A Novel Hybrid Method for Solar Power Prediction

Author(s): MD Rahat Hossain
Supervisor(s): Dr. Amanullah Maung Than Oo; Dr. ABM Shawkat Ali

Field of Research: Information and Computing Sciences
Research Organisation: School of Engineering and Technology

Abstract

Solar energy is judged as potential power producing resource because of its accessibility and geographical benefits in local power productions. Still a negative aspect, to solar choice, is its intermittent nature and dependence on climate variation. Solar energy resource, unlike dispatchable central station generation, produce power dependable on external irregular source and that is the incident solar radiation which does not always radiate when electricity is needed. This results in the variability, unpredictability, and uncertainty of solar energy supply. Consequently, the accurate or precise prediction of solar power presents a major challenge to distribution and transmission grid operators because knowing how much electricity installations will produce over the next certain specific period of time is the only way to optimally integrate large scale solar electricity into power grid operations. The involvement of renewable sources with storages make it mandatory to precisely predict the gains and the loads because based on that precise prediction control decision is made. None the less, such hybrid forecasting has significant impact on the optimum power flow, transmission congestion, power quality issues, system stability, load dispatch, and economic analysis. However, with the increased complexity in contrast to single power prediction systems, most hybrid prediction system, particularly heterogeneous regression algorithms or machine learning techniques based hybrid prediction systems turn out to be most challenging and complex. The thesis presents the detailed description and analysis of different experimental results to develop the hybrid prediction method for solar power to best possible accuracy. It is expected that the outcome of the research will provide noteworthy contribution to the relevant research field as well as to the Australian power industries in the near future.

Keywords: Computational intelligence, Heterogeneous regressions algorithms, Performance evaluation, Hybrid method, Mean absolute scaled error (MASE)
Timeline: 31 March 2013

Project Contacts

Name: MD Rahat Hossain
Contact Research Organisation: Institute for Resource Industries and Sustainability
Phone: 07 4923 2068
Email: m.hossain@cqu.edu.au