Open Access

ARTICLE

Evaluation of Local Thermal Environment in Operational Data Center

Yan Xiao¹, Ruoyi Xie¹, Song Gao², Jinhui Yang¹, Luyao Liang¹, Shaozhi Zhang^1,*

1 Institute of Refrigeration and Cryogenics, Zhejiang University, Hangzhou, China
2 Network Operation Department, China Telecom Hangzhou Branch, Hangzhou, China

* Corresponding Author: Shaozhi Zhang. Email:

Frontiers in Heat and Mass Transfer 2026, 24(4), 6 https://doi.org/10.32604/fhmt.2026.077396

Received 08 December 2025; Accepted 01 April 2026; Issue published 31 August 2026

Abstract

When redeploying servers or installing new racks within an existing data center, assessing the local rack-level thermal environment is crucial to ensure target areas meet stringent cooling requirements. To date, limited research has addressed this specific issue. In this study, experimental measurements and CFD simulations were conducted to evaluate the local thermal environment of an operational data center. The Index of Mixing (IOM) was selected as the primary metric for thermal evaluation. To closely mimic real-world conditions, a 2U server simulator featuring adjustable air volume and heating power was constructed. Experimental results indicated that the four 2U servers experienced superior cooling in the lower section of the target rack. Furthermore, operating without blind plates resulted in a 718% increase in the IOM of the target rack. Additionally, a CFD simulation methodology was developed to evaluate the rack-level thermal environment. The CFD model comprised three racks, designated hot and cold aisles, and an overhead space to simulate the mixing of hot and cold airflows. A flow resistance model was employed to approximate the servers as porous media, with resistance characteristics calibrated using the experimental data. Five representative server layout scenarios were simulated. The numerical results demonstrated strong consistency with experimental data regarding the rack inlet and outlet temperatures. These findings demonstrate that the proposed server simulator and CFD modeling approach serve as highly effective tools for assessing local thermal environments within data centers.

---

Keywords

---