The production of a CO2 flood in the Oklahoma panhandle led to severe corrosion of the carbon steel production tubing and casing. Traditional approaches to chemical corrosion inhibition were unsuccessful. A laboratory study was initiated to determine fmt the best corrosion inhibitor, and second the optimum effective inhibitor concentration in the produced fluids as a fimction of the production rate, CO2 partial pressure, and water to oil ratio. The tool used was the high speed autoclave test (HSACT) discussed in earlier publications. Statistical experimental designs were used to study the three major parameters. The results were expressed in terms of the inhibitor concentration necessary to achieve a desired corrosion rate (for example 1 mpy), and presented either in the form of response surfaces or linear multiple regression equations. While it was generally known that higher fluid velocities require a higher inhibitor concentration for equal target corrosion rates, it was less well appreciated that the CO2 partial pressure also has a significant effect on the effective inhibitor concentration. The model as
represented either by the response surface or the predictive equations is both inhibitor and field specific. Keywords: carbon dioxide, fluid velocity, partial pressure, corrosion inhibitor, modeling, statistical design, response surface, effective inhibitor concentration, target corrosion rate