Knowles Patent Steam Pumps

THE STANDARD.

Every Variety of Steam Pumping Machinery, viz.:

Knowles Steam Pump Works,

86 LIBERTY STREET, NEW YORK.

BLAKE'S IMPROVED STEAM PUMPS.

MORE THAN 13,000 IN USE.
Adapted to Every Situation.

SIMPLE,

POSITIVE,

COMPACT,

DURABLE.

Combined Pumps and Boilers for Railroad Water Stations a Specialty.

Improved Compound Duplex Pumping Engines for Water-Works.

GEO. F. BLAKE MANUFACTURING CO.,

june '81 ly

88 LIBERTY STREET, NEW YORK.

THE Franklin Institute is not responsible for the statements and opinions advanced by contributors to the JOURNAL.

EXPERIMENTS ON THE STRENGTH OF WROUGHT IRON AND STEEL AT HIGH TEMPERATURES.

By C. R. ROELKER, Passed Assistant Engineer U. S. Navy.

In the following paper an account will be given of experiments made by Dr. Julius Kollmann in 1877-78 on the strength of wrought iron and Bessemer steel at high temperatures, and the results obtained by him will be compared with those obtained in previous investigations of the same subject.

Kollmann's experiments, described by him in a prize essay crowned by the "Verein zur Beförderung des Gewerbfleisses," 1878, were carried on at the iron works of the "Gutehoffnungshütte," at Oberhausen, Rhenish Prussia. Besides numerous tests of the tensile strength of iron and steel at temperatures ranging between 70 and 2000 degrees Fahrenheit, Kollmann's experiments comprised investigations of the resistance of these materials to compression in the process of rolling them into bars and rails, and of various other questions of great practical importance in the manufacture of rolled iron. The present paper will be limited to a consideration of the tensile tests.

Two machines, a larger and a smaller one, were used in making the tensile tests, so that each might check any error in the results obtained WHOLE NO. VOL. CXII.-(THIRD SERIES, Vol. lxxxii.)

16

with the other one; both gave, however, closely agreeing results. They were always carefully examined and adjusted before the commencement of each experiment. In the smaller machine the strain was applied to the test piece by means of a hydraulic pump and was measured by weights on a scale beam. In the larger machine the strain was produced by the direct application of weights to a beam, the stretch being taken up by means of gearing, operating upon a screw connected with the lower holder of the test piece.

For the smaller machine the test pieces were made of round and square bars, turned or filed exactly to the following dimensions, viz.: 0.59 inch square, or 051 inch diameter in the body for a length of 9 inches, their total length being 11 inches.

For the larger machine the test pieces were cut out of broad flat iron, the strain being always applied in the direction of the fibre. They were made 1.57 inches wide by 0 39 inch thick for a length of 1:38 inches or 2:56 inches, the shorter specimens being used for the higher temperatures, because the stretch of the longer specimens could not have been taken up in the machine.

The test pieces for the smaller machine were heated throughout their whole length in a coke fire. Great care was taken to heat the specimens evenly throughout. A duplicate specimen, having exactly the same shape and dimensions, was heated each time in the same manner to the same degree as nearly as could be judged from photometric measurements by means of colored glasses. This duplicate specimen was used for determining the temperature of the test piece.

The larger test pieces were heated in a forge over a length of 1:38 inches or 2.56 inches, the ends being kept cool by wet coal. A piece of iron having the same dimensions as the heated portion of the test piece was heated at the same time to the same degree, as nearly as could be judged, and was used for determining the temperature of the test piece.

The elongation of the test pieces under a strain was measured by means of an instrument called by Kollmann a "multiplicator." It consisted of two levers connected, like a pair of tongs, by a bolt which served as a common fulcrum. The proportion of the short and long arms of each lever was as one to ten. The ends of the short arms were provided with jaws fitted with set-screws, by means of which the instrument could be attached quickly to the heated test piece after it was secured in the testing machine. One of the long arms carried at

its extreme end an arc on which a scale was marked, and which passed through an opening in the end of the other long arm. When the scale of the arc marked zero, the centres of the jaws of the two short arms were 1.38 inches apart, and any increase of this distance, magnified tenfold, was easily read off on the scale.

The temperature of the test pieces was determined by means of a calorimeter, consisting of a well-insulated copper cylinder containing exactly two litres of water and provided with a cover to prevent loss of water by vaporization. The duplicate test piece, heated as described, was dropped into the calorimeter at the exact moment when the strain was applied to the test-piece. The water was stirred till the temperature of the whole mass and of the test piece was equalized, and the rise of temperature was measured by a carefully tested thermometer. The calorimetric values of the copper cylinder, of the stirrer and of the thermometer had been determined and reduced to an equivalent weight of water. The loss of heat by radiation from the calorimeter was found, by experiment, to be too insignificant to affect sensibly the results obtained. To find the temperature of the test pieces from the rise in temperature of the calorimeter a set of tables were used, which were published in the "Zeitschrift Deutscher Ingenieure," 1875. These tables are based on the following formula for the mean specific heat, Cm, of wrought iron between the temperatures T and t in degrees of the centigrade thermometer, viz.:

C=0·1959074-0·00003269[T+t]+000000001108[T2+€ ÷ (T+t)2] The temperature of the test piece at the commencement of the experiment being thus found, it became necessary to determine how much it cooled off during the experiment. For this purpose numerous experiments were made with the two forms of specimens used in the testing machines. Their initial temperature and final temperature after exposure to the air for a known length of time were measured; the results thus obtained being plotted by laying down the lengths of the time of exposure as abscissæ and the corresponding temperatures as ordinates, a curve was obtained for each of the two forms of specimens, representing the law of cooling on exposure to the air. From these curves the final and mean temperatures of the specimens during the tests were deduced. In order to verify the results thus obtained, the final temperature was directly determined in a number of cases by dropping one of the broken parts of the test piece into the calori

meter.

The mean temperatures found in the above described manner were entered as the temperatures at which rupture took place. The rapid fall of the limit of elasticity and the great reduction of the cross area of the test pieces at high temperatures made it necessary to reduce the differences between the initial and final temperatures as much as possible, and for this reason the tests were made as rapidly as possible: the average duration of each test was but a little more than half a minute.

No allowance was made for any difference in the specific heats of wrought iron and Bessemer steel; the resulting errors in the determinations of temperatures, however, could be but trifling.

Three kinds of iron were tested, viz.: Fibrous wrought iron, finegrained wrought iron, and Bessemer steel. All the test pieces were taken from the ordinary stock manufactured at the works of the "Gutehoffnungshütte," and not from specially selected material. A chemical analysis of the specimens gave the following mean results, viz.:

The fracture of the fibrous wrought iron exhibited, as indicated by the name, a long fibre, but with traces of a coarse granular structure. The fracture of the fine-grained iron was rather fibrous than granular in appearance in consequence of the great reduction of the bars in the rolls. The Bessemer steel had a very even texture and a rather light shade of color.

The square and round iron bars for the small test pieces were rolled from single-piled hammered blooms. The broad flat bars for the large test pieces were rolled from double-piled hammered blooms. The steel bars were twice heated in rolling.