Rollover

Rollover is the ignition and burning of unburnt fuel gases in the hot smoke layer that collects under the ceiling of a room on fire. The SFPE Handbook of Fire Protection Engineering describes the same event as layer burning. This is the ignition of fuel-rich upper layer gases at the interface between the hot upper layer and the cooler lower layer. It can occur only when the fire is short of air, so that more fuel gas is produced than the incoming air can burn.

The unburnt pyrolysis gases in smoke can ignite when they mix with oxygen and are hot enough or meet an ignition source. When they do, flames spread through the hot gas layer, and radiation from flames under the ceiling heats everything below them. Rollover is therefore closely linked to flashover, yet the two are not the same event. The handbook chapter on compartment temperatures notes that the ignition of unburnt fuel in the hot fire gases, flames from openings and the ignition of all contents of a room may be different phenomena. Rollover also differs from a backdraft and a smoke explosion, which involve a sudden supply of air or a premixed gas cloud.

How unburnt gases collect under the ceiling

A fire in a room soon builds up two layers. The upper layer consists of hot combustion products that collect below the ceiling, and the lower layer consists mainly of ambient air that is drawn into the base of the fire. At first the fire burns with plenty of air, much like a fire in the open, and combustion is nearly complete. As the fire grows, the openings of the room limit the inflow of air. The fire then burns underventilated, or fuel-rich, and the fire plume usually reaches up into the upper layer.

In this state more fuel is pyrolysed than the available oxygen can burn. The surplus fuel gases, together with carbon monoxide and other products of incomplete combustion, are carried into the smoke layer. The layer then holds a mixture of fuel and inert gases at high temperature but with little oxygen. Such a mixture is a combustible gas that needs only oxygen and ignition energy to burn.

Oxygen reaches this mixture at the lower boundary of the layer, where the hot gases meet the fresh air of the lower layer. Layer burning starts there, at the interface between the two layers. The same fuel-rich gases can also flow out of the room below the top of a door. They then mix with air outside and burn as a separate flame zone, which is called external burning. According to the handbook, both forms of burning arise from a buildup of enough fuel in an atmosphere that can still mix with oxygen, and both occur only during underventilated burning.

Layer burning (rollover) in an underventilated room fire: fuel-rich gases in the hot layer burn where they meet fresh air, at the layer interface and outside the doorway. Based on Gottuk and Lattimer (2002), p. 2-55.

Conditions for ignition of the hot layer

Whether the gases in the upper layer can ignite depends on how much unburnt fuel they contain and on their temperature. A detailed method checks the composition of both layers against the flammability limits of each fuel gas. In a worked example in the handbook, a layer at 700 K with 10 per cent hydrocarbons, 2 per cent carbon monoxide and 1 per cent hydrogen was mixed with air at 300 K. The calculation gave an index of 1.07, above the limit of 1, so the hot layer was predicted to ignite and burn. The handbook reports that this approach has been used successfully to predict the flammability of the hot gas layer in enclosure fires.

A much simpler method uses the equivalence ratio. This number compares the rate at which fuel is supplied with the rate at which air is supplied, relative to the ratio needed for complete combustion. A value below 1 means that air is in excess and the products contain spare oxygen. A value above 1 means that fuel is in excess and unburnt fuel remains in the upper layer. For a single room, the ratio is calculated from the fuel mass loss rate and the air flow into the room:

$$\varphi =\frac{{\dot{m}}_f/{\dot{m}}_a}{r}$$

Symbol	Description	Unit
$\varphi$	equivalence ratio	dimensionless
${\dot{m}}_f$	mass loss rate (burning rate) of the fuel	kg/s
${\dot{m}}_a$	mass flow rate of air into the room or fire plume	kg/s
$r$	stoichiometric fuel-to-air mass ratio of the fuel	dimensionless

The ratio is 1 when fuel and air are supplied in exactly the proportion needed for complete combustion. Some tables give the air-to-fuel ratio instead of the fuel-to-air ratio, so the definition of the tabulated value must be checked before use. Where layer burning occurs, or where air enters the upper layer directly through several vents, the ratio should be calculated for the whole compartment rather than for the plume alone.

The ratio at which layer burning begins falls as the upper layer becomes hotter. With typical property values and a stoichiometric fuel-to-air ratio of 0.07, the predicted ignition ratio drops from about 2.1 for a layer at 300 K to about 1.3 at 1300 K. Experiments by Beyler found ignition at an equivalence ratio of 1.7 for layer temperatures of 500 to 600 K. In experiments by Gottuk with layers at 900 to 1100 K, flashes of burning first appeared at a ratio of 1.4 ± 0.4. Sustained burning began at 1.9 ± 0.3, and burning at the layer interface was generally seen shortly after the first flashes.

Equivalence ratio at which the upper layer ignites, as a function of the upper layer temperature, with measured onset values. Based on Beyler (2002b), p. 2-186.

As a worked example, the handbook considers a piece of cushioned furniture, mainly polyurethane foam, burning in a room with one open doorway. A two-layer system develops in the room. The foam burns at 37 g/s, the doorway admits 217 g/s of air, the upper layer is at 700 °C, and the stoichiometric fuel-to-air ratio of the foam is 0.1139.

$$\varphi =\frac{37/217}{0.1139}=1.5$$

The fire is fuel-rich. The incoming air can burn only about 24.7 g/s of foam, so about a third of the fuel leaves the flame unburnt and enters the layer. The layer temperature of 700 °C, or 973 K, lies within the range of Gottuk's experiments, in which the first flashes appeared at 1.4 ± 0.4. The predicted ignition curve gives a ratio of about 1.5 to 1.6 at this temperature. The example is therefore close to the threshold at which burning of the layer begins.

Oxygen also sets a lower limit. Flames that are fully immersed in a vitiated hot layer can be extinguished by the lack of oxygen. Measurements found extinction at oxygen concentrations of 12.4 to 14.3 per cent, with the lower value for a 50 cm pool burner and the higher value for a 9 cm burner. In tests with diesel pans of 0.62 m and 0.84 m, oxygen concentrations below 14 per cent could not be reached during pool burning in a compartment. The layer must therefore meet fresh air before it can burn, which is why burning starts at the interface or at the openings.

Flames under ceilings and in corridors

Flames in the upper part of a room do not always come from the smoke layer. When a fire is tall enough to reach the ceiling, its unburnt gases spread out radially under the ceiling, mix with air and form a ring of flame there. Experiments by Heskestad and Hamada showed that the flame then extends along the ceiling by about the difference between its free flame height and the ceiling height. The total flame length from the fuel to the flame tip stays about the same as the free flame height. Smaller-scale tests by Yu and Faeth found about half this radial extent.

This ceiling flame and rollover share the same effect on the room below. The downward radiation from flames under the ceiling and from the hot ceiling itself preheats and ignites items that are not yet burning, and it strongly affects the rate of fire spread. Unburnt gases in smoke can likewise spread the fire through their own radiation once they burn.

Burning of fuel-rich gases can also continue outside the room of origin. In tests with a fire room connected to a hallway, the gases entered the hallway as a ceiling jet and burned along the hallway ceiling. In post-flashover hexane fires with an equivalence ratio of about 2.0 and a window of 0.50 m by 0.24 m, this flame reached on average 2.7 m along the hallway. For an opening to open surroundings, external burning was first noted at a compartment equivalence ratio of 1.6.

Fuel-rich gases from a fire room entering a hallway as a ceiling jet and burning along the hallway ceiling. Based on Gottuk and Lattimer (2002), pp. 2-72–2-73.

Temperature controls these reactions. Hydrocarbons begin to react efficiently with oxygen above about 700 K, while carbon monoxide does not readily burn to carbon dioxide below about 800 K. Efficient burning outside the fire room needs gas temperatures close to those at the onset of flashover, 775 to 875 K, and the burning itself produces local temperatures above 1300 K.

Relation to flashover

Flashover is generally defined as the transition from a growing fire to a fully developed fire in which all combustible items in the room are burning. Most definitions rely on the gas temperature at which radiation from the hot layer ignites the contents. Temperatures of 300 to 650 °C have been linked with its onset, and 500 to 600 °C are the most widely used. A layer above 500 °C radiates about 20 kW/m² to the floor, which makes many materials pyrolyse rapidly. The International Fire Engineering Guidelines assume flashover when sustained flames reach the ceiling and the hot layer reaches 600 °C, although it may already occur at 500 °C.

Rollover concerns the gases, whereas flashover concerns the whole room. The handbook stresses that ignition of unburnt fuel in the hot gases, flames emerging from openings and ignition of all contents may be different phenomena, although all three appear in definitions of flashover. Firefighters generally recognise flashover by flames coming out through the open doorway of the fire room. In compartment tests, this emergence of flames generally corresponds to a ceiling layer temperature of 500 to 600 °C.

The difference matters for how results are read. A burning layer adds heat close to the ceiling and so raises the radiation that drives flashover. On the other hand, a fuel-rich layer forms only once the fire is already short of air, which is typical of a fully developed fire. Before flashover a room fire is usually local and fuel-controlled, and after flashover it is usually ventilation-controlled.

Backdraft and smoke explosion are related but distinct events. In a backdraft, gases that built up in a closed room burn explosively when fresh air suddenly enters, for example when a door is opened. A smoke explosion needs smoke and air that are already reasonably premixed within the explosive limits, ignited by autoignition or by a spark or glowing particle. Rollover, as described above, is a burning of the layer at its boundary with air while the fire continues to burn under limited ventilation.

Four related events in a room fire: rollover (layer burning), flashover, backdraft and smoke explosion. Based on Gottuk and Lattimer (2002), p. 2-55; Walton and Thomas (2002), p. 3-171; Van Rossum (2027), §4.9 and §4.10.

Significance for firefighting

Conditions in and under the smoke layer matter for crews inside a building as well as for escaping occupants, because radiation from the smoke can be dangerous to people. For firefighting it matters whether a fire is still local or already involves the whole compartment, and whether crews can approach with the fire in view. A favourable result of a fire engineering analysis does not mean that the fire service will carry out an interior attack. That decision is always made by the crews on site.

Some design methods state explicit limits for the conditions under which crews can operate. In Australia, the Fire Safety Verification Method handbook (2020) lists firefighter criteria at 1.5 m above the floor, to be agreed with the fire brigade. Routine work is limited to 25 minutes at 100 °C and 1 kW/m², and hazardous conditions to 10 minutes at 120 °C and 3 kW/m². Extreme conditions, such as a snatch rescue or a retreat from a flashover, allow one minute at 160 °C in the lower layer, 280 °C in the upper layer and 4 to 4.5 kW/m². These criteria assume full protective clothing and breathing apparatus.

In the Netherlands, the standard NEN 6093 for smoke and heat exhaust systems gives limits for a fire service attack in a space with such a system. They include a radiation of at most 3.5 kW/m² from the smoke layer at 1.8 m above the floor. The calculated smoke layer temperature may not exceed 200 °C where goods are in the layer and 500 °C in all other cases. A burning layer, with local temperatures above 1300 K, lies far outside all of these limits.

Modelling and limitations

Two-zone fire models assume that the temperature and composition of each layer are uniform. The handbook chapter on compartment fire modelling states that the question of what makes a gas layer flammable, and how combustion proceeds in a vitiated layer, had not been resolved for zone models. Results of zone models must therefore be interpreted with care once a fire becomes ventilation-limited.

The zone model CFAST handles the surplus fuel in a simplified way. When oxygen limits the fire, only part of the pyrolysed fuel burns, and the unburnt hydrocarbons are transported to the upper layer. If enough oxygen later becomes available in the upper layer, or the gases spill into another room, the model burns them as a substitute fuel. By default, burning stops below an oxygen mass fraction of 0.15.

The ignition criteria themselves have limits. The detailed method requires a very detailed description of the upper and lower layers. The simple equivalence ratio method rests on typical property values. The experiments behind it found the onset of burning over a band of ratios rather than at a sharp value. In Gottuk's tests, burning at the interface was also hard to see because soot covered the viewing ports. Predictions of rollover should therefore be read as an indication of the conditions under which it becomes likely, not as an exact moment.

See also

• Ventilation-controlled fire
• External flaming
• Flashover criteria
• Fire development stages
• Post-flashover fire

References

• Alpert, R. L. (2002). Ceiling jet flows. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 2-18–2-31). National Fire Protection Association; Society of Fire Protection Engineers.
• Australian Building Codes Board. (2020b). Handbook: Fire Safety Verification Method (Print version 1.1).
• Beyler, C. L. (2002b). Flammability limits of premixed and diffusion flames. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 2-172–2-187). National Fire Protection Association; Society of Fire Protection Engineers.
• Gottuk, D. T., & Lattimer, B. Y. (2002). Effect of combustion conditions on species production. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 2-54–2-82). National Fire Protection Association; Society of Fire Protection Engineers.
• Iqbal, N., & Salley, M. H. (2004). Fire Dynamics Tools (FDTs): Quantitative fire hazard analysis methods for the U.S. Nuclear Regulatory Commission Fire Protection Inspection Program (NUREG-1805). U.S. Nuclear Regulatory Commission, Office of Nuclear Reactor Regulation.
• National Research Council of Canada, International Code Council, Department of Building and Housing, & Australian Building Codes Board. (2005). International fire engineering guidelines (2005 ed.). Australian Building Codes Board.
• Peacock, R. D., McGrattan, K. B., Forney, G. P., & Reneke, P. A. (2026). CFAST – Consolidated Fire and Smoke Transport: Vol. 1. Technical reference guide (Version 7, NIST Technical Note 1889v1). National Institute of Standards and Technology. https://doi.org/10.6028/NIST.TN.1889v1
• Quintiere, J. G. (2002a). Compartment fire modeling. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-162–3-170). National Fire Protection Association; Society of Fire Protection Engineers.
• Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.
• Walton, W. D., & Thomas, P. H. (2002). Estimating temperatures in compartment fires. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 3-171–3-188). National Fire Protection Association; Society of Fire Protection Engineers.