Detroit Drill School Explosion
FOLLOWING exhaustive tests, a report was submitted by the Detroit Fire Department containing the opinions of technical experts regarding the explosion which occurred in a room at the department training school on July 31 and which resulted in the death of Assistant Chief Archie McClellan, of Port Huron. Mich., and injuries to forty-three spectators.
The room in which the explosion occurred is about the center of the west side of the six-story section of the building, on the first floor. It is fifteen feet by ten feet ten inches, by thirteen feet five inches high and contains about 2,183 cubic feet. The ceiling and floor are of concrete, with two vents in the ceiling, an 8-inch one in the center, covered bv a close fitting cast iron cover and one in the northwest corner of the room. The walls are of solid brick. The north and south walls are without openings, and the east and west walls have a window and door each. The room is ordinarily used as a demonstration room for gas masks, chemical extinguishers, automatic sprinklers, etc.
Description of Test Room
It was provided with a carbon dioxide automatic installation consisting of two drawn cylinders, each 8½ cinch outside diameter and 57-inch high. These cylinders were connected to sixty-eight feet of heavy brass 1/2inch tubing installed around the walls of the room. The tubing was fitted with two automatic door releasing devices and six discharge nozzles with openings of 5/32inch. Two of the nozzles were on the north and south walls, and one each on the east and west walls, thirtyfour inches from the floor and pointing toward the floor at an angle of about forty-five degrees. Each of the cylinders contained fifty pounds of liquid carbon dioxide.
The room was also provided with an Aero automatic fire detecting system, and a sprinkler system consisting of four 165° heads, located about one foot from the ceiling. The lighting equipment consisted of four 100watt lamps, enclosed in vapor-proof globes, mounted on the north and south walls, 6½ feet above the floor.
For the demonstration, a quantity of gasoline was poured into a galvanized pan, two by three feet and 1 1/2inch deep, and weighing about eight pounds when empty. The average distance from the pan to the nozzles was about five feet. From two to three minutes elapsed from the time the quart of gasoline was poured into the pan until the liquid was ignited. After the match was tossed by the demonstrator into the pan. he stepped out of the east door of the room, closed it, and stood outside the glass panel in the door to watch operations.
Flames rose six to eight feet. In about seven seconds, the carbon dioxide gas was released. About five seconds (various witnesses stated up to two minutes) later the explosion occurred. Some observers said that the gasoline pan rose about two feet from the floor just prior to the explosion. The blast forced out all panes in the east window, some of the panes in the west window, and cracked two others. Glass was scattered for a distance of seventy-five feet, while flames were projected out fifteen to twenty feet.
Is Explained by Subsequent Tests
In order to find the cause for the explosion, tests simulating the original demonstration were conducted on August 16 by the following investigators: A. H. Nuckolls, Chemical Engineer, Underwriters Laboratories, Chicago; Alfred H. White, Professor of Chemical Engineering, University of Michigan; Frank Burton, Consulting Engineer, Detroit; Ormund Rugg, Chief Chemist, Glass Plant, Ford Motor Company: Harold Mercer, Superintendent, Glass Plant, Ford Motor Company; H. H. Mills, Chief Safety Engineer, City of Detroit: Robert Loughead, Chief Engineer, Michigan Inspection Bureau, Detroit, and Fred DeVoist. Chief Chemist, Department of Health, City of Detroit.
During the tests, considerable turbulence in the air was created by the discharge of carbon dioxide, and at some of the tests, the pan was lifted as high as three feet.
The findings of some committee members follow:
H. H. Mills
“Under ordinary conditions the mixture of unconsumed gasoline vapors or CO2, with the air in the room, would be of such uniformity that the mixture would be too lean to ignite, but with the blast of CO2 on the floor, the unburned gasolinevapors or CO2 would be forced upward and probably stratified with the result that an explosive mixture was formed and an explosion made possible, although the quantity of gasoline vapors or CO2was not ample to fill the room to cause a major explosion. The conditions for such stratification of gases, were ideal considering the size and dimensions of the room.
“The accident and the tests clearly demonstrate the impossibility of predetermining the action of flammable or explosive mixtures.”
“It is my opinion that, if the pan of gasoline had not been lifted by the CO2pressure from the jets and violently spilled or splashed there would not have been an explosion. The unburned gases caused by the spraying of the gasoline mixing with the oxygen in the air reached a point of saturation where they became combustionable.”
“When the carbon dioxide was released from the jets pointing down at an angle of 45° the various streams of gas met near the floor and were then deflected upwards, since each jet balanced the force of the one opposite to it. In consequence, the pan containing the gasoline was lifted bodily upwards and the burning gasoline was sprayed into the upper part of the room. Although some of the gasoline burned there must have been a portion of the gasoline which was vaporized in an atmosphere containing so much carbon dioxide that it could not burn. When, however, this reached the upper part of the room where the carbon dioxide content was lower and the oxygen content higher, there was a sudden combustion of the gasoline resulting in an explosion.
“This explosion was the result of a number of conditions, such as the size of the room, the height and angle at which the jets were set, the size of the gasoline pan, etc. If any of these conditions were materially changed, it would probably be impossible to reproduce the explosion. During the tests made yesterday there was no explosion when the angle of the jets was altered.
“It seems so impossible that an explosion could be produced in this way that I was not inclined to believe it possible until I witnessed the reproduction of the phenomena.”
Professor Alfred H. White
“When the jets of carbon dioxide first struck the floor and were carried to the pan of gasoline the pan jumped at least two feet in the air with the gasoline still blazing. The flame was not extinguished but in about two seconds there was a bright flash in the room and the windows were blown out.
“I had not anticipated such a definite explosion. I had thought that the increase in pressure caused by the expansion of the hot gases, together with the added effect of the carbon dioxide released, might cause the windows to be blown out by the gradual accumulation of pressure. What took place, however, was a definite explosion of a mixture of gasoline vapor and air. The only explanation I have for this is that when the pan containing the blazing gasoline was thrown in the air by the jets of carbon dioxide gas from the fire extinguisher that the liquid gasoline was scattered as spray. This spray volatilized almost at once and formed a localized pocket of explosive mixture which flashed from the flame of the gasoline which was still burning in the air.”
A. H. Nuckolls
“It appears from the foregoing that the explosion which occurred in the Smoke Room of the Firemen’s Training School, July 31, was caused by the combustion of gasoline. It was evidently a localized gasoline explosion of a low order. The vaporized gasoline at the moment of the explosion was not distributed throughout the room, a quart of gasoline being capable of rendering only about 400 cubic feet of air flammable or explosive, but was concentrated in a mixture with air in a space probably near the center of the room. As would be expected under such circumstances the pressure effect was comparatively low, being of the order of about one to two pounds per square inch, which was sufficient to blow out the windows. The windows of ¼ inch wired glass were probably not capable of withstanding a pressure much, if any, above 135 pounds to the square foot.
“It may be inferred from a study of the flame phenomena observed when the explosion occurred during the demonstration July 31 that a considerable portion of the gasoline had not burned at the time the windows were blown out. The large amount of flame projected from the windows extending a horizontal distance of about ten feet, which severely burned several observers, was probably due mainly to vaporized gasoline and not carbon monoxide1 from the partial combustion of the gasoline. The conditions existing in a room containing 2,200 cu. ft. of air and about a quart of gasoline would preclude the formation of such a large volume of carbon monoxide in such a short period of time even with carbon dioxide being rapidly discharged into the room.
“It appears therefore that the pressure effect which ruptured the windows was due to the combustion of much less than a quart of gasoline probably about a pint. This inference is confirmed by the results of the third test in which the pressure effect was approximately reproduced by the combustion of about a pint of gasoline.
“It is clear that there was sufficient oxygen in the air for the combustion of the gasoline2. The dilution of the air by the carbon dioxide discharged into the room, near the floor level, during the small interval of time between the operation of the system and the explosion, was evidently not sufficient to retard combustion to any marked extent.
“The discharge of the carbon dioxide at high velocity from the six 5/32-in. outlets evidently caused a marked turbulence in the air of the room. In addition there was an upward movement of air and gas near the center of the room. This movement was probably at a maximum when the outlets were at an angle of 45°. Naturally gasoline in both the liquid and vapor phases was blown upward with the result that the liquid was vaporized somewhat more rapidly, and the vaporized gasoline was rapidly diluted with air forming a more combustible mixture and thus mechanically bringing about conditions which facilitated the communication3 of the flame from the burning portions of vapor to the adjacent unburned portions. These conditions, being highly favorable to a sudden increase in the rate of combustion with resulting increase in the maximum pressure. appear to be the underlying cause of the explosion which occurred in the Smoke Room July 31 and also of the explosion reproduced in the third test.
1Note:Carbon monoxide is lighter than air and at least 12 1/2 per cent by volume is required to form a flammable mixture with air.
2Note: Theoretically, 32 cu. ft. of oxygen would be sufficient for the combustion of a pint of gasoline. The room contained about 460 cu. ft. of oxygen at the start of the combustion. The partial pressure of the oxygen remaining after combustion would be 0.195 which is ample for combustion.
Other factors including the elevated temperature produced by the burning gasoline prior to the discharge of the carbon dioxide gas and the small pressure rise caused by the combustion and the discharge of the carbon dioxide gas contributed to the increase in the combustion rate but the underlying cause as indicated above, was the turbulence together with the dilution of the gasoline by the air as it was blown upw’ard.
“It is not thought that any minor difference in the conditions or in the time of operation of the release device would have prevented the explosion in the Smoke Room.
“The movement of the pan, containing the gasoline, upward undoubtedly caused a more rapid formation of the explosive mixture of gasoline vapor and air, thus contributing to the intensity of the combustion and the resulting explosion.
“It is generally supposed by firemen and others trained in practical fire prevention work that when the quantity of a flammable liquid, such as gasoline, is not sufficient to give enough vapor to form a flammable mixture with the air in an enclosure that there is little, if any potential explosion hazard. It is not surprising, therefore, that the explosion hazard from the combustion of a quart of gasoline in the Smoke Room of the Firemen’s Training School, the net capacity of which is of the order of 2,200 cu. ft., was not recognized in advance. Further the two fire extinguishing systems were considered ample to cope with the situation. Thousands of demonstrations apparently similar to the one which resulted in an explosion have been safely conducted.
“It is clear that the essential conditions bringing about an explosion hazard in this case relate to the rapid dissemination of the vaporized gasoline in the air together with a marked turbulence in advance of the formation of the nonexplosive atmosphere by the discharge of carbon dioxide, a situation that is unusual and evidently complex in its nature.”
1Note: Turbulence in a flammable mixture increases the frequency of the effective collisions of the pas or vapor and air molecules and thus accelerates the combustion particularly in case of a lean mixture.