Abstract
The paper points out that in the use of temperatures above normal, characteristics of materials must be carefully studied; failure in such cases generally has been due not to the quality of the metal or alloy but to improper application. The testing of metals at high temperatures therefore becomes a very fruitful field for research. Experiments of this character date back to 1820, and extensive literature on the subject is already available, although the modern line of investigation dates back only to 1912. Early methods consisted in heating a specimen in a furnace and then transferring it quickly to a testing machine. Such methods were necessarily crude. Also temperatures in early experiments were taken at points distant from the test specimen, thus introducing further error.
In the past ten years, methods of furnace construction, location of the thermocouple, calibration of the pyrometer, and general checking up of the work, have led to much greater accuracy. In work by French, the electric furnace was incorporated in the testing machine itself, thermal equilibrium was very carefully obtained, and temperature determinations were made under actual test conditions, by placing thermocouples in holes located at various points in the specimen, carrying the entire auxiliary apparatus. Bregowsky and Spring extended their studies to determine variation of temperature of test specimens at several internal and external points, using specimens drilled axially. Priester and Harder gradually heated specimens to the desired temperature and held the temperature steady for thirty minutes to obtain thermal equilibrium. Spooner noted that with addition of nickel, chromium, and tungsten the color of fractured bars changed considerably. Other investigators, both in America and abroad, proceeded on similar, constantly more refined lines. In experiments conducted by the author himself, interesting results were obtained showing variation both from point to point along the surface and from surface to center. Jeffries and Sykes extended the range of the work to copper, tungsten, Armco iron, nickel, and molybdenum wires, with appropriate variations in methods of test, sometimes using boiling water, and sometimes crisco, which can be heated to 480 deg. fahr. without volatilization. In impact tests, Guillet and Revillon used representative methods, heating specimens slightly above the temperature required in an electric furnace, placing them on the anvil, and measuring temperatures a short distance from the cross-section to be fractured. Methods similar to those in tensile tests have been applied to torsion experiments, but very little has been done in alternating stress tests. X-ray tests and “long-time” tests are in process of development.
In conclusion, the author points out that the effect upon the physical properties of metals of raising the temperature cannot yet be stated in terms of a definite law, although it may be generally assumed that the tensile strength and elastic limit of steel decrease, and elongation and reduction of area increase, as the temperature is raised. Disagreement between tests on steel and results with alloys in practice are explained on the ground that alterations in physical properties of metals and alloys due to variations in temperature are not always of the same nature. In particular, the author stresses the importance of directing research toward such tests as approximate working conditions rather than allowing it to retain an academic character.
