Numerical reconstruction of localized corrosion mechanism in a marine-exposed reinforced concrete beam: Inverse analysis of self-protection

ABSTRACT A multiphysics modeling approach coupling chloride transport and corrosion electrochemistry with a potential-dependent chloride threshold framework was used to reconstruct the localized corrosion in a reinforced concrete beam exposed to a marine environment for 25 years. The experimentally observed localized corrosion, attributed to self-protection mechanisms arising from cathodic polarization of the passive reinforcement surface due to sustained corrosion at existing pits, forms the basis of the approach. The reconstruction was constrained by two observations: (1) half-cell potential, reflecting the state of corrosion at the time of investigation; and (2) pit depth, reflecting the cumulative effect of corrosion over propagation period. Literature-based kinetic parameters failed to reproduce the half-cell potential distribution and pit depths, and literature-based chloride threshold distributions alone could not replicate localized corrosion. To address this limitation, a surrogate-based optimization framework combining Radial Basis Function with Sequential Least Squares Quadratic Programming was applied to a numerically generated dataset. This approach enabled calibration of pit-specific kinetic parameters and a time-invariant potential-independent chloride threshold distribution, allowing reconstruction of the observed localized corrosion pattern. The calibrated parameters produced pit depth and half-cell potential predictions at the pit locations within the experimentally observed range. Furthermore, the results indicate that localized corrosion is jointly governed by cathodic polarization of the passive reinforcement surface and a high chloride threshold associated with a high-quality steel-concrete interface, while corrosion activity at pits, pit size, and concrete resistivity control the extent of cathodic polarization.