Considering the development of low-carbon and sustainable alternatives to traditional energy sources is a global goal

many countries aim to have non-fossil energy sources account for over 80% of their installed capacity and achieve carbon neutrality by 2060. However

to ensure the stable operation of power grids with a high penetration of renewable energy

it is essential to precisely predict output fluctuations

which requires accurate forecasting of parameters such as wind speed and solar radiation intensity. Currently

wind speed forecasting is largely based on relatively inaccurate numerical weather prediction models. We aim to address this limitation by developing an intelligent wind speed forecasting method based on image generation using generative adversarial networks (GANs). First

a target dataset is created using thin plate spline (TPS) interpolation to correct initial gridded forecasts of the wind speed based on observed wind speed data. The wind speed data are then transformed into gridded data images

and a GAN is used to generate intelligent wind speed forecasts. To validate the proposed method

a simulation analysis is performed using actual data from a region in North China

and the results are found to be 21.7% more accurate than those of the European Centre for Medium-Range Weather Forecasts.