24th EANN 2023, 14 - 17 June 2023, León, Spain

Wind Energy Prediction Guided by Multiple-Location Weather Forecasts

Charalampos Symeonidis, Nikos Nikolaidis


  In recent years, electricity generated from renewable energy sources has become a significant contributor to power supply systems over the world. Wind is one of the most important renewable energy sources, thus accurate wind energy prediction is a vital component of the management and operation of electric grids. This paper proposes a novel method for wind energy forecasting, which relies on a novel variant of the scaled-dot product attention mechanism, for exploring relations between the generated energy and a set of multiple-location weather forecasts/measurements. The conducted experimental evaluation on a dataset consisting of the hourly generated wind energy in Greece along with hourly weather forecasts for 18 different locations, demonstrated that the proposed approach outperforms competitive methods.  

*** Title, author list and abstract as seen in the Camera-Ready version of the paper that was provided to Conference Committee. Small changes that may have occurred during processing by Springer may not appear in this window.