Recently, experiments have been performed to determine the electrochemistry of mild steel at high pressures of carbon dioxide (CO2) and in the presence of acetic acid (HAc). An electrochemical model which has been validated with data collected in glass cells will be compared to the new data which has been gathered in a high pressure flow loop Furthermore, the de Waard corrosion model has been modified to account for the presence of acetic acid and the predicted corrosion rated from the de Waard model will also be compared to the experimental corrosion rate data found by using linear polarization resistance (LPR) and weight loss measurements.
Numerous corrosion prediction models have been developed throughout the years in an attempt at quantifying the risk of CO2 corrosion.1-8 These models are largely empirical or semi-empirical correlations from either field data, laboratory data or a combination of the two. A limitation to these models appears when additional data about existing or new phenomena become available (such as the effect of organic acids discussed below) and then cannot easily be incorporated into the models without recalibration of the entire model. This can be difficult and often a time consuming exercise with an uncertain outcome. Another approach to CO2 corrosion modeling is exemplified by the so called mechanistic models9-10 which are developed on solid engineering/scientific grounds. Mechanistic models also include unknown constants, which must be estimated by comparison to experimental data, however these constants have clear physical meaning and usually do not need as frequent adjustments whenever new data emerge or new phenomena are incorporated into the model. The main drawback with mechanistic models is that they cannot be formulated successfully unless there is sufficient understanding about the phenomena being modeled, a problem not encountered by the purely empirical models. As the name suggests, semi-empirical models are somewhere in between. A limitation of all existing CO2 corrosion models is that they are predominantly based on laboratory data, generated in glass cell or small diameter flow loop experiments. While a wide range of experimental parameters can be covered in such equipment (temperatures, pH, flow rates) most in depth electrochemical studies have been limited to glass cells and low partial pressures of CO2. (typically 1 bar). Since limited experimental data exist at high CO2 partial pressures, the models based on low pressure data have to extrapolate to these conditions. Their application to complex field conditions is an even more uncertain exercise. Recently, new experimental data have been generated which offer an insight into the mild steel corrosion mechanisms and rates at high partial pressures of CO2 and in the presence of HAc.11 These data were generated with the primary intention of verifying and adjusting the existing mechanistic models for better corrosion rate prediction under these conditions. The experimental procedures and discussion of the results are described in the original publication11 and will not be repeated here. The details of the electrochemical model and the modified de Waard (GDN) model has also been given elsewhere12 and only the outline of the models will be shown below followed by the predictions and the discussion.
MODELS
Electrochemical model
In the model briefly described below it has been assumed that the main cathodic reaction is H+ reduction and that H2CO3 and HAc act primarily as additional sources of H+ ions (through dissociation). The only anodic reaction considered below is iron dissolution. Hence one can write the current density v
