Fatigue life estimation of notched aluminum sheet specimens subjected to periodic tensile overloads

Abstract

Structural elements are subjected to variable amplitude/spectrum fatigue loading where tensile overloads and compressive underloads may occur randomly. Depending on the magnitude of these loads, localized plastic deformation may occur at notches leading to residual stresses. These residual stresses have been known to significantly alter the fatigue life. In a spectrum, overloads may seldom occur in blocks and the effects of single overload cycles on the fatigue life must be investigated. In the present investigation, 2024-T3 clad aluminum sheet specimens with a gage width of 0.75” and circular hole diameter of 0.161” were subjected to constant amplitude fatigue loading with periodic overloads. The theoretical net stress concentration factor based on net section stress for each of the specimens is 2.48. Constant amplitude fatigue tests were conducted to determine SN curves at mean remote stress levels of 3ksi and 6ksi. Periodic tensile overloads of 44ksi, 35ksi and 20ksi were applied to constant amplitude loading blocks of 40ksi, 30ksi, 20ksi and 15ksi, each with a mean stress of 6ksi. The overload cycles reached the compressive load level prior to transitioning to the constant amplitude portion of the loading. The period between the overloads was varied at 10 cycles, 50 cycles, 100 cycles, 1,000 cycles and 5,000 cycles of constant amplitude loading. The experimental results showed an increase in the number of cycles to failure when the periodic tensile overloads were applied after every 100 cycles, 1,000 cycles and 5,000 cycles, and if frequent reverse yielding did not occur, as compared to the number of cycles to failure in the case of constant amplitude loading without overloads. In the analytical approach, Miner‟s rule was first used to estimate the cycles to failure. The study incorporated notch residual stresses into the Miner‟s rule in an effort to improve the life estimation. Neuber‟s rule was used to estimate residual stresses due to overloads. Two ideal conditions of strain hardening were considered – isotropic hardening and kinematic hardening – to establish a stress-strain curve for one cycle of loading. The results showed improved estimation of cycles to failure, when residual stresses were included in Miner‟s rule and if no reverse yielding occurred at the notch while fatigue loading. The loading that showed such a trend was maximum constant amplitude stress of 15ksi and a mean stress of 6ksi on which an overload of 20ksi was periodically applied. An improvement of 17% - 19% was obtained for a periodicity of 50 cycles or greater. In other cases of loading, if reverse yielding occurred at every cycle or frequently between cycles (periodicity of 10 cycles to 100 cycles) during fatigue then this caused detrimental damage and the specimen experienced a reduction in life.