The tests show that water-based systems extinguish and cool burning lithium-ion batteries, but the extinguishing agent needs to be injected directly into a battery module/pack in order to be efficient. Gaseous extinguishing agents put out a fire but provide less cooling.
There is a lack of standardised fire test methods and design and installation guidelines for extinguishing systems in battery rooms on ships, but the results of the project provide a good basis for starting to develop them.
Vessel emissions can be reduced through electric or electric hybrid operation. Electrified ship operations now focus on ferries and vessels in coastal traffic, but in the near future passenger ships, cruise ships and cargo vessels for shorter routes and supply vessels for oil rigs may also become relevant. Vessels also use batteries and electrical propulsion as a support system, to supplement diesel operation and to optimise diesel engines through a power boost by switching on electrical propulsion when extra power is needed, such as in severe weather or when disengaging diesel operation during port visits.
Battery-based energy storage introduces new risks. Lithium-ion (li-ion) batteries are most common due to their high energy density and a long life. However, li-ion batteries remain stable only within a certain operating range, and are also sensitive to external mechanical influences and heat. If the battery exceeds its operating range or is subject to abuse such as mechanical impact, it can enter a state of self-heating, known as thermal runaway, which can cause fire, explosion and emissions of combustible and toxic gases. Such occurrences naturally expose a ship at sea to specific hazards, so quick and effective intervention becomes important.
Great need for new knowledge
Knowledge about the effectiveness of various extinguishing agents and extinguishing methods to combat fires in li-ion batteries is limited. One of the main reasons for this is the absence of reliable test data based on well-documented fire tests. Extensive fire testing of batteries costs a lot and involves dealing with specific hazards. After-treatment and handling of the waste is also complicated and requires special expertise.
A literature study (see SP Report 2017:34) indicates that there are contradictory recommendations about suitable extinguishing agents for batteries, such as the use of water as an extinguishing agent. One explanation for this may be a tendency to confuse lithium-ion batteries with lithium-metal batteries, both of which are often called lithium batteries, but have very different properties. Descriptions of tests conducted and made public generally do not provide sufficient detail to enable a reliable analysis of the results presented. For example, in some commercial videos showing fire extinguishment of a single battery cell, extinguishing agents are applied in the final stage of the fire or in excessive amounts.
Simulated battery module in a test room
The aim of the project was to develop knowledge based on a relatively simple test set-up. For this reason, extinguishing tests were conducted with a Li-ion battery cell placed in a simulated battery module in a small test compartment. The test used a commercially available, pouch type battery classified for applications such as electric vehicles. The nominal capacity was 20 amp hours (Ah) and the cell included a carbon anode, a cathode of lithium iron phosphate (LFP, LiFePO4) and an organic electrolyte. Fire was started by heating the battery cell with an electric heating coil in contact with the bottom surface area of the cell. The heating led to a thermal runaway in the cell, increasing pressure in the cell enclosure and causing the cell to either open or burst. A small propane gas flame placed above the cell ignited the combustible gases flowing out of the cell. The test set-up and procedure was chosen in order to provide a repeatable test to allow for comparison of different extinguishing media.
The cell was installed in a simulated battery module, a steel sheet cube with a side dimension of 400 mm (see Figure 1). The module had four layers of perforated plates to emulate the difficulty in reaching a specific cell due to the dense packaging within a module.
The simulated battery module was placed on a table in a 3.7-by-3.7-metre test compartment with a ceiling height of 2.5 metres. The compartment was sealed closed in all tests. Figure 2 shows the test room and the location of the simulated battery module.
Tests with different extinguishing agents and application methods
The extinguishing agent was applied either in the room (total compartment protection) or inside the battery module (local application), and several different extinguishing agents were tested. Temperature measurements on and near the battery cell and video observations were used to evaluate the effectiveness.
The experiments with total compartment protection was carried out with three different nozzles: a traditional water spray nozzle and two water mist nozzles (low and high pressure nozzles respectively). None of the nozzles had any noticeable effect on the fire, which can be explained by the fact that the water droplets did not reach it. However, any of the systems could be used to prevent or delay the spread of fire from a battery compartment to adjacent compartments because the water droplets helped cool the space. Unfortunately, no total compartment protection tests could be conducted with gaseous agents because that exceeded the project’s budget.
Tests in which the extinguishing agent was injected directly into the simulated battery module were more successful. The following extinguishing agents were tested:
- Water applied with different types of swirl nozzles from the top of the module.
- Class A foam (0.5% admixture) applied with the same type of swirl nozzles used for clean water.
- Class F foam (premixed solution) applied with the nozzles as above.
- Compressed air generated foam (CAFS) with various foam consistencies.
- Aqueous Vermiculite Dispersion (AVD), an extinguishing agent consisting of vermiculite dissolved in water.
- Nitrogen gas.
The results showed that the water-based systems (with or without additives) quickly extinguished the fire and effectively cooled a plate thermometer that simulated adjacent cells. Class A and F foams did not improve fire extinguishing ability compared to pure water. However, Class A foam reduces the surface tension of water, which could improve the penetration ability in a real battery module. Class F foam has the potential to improve the cooling capacity because it consists of a saline solution. It has a disadvantage, though, as the salts increase the risk of short-circuiting battery cells that are not directly affected by the fire.
The compressed air foam system (CAFS) extinguished the fire, but the tests show that the foam expansion ratio is important for the foam to flow out and quickly fill up a volume. When the foam contained too much air, it did not flow as well, took longer to fill the volume and had an inferior cooling capacity.
Nitrogen gas extinguished the fire, but the cooling capacity was significantly inferior to the other extinguishing agents. This poses a risk that cells affected by the heat from a fire will not cool down and end up in self-heating mode even if the initial fire has been extinguished. As a result, it is crucial that a gaseous extinguishing agent is added for a long time to allow cells heated by a fire to cool.
Only one experiment was conducted with Aqueous Vermiculite Dispersion (AVD). According to the manufacturer, the extinguishing agent should form an insulating layer over a fire, thus limiting the supply of oxygen. The extinguishing agent extinguished the fire in the test, but its viscosity affects the flow, which meant that the extinguishing time was longer than for the other extinguishing agents. The performance can probably be improved with a slightly lower viscosity.
The results are indicative
The results of the study are indicative to the extent that experiments were conducted with only one (1) li-ion cell. The battery module was designed to simulate the problems of distributing and reaching the cell on fire with extinguishing agents, but experiments with an entire battery module would need to be carried out to properly reflect real installations. A fire detection system that is able to identify the battery module on fire is also required in practise in order to apply the extinguishing agent in the module on fire. The project provides a good basis for the development of standardised fire testing methods and design and installation guidelines for extinguishing systems in battery rooms, and funding for continued efforts in this area is sought.
The project is described in RISE Report 2018:77, "Lion Fire: Extinguishment and mitigation of fires in Li-ion batteries at sea”. The Swedish Transport Agency funded the project.
Petra Andersson, email@example.com
Magnus Arvidson, firstname.lastname@example.org
Franz Evegren, email@example.com
Mourhaf Jandali, firstname.lastname@example.org
Fredrik Larsson, email@example.com
Max Rosengren, firstname.lastname@example.org
RISE Research Institutes of Sweden