During nightshift on 16th September, there was a production low pressure trip on the oil export line, due to mechanical seal failure on `B` Crude Pump and subsequent trips on the `C` and `A` pumps. As part of this production shutdown, Bravo and Jacky wells were manually shut –in by the Alpha CCR. Beatrice Alpha production wells were checked by the operators to ensure they had reverted to their tripped status of the hydraulic wing valve being closed and the well chokes were closed in.
While reinstatement of the crude pumping system was carried out, it was discussed between the OIM, OTL and night shift operations team that production should be restarted on a single train (`A`) once the crude pumps were available. Steps were then taken to restart the production plant by using N2 on the `A` separator to provide a pressure of 1.5barg to give start up permissives for the repaired crude pumps. This was noted to be taken longer than normal however the N2 system continued to be used to see if the pressure could be achieved. Total duration of N2 pressurisation attempt is estimated to be 1.5hrs. Whilst this pressurisation was ongoing, well slots 6, 17 and 24 were prepared for start-up. Since the well SSSV and master valves were not automatically closed by the initial production trip, this well alignment consisted of resetting the hydraulic wing valve to the open position resulting in the well pressure being held at the choke ready for production start–up. In order to achieve this alignment PSHL10006 on each well was manually overridden to prevent automatic closure of the hydraulic wing valve upon initiation of the high or low pressure trip (50 and 3 barg respectively). All production is routed to the production header in normal operation, as in this case, thus the tester was offline. Investigation of the failure to pressure up the ‘A’ separator using N2 was traced back to PCV30088A (separator ‘A’ fuel gas valve) being stuck in the open position. This could not be resolved on the nightshift; therefore the decision was made to abandon the start up on the ‘A’ train. ‘B’ train could not be used due to the faults with the ‘B’ and ‘C’ crude pumps. At this point (0010hrs, 17 September), well alignment was left with the well pressure being held on the chokes. The wells remained in this state for the remainder of the night shift and were handed over to the oncoming dayshift in this condition. Prior to closing the hydraulic wing valves on well slots 6, 17 and 24, as agreed during the shift handover, a leak was detected on a 1.5” line on AD platform (well bay area, Cellar Deck, West). The leak was initially contained and the CCR was informed, at which point additional personnel attended the scene and aided the clean-up. An initial estimate of 0.5m x 0.3m of oil was noted to be in the sea at this point with further oil evident on the risers and the walkway below. Due to the weather conditions at the time (Westerly 30knots) this oil dispersed in a short period of time and no further evidence of oil going to sea were reported. The line was initially traced back to the test header PSV, which was isolated but this did not stop the flow. Further investigation identified the production header PSV also tied into the line and isolation of this device led to the leak stopping. At this point the wells were shut in at the hydraulic wing valves. Investigation shows that the line in question was 1.5”-DC-54005-A2-1T closed drains line from the production and test header PSV’s (PSV 10032 & 10033) to closed drains collection tank T-5301. No H2S or F&G alarms were initiated during the leak. The source of the pressure lifting the PSV was identified as originating from slot 6 choke valve which was passing. This has since gone into maintenance due to a broken gearing mechanism. The production header was depressured through the production separators to 5barg. A subsequent check for pressure build up in the production header did not show any increase following the isolation of the wells. The PON1 was submitted after the initial response to the event; however this was subsequently changed following this investigation due to more clarity around the timescales of events.
Failed line was not part of inspection program
Design of drain line is not continuously sloped from PSV discharge point to closed drains tank resulting in stagnant fluids gathering with resulting corrosion.
Ensure all operations personnel are aware of the requirement to have an available flow path to a production separator prior to aligning a well start up, as per `Process Systems Procedure`
Review current well start up alignment to prevent shock loading of choke valves. (In this event the current method leads to CITHP being quickly opened up to choke valves on reset of hydraulic wing valve holding pressure at the manual wing valve instead of choke.)