Evaluation of linear elastic fracture mechanics predictions of one and two-dimensional fatigue crack growth at cold-expanded holes

This work assesses the ability of linear elastic fracture mechanics (LEFM) with superposition to correlate the growth of one and two-dimensional fatigue cracks at cold-expanded open holes under constant amplitude loading. Care is taken in the work to accurately: control the test setup to ensure growth of a single crack, determine residual stress in the coupons, measure crack growth, determine the fatigue crack growth rate (FCGR), compute stress intensity factors, and correlate fatigue crack growth rate with stress intensity factor range DeltaK and stress ratio R. The work used long dog-bone coupons made from 2.03 mm and 4.83 mm thick 7075-T6 sheet having a gage section 38.1 mm wide, a centrally located 7.09 mm diameter hole, with through thickness notches for thin and corner notches for thick coupons. Quarter-elliptical stress intensity factors are used for thick coupons while one-dimensional stress intensity factors are used for thin coupons. Residual stress due to cold expansion (CX) was measured using the contour method as a function of position on the crack plane and gave typical results. Measurements on multiple coupons showed +/-10% variability in residual stress. Crack growth behavior of multiple as-machined (AM) coupons (without CX) tested at R of 0.1 or 0.5 agreed with earlier results published in the literature for both one and two dimensional cracks. Crack growth behavior of multiple thin CX coupons tested at the same two applied stress ratios was consistent with predictions by linear superposition, with scatter in lifetime consistent with the observed 10% variability in residual stress. LEFM based on quarter elliptical models was found to be ineffective to predict crack growth in thick CX coupons because the quarter elliptical shape assumption is inaccurate. The work therefore demonstrates the ability of LEFM with superposition to accurately correlate the behavior of coupons with and without residual stresses for one-dimensional crack growth, and only for AM for two-dimensional crack growth.