D.L. DuQuesnay

The effective stress range as a mean stress parameter

Abstract This investigation examines the hypothesis that variations in the crack opening stress level of short cracks can account for the observed variations in fatigue strength with mean stress under constant-amplitude cyclic loading. Support for the hypothesis is provided by the experimental data generated for a 2024-T351 aluminium alloy and a SAE 1045 steel. Based on the observed variations in crack opening stress with mean stress an effective stress range is postulated as a mean stress parameter. The effective stress range successfully reduces the fatigue data for all levels of mean stress to a unique fatigue-life curve for each material.

The effect of compressive underloads and tensile overloads on fatigue damage accumulation in SAE 1045 steel

Abstract Block histories containing either compressive underloads or tensile overloads followed by smaller fully reversed cycles were applied to smooth specimens to determine the load interaction effect of large stress cycles on the fatigue behaviour of subsequent smaller cycles. Interactive damage was obtained by subtracting the damage due to the overload and the steady state damage due to the small cycles from unity. As in previous tests on an aluminium alloy, the interactive damage per cycle decays as a power law function of the number of cycles following an underload or overload. The results indicate that small cycles, including those below the constant-amplitude fatigue limit, can contribute significantly to damage accumulation, and therefore small cycles should not be ignored when predicting fatigue lives for variable-amplitude histories.

Notch fatigue behaviour of SAE1045 steel

Abstract The effects of notch diameter, material heat treatment and applied stress ratio on the notch fatigue behaviour of an SAE1045 steel were investigated. The fatigue notch factor increases with notch diameter for sharp notches, but it decreases with notch diameter for blunt notches. The notch sensitivity of the quenched-tempered martensitic steels was higher than that of the as-received ferritic-pearlitic steel, especially for small notches. However, the notch sensitivity of the quenched-tempered materials was not significantly affected by the tempering temperature. The fatigue notch factor was found to be higher for R = −1 tests than for R = 0 tests for a given notch size in the as-received material. Short crack fracture mechanics was applied to predict the fatigue life for sharp and blunt notches. A reasonable agreement between the experimental and predicted data was observed.

The effect of compressive peak stress on fatigue behaviour

Abstract The effect of compressive peak stress on the maximum stress at the endurance limit, crack propagation rate, threshold stress intensity and crack closure was studied in a laboratory environment using two steels (SAE1045 and SAE1010) and two aluminium alloys (2024-T351 and 7075-T651). As the compressive peak stress, S cp, was increased in magnitude, the maximum stress at the endurance limit, S fa, decreased linearly. Compression-compression cycling did not initiate any cracks in a centre-notched SAE1010 steel specimen but initiated cracks, which gradually became non-propagating, in the notched 2024-T351 aluminium alloy specimens. In compression-tension tests, the crack propagation rate increased, and the threshold and the crack opening stress intensities decreased linearly with increasing compressive peak stress. During compression-compression cycling the load/displacement curves were not linear, indicating that the crack was not fully closed.

Notch-size effects in fatigue based on surface strain redistribution and crack closure

Abstract A fracture mechanics solution for the stress range at the fatigue limit and the depth of non-propagating cracks in smooth and notched specimens is presented. The solution is based on the non-uniformity of strains at the surface of a specimen and on the development of crack closure. Surface grains oriented for easy slip experience an inherent microstructurally dependent strain concentration, which decays with depth into the material at a rate that is inversely proportional to the grain size. For cracks in smooth specimens the threshold stress range is separated into a crack opening stress component and an intrinsic stress component. The latter is the stress range that is just sufficient to grow a fully open crack and it is estimated by considering the near-surface strain concentration for a variety of crack depths. The crack opening stress range is added to the intrinsic component to obtain the nominal threshold stress range as a function of crack depth. The maximum threshold stress range defines the fatigue limit of the specimen. For notched specimens a geometrical strain concentration is considered in addition to the inherent strain concentration. The predicted fatigue limits agree well with the observed values for a variety of notch sizes and stress ratios in the aluminium alloy 2024-T351. The proposed solution also offers a prediction for the depth of non-propagating cracks wa both smooth and notched specimens. In the latter, the depth of non-propagating cracks was found to be dependent on the notch radius and the stress level. The predictions of non-propagating crack depths agree with the experimental data for the aluminium alloy BSL65.

Fatigue damage accumulation in 2024-T351 aluminium subjected to periodic reversed overloads

Abstract The variation in damage accumulation rates which occur when periodic tension-compression and compression-tension overloads are introduced into an otherwise constant-amplitude loading history are examined in this investigation. Block loading histories consisting of either compression-tension or tension-compression overloads followed by smaller constant-amplitude fully reversed cycles were applied to smooth, axial, 2024-T351 aluminium specimens to determine the increase in damage accumulation rates for the fully reversed cycles following such overloads. The results show that the interactive damage per cycle remains constant for approximately 450 cycles following an overload and then it decays as a power law function of the number of cycles. The periodic application of a compression-tension or a tension-compression overload contributes significantly to damage accumulation during subsequent small fully reversed cycles, either above or below the constant-amplitude fatigue limit. Consequently, cycles with stress amplitudes below the constant-amplitude fatigue limit, which are usually omitted damage from calculations, should be accounted for in fatigue life prediction analyses for variable-amplitude histories.

Effects of compression and compressive overloads on the fatigue behavior of a 2024-T351 aluminum alloy and a SAE 1045 steel

Smooth cylindrical specimens of 2024-T351 aluminum alloy and SAE 1045 steel were tested under constant amplitude cycling to study the effect of compressive stress on fatigue life. The results show a reduction in the fatigue life at constant maximum stress as the compressive portion of the stress cycle is increased. Tests performed to study the effect of a periodic compressive overload on the order of the yield strength show an increasing reduction in fatigue life at constant maximum stress as the frequency of application of the overloads is increased. The compressive overload cycle significantly increases the damage done by subsequent smaller cycles, including those far below the constant amplitude fatigue limit. The results indicate that the present techniques used for damage summation for variable amplitude service histories may give grossly unconservative fatigue life predictions, especially for histories containing large compressive load cycles accompanied by a relatively large number of small cycles. It is suggested that conservative life predictions for such histories can be made using the stress-life curve from constant amplitude tests with compression on the order of the yield stress accompanying every cycle.

THE EFFECTIVE STRESS RANGE AS A FATIGUE DAMAGE PARAMETER

This paper shows that the changes in crack closure and the associated effective stress range induced by mean stresses and overloads are responsible for the accompanying changes in damage accumulation rates and fatigue limits. When crack closure effects are accounted for, fatigue damage can be calculated using a single effective stress range versus fatigue life curve. Experimental evidence to support the effective stress range as a fatigue damage parameter is provided for a 2024-T351 aluminum alloy and a CSA G40.21 structural steel.

MECHANISMS AND MECHANICS OF FATIGUE CRACK INITIATION AND GROWTH

ABSTRACT This paper reviews the mechanisms by which a fatigue crack grows and the metallurgical and loading variables that influence these mechanisms. Fatigue crack growth regimes for constant amplitude fatigue are characterized as a metallurgically short crack regime, a mechanically short crack regime and a long crack regime. In the metallurgically short crack regime the crack is short compared to metallurgical variables, such as the grain size, and growth is strongly affected by microstructure. In the second regime the crack is long enough that the resistance to growth by microstructural barriers is averaged out, but it is not long enough for crack closure due to the plastic wake of the crack to have reached a stable level. During this stage the crack grows faster than a long crack having a similar linear elastic stress intensity. In the final long crack regime crack closure has reached a steady state level and the crack growth rate can be characterized by the range of stress intensity and the stress ratio independent of crack length. The effect of stress or stress intensity level, mean stress or stress ratio, microstructure and crack closure on crack growth in each of these regimes is examined. The manner in which crack growth resistance to short cracks develops is shown to lead to short cracks in sharp notches which stop propagating at low stress levels, and to reduced fatigue concentration factors for small notches and defects. Work illustrating some of the ways in which variable amplitude loading changes crack closure, thresholds, and other characteristics of fatigue behaviour is reviewed. Particular attention is paid to closure changes that cause current fatigue analysis procedures to give unconservative life predictions.

Deformation and fatigue behaviour of a cold-rolled SAE 1010 steel

Abstract An SAE 1010 low-carbon steel was cold rolled to 22, 56 and 76% thickness reductions. Monotonic tensile tests, smooth and notched specimen fatigue tests and crack propagation tests were performed. The effect of loading direction on fatigue behaviour was examined. The monotonic and cyclic yield strengths increased and the ductility decreased as the degree of cold rolling increased. The increase in yield strenth and the decrease in ductility were more pronounced in the transverse direction than in the longitudinal direction. Under strain control, the fatigue life at high strain amplitudes was lower for loading in the transverse direction than for loading in the longitudinal direction. At low strain amplitudes the fatigue lives in both directions were approximately equal. The notched specimen fatigue strength was only slightly increased by cold rolling since two opposing factors, the smooth specimen fatigue strength and the notch sensitivity, were both increased by cold rolling. The threshold stress intensity decreased and the crack propagation rate, at a given stress intensity, increased as the degree of cold rolling increased.