AI algorithm can now optimize solar power performance- UNM Researchers

The performance of solar power has always been a debatable topic across various industrial spheres over a while now. However, in one of the latest accomplishments in the energy industry, researchers of the University of New Mexico have come up with an artificial intelligence algorithm that enhances the performance of solar power by forecasting cloud cover.

This step comes on heels of the Obama administration’s 2016 pledge of imposing about 26 to 28% reduction in GHG emissions by 2025. In pursuance of this goal, Department of Electrical and Computer Engineering Ph.D. candidate Guillermo Terren-Serrano in association with Professor Manel Martinez Ramon of University of New Mexico and NM EPSCoR, are closely working on addressing some of the challenges associated with solar energy.

As per credible reports, the algorithm was trained using solar radiation sensor and cameras mounted on the UNM campus. Reportedly, the camera system was developed by Martinez Ramon and Terren Serrano to follow the sun throughout the day, just like a sunflower, gathering data on both cloud cover and solar radiation at the same time.

Furthermore, managing the information collected by these cameras require enhanced storage capabilities than what a personal computer is capable of providing.

Moreover, both the researchers have been using the resources offered by the Center for Advanced Research Computing to process the data accumulated from the camera system in a bid to train the AI algorithm.

Once fully trained using the collected camera data, the AI algorithm would be able to estimate future solar output based on the current weather conditions and situations. This would eventually give users enough time to allow a backup source of energy in case power outage occurs.

Importantly, both, Martinez Ramon and Terren Serrano, are currently preparing three more journal papers for submission on this groundbreaking invention. In fact, they plan to launch a website later this year which would anyone to view the data from their cameras in real-time.

Source Credit: