A Machine Learning-based Short- Term Prediction Model for a Solar Plant: A South African case study

The research presented in this paper examines and develops a machine learning-based model to predict the hourly output of a rooftop solar installation in a petrochemical industrial facility in South Africa. This research attempts to thoroughly examine the factors impacting solar energy generation and then create a reliable predictive model for solar power. This work uses machine learning to produce an accurate and dependable model to increase the effectiveness of energy management systems. It was established that solar power output and efficiency depend on various environmental factors and the year's season. The results show RMSE values of 4.30, 4.22, and 11.89 for the 1D CNN, LSTM and hybrid CNN-LSTM models, respectively.