Related Experiment Videos
Mid-term power load forecasting using an ensemble deep learning model with BKA and CWGAN-GP enhancements.
Shucheng Luo1, Xiaohong Chen2, Xinfu Pang1
1Key Laboratory of Energy Saving and Controlling in Power System of Liaoning Province, Shenyang Institute of Engineering, Shenyang, 110136, People's Republic of China.
Scientific Reports
|April 21, 2026
Summary
This study introduces a hybrid deep learning model for mid-term electric load forecasting, enhancing accuracy with data augmentation and ensemble methods. The novel approach improves predictions for power systems with increasing volatility.
Area of Science:
- Electrical Engineering
- Artificial Intelligence
- Data Science
Background:
- Electric load forecasting is vital for power system operations, including scheduling, cost reduction, and planning.
- Increased use of electric vehicles and power electronics causes greater load volatility and nonlinearity.
- Scarcity of historical data for new loads challenges accurate forecasting.
Purpose of the Study:
- To propose a hybrid deep learning model for mid-term electric load forecasting.
- To address challenges posed by data scarcity and load volatility.
- To improve the accuracy and robustness of electric load predictions.
Main Methods:
- Data augmentation using Conditional Wasserstein Generative Adversarial Networks with Gradient Penalty (CWGAN-GP).
- Independent training of CNN-BiTransformer-BiLSTM and BiTCN-BiGRU-Attention models on semi-monthly datasets.
- Ensemble modeling using XGBoost to combine predictions from individual models.
Main Results:
- The hybrid model demonstrated significant improvements in forecasting accuracy.
- Root Mean Square Error (RMSE) was reduced by at least 3.02%.
- Mean Absolute Percentage Error (MAPE) was lowered by at least 17.7%.
Conclusions:
- The proposed hybrid deep learning model effectively enhances mid-term electric load forecasting accuracy.
- CWGAN-GP data augmentation and ensemble methods contribute to model robustness.
- The model offers a reliable solution for power systems facing increasing load complexity.
