Accurate solar radiation forecasting is crucial for the stability of photovoltaic power systems, yet current models often blur as prediction time increases. Addressing this challenge, researchers led by Nanjing University of Information Science and Technology have unveiled an AI-based solution called GAN-Solar, designed to generate sharper, more reliable forecasts for solar energy management. by Riko Seibo
Tokyo, Japan (SPX) Oct 16, 2025
Accurate solar radiation forecasting is crucial for the stability of photovoltaic power systems, yet current models often blur as prediction time increases. Addressing this challenge, researchers led by Nanjing University of Information Science and Technology have unveiled an AI-based solution called GAN-Solar, designed to generate sharper, more reliable forecasts for solar energy management.
The model harnesses the principle of Generative Adversarial Networks (GANs), which pit two neural networks against each other in a process likened to a contest between a "master painter" (the generator) and a "keen art critic" (the discriminator). The generator produces simulated future radiation maps from historical data, while the discriminator learns to detect whether the images are genuine or generated.
"Through this continuous adversarial training, the 'painter's' skills are constantly honed, ultimately enabling it to produce high-definition, accurate forecasts that are nearly indistinguishable from reality," explained Chao Chen, lead author of the study published in the International Journal of Intelligent Networks.
Unlike conventional models that lose detail over time, GAN-Solar provides higher fidelity in both global distribution and local features of solar radiation. "Traditional models 'see' less clearly over longer prediction times. GAN-Solar is like equipping the forecast system with a pair of high-precision glasses," Chen said. The improved accuracy supports smoother operation of solar power grids and satellite-linked networks that rely on stable energy input.
Experimental validation shows that GAN-Solar raised the Structural Similarity Index (SSIM) of predicted images from 0.84 to 0.87, outperforming other advanced models. The results demonstrate its capacity to deliver high-precision, low-distortion forecasts essential for real-time solar energy applications and satellite communication networks.
Research Report:GAN-based solar radiation forecast optimization for satellite communication networks
Related Links
Nanjing University of Information Science and Technology
Solar Science News at SpaceDaily