Clean Agent enclosure design for AS 4212

Copyright 2013 Retrotec Inc. All rights reserved.

By: Colin Genge

Light blue highlight shows sections changed from previous versions.

Article is based on ISO 14520 but where the Australian Standard AS 4212 diverges, the AS 4212 is noted
in the text.

The Clean Agent Discharge

Clean agent fire suppression systems are required in enclosures where a sprinkler system would cause
damage to sensitive contents such as computer servers or historical artifacts. Upon fire detection the
compressed agent, which can be a halocarbon or an inert gas, is released into the enclosure causing a
peak pressure of around 250 to 1250 Pascals to occur for a fraction of a second, the magnitude of which
is dependent upon total enclosure leakage area. Once the enclosure is completely flooded, the agent
will begin to leak out at a rate that primarily
dependent upon lower enclosure leakage area.
The distribution of the remaining agent will either
be constant throughout the enclosure due to
continual mixing or will establish an interface with
air above and agent below an interface that
descends in time. Up until 1988, enclosures
protected by Clean Agents used full discharge tests
to determine the Hold Time but since 1988, Door
Fans have been used for measuring the leakage
area which is entered into formulae in Annex E of
ISO 14520 to predict how long the agent will stay in
the enclosure (Hold Time). Figure 1: Clean Agent discharge causes cooling due to
decompression.
Peak Pressure develops during discharge
It is common practice for peak pressure
calculations to be done for inert agents but
these calculations have rarely been done for
halocarbon agents and that is a big problem
since they can, in some cases, produce as much
peak pressure as inerts. A 5 year research
project, in which the author played a major
role, involved many of the industry
manufacturers (including: Fike, 3M, Ansul,
Kidde Fenwal, Chemetron, Retrotec) and
uncovered many important facts.
1. Existing inert agent formulae under-
predicted peak pressure Figure 2: Peak Pressure increases with enclosure tightness
2. Under certain conditions, halocarbon
agents could produce as much peak pressure as inerts
3. Peak pressure from Halocarbons was extremely dependent upon humidity

Sufficient data was gathered to accurately predict the peak pressure for all agents and to develop a
more refined Hold Time model.

M:\Masters\Articles-Fire by Rt\Clean Agent enclosure design for AS 4212.docx Page 1 of 14 2013-11-06

 200 to 400 % error at 10 min. hold time

Figure 3: Peak pressure is a function of LVR (Leakage to Volume Ratio). Notice how the existing
formulae
Annex A of ISO 14520 Section A.2 Working Documents lists:
Peak Pressure Calculations must be performed
Annex A of ISO 14520 Section A.2 states “Working documents shall include” under section n) “enclosure
pressurization and venting calculations”. This dictates the Minimum Allowable Leakage Area the
enclosure must have according to the new peak pressure equations that have come out of the research
project. This Leakage Area can be provided by accidental enclosure leakage and/or the area of any
pressure relief vents (PRV) that will be open during the discharge period although relying on the open
area of the PRVs is the most common method. The Enclosure Integrity Procedure in Annex C is ideal for
creating two leakage area values, one used for the calculation of the Hold Time and another used for
evaluating Peak Pressure. These values must be measured after the enclosure has been completed.
Clearly it would be extremely bad news to find out that a completed enclosure needed to have a PRV
installed a few days before occupancy, but fortunately the designer can run calculations in advance
using the “Retrotec Peak Pressure and Hold Time Calculator” (available at Retrotec.com) to determine
whether a PRV is needed. The design may then be altered to optimize the enclosure using the tips
presented in this article.

It is not wise to simply specify a Pressure Relief Vent (PRV) of the correct size but its leakage rate must
be measured after installation to ensure the vent both opens at the correct pressure and has a large
enough leakage path to outdoors to prevent the Peak Pressure from exceeding the specified limit. This
will now become an additional part of the Enclosure Integrity Procedure which is being done already and
requires the same type of door fan equipment, albeit with more power in many cases, but very little in
added time.

The new peak pressure values are now in the FIA Guidance on Pressure Relief in the UK as well as the
FSSA Guidance in the US. For the first time, comprehensive formulae have been published and these
formulae are currently embedded in Retrotec’s FanTestic Integrity software.

Page 2 of 14
Selection of an appropriate Hold Time
After a typical 10 second discharge for halocarbons or 60 seconds for inerts, the Hold Time begins. Even
though this time was almost always specified as 10 minutes, this value is not always the correct Hold
Time. The designer must consider what the “10 minutes unless otherwise specified by the authority”
will actually be because much longer Hold Times are required for remote sites or those with heavy fuel
loads while much shorter Hold Times should be considered for small enclosures that are manned 24-7
but the Standard makes no provision for shorter retention times. Reducing this Hold Time to 6 minutes
for a small 35 cubic meter enclosure and to 3 minutes for 10 cubic meter enclosure would solve one of
the most costly and pernicious problems that installers face, where getting these enclosures tight
enough to pass the 10 minute requirement becomes virtually impossible.

Figure 4: ISO 14520, 2000 edition model for descending interface where 100% of the initial concentration leaks out the bottom
of the enclosure which replaces the lost volume with 100% above the interface.

Figure 5: The current version uses the Wide Interface which decreases retention time by about 40% forcing
enclosures to be 40% tighter and more difficult to pass.

The Wide Descending Interface was introduced into ISO 14520 Annex E in 2003 to allow for predictions
of Hold Times for any drop in initial concentration. It was based on a theory that the original NFPA
equations predicted the center of the interface. Instead of having a Sharp interface with the initial
concentration below and 0% concentration above the interface, it assumed that the concentration at
the interface was half the initial which slowly decreased to 0% at the top of the enclosure. The interface
thickness was then double distance from the predicted interface to the ceiling. The concentration was
then assumed to increase proportionally from 0% at the ceiling to 100% of the initial concentration at
the bottom of the interface. Since the point of interest was typically 85% of the initial, it was necessary

Page 3 of 14
for the 50% point in the interface to fall much less for any given concentration where the final was
greater than 50% of the initial concentration. This math was accomplished by cleverly calculating an
effective height of interest, He that was entered into the NFPA equations to trick them into calculating a
Wide Interface Hold Time. The effect of this was to decrease predicted Hold Times by up to 40% which
required enclosures to be much tighter than before, adding air sealing costs and increasing the size of
PRVs that would then have to handle the increased pressure.

A saving grace that came out of this was that the research also showed that the NFPA Sharp Interface
model previously followed in the CA2001 software and referenced in AS 4212, was actually much closer
to the test data than the ISO Wide Interface Model as shown in the proceeding graph for a typical inert.
The dots are the experimental data and the solid line is the old equation while the new proposed (and
accepted but not yet published in NFPA 2001 or ISO 14520) equation is the dashed line in the middle.

When it’s considered that this is the worst case for leakage and most often the actual data will show
much more retention, there is a good case for keeping something closer to the original NFPA Sharp
Interface equation which is shown by the solid line. Halocarbons such as HFC227ea (FM-200) are even
closer to the Sharp Interface equation. A compromise solution was arrived at that reduced the
conservatism of the existing standard and accurately represented the recent hold time test data while
maintaining a reasonable failsafe margin.

Page 4 of 14
 Figure 6 Halocarbons show a good correlation to the old formula

A graphical representation of the two interfaces is shown in Figure 7. The ISO Wide Interface on the
right is the most conservative while the Actual interface is much thinner and is closer to the NFPA Sharp
Interface which would be positioned on a line at the 50% point on the ISO graphic. In actual discharges
the Sharp Interface prediction used by NFPA is much closer to what can be expected although it is
theoretically possible for the interface to occur on the Proposed line but unlikely. ISO offers a wide
margin of safety with respect to Hold Time but loses out because it forces enclosures to be much tighter
which presents a larger danger of overpressure.

Figure 7 Visualizing the existing Wide interface and the new interface uncovered during the
research.

Page 5 of 14
Figure 8. Graphical demonstration of the three descending interface models. From left to right: Sharp Interface, Wide Interface,
and Thick Interface. The x-axis represents fraction of agent remaining.

Optimizing Peak Pressure and Hold Time performance

Clean agent discharges can produce damaging peak enclosure pressures that increase as total enclosure
leakage area decreases. Simply providing a lot of enclosure leakage area to solve the peak pressure
problem creates another problem because Hold Times decrease as the leakage area increases. One
solution is to add a pressure relief vent (PRV) that will provide increased leakage to reduce the peak
enclosure pressure; which can then be made tight to provide the specified Hold Time. Another solution
is to carefully consider the design parameters that affect peak pressure and Hold Time so that both
requirements are met without using PRVs. Even if this effort still requires the use of PRVs, optimizing the
enclosure will increase the degree of fire protection and possibly allow the use of smaller PRVs with
more passive protection built in.

Currently, many inert agent protected enclosures have PRVs installed where they are not needed but
other enclosures (both inert and halocarbon) need PRVs but they are not installed. This situation should
be resolved by using the new Enclosure Integrity evaluation procedure along with the new Peak
Pressure formulae.

Understanding the factors that affect the relationship between Peak Pressure and Hold Time will allow
for designs without PRVs that easily pass both criteria. Invariably a few simple changes to the enclosure
will dramatically improve the suppression system’s performance and also save the installer from having
to resolve difficult design problems in a last minute panic when the enclosure fails one or more of the
acceptance criteria which typically occurs just prior to occupancy.

The fail safe approach is to ignore enclosure leakage and to rely completely on venting to relieve the
pressure (seal tight and vent right). While this is the best solution in theory because containment of the
hazard is maximized, it does leave the enclosure vulnerable in most cases to over pressure in cast the
PRV fails which might be more damaging because of the frequency of false discharges than the fire
hazard itself. Potentially motorized vents are likely to fail compared to gravity weighted vents. Either
way, an evaluation done on the enclosure will identify the degree of risk in case of PRV failure. This also
points to the importance of testing the relief vent path annually to ensure it is not blocked (as if often is)
after checking to see if the PRV opens as its designed to do (as it often doesn’t).

Page 6 of 14
Vent selection
Not all vents are created equal, and the procedure of testing vents during commissioning and annual
inspections will start to highlight the differences in performance between various vent types.

The selection of a quality vent is a critical part of pressure relief that is often overlooked. Simply
knowing the required Free Vent Area (FVA) is not enough. Many passive (non-powered) vents are not
balanced, and will require a tremendous pressure to actually achieve their advertised FVA, which can
result in damaging pressure buildup before the vent has a chance to fully open. A low (50-100Pa)
opening pressure is ideal for this application, and a balanced blade system will ensure that the vent
opens fully soon after the opening pressure is reached. When selecting a passive vent, make sure to ask
the manufacturer for the following information:
 Free vent area
 Opening pressure of the vent blades (ideally between 50-100Pa)
 Pressure needed to fully open the vent (ideally not more than 150Pa)

Enclosure design rules

Specify sealing of the walls to the upper slab. Extending walls to the upper slab and sealing them airtight
is the only defense from fire and smoke entering the enclosure from the outside. Refer to C-1.2.1 (2) in
NFPA2001 which states “…enclosures absent of any containing barriers above the false ceiling, are not
within the scope of Annex C” meaning the enclosure will be difficult to test and verify. This statement
does not specifically appear in the ISO, EN and AS standards but can be inferred by the tone of the
document.

Flooding to the maximum reasonable height will reduce the need for airsealing and reduce the size of
the PRV, saving on installation costs.

Reduce the height of the protected equipment.

Doors often get wedged or propped open when the enclosure is in use. This practice impairs the clean
agent systems ability to put the fire out. A better solution is automatic door release mechanisms that
will close the doors whenever the first alarm sounds. Choose a door opener that will close the door
when it is de-energized so it is failsafe.

If a False Ceiling is specified, seal lower leaks first until the specified Hold Time is reached to reduce
reliance on the PRV. Use Section E.3.4.3 where the leaks are “temporarily sealed” using a second fan for
neutralization which will increase the Hold Time but not the Peak Pressure.

Ensure the False Ceiling has at least 5% open area to prevent it from being dislodged as the discharge
vents upwards.

Page 7 of 14
Figure 9: One Door Fan depressurizes the room while the second depressurizes above the ceiling so the pressure across the
ceiling is zero which effectively seals off the upper leaks allows the lower fan to measure the room leaks separate from above
ceiling leaks.

Determine whether it’s likely that the agent will be continually mixed during the Hold Time. If air
handlers continue to run, then continual mixing is certain but even equipment cooling fans or thermal
effects can be sufficient to cause continual mixing and in those cases the initial concentration must be
increased. The standard specification requires that the final concentration at the end of the Hold Time
is 85% of the design concentration or in Australia cannot fall below “Cmin” which is the agent’s minimum
extinguishing concentration. For continual mixing the initial concentration must be increased. Enough
agent must be discharged so the initial concentration is at least 15% over the design concentration.

If no mixing will occur, keep the height of the protected equipment to a

minimum. If the equipment height exceeds 75% of enclosure height,
continual mixing may be the only way to ensure a reasonable retention time.

Figure 10 As agent is lost, air

continually mixes with the
Pressure Relief Vent (PRV) tips agent to provide the same
If PRVs must be installed, there are several rules to follow to optimize their concentration everywhere in
the enclosure.
performance. Install vents as high as possible so that the lighter air, not the
more dense agent, is vented. Vents should open at pressures no lower than
50 Pa to ensure they don’t open unintentionally under normal HVAC pressures and no higher than 100
Pa so the pressure is vented early enough to prevent it from building up. Inert discharges always create
positive pressures and must vent out of the enclosure but halocarbons may create positive and/or
negative pressures creating a need to be vented in either direction or both depending on the agent and
the humidity. Ensure the correct direction for venting with the PRV is specified. All PRVs should be
inspected annually to confirm they will open according to their specifications and to verify that the vent
path to outdoors has not been accidently restricted which is quite common.

Page 8 of 14
Figure 31: the sign says "DO NOT OBSTRUCT", because it is very likely the vent path will be
obstructed which is why the vent path must be checked regularly.

When selecting a vent always ensure that the vent manufacturer has test data behind their performance
statements. Venting to atmosphere may not be straight forward and venting through additional rooms
before reaching atmosphere may be the only option available to designers. Cascading venting can be
planned for and the FIA UK guidelines give designers guidance on sizing vents for this specific situation.

Peak Pressure evaluation tips

PRVs that are designed to open at a certain pressure must be tested prior to and/or after installation to
verify the open at the prescribed pressure. This pressure can be imposed upon the damper in a test box
or the entire enclosure can be pressurized or a temporary pressure box can be constructed around the
damper for testing purposes. A large flow at a fairly high pressure will be required to test these vents in
their open position, so consider testing them in a test box. Once the position at test pressure of 125 Pa
is determined, the vanes must be locked in that position while the damper leakage area is tested. If
installed in a test box where there are no bias pressures, it can be tested in the direction of intended
venting. If installed in the enclosure, it should be tested in both directions to compensate for any bias
pressures and to achieve a more accurate test due to the increased amount of data collected. Ensure
the PRV is tested in the flow direction that will occur during discharge. There are dual acting PRVs that
will open in both directions but there free vent area differs with respect to direction so they must be
tested in both directions to see how open they are at 125 Pa.

Specified Enclosure Pressure Limit

Formulae have been used for over a decade to predict peak pressures and to size PRVs for thousands of
enclosures without damaging those enclosures. Since the 5 year research project showed that the
actual pressures exceeded those of the previously used formulae by at least 100%, and many of those
enclosures were discharge tested with inert agents, it is safe to say that a wide range of enclosures

Page 9 of 14
handled 500 Pa of peak pressure with ease. This can also be verified by using a high power fan to
pressurize enclosures where we have noticed no effects at 500 Pa. We can therefore assume that a
double sided wall that is securely fastened top and bottom will handle 500 Pa and that this can be used
as a “specified enclosure pressure limit”. If in doubt, test a wall section under the specified enclosure
pressure limit but this will require a specially designed fan or high power Door Fans can be used in
series.

While thicker walls can take more pressure, False Ceilings can only take about 50 Pa so they must be
protected from pressures higher than that with vented tiles.

Figure 42: The Wall Strength from 1997 HOTWC paper. One PSF = about 48 Pascals.

The previous edition of AS 4212 provided the following conservative enclosure pressure limits. Although
it has not been re-published in AS ISO 14520-2009, it can still be used as a guideline.

Figure 13: The Wall Strength published in AS 4212 and taken from the University of Manchester and the UK building code

Page 10 of 14
Figure 54: Peak pressure curves for all tested inert agents.

Page 11 of 14
Removal of formulas for extinguishants lighter than air

Testing has shown that the assumption used in the current ISO standard that a lighter than air agent
such as Nitrogen would rise, was incorrect. Because of cooling during discharge, the data clearly shows
a descending interface for Nitrogen [IG- 100] up to at least 0.75 of the initial concentration remaining.
After the concentration degrades beyond 0.75, the concentration in the middle of the enclosure stays
more or less constant with a slight degradation in concentration at the top and bottom. Finally, as the
concentration degrades below 0.5 of agent remaining, the interface begins to ascend but by this time
the remaining concentration is well below the point of interest. It must therefore be assumed that
either the interface is descending or that no reliable interface is established at all which would require a
continual mixing calculation to be performed making current ISO interface calculation irrelevant. In
either case, there is ample justification to remove all of the equations that refer to an ascending
interface because there is simply no evidence of one occurring.

IG-100 Discharge Data

(Test ID: 42) 75% remaining
4.5

4.0

3.5
Height (m)

Proposed
ISO
3.0 Experimental Data
Mixing

2.5

2.0
0.0 500.0 1000.0 1500.0 2000.0 2500.0 3000.0
Time (s)

Notice that in both cases the existing ISO formula actually goes in the opposite direction. The vertical
green line on the above graph shows how the continual mixing equation would relate to the actual
experimental data. The driving force is the difference in density between the nitrogen mixture and air
with no consideration as to the location of the leaks.

Retrotec’s FanTestic Integrity software does the analysis using the density difference of Nitrogen versus
air but does not provide an Ascending (or Descending) calculation because that will not occur. The ISO

Page 12 of 14
committee has accepted the removal of the Interface calculation but has maintained Continual Mixing
for the next version of the standard. The 2012 Edition of NFPA2001 has already published the removal
of the interface calculation and acceptance of continual mixing for all lighter than air agents. Retrotec’s
software reflects this development even though it contradicts the existing text but it would make no
sense to apply that part of the standard when it has been proven incorrect. Retrotec maintains the
calculation for Continual Mixing which has been proven correct. Leakage split is restricted to 50/50
(F=0.5) because the agent will alternately leak out of any opening regardless of position.

Conclusion
Currently, peak pressure equations are under-predicting peak pressures experienced during discharges.
The FSSA and FIA equations solve that.

The ISO Hold Time models are too conservative whereas the NFPA equations are much closer. Using ISO
equations but leaving out the effective height calculation (He) gives the same results as NFPA.

Door fan data can accurately measure Pressure Relief Area to determine whether the enclosure is safe
from the greatest hazard which is usually over pressure during discharge.
References
1. NFPA 2001 Edition2012, "Standard on Clean Agent Fire Extinguishing Systems”, National Fire
Protection Association, Quincy, MA, 2001.
2. “Clean Agent enclosure design optimization for peak pressures and agent retention”, SFPE
Conference Portland, 2011-10-23
3. “ISO 14520 Gaseous media fire extinguishing systems”
4. “Guidance on the pressure relief and post discharge venting of enclosures protected by gaseous
firefighting systems”, Fire Industry Association, Tudor House, UK
5. AS 4212 – 2002, Gaseous Fire Extinguishing System.

Colin Genge is the founder and owner of Retrotec Ltd, a Canadian company founded
in 1980 that does research for air leakage standards and designs for air leakage test
equipment and software. He is also the owner of Retrotec Inc, a US company that is
the world’s largest manufacturer of air leakage measurement systems with
customers in over 60 countries.

Page 13 of 14
Page 14 of 14