Instantaneous 2D extreme wind speed prediction using the novel Wind Gust Prediction Net based on purely convolutional neural mechanism

Abstract

Accurate prediction of spatial–temporal extreme wind gust is vital for the wind farm dynamic regulation, the floating wind turbine deployment and its early warning. Deep-learning approaches have been applied for wind prediction to alleviate the computational challenges of traditional numerical models. Yet, most previous studies emphasized the prediction accuracy only employing location-specific dataset, such methodologies are site-specific and ignore the importance of spatial–temporal fidelity. Furthermore, the Recurrent Neural Networks (RNN)-based approach previous employed exhibit low efficiency in terms of model convergence and on the aspect of practical engineering purposes. This study firstly proposed the wind gust prediction net (WGPNet), using residual learning with attention modulations to predict the instantaneous spatial–temporal wind gust in the West Pacific region with great potential wind-energy. And a public reanalysis dataset with very high resolution (0.25° x 0.25°) was employed to verify the proposed method under different criteria. The overall RMSE of predicted gust fields obtained by the proposed method dropped to 0.18 m/s. Comprehensive discussions with both temporal and spatial perspective, revealing that the proposed model can offer an accurate 2D wind gust prediction along timeline (the PCC equals to 0.98).

Keywords

• Spatial–temporal wind gust forecasting
• onshore and offshore gust
• 2D convolutional networks
• autocorrelation analysis
• spatial–temporal distribution