Machine learning for floating wind foundations

Overview

The analysis of floating wind turbine foundations requires numerous complex simulations to capture full system performance. A floating wind turbine system includes several varying inputs, including turbulent wind and random wave spectra combined with the interaction of the floater, turbine, turbine controller and mooring. This time domain analysis is computationally intensive and requires repetition of hundreds of load cases, each repeated with many wave seeds to capture the full array of possible metocean conditions.

At this event you can find out how machine learning algorithms were used to develop surrogate models that predict selected performance indicator(s) e.g. damage equivalent load (DEL) and how they were taught the behaviour of a floating wind turbine by performing time domain Orcaflex analysis of a range of floater configurations each varying in a key dimension.

You'll discover how, through this surrogate modelling, using a limited data set, results closely aligned with those obtained from OrcaFlex using a full data set can be achieved at a fraction of the computational cost, making this method highly efficient for practical engineering applications.

Please note: booking is not required to attend in person. Please register to attend online.

Programme