The mechanisms of CO2 corrosion are complex and the effect of several parameters was recognised as important in the damage prediction. Predictive models are validated using both field observation and laboratory tests, where the quantitative information about flowrate and flowpattern is somehow lacking. With the aim at obtaining a better correlation between laboratory test results and field observation, a computational fluodynamic code was used in this work, to model the situation inside a cylindrical cell with rotating specimens. Preliminary results show that this technique can be useful to a better understanding of the experimental data. The morphology of localised corrosion (mesa attack) was reproduced in samples tested in the cell.

You do not currently have access to this content.