← Survey overview · Per-Facility Analysis · Updated May 13, 2026

HED — NTT VA3 (Ashburn, VA)

Per-facility exterior plume analysis. Public-source inputs. Screening-level.

What this page covers

This page presents a screening-level exterior CFD analysis of NTT VA3, a 16 MW, six-vault data center in NTT Global Data Centers' 78-acre Ashburn campus, publicly attributed to HED for full architecture and MEP design. The analysis examines exterior thermal-plume behavior — rooftop heat-rejection discharge, ground-level generator stack and radiator exhaust during emergency operation, and the resulting plume trajectories, recirculation patterns, and intake exposure — under a directional sweep of wind cases at ASHRAE 0.4% summer design conditions.

Facility context

The facts below are derived entirely from cited public sources: the NTT-published facility data sheet, NTT's March 2018 campus announcement, DataCenterMap.com facility records, and HED's published mission-critical project attribution. No proprietary drawings, specifications, or operational data are used.

Address.

44245 Gigabit Plaza, Ashburn, VA 20147 (NTT-published mailing).

Owner-operator.

NTT Global Data Centers Americas (formerly RagingWire).

Critical IT capacity.

16 MW across six vaults — V1/V2/V4/V5 at 2 MW each, V3/V6 at 4 MW each.

Data floor.

~112,000 sq.ft.; two-story building; opened March 2018.

Cooling.

Waterless pumped-refrigerant cooling, N+2 redundancy. Rooftop condenser arrays. No cooling towers, no chilled-water loop.

Backup power.

Approximately 18 diesel generators in two pods (N+2 redundancy, "9 to make 8" topology per the NTT data sheet). Ground-level generator yard along the south building face.

Electrical service.

34.5 kV distribution, two redundant Dominion Energy feeds.

Source attribution.

HED — full A+E+MEP (BD+C Top 30 Data Center Architecture Firms 2024 listing; HED firm-published project attribution).

Wind cases modeled

Case 1 — Prevailing summer.

S/SSW (~210°), 8 mph (Dulles IAD wind rose mean summer). Pushes south-yard genset plume toward the building's south face and rooftop equipment.

Case 2 — Low wind / buoyancy-dominated.

~2 mph, calm. Tests rooftop condenser self-recirculation and yard-internal genset radiator entrainment.

Case 3 — Prevailing winter.

NW (~315°), 10 mph. Reverses dominant transport direction.

Cases 4–8 — Directional sweep.

N, NE, E, SE, W at design wind speed to identify worst-case rooftop intake exposure.

Observed plume behavior (summary)

The exterior CFD identified four phenomena to evaluate against the as-built facility. The qualitative observations below are derived from the public-source model only; quantitative results — temperature distributions, re-ingestion magnitudes, recirculation values — are not published here.

1. South-yard genset plume entrainment under prevailing summer wind. With the generator yard located along the south building face and prevailing summer wind from S/SSW, vertical exhaust columns from the genset pods loft to roof level and trend toward rooftop heat-rejection equipment intakes along the south roof edge. Worth verifying against the as-built generator-yard layout and rooftop equipment positions.
2. Rooftop self-recirculation under low wind. The waterless pumped-refrigerant architecture produces a dry, sensible-only rooftop plume. Under calm conditions, individual condenser bank discharge can re-enter neighboring banks if spacing is below the minimum dispersal distance for the discharge velocity. Bank-to-bank spacing on the as-built roof determines whether this is operational concern.
3. Cross-pod radiator plume interaction within the yard. The two 9-unit generator pods produce paired horizontal radiator discharges that interact spatially. The yard layout (single linear row vs. parallel rows) determines whether one pod's radiator discharge enters another pod's intake.
4. Campus-scale wake effects from adjacent NTT VA-series buildings. VA3 sits within a 7-building campus. Upwind buildings (VA1, VA2, VA4, VA6, VA7 depending on wind direction) generate wake structures that change the effective wind speed and turbulence intensity arriving at VA3. Worth confirming against the as-built campus layout.

Methodology

The methodology applied to this facility is the same standardized exterior CFD approach applied to every facility in the cohort — cylindrical far-field domain, logarithmic atmospheric boundary layer inlet, polyhedral mesh in Siemens STAR-CCM+, realizable k-ε RANS baseline. Full domain setup, boundary conditions, solver choices, and stated limitations are documented at the Methodology page. Key terms used here are defined at the Key Terms and FAQ page.