Prediction of Solar Photovoltaic Energy Output Based on Thin-Film Technology Utilizing Various Machine Learning Techniques

This paper presents solar photovoltaic (PV) energy prediction based on thin-film technology utilizing various machine learning (ML) models. Several ML models like Support Vector Machine (SVM), Extra Tree Regression (ETR), Decision Tree Regression (DTR), K-Nearest Neighbour (kNN) and Feed-Forward Neural Network (FFNN) were utilized to evaluate each model's performance according to performance metrics. The primary input parameters such as time, solar radiation, wind speed, ambient and PV module temperatures, and the actual power generated by the thin-film PV panel based on the 2018 data set were considered for predicting solar PV output power. The ETR is proposed to predict the PV power output in this work and compared with other ML models. The results showed that ETRs outperformed the different ML algorithms, whereas DTR performed the poorest. The ETR model had the best performance, with Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) values of 59.17 and 39.07, respectively. On the other hand, the DTR model performed poorly, with an RMSE of 81.83 and an MAE of 52.9, respectively.