It’s common knowledge that to put out a fire, you take away the heat, the fuel, or the oxygen. A lithium-ion battery fire does not play by those rules. Each cell carries its own fuel, and some chemistries even release their own oxygen when they break down.
This means that battery fires behave differently than ordinary fires. They can reignite after being extinguished, for example. They can begin releasing toxic or flammable gases even before an open flame appears. This can have serious consequences for first responders.
In Arizona, in 2019, several firemen were injured when they opened the door to a battery room that had been smoking for a while. Fresh air met the gases inside and caused an explosion.
To use battery storage solutions safely, we need to know how to deal with fires. How to prevent them, how to contain them, and how to put them out. This was KSP SafeBESS’ mission.
When safety systems fight each other
To find out what actually works in a battery room, research scientists at RISE Fire Research set fire to battery modules at a testing facility in Trondheim. They tested four suppression systems: sprinklers, high-pressure water mist, low-pressure water mist, and the inert gas IG-541. Two further tests let the fire burn and tested venting out the gases.
The results show how one safety system is not enough. The inert gas was the only system that successfully extinguished the flames and stopped the fire from spreading from one module to the next. But it did nothing about the flammable gas, which continued accumulating in the room.
This means that once the gas has been released, someone has to decide how long to keep the room sealed, and how to get the gas out without igniting it. This is not a problem fire crews can guess the answer to when they arrive on site. The hold time and venting strategy have to be worked out in advance.
The water-based systems cooled the modules well, but in the low-pressure mist test, hot spots survived and the fire spread between modules. Ventilation prevented an explosion, but exactly how much airflow is needed is still an open question.
The particularities of battery fires mean that different safety systems can counter each other. Inert gas and forced ventilation, for example, can cancel each other out, since one needs the room sealed and the other needs air moving through it. Flame and heat detection, suppression, gas detection, ventilation, and explosion protection all have to be designed together as a system, and adapted to the specific battery technology that is in the room.
Problems in Norwegian battery rooms
The project reveals that not all battery rooms in Norway were designed with proper fire safety measures. Research scientists at SINTEF Community found several with insufficient risk analyses, missing fire suppression systems and no protection against flammable or toxic gases. Some of these rooms have since been decommissioned.
The SafeBESS risk and vulnerability analyses points to several mitigation measures to reduce the risk to an acceptable level, including placing the battery room in a suitable location, choosing the correct materials, properly monitoring the battery system, having smoke and carbon monoxide detection systems, installing an extinguishing system, and having a strategy for relieving pressure and venting out harmful gases.
Answers on the way
Results were presented at a final conference on 28 October. The event was open and attended by stakeholders from outside the project, including representatives from fire departments, fire safety consultants and government agencies. In addition to sharing results, project leaders had planned time for discussion, and a lively Q&A concluded the conference during the last hour.
The project is in the process of finalising guidelines for the fire safety of battery rooms in buildings. These guidelines will be published soon.
“We want these guidelines to be useful to building owners, consultants and fire services alike, so that the important safety decisions about a battery room are made before it is built; not after something goes wrong,” says project lead Fride Vullum-Bruer.
