Sky Images and PV Power Measurements for Irradiance Forecasting

Published: 01-02-2021| Version 2 | DOI: 10.17632/cb8t8np9z3.2
Duane Robinson,
Ashish Agalgaonkar,
Roshan Godaliyadda,
Parakrama Ekanayake,
Sarath Perera,
Janaka Ekanayake


Ground-based sky images related to sunny, partially cloudy, and overcast sky conditions are available in this data set with corresponding power measurements that were obtained from two rooftop PV systems. The sky camera systems were made using Raspberry Pi mini-computers (3rd Generation Model B) and using programmable high-resolution Pi-Cameras with fisheye lenses. The sky camera systems were operated remotely and were programmed to automatically capture images from 8.00 am to 4.45 pm at a rate of 10 seconds and store them in JPEG format with a resolution of 1024x768 pixels. The RPI camera was assumed to have an equidistant mapping function. To collect the data, one camera was installed near a rooftop PV system (Location 1) at the Sustainable Buildings Research Centre (SBRC) at the Innovation Campus, University of Wollongong, Australia. The second camera was installed 1.9 km away (Location 2) from the first camera and it was installed at the roof of "Building 35" at the Main Campus of the University of Wollongong, Australia. Further corresponding PV power measurements were collected from Location 1 and Locations 3 ([34.41°N, 150.88°E] rooftop PV system at "Building 28", the main campus, University of Wollongong) which is 2 km away from camera 1 (Location 1 [34.40°N, 150.90°E]). Power measurements were acquired from both locations with a 1-minute resolution. Data sets obtained from camera 1 can be used for onsite PV forecasting, whereas using both data sets obtained from camera 1 and camera 2, PV forecasts for multiple PV sites can be derived via calculating cloud base heights.