Cleveland Paper - Explosions from Unknown Causes. [Discussion of the Paper by Mr. Bayles, Transactions, xix., p. 18]

[In discussion of the paper of Mr. J. C. Bayles, read at the New York meeting of September, 1890, Trans., xix., p. 18.1 It is often so difficult to locate the causes of failures of machinery and apparatus which can be examined and studied at leisure, that it is not surprising that the causes of explosions should escape detection. Usually the explosion is the first evidence that anything is amiss, and the time for observing the conditions is necessarily brief. An incident in a welding-works near Essen, Germany, might be cited as exemplifying what trifling causes may lead to disaster, and how difficult, under the most favorable conditions, it may be to discover them. A blowing-engine was arranged with two suction-pipes at each end of the blowing-cylinder—one leading to the air, and the other to the gas-holder. Each pipe was provided with a valve, and all were so adjusted that the cylinder drew in gas and air in the proportions found most advantageous. The gases from the cylinder were carried directly to the furnace through suitable pipes, and while the cylinder was provided with relief-valves, every effort was made to prevent explosions, principally by compressing the gases to a point where the rate of efflux of the mixture was greater than the velocity of combustion, if the expression is here allowable. It was found that the apparatus would work satisfactorily for some time, and then the relative proportions of the mixture would change, and an explosion might have been the result. Experiments consuming days were made to discover a method of averting this danger, but they were unsuccessful, until it was observed that the change in the composition of the mixture of gas and air took place when the gasometer began to rise or fall—the back-pressure due to friction when the holder was rising or its absence when falling being sufficient to change the conditions. As soon as this was discovered,