KLI

ART1 신경망을 이용한 다음날의 전력수요 예측

Metadata Downloads
Alternative Title
One day ahead load forecasting using the ART1 neural network
Abstract
ART1(Adaptive Resonance Theory 1) 신경망을 이용하여 다음날의 시간별 전력수요를 예측하는 산법을 제안하였다. 제안한 산법은 과거 10일간의 실측전력수요로부터 수요모형을 인식한 후, 이 모형을 이용하여 다음날의 시간별 수요를 예측한다. 예측정확도를 높이기 위하여 수요모형을 평일, 일요일, 월요일, 토요일 및 특수일의 5가지로 분류하였으며, 특수일은 다시 신정, 설, 추석 및 기타 공휴일로 구분하였다.

1993년 우리나라의 수요성적 데이터를 사례로 검토해 본 결과, 지수평활화법에 비하여 우수한 예측결과를 보여 주었다.
A load forecasting method using ART1(Adaptive Resonance Theory 1) neural network is proposed. The proposed method analyzes the power demand pattern of past 10 days, then predicts the hourly load of next day. To improve the accuracy of forecasting, five different day groups are defined and investigated according to the power demand, such as Sunday, Saturday, ordinary weekday, the day after holiday, and special day. The special day group includes New Year's day, Thanks giving day, New moon Year's day, and other holidays.

KEPCO's hourly load data recorded between Nov. 15, 1992 and Dec. 31, 1993 are used as case study references. It is shown that the proposed method provides less forecasting error compare to the conventional exponential smoothing method.
A load forecasting method using ART1(Adaptive Resonance Theory 1) neural network is proposed. The proposed method analyzes the power demand pattern of past 10 days, then predicts the hourly load of next day. To improve the accuracy of forecasting, five different day groups are defined and investigated according to the power demand, such as Sunday, Saturday, ordinary weekday, the day after holiday, and special day. The special day group includes New Year's day, Thanks giving day, New moon Year's day, and other holidays.

KEPCO's hourly load data recorded between Nov. 15, 1992 and Dec. 31, 1993 are used as case study references. It is shown that the proposed method provides less forecasting error compare to the conventional exponential smoothing method.
Author(s)
鄭勝敎朴源深
Issued Date
1995
Type
Research Laboratory
URI
https://oak.ulsan.ac.kr/handle/2021.oak/4057
http://ulsan.dcollection.net/jsp/common/DcLoOrgPer.jsp?sItemId=000002025192
Alternative Author(s)
Chong,Sung-KyoPark,Won-Sim
Publisher
공학연구논문집
Language
kor
Rights
울산대학교 저작물은 저작권에 의해 보호받습니다.
Citation Volume
26
Citation Number
1
Citation Start Page
227
Citation End Page
242
Appears in Collections:
Research Laboratory > Engineering Research
Authorize & License
  • Authorize공개
Files in This Item:

Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.