Multiphysics modelling for ultimate strength assessment of damage to ships induced by corrosion

Corrosion is recognised as a significant degrading mechanism which influences the marine structural integrity. The structural deterioration affected by corrosion can significantly degrade the ultimate strength. Even though some countermeasures of corrosion, such as coatings and cathodic protection have been applied, in-service observations have revealed that these protections are not always sufficient. To minimise the failure risk due to corrosion, an accurate prediction of the corrosion is essential. However, most of the corrosion prediction models rely only on historical datasets which are taken from a small range of ships in a limited range of conditions. The models give little insight into what happens when the conditions change. More importantly the factors causing corrosion are not linked, especially the interaction between mechanical stresses and corrosion kinetics, termed mechanoelectrochemical corrosion. To better understand how structural response changes based on corrosion under service loads, this thesis presents a series of finite element analyses which consider the coupling relationship between the surface mechanical stresses and the resulting change of corrosion rate. The coupling provides a complex corrosion-stress interaction depending on the experimental datasets. Quantification of this interaction is based on in situ experimental measurements of corrosion kinetics at different stress levels. Three studies, performed at different structural scales, compressive loadings, and corrosion-stress relationships have been performed to provide an understanding of the coupled corrosionstress effects on ultimate strength and corrosion surface topography. Study 1: utilises a stiffened plate model (1.5 m 0.95 m) with an explicit stress factor to assess the feasibility of corrosion-structural response under different loadings (20%, 50% and 80% proportion of yield stress) and various flange corrosion location, i.e., centre and end flange corrosion, with the same corrosion area (0.75 m 0.045 m). The study results in an irregular thickness reduction and local buckling, with failure mode changes triggered even with slight material loss, due to the non-uniform surface featureand stress concentration. The corrosion location provides different structural and corrosion responses. Study 2: utilises a two-bay/two-span stiffened panels (17.1 m 9.5 m) to assess two different corrosion locations: central corrosion and two-patch corrosion, with identical corroded surface areas (25.46 m2). This study aims to confirm the findings of the stiffened plate study by analysing larger structures, reducing the simplicity of boundary conditions and introducing realistic maritime structural loading scenarios. Distinct changes in the corrosion pattern are observed and up to 23% loss in ultimate strength, or 8% compared to uniform thickness reduction. Study 3: establishes a fuller understanding of the mechano-electrochemical corrosion, where the Study 1 stiffened plate model is coupled with a more complex corrosion-stress interaction through a multiscale multiphysics mechano-electrochemical corrosion model. A surface stress-based anodecathode relationship is defined with a level-set function at a single boundary to solve the Nernst-Planck equation with the help of moving boundary technique based on the Arbitrary Lagrangian-Eulerian method. The results demonstrate the evolution of broad pit corrosion features (1 cm to 10 cm width) with benching, closely resembling that observed in actual inspections/surveys, with up to 6% ultimate stress changes