Underground firefighting: How EV fires push responders to the limit

February 5, 2025

Brent Brooks offers insights into the difficulties of responding to EV fires in confined underground environments

We have ignition… A panicked occupant of a high-rise building calls 911 while fleeing an underground parking lot. They reported a car fire on the P3 level, describing extensive smoke and requesting urgent assistance.

Simultaneously, the alarm monitoring company notified that an alarm had been activated at the same address. A security guard from the building also called 911 to report the car fire underground.

Alarm bells ring, alerting the building occupants to the emergency. The security desk received numerous phone calls from concerned residents seeking more information and guidance.

Some decided to evacuate, others preferred to shelter in place, and some wanted to investigate the situation firsthand. The building has awakened, bustling with activity.

Emergency services on the clock

Emergency services have been dispatched and are en route. Two minutes have passed since the first report of the fire.

Firefighters arrived on the scene in four minutes, investigated the building’s alarm panel, and received keys and information from the building representatives and the original caller.

Six minutes have passed. Firefighters have determined the location of the fire by zone. They proceed to the area via the stairs or ramp, making one final radio transmission before they lose communication below grade.

Firefighters must locate the fire while wearing full protective gear and carrying heavy equipment in an unfamiliar environment.

 Their working time downrange is limited by their self-contained breathing apparatus (SCBA). To find the fire, they use thermal imaging and listen for activated sprinkler heads.

On average, it takes six minutes to locate the fire after arrival. The fire has a head start and now has a twelve-minute burn time. Crews need to locate a water source and begin applying water immediately.

Their working time is diminishing because their SCBA air is depleting. The fire is rapidly growing. Neighboring vehicles and the concrete ceiling are at risk. The window of opportunity closes quickly.

Firefighters are preveting the chain reaction from occurring. If this happens, recovery can become mathematically impossible.

Water application usually takes about six minutes, including troubleshooting and stretching hose lines from a water source. We are now eighteen minutes from the first 911 call, twelve of which were after arrival.

An internal combustion engine (ICE) is typically extinguished with aggressive water placement, preventing the ripple effect. Firefighters cool the ceiling and other exposures. Additional crews relieve the first arriving crews.

The second wave of resources must conduct overhaul operations and confirm the complete extinguishment.

 

The fight against smoke and time

Smoke-fighting operations are focused on clearing smoke and searching the underground for occupants. From experience, searching underground in zero visibility is almost impossible to do rapidly and efficiently.

At the same time, smoke is continuously produced—the continuation of smoke production in a large, un-compartmentalized, mutiny-level environment expands the search area beyond most fire department capabilities.

If we are lucky enough to find a victim, the speed at which we can get them to a fresh air environment is nearly impossible for their survival.

We cannot outrun smoke. However, we can stop its production and get it to lift, providing a survivable space for occupants. Ideally, if staffing permits, we would do this simultaneously with a fire attack.

Again, very few fire departments can handle large-area searches; radio communications underground are poor at best. Firefighters are continuously fighting against the clock.

Dispatch time, response time, locating the fire time, deployment time, and application of agent time—all while the fire grows and smoke expands its reach throughout the structure.

Electric vehicle fires have become a significant global issue, and new information emerges daily. Multiple theories and hypotheses are proposed, necessitating comprehensive research and fact-checking.

Firefighters currently combat these fires without knowledge of the underlying issues. This often leads to testing unverified tactics in hazardous situations, resulting in unforeseen outcomes. Firefighters actively seek solutions to these challenges.

Change is achieved by reflecting on our firefighting experiences and examining our actions constructively. Firefighters question why a fire was stubborn to extinguish and seek safer, more efficient ways to handle similar situations in the future.

Firefighters strive to comprehend the underlying science of their work. This evaluation process is crucial for driving progress and improvement in fire services. Remember, first responders do this for them.

Lessons from the past, solutions for the future

A significant example of this review process occurred after the 12-alarm Meridian Plaza fire in Philadelphia (1991), which claimed the lives of three firefighters: David Holcombe, Phyllis McAllister, and James Chappell.

Firefighters nationwide were unexpectedly introduced to pressure-restricting and reducing devices (PRVs, PRDs), as well as flow rates from 38 mm (1.5”) hoses and 700 kPa (100 psi) fog nozzles operated from a standpipe.

Following an investigation, NFPA 14 was revised in 1993. Firefighters often refer to NFPA 14 Buildings as pre-1993 and post-1993.

While designers, engineers, architects, and installers possessed knowledge of the science, this information was not effectively communicated or understood by the firefighters at that time.

Again, we are in a similar predicament involving EV fires. The technology was introduced and used before firefighters knew how to deal with it. Now, we are reacting to the problem and developing tactics.

What we do know about EV fires is that they develop rapidly. In most cases, the firefighters will encounter a well-involved car fire when they arrive and intervene. Firefighters will not be aware if the car is an EV until late into the firefight.

Car fires are stubborn because the car’s exterior shields the fire from our hose streams. The stream placement requires moving around the vehicle completely 360° and targeting the fire from all angles.

The trunk and engine compartment are locked and will continue to burn until firefighters force entry into these locations.

The bumpers, airbags, seatbelt retention system, and hydraulic rods can explode. Magnesium components react violently with water, requiring copious amounts of water to suppress. Now, let’s add the EV component to the mix.

Firefighting crews will respond to the building’s address and investigate the fire alarm panel to determine the location of the fire by zone and level.

Investigating the fire alarm panel prepares firefighters by providing a clearer picture of how the fire and smoke are behaving and how much of the building is already affected.

Firefighters will also gain insight into the building’s overall health and the history of the current fire’s origin.

They do this by reviewing which detectors are activated and in what order they activate. Did the notifications begin as a smoke alarm and then progress to multiple smoke detectors, heat detectors, and eventually to sprinkler activation? How many sprinkler heads are activated?

The sequence gives us an idea of the direction and speed at which the problem escalates, allowing us to deploy resources accordingly.

Decoding the data behind EV fires

Data shows the intensity at which EV fires can burn measured in megawatts, BTUs, heat release rates, and heat flux. The unpredictability of the duration for which EVs will burn is also noted.

Additionally, research shows how much smoke is produced by volume and the toxins within that smoke. Studies inform us how long firefighters’ PPE can protect them in hostile environments, and the statistics indicate the duration of their SCBA usage.

Engineers know the temperature at which concrete spalls and metal melts. With this information, we can predict the amount of water to use and forecast the response and deployment model based on timing. Our group is currently trying to connect these dots. If we can predict it, we can prevent it.

This article is the first of a two-part series. The second part will explore the deployment model and resource management strategies for addressing the challenges of underground firefighting in hostile environments.

This article was originally published in the February 2025 issue of Fire & Safety Journal Americas. To read your FREE digital copy, click here.

Read Next

Subscribe Now

Subscribe