OPTIMISATION OF ENERGY CONSUMPTION OF THE 5G NEW RADIO NETWORK INFRASTRUCTURE IN CONDITIONS OF LIMITED POWER SUPPLY

DOI: 10.31673/2412-4338.2026.019018

Authors

Abstract

The article addresses the urgent problem of increasing the energy resilience of Ukraine’s telecommunications infrastructure under martial law and systemic power deficits in the national grid. It is substantiated that the deployment of 5G New Radio (NR) networks is accompanied by a significant increase in energy consumption, necessitating a transition from the extensive growth of battery fleets to intelligent management of base station sleep states.

The research aims to develop a conceptual framework for adaptive energy management in 5G NR networks to ensure maximum autonomy during blackouts. The paper analyses 3GPP specifications, specifically the "Lean Carrier Design" concept and multi-level «Advanced Sleep Modes» (Micro, Light, Deep Sleep, and Hibernation), which allow for a reduction in the power consumption of the Radio Access Network (RAN) subsystem, accounting for approximately 70-80% of the total facility energy expenditure.

The research methodology is based on the synthesis of machine learning methods and the analysis of large-scale statistical data from the NCECR for the 2022-2024 period. A decentralized Federated Learning architecture is proposed, allowing critical infrastructure objects to learn collectively based on local traffic profiles while maintaining data confidentiality. For accurate State of Charge (SOC) prediction of lithium-iron-phosphate (LiFePO4) batteries, recurrent neural networks (ResLSTM) are utilized to account for non-linear discharge curves and environmental factors.

The research results confirm that the proposed model enables an optimized discharge profile for the 385,551 battery groups recorded in Ukrainian networks. Comparative analysis shows that implementing Federated Learning provides an increase in autonomous operation time by 44-180% in critical scenarios, extending service availability to over 20 hours during blackouts. This not only allows for meeting the regulatory requirements of 72-hour resilience, but also significantly extends the battery life cycle by reducing the Depth of Discharge (DoD).

The conclusions emphasize that the 5G NR technology stack, combined with intelligent algorithms, becomes a strategic element of the national telecommunications network's energy security. Future research is aimed at integrating renewable energy sources and adapting the model to Beyond 5G/6G architectures.

Keywords: energy efficiency, federated learning, LiFePO4 batteries, Lead-Acid batteries, infrastructure resilience, ResLSTM, QoS.

References

  1. Lin, X. (2023). An overview of the 3GPP study on artificial intelligence for 5G new radio. ArXiv. https://doi.org/10.48550/arXiv.2308.05315
  1. Lin, X. (2023). An overview of the 3GPP study on artificial intelligence for 5G new radio. ArXiv. López-Pérez, D., Domenico, A.D., Piovesan, N., Baohongqiang, H., Geng, X., Song, Q., & Debbah, M. (2021). A Survey on 5G Energy Efficiency: Massive MIMO, Lean Carrier Design, Sleep Modes, and Machine Learning. ArXiv, https://api.semanticscholar.org/CorpusID:231718616 
  2. López-Pérez, D., Domenico, A.D., Piovesan, N., Baohongqiang, H., Geng, X., Song, Q., & Debbah, M. (2021). A Survey on 5G Energy Efficiency: Massive MIMO, Lean Carrier Design, Sleep Modes, and Machine Learning. ArXiv, https://api.semanticscholar.org/CorpusID:231718616 
  3. Samorzewski, A., Deruyck, M., & Kliks, A. (2023). Energy consumption in RES-aware 5G networks. In Proceedings of the 2023 IEEE Global Communications Conference (GLOBECOM) (pp. 1024-1029). IEEE. https://doi.org/10.1109/GLOBECOM54140.2023.10437451 
  4. Ichimescu, A., Popescu, N., Popovici, E. C., & Toma, A. (2024). Energy Efficiency for 5G and Beyond 5G: Potential, Limitations, and Future Directions. Sensors (Basel, Switzerland), 24(22), 7402. https://doi.org/10.3390/s24227402 
  5. NKEK Annual Report for 2022. NKEK official website. https://nkek.gov.ua/pro-nkek/zvity-nkek/richnyi-zvit-za-2022-rik 
  6. NKEK Annual Report for 2023. NKEK official website. https://nkek.gov.ua/pro-nkek/zvity-nkek/richnyi-zvit-za-2023-rik 
  7. NKEK Annual Report for 2024. NKEK official website. https://nkek.gov.ua/pro-nkek/zvity-nkek/richnyi-zvit-nkek-za-2024-rik 
  8. Li, X., Li, B., Guo, S., Sun, Z., Wang, Q., Du, T., Lin, P., & Zhang, D. (2023). Machine learning and IoT-based Li-ion battery cloud monitoring system for 5G base stations. Fractals, 31(06), 2340110. https://doi.org/10.1142/S0218348X23401102 
  9. Li, X., Li, B., Guo, S., Sun, Z., Wang, Q., Du, T., Lin, P., & Zhang, D. (2023). Machine learning and IoT-based Li-ion battery cloud monitoring system for 5G base stations. Fractals, 31(06), 2340110. https://doi.org/10.1142/S0218348X23401102 
  10. Choudhary, A., Kumar, G., Dhariwal, S., & Srivastava, G. (2024). Energy management of base station in 5G and B5G: Revisited. In 2024 International Conference on Knowledge Engineering and Communication Systems (ICKECS) (Vol. 1, pp. 1-5). IEEE. https://doi.org/10.1109/ICKECS61492.2024.10617248 
  11. Emma. (2024, April 13). LiFePO4 battery voltage chart: A comprehensive guide. Vatrer. https://www.vatrerpower.com/blogs/news/lifepo4-voltage-chart-a-comprehensive-guide?srsltid%3DAfmBOorbBn3ZwaUQBDBJPbOX-Rk-GoF5rrSUAnk8OGe3EkTBMlet8WP8 

Downloads

Published

2026-04-01

Issue

No. 1 (2026)

Section

Articles