Fire Behavior

A study was performed on the effect of the ignition characteristics of secondary fuel packages in a compartment with a burning bunk bed. It was found that fuel packages for which heat release is slower can consume the oxygen in a compartment and lead to an oxygen-starved situation before flashover occurs, while fuel packages which release heat more quickly are more likely to lead to flashover in a room before oxygen depletion smothers the fire.

Basic Setup

The basic setup is the standard compartment fire defined for the ME382R fire science class: a 12' x 12' room with a 10' ceiling contains a bunk bed and two dressers. The data on the free-burning heat release rate for each type of furniture is taken from the NIST "Fire on the Web" internet site, http://fire.nist.gov/fire/fires/fires.html. The composition of the furniture, needed to calculate the heat of combustion and mass release rate for the furniture, was estimated to be 75% cellulosic and 25% hydrocarbon. The walls and ceiling of the room are composed of 20% gypsum board, 0.75 inches thick, and 80% glass fiber insulation, 4 inches thick. There was no sprinkler system or HVAC system operating in the compartment, and there were no open doors or windows. The other parameters used in the FPEtool software package were left at their default values. The bunk bed was set ablaze at time zero. The dressers were 6 and 10 feet from the bed, and 11 feet from each other.

Problem Studied

The problem studied dealt with the successive ignition characteristics of the dressers, using the freeburn module of FPEtool. The default characteristics defined in the problem statement were that the dressers would ignite when their surface temperature reached 390 ºC. The value of krc was set at 0.6 (kW/m2K)2s. In this study, the ignition temperature and time was altered, changing the time history of the heat release rate, which then is used as an input to the fire simulator module. The first finding in studying this problem was that the default dresser ignition characteristics would not lead to ignition of the dressers in a free burning situation. It was necessary to depress either the ignition temperature or the krc value substantially. Specifically, when the ignition temperature was set to 65ºC (150ºF), a very low value, the dressers would ignite at 223 seconds. This particular time is significant, because the fire simulator predicts that flashover will occur in the compartment at 223 seconds as a result of the burning bunk bed alone. This essentially indicates the great ease with which the hot gas layer at the ceiling can radiate to room furnishings, compared to direct radiation from a burning object to another object in the room. The fact that the hot layer is such an effective radiator is not surprising when one considers the vastly greater view factor for radiation from the hot layer to an object on the floor, than for radiation from a single plume to the same object. As a result of the discovery that direct ignition of the dressers by the bunk bed is unlikely, two other cases were examined. One was the effect of burning one of the dressers alone in the compartment. The other was burning the bunk and both dressers together, as might happen in an arson situation.

Results

Figure 1 shows the heat release rates versus time for the various burn situations. Note that the bunk bed displays a build-up of about 190 seconds with very little heat released, followed by a sharp spike of heat release peaking at 237 seconds at over 4 MW, and then a long decay to burnout. The dresser, by comparison, exhibits higher heat release rates until just before 200 seconds, when the bunk heat release rate spikes up. The dresser heat release rate peaks at 423 seconds at a lower 1.8 MW. The simultaneous burn profile is then the sum of the bunk and two dressers, leading to significant heat release before 190 seconds, a strong peak at 237 seconds, and a second strong peak at 423 seconds.

When free radiation from the bunk is used to ignite the dressers, the behavior is somewhat more complex. When the ignition temperature is set to 65 to 100ºC, both of the dressers ignite at 223 seconds, giving a curve which follows the bunk bed profile closely through the first peak, but has a second strong peak at 659 seconds. If the dresser ignition temperature is increased to 150ºC, only the dresser nearer the bunk ignites, but it ignites at about the same time, so that the second peak in the heat release curve is still around 650 seconds, but the peak is only at 2.2 MW. Between 200 and 300ºC, the dresser is on the verge of ignition, which the freeburn module has difficulty processing. At 300 ºC and above, neither dresser ignites, and the heat release curve is simply that of the bunk bed.

As noted previously, the bunk bed alone will produce flashover in the compartment at 223 seconds. Because the dresser will not ignite before that in freeburn, the temperature, layer height, and gas concentrations, shown in figures 2 through 6, produced by the bunk bed heat release profile and the dresser ignition cases are nearly identical. In each case, the temperature reached 600ºC (1112ºF), the flashover condition, at 223 ±1 seconds, a negligible variation. The hot gas layer reaches floor level just before that, at 210 seconds. The oxygen concentration decreases to 16% in the room, the CO reaches 100 ppm, and the CO2 reaches 3%.

While free ignition of the dressers has little effect on the time history of the burn, igniting the dresser at time zero, either alone or with the bunk bed, leads to substantially different behavior from the flashover above. The significant, but not large, heat release slowly heats the room, but also depletes the oxygen in the room. The simultaneous burning case reaches an oxygen concentration below 10%, which is considered low enough to smother the open flames, at 190 seconds, when the room temperature is 1050ºF. This is very near to the flashover temperature, but does not in fact cross that threshold, so that the room is left filled with hot gases. The room's contents include 120 ppm of carbon monoxide and 6.7% carbon dioxide. The hot layer reaches the floor at 105 seconds. The case where one dresser burns alone heats more slowly, reaching a temperature of 1030 ºF at 270 seconds before the oxygen concentration drops below 10%. The gases in the room also contain 6.8% CO2 and 103ppm CO. When only one dresser burns, the hot layer reaches the floor at 150 seconds. In no case when at least one dresser is lit at time zero does flashover occur. The room is left in a dangerous, oxygen depleted state, with a hot atmosphere that may explode if further oxygen is allowed to enter.

Conclusions

The very different time histories of the heat release rates from the bunk bed and dresser examined here lead to substantially different phenomena when they burn in a compartment. A bunk bed releases very little heat at first, then rapidly spikes to a peak. This heat release profile tends to encourage flashover in a room. By contrast, a dresser burns more material early in the burn, but does not peak as high or as fast. That profile does not tend to flash over, but it does tend to consume more oxygen and lead to higher concentrations of CO and CO2 in the compartment.

Igniting the dressers via direct radiation from the burning bunk was found to be very difficult. Even when the ignition temperature of the dressers was dropped to 65ºC, it took as long to ignite the dressers via direct radiation from the bunk as it does for the compartment to flash over as a result of the bunk burning.

Return to top