Fire Safety Research Institute Finds Water Best, but EV Battery Reaction Continues

Water remains firefighters’ best tested tool for controlling an electric-vehicle fire, but it cannot necessarily stop the chemical reaction inside a burning battery. That is the central tension in a new 150-page report from UL Research Institutes’ Fire Safety Research Institute, based on 18 full-scale experimental vehicle burns.

Nine of the burns received different suppression treatments. Conventional water provided the strongest flame control and battery-enclosure cooling among the tested options, outperforming vehicle fire blankets and chemical additives mixed with water. Yet even thousands of gallons may manage only the visible and external parts of the event while thermal runaway continues within lithium-ion cells.

Flame control is not the same as stopping thermal runaway

That distinction matters because an EV battery fire is not simply a larger version of a burning passenger compartment. Thermal runaway is a self-propagating chemical process in which a failing cell generates heat that can drive additional cells into failure. Overheating, punctures and overcharging are among the conditions identified as possible initiators.

Once propagation is underway, the battery pack’s structure creates a difficult heat-transfer problem. Cells are enclosed and protected beneath the vehicle, while hose water is generally applied from outside. Water can cool accessible surfaces, suppress flames and protect surrounding objects, but external cooling does not guarantee that it will reach the cells sustaining the reaction.

The researchers therefore warn that “applying copious amounts of water to the battery enclosure is wasteful and will not stop thermal runaway.” This is not a finding that water is ineffective. It defines what water can accomplish: control the fire and reduce external heat without serving as a reliable off switch for the internal battery process.

Blankets and additives did not solve the access problem

Fire blankets can restrict flames and smoke around a vehicle, but the experiments did not establish them as a better replacement for water when a battery was involved. Related blanket testing also identified a potential explosion hazard during incidents with active battery involvement, particularly significant for enclosed or confined environments.

Chemical additives likewise failed to outperform ordinary water in the tested conditions. That result has an important equipment and logistics implication: specialized agents do not automatically overcome the battery pack’s physical barriers or the self-heating chemistry inside its cells. The institute’s broader response guidance continues to identify water from a standard handline as the first-choice suppression method, without requiring a specialized additive.

Water also carries a less intuitive tradeoff. The report cautions that introducing it into an uninvolved section of a battery may increase the risk of ignition or reignition. Responders must consequently balance immediate flame control, cooling and exposure protection against the battery’s uncertain internal condition and its remaining stored energy.

Long incidents remain a systems and infrastructure problem

The findings help explain why an EV fire can demand prolonged monitoring even after flames have been reduced. A damaged pack may retain energy, and thermal activity can persist or resume. Incident management therefore extends beyond initial suppression to include assessing continued thermal runaway and possible reignition before the vehicle is released or moved.

A 2024 Tesla Semi fire illustrates the potential upper end of that burden: it closed a freeway for 15 hours and involved 50,000 gallons of water. One incident does not establish a typical water requirement, particularly across vehicles with different battery sizes, pack designs and fire conditions. It does show how suppression limits can translate into road closures, water demand and a lengthy emergency response.

The comparison with combustion vehicles also requires care. EVs reportedly catch fire less often than gasoline-powered vehicles, but a lithium-ion battery fire can be more difficult to extinguish once it occurs. Frequency and suppression difficulty are separate safety questions; evidence of a challenging response does not establish that EVs burn more often.

For vehicle engineers, fire departments and standards developers, the unresolved challenge is access to the reaction rather than a simple shortage of suppression capacity. Water still performed better than the tested blankets and additives, but the experiments show why more water does not automatically terminate thermal runaway. The institute plans to use the findings in a Fire Safety Academy training course, while the experimental dataset is intended to support future fire-protection research, standards and regulation.

By Thomas Caldwell — AMI’s senior editor for mechanical and mobility engineering, covering vehicle electronics, systems integration, electrification, chassis systems, propulsion, and safety policy.