For a general introduction into the basic design of steam supply systems and on the layout of the parts of the steam system at the Complex relevant to this incident see Appendix 1
Steam is generated in the power station situated on the North Side of the Complex and is distributed to steam users on the South Side of the Complex through steam distribution mains that run through culverts (service tunnels) that go beneath the A904 Bo’ness road that runs through the middle of the Complex.
A number of different steam mains and routes are available in order to give flexibility and to allow sections of the pipework to be isolated for maintenance. Many isolation valves, pressure relief safety devices and steam traps are located throughout the distribution system in order to ensure the correct routing and to maintain the correct conditions within the steam distribution system (for example the correct pressure, removal of hot liquid condensate etc.).
The mains located in the culverts represent a low point in the distribution system and rainfall collected in the culverts must be pumped out to ensure that the culverts do not flood and external damage to the pipework/lagging etc. does not occur.
For a period of at least three years preceding May 1999 the pipework through the West Gemec culvert was not being used and had been isolated by closing a number of valves on both the northern and southern sides of Bo’ness road. The steam supply to the South Side was via the East Gemec culvert for this period.
In May 1999 as part of the tie-in between the existing power station (northern side of Bo’ness road) and the proposed new Combined Heat and Power plant located on the South Side it was decided to re-utilise an existing medium pressure steam line which ran through the West Gemec culvert.
As a new project it required the preparation of a Plant Modification Proposal (PMP) which required consideration to be given to the safety issues associated with the proposed modification.
In order to test the line one of the valves on the North Side was opened and steam was used to test the line overnight by holding the line at the MP steam working pressure of 200 psig. Following completion of the tests some minor repair work was carried out including renewal of a steam trap and drain line. The line was then re-tested to check the modifications before being isolated again on the North Side in June 1999. The line then remained isolated until the following March 2000.
In order to connect the new Combined Heat and Power plant (CHP) and the Above Ground Gas Installation (AGI) units into the South Side steam supply main a new connection (300mm/12") was installed and new valving arrangements were added on the South Side in the vicinity of the West Gemec culvert. Commissioning of the new line started on 22nd March and this involved many operations of the valves in the area around the culverts in order to test the new system. The valve isolating the MP steam supply through the West Gemec culvert on the North Side (V1) was then opened. During testing some of the isolation valves on the South Side (V2/V3) were also routinely opened for testing which effectively brought the MP steam line through the West Gemec culvert back into use.
At 18:07 p.m. on 29th May the Complex suffered a power loss as described earlier (power distribution failure incident) which affected many of the process plants and utility systems on-site. As a result of the power failure, there was no power available to pump water out of drainage pits, and this water overflowed into the culverts, whose electrically-driven drainage pumps were not available. Around 12 hours later one of the safety valves on the new line to the CHP plant was reported to be lifting and venting condensate. Operators described it as being in powerful spurts every 5 minutes over a 30 minute period. Operators also described very loud banging noises and movement of the pipework in the vicinity of a manifold installed on the South Side.
In response the steam main was re-isolated on the southern side of the culvert but the MP steam main beneath the road in the West Gemec culvert was left connected to the North Side MP steam supply. Condensate was drained from some sections of the steam pipework and the MP steam supply re-established to the South Side. No report was raised regarding this incident (a Total Loss Control (TLC) report was normal practice on-site) and the immediate and root causes of the incident were not determined. The BP site inspection department were not informed that the pipework had experienced water hammer and no formal assessment was made to confirm the integrity of the pipework.
Steam was identified as blowing across the Bo’ness road from the vicinity of the East Gemec culvert on 30th May. Investigation revealed that both culverts were flooded and action was then taken to pump out the East Gemec culvert followed by the West Gemec culvert. Overall the culverts had been flooded for a period of 36 hours. Initially it was suspected that there was a steam leak in the East Gemec culvert. However on further investigation it was concluded that heat lost from the submerged MP steam pipeline (which provided a steam flow to the South Side of the Complex) boiled the water in the East Gemec culvert and generated the steam that was observed. The West Gemec culvert was isolated on the South Side even though it was open to the MP steam supply on the North Side. This effectively created a "dead-leg" in the steam supply system with MP steam able to enter at the North Side but not to exit at the South Side. The rate of steam flow into the West Gemec culvert and heat loss was insufficient however to boil the water in a similar fashion to that observed for the East Gemec culvert.
Following the pumping out of the water from the culverts the steam mains in the East Gemec and West Gemec culverts were inspected on 2nd and 5th June respectively. The BP Inspector was not told of the serious system upset, nor were those responsible for arranging the confined entry permits told of the severe water hammering and relief valve discharge of condensate in adjacent pipework. Prior to the inspection one of the steam traps in the West Gemec culvert, which discharged directly to atmosphere, was isolated for safety during the inspection, because it was continually "firing". The steam main however was not isolated on the North Side and remained connected to the section of MP steam main located in the West Gemec culvert
The inspections found that although some of the lagging was dislodged there was no evidence of structural damage having occurred as a result of both mains having been totally submerged.
Following completion of the inspections the steam trap in the West Gemec culvert that had been closed was not re-opened and therefore this condensate removal device serving the low point of the system continued to be isolated.
From the 30th May to 7th June in the period immediately prior to the subsequent incident the length of line in the West Gemec culvert was effectively a "dead leg" since the MP steam supply system had been isolated on the South Side but was open on the North Side. This allowed steam to continuously enter the section of MP steam main in the culvert. In the absence of a pathway through on the South Side MP steam entering the system therefore gradually cooled and condensed. The resulting change in state from vapour (steam) to liquid (hot condensate) resulted in a reduction in volume. This allowed more MP steam to enter the culvert section and hence to cool and condense. As a result steam was continuously entering the culvert section and condensate was continuously being produced. As the steam trap at the lowest point in the system had been isolated it allowed condensate to accumulate in the system.