Topic guide

Cooling

Explore technical publications on data centre cooling systems, thermal loads and operating conditions. Review air and liquid approaches, energy use and operational requirements within each source’s scope.

Sources
339
Downloadable PDFs
92
Years
2006–2026

All sources

169–192 / 339

Research paperPDF

Sensors (MDPI AG) 2022

Prediction of Overall Energy Consumption of Data Centers in Different Locations

In siting, do not use PUE as the only criterion; evaluate it together with electricity cost (tariff) and carbon (CUE, grid mix) estimates derived from PUE, since a low-PUE site may have a high tariff or carbon-intensive grid.

  • For an assumed 100 million kWh/yr of IT energy, estimated annual electricity cost ranged from $8.1376M (Delinha) to $9.6742M (Yinchuan) among five regions.
Research paperCitation

Energy and Buildings (Elsevier) 2021

Prototype energy models for data centers

Use the prototype densities and the approach-temperature concept to calibrate the tool's hourly cooling model and containment effectiveness.

  • The models follow ASHRAE 90.1 minimum requirements across vintages.
Technical guideFree to read

Uptime Institute Journal 2020

Rack Density is Rising (Andy Lawrence)

Ask users for modal and peak rack density separately, not just the average, and size cooling and power for the peak cluster.

  • Mean density was calculated after excluding respondents above 30 kW as outliers.
Technical guidePDF

Open Compute Project Foundation 2022

Reservoir and Pumping Unit (RPU) Specification, Ver 1.0

Model AALC (RPU plus door or rack liquid-to-air heat exchanger) as a cooling option for air-cooled halls, counting the RPU's 4RU space, 75 kg mass and the 2%/5% parasitic power caps against rack U-space, weight and power budgets.

  • AALC total power is at most 5% of rated cooling capacity, of which the RPU takes at most 40%.
Research paperPDF

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.
Research paperPDF

Building Simulation (Springer Science and Business Media LLC) 2020

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

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.

  • Cooperation with ATES improves average PUE from 1.78 to 1.45.
Research paperPDF

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.
Research paperPDF

Sustainability (MDPI AG) 2021

Strategies for Improving the Sustainability of Data Centers via Energy Mix, Energy Conservation, and Circular Energy

Calculate UPS load ratio (80% for 800 kW in a modular arrangement, 53% in parallel-redundant) and losses at design stage; consider specifying UPS efficiency at 30% load and compare a SiC-based power stage (above 98%).

  • CUE is defined as total CO2 emissions over ICT energy, or equivalently the CO2 emission factor × PUE.
Research paperPDF

Journal of Cleaner Production (Elsevier BV) 2025

Sustainable AI infrastructure: A scenario-based forecast of water footprint under uncertainty

In water-footprint planning, calculate on-site (WUE), off-site electricity (PUE × EWIF) and embodied hardware water as separate items.

  • Under BAU, water demand rises nearly sevenfold between 2030 and 2050; improvements in WUE, EWIF and embodied water could cut 2050 use by up to 95% against BAU.
Research paperFree to read

arXiv preprint (Microsoft Azure Research / University of Illinois) 2025

TAPAS: Thermal- and Power-Aware Scheduling for LLM Inference in Cloud Platforms

Couple thermal and power headroom in the design tool: show how inlet temperature rise with outdoor conditions (for economiser or adiabatic systems) erodes GPU thermal margin and can trigger throttling.

  • GPU temperature must stay below a threshold (for example 85 C) beyond which hardware throttles computation.

All 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.

Monthly newsletter · free

Keep your reading list current

Once a month: new library sources, changes to Turkish regulations and upcoming data centre events.