A HYBRID TCN-LSTM MODEL FOR PREDICTIVE MAINTENANCE OF AWS POWER SUPPLIES
DOI:
10.29303/ipr.v9i3.544Downloads
Additional Files
Abstract
Reliable power supply units are essential for Automatic Weather Stations (AWS) to maintain continuous data collection. However, traditional maintenance schedules often fail to prevent sudden equipment downtime. While machine learning can enable predictive maintenance, standard standalone models typically struggle to capture both immediate short-term anomalies and slow, long-term degradation. To address this gap, this study aims to evaluate and propose a hybrid Temporal Convolutional Network (TCN) and Long Short-Term Memory (LSTM) architecture specifically designed for AWS power supply forecasting. Using empirical time-series data, we monitored five operational parameters at 10-minute intervals from September 2023 to November 2024. Correlation analysis established battery temperature as a primary health indicator due to its strong inverse relationship with voltage (r = –0.87). Comparative evaluations demonstrated that while individual TCN and LSTM models exhibited architectural trade-offs, the proposed hybrid TCN-LSTM model achieved the highest predictive accuracy (R² = 0.9497; MAPE = 0.05%). The findings confirm that integrating these networks effectively balances rapid anomaly detection with stable long-term trend forecasting. Practically, this hybrid model can be integrated into AWS telemetry systems as a robust diagnostic tool, providing automated early warnings to prevent critical power failures.
Keywords:
Forecast Power Supply TCN LSTMReferences
[1] D. Hikmah, L. E. Arisanti, and D. Irmawan, “Tipe Pasang Surut di Pelabuhan Benoa Bali dengan Metode Admiralty Berdasarkan Data Automatic Weather Station (AWS),” J. Widya Climago, vol. 2, pp. 86–95, 2020, [Online]. Available: https://ejournal-pusdiklat.bmkg.go.id/index.php/climago/article/view/28
[2] N. H. A. Wahab et al., “Systematic review of predictive maintenance and digital twin technologies challenges, opportunities, and best practices,” PeerJ Comput. Sci., vol. 10, 2024, doi: 10.7717/PEERJ-CS.1943.
[3] H. H. Hosamo, P. R. Svennevig, K. Svidt, D. Han, and H. K. Nielsen, “A Digital Twin predictive maintenance framework of air handling units based on automatic fault detection and diagnostics,” Energy Build., vol. 261, p. 111988, 2022, doi: 10.1016/j.enbuild.2022.111988.
[4] S. P. S. Rathore, A. Gupta, A. Parashar, D. Upadhyay, R. K. Deb, and A. Gupta, “Machine Learning in Predictive Maintenance for Industrial Equipment,” Proc. - IEEE 2024 1st Int. Conf. Adv. Comput. Commun. Networking, ICAC2N 2024, pp. 1537–1541, 2024, doi: 10.1109/ICAC2N63387.2024.10895271.
[5] L. Lin, C. Walker, and V. Agarwal, “Explainable machine-learning tools for predictive maintenance of circulating water systems in nuclear power plants,” Nucl. Eng. Technol., vol. 57, no. 9, p. 103588, 2025, doi: 10.1016/j.net.2025.103588.
[6] H. Li, W. Zhao, Y. Zhang, and E. Zio, “Remaining useful life prediction using multi-scale deep convolutional neural network,” Appl. Soft Comput. J., vol. 89, p. 106113, 2020, doi: 10.1016/j.asoc.2020.106113.
[7] L. Ren, Y. Liu, X. Wang, J. Lu, and M. J. Deen, “Cloud-Edge-Based Lightweight Temporal Convolutional Networks for Remaining Useful Life Prediction in IIoT,” IEEE Internet Things J., vol. 8, no. 16, pp. 12578–12587, 2021, doi: 10.1109/JIOT.2020.3008170.
[8] C. Y. Hsu, Y. W. Lu, and J. H. Yan, “Temporal Convolution-Based Long-Short Term Memory Network With Attention Mechanism for Remaining Useful Life Prediction,” IEEE Trans. Semicond. Manuf., vol. 35, no. 2, pp. 220–228, 2022, doi: 10.1109/TSM.2022.3164578.
[9] A. H. Almaliki and A. Khattak, “Short- and long-term tidal level forecasting: A novel hybrid TCN + LSTM framework,” J. Sea Res., vol. 204, no. January, p. 102577, 2025, doi: 10.1016/j.seares.2025.102577.
[10] A. Boujamza and S. Lissane Elhaq, “Attention-based LSTM for Remaining Useful Life Estimation of Aircraft Engines,” IFAC-PapersOnLine, vol. 55, no. 12, pp. 450–455, 2022, doi: 10.1016/j.ifacol.2022.07.353.
[11] J. Bi, X. Zhang, H. Yuan, J. Zhang, and M. C. Zhou, “A Hybrid Prediction Method for Realistic Network Traffic With Temporal Convolutional Network and LSTM,” IEEE Trans. Autom. Sci. Eng., vol. 19, no. 3, pp. 1869–1879, 2022, doi: 10.1109/TASE.2021.3077537.
[12] X. Wang, Y. Liu, X. Liang, C. Zhang, C. Yang, and W. Gui, “Learning an Enhanced TCN-LSTM Network for Temperature Process Modeling in Rotary Kilns,” IEEE Trans. Autom. Sci. Eng., vol. 22, pp. 3056–3067, 2025, doi: 10.1109/TASE.2024.3388709.
[13] S. Gopali, F. Abri, S. Siami-Namini, and A. S. Namin, “A Comparative Study of Detecting Anomalies in Time Series Data Using LSTM and TCN Models,” pp. 1–15, 2021, [Online]. Available: http://arxiv.org/abs/2112.09293
[14] F. Paliling and Z. Sudirman, “Machine Learning untuk Perawatan Prediktive Mesin Berbasis Random Forest,” INFINITY, vol. 3, no. 2, pp. 80–84, 2023, doi: 10.34148/infinity.v9i1.xxx.
[15] B. Santoso et al., “Predictive Maintenance Automatic Weather Station Sensor Error Detection using Long Short-Term Memory,” Ultim. Comput. J. Sist. Komput., vol. 15, no. 2, pp. 41–51, 2023, doi: 10.31937/sk.v15i2.3403.
[16] S. Zeb and S. K. Lodhi, “AI FOR PREDICTIVE MAINTENANCE: REDUCING DOWNTIME AND ENHANCING EFFICIENCY,” Enrich. J. Multidiscip. Res. Dev., vol. 3, no. 1, pp. 135–150, 2025, doi: 10.55324/enrichment.v3i1.338.
[17] D. Pagano, “A predictive maintenance model using Long Short-Term Memory Neural Networks and Bayesian inference,” Decis. Anal. J., vol. 6, no. January, p. 100174, 2023, doi: 10.1016/j.dajour.2023.100174.
[18] H. Taoufyq, K. El Guemmat, K. Mansouri, and F. Akef, “Predictive Maintenance Approaches : A Systematic Literature Review,” J. Ind. Eng. Manag., vol. 18, no. 3, pp. 427–458, 2025, doi: https://doi.org/10.3926/jiem.8537.
[19] S. Lourensius, N. H. Djanggu, and Y. E. Prawatya, “Implementasi Predictive Maintenance Untuk Mesin Pengupas Buah Pinang Dengan Mikrokontroller,” Integr. Ind. Eng. Manag. Syst., vol. 7, no. 2, pp. 1–6, 2023, [Online]. Available: https://jurnal.untan.ac.id/index.php/jtinUNTAN/issue/view/2162
[20] C. T. N. Siregar, P. Kindangen, and I. D. Palandeng, “Evaluasi Pemeliharaan Mesin dan Peralatan Produksi PT. Multi Nabati Sulawesi (MNS) Kota Bitung,” J. EMBA J. Ris. Ekon. Manajemen, Bisnis dan Akunt., vol. 10, no. 3, p. 428, 2022, doi: 10.35794/emba.v10i3.42362.
[21] J. Yao, Z. Cai, Z. Qian, and B. Yang, “A noval approach based on TCN-LSTM network for predicting waterlogging depth with waterlogging monitoring station,” PLoS One, vol. 18, no. 10 October, pp. 1–19, 2023, doi: 10.1371/journal.pone.0286821.
[22] S. Bai, J. Z. Kolter, and V. Koltun, “An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling,” 2018, [Online]. Available: http://arxiv.org/abs/1803.01271
[23] Y. Zuo et al., “Short text classification based on bidirectional TCN and attention mechanism,” J. Phys. Conf. Ser., vol. 1693, no. 1, 2020, doi: 10.1088/1742-6596/1693/1/012067.
[24] Y. Yu, X. Si, C. Hu, and J. Zhang, “A Review of Recurrent Neural Networks: LSTM Cells and Network Architectures,” Neural Comput., vol. 31, no. 7, pp. 1235–1270, 2019, doi: https://doi.org/10.1162/neco_a_01199.
[25] N. Somu, G. Raman M R, and K. Ramamritham, “A deep learning framework for building energy consumption forecast,” Renew. Sustain. Energy Rev., vol. 137, no. April 2020, p. 110591, 2021, doi: 10.1016/j.rser.2020.110591.
[26] K. Smagulova and A. P. James, “A survey on LSTM memristive neural network architectures and applications,” Eur. Phys. J. Spec. Top., vol. 228, no. 10, pp. 2313–2324, 2019, doi: 10.1140/epjst/e2019-900046-x.
[27] W. Nugraha and A. Sasongko, “Hyperparameter Tuning on Classification Algorithm with Grid Search,” Sistemasi, vol. 11, no. 2, p. 391, 2022, doi: 10.32520/stmsi.v11i2.1750.
[28] K. Karthika, P. Balasubramanie, K. Dharshini, P. Shanmugapriya, and T. E. Ramya, “`,” 2024 15th Int. Conf. Comput. Commun. Netw. Technol. ICCCNT 2024, pp. 1–8, 2024, doi: 10.1109/ICCCNT61001.2024.10723986.
License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Authors who publish with Indonesian Physical Review Journal, agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution-ShareAlike 4.0 International Licence (CC BY SA-4.0). This license allows authors to use all articles, data sets, graphics, and appendices in data mining applications, search engines, web sites, blogs, and other platforms by providing an appropriate reference. The journal allows the author(s) to hold the copyright without restrictions and will retain publishing rights without restrictions.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in Indonesian Physical Review Journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
