REPORT OF THE COMMITTEE OF THE FRANK LIN INSTITUTE ON THE EXPLOSIONS OF The sub-committee, to whom was referred the examination of the strength of materials employed in the construction of Steam Boilers, beg leave to submit the following report : While it is important to know the causes which may produce a dangerous developement of elastic forces in the interior of steam boilers, it is obviously not less so, to understand aright the efficacy of those means on which we rely for confining or controlling their energies. Hence, in investigating the causes of explosions, it is both natural and expedient to examine separately those facts and principles which concern the divellent and the quiescent forces respectively. The number and variety of circumstances, which affect the character and durability of materials of which steam boilers are formed, are probably not less than of those which tend to modify the action of the fluids which they contain. In this view of the importance to be attached to the subject of the strength of materials, it may be considered remarkable, that while numerous investigations have been made as to the causes of danger, so little should have been attempted in regard to the most direct and obvious means of security. Before the series of experiments here detailed had been commenced, the necessity for such an investigation had been repeatedly pointed out, in public and private lectures on the steam engine; the reasons assigned for instituting the inquiry, being the very general and unsatisfactory nature of those results, which are given in practical treatises, respecting the strength of metals, as dependant on the mode of ma. nufacture, and on the different temperatures and other circumstances to which they are exposed. We had, it is true, a considerable number of results, obtained at different periods, by experiments on the direct cohesion of wrought iron,* they were, however, in *The following brief Table contains some of the general results, obtained by different authors, as the strength of wrought iron. general, undertaken for purposes very dif. erent from those which prompted the present investigation. Few of the experimenters had in view the influence of temperatures on tenacity; and even those data which they furnish for calculating the proper thickness of metal to be employed at ordinary temperatures, in constructing steam boilers, are liable to much uncertainty, owing to the diversity in the results themselves. Laborious and protracted as has been this investigation, still the practical importance of the subject has appeared to warrant a careful survey, and a diligent comparison of the various facts which might influence the practice of those who desire to attain a secure action in the steam boiler. Without entering, therefore, into all the delicate questions, which, had a mere scientific view been indulged, we night have been prompted to examine, it has been the aim of the committee to obtain and present such classes of facts as both scientific and practical men may make subservient to their respective purposes. The questions, which in the course of this inquiry, it has been found necessary to investigate, may be classed under three general divisions. 1. Principal. 2. Incidental. 3. Subsidiary. I. Principal.-1. What is the absolute tenacity per square-inch bar of rolled boiler iron, at ordinary temperatures, and to what irregularities is it liable? 2. The same for rolled copper? 3. What is the effect of increased temperature on the tenacity of iron and copper? 4. What in the tenacity of wrought iron, manufactured by other means than rolling into plates;-as by rolling it into bars or rods, by hammering and wire-drawing? 5. What are the relative advantages of iron made by refining from different sorts of pig metal and their mixtures? lbs. per sq. inch. Name of the Experimenter. Strength in ON THE STRENGTH OF STEAM BOILER MATERIALS, 6. What is the comparative value of sheet iron manufactured by the processes of puddling, blooming, and piling respectively; and in the last case, what influence have repetitions of the process? 7. What is the effect of piling into the same slab, iron of different degrees of fineness? 8. What is the comparative tenacity of rolled iron, in the longitudinal, diagonal, and transverse directions of the rolling respectively? 9. What influence may be produced, by long and repeated use, towards modifying the character of boiler iron? II. Incidental.-1. What is the specific gravity of the specimens submitted to examination? 2. What elasticity is found in the metals under differnt circumstances of the trial? 3. What relation exists between the force which will produce a permanent elongation in a bar, and that which will entirely overcome its tenacity? 4. What amount of elon ation may the several kinds of metal undergo before fracture ? 5. Does the amount of constriction or diminution of area, at the section of fracture, bear any relation to the absolute strength of the metals, to the direction in which the strips are cut from the plate, to the breadth and thickness of the strips themselves, or to the temperature under which the trial is made? 6. What is the effect of the rivets on the total strength of a boiler? III. Subsidiary.-1. What is the friction of the apparatus employed to determine tenacities? 2. What is the amount of its elasticity? 3. What is the latent heat of the vapour of water? 4. What is the specific heat of iron, copper and glass, respectively? 5. What is the rate of heating of a given mass of liquid, when subjected to the direct action of a solid of higher temperature? 6. At what rate will the same mass of liquid change its temperature by the action of air alone? From the foregoing statement, it will be seen that more than 20 distinct topics have demanded the attention of the committee. They have felt strongly inclined to embrace some other points of great practical and scientific importance, but the time already unavoidably consumed, and the very limited means which the other branches of inquiry and experiment on explosion left to be appropriated to the purposes of this sub-committee, compelled the relinquishment, for the present, of those objects which do not 19 immediately concern the construction and use of steam boilers. The discussion of the questions above enumerated, will necessarily follow an order somewhat different from that in which they are here stated. A view of the apparatus, employed by the committee, claims the first notice. The origin and preparation of the materials to be tested, will also precede the detail of experiments. Machine for proving the strength of materials.-The apparatus used, by the committee, for the direct determination of the principal questions regarding the strength of the specimens submitted to examination, is represented in plate I. (See front page). Mis a strong frame of oak timber, the two longer sides five feet in length, 14 inches deep, and 6inches thick. The two shorter, or end pieces, which project beyond the sides to the distance of 3 inches, are each two feet 8 inches long, seven and a half inches thick, and 14 inches deep. Between the two side pieces, (one of which is in the figure removed, to exhibit the interior or working parts,) is a space 14 inches wide, affording room for a screw, cross-head, guide rods, connecting blocks and wedges, to hold the specimens under trial; and also for the heating apparatus in experi ments at high temperatures. These four massive blocks or beams of timber are held together by strong screw bolts, passing through mortises in the end pieces, along tenons into screw nuts imbedded in the timber of the longitudinal bean.s. The frame is supported, as represented in the figure, by four firm trussel legs, 6 inches square, tied together near the bottom, and fastened as well to the ties as to the frame above, by mortising and bolting. The top of the frame is 3 feet 8 inches above the floor on which the machine rests. Through one end A, of the frame M, about 6 inches below the top, and centrally between the two side beams of the frame, passes the s rew S, 24 inches in diameter, and 3 feet long, cut into threads ths of an inch apart. Near the head of the screw is a neck turned rather deeper than the threads, to allow a clamp collar to embrace it; which, together with a strong cast iron plate, against which the head of the screw works, prevents any longitudinal motion of the screw itself. N is the box or nut of this screw, which by the revolution of S, either approaches to or recedes from the end A of the frame; 8 8, are two guide rods, one on each side of the screw, level with its axis and near the inner faces of the longitudinal beams of the frame, serving to support a cross head that contains in its central ring the nut N, and embraces by loops at its extremities the two guide-rods. The purpose of these loops is to prevent the nut from turning by the revolution of the screw. The cross head thus secured is united by two strong straps or bars of iron i i, 2 inches wide by half an inch thick, to a block of iron b, which is also furnished with two projecting arms that rest on the guide-rods already described. This block as well as the two others b' and b", is 4 inches long, 4 inches deep, and 1 inches thick, being perforated centrally in the direction of its thickness with a hole in the form of the frustum of a square pyramid, the purpose of which is to admit of wedges placed within them to hold the bars of metal under trial. A more detailed description of these will be given hereafter. The block b' is connected to b by a separate pair of straps i i, and has arms reposing on the guide-rods, or when necessary, admitting a vertical semi-revolution, so as to be laid over backward between the straps ii. This latter disposition of the block b was made whenever specimens of 20 or 30 inches in length were to be tried; but when those of only a few inches in length were under trial, b' was used in the position represented in the figure. The block b" is connected by the strong iron straps "i", which pass freely through a suitable opening in the head B of the frame, to the lever L. One of these straps is seen at e, the other being on the posterior side of the lever, with which they are united by means of a steel bolt turned with care and well polished. The straps are kept in place by a head, screw nut and washers, on the bolt. This lever is of the rectangular kind, the longer arm being horizontal, the shorter vertical, and the angular point being in the axis of a second or lower bolt which serves as a fulcrum. At the end next the frame, the lever bas a breadth or depth of 7 inches and a thickness of 1 inch. Towards the opposite extremity or that on which the weights are placed, it diminishes to a breadth of 4 inches, and a thickness of ths of an inch. The upper edge of the beam is straight to within 24 inches of the broader end, where it curves upwards, affording a massive support for the upper bolt already described. In a vertical direction beneath the bolt, and in the prolongation of the upper straight edge of the lever, is the position already indicated, of the second steel bolt, serving for a gudgeon, on which the lever turns. The distance betwen the axes of the two bolts, is 2.914 inches, which is therefore the length of the shorter arm of the lever. The bolts are very nearly of the same diameter, being each about 1.086 inches. The lower bolt rests against a plate of cast iron, having suitable projecting cheeks, with bearings adapted for its reception. A strap from the top of each cheek comes down over the bolt, and is fastened with a thumb screw, to prevent the lever being thrown out of place by the recoil of the machine. The two guide-rods & 8, pass through this cast iron plate, as well as through that which serves as a collar to the screw head S, on the opposite end of the frame. The lever is formed of the best wrought iron, and weighs 164 pounds, the matter being so distributed that if not neutralised by counter weights, its effect in straining any bar attached horizontally to the upper bolt would have been equal to 2495 lbs. To obviate this, and to prevent the weight of the lever from adding anything to the friction, it is accurately counterpoised by means of weights C and C,' corresponding to the parts of its mass which they are respectively required to sustain. Thus the weight of C, the larger counterpoise, is 103 pounds 12 ounces, that of C' 60 pounds 7 ounces. The former is, however, increased to counterpoise likewise, one-half the weight of the two straps i" i", the other half resting, as will be seen, on the horizontal guide-rods s s' The axes of the pulleys p p', over which the cords pass, are furnished with cavities to receive steel pivot-points, in order to reduce, as far as practicable, the friction of these parts. The diameter of these pulleys is 12 inches. The iron stirrup, to which the cord r is attached, is applied to the lower bolt or fulcrum of the lever, the projecting ends of which roll on straight, horizontal edges, forming the bottom of two loops with which the stirrup is furnished. By means of the suspending apparatus above described, the lever is enabled to obey any force acting vertically on its longer arm, with the advantage of ample strength and stiffness, combined with the condition of a theoretical lever, in respect to the gravity of parts. There are two modes of operating by which a bar of metal, placed in the machine between band b", might be broken, so as to ascertain the tenacity. The first is to apply the force of the screw S to strain the bar in raising a weight W suspended at any convenient point on the arm h of the lever; the second is to employ the screw only to regulate the height of that arm, and to restore it when relieved of the weights, to the horizontal position, whenever the extension of the bar had allowed it to fall below that position. ON THE STRENGTH OF STEAM BOILER MATERIALS. The latter method was, with very few exceptions, adopted by the committee,-both because it allowed of a more exact determination of the breaking weights, by a small addition at a time, and because it rendered the effect of the friction constant in its kind, being always in opposition to the gravitating force of the weight W, and subtractive, in the calculation. In order to apply this mode of action without requiring correction for the stiffness of the cord and the friction of the pulley p', it was only necessary, after adjusting the weight C', to remove so much as would allow the arm h of the lever to descend upon the slightest jarring of the machine. The tenacity of the bar, and the friction at the fulcrum, were then the only resistances to the motion of the weight W. The purpose of the tackle of pulleys P, is to elevate the scale pan and weight after they have descended to the floor, in order, by turning the screw S, to counteract the elongation of the bar under trial, and again to commence operations with the descending motion of the arm h. The power of the operator is applied to the tackle by means of the windlass w, furnished with crank, ratched wheel and click. The upper edge of the lever was graduated into parts distant from each other just ten times the length of the shorter arm. By the aid of these several appendages, the machine allows the most gradual additions to be made to the divellent force applied to the specimens, breaking each with a descending movement, and consequently, rendering the friction definite in the direc tion of its influence, being, as before stated, always subtractive. The very few cases in which the mode of operation rendered it additive, are particularly mentioned in the tables. At the outer extremity of the lever, and in the prolongation of its upper edge, is placed a style z, serving as an index to the graduated arc a, which is divided into minutes of a degree. The point of the style is 10 feet 3 inches from the axis of motion in the lever, and the length of the entire circumference which it would describe 772.8276 inches. Hence each degree is 2.14674 inches, and each minute, as measured on the arc, is .035779 of an inch. The whole extent of the arc a is about 5°, the zero, or point of horizontal position being placed 3 degrees from the upper extremity. The chief use of this arc was to determine approximately the elasticity of the bars, and of the machine itself, as preliminary to that inquiry. The weights W, in the scale-pan, (which, with its suspending chains, cross-bars, &c. weighs 56 pounds,) were, in 21 every case, applied on the lever, at the third mark, a distance from the axis of motion 30 times as great as that between the axes of the two bolts, or 30 x 2.914-87.42 inches. Friction of the Machine.-The amount of friction of the machine already described for testing the tenacity of metals, was an object requiring particular investigation before any thing more than a comparative value could be assigned to the results which were afforded by the experiments. To determine this point, it was deemed advisable to ascertain under various loads what proportion the weight which was sustained by the machine, after it had been raised by the screw S till the index stood at zero on the arc a, bore to that which, after the lever was relieved and then loaded again with the same weight, would cause it once more to descend to zero. Between the heads b' and b" was placed a strong bar of iron 1 inch wide by 4ths of an inch thick. Two methods were then pursued for the purposes of mutual verification. 1. A certain weight was placed in the pan suspended at h, and the screw S turned until, as before mentioned, it was raised to a level so that z stood at zero on the graduated arc. The windlass w was then employed to raise the scale-pan and entirely relieve the lever. On again restoring the weights, the index remained some minutes of a degree above 0, and an additional quantity of weight was necessary to bring it once more down to that level. As, in this case, the weight added served to increase the friction, it is manifest that the comparison of it with the whole weight, itself included, must be necessary in order to show the relation between the weight at first raised and that part of it which represented the friction of the machine. When the weight was raised by the screw, the bar which connected the heads 'b', must have sustained a strain composed of the weight raised added to the friction of the machine; whereas, when the weight was let on by the windlass while the index was at some distance above 0, the bar sustained a strain represented by the weight borne, diminished by the friction. II. The lever was caused to rest on a solid support near the extremity, the index being opposite to zero on the arc, and in that position the scale-pan was loaded with the weight under which it was intended to try the friction. The screw S was then carefully turned to strain the bar and bring the loaded arm of the lever barely off of the support. The weights were next raised by the windlass, and the recoil of the machine raised the lever to a certain clevation; from which it was once more depressed by replacing the weights upon it, and adding such an amount as would just depress the arm to the level of its original support. The first of the above methods gives the double, and the second the single, friction of the machine. The following table exhibits the weights in the scale pan, the weight representing the friction when the first method was employed; the same for the second method; and the ratio of the fiction to the total weight sustained by the lever. A correction is required particularly at the higher pressures, on account of the increased elas. ticity in the machine under the added weights, which actually brought the index down to zero sooner than it would have arrived at it, by the simple effect of a strain upon the lever regarded as inflexible. The machine was kept constantly well oiled, still a trifling difference may possibly have existed in regard to its condition at different times; but no influence of this sort was ever found sufficient to determine the rupture of a bar, after the weight had been taken up, the gudgeons newly oiled, and the same weight replaced which it had borne previous to that operation. From the above Table it appears that the second method gave results more nearly in accordance with each other than the first, but it will also be noticed that forty-four observations with the first gives a mean value sensibly identical with that obtained from forty-seven experiments with the second method of trial above described. We were hence led to adopt 5 per cent. on the weight as the effect of the friction of the machine. The bolts are of well polished steel and the lower bearing of cast iron, and the upper one or the eyes of the straps i" i", of wrought iron. This relation of friction to pressure between these substances, as deduced from the experiments of the committee, will be found to correspond very nearly with that obtained by Mr. Wood when operating on railway carriages.* *See Wood on Railways, Smith's edition, Philadelphia. 1832, p. 202. The mean of nine out of twelve experiments here detailed is exactly 5 per cent. for the friction between steel and cast iron. Total 47 Elasticity of the Machine.-In order to determine, in particular cases, the amount of elasticity exhibited by the bars under trial, it became necessary to ascertain the elasticity of the machine, when loaded with different weights. Several series of trials were accordingly made, expressly with a view to this object. Putting into the machine, in place of a bar to be broken, one which was intended not to yield sensibly to the strains applied, and not in any case to be permanently elongated by them, different weights were appended to the lever, and allowed to remain until the latter had become stationary. They were then carefully raised by the windlass, and the lever allowed to rise by the recoil until it became entirely free from strain. The number of degrees and minutes on the arc a, traversed by the index, was then noted, the weight replaced, and the trial repeated until it was ascertained that no error of observation had occurred. Three series of operations were performed, |