Figure 1a (left): Plume impact on taller downwind building. Figure 1b (right): Plume impact on taller upwind building.

Specifying Exhaust and

Intake Systems
By Ronald L. Petersen, Ph.D., Member ASHRAE, Brad C. Cochran, Member ASHRAE,
and John J. Carter, Member ASHRAE

T
he design of exhaust stacks and air intakes needs careful consider- is being designed is shorter than sur-
ation. Public concern has increased regarding air pollution in gen- rounding buildings, it is difficult to de-
sign a stack so the exhaust will not
eral. In addition, adverse exposure to air pollutants in the workplace impact neighboring buildings.
can affect employee health and productivity. In some cases, releases The effect of a taller downwind or up-
of toxic pollutants may lead to litigation. The following newspaper ar- wind building is illustrated in Figure 1.
The figure shows how the plume hits the
ticle excerpts illustrate some of these issues. face of the taller building when it is down-
wind and how, when it is upwind, the
Business Weekly (May 2, 1988): “Lo- 1998): “Suspicions confirmed. Public wake cavity region of the taller building
cal residents were frightened. New phar- health off icials say brain tumors at traps the exhaust from the shorter build-
macology laboratories at the University Amoco center more than coincidence.... ing. In either case, the plume impacts the
of California at San Francisco were in- A study of Building 503 at the Amoco face of the taller building.
vestigating everything from AIDS to Research Center in Naperville indicates Figure 2 further illustrates problems
parasitic diseases. Could disease organ- a rash of malignant brain cancers…. that can be created by poor stack design.
isms or toxic chemicals from those labs Eighteen Amoco Research Center em- Fumes from the exhaust may reenter the
escape and harm citizens?” ployees have developed brain tumors in building, enter adjacent buildings, or
San Francisco Chronicle (September the last 28 years.”
5, 1996): “A barrage of letters and con- Some challenges to specifying a good About the Authors
cerns about toxic chemicals have forced stack design include the existing build- Ronald L. Petersen, Ph.D., is vice president of
a circuit board manufacturer to drop, at ing environment, aesthetics, building Cermak Peterka Petersen Inc., Fort Collins, Colo.
Brad C. Cochran is a senior project engineer for
least temporarily, plans to move next door design issues, chemical use, source types,
Cermak Peterka Petersen Inc.
to a peninsula high school.” local meteorology and topography. For John J. Carter is a project engineer for Cermak
Chicago Daily Herald (April 17, example, if a new laboratory building that Peterka Petersen Inc.

Reprinted by Permission
 Stack Design

impact pedestrians at unacceptable concentration levels. To

avoid adverse air quality, taller stacks, higher volume flows
and/or optimum locations on the roof may be necessary.
In most cases, laboratory stack design is a balance between
various constraints and obtaining adequate air quality at sur-
rounding sensitive locations (air intakes, plazas, operable win-
dows, etc.). The lowest possible stack height is desired for
aesthetics, while exit momentum (exit velocity and volume flow
rate) is limited by capital and energy costs, noise, and vibration.
To determine the optimal exhaust design, predictions of ex-
pected concentrations of exhausted pollutants at sensitive loca- Figure 2: Illustration of potential air quality problems due

tions are needed to compare against health limits and odor to laboratory emissions.

thresholds. These predictions can be accomplished with vary-

ing degrees of accuracy using three different methods: 1) a full- 4. Locate the air intake at the base of a relatively tall stack or
scale field program; 2) a mathematical modeling study; or 3) a tight cluster of stacks, if this location is not adversely affected by
reduced-scale study conducted within an atmospheric-bound- exhaust from nearby buildings. Intakes should not be located
ary-layer wind tunnel. A full-scale field program may provide near the base of highly toxic stacks due to potential fan leakage.2
the most accurate prediction of concentration levels but can be 5. Avoid locating intakes near vehicle loading zones. Cano-
expensive and time consuming. In addition, it is impossible to pies over loading docks do not prevent hot vehicle exhaust
evaluate designs before construction is completed. from rising up to intakes above the canopy.2
Numerical models can be divided into two categories, ana- 6. Use High Efficiency Particulate Air (HEPA) filters or Ul-
lytical models and computational fluid dynamics (CFD) mod- tra Violet Germicidal Irradiation (UVGI) systems of similar ef-
els. Analytical models assume a simplif ied building ficiency in isolation room exhaust streams.3
configuration and provide concentration estimates based on 7. Combine several exhaust streams internally to dilute in-
assumed concentration distributions, i.e., Gaussian. These termittent bursts of contamination from a single source, as
models do not consider site-specific geometries that may sub- well as producing an exhaust with greater plume rise. Addi-
stantially alter plume behavior. tional air volume may also be added to the exhaust at the fan
CFD models attempt to resolve the plume transport by solv- to achieve the same end.2
ing the Navier-Stokes equations at finite grid locations. Wind- 8. Group separate stacks together (where separate exhaust
tunnel modeling, on the other hand, is much like conducting a systems are mandated) in a tight cluster to take advantage of
field experiment where the concentrations are measured in a the increased plume rise from the resulting combined jet.2 Note
simulated flow at the points of interest over a scale model of that all the exhausts must operate continuously to take full
the buildings under evaluation. advantage of the combined jet.
This article describes a quantitative approach to accurately 9. Avoid rain caps or other devices that limit plume rise on
evaluate exhaust and intake designs to ensure acceptable air exhaust stacks. Conical rain caps often do not exclude rain,
quality inside and around buildings. Also described for back- because rain does not fall straight down. Alternate design op-
ground purposes are various exhaust and intake design issues tions are provided in Chapter 43 of the ASHRAE Handbook—
such as applicable standards and recommendations, analyti- HVAC Applications.2
cal methods, plume rise, architectural screens, and entrained 10. Consider the adverse effect of architectural screens. A
air exhaust stacks. solid screen effectively decreases the stack height by 80%.4

Exhaust/Intake Design Issues Analytical Methods

Applicable Standards and Recommendations Chapter 43 of the ASHRAE Handbook—HVAC Applications

Several organizations have published standards or recom- discusses exhaust stack design in some detail.2 The chapter
mendations regarding laboratory exhaust stack design as sum- contains two primary types of information regarding stack de-
marized here. sign: 1) a geometric method of determining stack height; and
1. Maintain a minimum stack height of 10 ft (3 m) to protect 2) mathematical equations for predicting rooftop concentra-
rooftop workers.1 tions. In the geometric method, the recommended stack height
2. Locate intakes away from sources of outdoor contamina- is that for which the bottom edge of the exhaust plume will be
tion such as mobile traffic, kitchen exhaust, streets, cooling above various recirculating and high turbulence zones. In gen-
towers, emergency generators and plumbing vents.2 eral, this method is entirely inadequate for exhaust streams
3. Do not locate air intakes within the same architectural that contain toxic or odorous material, as it does not provide
screen enclosure as contaminated exhaust outlets.2 an estimated concentration at an air intake or other sensitive
Figure 3: Minimum recommended stack height above roof-

top air intake using A SHRAE methods.

location. Hence, no information on the adequacy of the stack Figure 4: Plume centerline height for conventional and en-

trained air exhaust systems.

to avoid concentrations in excess of health or odor limits is
provided. The analytical equations tend to be conservative for
an isolated building or one that is significantly taller than the ASHRAE provides a listing of 1% wind speeds for various
surrounding buildings and for air intakes on the roof level. metropolitan areas around the world.7
Also, they are not appropriate for complex building shapes or For a given stack design and receptor location, there is a
when buildings of similar or taller height are nearby. “critical wind speed” causing the maximum concentration.
Using the ASHRAE dispersion equations and a 400 mg/m3 per Wind speeds lower than this critical speed result in greater
g/s ASHRAE design criterion, a graph can be generated giving plume rise; higher wind speeds provide more dilution due to
the minimum recommended stack height to ensure that the de- the greater volume of air passing the exhaust stack. The criti-
sign criterion is met, as shown in Figure 3.2,5 For example, as- cal wind speed increases with exit velocity, exhaust volume
sume an air intake is located 60 ft (18.3 m) away from an exhaust flow rate and stack height.
stack. The figure shows that a 20 ft (6.1 m) stack is needed if the
volume flow rate is 1,000 cfm (0.47 m3/s), and a 10 ft stack (3 m) Architectural Screens

is needed with a 25,000 cfm (11.8 m3/s) volume flow rate. The Architects or building owners often want to hide their ex-
figure clearly shows the benefit of higher volume flow rates. haust stacks using screening material. An ASHRAE funded
research study was conducted to evaluate the effect of archi-
Plume Rise and Dispersion tectural screens on rooftop concentration levels.4 The study
Adequate plume rise is important to ensure that the exhaust found that screens can significantly increase concentrations
escapes the high turbulence and recirculation zones on the on the roof and, in effect, reduce the effective stack height. The
building roof. Plume rise increases with increased exit mo- study evaluated various enclosure sizes and heights but found
mentum and decreases with increased wind speed.6 Reducing that the main parameter affecting rooftop dispersion was the
the diameter to increase exit velocity will enhance plume rise. screen porosity. The results of the study provide a quantitative
However, a high exit velocity in itself does not guarantee ad- relationship between screen porosity and stack height.
equate plume rise since the volume flow rate, and thus momen-
tum, are factors as well. Plume rise is also degraded by increased Entrained Air Exhausts

atmospheric turbulence since the vertical momentum of the Entrained air exhaust manufacturers often quote an effective
exhaust jet is more quickly diluted. stack height for their system, which many designers consider
If the ratio of exit velocity to approach wind speed is too when choosing the appropriate system. The effective stack height
low, the plume can be pulled downwards into the wake of the specification is based on a mathematical equation that predicts
stack structure creating negative plume rise, a condition re- the height of the centerline of the emitted exhaust stream versus
ferred to as stack-tip-downwash. This downwash defeats some downwind distance.6 The effective stack height that is often
of the effect of a taller stack and can lead to high concentra- presented is, in reality, the maximum height of the exhaust plume
tions at the building surface. A rule of thumb for avoiding centerline at some large distance (say, 100 to 200 ft [30 to 61 m])
stack-tip-downwash is to have the exit velocity be at least 1.5 downwind of the stack and is not an effective stack height. What
times the wind speed at the top of the stack.2 The wind speed the manufacturers should supply as a specification is the “effec-
exceeded 1% of the time is commonly used for estimating the tive stack height improvement” over a conventional exhaust
minimum exit velocity required to avoid stack-tip-downwash. system. The stated improvement may not be as great as might be
 Stack Design

expected, as shown in the following analysis.

Figure 4 shows the predicted plume centerline height (called
effective stack height by some entrained air system suppliers)
versus distance from the stack for a conventional exhaust sys-
tem with a 15,000 cfm (7.1 m3/s) volume flow rate and a 3,000
fpm (15.2 m/s) exhaust velocity, and a typical entrained air
system with a 25,000 cfm (11.8 m3/s) total volume flow rate and
a 4,000 fpm (20.3 m/s) exit velocity. Plume centerline heights
were calculated for 10 mph (4.5 m/s) and 20 mph (8.9 m/s)
stack height wind speeds. The figure shows that the increase in
plume centerline height (effective stack height improvement)
for the entrained air system versus the conventional exhaust
Figure 5: Concentration distribution along building roof and
system is only 1 to 2 ft (0.3 m to 0.61 m) near the stack; and
sidewall.
increases to 7 to 10 ft (2.1 to 3 m) at 100 ft (30.4 m) downwind.
This analysis shows why the effective stack height specifica- toxic chemical with a low emission rate may be of less concern
tion is misleading. The manufacturers should be encouraged to than a less toxic chemical emitted at a very high rate.
delete this specification and add the specification of the “ef- In practice, a chemical inventory for each exhaust type is
fective stack height improvement” over a conventional system. examined to determine the appropriate values of (C/m)health
and (C/m)odor for any released chemicals. Dispersion modeling
Recommended Analysis Approach is performed to determine (C/m)max for all stack designs stud-
The Basic Approach ied. Those designs that yield concentrations lower than the
The recommended approach to evaluating the air quality design goal, i.e., (C/m)max < (C/m)goal, are the recommended
aspects of exhaust stacks is to perform dispersion modeling to exhaust stack designs.
demonstrate that expected concentrations do not exceed health
limits or odor thresholds. The design recommendations and Formulating a Concentration Design Goal

standards discussed earlier can be helpful in the design pro- Three types of information are needed to develop normal-
cess, but they do not guarantee adequate air quality. ized health limits and odor thresholds: 1) a listing of the toxic
or odorous substances that may be emitted, 2) health limits
The air quality acceptability question can be written: and odor thresholds for each emitted substance, and 3) the
Cmax < Chealth ? (1) maximum potential emission rate for each substance.
and Substances Emitted. A list of toxic and odorous chemicals
Cmax < Codor ? (2) is usually obtained from the building owners. The list may be
where Cmax is the maximum concentration expected at a sen- a chemical inventory or a list prepared to meet environmental
sitive location (air intakes, operable windows, pedestrian ar- regulations. Storage amounts are useful for obtaining an up-
eas), Chealth is the health limit concentration and Codor is the per-bound estimate of the largest amount released.
odor threshold concentration of any emitted chemical. Health Limits. Recommended health limits, Chealth, are based
When a large number of potential chemicals are emitted on the ANSI/AIHA Standard Z9.5 for Laboratory Ventilation,
from a building, a variety of mass emission rates, health limits which specifies air intake concentrations no higher than 20%
and odor thresholds are examined. It then becomes operation- of acceptable indoor concentrations.8 Acceptable indoor con-
ally simpler to recast the acceptability question by normaliz- centrations are taken to be the minimum short-term exposure
ing (dividing) Equations 1 and 2 by the mass emission rate, m: limits (STEL) from the American Conference of Governmental
C  C  Industrial Hygienists (ACGIH), the Occupational Safety and
  <  ? (3) Health Administration (OSHA), and the National Institute of
 m  max  m  health Occupational Safety and Health (NIOSH), as listed in ACGIH.9,10
and Odor Thresholds. ACGIH provides a good source for odor
C  C  thresholds, Codor.11 ACGIH critically reviews previous experi-
  <   ? (4)
 m  max  m  odor mental data and lists geometric means of accepted data. For
chemicals not listed in ACGIH, geometric means of high and
The left side of each equation, (C/m)max, is only dependent low values provided in Ruth, 1986 are recommended.11,12
on external factors such as stack design, receptor location, and Emission Rates. For laboratories, the emission rates are typi-
atmospheric conditions. The right side of each equation is cally based on small-scale accidental releases, either liquid spills
related to the emissions and is defined as the ratio of the health or emptying of a lecture bottle of compressed gas. The actual
limit or odor threshold to the emission rate. Therefore, a highly emission rates from experimental procedures are difficult to quan-
tify, especially at large laboratories with diverse research. Small on the building roof and sidewall for each condition. One test
accidental releases have two advantages: 1) they can be consid- was run with a 10 ft (3 m) solid screen positioned around an 11
ered to be the upper limit of the largely unknown release rates ft (3.4 m) stack. All other tests had an unobstructed roof.
occurring in laboratories; and 2) they can be quantified. Evapo- Figure 5 shows the normalized concentrations (C/m) on the
ration from liquid spills is computed from EPA equations based building roof and sidewall for the various configurations. It
on a worst-case spill within a fume hood.13 Typically, the worst- should be noted that string distances between 0 ft and 50 ft
case spill is defined as the complete evacuation of a 1 L (0.26 (15.2 m) are on the building roof, and string distances between
gal) beaker over a 1 m2 (11 ft2) area. Appropriate adjustments to 50 and 100 ft (15.2 and 30.4 m) are on the sidewall. The figure
the worst-case spill volume can be made to account for maxi- shows the expected trend that as stack height increases, the con-
mum storage quantities less than 1 L (0.26 gal) and for fume centrations on the roof decrease with the point of maximum C/m
hood counter surface areas that are less than or greater than 1 m2 moving farther away from the stack location. Concentrations on
(11 ft2). the building sidewall are much lower than those on the roof for
Compressed gas leaks typically assume the emptying of a each configuration evaluated, which shows the advantage of
fractured lecture bottle in one minute. For other sources, such as locating air intakes on building sidewalls versus the roof.
emergency generators, boilers, and vehicles, chemical emissions The results in Figure 5 can be used to assess the adequacy of
rates are often available from the manufacturer. the ASHRAE mathematical method illustrated in Figure 3. Us-
Concentration Design Goal Selection. Once the informa- ing Figure 3, a 13 ft (4 m) stack is recommended when the air
tion on chemical usage, health limits, odor thresholds, and intake is 50 ft (15.2 m) from a 5,000 cfm (2.4 m3/s) exhaust with
emission rates is gathered, the normalized health and odor 3,000 fpm (15.2 m/s) exit velocity. Figure 5 shows that a 10 ft (3
limits, (C/m)health and (C/m)odor, are computed. The concentra- m) stack would be adequate to meet the 400 mg/m3 per g/s
tion design goal, (C/m)goal, for the stack/receptor design will ASHRAE design criterion with a 2,000 fpm (10 m/s) exit veloc-
ideally be the minimum value of these limits for all of the ity. If the air intake is on the sidewall, there is presently no
chemicals. As new processes and chemicals are used in the reliable mathematical method to account for the concentration
laboratory, the quantities (C/m)health and (C/m)odor can be evalu- decrease. Thus, a conservative approach would have to be taken,
ated for each chemical added to the inventory. If these values i.e., use Figure 3, which would result in a 13 ft (4 m) stack.
are less than (C/m)goal, air quality problems may arise, and us- Figure 5 shows that a 1 ft (0.3 m) stack would meet the ASHRAE
age or emission controls may be warranted. criterion if the air intake were on the building sidewall.
Often for a facility with intensive chemical usage, the mini- Figure 5 also shows that with a 10 ft (3 m) solid screen
mum normalized concentration value is below the minimum positioned around an 11 ft (3.4 m) stack, the concentrations on
normalized concentration achievable with a reasonable stack the roof are similar to those for a 5 ft (1.5 m) stack without a
design. Usage controls can reduce the worst-case emission rates, screen present. Hence, the solid screen has reduced the effec-
m, and raise the concentration design goal to achievable levels. tive stack height by a factor of 0.4. This result illustrates the
Such usage controls can include diluted mixtures, smaller liq- adverse effect of rooftop features, which are often not accounted
uid storage quantities or smaller gas bottles. for using the simplified methods discussed previously. Hence,
stack heights can be specified that are not tall enough to en-
Wind-Tunnel Modeling sure acceptable air quality. The effect of screens on rooftop
In the recommended approach, wind-tunnel modeling is used concentrations is discussed in detail in Petersen, et al.4
to predict maximum concentrations, normalized by emission
rate, (C/m), for the stack designs and locations of interest. Discussion and Conclusions
ASHRAE provides more information on scale model simula- This article has provided general information regarding the
tion and testing methods.14 Wind tunnel modeling is recom- need for good stack design and discusses issues that should be
mended because it provides the most accurate estimates of considered when specifying exhausts and intakes. No matter
concentration levels in complex building environments.15 what type of exhaust system is used, the important parameters
As part of ASHRAE research project RP-805, a simple rectan- are the physical stack height, volume flow rate, exit velocity,
gular building 50 ft high, 50 ft wide and 100 ft long (15.2 by expected pollutant emission rates and concentration levels at
15.2 by 30.5 m) was modeled and positioned in a boundary sensitive locations. Whether conventional or entrained air ex-
layer wind tunnel with a simulated suburban approach wind haust systems are used, the overall performance should be evalu-
condition.4 A tracer gas mixture was released from a stack in- ated using the appropriate criterion, i.e., ensuring acceptable
stalled on the roof of the model building at the building center. concentrations at sensitive locations. Selecting an exhaust sys-
The simulated parameters were: 5,000 cfm (2.41 m3/s) volume tem based on an effective stack height alone is not sufficient to
flow rate, 2,000 fpm (10.2 m/s) exit velocity, stack height vary- ensure an adequate exhaust system design.
ing from 1 ft to 12 ft (0.3 m to 3.7 m), with a 16 mph (7.2 m/s) The article also presented a quantitative approach to evaluate
wind speed at 33 ft (10 m). Concentration levels were measured the air quality aspects of exhaust stack design. The approach
 Stack Design

includes dispersion modeling, specifically wind-tunnel model- Transactions 105(1).

5. 1999 ASHRAE Handbook—HVAC Applications, Chapter 13.
ing, to predict maximum concentrations at sensitive locations
6. Briggs, G.A. 1969. Plume Rise. Critical Review Series, U.S. Atomic
such as air intakes, operable windows, and pedestrian areas. Con- Energy Commission, NTIS #TID-25075.
centration goals for design acceptability are based health limits, 7. 2001 ASHRAE Handbook—Fundamentals, Chapter 27.
odor thresholds, and emission rates of chemicals likely to be 8. AIHA. 1992 ANSI/AIHA Standard Z9.5 Laboratory Ventilation.
American Industrial Hygiene Association.
used at the facility. Using the approach for a simple building
9. ACGIH 2001. Guide to Occupational Exposure Values – 2001.
geometry demonstrated that the mathematical methods tend to American Conference on Governmental Industrial Hygienists.
give unnecessarily tall stack heights for an unobstructed roof 10. ACGIH. 2001 Threshold Limit Values for Chemical Substances
and give stacks that are not tall enough for a roof with obstruc- and Physical Agents.
11. ACGIH. 1989. Odor Thresholds for Chemicals with Estab-
tions. lished Occupational Health Standards.
References 12. Ruth, J.H. 1986. “Odor threshold and irritation levels of several
1. NFPA. 1996. Fire Protection for Laboratories Using Chemicals. chemical substances: a review.” Journal of American Industrial Hy-
ANSI/NFPA Standard 45-96. National Fire Protection Association, giene Association 47:A-142.
Quincy, Mass. 13. EPA. 1992. Workbook of Screening Techniques for Assessing Im-
2. 1999 ASHRAE Handbook—HVAC Applications, Chapter 43. pacts of Toxic Air Pollutants (Revised), U.S. Environmental Protection
3. CDC. 1994. “Guidelines for preventing the transmission of Myco- Agency, Office of Air Quality, Planning and Standards, Research Tri-
bacterium tuberculosis in health-care facilities.” U.S. Department of Health angle Park, N.C., EPA-454/R-92-024.
and Human Services, Public Health Service, Centers for Disease Control 14. 2001 ASHRAE, Handbook—Fundamentals, Chapter 16.
and Prevention. 15. EPA. 1981. Guideline for Use of Fluid Modeling of Atmospheric
4. Petersen, R.L., et al. 1999. “Influence of architectural screens on Dispersion. U.S. Environmental Protection Agency, EPA-600/8-81-
rooftop concentrations due to effluent from short stacks.” ASHRAE 009.