Forecasting of Power Output of a PVPS Based on Meteorological Data Using RNN Approaches

Mohsen Beigi, Hossein Beigi Harchegani, Mehdi Torki, Mohammad Kaveh, Mariusz Szymanek, Esmail Khalife and Jacek Dziwulski
Additional contact information
Mohsen Beigi: Department of Mechanical Engineering, Tiran Branch, Islamic Azad University, Tiran 8531911111, Iran
Hossein Beigi Harchegani: Institute for Higher Education, Academic Center for Education, Culture, and Research (ACECR), Ahvaz 6139688839, Iran
Mehdi Torki: Department of Computer Engineering, Faculty of Electrical and Computer Engineering, Technical and Vocational University (TVU), Tehran 1435761137, Iran
Mohammad Kaveh: Department of Petroleum Engineering, College of Engineering, Knowledge University, Erbil 44001, Iraq
Mariusz Szymanek: Department of Agricultural, Forest and Transport Machinery, University of Life Sciences in Lublin, 20-612 Lubin, Poland
Esmail Khalife: Department of Civil Engineering, Cihan University-Erbil, Kurdistan Region, Erbil 44001, Iraq
Jacek Dziwulski: Department of Strategy and Business Planning, Faculty of Management, Lublin University of Technology, 20-618 Lublin, Poland

Sustainability, 2022, vol. 14, issue 5, 1-12

Abstract: Artificial intelligence (AI) has become increasingly popular as a tool to model, identify, optimize, forecast, and control renewable energy systems. This work aimed to evaluate the capability of the artificial neural network (ANN) procedure to model and forecast solar power outputs of photovoltaic power systems (PVPSs) by using meteorological data. For this purpose, based on the literature review, important factors affecting energy generation in a PVPS were selected as inputs, and a recurrent neural network (RNN) architecture was established. After completing the trained network, the RNN capability was assessed to predict the energy output of the PVPS for days not included in the training database. The performance evaluation of the trained RNN revealed a regression value of 0.97774 for test data, whereas the RMSE and the mean actual output power for a sample day were 0.0248 MJ and 0.538 MJ, respectively. In addition to RMSE, an error histogram and regression plots obtained by MATLAB were employed to evaluate the network’s capability, and validation results represented a sufficient prediction accuracy of the trained RNN.

Keywords: artificial intelligence; clean energy; historical data; short-term forecasting; recurrent neural network (search for similar items in EconPapers)
JEL-codes: O13 Q Q0 Q2 Q3 Q5 Q56 (search for similar items in EconPapers)
Date: 2022
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (2)

Downloads: (external link)
https://www.mdpi.com/2071-1050/14/5/3104/pdf (application/pdf)
https://www.mdpi.com/2071-1050/14/5/3104/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jsusta:v:14:y:2022:i:5:p:3104-:d:765785

Access Statistics for this article

Sustainability is currently edited by Ms. Alexandra Wu

More articles in Sustainability from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().