Extraits du rapport du National Transportation safety Board des États-Unis après le déversement du pipeline 6B, à Marshall, Michigan
Rapport NTSB 2012, p. 93:
2.5 Mischaracterization of the Crack Feature
According to PII [PII Pipeline Solutions, specializing in In-Line Inspection], a “crack-like” characterization was indicative of a single linear crack whereas a “crack-field” characterization implied that the feature was made up of a cluster of small cracks typically associated with SCC. All six features identified on the ruptured segment, including the 51.6-inch-long feature that grew to failure, were initially characterized as “crack-field” features by the junior analyst; however, a supervisor changed the final report to read “crack-like” features. When PII identified a feature as a “crack-field,” PII also reported the length of the longest individual crack within the cluster. Enbridge used a criterion of 2.5 inches for the longest crack as a trigger for excavation of “crack-field” defects.
After the Marshall accident, PII reexamined the in-line inspection data and determined that the features were misclassified. Based on this examination of the failure defect, the rupture feature would have had a longest indication that measured 3.5 inches. Because this longest indication within the cluster was greater than the Enbridge excavation criteria for “crack-field” features, the 51.6-inch feature would likely have been excavated by Enbridge in 2005. Therefore, the NTSB concludes that PII’s analysis of the 2005 in-line inspection data for the Line 6B segment that ruptured mischaracterized crack defects, which resulted in Enbridge not evaluating them as crack-field defects.
Rapport NTSB 2012, p. 88 :
2.4.1 Engineering Assessment of Cracks and Margin of Safety
Enbridge applied a lower margin of safety when assessing crack defects versus when assessing corrosion defects. The Enbridge integrity crack management group calculated the predicted failure pressure for each reported defect from data supplied following in-line inspections. From these calculations, Enbridge would select and prioritize pipeline segments for excavation. To Enbridge, the excavation of a pipeline segment would expose the segment and would include a visual inspection and a nondestructive examination for cracks (including SCC) and corrosion. The results from these field assessments were sent to the integrity crack management group and used to assess tool accuracy and to make decisions for repairing the defect.
All crack-like features that had a predicted failure pressure that was calculated to be less than the hydrostatic test pressure of the pipeline segment were scheduled to be excavated. Hydrostatic test pressure is defined by 49 CFR 195.304 as a minimum pressure of 1.25 times the MOP of the pipeline. The Line 6B rupture segment had a MOP of 624 psig with a stated hydrostatic test pressure of 796 psig (or 1.28 times the MOP). By comparison, the corrosion defects on Line 6B were required to be excavated and remediated in accordance with 49CFR195.452(h)(4)(i)(B) when calculated predicted failure pressures were less than 1.39 times the MOP of the pipeline or SMYS (867 psig, the pressure that equates to a circumferential stress equivalent to the SMYS of the pipe). Therefore, the calculated margin of safety for a corrosion feature was 11 percent higher than that of a crack feature.
The use of a lower safety factor for crack defects is inconsistent with the growth rate assumptions used by the Enbridge crack management and corrosion management groups. The crack growth rate used in the engineering assessments of cracks is greater than the maximum corrosion growth rate assumption. Furthermore, Enbridge has stated that a greater range of possible errors is associated with crack tools and that a higher reliability exists with corrosion tools. However, neither of these factors was reflected in the lower safety margin used by Enbridge when assessing cracks than when assessing corrosion. A larger margin of safety would have resulted in a larger number of crack defects being eligible for excavation and examination.