Photovoltaic Forecasting using a Sensor Network of Wireless Optical Cameras

Format

SOECS Senior Project Demonstration

Faculty Mentor Name

Rahim Khoie

Faculty Mentor Department

School of Engineering and Computer Science

Abstract/Artist Statement

As solar energy becomes a more popular means of power generation, the need for solar energy prediction tools also grows. Having the ability to predict future power generation capacity is an important step in developing a smarter power grid. An early warning system for times of low power could help to keep away blackouts by allowing electricity consumers to lower their power usage ahead of time. This project demonstrates the feasibility of a solar prediction system. The project is a working prototype which tracks clouds in order to predict future shadows over a photovoltaic panel. These shadows are what cause drops in power generation ability. The system uses two optical, wireless cameras to take pictures of the sky. Each camera is connected to a rechargeable battery and 10W solar panel to power the camera. The cameras transmit over wireless standard 802.11g to a router. These images are then received by a central computer which uses our own custom image detection software in order to predict the locations of clouds in the future. This software uses both C++ for the image detection and C# for the graphics user interface as well as OpenCV an open source computer vision library. It is our hope that the development of our small scale system will help to foster interest in such prediction systems and further help to incentivize green power production in the future.

Location

School of Engineering & Computer Science

Start Date

30-4-2011 2:00 PM

End Date

30-4-2011 3:30 PM

This document is currently not available here.

Share

COinS
 
Apr 30th, 2:00 PM Apr 30th, 3:30 PM

Photovoltaic Forecasting using a Sensor Network of Wireless Optical Cameras

School of Engineering & Computer Science

As solar energy becomes a more popular means of power generation, the need for solar energy prediction tools also grows. Having the ability to predict future power generation capacity is an important step in developing a smarter power grid. An early warning system for times of low power could help to keep away blackouts by allowing electricity consumers to lower their power usage ahead of time. This project demonstrates the feasibility of a solar prediction system. The project is a working prototype which tracks clouds in order to predict future shadows over a photovoltaic panel. These shadows are what cause drops in power generation ability. The system uses two optical, wireless cameras to take pictures of the sky. Each camera is connected to a rechargeable battery and 10W solar panel to power the camera. The cameras transmit over wireless standard 802.11g to a router. These images are then received by a central computer which uses our own custom image detection software in order to predict the locations of clouds in the future. This software uses both C++ for the image detection and C# for the graphics user interface as well as OpenCV an open source computer vision library. It is our hope that the development of our small scale system will help to foster interest in such prediction systems and further help to incentivize green power production in the future.