@Article{fdmp.2026.087557,
AUTHOR = {Jiabin Wang, Rundong Wang, Xingda Tong, Shaopeng Hao, Bo Yan, Zhihui Wang},
TITLE = {A Unified Semi-Empirical Model for Low-Velocity Corrosion Mitigation and High-Velocity Erosion Enhancement in CO<sub>2</sub>-Containing Environments},
JOURNAL = {Fluid Dynamics \& Materials Processing},
VOLUME = {},
YEAR = {},
NUMBER = {},
PAGES = {{pages}},
URL = {http://www.techscience.com/fdmp/online/detail/28309},
ISSN = {1555-2578},
ABSTRACT = {This study develops a unified, material-specific semi-empirical framework to describe both low-velocity corrosion mitigation and high-velocity erosion-corrosion enhancement in CO<sub>2</sub>-containing environments. L360, 20# and X65 steels were investigated using static weight-loss and electrochemical tests, together with dynamic erosion-corrosion experiments and SEM/EDS characterization. Static tests were conducted at total pressures of 3–7 MPa, CO<sub>2</sub> partial pressures of 4–60 kPa, and temperatures of 10–60°C, while dynamic tests were performed at 3.5 MPa total pressure, 40 kPa CO<sub>2</sub> partial pressure, and 25°C, over velocities of 0–50 m/s and impingement angles of 0–80°. Static corrosion rates increased with increasing CO<sub>2</sub> partial pressure and temperature. Under dynamic conditions, corrosion rates initially decreased with velocity, indicating a low-velocity mitigation effect, before increasing at higher velocities as erosion became increasingly influential, with the maximum measured rate occurring at an impingement angle of 45°. Because SEM/EDS evidence did not establish the presence of a crystalline FeCO<sub>3</sub>-dominated protective film, the low-velocity behavior was represented phenomenologically through a surface-coverage/deposit effect rather than attributed to a specific scale-growth mechanism. The resulting model combines a de Waard-type static corrosion baseline with exponential coverage-induced mitigation, a critical-velocity erosion enhancement term, and a modified Finnie angular function. Applied to the present dataset, the framework achieved an R<sup>2</sup> of 0.9308 and a Root Mean Square Error (RMSE) of 0.0319.},
DOI = {10.32604/fdmp.2026.087557}
}