6

Installation in Process Systems

6.1. Design Considerations with Respect to Other

System Components
When a flame arrester is installed in a system, consideration must be given
to how it may adversely affect other system components. For example,
when an end-of-line or in-line deflagration flame arrester is installed on
the vent line of a low-pressure storage tank (designed per API 650 or UL
Std. 142), it is very important to ensure that the flame arrester pressure
drop is low enough so that the tank design pressure is not exceeded. Otherwise, the tank will be overpressured and may be deformed or possibly
rupture, which could result in the release of hazardous chemicals. This may
require the selection and installation of a larger size flame arrester, or it
may also be necessary to oversize the vent nozzle (and any associated
piping) to reduce the pressure drop.
Consideration should also be given to the possibility that the flame
arrester may plug, which could produce a vacuum condition in a low-pressure tank when the tank is pumped out, and implode (collapse) the tank.
This may require the installation of a vacuum breaker or a pressurevacuum conservation valve. If the tank contents are flammable and admission of air may result in an ignitable mixture, it may be necessary to install
an inert gas blanketing system on the tank, actuated by a pressure controller, which would admit a sufficient flow of inerting gas when a vacuum condition is detected.
139

140

6. Installation in Process Systems

For detonation flame arresters that are installed in manifolded vent

systems, similar considerations as above apply to prevent overpressure or
vacuum damage of atmospheric pressure tanks.

6.2. Piping and Flame Arrester System Design Considerations

When in-line detonation flame arresters are installed in single or
manifolded vent systems, consideration must be given to the design of the
system (flame arrester and piping). The following design practices should
be considered:
1. Piping and flame arresters should be designed for the maximum
anticipated temperature and pressure.
2. Piping supports should be designed to withstand the maximum
anticipated pressure transients.
3. Proper materials of construction for the flame arrester and piping
must be selected to minimize corrosion, which if uncontrolled,
could possibly result in catastrophic failure and subsequent loss of
containment.
4. Piping should be designed to avoid pocketing and liquid accumulation. If pocketing cannot be avoided, proper drainage should be
provided at low points and the piping support system designed for
the additional weight of accumulated liquid.
5. If there is a possibility of accumulated liquid freezing, the flame
arrester and piping should be heat traced and insulated.
6. If a flame arrester is preceded by piping larger than the arrester
inlet nozzle, it is necessary to install a pipe of at least 120 pipe diameters long of the same size as the flame arrester inlet nozzle
upstream of the arrester (see Section 5.3.15).
Although the peak pressures associated with detonations are very high,
they have very short duration (typically 12 ms). During this time the structural loading seen by the piping is minimal since the natural period of the
piping components is not reached. However, it is important to ensure that
the detonation flame arrester is designed and tested (hydrostatic or pneumatic pressure tests) to withstand the maximum line pressure that can
occur. It is also critical to specify and install piping that will withstand the
detonation. At initially low pressures of a few psig, and depending on factors which include size and configuration of the piping system, reactivity of
the gas, strength and location of the ignition source, detonation velocities
can approach 2000 m/s, and incident pressure (pressure on the side-wall of
the pipe) can exceed 4000 psig. These incident pressures can be reflected

6.2. Piping and Flame Arrester System Design Considerations

141

from elbows or the ends of the piping system, or flat appurtenances perpendicular to the flow path of the gases and flame front in the piping, and
the reflected shock can approach 9000 psig. Because the extremely high
pressure portion of detonation loads are only for short periods of time,
they only affect a short section of the piping at a given time. Because of the
response of ductile materials to loads of short duration, designing a
manifolded piping system for the peak pressures of a detonation is not
required. Piping constructed of materials with a ductility of elongation
12% at rupture, a design tensile strength of 50,000 psig, and a design
pressure rating of 350 psig should be able to contain detonations in systems which are operating at a maximum of 6 psig, and at temperatures
between 20F and 400F (Henley 2000).
When a flame arrester is installed in a low pressure vapor collection
manifold that may see high pressures during venting associated with an
emergency shutdown condition (ESD), the flame arrester must be
designed so that the housing is able to withstand this ESD venting situation. The same requirement applies to the manifold piping.
Fittings, elbows, and tees should be minimized since these affect flame
propagation via the influence of flow turbulence and shock reflection.
Experimental work by White and Oswald (1992) showed that elbows and
fittings increase the flame arrester loading and may affect the arrester
flame quenching capability. The capability of an end-of-line and in-line
deflagration arrester will be exceeded, but a detonation arrester usually
will be unaffected. Aside from the effect that the shock wave has on a flame
arrester, the reaction of the pipe itself should be accounted for in the
design of pipeline supports to withstand the bending-related forces and
moments that are produced.
Piping manifolds should be designed to avoid pocketing and accumulation of transported and condensed liquids. They should be sloped so that
the piping is inclined towards the low point. Some companies slope the
piping toward a vessel in the system, or if this is not possible, then heat tracing is provided to prevent condensation from occurring at the low point.
Low points in the piping should be provided with a means of drainage. It is
very important to remove any accumulated liquid that can become a slug
and impact on the flame arrester face if a deflagration or detonation
occurs. Such impact can damage or plug the arresting element. Manual
drainage can be considered if it is addressed by a formal procedure. Alternatively, it may be desirable to provide an automatic drain system at the
low point to ensure the prompt draining of accumulated liquid.
For installations in cold climates, as well as applications in which product freezing, polymerization, or crystallization is a concern, the flame
arrester and piping should be insulated and, and where required, heat

142

6. Installation in Process Systems

traced. Where polymerization is a concern, the temperature of the tracing

should be kept as low as possible. It is important to ensure that the temperature of the heat tracing be kept below the accepted operating range of the
flame arrester. This can be achieved by using tempered water (often with
glycol added), or low pressure steam. Electrical tracing with a temperature
control system or self-limiting electric tracing is also often used.

6.3. Maintaining Reliability

Maintaining flame arrester reliability is critical to successful operation. Modifications to the system which may change gas composition should be carefully evaluated to ensure that the mixture with the lowest MESG is still within
flame arrester certification. Other management of change issues include
piping modifications, increased or decreased pressures, and the introduction of corrosive, condensable, or other potentially harmful materials.
To ensure reliability, dry-type flame arresters should be inspected on a
regular basis, as mentioned previously. Also, as discussed in Section 5.3.18,
it is often the practice to provide pressure drop and temperature instrumentation to monitor the flame arrester condition in service. This will provide indication of impending problems that can affect the performance of
the flame arrester.
Hydraulic (liquid seal) flame arresters require attention on a regular
basis similar to that for dry-type flame arresters. It is critical to ensure that
the liquid level is at the required height, and level instrumentation with
alarms is recommended. Automatic addition of makeup liquid (to replace
evaporation and entrainment losses) is desirable. Temperature instrumentation is also recommended to monitor the occurrence of sustained burning (a stabilized flame). See Section 5.3.18 for more details.

6.4. Optimum Location in System

A number of factors should be considered when determining the optimum
location of a flame arrester, such as:
ease of inspection and maintenance
minimization of run-up length to avoid high flame speed and failure
of the flame arrester (flame breakthrough)
emission of vent vapors which could impact on personnel
As mentioned in Section 5.3.16, one of the primary criteria for flame
arrester location is that it be installed in a location that affords easy access

6.5. Supports for Static and Dynamic Forces

143

for inspection and maintenance, including ease of removal. End-of-line

flame arresters that are installed on outdoor tanks or process vessels
should be mounted directly on the vent nozzle, and the vent nozzle should
preferably be located near the edge of the tank roof for easy access. If the
vent nozzle is not near the tank roof edge, the flame arrester should be an
in-line deflagration arrester so that the emitted vapors do not impinge on
the tank roof. For tanks or process vessels located inside buildings, the vent
line should be routed above the roof, and an end-of-line flame arrester
installed at the end of the vent pipe. In-line detonation flame arresters in
manifolded vent systems may be mounted as needed for accessibility subject to piping constraints.
In-line deflagration flame arresters must have piping on the unprotected side as short as possible to minimize run-up length to avoid the
potential for significant flame acceleration and possible DDT in the pipe.
The run-up length is affected by pipe diameter, and the number and types
of fittings, and not just piping straight length. The maximum allowable
run-up length is strictly limited by the recommendations of the vendor.
For Group D gases, the distance between the flame arrester and the end of
the discharge pipe that vents directly into the atmosphere is sometimes no
more than 20 feet. For gases in other NEC groups the allowable run-up
length may be appreciably less, and it is strongly recommended that the
flame arrester manufacturer be consulted for the maximum run-up
lengths for their specific flame arresters. The allowed distance must be
established by proper testing with the appropriate gas mixture and the
actual pipe diameter to be used. Turbulence-promoting irregularities in
the flow (bends, tees, elbows, valves, etc.) cannot be used unless testing has
addressed the exact geometry.
The flame arrester discharge point should be located at a minimum
height above grade, building roof, or an operating level, and pointed away
from areas where personnel can be impacted by the vapors or flame. Generally, a height of seven to ten feet above any level where personnel can be
exposed is sufficient. Nothing should be above the discharge point such as
a platform, roof overhang, piperack, etc.

6.5. Supports for Static and Dynamic Forces

Vessel nozzles and piping connected to flame arresters must be designed to
provide adequate support for static and dynamic forces. Static forces are
primarily due to the weight of the arrester. Large end-of-line and in-line
flame arresters can be quite heavy, so vessel nozzles must be strong enough
to carry the weight of the arrester and to resist any bending moments. It

144

6. Installation in Process Systems

may be necessary to reinforce the nozzle with gussets, or an independent

support structure may be required. Some large in-line flame arresters may
have to be installed horizontally, and the arrester and attached piping may
need support to carry the weight and remove stresses on the piping.
Nichols (1999) describes a number of reaction forces that are generated during deflagrations and detonations in piping systems such as:
1. A radial force on the pipe wall ahead of the deflagration wave. There is a
varying pressure between the acoustic wave and the flame front
where the pressure builds from near atmospheric pressure, P1 (step
change at the wave front) to eight times P1 (or higher) at the flame
front. The pressure ratios depend on the flame acceleration. There
is no such effect with a detonation.
2. A local radial force on the pipe wall at the flame front where there is a step
change in pressure and hence radial stress in the wall. This pressure rises
to about 8 times P1 for a deflagration and about 20 times P1 for a
detonation. However, since the deflagration velocity is subsonic, the
lower stress acts for a greater time compared to the higher stress for
a detonation, and may or may not give greater potential pipe material damage.
3. A radial force on the pipe wall behind the flame front. For a closed pipe,
this remains constant (at about 8P1) everywhere for a deflagration,
but decays from about 20P1 to 8P1 behind the detonation over most
of the pipe length.
4. A pressure wave at bends which gives a longitudinal force acting to move the
pipe assembly on the pipe hangers. For a deflagration this reaction force
is the pressure change times the pipe area. For a detonation there is
an additional momentum effect which doubles the apparent pressure to about 40P1. The duration of the force depends on the pressure rise and decay profile, but is usually below the pipe assembly
resonant frequency.
5. Longitudinal out-of-balance forces. For a Z-configuration pipe layout
there is a steady force between the two bends generated by the pressure difference in the gas while the pressure front moves between
the two bends. This gives a longer duration to the force which can
approach the resonant frequency of the pipe.
Detonation flame arresters and associated piping must be able to withstand the effects of explosion transients that include the pressure or pressure-related force, the specific impulse (integral of the time versus
overpressure) or net impulse (upstream minus downstream impulse) highspeed gas momentum transfer and flux, and temperature-, heat-, and thermal-flux related loads (White and Oswald 1992). White and Oswald con-

6.5. Supports for Static and Dynamic Forces

145

ducted a series of experiments to obtain information on the effect of

detonation flame arrester structural response to detonations traveling in a
long pipe system, as would be encountered in marine vapor control systems. The work involved a theoretical engineering analysis, testing with a
small-scale and large-scale pipe, and evaluation of the USCG test procedure. The small-scale tests were conducted in a 6-inch pipe using a USCGapproved detonation arrester and a simulated arrester (the arresting element was tightly wound wire cloth). The large-scale tests were conducted in
an 18-inch pipe with a simulated detonation arrester (the arresting element was a section packed with 1-inch carbon steel Pall rings). The test
gases were three concentrations of propane in air (4.3%, 6.5%, and 7.5%).
Based on these tests, the following were concluded:
l. Overdriven detonations, not long-pipe stable detonations, provide a
greater potential for mechanical damage to detonation flame
arresters.
2. For describing structural loading functions needed for design analysis, the use of overdriven detonation data representing the net
overpressure (run-up side pressure less protected side overpessure)
on the flame arrester element and supporting structure is preferable to data representing only the run-up side, side-on overpressure.
However, the run-up side transient history of side-on overpressure
for overdriven detonations should provide a conservative estimate
for design purposes.
3. For deflagrations, the dependence of overpressure and its effect on
a flame arrester do not scale in proportion to pipe length, pipe
diameter, or pipe length to diameter ratio.
4. There are no clearly discernable, broadly applicable, correlations
between the 6-inch and 18-inch deflagration and detonation experiments. Therefore, comparisons were done on a parameter-byparameter basis. However, comparisons of data taken during
experiments with the two pipe sizes reveal that enough scalerelated differences exist that interpolation between the two scales
for an intermediate size should be done only where conditions are
very similar. Then, overpressure and specific impulse can be estimated based on L/D.
5. Venting an explosion ahead of a flame arrester can reduce the thermal flux and the impulse to which the arrester is subjected. Test
results indicate that peak side-on overpressure is halved, specific
impulse is reduced by a factor of three, and the temperature is substantially reduced. However, overpressure and flame speed at the
flame arrester do not appear to be changed significantly.

146

6. Installation in Process Systems

6. Based on testing and thermal modeling, the potential for reignition

on the protected side of an approved flame arrester under detonation conditions is minimal. A detonation flame arrester that is subjected to a sustained, stable detonation through a very long pipe will
receive a higher thermal flux than it would receive if subjected to an
overdriven detonation. However, thermal modeling indicates that
asymptotic conditions are obtained for some length of pipe which,
if longer, will yield no significant increase in thermal flux.
White and Oswald present two computer programs in the report,
which can be used to obtain estimates of the structural response and the
thermal response of a detonation flame arrester. However, they point out
that neither of the programs will reveal completely what is needed to
design a successful flame arrester support. The authors also evaluated the
USCG test procedure and present recommendations on how this test procedure should be modified to improve it.

6.6. References
6.6.1. Regulations, Codes of Practice, and Industry Standards
HSE (Health and Safety Executive). 1980. Guide for Flame Arresters and Explosion
Reliefs. Booklet HS(G)11. Health and Safety Executive, Her Majestys Stationery Office, London, England.

6.5.2. Specific References

Henley, R. 2000. Draft of proposed new NFPA Guide for Mitigation of Explosion
Hazards in Manifold Piping Systems.
Howard, W. B. 1992. Use Precautions in Selection, Installation, and Operation of
Flames Arresters. Chemical Engineering Progress, 88(4), 6975.
Nichols, F. 1999. Personal communication from Fred Nichols (retired from ICI),
Aston, Chester, England to S. S. Grossel (September 7, 1999).
White, R. E. and Oswald, C. J. 1992. Mitigation of Explosion Hazards of Marine
Vapor Control Systems. Southwest Research Institute, Final Report, SWRI
Project No. 06-4116, for the American Petroleum Institute, Washington, DC
(October 1992).