Acoustic & Emissions Engineering Design

Acoustic and emissions engineering is where the abstractions of the siting and permitting chapters become steel, catalyst, and concrete — and where a number written into a permit condition two years earlier either gets met or gets the project a stop-work order, a consent decree, or a midnight phone call from a county supervisor. The discipline is unusual in this guide because its design targets are not engineering optima — they are externally-imposed thresholds: the property-line dBA/dBC limit set by the noise ordinance (Chapter 3.11), and the ground-level NOx/CO/PM concentration and stack parameters set by the air permit (Chapter 3.9). The engineer's job is not to decide how quiet or how clean to be. It is to hit a number someone else chose, at the lowest capital and footprint cost, on equipment whose noise and emissions are a byproduct of the power and cooling decisions made everywhere else in Part 6.

That framing matters because the cheapest variable is almost always the one nobody wants to spend: land. Distance is the most reliable decibel in acoustics — sound spreads geometrically, and a doubling of distance from a point source buys roughly 6 dB with no hardware, no maintenance, and no acoustic residual that drifts out of tune. Every dollar you decline to spend on setback you spend twice on enclosures, attenuators, and the low-frequency mitigation that enclosures do not deliver. The same logic governs emissions: the cheapest NOx is the NOx you never combust (grid power, fuel cells), and after that, the SCR you sized into the generation island at design time rather than retrofitting under a consent decree. This chapter walks the four physical systems: acoustic enclosure and attenuation, emissions control, stack and dispersion, and vibration isolation.

Before any equipment is selected, the design basis must capture two externally-fixed numbers and one increasingly-important nuance. The first is the property-line sound limit, typically expressed as a day band and a stricter night band — the 2025–2026 norm is roughly 60 dBA day / 55 dBA night, and a growing number of jurisdictions have collapsed these into a single 24-hour limit (Prince William County, Virginia, extended its 55 dBA limit to 24 hours in February 2026). The second is the air-permit envelope: the allowable mass-emission rates and the ground-level concentration limits that a dispersion model must demonstrate compliance against. The nuance — and it is the one that wrecks projects designed to the old rules — is the migration from A-weighting to C-weighting and octave-band limits.

A-weighting deliberately discounts low-frequency energy because the human ear is less sensitive to it at moderate levels. But the dominant annoyance from a data center is precisely low-frequency: the 20–200 Hz tonal hum of large axial fans, dry-coolers, and transformers. That hum travels farther than mid-band noise (low frequencies attenuate less with distance and pass through barriers and walls that stop high frequencies), and it sits below the reliable detection floor of an A-weighted meter. A facility can pass a 55 dBA night limit at the fence and still generate a complaint stream, because the neighbors are not hearing dBA — they are hearing a 63 Hz tone that the A-filter subtracted out. The 2026 ordinance wave answers this with C-weighted (dBC) and 1/3-octave limits that the hum cannot hide from. If your jurisdiction has made that move, an A-weighted-only acoustic model is not conservative — it is wrong, and it will certify a design that fails the actual test.

The mechanical and electrical plant is a stack of point and line sources, each with a characteristic spectrum. Gensets and prime-power engines are broadband and loud — engine casing, exhaust, radiator/cooling air, and intake all radiate — and they run during the exact night-time test window when permits are tightest (a black-start or load test at 2 a.m. is the worst-case audit). Chillers and dry-coolers are the steady-state, always-on sources whose large low-speed fans produce the dominant low-frequency tone. Cooling towers add water-fall noise and fan noise. Air-cooled chillers and the fan walls of a liquid-to-air heat-rejection plant are the modern density driver: as the heat load per hall climbs with the rack ramp (Chapter 6.7), the installed fan power — and therefore the radiated acoustic power — climbs with it.

The mitigation toolkit is a hierarchy, ordered by cost-effectiveness and by how much of the low-frequency problem each tool actually solves. Setback is first and best: it buys broadband and low-frequency attenuation with no hardware. Source selection is next: low-noise / EC (electronically commutated) fans run more slowly and quietly for the same airflow, and selecting a fan whose blade-pass frequency avoids a resonance shifts the tone out of the worst annoyance band. Attenuators (parallel-baffle silencers on intake and discharge) and acoustic louvers deliver 5–15 dB but cost static pressure — which the fans must overcome with more power, partially re-radiated as noise, and which the cooling plant must budget for. Barriers and berms deliver 10–20 dB at the receptor for mid-and-high frequencies but diffract around their edges and are nearly transparent to low frequency. Full enclosures are the last and most expensive resort, and they create their own ventilation and heat-rejection problem.

The table is a sequencing rule. You spend land first, source-select second, and treat hardware mitigation as the expensive remainder — because each row below setback either costs static pressure (which the fans pay back partly as noise and the cooling plant pays as energy), creates a heat-rejection problem of its own (enclosures), or simply fails on the band that matters (barriers and low frequency). The most common and most expensive mistake is to skimp on setback at scoping time — to buy the smaller, cheaper parcel — and then discover at commissioning that no achievable combination of attenuators and barriers closes the night-time low-frequency gap. At that point the only remaining moves are operational curtailment (running the plant below capacity at night, i.e. stranding the power you came for) or buying adjacent land at a distressed-buyer premium. The acoustic budget is set on the site plan, years before the first measurement.

The emissions problem arrived with the same force as the noise problem and from the same root cause: on-site generation. When the grid could not energize the load on an AI timeline, operators brought their own power — reciprocating engines and aeroderivative/industrial gas turbines — and in doing so converted a passive electrical load into a stationary combustion source with a Clean Air Act footprint (the permitting side lives in Chapter 3.9; the fuel-process and gas-handling engineering in Chapter 4.9). This chapter owns the engineering response: the post-combustion catalyst train that takes raw exhaust and brings it under the permit limits.

The two workhorses are Selective Catalytic Reduction (SCR) for NOx and an oxidation catalyst for CO and unburned hydrocarbons (VOC). SCR injects a reductant — aqueous urea or anhydrous/aqueous ammonia — upstream of a catalyst bed, where NOx is reduced to nitrogen and water vapor. Modern SCR on data-center turbines is engineered to very low outlet levels: the regulatory bar for new high-utilization natural-gas turbines with combustion controls plus SCR is on the order of 5 ppm NOx, and 2026-vintage platforms market sub-2 ppm performance. The oxidation catalyst sits in the same housing and handles the CO that lean combustion and SCR can otherwise leave high. The engineering is not free: SCR adds backpressure (a turbine derate), demands a reductant storage-and-dosing system (an ammonia inventory with its own process-safety and setback implications), requires the exhaust to sit in a catalyst-favorable temperature window, and produces a small ammonia slip that is itself a permit-limited pollutant. Get the temperature window or the dosing control wrong and you trade a NOx exceedance for an ammonia-slip exceedance.

The air permit is not granted against mass-emission rates alone — it is granted against ground-level concentrations at and beyond the property line, demonstrated by a dispersion model (AERMOD in the US, or the regional equivalent). That model is parameterized by physical stack attributes the engineer controls: stack height, inside diameter, exit temperature, and exit velocity. Plume rise — how high the hot, fast exhaust climbs before it dilutes — is a strong function of buoyancy (temperature) and momentum (velocity). A stack that is too short, too wide (low velocity), or too cool gives up plume rise, and the modeled ground-level concentration at the fence rises accordingly. The subtle detail is the building-downwash effect: a stack that does not clear the aerodynamic wake of the data hall itself (the 'good engineering practice' stack-height rule) sees its plume dragged down into the building's recirculation cavity, spiking concentrations exactly where the receptors are.

The coupling here is tight: the stack you build must match the stack the model assumed. If the permit was modeled on a 30 m stack at 25 m/s exit velocity and value engineering later shortens it or the as-built turbine derate drops the exhaust temperature, the as-built source no longer disperses the way the permit certified — and a fully compliant mass-emission rate can produce a modeled (and real) exceedance at the boundary. Treat the dispersion model's stack parameters as permit conditions, not design suggestions: changing them is a permit modification, not a field decision. This is also where the rain-cap fight lives — a horizontal or capped discharge that protects the catalyst from rain destroys vertical exit velocity and plume rise; the model nearly always demands a vertical, uncapped (or louvered-cap) discharge, and the mechanical design must accommodate it.

Not all of the low-frequency problem travels through the air. The rotating and reciprocating plant — chillers, pumps, dry-coolers, engines — feeds vibration into the slab, the structure re-radiates it as low-frequency sound (and transmits it as perceptible vibration into adjacent occupied space and, occasionally, into neighboring buildings), and no exterior barrier or enclosure touches this structure-borne path because it never went through the air. This is the mechanism behind the most stubborn hum complaints: the energy couples into the building at the equipment feet and re-emerges as a tone everywhere the structure connects to ground.

The fix is isolation at the source, designed in at equipment set, not retrofitted: spring isolators or inertia bases under rotating equipment, tuned so the isolator's natural frequency sits well below the equipment's forcing frequency (the lower the isolation frequency relative to the disturbance, the more energy is rejected); flexible connectors on every pipe and duct penetration so the charged liquid-cooling loop (Chapter 5.7) does not become a rigid bridge that telegraphs pump vibration across the building; and seismic-rated isolation that survives the anchoring requirements of Chapter 6.7 without short-circuiting the isolation with a rigid restraint. The consequence of skipping it is the worst kind: a complaint that no acoustic barrier, attenuator, or setback can cure, because the path is through the ground — discovered only after the plant is running and the easy interventions are already exhausted.

Acoustic and emissions engineering is a backward-solved problem: the targets are fixed externally (the ordinance in Chapter 3.11, the air permit in Chapter 3.9), and the engineer's degrees of freedom are the placement, hardware, and source-selection choices that meet them at the lowest footprint and capex. Three rules fall out. Spend land before hardware — setback is the cheapest dB and the only one that works on low frequency at scale, so the site plan, not the silencer catalog, is where the acoustic battle is won. Design the controls in, not on — SCR, oxidation catalyst, and vibration isolation are cheap as integrated line items and punishing as retrofits, and the 'temporary' and 'emergency' classifications that let you defer them are narrowing under the 2026 NSPS. And match the as-built to the model — the dispersion model's stack parameters and the acoustic model's octave-band assumptions are permit conditions, so value-engineering a stack shorter or skipping the dBC analysis is not a cost saving, it is a re-permit waiting to happen.