Prediction of Short-Term and Long-Term Hourly Global Horizontal Solar Irradiation Using Artificial Neural Networks Techniques in Fez City, Morocco

摘要

The integration of renewable energy plants into the energy mix leads to serious problems in maintaining the balance of electricity grids. Indeed, renewable energy plants can produce electricity when there is not much need. Therefore, predicting renewable energy potentials and then the output of power plants can allow grid operators to prepare decision scenarios in advance. In this work, we are interested in predicting hourly short-term (h + 1) and long term (h + 48) global horizontal solar irradiation (GHI) by applying two types of Artificial Neural Networks (ANN): Multilayer Perceptron (MLP) and a Nonlinear AutoRegressive neural network with exogenous inputs (NARX).