Webinar Series Part 23:
When AI meets the sky: All sky imagers – from basic research to energy applications
This webinar showcases innovative cloud observation and classification methods and their applications in renewable energy forecasting. Mehdi Ben Slama from Reuniwatt presents advances in cloud type classification using all-sky imagers, including novel approaches to estimate multi-layer cloud amounts from single images. Emanuele Giovanni Carlo Ogliari from Politecnico di Milano discusses machine learning methods for short-term solar radiation nowcasting based on sequences of infrared all-sky images, highlighting improvements over traditional persistence models. Prof. Stephanie Fiedler leads the event, emphasizing how these advanced techniques enhance solar radiation forecasts critical for electric grid stability and management. The webinar offers insights into cutting-edge atmospheric science developments and their practical implications, with a final discussion on cloud observation, modeling, and solar radiation prediction.
Topics discussed
- Innovative methods for cloud observation and classification using all-sky imagers, including techniques for estimating multi-layer cloud amounts from a single image
- Machine learning-based approaches for nowcasting global horizontal irradiance 5-15 minutes ahead
- How advanced cloud observation and prediction techniques contribute to more accurate solar radiation forecasts, crucial for managing electric grids and renewable energy systems
Moderator: Prof. Stephanie Fiedler, Professor, Directorate Institute of Environmental Physics @Universität Heidelberg
Speakers:
Emanuele Ogliari, PhD, Associate Professor, Department of Energy @Politecnico di Milano
Mehdi Ben Slama, Engineer – R&D Computer Vision @Reuniwatt