Analysis of sweet corrosion profiles

Abstract

The sweet corrosion profiles of X-52 carbon steel in 0.l% NaCl at 30, 40, 50 and 60°C were analysed. The corrosion tests lasted for a period of one month. They were performed using a flow loop under a total gas pressure of 1 atm. containing carbon dioxide (99.95% & 100 vpm O2). The corrosion profiles were obtained by scanning the corroded surfaces with a stylus. The results were analysed and compared with linear polarisation resistance (LPR) data. lt has been found that the (LPR) data at 40°C, i the absence of significant pitting, progressively underestimated the actual metal loss values. It has been proposed that the above discrepancy is due to the occurrence of small scale localised corrosion. The above effect was found to be exaggerated at 60°C, where extensive pitting occurred. Attempts to find a correlation between the (LPR) measurements and the actual metal loss distributions were unsuccessful. Furthermore, although sudden increase of the instantaneous corrosion rates for passive samples appeared to signal the onset of localised attack, i the absence of prior knowledge of the relative spread of the localised corrosion, it was not possible to quantify the observed errors. The range of the metal loss distributions was found to increase with time. This was most notíceable at 60°C. The progress of localised corrosion therefore, appears to have extended the distributions towards the extreme values. It has been proposed that, in the absence of severe pitting, the sweet corrosion profiles can be approximated by the Weibull statistical function. For pitted samples, where the localised attack resulted i metal loss on different scales, the upper 10% of the values were also found to be approximated by the Weibull function. The Weibull function was used for predicting the progress of the metal loss values with time. The analysis of the top 10% values, for pitted surfaces at 60°C, however, suggested the occurrence of an accelerating rate of attack. The original model was therefore improved to provide a time dependent sweet corrosion model. It has been estimated that the corrosion rate at 60°C increases with time at intervals of approximately 8 days. The effect of methanol on the sweet corrosion of X-52 carbon steel in 1% NaCl was also studied under stagnant conditions at 4, 40 and 60°C. The partial pressure of carbon dioxide was maintained at 0.3 bar. lt has been found that the corrosion rate decreases with increasing methanol concentration. The data was used to develop a predictive model within the range of 10 to 60% vol. methanol.