Transforming Electroluminescence Images to Current-Voltage (IV)Curves Using Deep Learning
3 page evaluation abstract for PVSC 52
3 page evaluation abstract for PVSC 52
Conference Record of the IEEE Photovoltaic Specialists Conference
PV systems using horizontal single axis trackers (SATs) generate more energy than PV systems on fixed racking. The most common method of positioning a SAT minimizes the angle of incidence between the PV and the sun while avoiding row-to-row shading, thereby increasing energy production. Another approach, optimal position tracking, uses beyond sun position to attempt to capture the most irradiance possible under conditions where sun tracking is irrelevant (e.g., when a cloud obscures the sun). We present a study displaying the function of our SunDelay algorithm for optimal position tracking in Albuquerque, New Mexico and Eugene, Oregon. The SunDelay algorithm utilizes sun detection from sky imagery, location, and time data to position the panel optimally. The SunDelay algorithm tracks the sun naively throughout the day when the sun is not obscured. Once the sun is obscured in a series of sky images SunDelay makes a decision to either stay tracking naively or position the panel horizontally (towards zenith) to collect more diffuse irradiance. The SunDelay algorithm is compared against naive tracking through simulation, using data obtained from a Multi Planar Irradiance Sensor. The SunDelay algorithm collecting up to 0.2% and 4.7% more irradiance in NM and OR respectively over a month.