The NO₂ and CO limits that an exhaust study is judged against come from the U.S. Environmental Protection Agency, not from ASHRAE. They are the National Ambient Air Quality Standards (NAAQS). A CFD study gives you a concentration at an intake for one chosen wind and load case. That is a conservative screen against those numbers, not a demonstration of regulatory compliance. This article explains what the numbers are, how to read a study against them, and why the near field needs CFD at all.
The limits
EPA’s NAAQS table lists the primary standards for the two gases that generator and boiler exhaust studies report. Each is written in the Code of Federal Regulations, in 40 CFR 50.8 for CO and 40 CFR 50.11 for NO₂, where the rule text sets the level and the averaging time.
| Pollutant | Averaging time | Level | Form |
|---|---|---|---|
| Nitrogen dioxide (NO₂) | 1 hour | 100 ppb | 98th percentile of each year’s daily maximum 1-hour concentration, averaged over 3 years |
| Nitrogen dioxide (NO₂) | 1 year | 53 ppb | Annual mean |
| Carbon monoxide (CO) | 1 hour | 35 ppm | Not to be exceeded more than once per year |
| Carbon monoxide (CO) | 8 hours | 9 ppm | Not to be exceeded more than once per year |
The form matters as much as the level. The NO₂ standard is a statistic over three years of measurements: pick the worst hour of each day, take the 98th percentile of those daily maxima for each year, and average the three years. A few bad hours a year do not break it. The CO standards allow one exceedance a year.
What ASHRAE 62.1 says
ASHRAE Standard 62.1, the ventilation standard, does not set these numbers. Its Section 4 on outdoor air quality asks the designer to do two things before finishing the ventilation design: find out whether the site’s region is in attainment or nonattainment with the NAAQS (4.1), and survey the building site for local contaminants such as nearby exhaust sources (4.2). The NAAQS values are reproduced in an informative appendix of the standard, for reference. That is the wording of the 2016 addendum to the 2016 edition. Section numbers and appendix letters can move between editions, so check the one your project uses. When a study quotes 100 ppb or 35 ppm, the source of the number is EPA.
Reading a study against a limit
EPA writes the standards for outdoor air as measured by monitors over years. Designers borrow the same levels as a target for the air an intake draws, because they are the best-known public benchmark for what is acceptable to breathe. That is a design practice, not a legal test of the intake.
A CFD dispersion study works differently from a monitor. It solves the flow for one wind direction, one wind speed and one set of running equipment, and reports the concentration it settles to at each intake. For that case the answer is a steady number. So a study can tell you two things:
- If the cases include the worst winds you can find and the concentration is well below the limit, the intake has room to spare, because the other hours are no worse.
- If a wind case exceeds the limit, the study has not shown a violation. It has shown a wind and load that would need to happen often enough, and for long enough, to matter. How often is a separate question about the local wind record and how many hours the equipment runs. A generator that runs only short test cycles is a different problem from a boiler that runs all winter.
That is why we call a study result a screen. It is conservative when the cases are chosen to be the bad ones. It does not reproduce a three-year, 98th-percentile statistic, and it does not claim to.
Why the near field needs CFD
EPA’s regulatory plume model is AERMOD, with a building treatment called PRIME. It handles a building as a single idealized box, and EPA documents its limits in its own downwash white paper:
- AERMOD “can only model the equivalent of a single building or tier”, so a site with several buildings at different heights is reduced to one.
- The downwash algorithms are “based on solid, square and rectangular ground-based buildings”.
- EPA says analyses show AERMOD to “both overpredict and underpredict ground-level concentrations in the building wake”, depending on building dimensions, stack height and location, and wind direction.
An air intake on the roof of the same building that the stack serves sits inside that wake and recirculation zone. That is the case where the idealization matters most. CFD solves the flow around the actual buildings, so the gaps between them, the roof-level equipment and the recirculation behind a parapet are in the model. For a receptor a kilometer away across open ground, a plume model is a reasonable and much cheaper tool. For an intake 30 feet from the stack, the building is the whole story.
Pathlines are not concentrations
A pathline shows where a parcel of exhaust goes. If a pathline from a stack reaches an intake, exhaust gets there. It does not say how much.
The number depends on dilution: what fraction of the air entering the intake came from the stack. Here is arithmetic with stated inputs, not a result. Say the exhaust gas contains 500 ppm of NOx (an assumption for illustration, not a rated figure). If 1 part in 10,000 of an intake’s air is exhaust, the intake sees ppm, which is 50 ppb. If it is 1 part in 5,000, it sees 100 ppb, which is the NO₂ limit. If it is 1 part in 1,000,000, it sees 0.5 ppb. The same pathline connects the stack to the intake in all three cases, and only the last one is harmless.
So a study that shows pathlines only can say that exhaust does or does not reach an intake. It cannot say whether the concentration is under 100 ppb. A study that carries the exhaust as a tracer through the flow can report the dilution at each intake, and then multiply it by the pollutant concentration. For a concentration that is a small fraction of the intake’s air, the pollutant at the intake is proportional to the emission rate, so changing the emission (aftertreatment on or off, a different generator rating) does not need a new flow solve. Treating all of the NOx as NO₂ is the usual screening assumption, and it leans toward overstating the NO₂ at the intake. Whatever is already in the outdoor air is added on top, and EPA publishes monitor data for that.
What moves the answer
Four inputs change where the plume goes, and a study varies the ones that matter on the site.
- Stack height. A stack above the roof line releases into faster, cleaner air than one below it. The pair of diagrams below shows the same generator, wind and building with two stack heights.
- Exit velocity and exhaust temperature. A fast, hot jet rises before the wind bends it over. Cool exhaust at a low speed does not, and it falls into the building’s wake.
- Wind direction. A building has a few directions in which a given intake is downwind of a given source. Those are the ones to find.
- Wind speed. More wind dilutes more, but low wind lets a plume linger and removes the push that carries it past the roof, so the worst case is not always the strongest wind. It has to be tested.
A — short stack
Released below the roof line, the plume is carried onto the roof and into the air-handler intake.
B — tall stack
Released above the roof, the plume passes over the air handler.
What a dispersion study delivers
Our engineers have delivered 14 exhaust dispersion studies since 2023, most of them for hospitals. Each reports a table of concentrations at every intake and louver, for each wind case, in parts per billion for NO₂ and parts per million for CO. Each value sits beside its limit, with EPA named as the source. Maps and pathlines show where each plume goes and why. What-if cases test a taller stack, a moved intake, aftertreatment, a parapet or a wind screen.
If you are placing a generator, boiler or stack near an intake, the exhaust dispersion studies page describes what we need from you and how a study runs.