Integrated Correction Method of Multi-point Numerical Weather Prediction Wind Speed and Irradiance

Abstract

Numerical weather prediction (NWP) correction is one of the key technologies to improve the accuracy of wind and solar power forecasting. However, there are few researches on the correction of NWP irradiance. At the same time, most of the existing NWP wind speed correction methods only consider a single location, ignoring the spatio-temporal coupling of wind speed, which affects the correction effect. Aiming at this problem, this paper proposed an integrated correction method for multi-point NWP wind speed and irradiance considering the spatial and temporal correlation of regional wind and solar resources. Taking the measured wind speed and irradiance data of multiple wind farms and photovoltaic power plants in the region as the learning objective of the correction model, an integrated NWP correction model based on attentional neural network was established to simultaneously correct multiple station-level NWP data with certain correlation. The proposed method was verified by the NWP data and historical wind speed / irradiance data of 8 wind farms and 7 photovoltaic power plants in a region. The calculation results show that compared with the traditional single-point NWP correction method, the proposed method can effectively improve the accuracy of NWP.

Keywords

• wind farm
• photovoltaic power plant
• spatial and temporal correlation
• numerical weather prediction (nwp)
• power forecasting
• attentional neural network