Dynamic Bayesian Network-based online fatigue assessment for mooring lines of floating offshore wind turbines

Abstract

This paper presents a novel approach for the online fatigue assessment of mooring lines of Floating Offshore Wind Turbines (FOWTs) by integrating Dynamic Bayesian Networks (DBN) and Particle Filters (PF). The proposed DBN-PF methodological framework aims to address the complexities of sequential data analysis inherent in monitoring fatigue accumulation under dynamic offshore conditions. Validated against experimental results derived from fatigue loading tests performed on mooring line materials, the DBN-PF framework demonstrates its applicability to both polyester ropes and mooring chains under constant and variable amplitude fatigue loading. This study applies this method to the online fatigue assessment of the mooring lines of the International Energy Agency 15 MW FOWT. This involves a comprehensive aero-hydro-elastic-servo dynamic analysis to simulate operational conditions and generate input data. The results indicate that the DBN-PF methodology provides a reliable diagnosis and prognosis of fatigue damage, with predictions showing good agreement with the conventional code-based method. The significant finding highlights that the DBN-PF framework demonstrated strong predictive performance for cumulative fatigue damage even when provided with limited data, underscoring its potential for efficient real-time fatigue monitoring and its contribution towards enhancing the reliability and maintenance strategies for FOWTs.