Application of SADA method on full-scale measurement data for dynamic responses prediction of Hywind floating wind turbines

Abstract

Floating offshore wind turbines (FOWTs) are promising solutions for offshore renewable energy harvesting, with the successful installation and operation of the world's first commercial floating wind farm, Hywind Scotland, in 2017. However, both academia and industry are still constantly facing challenges in the aspects of cost reduction, monitoring, safety and sustainability improvement for the design and maintenance of FOWTs. The purpose of this paper is to demonstrate an engineering application of a novel Artificial Intelligence knowledge-based method, named SADA, on the full-scale measurement data of an Hywind FOWT. The SADA method was applied to perform numerical optimization and dynamic responses prediction of the FOWTs, based on the full-scale data from one Hywind FOWT in Scotland. The methodology of SADA and the key technology of the application are introduced firstly. Then, the selection of Key Discipline Parameters (KDPs) is introduced, followed with the training of AI-based numerical models with full-scale measurement data, including Floater motions, wind, wave and current data. After that, the numerical model imbedded in SADA is trained to be intelligent for the objective Hywind FOWT under different sea states. The intelligent SADA model is used to do comparisons and predictions. The comparison results show that using SADA method, the AI-trained numerical model can predict the motions of Hywind supporting Floater in higher accuracy. In addition, other physical quantities that cannot be obtained directly in full-scale measurement easily but are of great concern by industry, can also be obtained from a more believable perspective. This AI-based SADA method brings an innovative vision for FOWTs' full-scale measurement technology in the future.