Prediction of Solar Energy Generation Using Machine Learning

Establishing solar power plants has become more viable option in the energy sector and make us to less reliant on fossil fuels for the development of the economy and society. We require a precise forecast of solar energy generation in order to manage the use of solar energy. An approach for forecasting solar energy is presented in this paper which is based on automated methods. The usefulness of the investigated models was assessed in order to forecast continuous energy produced by the solar plant while taking weather conditions into consideration. The datasets we used in the study include one that shows solar generation data for 34 days starting in 2020 and another that shows weather conditions on those days relative to time. The model considers every examination of faults and anomalies that arise during the production of solar energy. LR and RF shows how well the model fits the data by evaluating the R-square value, which is a measurement of the proportion of the target variable’s fluctuation that can be predicted from the variables that are independent (features) included in a regression model. High R-squared values for both models indicate that they are able to adequately account for the majority of the variability in the target variable. The model demonstrates a strong relationship between the independent variables and the target variable when it captures and predicts a significant portion of the variance of the variable being targeted.