Techno-Economic Comparison of Data Center Cooling Using Magnetic Bearing Chillers and Aquifer Thermal Energy Storage
Energies (MDPI AG) · Apurva Malpure, Andrew Stumpf et al.
An hourly techno-economic comparison of three chilled-water plants for the same data-centre IT load profile: a water-cooled centrifugal chiller baseline, magnetic-bearing chillers (MBC), and MBC with a screening-level aquifer thermal energy storage (ATES) discharge.
CC BY 4.0 — share with attribution
On this page
Overview
Data centres are large and rapidly growing electricity consumers, and cooling systems account for a substantial share of their energy demand. The study presents a climate-sensitive, hourly techno-economic comparison of three cooling configurations under consistent operating assumptions. All share a waterside economiser and two 4058 kW chillers, and Phoenix and Fairbanks bracket US cooling climates. The plant uses 7 °C chilled-water supply with a 5 K delta-T, and a cooling tower with a 4 K approach and a 2 °C minimum condenser-water temperature.
Phoenix baseline operating modes: 204 h full free cooling, 1748 h partial free cooling and 6808 h mechanical only; Fairbanks: 5350 h, 1297 h and 2113 h. MBC cut annual cooling electricity from 1169.4 to 957.4 MWh in Phoenix (18.1%) and from 361.6 to 319.4 MWh in Fairbanks (11.7%). Cooling-only PUE' was 1.095 (baseline) and 1.078 (MBC) in Phoenix, and 1.029 and 1.026 in Fairbanks.
With a premium capex of USD 123,527, MBC paid back in 5.8 years in Phoenix but 36.6 years in Fairbanks. Peak cooling demand fell by 119.4 kW and 71.6 kW respectively. MBC+ATES gave no extra electricity savings over MBC-only, and its better discounted savings disappear (NPV negative) if its O&M is set equal to MBC-only.
Limits: the authors stress that the ATES case is a simplified screening scenario and that results depend on tariff and O&M assumptions; ATES should not be read as a complete seasonal design or full techno-economic assessment. Only two climates were examined.
Key points
- The plant uses 7 °C chilled-water supply with a 5 K delta-T, and a cooling tower with a 4 K approach and a 2 °C minimum condenser-water temperature; tower water temperature = max(outdoor wet-bulb + approach, minimum).
- Phoenix baseline operating modes: 204 h full free cooling, 1748 h partial free cooling and 6808 h mechanical only; Fairbanks: 5350 h, 1297 h and 2113 h.
- MBC cut annual cooling electricity from 1169.4 to 957.4 MWh in Phoenix (18.1%) and from 361.6 to 319.4 MWh in Fairbanks (11.7%).
- Cooling-only PUE' was 1.095 (baseline) and 1.078 (MBC) in Phoenix, and 1.029 and 1.026 in Fairbanks.
- With a premium capex of USD 123,527, MBC paid back in 5.8 years in Phoenix but 36.6 years in Fairbanks; peak cooling demand fell by 119.4 kW and 71.6 kW.
- MBC+ATES gave no extra electricity savings over MBC-only; its better discounted savings disappear (NPV negative) if its O&M is set equal to MBC-only.
Using it in design
- Evaluate magnetic-bearing chiller investment by climate: payback was 5.8 years in a hot climate relying more on mechanical cooling, 36.6 years in a cold climate dominated by free cooling.
- Model chiller staging and operating modes (full, partial free cooling, mechanical) hourly and report cooling-only PUE and peak demand.
- Show storage-option savings such as ATES with sensitivity to O&M and tariff assumptions; NPV turned negative when O&M was equalised.
Monthly newsletter · free
Keep your reading list current
Once a month: new library sources, changes to Turkish regulations and upcoming data centre events.
SoSo CAD
Check cooling capacity alongside your design
Review zone heat loads, available capacity and the capacity balance after the largest unit is lost in SoSo CAD.
Request a demo
Who should read it
Designers and investment decision makers comparing chilled-water plant chiller technology and storage options techno-economically.
Document contents
- 1. Introduction
- 2. Materials and Methods (2.1 Assumptions and Boundary Conditions, 2.2 System Description, 2.3 Data Inputs and Parameters: weather and humidity data, IT load profile, indoor environmental conditions; 2.4 Chiller and Auxiliary Power Model, 2.5 ATES Dispatch and Storage Balance, 2.6 Hourly Simulation Workflow, 2.7 Financial Evaluation, 2.8 Model Verification and Quality Assurance)
- 3. Results (3.1 Energy Performance: operating modes, energy consumption and PUE'; 3.2 Economic Performance)
- 4. Discussion (4.1 Benchmarking Against Previous Studies, 4.2 Limitations and Future Work)
- 5. Conclusions
Related standards
- ASHRAE 90.4
- ASHRAE TC9.9 Thermal Guidelines
Read the document
magnetic-bearing-chiller-ates-dc-cooling-2026.pdf
CC BY 4.0 — share with attribution
The document is provided unaltered because its licence permits redistribution.
Publication details
- Authors
- Apurva Malpure; Andrew Stumpf; Upasana Pandey; Yu-Feng Lin; Craig Bradshaw
- Publisher
- Energies (MDPI AG)
- Year
- 2026
- Volume
- 19
- Issue
- 17
- Pages
- 3947
- Document language
- English
- Source type
- Scientific
- Access
- Open licence — downloadable
- Licence
- CC BY 4.0 — share with attribution
- Licence evidence
- www.mdpi.com
Original licence note
p.1 states the article is open access under the Creative Commons Attribution (CC BY) licence.
Citation
Apurva Malpure; Andrew Stumpf; Upasana Pandey; Yu-Feng Lin; Craig Bradshaw (2026). Techno-Economic Comparison of Data Center Cooling Using Magnetic Bearing Chillers and Aquifer Thermal Energy Storage. Energies (MDPI AG). 19(17), 3947. https://doi.org/10.3390/en19173947
Events on this subject
Browse the events calendarAll content and rights belong to the respective authors and publishers. SoSo CAD provides only the citation, its own summary and a link to the source, and is not responsible for the sources' content.
Related sources
All sources on this topic
Energies (MDPI AG) 2026
Smart Reuse of Waste Heat from Data Centres: Energy and Exergy Analysis for a District-Heating Network in Bulgaria
Set data-centre setpoints (air temperature, free-cooling threshold) together with waste-heat sales: warmer setpoints and more free cooling cut cooling electricity by about 45% but also cut recoverable heat to about 45%.
- Centralised DH heat pumps are cited as giving COP 3-5 for supply temperatures up to 70-80 °C.
Buildings (MDPI AG) 2026
Load-Adaptive Operational Control of an Indirect Evaporative Cooling System for Coordinated Energy and Water Use in a Modular Data Center in a Cold Region
In IEC control, stage airflow to follow load rate and spray to follow room temperature, and prefer a low fresh/return air ratio.
- Going from load rate 0.7 to 1.0 raised annual electricity by 16.2% but water by 74.8% and WUE by 70.6%.
Applied System Innovation (MDPI AG) 2026
Proactive Cooling Control Algorithm for Data Centers Based on LSTM-Driven Predictive Thermal Analysis
For predictive cooling control, plan at least one inlet and one outlet temperature sensor per rack; below that, prediction error rose by about 22%.
- The LSTM uses a 60 min input window and a 30 min prediction horizon, with a reported RMSE of 0.25 and R² of 0.985.
Sustainability (MDPI AG) 2026
Retrofit Optimization of Raised-Floor Plenum Thermal Performance for Energy-Efficient and Sustainable Operation of Non-Standard Campus Data Centers
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%).
- 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.
Buildings (MDPI AG) 2025
Energy Performance Study of a Data Center Combined Cooling System Integrated with Heat Storage and Waste Heat Recovery System
Assess heat recovery from chilled-water return with a water-source heat pump and hot-water storage in annual-hour simulation; match the winter and summer outlet temperature difference (45 and 30 °C) to the user's need.
- Free natural-resource cooling covers about 28% of the cooling demand.
Energies (MDPI AG) 2025
Energy Efficiency Measurement Method and Thermal Environment in Data Centers—A Literature Review
In PUE comparisons state the facility type (edge, SMB, enterprise, colocation, hyperscale) and assign IT/facility power categories per The Green Grid definitions in detail; PUE across types is not directly comparable.
- No study of staff thermal comfort in data centres was found; in one computer lab PMV was 1.1-1.4, outside the ISO ±0.5 band.