Separable Convolutional Hierarchical Decomposition for Lightweight Residential Load Forecasting in Smart Grids

Authors

  • Satriawan Rasyid Purnama Universitas Diponegoro Indonesia
  • Henri Tantyoko Universitas Diponegoro Indonesia
  • Adi Wibowo Universitas Diponegoro Indonesia
  • Yesaya Rudolf Susanto Widyanto Universitas Diponegoro Indonesia

DOI:

https://doi.org/10.26877/asset.v8i3.2869

Keywords:

Residential load forecasting, smart grid, time-series forecasting, hierarchical decomposition, edge computing

Abstract

Residential load forecasting is essential for maintaining grid stability and energy management in smart grids. However, achieving accurate real-time forecasting under resource constraints remains challenging because Transformer and LSTM models can be computationally demanding, while lightweight linear models such as DLinear have limited modeling flexibility. This study investigates whether a hierarchical separable convolutional framework can provide accurate and efficient residential load forecasting. To address this, SeparableCLF, a lightweight hierarchical decomposition model using depthwise separable convolution, is proposed and evaluated on hourly OpenEI residential load data from 20 U.S. states (2012) at forecast horizons of 6, 12, 24, 48, and 96 h. Relative to DLinear, SeparableCLF reduced MAPE by 0.93, 1.54, and 0.79 percentage points at 24, 48, and 96 h, respectively while requiring substantially fewer parameters than Transformer and LSTM models.SeparableCLF achieved the lowest MAPE at 12, 24, and 48 h. DLinear achieved the lowest errors at 6 h, whereas LSTM achieved the lowest MAPE at 96 h; at 96 h, SeparableCLF retained the lowest MAE, MSE, and RMSE among the compared models, indicating suitability for real-time forecasting on smart meters and edge-based smart grid devices.

Author Biographies

  • Satriawan Rasyid Purnama, Universitas Diponegoro
    Departement of Informatics, Faculty of Science and Mathematics, Universitas Diponegoro, Jl. Prof. H. Soedarto, S.H., Tembalang, Semarang 50275, Central Java, Indonesia
  • Henri Tantyoko, Universitas Diponegoro
    Departement of Informatics, Faculty of Science and Mathematics, Universitas Diponegoro, Jl. Prof. H. Soedarto, S.H., Tembalang, Semarang 50275, Central Java, Indonesia
  • Adi Wibowo, Universitas Diponegoro
    Departement of Informatics, Faculty of Science and Mathematics, Universitas Diponegoro, Jl. Prof. H. Soedarto, S.H., Tembalang, Semarang 50275, Central Java, Indonesia
  • Yesaya Rudolf Susanto Widyanto, Universitas Diponegoro
    Departement of Informatics, Faculty of Science and Mathematics, Universitas Diponegoro, Jl. Prof. H. Soedarto, S.H., Tembalang, Semarang 50275, Central Java, Indonesia

References

[1] Muqtadir A, Li B, Ying Z, Songsong C, Kazmi SN. Nowcasting the next hour of residential load using boosting ensemble machines. Sci Rep 2025;15:7157. https://doi.org/10.1038/s41598-025-91767-6.

[2] La Tona G, Luna M, Di Piazza MC. Day-ahead forecasting of residential electric power consumption for energy management using Long Short-Term Memory encoder–decoder model. Math Comput Simul 2024;224:63–75. https://doi.org/https://doi.org/10.1016/j.matcom.2023.06.017.

[3] G R H, Sreedharan S, C NB. Advanced short-term load forecasting for residential demand response: An XGBoost-ANN ensemble approach. Electric Power Systems Research 2025;242:111476. https://doi.org/https://doi.org/10.1016/j.epsr.2025.111476.

[4] Kaur S, Bala A, Parashar A. A multi-step electricity prediction model for residential buildings based on ensemble Empirical Mode Decomposition technique. Science and Technology for Energy Transition (STET) 2024;79. https://doi.org/10.2516/stet/2024001.

[5] Saini VK, Singh R, Mahto DK, Kumar R, Mathur A. Learning Approach for Energy Consumption Forecasting in Residential Microgrid. 2022 IEEE Kansas Power and Energy Conference, KPEC 2022, 2022. https://doi.org/10.1109/KPEC54747.2022.9814744.

[6] Gupta A, Kumar A. Mid Term Daily Load Forecasting using ARIMA, Wavelet-ARIMA and Machine Learning. 2020 IEEE International Conference on Environment and Electrical Engineering and 2020 IEEE Industrial and Commercial Power Systems Europe (EEEIC / I&CPS Europe), 2020, p. 1–5. https://doi.org/10.1109/EEEIC/ICPSEurope49358.2020.9160563.

[7] Beydoun H, Khan A, Arefifar SA. Comparative Analysis of Time Series and Artificial Intelligence Algorithms for Short Term Load Forecasting. 2021 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE), 2021, p. 1–7. https://doi.org/10.1109/CCECE53047.2021.9569166.

[8] Inala KP, Gaddam S, Etti S, Kashetty P, Karangula J, Anaparthi N. Machine Learning Algorithms for Load Forecasting in Smart Grid. In: Panda G, Basu M, Siano P, Affijulla S, editors. Proceedings of Third International Symposium on Sustainable Energy and Technological Advancements, Singapore: Springer Nature Singapore; 2024, p. 487–99.

[9] Wang L, Lin Y, Song T, Chen Y, Li K, Ran J. Short-term residential electricity consumption forecast considering the cumulative effect of temperature, dual decomposition technology and integrated deep learning. Energy Informatics 2025;8:94. https://doi.org/10.1186/s42162-025-00552-2.

[10] Tantyoko H, Nurjanah D, Rusmawati Y. Utilizing Sequential Pattern Mining and Complex Network Analysis for Enhanced Earthquake Prediction. Advance Sustainable Science, Engineering and Technology 2024;6. https://doi.org/10.26877/asset.v6i4.1003.

[11] Lim B, Zohren S. Time-series forecasting with deep learning: a survey. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 2021;379:20200209. https://doi.org/10.1098/rsta.2020.0209.

[12] Waheed W, Xu Q, Aurangzeb M, Iqbal S, Dar SH, Elbarbary ZMS. Empowering data-driven load forecasting by leveraging long short-term memory recurrent neural networks. Heliyon 2024;10. https://doi.org/10.1016/j.heliyon.2024.e40934.

[13] Bartouli M, Hagui I, Msolli A, Helali A, Hassen F. Smart Grid Load Forecasting Models Using Recurrent Neural Network and Long Short-Term Memory. Jordan Journal of Electrical Engineering 2025;11:35–51. https://doi.org/10.5455/jjee.204-1703066445.

[14] Kong W, Dong ZY, Jia Y, Hill DJ, Xu Y, Zhang Y. Short-Term Residential Load Forecasting Based on LSTM Recurrent Neural Network. IEEE Trans Smart Grid 2019;10:841–51. https://doi.org/10.1109/TSG.2017.2753802.

[15] Abbasimehr H, Shabani M, Yousefi M. An optimized model using LSTM network for demand forecasting. Comput Ind Eng 2020;143:106435. https://doi.org/https://doi.org/10.1016/j.cie.2020.106435.

[16] Zhang Q, Ma Y, Li G, Ma J, Ding J. Short-Term Load Forecasting Based on Frequency Domain Decomposition and Deep Learning. Math Probl Eng 2020;2020:7240320. https://doi.org/https://doi.org/10.1155/2020/7240320.

[17] Ribes S, Trancoso P, Sourdis I, Bouganis C-S. Mapping Multiple LSTM models on FPGAs. 2020 International Conference on Field-Programmable Technology (ICFPT), 2020, p. 1–9. https://doi.org/10.1109/ICFPT51103.2020.00010.

[18] Taleblou R. Comparing the performance of different deep learning architectures for time series forecasting. Journal of Mathematics and Modeling in Finance 2025;5:63–87. https://doi.org/10.22054/jmmf.2025.83410.1157.

[19] Zhou H, Zhang S, Peng J, Zhang S, Li J, Xiong H, et al. Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting. CoRR 2020;abs/2012.07436.

[20] Wu H, Xu J, Wang J, Long M. Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecasting. CoRR 2021;abs/2106.13008.

[21] Nie Y, Nguyen NH, Sinthong P, Kalagnanam J. A Time Series is Worth 64 Words: Long-term Forecasting with Transformers. International Conference on Learning Representations 2023.

[22] Demiss BA, Elsaigh WA. Application of novel hybrid deep learning architectures combining Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN): construction duration estimates prediction considering preconstruction uncertainties. Engineering Research Express 2024;6:032102. https://doi.org/10.1088/2631-8695/ad6ca7.

[23] Huang Z, Wang P, Wang J, Miao H, Xu J, Zhang P. Improving Transformer Based End-to-End Code-Switching Speech Recognition Using Language Identification. Applied Sciences 2021;11. https://doi.org/10.3390/app11199106.

[24] Bai S, Kolter JZ, Koltun V. An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling. CoRR 2018;abs/1803.01271.

[25] Li Y, Li K, Chen C, Zhou X, Zeng Z, Li K. Modeling Temporal Patterns with Dilated Convolutions for Time-Series Forecasting. ACM Trans Knowl Discov Data 2021;16. https://doi.org/10.1145/3453724.

[26] Tantyoko H, Purnama SR, Vianita E. Multi-Horizon Short-Term Residential Load Forecasting Using Decomposition-Based Linear Neural Network. Advance Sustainable Science, Engineering and Technology 2025;7. https://doi.org/10.26877/asset.v7i3.2033.

[27] Zeng A, Chen M, Zhang L, Xu Q. Are transformers effective for time series forecasting? Proceedings of the Thirty-Seventh AAAI Conference on Artificial Intelligence and Thirty-Fifth Conference on Innovative Applications of Artificial Intelligence and Thirteenth Symposium on Educational Advances in Artificial Intelligence, AAAI Press; 2023. https://doi.org/10.1609/aaai.v37i9.26317.

[28] Howard AG, Zhu M, Chen B, Kalenichenko D, Wang W, Weyand T, et al. MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications. CoRR 2017;abs/1704.04861.

[29] Chollet F. Xception: Deep Learning with Depthwise Separable Convolutions. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, p. 1800–7. https://doi.org/10.1109/CVPR.2017.195.

Downloads

Published

2026-07-10

Issue

Vol. 8 No. 3 (2026): May - July

Section

Articles

License

Copyright (c) 2026 Advance Sustainable Science Engineering and Technology

Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.