OPERATIONAL SAFETY RISK ASSESSMENT IN OFFSHORE OIL WELLS

Identifying and analyzing hazards and their associated risks have been a major concern of the oil industry since the Piper Alpha incident in 1988, but techniques to assess risks were only explored further in the last decade after the incident in Macondo. It became clear that there were many hazards associated with oil exploration and exploitation activities; therefore, these hazards had to be identified and proper measures to contain their damage had to be put in place. This led to a development in existing risk analysis techniques such as bowtie diagrams, fault trees, Bayesian networks and Markov methods, but also hybrid techniques such as Markov chain Monte Carlo methods and fuzzy fault trees. In this work, we presented and discussed a literature review on these techniques, focusing on their contributions to the oil industry, as well as existing legislation and standards concerning safety. After comparing advantages and disadvantages of these techniques, we proposed a methodology using graphs to assess the safety level associated to a standardized sequence of operations. This methodology is based on an ontology of operations (which provides standardization) and the concept of BIS (barrier integrated sets), which are sets of elements from a well that ensure safety to perform each operation. With available statistical data, or by using fuzzy sets to cover their absence, this methodology is able to quantify reliabilities and determine if a certain operation can be performed safely based on a risk acceptance criterion. In order to demonstrate this concept, we calculated the reliability of two primary BIS during the intermediate drilling phase and showed how the reliability changes based on the corresponding drilling operational sequence.