What is it about?

Imagine a future where we can accurately predict how much sunlight will reach us, making solar power more efficient than ever. We've developed a smart method that does just that! It's easy to understand, considers all important factors, and even provides a range of possible results for better planning. Tested in various climates and times, it consistently shines brighter than other methods. Simply put, our work is lighting the way for improved solar energy use.

Featured Image

Why is it important?

We've pioneered a method that revolutionizes solar irradiance prediction, a key to efficient solar power plants. Our model stands out by providing accurate and easily understandable forecasts. It considers often overlooked factors like cloud cover and clear sky conditions, and offers a range of potential outcomes, making it a game-changer for planning in the renewable energy sector.


Solar forecasting has been slower to adopt AI and deep learning compared to fields like NLP. Researchers often overlook key factors like weather and geography, affecting prediction accuracy. Our study uses a sophisticated AI algorithm that considers the 'clear sky model', a crucial factor in solar forecasting. This algorithm is accurate and interpretable, making AI forecasts finally understandable.

Institute of Electrical and Electronic Engineering, University M’hamed Bougara, Boumerdes 35000, Algeria

Read the Original

This page is a summary of: Explainable forecasting of global horizontal irradiance over multiple time steps using temporal fusion transformer, Journal of Renewable and Sustainable Energy, September 2023, American Institute of Physics,
DOI: 10.1063/5.0159899.
You can read the full text:



The following have contributed to this page