| Title |
Comparative Analysis of Artificial Intelligent Prediction Models for Nationwide Short-Term Electricity Demand |
| Authors |
오병찬(Byeong-Chan Oh) ; 서혁준(Hyuk-Jun Seo) ; 강혜겸(Hye-Kyeom Kank) ; 김성열(Sung-Yul Kim) ; 모제스 아모아시 아쿠아(Moses Amoasi Acquah) |
| DOI |
https://doi.org/10.5370/KIEEP.2020.69.4.253 |
| Keywords |
Support Vector Regression; Long Short Term Memory; Gate Recurrent Unit; Electrical Load forecasting |
| Abstract |
Uncertainty in the output of renewable energy can lead to an imbalance in power supply and demand. It can lead to disruptions in power supply and demand plans such as a decrease in the supply reserve ratio and an increase in the system marginal price.
Therefore, for stable power supply and demand, accurate power demand forecasting is essential. In this paper, the accuracy of nationwide short-term power demand forecasting was compared using artificial intelligence based forecasting model. In order to evaluate the prediction error, MAPE and RMSE error functions were used, and the power demand prediction was performed according to weekdays and weekends. Also input variables were selected through correlation analysis with meteorological factors.
Data from 2010 to 2018 were used for model training, and data from 2019 were used as test data. Finally the performance of each prediction method was compared through a case study. |