Simulation-based assessment of data center waste heat utilization using aquifer thermal energy storage of a university campus

Overview

Practical use of data-centre waste heat is very challenging, and its possible benefits are uncertain and difficult to quantify with common methods. The study is a feasibility study using dynamic simulation tools to predict the campus energy performance resulting from integrating a proposed data centre with an existing ATES. Uncertainty in the data centre's thermal management, chiefly supply-air setpoints and air management, was propagated through the cooling energy results; supply-air temperature setpoints of 13, 17, 21 and 25 °C (within ASHRAE guidelines) were simulated.

The reference data centre has about one hundred racks on a 23 m × 30 m raised floor, averaging 10.7 kW per rack, for a total IT load of 1090 kW. The campus ATES has 32 wells, 20-80 m deep, in three warm and three cold clusters linked by about 2 km warm and cold rings; cold wells hold 6-9 °C water and warm wells 14-17 °C, against a natural subsurface temperature of 10-13 °C. Because the ATES had mainly supplied cooling, imported heat exceeded exported heat; keeping the yearly heat balance is identified as the potential barrier to connecting a continuously cooling data centre.

Three cases were compared with a conventional chiller plant: the isolated data centre without ATES, ATES in the current campus with its import/export balance policy, and ATES in a future campus. Cooperation with ATES improves average PUE from 1.78 to 1.45. Integrated operation cuts data-centre CO2 emissions by 680 tons per year on average under current ATES policy and by 1000 tons per year in the future scenario.

The conclusion is that linking with ATES lowers PUE and CO2, and campus heat export also helps rebalance an aquifer drifting toward net cooling. Limits: a simulation feasibility for a single campus and a planned data centre; results depend on this ATES's heat balance policy.

Key points

1. The reference data centre has about one hundred racks on a 23 m × 30 m raised floor, averaging 10.7 kW per rack, for a total IT load of 1090 kW.
2. The campus ATES has 32 wells, 20-80 m deep, in three warm and three cold clusters; cold wells hold 6-9 °C water and warm wells 14-17 °C, against a natural subsurface temperature of 10-13 °C.
3. Because the ATES had mainly supplied cooling, imported heat exceeded exported heat; keeping the yearly heat balance is the potential barrier to connecting a continuously cooling data centre.
4. Supply-air temperature setpoints of 13, 17, 21 and 25 °C (within ASHRAE guidelines) were simulated in the uncertainty analysis.
5. Cooperation with ATES improves average PUE from 1.78 to 1.45.
6. Integrated operation cuts data-centre CO2 emissions by 680 tons per year on average under current ATES policy, and by 1000 tons per year in the future scenario.

Using it in design

• When considering connecting a data centre to an ATES, make the yearly heat balance (imported versus exported heat) the first feasibility check; a continuously cooling data centre can shift it toward heat rejection.
• Propagate the supply-air setpoint (13-25 °C) uncertainty through waste-heat potential and PUE results.
• Compare the PUE and CO2 gain of ATES integration (1.78 to 1.45, 680-1000 tons per year) against a conventional chiller plant on the same model.

Who should read it

Planners and energy engineers evaluating data-centre connection to campus and district ATES systems.

Document contents

1. 2 DC waste heat utilization system (2.1.1 ATES district system, 2.1.2 DC system, 2.2 Utilization scenarios)
2. 3 Modelling (3.1 Modelling method and assumptions)
3. 4 Discussion (4.2 Recommendation for ATES management: current utilization scenario, reduction of DC electricity; future utilization scenario, benefits of using DC waste heat)
4. 5 Conclusion

Related standards

• ASHRAE TC 9.9 thermal guidelines

Read the document

Publication details

Authors

Vojtech Dvorak; Vojtech Zavrel; J. I. Torrens Galdiz; Jan L. M. Hensen

Publisher

Building Simulation (Springer Science and Business Media LLC)

Year

2020

Volume

13

Issue

4

Pages

823-836

Document language

English

Source type

Scientific

Access

Open licence — downloadable

DOI

10.1007/s12273-020-0629-y

Licence

CC BY 4.0 — share with attribution

Licence evidence

doi.org

Citation