Publication detail

FORECASTING WEEKLY ELECTRIC LOAD USING A HYBRID FUZZY-NEURAL NETWORK APPROACH

KHAN, M. ŽÁK, L. ONDRŮŠEK, Č.

Czech title

Předpověď týdenní spotřeby elektrické energie s využitím fuzzy neuronové sítě.

English title

FORECASTING WEEKLY ELECTRIC LOAD USING A HYBRID FUZZY-NEURAL NETWORK APPROACH

Type

journal article - other

Language

en

Original abstract

A hybrid approach utilizing a fuzzy system and artificial neural network (ANN) for short-term load forecasting of the Czech Electric Power Company (ČEZ) is proposed in this paper. Expert knowledge represented by fuzzy rules is used for preprocessing input data fed to a neural network. For training the neural network for one-week ahead load forecasting, fuzzy ‘If-Then’ rules are used, in addition to historical load and temperature data that are usually employed in conventional supervised training algorithms. The fuzzy-neural network is trained on real data for the years 1994 through 1998 and evaluated on the data for the year 1999 for forecasting next-week load profiles. A very good prediction performance is attained as shown in the simulation results, which verify the effectiveness and superiority of the modeling technique.

Czech abstract

Článek popisuje využití fuzzy neronové sítě pro předpověď spotřeby elektrické energie na další období.

English abstract

A hybrid approach utilizing a fuzzy system and artificial neural network (ANN) for short-term load forecasting of the Czech Electric Power Company (ČEZ) is proposed in this paper. Expert knowledge represented by fuzzy rules is used for preprocessing input data fed to a neural network. For training the neural network for one-week ahead load forecasting, fuzzy ‘If-Then’ rules are used, in addition to historical load and temperature data that are usually employed in conventional supervised training algorithms. The fuzzy-neural network is trained on real data for the years 1994 through 1998 and evaluated on the data for the year 1999 for forecasting next-week load profiles. A very good prediction performance is attained as shown in the simulation results, which verify the effectiveness and superiority of the modeling technique.

Keywords in English

One-week ahead load forecasting, Multilayer neural networks and Hybrid fuzzy-neural networks (FNN)

RIV year

2004

Released

26.11.2001

ISSN

1210-2717

Journal

Inženýrská mechanika - Engineering Mechanics

Volume

2001

Number

8

Pages count

55

BIBTEX


@article{BUT40535,
  author="Muhammad R {Khan} and Libor {Žák} and Čestmír {Ondrůšek},
  title="FORECASTING WEEKLY ELECTRIC LOAD USING A HYBRID FUZZY-NEURAL NETWORK APPROACH",
  journal="Inženýrská mechanika - Engineering Mechanics",
  year="2001",
  volume="2001",
  number="8",
  month="November",
  issn="1210-2717"
}