A stochastic model of fire spread in a building is shown. Fire spread is assumed to propagate discretely from one space to another along the network routes in a building. A calculation method is derived for the probability that specific spaces become involved in fire at an arbitrary time after ignition by using given fire spread probabilities between the spaces. Also, for the case that extinguishment is attempted, such probabilities can be gotten by this method if extinguishment probabilities are given corresponding to the scale of the fire. The method is applied to a hypothetical small house for illustrating the effectiveness of this method.
Transient fire behavior in a compartment was investigated using a reduced scale model with a PMMA specimen. The fire growth process from ignition to full involvement was presented in the floor, wall and ceiling fire situations. Flashover time depended on the ventilation parameter A√H, and had a minimum at about A√H = 0.01m5/2 in the floor and wall fire situations. But the behavior of ceiling fires was quite different from those of floor and wall fires. The ceiling may have a considerable resistance to self-sustained burning compared with the floor and wall. Sustained burning of the ceiling fire was observed only in the small A√H region of the small compartment.
The spontaneous combustion of coal is a typical reaction of exothermic oxidizing reaction. The rate of oxygen consumption of coal has been measured as a function of temperature from 80°C to 380°C. The kinetic parameters of the exothermic oxidizing reaction of coal in oxygen have been determined and the reaction was found to act the three stage reactions in the temperature range of 80-380°C.
Several simple solution methods of the smoke control problems that can be reduced to simple systems with the use of conservative assumptions were presented in an attempt to provide practical means for assessing the efficacy of smoke control measures for large scale spaces, such as theaters, auditoriums and large assembly rooms. Some of the methods were given in the forms of analytical solutions and the others were formulated into the forms of trial and error methods.
A series of full scale experiments were conducted using the BRI fire test facility to investigate the smoke filling behaviors in large scale spaces under various smoke control conditions. These experiments also served as the validation study of the simple analytical theories presented in Part 1 of this study. As a result, the predictive capabilities of the simple theories were proved to be faily good.
Thermal degradation of wood has been studied by the measurements of oxygen consumption (O.C.) and weight loss (W.L.) for wood and its constituents (cellulose, xylan, and lignin) under isothermal condition in pure oxygen atmosphere in the range 160-260°C. Lignin consumed a large amount of oxygen even at 160°C, nevertheless its W.L. value was extremely small. On the ohter hand, O.C. as well as W.L. of cellulose and xylan increased remarkably above 220°C and 180°C, respectively. The plots of W.L. vs. O.C. for wood components, which were evaluated on the basis of the component weight ratio of 2:1:1, gave a straight line up to 200°C. The plots for wood coincided fairly well with the evaluated plots up to 200°C. Above 220°C, however, the plots of W.L. vs. O.C. of wood increased largely as compared to those of 200°C and a breaking point occurred at an exposure time of 4h at 260°C. IR spectral measurements for these systems suggests that autoxidation of the amorphous region, due to the presence of peroxide and OH radicals, must occur untill 200°C is reached, while above 220°C the oxidation and degradation prevail more definitively, where loosening and collapsing of the crystalline fibril region of wood occurs.
This paper describes two series of fire tests of upholstered chairs for the investigation of their burning behaviour and for the comparison of results between the two series. The two series were (1) C-series, where the test specimen was located at the center of a test room, and (2) W-series, where it was located at the side of a wall. Measurements were made of the following items: mass burning rate, radiative heat transfer to the surroundings, heat transfer to the wall, optical smoke density, etc. In comparing the two series, no obvious difference between mass burning rate could be seen. For an engineering point of view it could be assumed that the burning rate of chair at the side of the wall is the same as at the center of the room. Also the radiant heat flux from flame can be expressed as a function of the mass burning rate and the distance from flame.
The product of visibility and the smoke density as expressed by the extinction coefficient is almost constant. This product depends mainly on the brightness of the object, the illuminating light in the room, and the smoke properties. However, with irritant smoke this simple relation can apply only to low smoke densities. Experiments were made to elucidate the effects of irritant smoke on visibility. Twelve subjects were evaluated on their visual acuity with Landolt's ring test chart in a highly irritant smoke filling room and eye blink rate was measured, as an objective indication of eye irritation. It is concluded that the lower visibility through fire smoke is attributed to the running tears due to irritants, as well as to the extinction coefficient. Based on this new understanding, the degree of physiological and physical effect is discussed.
A zonal model of compartment fire was developed to predict the heat flux distribution inside the room of origin. The feature of this model is expressed ; the room interior wall is divided into elements and the net heat flux of each part is estimated by means of net radiation method, in which the radiative interaction of heat between walls smoke and flame is considered. In order to examine the validity and limitation of the proposed zone model, a series of experiments were conducted with a small scale model compartment under 32 different conditions consisting of several kinds of fire sources and ventilation openings. The macroscopic values of average heat flux, hot layer temperature and hot layer depth etc, and net heat flux distributions were in good agreement between the experiment and the prediction.