White Paper
“What A Coincidence! –
Very Early Warning Smoke
Detection Together With
Suppression Actuation”
Overview
The early detection of smoke becomes significantly more difficult in
areas, such as computer rooms and telecommunications facilities,
where the air, and hence any smoke, is being continuously circulated
and filtered by air handling systems. It is particularly crucial in these
areas that smoke be detected as early as possible in order to minimize
the spread of smoke to other areas, electronic equipment damage,
disruption to business continuity, staff exposure to toxic smoke and the
many other consequential losses incurred as a result of a fire. Airsampling Smoke Detection (ASD) systems have the capacity to reliably
detect smoke in high air-flow environments, where other technologies
fail to do so. For this reason, ASD systems are commonly used in
datacom installations to provide very early warning.
In addition to smoke detection, there is usually some form of fire
suppression that can be released when a fire event reaches the stage
at which this is justified. Although ASD systems are not commonly used
to trigger suppression, there is no real reason why they should not be.
To prove this point, we conducted a computer modelling study,
supported by testing, which demonstrates the equivalence of ASD to
the accepted practice of using point-type smoke detectors for controlled
initiation of suppression release. Using the results of this study, we
developed a software tool that will assist fire protection system
designers to set appropriate ASD alarm thresholds that achieve this
equivalency.
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
Introduction
What’s It All About?
The purpose of this white paper is to present a fire protection solution,
using ASD as both a very early warning smoke detector and as a code
compliant triggering mechanism for the actuation of fire suppression
systems.
We conducted a series of computer simulated and empirical tests to
establish what ASD alarm sensitivity settings would be required to
produce a suppression triggering performance equivalent to that of the
coincidence detection technique using two standard point-type smoke
detectors. The results of this study were then used to develop a software
tool for calculating the ASD alarm thresholds needed to provide
equivalency with a variety of different sensitivity point detectors under
differing area of coverage, room volume and air-flow conditions. Fire
design engineers given the task of setting up this type of fire protection
system can use this software tool, known as the ASD Suppression
Actuation Threshold (ASAT) Calculator, to appropriately set the ASD
alarm levels for both very early smoke detection and timely suppression
release.
Smoke Detection
Challenges
Computer installations, data centres and telecommunications facilities
(otherwise known as datacom facilities) present a significant challenge to
those set the task of designing their fire protection systems. The
increasing computing and communications equipment density in such
facilities generates enormous amounts of heat. As the temperature rises,
it becomes more difficult to keep the datacom facility cool. Therefore, the
risk of runaway temperatures causing a fire also increases.
As the increasing heat density trend continues, the struggle to keep these
mission-critical spaces cool drives air speeds and recycle rates up. The
air handling units circulate cooled air to control the temperature but, in
doing so, they also impede smoke detection. Smoke is drawn along with
the moving air, potentially away from the smoke detectors. Air movement
also has a dilution effect on the smoke. Under such circumstances, smoke
could go undetected for some time and its concentration could be allowed
to reach undesirable levels.
Smoke detection is further impaired by the High Efficiency Particle
Arrestor (HEPA) filters that form part of the airconditioning system. During
the airconditioning cycle, these filters remove much of the smoke from the
air before it is returned to the room. This prevents the normal smoke buildup which would otherwise aid detection.
In datacom environments, fires are commonly the result of electrical
overheating in cables, connectors or circuit components and involve a
large quantity of plastics. Fires of this nature usually begin as low energy,
cool smoke fires which can take long periods of time to build up to the
point at which flames appear. The detection of fires in the early incipient
(smouldering) stage of a fire is extremely difficult.
Although plastics fires generate large amounts of smoke, electrical cables
and components are usually located within enclosures, ducts or conduits.
It may, therefore, take some time for the smoke to enter the open area
where conventional smoke detectors would be placed, in accordance with
the relevant fire standards.
Page 2 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
The earliest possible detection is critical for the following reasons:
•
•
•
Fire Suppression
Challenges
To minimize smoke damage – This has an impact on business
continuity and the protection of assets. According to the USA Federal
Commission of Communications (FCC)[1], around 95% of all damage
done by fires in computer rooms is actually the result of the highly
corrosive substances released in the smoke of plastics fires; very little
damage being due to flame.
The adverse effects of plastics smoke – The most important impact of
this is the risk to personnel. The substances released by burning
plastics are highly toxic as well as corrosive.
Suppression actuation – Where the presence of smoke is used to
activate fire suppression systems, both a suppression release and the
clean up process will disrupt business continuity; some suppressants
may even damage assets.
In addition to smoke detection, most datacom facilities would require
some form of fire suppression to be activated should a fire get out of
control; as a last resort as it were. Suppression systems permitted by the
NFPA 76[2] regulations for telecommunications installations are: sprinklers,
water mist and clean agents.
Sprinklers are the most popular suppression system. However, in the case
of IT or Telco facilities the electrically conductive water from a traditional
sprinkler system would do considerable damage to the electronic
equipment being protected.
The requirement for a clean electrically inert suppressant, non-injurious to
personnel, has driven the development of a number of commercial clean
agent suppressants. Some are complex chemical compounds which act
on the fire in various ways, such as reducing the heat, while others are
pure inert gases which displace the oxygen to inhibit combustion. The
inert gas elements (helium, neon, argon, krypton and Xenon), due to their
elemental nature and chemical stability, are able to extinguish fire without
seriously damaging equipment or furniture. Carbon Dioxide (CO2) is also
used as a clean agent suppressant but in high concentrations it is toxic if
inhaled.
The type of gas chosen depends on the enclosure being protected. In the
case of datacom environments, for example, it is advisable not to use a
suppressant which breaks down into corrosive by-products such as
Hydrogen Chloride (HCl) and Hydrogen Fluoride (HF). These chemical
compounds are extremely damaging to electronic equipment.
Regardless of the type of suppressant used, a crucial issue is, at what
stage during a fire should we trigger the release of fire suppressants?
Unnecessary or poorly-timed suppression dumps must be avoided since
the cost of such mistakes is enormous. Consider the following:
•
Delayed fire suppression will expose the facility to unnecessary
smoke and heat damage; together with all of the associated risks to
life safety, business continuity and assets. Releasing suppression too
late also drastically increases the possibility that the fire will not be
brought under control by the suppressant.
Page 3 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
•
•
The Consequences Of
An IT Or Telco Fire
Unnecessary suppression release is an extremely expensive mistake;
a volume of clean agent sufficient for even a modestly sized area can
cost tens of thousands of dollars to replace.
A premature release of suppression (for example, before the flaming
stage of the fire) creates the risk of there being no suppressant left to
minimize damage should it be required later on.
So, what would be the consequences should a fire start in a computer
room or telecommunications facility? There would be two main
consequences: the significant health and safety risks to personnel and the
extensive financial costs incurred.
Plastics, such as Polyvinyl Chloride (PVC) and Polymethyl Methacrylic
Acid (PMMA) can be found in abundance in these facilities. PVC is
commonly used as insulation around electrical wiring and, along with other
plastics, also forms a large percentage of the casings and furniture
materials commonly found in datacom facilities.
Chemical decomposition is the name given to the normal aging process of
synthetic polymers (plastics). During combustion, this process is greatly
accelerated and results in the release of several toxic and corrosive
substances; the most common of which being Hydrogen Chloride gas
(HCl). This highly toxic and corrosive acidic gas, known as Hydrochloric
acid in its liquid form, will not only burn any tissue that it comes in contact
with, it will corrode the metallic components of electronic circuits on which
it is deposited and ignite if the temperature continues to rise. Lethal levels
of HCl will be given off in a matter of two to three minutes after even a
very small amount of plastic has reached its decomposition temperature,
that is, in the early stages of combustion long before any flames appear.
Hence, there is a high likelihood that staff and equipment will be exposed
to this gas before the smoke is detected. Exposure to small amounts of
HCl will cause a burning throat and watering eyes which will alert staff to
the potential presence of other lethal gases such as Carbon Monoxide
(CO) and Hydrogen Cyanide (HCN). Should HCl ignite, the resulting
flames will spread extremely rapidly.
Other toxic substances are also found in the smoke and soot from plastics
fires. For example, Chlorinated Dioxins, Dibenzofurans and Benzine are
given off by burning PVC. These substances are all known carcinogens
(cancer causing). The neurological poison methylmethacrilate will be
released in the smoke of a PMMA fuelled fire.
The physiological effects of exposure to toxic combustion bi-products and
smoke inhalation, are long-term contributors to the overall financial cost.
Loss of productivity due to increased taking of sick leave, less efficient
workers due to poor health and compensation claims on the grounds of
death or permanent injury are all realistic possible outcomes.
The corrosive effect of HCl, deposited on electronic equipment, is another
long-term cost contributor. Eventually it would be necessary to purchase
replacements for affected components or perhaps entire pieces of
equipment. It is widely recognised that even minute amounts of HCl per
cm2 causes a level of corrosion capable of rendering equipment
inoperable[3].
Page 4 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
More immediate expenditure includes the following:
•
•
•
The Two levels of
protection
The loss of revenue as a result of the business down time required to
repair damage and clean up soot. This could possibly amount to
millions of dollars per hour depending on the extent of the damage
and size of the business.
The considerable expense of the clean-up process itself. Plastics fires
are very messy so this could also amount to many thousands of
dollars.
The enormous expense of replacing any clean agent fire suppressant,
if released during the fire event. As discussed earlier, this too could
potentially cost thousands of dollars depending on the amount of
clean agent required.
Now that we have considered the difficulties faced by smoke detection
devices and the risks involved with fire suppression within a datacom
facility, it is evident that the earliest detection of fire possible is the best
practice. We next need to decide which smoke detection device to use.
Logically, for early warning, we would need very sensitive detectors
capable of detecting very low concentrations of smoke (obscuration
levels). Obscuration is the term given to the amount of smoke which
obscures vision to a particular degree and is measured in percent
obscuration per metre or foot (%Obs/m or %Obs/ft).
We also need to determine a suitable method for triggering the release of
clean agent suppression. The clean agent must be dumped at the most
appropriate point during the fire event; not too early when it would be both
ineffective and a waste of money and not so late that the fire has caused
damage which could have been avoided. Logically, in this case, we would
need a smoke detection device with a lower sensitivity than for very early
warning, to prevent unnecessary suppression releases at relatively low
obscuration levels.
So, our fire protection system must incorporate the following:
•
•
Sensitive smoke detectors for very early warning.
Less sensitive smoke detectors for appropriate suppression release,
at some stage after the very early warning device has issued an
alarm.
Fire Protection Technologies
Photoelectric
Point Detectors
Photoelectric point detectors, like those shown below, were originally
designed for the detection of the large smoke particles produced by
plastics fires. This type of detector is passive; in other words the smoke
must make its own way into the detection chamber. To do so, it has to
overcome several obstacles.
Typical photoelectric point
detectors.
Page 5 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
In order to reach ceiling mounted detectors, smoke particles must either
obtain sufficient thermal energy from the fire to ascend to ceiling height or
rely on simple diffusion.
Once there, they must then penetrate the insect-proof cage which
surrounds the detector chamber. If smoke makes it this far, it must then
reach a preset density level or obscuration level in order to cause an
alarm.
Inside the detector chamber there is a light beam. When smoke particles
drift into the normally straight path of this light beam, they scatter it in all
directions. Some of this scattered light will strike a photoelectric sensor
causing an electric current to flow and an alarm to be issued.
In ideal still air conditions, point detector sensitivities range from 2 to
12%Obs/m (0.6 to 3.7%Obs/ft). As discussed earlier, the airconditioning
required to prevent the overheating of electronic equipment commonly
interferes with the diffusion process by diluting or carrying smoke away
from this type of detector.
Although plastics fires produce large amounts of smoke, during their
incipient (smouldering) stage, this smoke may not possess sufficient
thermal energy to rise to the ceiling regardless of air movement or cooling.
Hence, photoelectric detectors mounted on the ceiling may not be able to
detect smoke early enough even in a still air environment.
Air-sampling Smoke
Detection (ASD)
Systems
ASD systems would be more likely to detect smoke, in a high air-flow
environment, than the point type detectors discussed above. This is partly
due to their higher sensitivity 0.005%Obs/m (0.0015%Obs/ft) and partly
due to the way ASD devices operate.
A typical ASD device consists of sections of small diameter pipe with
sampling holes drilled at regular intervals along their lengths. An aspirator
(fan) in the detection unit, at one end of these pipes, actively draws in air
and smoke through these sampling holes towards the smoke detector.
Once inside the detection chamber, a laser light scattering technique is
used to determine the amount of smoke present in the air sample. The
ability to actively collect air samples from the vicinity of the sampling holes
and the sensitivity of the smoke detection technique, allows ASD systems
to detect smoke very early - in the incipient (smouldering) stage of a fire
event. An illustration of an ASD installation is presented below.
Example of an air-sampling smoke detection system
Page 6 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
This very early warning capability is exactly what we are looking for, but
how would an ASD cope in a high air-flow environment? Since the ASD
device actively draws air into its detector chamber, it will actually
counteract the effects of air-flow due to the airconditioning; moving air can
be drawn in just as easily as still air. In fact, smoke that has been mixed
and diluted by rapidly moving air has a better chance of passing close to
and being collected by numerous sampling holes, thereby, increasing the
amount of smoke reaching the detection chamber. The more
homogeneous (evenly distributed) the smoke in the room is, the higher the
number of sampling holes likely to take in smoke. This results in an
increase in the obscuration of the smoke delivered to the ASD and is
known as the “cumulative effect”. This phenomenon provides ASD with a
significant advantage over point detectors in environments with high airflow.
Another advantage of this type of smoke detection device, especially in IT
or Telco installations, is its flexibility with respect to where the pipes can
be located. ASD sampling pipes can be run along the ceiling to protect the
open area and/or be put under the floor void where most of the electrical
cabling is confined. It is also feasible to sample air at the return air vent of
the airconditioning system. Placing an ASD pipe here ensures that if any
smoke was diverted from the ceiling mounted or under floor void sampling
holes by the airconditioning, it will be detected before the filtered air is
circulated back into these areas. The illustration below shows ASD
sampling pipes on the ceiling, under the floor void and at the return air
vent of the airconditioning unit.
Illustration of ASD sampling pipes on the ceiling, under the floor void and at the return air vent of
the airconditioning unit.
Page 7 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
Other advantages of ASD are its wide sensitivity range 0.005 to
20%Obs/m (0.0015 to 6%Obs/ft) and the fact that multiple alarm
thresholds can be set. These multiple alarm thresholds can be used to
ensure that responses to various situations are proportionate to the level
of risk. A low level alarm might be issued where only a low priority
investigation is required, while a higher level alarm might be used to call
the fire brigade or to release suppression.
Triggering Suppression
In order to minimize the possibility of unnecessary suppression release as
a result of a single detector issuing a false alarm, it is common practice
that two or more detectors have to issue an alarm before a suppression
dump can occur. The two detector method is known as a coincidence
detection scheme, while more than two alarms constitutes a ‘counting
system’.
In the past, fire regulations have stipulated that detectors used in the
coincidence detection method must be of different technologies. However,
due to the increasing trend against the use of ionisation type point
detectors, because of their radioactive component, this is now not
mandatory. Two detectors of the same type can now be used to trigger
suppression provided that they are independent of one another.
The following point type detection devices are commonly used in a
coincidence detection scheme; either two of the same type or two of
different types:
•
•
•
Optical detectors - Photoelectric point detectors fall into this category.
Ionisation point detectors – this type of detector is being phased out
as mentioned previously.
Heat detectors – This type of detector can detect a particular
temperature or a rate of rise in temperature.
The illustration below shows a popular technique where two photoelectric
point detectors are used for coincidence detection.
Example of the use of two photoelectric point detectors in a coincidence detection scheme for
triggering suppression.
Page 8 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
The major objection to the use of heat detectors for suppression release
arises from the fact that we require very early warning of the presence of
smoke, not the presence of a flaming fire. Heat detectors would only alarm
once a flaming fire had caused a significant rise in temperature. The
prolonged incipient stage (smouldering) plastics fires that we expect in
datacom facilities may produce an abundance of smoke but insufficient
heat to set off the heat detectors. The cooling effects of the airconditioning
system would further compound this problem.
We decided to explore the feasibility of using ASD devices for the dual
purpose of very early warning and triggering suppression release. The
wide sensitivity range of ASD and its multiple alarm levels means that a
sensitive alarm threshold can be set to provide very early warning smoke
detection, while a second independent less sensitive alarm level can be
set for suppression release. Even where ASD systems are being used for
very early warning smoke detection, in datacom facilities, suppression is
usually triggered by a second detection scheme, based on photoelectric
point or heat detectors, which is an expensive duplication of fire protection
systems. Using ASD devices for both would mean that only one
technology need be installed and maintained.
About Our Research
Research Goals
We conducted this project with the following objectives in mind:
•
•
•
The Computer Models
Since many current fire codes and standards require that suppression
actuation be governed by coincidence detection, based on
conventional detection technologies, we wished to show that ASD
devices could also be used for this purpose and that their
performance would be equivalent to that of the conventional
technologies. In particular, we wanted to prove equivalency with
photoelectric point detectors.
Since the cumulative affect that benefits ASD devices is not observed
with conventional technologies, we also wished to show how this
phenomenon can assist the detection of smoke within high air-flow
environments.
Although ASD system specifiers and designers learn by experience at
what values alarm thresholds must be set for suppression actuation,
we wished to provide them with a more scientific and consistent
means of determining the appropriate alarm settings for point detector
equivalent coincidence detection.
We used the Fire Dynamics Simulator (FDS) developed by the National
Institute of Standards and Testing (NIST) to generate a series of
Computational Fluid Dynamics (CFD) computer models of a variety of
computer rooms, detector configurations, air change rates and fire
locations.
The fires were modelled in the following three locations:
•
•
•
Next to the return air vent of the airconditioning unit.
Under the floor void next to the return air plenum.
Far away from the return air vent.
Detection points, representing either point detectors or ASD sampling
holes, were positioned as shown in the illustration below.
Page 9 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
Illustration of the detection point locations (ASD sampling holes or
photoelectric point detectors) in the simulated computer rooms.
A PVC fire of a reasonably large size (90 kW) was modelled to ensure that
the simulated point detectors would reliably detect smoke.
The airconditioning was closed loop, that is, no clean air was introduced.
The return air vent was on the wall and the supply vents on the floor.
Three common point detector sensitivities were modelled: 1%Obs/m
(0.3%Obs/ft), 3.8%Obs/m (1.2%Obs/ft) and 9%Obs/m (2.3%Obs/ft).
The Data Collected
For each of the room sizes, ceiling heights, detection point sensitivities,
fire locations and detection point locations, VESDA equivalent sensitivities
were determined as follows:
1. At the exact moment that the second of the simulated coincidence
detectors issued an alarm, the smoke obscuration level at every
detection point was recorded.
2. These values were summed and divided by the number of detection
points to account for the cumulative effect and obtain an average
obscuration for the entire room.
3. The average obscuration level for all detection points was taken as
that which the VESDA alarm threshold must be set at to provide
equivalency to point detectors.
This study generated more data than it is practical to show here. However,
we have included two examples below; one for ceiling mounted and one
for floor void detectors.
VESDA equivalent sensitivity results for ceiling mounted
detectors in a 10.75 ft high, medium size computer room
VESDA equivalent sensitivity results for floor void detectors in
a 10.75 ft high, medium size computer room
Page 10 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
The above curves provide an estimate of the VESDA alarm sensitivities
which would be required, in a room with the particular parameters stated,
to detect smoke at the obscuration level where suppression release
becomes necessary. These curves do not account for variables such as
scrubbing rate (removal of smoke by filters) or clean air makeup
(percentage of clean air added before air is returned), since our results
indicated that these factors would not have any affect.
The entire set of ASD obscuration level curves produced was used to
generate the ASAT Calculator software tool which can be used to
calculate the appropriate ASD alarm sensitivities for suppression release.
The sensitivities calculated being equivalent to those for point detectors
and being calculated to initiate suppression at the same stage of the fire.
More details about the ASAT Calculator, can be obtained from Xtralis
(refer to the contact details at the end of this paper).
Sensitivity values, generated by the ASAT Calculator, should be verified
using the ASAT Verification and Commissioning Smoke Test Procedure
(Document Number 12746).
Discussion and Conclusions
Smoke Dilution or Smoke
Cumulation?
In general, as air-flow rate increases, we would expect smoke in the room
to become more diluted. This expectation is reflected in the NFPA 72[4]
requirement for more closely spaced detectors as air change rate
increases. Obviously, reducing detector separation results in a need for a
larger number of detectors. In the case of a VESDA system, this means
additional sampling holes which can cause one of the two effects
described below:
•
•
Dilution – smoke drawn into the VESDA pipe via sampling holes close
to the smoke source may be diluted by clean air entering sampling
holes further away.
Cumulation – a smoke build up may occur in the VESDA pipe due to
the fact that the reduction of the distance between sampling holes
brings more sampling holes closer to the smoke source which, in turn,
increases the number of sampling holes drawing in smoke.
Which of the above effects occurs, is highly dependent on the geometry of
the room. Air movement patterns relative to detector location will also
influence the behaviour of smoke near the sampling holes.
The results for the ceiling mounted detector illustrate the effect of dilution;
the obscuration level needed to cause an alarm decreases as air change
rate increases, that is, an increase in VESDA sensitivity is required to
compensate for smoke dilution.
Conversely, the results for the floor void detector indicate cumulation; the
obscuration level needed to cause an alarm increases as air change rate
increases, that is, a decrease in VESDA sensitivity is required to
compensate for the build up of smoke in the VESDA sampling pipes.
The exact time to detection is impossible to estimate accurately as it will
be affected by room geometry, air-flow, fire location, dilution or cumulation
etc.
Page 11 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
The Advantage of the
Cumulative Effect
In a high air-flow environment, smoke distribution within that environment
becomes homogenous (the same concentration at all points in the
enclosure) more quickly than in a still air environment. It is this fact that
gives ASD devices the detection edge over point detectors. Point
detectors detect locally, that is, at the exact point at which they are
located. If a point detector is of a particular sensitivity 9.8%Obs/m
(3%Obs/ft), for example, but the homogenous smoke concentration is less
than this value 4.9%Obs/m (1.5%Obs/ft), for example, no alarm will be
issued. No matter how many additional detectors are added, none will
issue an alarm.
Conversely, if we consider a 10 sampling hole ASD system with an alarm
setting at the equivalent sensitivity of one tenth of 9.8%Obs/m (3%Obs/ft),
that is, 1%Obs/m (0.3%Obs/ft), the result is quite different. Since ASD
devices draw in air and smoke at several sampling holes (the cumulative
sampling affect), the 4.9%Obs/m (1.5%Obs/ft) homogenous smoke
concentration, being higher than the ASD threshold of 1%Obs/m
(0.3Obs/ft), will cause an alarm.
The National Fire Protection Association (NFPA) in the United States has
produced the NFPA 72[4] and 76[5] code to promote industry best practice
and performance-based design methodology for the detection of smoke in
high air-flow Telecommunications environments. It is commonly believed,
around the world, that the higher the air change rate the closer together
should detection points be placed. For example, the NFPA 72[4] code
specifies maximum areas of coverage for a single point detector, and
maximum permissible point detector spacing at various air change rates
as shown below.
NFPA chart showing the regulation maximum
areas of coverage per point detector at various
air change rates in a room (excluding under-floor
or above-ceiling spaces).
ASD sampling holes are generally treated in the same manner as point
detectors, however the NFPA 72 code allows a relaxation of the above
spacing requirement for “Air-sampling …smoke detectors installed in
accordance with the manufacturer’s documented instructions”[5].
This is in part recognition of the fact that ASD has an inherent advantage
over point detectors due to the cumulative effect which compensates for
smoke dilution.
Page 12 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
Where smoke is homogenously distributed, installing larger numbers of
point detectors placed closer together will not increase the chance of
detection. Only an increase in smoke concentration, above the point
detector sensitivity, will do so. For this reason, we do not recommend any
alteration in VESDA sampling hole spacing to compensate for air change
rate, unless required by local codes and standards. More information on
designing VESDA systems in high air-flow environments can be found in
our “Telecommunications and Data Processing Facilities Design Guide”
(Document Number 11782).
In cases where smoke is not homogenously distributed throughout an
area, statistically, there is a higher probability of detection with more point
detectors located closer together.
Good design practice can assist in overcoming any anticipated detection
difficulties. Placing sampling holes above high risk equipment, for
example, or assessing the likely air-flow paths and how this will impact on
smoke distribution and/or dilution will all improve detection performance.
Even where there is only a small amount of smoke present, and this
smoke is being diluted by fresh air entering other sampling holes on the
VESDA sampling pipe, the high sensitivity of the VESDA detector will still
allow it to detect the diluted smoke.
ASD Systems For
Very Early Warning
Smoke Detection
As we determined earlier, in datacom facilities, we need the earliest
warning smoke detection possible to minimise business disruption,
equipment damage, risk to staff safety and overall financial loss. There
are several very good reasons for using high quality ASD devices such as
our VESDA detectors:
•
•
•
•
•
•
Their high sensitivity, 0.005%Obs/m (0.0015%Obs/ft), allows them to
detect minute amounts of smoke. VESDA detectors can detect the
initial smouldering stage of the combustion of a short length of wire,
providing very early warning and the opportunity to investigate.
Photoelectric point detectors are unlikely to detect such small
amounts of smoke under most practical conditions.
The wide sensitivity range of VESDA detectors allow them to be used
for very early warning as well as to initiate suppression at the very
high obscuration levels which would saturate point detectors.
The pre-programmable multiple alarm thresholds (Alert, Action,
Fire 1 and Fire 2) allow investigation and response activities
appropriate to the level of risk long before the release of suppression
would even be a consideration.
Their proven reliability ensures that there are no false alarms leading
to unnecessary suppression release.
VESDA sampling pipes can be placed on the ceiling, under the floor
void or at the return air vent which provides the maximum amount of
protection under all conditions (airconditioning on or off) and greatly
increases the potential for reliably detecting a fire event very early.
The maintenance of VESDA detectors protecting difficult locations is
easier than for point detectors in the same locations. This is because
only the VESDA sampling pipes need be mounted in inaccessible
places; the detector itself being positioned somewhere more
convenient.
Page 13 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
For the fire sizes modelled, in a small room with a low ceiling, an initial
alarm could occur as soon as 20 seconds after ignition, when the fire is
only 1.2 kW; the heat given out by a fire of this size is approximately the
same as that produced by an electric kettle. For large areas with high
ceilings, detection is more likely to be somewhere in the 30 to 50 seconds
range when the fire is between 5 and 7 kW.
The benefits of having time to investigate, thereby avoiding any
unnecessary use of suppression, can only be realised if the earliest
possible warning is provided by the detection system and if that system is
capable of dealing with the smoke dilution caused by air movement.
ASD System Design For
Suppression Release
Following the study, six coincidence detection options using ASD systems
were developed and tested. These provide maximum flexibility for
designers of fire protection systems:
1. In larger installations, where more than one VESDA unit is required for
smoke detection, it may be possible to arrange the pipe networks of
these detectors such that they can provide coincidence detection. An
example of an arrangement where VESDA sampling pipes are
alternated, in a manner which maintains the required sampling hole
spacing while providing two independent detectors for coincidence
detection, is illustrated below. This method requires the installation of
only one technology, which drastically reduces both installation and
maintenance costs.
VESDA Detector 1
VESDA Detector 2
Example of a VESDA alternating sampling pipe network for
coincidence detection suppression actuation.
2. In cases where there are several VESDA detectors installed but it is not
appropriate to alternate the pipe networks, one VESDA detector can be
used to sample the exhausts of several other detectors. This
arrangement is also illustrated below and again requires a single
technology. The cost effectiveness here will vary from system to system
depending on the number of detectors to be installed and maintained.
Page 14 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
VESDA
Detector 1
VESDA
Detector 2
Coincidence
Detection
Cylinder
Coincidence
Detection Detector
(CDD)
Example of VESDA exhaust sampling for coincidence
detection suppression actuation.
3. If a particular area contains VESDA detectors on the ceiling and across
the return air vent of the HVAC system, these detectors can be used to
provide coincidence detection. This is also a very cost effective method
since the existing smoke detection equipment is being used for both
smoke detection and suppression actuation.
4. In cases where point detectors are already installed, they can be
combined with VESDA detectors to provide coincidence detection as
illustrated below. Using a VESDA system with point detectors
corresponding to each sampling hole, although more costly, would
provide an acceptable balance between the risk of false alarms and
very early warning smoke detection.
VESDA Detector
Example of dual technology, coincidence detection,
suppression actuation.
Page 15 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
5. In the case of a VESDA LaserSCANNER, the detector capable of
determining how much smoke is entering each of its four sampling
pipes, two adjacent sampling pipes can be used to provide a
coincidence detection. This option allows the fire protection equipment
to be used for suppression actuation as well as smoke detection; the
purchase of additional equipment being unnecessary.
6. For very small facilities where only one VESDA detector is required to
cover the protected area, it is possible to use two of the multiple alarm
levels to provide coincidence detection. Doing this avoids having to
purchase and maintain additional equipment for suppression actuation.
The ASAT Calculator can be used to set appropriate alarm thresholds for all
VESDA detectors involved. Naturally, the choice of which method to use will
depend upon a number of factors such as the risk profile. All local codes
and standards should also be consulted before making any decisions.
The illustrations presented below are examples of possible relationships
between the two alarm thresholds for a couple of the coincidence detection
options outlined above. In all cases, both alarms indicated by the ‘AND’
must be issued before suppression will be released.
Alternating VESDA sampling pipes option. Both detectors must reach
the specified suppression actuation threshold in this
coincidence detection scheme.
Conclusions
Single VESDA detector multiple alarm thresholds option.
The VESDA detector must reach an initial alarm
threshold followed by the specified alarm
threshold for suppression actuation.
Our study provided a convenient method for benchmarking the detection
equivalence of two totally different technologies.
The following conclusions can be drawn about the use of a VESDA system
for both the very early detection of smoke in high air-flow environments and
the control of suppression release:
•
•
•
The VESDA system is able to overcome the smoke detection
challenges presented by air movement and the scrubbing of filters.
The very early warning capability of the VESDA system allows it to
detect smoke early enough to allow time for investigation and action,
long before a fire event reaches the stage where it will disrupt business
continuity, destroy assets, endanger life or require a very costly
suppression release.
A VESDA system can also be designed to provide a reliable, cost
effective, suppression actuation mechanism using coincidence
detection.
Page 16 of 18
“What A Coincidence!” – Very Early Smoke Detection Together With Suppression Actuation
We expect that the ASAT Calculator will greatly assist fire system designers
in the future, by facilitating the combination of both very early warning
smoke detection and coincidence detection for suppression release with a
single smoke detection technology.
References
[1] FCC (1993) Network Reliability: A Report to the Nation.
[2] NFPA (2002) Recommended Practice for the Fire Protection of Telecommunications Facilities.
[3] The proceedings of the 2nd International Fire Protection Seminar Buenos Aires 1999 –Telecommunication
Facilities. (ref The effects of various levels of exposure to the corrosive by-products of plastics fires.)
[4] NFPA (2002) National Fire Alarm Code.
[5] NFPA (2002) National Fire Alarm Code 2-3.6.6.3.
Page 17 of 18
www.xtralis.com
The Americas +1 781 740 2223
Continental Europe +41 55 285 99 99
Asia +852 2297 2438
UK and the Middle East +44 1442 242 330
Australia and New Zealand +61 3 9936 7000
The contents of this document are provided on an “as is” basis. No representation or warranty (either express or implied) is made as to the
completeness, accuracy or reliability of the contents of this document. The manufacturer reserves the right to change designs or specifications
without obligation and without further notice. Except as otherwise provided, all warranties, express or implied, including without limitation any
implied warranties of merchantability and fitness for a particular purpose are expressly excluded.
This document includes registered and unregistered trademarks. All trademarks displayed are the trademarks of their respective owners. Your use
of this document does not constitute or create a licence or any other right to use the name and/or trademark and/or label.
This document is subject to copyright owned by Xtralis AG (“Xtralis”). You agree not to copy, communicate to the public, adapt, distribute, transfer,
sell, modify or publish any contents of this document without the express prior written consent of Xtralis.
Doc. 11843_05