| Air-cooled chillers (ACCs) have emerged as the dominant mechanical cooling solution for data centers of all scales, offering meaningful advantages in water conservation, reduced infrastructure complexity, and faster deployment timelines compared to tower-based water-cooled alternatives. However, these benefits carry tradeoffs in thermodynamic efficiency — particularly at elevated ambient temperatures — that must be rigorously quantified and managed through strategic engineering. Correct ACC selection is not merely a capital procurement decision; it is a foundational determinant of energy efficiency, uptime resilience, and long-term sustainability performance aligned with increasingly demanding green data center targets. |
| ▶ Key Takeaways • Site environment drives capacity: Ambient design-day temperature, elevation, humidity, and corrosive atmospheres must inform chiller selection before any equipment shortlist is developed. • Recirculation is a critical risk: Hot condenser exhaust re-entering unit inlets raises effective entering air temperature (EAT), degrades capacity, and can trigger high-pressure lockouts — a threat mitigated by CFD site modeling. • Derate early and generously: Apply a minimum 10–15% capacity buffer for mission-critical facilities, accounting for ambient temperature exceedance, altitude, and fouled-coil scenarios. • Raise chilled water supply temperature: Increasing CHWST from 44°F (6.7°C) to 50°F (10°C) or higher can improve chiller COP by 12–24%, dramatically expanding economizer hours. • Evaluate IPLV, not just full-load efficiency: Data centers operate predominantly at partial load; Integrated Part Load Value (IPLV) is the most operationally relevant efficiency metric and should govern final equipment selection. |
1. Environmental Factors
The thermal and atmospheric environment at a proposed data center site is the single most consequential input to ACC selection. Designers must obtain ASHRAE design-day data — specifically the 0.4% or 1.0% summer dry-bulb design conditions from ASHRAE Fundamentals Handbook (Chapter 14 / ASHRAE 169) for the project location. The 0.4% value represents the dry-bulb temperature exceeded only 0.4% of annual hours (~35 hours per year) and is the recommended design basis for mission-critical cooling plants where high-pressure compressor trips must be avoided even during peak heat events.
Altitude derating is mandatory for sites above sea level. Reduced air density diminishes condenser airflow effectiveness at a rate of approximately 1–2% capacity reduction per 300 m (approximately 1,000 ft) above sea level. A facility at 1,500 m (4,921 ft) — such as Denver, Colorado — may require 5–10% additional chiller tonnage relative to sea-level ratings to deliver identical cooling output.
Humidity and corrosive atmospheres govern coil material specification. Coastal sites within 1.5 km (approximately 1 mile) of saltwater, or industrial sites near chemical processing or heavy vehicle traffic corridors, present accelerated fin-and-tube corrosion risks. Standard copper-tube, aluminum-fin condenser coils are inadequate in these environments; designers should specify epoxy-coated, E-coated (electro-deposited), or phenolic resin-coated fins, and may require copper fins or cupronickel alloys for the most aggressive marine exposures. Dust accumulation in arid climates reduces coil heat transfer and requires more frequent cleaning cycles.
Solar radiation increases the entering air temperature (EAT) at condenser inlets on exposed rooftop or ground-level installations. South- and west-facing unit inlets can experience EAT values 3–7°F (1.7–3.9°C) above recorded ambient dry-bulb due to solar loading on surrounding surfaces and radiant heat from rooftop membranes. This effect must be added to the design-day dry-bulb temperature when establishing the rated operating point.
Seismic zone requirements per ASCE 7 and IBC must be evaluated for rooftop or elevated pad installations. Chillers in Seismic Design Categories C–F require equipment anchorage calculations, seismic-rated spring isolators or snubbers, and flexible piping connections. These requirements add structural engineering scope and can affect chiller pad design, unit weight distribution, and clearance layouts.
2. Site Layout and Clearance Requirements
ACC placement is governed by clearance requirements that directly affect thermal performance, serviceability, and code compliance. Manufacturer-recommended minimum side clearances typically range from 2 to 6 ft (0.6–1.8 m) between unit casings and adjacent walls, structures, or other chillers. Top clearances for axial discharge fans must remain unobstructed — typically a minimum of 1.5× the fan discharge diameter above the fan stack to prevent discharge pressure buildup and recirculation. Parapet walls and mechanical penthouse overhangs that project above the fan discharge plane are a primary recirculation cause and must be accounted for in siting.
Prevailing wind direction at the site must be studied using local meteorological data (wind roses from NOAA or local airports). Units should be oriented so that condenser air discharge is directed away from adjacent unit inlets and away from building air intakes. Units in a row should be arranged with their discharge perpendicular to or aligned with the prevailing wind to assist exhaust dispersion rather than allow re-entrainment.
The table below summarizes general installation siting guidance:
| Siting Parameter | Recommended Minimum | Notes |
| Side clearance (unit-to-unit) | 4–6 ft (1.2–1.8 m) | Increase for larger tonnage units; follow OEM specs |
| Side clearance (unit-to-wall) | 3–5 ft (0.9–1.5 m) | Solid walls create recirculation pockets |
| Fan discharge clearance (vertical) | Unobstructed above fan plane | No overhead structures within fan discharge radius |
| Distance from building air intakes | ≥ 25 ft (7.6 m) | Per ASHRAE 62.1; greater preferred for high-capacity plants |
| Distance from exhaust louvers | ≥ 15 ft (4.6 m) | Building exhaust increases EAT if closer |
| Service aisle (compressor/coil access) | ≥ 6 ft (1.8 m) | Required for coil pull and compressor changeout |
Table 1. ACC Siting Clearance Reference (manufacturer specifications take precedence).
Ground-level vs. rooftop installation involves fundamental tradeoffs. Rooftop placement benefits from better exhaust air dispersion and reduced acoustic impact on occupied areas at grade, but imposes structural loading constraints (typically 100–200+ psf [4.8–9.6 kPa] for large chillers), rigging complexity, and restricted service access. Ground-level installations simplify maintenance and equipment replacement but require fencing, security provisions, and careful exhaust dispersion planning in constrained urban sites.
3. Recirculation Risks
Hot air recirculation is defined as the condition in which warm condenser exhaust air — discharged upward or horizontally from ACC fans — re-enters the condenser coil inlet of the same or an adjacent unit, raising the effective Entering Air Temperature (EAT) above the actual ambient dry-bulb temperature. Even a modest 3–5°F (1.7–2.8°C) EAT elevation can reduce chiller capacity by 3–6% and increase compressor power consumption proportionally, with cascading effects on overall plant efficiency.
Primary causes of recirculation include: insufficient unit-to-unit spacing; solid walls or parapets acting as deflectors that redirect exhaust air back toward inlets; adjacent buildings creating wind shadow zones that trap warm discharge plumes; and poor orientation relative to dominant wind directions. Still-air conditions (calm nights with no prevailing wind) represent worst-case recirculation scenarios.
| ⚠ Warning — High-Pressure Lockout Risk Sustained recirculation sufficient to raise condenser EAT by 8–12°F (4.4–6.7°C) above design conditions can drive compressor head pressure above the high-pressure cutout setpoint, causing automatic compressor shutdown. In a poorly sited data center cooling plant, this failure mode may cascade across multiple chillers simultaneously, resulting in complete loss of cooling capacity. |
Mitigation strategies include conducting 3D Computational Fluid Dynamics (CFD) modeling of the proposed chiller plant layout under multiple wind speed and direction scenarios before construction. CFD analysis quantifies EAT rise at each unit’s inlet under worst-case conditions and enables optimization of unit spacing, orientation, and discharge deflector design. Discharge air deflectors — hoods or directional vanes fitted to fan stacks — can redirect exhaust plumes vertically or away from adjacent units. Wind screens and baffles between units can interrupt recirculation pathways, though they must be designed not to restrict primary inlet airflow.
4. Derating Considerations
Air-cooled chillers are factory-rated at AHRI Standard 550/590 conditions: 85°F (29.4°C) entering condenser air temperature, 44°F (6.7°C) chilled water supply temperature (CHWST), and 54°F (12.2°C) chilled water return temperature (CHWRT). Real-world operating conditions at many data center sites deviate substantially from these reference points, necessitating deliberate capacity derating.
| Derating Factor | Typical Magnitude | Design Response |
| Elevated ambient temperature (above 85°F / 29.4°C) | ~0.5–1.0% capacity loss per 1°F (0.56°C) rise; power increases ~0.5–0.8%/°F | Use ASHRAE 0.4% design-day temp as rated condition basis |
| Altitude above sea level | ~1–2% capacity loss per 300 m (1,000 ft) | Apply altitude correction factor from OEM performance data |
| Fouled condenser coil | 3–8% capacity reduction (deferred maintenance) | Establish preventive maintenance schedule; include fouling margin |
| Hot air recirculation | 1–6% capacity reduction (site-dependent) | CFD modeling; include EAT rise in design ambient |
| End-of-refrigerant-charge degradation | 2–5% if not detected promptly | Continuous refrigerant monitoring; annual charge verification |
Table 2. Air-Cooled Chiller Derating Factors and Recommended Design Responses.
For mission-critical Tier III and Tier IV data centers, a minimum 10–15% capacity buffer above peak design load — applied after all derating factors — is recommended practice. This buffer accommodates simultaneous derating conditions (high ambient + partial coil fouling + EAT rise) that can coincide during the most demanding cooling events of the year. Facilities that are sized to exactly meet peak load under AHRI reference conditions, without derating margins, are statistically likely to experience capacity shortfalls during actual peak operating conditions.
5. Cooling Requirements and Load Profiling
Accurate IT load assessment is the foundation of chiller plant sizing. A common and consequential error is designing cooling capacity to nameplate IT equipment ratings, which overestimates actual heat rejection. In legacy enterprise data centers, actual server utilization typically ranges from 40–60% of nameplate power draw. Hyperscale and cloud-optimized facilities operating at higher consolidation ratios may approach 70–85% utilization — a distinction that materially affects cooling tonnage requirements.
The total thermal load imposed on the chilled water plant encompasses not only IT equipment heat rejection but also UPS system losses (typically 3–6% of IT load for double-conversion UPS), lighting heat gain, building envelope transmission (more significant for surface-area-intensive single-story designs), and power distribution infrastructure. A complete heat balance is required before chiller tonnage is established.
Cooling plants must be designed not only for current peak load but also for future IT capacity expansion. Phased expansion planning should influence chiller pad sizing, chilled water header sizing, and electrical infrastructure even if initial chiller installation is staged. Common redundancy configurations include:
- N+1: One standby chiller available across the total plant; adequate for Tier II–III facilities with acceptable planned maintenance windows.
- 2N: Full parallel redundant plant; each half serves 100% of load independently; standard for Tier IV mission-critical operations.
- N+2: Two standby chillers; applied where simultaneous maintenance and unplanned failure scenarios must both be accommodated without load reduction.
Power Usage Effectiveness (PUE) — defined as total facility power divided by IT equipment power per ASHRAE 90.4-2019 and the Green Grid standard — establishes the efficiency benchmark against which cooling plant selections are evaluated. A PUE of ≤ 1.4 is the ASHRAE threshold for “efficient” facilities; hyperscale operators routinely achieve PUE of ≤ 1.2. The chiller plant, including pumps and ancillary loads, typically accounts for 30–40% of total overhead power in air-cooled facilities, making compressor and fan efficiency central to achieving PUE targets.
Full-load chiller efficiency (kW/ton or COP) is insufficient as a selection criterion for data centers, where IT loads vary continuously. Integrated Part Load Value (IPLV), as defined in AHRI 550/590, weights chiller efficiency at 25%, 50%, 75%, and 100% load conditions and is the operationally representative efficiency metric. Chillers with strong IPLV performance — particularly those using Variable Speed Drive (VSD) compressors — deliver substantially better annual energy efficiency than their full-load ratings suggest.
6. Chilled Water Temperatures
The conventional chilled water standard — 44°F (6.7°C) supply / 54°F (12.2°C) return, representing a 10°F (5.6°C) delta-T (ΔT) — was established for comfort HVAC applications and does not reflect the thermodynamic potential available in modern data center environments. This conservative approach imposes unnecessary compressor lift and limits economizer hours in moderate climates.
Raising the Chilled Water Supply Temperature (CHWST) to 50°F (10°C), 55°F (12.8°C), or higher directly reduces the compressor pressure ratio required to reject heat to the condenser, improving Coefficient of Performance (COP) at approximately 2–4% per 1°F (0.56°C) increase in CHWST. A shift from 44°F to 50°F CHWST yields a 12–24% COP improvement — a substantial efficiency gain achievable purely through set-point optimization without capital expenditure on new equipment.
IT equipment compatibility with elevated CHWST depends on the thermal class of the servers being cooled via Computer Room Air Handlers (CRAHs). ASHRAE TC 9.9 classifies standard enterprise servers as Class A1 (allowable supply air 59–77°F / 15–25°C) and Class A2 (50–95°F / 10–35°C). Most modern CRAH units can deliver adequate supply air temperatures for ASHRAE A1/A2 class equipment using chilled water at 55–59°F (12.8–15°C), validating the case for higher CHWST set points.
| ▶ Engineering Note — Low ΔT Syndrome Low delta-T syndrome occurs when the actual chilled water ΔT across the system falls significantly below design (e.g., 4–5°F instead of 10°F), forcing excessive chilled water flow to transfer the required cooling load. This increases pump energy, reduces chiller efficiency, and can cause control instability. Root causes include CRAH valve hunting, bypassed coils, and oversized air handling equipment. Variable-flow primary–secondary or primary-only chilled water systems with variable frequency drives (VFDs) on pumps are essential to managing ΔT across variable load profiles. Minimum chiller evaporator flow rates must be maintained to avoid evaporator freeze or laminar flow heat transfer degradation. |
7. Sizing Strategies
The total cooling load calculation must aggregate all heat sources within the thermal envelope: IT equipment (server, storage, networking), UPS losses, power distribution unit (PDU) losses, lighting, and building envelope gains. IT load is most accurately quantified via branch circuit monitoring data for existing facilities, or through power density modeling (W/ft² or W/rack) combined with utilization factors for new construction. The industry-standard approach aligns with ASHRAE TC 9.9 Data Center Power and Cooling guidelines.
Chiller tonnage selection must balance three competing constraints: current peak load coverage with appropriate derating margins; near-term expansion capacity without over-investing in immediately idle plant; and long-term scalability through chiller pad and piping infrastructure sized for full buildout tonnage even if initial installation is staged.
Modular vs. large monolithic chillers present a meaningful design decision for data center applications:
| Criterion | Modular (Smaller) Units | Large Monolithic Units |
| Typical capacity range | 50–300 tons (176–1,055 kW) | 300–2,000+ tons (1,055–7,034 kW) |
| Part-load staging | Excellent — units staged on/off or at part load | Dependent on compressor unloading capability |
| Maintenance access | Simpler; individual unit taken offline with less impact | More complex; larger service zone required |
| Phased deployment | Well-suited; add units as IT load grows | Less flexible; full unit installed at once |
| Full-load efficiency | Moderate; may trail larger centrifugal units | Highest COP achievable at design load |
| Structural footprint | Distributed across larger area | Concentrated; requires robust structural support |
Table 3. Modular vs. Large Monolithic Chiller Comparison for Data Center Applications.
8. Compressor Types
Compressor selection is the primary driver of chiller efficiency, capacity range, part-load behavior, and lifecycle maintenance cost. Four compressor technologies are commercially deployed in data center ACC applications:
Scroll Compressors
Common in smaller chillers ranging from 20 to 150 tons (70–527 kW), scroll compressors are hermetically sealed, mechanically simple, and highly reliable. They are available in fixed-speed and variable-speed configurations. Their primary limitation is a restricted unloading range (typically minimum 25–50% of rated capacity), making them less efficient than screw or centrifugal compressors at very low partial loads. Multi-scroll circuit configurations within a single chiller frame improve staging capability and provide internal redundancy.
Screw Compressors
The dominant compressor technology in mid-to-large ACC applications, ranging from 100 to 500+ tons (352–1,759+ kW). Twin-rotor helical screw compressors offer excellent part-load performance via slide valve unloading, achieving smooth capacity modulation from 10–100% of rated output. Semi-hermetic construction permits field serviceability of motor and rotor assemblies. VSD-equipped screw compressors deliver class-leading IPLV performance and are widely specified for primary data center chiller plants where load variability is significant.
Centrifugal Compressors
Applied in large-capacity ACCs from 200 to 2,000+ tons (703–7,034+ kW), centrifugal compressors achieve the highest full-load efficiency of any compressor type, with COP values routinely exceeding 5.0–6.5 at AHRI standard conditions. They are velocity-based machines sensitive to changes in compression ratio (lift); reduced condenser EAT significantly improves performance, while elevated ambient conditions can cause surge at low loads. Magnetic bearing centrifugal compressors eliminate oil systems entirely, removing one of the most maintenance-intensive chiller subsystems, reducing friction losses, and simplifying oil-contamination management in the refrigerant circuit.
Variable Speed Drive (VSD) Compressors
VSD control — applied to screw or centrifugal compressors — is the single most impactful efficiency enhancement available in ACC selection for data center applications. By continuously matching compressor speed to instantaneous cooling demand, VSD compressors avoid the energy waste of hot-gas bypass or slide-valve unloading at low loads. Given that data center IT loads fluctuate continuously and many facilities operate at 40–70% of design capacity for the majority of annual hours, VSD compressors deliver disproportionately large energy savings relative to their incremental cost premium.
Multi-compressor circuits — two or more compressor stages operating on independent refrigerant circuits within a single chiller chassis — provide staged capacity control and internal redundancy. A dual-circuit 400-ton chiller can, for example, deliver approximately 200 tons on one circuit if the second experiences a fault, preventing a complete loss of cooling rather than requiring an N+1 standby unit to carry the full load.
9. Maintenance Requirements
ACC maintainability is a lifecycle performance factor that is frequently underweighted in initial equipment selection. Deferred maintenance directly translates to capacity derating, elevated energy consumption, and accelerated equipment failure — consequences that are operationally unacceptable in critical facilities.
- Condenser coil cleaning: Monthly visual inspection; quarterly pressure washing at minimum. Coastal and industrial sites with corrosive or particulate-laden air may require monthly coil washing and annual protective coating reapplication.
- Fan motor inspection: Vibration analysis per ISO 10816 standards semi-annually; direct-drive fan motor bearing lubrication or replacement per OEM schedule; belt tension and alignment checks for belt-drive fan configurations quarterly.
- Refrigerant management: Leak detection and charge verification per EPA Section 608 regulations, including annual leak rate reporting for systems with ≥ 50 lb (22.7 kg) charge. Continuous electronic leak detection sensors are recommended for critical ACC plants, particularly as A2L lower-GWP refrigerants are adopted.
- Compressor oil analysis: For screw and conventional-bearing centrifugal compressors, oil sampling and spectroscopic analysis semi-annually detects wear metals, moisture, and acid contamination. Oil-free magnetic bearing centrifugal compressors eliminate this task entirely.
- Filter-drier and sight glass inspection: Annually at minimum; replacement triggered by moisture indicator or following any refrigerant-side service event.
- Controls and safety device verification: Annual calibration check of high-pressure cutout, low-pressure cutout, motor overloads, and flow switch setpoints.
| ▲ Important — Service Aisle Requirement A minimum clear service aisle of 6 ft (1.8 m) must be maintained on the condenser coil access face of each chiller to permit coil pull for tube replacement and facilitate compressor changeout. Failure to preserve this clearance in final site layout — often consumed by piping or electrical routing — results in disproportionately expensive field workarounds during major maintenance events. |
For Tier III and Tier IV data centers, OEM-backed comprehensive maintenance agreements with contractually guaranteed response times (typically 4-hour emergency dispatch) are recommended. Service agreements should include access to OEM spare parts inventory, factory-trained technicians, remote monitoring integration, and annual performance certification testing against AHRI 550/590 rated conditions.
10. Operating Limitations
Every ACC design has defined environmental and operational boundaries that, if exceeded, trigger protective shutdowns or accelerate component degradation. Understanding these limits is essential for both design-day equipment sizing and for operational protocol development.
| Operating Parameter | Standard Range | Extended Range (with options) |
| Minimum ambient temperature | 32–40°F (0–4.4°C) | Down to -20°F (-29°C) with low-ambient kit (head pressure controls, fan VFDs, condenser fan cycling) |
| Maximum ambient temperature | 105–115°F (40.6–46.1°C) | Above this range: high-pressure cutout trips; evaporative pre-cooling can mitigate |
| Acoustic output | 70–85 dB(A) at 30 ft (9.1 m) | Acoustic enclosures, sound-attenuating fan stacks available; adds cost and static pressure |
| Voltage tolerance | ±10% of nominal supply voltage | Phase imbalance < 2% required for VSD compressors; voltage sags trigger protective relay |
Table 4. ACC Operating Envelope — Standard vs. Extended Configurations.
Refrigerant selection is a rapidly evolving regulatory domain. R-410A — the dominant ACC refrigerant for the past two decades — is being phased down under the U.S. AIM Act (American Innovation and Manufacturing Act), which mandates an 85% reduction in HFC production and consumption by 2036. Replacement refrigerants now entering the data center market include R-32 (GWP 675), R-454B (GWP 466), and R-513A (GWP 573). R-32 and R-454B are classified as A2L (mildly flammable) under ASHRAE 34, requiring enhanced refrigerant containment design, continuous leak detection systems, and compliance with NFPA 70 (NEC) and updated ASHRAE 15 safety code provisions for the occupied building. Engineers specifying new ACC equipment in 2026 and beyond should confirm refrigerant compliance with current and projected AIM Act phase-down schedules.
Evaporative pre-cooling of condenser entering air — through direct evaporative media pads or high-pressure misting systems upstream of condenser coils — is an effective hot-weather performance mitigation strategy for sites where peak ambient temperatures challenge standard ACC operating limits. This approach introduces water consumption, requiring water treatment to manage biological growth and mineral scale on coil surfaces, but can extend ACC useful operating range by 10–15°F (5.6–8.3°C) of effective ambient temperature reduction.
11. Additional Selection Factors
| Factor | Key Considerations |
| Free Cooling / Economizer Modes | Integrated plate heat exchanger economizers or dedicated free-cooling circuits allow heat rejection directly to ambient air when outdoor conditions are favorable (typically EAT < CHWST + 5–10°F / 2.8–5.6°C). Cooler climates (e.g., northern U.S., Nordics) can achieve 2,000–5,000+ economizer hours annually. ASHRAE 90.4 explicitly requires evaluation of economizer modes in data center energy design analysis. |
| BMS/BAS Controls Integration | Chillers must communicate via BACnet, Modbus, or LonWorks protocols with the facility Building Management System (BMS). Demand-based chilled water reset, remote diagnostics, predictive fault alerts, and automated load sequencing are essential capabilities for large ACC plants. |
| Refrigerant Containment | For A2L refrigerants, ASHRAE 15-2022 and local Authority Having Jurisdiction (AHJ) requirements mandate refrigerant detectors at low-point locations, automatic unit isolation on leak detection, and ventilation system interlocks for enclosed equipment rooms. |
| Footprint and Weight | Large ACCs (500+ tons) can weigh 30,000–80,000 lb (13,608–36,287 kg) and require structural engineering analysis for any elevated installation. Rigging access for crane placement must be incorporated into site planning, including temporary crane pad locations and overhead obstruction clearances. |
| Total Cost of Ownership (TCO) | First cost represents a minority of 15–20-year lifecycle cost. Energy — the dominant cost component — combined with maintenance labor, refrigerant, and component replacement typically constitutes 75–85% of TCO. Chillers with superior IPLV performance, lower refrigerant charge (smaller leak risk), and longer MTBF compressor assemblies consistently outperform lower-capital alternatives on a lifecycle basis. |
| Warranty and Service Availability | Standard manufacturer warranty is typically 1 year parts/labor, with compressor warranties of 3–5 years. Evaluate OEM local service coverage: proximity of factory-trained technicians and regional spare parts inventory are critical for Tier III/IV uptime commitments. International facilities must verify in-country service agreements prior to equipment selection. |
Table 5. Additional ACC Selection Factors for Green Data Center Facilities.
CONCLUSION
A Holistic, Site-Specific Engineering Imperative
Selecting air-cooled chillers for green data centers demands a holistic, site-specific engineering process that integrates environmental analysis, thermodynamic performance modeling, layout optimization, redundancy planning, and lifecycle cost evaluation into a single, coherent decision framework — no single factor can be addressed in isolation without introducing risk across the others. The upfront investment in ASHRAE design-day data acquisition, 3D CFD recirculation modeling, accurate IT load profiling, refrigerant regulatory compliance planning, and total cost of ownership analysis consistently produces cooling infrastructure that is simultaneously more energy-efficient, more resilient to climate and operational extremes, and more aligned with the sustainability commitments — including PUE targets, water conservation goals, and low-GWP refrigerant mandates — that define the green data center standard in 2026 and beyond.
Referenced Standards & Guidelines: ASHRAE Standard 90.4-2019 (Energy Standard for Data Centers) | AHRI Standard 550/590 (Performance Rating of Water-Chilling and Heat Pump Water-Heating Packages) | ASHRAE TC 9.9 (Thermal Guidelines for Data Processing Environments, 5th Ed.) | ASHRAE Standard 15-2022 (Safety Standard for Refrigeration Systems) | ASHRAE Standard 34 (Designation and Safety Classification of Refrigerants) | EPA Section 608 (Refrigerant Management Regulations) | U.S. AIM Act (2020) | ASCE 7-22 (Minimum Design Loads for Buildings and Other Structures) | ASHRAE Standard 62.1 (Ventilation for Acceptable Indoor Air Quality)
Green Data Center Industry Guide | Mechanical Systems & Thermal Management Series | Published June 30, 2026 | This article is intended for use by qualified mechanical engineers, licensed design professionals, and experienced facility managers. All design decisions must be validated against applicable codes, local Authority Having Jurisdiction (AHJ) requirements, and manufacturer documentation for specific equipment.