(Paper No. 3582.) "Fracture of Axles Originating in Drilled Holes." By LUCIEN ALPHONSE LEGROS, Assoc. M. Inst. C.E. It is well known that in many materials a drilled hole will prevent the spreading of a crack, and the extension of a crack in a casting or forging is frequently stopped by this means; but the fact that in some cases a drilled hole may cause the commencement of a crack does not appear to be so generally appreciated. Cracks may be expected to appear in a material subject to alternating stresses, at places where there is an abrupt change of section. or where a notch or groove has been made by a cutting-tool in a turned surface. Locomotive-engineers have for many years past recognized the importance of these details and their influence on the life of both cranked and plain axles. In 1897 the Author examined some tramway-car axles which had been fractured through the centres of shallow drilled holes. Photographs of two of these fractures are shown in Figs. 1 and 2, Plate 4; the axle shown in Fig. 1 failed while at work, after running 15,380 miles, whereas that shown in Fig. 2 was removed from service after running 13,900 miles and was subsequently broken, the crack having been detected before fracture took place. The steel from which these axles were forged was made to British Railways standard specification, the mean axle-load was 5 tons and the diameter of the axle at the fracture was in each case 3 inches; the cracks occurred several inches from the key-seat and at points where the stress would not be the maximum. The drilled hole was made to receive the point of a set-screw to prevent endwise movement of a chain-wheel. After these failures the set-screws were replaced by pairs of clamp-plates gripping the axle and secured to each other by bolts. In 1902 the Author witnessed a failure of an axle in service on a suburban passenger-train near London, and on examining the fracture found it to be similar in character to those shown in Plate 4, the crack having originated in a drilled hole for a set-screw to prevent movement of part of the driving-gear for the electric-lighting plant carried on the axle. This failure on a passenger-carriage of one of the most important British railways has led the Author to believe that the danger from fractures originating in drilled holes is not fully appreciated. In addition to the foregoing cases the Author has also met with failures which have originated in an oil-hole drilled in a shaft at the centre of a journal to meet an axial hole conveying lubricant from the end of the shaft. With some forms of forced lubrication such drilled holes occur in crank-pins and journals, and may readily produce cracks of the type indicated, although probably the risk could be greatly reduced by rounding off the top of the hole so that no sharp edge be left. In conclusion, the Author desires to direct attention to the risk incurred by drilling holes in axles and shafts, and to emphasize the importance of specially examining any shafts or axles so drilled with a view to detecting any signs of fracture. The Paper is accompanied by two photographs from which Plate 4 has been prepared. "A (Paper No. 3585.) Simplified Method of laying out Transition-Curves." By THOMAS ALEXANDER ROSS, Assoc. M. Inst. C.E. IN countries where the engineer is entirely dependent on native chainmen, the ordinary routine-work of laying out transition-curves is considerably increased by the difficulty of explaining to the men the odd feet and inches into which the chain may have to be divided at the half, quarter and three-quarter lengths. In such countries, therefore, it is desirable that the "length" of the transition-curve should be represented by a whole number readily divisible into the fractions desired. The length of the transition-curve is based upon the adopted superelevation of the outer rail, for various degrees of curvature, and on the rate at which that superelevation is run out. Both of these factors depend upon the probable speed of the trains, but as this varies between wide limits an approximate average speed has been adopted. For use on railways in China, where the ordinary speeds for many years to come are likely to range between 15 and 45 miles per hour, the Author has worked out the Transition-Curve Table given in the Appendix, which gives values of L, the "length" of the transition-curve, in exact multiples of 10 feet for curvatures expressed in even multiples of quarter-degrees. The curve adopted is the cubic parabola, and the Table has been compiled for an average train-speed of about 30 miles per hour, with superelevation GM2 calculable by the usual formula, E = 1.25 R' inch in 60 feet. run out at the rate of In practical field-work, however, it is generally sufficient, at first, to mark only the beginning of the transition-curve, its middle point, and the point where it joins the circular curve, i.e., the point of osculation. These can be found in the following manner:— -With the instrument at the apex-point, lay off the B.C. and E.C. (beginning and end of curve) for a circular curve in the usual manner (Fig. 1). A further distance equal to L measured out on each tangent gives the P.T.C. (point of transition-curve) and E.T.C. (end of transition-curve). At B.C. set off at right-angles to the tangent a distance b, equal to the ordinate at L, or to half the "shift." This gives the middle point of the transition-curve; continue this offset a further distance equal to b. This will give P.C. (the point from which the circular curve can be set out). With the instrument at this point, and a backsight parallel to the original line, lay off the circular curve, the first deflection being worked out for L. This will give P.O. (the point of osculation), or the point at which the Fig. 1. transition-curve and the circular curve join. The three pegs thus located are all that is required until the formation is finished, when reference-points (generally stones having crosses cut on their tops) can be laid out by ordinates or deflections at the quarter and threequarter lengths. It will be seen that the length in feet of the transition-curve given in the Table is equal to the deflection of the curve in degrees multiplied by 40. This length can therefore be obtained by a simple mental computation, at any time, without reference to the Table. The Paper is accompanied by a diagram, from which the Figure in the text has been prepared; and by the following Appendix. [APPENDIX. |