Many utility-scale solar farms use horizontal single axis tracking to follow the sun throughout the day and produce more energy. Solar farms on skewed topography produce complex shading patterns that require precise modeling techniques to determine the energy output. To accomplish this, MATLAB was used in conjunction with NREL weather predictions to predict shading shapes and energy outputs. The MATLAB models effectively predicted the sun’s position in the sky, panel tilt angle throughout the day, irradiance, cell temperature, and shading size. The Cal Poly Gold Tree Solar Farm was used to validate these models for various lengths of time. First, the models predicted the shading and power output for a single point in time. Four points of time measurements were taken; resulting in 6 to 32 percent difference in shade height, 5 to 60 percent difference for shade length, and 29 to 59 percent difference for power output. This shows the difficulty of predicting a point in time and suggests the sensitivity of numerous variables like solar position, torque tube position, panel tilt, and time itself. When predicting the power over an entire day, the power output curves for a single inverter matched almost exactly except for in the middle of the day due to possible inaccurate cell temperature modeling or the lack of considering degradation and soiling. Since the backtracking region of the power curve is modeled accurately, the optimization routine could be used to reduce interrow shading and maximize the energy output for a single zone of the solar field. By assuming every day is sunny, the optimization routine adjusted the onset of backtracking to improve the energy output by 117,695 kilowatt hours for the year or 8.14 percent compared to the nominal settings. The actual solar farm will likely never see this increase in energy due to cloudy days but should improve by a similar percentage. Further optimization of other zones can be analyzed to optimize the entire solar field.
| Identifer | oai:union.ndltd.org:CALPOLY/oai:digitalcommons.calpoly.edu:theses-3856 |
| Date | 01 June 2021 |
| Creators | Smith, Logan J |
| Publisher | DigitalCommons@CalPoly |
| Source Sets | California Polytechnic State University |
| Detected Language | English |
| Type | text |
| Format | application/pdf |
| Source | Master's Theses |
Page generated in 0.002 seconds