Hydrogen/plasticity interactions at an axial crack in pipeline steel: Fid-Bau Portal

Hydrogen/plasticity interactions at an axial crack in pipeline steel

Dadfarnia, M. / Sofronis, P. / Somerday, B.P. / Robertson, I.M.

The present work attempted to provide a measure of the magnitude of the hydrogen concentration ahead of a crack as well as the time at which steady state is reached. It is very likely that both parameters are associated with the embrittlement of the material. Certainly, one can argue that hydrogen reaches steady state conditions very quickly in comparison to the long term service of a pipeline. Precise identification of the embrittlement conditions requires the coupling of the present simulations with a micromechanics model for fracture. Experimental work and theoretical calculations of fracture initiation and growth originating from an axial crack in a pipeline is the subject of an ongoing investigation. Experimental studies of hydrogen transport and trapping at crack tips are scarce, and one likely reason is because of the complexity of such measurements. For example, the most accurate experiments must have high spatial resolution, probe the high-constraint region in the interior of the specimen, and be conducted in situ with high-pressure hydrogen (or its isotopes). The authors are not aware of any attempts to measure the transient evolution of hydrogen concentration fields at a crack tip. There have been some attempts to quantify the magnitude of the hydrogen concentration at a crack tip. Results from these studies must be considered qualitative, due to complications associated with spatial resolution and surface measurements. However, these studies confirm that hydrogen concentrates in the hydrostatic stress field and plastic zone ahead of the crack tip. Since conditions of small scale yielding were prevalent in the calculations for the full-field solution, the modified boundary layer (MBL) approach was used to solve the coupled elastoplasticity/diffusion problem in the neighborhood of a crack. It has been demonstrated that using the MBL approach with a diffusion domain of size equal to the uncracked wall ligament, one can predict the peak normalized hydrogen concentration in NILS at steady state within 2 % deviation from the actual value in the full-field solution. Thus the modified boundary layer approach offers an expedient way to study hydrogen transport around cracks in pipes. Lastly, attention should be drawn to the effect of the hydrogen diffusion coefficient on the time to steady state. Looking at the nondimensionalized hydrogen transport, one notices that there is no explicit dependence or the hydrogen diffusion coefficient D. Therefore, the solution is independent of the hydrogen diffusion coefficient D. This implies that the effective nondimensionalized time tau_ss to steady state is also independent of D. Then, using the relationship t_ss = tau_ssL²/D between tau_ss and t_ss of Eq 2, one concludes that the actual effective time to steady state t_ss is inversely proportional to the diffusion coefficient D (i.e., t_ss ~ 1/D). This is an important result in view of the existing ambiguity around the magnitude of the diffusion coefficient in bcc ferritic systems in which a large range of diffusivities have been reported around room temperature.