Retrofit Optimization of Raised-Floor Plenum Thermal Performance for Energy-Efficient and Sustainable Operation of Non-Standard Campus Data Centers

Overview

The study asks how disordered airflow distribution and prominent local hotspots in non-standard campus data centres can be fixed with plenum retrofits. The room is 18.0 m × 8.4 m × 3.6 m with a 0.4 m raised floor, 36 racks at 0.35-2.94 kW, no containment, and three precision units totalling 8 m³/s at 18 °C supply. Validation: temperature MAE no more than 0.9 °C in all zones and velocity MAE no more than 0.20 m/s, with point velocity deviations reaching about 18-30% near tile joints and pipe obstructions. Tiles were modelled as 50% porous media.

Pre-screening found the best perforation ratio and position for a transverse plenum baffle. Perforation screening (0.3 m from the outlet): 30% open gave uniformity index 0.62 with an 18.7% resistance increase; 50% open gave 0.58 with 8.5%; 70% was worse. A 50% baffle was selected. Position screening at 50% perforation: 0.3 m from the unit outlet was best (uniformity 0.58, local velocity 4.2 m/s); at 0.9 m the uniformity index was only 0.52.

A factorial set of eight scenarios then combined baffle types (none, perforated, solid partitions, composite) with uniform or graded variable-opening tiles. The optimal scenario S4 (perforated baffle plus graded tile openings) cut the tile net-flow range from 0.100 to 0.077 m³/s, lowered CV from 12.8% to 10.8% and raised the uniformity index by 10.7%; solid partition baffles alone (S2) widened the range to 0.132 m³/s. For S4, SHI fell from 0.42 to 0.35, maximum rack inlet temperature fell from 27.2 °C to 26.7 °C, and all racks met the 27 °C ASHRAE recommended limit.

Energy estimate: about 1.5 °C higher supply setpoint at 2-3% COP gain per °C (3-4.5% refrigeration saving), minus an 8.5% fan-power penalty (about 0.4% of total), gives a PUE change from about 1.90 to 1.87. Limits: a single small, asymmetric room; results depend on this geometry and the energy estimate rests on simple assumptions.

Key points

1. The room is 18.0 m × 8.4 m × 3.6 m with a 0.4 m raised floor, 36 racks at 0.35-2.94 kW, no containment, and three precision units totalling 8 m³/s at 18 °C supply.
2. Validation: temperature MAE no more than 0.9 °C in all zones and velocity MAE no more than 0.20 m/s; point velocity deviations reached about 18-30% near tile joints and pipe obstructions; tiles modelled as 50% porous media.
3. Baffle perforation screening (0.3 m from the outlet): 30% open gave uniformity index 0.62 with an 18.7% resistance increase; 50% open gave 0.58 with 8.5%; 70% was worse; 50% was selected.
4. Baffle position screening at 50% perforation: 0.3 m from the unit outlet was best (uniformity 0.58, local velocity 4.2 m/s); at 0.9 m the uniformity index was only 0.52.
5. Optimal scenario S4 (perforated baffle plus graded tile openings) cut the tile net-flow range from 0.100 to 0.077 m³/s, lowered CV from 12.8% to 10.8% and raised the uniformity index by 10.7%; solid partition baffles alone (S2) widened the range to 0.132 m³/s.
6. For S4, SHI fell from 0.42 to 0.35, maximum rack inlet temperature fell from 27.2 °C to 26.7 °C, and all racks met the 27 °C ASHRAE recommended limit.
7. Energy estimate: about 1.5 °C higher supply setpoint at 2-3% COP gain per °C (3-4.5% refrigeration saving), minus an 8.5% fan-power penalty (about 0.4% of total), gives PUE from about 1.90 to 1.87.

Using it in design

• For plenum imbalance, consider a 50%-perforated transverse baffle about 0.3 m from the unit outlet; 50% open gave a smaller resistance increase than 30% (8.5% against 18.7%).
• Combine the perforated baffle with graded variable-opening tiles (restricting near-end tiles, opening far-end tiles); a solid baffle alone made distribution worse.
• Calculate the fan-power penalty (8.5%) together with the energy saving and report the net PUE effect (1.90 to 1.87).
• Show with a full-room thermal verification that all rack inlets stay below 27 °C.

Who should read it

Cooling engineers retrofitting raised-floor plenums in small, irregular campus server rooms.

Document contents

1. 1. Introduction
2. 2. Materials and Methods (2.1 Test Facility, 2.2 Field Measurements, 2.3 CFD Simulation: boundary conditions, solution control, mesh generation and independence verification)
3. 3. Model Validation (3.1 Cabinet Inlet/Outlet Temperature Validation, 3.2 Perforated Floor Tile Outlet Validation, 3.3 Cold/Hot Aisle Temperature Validation, 3.4 Conclusions on Model Reliability)
4. 4. Scenarios and Results (4.1 Scenario Design, 4.2 Scenario Analysis, 4.3 Evaluation Metrics, 4.4 Full-Room Thermal Verification)
5. 5. Discussion (5.1 Analysis of Energy Consumption and PUE Impact, 5.2 Discussion of Engineering Applicability and Limitations)
6. 6. Conclusions

Related standards

• ASHRAE TC 9.9 Thermal Guidelines for Data Processing Environments

Read the document

Publication details

Authors

Jinuo Zhang; Zhiyi Wang; Guoming Jiang

Publisher

Sustainability (MDPI AG)

Year

2026

Volume

18

Issue

16

Pages

8144

Document language

English

Source type

Scientific

Access

Open licence — downloadable

DOI

10.3390/su18168144

Licence

CC BY 4.0 — share with attribution

Licence evidence

www.mdpi.com

Citation