A 2000-liter, large scale flow loop with 10 cm I.D. pipeline and three different test sections has been used over several years with the goal of understanding the mechanisms involved in H2S/CO2 corrosion environments. Mild steel from an API 5L X65 pipeline was used for the flush-mounted weight loss samples in the flow loop, each was exposed to consistent test conditions for up to 3 weeks at a time. The corrosion product morphology and resulting uniform or localized corrosion are here reviewed from experiments conducted at 60°C, with partial pressures of CO2 up to 8 bar, partial pressure of H2S up to 10 mbar and NaCl concentrations up to 10 wt.%; tests were conducted at pH 4, 5 and 6.
The ability to predict corrosion relies upon using data from the laboratory and from the field in order to develop models of the type of corrosion observed. Empirical models provide a solid starting point for capturing the effect of major parameters (pCO2, pH2S, acetic acid, chlorides, temperature, flow, etc.) and indicate how each of those can affect the uniform corrosion rate in upstream oil and gas pipelines. However, it is more important to be able to predict the likelihood and magnitude of localized corrosion in H2S and CO2/H2S environments, which is often related to the nature or breakdown of a protective corrosion product layer and therefore it’s role in development of localized attack needs to be elucidated. As more knowledge emerges and fills in gaps in a theoretical framework, development of a mechanistic model becomes possible and is the best way to describe the effect of multiple parameters involved. The mechanistic uniform H2S and H2S/CO2 corrosion prediction model used currently in MULTICORP* is based on fundamental theories and experimental results1 aimed at describing specific mechanisms of corrosion.
