The carbon steel lines carrying brackish water associated with a heavy oil SAGD (Steam Assisted Gravity Drainage) operation in Northern Alberta experienced severe localized corrosion attack specific to bends, welds, and other locations containing a gas phase. Pinhole leaks occurred after only seven years of service. A comprehensive study was conducted which included consideration of possible damage mechanisms based on flow regime, water chemistry, gas composition, piping elevation changes, and various corrosion product/material analysis.
This paper is a continuation of a previous study3 which was based on similar failures which occurred at similar facility. New findings are discussed regarding how the pipes are damaged. It is concluded that CO2 breakout is the major cause of severe pitting on the bends and preferential weld attack despite a low CO2 concentration of < 30 mg/L in the water. The CO2 breakout under turbulent flow and/or piping elevation changes resulted in corrosion rates up to 1.125 mm/y (45 mpy) at 10°C. In comparison, the De Waard-Milliams model predicted a corrosion rate of only 0.0425 mm/y (1.7 mpy)!
Turbulent flow (Reynolds number is 1.7 × 105) and depressurization contributed to the CO2 breakout from the water which caused severe corrosion with the protective siderite readily removed. Corrosion occurring at CO2 gas/water/metal boundaries resulted in high pitting rates and corrosion was aggravated with the increase in mass transfer when gas bubbles collapse.
Oxygen corrosion and microbial induced corrosion as secondary damage mechanisms are also discussed in this paper. The corrosion damage mechanism study provides guidance for BW piping/pipeline designs and material selection.
In SAGD facilities, brackish water (BW) containing trace amounts of free CO2 accounts for 15% of the boiler feed water used for steam generation. In 1975, De Waard and Milliams1 developed a now well-known CO2 corrosion model. This model has been widely used to predict CO2 corrosion of carbon steel (CS) in the oil and gas industry. The De Waard-Milliams model was based on tests conducted with 0.1% NaCl solution under oxygen free conditions with the liquid flowing around the electrodes at 1 m/s. The corrosion rate (CR) was determined by means of weight loss and polarization resistance measurements.