Data Center Cooling System Design: From Heat Load to System Selection
Data center cooling system design should be driven by the sensible heat load and the airflow path, not by room area or the number of air conditioners. The process starts with a load profile, then selects capacity, airflow, redundancy, controls, and site-specific systems that can keep equipment within its specified inlet temperature and humidity range.
Build the Heat Load Profile
- IT equipment measured or projected load in kW.
- UPS, power distribution, lighting, and people loads.
- Solar and envelope heat gain for the room or module.
- Peak load and normal operating load.
- Growth assumptions and future capacity.
Each kW of IT load produces roughly 1 kW of sensible heat, but the total load also includes power conversion and environmental heat. A cooling design should state which loads are included and what load growth has been assumed.
Specify Sensible Cooling Capacity
Precision cooling is rated by both total and sensible capacity. For a sealed data center, the sensible capacity is the useful capacity because most equipment heat does not add moisture to the air. The design should compare the room’s sensible load with the equipment sensible capacity at the intended entering air and airflow conditions.
Design Airflow and Rack Layout
Cooling only works if air reaches the equipment intakes and returns to the cooling unit. The design should define cold and hot aisles, supply and return airflow, rack layout, blanking panels, cable openings, and any containment or close-coupled cooling used for high-density rows.
If rack power density varies, separate the cooling design by zone rather than treating the whole room as one average density. High-density racks may need in-row or rack-level cooling while lower-density areas use room-level units.
Select Redundancy and Control Strategy
Cooling redundancy should cover capacity, power, and any shared water or refrigerant paths. N+1 is common for facilities that need one unit out of service during maintenance. The control strategy should describe supply temperature setpoints, fan response, alarm thresholds, and behavior when one unit or path is unavailable.
Include Site-Specific Systems
Direct expansion cooling, chilled water, free cooling, and close-coupled systems each depend on local climate, available utilities, space, and operating cost. The design should identify the site conditions used for selection and state what happens at the maximum design ambient temperature.
Deliver a Verifiable Design Package
A complete package should include load calculations, airflow diagrams, equipment schedules, redundancy diagrams, control descriptions, test requirements, and acceptance criteria. The owner should be able to verify the installed system against the design basis before handover.
Frequently Asked Questions
How do I design cooling for a data center?
Start with the sensible heat load and airflow path, then select capacity, redundancy, controls, and site-specific cooling systems that can maintain the required equipment environment.
Why is room area not enough for cooling design?
Two rooms of the same size can have very different IT loads and airflow requirements. Power density and heat load are more important than floor area.
Should cooling design include power and water paths?
Yes. Redundancy is only meaningful if the electrical and water or refrigerant paths feeding the cooling units are also included in the design.
Need a cooling system design review or an RFQ for a new facility? Provide the heat load, room layout, and site conditions to VERHI.
This guide is based on engineering selection principles and VERHI's published scope. Site-specific values should be confirmed with the VERHI engineering team.
