Introduction
Many of the fire safety regulations in SOLAS
stand in correlation with performance in fire tests. Some relevant characteristic
parameters which are currently measured are:
These parameters are measured in different ways depending on the represented fire
risk scenarios and with various criteria depending on the hazards involved. The
different tests have not developed with particular attention to FRP composite
constructions but may still be applicable, even if certain considerations may be
necessary. However, there is already a market for FRP composite constructions in
naval and commercial maritime applications, particularly for high-speed crafts
(HSC). For this purpose, new regulations and standardized tests applying to such
materials have been implemented in the International Code of Safety for High-Speed
Craft (HSC Code). It includes several significant differences with
regard to the safety organization, available egress time and requirements for the
materials, but it may still be relevant to refer to the related fire tests when
considering FRP composite structures in SOLAS
ships. Any standardized or experimentally set up test may be referred to as a SOLAS
regulation
II-2/17 assessment but may require evaluations of the test results.
Assessments by experts may also allow the transfer of test results from one FRP
composite composition to another.
Below follows a discussion on the limitations of safety validation through tests in
general and on uncertainties that need to be considered when using current fire test
procedures to validate FRP composite in particular. Thereafter, the most relevant
fire tests prescribed by SOLAS
and the HSC Code are briefly described, with focus on the
particularities with testing FRP composite. For some FRP composite constructions it
may be necessary to look beyond the approved fire test procedures and consider other
standardized tests or tailored experimental tests, which are discussed at the end of
this chapter.
D.1 Uncertainties when using tests to validate FRP composite
1 Testing is a good tool to evaluate whether a construction performs satisfactorily
in a certain situation. Full-scale testing is the method that typically will give
the most accurate results on how a design will perform, even if natural variations
are always present. Since it would be very costly to evaluate all possible scenarios
in full-sized tests, some characteristic parameters are generally investigated in
certain ways during exposure to plausibly worst-case scenarios. The overall safety
performance is therefore assumed to stand in correlation with the performance in
these characteristic tests, derived from knowledge of fire dynamics and behaviour of
materials when exposed to fire.
2 However, FRP composite and steel, which it generally replaces, are inherently very
different. Some general particularities with FRP composites are the anisotropy and
inhomogeneity, which may give variations in test results depending on the
positioning. Another potential difficulty is that the different plies of resin
impregnated fibre cloths might delaminate during testing. Produced gases will
strengthen this tendency as they seek the outlet of "least resistance". The latter
effect will not be captured in a small-scale test since the maximum travel distance
for gases in the real fire will be much longer than in, e.g. a "Cone Calorimeter"
test where the maximum "travel" distance is 5 cm. The "edge" effect will therefore
be much more important in a small-scale test than in a full-sized test. Different
remedies to problems related to scale are given in the literature and they include
edge protection, which in the Cone Calorimeter could be the use of a sample holder
that covers the edges completely or to vary the sample size or orientation.
3 Evaluation of two such diverse construction materials through the same tests may be
claimed to be quite obtuse. Today's fire tests are generally constructed to measure
some key properties reflecting different disadvantages of traditional (steel)
constructions and ideally represent the performance of such constructions when
exposed to a severe fire. However, some characteristics are left out in the tests
because of the implicit benefits of traditional solutions. Therefore, implicit
advantages may not be represented in the tests and may not be possible to evaluate.
What must be considered further is also the uncertainty associated with performance
criteria generally being binary, i.e. pass or fail. When evaluating designs through
tests there is always a lowest level for passing the test, an acceptance criterion.
Assurance of identical set-ups and measurements are obviously of greatest
significance when tests are carried out by different people and at different labs in
countries throughout the whole world. However, even without those uncertainties, a
test says nothing about the performance not represented in the test, i.e. the
performance of the sample if the load, temperature or time in the test increases by
10%, 20% or 50%. In general, the prescriptive fire tests of the International Code
for Application of Fire Test Procedures, 2010 (2010 FTP
Code) only give pass or fail. Therefore, no information is given on how
the construction performed during the test or how long it could have performed with
satisfaction. An example of this is the ability of steel bulkheads to withstand high
temperatures before structural deterioration. It is because of the implicit
advantages of steel, not visible in standardized tests, that there is an additional
requirement for many structures to be made of steel or other equivalent material.
However, when aluminium was introduced to merchant shipbuilding, it was necessary to
address this in a better way. Aluminium was, according to regulations, considered as
an alternative non-combustible material to steel. However, the relatively poor
structural behaviour at elevated temperatures (aluminium does not burn but
nevertheless melts in the non-combustibility test) highlighted the simplistic nature
of the non-combustibility requirement. Aluminium structures were therefore generally
required to be fitted with double sided insulation and were thereby considered
equivalent to steel in this regard. Furthermore, when non-metal load-bearing
structures are considered for HSC, they are subjected to an additional load during
structural fire resistance tests in order for the structure to be considered
equivalent to a metal construction. Therefore, there may be reason to assess whether
the standardized tests fully reflect the risks and benefits of FRP composite
structures in case of fire. Implicit properties beyond the tests need to be
identified, which is one of the objectives behind these guidelines, and may require
verification through additional tests.
D.2 Low flame-spread characteristics
1 The potential for flame spread of a material is tested in equipment where an
irradiating panel provides heat input to a surface in order to initiate flaming
combustion. The IMO typical example of such equipment is shown in figure 5. Fire is
initiated where the distance between panel and sample is the shortest, i.e. where
the irradiation intensity is the highest. The radiation level decreases at the test
specimen from left to right in figure 5, and the extreme burning point to the right,
i.e. the point with the lowest irradiation level for sustained combustion, is given
as a measure of flame spread for the material. The speed of the flame front movement
is also quantified in an appropriate way. There are also criteria regarding the peak
heat release as well as of the total evolved effect.
Figure 5: Test for flame spread according to part 5 of the 2010 FTP Code
2 When testing FRP composite according to this procedure, it is important to comply
with the requirement to test the specimen with end-use conditions. The material
behind the tested surface material will significantly affect the fire behaviour. A
well-insulating material behind a thin ply will keep much more of the heat at the
surface and generally worsen the conditions for the tested surface material.
Therefore, if the end-use is a sandwich panel, it is not appropriate to test only
the surface laminate on a steel plate or directly in the sample holder. The
equipment normally fits a 50 mm thick sample and for FRP composite it is
recommendable to include as much of the composite material as possible in the sample
holder.
D.3 Generated effect and smoke in small scale
1 The HSC Code includes regulations for furniture and other
components which require investigating fire behaviour on a small scale in the "Cone
Calorimeter" test equipment defined in the standard ISO 5660 (shown in the schematic
picture in figure 6). The 0.1 x 0.1 m specimen is horizontally positioned and
subjected to irradiation from electrically heated surfaces above the tested
material. Irradiation levels are typically in the range of 25 to 50
kW/m2.
Figure 6: Schematic picture of a Cone Calorimeter
2 Except from time to ignition, the standard ISO 5660 Cone Calorimeter test includes
measurement of smoke (obscuration) and heat release under different radiant fluxes.
There is a criterion for the peak heat release rate. The time integrated HRR signal
provides the total heat release (THR), which must be limited and is a very important
material fire characteristic. The HRR curve for such an experiment on a carbon fibre
based composite laminate is shown in figure 7.
Figure 7: Small-scale experimental results from carbon FRP composite
material
D.4 Generated effect and smoke on a large scale
1 The criteria for the Cone Calorimeter are designed to correlate with a large scale
"Room Corner" test scenario according to the standard ISO 9705. It is an important
standardized piece of equipment for testing material potential for HRR and smoke,
schematically pictured in figure 8. In this test, the material to be tested is
mounted on walls and ceiling and a propane gas burner positioned in a corner of a
full-scale room provides a 100 kW power output for 10 min, followed by a 300 kW
output for an additional 10-minute period. The HRR and smoke production rate are
continuously measured and the criteria that apply are similar to those in the Cone
Calorimeter.
Figure 8: Schematic view of ISO 9705 Room-Corner experimental set-up
2 The standard ISO 9705 test is important for marine applications as it is used in
the 2010
FTP Code for experimental verification of FRM, "Fire Restricting
Materials", used on HSC. On SOLAS
ships requirements are more relaxed and surfaces are often coated with combustible
paints that would not pass FRM requirements, in particular if applied to an FRP
composite surface. The material behind the surface finish has a major impact on the
test results and, due to the high thermal conductivity of FRP composite, this test
is therefore rather challenging for FRP composite systems. Furthermore, droplets and
debris must also be considered according to the test requirements. It is thus
crucial that FRP composite materials are tested in end-use conditions. It should be
noted that, in comparison with the test for surface flammability, the room corner
test is not only full scale but also includes further complexities, in particular
with regard to effects of enclosure fire dynamics. Flames and smoke are collected in
the room and heat up surfaces in a different way. These reradiate between each
other. The effects from enclosure fire dynamics also generally make the test harder
to pass than the test for spread of flame; that is, materials that pass the room
corner test generally also pass the test for spread of flame. For exterior
combustible surfaces, the ability to manage effects from enclosure fires could be
claimed irrelevant, as these effects will not appear out in the open on exterior
surfaces. Therefore, for such areas a different test could be more suitable.
D.5 Non-combustibility
The previously described test methods have been presented in an approximate order of
difficulty with regard to fire behaviour of the materials. The ultimate fire-related
material quality is non-combustibility, determining whether the material is at all
considered combustible. An accepted method for measuring combustibility is the fire
test given in part 1 of the 2010 FTP
Code (see figure 9). A specimen is exposed to 750°C in a cylindrical
furnace where temperature increase, flames and weight loss are measured to determine
combustion.
Figure 9: Combustibility test equipment according to part 1 of the 2010
FTP Code
D.6 Smoke generation and toxicity
1 In evaluations of materials it is often relevant to combine properties of fire
behaviour (fire growth, fire spread, etc.) with materials' potential for smoke
generation and toxicity. For maritime applications, the "smoke box" is used for
smoke and toxicology measurements, based on part 2 of the 2010 FTP
Code. For SOLAS applications this test is only required if results in
the test for spread of flame are insufficient. In this method, a 0.5 m3 closed cubic
box (figure 10) is used for exposing a small (75 mm x 75 mm) sample for irradiation
and measuring continuously gases and smoke opacity in the box. Criteria concern
maximum amount of smoke produced and maximum concentrations of the following gaseous
species: CO, HCl, HF, NOx, HBr, HCN and SO2, as given in the
2010
FTP Code. The test proceeds for 10 minutes if a maximum has been observed
in the smoke obscuration level; otherwise the test proceeds for another 10 minutes.
The toxicity levels when the smoke obscuration reached its peak value are used as
the result from the test.
Figure 10: Smoke box equipment
2 In this test, materials generally produce more smoke before ignition than after
they have ignited. The same applies to most gases, in particular CO levels which are
significantly higher before ignition (the opposite applies for HCN). Therefore, FRP
composite materials that have been treated to impede ignition and flame spread
generally produce smoke and toxic gas in levels which may make it challenging to
pass the test.
3 There is no requirement to test insulations, bulkhead panels and similar items for
smoke and toxicity, since they are assumed to be non-combustible. However,
regardless of whether a fire restricting material is used on top of an FRP composite
panel, if a surface with low flame-spread characteristics is applied or if the FRP
composite panel is left bare it could be claimed that it is the surface of the
compartment which should be tested. End-use conditions apply also in this test
method and as much of the FRP composite that fits in the 25 mm sample holder should
then be included in the test. The long and significant heat exposure will cause
materials underneath the potentially burning surface to thermally decompose. Even if
the result is not the same as if the underlying materials were directly exposed,
they will contribute to the generated smoke and toxic gases to an extent that is
representable to the heat exposure in the test and in a fully developed fire.
D.7 Structural resistance
1 For load-bearing structures on SOLAS
ships, structural resistance to fire is tested by exposing the sample to a
well-defined temperature that increases over time. Typical standardized
time-temperature curves are used as reference for the temperature in the furnace as
depicted in figure 11.
Figure 11: Time-temperature curves used for testing of structural
resistance
2 In the structural resistance test the sample insulation properties are tested, i.e.
its ability to withstand heat while keeping the temperature down at the unexposed
side of the sample. The required performance time in a test and the demand for the
backside temperature depends on type of test and type of classification. An example
of a structural resistance test, used e.g. for walls, doors, bulkheads, etc., is
illustrated in figure 12, where a load-bearing wall with a window is exposed to
heat. Another test for a door construction is shown in figure 13.
Figure 12: Large-scale structural fire resistance test of a window
Figure 13: Insulation test of a door where thermocouples measure the temperature
of the unexposed side during the heat exposure
3 As discussed above, for SOLAS
applications there is no requirement in the test procedures to evaluate the
construction load-bearing capabilities. In the HSC Code the divisions corresponding to A-class divisions in SOLAS
are referred to as Fire Resisting Divisions (FRD). The main difference is the
requirement for an A-class division to be constructed with non-combustible material,
which does not apply to an FRD. The structural fire resistance test is basically
identical to the test required for A-class divisions, except for an additional
load-bearing requirement. This requirement implies that FRD decks and bulkheads
shall withstand the standard fire test while subjected to transverse and in-plane
loading, respectively. A FRD deck or bulkhead structure must sustain the specified
static loading whilst exposed to fire in a large-scale furnace for 30 minutes or 60
minutes in order to be certified as an FRD30 or FRD60 division, respectively.
4 Loading during fire resistance tests may be highly relevant when evaluating FRP
composite constructions for SOLAS
ships. However, research has shown that it is more suitable to apply the design load
than the relatively low static load (in accordance with part 11 of the FTP Code)
when testing insulated FRP sandwich panel bulkheads. It is likely that a similar
fire test procedure is suitable also for other FRP composite design concepts (e.g.
non-insulated FRP bulkheads, different deck concepts) but this must yet be verified.
Penetrations in FRP composite structures could reduce the load-bearing capacity and
may call for testing of penetrations in load-bearing structures as well. Tests have
been performed with certain FRP composite panels with holes that did not show any
such effects. However, effects clearly depend on the made penetrations and on the
safety margins included in the design. Fire resistance tests for penetrations on HSC
are not performed with applied load.
Figure 14: Small-scale furnace for structural resistance tests
4 Small-scale test methods for structural resistance exist but are used in R&D
projects or for product quality control. The maximum size of the tested sample in
the small-scale furnace is 0.5 m x 0.6 m (figure 14), which is to be compared to a
typical full-sized test as shown in figure 12, where a 3 m x 3 m sample is being
tested. In a SOLAS regulation II-2/17 assessment it may be relevant to refer to
standards other than IMO test standards to evaluate fire-resistance (e.g. ISO 834-12
Fire resistance tests – Elements of building construction, Part 12: Specific
requirements for separating elements evaluated on less than full scale furnaces and
ISO 30021 Plastics – Burning behaviour – intermediate-scale fire-resistance testing
of fibre reinforced polymer composites).
D.8 Additional testing
1 Throughout different research projects many experimental tests have been carried
out. Except for tests according to all of the standardized test procedures described
above, tests have, for example, been carried out for divisions' structural integrity
in vertical and horizontal furnaces with various time, integrity requirements and
loads (nominal load according to the HSC Code, design load and realistic load). Many solutions for doors,
windows and penetrations have also been certified in such tests and different
outfitting solutions have been tested in experimental tests with corresponding fire
exposure. Fire growth has been evaluated for external combustible FRP composite
surfaces based on a standardized test method for testing reaction to fire properties
of building façade systems.footnote In the tests, the performance of FRP composite
surfaces protected with different passive or active measures were compared with a
completely non-combustible surface (hence the multiple layers of paint on a steel
ship were ignored). Performance criteria have been developed for external drencher
systems to determine under which conditions a drencher may be effective when using
FRP composite on external surfaces. Tests have also been performed based on the
Guidelines for the approval of fixed pressure water-spraying and water-based
fire-extinguishing systems for cabin balconies (MSC.1/Circ.1268) which showed that a balcony sprinkler prevented a fully
developed cabin fire from spreading to FRP composite surfaces on the balcony and on
outboard sides of the ship.
2 Depending on the intended use of FRP composite further tests may be relevant,
e.g.:
-
.1 a joint between steel and FRP composite could be fire tested to ensure
that collapse will not occur due to heat conduction from fire in an
underlying steel compartment;
-
.2 if insulation is used, it may be relevant to test FRP composite which is
insufficiently insulated, e.g. a small or large-scale furnace test with 0.1
m x 0.1 m or 0.5 m x 0.5 m lack of insulation, or emergency
repaired/modified; and
-
.3 structural integrity test of a composite deck exposed to fire from
above.
3 It may also be claimed necessary to prove that an FRP composite material is not
easily ignited. Even though restricted ignitability is required by functional
requirements in SOLAS regulations, there is no IMO certifying test to show
this property. However, EN ISO 11925-2, Reaction to fire tests – Ignitability of
building products subjected to direct impingement of flame – Part 2: Single-flame
source test or the Guidelines on fire test procedures for acceptance of
fire-retardant materials for the construction of lifeboats (MSC/Circ.1006) provide possible test methods. EN ISO 11925-2 specifies a
test method which measures the ignitability of building products when exposed to a
small flame. Based on numerous fire tests with various FRP composite materials,footnote it has, however, been judged very likely that most
exposed surfaces of untreated FRP composite (i.e. the laminate) would pass such a
test. This can also be distinguished from the Cone Calorimeter test data in figure
7. The graph does not only show that the FRP composite may become involved in a
significant fire but also that it resists the rather significant irradiation of 50
kW/m2 for at least one minute before becoming involved in a large
fire. For reference, 15 to 20 kW/m2 towards the floor is often referred
to as a criterion for when flashover is determined in an enclosure fire. A Molotov
cocktail has, for example, been concluded not to be able to ignite the particular
FRP composite surface tested in figure 7. In the aforementioned test method for
ignitability of building products, the material is exposed to a flame the size of a
match for 15 or 30 seconds. It can thereby be concluded that FRP composite surfaces
generally have restricted ignitability and what could rather be a problem is fire
spread if the surface is exposed to an already established fire. If considered
relevant, the ignitability of various FRP composite surfaces may be evaluated
through a test, e.g. according to the standard EN ISO 11925-2 or MSC/Circ.1006.