Neural network with genetic algorithms for the monthly electric energy consumption and peak power middle-term forecasting

• Autorzy: Piotrowski P.
• Języki publikacji: EN

Abstrakty

EN

This paper presents efficiency of neural networks with genetic algorithms for solving monthly electric energy consumption and peak power middle - term forecasting problem. Two problems were solved: - middle-term forecasting of monthly electric energy consumption for alt STOEN S.A. consumers - middle-term forecasting of peak power and monthly electric energy consumption for selected groups of big electric power consumers in STOEN S.A. The paper include analysis of the performance, the optimization process of artificial neural network (number of layers, number of neurons in hidden layer(s), types of activation functions, number and type of input - output data) and the efficiency comparison of forecasting using the standard artificial neural network and the artificial neural network with genetic algorithms. At the end of the article there are forecasting results with conclusions. Analysis shows many benefits from precise forecasting for Electrical Plants, Power Regions and groups of electric power consumers as well.

Słowa kluczowe

EN

neural netwok; genetic algorithm; peak power; middle term forecasting; electric energy consumption

PL

sieć neuronowa; algorytm genetyczny; moc szczytowa; prognozowanie średnioterminowe; zużycie energii elektrycznej

• Wydawca: Wydawnictwo Politechniki Łódzkiej
• Czasopismo: Journal of Applied Computer Science
• Rocznik: 2002
• Tom: Vol. 10, nr 1
• Strony: 105--115
• Opis fizyczny: Bibliogr. 12 poz.

Twórcy

autor

Piotrowski P.

• Electrical Power Engineering Institute, Warsaw University of Technology, Koszykowa 75, 00-662 Warsaw, Poland

Bibliografia

• [1] Helt P., Parol M., Piotrowski P.: Metody sztucznej inteligencji w elektroenergetyce, Warszawa, WPW, 2001.
• [2] Iwamoto S., Yasuhiro H.: Long - Term Forecasting Using Improved Recurrent Neural Network, Electrical Engineering in Japan, Vol.114, No.8, 1994, 41-53.
• [3] Makino K, Shimada T.: Daily Peak Electric Load Forecasting Using an Artificial Neural Network and an Improvement Method for Reducing the Forecasting Errors, Electrical Engineering in Japan, Vol.116, No.5, 1996, 28-41.
• [4] Malko J., Mikołajczyk H., Skorupski J.: Sztuczne sieci neuronowe - nowe narzędzie prognozowania elektroenergetycznego, VI International Conference - Aktualne Problemy w Elektroenergetyce, Jurata 1995, 98-108.
• [5] Malko J.: Wybrane zagadnienia prognozowania w elektroenergetyce, Wrocław, Oficyna Wydawnicza Politechniki Wrocławskiej, 1995.
• [6] Osowski S.: Sieci neuronowe, Warszawa, Oficyna Wydawnicza Politechniki Warszawskiej, 1994.
• [7] Piotrowski P.: The monthly electric energy consumption middle - term forecasting using an artificial neural network with genetic algorithms for STOEN joint - stock company, 10lh International Conference on System-Modelling-Control, Zakopane 2001.
• [8] Piotrowski P.; Sztuczna inteligencja - nowe narządzie w elektroenergetyce, miesięcznik STOEN S.A. - “Pod Prąd”, nr 8/1999.
• [9] Piotrowski P., Górski M.: Ekonomiczne aspekty krótkoterminowego prognozowania zużycia energii elektrycznej wybranych grup odbiorców przy wykorzystaniu techniki sieci neuronowych, VIII International Conference “Aktualne problemy w Elektroenergetyce”, Jurata 1999.
• [10] Piotrowski P., Kujszczyk S.: The peak power and energy consumption short - term forecasting using artificial neural network in selected groups of energy consumers, Fourth International Conference on Unconventional Electromechanical and Electrical Systems -UEES'99, Petersburg 1999.
• [11] Zielinski S.: Survey of short - term electrical load forecasting methods, International Conference - Aktualne Problemy w Elektroenergetyce, Jurata 1997,122-129.
• [12] Zielinski S.: Artificial Intelligence in load forecasting, International Conference - Aktualne Problemy w Elektroenergetyce, Jurata 1995, 109-115.