Coupled High-Resolution Atmosphere Wave Interactions. Evert Wiegant1, Peter Baas1, Remco Verzijlbergh1, Bas Reijmerink2, Sofia Caires2
Total Page:16
File Type:pdf, Size:1020Kb
The new frontier in numerical metocean modelling: PO.146 coupled high-resolution atmosphere wave interactions. Evert Wiegant1, Peter Baas1, Remco Verzijlbergh1, Bas Reijmerink2, Sofia Caires2 1Whiffle Weather Finecasting, 2Deltares Abstract Objectives Numerical modelling of wind and wave parameters are crucial input for design and operation of offshore wind farms. We have developed the world’s first • Show development of a fast coupled LES - spectral wave model coupled atmospheric Large Eddy Simulation (LES) – spectral wave model. This • Explain how coupling is implemented innovative model is capable of capturing wind-wave interaction on the wind farm scale. Influences of bathymetry in the wave field can be observed in the • Illustrate the effects of coupling on wave fields wind field and wind farm wake effects are seen in the wave fields. In a case • Present validation results that show improvements in wave modelling study for a large offshore wind farm, the coupled model shows significant improvements in the wave model results compared to the traditional modelling approach. SWAN: Simulating Waves Near Shore GRASP: GPU-Resident Atmospheric Simulation Platform Model Model State of the art spectral wave model, Whiffle’s operational Large Eddy developed and maintained by TU Delft Simulation (LES) model (Schalkwijk et and Deltares (Cavaleri et al. 2018) al. 2015) Strength Strength Accurately describes diffraction, • GPU based: fast enough for allowing for computation over strong operational forecasting and year- gradients long simulations • Sub 100m resolution: resolves Limitations addressed turbines and turbulence Currently lacks heterogeneous features at high resolution due to coarse wind Figure 2 : Snapshot of a 70m wind field Limitations addressed Figure 1 : Snapshot of the wave field as Currently simple wind-wave interaction input modeled by SWAN. The small cube around the Belgium Offshore Wind denotes the high resolution domain. Farm zone modeled by GRASP. (no sea state) CHASM project: Coupled High-resolution Atmosphere Sea Modelling l The world’s first fully coupled LES – spectral I.C. & forcing GRASP GRASP ECMWF wave model Large-eddy simulation B.C. Large-eddy simulation ERA5 precursor cursor l Resolves wind turbines, turbulence and I.C. & wind-wave interactions B.C. SWAN SWAN Spectral wave model B.C. Spectral wave model large domain small domain l Captures effects of wind farm wakes on waves EMODnet Bathymetry l Consistent wind and wave fields for metocean forecasts and desk studies Figure 3 : Coupled model set-up. ECMWF fields provide initial and boundary conditions. SWAN and GRASP use a nesting structure and exchange wind and wave information in the high-resolution nest. Figure 4 : Snapshot of the wave field around the l Computationally efficient implementation Belgium Offshore Wind Farm zone. on hybrid CPU-GPU system Results Conclusions Case study at large offshore wind farm shows significant improvements in wave • Present day high-performance computers allow for high resolution (~100m) modelling inside the wind farm (Van Eck, 2019) coupled atmosphere sea modelling 20% • The world’s first coupled LES – spectral wave model has been developed 17.74% 16.81% 16.05% and implemented on a hybrid CPU-GPU system 15% 12.35% 11.67% 11.43% • Wind-wave interactions are now modelled in a physically consistent way. 10% Effects of bathymetry, wind farm wakes, turbulence and stability are all accounted for. 5% • Wave model accuracy is improved in and around wind farms 1.13% 0% bias RMSE MAE References -3.72% -3.46% -5% 1. Van Eck, J. G. F. (2019). Large-eddy simulations coupled to a spectral wave model for enhanced metocean modelling 1-way ERA5 wind 1-way GRASP wind 2-way GRASP wind in offshore wind farms. MSc Thesis, Delft University of Technology. 2. Cavaleri, L., Abdalla, S., Benetazzo, A., Bertotti, L., Bidlot, J. R., Breivik, … van der Westhuysen, A. J. (2018). Wave Figure 5 : Model error statistics over a two month period of SWAN wave model against modelling in coastal and inner seas. Progress in Oceanography, 167(February), 164–233. 3. Schalkwijk, J., Jonker, H. J. J., Siebesma, A. P., & Van Meijgaard, E. (2015). Weather forecasting using GPU-based observations of significant wave height inside a large offshore wind farm. large-Eddy simulations. Bulletin of the American Meteorological Society, 96(5) MEET US AT E-C10 windeurope.org/offshore2019 Download the poster #Offshore2019.