Checks on factors in characterizing fire-hazard of several organic building materials with respect to the criterion on ignition and surface flame-spread have been pursued by radiation-predominant present JIS A 1321-type furnace and convection-predominant vertical furnace (used to be JIS A 1321-type furnace) and Madison tunnel-type furnace. Following measurement have been conducted particularly for plywoods in ISO-type furnace ;
1) relative rate of heat-produced or dQ/dt.
2) decomposition-rate or dW/dt. (dL/dt in tunnel-furnace)
3) smoke produced and its production-rate or dC_A/dt.
4) temperature-disribution inside of the samples or T vs. thickness d.
5) radiation-intensity and temperature distribution on the surface of the samples in the respective furnaces.
Counter-charts representing the relations among those factors have been obtained and illustrated in the report for the further analysis on fire-hazard of the samples, which suggest the predominant effect of radiation from the fire-source in the furnace super-imposed on the convectional heat-transfer with respect to criterions on the ignition and combustion patterns of the samples. Secondly, distinctive difference concerning item 4 (based on the difference of the amount of heat-flux through the surface) have been recognized with the nearly same surface temperature between ISO-type furnace (presently available) and used to be JIS A 1321-type.
Thirdly linear relation has been observed between weight-loss of the samples or (W₀−W )/W₀ and sample-thickness d for a series of plywoods and coincidence has been recognized at d of ca. 12mm concerning the pattern on dQ/dW of test-samples vs. thickness d between present JIS A 1321-type and the convection predominant furnaces, which illustrated opposite patterns particularly for thin samples. Finally, similar patterns have been observed in dC_A/dt vs. dQ/dW and dC_S/dt vs. dQ/dW diagrams between present JIS A 1321-type furnace and Madison-type tunnel furnace (ASTM D E 286-69) for the surface flame-spread. We have reached to the following pattern concerning the fire-hazard of the organic building materials based on the above data ; dW/dt or dL/dt of the samples in the vertical and tunnel samples which are proportional to dQ/dt, will be dependent on the radiation-intensities through the surface of the samples.
Therefore, the appropriate setting of the measure of the radiation amount in the test-furnace corresponding to those as estimated by the respective phase in the modeling of fire will be the first preference in the criterion on the classification in the firehazard of organic interior building materials.

View full abstract

As a theoretical checks on the conclusion for the possibilities in the spontaneous ignition of wood pillar as a cause of fire in Happokan-hotel case which had been refered to in Part I of the report, computer simulations have been pursued with respect to the critical conditions for the spontaneous ignition of wood spheres of various size. Above checks have been primarily concerned with the anisotropies in the thermal conductivity and diffusion constant or permeability of water vapor in the real wood-system and with the circumference-temperature of the heat-bath and radii of the wood spheres. Secondly, the checks have been concerned with the problems on the transfer the ignition-point for the spontaneous ignition in wood-spheres induced by the change in the surface-temperature of the sphere or heat-flux and in the activation energy of the thermal decomposition. Thirdly, checks have been pursued with the change in the critical condition for the ignition as induced by the presence of neighboring materials such as insulating materials and mortars, which wrapped the sphere.
The effect of nominal moisture-contents of wood as a function of relative humidity of the environment to retard the ignition has been discussed with respect to the quasi-stationary state problems as revealed in the solutions of non-linear differential equations. Finally, the numerical computations for the ignition of wood materials has been pursued with the case of arbitrary shape and size and with different activation energies in the respective part of the material as an example of the real cases.
Very good coincidences have been observed between calculated values and observed ones in the respective cases, which support our conclusion in Part I of the report. The numerical computations have been conducted by I.B.M. 1620-type computer.

View full abstract

The visibility of signs in smoldering smoke of filter-paper was reported in the previous issue¹⁾ by the author. In this paper, the visibility of signs in various smoke such as generated from wooden and various plastic building materials is studied.
The relation between the visibility and the smoke density at the obscuration threshold of the sign in smoldering smoke (white) or in flaming smoke (black) are shown in Fig. 1 and Fig. 2. The product of the visibility (V ) and the smoke density (σ) at the obscuration threshold is almost constant in the case of either white smoke or black one.
Relative scattered luminous flux due to smoke, which influences the visibility of sign, is measured (see Fig. 3). There is no difference between the scattered luminous flux due to smoldering smoke of Japanese cedar and that of various plastics. However, the scattered luminous flux due to flaming smoke varies with the kind of building materials or the quantity of supplied air for burning. Assuming that the value of k for smoldering smoke (white) is 1.0, the values of k for flaming smoke is obtained from the experimental results to be about 0.3∼1.0 as shown in Table 1 and Table 2.
The sizes of smoke particles, which influence strongly the scattered luminous flux, are measured by taking microscopic photographes (see Fig. 4). The particles of various smoldering smoke are sphere of about 1 μ as shown in Photo. 1, Photo. 2, and Fig. 5, while those of flaming smoke are consists of non-sphere particles of about 1∼20 μ, and some sphere particles which are considered smoldering smoke particles. Also, an experiment is carried out to take microscopic photographes on suspended smoke particles (see Photo. 4 and Photo. 5).
The threshold contrast (δ_c) of a back-lighted sign varies depending on the visibility, the illuminance in corridor and the properties of smoke, but it is presumed to be δ_c = 0.01∼0.02 under the conditions of the visibility of 5∼15 m and usual corridor-illumination lights (see Fig. 8).
The value of L in white smoke is given by the mean illuminance (Mean illuminance from all direction multiplied by 1/π ) without smoke. However, L in black smoke requires tremendous calculation because it depends on the smoke density and the properties of smoke. By assuming that the illumination is given by point light source, and reflection from wall surface can be neglected, L in black smoke will be given approximately by Eq. (4).

View full abstract

Small tunnel-type test furnace of 85 cm length and 25 cm width with variable inclined angle was developed for the checks on the surface flame-spread rate and smoke-evolution with respect to the organic building materials as wood panels and plywoods.
Following results have been obtained for the stated purpose ;
A) Mean surface flame spread rate dL/dt and heat evolution rate dQ/dt were approximately propotional to cos^1/2θ and cos^1/4θ with respect to the inclination-angle of the furnace θ, where the plot of flame spread trends vs. time or mean dL/dt were approximated to be linear macroscopically.
B) Oscillative correspondence were recognized in the relations among smoke-evolution rate dC_s/dt, hest-evolution rate dQ/dt and weight-loss rate vs. flame-spread rate dL/dt.
Linear correspondence were also recognized in the relations among dC_s/dt, dQ/dt and dW/dt.
Smoke evolved C_s were proportional to heat evolved Q [deg·min] before the flame had been reached to vent-end (surface combustion area I), while C_s were inversely proportional to Q after the flame had been reached to vent-end (volume combustion to thickness-or Z-direction area II).
Above results will suggest the following pattern on factors concerning the surface flame spread of organic building materials as wood panels or plywoods ;
A) Dominationg factors on the mean surface flame spread rate of the above materials were the flame shape (lenghth and thickness) on test materials which were subjected and proportional to Reynolds number Re and total heat-transfer coefficient involving heattransfer by radiation from flame or h_X (=h +h_R ) for the system concerned in the laminar flow approximation. Secondly, decomposition rate of the material or Arrhenius term will support the acceleration of the spread which depends on the chemical nature of the materials.
B) Oscillative relations were recognized among factors representing the flame spread on the microscopic view-point or dC_s/dt, dQ/dt and dW/dt which were due to the oscillative heat equilibria induced successively by the competitive combustion among lenghth- or x-direction along the surface and those of thickness- or z-direction of the test samples.

View full abstract