Lithium-Ion Energy Storage Battery Explosion
Incidents
Robert Zalosh a, Pravinray Gandhi b & Adam Barowy b
a
Firexplo LLC, Wellesley, USA
b
UL LLC, Northbrook, USA
E-mail: [email protected]
Abstract
Utility-scale lithium-ion energy storage batteries are being installed at an accelerating rate in many
parts of the world. Some of these batteries have experienced troubling fires and explosions. There
have been two types of explosions; flammable gas explosions due to gases generated in battery
thermal runaways, and electrical arc explosions leading to structural failure of battery electrical
enclosures. The thermal runaway gas explosion scenarios, which can be initiated by various
electrical faults, can be either prompt ignitions soon after a large flammable gas mixture is formed,
or delayed ignitions associated with late entry of air and/or loss of gaseous fire suppression agent.
The electrical explosions have entailed inadequate electrical protection to prevent high energy arcs
within electrical boxes vulnerable to arc induced high pressures and thermal loads. Estimates of both
deflagration pressures and arc explosion pressures are described along with their incident
implications.
Keywords: battery explosions, thermal runaways, arc flash, deflagration protection.
1. Introduction
According to the International Energy Agency (2020), worldwide energy storage system capacity
nearly doubled from 2017 to 2018, to reach over 8 GWh. The total installed storage power in 2018
was about 1.7 GW. About 85% of the storage capacity is from lithium-ion batteries.
U.S. Energy Information Administration (2019) projections are that megawatt-scale battery capacity
will approximately triple from 2018 to 2021. Based on current utility plans, EIA projects most of the
additional capacity to come from increasingly large lithium-ion energy batteries. Many such
installations are now in the range 2 MW to 20 MW, but several planned installations have capacities
greater than 100 MW. A major reason for these expansions is that the cost for lithium-ion batteries
lowered from approximately $1200 per kWh in 2010 to less than $200 per kWh in 2018 (Bloomberg,
2019).
Figure 1 shows a simplified layout of a utility-scale lithium-ion Energy Storage Battery (ESB)
installation unit. Lithium-ion cells, which can be any one of the three different geometries shown in
the diagram, are arranged in a module. These cells almost always have vents to prevent internal over-
pressurization. Modules are equipped with electrical protection (fuses) and sensors for monitoring of
voltages and (sometimes) temperatures, and either passive or active ventilation provisions.
Modules are placed and electrically interconnected in vertical racks as indicated in Figure 1. Each
rack has a rack-level battery management system that communicates with the module sensors, and
also has one or more DC connectors and fuses. A typical rack has a voltage of about 1000 VDC. The
racks are installed in an enclosure, sometimes called a Battery Energy Storage Unit, equipped with
13th International Symposium on Hazards, Prevention, and Mitigation of Industrial Explosions
Braunschweig, GERMANY – July 27-31, 2020
system level Battery Management System (BMS) for electrical control, a Heating Ventilation Air
Conditioning (HVAC) system, and a fire detection and suppression system.
Figure 1. A simplified Battery Energy Storage Unit example.
Battery Energy Storage Units have doors for operating and maintenance personnel and for installation
and replacement of equipment. A variety of Energy Storage Unit sizes have been used to accommodate
the varying electrical energy and power capacities required for different applications. Several designs
are variations or modifications of standard ISO freight containers, with nominal dimensions of 2.4 m x
2.4 m x 6 m, and 2.4 m x 2.4 m x 12 m. Other designs are up to 16 m in length.
Installations are being located in rural, urban, and suburban areas, often adjacent to a solar power or
wind turbine generator for charging the battery. There are also many behind-the-meter installations
in which the energy storage battery can be charged by the utility electricity grid and often has
electrical controls to allow discharging through an inverter and utility interface back into the grid.
Unfortunately, there have been a large number of energy storage battery fires in the past few years.
For example, in South Korea, which has by far the largest number of energy storage battery installations,
there were 23 reported fires between August 2017 and December 2018 according to the Korea Joongang
Daily (2019). A Korean government led investigation of these incidents found that one important
cause of the fires was defective battery protection systems. The failure of these protection systems in
some incidents caused components to explode. Other fires in South Korea and elsewhere have involved
explosions from other causes, including a vulnerability of some batteries to operate at abnormally high
temperatures under certain fault conditions (Yonhap News Agency, 2020).
The objectives of this paper are 1) to describe some generic scenarios of energy storage battery fire
incidents involving explosions, 2) discuss explosion pressure calculations for one vented deflagration
incident and some hypothesized electrical arc explosions, and 3) to describe some important new
equipment and installation standards and regulations intended to prevent energy storage battery
explosions.
2. Thermal Runaway Gas Explosion Incidents
Various recent papers, for example Guo et al (2018) and Li et al (2019), describe how any one of
several fault conditions, including electrical faults, overcharging, and particulate/moisture
contamination, can lead to an escalated temperature in one lithium-ion cell, causing deterioration and
eventual failure of the cell separator, with subsequent electrolyte decomposition and elevated vapor
pressure. This leads to a thermochemical runaway venting in the cell that can then propagate to
many other cells in an energy storage battery module. The vented thermal runaway causes
flammable gas to be emitted into the battery enclosure, where the resulting flammable
mixture can be ignited by hot module casings, electrical connectors, or ejected sparks from the involved
module. This generic explosion hazard has been discussed previously by Marr et al (2013) and Baird
et al (2020), among others.
Two or three different runaway gas explosion hazard scenarios can occur. In one scenario, the
flammable gas mixture is ignited soon after it is formed near the initiating module, such that there is
only a minor deflagration and a subsequent fire. In the second scenario, batteries in thermal runaway
release flammable gases without igniting initially, and a delayed explosion associated with the
accumulation of additional flammable atmosphere then occurs. Another variation of the delayed
explosion incident described below involves an initial fire with accumulation of incomplete combustion
products and possibly fire suppression agent, until something happens, e.g. oxygen addition to the rich
gas mixture, to suddenly render the mixture ignitable.
2.1 Prompt explosion incidents
One particular Korean energy storage battery incident in which a prompt thermal runaway occurred
was investigated and described by Kim et al [2019]. The battery portion of the 1.0 MWh Energy
Storage System (ESS) consisted of 15 racks, each containing nine modules, which in turn contained
22 lithium ion 94 Ah, 3.7 V cells. A 250 kW Power Conversion System (PCS) was connected to the
battery through a Battery Control Panel. Electrical protection included module fuses and rack fuses.
The initiating event in this incident was a series of PCS faults. As indicated in Figure 2, the faults
occurred shortly after the battery was fully (95%) charged and began to discharge. However, t h e
P C S f a u l t s p r o d u c e d a c u r r e n t r e v e r s a l a n d f o r c e d b a t t e r y c h a r g i n g , with
a thermal runaway when the battery was at a 90.8% State of Charge (SOC).
Figure 2. Battery state of charge preceding thermal runaway (from Kim et al, 2019)
The electrical faults in this incident caused a cell voltage imbalance and ground fault in one of the
battery modules. Data records indicate the cell voltage imbalance rose to 3 V, whereas the battery
manufacturer recommends the system be tripped when voltage imbalance exceeds 0.10 V. The
corresponding recorded cell temperature rise was at least 64 oC, producing a cell temperature in
excess of the 80 oC threshold for the onset of cell membrane deterioration and exothermic
decomposition reactions (Guo et al, 2018). The absence of any battery data after the onset of thermal
runaway suggests that the flammable electrolyte decomposition gases were ignited soon after thermal
runaway in at least one module.
Recommended safety improvements in Korean ESS installations include Surge Protection Devices (Kim et
al) and limitations on the charging rate and the maximum allowable State of Charge (Yonhap News, Feb 6,
2020).
In other battery ESS thermal runaway incidents, the flammable gases accumulated in the battery
enclosure for a while because of inadequate enclosure ventilation. Ignitions eventually occurred
producing varying levels of enclosure explosion damage.
2.2 Delayed explosion incident in Arizona
One delayed explosion battery ESS incident is particularly noteworthy because the severe firefighter
injuries and unusual circumstances in this incident were widely reported (Renewable Energy World,
2019). This incident occurred at the Arizona Public Service (APS) McMicken Energy Storage facility
in Surprise, Arizona, 28 miles northwest of Phoenix. As shown in Figure 3, the facility is adjacent to
an APS substation. According to the APS posted updates (APS 2019), this is a 2 MW, 2MWh facility
put into service in March 2017.
Figure 3. APS ESS in Surprise, Arizona
Figure 4 is a view into the APS ESS (after the incident) showing the racks and a module transporter
in the aisle. Besides a HVAC system, the ESS was equipped with an automatic fire suppression
system that was designed to discharge a gaseous suppression agent, probably FK-5-1-12
(CF3CF2C(O)CF(CF3)2), upon fire detection.
Figure 4. View into APS ESS after incident (APS, 2019).
On the afternoon of April 19, 2019, alarms triggered by the Battery Management System were
received by battery monitoring personnel. At about 5 PM, smoke was seen b y a p a s s e r b y venting
from the ESS (APS, 2019). The Surprise Fire Department was called and responded to the scene.
Upon arrival, firefighters also observed a dense ground level white smoke or fog as can be seen in
Figure 5.
Figure 5. Firefighters observing white smoke or fog around APS ESS in Surprise, AZ (Chief Ruiz)
The Surprise firefighters were concerned about possibility of hazardous materials inside and venting
from the ESS and requested assistance from a four-man hazmat team at the nearby Peoria Fire
Department. The four Peoria firefighters approached the still closed ESS structure with handheld gas
detectors. The readings, including very high carbon monoxide levels, and telephone conversations with
APS personnel caused them to pause for several hours before attempting entry into the ESS.
According to the Surprise Fire Department incident report, when the external gas concentration and
temperature readings subsided, the firefighters did open the ESS door. S o o n a f t e r t h e
d o o r w a s o p e n e d a t about 8 PM, an explosion occurred! The blast wave pressures and
velocities caused two firefighters to be blown approximately 23 m away from the door, passing under
the chain link security fence shown in Figures 3 and 5. The emerging flames from the ESS produced
severe burn injuries. All four Peoria firefighters were hospitalized, two with severe burn and trauma
injuries, including facial and head injuries despite their wearing helmets and face shields. Four
firefighters from the Surprise Fire Department experienced minor injuries.
Figure 6 shows the front of the APS ESS after the explosion. There is some bulging distortion of the
wall and there is char above the open doorway. Pressure induced deformations of the ESS end wall
and ceiling are also visible in Figure 4. A door on the end wall was blown off.
Figure 6. APS Surprise ESS after incident.
An incident investigation led by APS is still underway with occasional progress reports posted on
an APS web page (APS, 2019). As of this writing, the following investigation results were posted
in the most recent (November 2019) APS progress report.
A single rack of modules was compromised by the initiating thermal event; the fire did not
spread to surrounding racks.
After the initiating event, the fire suppression agent was discharged.
The compromised modules emitted a mixture of explosive gases, which built up in the
container.
The battery modules did not themselves explode; the gas mixture reached certain
concentrations, came in contact with a thus far unidentified ignition source and subsequently
exploded.
The APS updates so far have not provided an explanation for the source of the white smoke or fog
seen emanating from the ESS during this incident, but there has been a lot of conjecture about this. A
similar looking fog is often observed during the discharge of gaseous agent fire suppression systems.
The fog induced by the gaseous agent discharge is usually due to atmospheric water vapor condensing
because of the cooling effect of the gas expansion as it is discharged from the suppression system
nozzles. In the APS ESS incident, the ESS atmosphere probably also contained condensable flammable
vapors, such that the observed fog could be a combination of condensed flammable vapors and water
droplets, possibly together with some FK-5-1-12.
The explosion analysis described below is not part of the APS investigation, nor has it been reviewed
by the investigation team. The objective of this analysis is to develop an understanding of the flammable
vapor mixture size and burning velocity that was ignited upon opening the door. The premise is that
the ignition source had been present all along, but the concentration of FK-5-1-12 in
the atmosphere had been sufficiently high to prevent re-ignition and provide some inerting. The FK-
5-1-12 minimum design concentration for flammable liquid fires, per NFPA 2001 (2015) guidelines,
is 5.9 %, with higher concentrations probably being needed depending on the electrolyte cup burner
Minimum Extinguishing Concentration (MEC). FK-5-1-12 cup burner MECs measured by Ingram
(2013) for several electrolytes ranged from 6.2% to 6.7%. Since the NFPA 2001 minimum design
concentration is equal to the cup burner MEC multiplied by 1.3, it is quite possible that the FK-5-1-
12 design concentration for the ABS ESS was about equal to or higher than the NFPA 2001 specified
propane inerting concentration of 8.1%; thereby providing inerting to prevent re-ignition. When the
APS ESU door was opened, a significant amount of FK-5-1-12 and other gases flowed out and probably
were replaced with air to form a flammable mixture ignited by the still hot modules or by flame from
a suddenly re-ignited fire in the rack.
Deflagration vent areas required to comply with NFPA 68-2018 have been calculated for the APS
BSS enclosure using different combinations of laminar burning velocity, S u, ratio of flammable mixture
partial volume fraction, Xpv, i.e. ratio of flammable mixture volume to enclosure volume, and the
reduced deflagration pressure, Pred. The NFPA 68-2018 vent sizing methodology includes adjustments
for turbulence enhancement of the laminar burning velocity and partial volume effects. The calculated
vent areas have been compared to the actual doorway opening (2 m 2), and also to an estimated vent
area of 2.5 m2 corresponding to the total vent area of the doorway together with partial ventilation
openings. A second door was blown open at some time during the deflagration and provided some
additional obstructed vent area that is difficult to estimate because of the door inertia. The normal ESS
ventilation system is designed to be shut down prior to the discharge of the gaseous fire suppression
agent, but there clearly was some leakage causing the emission of the observed fog before the door
was opened.
Based on explosion testing with structures of similar size and construction (including ISO mobile
containers), the observed damage in Figures 4 and 5 are estimated to be caused by an internal
pressure in the range 0.10 bar to 0.17 bar. These P red values together with various values of S u and
Xpv that result in a calculated required vent area of either 2 m 2 or 2.5 m2 are shown in Table 1. The
APS ESS enclosure dimensions used in these calculations are 2.1 m x 3.5 m x 6.1 m.
Table 1: NFPA 68 Parameters Resulting in Observed APS ESS Av
Pred (bar) Su (m/s) Xpv Av (m2)
0.10 0.20 0.17 2.00
0.10 0.25 0.063 2.00
0.10 0.20 0.5 2.51
0.10 0.25 0.122 2.49
0.14 0.20 0.36 2.01
0.14 0.31 0.046 2.00
0.14 0.28 0.063 2.00
0.14 0.22 0.19 2.00
0.14 0.31 0.075 2.50
0.17 0.20 0.60 2.00
0.17 0.22 0.31 2.00
0.17 0.22 1.0 2.49
Results in Table 1 show that NFPA 68 calculated areas are roughly equal to the actual enclosure vent
area for a mixture burning velocity no greater than 0.22 m/s if most of the ESS enclosure was filled
with a flammable gas mixture when the explosion occurred. A burning velocity of about 0.20 m/s
has been calculated for methane-hydrogen mixtures with an equivalence ratio of about 0.60 providing
the ratio of hydrogen-to-methane is less than about 0.60 (Sarli and Benedetto, 2007). It also is
applicable to mixture equivalence ratios of 1.5 with hydrogen-to-methane ratios less than about 0.50.
Similarly, a burning velocity of about 0.20 m/s has been calculated for vented runaway lithium ion
battery gases at equivalence ratios of about 0.60 and about 1.6 to 1.8 (Johnsplass et al 2017).
If only a small fraction (less than 10%) of the enclosure had a flammable mixture at the time of the
explosion, the mixture burning velocity may have been as large as 0.31 m/s, which is a representative
burning velocity calculated (Johnsplass et al, 2017) for the flammable mixtures produced in lithium-
ion cell runaways at equivalence ratios of about 0.7 and about 1.5 to 1.7. Since the flammable mixture
was formed so soon after the door was opened, and areas away from doorway probably retained
significantly higher concentrations of FK-5-1-12 so as to prevent flame propagation, it is more likely
that Xpv is not much greater than 0.10. The results in Table 2 with values of X pv ≤ 0.10 suggest that
the mixture burning velocity was in the range 0.25 to 0.31 m/s.
3. Thermal Runaway Explosion Prevention Measures
The lithium-ion energy storage battery thermal runaway issue has now been addressed in several
recent standards and regulations. New Korean regulations are focusing on limiting charging to less
than 90% SOC to prevent the type of thermal runaway conditions shown in Figure 2 and in more
recent Korean battery fires (Yonhap News 2020). The new NFPA 855 standard for energy storage
systems requires that “a listed device or other approved method shall be provided to preclude, detect,
and minimize the impact of thermal runaway.” A footnote elaboration explains that the approved
method can be a battery management system evaluated to UL 1973 or UL 9540.
Another NFPA 855 requirement for lithium-ion systems is for explosion control, specified to be either
explosion prevention systems in accordance with NFPA 69 or deflagration venting in accordance
with NFPA 68. There is an exception for systems that have had large-scale fire tests demonstrate
“that flammable gas concentrations in the room, building, or walk-in unit cannot exceed 25 percent
of the LFL in locations where the gas is likely to accumulate.” An NFPA 855 Annex paragraph
explains that “a UL 9540A test or equivalent methodology should evaluate the fire characteristics of
an ESS that undergoes thermal runaway, such as might occur due to a fault, physical damage, or
exposure hazard. The data generated by the large-scale fire testing is intended to be used by
manufacturers, system designers, and AHJs [Authorities Having Jurisdiction] to determine the fire
and explosion protection required for an ESS installation.” Use of the UL 9540A tests for
characterizing thermal runaway gas explosibility properties and release volumes is described in
another paper by the authors for this symposium (Barowy, Gandhi, and Zalosh 2020).
A recent New York City (2019) Fire Department regulation for outdoor battery energy storage
systems also requires thermal runaway fire testing evaluations and has two additional requirements
for explosion mitigation that are analogous to the NFPA 855 requirements. It is also required that
venting is positioned and oriented so that blast waves and any potential shrapnel are directed up and
away from personnel. The new NYC regulation also requires a manually-operated purge system to
exhaust smoke and gases by a switch installed at the fire department or utility connection. This
requirement is intended to prevent firefighter injuries of the type that occurred in the APS Surprise
AZ incident described above.
4. Arc Flash Explosion Incidents
Several lithium-ion battery energy storage system incidents involved electrical faults producing an
arc flash explosion. The arc flash in these incidents occurred within some type of electrical enclosure
that apparently could not withstand the thermal and pressure loads generated by the arc flash. One
example of an electrical enclosure that is designed to withstand a limited/controlled arc flash is a DC
contactor. Each rack in the ESS enclosure is usually equipped with at least one high-voltage DC
contactor to make and break electrical connections to the rack modules. The Battery Management
System also has a DC contactor for overall system charging and discharging current.
Figure 7 shows an example of a DC contactor design and operation. The stationary contact connectors
at the top of the contactor are connected to the rack electrical circuit. The movable contact bridge
position is dependent on the presence or absence of current in the contractor coil. When the contact
bridge approaches or disengages from the stationary contact an electric arc is formed as shown in the
diagram. Since the arc can transmit a lot of heat to the contacts, there is often a gas within the
contactor housing to provide cooling. In some cases, as shown in this diagram, hydrogen is used. In
other cases, nitrogen or sulfur hexafluoride is used in these contactors.
Figure 7. DC Contactor operating diagram (from www.epcos.com High-Voltage Contactors)
DC contactors are rated for their maximum operating voltage and nominal current load. They can
also withstand temporary over-currents for a short period of time. Over-currents produce more
energetic arcs upon opening and closing contacts, and these can challenge the strength and thermal
resistance of the ceramic housing, which often has an epoxy resin sealant. Some of the Korean ESS
fire incidents were reportedly attributed to catastrophic failure of DC contactors due to large electrical
surges (associated with ground faults and short circuits) that fuses were not able to interrupt
sufficiently fast to prevent excessive currents in the DC contactors.
The issue addressed in this section of the paper is the explosion pressures generated from energetic
arcs within either a DC contactor or another electrical enclosure, such as a battery module or a rack
protection enclosure (within which the DC contactors are situated). Chévrier et al (1993) have shown
via Computational Fluid Dynamic modelling and experiments that the arc current causes joule heating
and thermal radiation of the gas between the electrodes such that the ionized arc gas temperatures are
approximately 15000 oK. A combination of expansion of the hot ionized gas and radiant heat
absorption of the surrounding gas is responsible for the arc induced pressure rise in a contactor
enclosure. Chévrier et al measured and calculated a peak pressure of about 1 bar for an 8 kA arc of
9 ms duration in a thermal expansion circuit breaker.
Thoren (1962), using results reported by Hennecke and Horn (1960), derived a simple equation for
estimating the peak pressure from a confined arc flash assuming half of the arc flash energy heats the
air uniformly in the electrical enclosure. The equation is
2 NUI arc t arc
P= (1)
V
where
P is the arc peak pressure in bar-g,
N is the number of arcs,
U is the voltage drop per arc in kilovolts,
Iarc is the arc current in kiloamps,
tarc is the arc duration in seconds,
V is the enclosure volume in m 3.
Equation 1 was used by Crawford et al (1993) to calculate a pressure of 3.2 bar-g that would be
generated from an 8.1 kA, 140 V, three cycle arc, in an 0.035 m 3 motor terminal box, similar to some
that had experienced ruptures during several chemical plant incidents. They also estimated that the
burst pressure of many motor terminal boxes is in the range 0.7 bar-g to 3 bar-g, and that the cover
of one typical box would fracture at a pressure of 0.72 bar-g. Motor terminal boxes are sturdier than
the electrical boxes used in energy storage systems, which would likely have lower fracture or rupture
pressures.
Hoagland et al (2017) conducted arc flash tests in a 0.13 m 3 enclosure and found that their measured
pressures agreed very well with pressures calculated using Equation 1. This implies that Equation 1
can also be used to estimate the arc flash pressures that can be produced in various hypothesized arcs
in representative energy storage battery electrical enclosures. Table 2 shows the results of some such
calculations for arc parameters that are characteristic of (albeit not directly attributable to any
particular ESS) arc flash parameters specified by battery manufacturers for determining required
thermal personnel protective equipment to be used by electrical service personnel working on or near
circuits while the battery is near 100% SOC. The calculated pressures are conservatively high estimates
for enclosures which have ventilation openings. Not shown in Table 2 are the much larger explosion
pressures calculated for hypothesized arc faults in the much smaller DC contactors.
Table 2: Calculated Arc Flash Electrical Enclosure Pressures
Enclosure V (m3) U (kV) Iarc (kA) tarc (s) P (bar-g)
Rack Encl. 1 0.037 0.9 2.5 0.025 3.1
Rack Encl. 2 0.036 1.1 6.5 0.021 8.4
Module 1 0.033 0.09 3.5 0.026 0.49
Module 2 0.034 0.8 0.38 0.050 0.90
Module 3 0.053 0.08 0.60 0.026 0.047
The calculated arc flash pressures for rack enclosures are an order-of-magnitude larger than those for
module enclosures, and are higher than the enclosures can withstand. However, these enclosures
usually are vented, which make actual arc pressures significantly lower than the values shown in
Table 2. Another significant factor affecting the enclosure strength in an actual arc incident is the
radiant heat transfer to the enclosure walls, thereby weakening them significantly. Therefore, it is
likely that the structural integrities of both rack and module enclosures are compromised in energetic
arc explosions. This is probably why at least some of the fire incidents involved arc induced failures
of electrical enclosures.
5. Conclusions
Several large-scale lithium-ion energy storage battery fire incidents have involved explosions. The
large explosion incidents, in which battery system enclosures are damaged, are due to the deflagration
of accumulated flammable gases generated during cell thermal runaways within one or more modules.
Smaller explosions are often due to energetic arc flashes within modules or rack electrical protection
enclosures. These smaller explosions can either initiate or exacerbate energy storage system fires.
Estimates of the vented deflagration pressure in one incident have been used, along with various
assumptions of gas mixture volume and laminar burning velocity, to calculate NFPA 68 deflagration
vent areas for the energy storage system enclosure. Comparisons of calculated and actual incident
vent areas suggest that only a small fraction of the enclosure volume may have had a flammable
mixture, with a laminar burning velocity in the range 0.25 to 0.31 m/s, at the time of the explosion. It
is likely that the rest of the enclosure had a sufficient residual concentration of suppression agent gas
to prevent more extensive flame propagation.
Recently issued standards and regulations for lithium-ion storage battery systems now explicitly require
protection to prevent and/or control thermal runaways leading to possible deflagrations. Other recently
imposed measures are intended to limit electrical fault energies in battery control units. By limiting the
arc flash energy, these new measures may allow modules and other electrical enclosures to withstand
otherwise excessive arc explosion pressures.
References
APS, (2019) McMicken investigation continues, aps.com/en/About/Our-
Company/Newsroom/Articles/Equipment-failure-at-McMicken-Battery-Facility.
Baird, A., Archibald, E., Marr, K., and Ezekoye, O. (2020) Explosion hazards from lithium-ion
battery vent gas, J. Power Sources, v 446, 227257.
Barowy, A., Gandhi, P., and Zalosh, R., (2020) Explosibility properties of gases from lithium-ion
storage battery thermal runaways, for presentation at ISHPMIE 2020.
Bloomberg (2019) Battery Reality: There’s Nothing Better Than Lithium-Ion Coming Soon,
https://www.bloomberg.com/news/articles/2019-04-03/battery-reality-there-s-nothing-better-
than-lithium-ion-coming-soon.
Chévrier, P., Barrault, M., and Flévet, C. (1996) Hydrodynamic model for electrical arc modelling,
IEEE Trans. Power on Power Delivery, v 11, pp 1824-1829.
Crawford, K., Clark, D., and Doughty, R. (1993) Motor terminal box explosions due to faults, IEEE
Trans. On Industry Applications, v 29, pp 72-81.
Di Sarli, V. and Di Benedetto, A. (2007) Laminar burning velocity of hydrogen-methane premixed
flames, International Journal of Hydrogen Energy 32, pp 637-646.
Guo, D. et al, (2018) The causes of fire and explosion of lithium ion battery for Energy Storage, 2nd
IEEE Conference on Energy Internet and Energy System Integration, ISBN-13: 978-1-5386-
8549-5.
G. Hennecke and W. Horn, (1960) Pressure development and radiation in the case of arc short circuits
in switchgears,” Conti Elektro Berichte, vol. 6, no. 3, pp. 119-128.
Hoagland, E., Maurice, C., Haines, A., and Maurice, A. (2017) Arc flash pressure measurement by
the physical method, effect of metal vapor on are blast, IEEE Trans. On Industry Applications, v
53, pp 1576-1582.
Ingram, J. (2013) Lithium-Ion Batteries: A Potential Fire Hazard, Power Magazine.
International Energy Agency (2020), Energy storage – Tracking Energy Integration – Analysis – IEA,
https://www.iea.org/reports/tracking-energy-integration/energy-storage.
Johnsplass, J. et al (2017) Simulation of burning velocities in gases vented from thermal runaway
lithium ion batteries, Proceedings of the 58th SIMS, DOI: 10.3384/ecp17138157.
Kim, H., Kim, S.H., and Song, H-Y. (2019) DC Ground fault induced fire accident in lithium ion
battery ESS, Transactions Korean Institute of Electrical Engineers, v 68, pp 1470-1476.
Korea Joongang Daily, ESS units caught on fire due to bad management, June 12, 2019.
Li, W. et al (2019) Theoretical and experimental analysis of the lithium-ion thermal runaway
process based on the internal combustion engine combustion theory, Energy Conversion and
Management, v 185, pp 211-22.
Marr, K., Somendepalli, V., and Horn, Q. (2013) Explosion hazards due to failures of lithium-ion
batteries, Proceedings AIChE Spring Meeting and 9th Global Congress on Process Safety.
NFPA 68 (2018) Explosion Prevention by Deflagration Venting, National Fire Protection
Association.
NFPA 69 (2019) Standard on Explosion Prevention Systems, National Fire Protection Association.
NFPA 855 (2020) Standard for the Installation of Stationary Energy Storage Systems, National Fire
Protection Association.
NFPA 2001 (2015) Standard on Clean Agent Fire Extinguishing Systems, National Fire Protection
Association.
New York City Fire Department Rule 3 RCNY 608-01 (2019) Outdoor Energy Storage Systems,
https://rules.cityofnewyork.us/tags/energy-storage.
Renewable Energy World (2019), APS battery energy storage facility explosion injures four
firefighters; industry investigates, https://www.renewableenergyworld.com/2019/04/23/aps-
battery-energy-storage-facility-explosion-injures-four-firefighters-industry-investigates/#gref.
Thoren, B. (1962) Arcs on power lines and in high voltage switchgear, Elteknik, vol. 139-143, pp.
139-143.
UL 1973 (2018) Standard for Batteries for Use in Stationary, Vehicle Auxiliary Power and Light
Electric Rail (LER) Applications, Underwriters Laboratories Inc.
UL 9540 (2017) Standard for Energy Storage Systems and Equipment, Underwriters Laboratories
Inc.
UL 9540A (2019) Standard for Test Method for Evaluating Thermal Runaway Fire Propagation in
Battery Energy Storage Systems, Underwriters Laboratories Inc.
U.S. Energy Information Administration (July 2019) U.S. utility-scale battery storage power capacity
to grow substantially by 2023, https://www.eia.gov/todayinenergy/detail.php?id=40072
Yonhap News Agency, Faulty batteries blamed for ESS fires: panel, February 13, 2020.
