Predicting Linearised Wind Resource Grids using Neural Networks

Modelling the flow over terrain is a key element of wind resource assessments within energy industry. Existing modelling methods range from fast, low fidelity analytical models to time-consuming and computationally expensive high-fidelity Computational Fluid Dynamics (CFD) software. In this work, Grid-Kernel Neural Network approach has been developed used create surrogate emulate WAsP software, by calculating changes in speed direction due orography roughness terrain. This data-driven proven be successful predicting orographic at multiple heights above ground. At 100 m ground, mean absolute error values were 1.6% speedup 0.4° for changes, respectively. Although model linear, potential solver, findings here can counted as first step towards creating fully CFD model.