Abstract
Several models are in general use throughout the oil and gas production industry to predict corrosion rates in carbon steel flowlines and pipelines, mainly focusing on CO2 corrosion. New models are continuously announced, and old ones are updated to obtain ever more accurate predictions. The models generally emphasize specific features of the corrosion mechanism, such as the effect of flow, scale formation, pH, etc. The focus of this paper is on the application of the models to optimize the design, materials selection and operations of a pipeline. One common observation based on the extensive use of models for these applications is that the accuracy of the corrosion rate predictions is limited in the first place by the lack of information on the availability of corrosion control methods, secondly by the quality of the input data, and only by exception by the corrosion models. In consequence, to gain more confidence in the use of models it would be more effective to put more effort and resources into improving the quality of the input data than in increasing the sophistication of the models.
