A Data-Driven Approach for Localization and Power Generation Estimation of Invisible Solar Resources

In recent years, due to the rapid increase in the number of solar panels, a portion of the daily power demand is being met by the energy generated by these solar panels. This study presents a data-driven approach to estimatehow much of the demand can be met by off-grid solar panels. The research consists of two phases. In the first phase, the detection of rooftop solar panels from aerial photographs is worked on using U-net image segmentation. The use of U-net for solar panel detection is evaluated with various efficiency parameters. In the second phase, the goal is to predict the power generation of these solar panels using machine learning algorithms and climate data. Power prediction is performed using five different machine learning algorithms, and these algorithms are compared. In conclusion, this study takes an important step in evaluating the role of solar panels in energy production and understanding the future potential of solar energy usage. Such technological advancements can enhance sustainability in the energy sector and reduce environmental impacts.

Author

• Sahin, Esma

Language

• en

Keyword

• invisible
• Localization
• solar
• off-grid
• power
• estimation

Date created

• 2024

Type of Work

• Master's Thesis

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Electrical Engineering

Degree

• M.S.

Advisor

• Dehghanian, Payman

Committee Member(s)

• Korman, Can
• Ahmadi, Shahrokh

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/1z40kt66c