DBPapers
DOI: 10.5593/SGEM2014/B41/S17.019

DEVELOPING A MATLAB TOOL FOR PV SYSTEMS ENERGY PRODUCTION FORECASTING USING ANFIS

O. Dragomir, F. Dragomir
Wednesday 1 October 2014 by Libadmin2014

References: 14th International Multidisciplinary Scientific GeoConference SGEM 2014, www.sgem.org, SGEM2014 Conference Proceedings, ISBN 978-619-7105-15-5 / ISSN 1314-2704, June 19-25, 2014, Book 4, Vol. 1, 141-148 pp

ABSTRACT
Providing integrated solutions dedicated to optimizing management of microgrids with distributed power from renewable energy sources (RES), is an important contribution to promoting of clean energy technologies. The solutions involve, among others, the integration of artificial intelligence techniques able of: monitoring, assessing (diagnosis) and estimating (prediction) periods of overload or under load etc. and propose plan effective action in terms of: reducing costs, improving profits or reducing the microgrids vulnerability. Within this scope, a Matlab object oriented application based on adaptive neuro- fuzzy inference systems (ANFIS) was developed to facilitate forecasting energy production from RES. Firstly, the characteristics of ANFIS are briefly described in order to underline the advantages and disadvantages of these types of neuro- fuzzy systems in forecasting approaches. In the second part, are described the methodology and the flowchart used for GUI modeling, as well as, the data used for forecasting of produced energy from RES pre-processing using statistical methods and Matlab. The result of these computations is a data base used, in the third part, for ANFIS training and testing. The proposed graphical user interface (GUI) is tested in order to forecast de energy generation for short term. The effect of ANFIS parameters on the forecasting performances are underlined using root mead square error (RMSE).

Keywords: ANFIS, renewable energy sources- RES, forecasting, PV systems, Matlab