| Title |
Development of an Optimal Model for Day-Ahead Curtailment Prediction : A Focus on the Jeju Electricity Market |
| Authors |
이정범(Jeong-Beom Lee) ; 김지형(Ji-Hyoung Kim) ; 김규태(Kyu-Tae Kim) ; 최은호(Eun-Ho Choi) ; 강동주(Dong-Joo Kang) |
| DOI |
https://doi.org/10.5370/KIEE.2025.74.3.417 |
| Keywords |
curtailment; Jeju electricity market; NWP; prediction; XGBoost |
| Abstract |
This study presents the development of a day-ahead curtailment prediction model for the newly introduced Jeju electricity market in 2024. Using NWP (Numerical Weather Prediction) data, electricity demand and renewable energy supply were predicted. Then, based on the predicted demand and supply data, the initial curtailment was predicted using PCA (Principal Component Analysis) and MLR (Multiple Linear Regression) models. The final curtailment was then predicted using the XGBoost model. The developed model, DR-XGB (Dimension Reduction XGBoost), showed a prediction performance with an R2 of 0.41 and an RMSE of 11.0MWh, outperforming other models. Because the DR-XGB model was developed based on NWP data, it can be utilized in the future land power market. Furthermore, it is distinct from other studies because it used data from the actual electricity market. The day-ahead curtailment prediction model is expected to contribute to establishing day-ahead plans for market participants in the electricity market. |