Embedded Files
Tropical Cyclones
GANs for typhoon predictions trained with satellite data
GANs for typhoon predictions trained with satellite data
A generative adversarial network was developed to predict the future evolution and trajectory of typhoons directly from satellite images. Combining satellite imagery with atmospheric velocity fields significantly improved the prediction of typhoon tracks, demonstrating the potential of physics-informed data-driven methods for environmental flow prediction.
GANs for typhoon prediction trained with satellite and flow data
GANs for typhoon prediction trained with satellite and flow data
A generative adversarial network was developed to predict both the track and intensity of typhoons by combining satellite observations with meteorological data. The approach enabled rapid 6- and 12-hour forecasts and demonstrated how combining observational and reanalysis data can improve data-driven prediction of complex atmospheric flows.
Atmospheric Winds
Neural networks for improving wind power efficiency
Neural networks for improving wind power efficiency
Neural-network approaches were reviewed for wind-turbine control, wind-farm optimization, and aerodynamic blade design, highlighting how data-driven models can improve wind-energy efficiency and accelerate CFD-based engineering.
Exploiting spatiotemporal correlations in wind data
Exploiting spatiotemporal correlations in wind data
It was shown how spatial and temporal correlations in atmospheric flows influence neural-network-based wind prediction, linking regional and seasonal flow characteristics to the learnability.
Page updated
Google Sites
Report abuse