SeaFD: Realistic CFD wind load computations for offshore wind turbines

The goal of SeaFD is to identify and predict the type of extreme weather events characteristic of the North Sea that are potentially leading to failures of bearing in wind turbines. A new high-resolution weather model will be developed and validated against true scale measurements on offshore wind turbines. It will allow the prediction of local weather conditions at each rotor of a wind farm 36 hours ahead.

SeaFD will aim at extending the lifetime of the offshore wind turbines by:

1. Identifying weather systems leading to degradation of drivetrain bearing material. Historical and new wind and wave data near the Northwind offshore wind farm will be analysed and correlated to rotor torque, rotor bending and extreme bearing stresses of the MaSiWEC turbine in order to identify the type of weather phenomena that needs to be addressed in the risk assessment of the operation of offshore wind farms.
2. Creating a high-resolution weather model tuned for North Sea wind farms. To improve the understanding of extreme weather events and their interaction with wind farms, global historical weather datasets from ECMWF and NCAR will be coupled to mesoscale weather forecasting modelling (WRF), and to Large Eddy Simulation (LES), to compute wind fields approaching wind turbines, and wind fields within wind farms, at turbine scales. Wind turbines will be modelled as actuator lines to include wind turbine wake effects, and to allow for simulating the wind loading per turbine.
3. Validating the new weather model by measurements in the Northwind offshore wind farm. Several weather events will be selected by the steering group, recomputed by the new weather model and compared against field wind data. Parkwind will provide the basic data of all turbines in the Northwind farm. VUB and VKI will install LIDARs, in collaboration with ENGIE and 3E. The Flemish MOW Agentschap Maritieme Dienstverlening en Kust will make available wave data from Sea Buoys near the wind farm. Once the optimum models settings have been found, the high-resolution weather model is ready to be used for weather predictions at long and at short term (1.5 days ahead), thus improving the mapping of new and risk management of running offshore wind farms, respectively.
4. Predicting real-time wind turbine loading, using a lower-order model for wind park operation. With the help of the weather model, CFD RANS simulations, SCADA 1 sec data, LIDAR and wave buoy data, a reduced-order modelling for prediction of wind fields and wind turbine loading will be set up for each wind turbine using machine-learning approach. The resulting tool is aimed to be used for wind park operation.