Global horizontal irradiance and wind speed prediction using ANN: Comprehensive Study

Nowadays renewable energy resources such as wind and solar fulfill the requirement of electricity without the emission of greenhouse gases. Wind and solar have a highly irregular and fluctuating nature. It is very difficult to predict them accurately. In this study, we are predicting two meteorological parameters; global horizontal irradiance and wind speed. Additionally, a comparison is carried out among different Artificial neural network (ANN) algorithms; Levenberg marquardt (LM), bayesian regularization (BR), one step secant (OSS), Polack Riebere conjugate gradient (CGP) and Resilient Propagation (RP) in terms of mean square error (MSE), root mean square error (RMSE) and correlation coefficient (R) value. The results show that BR outperformed in terms of MSE (0.3023), RMSE (0.54981), and R-value (0.83533). LM also provided less error with fast speed. The OSS algorithm provided least performance compared to other models i.e. MSE (0.34968), RMSE (0.591337), and R-value (0.80629).