July 12, 2001 11:57 AM

LABORATORY SPACE PRESSURIZATION

CONTROL SYSTEMS

Dale T. Hitchings, PE, CIH

INTRODUCTION:

Maintaining the proper differential pressure in laboratory spaces is one of the most

challenging tasks facing the laboratory environmental control engineer. Some of the
items which need to be addressed when designing a space pressurization control

system are:

Hazard Assessment

Constant-volume vs. VAV Systems

Differential Pressure vs. Differential Volume Systems

Negative vs. Positive Pressure Requirements
Control Signal to Noise Ratio
Control Stability and Speed of Response
Failure Mode Analysis

Building Construction impact on space pressure control

Duct Leakage impact on space pressure control

Laboratories and clean rooms may require that a differential pressure be maintained

between them and the adjoining spaces. This requirement may come from code
considerations or from the operational requirements of the space. For example, NFPA-
45 states that laboratory work units and laboratory work areas in which hazardous
chemicals are being used shall be maintained at an air pressure that is negative relative
to the corridors or adjacent non-laboratory areas.. 1 This is to prevent the migration of
Dale T. Hitchings, PE, CIH is president of SAFELAB Corporation His firm specializes in laboratory
planning, safety, and design; industrial ventilation; and indoor air quality consulting. He received is BSME
from Northwestern University in 1981 and is a Registered Professional Engineer and a Certified Industrial
Hygienist Mr. Hitchings is an ASHRAE member and Chairman of the Laboratory Systems technical
committee TC9.10.
Page 1
 July 12, 2001 11:57 AM

fire, smoke and chemical releases from the laboratory space. Labs containing radiation

hazards or biohazards may also be required by different agencies to maintain a

negative pressure to contain these hazards. Clean Rooms , on the other hand, are

normally operated at a positive static pressure to prevent infiltration of particulates.

Even if your building codes and regulatory agencies do not require pressurization you

may wish to include this feature in your facility anyway for the reasons described above.

CONTROL STRATEGIES:
The desired result of all space pressurization control systems is to control the infiltration
into or the exfiltration out of a space. Space pressurization control strategies can be
divided into two major categories: passive and active.

For constant-volume laboratories a passive method involves simply balancing the

system so that the desired space pressurization is achieved. This method has serious

limitations which should be considered carefully before choosing to design a constant-

volume system with passive space pressurization control. This type of system will
work only if: 1) all fume hoods remain on and at constant speed or volume at all times,

2) no exhaust sources (i.e. hoods) are added or removed, 3) the offsets are large

enough to mask changes in exhaust and supply system performance caused by filter

loading, etc., 4) the system is tested and balanced frequently to design conditions, and
5) the system is adequately maintained. If you cannot guarantee (or even desire) all

these restrictions then this design approach is inappropriate for your application.

An active method for use in a constant-volume laboratory involves the utilization of

pressure-independent, constant-volume control devices in the exhaust and supply ducts

Page 2
 July 12, 2001 11:57 AM

to actively and dynamically adjust the flowrates to keep them constant and decoupled

from system static pressure fluctuations.

VAV labs require active methods to control space pressure due to the continuously
changing exhaust volume from the fume hoods and other exhaust sources. Active VAV

space differential pressure control methods may be subdivided into two types: pure
differential pressure measurement/control (P) and differential volume or flow-tracking

(V).

DIFFERENTIAL PRESSURE SYSTEMS:

The P method of space static pressure control is relatively straightforward and a

schematic of it is shown in Fig. 1. In this method, the differential pressure is controlled
with a differential pressure sensor and a controller and the supply air volume is simply a

function of the P, the setpoint, and the PID constants and . Another similar
method of static pressure control utilizes the Bernoulli principle which states that a
pressure gradient will accelerate a fluid to a velocity proportional to the square root of
the pressure differential. These pseudo-P systems utilizes an air velocity probe

mounted in a tube inserted into a hole in the wall between the controlled space and the

reference space. The differential pressure will induce air to flow through the tube and
the velocity of the air is sensed by the velocity probe. A controller then varies the

supply air volume to the laboratory to maintain a velocity pressure setpoint.

FLOW-TRACKING SYSTEMS:

The V method of space pressurization control utilizes analog or digital electronic

controls to measure the real-time variables and solve the dynamic air balance equation.
A typical V system is shown in Fig. 2. The exhaust volume is either measured after

Page 3
 July 12, 2001 11:57 AM

convergence into a manifold or the individual sources are measured and summed as
shown. The supply volume is then controlled (tracked) to achieve the offset. The offset

is the desired infiltration or exfiltration in CFM. A negative offset will reduce the supply

volume below the exhaust volume and will result in a negative space pressure. A
positive offset will increase the supply volume above the exhaust volume and will result

in a positive space pressure. Although the equation in Fig. 2 implies that the offset is a
constant, in practice, it is a variable. As the volume sensors drift in accuracy, the actual

offset will change. It is necessary to choose an offset that is large enough to

compensate for tight vs. loose envelope construction, duct leakage, and the accuracy of
the flow measuring devices. You should choose an offset using Equation 1.

Offsetdesign = 2 S F max Equation 1.

Where:
= instrument error in % full scale or % of reading
S = safety factor: depends on tightness of envelope,
amount of unmeasured duct leakage and
degree of laboratory hazard present;
recommended range: 0.5 - 2.0
Fmax = design maximum exhaust or supply flowrate,
whichever is greater

This will assure (if the safety factor is greater than 1) that under worst case conditions
you still have some actual offset in the desired direction of flow. For example, in a

negatively pressurized lab, the exhaust flow rate will be higher than the supply flow rate,
so the Fmax is the maximum exhaust volume. If Fmax is 5000 CFM, is 5% (0.05), and S

is 110% (1.1) then the Offsetdesign = (2)(.05)(1.1)(5000) = 550 CFM. Therefore,

the worst case scenario would be a system where the exhaust volume reading is 5%
below actual, and the supply volume reading is 5% above actual, giving a total airflow
error of +500 CFM. If the design offset is -550 CFM then the actual offset will be -50

Page 4
 July 12, 2001 11:57 AM

CFM. Normally, airflow errors are random and will tend to cancel. In large labs with

multiple flow measuring instruments, you would expect the total error to be less than the
maximum cited in the example. If this is the case, and there is minimum duct leakage,

the laboratory envelope is very tight, and the laboratory hazard is low, then a safety
factor less than 1 may be appropriate. In any case, verification of actual operating Ps

and tuning of lab offsets should be done at a predetermined frequency based on the
level of hazard in the laboratory. A maximum of six months between offset calibrations

is recommended.

HYBRID SYSTEMS:

In laboratories containing extremely toxic or infectious agents such as Biosafety Level 3

or 4 laboratories, it may be prudent to utilize both a P and a V system to assure that

an adequate differential pressure is maintained at all times. The most common way of

doing this is to design a basic V system as previously described, and add a P

sensor and controller which is used to reset the V system offset. Here, the long time
delays required to produce an accurate average do not affect the speed of response of
the system. The P system can dynamically calculate an appropriate offset. As the
characteristics of the room change, such as duct leakage and envelope tightness, the

offset will change (it usually grows) to maintain the desired laboratory space pressure.
Once the offset has grown to a predetermined value it may be necessary to recalibrate

flow measuring instruments, seal ductwork, or seal bypasses in the laboratory envelope
to bring the system back into specification. Monitoring the offset in a hybrid system of

this type is a good way to monitor the integrity of the total duct/control/envelope system.

Page 5
 July 12, 2001 11:57 AM

STABILITY vs. SPEED OF RESPONSE:

P and pseudo-P control schemes have certain characteristics which the

designer/owner needs to be aware of. A reasonable pressure differential to maintain

using normal construction techniques is approximately 0.01 water gauge (w.g.) To put
this into perspective, 0.01 w.g. = 0.00036 PSI. This is an extremely small pressure

differential (signal) to measure and providing adequate calibration for the instrument is

also difficult. The fluctuations (noise) in this signal, which are caused by the opening
and closing of doors, people traffic, elevators, stack effects and atmospheric

disturbances like wind, are on the order of 0.1 w.g. This represents a signal to noise

ratio of approximately 1:10.

Imagine trying to determine the level of a lake to within an inch when the waves are a
foot high. To do so, it would be necessary to average out the wave crests and troughs.
It can be done, but it takes time. If you want great accuracy you have to average over a
long period of time. If you need to respond quickly to the signal then you cant be as

accurate. Accuracy and speed or response are in direct conflict. For true stability in a
P system, the response time is usually measured in minutes. Therefore, many of
these systems and instruments sacrifice stability for speed and can oscillate about the

setpoint for quite some time before settling down to stable control. Unfortunately, this

settling down period is often greater than the frequency of upsets and the controlled

device may oscillate all day long until everyone goes home.

pseudo-P systems which measure the air velocity are somewhat faster and more

stable because the velocity signals and noise are proportional to the square root of the

differential pressure. This improves the signal to noise ratio to approximately 1:3. This
simple change in the measured variable improves the system performance by a factor

Page 6
 July 12, 2001 11:57 AM

of three. However, the noise is still about three times as large as the signal and you still

may wait as long as 60 seconds for marginally stable output after an upset in the space

pressure. The performance of this type of equipment varies from manufacturer to

manufacturer and care should be exercised when selecting them for your facility.

Another undesirable characteristic of both of these pressure measuring devices is that

the measured variables (pressure or velocity) totally disappear when the laboratory door

is opened. Some controllers have the ability to freeze the output for a predetermined
time delay to compensate for this. However, if the door is left open long enough, the

pressure control system will start to shut down the supply volume in order to bring the
space back to a negative setpoint. When this occurs, the air from the hallway flows into
the open laboratory to replace the exhaust air and the hallway pressure may drop.
Other lab pressure controls which use this hallway as a pressure reference may also
start to close down the supply air to their labs thereby creating a cascade effect. As

more air is drawn into the affected labs from the hallway, the pressure will continue to
drop even more. As you can imagine, this can cause serious building pressure
problems. However, for facilities not requiring critical room pressure control and where

the effects of settling time and stability are not an issue from a hazard assessment
standpoint and where the HVAC system and architectural designs can minimize the

cascade failure effect mentioned above, this control system may be used with some
success.

In comparison, the signals measured using airflow stations by V systems are on the
order of 1000 feet per minute and the noise is on the order of 100 feet per minute

resulting in a signal to noise ratio of approximately 10:1. This allows much more
accurate, stable, and rapid response to changing inputs characteristic of a VAV

Page 7
 July 12, 2001 11:57 AM

laboratory. All airflow measuring devices are not created equal, however. There are

types which employ arrays of hot wire anemometers. These are very accurate and
have a relatively wide turndown but are sensitive to buildup of particulates and

corrosives on the sensors which affects their response time and reliability. Averaging
pitot-tube arrays have a similar problem with buildup plugging the sensing ports and

causing non-averaging response. They are also extremely inaccurate at low velocities
due to the exponential nature of the velocity pressure signal. A velocity sensing

technology which has been used for years to measure liquids is now being applied to
airflow stations. It is called vortex shedding, and is showing up in some newer

laboratory pressure control systems. There is some debate over the theoretical

accuracy of these devices at large turndowns due to reynolds numbers, but field testing
will prove if these devices can be effective in V control systems.

OTHER DESIGN CONSIDERATIONS:

Construction techniques can also influence the performance and effectiveness of space
pressurization controls. Loose construction makes it difficult to establish an effective
differential pressure in the space and large offsets are necessary to compensate. When
large offsets are used, pulling large amounts of secondary air in from adjoining spaces,

temperature and humidity control problems may result. Plugging all the holes and

bypasses in the laboratory envelope during the renovation or construction process may
be required to eliminate this problem.

Duct leakage may also effect the accuracy and performance of V systems. Air

leaking out of or into the duct system between the flow measurement device and the

laboratory envelope can result in significant error. In constant pressure systems this
error may be relatively constant, but if the system static pressure floats then the error

Page 8
 July 12, 2001 11:57 AM

will float also. The author recommends that supply and exhaust ductwork be specified

and leak tested to allow a maximum of 0.5% leakage. This is easily achievable by most
contractors with some practice and guidance in duct sealing and construction

techniques. Using welded and/or flanged and gasketed duct construction can make it
virtually leak-free. Placing flow measuring devices close to the wall penetration in a

tight section of duct is recommended because it minimizes errors by reducing the duct
length where leakage effects the flow measurement. Assuming that the flow measuring
device is located inside the laboratory envelope, leakage which occurs upstream of the

exhaust flow measuring device and downstream of the supply flow measuring station
has already been measured and should not effect the error.

COST COMPARISONS:
The P and pseudo-P systems, due to their simplicity will typically cost less than a
V system for the same laboratory. How much less depends upon the size and
complexity of the laboratory ventilation system and the type of fume hood controls. As

the number of measured and controlled devices increases so will the differential cost. If
exhaust and supply volumes are measured at only one position each, the cost
differential will decrease. However, if you have chosen VAV fume hood controls and

general exhaust controls that have flow measurement capability built into each, then
only a supply air flow measuring device and controller are needed to complete the

system. In this case the cost of the two systems may be negligible. It is difficult to
divorce the cost of the space pressure controls from the cost of the fume hood controls

since they are often integrated.

FAILURE MODES AND SYMPTOMS:

Page 9
 July 12, 2001 11:57 AM

The easiest way to assure proper space pressurization control system operation is to

monitor it with sensitive equipment which has been properly calibrated and to provide
alarms to alert personnel when conditions are outside specifications. A simple

qualitative measure of space pressurization control system effectiveness is smell. If you

can smell the chemicals which are used in the laboratory when you are in the corridor
this may be an indication that your space pressurization controls are ineffective. It may

also mean that contaminated air from the laboratory is be reingested back into the

building air supply due to inadequate stack design. In the case of a clean space,

excessive particulate counts may indicate a space pressure problem causing infiltration.
Another simple experiment which can be done quickly and easily is the foot-in-the-door

test. Open the lab door and place your foot in the doorway next to the jamb and allow
the door to close against it. Next, feel the airflow through this opening with your hand,
or use a smoke tube to determine its direction. Its not quantitative, but it will sure tell
you if your positive when you should be negative (or visa versa) or if there is an

excessive pressure differential. If either of these are detected then a more quantitative
approach should be used to diagnose the magnitude and cause of the problem.

During the design of the laboratory environmental control system, a careful analysis of

the failure modes of each component and the effect of a failure on the operation of the

system should be undertaken. There are many methods to do this, most of which are
beyond the scope of this article. One of the more popular methods is called a fault tree.

This method may be used on entire systems, subsystems, and individual components.

For example, if attempting to do a fault tree analysis on a room pressure control system,

choose the component you wish to analyze such as a through-the-wall velocity sensor.
List this at the top of the tree. Next consider all of its failure modes and list them

underneath. Next, for the most serious failure modes consider all the affects on

Page 10
 July 12, 2001 11:57 AM

controlled components and other coupled and decoupled systems and list them

underneath that mode. Repeat this process until all the paths are complete. It may be
necessary to do all the branches of the tree to discover which paths represent the worst

scenarios. When all the failure modes of all the components, subsystems and systems
have been completed, you can make some strategic design modifications to eliminate

or ameliorate the most serious scenarios by installing more reliable components in key
locations or installing redundant controls or systems to provide backup and thus

truncate the tree.

HAZARD ASSESSMENT: MAKING THE PRUDENT CHOICE

Before choosing a space pressure control system, the design team should assess the
hazards present in the laboratory and choose a system which is appropriate given its
failure modes, the level of hazard and the level of risk that the owner is willing to accept.
Chapter 14: Laboratories of the 1991 ASHRAE Applications Handbook is currently

undergoing a re-write for the 1995 volume. The first section if this chapter is entitled
Risk Assessment. During a discussion of this topic by the Handbook Subcommittee of
TC9.10 Laboratory Systems, it was decided that the words risk assessment and risk
analysis were often mixed up and misinterpreted. Furthermore, the work risk has

certain legal definitions associated with it. The general opinion of the committee was

that risk analysis, if interpreted literally, involves the assimilation of hundreds of pieces
of design and hazard information and the calculation of the probable frequency of

accidents and the probable results of those accidents using experience and/or actuarial
data. The committee felt that most mechanical engineers were not qualified to perform

an analysis of this type and that it should be undertaken by the owner who is probably

more familiar with the specific hazards and processes in their laboratory than the
laboratory designer could ever be. The designer should, of course, participate in the

Page 11
 July 12, 2001 11:57 AM

exercise to supply data about the facility and systems design and their impact on the

risk of operating the facility, including the space pressurization systems. The title of the
handbook section has now been changed to Hazard Assessment since the discussion

of risk was considered to be beyond the scope of the chapter (and this article).

CONCLUSION:

There are many types of space pressurization control systems, methods and equipment

to choose from. You can make your decision based on instinct, vendor data, your own
experience, objective testing, or by seeking advice from a laboratory consultant. The

instinct and vendor data methods involve more risk than some are comfortable with.
User experience is the least risky option, but only if you have the experience with the
equipment. Objective testing provides excellent results but is expensive and time
consuming. Seeking advice from a laboratory design consultant will cost less than
objective testing and much less than replacement of equipment which does not meet

owner specifications. However, the equipment is only one piece of a complex puzzle.
The design of the laboratory envelope, its layout, and the supply and exhaust systems

all interact in an intricate fashion. An experienced laboratory consultant can also lead
you around the pitfalls, help avoid common problems and show you how all the aspects

of the facility affect its performance and help assure that the laboratory facility operates

as safely and efficiently as possible.

REFERENCES:

1NFPA Standard Code No. 45, Fire Protection for Laboratories Using Chemicals,

National Fire Protection Association (NFPA), Quincy, MA, 1986.

Page 12