linings and flame propagation observations listed. describes the materials used.

Table 2

Draft conditions in these tests were either by natural convection or a forced draft of approximately 100 fpm. The forced draft used is within the range of the air flow

encountered in corridors serving as return air plenums in enclosed buildings.

While a variety of parameters have been tested in the program, only the experiments which have the necessary and sufficient data for analysis under the concepts of energy and radiation models are presented for discussion in this report. All of the experiments selected for these discussions will have the following standard conditions:

ignition

by four wood cribs; walls and ceiling of gypsum board; no forced draft; and floor covering as the test material in the corridor. Brick, a number of carpets, vinyl tile and

red oak have been used as floor coverings in experiments to date. All of the carpets passed the pill test. A rubberized hair felt pad used in some of the experiments did not pass the pill test. It is important to note here that the

current federal standard does not require pads to pass the pill test.

The sequence of events in an experiment involving floor coverings usually follows the same pattern. After ignition, cribs would be burning vigorously at about 3 minutes sending hot gases and light smoke into the corridor along the ceiling. At about 4 to 5 minutes floor covering in the burn room should be ignited. Smoke and gases in the corridor would then greatly increase in intensity. If there were no floor coverings in the burn room, as in the case of reference run Test 339 with brick floor, the crib burning rate would reach a maximum at 7 minutes and stay at approximately that level until 14 minutes at which time the crib fire would begin diminishing (figure 4). During the period of maximum crib burning rate, the fire plume would extend from the burn room and lick the corridor ceiling near the burn room doorway while light smoke continued to pour out from the corridor. The air temperature at the burn room doorway for the reference run are within about 200 °C of the ASTM E-119 time-temperature history (figure 5).

When there are carpets in the burn room and the corridor, the wood crib burning rate invariably has a higher initial rate of climb and the burn room top doorway air temperature generally climbs slightly faster than the ASTM E-119 curve (figures 6, 7, 8, 9, and 10). The air supply to the burn room is through the small vents in the burn room, and through the burn room doorway. An interesting phenomenon

associated with carpet in the burn room is that the wood crib burning rate invariably takes a sudden drop (figure 4) indicating oxygen depletion in the burn room prior to flame spread in the corridor. This is believed to be due to the large volume of pyrolysis gases given off by the carpets, resulting in a fuel rich atmosphere.

Table 3 compares the corridor gravimetric smoke data for a number of tests. Note the heavy smoke concentration for Tests 340, 341, and 342 and the drop in the respective burning rates in figure 4. Also note that Test 344 has low smoke concentration. In fact, the level of concentration is of an order of magnitude lower than the other carpet tests, and very comparable to the reference test. If we go back to the burning rate plots (figure 4), we find that Test 344 and the reference test are the only two tests where the crib burning was not affected by the choking or oxygen depletion phenomenon.

Shortly after the carpet ignited in the burn room a flame was observed to spread slowly over the carpet surface from the burn room door and then to accelerate. Then the whole corridor would "flame over" with flame following the entire surface of the carpet. The term "flame over" has been used to describe the observed rapid flame spread, since there is a preferred direction in the corridor fire to distinguish

it from the more general flashover phenomenon.

Table 4

lists the flame over initiation time for various tests.

Note with the exception of two tests, flame over initiation times cluster around 4-1/2 to 7 minutes. Figure 11 illustrates the flame over phenomenon by plotting the carpet surface flame spread distances against time. After a relatively slow flame initiation period, during which time the first 5 feet of the carpets became involved, the carpet flame would sweep to the end of the corridor. The slow down of the flame spread over the last few feet can be explained as an end effect. In Test 342 the smoke concentration is so high, choking took place inside the corridor during flame over as shown by figure 11.

The flame over phenomenon was observed with all carpets studied, including a wool carpet, an aromatic polyamide, and other synthetics. However, quantitatively each carpet fire burned differently when considering energy and flame initiation rate. These and other findings will be discussed in more detail in the following sections.

Examination of the carpet and combustible pad after tests indicates that the rapid flame spread is strictly a surface phenomenon involving only the top pile of the carpet. However, if the fire is allowed to continue to burn then the

pad and the subfloor become involved, contributing

additional smoke and fire.

Appendix A contains descriptive observations of significant events that took place in the corridor for Test 330 to Test 345. Test 331 and Test 332 were omitted because of high crib moisture content of 23% and 15% respectively.

5. Energy Balance Calculations

Due to the variation of the crib burning rates, caused by variations of the available oxygen, and due to the narrow spread of flame-over initiation times among some carpets, it is felt that carpet hazard evaluation by flame spread observation alone is not adequate.

A better definition of the hazard could be the ease with which the carpet will ignite and start to propagate a fire; and this brings us to the critical energy input concept. Before proceeding with the critical energy concept the energy balance calculation for the corridor system must be derived.

5.1 Energy Balance Equations

To perform the energy balance calculations it is necessary