Open AccessPhotovoltaic power forecasting using statistical methods: impact of weather data
Author(s) -
De Giorgi Maria Grazia,
Congedo Paolo Maria,
Malvoni Maria
Publication year - 2014
Publication title -
iet science, measurement and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.418
H-Index - 49
eISSN - 1751-8830
pISSN - 1751-8822
DOI - 10.1049/iet-smt.2013.0135
Subject(s) - kurtosis , skewness , photovoltaic system , artificial neural network , computer science , standard deviation , power (physics) , time series , meteorology , statistics , mathematics , engineering , artificial intelligence , machine learning , electrical engineering , quantum mechanics , physics
An important issue for the growth and management of grid‐connected photovoltaic (PV) systems is the possibility to forecast the power output over different horizons. In this work, statistical methods based on multiregression analysis and the Elmann artificial neural network (ANN) have been developed in order to predict power production of a 960 kW P grid‐connected PV plant installed in Italy. Different combinations of the time series of produced PV power and measured meteorological variables were used as inputs of the ANN. Several statistical error measures are evaluated to estimate the accuracy of the forecasting methods. A decomposition of the standard deviation error has been carried out to identify the amplitude and phase error. The skewness and kurtosis parameters allow a detailed analysis of the distribution error.