DNN과 LSTM 활용한 일일 전력수요모델 개발 및 예측Modeling and Predicting South Korea's Daily Electric Demand Using DNN and LSTM
- Other Titles
- Modeling and Predicting South Korea's Daily Electric Demand Using DNN and LSTM
- Authors
- 김영수; 박호정
- Issue Date
- 2021
- Publisher
- 한국기후변화학회
- Keywords
- ANN; Covid-19; Deep Neural Network; Electriciy Demand; LSTM; Scenario Analysis
- Citation
- 한국기후변화학회지, v.12, no.3, pp.241 - 253
- Indexed
- KCI
- Journal Title
- 한국기후변화학회지
- Volume
- 12
- Number
- 3
- Start Page
- 241
- End Page
- 253
- URI
- https://scholar.korea.ac.kr/handle/2021.sw.korea/138763
- DOI
- 10.15531/KSCCR.2021.12.3.241
- ISSN
- 2093-5919
- Abstract
- Demand for electricity is influenced by factors, such as economic structure, industrial sector, and environmental volatility.
This paper aims to capture characteristics pertaining to electricity demand by developing a daily forecast model. Key variables were selected from socio-economic and environmental perspectives. Installed capacity alongside socio-economic metrics including the consumer composite sentiment index (CCSI), trade balance, unemployment rate, and day of the week were added. Environmental variables for the model were average daily temperature and COVID-19 case count.
The Deep Neural Network (DNN) model, an Artificial Neural Network (ANN) model, was used to compensate for lack of definitive linear relationships between variables for electricity demand. The trained model was performed with rRMSE of 3.74% and MAPE of 2.67%.
Further scenario analysis helped to shed light on the utility of this model. The scenario aims to explain how energy demand is affected by supply-centric electricity policies and demand management policies, as well as macro-economic expansion and contraction. A recurrent network model, LSTM, was used to forecast average daily temperatures and COVID-19 cases before the DNN electricity demand forecast model interpreted the forecast results.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - College of Life Sciences and Biotechnology > Department of Food and Resource Economics > 1. Journal Articles
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.