Dry Coolers in Data Center Cooling Systems

Dry Cooler in Data Center's Cooling System

In the data center cooling system, Dry Cooler is a kind of energy-saving equipment that uses air and coolant for sensible heat exchange, mainly cooling down the circulating coolant through the natural cold source (outdoor cold air), which is the core component of the “free cooling” technology, and is crucial for reducing the PUE (energy usage efficiency) of the data center. It is the core component of “free cooling” technology, which is crucial for reducing PUE (energy usage efficiency) in data centers.

The core value of the dry cooler is to utilize outdoor low temperature air to replace the mechanical cooling of the traditional chiller to reduce energy consumption, and its application mode depends on the local climatic conditions:

1. Independent Free Cooling Mode

Applicable Scenarios: Cool climate areas (e.g., northern winter, plateau area), with the outdoor temperature below 20℃ for a long period of time.

Operation logic: when the outdoor temperature ≤ coolant target temperature (such as 20 ℃), the dry cooler runs alone, completely replacing the chiller, through the fan will be cold air into the coil, directly for the coolant cooling. At this time, the data center only needs to consume fan power, PUE can be as low as 1.1-1.2 (traditional chiller PUE is 1.4-1.8). 2.

2. Linkage mode with chiller

Scenario: Temperate or subtropical areas (high temperature in summer and low temperature in winter).

Operation logic:

Transitional season (spring/autumn, outdoor 10-25℃): dry cooler is dominant, chiller is auxiliary (only started when dry cooler cooling is insufficient);

Summer (outdoor >25℃): dry cooler efficiency drops, switch to chiller dominant, dry cooler as pre-cooling equipment (first cool down the coolant to 30℃, and then enter chiller to further reduce to 15℃), reduce the load of the chiller.

3. Modular deployment

Large data centers mostly adopt the modular design of “multiple dry coolers in parallel”, which does not affect the overall operation when a single unit fails, and can flexibly start and stop part of the equipment according to the real-time heat load (e.g., turn off half of the fans at night when the load is low), thus further saving energy.

Selection and Design Points

Matching Heat Load: Calculate the required heat transfer area (㎡) according to the total heat load of the data center (kW), the formula reference: heat transfer area = heat load ÷ (heat transfer coefficient × logarithmic average temperature difference), in which the heat transfer coefficient is related to the material of the coiler and the density of the fins (copper fins > aluminum fins).

Climate parameters:

Design temperature: take the local summer ventilation outdoor temperature (such as Beijing 30.8 ℃, Guangzhou 33.5 ℃), to ensure that extreme high temperatures can still meet the heat dissipation needs;

wind speed: natural convection type is suitable for wind speed stabilization area, forced ventilation type need to calculate the fan air pressure (to overcome the resistance of the coiler + installation height of the wind resistance).

Corrosion resistance of materials: 316 stainless steel coiler + anti-corrosion coating shell should be selected for coastal areas (high salt spray) or industrial areas (high pollution) to avoid rust and corrosion affecting the life.

Intelligent control: Give priority to the dry cooler that supports linkage with data center BMS system, which can automatically adjust the fan speed (inverter fan) according to real-time outdoor temperature and coolant temperature to achieve “on-demand heat dissipation”.

Redundancy design: at least 10%-20% of the heat transfer margin is reserved (to cope with sudden load growth or equipment failure), and the fan adopts N+1 redundancy configuration.