Integrated optimization of multi-parameter cooling strategies in data centers using numerical simulation and response surface methodology

Data centers struggle with efficient cooling due to high server heat output, leading to excessive energy consumption and operational inefficiencies. Traditional cooling methods, such as CRAC units, often fail to optimize energy use and airflow distribution. This study addresses these issues by analyzing rack spacing, air velocity, and inlet air temperature using computational modeling and Response Surface Methodology. Key findings indicate that reducing rack spacing to 0.02 m improves heat dissipation, yielding a Return Heat Index (RHI) of 0.85. In contrast, increasing spacing to 0.5 m raises the Supply Heat Index (SHI) by 143 %, signifying reduced cooling efficiency. Additionally, SHI increases by 13.3 % as air velocity rises from 2 m/s to 3 m/s but stabilizes at 4 m/s, indicating diminishing returns. An inlet air temperature of 10 °C results in an 86.3 % Return Temperature Index (RTI), representing peak cooling efficiency. These findings highlight the importance of precise parameter adjustments and the effectiveness of response surface methodology in optimizing cooling efficiency, reducing energy consumption, and lowering operational costs in data centers.