SERIS conducts research on solar irradiance forecasting at different time scales (minutes, hours and days ahead) for PV applications. Such research is essential for anyone interested in optimizing the O&M activities of solar plants and, subsequently, stabilizing the electricity generation in Singapore.
In order to develop robust solar irradiance forecasting models, it is important to finely monitor the atmospheric properties at high spatial and temporal resolution. In this context, sky cameras are excellent candidates due to their ability to measure a wide range of atmospheric properties such as aerosol load, cloud coverage and cloud types. Combined with appropriate algorithms, a sky camera is also able to evaluate the sky radiance distribution and, ultimately, to estimate the variability of the solar resource at short notice on any oriented surface.
A sky camera can consequently cover a wide range of applications that usually require many specific instruments installed across the roof and façades of a building. A sky camera alleviates the need for separate instruments, reducing the overall cost of an advanced meteorological station.
SERIS has designed a high-quality sky camera to provide a robust, modular and reconfigurable system that meets the most demanding needs of scientific research and industrial applications. The instrument consists of an autonomous, continuously operating, digital imaging and software system designed to capture hemispheric sky images. The sky camera is capable of providing time series retrievals and on-the-fly analysis of sky conditions during daytime hours.
Today, 16 of these sky cameras have been deployed around Singapore. This network is considered as the most advanced high-quality sky camera network dedicated to solar potential and energy meteorology existing worldwide.
Sky images are collected from sunrise to sunset and a database of more than one million sky images are already available. This database and the resulting information are being used as input to AI approaches that learn about typical cloud patterns and their movement.
Therefore, investing in high-quality sky cameras is essential today for building the database of sky images that will be used for training the short-term solar forecasting algorithms of tomorrow.