MATHEMATICAL MODELS FOR PREDICTION OF CORROSION RESISTANCE OF PIPE STEEL GRADES IN CO2 ENVIRONMENT

Abstract

A brief characteristic and comparison of 14 mathematical models of CO2 corrosion prediction: de Waard model (DW), Norsok M-506 (NO), Hydrocor (HY), Corplus (CO), Cassandra (CA), KSC Model (KS), Multicorp (MU), ECE Model (EC), Predict (PR), Tulsa Model (TU), ULL Model (UL), CorPos (CP), OLI Model (OL), SweetCor (SW) were presented. The basic characteristics of models were given (the information about the developers, it is described if the model is open or commercial, if it is based on laboratory data, field data or a mechanistic model, if the H2S effect is taken into account). The effect and specifics of calculation in each of the models are considered. The Tulsa Model (TU) and Predict (PR) have the greatest effect. At that, the risk of local corrosion, the influence of organic acids and H2S on the corrosion rate (the partial pressure of CO2 in the model up to 70 bar; the model is commercial). Open Norsok M-506 (NO) and KSC Model (KS) have moderate effect in accounting for formation and condensed water and pH effect on corrosion rate. However, the KSC Model (KS) takes into account the risk of localized corrosion. The open DW model has great efficiency in accounting for crude oil wetting. For further application in development and improvement of pipe steels produced by PJSC “Severstal” it is worth to apply multi-factor models Tulsa Model (TU) and Predict (PR) to account for operating parameters. It is also worth noting that all the listed models don't take into account the rolled steel parameters (microstructure, resistance to hydrogen cracking HIC, general corrosion tests). That's why Severstal together with IMET UrO RAS is developing our own model based on standardized tests and production experience and will allow to produce the new types of pipe steel with specified properties for consumers.