The demand for reliable and efficient power systems has been increasing day by day due to the increasing complexity of electrical grids and the integration of renewable energy sources. Traditional power load forecasting methods are struggling to keep pace with these evolving demands, often resulting in significant power losses and inefficiencies for different scenarios. Existing load forecasting models often face challenges in terms of accuracy, adaptability and computational efficiency. These models typically do not fully utilize the intricate relationships within power systems, leading to suboptimal predictions and higher power losses. Furthermore, they often suffer from delays in processing and adapting to real-time changes in the grid, thereby reducing their practical utility. To address these issues, this paper introduces a novel load forecasting model based on Graph Q Networks (GQNs). GQNs leverage the power of graph-based learning to effectively model complex interdependencies in power systems. The GQN model exhibits MAPE of 1.2% and RMSE of 0.8 MW of Method [12], significantly outperforming the most efficient method. The proposed model outperforms MAPE of 2.4% and RMSE of 1.9 MW of Method [13] most inefficient method among the studies. The proposed model allows for more accurate and dynamic load forecasting, significantly minimizing power losses. By integrating these methods with graph theory, the model not only enhances forecasting accuracy but also reduces operational delays. Moreover, its adaptability makes it suitable for a wide range of power systems to emergingย smartย grids and those with high renewable energy integration, further reinforce its utility.
Keywords
Load Forecasting, Power Loss Minimization, Graph Q Networks, Machine Learning, Energy Efficiency.
The research presented in this paper successfully introduces an innovative load forecasting model based on Graph Q Networks (GQNs), tailored to address the burgeoning complexities and evolving demands of modern power systems. The findings from our comprehensive analysis, as detailed in the results section, unequivocally demonstrate the superiority of the GQN model over the existing Method [4], Method [12], and Method [13], in terms of accuracy, computational efficiency, and adaptability to real-time changes in diverse grid scenarios.
References
1.M. Madhukumar, A. Sebastian, X. Liang, M. Jamil and M. N. S. K. Shabbir, Regression Model-Based Short-Term Load Forecasting for University Campus Load. IEEE Access, vol. 10, pp. 8891-8905, 2022, doi: 10.1109/ACCESS.2022.3144206.
2.Y. Wang, N. Gao and G. Hug, Personalized Federated Learning for Individual Consumer Load Forecasting. CSEE Journal of Power and Energy Systems, vol. 9, no. 1, pp. 326-330, January 2023, doi: 10.17775/CSEEJPES.2021.07350.
3.J. Son, J. Cha, H. Kim and Y. -M. Wi, Day-Ahead Short-Term Load Forecasting for Holidays Based on Modification of Similar Daysโ Load Profiles. IEEE Access, vol. 10, pp. 17864-17880, 2022, doi: 10.1109/ACCESS.2022.3150344.
4.P. Zhao, D. Cao, Y. Wang, Z. Chen and W. Hu, Gaussian Process-Aided Transfer Learning for Probabilistic Load Forecasting Against Anomalous Events. IEEE Transactions on Power Systems, vol. 38, no. 3, pp. 2962-2965, May 2023, doi: 10.1109/TPWRS.2023.3256130.
5.F. Ziel, Smoothed Bernstein Online Aggregation for Short-Term Load Forecasting. IEEE DataPort Competition on Day-Ahead Electricity Demand Forecasting: Post-COVID Paradigm. IEEE Open Access Journal of Power and Energy, vol. 9, pp. 202-212, 2022, doi: 10.1109/OAJPE.2022.3160933.
6.C. Wang, Y. Zhou, Q. Wen and Y. Wang, Improving Load Forecasting Performance via Sample Reweighting. IEEE Transactions on Smart Grid, vol. 14, no. 4, pp. 3317-3320, July 2023, doi: 10.1109/TSG.2023.3269205.
7.H. Hou et al. Load Forecasting Combining Phase Space Reconstruction and Stacking Ensemble Learning. IEEE Transactions on Industry Applications, vol. 59, no. 2, pp. 2296-2304, March-April 2023, doi: 10.1109/TIA.2022.3225516.
8.Y. Guo et al. BiLSTM Multitask Learning-Based Combined Load Forecasting Considering the Loads Coupling Relationship for Multienergy System. IEEE Transactions on Smart Grid, vol. 13, no. 5, pp. 3481-3492, Sept. 2022, doi: 10.1109/TSG.2022.3173964.
9.Z. Zhou, Y. Xu and C. Ren, A Transfer Learning Method for Forecasting Masked-Load With Behind-the-Meter Distributed Energy Resources. IEEE Transactions on Smart Grid, vol. 13, no. 6, pp. 4961-4964, Nov. 2022, doi: 10.1109/TSG.2022.3204212.
10.J. Dong, L. Luo, Y. Lu and Q. Zhang, A Parallel Short-Term Power Load Forecasting Method Considering High-Level Elastic Loads. IEEE Transactions on Instrumentation and Measurement, vol. 72, pp. 1-10, 2023, Art no. 2524210, doi: 10.1109/TIM.2023.3304671.
11.W. Yanan, W. Jiekang and L. Zhen, A Hybrid Model for Short-Term Load Forecasting Based on Novel Input Sequence Selection and CSO Optimized Depth Belief Network. IEEE Access, vol. 11, pp. 116141-116152, 2023, doi: 10.1109/ACCESS.2023.3325671.
12.N. Ahmad, Y. Ghadi, M. Adnan and M. Ali, Load Forecasting Techniques for Power System: Research Challenges and Survey. IEEE Access, vol. 10, pp. 71054-71090, 2022, doi: 10.1109/ACCESS.2022.3187839.
W. Li et al. Firefly Algorithm-Based Semi-Supervised Learning With Transformer Method for Shore Power Load Forecasting. IEEE Access, vol. 11, pp. 77359-77370, 2023, doi: 10.1109/ACCESS.2023.3297647.
๐ How to Cite This Paper
Preeti Manke, Dr. Sourabh Rungta, Dr. Satyadharma Bharti, Sharad Tembhurne (2026). Smart Grid Efficiency Enhancement through Load Forecasting. International Journal of Engineering and Techniques, 12(4), 194โ199. ISSN: 2395-1303. DOI: https://doi.org/10.5281/zenodo.22097067
