Long-term energy performance forecasting of integrated generation systems by recurrent neural networks

The aim of this paper is to implement a soft computing strategy to improve the long-term energy performance forecasting of stand alone electric generation systems integrated by renewable energy systems as photovoltaic and wind energy. The paper describes the implementation of a dynamic recurrent neural network (RNN) to optimize the long-term energy performance forecasting of integrated generation systems (IGS) and shows its effectiveness in exploiting the large amount of data about an optimal operation of Diesel Groups (DGs) and of renewable generating units as well as on the operating experience of IGSs supplied by highly variable and site-specific renewable energy sources and coupled with different load demand patterns coming from extensive simulation by logistical model.