ABSTRACT
Forecasting the generation of renewable energy is necessary
mainly for the operators of these systems, in the fieeld of wind
energy it is no diffeerent, especially considering the inconstancy
and unpredictability of the wind behavior. Theerefore, this paper
proposes and evaluates diffeerent models based on
machine learning aiming at this problem, with emphasis on a
wind farm located on the coast of Cuba. Thee results of the work
were used to develop an operational monitoring system for
the plant capable of emittiing alarms when the actual energy
generation is diffeerent from the generation made by the model.