A Bayesian approach to risk modeling of autonomous subsea intervention operations

Abstract

The introduction of autonomy in subsea operations may affect operational risk related to Inspection, Maintenance, and Repair (IMR). This article proposes a Bayesian Belief Network (BBN) to model the risk affecting autonomous subsea IMR operations. The proposed BBN risk model can be used to calculate the probability of aborting an autonomous subsea IMR operation. The nodes of the BBN are structured using three main categories, namely technical, organizational, and operational. The BBN is tested for five unique scenarios using a scenario generation methodology for the operational phase of the autonomous IMR operation. The BBN is quantified by conducting a workshop involving industry experts. The results from the proposed model may provide a useful aid to human supervisors in their decision-making processes. The model is verified for five scenarios, but it is capable of incorporating and calculating risk for other combinations of scenarios.