| dc.description.abstract | The blowout preventer system is acting the secondary safety barrier in a hydrocarbon well, where the drilling mud column is defined as the primary safety barrier. However, in pratical industry, there is a new demand for improved methods of assessing the reliability and availability of blowout preventer systems. The objective of this master thesis is to propose the relatively new method for reliability and availability assessment based on Bayesian Network, focusing on the comparsion between the various blowout preventers stack and the influence of the external information in the case study.
The thesis starting with introduction of safety critical system including the basic terminology for reliability and availability assessment. The relevant standards regarding oil and gas industry are also introduced.
The brief review about the blowout preventer is presented next. The basic structure of blowout preventer and the three main subsystems are identified and introduced. The classification of possible failure and desired functions of main components are reported. Finally, the brief research review about the previous reliability assessment method of subsea blowout preventer is presented, pointing out some potential weakness of the traditional methods, which indicating the Bayesian Network is the one possible solution when new requirement of reliability and availability is demanded.
Then the introduction of Bayesian Network is mainly investigated for those who are not very familiar with this method. The possible allocation of Bayesian Network in reliability assessment and the comparsion between Bayesian Network model and the traditional method are mainly discussed. Then in the end of this chapter, there are two examples to show how the traditional method used in the assessment of blowout preventer can be transferred into Bayesian Network model without losing any details, in addtion, the advanced modeling powers regarding introduction of probabilistic gates, multiple states for variables and updating information when scanerio is known are revealed in both examples.
Finally the case study about reliability and availability assessment of different blowout preventers is created. There are mainly three different type of blowout preventer stacks withing different degrees of performing the desired functions under the most demanding situations. The Bayesian Network model is able to perform such assessment effectively and one addtional information is taken into account since the contribution of the wellbore pressure has significant implications on the blowout preventer's ability to seal around or seal off the wellbore.
Finally, the conclusion and discussion are provided. The main conclusion are summarized as three key findings. First, Bayesian Network is proven to carry out the reliability and availability assessment when there is the new requirement in pratical situations, especially for updating the information when the test data is available during the operation. Second, Bayesian Network based reliability and availability assessment is possible for applying in the large scale model since it can handle probabilistic gates and multiple states of components, where the traditional method such as Fault Tree Analysis may not be able to deal with such challenges. Third, according to the analysis results of the case study, the blowout preventer equipped with the Deepwater Horizon type of stack is considered as the most reliable one in the most demanding situation, if the correct repair strategy is applied. In addition, this kind of blowout preventer is relatively very stable under the high wellbore pressure condition even though withing lower redundancy of the pipe ram subsystem, due to inclusion of casing shear ram which improves the shearing ability significantly. | |
