Temperature chaining

Content sourced from Wikipedia, licensed under CC BY-SA 3.0.

Temperature chaining is a way to turn a datacenter’s electricity use into usable heat. Instead of cooling all IT gear with one big cooling loop, this approach uses a water-based cooling system with multiple temperature stages. Each stage uses a different liquid-cooling technology that tolerates a different temperature, and the stages are connected in sequence (serially), not simply in parallel. This creates large temperature differences with only a small amount of water, so the datacenter can produce warm water that can be reused elsewhere.

How it works
- A shared water infrastructure is split into several levels with increasing temperatures.
- Different liquid cooling technologies are used at each stage, chosen for how hot the water can be while keeping IT components safe.
- The cooling circuit is designed as a cascade: the output of one stage feeds into the next, instead of feeding all stages from the same water supply. This creates bigger temperature gaps (high delta-T) and reduces the amount of water needed.
- The result is a facility that can deliver stable, high-temperature water to a heat-reuse user, turning the datacenter from a pure electricity consumer into a thermal energy producer.

Applications and benefits
- Temperature chaining supports heat reuse in heating systems and even in district heating networks, increasing flexibility and future-proofing energy networks.
- By using liquid cooling, IT equipment can run more efficiently and at higher densities. Liquid carries away heat more effectively than air, so fans and power used for cooling can be reduced.
- A hybrid approach—mixing several liquid cooling technologies in the right configuration—lets a datacenter support a wide range of platforms and services, regardless of the specific design or footprint.
- Reusing heat can lower overall energy use and reduce cooling overhead. The open-circuit heat reuse model, where the datacenter passes warm water to an external user, is especially sustainable: the facility acts like a large water heater with minimal cooling infrastructure.

Liquid cooling technologies (quick overview)
- Cooling at room, rack, or chip level and immersion: four broad categories.
- CRAC/CRAH can be adapted to water cooling, absorbing heat in the space or at the rack level.
- Indirect Liquid Cooling (ILC): water-cooled racks with rear-door or in-row heat exchangers; heat from air-cooled IT is captured as it leaves the rack, reducing the need for separate room cooling.
- Direct Liquid Cooling (DLC): coolers with cold plates mounted on chips reduce fan energy and improve IT efficiency, though other components may still need some cooling.
- Total Liquid Cooling (TLC): immerses IT components fully in liquid, minimizing energy loss and eliminating most fans; requires a dielectric liquid (like oil) to insulate against electricity.
- No single solution fits all. Each platform should use a mix of the technologies that best fit its parts (storage, servers, specialized equipment).

Design notes
- Each cooling technology must be controllable (for example, via PLC) and compatible in fittings and liquids.
- Systems are often segmented, with different temperatures in different parts of the circuit. The output of one stage feeds the input of a higher-temperature stage, creating a controlled, multi-step temperature rise.
- Return loops, buffer tanks, and pumps help manage volume and pressure between stages.
- In practice, a straightforward, smaller setup can still achieve temperature chaining by using a loop with a mixing valve and buffer tank to gradually raise the cooling temperature and maintain a steady output.

Bottom line
Temperature chaining makes it possible to capture and reuse more of a datacenter’s heat by using a multi-stage, serial liquid-cooling system. This increases energy efficiency, boosts IT density, reduces cooling needs, and supports open opportunities for heat reuse beyond the data center walls.