Experimental and Numerical Investigation of Airflow Organization in Modular Data Centres Utilizing Floor Grid Air Supply

Overview

The study asks how to improve the thermal environment of a modular data centre's computer room, reduce power consumption and keep servers running safely. The hall measures 26.4 m × 16.2 m × 4.5 m with 160 standard 42U cabinets (600 mm × 1200 mm) in a 10 × 16 arrangement; the design has a 600 mm raised floor and 600 × 600 mm grilles at 40% open area. The CFD model was checked against return-air temperature and velocity measurements: average relative errors were 13.24% for temperature and 10.73% for velocity at return inlet 1, and 7.73% and 7.26% at return inlet 2, with maximum errors about 15%.

Raising the floor from 600 to 800 mm cut entransy loss to 10,578.8 kW·K, and going to 1000 mm gave only a further drop to 10,533.1 kW·K, so 600-800 mm is recommended as the cost-benefit range. Raising grille open area from 20% to 40% cut entransy loss from 16,642.0 to 10,933.8 kW·K, while going on to 60% raised it to 15,064.2 kW·K; at 20% or 60% open area, cabinet intake temperatures exceeded 27 °C. A combined pattern (lower opening near the cooling units, higher at the far end, 40% mid-aisle) gave the lowest uncontained entransy loss, 8238.3 kW·K.

With cold-aisle containment entransy loss was 4450 kW·K, and with both cold and hot aisles contained it was 3500 kW·K. The paper recommends 600-800 mm floors with 40% open grilles, cold-aisle containment for typical underfloor halls, and dual containment for high density or long throws.

Limits: validation errors are fairly large, around 7-13% on average at the return grilles; entransy loss measures irreversible mixing and cannot be compared directly with common standard indicators such as RCI or SHI. Results depend on one modular hall geometry.

Key points

1. The hall measures 26.4 m × 16.2 m × 4.5 m with 160 standard 42U cabinets (600 mm × 1200 mm) in a 10 × 16 arrangement; the design has a 600 mm raised floor and 600 × 600 mm grilles at 40% open area.
2. Average relative errors against measurements were 13.24% for temperature and 10.73% for velocity at return inlet 1, and 7.73% and 7.26% at return inlet 2; maximum errors about 15%.
3. Raising the floor from 600 to 800 mm cut entransy loss to 10,578.8 kW·K; 1000 mm gave only a further drop to 10,533.1 kW·K, so 600-800 mm is recommended.
4. Raising grille open area from 20% to 40% cut entransy loss from 16,642.0 to 10,933.8 kW·K, while 60% raised it to 15,064.2 kW·K; at 20% or 60%, cabinet intake temperatures exceeded 27 °C.
5. A combined pattern (lower opening near the cooling units, higher at the far end, 40% mid-aisle) gave the lowest uncontained entransy loss, 8238.3 kW·K.
6. With cold-aisle containment entransy loss was 4450 kW·K, and with both cold and hot aisles contained it was 3500 kW·K.

Using it in design

• Choose raised-floor height of 600-800 mm and grille open area of about 40%; a 1000 mm floor gave only a small extra gain, and 20% or 60% open area pushed cabinet intake above 27 °C.
• Make cold-aisle containment the default in typical underfloor halls and dual containment for high density or long throws; entransy loss fell to 4450 and 3500 kW·K.
• Consider the combined pattern with lower opening near the cooling units and higher at the far end (8238.3 kW·K).

Who should read it

Modular data-centre cooling designers deciding raised-floor height, grille opening and aisle containment.

Document contents

1. 1. Introduction
2. 2. Materials and Methods (2.1 Experiment: experimental subjects, experimental test; 2.2 Model Building and Verification: basic assumptions, mathematical model; 2.3 Grid Division and Verification; 2.4 Boundary Conditions and Operating Conditions; 2.5 Data Processing and Entransy)
3. 3. Results and Discussion (3.1 Model Verification, 3.2 Effect of Raised Floor Height, 3.3 Effect of Floor Grille Opening Rate, 3.4 Effect of Closed Cold/Hot Channels, 3.5 Evaluation: Entransy Loss)
4. 4. Conclusions

Related standards

• GB 50174 Code for design of data centers
• ASHRAE TC 9.9 Thermal Guidelines for Data Processing Environments

Read the document

Publication details

Authors

Jingping Zhao; Jianlin Wu; Mengying Li

Publisher

Buildings (MDPI AG)

Year

2024

Volume

14

Issue

9

Pages

2750

Document language

English

Source type

Scientific

Access

Open licence — downloadable

DOI

10.3390/buildings14092750

Licence

CC BY 4.0 — share with attribution

Licence evidence

www.mdpi.com

Citation