DISCUSSION.

The President.

Mr. Job.

THE PRESIDENT.-It occurs to me to ask what influence the size of the ingot which has been characteristic of the last fifteen years, has on the quality of steel in rails. We all know that the larger the ingot, the greater the segregation, and we have many times queried, whether more successful rails would not be obtained, from the same grade of steel, if it were cast into much smaller ingots, than is now obtained with the customary large ingot. We recently examined twelve rails drawn from track which had failed within the first six months after they were put down new. The failures consisted in splitting at the ends, breaking down of the head in certain parts of the rail, and in some cases a piece of the head three or four feet long breaking off. In the case of every one of these twelve rails, an analysis showed serious segregation. We were forced to the conclusion that in these rails, an attempt had been made to utilize the whole ingot in making rails, and we are confident this is a much more serious cause of failure than is generally believed.

MR. ROBERT JOB.-We have had instances of split rails such as described by the President, and have often found evidence of segregation, but in every case of such failure which we have investigated the steel proved to be radically unsound, and thus unwelded, and hence permitted crushing under relatively light pressure.

We have also examined rails which have been as badly segregated as those just mentioned, and in spite of it have given good service, but in these cases the steel was relatively free from unsoundness, particularly within one-half inch of the bearing surface. Hence it seems clearly indicated that lack of soundness is the main condition to be guarded against; it is the bête noirfrom the consumers' standpoint of present practice. Without it there would probably be few, if any, split rails, no slivered or flowed over rails, and no "soft spots," or mashing down. Moreover, with unsoundness removed, segregation to an injurious extent would be almost impossible.

RAIL SECTIONS AS ENGINEERING STRUCTURES.

By P. H. DUDLEY.

The mechanical properties, of stiffness and strength of a section, increase in a rapid ratio as the height is augmented, as shown by Table No. 1.

The 80- and 100-pound sections become more efficient engineering structures than the 60- and 65-pound which they replaced, by inducing a longer distribution of the passing wheel loads to the cross-ties and ballast, and lessening the deflection under the wheels. A greater portion of the wheel effects is absorbed by the constraining negative bending moments in the wheel spacing, reducing the positive moments under the wheels. This favorable action for a smoother running surface in the general depression under the wheels, however, increases the wheel-contact pressure intensity per square inch in the bearing surface of the section, and imposes a greater burden upon the metal of the entire head.

The stiff sections for a given unit-fiber stress carry larger bending moments than the limber rails and therefore are more efficient engineering structures for heavier axle and total loads.

In the 80- and 100-pound sections nearly the given maximum moments have been obtained in tests.

Unit-fiber stresses in tension from 0 to 30,000 pounds are those which occur daily under present locomotives, with some stresses of the higher figures and liable from a flat wheel to be exceeded.

Large unit-fiber stresses have been common since steel has been used for rails, for the limber sections had frequent sets, which indicated they were stressed beyond their elastic limits.

Under 65-pound rails in a yard track, a unit-fiber stress has been measured of 56,000 pounds.

The elongation for the unit-fiber stresses in Table No. 2, would be 0.00034, 0.00067, 0.001, and 0.00134 in. respectively, for the modulus of elasticity at 30,000,000 pounds, summer temperatures, though it is higher for the winter.

TABLE NO. 2.

POSITIVE BENDING MOMENTS IN INCH-POUNDS FOR THE GIVEN UNITFIBER STRESSES PER SECTION IN TABLE No. I.

The height of the section is increased to augment the mechanical properties, of stiffness and strength. This has a tendency to reduce the unit-fiber stresses in the base of the rails, but the subsidence of the road-bed under the wheel loads, the looseness of the cross-ties in the ballast, and the rails under the spikes, cause the bending moments carried by the stiff rails, to exceed those by the limber sections for the same wheel loads, though the percentage distributed per individual cross-tie is less. This increases the intensity of the wheel-contact pressures in the bearing surface. The wheel loads also have been doubled on the stiff rails, over what they were on the light sections, and the metal in the bearing surface therefore sustains two and three times the burden required by the former limber sections. See Vols. III and IV of the Proceedings, for Unit Fiber Stresses and Bending Moments.

To carry the stresses and distribute the wheel loads, the rail section has the top of the head or bearing surface shaped for the wheel treads, to receive their pressures and distribute the loads to the cross-ties, ballast, and road bed. The side of the head of

the rail is the guide for the passing wheel flanges, while the entire section becomes the girder to distribute the wheel loads.

The metal in the rail section has three functions to perform to receive, support, guide and distribute the wheel loads of the passing trains. 1. The metal in the bearing surface sustains its loads principally by its properties of cubic elasticity, and should be sound and homogeneous. 2. The side of the head resists abrasion of the wheel flanges by its toughness and tenacity. 3. To distribute the loads by the entire section its linear elasticity is exercised.

The factor of safety in the girder determines what physical properties may be used for the bearing surface and guide.

The metal in the head of the rail must receive and sustain the wheel contact pressures by its cubic elasticity, and distributes the loads through the section as a girder, by its linear elasticity.

The distortion of the rail head in service shows that the steel has low limits of cubic elasticity, and is not homogeneous. (See Figs. 1 and 2.)*

When the steel is solid, sound and of fine texture, the head does not become distorted under the service, though it wears in the bearing surface and on the side. When the ingot is unsound and spongy, then the head flattens and crushes under the wheel treads. Rails from the top of the ingots, where by liquation the upper portion contains a higher percentage of carbon and phosphorus, the central core of metal is not sound and strong, but fragile, and does not sustain the wheel contact pressures as well as the exterior portion of the section. (See Fig. 1.)

When a decided pipe in the ingot did not occur in cooling, the repeated pressures of the wheel contacts develop a check which is equivalent to a pipe, the metal immediately over it in the bearing surface stretching sidewise by its linear elasticity, the check widens until a portion of the head becomes detached from the web of the rail, unless removed from the track. (See Fig. 2.) The steel in the head is not homogeneous, either in quality or structure, and becomes distorted as a section, from inadequate physical properties of cubic or elasticity of volume, to sustain and distribute the wheel loads.

*Acknowledgment is made to the Iron Trade Review for the cuts used in this paper.—ED.

The splitting of the head in the earlier steel rails was in nearly all cases traced directly to a pipe in the ingot. These conditions still exist, yet there are numerous instances in which the pipe did not develop in cooling, but does in service, in the unsound metal of the central core of steel, as indicated in Figs. 1 and 2. Pieces break from the side of the head, in steel where so decided liquation

FIG. 1.-Section from Upper Portion of Ingot Showing Central Core and

Pipe.

FIG. 2. Check Developed in Rail-head by Service, 12 Feet in Length,

Sound at Ends.

has occurred in the ingot as to make two or more grades of steel in the head.