Open AccessHour‐ahead demand forecasting in smart grid using support vector regression (SVR)
Author(s) -
FattaheianDehkordi Sajjad,
Fereidunian Alireza,
GholamiDehkordi Hamid,
Lesani Hamid
Publication year - 2014
Publication title -
international transactions on electrical energy systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.428
H-Index - 42
ISSN - 2050-7038
DOI - 10.1002/etep.1791
Subject(s) - support vector machine , smart grid , electricity market , computer science , artificial neural network , demand response , kernel (algebra) , power grid , demand forecasting , electricity , process (computing) , grid , electric power system , data mining , power (physics) , mathematical optimization , artificial intelligence , operations research , engineering , mathematics , electrical engineering , physics , geometry , combinatorics , quantum mechanics , operating system
Summary Demand forecasting plays an important role as a decision support tool in power system management, especially in smart grid and liberalized power market. In this paper, a demand forecasting method is presented by using support vector regression (SVR). The proposed method is applied to practical hourly data of the Greater Tehran Electricity Distribution Company. The SVR parameters are selected by using a grid optimization process and an investigation on different kernel functions. Moreover, correlation analysis is used to find exogenous variables. Acceptable accuracy of load prediction is shown by comparing the result of SVR model to that of the artificial neural networks and the actual data, concluding that the method is applicable to day‐ahead spot pricing of electricity in the liberalized power market. Copyright © 2013 John Wiley & Sons, Ltd.