From the East-European Regional Day-Ahead Markets to a Global Electricity Market

被引:6
作者
Bara, Adela [1 ]
Oprea, Simona-Vasilica [1 ]
Tudorica, Bogdan George [2 ]
机构
[1] Bucharest Univ Econ Studies, Dept Econ Informat & Cybernet, 6 Piata Romana, Bucharest, Romania
[2] Petr Gas Univ Ploiesti, Dept Cybernet Econ Informat Finance & Accountancy, 39 Bucuresti Blvd, Ploiesti, Romania
关键词
Black swan; Electricity price forecast; Regional day-ahead markets; ML algorithms; PRICES;
D O I
10.1007/s10614-023-10416-0
中图分类号
F [经济];
学科分类号
02 ;
摘要
The so-called black swans, COVID-19 and the invasion in Ukraine, have led to an unprecedented increase in electricity prices. Since 2021, after lockdowns, the electricity price has started to increase due to economic recovery, rising prices of the tCO(2) and other primary sources that become unavailable or at much higher prices. In this context, we noticed that the variation of electricity prices in one country can be explained by the price fluctuations of the previous day in the neighboring countries. For instance, the prices for the current day (d) in the Romanian Day-Ahead Market is strongly correlated with the prices of the previous day (d-1) on DAMs of its neighboring countries. It is worth mentioning that the target can be switched by the rest of the variables. Not only the price in Romania can be estimated using the proposed Electricity Price Forecast (EPF) method, but also the prices in other neighboring countries can be a target for prediction because the regional prices on similar markets contain most of society's distress. Another interesting aspect is that the proposed forecasting methodology is robust, as proved by testing it on a varied and longer time interval (from January 2019 to August 2022). Furthermore, the proposed price forecasting methodology includes the adjustment of training interval according to the price standard deviation and weighing the results of the five individual Machine Learning (ML) algorithms to further improve the prediction performance. The set consists of data collected between 1st of January 2019-one year before COVID-19 pandemic outburst and 31st of August-several months after the war has started in the Black Sea region.
引用
收藏
页码:2525 / 2557
页数:33
相关论文
共 58 条
[1]  
Adolfsen J.F., 2022, European Central Bank
[2]   Day-Ahead Electricity Price Forecasting Based on Hybrid Regression Model [J].
Alkawaz, Ali Najem ;
Abdellatif, Abdallah ;
Kanesan, Jeevan ;
Khairuddin, Anis Salwa Mohd ;
Gheni, Hassan Muwafaq .
IEEE ACCESS, 2022, 10 :108021-108033
[3]  
AlmaBetter, 2024, Random Forest Algorithm in Machine Learning
[4]   Impact of fossil fuel prices on electricity prices in Mexico [J].
Bernal, Bruno ;
Carlos Molero, Juan ;
Perez De Gracia, Fernando .
JOURNAL OF ECONOMIC STUDIES, 2019, 46 (02) :356-371
[5]   Random forests [J].
Breiman, L .
MACHINE LEARNING, 2001, 45 (01) :5-32
[6]  
Brownlee J., 2020, How to Develop a Random Forest Ensemble in Python
[7]  
Brownlee J., 2020, How to develop voting ensembles with Python
[8]   Mid-term energy cost-oriented flow shop scheduling: Integration of electricity price forecasts, modeling, and solution procedures [J].
Busse, Jan ;
Rieck, Julia .
COMPUTERS & INDUSTRIAL ENGINEERING, 2022, 163
[9]  
Butler N., 2022, IMPACT UKRAINE WAR G
[10]   Forecasting Electricity Prices: A Machine Learning Approach [J].
Castelli, Mauro ;
Groznik, Ales ;
Popovic, Ales .
ALGORITHMS, 2020, 13 (05)