In part one of this two-part series, we introduce a comprehensive mechanistic model for predicting CO2 corrosion rates and chemical speciation throughout the boundary layer, developed using multiphysics software. Parametric studies are conducted to interrogate the model over an extensive range of temperatures, partial pressures, and bulk pH values. Contour plots are produced showing the variation in key outputs, including corrosion rate, surface pH, and surface saturation ratio with respect to iron carbonate (FeCO3).
The results display significant variation between the bulk and surface chemistry across the entire range of conditions examined, with local species concentration several times larger/smaller compared to the bulk. The observed output responses to the input conditions in many instances were found to vary depending on the other set conditions, demonstrating the difficulty in predicting corrosion behavior without the aid of either such computational models or reliable experimental data. The observed behaviors are discussed in detail in relation to the fundamental aspects incorporated into the model in order to provide greater understanding of the corrosion process.
Simulation and modeling of corrosion processes is an area of research that has seen significant growth in recent decades, with technological advancements drastically reducing the time required to solve the equations that underpin real-world physics. Predicting the behavior of a system computationally, when done accurately, provides great benefit complementing experimental testing to further explain what is happening within the corrosion process. There have therefore been multiple predictive models produced over the years to achieve this aim. Within the realm of carbon dioxide (CO2) corrosion, Kahyarian et al.1 identified three main categories for these types of models: empirical/semi-empirical, elementary mechanistic, and comprehensive mechanistic. Early models such as the work be de Waard and Milliams2 or the original model produced by Nesic et al.3 relied upon experimental data or simplification of the processes to predict corrosion rate. However, in more recent works4-8 the focus has been on developing comprehensive mechanistic models to fully describe and couple together the underlying equations. As outlined by Kahyarian et al.,1 these models have become more advanced over the years with more complete descriptions of the physicochemical laws driving the processes. However, despite their complexity, these models are generally created in such a way as to be unfavorable to the production of large data sets and instead are used to evaluate specified environmental conditions.
