Online tuned neural networks for PV plant production forecasting

Author

Ciabattoni, Lucio ; Grisostomi, Massimo ; Ippoliti, Gianluca ; Longhi, Sauro ; Mainardi, E.

Author_Institution

Dipt. di Ing. dell´´Inf., Univ.´´ Politec. delle Marche, Ancona, Italy

fYear

2012

fDate

3-8 June 2012

Abstract

The paper deals with the forecast of the power production for three different PhotoVoltaic (PV) plants using an on-line self learning prediction algorithm. The plants are located in Italy at different latitudes. This learning algorithm is based on a radial basis function (RBF) network and combines the growing criterion and the pruning strategy of the minimal resource allocating network technique. Its on-line learning mechanism gives the chance to avoid the initial training of the NN with a large data set. The performances of the algorithm are tested on the three PV plants with different peak power, panel´s materials, orientation and tilting angle. Results are compared to a classical RBF neural network.

Keywords

learning (artificial intelligence); load forecasting; photovoltaic power systems; power engineering computing; radial basis function networks; resource allocation; PV plant production forecasting; RBF neural network; growing criterion strategy; minimal resource allocating network technique; online self-learning prediction algorithm; online tuned neural networks; photovoltaic plants; power production forecasting; pruning strategy; radial basis function network; tilting angle; Artificial neural networks; Biological neural networks; Forecasting; Neurons; Prediction algorithms; Production; Minimal Resource Allocating Networks; Neural Networks; Production Forecasting; Self learning algorithm;

fLanguage

English

Publisher

ieee

Conference_Titel

Photovoltaic Specialists Conference (PVSC), 2012 38th IEEE

Conference_Location

Austin, TX

ISSN

0160-8371

Print_ISBN

978-1-4673-0064-3

Type

conf

DOI

10.1109/PVSC.2012.6318197

Filename

6318197