International Journal of Fracture最新文献

A novel failure criterion, named Strain-based Finite Fracture Mechanics, is proposed to predict the fatigue life of additively manufactured notched components under uniaxial loading conditions. The model relies on the simultaneous fulfillment of two conditions: a non-local strain requirement and the discrete energy balance. The inputs of the model are strain and the stress intensity factor at failure, which depend on the number of cycles according to power law equations. The inputs can be obtained based on strain-life and stress intensity factor-life data from plain and notched specimens. The present approach is comprehensively validated against experimental datasets on additively manufactured samples from the literature for different materials, raster angles, notch geometries and loading conditions. Predictions by other approaches, such as Finite Fracture Mechanics (in its original stress formulation) and the Theory of Critical Distances, are also considered, for the sake of completeness. Results show that, in general, the proposed strain-based model is more accurate and provides consistently precise predictions across different cases.

Modern phenomenological damage models use Lode parameter L and triaxiality (eta ) to describe the stress state of an isotropic material. Value pairs in the region between (L, eta = (0, 0)) and (L, eta = (0, frac{1}{sqrt{3}})) in plane stress condition can lead to ambiguous descriptions of the deformation. The case of simple shear is not defined separately. By using the difference in angles between the principal strain and principal stress axes, cases of coaxial stretch superposed with simple shear can be distinguished from cases of coaxial stretch without simple shear. In the case of anisotropic material or large elements, the distinction between these ambiguous cases can be utilized to optimize failure models. This study proposes a method to recover the deformation gradient and shear direction for proportional and non-proportional loading with an elastoplastic von Mises material. The deformation gradient is suitable for distinguishing stress states with simple shear from stress states without simple shear in plane stress condition.

Sharp kinks may be observed under shear loading or in materials containing weak directions, such as those produced by additive manufacturing. A better understanding of the fracture of these materials, both theoretically and experimentally, is required to deploy them in structural applications. This study focuses on the measurement of stress intensity factors (SIFs) around a sharp kink using digital image correlation (DIC). The performances of two DIC-based techniques, namely, integrated-DIC and post-processing of DIC-measured displacement fields, are assessed on a benchmark test using fused deposit modeling capabilities, and are compared to a reference finite element solution. It is shown that Williams’ expansion remains valid on a large enough region around the crack to extract reliable SIFs even very close to the crack kink. Both techniques are very trustworthy, provided the SIF identification zone is carefully defined to exclude the kink zone of influence.

This study proposes a fatigue life prediction method combining small-sample data expansion with the Weibull distribution function, incorporating the first order reliability factor (FOSM) to improve accuracy. Using Generalized Polynomial Chaos Expansion (GPC) and Latin Hypercube Sampling (LHS), small-sample fatigue data is expanded, followed by enhancing the two-parameter Weibull model with FOSM. Results show the generalized polynomial chaotic expansion method and Latin hypercube sampling are used to obtain the probability density curve when the stress level is 350 MPa, and the original data are all on this probability density curve, indicating that the expansion method is more credible. High prediction precision within a 1.5 × error range, with logarithmic safety life linearly related to stress level and decreasing with higher failure probability.

The rapid development of 3D printing of 316L stainless steel thin-walled structures obtained by direct energy deposition has generated an increased interest in the mechanical properties of such materials for use in applications; in particular, failure models are needed to ensure structural reliability. We consider the response of uniaxial tension specimens, with and without notches, to characterize the constitutive and failure behavior of the material. Specifically, we use numerical simulations of the notched tension experiment, achieved with a simple power-law strain hardening model and a failure criterion based on attaining a triaxiality-dependent critical strain-to-failure, to demonstrate that this model is capable of reproducing the material behavior accurately.

A particular failure mode of highly porous brittle materials consists in the propagation of cracks under uniaxial compressive loads. Such ’anticracks’ have been observed in a range of materials, from snow and porous sandstone to brittle foams. Here we present a computational model for the formation and propagation of anticrack-type failure in porous materials within the general computational framework of bond-based peridynamics. Random porosity is represented, on a scale well above the characteristic pore size, by random bond deletion (dilution disorder). We apply our framework to experimental data on anticrack propagation in silicate foams.

We introduce the fast convolution-based method (FCBM) for a peridynamic correspondence (cPD) model to simulate finite plastic deformations and ductile fracture in 3D. The cPD model allows the direct use of classical finite plasticity constitutive ductile failure models, like the Johnson–Cook (J-C) model used here. We validate the FCBM for the cPD model against experimental results from the literature on ductile failure in Al2021-351 alloy samples of various geometries. Notably, calibration of elastic and hardening material parameters is made only using the experimental data from the simplest geometry, a smooth round bar, and only up to the necking point. We then use that calibrated model beyond necking, through full failure, and for all the different sample geometries. The performance (speedup and memory allocation) of the new method is compared versus the meshfree method normally used to discretize PD models for fracture and damage. The proposed method leads to efficient large-scale peridynamic simulations of finite plastic deformations and ductile failure that are closer to experimental measurements in terms of displacement and plastic strain at failure than previous FEM-based solutions from the literature.

The fracture surfaces of 49 SE(B) toughness tests performed on five different geometries, were carefully investigated by SEM imaging and cross-section analysis. The specimens were extracted from a large multi-pass weld in T-S orientation. The failure characteristics were associated with three distinctly different zones of the weld. Transgranular fracture occurred primarily in the reheated zone and in the as-welded zone with a dendritic microstructure inclined relative to the crack plane. With a dendritic microstructure aligned with the crack plane intergranular fracture occurred. The toughness of the as-welded zone with a low inclination angle, was significantly lower than that obtained in the other two weld zones. Due to the relatively large size of the zones compared to the fracture process zones of the tests, it is appropriate to characterize the failure behavior as large-scale heterogeneity. Weakest-link modeling may be applied locally in each weld zone, giving rise to three different sets of model parameters. A new calibration technique is introduced and used to fit a local weakest-link model to the toughness distribution curves of the individual zones.