In collaboration with the Norwegian, Danish and US maritime authorities, battery manufacturers, system integrators, suppliers of fire extinguishing systems, shipyards and shipowners, DNV GL published a new report on battery safety in ships. The report analyzes explosion and fire risks in maritime battery installations and the effectiveness of fire extinguishing systems in case of a battery fire.
Two areas were prioritized to provide information.
The first key focus was quantifying off-gas content and explosion risks. Testing was performed at both the cell and multi-cell level, for different chemistries and form factors, and under different failure modes. These results provide reference and guidance on the amount of ventilation and the effectiveness.
According to the report, in general, the magnitude of possible consequence depends heavily on the number and size of the battery cells expected to be involved in an incident.
The second primary objective was evaluation of the capabilities of various fire suppression and extinguishing media with respect to lithium-ion battery fires. Each of the systems available has different strengths and weaknesses, and thus different systems may be more effective or necessary depending on the key risks posed by a particular battery arrangement or installation.
In general, fire suppression is more effective when detected and deployed early and if it can be released into the module. Key factors to evaluate as far as requirements are short term cooling, long term cooling, and gas absorption.
Results
Fire suppression systems
Tested fire suppression systems provide different benefits, with unique strengths and drawbacks, providing no ‘silver bullet’ solution.
Direct injection of foam shows the best heat mitigating performance compared with all tested methods. This method had the highest potential for module-to-module fire mitigation, especially when designed for sufficient capacity to flood the modules/racks over longer time periods.
High pressure water mist protection provides good heat mitigation at module level in addition to providing full battery space protection from external fires. It also has good gas absorption and gas temperature reduction capabilities.
NOVEC extinguish the battery fire flames, but performs poorer regards to heat mitigation, gas temperature reduction and gas absorption compared to water mist. Room ventilation needs to be closed for this suppression method to be functional. This can increase the toxic and explosive battery gas concentration in the room until ventilation can start again.
Sprinklers do not extinguish the visible flames, but record similar heat mitigation capabilities at module level as high-pressure water mist. Since water can displace the gas into pockets with high concentrations, the explosion risk is considered to become more severe with sprinklers.
Each battery installation will have to assess necessary barriers in consultation with the battery manufacturer to identify the application most suited for that project. Because of limited amount of available suppression media on board a vessel, the actual volumes and release rates needs to be calculated and depend on the battery system.
Heat and gas generation
The cell level and module level tests provided evidence that visual combustion produced more heat, but less gas compared to tests without visual combustion. Tradeoffs in the risk evaluation needs to be done between extensive heat generation in comparison to extensive explosive and toxic gas generation.
[smlsubform prepend=”GET THE SAFETY4SEA IN YOUR INBOX!” showname=false emailtxt=”” emailholder=”Enter your email address” showsubmit=true submittxt=”Submit” jsthanks=false thankyou=”Thank you for subscribing to our mailing list”]
The NMC cell which released significantly more volume was the one test that did not induce visible combustion external to the cell. It seems that the gas production is halved when there is visible combustion. However, additional tests are needed to quantify the exact number.
The amount of oxygen released is not enough to affect combustibility external to the cell. It is considered more likely that O2 is released internal to the cell and play a very central role in the onset of thermal runaway. This will also lead to more aggressive heat development and increased CO or CO2 production.
It is also noticed that limiting the oxygen supply will suppress the battery fire, but not be sufficient to cool down the battery. In these cases, the off-gassing is increased compared to fires where oxygen is fueled to the fire.
Toxicity
If the room is to be entered after an incident, all the identified toxic gases needs to be considered. The gases identified in this project are carbon monoxide, nitrogen dioxide, hydrogen chloride, hydrogen fluoride, hydrogen cyanide, benzene, toluene.
Very small gas concentrations will make the atmosphere toxic, and the gas will dilute fast. As a result, the sensor detecting the toxic gases can be placed in the normal breathing zone for people, 1-1.8 meters from the floor.
Personal Protection Equipment should be used when re-entering the battery space after a battery fire, also after deployment of fire suppression material.
When weighting the Immediately Dangerous to Life or Health (IDLH) values with the released gas amounts, CO, NO2 and HCL will first reach its IDLH values.
Off-gas detection
CO is the main component present for the longest period of time and is considered especially important for early stage detection.
Off-gas in the early stages of thermal runaway events will be colder than off-gas release in the later stages. The early off-gas can therefore become heavier than the air, collecting at floor level. Thus, the report notes that it should be considered if gas-detection related to room explosion risks should be applied at both levels, close to the floor and close to the ceiling.
Tests performed in this project show that relying only on Lower Explosion Limit sensor(s) and cell voltage levels to detect early stages of a thermal runway event is not enough.
Both the Li-ion Tamer sensor and smoke detector, when placed close to or inside the affected module, proved the most reliable means of pre-thermal runaway warning. The early detection of thermal runaway has also proven that a cell can be disconnected, effectively stopping the overheating process.
Ventilation
In order to achieve the most potential of a forced extraction duct, a high extraction point in the room has proven to be the key factor. This makes sure that the necessary air changes per hour stays low while still providing the necessary dilution of explosive gases in the space.
The explosion pressure limit is set to 0.5 barg. Above this pressure the bulkheads will be damaged. With a ventilation rate of 6 ACH it should be sufficient to avoid such pressure if 350 liters of battery gas released in the room is considered as a worst case. This corresponds to a cell or module of 115-175Ah. If cells of total 250Ah is failing, this requires 10 ACH, while failing 500Ah requires 22 ACH in a room of 25m3 of free space. The ventilation can be turned on demand based at early off-gas detection with sensors close to or inside the modules.
If batteries of 4000 Ah is failing, it will not be sufficient with 100 ACH to avoid an explosion magnitude of 0.5 barg.
A ventilation formula for a battery room is proposed. The formula calculates the air changes per hour (ACH) with the size of the failed batteries, the design bulkhead pressure, the room volume and the vent distance from the ceiling as input variables.
Temperature class and gas group
Based at the tests performed, the temperature class for battery off-gas explosion proof equipment is recommended to be T2 according to the IEC 60079 standard.
The gas group is identified as Group IIC according to the IEC 60079-20-1 standard.
Thermal runaway identification
Based at the tests performed, important difference was reported between the Nickel Manganese Cadmium (NMC) and Lithium Iron Phosphate (LFP) cells. The LFP cylindrical cells were much harder to force into thermal runaway compared to the NMC pouch cells.
For the NMC pouch cells, a temperature increase rate more than 10 °C/sec together with a max temperature above 450°C seems to be enough to identify the onset point for a thermal runaway with visual combustion.
For the LFP cells, a temperature increase of 4 °C/sec seems to be sufficient to identify the onset point for the thermal runaway. The chance of achieving this increase with increased state of charge, and it might be necessary to charge the LFP battery cells beyond 100% SOC to provoke visual combustion.
Quantitative Risk Assessment
Fire propagation protection and the Current Interruptive Device are two of the most significant safeguards to be installed in the battery system.
When comparing the battery fire risks with data registered in the HIS Fairplay database for fires in a diesel engine room, the report notes that the possibility of a battery fire is lower in comparison to a diesel fire.
Nevertheless, engine room fires registered in the HIS Fairplay database include fires of many different magnitudes not necessarily correlating to the fire scenario established in battery system QRA. This means that better data would be necessary to fully evaluate if a battery system is safer than a conventional combustion engine.
Battery system design
The required ventilation rate and the amount of fire suppression material depends on the number and the size of the battery cells involved in the fire. If the complete battery system catches fire, the suppression and ventilation will not be able to address the fire and explosion risks.
It is of most importance to design a battery system with fire propagation protection and Current Interruptive Devices to limit the fire to one part of the battery system, and to install a well-tested Battery Management System capable of preventing several modules being overcharged at the same time
the report concludes.
See more information in the following report.
