# WindDragon: Enhancing wind power forecasting with Automated Deep Learning

**Authors:** Julie Keisler, Etienne Le Naour  
**arXiv:** [2402.14385](https://arxiv.org/abs/2402.14385)  
**Date:** 2024-02-22  
**Summary page:** https://www.summarizepaper.com/en/paper/2402.14385v1/

## Key points

- Integration of wind power into power grids is crucial for achieving net zero carbon emissions by 2050
- Variability and uncertainty of wind power present challenges for system operators
- Accurate forecasting of wind power is essential for effective grid operation and system balancing
- Groundbreaking approach to short-term windpower forecasting developed at a national level using Automated Deep Learning technology and Numerical Weather Predictions wind speed maps
- Aim to provide accurate forecasts within a 1 to 6-hour horizon
- Research showcases how cutting-edge technologies can overcome challenges in integrating increasing amounts of wind power into existing grids
- Application of Automated Deep Learning enhances forecast accuracy and contributes towards sustainable energy goals and mitigating climate change impacts

## Summary

The integration of wind power into power grids is crucial in achieving net zero carbon emissions by 2050. However, the variability and uncertainty of this renewable energy source present challenges for system operators. Accurate forecasting of wind power is essential for effective grid operation and system balancing. To address this need, a groundbreaking approach to short-term windpower forecasting has been developed at a national level. The innovative method outlined in the paper "WindDragon: Enhancing wind power forecasting with Automated Deep Learning" leverages Automated Deep Learning technology and Numerical Weather Predictions wind speed maps. By combining these advanced techniques, researchers aim to provide accurate forecasts within a 1 to 6-hour horizon. Authored by Julie Keisler from EDF R&D OSIRIS and Etienne Le Naour from ISIR, this study showcases how cutting-edge technologies can be harnessed to overcome challenges in integrating increasing amounts of wind power into existing grids. The application of Automated Deep Learning not only enhances forecast accuracy but also contributes towards achieving sustainable energy goals and mitigating climate change impacts. This research represents a significant advancement in renewable energy forecasting and highlights the potential for leveraging technology to optimize grid operations and facilitate the transition towards a more sustainable future.