Prediction of solar radiation potential in Libya using artificial neural networks

This study explores the software of synthetic neural networks (ANNs) for predicting daily solar radiation in a specific Libyan city. Two famous ANN models-Back propagation Neural Networks (BPNNs) and Radial Basis Function Networks (RBFNs) -were implemented and compared to assess their performance in this area. The have a look at utilized a dataset comprising geographical and meteorological parameters, sourced from NASA's geo-satellite tv for pc database, covering 25 Libyan towns over a length of six years. The consequences validated that RBFNs outperformed BPNNs in phrases of accuracy, processing time, and blunders minimization, with RBFN1 achieving a regression ratio of 93.15% and a minimum suggest squared blunders (MSE) of zero.0090. This performance turned into superior to the first-class-appearing BPNN configuration, which attained a regression ratio of 93% and an MSE of 0.0124. The take a look at highlights the potential of ANNs, specially RBFNs, in growing correct and reliable fashions for solar radiation prediction. These findings make contributions to the wider software of gadget studying techniques in renewable energy forecasting, underscoring the significance of similarly studies to decorate version overall performance and generalization talents. Keywords: Artificial neural network, solar radiation, backpropagation, radial basis function, network, Libya.