Issue 35

G. Gobbi et alii, Frattura ed Integrità Strutturale, 35 (2016) 260-270; DOI: 10.3221/IGF-ESIS.35.30 268 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 0 0.5 1 1.5 2 2.5 Load, F [N] Displacement, V LL [mm] Experimental, No H Experimental, H Numerical plane strain, No H Numerical plane strain, H Numerical plane stress, No H Numerical plane stress, H Figure 6 : F-V LL curves from experimental tests and numerical simulations, for plane strain and plane stress. Solid lines plot data without hydrogen, dashed lines with hydrogen. By observing Fig. 6, it is possible to note that in plane stress condition the calibration process does not allow a good overlapping with the experimental results. Despite many simulations, it appeared that the combination of TSL parameters for this model never resulted in a good fitting with the experimental data. In terms of fracture toughness J value, that represents the area below the F-V LL curve, the value obtained by the calibration process simulating the test without hydrogen is a bit smaller than the experimental one (experimental fracture toughness without hydrogen = 215 N/mm). Then, this tiny discrepancy is reversed comparing experimental and numerical curves related to the hydrogen environment conditions (experimental fracture toughness with hydrogen = 22 N/mm). On the contrary, the plane strain model better fits experimental data in both hydrogen and no-hydrogen conditions. Therefore, between the two developed models, the 2D plane strain one seems to represent better the experimental evidence. However, as a future development, a 3D numerical model would give more information about the embrittlement depending on the local stress condition. C ONCLUSIONS he current work presents a cohesive zone model reproducing a toughness test in presence of hydrogen for both plane stress and plane strain configurations. The following conclusions can be drawn:  firstly a calibration of traction-separation law (TSL) that defines the constitutive properties of cohesive elements is developed for the steel AISI 4130;  a model running on three steps of simulations calculates the total hydrogen concentration present into the specimen including the interstitial lattice sites and the trapped amount. Based on the total content of hydrogen a scale factor is computed and used to reduce the TSL area previously calibrated in order to simulate the embrittlement effect of hydrogen;  for the present steel, AISI 4130, the considered relation to calculate trapped hydrogen concentration C T depending on plastic strain seems not appropriate. By the implemented equations, the numerical model is not able to estimate the hydrogen concentration in reversible traps. Therefore, the model computes hydrogen embrittlement only as a function of hydrogen in interstitial lattice sites. This is a consequence of a brittle behavior of the steel, according to the experimental test; T