INSTANTANEOUS PROBABILITY OF UNCONTROLLED EXTERNAL LEAKAGE DURING THE PRODUCTION PHASE OF ASUBSEA WELL

摘要

This work develops a methodology for evaluating the uncontrolled external leakage probability of a subsea well during the production phase. Based on a barrier diagram, an algorithm for possible leak paths identification is proposed, considering different operation modes: gas lift operation, free flowing or well closed at the subsea Xmas Tree. Considering the equivalency between these paths and the minimum cut sets from a fault tree modeling, the uncontrolled external leakage probability is calculated using the upper bound approximation. The effect of common cause failures is considered for the failure mode fail-to-close-valve. The instantaneous availability function of each component is modeled to represent the maintenance strategy applied. Non repairable, repairable and periodically tested items are used. For the latter, a nomenclature to distinguish two subtypes is introduced: the PT-R and PT-NR models, respectively Periodically Tested Repairable, and Periodically Tested Non Repairable. Probability distributions parameters are roughly estimated in order to make a case study. The failure rate functions determined are used as input for the proposed model, regarding the following failure modes: fail-to-close, external-leakage, and internal-leakage at the closed position. The objective of this section is to adjust a Weibull distribution, eliminate the usual assumption of constant failure rate and account for eventual wear-out effects. Finally, instantaneous probability results and sensitivity analysis are demonstrated for a base case study. Parameters like time between tests, inspections, and component reliability are varied in order to identify the impact on the uncontrolled external leakage probability. Therefore, the main objective is to propose a model that could support decision making on the well integrity management system during the production phase of a subsea well. To make this possible, reliable input data should be further considered.