The Effect of Grinding on the Fatigue Strength of Steels

Abstract

The paper is a description of an investigation to determine the effects of cylindrical grinding on the fatigue strength of a range of steels: a nickel-chromium-molybdenum alloy steel hardened and tempered to give a nominal ultimate stress of 80 tons per sq. in., the same steel hardened and tempered to give a nominal ultimate stress of 62 tons per sq. in. (Vibrac V30), a 0.6 per cent carbon steel (45 tons per sq. in.) and a 0.1 per cent carbon steel (27 tons per sq. in.). Fatigue tests were carried out in a rotating-beam machine giving uniform bending moment, and the limiting fatigue ratios (endurance limit/nominal ultimate stress) of 0.538, 0.545, 0.420, and 0.550 were determined for turned and longitudinally-polished specimens of the Ni-Cr-Mo steel (80 tons per sq. in.), Vibrac V30, 0.6 per cent carbon and 0.1 per cent carbon steels. Preliminary tests on Vibrac V30 showed that there was a small reduction (about 4 per cent) in the limiting-fatigue ratio due to grinding. Further experimental work, however, showed that there was no intrinsic reduction in the fatigue strength for this material. Fatigue tests on ground specimens stress-relieved ‘in vacuo’ after grinding indicated that the reduction found in preliminary tests was not due to surface tensile stresses, and it is suggested that the reduction was owing to poor grinding technique, causing burning of part of the surface at the bottom of the 3-inch transition radius. Some specimens of Vibrac V30 were ground by two external laboratories and tested by the author. The fatigue strength of these two batches of specimens was not completely determined owing to the small number of specimens available. In one case the grinding was very uniformly carried out, but indicated a reduction in the fatigue strength of at least 8–10 per cent. The other batch of specimens also showed a reduction in fatigue strength, but no definite conclusions can be drawn owing to the small number of tests carried out and the large scatter of results obtained. There was no reduction in the fatigue strength due to grinding the 80 tons per sq. in. steel, Vibrac V30, and 0.6 per cent carbon steel, although a small reduction (about 3 1/2 per cent) was obtained with the 0.1 per cent carbon steel. This may possibly have been caused by undetected burning of the surface during grinding, as the same speed, feed, and wheel were used for all four materials. It is concluded that grinding does not cause a reduction in the fatigue strength of the materials tested, provided the grinding is carried out carefully and under good grinding conditions. The degree of care required to avoid deleterious effects, however, is such that it seems not advisable to advocate the grinding of fatigue specimens generally, unless preliminary fatigue tests have been carried out to determine the true state of the ground surface. Finally, the grinding of parts that are to be subjected to high fluctuating stresses should be avoided, owing to the uncertain state of the ‘production’ ground surface.