Characterisation of fatigue crack tip field in the presence of significant plasticity

Abstract

Characterisation of a fatigue crack tip in the presence of significant plasticity has been challenging due to the lack of suitable tools and lack of knowledge of material constitutive information under cyclic loading. In this paper, Digital Image Correlation (DIC) and integrated finite element (FE) analyses have been used to characterise the crack-tip field beyond the small-scale yielding (SSY) regime in a stainless steel 316L of a compact-tension (CT) specimen under mode I loading conditions. The non-linear characteristics of the near-tip deformation field were verified by the poor fit to the William's regression and the overestimation of the stress intensity factor K. The extent of the crack tip plasticity was estimated using a detailed constitutive material model and compared with the estimated by Irwin. The displacement fields local to a stationary fatigue crack were mapped using DIC, and inputted into the FE model as boundary conditions so that an integrated FE analysis was carried out. Fatigue pre-cracking was simulated in the FE analysis prior to the full-field analysis of the fatigue crack tip, including stress/strain distributions ahead of the crack tip and the crack opening displacement (COD) under selected loading conditions. Although a distinct "knee" was captured from the compliance curves in both the DIC measurements and the FE analyses, consistent with the existing knowledge on the phenomenon of crack closure, it does not appear to correlate with the crack driving force measured by the J-integral.

3D predictions of the local effective stress intensity factor as the fatigue crack propagation driving force

“…At the final iteration, six nodes near the center are still opened (Figure 11b) meaning that they are not influenced by closure. This decrease has also been observed in the 2D study done by Lin et al [50] on 316L where the digital image correlation was used to measure the local displacement ahead of the crack tip: the closure was more pronounced in ΔK = 15MPa√m, R=0.1 than in ΔK = 20MPa√m, R=0.1 which is in turn greater than in ΔK = 25MPa√m, R=0.1. Furthermore, during his study, Vor [36] monitored that crack closure measured, globally, in specimens subjected to constant cyclic loading, is the same for different values of applied ΔK with the same loading ratio and tested that for ΔK = 12, 15, and 18 MPa√m at R=0.1.…”

The application of digital image correlation (DIC) in fatigue experimentation: A review

“…Lin et al 135 employed a commercial HR-2D-DIC system combined with FE modeling to investigate the effects of significant crack tip plasticity on the crack opening and crack driving force. Attempts to map the deformation F I G U R E 3 6 Full-field digital image correlation (DIC) strains on the gauge section of a tension-tension dog bone corrosion fatigue specimen at crack length 10 mm.…”

“…102 [Colour figure can be viewed at wileyonlinelibrary. com] F I G U R E 3 8 (A) Meshed computed tomography (CT) specimen for use in a finite element (FE) model to investigate plasticity beyond small-scale yielding (SSY) and (B) nearcrack-tip region where digital image correlation (DIC) displacements could not be evaluated surrounded by region where DIC displacements could be evaluated and applied to FE model nodes 135 [Colour figure can be viewed at wileyonlinelibrary.com] resulting DIC displacement data (away from the crack) lied on a regular grid. To identify the crack path and crack tip, a relatively small subset size was used that allowed segmentation of the crack based on detecting those subsets with abnormally high displacement gradients and/or a low correlation coefficient.…”

Determination of effective stress intensity factors under mixed‐mode from digital image correlation fields in presence of contact stresses and plasticity

“…The crack tip and path have to be determined beforehand using another approach. The displacement is extracted on a closed contour (containing the crack) on the DIC field, as shown in figure 3, and linearly interpolated (without any noise filtering) and imposed on the nodes along the contour of the FEM model, similarly to what is done in [29] or [30] (no interpolation in the later). The displacement field obtained by solving the FEM problem is then compared with the DIC field.…”