Catalog Home Page

Sky camera network and data acquisition system development for the purpose of Short-term Photo-Voltaic power output forecasting

Burton, Ashton (2017) Sky camera network and data acquisition system development for the purpose of Short-term Photo-Voltaic power output forecasting. Honours thesis, Murdoch University.

PDF - Whole Thesis
Download (6MB) | Preview


A sky camera network was installed at Murdoch University capturing high resolution whole sky images every ten seconds for a study, to be completed by the University of Oldenburg, investigating short term solar irradiance forecasting[1]. The study investigates the effectiveness of sky camera imaging and processing techniques used to determine forecasted cloud induced variations on photo-voltaic (PV) power systems based on images and validation data acquired at Murdoch University in Perth, Western Australia. This project demonstrates the effectiveness of two 180-degree fish-eye Vivotek network security cameras recording 1920x1920 pixel RGB images simultaneously at 10 second intervals to provide sufficient information for accurate solar irradiance and PV power output modelling. Results from the collaboration between Murdoch University and the University of Oldenburg were presented at the World Renewable Energy Congress which provide a detailed analysis of solar forecasting accuracy and the associated fuel savings in comparison to other forms of generation and PV management systems. A persistence method of forecasting was determined to properly utilize sky camera forecasting’s ability to predict future cloud events and reduce the spinning reserve (SR) that is required in PV and diesel power systems. The results showed that a total of 61% of the time between 10th November and 28th December 2016 the model identified clear sky irradiance. A total of 84 cloud events occurred during this period and 76 of these were correctly identified giving a total miss rate of ~8% [2]. Based on these findings, a further study was conducted by the University of Oldenburg which determined estimated fuel savings achieved with a sky camera based forecasting control method. The findings showed up to 5.5% reduction of fuel consumption in a PV diesel power network with 60% of total capacity being installed PV[3].

The forecasting method used was outlined in a report (West et al 2014) and variations of the image processing methods required were investigated at Murdoch University to explore potential alternatives to more complicated and computationally expensive methods. Various further research is proposed and the effectiveness of alternative methods is presented. The sky camera network and data acquisition system that was developed is discussed in detail and shown to be capable of providing the information required for the forecasting and validation of cloud induced power variations in PV systems.

Publication Type: Thesis (Honours)
Murdoch Affiliation: School of Engineering and Information Technology
Supervisor: Lee, Gareth
Item Control Page Item Control Page


Downloads per month over past year