Thomas Flashover Correlation

Flashover is a transition, not a number. Nothing on this page changes that, and the correlation should be read with the sentence in front of it.

What happens physically is well understood. As a compartment fire grows, the upper layer thickens and heats, and it radiates downward. Somewhere around 500 to 600 °C in the layer, corresponding to roughly 20 kW/m² incident on the floor, every exposed combustible surface in the room reaches its ignition temperature at about the same time. The room goes from a fire in it to a room on fire, over seconds. Nobody in it survives, and the fire's behaviour changes completely: it becomes ventilation-limited almost immediately, and flames appear at the openings.

Philip Thomas's 1981 correlation estimates the heat release rate at which that transition becomes LIKELY in a given enclosure. The two terms are the two ways heat leaves: $7.8A_T$ is what the boundaries absorb, $378A_0\sqrt{H_0}$ is what the opening carries out. When the fire exceeds their sum, the layer can reach the flashover condition. The constants 7.8 and 378 are dimensional — square metres, metres, kilowatts — and are the correlation. Nothing about them survives a change of units.

Reading the two terms separately is worth doing every time. A room where the ventilation term dominates is one where the opening governs, and where a window failing changes the answer abruptly. A room where the boundary term dominates is governed by its size and its linings, and there the material of the walls and ceiling is the lever. Fire investigators read glass breakage in exactly this way, because a closed room can sit well below the threshold with the fire it has, and an opening appearing in the heat can put it over the line within seconds.

What the correlation does not do. It does not predict that flashover will occur. It does not say when. It does not certify that a fire below the figure is safe. Real compartments make the transition across a band, and where in that band a particular room goes depends on the fuel arrangement, the lining materials, where the fire started and what breaks. Different authors put the same enclosure meaningfully far apart — Babrauskas's criterion and Thomas's can differ by tens of per cent on the same room, and neither is wrong, because they are fits to different data with different definitions of what counts as flashover.

One further check belongs with every answer. Compare it against what the opening can supply: roughly $3000A_0\sqrt{H_0}$ kilowatts of burning inside the room. If the flashover figure is HIGHER than that, the compartment cannot reach flashover on the air it admits. That is not good news. What such a room does instead is burn under-ventilated, accumulate unburned fuel in the layer, and flash violently when someone opens a door — the condition firefighters call a backdraught, and it kills firefighters specifically.

Small rooms flash over more easily than large ones, because there is less lining to absorb heat. That is a real and uncomfortable consequence of compartmentation, which trades a lower flashover threshold in each room for containment of the fire to that room. The trade is normally the right one. It is still a trade.