What RELI-FLOSS delivers
The project is developing a coherent reliability-based design optimisation (RBDO) framework tailored for floating offshore wind.
Key innovations include:
Machine-learning model for environmental data processing
A neural-network-based model was created to process wind and wave datasets supplied by the Offshore Renewable Energy (ORE) Catapult. The workflow covered data normalisation, training/testing separation, network architecture definition, training optimisation and prediction. This enabled accurate modelling of wind and wave data.
High-fidelity wave–structure interaction model
A high-fidelity wave–structure interaction model was developed using ANSYS AQWA to simulate realistic wave-induced responses of FWT support structures and generate dynamic hydrodynamic loads.
Stochastic structural modelling
A stochastic finite-element model was established to capture the combined effects of environmental and gravitational loads. Stochastic variables were introduced to represent material and loading uncertainties, enabling the estimation of stress distributions and identifying critical structural regions.
Reliability-based design optimisation framework
The RBDO framework automates the design iteration process by combining machine learning, stochastic analysis, reliability assessment and optimisation algorithms. It delivers optimised designs that meet a target reliability index while reducing material usage.
Project outputs included a peer-reviewed book chapter, Design Optimisation of Offshore Wind Turbine Support Structures (IntechOpen, November 2024), and a journal article currently in preparation.
The framework provides a systematic alternative to traditional partial safety factor methods.