Risk assessment of decommissioning options using Bayesian networks

For complex and costly decommissioning activities, it is beneficial, if not necessary, that all relevant risks are identified and assessed on an overall basis, treating and assessing all risks within the same theoretical framework. Only then may the different options be consistently compared and the risks associated with decommissioning demonstrated and documented to the different parties of interest. The present paper suggests an approach for assessing the risks associated with the decommissioning of offshore facilities. The approach takes basis in a discrete point in time representation of the considered decommissioning options where important phases of the options are represented in terms of event scenarios. Using the possibilities of Bayesian Probabilistic Networks (BPN), the failure probabilities and risk events involved in the modeling of an option may then be analyzed for each phase and added up time-wise over the entire decommissioning process. The principles of BPNs are shortly described, and the proposed approach is illustrated by an example linking the operational and structural risks in connection with a re-float decommissioning option for a concrete offshore platform. It is shown how the sensitivity may be evaluated on the basis of the BPNs, thus providing a valuable framework to first improve the risk model in terms of the representation of important scenarios, then for deciding where to apply additional safety measures most effectively, and last but not least, to demonstrate and document the contributions to the mission failure probability.