No abstract

Mild steel and an aluminium-zinc-magnesium alloy were tested axially under cycles of constant load range and constant total strain range at frequencies between 5 and 15 c/min for endurances from monotonic tension to 50 000 cycles. Mean loads or strains of zero and half the maximum in tension were studied. Hysteresis loops were recorded automatically at regular intervals throughout the life to failure. Two mechanisms of fracture were found for mild steel during load cycling, a necked-out type caused by cyclic ‘creep’, and conventional fatigue. The aluminium alloy did not exhibit this behaviour and could not in fact be broken in repeated tension at endurances below 600 cycles. For strain cycling a linear relationship was found to exist when the logarithm of the total or plastic strain range was plotted against the logarithm of the number of cycles, for each metal and mean strain. Load and strain cycling results cannot be compared on the basis of the setting-up cycle (that is, the conditions for the first cycle), but only in the steady cyclic condition after work hardening or softening has occurred. During a test each of the metals worked towards a steady cyclic condition. A smooth curve is obtained when the total energy to fracture is related to the endurance, for all cases of load and strain cycling studied.

No abstract

Part 1 of this paper describes the uniaxial fatigue testing of injection-moulded cylinders made of acetal copolymer at various frequencies between 0.167 and 10 Hz. The influence of mean stress and cycle control (stress or strain) has also been studied. Two mechanisms of failure were observed: cyclic thermal softening, above certain limits of stress and frequency, and traditional crack propagation fatigue, which was found to be independent of cyclic frequency. Strain cycle control tests were thermally stable and stress range became stable after the initial 1 per cent of the total endurance. Creep rupture tests were also performed. Part 2 considers the fatigue and creep rupture of tubular specimens containing geometrical stress raisers. These were injection moulded under the following conditions: (a) single-end gated, (b) side gated, (c) double-end gated, (d) ‘rough’ mould and polished mould. Moulding defects such as weld lines, presence of voids and sprue, poor surface finish, etc., are compared for fatigue-strength reduction with machined circumferential notches and transverse holes of known theoretical stress concentration.

No abstract