Premium
Regional midterm electricity demand forecasting based on economic, weather, holiday, and events factors
Author(s) -
Liu Yuxi,
Zhao Jiakui,
Liu Jian,
Chen Yuze,
Ouyang Hong
Publication year - 2020
Publication title -
ieej transactions on electrical and electronic engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.254
H-Index - 30
eISSN - 1931-4981
pISSN - 1931-4973
DOI - 10.1002/tee.23049
Subject(s) - mean absolute percentage error , electricity , electricity price forecasting , demand forecasting , autoregressive model , probabilistic forecasting , electricity market , cluster analysis , investment (military) , econometrics , time series , computer science , consumption (sociology) , electricity demand , operations research , artificial neural network , electricity generation , economics , engineering , power (physics) , artificial intelligence , machine learning , social science , law , sociology , quantum mechanics , political science , physics , politics , probabilistic logic , electrical engineering
Midterm electricity demand forecasting plays an important role in energy management, policy‐making, and investment decisions in regulated or deregulated electricity market. In this study, the proposed electricity demand forecasting method is composed of two stages. In the forecasting stage, historical electricity consumption data are utilized, while four diverse prediction models—support vector machine regression, L 1/2 sparse regression, back‐propagation neural network regression, and autoregressive integrated moving average time‐series analysis—are introduced according to the clustering results of electricity demand curves against 27 provincial power enterprises within State Grid Corporation of China. In the adjustment stage, a mathematical model is developed under a range of economic variables, climatic conditions, holiday periods, important social events, and other drivers, aiming at correcting prediction results in the first stage and improving prediction precision. Extensive experimental study shows that the proposed method can greatly improve the forecasting accuracy of electricity demand consumption in the next 12 months to some extent. The average mean absolute percentage error for the proposed method is approximately 2.56% in the forecasting over the period 2016. © 2019 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.