【正文】 nces shown between types A and B in the laboratory tests largely disappear in the larger fire tests. It is thought that this effect of scale is due in part to the way in which the wall of the tank affects the flow of foam in forming a foam blanket. With the small diameter laboratory fire tray, the foam quickly spread to the walls of the vessel forming a foam blanket upon which a significant depth of foam could be built. For the larger scale fires, a foam blanket appeared to be formed far less readily, and even when foam covered the surface of the fuel, it was more difficult to build up an impervious foam layer. The indications of a scale effect for this method of application are important since the practical problem of storage tank protection is concerned with fires many orders greater than those so far discussed. A small number of tests, therefore, were carried out on a much larger fire, 400 ft2, in a concrete dike to examine further the effects of scale.TESTS ON 400 FT2 Free Because of smoke pollution problems in the area where the tests weremade, it was not possible to carry out tests with long preburn times on the larger fires. The tests, therefore, were limited to applying foam from a standard foammaking nozzle to a 400ft2fire of 400 gal of regular grade gasoline (approximately 2 in. deep). A 2min preburn time was used in these tests. The nozzle provided a constant discharge of 40 gpm, an application density of gpm/ft2, which was directed as far as possible to the center of the test fire. The results obtained are shown in Table 3. TASLE 3. Results of the 18ft2Fire Test Foam type and properties Fire performanceHydrolyzed protein A expansion, 7。 critical shear stress, 180dynes/cm2。 25 per cent drainage time,2 min 30 sec. Hydrolyzed protein B expanmon, 。 critical shear stress,340 dynes/cm2。 25 per cent drainagetime, 3 rain 20 sec. No fire control. After 46 rnin of foam application, some reduction in flame intensity. No further reduction. Foam application stopped after 12 min. Fire reached full intensity again within an additional 2 min.Fluoroprotein . 70 expansion, 7。 critical shear stress, 90 dynes/cm2。 25 per cent drainage time, 2 rain 25 sec.Fire controlled in 45 sec. Foam application stopped after 1 min 15 sec. Fire out after 1 rain 30 sec.From the results of the laboratory tests and the 18ft2 fire, one might have expected that the 400ft 2 test fire would have been controlled readilyunder test conditions by the two hydrolyzed protein foam pounds applied at an application density of gpm/ft 2. One difference in the400ft2 test fire was that only a 2in. fuel depth was used as pared with deeper fuel layers on the smaller fires. It could, however, be an extension of the scale effect that was thought to explain the differences obtained between the small laboratory test fires and the 18ft 2 test fire results. LIGHT WATER Investigations of new foaming agents at the . Naval Research Laboratory led to the development of Light Water foam pound. There are a number of reports 57 relating its performance on shallow fuel layers associated with its use on aircraft crash fires. There are no test results reported on its performance on deep fuel layers。 therefore, a test was made on the 18ft2 test fire with a 30min preburn time at an application rate near the critical rate for the fluoroprotein foam . 70 used in the test program. The results are given in Table 4 together with the results for the other foaming agents used in the test program.TABLE 4. Results of 18 ft2 Fire Test of Regular Grade 90 Octane Gasoline, 17 in. Deep, after 30 rain Preburn Time (Foam Applied in Straight Stream to Center of Fire Area at a Rate of gpm/ft2)Foaming agent Fire performanceLight Water, 6 per cent solution Fluoroprotein . 70, 6 per cent solution Hydrolyzed protein A Hydrolyzed protein B90 per cent control in 1 rain. Extinction in 5 rain* 90 per cent control in 2 rain. Extinction in 5 rain * No control after 15rain foam application No control after 15rain foam application*Although Light Water had a more rapid control time than . 70, the long extinction times in both cases were due to persistence of flames near the hot tank wall.CONCLUSIONSThe following conclusions were drawn from the work described in this paper. ? Where normal proteinbased foams are applied forcibly to fires in Class A flammable liquids, their efficiency is greatly reduced pared with gentle application to the surface. ? Under operational conditions in the event of fire in a fuel storage tank containing gasoline, a hot zone at 70176。C to 80 176。 C is likely to form. Under these conditions, normal proteinbased foams are even less efficient, unless applied gently, and foam application at the present remended rates ( gpm/ft 2) would have no effect on the fire. ? Results on the 400ft2 test fire suggest that there could be scale effects for this type of foam application and point to higher values of the critical rate of application even for short preburn times. Further work is required at this and larger scales on fires in deep fuel layers with long preburn times in metal tanks. ? Foams made from the fluoroprotein foam and Light Water, applied at rates below the present remended rates of application, extinguished the test fires rapidly, even when a deep hot zone had been formed.REFERENCES[1] Tuve, R. L. and Peterson, H. B., A Study of Some Mechanical Foams and Their Use for Extinguishing Petroleum Fires, NRL Report 3725, 23 Aug. 1950, . Naval Research Laboratory [2] French, R. J., Hinkley, P. L., and Fry, J. F., The Surface Application of Foam to Petrol Fires, Fire Research Note No. 21, 1952, Department of Scientiiic and Industrial Research and Fire Offices39。 Joint Fire Research Organization. [3] Fry, J. F. and French, R. J., A Mechani