Research on energy optimization for liquid-cooled server cooling systems based on swarm intelligence algorithms and LSTM
Journal of Mechanics · 31.01.2025 · 41 · 690-702
DOI: 10.1093/jom/ufaf051 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Mechanics, online erschienen am 10.01.2026. Verfasst von Hsuan-Li Shih und 3 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 18 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Research of Control System for Liquid-Cooled Server Cooling Systems.
Abstract (Originalsprache)
With the growing development of artificial intelligence and the increasing importance of computational power, servers are required to be equipped with high-density graphics processing unit (GPU). The high-power consumption of GPU significantly increases the cooling demand, potentially reducing the operational efficiency of data centers. Therefore, how to optimize the energy efficiency of server cooling systems while ensuring the safe operation of GPU has become a crucial research topic. This study focuses on the energy optimization of fan and pump within the cooling distribution unit in a liquid-to-air server cooling system under high thermal load conditions. By introducing deep learning techniques, a temperature prediction model based on long short-term memory networks is constructed, and global optimization of control parameters is performed using particle swarm optimization. The proposed approach aims to minimize cooling system energy consumption while maintaining GPU temperatures within safe limits. Simulation and experimental results verify that the proposed method can effectively adjust fan and pump loads, reduce cooling energy consumption and improve the dynamic response and control stability of the system. This provides a feasible solution for future energy-efficient optimization of cooling systems in data centers.
Förderung
- Research of Control System for Liquid-Cooled Server Cooling Systems
Zitieren
APA
Shih, H., Lee, C., Tung, K., & Chen, R. (2025). Research on energy optimization for liquid-cooled server cooling systems based on swarm intelligence algorithms and LSTM. Journal of Mechanics, 41, 690-702. https://doi.org/10.1093/jom/ufaf051
Vancouver
Shih H, Lee C, Tung K, Chen R. Research on energy optimization for liquid-cooled server cooling systems based on swarm intelligence algorithms and LSTM. Journal of Mechanics. 2025;41:690-702. doi:10.1093/jom/ufaf051
BibTeX
@article{Shih2025,
author = {Shih, Hsuan-Li and Lee, Chien-Ming and Tung, Kai-Yang and Chen, Rongshun},
title = {Research on energy optimization for liquid-cooled server cooling systems based on swarm intelligence algorithms and LSTM},
journal = {Journal of Mechanics},
year = {2025},
volume = {41},
pages = {690--702},
doi = {10.1093/jom/ufaf051}
}
EUVIMED