Thermal Performance Evaluation of a Data Center Cooling System under Fault Conditions

Overview

The study asks how long server inlet temperature stays acceptable during a cooling system fault and how quickly backup cooling must be running. The 6SigmaDC room model has 192 racks at 4.0 kW with CAC, underfloor supply with side-wall return, a 900 mm raised floor, a 3.0 m false ceiling, 150 kW CRAHs and chilled water varied from 10 to 18 °C. The model finds the highest chilled-water and CRAH supply temperatures that still give compliant inlets, and a pipe-volume energy balance then converts those temperatures into elapsed time after the outage.

Server inlets stayed within 18-27 °C up to a CRAH supply-air temperature of about 24 °C, corresponding to 17 °C chilled water. Above 17 °C chilled water, CRAH capacity falls below 80% and room temperature rises rapidly; from 18 °C chilled water and 25 °C supply air, coil capacity fell below the load and inlet temperatures rose rapidly. The chilled-water piping held 234.3 m³ in total, about 45 m³ per typical floor, serving a 2500 kW floor cooling load at 6.75 m³/min with a fixed 5.5 °C water delta-T.

After the outage, chilled water took about 150 s to rise from 10 to 14 °C, and the whole 45 m³ reached 15.5 °C at 3.45 min, 21.0 °C at 10.36 min and 26.5 °C at 17.27 min. Compliant operation could continue for about 320 s (5 min 20 s) after the cooling outage, with CRAH coil capacity at 85%, so backup cooling must be active within 300 s. High-temperature, contained designs have only about five minutes before backup cooling must be running.

Limits: a single project (20 MW, Korea) and six fault conditions; the result depends on this plant's pipe volume, CRAH capacity and 4 kW per rack load. The authors also address organising and managing operational stability and energy-efficiency verification systematically through the commissioning process.

Key points

1. CFD module: 192 racks at 4.0 kW/rack with CAC, underfloor supply with side-wall return, 900 mm raised floor, 3.0 m false ceiling, 150 kW CRAHs and chilled water from 10 to 18 °C.
2. Server inlets stayed within 18-27 °C up to a CRAH supply-air temperature of about 24 °C, corresponding to 17 °C chilled water; above that CRAH capacity falls below 80% and room temperature rises rapidly.
3. The chilled-water piping held 234.3 m³ in total, about 45 m³ per typical floor, serving a 2500 kW floor cooling load at 6.75 m³/min with a fixed 5.5 °C delta-T.
4. After the outage, water took about 150 s to rise from 10 to 14 °C; the 45 m³ reached 15.5 °C at 3.45 min, 21.0 °C at 10.36 min and 26.5 °C at 17.27 min.
5. Compliant operation could continue for about 320 s (5 min 20 s) after the outage with CRAH coil capacity at 85%, so backup cooling must be active within 300 s.
6. From 18 °C chilled water and 25 °C supply air, coil capacity fell below the load and inlet temperatures rose rapidly.

Using it in design

• Calculate the permissible time for a cooling outage with a pipe-volume energy balance and set the backup-cooling start time accordingly (here 300 s); high-temperature, contained designs give less time.
• Limit chilled-water and CRAH supply temperature to the ASHRAE-compliant upper bound (17 °C water, 24 °C air) as the pre-fault operating point; above 17 °C capacity falls below 80%.
• Model UPS-backed pumps and pipe volume (45 m³ per typical floor) as thermal buffer and add them to fault-condition commissioning test scenarios.

Who should read it

Designers and commissioning engineers determining cooling fault resilience and backup-cooling start time.

Document contents

1. 1. Introduction
2. 2. Literature Reviews
3. 3. Major Issues in Data Center Planning (3.1 Root Causes and Scale of Fault Conditions, 3.2 Temperature and IT Reliability)
4. 4. Case Study: Data Center and Modeling (4.1 Data Center Cooling System, 4.2 IT Server Room Thermal Model, 4.3 Simulation)
5. 5. Temperature Control Approaches (5.1 Temperature Increase over Time After a Cooling System Outage)
6. 6. Conclusions

Related standards

• ASHRAE TC 9.9 Thermal Guidelines for Data Processing Environments
• Uptime Institute Tier Standard

Read the document

Publication details

Authors

Jinkyun Cho; Beungyong Park; Yongdae Jeong

Publisher

Energies (MDPI AG)

Year

2019

Volume

12

Issue

15

Pages

2996

Document language

English

Source type

Scientific

Access

Open licence — downloadable

DOI

10.3390/en12152996

Licence

CC BY 4.0 — share with attribution

Licence evidence

www.mdpi.com

Citation