DOI: 10.5593/sgem2017/42/S17.059


O. E. Dragomir, F. Dragomir
Tuesday 12 September 2017 by Libadmin2017

References: 17th International Multidisciplinary Scientific GeoConference SGEM 2017, www.sgem.org, SGEM2017 Conference Proceedings, ISBN 978-619-7408-07-2 / ISSN 1314-2704, 29 June - 5 July, 2017, Vol. 17, Issue 42, 467-474 pp, DOI: 10.5593/sgem2017/42/S17.059


This article proposes a software tool, based on artificial intelligence technics, to grid operators and energy companies, to optimize the power system. Precisely, it uses pattern recognition capabilities of neural networks, in order to enable a higher share of renewable energy to consumers. The applied analysis on PV power and other related electrical grid data, such as smart energy meter readings, are based on pattern recognition technics. The neural networks used for this approach will identify days of the week with similar energy consumption profiles.

In this respect, firstly are presented concepts of: pattern recognition and neural networks. Secondly, these intelligent tools are implemented using Matlab programming language, in order to develop a graphical user interface for data monitoring and pattern recognition of load profiles. Thirdly, the software demonstrator is tested using real monitored data provided by Multidisciplinary Science and Technology Research Institute of Valhi University of Targoviste, Romania.

Keywords: pattern recognition, self organisig maps-SOM, neural network, load, energy