Fatigue Crack Propagation in Metals

Abstract

PEARSON¹ has recently discussed the problem of the rate of growth of a fatigue crack. He tabulated values of the stress intensity factor K necessary to cause a crack to grow under a repeated tensile loading cycle at a certain velocity in a number of alloys and showed that the ratio of K to the Young's modulus E of the material was roughly constant. The fact that the fatigue crack growth rate characteristics bear no relation to any mechanical property other than E has been demonstrated previously by Frost^2,3, who pointed out, however, that any such relationship applies only to pure metals or to alloys having the same growth rate characteristics as the base metal and in which the growth rate is independent of the tensile mean stress in the loading cycle. For example, although the growth rate characteristics of annealed and cold-rolled copper are the same, those of the various aluminium alloys may differ widely^3,4. Complex alloys, such as the high-strength aluminium alloys, can have their crack growth characteristics affected markedly by microstructure; brittle second phase particles are known to provide easy fast fracture initiation sources. It is this introduction of elements of fast fracture into the growth process which causes the growth rate characteristics of the various aluminium alloys, having the same Young's modulus, to be different from each other and from the base metal and thus invalidates Pearson's generalization. Although the three copper-aluminium alloys tested by Pearson gave similar K/E values, a different value would certainly have been obtained for a zinc–aluminium alloy. To obtain similar K/E values for brass and aluminium alloy, for which he quotes similar K values, Pearson takes a value of E for brass similar to that for the aluminium alloy. Most authorities quote an elastic modulus of brass nearer to that of the titanium alloy (for which he quotes a K value different from that for brass) than to that of the aluminium alloy. It would seem, therefore, that the constancy of K/E obtained for the materials tested by Pearson must be considered fortuitous.