Drilling company Transocean had an incident on one of its North Sea rigs similar to the problem which caused the biggest oil spill in US history earlier this year, it has emerged.
Transocean was operating BP's Deepwater Horizon rig when it suffered a blow-out, killing 11 workers and releasing millions of barrels of oil into the Gulf of Mexico.
An internal company report obtained by the BBC shows that four months before the US disaster the Sedco 711 rig in the North Sea, which is leased by Shell and operated by Transocean, experienced similar problems.
In this case, however, the blow-out preventer - which is believed to have failed on the Deepwater Horizon - worked effectively, preventing oil and gas from spurting uncontrolled up the rig's pipe.
The incident on December 23 2009 was investigated by the Health and Safety Executive and Transocean drew up an internal report.
Oil expert Dr Greville Williams of the Institution of Chemical Engineers, who has seen the report, told BBC Radio 4's Today programme: "It is the story of a blow-out occurring during a completion exercise. They had drilled the well, they had got the end of the pipe sorted out.
"What's happened is that the flow valve at the bottom looks like it has been unseated during operations. When they started to displace the mud which is holding the reservoir back, the pressure started to move that mud and sea-water back up the column. That's the blow-out starting to happen."
Dr Williams said the report showed that key indicators that something was going wrong were misinterpreted or discounted, while undue weight was given to a positive pressure test at the base of the well.
"It seems to me from this that they were accepting that they didn't do enough pressure testing of the flow control valve and probably didn't allow for more than one barrier to prevent the reservoir from communicating," he said. "Normally, you would like to have at least two, but they were relying at this point on one barrier."
In a statement, Transocean told the BBC: "Any (safety-) related events that occur on a rig anywhere in the world, including the one on December 23 2009, are immediately reported to management, fully investigated and the valuable information gleaned from that investigation is used to improve existing safety systems across the fleet."