Zero-Carbon Development in Data Centers Using Waste Heat Recovery Technology: A Systematic Review

Overview

The advance of AI, big data and cloud computing is expanding data centres worldwide and worsening energy consumption and carbon emission challenges; according to IEA projections global data-centre electricity demand is expected to double by 2030. The review systematically assesses the role of waste heat recovery technologies in low-carbon development. The pathways are grouped as T1 heat collection and upgrading (heat pumps, liquid cooling, thermal storage), T2 waste-heat-to-power (ORC, Kalina cycle, thermoelectric generators), T3 waste-heat-enabled cooling and T4 coupling with district energy.

For each group it summarises operating temperature windows, typical applications and reported energy, carbon and payback results from case studies. Heat pumps for WHR are listed for source temperatures of 20-55 °C and absorption refrigeration for 30-85 °C. In a cited study, adding a water-tank thermal store saved 5% in annual energy costs, with payback under 15 years when storage efficiency stays above 80%. Borehole thermal energy storage raised utilisation of low-grade data-centre waste heat from 77% to 96%, cutting CO2 by up to 8% and annual energy cost by 6%, but payback extended to 19 years.

One reviewed WHR system saved 27,978 kWh of cooling electricity per year (23.9 tons CO2) with a payback of approximately 4.3 years. A TRNSYS 18 study compared connecting heat demand to the secondary side versus the primary side of the cooling distribution unit, and both schemes were profitable within one year.

The conclusion is that the best green-data-centre answer is an integrated electricity, heat, cooling and storage system matched to temperature levels, not a single technology, and the authors call for unified PUE/COP/LCA evaluation frameworks. Limits: as a systematic review it has no new measurements and the case figures come from different conditions.

Key points

1. Heat pumps for WHR are listed for source temperatures of 20-55 °C and absorption refrigeration for 30-85 °C.
2. In a cited study, adding a water-tank thermal store saved 5% in annual energy costs, with payback under 15 years when storage efficiency stays above 80%.
3. Borehole thermal energy storage raised utilisation of low-grade data-centre waste heat from 77% to 96%, cutting CO2 by up to 8% and annual energy cost by 6%, but payback extended to 19 years.
4. One reviewed WHR system saved 27,978 kWh of cooling electricity per year (23.9 tons CO2) with a payback of approximately 4.3 years.
5. A TRNSYS 18 study compared connecting heat demand to the secondary side versus the primary side of the cooling distribution unit, and both schemes were profitable within one year.
6. The review groups pathways as T1 heat collection and upgrading, T2 waste-heat-to-power, T3 waste-heat-enabled cooling and T4 coupling with district energy.

Using it in design

• In waste-heat use, first match technology to source temperature (heat pump 20-55 °C, absorption cooling 30-85 °C) and think of electricity, heat, cooling and storage together rather than a single technology.
• In thermal-storage decisions, weigh storage efficiency (above 80%) and payback (under 15 years for 5% cost saving, 19 years for borehole storage) together.
• Report waste heat recovery together with PUE, COP and LCA.

Who should read it

Planners and energy engineers evaluating data-centre waste-heat recovery options and technology matching.

Document contents

1. 1. Introduction
2. 2. Waste Heat Recovery Methods for Data Centers (2.1 Application of Heat Pump in Waste Heat Recovery, 2.2 Thermal Energy Storage, 2.3 Waste Heat Recovery by Absorption/Adsorption Cooling Technology)
3. 3. Power Generation Technology in Waste Heat Recovery (3.1 Organic Rankine Cycle System, 3.2 Kalina Cycle, 3.3 Thermoelectric Generators, 3.4 Electricity Production Realized by Other Technologies)
4. 4. Waste Heat Recovery for District Heating
5. 6. Conclusions

Read the document

Publication details

Authors

Lingfei Zhang; Zhanwen Zhao; Bohang Chen; Mingyu Zhao; Yangyang Chen

Publisher

Sustainability (MDPI AG)

Year

2025

Volume

17

Issue

22

Pages

10101

Document language

English

Source type

Scientific

Access

Open licence — downloadable

DOI

10.3390/su172210101

Licence

CC BY 4.0 — share with attribution

Licence evidence

www.mdpi.com

Citation