Global Horizontal Irradiance Prediction Using Clustering and Artificial Neural Network

Due to the limited supply of fossil fuels and their negative effects on the ecosystem, renewable energy sources, in particular solar energy, are becoming more and more significant. In order to maximize the use of solar energy and increase the effectiveness of solar energy facilities, solar energy forecasting is an essential area of research. In this study, we investigate the application of machine learning methods, particularly linear regression, polynomial regression, and artificial neural networks, for Global Horizontal Irradiance (GHI) prediction which would be important for effective solar energy facilities. Before applying the machine learning methods, we also examine the effects of clustering the data using the K-means algorithm. The accuracy of the different methods is evaluated using mean absolute error (MAE), root mean squared error (RMSE), and R-squared ( \(R^2\) ) values. According to the results, the accuracy of ANN along with K-Means performed better than the other methods, with an MAE of 53.941971, RMSE of 25.018180, and \(R^2\) of 0.966465. Our results indicate that machine learning methods, in particular ANN, can be helpful for precise forecasting of GHI, and clustering the data can further enhance the models’ accuracy. The renewable energy sector may be significantly impacted by these findings, especially in terms of optimizing solar energy use and improving the effectiveness of solar energy facilities.