Implementation and Numerical Exploration of Advanced Aqueous Carbon Dioxide Corrosion Models

When carbon dioxide (CO2) is dissolved in an aqueous solution, it forms carbonic acid (H2CO3) which can cause steel surfaces to corrode. This is a serious issue in many industrial pipelines which are often used to transport CO2 saturated solutions. Understanding exactly how the internal surface of a pipe is going to corrode is vital in effective corrosion control. In this work, the potential for simulating the corrosive environment via finite element modelling is explored. Comprehensive mechanistic models of CO2 corrosion have previously been developed, but not extensively interrogated. As a result, the associated limitations of these models, as well as their potential for further advancement, is not fully understood. It is therefore the intent of this thesis to conduct a thorough investigation into the current state of the CO2 modelling and explore pathways to their continued advancement via computational modelling, numerical exploration, and experimental testing.

Two one-dimensional (1-D) models of CO2 corrosion have been developed within the COMSOL Multiphysics® software. The models focus on the hydrodynamic boundary layer in the near-wall region and incorporate both chemical and electrochemical reactions throughout the boundary, as well as at the surface, and in the bulk flow. Extensive numerical exploration of these two models is conducted to investigate how changes in the environmental and hydrodynamic conditions influence important outputs, such as the corrosion rate, surface pH, and the likelihood of protective films forming. Significant differences in the response were identified based upon the controlling mechanism for the corrosion reactions; whether it is the rate of charge-transfer at the surface or the rate at which electro-active species are transported to the surface.

Computational fluid dynamics (CFD) is used to model common flow scenarios and then integrated into the corrosion models. A novel methodology is explored to expand the application of the models to higher dimensions and more complex flow fields. Rotating cylinder electrode (RCE) experiments are conducted alongside the simulations to assess the accuracy of the corrosion models and explore their various strengths and weaknesses. In doing so, it was found that the specific mechanisms CO2 corrosion included in the model, as well as the temperature dependencies, and the method by which flow effects are accounted for can all have a significant difference on the accuracy of the simulation.

Throughout the thesis, many avenues for improvement and further development are identified. These ideas have been collated and explored to lay the groundwork for future modelling projects based on the fundamental approaches detailed here. Approaches are summarised and preliminary work has been completed to clearly identify how the current models can be built upon, as well as the potential pitfalls and areas of complexity.