| dc.description.abstract | Investigating and establishing root causes of oil well kicks is a hot topic today, especially after the blowout of the Macondo well in the Gulf of Mexico in 2010. Lessons can be learned from previous cases and their known root causes. From previous cases, solutions to similar problems today can be found, implemented and remembered. The Petroleum Safety Authority in Norway issued in 2011 a report on trends and risks on the Norwegian Continental Shelf. One of the conclusions from the report was that there has been an increase in the number of leaks and well incidents on the Norwegian Continental Shelf since 2008. This negative trend has to be reversed and lessons learned needs to be considered and implemented.
Root causes were found for three different fields by applying cased based reasoning approach. The root cause found for Snorre A, well 34/7-P31A, was swabbing, for the case on Valhall, well 2/8-A08B, the root cause was lost circulation, and for West Wanguard, well 6407/6-2, the root cause found was due to abnormal pressure and swabbing. Observations from the fields were interconnected to each other by relations and entities. This allowed for a schematic model of kick causes to be built.
By applying parts of the knowledge based module, it is possible to find root causes of kick events.However, the model is more quantitative than qualitative. Specific observations have specific entities and are linked together by relations. In order find observations of kick causes, good knowledge of the subjected is required. The model tends to focus more on the amount of observations it can retrieve for comparing with new cases, rather than valuing the content in observations. Another problem is the quality of data available for analyzing. This was a great challenge when investigating the root causes of kick events. Although the analyses has established root causes, for one of the cases it was a struggle due to the lack of data. | nb_NO |
