Operational analysis of the cooling system in a direct liquid-cooled data center: a measurement and simulation study on the impact of supply water temperature

To further improve the energy efficiency of direct liquid-cooled data centers, advanced cooling control strategies capable of dynamic adjustments according to varying operational and environmental conditions are required. A critical parameter for optimization in the liquid cooling system is the supply water temperature. This study investigates the impact of the supply water temperature on data center energy performance by analyzing operational data from the direct liquid-cooled HAWK supercomputer, which was operated under different supply temperature settings. The analysis demonstrates that increasing the supply temperature from 17 to 25 ∘ C improves the liquid cooling system’s energy efficiency, achieving a power consumption reduction of 63.3 % at an outdoor wet-bulb temperature of 19 ∘ C. However, this also increased heat transfer to the server room air, with the proportion nearly doubling from approximately 2.6 % to 5.0 %, and increased IT equipment power consumption by about 3.15 %. Based on these insights, a data center-level power and thermal model was developed, incorporating the liquid cooling system, air cooling system, IT equipment, and the interactions among these systems. Simulation studies using this model further analyze the impact of supply water temperature on the energy performance of each system and the overall data center. The results highlight the considerable energy-saving potential of dynamic control strategies that adjust the supply water temperature according to outdoor conditions. The findings also highlight the importance of holistic data center modeling for developing effective cooling control strategies and enabling efficient heat reuse integration. • Analyzed 21 months of operational data from a direct liquid-cooled data center. • Identified interactions between liquid cooling, air cooling, and IT equipment. • Developed cross-system model for direct liquid-cooled data centers. • Evaluated impact of supply water temperature on data center power consumption. • Identified significant energy savings via dynamic supply temperature control.