Fire-resistant Tunnel Construction: Results of Fire Behaviour
Tests and Criteria of A pplication
H.-W. Dorgarten1, H. Balthaus2, J. Dahl1, B. Billig1
1
HOCHTIEF Construction AG, Consult, Essen, Germany
2
HOCHTIEF Construction AG, Infrastructure and Tunnelling, Essen, Germany
ABSTRACT
There is an increased public interest in the safety of tunnels in recent years due to the huge number of
fatalities and material damage from recent fires in tunnels. Due to the characteristics of burning
vehicles the fire impact in tunnels differs significantly from fires in buildings. Within the scope of
research and development HOCHTIEF has been consistently pursuing the objective of developing fire
resistant constructions especially for application in underground works and in tunnelling. In
cooperation with noted experts, fire resistant concrete using synthetic fibres and special aggregates has
successfully been developed. After a brief outline of the basic principles of the high temperature
behaviour of concrete, results from fire behaviour tests on highly loaded segments are presented. In
addition criteria of application are given.
1. INTRODUCTION
In recent years several incidents have tragically and clearly underlined the catastrophic effects that
may be caused by tunnel fires. For this reason, tunnels are to meet very stringent safety requirements,
with preventive measures gaining more and more importance. Frequently the focus is put on
organisational measures ensuring safe tunnel operation, technologies for early fire detection, improved
solutions for rescue routes as well as technical fire protection installations. Nonetheless, the potential
for innovation in the field of structural protection against fire must also be considered since the latter
constitutes the basis of all ensuing rescue measures, and is highly instrumental in reducing the
potential consequences of fires.
Thus, for several years, HOCHTIEF has been consistently pursuing the objective of developing fire-
resistant constructions especially for single and dual pass linings (Balthaus and Dahl, 2003). The
„System HOCHTIEF“ fire-resistant concrete was developed within the context of a comprehensive
research and development project involving the cooperation of the iBMB (Institute for Building
Materials, Structural Concrete and Fire Protection, Technical University of Brunswick, Germany).
2. STRUCTURAL FIRE PROTECTION: STATE-OF-THE-ART, REGULATIONS
Structural fire protection aims to retain load-carrying capacity and serviceability during and
subsequent to a fire. For this purpose, it is of particular importance to prevent extensive concrete
spalling and protect the main reinforcement from being exposed to high temperatures that would result
in irreversible strain.
B06 1
To attain this goal with regard to road tunnels, the structural components at risk are often covered with
mineral-based fire protection panels or fire protection facings. The problem of durability, in particular
due to dynamic effects brought about by tunnel operation and the impact of moisture, requires
increased maintenance and service. At the same time, the structure behind these mountings may no
longer be checked on the occasion of tunnel inspections. A further drawback consists in the fact that
these systems do not provide any fire protection during the construction phase.
The German code of practice ZTV-ING issued by BAST, 2003 alternatively provides for the
application of additional fire protection reinforcement for cut and cover tunnels in part 5, sec. 2. This
measure is to reduce concrete spalling to such an extent that the concrete cover in front of the main
reinforcement will remain and protect the latter from temperatures exceeding 300°C. So far, however,
the authors are not aware of any experience gained in the application of this type of fire protection
reinforcement in non-cut-and-cover tunnels and single pass linings, in particular.
3. DAMAGE MECHANISMS AND IMPACTING FACTORS
In the event of a fire in an unprotected tunnel, the main damage occurs by concrete spalling due to the
steep temperature gradient, in particular during the initial phase of the fire. Moreover, in case of
burning vehicles, there is the very long duration of the fire involving extremely high temperatures of
up to ca. 1,200°C. Concrete spalling develops due to various damage mechanisms, resulting from the
interaction between working loads and thermally induced additional strains (cf. e.g. Both et al., 2003).
In the course of the thermo-hydraulic process, the temperature gradient induces a transport of mass in
the form of water, steam and air through the pore structure of the concrete. Starting from the fire-
exposed concrete surface, the concrete develops: a dried-up, a drying and a quasi water saturated zone.
This is due to the steam that not only escapes via the fire-exposed surface, but also makes its way into
the concrete. The steam condenses in the cooler zone, increases the local water content up to the point
of water saturation and, thus, brings about an extreme reduction in vapour permeability. Subsequently,
the steam pressures generated in front of this zone become extremely high, up to the point where the
tensile strength of concrete is exceeded, giving rise to sudden spalling of sections of the concrete (as
shown in fig. 1).
Figure 1: Damage mechanisms due to the formation of steam pressure
The thermo-mechanical process describes the impact of thermally induced strains on the stresses and
strength of concrete. On the one hand, intrinsic stresses are caused by a non-linear temperature
distribution in the cross section. On the other hand, the different, and temperature-dependent,
properties of the reinforcement and the concrete components give rise to different temperature
B06 2
expansion behaviour. The aggregate will expand depending on the type of rock, while the cement
matrix is subjected to shrinkage processes caused by drying-out and chemical changes. Further
aggregate changes may be triggered by chemical reactions and mineralogical transformations, as well
as disintegration processes.
The first step towards developing a fire-resistant concrete consists in preventing the thermo-hydraulic
processes, i.e. the formation of a saturation zone and the high steam pressures linked with it. This may
be achieved by adding synthetic fibres to the concrete, as these fibres melt at temperatures of ca.
160°C, depending on the type of fibre used. The capillary pores and micro-cracks induced this way
form the preferred steam transportation path. Theoretical studies corroborate the experimental finding
that water vapour transportation and, thus, steam pressure equalisation via fine cracks is much more
effective than via larger individual cracks (Paliga and Schaab, 2002).
The selected aggregates are of central importance to the thermo-mechanical processes that occur in
concrete. A favourable aggregate size distribution and the imposition of a corresponding maximum
grain size may have a positive impact on the propensity for spalling. Furthermore, the strength
manifested after a fire very much depends on the type of rock. This is due to chemical and crystalline
changes and disintegration processes resulting from thermal impact. In quartziferous aggregates, for
instance, a mineral transformation combined with an increase in volume occurs at a temperature of
573°C. This entails a loosening of the concrete structure, leading to a decrease in concrete strength.
Apart from the impact of the (mainly local) mechanisms linked with the building materials, system-
based (global) effects on potential fire damage also need to be taken into account. In single pass tunnel
linings, for instance, the joint design has to be adjusted to allow for the thermal changes on the
concrete surface. To avert concrete contact due to thermal strain, specific precautions must be taken
with regard to the longitudinal und circumferential joints.
4. PLANNING AND EXECUTION OF FIRE BEHAVIOUR TESTS
4.1 Development of fire-resistant concretes
In a first stage of development, basic recipes for fire-resistant concretes were developed by conducting
fire behaviour tests with more than 50 test specimens. In this context, the RABT temperature-time-
curve in accordance with ZTV Tunnel (ZTV-ING part 5) was applied; it describes the anticipated fire
load in the event of road tunnel fires. Amongst other things, the concrete mix, reinforcement
arrangement, type of aggregate, fibre, size of fire-exposed surface and loading was varied during these
tests.
Figure 2: Test specimen on top of a fire chamber, fibre content trend line
B06 3
An evaluation of the test results obtained during this development phase revealed that the synthetic
fibre content has a significant impact on spalling behaviour. Subject to the concrete compressive
strength, fibre contents of between 2 and 4 kg/m³ are required. The higher the strength, the lower is the
capillary pore volume within the cement matrix. This suggests ´The higher the strength and density of
the concrete, the higher the fibre content needs to be´.
These experimental fire behaviour tests as well as additional calculation-based studies have
established that concrete spalling may definitely be prevented by adding synthetic fibres and by
selecting suitable aggregates. Both the carrying capacity as well as the serviceability of the tunnel
construction during and after fire exposure may then be guaranteed (Dahl and Richter, 2001; Haack,
2001).
4.2 Planning and conception of full-scale tests
Planning the full-scale tests involved the prior ascertainment of the actual state of stress in the tunnel
lining prior to and during a fire by conducting theoretical preliminary studies. The methods known up
to that point in time merely involved highly simplifying methods that only insufficiently captured non-
linear behaviour, in particular.
The nonlinear temperature distribution in the cross section that could be described as the result of an
unsteady heat transport is taken as a point of departure. Moreover the mechanical behaviour of the
concrete varies with temperature (fig. 3; cf. Kordina et al., 1999). The higher the temperature, the
lower the maximum compressive strength of the concrete, while ductility increases simultaneously.
Figure 3: Examples of the temperature-dependent material properties of concrete
For describing these nonlinear processes, the tunnel lining was mapped in a truss model by means of a
layer-based method (fig. 4). This method involves the subdivision of the cross section of a beam into a
certain number of concrete and reinforcing layers. For each layer, a mean temperature is calculated for
the period under review. Subsequently, the mechanical properties of the individual layers are defined
subject to the temperature of the respective layer. Based on the additional assumption that the cross
section remains even, it is now possible to determine the distribution of stresses in the cross section as
well as the axial force and transverse moment as a function of the cross-sectional strain and distortion.
It is then possible to use the standard nonlinear methods of determining internal forces, considering the
current level of stiffness. Other nonlinearities taken into account in this context are influences arising
from bedding, longitudinal joints (torsion springs) and circumferential joints (radial coupling).
B06 4
This model was used to realistically define the stresses occurring in full-scale tests before the tests are
conducted. It must be taken into account that, apart from the local constraint brought about by the
local increase in temperature in the tunnel construction, there is also a global constraint in the event of
a fire. The structure as such as well as the resistance from the surrounding ground give rise to
corresponding reaction forces. As the test specimen merely represents one section of the tunnel
construction, this global constraint must be calculated beforehand and applied to the test specimen
independent of the fire load.
Figure 4: Layer-based method
The tests were designed in such a way that the universal validity of the findings could be guaranteed.
This is why the stress level selected for the tests forms an upper limit of real conditions. In order to
ascertain this limit, values obtained from several projects were analysed. Amongst others, the stresses
that occurred in the River Weser Tunnel, the Herren Tunnel in Lübeck and the Katzenberg Tunnel (an
extreme load case due to swelling pressure) were analysed. The newly developed calculation model
was checked and verified by means of various comparative studies that have shown consistent results.
4.3 Full-scale tests
To develop a concrete ready for application, full scale fire behaviour tests involving heavily loaded
tubbing elements were conducted in 2003 (fig. 5). Amongst other things, they were designed to
experimentally ascertain the effect of external loads from rock and water pressure, as well as the
impact of the arched shape of the tunnel lining, on the load-carrying capacity in the event of a fire.
Figure 5: Test involving fire-exposed test specimens, experimental set-up
One fundamental problem concerning approval procedures consists in a current lack of national and
international standard test specifications applicable to fire behaviour tests. It is urgently recommended
B06 5
to standardise and harmonise test procedures. Against this backdrop, the test programme was
coordinated in advance with the DB AG and the EBA (Federal Railway Office) with a view to
possibly applying the developed method to future railway tunnels.
In the tests, the precast segments were exposed to a fire load in accordance with EBA Directive 07/97
(EBA, 1997). External hydraulic jacks exposed the test specimens to such a degree of strain that a
state of stress matching the real strain in a tunnel prior to and/or during the fire was built up. What is
decisive is the extreme stress at the edges of the cross section on the surface exposed to the fire. In line
with the statical preliminary studies, this stress level was 21 N/mm², plus fire-induced intrinsic
stresses.
The essential test results may be summarised as follows:
- Thermal load
Temperatures in the concrete cross section corresponded well to the preceding temperature
calculations. Maximum temperature of 365°C was measured in the lower reinforcement layer
(cover c ≈ 6 cm) (cf. fig. 6). This remained below the limiting temperature that might result in
irreversible strain, taking into account the loads applied to the steel in this particular case (Kordina
et al., 1999).
Figure 6: Temperature in the fire chamber, temperature distribution in the concrete cross section
- Mechanical load
To be on the safe side the maximum stresses expected in a tunnel lining during the first 60 minutes
of fire were applied in the test from the beginning. Even when subjected to such high external
loads, the tests specimens endured the high fire load in accordance with the EBA curve without
showing any sign of failure. Measured deformations such as deflection at the segment centre as
well as displacements of supporting plates were reasonably in line with the results of the
preliminary calculations, and thus corroborate the theoretical approaches taken.
- Spalling behaviour, strength properties and serviceability
No spalling occurred on the side that was exposed to fire. Even the surface of the concrete
structure revealed a very good level of consistency (fig. 7). Except for a thin layer close to the
surface, the concrete compressive strengths measured subsequent to the fire behaviour tests
exceeded the reference values of a C35/45 (fig. 7). On the whole, the load carrying capacity of the
overall cross section and, as a consequence, the stability of the construction was retained.
B06 6
Figure 7: Concrete surface after the fire behaviour test, profile of compressive strength
- Carbonation depth was about 2 cm. The adhesive pull strength measured on the immediate surface
exposed to the fire exceeded 1.5 N/mm², thus confirming the concrete compressive strength
established directly on the concrete surface. Taking all these parameters together results in the
conclusion that it will not be necessary to carry out repairs immediately after the fire. However, it
is recommended to carry out medium-term repairs of the concrete surface in areas exposed to high
thermal strain for optical reasons and for maintaining the life of the tunnel. After preparation of
the foundation, a surface protection system should be applied in order to slow down any further
progress of carbonation.
5. CRITERIA OF APPLICATION
The highly comprehensive test programme has revealed that concrete as a building material can attain
a high level of resistance to fire exposure. To provide fire-resistant tunnel construction, a combination
of measures is required. In terms of project-related utilisation, the following application criteria should
be taken into consideration:
- Principally, the risk of spalling is ascertained and/or eliminated by using appropriate synthetic
fibres, carefully assessing and selecting aggregates and by an appropriate grain size distribution,
with specified maximum grain size.
- Extensive additional static studies to assess the load-carrying capacity and serviceability within
the segmental lining multi-joint system, taking into account temperature strains and temperature-
dependent material behaviour, are indispensable.
- The tests have revealed that the use of steel fibres in concrete alone does not result in improved
spalling behaviour. To ensure sufficient load carrying capacity subsequent to a fire, additional
reinforcement is required.
As to joint design, the local thermal strains in the individual precast segments are to be taken into
consideration. An appropriate joint design must ensure that both the impermeability and the efficiency
of the system coupling in the circumferential joints are fully maintained after a fire. The coupling
between the rings may be secured by a deep tongue and groove joint to name but one example.
B06 7
6. SUMMARY AND OUTLOOK
The results of the development work have fully met the current requirements that have to be made
with regard to fire-resistant tunnel constructions. The new fire protection system does not require any
additional measures such as fire protection reinforcement or fire protection cover. The novel system is
capable of fully meeting the criteria applicable to a defined structural protection scheme both for roads
and for railway tunnels. Use of the system will ensure that the load-carrying capacity of the tunnel
lining as well as its serviceability is not impaired during a fire, not even in the event of major external
strains.
The criteria of application demonstrate that fire-resistant concretes are not only generated by adding
certain synthetic fibres, but require careful use of concrete technology as a whole. In addition, we need
to look beyond concrete as a construction material, taking into consideration the overall system in
order to really achieve a fire-resistant construction of the tunnel as a whole.
This newly developed and trendsetting fire protection system provides developers and planners with
an important and novel tool for making tunnel construction safer and longer-lasting. In view of future
infrastructural projects all those involved in construction schemes have to ensure an adequate further
development of the state-of-the-art.
REFERENCES
Balthaus, H., Dahl, J., 2003. Neue Wege beim Brandschutz im Tunnelbau (New avenues of fire
protection in tunnel construction). Proceedings Bautechniktag 2003, Hamburg
BAST (Bundesanstalt für Straßenwesen), 2003. ZTV-ING Teil 5 Tunnelbau (Additional technical
terms and conditions as well as guidelines for engineering constructions, part 5 Tunnel Construction).
Verkehrsblatt-Verlag (publisher), as of 01/2003.
Both, C., Wolsink, G.M., Breunese, A.J., 2003. Spalling of concrete tunnel linings in fire. (Re)
claiming the Underground Space, Saveur (ed.), Swets & Zeitlinger, Lisse.
Dahl, J., Richter, E., 2001. Brandschutz: Neuentwicklungen zur Vermeidung von Betonabplatzungen
(Fire protection: New developments for avoiding concrete spallings). Tunnel 6/2001, pp 10 – 22.
EBA (Eisenbahnbundesamt), 1997. Anforderungen des Brand- und Katastrophenschutzes an den
Bau und Betrieb von Eisenbahntunneln (Fire protection and disaster prevention requirements
applicable to the construction and operation of railway tunnels) Directive , Bonn.
Haack, A., 2001. Kommentar zum neuentwickelten Brandschutz im einschaligen Tunnelausbau
(Comment on the newly developed fire protection scheme in single path linings). Tunnel 6/2001, pp
23 – 31.
Kordina, K., Meyer-Ottens, C., Richter, E., 1999. Beton-Brandschutz-Handbuch (Concrete fire
protection manual), 2nd edition, Verlag Bau + Technik (publishers), Düsseldorf.
Paliga, K., Schaab, A., 2002. Vermeidung zerstörender Betonabplatzungen bei Tunnelbränden
(Avoiding destructive concrete spallings in tunnel fires), Bauingenieur volume 77.
B06 8
