Risk assessment of FPSO swivel stack systems using a multi-source data fusion approach

Abstract

This paper addresses the risk analysis of Floating Production Storage and Offloading (FPSO) swivel stack systems, crucial components of Single Point Mooring (SPM) systems. Traditional risk assessment methods face challenges in data fusion, uncertainty handling, and dynamic prediction. To overcome these limitations, a multi-source data fusion approach is proposed. A field monitoring system for the swivel stack systems captures statistical distribution patterns, enhancing prior data diversity. Dynamic Bayesian networks (DBNs) and interval type-2 fuzzy sets (IT2FS) model uncertainty. A SAM-IT2TrFN aggregation algorithm is proposed to aggregate expert fuzzy opinions and obtain prior probabilities. Additionally, the Noisy-OR gates method is combined with traditional AND-OR logic gates to construct conditional probability tables (CPTs). Weibull distribution is employed to construct state transition probability tables for dynamic nodes. Validation is conducted through a case study in the Bohai Bay, identifying key risk factors and proposing control measures. Dynamic probability analysis reveals the evolution of the risk system over time. The results demonstrate the method's comprehensive risk evaluation, with the SAM-IT2TrFN aggregation algorithm exhibiting superior computational performance, especially in handling divergent expert opinions.