Русская версия English version
Home / 2018 issue 6 / Choice of the method for solving problems of short-term forecasting of CHPP auxiliary power consumption

Choice of the method for solving problems of short-term forecasting of CHPP auxiliary power consumption

A.S. Vedernikov, E.A. Yarygina, A.V. Gofman

Vestnik IGEU, 2018 issue 6, pp. 32—38

Download PDF

Abstract in English: 

Background. The scientific problem of this research is prediction of power consumption of power plant auxiliary supply with a minimum error. Short-term forecasting problems have been previously solved at the level of electric power systems and industrial enterprises. As for forecasting auxiliary power consumption, retrospective data on power consumption have been used as forecast values. This problem remains relevant in accordance with RF Government Decree No. 1172 of December 27, 2010, which states that power plants assume responsibility for the consumed electricity exceeding the established limits. A 2 or more % deviation in consumption from the set value leads to additional costs. The aim of the study is, therefore, to select a method of predicting values of CHPP auxiliary power consumption with a low error.

Materials and methods. To solve the problems of short-range forecasting, we have chosen a method based on artificial neural networks and trained these networks by using numerical optimization methods: the Broyden-Fletcher-Goldfarb-Shanno learning algorithm, Conjugate gradient method, Gradient descent method which have been earlier used for solving various practical problems in the electric power industry. The software package Statistica Neural Networks has been used by us to determine the hourly values of the electric load of CHPP auxiliary needs.

Results. We have chosen the method based on the artificial neural networks «multilayer perceptron» and the Broyden-Fletcher-Goldfarb-Shanno algorithm for its training, which allows CHPPs to predict auxiliary power consumption with the mean absolute error of 0,43 %.

Conclusions. The proposed method of short-term prediction of auxiliary power consumption at CHPP has been tested and approved at the Branch of JSC «SO UES» of the Interregional Dispatching Office of the Middle Volga for estimating predicted power consumption values of power plants in the process of electricity balance planning.

References in English: 

1. Osovskiy, S. Neyronnye seti dlya obrabotki informatsii [Neural networks for information processing]. Moscow: Finansy i statistika, 2002. 344 p.

2. Barron, A.R. Statistical properties of artificial neural networks. Proceedings of the IEEE International Conference on Decision and Control. NewYork, 1989, рр. 280–285.

3. Alekseeva, I.Yu., Vedernikov, A.S., Skripachev, M.O. Prognozirovanie elektropotrebleniya s ispol'zovaniem metoda iskusstvennykh neyronnykh setey [Forecasting of power consumption using the method of artificial neural networks]. Vestnik SamGTU. Seriya Tekhnicheskie nauki, 2010, рр. 135–138.

4. Gofman, A.V. Usovershenstvovanie metodiki prognozirovaniya mnogonomenklaturnogo predpriyatiya s uchetom proizvodstvennoy deyatel'nosti [Improvement of the methods of forecasting a multiproduct enterprise taking into account its production activity]. Sbornik nauchnykh trudov Mezhdunarodnoy nauchno-tekhnicheskoy konferentsii «Elektroenergetika glazami molodezhi» [Collected scientific works of the International scientific and technical conference “Electric Power Engineering in the Youth’s Eyes]. Samara: SamGTU, 2011, vol. 2, pp. 153–157.

5. Barron, A.R. Complexity regularization with applications to artificial neural networks. In G. Roussas (ed.) Nonparametric Functional Estimation. Boston, MA and Dordrecht, the Netherlands: Kluwer Academic Pub-lishers, 1990, pp. 561–576.

6. Solov'eva, I.A., Dzyuba, A.P. Prognozirovanie elektropotrebleniya s uchetom faktorov tekhnologicheskoy i rynochnoy sredy [Prediction of power consumption taking into account factors of the technological and market environment]. Nauchnyy dialog, 2013, no. 7(19), pp. 97–113.

7. Vedernikov, A.S., Balukova, E.A. Opredelenie korrelyatsionnykh zavisimostey mezhdu faktorami, vliyayushchimi na elektropotreblenie sobstvennykh nuzhd TETs [Determination of correlation dependences between factors influencing CHPP electric power consumption]. Izvestiya vuzov. Elektromekhanika, 2016, no. 6, pp. 110–113.

8. Galushkin, А.I. Teoriya neyronnykh setey. Kn. 1 [Theory of neural networks. Book 1]. Moscow: IPRZhR, 2001. 385 p.

9. Vedernikov, A.S., Balukova, E.A., Komasova, E.A. Postroenie neyronnykh setey dlya prognozirovaniya elektropotrebleniya sobstvennykh nuzhd TETs [Construction of neural networks for forecasting CHPP auxiliary power consumption]. Materialy VIII Mezhdunarodnoy nauchno-tekhnicheskoy konferentsii: «Elektroenergetika glazami molodezhi». V 3 t., t. 3 [Proceedings of the VIII International Scientific and Technical Conference: «Electric Power Engineering in the Youth’s Eyes». In 3 vol., vol. 3]. Samara: Samarskiy gosudarstvennyy tekhnicheskiy universitet, 2017, pp. 72–78.

10. Benn, D.V., Farmer, E.D. Sravnitel'nye modeli prognozirovaniya elektricheskoy nagruzki [Comparative models of electric load forecasting]. Moscow: Energoatomizdat, 1987. 200 p.

11. Cybenko, G. Approximations by superpositions of a sigmoidal function. Mathematics of Control, Signals, and Systems, 1989, no. 2, pp. 303–314.

12. Kurbatskiy, V.G., Tomin, N.V. Prognozirovanie elektricheskoy nagruzki s ispol'zovaniem iskusstvennykh neyronnykh setey [Forecasting of electric load using artificial neural networks]. Elektrika, 2006, no. 7, pp. 26–32.

13. Barskiy, A.B. Neyronnye seti: raspoznavanie, upravlenie, prinyatie resheniy [Neural networks: recognition, management, decision-making]. Moscow: Finansy i statistika, 2004. 176 p. (Prikladnye informatsionnye tekhnologii).

14. Borovikov, V.P. Neyronnye seti. Statistica Neural Networks. Metodologiya i tekhnologii sovremennogo analiza dannykh [Neural networks. Statistica Neural Networks. Methodology and technologies of modern data analysis]. Moscow: Goryachaya liniya – Telekom, 2008. 392 p.

15. Khalafyan, A.A. STATISTICA 6. Statisticheskiy analiz dannykh [STATISTICA 6. Statistical analysis of data]. Moscow: Binom-Press, 2007. 512 p.

16. Vedernikov, A.S., Yarygina, E.A., Khamitov, R.N. Programma «Prognozirovanie elektropotrebleniya sobstvennykh nuzhd TETs na osnove neyronnoy seti» [Computer program «Forecasting of thermal power plant auxiliary power consumption based on neural networks»]. Svidetel'stvo o registratsii elektronnogo resursa № 23603 [Certificate of registration of digital resource No. 23603]. Moscow: OFERNIO, 2018.

 

Key words in Russian: 
прогнозирование электропотребления, собственные нужды ТЭЦ, искусственные нейронные сети, алгоритмы обучения, ошибка прогноза
Key words in English: 
power consumption forecasting, CHPP auxiliary power, artificial neural networks, learning algorithms, forecast error
The DOI index: 
10.17588/2072-2672.2018.6.032-038
Downloads count: 
33