PSI - Issue 13

Letícia dos Santos Pereira et al. / Procedia Structural Integrity 13 (2018) 1985–1992 Author name / Structural Integrity Procedia 00 (2018) 000–000

1989

5

with 0.1 and 2.8 tons respectively for the Charpy and DWTT. The impact speeds were 4.85 m/s for both cases. MSC Patran 2013 was used as pre-processor and Abaqus CAE 6.13 was used as processor and post-processor. The simulations employed 3D hexahedral elements of reduced integration and linear interpolation with a greater refinement of the mesh in the contact regions and where the crack propagation would occur ( Fig. 3(a) ). In average, elements in the remaining ligament are 0.25 x 0.50 x 1.0 mm.

(a) (b) Fig. 3. (a) GTN model, mesh and symmetry for the DWTT specimen; (b) illustrative stress fields and studied domains.

All the GTN parameters were calibrated using as a reference the recommended parameters showed in Tab. 1 . The calibration was made changing one parameter at a time until the experimental load-displacement curves were well reproduced. Once calibrated, the same parameters (Tab. 2) were employed for both Charpy and DWTT specimens. The same was conducted for XFEM, however, the parameters were not the same for the two geometries – for both specimens the maximum cohesive stress was 1100 MPa, but cohesive energy was, respectively, 9 and 5,8 N/mm 2 for Charpy and DWTT. The difference comes from the different stress triaxiality found in such specimens - the higher the triaxiality, the smaller is the cohesive energy. Table 2. Calibrated parameters for GTN model (mm) 0.00015 0.0015 0.3 0.1 0,02 4,0 1,5 1,0 2,25 0.25 3.3. Energy separation The basis for the energy separation methodology employed in this paper is to isolate domains where the stress field deviates from the hypothetical bending field that would occur if there were no hammer, support and propagating crack. The idea is that the energy absorbed by elements within this specific domain is associated with the different processes occurring in the specimen, such as deformation, crack initiation and crack propagation, as discussed previously. The criteria for determining such domains are stress-based (evaluating von Mises equivalent stresses) as explained earlier and Fig. 3(b) illustrates some domains considered for one DWTT geometry – in such domains, energies associated with crack initiation, crack propagation, contacts and specimen’s deformation can be computed. Essentially similar results in terms of domains’ geometries were found for Charpy geometry. All additional details can be found in the works of Moço (2017) and Pereira (2017). 4. Results 4.1. Stress state analyses The stress fields found in fracture specimens (e.g.: Charpy and DWTT) and real gas pipelines strongly differ. While in pipelines tensile membrane stresses prevail, in both specimens studied here a 3-point bending loading takes place. Even considering that equivalent stresses and strains can be easily computed, such results do not completely support