Data centers are, in thermal terms, machines that turn electricity into heat. Every kilowatt delivered to the servers has to be removed again, around the clock, in the hottest week of the year as well as in winter. As rack densities rise and liquid cooling moves from niche to mainstream, the outdoor heat-rejection equipment has become a central design decision. Radiators play a part in three places: rejecting the heat of liquid-cooled IT, providing free cooling alongside chillers, and cooling the standby generator sets that protect the facility.
Liquid Cooling Changes The Heat-Rejection Picture
Direct-to-chip and rear-door cooling carry the server heat away in a water or glycol loop at temperatures far higher than traditional chilled water. That warmer loop can often be cooled by outdoor air for most - or all - of the year without running a compressor. A radiator with electric fans then becomes the main heat-rejection device: the fans run only as fast as the load and the ambient require, and the energy saved directly improves the facility's efficiency.
TECS designs V-type radiators for exactly this kind of duty, where many units stand side by side on a roof or a yard, and horizontal radiators where the layout allows.
Free Cooling Alongside Chillers
Where chillers remain in the system, a radiator in the chilled-water or condenser-water circuit can take load off them whenever the outdoor air is cold enough - known as free cooling or a chiller economiser. The number of free-cooling hours depends on the climate, the water temperatures and the control sequence, so the radiator is sized with the seasonal weather data of the site rather than a single design point.
Cooling The Backup Power
Every serious data center keeps generator sets ready to take over from the grid. Their engines need radiators that reject the full jacket-water and charge-air heat at the hottest design ambient, often while many sets run side by side. Remote radiators let the generator hall stay compact and quiet, and LT/HT designs cool both engine circuits in one unit. The same rules apply as for any power plant: the engine manufacturer's heat rejection, flow and pressure-drop limits, and an honest view of recirculation between neighbouring units.
What To Specify
- The heat load and the fluid - water or glycol, flow rate, supply and return temperatures.
- The climate data - design ambient for peak load and the annual temperature profile for free cooling.
- Redundancy - how many units or fans may be out of service while the load is still met.
- Sound - the limit at the property line, which matters for urban and edge sites.
- Control - fan speed control, sequencing and the interface with the building management system.
- The site atmosphere - corrosion category, dust and coastal exposure for fins and casings.
As data centers move towards liquid cooling and higher densities, outdoor heat rejection with radiators becomes more efficient and more important at the same time. Designing the radiators for the real climate, the redundancy and the sound limits of the site keeps both the IT load and the backup power secure. See how TECS radiators serve data centers, or contact our engineers.
