Issue 48

R. Baptista et alii, Frattura ed Integrità Strutturale, 48 (2019) 257-268; DOI: 10.3221/IGF-ESIS.48.27 260 Hz. A second derivative based Digital Image Correlation method (DIC) was used to measure the crack length evolution throughout the fatigue test (Lavision Imager Pro X camera system; speckle pattern produced with spray paint). Collected the crack length data, the Paris law was determine based on the seven point incremental polynomial technique presented in ASTM E 647 [8] (Fig. 2). Algorithm for fatigue crack propagation Previous work by the authors [9] showed that is possible to use closed FEM software’s to calculate fatigue crack growth paths on complex geometries. Baptista et al. [9] have demonstrated that is possible to use ABAQUS XFEM module to calculate the Stress Intensity Factor (SIF) and automatically predict crack propagation direction and incrementation. Unfortunately, this commercial solution works as a black-box, therefore it is not possible to control the necessary parameters of crack propagation simulation procedure. In ABAQUS the user cannot input the experimentally determined Paris Law parameters. These must be converted, as ABAQUS uses the Paris Law in the Strain Energy Release Rate form. The process is also very computer intensive, and the final results can still be fine-tuned. In order to solve the previously reported problems, several authors have developed alternative solutions, in order to calculate the fatigue crack growth path and fatigue life automatically. Dhondt [10] developed a solution using only freeware software. The algorithm used is common to other authors. First it is necessary to calculate the linear elastics fracture mechanics parameters for the different crack opening modes, for an initial crack. Next the Paris Law [11] is used to determine the crack propagation rate, and to predict the next crack increment. Finally, the crack can be updated, and the process is repeated. The algorithm developed by Dhondt [10] must therefore remesh the part in every increment. This method is more accurate but is also more computer intensive. As an alternative, other solutions have been developed. Rabold et al. [12] have developed a solution using ABAQUS and Python. ProCrack [12], uses a global mesh with a XFEM crack to calculate the crack displacements for each increment, and a sub-model of the crack front to calculate the linear elastics fracture mechanics parameters using the contour integral technique. Therefore, it is possible to model complex geometries, but there is no need for complex meshing procedures around the crack front, as only a simple material tube is modeled in the sub-model. Shi et al. [13], have solve the problem using only XFEM. There is no need to remesh the part, but in order to achieve the necessary quality for the SIF solution, the mesh in the crack propagation area must use an element size of as low as 0.3 mm. The obtained results from Shi et al. [13] were satisfactory, but the process is also very computer intensive [13]. Figure 3 : Flowchart of the algorithm used for fatigue crack growth simulation. To avoid the problems encountered, the new algorithm developed in this paper does not use sub-models or constant meshes with high element densities. As proposed by Dhondt [10], requires part remeshing in each increment, but is not