To identify more precisely the hazard of floor covering materials, attention has been focused on full-scale fire experiments in which floor covering materials (i.e., an assembly placed over a subfloor) have been examined in a room -- corridor environment. In particular, these experiments involve a large fire in a room connected to the corridor by an open doorway. The corridor is lined with negligibly combustible materials along the walls and ceiling but with a combustible flooring assembly. This simulates the least combustible class of assemblies in which corridor fire spreads. Corridors lined with combustible wall and ceiling materials as well as floor covering represent a more severe fire hazard. Several organizations have conducted room/corridor fire experiments to investigate the hazard of combustible lining material in corridors. They include the National Research Council of Canada (NRC), Illinois Institute of Technology Research Institute (IITRI), and the National Bureau of Standards (NBS). Although the room/corridor arrangement has been similar in these experiments, a variety of ignition sources have been used, ventilation conditions have varied, and overall construction materials have been different. Moreover, recorded measurements have ranged from visual observations to extensive data from automatic record instrumentation. At a minimum, these experimental results serve to establish a spectrum of corridor building fires and serve to provide a definition of the hazards.

1.2.1. NBS Corridor Fires

The NBS corridor system is depicted in figure 1. A total of 18 experiments involving floor covering materials have been made. In 14 experiments the principal corridor combustible lining has been the flooring assembly, of which, 11 involved sustained rapid flame propagation to the exhaust window ("flameover"); and the remaining 3 experiments resulted in partial or no sustained flame spread. A summary of results from these experiments is shown in table 5. Recent publications have described this program [6,7,8].

A typical corridor fire test resulting in flameover is composed of 7 distinct stages of fire development. (1) The test commences with the ignition of four 40 pound wood cribs in the burn room. In several minutes the crib fire reaches a state of fairly steady burning. It appears that approximately half of the air required to support crib combustion flows into the burn room through the two floor level vents, while the remaining air supply enters from the corridor. (2) Products of combustion flow from the room along the corridor heating its walls and ceiling. Inlet air flows over the corridor floor which is heated by radiative transfer from upper hot walls and ceiling and by hot smoke products. (3) Once the crib fire builds-up, ignition occurs on à 2 1/2 ft wide floor covering runner which extends from the corridor into the burn room. This ignition results from a high radiant heating exposure over this runner and from flaming crib embers falling to the floor. Once ignition occurs at a location on the runner rapid flame spread follows over the entire runner in the burn

room.

burn room.

(4) This flooring fire then begins to emerge and spreads slowly into the corridor. The flame fans out from the doorway advancing upwind against the incoming air flow and is driven by radiant heating of the floor. (5) As this fire advances it depletes the oxygen of the incoming air and preheats this air. This results in a buoyant force which diverts some air flow from entering the (6) It appears that the occurrence of flameover is preceded by a reduction in air supply from the corridor to the burn room such that the crib fire becomes fuel rich. The crib fires continue to produce pyrolysis products at the same rate but complete combustion is not possible within the burn room. Thus, a cloud of combustion products, probably including aerosols and soot, enter the corridor flowing out over the floor fire which extends not more than about 5 ft down the corridor. This is followed by the onset of flameover or rapid flame spread within the corridor. (7) Flameover occurs as a wave of flames and hot combustion products advance down the corridor producing a large increase in corridor temperatures and heat flux. In the order of 1 minute the flames emerge from the window completing the flameover process. will be presented to further illustrate these processes.

More detailed data

During steady burning rate of the wood cribs a maximum energy release rate of 80,000 Btu/min occurs. Some of this energy is lost to the burn room and the rest enters the corridor. The convective energy flow values were calculated by Fung, et al [8], based on two to three velocity and temperature

measurements in the doorway. These are plotted against the initiation time of flameover, to shown as open symbols in figure 2. (These flameover times do not agree with values given by Fung, et al [8] since they termed flameover the first indication of observed flooring fire in the corridor.) A straight line drawn through these points indicates a nominal convective energy flow rate to the corridor of 60,000 Btu/min and an average induction time of 220 seconds before this rate of energy was released. The time preceding steady crib burning generally occurs after 5% weight loss results (t5g). If the flameover time is adjusted for each test by subtracting t5% from tro, then the solid symbols in figure 2 yield the same burning rate with the intercept through zero. These results indicate the degree of reproducibility for the crib fire source, and the extent to which energy input rate may be regarded as a step function of time.

When the floor covering runner ignites in the burn room it releases an amount of energy which can become significant with respect to the crib fire. This sudden release of energy is indicated by an increase in the gas temperature near the corridor ceiling. Figure 3 shows this effect for a nylon carpet (N-3/U) which had a large energy release rate as compared to a vinyl sheet flooring (R-4) in which there was negligible energy release rate. This pair of tests (348 and 346) tends to bracket the range of conditions which result from exposure to the crib fire source in the burn room. They will be used to illustrate other phenomena which result during the fire development in order to contrast the range of resulting conditions.

During the relatively slow advancement of flooring flame along the corridor, radiant heat transfer to the floor tends to promote continued flame spread. The radiation levels developed depend strongly on the primary crib fire source but are also influenced by the rate of energy release of the flooring material. The exiting hot gases and ceiling temperatures along the corridor determine the extent of radiant heat transfer. Figures 4 and 5 illustrate typical radiation flux distributions along the corridor floor before flameover is initiated. In test 346 the flux was essentially due to the crib energy release; while, in test 348 the radiant flux (at 300 sec) includes the effect of energy release (and most likely smoke) from the sudden ignition of the nylon runner in the burn room.

Before flameover the corridor temperature distribution between the floor and ceiling was greatly stratified with hot products of combustion at the top and cool air flow at the bottom. At the inception of flameover, temperatures at all heights increased rapidly and significantly throughout the corridor space. This is illustrated in figure 6.

As pointed out earlier, air flow between the corridor and burn room appeared to become affected by the advancing floor fire. Eventually, this reduction of air supply to the burn room affected the crib burning rate. In fact, after flameover the crib burning rate always dropped. However, before flameover significant changes occurred in air supply to affect the nature of combustion in the burn room. Figure 7 illustrates this phenomenon by examining the measured crib weight loss rate and the maximum (estimated) burning rate which could occur within the burn room. This potential maximum rate was estimated from velocity and temperature measurements of the exhaust products at the burn room doorway (an assumed air to fuel ratio of 4 was used). No flameover occurred in test 346 and no drop in air supply is seen in figure 7. In test 348 the flooring fire had spread into the corridor by 320 seconds, and near 440 seconds (the onset of flameover), the figure suggests that most of the fuel generated in the burn room probably burned within the corridor space. Thus, the release of fuel-rich combustion products from the burn room to the corridor can be a contributing factor in causing flameover in addition to radiant heating of the floor covering.