist spray was applied to fires produced by single replica 70 NA 2.5±1 4±1 5±1
LiBs. The replica batteries were heated from below with a propane flame, 80 NA NA NA NA
inducing electrolyte venting. When ignition occurred, the water mist 90 NA NA NA NA
system was activated. Fig. 9 shows an example of the various stages of a 100 NA NA NA NA
replica LiB fire and interaction with the horizontal WMFSS located at
70 mm vertical and 400 mm horizontal from the top of the LiB. The short
flame lift-off zone observed in Fig. 9c) extends on the activation of water 100
mist and appears to reach approximately 65 mm in Fig. 9e) before
extinguishment. It should be noted that, for an unsuppressed LiB fire,
toward the end of the battery venting process, the speed of the outgoing 80
gases decreases and combustion changes from a jet fire to a diffusion Maximum Flame Lift-off length

Vertical Distance (mm)

flame and the flames emanate from the top of the battery without a flame
lift-off zone. 60

a b c
40 EXTINCTION

20

d e f
0
100 200 300 400
Horizontal Distance (mm)

Fig. 10 The nozzle-to-flame configurations producing effective

Fig. 9 Images show the evolution of the replica battery fire suppression by the horizontal WMFSS for different nozzle vertical and
extinguishment exposed to horizontal water mist positioned at 70 mm horizontal distances from the top of the replica LiB.
and 400 mm vertical and horizontal distances respectively. a) the
flammable gas (electrolyte vapor) venting; b) onset of fire; c) fire It is worth noting that the two WMFSS (vertical and horizontal)
development; d) activation of the WMFSS; e) interaction of the water examined are quite different in terms of nozzle type and flow rate, due to
mist and fire; and f) fire extinguishment. the reasons explained previously, thus it makes a direct comparison on
the efficacy of the systems difficult. The investigation into the efficacy
The effect of varying the vertical and horizontal distances between of vertical and horizontal WMFSS with the same flow rates is the subject
the water mist nozzle and the top of the replica LiB are presented in Table of future research. However, the dominant mechanisms involved in both
5 and Fig. 10. The results show that immediate flame suppression on systems of water mist suppression of Bunsen burner and replica LiB fires
activation is achieved for vertical distances less than 70 mm. The vertical are thought to be:
distance appears to be the critical factor in this technique, while there is 1. Cooling; reducing the heat feedback from the flame to the gas
no change in extinguishing behavior up to the maximum tested horizontal mixture exiting the Bunsen burner and replica LiB to a level
length of 400 mm. The mechanism associated with fire extinguishment below that required for the flame to be sustained;
appears to be similar to the Bunsen burner flame extinguishment, water 2. Dilution of the gas preheat zone primarily by water vapor. This
mist droplets moving into the combustion zone. At vertical distances of mechanism has the potential of reducing the gas concentration
80 mm and above, the majority of the water droplets have passed through below its Lower Flammability Limit.
the flame rather than the lift-off zone at the base of the flame and Water mist can also provide cooling to the battery after the
extinguishment did not occur. The successful extinguishments at 70 mm electrolyte combustion has been extinguished, potentially providing
vertical distance and greater than 100 mm horizontal distance (just above control of thermal runaway and reducing the likelihood of cascading heat
the lift-off zone) will be due to the effects of gravity; the droplets falling transfer to adjoining cells.
from the horizontal axis into the lift-off zone. The non-extinguishment at In summary, it can be seen that water mist can extinguish a LiB fire
70 mm vertical and 100 mm horizontal distance is likely to be due to the by passing through the lift-off zone. For a real LiB, if the battery surface
momentum of the water mist droplets being too large to allow is not sufficiently cooled at the end of water mist application, the internal
entrainment into the combustion zone at the base of the flame so that the reactions of the cell will not be interrupted and the LiB is likely to reignite
cooling and gas dilution did not occur. or even cause the thermal runaway of the adjacent battery. Spraying
water mist directly on the surface of the cell can effectively cool the cell
and inhibit the internal reactions (Wang et al. 2019). Thus, there is a
6
Frontiers in Heat and Mass Transfer (FHMT), 17, 13 (2021) Global Digital Central
DOI: 10.5098/hmt.17.13 ISSN: 2151-8629

nozzle position that allows the droplets to pass through the lift-off zone salts', Journal of Power Sources, 269: 804-11.
and maintain a large contact area with the surface of the cell (through https://doi.org/10.1016/j.jpowsour.2014.07.065
droplets falling from the horizontal axis due to gravity), which provides
the best extinguishing and cooling effect. This is of practical significance Feng, Xuning, Minggao Ouyang, Xiang Liu, Languang Lu, Yong Xia,
for real LiB fires and determining the optimum position will be addressed and Xiangming He. 2018. 'Thermal runaway mechanism of lithium ion
in future studies. battery for electric vehicles: A review', Energy Storage Materials, 10:
246-67. https://doi.org/10.1016/j.ensm.2017.05.013
4. CONCLUSIONS Galatro, Daniela, Maan Al-Zareer, Carlos Da Silva, David Romero,
Cristina %J Frontiers in Heat Amon, and Mass Transfer. 2020. 'Thermal
Lithium-ion batteries pose significant fire risks and the development of behavior of lithium-ion batteries: aging, heat generation, thermal
fire extinguishment systems for LiBs has not been sufficiently management and failure', 14. https://doi.org/10.5098/hmt.14.17
established to provide a satisfactory level of security in the event of a
fire. This paper highlights that water mist may be an effective method of Ghiji, Matt, Shane Edmonds, and Khalid %J Applied Sciences
extinguishment of LiB fires. Moinuddin. 2021. 'A Review of Experimental and Numerical Studies of
A water mist fire suppression technique has been introduced to Lithium Ion Battery Fires', 11: 1247.
suppress LiB fires. In this technique, water droplets are produced by a https://doi.org/10.3390/app11031247
fan nozzle and sprayed horizontally toward the flames of a replica 18560
Ghiji, Mohammadmahdi, Vasily Novozhilov, Khalid Moinuddin, Paul
LiB. The water droplets are entrained into the gas mixture at the base of
Joseph, Ian Burch, Brigitta Suendermann, and Grant. J Energies Gamble.
the flames, cooling the gas and depleting the oxygen concentration. This
2020. 'A Review of Lithium-Ion Battery Fire Suppression', 13: 5117.
horizontal water mist fire suppression technique has been tested at
https://doi.org/10.3390/en13195117
different vertical and horizontal distances to highlight the mechanisms
associated with water droplets interacting with fires. The results show Guo, Feng, Yu Ozaki, Katsunori Nishimura, Nozomu Hashimoto, Osamu
fires are suppressed rapidly on activation of the WMFSS for geometries J Combustion Fujita, and Flame. 2019. 'Experimental study on flame
that enable the water mist direct access to the lift-off zone, between the stability limits of lithium ion battery electrolyte solvents with
gas source and base of the flame. Further investigation of water mist to organophosphorus compounds addition using a candle-like wick
provide cooling for the control of thermal runaway reactions and combustion system', 207: 63-70.
cascading heat transfer to adjacent LiBs is recommended. Additionally, https://doi.org/10.1016/j.combustflame.2019.05.019
having characterized the thermal behavior of the electrolyte, this
information can be used to validate numerical simulations of LiB fire and Guo, L. S., Z. R. Wang, J. H. Wang, Q. K. Luo, and J. J. Liu. 2017.
suppression. 'Effects of the environmental temperature and heat dissipation condition
on the thermal runaway of lithium ion batteries during the charge-
discharge process', Journal of Loss Prevention in the Process Industries,
ACKNOWLEDGEMENTS 49: 953-60. https://doi.org/10.1016/j.jlp.2017.05.029
This work was supported by the Defence Science and Technology, Hill, Davion. 2017. "Considerations for Energy Storage Systems (ESS)
Department of Defence, Australia and Victoria University. Fire Safety." In. Dublin, OH, U.S.A: Det Norske Veritas (U.S.A.), Inc.
(DNV GL).
REFERENCES
Julien, Christian, Alain Mauger, Ashok Vijh, and Karim Zaghib. 2016.
Arora, Pankaj, and Zhengming Zhang. 2004. 'Battery separators', 'Lithium batteries.' in, Lithium Batteries (Springer).
Chemical reviews, 104: 4419-62. https://doi.org/10.1021/cr020738u https://doi.org/10.1007/978-3-319-19108-9_2
Babrauskas, Vytenis, Daniel T Gottuk, John R Hall Jr, Kazunori Harada, Karpov, A, V Bulgakov, A Galat, and VB Novozhilov. 2004. "Numerical
Erica D Kuligowski, Milosh Puchovsky, John M Watts Jr, and Modeling of Water Mist Suppression of Small Scale Flame Spreading
Christopher J Wieczorek. 2015. The Cone Calorimeter (Springer). over Solid Combustibles." In Sixth Asia-Oceania Symposium on Fire
https://doi.org/10.1007/978-1-4939-2565-0 Science and Technology, 922-34. https://doi.org/10.3801/iafss.fss.8-753
Babrauskas, Vytenis, and Richard D Peacock. 1992. 'Heat release rate: Kong, Lingxi, Chuan Li, Jiuchun Jiang, and Michael G J Energies Pecht.
the single most important variable in fire hazard', Fire Safety Journal, 2018. 'Li-ion battery fire hazards and safety strategies', 11: 2191.
18: 255-72. https://doi.org/10.1016/0379-7112(92)90019-9 https://doi.org/10.3390/en11092191
Blum, Andrew F, and R Thomas Long Jr. 2016. Fire Hazard Assessment Lamb, Joshua, Christopher J. Orendorff, Leigh Anna M. Steele, and Scott
of Lithium Ion Battery Energy Storage Systems (Springer). W. Spangler. 2015. 'Failure propagation in multi-cell lithium ion
https://doi.org/10.1007/978-1-4939-6556-4 batteries', Journal of Power Sources, 283: 517-23.
https://doi.org/10.1016/j.jpowsour.2014.10.081
Depetro, Aidan. 2016. 'Future Submarine Fire Safety Study', Victoria
University. Larsson, Fredrik. 2017. 'Lithium-ion Battery Safety-Assessment by
Abuse Testing, Fluoride Gas Emissions and Fire Propagation', Chalmers
Ditch, Benjamin, and Jaap De Vries. 2013. Flammability
University of Technology.
characterization of lithium-ion batteries in bulk storage (FM Global).
Larsson, Fredrik, Petra Andersson, Per Blomqvist, Anders Lorén, and
Egelhaaf, Markus, David Kress, Dieter Wolpert, Thomas Lange, Rainer
Bengt-Erik Mellander. 2014. 'Characteristics of lithium-ion batteries
Justen, and Hartung Wilstermann. 2013. 'Fire Fighting of Li-Ion Traction
during fire tests', Journal of Power Sources, 271: 414-20.
Batteries', SAE International Journal of Alternative Powertrains, 2: 37-
https://doi.org/10.1016/j.jpowsour.2014.08.027
48. https://doi.org/10.4271/2013-01-0213
Larsson, Fredrik, Petra Andersson, and Bengt-Erik Mellander. 2016.
Eshetu, Gebrekidan Gebresilassie, Jean-Pierre Bertrand, Amandine
'Lithium-ion battery aspects on fires in electrified vehicles on the basis
Lecocq, Sylvie Grugeon, Stephane Laruelle, Michel Armand, and Guy
of experimental abuse tests', Batteries, 2: 9.
Marlair. 2014. 'Fire behavior of carbonates-based electrolytes used in Li-
https://doi.org/10.3390/batteries2020009
ion rechargeable batteries with a focus on the role of the LiPF6 and LiFSI

7
Frontiers in Heat and Mass Transfer (FHMT), 17, 13 (2021) Global Digital Central
DOI: 10.5098/hmt.17.13 ISSN: 2151-8629

Larsson, Fredrik, Simon Bertilsson, Maurizio Furlani, Ingvar Albinsson, Reaction-To-Fire Tests; Heat Release, Smoke Production and Mass Loss
and Bengt-Erik Mellander. 2018. 'Gas explosions and thermal runaways Rate." In. 2015. ISO 5660, International Standard. Geneva: International
during external heating abuse of commercial lithium-ion graphite- Organization for Standardization, ISO/TC 92/SC 1 Fire initiation and
LiCoO2 cells at different levels of ageing', Journal of Power Sources, growth. https://doi.org/10.3403/bsiso5660
373: 220-31. https://doi.org/10.1016/j.jpowsour.2017.10.085
Ribière, Perrine, Sylvie Grugeon, Mathieu Morcrette, Simeon Boyanov,
Li, Y, D Yu, S Zhang, Q Hu, X Liu, and J Wang. 2015. 'On the fire Stéphane Laruelle, and Guy Marlair. 2012. 'Investigation on the fire-
extinguishing tests of typical lithium ion battery', Journal of Safety and induced hazards of Li-ion battery cells by fire calorimetry', Energy &
Environment, 15: 120-25. Environmental Science, 5: 5271-80. https://doi.org/10.1039/c1ee02218k
Liu, Yujun, Qiangling Duan, Jiajia Xu, Haodong Chen, Wei Lu, and Rodriguez, Branelle R. 2013. "Development of the International Space
Qingsong Wang. 2018. 'Experimental study on the efficiency of Station (ISS) Fine Water Mist (FWM) Portable Fire Extinguisher." In
dodecafluoro-2-methylpentan-3-one on suppressing lithium-ion battery 43rd International Conference on Environmental Systems, 3413.
fires', RSC advances, 8: 42223-32. https://doi.org/10.1039/c8ra08908f
Ruiz, V., A. Pfrang, A. Kriston, N. Omar, P. Van den Bossche, and L.
Long, Richard Thomas, Andrew F Blum, Thomas J Bress, and Benjamin Boon-Brett. 2018. 'A review of international abuse testing standards and
RT Cotts. 2013. "Best practices for emergency response to incidents regulations for lithium ion batteries in electric and hybrid electric
involving electric vehicles battery hazards: A report on full-scale testing vehicles', Renewable and Sustainable Energy Reviews, 81: 1427-52.
results." In. Quincy, Massachusetts, U.S.A.: National Fire Protection https://doi.org/10.1016/j.rser.2017.05.195
Research Foundation.
Singh, Randeep, Gero Lapp, Jason Velardo, Phan Thanh Long, Masataka
Luo, Wei-tao, Shun-bing Zhu, Jun-hui Gong, and Zheng Zhou. 2018. Mochizuki, Aliakbar Akbarzadeh, Karsten Mausolf, Kristin J Frontiers
'Research and development of fire extinguishing technology for power in Heat Busse, and Mass Transfer. 2021. 'battery Cooling Options in
lithium batteries', Procedia engineering, 211: 531-37. Electric Vehicles with Heat Pipes', 16. https://doi.org/10.5098/hmt.16.2
https://doi.org/10.1016/j.proeng.2017.12.045
Summer, Steven M. 2010. Flammability assessment of lithium-ion and
Maloney, Thomas. 2014. "Extinguishment of lithium-ion and lithium- lithium-ion polymer battery cells designed for aircraft power usage (US
metal battery fires." 46-51. Washington, DC, USA.: US Federal Aviation Department of Transportation, Federal Aviation Administration).
Administration.
Tarascon, J-M, and Michel Armand. 2001. 'Issues and challenges facing
Mawhinney, JR, BZ Dlugogorski, and AK Kim. 1994. 'A closer look at rechargeable lithium batteries', Nature, 414: 359-67.
the fire extinguishing properties of water mist', Fire Safety Science, 4: https://doi.org/10.1038/35104644
47-60. https://doi.org/10.3801/iafss.fss.4-47
Wang, Qingsong, Peifeng Huang, Ping Ping, Yulong Du, Ke Li, and
Novozhilov, V. 2003. 'Some aspects of the mathematical modelling of Jinhua Sun. 2017. 'Combustion behavior of lithium iron phosphate
fireballs', Proceedings of the Institution of Mechanical Engineers, Part battery induced by external heat radiation', Journal of Loss Prevention in
E: Journal of Process Mechanical Engineering, 217: 103-21. the Process Industries, 49: 961-69.
https://doi.org/10.1243/095440803766612775 https://doi.org/10.1016/j.jlp.2016.12.002
Ouyang, Dongxu, Mingyi Chen, Que Huang, Jingwen Weng, Zhi Wang, Wang, Qingsong, Binbin Mao, Stanislav I. Stoliarov, and Jinhua Sun.
and Jian Wang. 2019. 'A Review on the Thermal Hazards of the Lithium- 2019. 'A review of lithium ion battery failure mechanisms and fire
Ion Battery and the Corresponding Countermeasures', Applied Sciences, prevention strategies', Progress in Energy and Combustion Science, 73:
9: 2483. https://doi.org/10.3390/app9122483 95-131. https://doi.org/10.1016/j.pecs.2019.03.002
Pacala, Stephen, and Robert Socolow. 2004. 'Stabilization wedges: Wang, Qingsong, Ping Ping, Xuejuan Zhao, Guanquan Chu, Jinhua Sun,
solving the climate problem for the next 50 years with current and Chunhua Chen. 2012. 'Thermal runaway caused fire and explosion
technologies', science, 305: 968-72. of lithium ion battery', Journal of Power Sources, 208: 210-24.
https://doi.org/10.1126/science.1100103 https://doi.org/10.1016/j.jpowsour.2012.02.038
Ping, Ping, QingSong Wang, PeiFeng Huang, Ke Li, JinHua Sun, Wang, Qingsong, Guangzheng Shao, Qiangling Duan, Man Chen,
DePeng Kong, and ChunHua Chen. 2015. 'Study of the fire behavior of Yongqi Li, Ke Wu, Bangjin Liu, Peng Peng, and Jinhua Sun. 2016. 'The
high-energy lithium-ion batteries with full-scale burning test', Journal of efficiency of heptafluoropropane fire extinguishing agent on suppressing
Power Sources, 285: 80-89. the lithium titanate battery fire', Fire technology, 52: 387-96.
https://doi.org/10.1016/j.jpowsour.2015.03.035 https://doi.org/10.1007/s10694-015-0531-9
Prasad, K, C Li, and K Kailasanath. 1998a. "Numerical Modeling of Fire Wilkens, Konrad, Bjarne Johnsen, Abhishek Bhargava, and Anders
Suppression Using Water Mist. 2. An Optimization Study on Jet Dragsted. 2017. "Assessment of existing fire protection strategies and
Diffusion Flames." In.: Naval Research Lab Washington Dc Center for recommendation for future work." In Project BLUE BATTERY, Part II.
Reactive Flow and Dynamical Systems. Denmark: Danish Institute of Fire and security Technology.
https://doi.org/10.21236/ada349379
Ye, Jiana, Haodong Chen, Qingsong Wang, Peifeng Huang, Jinhua Sun,
Prasad, Kuldeep, Chiping Li, and K Kailasanath. 1998b. "Optimizing and Siuming Lo. 2016. 'Thermal behavior and failure mechanism of
water-mist injection characteristics for suppression of coflow diffusion lithium ion cells during overcharge under adiabatic conditions', Applied
flames." In Symposium (International) on Combustion, 2847-55. energy, 182: 464-74. https://doi.org/10.1016/j.apenergy.2016.08.124
Elsevier. https://doi.org/10.1016/s0082-0784(98)80143-4
Zheng, Li Gang, Ming Gao Yu, Shui Jun Yu, and Chang Lu. 2011.
Rao, Hui, Zhaoxia Huang, Hui Zhang, and Shuming Xiao. 2015. "Study 'Measurement of Flame Height by Image Processing Method', Advanced
of fire tests and fire safety measures on lithiumion battery used on ships." Materials Research, 301-303: 983-88.
In 2015 International Conference on Transportation Information and https://doi.org/10.4028/www.scientific.net/amr.301-303.983
Safety (ICTIS), 865-70. IEEE.
https://doi.org/10.1109/ictis.2015.7232158