Numerical and experimental dataset for a retrofitted data center

Area of Science:

• Mechanical Engineering
• Thermal Sciences
• Computational Fluid Dynamics

Background:

• Data centers generate significant heat, requiring efficient cooling solutions.
• Accurate thermal management is crucial for data center reliability and performance.
• Air-cooling is a common method, but its effectiveness depends on precise flow and temperature distribution understanding.

Purpose of the Study:

• To provide experimental and numerical data for flow and thermal distributions in an air-cooled data center.
• To offer validated OpenFOAM simulation files for thermal structure analysis.
• To facilitate the validation of numerical models for data center thermal management.

Main Methods:

• Experimental measurements using temperature sensors at the rear of server racks (2 kW load).
• Numerical simulations using OpenFOAM based on experimental data.
• Development of script files for data processing and simulation.

Main Results:

• Detailed exhaust temperature profiles from experimental campaigns.
• Validated numerical model showcasing flow and thermal structures within the data center.
• Provision of publically available experimental data and numerical models for research.

Conclusions:

• The provided data and models enable validation of numerical simulations for specific thermal scenarios.
• The validated model accurately represents flow and thermal distributions in an air-cooled data center.
• This work supports the development of more efficient data center cooling strategies.