A hybrid dynamic and fuzzy time series model for mid-term power load forecasting

Woo Joo Lee, Jinkyu Hong

Research output: Contribution to journalArticle

66 Citations (Scopus)

Abstract

A new hybrid model for forecasting the electric power load several months ahead is proposed. To allow for distinct responses from individual load sectors, this hybrid model, which combines dynamic (i.e., air temperature dependency of power load) and fuzzy time series approaches, is applied separately to the household, public, service, and industrial sectors. The hybrid model is tested using actual load data from the Seoul metropolitan area, and its predictions are compared with those from two typical dynamic models. Our investigation shows that, in the case of four-month forecasting, the proposed model gives the actual monthly power load of every sector with only less than 3% absolute error and satisfactory reduction of forecasting errors compared to other models from previous studies.

Original languageEnglish
Pages (from-to)1057-1062
Number of pages6
JournalInternational Journal of Electrical Power and Energy Systems
Volume64
DOIs
Publication statusPublished - 2015 Jan 1

Fingerprint

Time series
Dynamic models
Air
Temperature

All Science Journal Classification (ASJC) codes

  • Energy Engineering and Power Technology
  • Electrical and Electronic Engineering

Cite this

@article{bf268e65215942e2ae08ad8555b2666c,
title = "A hybrid dynamic and fuzzy time series model for mid-term power load forecasting",
abstract = "A new hybrid model for forecasting the electric power load several months ahead is proposed. To allow for distinct responses from individual load sectors, this hybrid model, which combines dynamic (i.e., air temperature dependency of power load) and fuzzy time series approaches, is applied separately to the household, public, service, and industrial sectors. The hybrid model is tested using actual load data from the Seoul metropolitan area, and its predictions are compared with those from two typical dynamic models. Our investigation shows that, in the case of four-month forecasting, the proposed model gives the actual monthly power load of every sector with only less than 3{\%} absolute error and satisfactory reduction of forecasting errors compared to other models from previous studies.",
author = "Lee, {Woo Joo} and Jinkyu Hong",
year = "2015",
month = "1",
day = "1",
doi = "10.1016/j.ijepes.2014.08.006",
language = "English",
volume = "64",
pages = "1057--1062",
journal = "International Journal of Electrical Power and Energy Systems",
issn = "0142-0615",
publisher = "Elsevier Limited",

}

A hybrid dynamic and fuzzy time series model for mid-term power load forecasting. / Lee, Woo Joo; Hong, Jinkyu.

In: International Journal of Electrical Power and Energy Systems, Vol. 64, 01.01.2015, p. 1057-1062.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A hybrid dynamic and fuzzy time series model for mid-term power load forecasting

AU - Lee, Woo Joo

AU - Hong, Jinkyu

PY - 2015/1/1

Y1 - 2015/1/1

N2 - A new hybrid model for forecasting the electric power load several months ahead is proposed. To allow for distinct responses from individual load sectors, this hybrid model, which combines dynamic (i.e., air temperature dependency of power load) and fuzzy time series approaches, is applied separately to the household, public, service, and industrial sectors. The hybrid model is tested using actual load data from the Seoul metropolitan area, and its predictions are compared with those from two typical dynamic models. Our investigation shows that, in the case of four-month forecasting, the proposed model gives the actual monthly power load of every sector with only less than 3% absolute error and satisfactory reduction of forecasting errors compared to other models from previous studies.

AB - A new hybrid model for forecasting the electric power load several months ahead is proposed. To allow for distinct responses from individual load sectors, this hybrid model, which combines dynamic (i.e., air temperature dependency of power load) and fuzzy time series approaches, is applied separately to the household, public, service, and industrial sectors. The hybrid model is tested using actual load data from the Seoul metropolitan area, and its predictions are compared with those from two typical dynamic models. Our investigation shows that, in the case of four-month forecasting, the proposed model gives the actual monthly power load of every sector with only less than 3% absolute error and satisfactory reduction of forecasting errors compared to other models from previous studies.

UR - http://www.scopus.com/inward/record.url?scp=84907743126&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84907743126&partnerID=8YFLogxK

U2 - 10.1016/j.ijepes.2014.08.006

DO - 10.1016/j.ijepes.2014.08.006

M3 - Article

AN - SCOPUS:84907743126

VL - 64

SP - 1057

EP - 1062

JO - International Journal of Electrical Power and Energy Systems

JF - International Journal of Electrical Power and Energy Systems

SN - 0142-0615

ER -