Gå til hovedinnhold

Investigation of gas leak on Valhall

On Monday, 12 January 2004, a gas leak occurred on the Valhall production and compression platform (PCP).
The leak came from a grease nipple on a 20" sectioning valve downstream of the third stage scrubber.
The Petroleum Safety Authority Norway (PSA) has conducted an investigation of the incident.


Course of events

A hired valve technician was to lubricate a sectioning valve and unscrewed the cap of the grease nipple. He gradually discovered a minor leak that rapidly escalated (approximately 17.53 hours).

He first attempted to replace the cap, but gave up this effort due to a lot of gas in the area and the low temperature affecting his hand as a result of the gas that was flowing out.

He left the work site, climbed down a ladder and ran to the control room, which is about 20-30 meters away, to report the leak.

The control room operator contacted the area technician and asked him to check the leak. The control room decided to implement a level 2 emergency shut down (ESD2) at 18.06.31 hours, and an announcement was made over the PA system for personnel to muster in the lifeboats.

Depressurization was completed at about 18.24 hours. The deluge system was triggered from about 18.33-18.36 hours as a result of the power cut (ESD2) and subsequent loss of compressed air.

The personnel roll-call was completed at 18.46 hours, and normalization commenced from that time. Personnel were allowed to leave the lifeboats at about 19.00 hours.

Actual consequences

The actual consequences of the incident are limited to the gas emission from the grease nipple. BP has estimated the total volume of the leak to be approx. 25 m3, with a maximum rate of about 0.18 kg/s. No personal injuries were reported.

The leak caused a shutdown of about five hours before personnel could start running up the facility.

Potential consequences

The potential consequences of this incident are a somewhat larger gas leak. The hole in the grease nipple is about 1.5 mm in diameter. The maximum opening in such a nipple has a diameter of approx. 5 mm.

Triggering factor / direct causes

Leaky nipple

One of the grease nipples on USDV 3870 did not remain tight against the process pressure (378 psig). It is unclear why the nipple started to leak. The nipple was subsequently dismantled and there were no visible mechanical defects on the spring, ball or seat.

Underlying causes - barriers and non-conformances

Postponement of the replacement program for grease nipples

BP has identified weaknesses in the old nipples and a replacement program was prepared in 2001/2002. The plan is to carry out the replacements during planned shutdowns, but this work has not been prioritized in the last two planned shutdowns. Only the most critical valves, which require leak testing, have received new nipples.

The current plan is to replace the remaining nipples during the 14-day shutdown of Ekofisk in 2004.

Work on pressurized systems

The established practice on Valhall is to lubricate pressurized valves through nipples with a simple check-valve (one barrier).

This does not conform with the company's own instructions for work on pressurized systems, cf. BP procedure 1.70.033 Chap. 2.2 on positive safety measures.

According to this procedure, a Safe Job Analysis (SJA) should also have been implemented to clarify actions in the event of a leak.

Furthermore, there is a requirement for a guard until positive safety measures are established. In addition, check-valves are not regarded as being satisfactory barriers.

Safety clearance

A work permit was issued for the work in question (WOP-0643 14), but this was only valid for cold work and thus did not include opening of HC systems (was not ticked).

After the incident, this permit was expanded to include replacement of the grease nipples and the following was added "gas sniffing during leak testing" under the box for special precautions.

This expanded permit applied to the time period when the unit was still depressurized after the shutdown.

The work permit was otherwise not expanded to cover opening of HC systems, or as regards to whether an SJA or safe work review (SAG) was to be performed.

Handling of hazard and accident situations

It took about 13 minutes from the time when the control room was informed about the leak until the emergency shutdown (ESD2) was implemented.

The control room received information about the leak from both the technician who carried out the grease job and from the area operator.

The work site was clearly laid out, there was just a short distance to the control room and there was also an alarm from a gas detector.

On this basis, we believe that it took too long before the necessary actions were implemented.

Summary

We have not identified serious failure of barriers, but we have found a number of minor weaknesses in barrier elements linked to both human, technical and organizational factors.

This led to an accidental emission of gas before the safety systems, together with operative intervention from the control room, halted further development.

The following weaknesses were identified (short form):

  • This specific type of nipple should have been replaced in a campaign about two years ago, but the replacement was not defined as being sufficiently critical, see Section 9 of the Management Regulations on planning and Section 43 of the Activities Regulations on classification.

  • An old type of nipple was installed after a new, improved type became available on the installation, see Section 12 of the Management Regulations on information.

  • Grease jobs were performed on a pressurized system with just one check-valve as a barrier, see Section 2 of the Management Regulations on barriers.

  • Procedures and instructions for greasing valves are not clear and do not fully describe the potential hazard, see Section 22 of the Activities Regulations on procedures.

  • The job in question was not sufficiently planned based on the error scenarios that could arise, see Section 28 of the Activities Regulations on measures in connection with execution.

  • The control room operators took 13 minutes to implement ESD2 and did not have a sufficient overview to depressurize the right area. This meant that the leak lasted longer, see Section 11 of the Management Regulations on manning and competence and Section 19 of the Activities Regulations on competence and Section 68 on handling of hazard and accident situations.

  • It took 46 minutes to complete the personnel roll-call, see Section 68 of the Activities Regulations on handling of hazard and accident situations.

BP's own investigation report addresses some corresponding factors.

The PSA has asked the company to provide an overall response as to what measures BP will implement after this incident, both on Valhall and on any other fields.

Contact person in the PSA:
Inger Anda