Climate Artificial Intelligence for Hydropower
Smart Climate Hydropower Tool is a web service intended to support Hydropower energy producers facing the problem of energy forecast.
Their decision making processes (either concerning plants managing issues or energy trading) will benefit from enhanced seasonal energy forecasts.
We embedded Artificial intelligence algorithms (supervised learning techniques) at the core of the service to feed energy forecast with available state of art seasonal forecast, and guide users through a user friendly web interface. Working in tied connection with end users for an effective codesign process, developed service will exploit value of seasonal forecast, clearly show performances and added value of the provided energy forecasts and ideally pave the road for highly scalable and worldwide replicable similar services.
The tool is actually developed under EU H2020 funded project called CLARA (EU FP7 project No 730482) coordinated by CMCC (http://www.clara-project.eu/) and with the support of Enel Green-Power
SCHT is a web-based climate service aimed at supporting hydropower suitability and production assessment. The climate service intends to improve the decision-making process for two objectives:
a) energy production and trading
b) hydropower management energy operations.
Also, the service identified the regional and national institutional settings, as well as regional environmental agencies (ARPAs) as potential users
The forecast system relies on data driven artificial intelligence (AI) algorithms, partly already applied in published research (Callegari, et al., 2015, De Gregorio et. al 2017) with encouraging results.
Currently SCHT exploit a wide set of AI algorithm from machine learning-ML predictors (support vector regression- SVR and gaussian process-GP) to deep learning (LSTM and RNN).
Major operative advantages of AI with respect to mechanistic hydrological models include limited to none a priori knowledge of involved physical phenomena, high level of flexibility when managing heterogeneous sets of variables related to discharge generation, and quick setup time of the forecast system.