Digital twin modelling of floating offshore wind turbine with fully coupled aero-hydrodynamic simulation

Yung, Kobe Hoi-Yin and Xiao, Qing and Incecik, Atilla and Thompson, Peter (2023) Digital twin modelling of floating offshore wind turbine with fully coupled aero-hydrodynamic simulation. In: CENSIS (Centre for Sensor and Imaging Systems) Tech Summit 2023, 2023-11-02 - 2023-11-02, Glasgow Royal Concert Hall.

[thumbnail of Yung-etal-CENSIS-2023-Digital-twin-modelling-of-floating-offshore-wind-turbine]
Preview
 Text. Filename: Yung-etal-CENSIS-2023-Digital-twin-modelling-of-floating-offshore-wind-turbine.pdf
Final Published Version
License: Strathprints license 1.0
Download (2MB)| Preview

Abstract

Offshore wind energy demonstrated as reliable energy source over the past decade and contributed to the road of decarbonisation Net Zero 2050. There was 48.2 GW offshore wind capacity already completed by 2021. Particularly, the Floating Offshore Wind Turbines (FOWTs) have received great attention, and it is expected to have higher potential to harvest wind energy than traditional fix-bottom type offshore wind turbines. In terms of the structural integrity and operational requirements, the mooring systems dictate survivability of FOWT under the extreme wave loading and the required station-keeping performance. Due to the extreme cost problems facing in the offshore wind industry, there is a niche for developing a comprehensive Digital Twin (DT) Model that can safeguard the FOWT. Machine Learning (ML) is a prevailing tool for predicting and estimating highly nonlinear dynamics. Yet it lacks physical interpretation between the training input data and target output data. The offline training ML requires a big data library to form the regression and the extreme sea conditions may not be always available. Therefore, a new hybrid Physics-ML DT is proposed here. The Multi-Layer Perceptron Neural Network is trained online to learn the unknown dynamics, with tanh activation function.

CORE (COnnecting REpositories)