Lithium-ion (Li-ion) batteries are widely used in a range of applications, from portable handheld devices, electrical bikes, scooters and vehicles, through to large-scale energy storage systems.
Compared to lead acid batteries, the energy density of Li-ion batteries is close to 300% higher and can sustain up to 20 times as many charging cycles. These batteries are also more environmentally friendly and are practically maintenance free.
This type of battery, although generally safe to use, does present a considerable fire hazard. Li-ion battery fires occur when the battery undergoes a process known as thermal runaway, which leads to rapid and uncontrolled release of heat, flame and gases.
This paper will give a brief overview of Li-ion battery failure modes and will then discuss their relevance to the use of optical flame detection. We will review the test methods that can initiate thermal runaway and their applicability to applications in which optical flame detectors would be used. We will ultimately settle on a method for quick and reproducible initiation of thermal runaway for use with optical flame detectors.
Lithium-ion battery failure modes
This section presents a brief overview of Li-ion battery failure modes, but it does not discuss different battery chemistry or packaging techniques. The failure modes typically fall into two categories, these are: 1) Internal causes and 2) External causes.
Internal causes relate to manufacturing defects in the battery that may lead to self-heating ignition.
External causes relate to electrical, mechanical or thermal abuse. This can take the form of over-voltage charging, a short-circuited battery, mechanical puncturing or crushing or prolonged exposed to high temperatures.
The net result of the internal or external causes is a reaction that results in the phenomenon called thermal runaway. Thermal runaway refers to a chain reaction of events within the battery that causes a rapid and uncontrolled increase in temperature. This uncontrolled temperature increase is of critical concern as it can lead to catastrophic consequences such as fires and explosions causing widespread damage and revenue loss.
A survey released in May 2020 looking at the characteristics of fires caused by batteries in electrical and electronic equipment (WEEE) showed that ‘the average cost of all incidents in 2018 was estimated at €190,000, which can represent a significant burden for an individual company. The most severe fires occurring at respondents’ facilities in the last four years gave rise to an average reported cost of damages of €1.3 million.’1
Relevance of failure modes to optical flame detectors
Optical detectors work on a ‘line of sight’ basis and can ‘see’ a fire from a long distance without the smoke or heat needing to reach the detector first. The detectors can respond to an explosion event in as little as a few milliseconds at short distances indoors and extremely quickly at long distances. For example, a flaming 1 sq. ft (0.1m2) of n-heptane fire can be detected up to 80m away in only 7.1 seconds2. This capability makes optical flame detection highly effective in a range of applications for both indoor and outdoor locations where conventional fire-detection techniques would not work.
If we therefore think of the battery failure modes already discussed, we can link these to the applications where optical flame detection may be used.
However unlikely, it is quite probable that the failure modes by internal causes, in other words manufacturing defects, can occur at any time during the life of the Li-ion battery. For this reason, it is probably more useful to consider the failure modes by external influences.
Electrical abuse, e.g. over charging
Applications that have seen the use of optical flame detection tend to be where electric vehicles, e.g. forklift trucks, are charged. Typically, this would be in an outdoors location or in a large volume building such as a warehouse. Other Li-ion battery applications that have seen a significant increase in the use of optical flame detectors are Battery Energy Storage Systems (BESS). NFPA 8553 describes how ‘Flame detection can be applied internal or external to an installation. Internal application would be to the container, enclosure, or building. It would not traditionally be applied inside a cabinet. For example, it can be used to monitor a hot isle. External application would be to ESS facilities with multiple containers. It would provide a detection if internal measures failed. It can also be tied to video cameras to provide situation information to first responders of an incident.’ The FLS-IR3-HD, with its colour video feed, has been used widely in many BESS facilities.
Mechanical abuse, e.g. crushing or puncturing.
Optical flame detectors are widely employed in a range of recycling facilities. Many fires have their origin on the concrete ‘tipping floor’ because the vehicles used to move the waste materials about, damage Li-ion batteries that have been discarded. Fire escalation in recycling facilities have seen incidents requiring dozens of firefighters4 to attend and people living near the facility to be evacuated. In waste-to-energy plants, the waste material is the process fuel and so power generation may need to be shut down leading to a loss in revenue. The FLS-IR3, has been used widely in different type of recycling facilities.
Flame detector testing
Today, there is no standard method employed by Factory Mutual (FM3260) or the European Union, as part of their construction products regulation and it’s EN54-10 standard, to initiate thermal runaway in a Li-ion battery for testing optical flame detectors. Fire and Gas Detection Technologies Inc therefore conducted a study looking at all available techniques used to initiate thermal runaway to develop a repeatable test protocol that makes practical sense for applications where optical flame detectors are used.
The three test methods reviewed were:
a) Over voltage
b) Temperature increase
c) Mechanical damage
Testing over voltage was found to be time consuming and unpredictable. Failure modes under this test condition also found that gases and smoke emitted could be detected earlier than flame in relatively small, enclosed spaces.
Testing to temperature increase could be undertaken in a few ways. One such test method employed a Li-ion battery being dropped into a burning fire that was too small for the flame detector to respond to in isolation. This test had the desired result of quickly initialising thermal runaway, and triggering an alarm, but it was unclear if the flame detector was responding to the Li-ion battery alone or the cumulative energy from the small fire and the Li-ion battery fire. Due to this uncertainty this test method was abandoned as being unreliable.
Another possibility for abuse by temperature used a hot plate to increase the temperature over time until thermal runaway occurred. As with the over-voltage test this was found to be time consuming and unpredictable with gas and smoke produced before flame. It was therefore felt the focus for testing should switch to mechanical damage as this was the more likely failure mode in applications where optical flame detection would be typically used.
A reproducible test method
Many factors were considered and tested when developing the mechanical damage protocol. Such factors included the type of mechanical damage inflicted, e.g. crush versus puncture; how to securely hold the battery or batteries; and should the batteries be impacted horizontally or vertically through the body of the battery. What evolved, through testing, was a reproducible method for initiating thermal runaway.
The reproducible method employed a guillotine with a saw-like blade suspended approximately 30cm above the battery(ies), although this height could be varied. A locking pin was used to prevent the accidental release of the blade, and in turn the locking pin was attached to a long length of cable. Once the testing technician was a safe distance from the test installation the pin was released, and the guillotine blade fell piercing the Li-ion cell(s).
It was found that puncturing a single Li-ion cell produced a highly energetic flame that lasted just a few seconds. To increase the fire burn time and detection distance additional batteries were added with the blade piercing all simultaneously.
Testing over time demonstrated the reliable and repeatable nature of the test method.
Summary
This article has presented a brief overview of Li-ion battery failure modes and their relevance to the use of optical flame detection. We also discussed various test methods used to initiate thermal runaway and ultimately, we developed a method for quick and reproducible testing for use with optical flame detectors. If you would like to see further details of the testing arrangement, please contact us.
References
2 https://www.fg-detection.com/flamespec-ir3 FlameSpec IR3 product datasheet version F120V0010.09 May 2023.
3 https://www.nfpa.org/codes-and-standards/8/5/5/nfpa-855 Standard for the Installation of Stationary Energy Storage Systems
4 https://www.letsrecycle.com/news/latest-news/viridors-kent-plastics-plant-reopens-after-fire/
About the Author
Dr Eliot Sizeland C.Eng MInstMC is Vice President of Business Development, Fire & Gas Detection Technologies, Inc.
