Liquid cooling

White paperPDF

Open Compute Project Foundation · Year not stated

Guidelines for Using Propylene Glycol-Based Heat Transfer Fluids in Single-Phase Cold Plate-Based Liquid Cooled Racks (OCP white paper)

Write the PG25 or PG55 choice into the fluid specification according to the freeze protection the climate needs, and record that keeping glycol above 25% by volume removes the need for biocide.

• Maximum corrosion rates under ASTM D8040-17: copper 0.22 mpy (5 mg/coupon), brass 0.46 mpy (10 mg/coupon).

Liquid coolingCoolingCommissioning

Read / Download

White paperFree to read

Schneider Electric · 2026

How 6 AI Attributes Change Data Center Design

Add an AI cluster design profile with rack defaults of 750x1,200 mm, 48U, static 2,270 kg and dynamic 1,600 kg.

• GPU TDP trend: V100 300 W, A100 400 W, H100 700 W, B200 1,000 W, B300 1,400 W.

Liquid coolingRacksElectrical

Technical guidePDF

Open Compute Project Foundation · 2023

Immersion Requirements Rev 2.0, First Amended (Revision 2.10, August 18, 2023)

Compare vendor offers on the same template: request W32 footprint density (kW/m2), full-solution static floor load, ceiling clearance and non-IT power per kW of IT in writing.

• Containment must hold at least 100% of the fluid volume of any single system without breach.

Liquid coolingCoolingFire safety

Read / Download

White paperPDF

Open Compute Project Foundation · 2022

Leak Detection and Intervention, Revision 1.0 (OCP Advanced Cooling Solutions white paper)

Run a probability-and-impact leak risk analysis at project start, specify BSPP/O-ring joints for serviceable connections and check the O-ring compound against the coolant chemistry using the full wetted-materials list.

• A UQD2 dry-break quick disconnect is specified to drip less than 0.063 cc when disconnected at 200 psi.

Liquid coolingReliabilityCommissioning

Read / Download

Technical guideCitation

NREL · 2018

DCDM1: Lessons Learned from the World's Most Energy Efficient Data Center (NREL/PR-2C00-70433, March 12, 2018)

Fix two design temperatures early, the coolest supply you can produce on the hottest day and the return temperature needed for heat reuse, and select IT equipment against them.

• Ten 100-kW racks take much less space than the equivalent fifty 20-kW air-cooled racks.

Liquid coolingWaterPerformance metrics

White paperFree to read

Schneider Electric · 2019

Liquid Cooling Technologies for Data Centers and Edge Applications

Create a coolingMethod enum (air, rdhx, dtc_1p, dtc_2p, immersion_chassis, immersion_tub_1p, immersion_tub_2p).

• Immersion removes more than 95% of heat via liquid.

Liquid coolingCoolingEnergy

Technical guideCitation

The Green Grid · 2016

Liquid-cooled IT for the Masses: Achieving a Data Center PUE of 1.1

For designs capturing at least 70-80% of heat to liquid with chiller-less warm-water loops, set the PUE target near 1.1.

• Format: 2016 webcast; related framework ASHRAE TC9.9 W classes.

Liquid coolingEnergyPerformance metrics

Technical guidePDF

Open Compute Project Foundation · 2025

Modular Technology Cooling System for "Cloud Scale": Design & Delivery of Liquid to Rack Distribution Systems, Version 1.0 (25 August 2025)

When sizing TCS capacity against electrical capacity, state the cold-plate capture fraction (70-80%) and target delta-T (10 or 20 C) explicitly, and use the roughly doubled delivery at 20 C (4,845 versus 2,422 kW) as design headroom.

• Most IT equipment is designed for a 10 C coolant delta-T, with some designs considering 20 C.

Liquid coolingCoolingReliability

Read / Download

White paperFree to read

Schneider Electric · 2025

Navigating Liquid Cooling Architectures for Data Centers with AI Workloads

Model CDUs as catalog items with exchange type (LA, LL, LR), form factor (rack, floor), capacity kW, approach temperature, flow L/min and redundancy.

• Rack-mounted CDUs: about 20 to 40 kW (L-A) or about 40 to 80 kW (L-L).

Liquid coolingCoolingRacks

Technical guidePDF

Open Compute Project Foundation · 2019

OCP ACS Liquid Cooling Cold Plate Requirements Document, Revision 1.0 (October 9, 2019)

Use the parameter tables for cold plates, CDUs, manifolds, couplings and pumps as the data sheet checklist when comparing vendor solutions.

• CPU/GPU cold plates commonly capture 70-75% of IT equipment heat in the hybrid basic class.

Liquid coolingCoolingRacks

Read / Download

Technical guidePDF

Open Compute Project Foundation · 2026

OCP Cooling Environments: Cold Plate Sub-Project Design Specification, Universal Quick Disconnect (UQD), v2.0.0 (effective September 22, 2026)

Select size (02-08) by rack or server loop flow: rated flows run from 2.27 to 17.79 L/min; estimate pressure drop from the Cv range and ask for correction or extra testing where the fluid differs from water.

• Cv ranges: 0.30-0.43 for size 02, 1.1-1.6 for 04, 2.2-3.3 for 06 and 3.6-5.6 for 08.

Liquid coolingRacksCommissioning

Read / Download

Technical guidePDF

Open Compute Project Foundation · 2020

Design Guidelines for Immersion-Cooled IT Equipment, Revision 1.0 (December 2020)

Compare the bill of materials of a server to be immersed (cable jackets, heat-shrink, EPDM capacitors, CMOS battery, relays, thermal paste) against the paper's problem table and obtain a compatibility statement from the fluid supplier for each item.

• In Soxhlet extraction testing, a mass loss below 2% indicates good material compatibility; the 48-hour test gives worst-case results.

Liquid coolingCoolingRacks

Read / Download

Technical guidePDF

Open Compute Project Foundation · 2026

Open Data Center: Next-Generation ML Infrastructure Design Principles, Revision 0.7.0 (Design Specification, effective Aug 27, 2026)

Base floor and transport-route load checks on the 8,000 lb rack and the 15,000 lb rack-plus-sidecar combination, and include the weight of the moving equipment.

• Available quick connects currently top out at 2 in.

RacksLiquid coolingSite planning

Read / Download

White paperFree to read

Schneider Electric · 2026

Prefabricated White Space Module Time Study: Lessons from a High-Density Liquid-Cooled Deployment

Offer a 'prefabricated IT pod' option in the white-space designer, with catalog defaults 1.2 MW, 42 racks, 93 m2.

• Learning curve: earlier pods 18 h on site versus 15 h later, a 17% improvement.

Liquid coolingRacksCommissioning

Technical guideFree to read

Eaton / Trane · 2025

Reference Design #505: Eaton 1 GW Power Architecture Meets Trane Thermal Management

Voltage-selection rule: motors above about 200 kW (large chillers and pumps) default to MV (4.16/6.6/11 kV) feeders; compute I = P/(sqrt(3) x V x PF) to show cable savings.

• 1,500 HP pump: 1,420 A at 480 V vs 164 A at 4.16 kV

ElectricalUPSLiquid cooling

Technical guidePDF

U.S. DOE FEMP · 2021

Sandia's Liquid-Cooled Data Center Boosts Efficiency and Resiliency (FEMP success story, DOE/GO-102021-5582, July 2021)

Consider negative-pressure (vacuum) distribution as an option for reducing leak risk in direct liquid cooling; flexible hoses give layout flexibility.

• The system kept CPUs below 50 °C (122 °F); each CDU draws 3.4 kW at full load, about 1.6% of total system power.

Liquid coolingCoolingReliability

Read / Download

Technical guidePDF

Open Compute Project Foundation · 2026

Single-Phase Direct-to-Chip Cold Plate Test Specification, Version 1.0.0 (OCP Cooling Environments: Cold Plate Sub-Project, effective September 3, 2026)

Reference the specification in a cold plate purchase requirement and state which tests, limits and sampling the customer requires, since many limits are examples.

• Flow rate testing goes up to velocities below 1.5 m/s to ensure no erosion or leaks.

Liquid coolingReliabilityCommissioning

Read / Download

Technical guideFree to read

Microsoft · 2024

Sustainable by design: Next-generation datacenters consume zero water for cooling

Add 'closed-loop DLC + dry cooler / air-cooled economizing chiller (zero-evaporation)' to the cooling-option catalog with WUE_cooling about 0.

• Pilots in Phoenix and Mt. Pleasant were planned for 2026.

WaterLiquid coolingCooling

Technical guideCitation

NREL · 2017

Thermosyphon Cooler Hybrid System for Water Savings in an Energy-Efficient HPC Data Center: Modeling and Installation (NREL/CP-2C00-66690, February 2017)

Place a dry stage upstream of the tower and size it against the share of hours where warm return water exceeds dry-bulb temperature by enough margin, as the 9.4 °C transition point does here.

• Building heat reuse peaks at 0.434 MW of a 0.910 MW load, leaving 0.496 MW to reject to atmosphere.

CoolingLiquid coolingWater

Technical guideCitation

NREL · 2018

Thermosyphon Cooler Hybrid System for Water Savings in an Energy-Efficient HPC Data Center: Results (NREL/TP-2C00-72196, September 2018)

Check whether the warm return water from liquid cooling (here 84 °F after energy recovery) leaves enough temperature difference to ambient for a dry or hybrid stage, and size it against local dry-bulb hours.

• First-year heat rejection: 10.5% building heat reuse, 42.5% thermosyphon and 47% cooling towers.

CoolingWaterLiquid cooling

Technical guideFree to read

Vertiv · 2021

Understanding Data Center Liquid Cooling Options and Infrastructure Requirements

Rule: rack kW > 20 suggests containment plus in-row cooling or rear-door heat exchangers; rack kW > 30-40 recommends liquid.

• Single CDUs are available above 1 MW

Liquid coolingCoolingRacks

Technical guidePDF

U.S. DOE FEMP / NREL · 2019

Using Thermosyphon Hybrid Cooling System to Optimize Data Center Water Efficiency (DOE/GO-102019-5172, May 2019)

Arrange the heat-rejection hierarchy as in the scheme: heat reuse first, then dry cooling, evaporative towers last; write the control logic in that order.

• Installing the TSC required only a 1-day data center outage.

CoolingWaterLiquid cooling

Read / Download

Technical guideFree to read

Vertiv · 2025

Vertiv 360AI High Density Reference Design #006: 2x 1.28 MW NVIDIA Blackwell DGX SuperPODs

Seed an AI-hall wizard template with: compute rack 73 kW, support rack 14 kW, liquid fraction 0.78, power 4N/3, cooling N+1, mechanical UPS 2N.

• Total IT load 2,560 kW = 2 pods x 1,280 kW.

ElectricalUPSLiquid cooling

Technical guideFree to read

Vertiv · 2024

Vertiv 360AI High Density Reference Design #014: 2x 1.2 MW NVIDIA Blackwell DGX SuperPODs, 2N

Provide a "2N AI hall" preset to compare with the 4N/3 preset: 2N makes installed UPS 200% of load against 133% for 4N/3; show the capex and efficiency deltas.

• Row PDUs 600 kVA, 2 per pod, fed via STS.

ElectricalUPSLiquid cooling

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.