A research team based in the United States has pioneered an innovative solar forecasting methodology that leverages a blend of infrared (IR) imagery and global solar irradiance measurements. This groundbreaking approach is touted to enhance the accuracy of solar nowcasting and intra-hour forecasting, making it particularly relevant for photovoltaic (PV) real-time markets and the optimization of energy distribution within microgrids.
Guillermo Terrén-Serrano, the corresponding author of the research, pointed out that this method offers an economical alternative to traditional ceilometers. He mentioned that while typical all-sky imagers primarily operate in the visible light spectrum, they fall short when it comes to approximating cloud heights. The researchers' solution, which involves the use of a radiometric infrared camera, data logger, high-resolution solar tracker, pyranometer, outdoor computer, weatherproof casing, visible light fisheye, weather sensors, and camera lenses, costs less than $2,000. In comparison, ceilometers can set you back around $20,000.
Ordinarily, visible light cameras are employed to capture ground-based sky images, aiding PV models in responding to changing cloud conditions. Nonetheless, these cameras suffer from pixel saturation caused by the intense sunlight, rendering them less effective for accurate solar forecasting. The team's alternative employs IR cameras that mitigate the pixel saturation issue associated with visible light cameras.
Nevertheless, IR-based forecasting introduces its own set of challenges, including a lower signal-to-noise ratio, primarily due to potential distortions caused by solar irradiance. The research addresses this by introducing efficient data processing methods that remove the deterministic component of global solar irradiance from pyranometer measurements and IR images.
To mitigate the impact of irradiance, the novel method begins by employing machine learning techniques to identify biases affecting the clear sky index (CSI). CSI measures the influence of clouds on global solar irradiance, and refining the accuracy of this measure results in more precise forecasting. Subsequently, another algorithm categorizes IR sky images into four distinct sky conditions: clear sky, cumulus clouds, stratus clouds, and nimbus clouds. Using this classification, the algorithm interacts with GSI data to calculate the impact of irradiance on the image, effectively enhancing its suitability for forecasting.
Additionally, the algorithm is designed to counter the influence of camera dirt. Recognizing that sky imagers may not undergo daily cleaning during operation, the researchers propose an image processing-based method to eliminate radiation emitted by debris on the outdoor germanium camera window from IR images.
The algorithms were trained and evaluated using data from Albuquerque, New Mexico, which boasts an arid semi-continental climate with minimal precipitation. The research team acknowledges the need for future research to develop a global model that's applicable to various locations.
In conclusion, the researchers assert that their proposed method is efficient and could pave the way for low-cost radiometric IR cameras to replace expensive ceilometers in the future. They emphasize the significance of streamlined data processing in simplifying learning algorithms, increasing forecasting accuracy, and reducing computing time, particularly in real-time applications like nowcasting and intra-hour solar energy forecasting.
These groundbreaking findings are presented in the study titled "Processing of global solar irradiance and ground-based infrared sky images for solar nowcasting and intra-hour forecasting applications," recently published in Solar Energy. The research was conducted by a team from the University of California Santa Barbara and the University of New Mexico.
Post a Comment