Regresar

Short-term active power forecasting of a photovoltaic power plant using an artificial neural network

Abstract:

The increasing use of solar energy as a renewable source for electricity generation through photovoltaic power plants has led the interest of their power production forecasts tending to facilitate the management and optimization of this valuable renewable resource. This paper presents a short-term active power forecasting model based on an artificial neural network (ANN). The data used correspond to time series of meteorological variables, such as wind speed, radiation, relative humidity, among others, and electrical, such as power. The raw data are preprocessed to account for missing values and outliers, and the design of the artificial neural network considers variable selection to utilize the best input variables to the model, and a suitable number of layers, number of neurons, learning algorithm and transfer function.