Prediction of Power Generation of a Photovoltaic Power Plant Based on Neural Networks

Photovoltaic energy production is an important factor for increasing the electricity supply. The ability to predict electric power (EPP) of a photovoltaic (PV) farm supports from management process grid trade in market and much more. Also, by predicting PV (PVP), it possible monitor lifetime solar cells that form backbone any system. As critical result, sudden failures plant can be avoided. Using long short-term memory recurrent neural network (LSTM-RNN) model, this work evaluates prediction accuracy two forecasting strategies: recursive strategy non-recursive Multiple-Input Multiple-Output, respectively. dataset consists 5-years in-filed data measurements collected CETATEA plant, research site facility renewable energies located Cluj-Napoca, Romania. high granularity values recorded each 1 hour guarantees overall impact size, number previous observations, forecast horizon on evaluated strategy. performance metrics used evaluate are root mean square error, bias average error. results analysis demonstrates implemented machine learning models production, as well their importance loss process.