ISO/TR 13086-2:2017

TECHNICAL ISO/TR
REPORT 13086-2
First edition
2017-12
Gas cylinders — Guidance for design
of composite cylinders —
Part 2:
Bonfire test issues
Bouteilles à gaz — Recommandations pour la conception des
bouteilles en matière composite —
Partie 2: Aspects concernant les essais à la flamme vive
Reference number
©
ISO 2017
© ISO 2017, Published in Switzerland
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized otherwise in any form
or by any means, electronic or mechanical, including photocopying, or posting on the internet or an intranet, without prior
written permission. Permission can be requested from either ISO at the address below or ISO’s member body in the country of
the requester.
ISO copyright office
Ch. de Blandonnet 8 • CP 401
CH-1214 Vernier, Geneva, Switzerland
Tel. +41 22 749 01 11
Fax +41 22 749 09 47
copyright@iso.org
www.iso.org
ii © ISO 2017 – All rights reserved

Contents Page
Foreword .iv
Introduction .v
1 Scope . 1
2 Normative references . 1
3 Terms and definitions . 1
4 Background . 1
5 Statement of safety . 2
6 Components of fire testing . 2
6.1 Composite materials . 2
6.2 Fire . . 3
6.2.1 General. 3
6.2.2 Fire tests in standards . 5
6.2.3 Standardized fire test . 7
6.2.4 Considerations for future standardized fire tests . 8
6.3 Pressure relief devices. 9
6.4 Venting .11
6.5 Interaction.12
6.6 Availability of reports .16
6.7 Optimized test method using thermally activated pressure relief devices .17
6.7.1 Explanation of optimized test method .17
6.7.2 Procedures for optimized test method .20
7 Summary .23
Annex A (informative) Comparison of fire tests in standards and reports .24
Annex B (informative) Standardized test requirements using thermally active pressure
relief devices .28
Bibliography .33
Foreword
ISO (the International Organization for Standardization) is a worldwide federation of national standards
bodies (ISO member bodies). The work of preparing International Standards is normally carried out
through ISO technical committees. Each member body interested in a subject for which a technical
committee has been established has the right to be represented on that committee. International
organizations, governmental and non-governmental, in liaison with ISO, also take part in the work.
ISO collaborates closely with the International Electrotechnical Commission (IEC) on all matters of
electrotechnical standardization.
The procedures used to develop this document and those intended for its further maintenance are
described in the ISO/IEC Directives, Part 1. In particular the different approval criteria needed for the
different types of ISO documents should be noted. This document was drafted in accordance with the
editorial rules of the ISO/IEC Directives, Part 2 (see www.iso.org/directives).
Attention is drawn to the possibility that some of the elements of this document may be the subject of
patent rights. ISO shall not be held responsible for identifying any or all such patent rights. Details of
any patent rights identified during the development of the document will be in the Introduction and/or
on the ISO list of patent declarations received (see www.iso.org/patents).
Any trade name used in this document is information given for the convenience of users and does not
constitute an endorsement.
For an explanation on the voluntary nature of standards, the meaning of ISO specific terms and
expressions related to conformity assessment, as well as information about ISO's adherence to the
World Trade Organization (WTO) principles in the Technical Barriers to Trade (TBT) see the following
URL: www.iso.org/iso/foreword.html.
This document was prepared by Technical Committee ISO/TC 58, Gas cylinders, Subcommittee SC 3,
Cylinder design.
iv © ISO 2017 – All rights reserved

Introduction
Composite reinforced cylinders have been used in commercial service for about 40 years. Common
fibres used in composite cylinders include glass, aramid, and carbon. Resin matrix materials are
commonly epoxy or vinyl ester.
Composite cylinders are known to be exposed to the action of fire, ranging from radiant heating to
full engulfment in the fire. Cylinder performance during exposure to fire might depend on the cylinder
materials of construction, size of the fire, dimensions of the cylinder, its orientation, its contents, and
the use of temperature or pressure activated relief devices.
Fire exposure tests are often included in composite cylinder standards, sometimes as a mandatory
test and sometimes as an optional test. This document addresses issues related to composite cylinders
exposed to fire, summarizes test requirements, and offers a new approach to qualifying cylinders with
relief devices.
TECHNICAL REPORT ISO/TR 13086-2:2017(E)
Gas cylinders — Guidance for design of composite
cylinders —
Part 2:
Bonfire test issues
1 Scope
This document addresses the topic of safety and performance of composite cylinders in a fire situation.
A statement of safety addresses the topics which should be understood in order to operate cylinders
safely in service. The remainder of this document provides a basic level of understanding of these topics.
2 Normative references
There are no normative references in this document.
3 Terms and definitions
No terms and definitions are listed in this document.
ISO and IEC maintain terminological databases for use in standardization at the following addresses:
— IEC Electropedia: available at http://www.electropedia.org/
— ISO Online browsing platform: available at https://www.iso.org/obp
4 Background
Composite cylinders began service in the 1950s, initially as rocket motor cases with glass fibre
reinforcement. This led shortly to glass fibre pressure vessels with rubber liners, and then to glass fibre
pressure vessels with metal liners. Metal liners were typically either aluminium or steel. Eventually,
new structural fibres, such as aramid and carbon, came into use for reinforcing pressure vessels. Today,
typical reinforcements are glass and carbon, either individually or together as a hybrid. Typical liner
materials are steel, aluminium, or polymers, often high density polyethylene (HDPE) or a polyamide (PA).
Composite cylinders offer certain advantages, particularly light weight and corrosion resistance.
However, there are some performance requirements that tax the abilities of composite cylinders. One
of these is the ability to withstand exposure to fire conditions without rupture. Fire conditions might
include both direct exposure to fire, and to the elevated temperatures resulting from a fire. Direct
exposure might include localized flames, or an engulfing fire.
Sources for a fire could include discharge of flammable gases from nearby cylinders, spilled liquid fuel
from motor vehicles, car fires, house or building fires, and grass or forest fires, to name a few. There is
significant variation in the fire conditions that arise from each of these causes, and there are issues on
reproducibility of any of these types of fires.
Composite cylinders might be able to withstand a certain level of fire exposure on their own. However,
it is more common in certain applications to use a system approach that could include isolation from
fire, insulation, pressure activated relief valves or devices, and/or thermally activated relief devices.
However, there might be conditions where the risk of rupture is less than the risk and consequence of
leakage, and a pressure relief device (PRD) or similar device would not be used. Individual cylinders
might be tested without any type of protection, but it is also common for the cylinder to be tested as part
of a system that contains some means of protection. Regardless, the cylinder should be representative
of a production cylinder and the test should address hazards which might occur.
5 Statement of safety
Composite cylinders, and assemblies of composite cylinders, can be used safely in conditions where
there might be exposure to fire conditions if there is an:
— understanding of composite materials, including the liner;
— understanding of fires;
— understanding of PRDs, if used;
— understanding of insulation, if used;
— understanding of valves and their failure mechanisms;
— understanding of venting;
— understanding of single cylinder vs. multiple cylinder systems;
— understanding of interaction of the above elements;
— optimized test, which is developed, based on above understandings.
Clause 6 addresses the elements of the statement of safety, and provides some understanding for each
of the elements.
6 Components of fire testing
6.1 Composite materials
The reinforcement of a composite cylinder consists of reinforcing fibres in a resin matrix. There might
be resins or additives in the resin that affect structural or thermal performance. There might also be
external coatings that protect the composite, such as intumescents. When exposed to fire, intumescents
form a char layer which has low conductivity and protects the underlying material. There might also be
ablative layers, which could remove heat of the fire as they ablate.
Reinforcing fibres primarily include glass and carbon, and occasionally aramid. E-glass properties
[1]
on matweb.com show glass has a melting point of about 1 725 °C (3 137 °F), and therefore might
soften or melt in a bonfire test, where temperatures might reach 1 960 °C (3 500 °F) which is the flame
1)
® [2]
temperature of the combustion of natural gas in air. Kevlar properties on matweb.com show
aramid fibres begin to lose strength above 425 °C (797 °F), and might decompose and burn at 500 °C
(932 °F). Carbon fibre might oxidize in the fire and lose strength at temperatures of 600 °C (1 112 °F).
The onset of pyrolysis, affecting organic materials such as epoxy resin, can be as low as 300 °C (572 °F).
Resins are typically epoxy or vinyl ester. These materials might burn in a fire. The resins might contain
additives that are also attacked by fire, but some additives might be fire retardants.
A liner is generally used to prevent gas from leaking through the composite, and also serves as a
winding mandrel for the composite. The liner is typically steel, aluminium alloy, HDPE, or PA. Polymer
liners generally have a metallic end boss, either on one or both ends, centred on the longitudinal axis.
The composite reinforcement is wound in layers on top of the liner. The composite reinforcement
typically ranges from 3,2 mm to 50 mm (0,13 in to 2 in) thick and is dependent upon factors such as
vessel diameter, working pressure, and regulations. Curing of the laminate is achieved by cross-linking ®
1) Kevlar is an example of a suitable product available commercially. This information is given for the convenience
of users of this document and does not constitute an endorsement by ISO of this product.
2 © ISO 2017 – All rights reserved

of the resin, involving a combination of time and temperature. This time and temperature depends
on the resin materials and the thickness of the laminate. These layers might be of a single material,
multiple materials in a layer, or alternating layers of material. External insulation or protective layers
might be wound onto the cylinder.
The thinner the laminate, the faster the degradation of the laminate in the fire, and the faster the gas
inside is heated. A thicker laminate takes longer to fail, and longer to transfer heat through the wall. Due
to cylinder geometry and winding, the structural composite is generally thickest in the cylindrical section
and near the end bosses, and thinner in other sections of the pressure vessel such as the end domes.
Composite materials generally have a low thermal conductivity, and are often considered to be an
insulating layer. Degradation or failure of the composite laminate can occur from three causes. First,
the fibre might be directly reduced in strength by the fire. Second, the fire might burn resin out of the
laminate. When the resin is removed from the laminate, the load is not efficiently transferred from
inner layers to outer layers. This might cause the inner layers to be overstressed and fail. Third, there
might be heat from the fire that increases pressure within the cylinder, and decreases the strength of
the laminate, even though there is no direct flame contact. When directly exposed to a fire, charring
and burnout of the resin is the primary factor in degradation of the composite. It is also possible for
heat from the fire to be transferred to the liner, which in the case of polymer liners, might melt and
allow the cylinder contents to vent.
Testing is generally conducted on full scale cylinders filled with the gas that the cylinder has been
designed to contain and where appropriate equipped with the PRDs designed to protect the pressure
vessel from bursting. There might be times when a full diameter, short length cylinder would provide
an accurate measure of performance. It is generally accepted that similar sizes would have similar
performance in a fire test, so a fire test is not always required with a change of design. However, studies
have shown that time to burst in a fire is only the same if the composite materials, thickness, stresses,
and winding sequences are the same, so good judgment should be used when qualifying new designs.
6.2 Fire
6.2.1 General
Fire exposure in service can be from a number of sources, including diesel fuel, kerosene, gasoline,
propane, natural gas, hydrogen, tyres, wood, or other combustible materials. In some cases, fires involving
cylinders containing flammable materials release their contents, which then adds to the fire. The flame
temperature varies by fuel source, and can be affected by wind conditions and ambient temperatures.
The size of the fire and length of time in the fire would depend on the amount of fuel for the fire and its
distribution around the cylinder. In some cases, the cylinder might only be exposed to a heat flux, and not
directly to a fire. Different applications might have different fire sources and risk levels.
At one extreme, the fire might be focussed on a very small area (such as with a propane torch). In this
case, it might be unlikely that a PRD is activated. A localized fire might be caused by a small pool of
liquid fuel, a burning tyre, or an engine fire. In this case, it is more likely that a PRD would be activated,
but activation is not absolutely certain. If the relief device was not activated, a cylinder rupture would
be likely. An engulfing or global fire would involve exposure of the entire cylinder. Newhouse and
[3]
Webster indicate that, in this case, activation of the PRD is virtually assured. If the cylinder was
exposed to a heat flux, it might be at a temperature that would degrade the composite strength directly,
in which case, a thermally activated PRD would likely be activated and release the contents safely. If the
temperature was below the activation temperature of a thermally activated PRD, it might be possible
for a pressure activated PRD to be triggered.
Test methods have been developed that reflect fires that might actually occur in service. However, such
fires are not precise or particularly repeatable. Most standards use a somewhat localized fire, typically
limited in length to 1,65 m. Some standards are developing tests with a more localized fire that acts for
a given time, followed by exposure over a larger area. In some standards, localized fire is achieved by
testing the cylinder in a vertical orientation. Fires generally continue until the cylinder is vented or the
fuel is consumed. A fire would typically last 20 minutes or more. The length of time a fire burns, vs. the
time for a PRD to activate, or the cylinder to rupture, are of particular interest to first responders.
Fuels used in bonfire tests typically have included kerosene, diesel fuel, or kerosene soaked wood. More
recently, natural gas or propane has been used in burners that are oriented to cover a given portion
of the cylinder. Hydrogen/oxygen burners have also been used experimentally. The flow of gas can be
controlled to give a range of thermal input and temperature level. Thermal input to the cylinder includes
heat from convection and from radiation. Wind can affect real fires, and similarly those made with
wood, kerosene, or diesel fuel. There will be flames that “lick” the surface of the cylinder, essentially
moved back and forth by the wind. If natural gas or propane burners are used, the fire should be more
controlled than would be the case in actual field service, that is, control the area of the cylinder in the
fire is better achieved, including the height of flames, the evenness of the flames, and the total heat flux,
and thereby a more consistent test is achieved. In some cases, these gaseous fuelled fires are conducted
in tubes or sheltered areas, such that wind is not a factor.
[4]
The combustion temperatures of the different materials, as reported by Murphy are given in Table 1.
Table 1 — Adiabatic flame temperature (burning a stoichiometric mixture of fuel)
Fuel °C °F
Gasoline 1 977 3 591
Diesel 2 054 3 729
Natural gas 1 884 3 423
Propane gas 1 990 3 614
Hydrogen gas 2 115 3 839
The effective temperature would be lower in the case that wind pushes the flame around on the
surface of the cylinder. Figure 1 shows how measured diesel fuel flame temperature varies during one
particular test.
Key
X minutes
Y degrees °F
Y´ PSIG
Figure 1 — Measured flame temperature vs. time
4 © ISO 2017 – All rights reserved

One advantage of using solid or liquid fuels is that the test set-up can be accomplished without too much
set-up equipment being required. Also, fires in actual service would more likely be solid or liquid fuels,
making this a more valid test. One advantage of using gaseous fuel is that the test is more repeatable,
making the test more reliable as a predictor of performance in the field. If a cylinder ruptured during
the test, there would be no need for environmental cleanup, while solid or liquid fuels might be spread
around the test area and contaminate the soil.
6.2.2 Fire tests in standards
Table 2 lists some common standards that address fire testing, and how they define the fire. Further
detail from these standards is provided in Annex A. The fire test is designed to demonstrate that
finished cylinders, complete with the fire protection system (cylinder valve, PRDs and/or integral
thermal insulation) specified in the design, prevents the rupture of the cylinder when tested under the
specified fire conditions.
Table 2 — Examples of standards or reports that address fire testing
[5] [6]
Item Requirements/reference ISO 11119 ISO 11439 EN ANSI DOT DOT RPT
[7] [8] [9]
12245 NGV2 FRP1 HS 811
[10]
Application Transportable cylinder X X X
Vehicle fuel container X X X
Fire source Any fuel with uniform, sufficient
heat to maintain temperature
X X X
Wood or kerosene X
Kerosene-soaked wood,
X
gasoline, or JP-4 fuel
LPG (propane)
X
Fire size 1,65 m length × cylinder
X X X
diameter
Total engulfment X X
250 mm length × cylinder
X
diameter
Fire ≥ 590 °C on cylinder surface
X X
temperature
≥ 590 °C within 25 mm of
X
cylinder surface
≥ 430 °C within 25 mm of
X
cylinder surface
900 °C to 950 °C on cylinder
X
surface
Not specified X
Cylinder Horizontal
X X X X
orientation
Horizontal and vertical X X
Cylinder con- Intended contents; or air or
X X
tents nitrogen
Natural gas (or methane)
X X
Air or nitrogen X
Hydrogen X
Fire protection Required
X X X X
device
Table 2 (continued)
[5] [6]
Item Requirements/reference ISO 11119 ISO 11439 EN ANSI DOT DOT RPT
[7] [8] [9]
12245 NGV2 FRP1 HS 811
[10]
Not required X X
Pressure Service X X X X X X
And 25 % of service if a
thermally activated pressure
X X
relief device (TPRD) not used
Test time with Until vented
X X X X X
protection
30 minutes or until vented
X
Test time with Two minutes
X X
no PRD
The standards listed above have a similar approach to fire testing. However, the case where multiple
cylinders are being vented by a single PRD is generally not well covered in most standards. Standards
might be updated to recognize the increased time to vent multiple cylinders, even if only one cylinder is
actually in a fire. Standards being developed for hydrogen vehicle fuel containers, including SAE J2579,
and the UN Global Technical Regulations for hydrogen fuel cell vehicles – ECE/TRANS/180/Add.13,
have taken a modified approach to the fire test, beginning with a smaller, localized fire for the first 10
[11]
minutes of the test, then progressing to an engulfing fire, as reported by Scheffler . Figure 2 shows
the development of the fire from local to global.
Key
X minutes 3 localized area
Y minimum temperature 4 engulfing region outside localized area
(burner ramp rate)
1 localized fire exposure
2 engulfing fire 5 ignite main burner
Figure 2 — Fire Test approach for hydrogen fuel containers
6 © ISO 2017 – All rights reserved

6.2.3 Standardized fire test
From these standards that address fire testing, it follows that a common approach is currently:
1) Cylinder preparation for fire test:
a) The cylinder is pressurized with nitrogen or the intended gas at working pressure. Air is not
recommended as a test gas. A test at 25 % of working pressure is recommended if the cylinder
is protected by a pressure activated relief device, but is not required if a thermally activated
relief device is used. Tests may be conducted at ambient temperatures between −7 °C and 43 °C
(20 °F and 110 °F), but the container settled pressure is to be temperature compensated to
15 °C (60 °F). Some standards limit wind to 2,25 m/s to avoid problems with the flame.
b) Metallic shielding of a minimum 0,4 mm (0,016 in) thickness is used to prevent direct flame
impingement on cylinder valves, fittings, and/or PRDs. The metallic shielding is not to be in
direct contact with the specified fire protection system (PRDs or cylinder valve).
NOTE Some consideration has been given to removal of the shielding requirement, provided that fire is
not directed onto the PRDs.
c) Thermocouple location: At least three thermocouples are placed along the bottom of the
cylinder under test not more than 0,75 m apart and within 25 mm (1 in) of the surface. Each
thermocouple may be attached to a steel cube up to 25 mm (1 in) on a side. Thermocouples are
protected from direct flame impingement by attaching to steel cubes as mentioned above or by
metallic shielding of a minimum of 0,4 mm (0,016 in) thickness.
NOTE The purpose of the 25 mm steel cube is to average out temperatures over time, i.e. to average
over the time period while alternately being directly in the flame vs. in air.
d) Cylinder internal pressure is monitored.
2) Cylinder orientation
a) Cylinders that are mounted vertically may be tested in a vertical orientation.
b) Cylinders less than or equal to 1,65 m (65 in) long should be horizontally oriented with a fire
source below the centre of the long axis of the cylinder for the entire length of the cylinder.
c) For cylinders of length greater than 1,65 m, the cylinder is to be positioned in accordance with
the following procedure.
i) If the cylinder is fitted with a PRD at one end, the fire source is to commence at the opposite
end of the cylinder.
ii) If the cylinder is fitted with PRDs at both ends, or at more than one location along the
length of the cylinder, the centre of the fire source is to be centred midway between the
pressure relief devices that are separated by the greatest horizontal distance.
3) Fire sources: The fire source can be any fuel provided it supplies uniform heat sufficient to maintain
the specified temperature until the cylinder has vented. Gaseous fuel sources are recommended
based on ease of control and avoidance of environmental contamination in the event of cylinder
rupture. A uniform fire source of 1,65 m (65 in) length provides direct flame impingement on the
container surface across its entire diameter. The bottom surface of the cylinder is to be about
0,10 m above the base of the fire.
4) Fire test requirements: Thermocouples and internal cylinder pressure are to be recorded at least
every 30 seconds for the duration of the fire test. Fire test temperature is to be ≥ 590 °C (1 094 °F)
as measured using at least three thermocouples. Within five minutes following fire ignition the
temperature of two of the three thermocouples are to be ≥ 590 °C (1 094 °F) and two of the three
thermocouples are to be ≥ 590 °C (1 094 °F) throughout the duration of the test. It is noted that
the temperature is generally the only measure of the fire intensity; however, heat flux is also
important as a measure of fire intensity, and consideration should be given to its use, particularly
as consideration to localized vs. engulfing fires. However, specification of the fire source, in addition
to fire size, does give some measure of control over heat flux.
5) Fire test success: A successful fire test is one in which the cylinder vents its contents down to a
pressure less than 7 bar through a PRD without bursting. In the event that complete venting occurs
in less than five minutes, the minimum temperature requirements do not apply and the fire test is a
success.
NOTE Some standards have not required venting, but instead require a fixed number of minutes in the
fire without rupture. The number of minutes has ranged from two minutes to 20 minutes.
6) Fire test failure: If the cylinder bursts the fire test is a failure in the case where a PRD is used. If
the cylinder bursts before the specified time limit the fire test is a failure in the case where a PRD
is not used. Any failure during the test of a valve, fitting or tubing that is not part of the intended
protection system design invalidates the result.
7) A single test is generally considered adequate for qualification, based in part on prior testing
of similar cylinders. If a non-typical cylinder or PRD is used, consideration should be given to
additional tests for characterization.
8) Design changes: It is generally accepted that a test should be conducted if a design change is likely
to change the results of the test. For a bonfire test, such design changes include:
a) Significant change in internal volume. This includes an increase in length of 20 % or more, or a
change in diameter of 20 % or more.
b) Significant change in pressure. This includes an increase in pressure of 20 % or more.
c) If a new PRD is used.
d) If the fibre type is changed. This includes changes between carbon, aramid, and glass fibre, but
generally not for changes in manufacturer for the same type of fibre.
e) If the liner material is changed. This includes changes between steel, aluminium, and polymer
liners, but generally not for changes in supplier.
6.2.4 Considerations for future standardized fire tests
From recent developments, however, the following test approach issues should be considered:
— should the fire be localized, engulfing, or some combination based on application;
— should the size of the fire be tied to the size of the cylinder;
— should the cylinder be tested by itself, or in a package representative of the actual application.
[10]
Webster, 2010 , reported:
The design of compressed hydrogen fuel systems for hydrogen vehicles has been largely based on the
compressed natural gas (CNG) vehicle experience. In addition to installation requirements, the test
procedures used to qualify on-board hydrogen fuel systems for service use were based upon the test
protocol developed for the CNG industry.
Between 2000 and 2008, there have been over 20 failures of CNG cylinders onboard vehicles. The
single largest cause of these failures (over 50 %) was fire. These CNG cylinder failures have occurred
on OEM passenger vehicles (Ford Crown Victoria, Honda Civic), as well as on OEM transit buses
(Heuliez, MAN Bus).
Note that the effect of localized fires is more pronounced on cylinders of longer length, as Thermally
activated Pressure Relief Device (TPRD) locations are typically spaced far apart. Some of the fire
failures could be attributed to slow reacting TPRD designs, but the majority of the failures were caused
8 © ISO 2017 – All rights reserved

by localized fire effects where the flame exposure was at a location on the cylinder remote from the
TPRD location.
Note that TPRDs do not tend to activate unless they are sufficiently exposed to a high heat source, or
direct flame impingement.
All CNG or draft compressed hydrogen cylinder standards worldwide only specify a bonfire test of a
cylinder where the fire source is a standard 1,65 m length. This fire length is derived from a US DOT fire
test developed in the 1970s for application to composite air-breathing cylinders of relatively small size.
[10]
The Webster 2010 report concluded that a localized fire test was required that would be severe
enough (900 °C for 30 minutes in the same location of approximately 250 mm in length with a flame
impinging on the cylinder surface) such that a cylinder design able to survive the test is safe for service
even if there are small differences in repeatability. The burner design is specified and the fuel is propane.
The cylinder is filled to its service, or working, pressure with the fuel under test (CNG, hydrogen).
Cylinder orientation was also specified depending on the location and number of TPRDs used.
The surface temperature of 900 °C was decided based upon a review of several vehicle fire tests, as
[12]
reported by Gambone, 2008 . None of the vehicle fire cases reviewed had a fire temperature that
exceeded 900 °C for 30 minutes, therefore a fire of 900 °C intensity for a continuous 30 minutes likely
exceeds any continuous source of heat in a vehicle fire.
One approach to assessing the consistency, and validity, of different fires might be to measure the
depth of a resin char layer, or decomposition of a fibre, vs. time and vs. fuel use. This could be done
with a standard specimen so that the only variable is the fire itself. Thermocouples should measure
temperature of the outer surface and within the cylinder contents. Another approach might be to put
thermocouples in a standard test specimen and monitor the temperature vs. time and fuel use for
different fires, and also internal pressure vs. time using a standard contained gas, such as nitrogen.
Thermocouples would be on the inner and outer surface of the composite, within the laminate, and
within the cylinder contents. In fires fed by gas burners, the fuel delivery rate and fuel/air ratio would
both be variables. Wind velocity would be limited or controlled in these studies.
It is recommended that the fire test be conducted with the gases to be contained, if known. Different
gases have different temperature rise rate, pressure vs. temperature, flow characteristics through the
PRD and vent piping, and flammability. It is recommended that air not be used as a test gas unless it is
the intended content. Under some conditions, the air might react with the liner, creating an internal fire
that would intensify under pressure and cause the vessel to rupture. If the contained gas is a toxic gas,
it is generally recommended that a PRD not be used, so a fire test might not be appropriate, except to
determine time for the cylinder to burst in a fire as part of a system risk analysis.
6.3 Pressure relief devices
A PRD is a device that, when activated, vents the contents of the cylinder, and cannot be reclosed. A
pressure relief valve (PRV) is a device that, when activated, vents a portion of the contents, but can be
reclosed once a certain lower pressure is reached. PRVs are generally not used on composite cylinders,
because composite cylinders lose strength in a fire, and a PRV will not totally vent the cylinder, which
could result in a rupture of the cylinder.
There are two fundamental types of PRDs, those that are pressure activated and those that are
thermally activated. There are also hybrids that might need a combination of heat and pressure to
activate.
Pressure activated PRDs generally use a rupture disk made of metal. The rupture disk is held in a
carrier, and the carrier is installed in a valve or other component that is exposed to cylinder pressure.
The rupture disk would have a burst pressure that is lower than the burst pressure of the cylinder.
This would be true at any temperature, or any condition of fire exposure, to provide proper protection.
However, this condition might be difficult to meet if the cylinder is not at full pressure, and the fire is
sustained. In this case, the pressure might not reach the activation pressure of the rupture disk, yet the
cylinder strength is continually degraded by the fire, so a burst is likely. Another issue with rupture
disks is that, since they have a lower safety factor than the cylinder, they operate at a high stress level,
and they are at risk of failing by cyclic fatigue.
Rupture disks have been allowed in some composite cylinder standards, particularly in applications
where the cylinder is either full or empty. The thermal insulating properties of composite reinforcement
limits the increase in internal pressure during a fire, in which case the use of pressure activated PRDs
is not recommended. Pressure activated PRDs might be of value when working with liquefied gases
such as propane. The use of rupture disks for composite cylinders is subject to further discussion, and
including consideration of the application and related performance requirements. Some standards for
some applications have not allowed the use of pressure activated PRDs. However, rupture disks have
been used safely in some applications – for example, several million emergency breathing cylinders
using rupture disks for protection have a safe service record.
There are several methods that may be used to design and manufacture a thermally activated PRD.
The earliest method was to use fusible materials, specifically eutectic metal alloys. These thermally
activated PRDs generally activate when triggered regardless of the pressure in the cylinder. This makes
them a safer option than pressure activated PRDs if the application uses cylinders when they do not
have full pressure. Care would be taken in the design of PRDs with fusible materials such that they do
not extrude prior to being in a fire, and do not freeze shut once gas begins to flow when activated.
There are other materials that may be used in fusible PRDs besides eutectic alloys. Thermoplastic
polymers would be an example. Glass globes, similar to those in fire sprinkling systems, are being
used successfully. As temperature increases, a liquid inside the globe expands, breaking the globe
and activating the gas release mechanism. Shape memory metals can be used to activate a gas release
mechanism when a given temperature is reached. A gas release mechanism can be triggered if a
pressurized gas is released from a fusible trigger. An ignitable device, such as a detonation (ignition)
cord, can trigger a gas release mechanism. An electrical device, such as a thermocouple, can trigger
a gas release mechanism. However, electrically powered sensors and PRDs are discouraged from use
unless they have an uninterruptable supply of electrical power.
Other methods are expected to be developed over time. Any method should be suitable as long as it
properly activates a mechanism that releases the cylinder’s pressurized gas in a fire situation, and is
sufficiently robust to handle the environmental conditions without degrading over the lifetime of the
cylinder. Test methods should be developed to allow new PRD methodologies to be used, as they might
be more effective and reliable.
A polymeric liner that melts and releases gas might be an acceptable pressure relief mechanism.
However, it is necessary for the liner to reach its melting temperature before the composite is degraded
sufficiently to rupture. This might be possible in some, but not all, applications.
Insulation, intumescent coatings, or heat deflectors may also be used to protect cylinders in a fire.
While they might not release the contained gas, they can delay heat input to the cylinder and related
[10]
damage to the composite, as reported by Webster, 2010 . The use of means to delay the effects of
fire might be useful if the application or regulation puts a time limit on exposure to the fire, or allows
a rupture after a specified time. If thermally activated PRDs are used in combination with methods to
delay heat input, the heat should be allowed to flow to the PRDs with minimal restriction.
Combination PRDs are sometimes used where a pressure activated and thermally activated PRD are
used together, either in series or parallel. An example of a series device would be a rupture disk backed
by a fusible material. The fusible material might provide some structural support for the rupture disk,
so that there is less cyclic fatigue over its life. The disk would not rupture until after the fusible material
is activated. However, issues about venting partially charged cylinders would remain. A parallel device
would be activated by either high pressure or high temperature. This would increase the opportunity
to activate the PRD, but there would still be issues relating to fatigue life of the disk.
PRDs are often installed in one or both ends of a cylinder, particularly when the cylinder is of limited
length. When longer cylinders are protected by discreet PRDs, it might be necessary to use a manifold
and space the PRDs along its length. Some systems have multiple cylinders connected by pressurized
manifolds, with one or more PRDs connected. When a PRD is activated, all of the cylinders on the
manifold are vented. It is necessary to assure that all cylinders on the manifold can safely be vented
10 © ISO 2017 – All rights reserved

through a single PRD before cylinder strength is degraded in the fire to the point where one might
rupture. Care should be taken not to choke the flow when venting, either by having a small orifice to
flow through, or by having an excessive length on a vent line, as this could prevent the cylinders from
venting in a timely manner.
There are two fundamental modes in which a PRD can fail. A type 1 failure is when the device fails to
operate when it should. A type 2 failure is when the device operates when it should not have. There are
risks involved in either type of failure which should be considered when assembling a system. A subset
of the type 1 failure is when the device begins to operate when it should, but fails to properly vent the
cylinder, resulting in rupture. Examples of this include insufficient vent flow area, freezing of a eutectic
that plugs the line after flow begins, or plugging of the flow area by contamination. The standards for
the PRDs should have testing to reduce the risk of both types of failure of the PRD.
[13]
A 1997 report by Gambone conducted a safety analysis using PRD performance statistics
...