Návrh neurónovej siete pre predikciu spotreby elektrickej energie

Article Sidebar

PDF
Published Nov 23, 2020

Main Article Content

Marek Novák
Žilinská univerzita v Žiline, Katedra elektroenergetiky a elektrických pohonov, Univerzitná 1, Žilina, SK 01026, Slovenská Republika
Pavol Belány
Žilinská univerzita v Žiline, Katedra elektroenergetiky a elektrických pohonov, Univerzitná 1, Žilina, SK 01026, Slovenská Republika
Peter Braciník
Žilinská univerzita v Žiline, Katedra elektroenergetiky a elektrických pohonov, Univerzitná 1, Žilina, SK 01026, Slovenská Republika
Marek Roch
Žilinská univerzita v Žiline, Katedra elektroenergetiky a elektrických pohonov, Univerzitná 1, Žilina, SK 01026, Slovenská Republika

Abstract

Predikcia spotreby elektrickej energie je v posledných rokoch významnou témou. Nárast množstva obnoviteľných zdrojov energie má dopad na disharmóniu medzi výrobou a spotrebou. Preto predikčné modely či už spotreby alebo výroby elektrickej energie nachádzajú svoje uplatnenie stále častejšie. Hlavným cieľom tohto článku je predstaviť návrh predikčného modelu spotreby elektrickej energie založenom na neurónovej sieti. Model overovaný na základe dát z inteligentných meracích systémov vykazuje veľmi presnú predpoveď pri krátkej dobe tréningu a agregovaní spotreby elektrickej energie odberných miest.
Kľúčové slová: predikcia; neurónová sieť, spotreba elektrickej energie
Prediction of electricity consumption has been an important topic in recent years. Increasing amount of renewable energy sources has an impact on the disharmony between production and consumption. Therefore, prediction models of consumption or production of electricity are finding their application more and more often. The main goal of this paper is to present a proposal for a predictive model of electricity consumption based on a neural network. The model verified on the basis of data from intelligent metering systems shows a very accurate prediction with a short training time and aggregation of electricity consumption of consumption points (Design of a neural network for prediction of electricity consumption)
Keywords: prediction; neural network, electricity consumption

Article Details

Issue
Vol 13 No 1 (2020)
Section
Smart Grids and Electromobility
Authors of the contribution acknowledge that the contributed article has not been up to the day published in any journal, neither has it been submitted for publishing in another journal. At a time, Authors delegate to the editorial board of ELEKTROENERGETIKA journal the authority to publish the contributed article on the journal Internet WEB page (http://jeen.fei.tuke.sk) at understanding that the journal editorial board assumes no responsibility for damages suffered due to use of acts, methods, products, instructions for use or other ideas published by the article authors, neither for by them advertised information. Authors represent that they are fully aware of the fact that articles are not being linguistically corrected by the journal editorial staff, and that they hence assume unlimited responsibility for lingual correctness.