Modeling Crack Initiation in BWR and PWR Primary Coolant Circuits

Models are described for calculating the crack initiation times for Alloy 600 and Type 304 SS in PWR and BWR primary coolant circuits, respectively. In PWRs, initiation is defined in terms of the grain boundary oxidation concept of Scott and Le Calvar, whereas in BWRs, cracks are envisioned to nucleate from corrosion pits. In contrast, in BWRs, we envision cracks to nucleate from corrosion pits, with the difference in the two systems being primarily due to electrochemical factors. Thus, in BWR primary coolant and the absence of hydrogen water chemistry (HWC), the oxidizing conditions due to the radiolytic production of H2O2 cause the ECP to be significantly more positive than the critical pitting potential. Accordingly, the nucleation and growth of pits due to passivity breakdown and the establishment of differential aeration between the pit nucleus’s internal and external environments, which results in growth of pits to the critical size necessary to satisfy the Kondo criteria for transition of a pit into a crack, is judged to be a realistic scenario. Contrariwise, in PWR primary coolant, the ECP is so negative [≈−1.0 Vshe] due to the large amount of pressurizing H2 present in the circuit [20–60 cm3(STP)/kg H2O] that the nucleation and growth of pits is not possible. However, Totsuka and Smialowska found that MA Alloy 600 suffers hydrogen-induced cracking (HIC) at an ECP < −0.85 Vshe, demonstrating that, in service with a high hydrogen concentration, brittle fractures will occur. The initiation sites were not identified. The crack initiation models for Alloy 600 in PWRs and Type 304 SS in BWRs reproduce the effects of the following independent variables: applied stress, temperature, cold work, grain boundary segregations, water chemistry, pH, and electrochemical potential. The origins of the observed scatter in experimentally measured crack initiation times are discussed, and the challenges of developing a more general crack initiation model (GCIM) are identified. From a mathematical viewpoint, the most significant challenge arises from the nested distributions involving the many parameters and expressions within the GCIM that are either distributed because of an imprecise definition or because some experimentally determined input parameters are experimentally scattered. Additionally, the evolution of semi-elliptical surface cracks resulting from the electrochemical crack length (ECL) being shorter than the classical mechanical crack length (MCL) must be incorporated if the GCIM is to find utility in the water-cooled nuclear power industry where semi-elliptical surface cracks are normally observed.