COLONIAL PIPELINE CO

Incident Narrative

ON MAY 1, 2018, AT AROUND 20:00, THE ""B"" OPERATOR AT MITCHELL JUNCTION BEGAN MAKING HIS YARD MANIFOLD INSPECTIONS. WHEN HE GOT TO THE UPPER POND NEXT TO THE LINE 3 AND LINE 4 CROSSOVER VALVES, HE NOTICED A SHEEN ON THE POND AND CONTACTED THE ""A"" OPERATOR ON DUTY. THE ""A"" OPERATOR CONTACTED THE OPERATIONS MANAGER (OM) AROUND 20:30, TO INFORM HIM OF INDICATIONS THERE WAS A SHEEN ON THE UPPER SECONDARY CONTAINMENT POND NEAR THE CROSS OVER VALVES AT MITCHELL JUNCTION. THE SENIOR OPERATOR WAS CALLED OUT TO DETERMINE THE NATURE OF THE SHEEN AND CONFIRMED IT WAS GASOLINE. AROUND 21:00 THE OM INSTRUCTED THE MITCHELL OPERATOR TO CONTACT THE COLONIAL CONTROL CENTER TO SHUT DOWN LINES 03, 04 AND 27. THERE WAS NO KNOWN VOLUME TO DISCLOSE AT THIS TIME BUT COURTESY NOTIFICATIONS WERE MADE LATER THAT EVENING TO THE FOLLOWING AGENCIES: VDEA, VASCC, VDEM, AND CUMBERLAND COUNTY EMERGENCY MANAGEMENT/FIRE CHIEF. PROJECT AND DISTRICT PERSONNEL WERE CONTACTED AROUND 21:00. CONTRACTOR AND MAINTENANCE PERSONNEL WERE CONTACTED TO FURTHER EXPLORE THE SOURCE OF THE SHEEN AND LATER FOUND A 3' X 3' AREA OF PRODUCT STAINED SURFACE SOIL LOCATED OVER THE TOP OF LINE 27 AND ADJACENT TO A CULVERT PIPE THAT DISCHARGES TO THE SECONDARY CONTAINMENT POND. BOOM WAS PLACED ON THE SECONDARY CONTAINMENT POND AND, AS A PRECAUTION, THE DOWNSTREAM STORM WATER RETENTION POND. ADDITIONAL EQUIPMENT AND PERSONNEL WERE CALLED IN TO ASSIST WITH EXCAVATING THE AREA THE FOLLOWING MORNING. BY THE NEXT MORNING (MAY 2), THERE WAS NO VISIBLE SIGN OF PRODUCT ON THE SURFACE OF THE SECONDARY CONTAINMENT POND. SOIL WAS EXCAVATED AND STOCKPILED ON PLASTIC PENDING OFF-SITE DISPOSAL TO EXPOSE LINE 27 FOR INSPECTION. THE SOURCE OF THE LEAK WAS DETERMINED TO BE A HAIRLINE CRACK ON THE WELD OF A HISTORICAL REPAIR. THE PRIOR REPAIR WAS A BOLT-ON-CLAMP THAT WAS INSTALLED IN 1986 (BASED ON PRIOR BREAK & LEAK REPORT) TO ADDRESS HAIRLINE CRACK IN THE PIPELINE. THE LEAK OCCURRED ALONG THE LONGITUDAL SEAM WELD OF THE PRIOR REPAIR. OBSERVATIONS BY NED FIELD PROJECT MANAGER AND PROJECT INSPECTOR INDICATED THAT THE WELD WAS A FILLET WELD. AFTER DETERMINING THE REPAIR THAT WOULD BE NECESSARY (SEE BELOW), COLONIAL CONFIRMED THAT COSTS TO RESPOND TO THE INCIDENT WERE GREATER THAN $50K, AND MADE NOTIFICATIONS TO THE NRC AND PHMSA AID ON MAY 2, 2018. THE LEAK WAS REPAIRED BY PLACING A TYPE B PRESSURE CONTAINING SLEEVE OVER THE PRIOR REPAIR TO CONTAIN THE LEAK. THE TYPE B SLEEVE WAS A STACKED SLEEVE REPAIR. FINAL REPORT UPDATE: AN INCIDENT ANALYSIS WAS COMPLETED. WHILE A METALLURGICAL ANALYSIS WAS NOT COMPLETED ON THE FAILED WELD, AN ENGINEERING REVIEW OF AVAILABLE INFORMATION INDICATES THAT THE CRACK IN THE LONGITUDINAL FILET WELD OF THE BOLT-ON-CLAMP LIKELY FORMED DUE TO A COMBINATION OF FACTORS INCLUDING THE INSTALLATION METHODS AND TECHNIQUES OF THE PRIOR REPAIR, THE DESIGN OF THE REPAIR (I.E., USE OF A LONGITUDINAL FILET WELD AND BOLT-ON-CLAMP VS TYPE B PRESSURE CONTAINING SLEEVE) AND FATIGUE OF THE WELD DUE TO PRESSURE CYCLING ON LINE 27.