Manuscript received October 20, 2022; revised December 11, 2022; accepted Mach 7, 2023.
Abstract—Energy is essential to facilitate the social and economic growth of a society. But this energy generation using fossil fuels results in a tremendous amount of greenhouse gas emissions. As a renewable alternative, the increased competitiveness of solar PV panels has increased the number of solar energy generation stations in recent years. Since solar power generation is highly intermittent and dependent on local weather characteristics, AI can be implemented to predict solar energy output from a solar power plant. As the need to predict solar photovoltaic (PV) energy output is essential for many actors in the energy industry, Statistical Data Analysis and Machine Learning (ML) can be employed towards this end. In this study, comparative analysis of different machine learning models is performed to estimate power-plant solar energy generation from historical meteorological data. A variety of supervised machine learning techniques are implemented to predict and forecast solar energy. The implemented models include Weighted Linear Regression (WLR) with and without dimensionality reduction, Gradient Boosting Model (GBM), and Artificial Neural Networks (ANN). Findings indicate that, both the ANN and GBM models performed significantly well in short-term prediction, whereas Long-Short Term Memory (LSTM) Recurrent Neural Network (RNN) achieved reliable performance in forecasting. The trained models, therefore, may provide a way for grid-operators to Predict and balance energy generation and consumption.
Index Terms—Machine learning, solar energy, prediction, forecasting
Nifat Sultana is with University of Dhaka, Bangladesh.
Tasnim Ahmed is with Rakuten Mobile, Inc., Japan.
*Correspondence: firstname.lastname@example.org (T.A.)
Cite: Nifat Sultana and Tasnim Ahmed, "Performance Analysis of Machine Learning Models in Solar Energy Forecasting," International Journal of Machine Learning vol. 13, no. 3, pp. 131-135, 2023.Copyright @ 2023 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).