HVAC Sizing for Data Center Support Spaces
A Step-by-Step Engineering Guide for Gray Space, Administrative, Operational, Storage, and Burn-In Areas
Executive Summary
Heating, ventilation, and air-conditioning (HVAC) design for the non-white-space areas of a data center is consistently underestimated. Most engineering teams over-rotate on the Information Technology hall, where Computer Room Air Handler (CRAH) and Computer Room Air Conditioner (CRAC) sizing is highly proceduralized, while the gray space, administrative space, operational support space, storage, loading dock, and burn-in areas receive inconsistent treatment. The result is a class of operational risk that surfaces during commissioning or, worse, during a real failure event. Battery rooms over-pressurize. Storage rooms reach summer temperatures incompatible with spare-parts shelf life. Burn-in rooms trip on high-temperature alarms in the middle of a critical test campaign. Loading docks freeze in winter. None of these failures should ever happen, and none of them require new technology to prevent. They require disciplined application of well-established sizing methodology.
This white paper presents a step-by-step methodology that thermal engineers, mechanical plant engineers, and architectural designers can apply consistently across every non-white-space area of a data center. It walks the reader through space classification, environmental envelope selection, internal heat-gain calculation, ventilation rate determination, exhaust and pressurization strategy, redundancy selection, equipment selection, distribution sizing, and commissioning. It anchors every design decision to the controlling codes and standards — the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) standards 55, 62.1, 90.1, 90.4, and Technical Committee 9.9 thermal guidelines; the National Fire Protection Association (NFPA) standards 75, 76, 70, and 855; the International Energy Conservation Code (IECC); and the International Building Code (IBC).
The audience for this paper is the working thermal engineer, not the code official and not the executive. The structure is procedural. The expected outcome is that an engineer who has read this document can size any non-white-space area of an enterprise or AI-class data center using a defensible, repeatable, and reviewable approach that aligns with the practice’s Phase 1 Thermal Design Guidance and Mechanical Plant Design Guidance reference framework.
Key conclusions:
· Sizing must begin with the room program, not with equipment selection. Every space has a specific function, occupancy, equipment population, and environmental envelope. Size from the function down, not from the equipment up.
· Use ASHRAE 55 comfort criteria for occupied non-process spaces (offices, NOC, SOC, conference, break rooms). Use ASHRAE TC 9.9 envelopes only for spaces that house IT equipment. Do not conflate the two.
· Use ASHRAE 62.1 ventilation rate procedure for occupied spaces and process-driven dilution for battery, generator, fuel, and burn-in spaces. These two are not interchangeable.
· Pressurization is a deliberate design decision, not an accident. Establish a documented pressurization cascade across every door and verify it during commissioning.
· N+1 redundancy is the floor for any space that holds critical IT-adjacent infrastructure. Comfort spaces can use N. Process spaces require dedicated exhaust treatment that cannot be shared with comfort systems.
· Document every assumption in a load schedule. Reviewers, commissioning agents, and operators must be able to trace every Btu and every cubic foot per minute back to a specific source.
Full white paper below
