Consistent Endurance Fatigue Knockdown Factors for Sour Service From Industry-Wide Database

The embrittlement of steel in the presence of water and hydrogen sulfide is a well known phenomenon. For the resulting degradation in fatigue performance, the industry today applies a knock-down factor onto in-air S-N endurance curve that relates the sour-service life to the in-air performance. Several published results are available, and most of these, although rigorous in test approach, report knock-down factors that include unspecified levels of conservatism, consistent with that particular author’s engineering intuition. (For example, typically a lower bound S-N curve in sour-service testing is related to a mean S-N curve in-air). This paper summarizes all of the published small scale sour service testing results, and applies a consistent knock-down factor calculation. Analyses of the data show that sour degradation is highly correlated to H2S concentration and solution pH value. In practice frequency scanning test is highly recommended since sour fatigue test results are highly dependent on loading frequency. Although the database is small, some trends are discernable. In particular, observations indicate that sour service exposure may act as an equalizer, removing the initiation life associated with the time for initial micro-defects at the weld toes to become macro-cracks and leaving only aggravated propagation due to sour service. In this paper, we use initiation life (for welds) to describe the life for the micro-defects (∼0.1mm height) to become macro-cracks (∼1.0 mm height), and a postulation is made that associates the fatigue performance of girth welds (F2, E, D, etc.) with the size and magnitude presence of these micro-defects. The metal surface attack of the sour environment is postulated to provide pitting-like initiation sites for the macro-crack for fatigue propagation. As a base we can use the F2 level performance as the performance due to presence of macro-cracks, and any margin for the D and E level fatigue performances then is associated with more benign initial micro-defects. Once we remove the differences in initiation life; all of the sour performance converges on a single lower performance curve. In this scenario, the knockdown factor is more consistently computed from a standard performance S-N curve rather than the same girth weld’s in-air performance since the in-air data may include significant initiation life. Furthermore, if project sour condition is less severe than NACE TM0177, Solution B with pH = 3.5 and H2S partial pressure = 70mbar, a knock down factor of 45 indexed to BS7608 E design curve is supported from the current database. This provides a design recommendation which can be used for preliminary design in sour environments.