Identification of clouds using an all-sky imager

With the increasing number of utility-scale photovoltaic power plants, the uncertainty and volatility in power generation increases in the electricity system. In the range of long to medium term forecasting, the literature and the industry have several tools and methods in process. Nevertheless, the current and future energy transition, for instance the shorter gate closer of energy markets, is increasing the need for tools to support the ultra-short-term forecasting. This paper proposes a tool to support the ultra-short-term forecasting using an all-sky imager to map the clouds in the sky and estimate a Cloud Index. Three threshold-based methodologies are presented and tested, they are able to detect the clouds and categorize them as thick clouds and thin clouds.