Objective. Students will identify cooling infrastructure topology elements — chillers, cooling towers, CRAC/CRAH units, and piping loops — and evaluate whether a given mechanical design satisfies the redundancy requirements of a specified Tier.
This lesson, its worked examples and its practice questions has not yet been verified by an industry-certified professional or by a certification or approval body. It is a study draft and may contain errors. Do not use it as instructions for work on a job site or on live equipment. Always follow your instructor, your supervising professional, and current official references and regulations.
VoTechWorks program approval and accreditation are still in progress. Credentials are issued only by the certifying bodies named in each program, after their own exams.
Data centers generate substantial heat from IT equipment, and that heat must be removed continuously to prevent equipment failure. The mechanical cooling infrastructure is the system responsible for this removal. Understanding how cooling components are arranged — their topology — and how that topology must change across Uptime Institute Tier levels is essential for anyone working in data center facilities.
A typical chilled-water cooling plant contains several major elements that work together as a system.
Chillers are refrigeration machines that remove heat from water, producing cold water (typically in the range of 40–55 °F supply, though actual setpoints vary by design). That chilled water is then distributed to cooling equipment inside the data center.
Cooling towers (or dry coolers in some configurations) reject the heat absorbed by the condenser water to the outdoor atmosphere. They work in a cycle with the chillers.
CRAC units (Computer Room Air Conditioners) use a direct-expansion refrigerant circuit built into the unit itself. They cool return air from the room and discharge conditioned air back into the space without relying on a central chilled-water loop.
CRAH units (Computer Room Air Handlers) use chilled water supplied from a central plant rather than a self-contained refrigerant circuit. A CRAH unit is essentially a fan coil: warm air passes over a chilled-water coil, the coil absorbs the heat, and cooled air returns to the space. CRAH units are common in larger facilities with central chilled-water plants.
Piping loops distribute chilled water from the plant to CRAH units and return warmer water back. A well-designed loop includes isolation valves at every major branch so that individual components can be isolated for maintenance without interrupting flow to the rest of the system.
Primary and secondary pumping is a common arrangement in which primary pumps circulate water through the chillers at a constant flow rate while secondary pumps distribute water through the distribution loop to CRAH units, often at variable flow rates controlled by demand.
Air-side economizers bring in outdoor air directly when conditions (temperature and humidity) allow, reducing or eliminating the mechanical cooling load. Their configuration must account for redundancy just as water-side equipment must.
Direct expansion (DX) systems use refrigerant rather than chilled water as the heat-transfer medium at the point of cooling. CRAC units are the most common example. DX systems are simpler to deploy but can be harder to scale with full redundancy.
The Uptime Institute Tier Standard describes four tiers of data center infrastructure. For mechanical systems, the key distinction between Tier II and Tier III is the concept of concurrent maintainability.
Tier II provides redundant capacity components (N+1 or better for individual components), but the site may require a planned shutdown to perform maintenance on distribution paths. A Tier II site has redundant chillers, pumps, and CRAH units, but the distribution piping and paths may be single-routed. If a pipe segment requires maintenance, the entire cooling loop may need to be shut down.
Tier III requires concurrent maintainability: any single component — including distribution paths — can be taken out of service for planned maintenance without causing a shutdown of IT load. This requires at least two independent distribution paths (not just two components at a single point). Each path must be capable of supporting the full load on its own, and each must be independently isolatable via valves.
In practice, Tier III mechanical designs include:
A common compliance gap is providing N+1 chillers (a Tier II concept) without providing a second independent distribution path. Adding a spare chiller does not satisfy concurrent maintainability if both chillers feed a single distribution loop with no bypass or alternate path.
When evaluating a mechanical diagram, ask these questions for each subsystem:
1. Is there a redundant component? (N+1 minimum)
2. Is there a redundant path? (Tier III requires this for distribution)
3. Can each component and each path segment be independently isolated? (Valves present at all branch points?)
4. Can maintenance be performed on any single element while the system continues to operate at full capacity?
If the answer to question 2 or 3 is no, the design does not meet Tier III concurrent maintainability regardless of how many spare chillers or CRAH units are present.
Working near mechanical cooling systems carries real hazards. Always observe lockout/tagout (LOTO) procedures before working on any pump, valve actuator, or chiller under maintenance — stored hydraulic pressure and rotating machinery are both serious hazards. Wear appropriate PPE: at minimum, safety glasses and gloves when working near chilled-water connections, and hearing protection near operating chillers and cooling towers. Cooling towers may carry biological hazards (Legionella) and require specific safety and water-treatment protocols — never perform maintenance inside a cooling tower without reviewing the facility's confined-space and water-treatment procedures. Refrigerant handling on DX and CRAC systems requires EPA Section 608 certification and the correct PPE for refrigerant exposure.
Answer each question, then check it. Nothing is recorded; this is your own practice.
A free tech program — complimentary, as our commitment to get America working. Help us build VoTechWorks. Tell us what works, what doesn’t, and what you want to learn.