Short-term PV power forecasting using empirical mode decomposition in integration with back-propagation neural network. Issue 1 (2nd January 2020)

• Record Type: Journal Article
• Authors: Yadav, Harendra Kumar; Pal, Yash; Tripathi, Madan Mohan
• Is Part Of: Journal of information & optimization sciences. Volume 41:Issue 1(2020)
• Page Start: 25
• Page End: 37
• Publication Date: 2020-01-02
• Subjects: 49QXX; Solar radiation -- PV power -- Forecasting -- EMD -- Back-propagation neural network -- Optimization -- PV power forecasting; Electronic data processing -- Periodicals; Information science -- Periodicals; Mathematical optimization -- Periodicals; 519.6
• DOI: 10.1080/02522667.2020.1714181
• Languages: English
• ISSNs: 0252-2667

Abstract

Abstract: Solar panel photovoltaic (PV), grid-connected and off-grid connected systems are promptly increasing in India, to enrich the solar power generation. Solar power generation is one of the furthermost encouraging sources of renewable energy sources. Accurate PV-power forecasting is an essential requirement for electricity companies and grid operators to increase the commitment of units, profits, planning of energy transmissions, scheduling maintenance and planning of supply and demand in the electricity grid. The perfect prediction of PV energy is an important task because it is dependent on solar radiation, which is uncontrollable and weather dependent. In this paper, an innovative PV power forecasting method has been developed by integrating empirical mode decomposition (EMD) and back-propagation neural network (BPNN).EMD is used to decompose PV time series into five intrinsic mode functions (IMF's) and a residue. Then, each IMF and residue is used to train the back-propagation neural network (BPNN). The proposed EMD-BPNN method is estimated on PV power dataset collected from 100 kW roof-top grid-connected solar plant. The proposed EMD-BPNN method is progressing better than some recently reported methods for predicting PV power in terms of computational accuracy and complexity.