Accurate short-term and long-term power load forecasting plays a pivotal role in ensuring the reliability and economic efficiency of modern smart grids. To address the challenges of complex temporal patterns, multi-scale periodicity, and dynamic variability in load data, we propose a novel deep forecasting framework named TPA-Net (Temporal Pyramid CNN–BiLSTM–Attention). The proposed architecture consists of three key components: a temporal pyramid multi-scale convolutional module to extract hierarchical periodic features across hourly, daily, and weekly levels; a bidirectional LSTM (BiLSTM) module to capture global temporal de-pendencies in both forward and backward directions; and an attention mechanism to dynamically emphasize critical time steps. Extensive experiments conducted on real-world power consumption datasets demonstrate that TPA-Net consistently outperforms state-of-the-art baselines across multiple forecasting horizons (1-h, 6-h, and 24-h), achieving significant improvements in RMSE, MAPE, and R2 metrics. These results highlight the effectiveness and generalizability of TPA-Net in complex load forecasting scenarios.
AkhtarSShahzadSZaheerA, et al.Short-term load forecasting models: a review of challenges, progress, and the road ahead. Energies2023.
2.
DongQHuangRCuiC, et al.Short-term electricity-load forecasting by deep learning: a comprehensive survey. arXiv preprint arXiv:2408.162022024.
3.
BenidisKRangapuramSSFlunkertV, et al.Deep learning for time series forecasting: tutorial and literature survey. Comput Surv2022.
4.
TarmaniniCSarmaNGezeginC, et al.Short term load forecasting based on arima and ann approaches. Energy Rep2023; 9: 550–557.
5.
ErenYKucukdemiralIB. A comprehensive review on deep learning ap- proaches for short-term load forecasting. Renew Sustain Energy Rev2024; 189: 114031.
6.
BianHWangQXuG, et al.Load forecasting of hybrid deep learning model considering accumu- lated temperature effect. Energy Rep2022.
7.
ChengLZangHXuY, et al.Probabilistic residential load forecasting based on micrometeorologi- cal data and customer consumption pattern. IEEE Trans Power Syst2021; 36(4): 3762–3775.
8.
AseeriAO. Effective rnn-based forecasting methodology design for improving short-term power load forecasts: application to large-scale power-grid time series. Journal of Computational Science2023.
9.
DedinecAFiliposkaSDedinecA, et al.Deep belief network based electricity load forecasting: an analysis of Macedonian case. Energy2016; 115: 1688–1700.
10.
DudekG. Neural networks for pattern-based short-term load forecasting: a com- parative study. Neurocomputing2016; 205: 64–74.
11.
ArastehfarSMatinkiaMJabbarpourMRShort-term residential load forecasting using graph convolutional recurrent neural networks. Eng Appl Artif Intell2022.
12.
DongXQianLHuangL. Short-term load forecasting in smart grid: a combined cnn and k-means clustering approach. Proc. of IEEE BigComp, 2017, pp. 119–125.
13.
FahimanFErfaniSMLeckieC. Robust and accurate short-term load forecasting: a cluster oriented ensemble learning approach. Proc. of International Joint Conference on Neural Networks (IJCNN), 2019, pp. 1–8.
14.
DograRAnandABediJ. Consumers profiling based federated learning approach for energy load forecasting. Sustain Cities Soc2023; 98: 104815.
15.
DongZDongZZhaoT, et al.Short term load forecasting with markovian switching distributed deep belief networks. Int J Electr Power Energy Syst2021; 130: 106942.
16.
AhajjamMALiceaDBGhoghoM, et al.Experimental investigation of variational mode decomposition and deep learning for short-term multi-horizon residential electric load forecasting. Appl Energy, 2022.
17.
BayramFAupkePAhmedBS, et al.Da-lstm: a dynamic drift-adaptive learning framework for in- terval load forecasting with lstm networks. Eng Appl Artif Intell2023.
18.
CheungCMKuppannagariSKannanR, et al.Leveraging spatial information in smart grids using stgcn for short-term load forecasting. 13th Int. Conf. on Contemporary Computing, 2021, pp. 159–167.IC3.
19.
ChoiERyuSKimH. Short-term load forecasting based on resnet and lstm. Proc. of IEEE SmartGridComm, 2018.
20.
DaiYYangXLengM. Optimized seq2seq model based on multiple methods for short-term power load forecasting. Appl Soft Comput2023; 142: 110335.
21.
DeepanrajBSenthilkumarNJarinT, et al.Intelligent wild Geese algorithm with deep learning driven short term load forecasting for sustainable energy management in microgrids. Sustainable Computing: Informatics and Systems2022; 36: 100813.
22.
DinGMUMarneridesAK. Short term power load forecasting using deep neural networks. Proc. of International Conference on Computing, Networking and Communications (ICNC), 2017, pp. 594–598.
23.
ChiuMCHsuHWChenKS, et al.A hybrid cnn-gru based probabilistic model for load forecasting from individual household to commercial building. Energy Rep2023; 9: 94–105.
24.
AouadMHajjHShabanK, et al.A cnn-sequence-to-sequence network with attention for residential short-term load forecasting. Elec Power Syst Res2022.
25.
EskandariHImaniMMoghaddamMP. Convolutional and recurrent neural network based model for short-term load forecasting. Elec Power Syst Res2021; 195: 107173.
26.
GaoJWangKKangX, et al.Ultra-short-term electricity load fore- casting based on improved random forest algorithm. AIP Adv2023; 13(6): 065208.
27.
BashirTHaoyongCTahirMF, et al.Short term electricity load forecasting using hybrid prophet-lstm model optimized by bpnn. Energy Rep2022.
28.
DongHLuoLLuY, et al.A parallel short-term power load forecasting method considering high-level elastic loads. IEEE Trans Instrum Meas2023; 72: 1–10.
29.
ChangCWangW-YPengW-C, et al.Llm4ts: two-Stage fine-tuning for time-series forecasting with pre- trained llms, 2023. arXiv preprint arXiv:2308.08469.
30.
DasAAnnaqeebMKAzarE, et al.Occupant-centric miscellaneous electric loads prediction in buildings using state-of-the-art deep learning methods. Appl Energy2020; 269: 115135.
31.
FanGDingCZhengJ, et al.Empirical mode decomposition based multi-objective deep belief net- work for short-term power load forecasting. Neurocomputing2020; 388: 110–123.
32.
BuXWuQZhouB, et al.Hybrid short-term load forecasting using cgan with cnn and semi- supervised regression. Appl Energy2023.
33.
LeCunYBotouLBengioY, et al.Gradient-based learning applied to document recognition. Proc IEEE1998; 86(11): 2278–2324.
34.
SelvinSVijayakumarRGopalakrishnanEA, et al.Stock price prediction using lstm, rnn, and cnn-sliding window model. In: Proceedings of the International Conference on Advances in Computing Communications and Informatics (ICACCI), IEEE. p 1643–1647, , 2017.
35.
GravesASchmidhuberJ. Framewise phoneme classification with bidirectional lstm and other neural network architectures. Neural Netw2005; 18(5-6): 602–610.
36.
HochreiterSSchmidhuberJ. Long short-term memory. Neural Comput1997; 9(8): 1735–1780.
37.
VaswaniAShazeerNParmarN, et al.Attention is all you need. In: Proceedings of the 31st International Conference on Neural Information Processing Systems (NeurIPS). Curran Associates Inc, 2017, pp. 5998–6008.
38.
DzmitryBKyunghyunCYoshuaB. Neural machine trans- lation by jointly learning to align and translate. arXiv preprint arXiv:1409.04732014.
39.
SongQJunhuanLLuoF, et al.A short-term electricity load forecasting method based on woa-bilstm-attention. J Phys: Conf Ser2023; 2532: 012003.
