File(s) under permanent embargo
Short-term PV power forecasting using hybrid GASVM technique
journal contribution
posted on 2019-09-01, 00:00 authored by W VanDeventer, E Jamei, Gokul Thirunavukkarasu, Mohammadmehdi Seyedmahmoudian, T K Soon, Ben HoranBen Horan, S Mekhilef, A StojcevskiThe static, clean and movement free characteristics of solar energy along with its contribution towards global warming mitigation, enhanced stability and increased efficiency advocates solar power systems as one of the most feasible energy generation resources. Considering the influence of stochastic weather conditions over the output power of photovoltaic (PV) systems, the necessity of a sophisticated forecasting model is increased rapidly. A genetic algorithm-based support vector machine (GASVM) model for short-term power forecasting of residential scale PV system is proposed in this manuscript. The GASVM model classifies the historical weather data using an SVM classifier initially and later it is optimized by the genetic algorithm using an ensemble technique. In this research, a local weather station was installed along with the PV system at Deakin University for accurately monitoring the immediate surrounding environment avoiding the inaccuracy caused by the remote collection of weather parameters (Bureau of Meteorology). The forecasting accuracy of the proposed GASVM model is evaluated based on the root mean square error (RMSE) and mean absolute percentage error (MAPE). Experimental results demonstrated that the proposed GASVM model outperforms the conventional SVM model by the difference of about 669.624 W in the RMSE value and 98.7648% of the MAPE error.
History
Journal
Renewable energyVolume
140Pagination
367 - 379Publisher
ElsevierLocation
Amsterdam, The NetherlandsPublisher DOI
ISSN
0960-1481eISSN
1879-0682Language
engPublication classification
C1 Refereed article in a scholarly journalCopyright notice
2019, Published by Elsevier Ltd.Usage metrics
Keywords
Genetic algorithm (GA)Genetic algorithm based support vectormachine (GASVM)Photovoltaic (PV)Short-term forecastingSupport vector machine (SVM)Science & TechnologyTechnologyGreen & Sustainable Science & TechnologyEnergy & FuelsScience & Technology - Other TopicsGenetic algorithm based support vector machine (GASVM)SUPPORT VECTOR MACHINESOLAR-RADIATIONNEURAL-NETWORKPHOTOVOLTAIC SYSTEMOUTPUTPREDICTIONUTILITYMechanical Engineering
Licence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC