Understanding the criticality of volumetric defects on the fatigue behavior of additively manufactured aluminum alloys

Abstract

This study investigates the criticality of volumetric defects on the fatigue behavior of additively manufactured (AM) Al alloys. The role of various microstructures in determining sensitivity to defects, in particular their effect on the fatigue crack initiation and growth, is examined. Furthermore, the effectiveness of both destructive and non-destructive evaluation methods in characterizing the critical defect features for fatigue life prediction is assessed. The uniaxial fatigue behavior of two Al alloys, AlSi10Mg and Scalmalloy, each possessing distinct microstructures and thus likely to exhibit varying defect sensitivities, is investigated. It is shown that while applying post-process thermal treatments can enhance the tensile properties of AlSi10Mg by altering the microstructures, the fatigue properties are still largely governed by the presence of volumetric defects. In contrast, Scalmalloy, which possesses a bimodal grain structure, exhibits similar sensitivity to crack-initiating defects. However, microstructural effects become more apparent in the near-threshold regime. Specifically, the location of defects within the microstructure influences fatigue behavior in Scalmalloy, and microstructural features become increasingly influential approaching the crack growth threshold. Nevertheless, it is found that by considering both the short and long crack growth rates, the fatigue behavior can be sufficiently described within scatter bands of ±3 for both alloys. Moreover, non-destructive part qualification methods are explored and the efficacy of x-ray computed tomography (XCT) examination in capturing critical defect features is quantified. The results show that the probability of detection (POD) of defects is a function of the XCT scan parameters, coupon geometry and AM defect types. At a given scan resolution and defect size, lack-of-fusions (LoFs) exhibit a lower POD compared to spherical defects such as keyholes (KHs) and gas-entrapped pores (GEPs). While LoFs are more difficult to capture in lower resolutions, even when detected, the errors in their size measurements are greater compared to spherical defects which can be reliably detected with good accuracy in their sizing. A distance-criterion based correction is proposed, which allows critical information for LoFs to be recovered in low-resolution scans by reconstructing the defect geometry. It is demonstrated that the procedure can significantly reduce the sizing errors, allowing for rapid and accurate defect characterization. To assess the effectiveness of the distance-criterion approach for fatigue life predictions, AlSi10Mg and Scalmalloy fatigue specimens are examined by XCT prior to testing and the representations of the critical defects are identified in the XCT scans. The XCT data is post-processed by applying the distance-criterion based correction and incorporated into crack-growth based fatigue models. While the uncorrected data provides mostly non-conservative predictions due to the underestimation of defect sizes, the corrected data input into the fatigue models can provide accurate, moderately conservative fatigue life predictions within scatter bands of ±3.