Nuclear corrosion modeling can help engineers connect reactor coolant conditions to the initiation and growth of localized damage in light water reactor (LWR) components. The challenge is that corrosion is not governed by a single operating variable: electrochemistry, temperature, coolant chemistry, flow, material microstructure and mechanical loading all interact. A modeling program that developed and integrated corrosion-degradation codes with Grizzly illustrates how these factors can be brought together to support component-aging assessments in boiling water reactors (BWRs) and pressurized water reactors (PWRs).
The practical objective is to estimate damage over time using the component’s operating history and anticipated future conditions. That requires models to represent both the environment at a location in the primary coolant circuit and the material’s response to it, including crack initiation and propagation.
Why nuclear corrosion modeling must couple environment and mechanics
Localized corrosion includes pitting, stress corrosion cracking (SCC) and corrosion fatigue. In LWR primary coolant circuits, these processes are electrochemical phenomena influenced by mechanical loading and material structure. Relevant environmental inputs include electrochemical corrosion potential (ECP), temperature, pH, solution conductivity, flow velocity and concentrations of redox-active species such as oxygen, hydrogen peroxide and hydrogen. Stress intensity at a crack and characteristics such as grain size and precipitates also affect predicted damage.
These inputs are linked. For example, radiolysis produces chemical species in reactor water, and their concentrations influence ECP. ECP, conductivity and local crack or pit conditions then affect corrosion and crack-growth calculations. A model that treats these processes as disconnected inputs may miss how changes in coolant conditions alter damage accumulation.
The developed BWR_MASTER and PWR_MASTER codes were upgraded to calculate radiolytic species, ECP and crack growth rate (CGR) under changing reactor conditions. Their purpose is to estimate a corrosion evolutionary path: how damage may accumulate as operating conditions and history change.
From electrochemical potential to crack growth
ECP is a key link between coolant chemistry and corrosion behavior. In BWRs, it depends strongly on oxygen, hydrogen peroxide and hydrogen concentrations. The models represent how these species and other operating variables affect ECP along the coolant circuit, then use the resulting environmental conditions in damage calculations.
The work also advanced the Mixed Potential Model (MPM) used to calculate ECP. A quantum-mechanical correction accounts for electron and electron-hole tunneling through a passive surface film. The reported result is that a single set of kinetic parameters for each redox reaction can be used across different environmental conditions, rather than requiring a separate set for each ECP. This reduces the number of parameters needed to calculate ECP and CGR as conditions vary.
Crack-growth models also require a representation of processes at the crack tip. For stainless steels, the Coupled Environment Fracture Model (CEFM) connects local electrochemical conditions with fracture behavior. For mill-annealed Alloy 600 in PWR environments, the work developed models based on the CEFM and on hydrogen-related pressurization of grain-boundary voids. Their calculated growth rates were reported to agree with experimental results. The models also predict that CGR depends on variables including temperature, ECP and stress intensity.
Predicting crack initiation in BWRs and PWRs
Crack initiation is distinct from propagation, so the modeling program developed approaches for both stages. In BWR stainless steels, the initiation model treats a corrosion pit as a potential starting point. It estimates the time required for a pit to grow until its stress intensity exceeds the threshold for stress corrosion cracking under the prevailing stress and environment.
In mill-annealed Alloy 600, the proposed PWR initiation model focuses on emergent grain boundaries. Water penetration and internal oxidation are postulated to form chromium oxide that wedges a grain boundary open. The model represents effects including applied stress, temperature, cold work, water chemistry, pH and ECP.
These approaches make clear why component assessments need more than a single crack-growth rate. Engineers must consider whether damage can initiate, how long initiation may take, and how an established crack could then grow under changing conditions.
Integrating corrosion predictions with structural analysis
Grizzly provides structural-analysis capabilities that can supply stress intensity factors for crack-propagation calculations. The BWR_MASTER code was made compatible with macOS, and the work connected it with Grizzly so that Grizzly-generated stress intensity factors could be used in predicting crack growth. Simulations examined a BWR core-shroud weld and compared stress intensity and propagation calculations using different approaches, including a reduced-order fracture model and the J-integral method.
This connection is important because corrosion models need mechanical inputs that reflect the component and its loading. A workflow that combines environmental calculations with structural analysis can support estimates of damage accumulation and component lifetime for specified operating conditions. The report also demonstrated why startups and shutdowns matter: changes in power, temperature and coolant conductivity can produce chemistry and crack-growth transients that influence accumulated damage.
Uncertainty, sensitivity and model inputs
Predictions are only as useful as their underlying inputs. Sensitivity analysis identified radiolytic yields, Tafel constants, exchange current densities and standard potentials as influential parameters for calculated species concentrations, ECP and CGR. Comparisons with plant measurements showed improved agreement after updated radiolytic-yield data were incorporated into BWR_MASTER. The analysis also indicated that the choice of chemical reaction-rate dataset matters; alternative datasets did not produce equally satisfactory results for all species and conditions.
Uncertainty also arises because crack initiation times and crack growth rates vary. The work developed a theoretical framework for distributed initiation times and showed that normally distributed independent inputs can produce a log-normal distribution in CGR. This matters for engineering interpretation: a deterministic calculation provides a modeled outcome for specified inputs, but variability in those inputs limits the precision that should be expected in the predicted rate.
Practical implications for plant-life assessments
For plant engineers and technical decision-makers, the central lesson is to evaluate corrosion damage as a coupled, time-dependent problem. Useful assessments should:
- Represent coolant chemistry and electrochemical conditions at the component location.
- Connect ECP and other environmental variables to initiation and crack-growth models.
- Account for mechanical loading and stress intensity from structural analysis.
- Include operating transients, such as startup and shutdown, where they affect chemistry or loading.
- Review model sensitivity and the quality of kinetic and radiolysis inputs.
Integrated nuclear corrosion modeling does not remove uncertainty, but it offers a structured way to relate reactor operating history to localized damage. By combining electrochemistry, materials behavior and fracture mechanics, these methods can support more informed assessments of LWR component aging and the conditions that influence predicted service life.