ON THE STRENGTH OF STEAM BOILER MATERIALS. ness, the operation was extended so far as to reduce below a proper point the temperature of the copper, and thus to injure its texture. It will be seen that the highest results obtained by the committee, are almost identical with that given by Mr. Rennie. In every calculation of the strength of meterials for a steam boiler, the least strength known to be possessed by any part of the sheet, is that which alone can be relied on for fixing the pressure to which it may be subjected. For copper, at ordinary temperatures, the lowest result obtained by the committee was 30406 lbs, per square inch, and the mean minimum for the 8 bars 32146 pounds. To other temperatures subsequent developments apply. Effect of increased temperature on Copper. -The effect of temperature on tenacity, has been hitherto but slightly examined, either for theoretical or practical purposes. The general truth that heat diminishes, and eventually overcomes cohesion, is too well established by daily observation to admit of question. The temperature of no tenacity, is generally supposed to be that at which the fusing point of the given substance is placed, and the point of maximum tenacity ought, upon general principles, to be found at the point where least heat prevails, that is, at the natural zero, or point of absolute cold, if such a point exist in nature. Between these two extremes, it might be supposed that the tenacities of different substances, particularly such as are capable of passing immediately from the solid to the liquid state, would be found to obey certain laws. As the total cohesion at the maximum would present to a mechanical agent tending to overcome it, the whole of its resistance, and as, at more elevated temperatures, a part of that tenacity would be overcome by heat, and the rest must be destroyed by the mechanical force, it is evidently a question of experiment, to decide what relation the two forces have to each other at the several temperatures between the two extremes to which we have just alluded. To decide the theoretical question, or in other words, to deduce, from the experiments, a law which might be expressed in an abstract form corresponding to all the possible phenomena, would require a state of the materials different from that usually found in commerce or employed in the arts. It would also, as we have seen, require a knowledge of that, about which philosophers no less than practical men, are far from being agreed;namely, the point of absolute cold. As the purposes of this committee did not lead 46-5 them to investigate the problem in all its possible bearings, but only in view of the limits which practice assigns, and with the conditions commonly given to the materials, it will not perhaps, be easy, to construct from the tables a formula in all respects unexceptionable. The general course of experiments involved the necessity of operating, at the different temperatures, on different bars of copper, and as all the bars are not found to give, even at ordinary temperatures, the same strength, for equal areas of section, it became necessary to deduce from experiments on each bar, at some assumed low temperature, a standard tenacity with which to compare its strength at every other point. The part of this standard tenacity which was taken away by the heat at the higher temperatures, becoming known by the experiment, a comparison was furnished for deciding approximately the relation between the temperature given, and the portion of tenacity which it had overcome. It will be found on an inspection of table XXIV. containing the comparison of these experiments, that on the eight different bars, the whole number of trials which furnished standards of comparison, at ordinary temperatures, was sixty-six, and consequently on an average about eight trials to each bar; while at the elevated temperatures there were made thirty-nine different experiments at nineteen different points on the scale, the greater number of points, however, having but one experiment each. An inspection of plate IX., where these experiments are represented, will show that at nearly all parts of the scale, within which the trials were made, the strength diminishes more rapidly than the temperature increases, but some of the higher experiments indicate that the conditions of the law are such as to be represented by a curve, having a point of inflection. It will also be noticed that the three experiments which appear anomalous, and which in the plate are narked with queries, are all found in trials of the same bar of copper, (No. 7,) and that all these might be referred to a curve, varying but little in form from that which we have traced. It is not however necessary to suppose that these experiments belong to a different curve, for upon recurring to the table of bar, No. 7, (table XXII,) it will be found that one of the anomalies is satisfactorily accounted for by a delay in taking the temperature after the fracture had occurred, and t at one of the others and probably both, were cases of weakening by a slight alloy of the copper by the melted metal through which it passed, in consequence of not having been defended by oxide. The other bars tried at high tem peratures were treated with dilute nitric acid, creating a thin film of oxide, which effectually defended the surface, without sensibly diminishing even the smoothness of the bar. It will be observed that the difference of tenacity, at the lower temperatures, for a difference of from 60 to 90 degrees, is scarcely greater than the actual irregularities of structure in the metal at common temperaPlate IX. Observed diminutions of tenacity. Calculated diito. tures, and consequently, it was not practicable from these experiments alone to deduce sa law which should express the tenacities at all points between the maximum above referred to, and the melting point of the metal. Nor would much confidence probably have been reposed in results thus obtained. In laying down the results in plate. IX, the line ab is made to rep esent the total tenacity of copper at 32. The horizontal ON THE STRENGTH: OF STEAM BOILER MATERIALS. dotted lines express the observed tempera.. tures above 32°, and the vertical ones, the diminutions of tenacity at the respective points. In examining the eleventh and twelfth columns of Table XXIV. (Frank. Journ., vol. xix, p. 242), with a view to a relation which may afford a practical rule for cal. culating the strength of copper at any given temperature, it will be found, that with the exception of the three anomalous cases, Nost 11, 14, and 17 of the table, they may be referred to a species of. parabola, of which the ordinates representing the temperatures above 32°, have to the abscissas representing the diminutions of tenacity, a relation expressed by the cube roots of the squares of the latter quantities; or, in other words, that the squares of the diminutions are as the cubes of the temperatures*. By applying the law above statedt, and assuming the greatest diminution observed, or that obtained at 1,000° above the freezing! point, as a true standard of comparison, we get the calculated results contained in the 13th column of the table, and a comparison of that with the twelfth, furnishes the dif ferences in column 14th. This last, com pared with the 9th, shows that the greatest deviations, even of the anomalous experi ments already noticed, do not amount to so much as the actual irregularities sometimes found in the metal at common temperatures, for while the highest numbers in the 14th column are less than four and one-tenth per cent. of the total strength, several of those in the 9th amount to more than four and a half per cent. of the same sum. The curve traced (Plate IX.), represents the column of calculated results, and is con.. *To determine whether any, and if any, what, single function of the temperature will at any point expres the diminution of strength, as compared with that observed at other points, it was not deemed expedient to rely on a single comparison.. The following method was therefore employed to obtain an expression corresponding with each of fourteen diffe ent points, compared with thirteen others. Putting/=any observed temperature above32 deg. ; t = any other temperature above the same; d= the diminution of tenacity, by the foriner temperature, and d' = that by the latter: also r the power of the temperature, according to which the diminution of tenacity varies, we have t'e d' ta: t'x::d: d' whence-from which we get a = to d; Log d' Log·d', Thus at a temperature of 984 deg, Log - Log t Fab., the tenacity was found by experiment to have been diminished .6691, its amount at 32 deg. being 1.0000; and its diminution at 492 deg. was .2183; hence by the above formula, Log 6691-Log .2133 Log (984-32)—Log (492—32) =1.503. The following table exhibits the mean results of the several sets of comparisons with the tem 467 fully established; and in this granular state they, in some cases, continue through a considerable range of temperature. The melting points are those at which fluidity is clearly established. But, notwithstanding this fact, and the very close accordance of the law above mentioned, with the observed diminutions of tenacity, we do not venture to assert that the theoretical law which might be derived fron operating on copper absolutely pure, and of uniform tenacity throughout the specimen, would not give a form so varied as to change the parabolic curve into one possessing a point of inflection. An inspection of the figure, as well as a reference to the table in the preceding note, will be found to favour the supposition that the rate of increase in temperature corresponding to a given decrease of tenacity, does in fact pass through a minimum near the point where one-half of the absolute tenacity is overcome. The right hand branch of the curve indicates the probable course after inflection. Extensibility of Copper.-In producing the rupture of bars of copper it became evident that this metal undergoes during the mechanical strain to which it is subjected, a degree of elongation, dependent in some measure on the temperature to which it is raised. The mode of asc. rtaining this point consisted in measuring, after the trial of each bar had been completed, the united lengths of all its fragments. In reconstructing the bars for this purpose, care was taken to bring the corresponding portions into as close a contact as possible, and also to allow by estimation for any imperfection in the same from roughness of the fracture. A second mode was, to select from among the fragments of each bar one or more which retained the original inch-marks, and which had at the same time been apparently strained to the full extent of its resistance without actually parting. By this latter method of trial it was ascertained that the extensibility of all the 8 bars, with the exception of Nos. 6 and 7, was nearly uniform, varying only between 40 and 44 per cent. of the original length. A section measured on No. 6, gave the length between two-inch marks only 1.25 inch, and on No. 7, 1.28. The trials on both these sections had been made at ordinary temperatures. When comparing the total lengths after fracture, with the original length of each bar, we obtained as a general result, very nearly the same extension as when employing the several inch-marks as just stated. The mean elongation of the whole after 116 fractures, was 43.5 per cent. of the original length. Other things being equal, the bars of least area appeared to have been most extensible. No. 2 was stretched, by 18 fractures, from 30 to 46 inches. No. 8, by 14 fractures, from 30 to 43 inches. But the circumstance of most importance is the temperature of the bar at the moment of trial. Thus, on bar No. 7, (Table XXII.) the first fracture was made at 912° and the area of section afterwards was .744 x .244.181536 square inch, and the diminution from its original size only .002571, while at the thirteenth fracture, when the temperature was 81.5°, the area, after trial, was .550 x .174.095700, a diminution of .088316, or 34 times as much elongation as before. Strength of Boiler-Iron at ordinary temperatures.-The results of experiments on boiler-iron, at ordinary temperatures, will be found included in thirty-two tables, from XXVI. to LVII., inclusive.—(Frank. Jour. pp. 246 to 275, and 326 to 360.) On some few of the specimens, the strength of which is exhibited in these tables, all the experiments were made with a particular view to the irregularities of the metal, and at, or near the same temperature, while on other bars much diversity in the objects of the experiments prevailed, and consequently of these, only a few trials can be selected which may be considered entirely appropriate to the present topic. When making comparisons with a view to the mean strength of sheet iron, even from the same plate, it is necessary to consider that the question may be answered differently according to the direction in which the specimen was cut off; to the condition in which it was submitted to trial, whether rough from the shears. filed to a uniform size and smooth surface, or filed away in notches to overcome the influence of the shears; or, finally, according to the previous treatment of the specimens, whether subjected or not to annealing or other influences of heat after leaving the rolls. The tables furnish, under appropriate heads, the information necessary to answer, separately, the several questions arising out of these different aspects of the subject. With regard to the method of preparing the specimens, by reducing them to an uniform size throughout the whole extent of the bar, it may be remarked, that on the 41 bars of iron, which in the course of this report are described as having undergone that preparation, there were measured 1049 points, or sections, and there were made 517 fractures, showing on an average but little more than two inches between two adjacent points of fracture. It also appears that on only two of those bars (Nos. 220 A. and 224 B.) did the mean area, of all the points measured, correspond exactly with that of all the sections of fracture. On 22 ON THE STRENGTH OF STEAM BOILER MATERIALS. bars the mean area of the fractured parts is less than that of the measured sections, by an average of .000340 of a square inch, and on 17 bars he mean section of fracture is greater than the mean measured sections by an average amount of .000187. This proves what might indeed have been anticipated, that the fractures would, in general, take place at the smaller sections, and as the mean area was about .175 sq. inch, it appears that the difference between the measured and the fractured sections, due solely to irregularities of filing, that is, between the condition of our specimens and that of others which should be absolutely uniform in size, amounts to not more than or Te part of the total strength. This portion is less than the irregularities in the structure of rolled iron, as may be shown by referring to tables LV. and LVII. Methods of manufacturing Boiler Iron.— For the information of the general reader, who may not be familiar with the several processes in the manufacture of iron referred to in some of the foregoing, and in several of the subsequent tables, it may be proper here to make a few remarks explanatory of the methods pursued in the United States for producing wrought iron of the descriptions embraced in this part of the report. It has already been stated that the iron furnished to the committee was, with a single exception, manufactured by the aid of charcoal. This remark applies, of course, to the first process, that of smelting it from the ore. which is, for the most part, performed in the usual blast furnaces, from 30 to 40 feet in height, and about 8 feet in their greatest interior diameter, producing the different varieties of pig metal. It has been mentioned to us, that in Missouri this process is sometimes dispensed with, especially when working the ore of the iron mountain," a rich, heavy, magnetic oxide, of a bluish or iron-grey colour, and of the extraordinary specific gravity of 5.36. The ore is there put into open forgefires, resembling the Catalan forges of the south of Europe, and by a similar treatment to that which there prevails, brought at once to the condition of maileable iron, without passing through the state of cast or pig metal. The process of manufacturing blooms, or as they are, when intended for boiler plate, technically termed, "blocks," is to subject the pig metal of the blast furnace to the combined action of heat and air in an open forge-fire of charcoal, drawing off the melted cinder, or “ slag,” by a suitable opening; and after stirring and compacting the iron as it begins to agglutinate, or "come 469 round to nature." to carry the ball to the heavy forge-hammer, and form it into a prismatic mass, from 15 to 20 inches in length, and from 5 to 9 inches in diameter, according to the weight of the plate intended to be obtained from it. These blocks when taken to the rolling mill are heated in an air furnace, supplied generally with bituminous coal as a fuel, and at the first heat are reduced by a heavy hammer into slabs, two or three inches thick, and of a length nearly corresponding with that of the blocks. This operation discharges much of the remaining cinder, and other impuri ties left in the block by the bloomery treatment. At the second process they go to the rolls where they are placed first, with the length of the slab corresponding in direction with that of their axis; secondly, with the length of slab across the diameter of the rolls, until it has been increased to the required breadth of the finished sheet; and, finally, by placing the original length of slab once more parallel to the axis, and extending the plate till it has been reduced to the requisite thickness. Sheet iron by the process of puddling, is, for some purposes, manufactured from pig metal into malleable iron, without the intervention of any other process of refining, than that which takes place in the puddling furnace itself. But for the boiler plate, it is believed to be customary, first to subject it to the action of the "run-out' refinery fire, in an open charcoal, or coke furnace urged by a powerful blast. As, in this fire, a large mass of metal is melted down at a time, and the cinder drawn off separately, the earthy impurities which in simple puddling would be retained in the balls, are at once removed, and by the aid of a small stream of water which is occasionally made to accompany the blast, a partial decarbonization of the metal is probably effected. When in full fusion, the metal is drawn off or "run out" into an oblong bed, and while still hot, is broken up into blocks of a few pounds weight each, to be conveyed to the puddling surface. In the latter it undergoes a second fusion, and the usual operation till agglutinated into "balls." minous coal is the fuel here employed, and the furnace is of the reverberatory form. The balls pass from this furnace first to the large hammer, by which they are moderately compacted; and immediately after to the rolls by which they are reduced into broad bars or slabs. The latter are reheated and at once rolled into plate, the former cut up into lengths of about 15 or 18 inches and piled, three high, to be reheated and welded into slabs of sufficient magnitude for plates of boiler iron. Bitu |