Napovedovanje cen in povpraševanja po električni energiji

magistrsko delo

Katarina Brilej (Author), Mihael Perman (Mentor)

Abstract

Napovedovanje porabe električne energije je priljubljena tema že kar nekaj časa, elektrike se namreč v večjih količinah ne da shraniti, stabilnost elektroenergetskega sistema pa zahteva stalno ravnovesje med proizvedeno in porabljeno električno energijo. Z deregularizacijo trga električne energije je dragoceno orodje postalo tudi napovedovanje cen. Magistrsko delo ponuja pristop k napovedovanju cen in povpraševanja po električni energiji na podlagi modelov ARMA-GARCH (področje časovnih vrst) in metode podpornih vektorjev (področje strojnega učenja). Pri tem gre za ceno pri trgovanju za dan vnaprej (ang. day ahead trading). Tako v primeru cene kot porabe električne energije so podatki dani na urni ravni, napovedujemo pa jih za 24 ur naprej. Ključni del predstavlja izbira in identifikacija modela, s katerim kasneje izdelamo napovedi. Tako ceno kot porabo napovemo le z uporabo preteklih vrednosti cene in porabe, brez dodatnih zunanjih spremenljivk. Identifikacija poteka po Box-Jenkinsovi metodologiji, ki predstavlja standardno orodje v analizi časovnih vrst. Pri vsakem pristopu se osredotočimo na nekaj različic modelov, optimalne parametre pa izberemo s pomočjo različnih kriterijev. Pri modelih časovnih vrst sta to kriterij AIC in Ljung-Boxov test, pri metodi podpornih vektorjev pa prečno preverjanje. Na koncu z izbranimi modeli napovemo prihodnje vrednosti cene in porabe električne energije. Modele med seboj primerjamo na podlagi različnih mer kakovosti modelov (MAPE, RMSE). V obeh primerih rezultati kažejo, da so za dane podatke napovedi modelov ARMA-GARCH boljše. V primeru porabe električne energije izrazitega favorita ni, v primeru cene pa izstopa model ARMA-GARCH-t.

Keywords

napovedovanje povpraševanja po električni energiji;napovedovanje cene električne energije;modeli ARMA;modeli GARCH;metoda podpornih vektorjev;

Data

Language:	Slovenian
Year of publishing:	2021
Typology:	2.09 - Master's Thesis
Organization:	UL FMF - Faculty of Mathematics and Physics
Publisher:	[K. Brilej]
UDC:	519.22
COBISS:	75691011
Views:	1227
Downloads:	315
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Electricity price and demand forecasting
Secondary abstract:	Electricity demand forecasting has been a popular topic for quite some time, as electricity cannot be stored in large quantities, whereas the stability of the electric power system requires a constant balance between electricity produced and consumed. With the deregulation of the electricity market, price forecasting has become valuable. This master’s thesis offers an approach to electricity price and demand forecasting based on ARMA-GARCH models (time series) and support vector machines (machine learning). We focus on the day-ahead electricity prices. In both electricity price and demand, the analyzed data are available hourly, and our models provide 24-hour forecasts for the next day. The crucial part in the elaboration of the forecasts is the model selection and identification. Both price and demand are predicted using only past values of price and demand, without additional external variables. The selection of models follows the Box-Jenkins methodology, a standard tool in time series analysis. We focus on a few (similar) models in each approach, and the optimal parameters are selected using different criteria. For time series models, these are the AIC criterion and the Ljung-Box test, and for the support vector machines, we use cross-validation. Finally, we predict the future values of electricity price and demand with the previously selected models. We compare the models based on different measures for forecast accuracy (MAPE, RMSE). In both cases, the results show that the predictions of the ARMA-GARCH models are better for the given data. In the case of electricity demand, no model is considerably better than the rest, and in the case of price, the ARMA-GARCH-t model stands out.
Secondary keywords:	electricity demand forecasting;electricity price forecasting;ARMA models;GARCH models;support vector machines;
Type (COBISS):	Master's thesis/paper
Study programme:	0
Thesis comment:	Univ. v Ljubljani, Fak. za matematiko in fiziko, Oddelek za matematiko, Finančna matematika - 2. stopnja
Pages:	XII, 86 str.
ID:	13349324