Artificial Intelligence-Based Solar Power Generation Prediction and Analysis

Integration of renewable energy sources with the conventional grid is the key way to sustain the world with the limited conventional sources and also open the door to the green energy system. Solar, wind, hydro and many more are the renewable energy sources connected with the grid; however, its optimization is one of the crucial aspects as the input to these sources is variable and intermittent in nature. This paper deals with the prediction of solar power plant located at Ahmedabad. The simulation work has been carried out with System Advisor Model (SAM) software. Various parameters like temperature, solar irradiance and generated output power have been analysed through this study. Various Artificial Intelligence (AI) algorithms, Linear regression (LR) model, artificial neural network (ANN) model, support vector machines (SVM) model, random forest (RF) model, gradient boosting (GB) models has been simulated, analysed and compared in results w.r.t the root means square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE) and correlation coefficient (CC) and fond that for small amount of the power plant data provided here, ANN and SVM fits best as they provide 1 to 10 % error as compared to rest of all errors. CC gives result 1 except SVM algorithm and LR model provides linear relations.