Analysis of cracking in tunnel lining segments

A. Focaracci
Prometeoengineering.it srl, Rome, Italy

SUMMARY: The article describes the study carried out with the purpose to seek the probable causes that
have provoked an overall systematic description of the cracks in the works of covering for the tunnel dug with
TBM and covered throw prefabricated structures in armed concrete. First of all, it is analysed carefully the
overall systematic description of the cracks and in addition the operating methods of excavation and covering.
This analysis allows to underline that the reasons for the cracking phenomenon are to return several factors
connected to ground conditions and to the presence of layer, to work conditions and to static aspects of the
covering. In underground works realized with cutter, these factors are closely related and strongly linked to-
gether and lead to undesired effects, in the adverse coincidence

many fields, but mainly with a view to providing solu-

tions for priorities set by regulations which, following a
1 INTRODUCTION totally prescriptive approach, had been completely for-
gotten in the past. I refer in particular to systems to facili-
tate evacuation, to improvements in the fire resistance of
structures and to fire fighting systems. Some of these de-
sign solutions are illustrated here.
Passive measures for the evacuation of persons from
a tunnel include the fundamental role played by structural
considerations and by the resistance of structures to fire
in particular.

Figure 1 Fire experiments at Montelibretti – A test tunnel with

the body of a carriage from Line C of the Rome Metro

The safety design for a road or rail tunnel must take

full consideration of studies of thermal fluid dynamic
phenomena resulting from an accident in a tunnel in order
to understand the possible effects on people exposed to
danger within the tunnel over a determined period of time
as a result of vehicles flowing through it.
The formulation of safety designs for many new or
existing tunnels has led to the development of a large Figure 2 Longitudinal cracks in the prefabricated segments of
number of innovative solutions. Work has been done in TBM driven tunnels
 European road network”, which implements that direc-
A study was conducted to find the probable causes tive. The legislation governing the rail network is the De-
which often lead to cracking phenomena in tunnels exca- cree of the Ministry of Infrastructures and Transport of
vated by TBMs and lined with prefabricated rc segments 28/10/2005 in Italy and the EC Directive 2008/163/EC
(a system employed mainly in the construction of under- entitled “safety in rail tunnels in the conventional and
ground works near the surface in urban contexts such as high speed trans European rail network” in Europe.
those for metropolitan railways) . These regulations identify fire as an initiating
The phenomenon of cracking is complex and in all event and set the following objectives to be achieved by
probability is the combined result of geomechanical fac- the infrastructure: “The integrity of a structure must be
tors, the methods used to place concrete and also of the maintained in the event of fire for a sufficiently long pe-
statics. These aspects are common in the construction of riod to allow first-aid to be given and passengers and
tunnels using full face tunnel boring machines with and personnel to be evacuated and rescue teams to operate
without earth pressure at the face. without the risk of structural collapse.
The analysis was then broadened to include the The fire experiment on the body of a vehicle
possibility of reducing cracking phenomena by using ma- from Line C of the Rome Metro at the fire fighting school
terials such as fibres, both metallic and synthetic. At the at Montelibretti was conducted in this context. It allowed
same time the response of concrete segments to fire was information on all the necessary factors concerning struc-
studied using data acquired from specially designed mod- ture, systems, plant and thermal fluid dynamics to be ac-
els. quired for the formulation of a specific and innovative
design for structures in tunnels and for prefabricated
segments in particular, because it was obtained directly
from full scale tests.

2 STRESS-STRAIN ANALYSIS OF
PREFABRICATED CONCRETE
SEGMENTS FOR TUNNELS

2.1 Background

A study was conducted on the structural behav-

Figure 3 Cracks in the prefabricated concrete segments of iour of prefabricated concrete segments used in the con-
TBM driven tunnel
struction of tunnels with reinforcement distributed
throughout the entire volume. Finite element analysis
Concretes reinforced with both metallic and (FEM) was performed on the stress-strain state under the
polypropylene fibres, termed fibre reinforced concrete different load conditions to which the prefabricated com-
(FRC) are innovative products developed in the nineteen ponents are subject during their normal life.
seventies. They can increase the fire resistance of con- The forces exerted on the segments in a tunnel
cretes in conditions where there is a danger of fire. are the result of both transitory conditions such as those
The idea of inserting fibre reinforcement in a ma- of the transport, placing and thrust of the jacks and per-
terial to improve its properties dates back a long time. manent forces such as the earth pressure when in service
The use of short discontinuous fibres distributed evenly or the fire load generated as the result of an accident.
within a concrete matrix confers greater strength to the The assumptions made to construct the model
concrete to withstand cracking and as a consequence im- simulate the geometry of the curve of the segment and
proved performance in terms of durability, even when the constraint conditions, even for faults in assembly,
subject to high temperatures. where a yielding support was modelled to simulate an
These characteristics of fibre reinforced concrete imperfect joint between one ring of segments and the
have been used in a study on prefabricated segments for next. As concerns earth pressure, it was decided to apply
the construction of tunnels using TBMs under all the load a load distributed on the edge of the segment by studying
conditions to which the prefabricated components are the effect of a change in this load resulting from correc-
subject during their normal life. tions to the direction of the cutter.
Italian safety regulations for road tunnels are con- The model constructed was designed to identify
tained in Legislative Decree No. 264 of 5th October 2006 possible concentrations of stresses and to understand the
entitled "The implementation of EC Directive possible improvements that changes to the normal rein-
2004/54/EC concerning the safety of tunnels in the trans forcement used in these structures might have on the dis-
tribution of the stress state in the lining rings.
 An analysis of the operations connected with the The considerations made led to the conclusion that
service stage found that conventional reinforcement dis- three courses of action needed to be taken to solve the
tributed evenly within prefabricated segments does not problem of cracking:
withstand the stress state that is generated as well as it − reduction of the thrust, which can be achieved by in-
might. The stress state in the operations for the removal creasing the foam injections through the nozzles at
of the forms, handling, storage and transport is typical of the cutter head, injecting bentonite along the shield
simple bending with minimal loads compared to those and increasing the excavation diameter of outermost
encountered during assembly and service. The strength of tool accompanied by an increase in the bi-component
the conventionally reinforced section is much higher than mortar injection to limit the increase in the loss of
the performance actually required of it. volume due to over break;
One of the particularly onerous load conditions is − greater care in the erection of the rings, with attention
generated by the thrust of the jacks for TBM advance. In paid to see that the face of the ring to which the thrust
this situation the concrete segments are subject to very is applied is free from sunken and raised segments;
strong loads which are borne under poorly defined condi- − reinforcement of the edge of the concrete segments on
tions due to inevitable irregularities in the plane in which the side opposite to that which bears the thrust, by in-
they bear against the previous ring and to possible as- creasing the existing reinforcement to withstand the
sembly faults already mentioned. tendency of cracks to form more effectively.
It is clear from the analysis performed that one of Finite element method (FEM) analysis of the
the main causes of cracks which are often found in the thrust phase under theoretical and therefore optimum
concrete segments of tunnels excavated using TBMs, as conditions shows that the stresses along the tunnel align-
already mentioned, is faulty assembly when the fit with ment are all compressive stresses, while those perpen-
the previous ring, especially with regard to the key seg- dicular to it are generated in the tensile zones concen-
ment, is not implemented properly leaving it raised (the trated on the side on which the thrust is applied.
key segment is slightly higher than the other segments in In the absence of assembly defects, the stress
the ring) or depressed (more likely when the key segment values produced are generally compatible with the
sinks too deeply into the ring). In addition, frequent strength characteristics of the materials employed, even
changes of direction are required to follow changes in the when the pressure is greater than that of the design speci-
alignment of the tunnel and to maintain the shield on the fications. The compressive stresses fall within the limits
alignment and these are effected by adjusting the hydrau- of the class of concrete employed and the tensile stresses,
lic pressure in the jacks which are divided into four which are always lower than those specified by regula-
groups each of which with four jacking units. tions, can be easily withstood by the reinforcement.
Various analyses were therefore conducted to
simulate the imperfect alignment of the key segments by
reducing the elastic modulus of the bearing surface, lim-
ited to one part of the elements into which it was divided.

Figure 5 Ideal thrust conditions – Deformation a stress perpen-

dicular to the direction of thrust

Figure 4 Diagram of the conditions analysed In the presence of “assembly defects” and/or ex-
ceptionally high thrust values, the compressive and ten-
sile stresses on the surface bearing against the previous
The strong forces of friction that must be over- ring rise to adjust the stress conditions to the absence of
come may depend basically on two factors, the fact that an area on which to bear and to absorb the excess thrust.
the pozzolana that is excavated, ground to a powder and The stress peaks generated on the edges of the zone
mixed with foam tends to grip around the shield and that where the support is missing, and that is opposite those of
the machine fits very tightly in the excavated tunnel. the thrust, can cause the origin and propagation of cracks.
 These types of stresses inside a concrete segment
act by means of a three dimensional stress state which
can be countered by placing longitudinal and transverse
reinforcement (i.e. tied bars which also generate a benefi-
cial confinement effect on the concrete) designed to “sew
together” the zones in which cracks form.
Subsequently, on the basis of the data considered
so far, the Montelibretti fire experiment was used to
study the possibility of improving the conventional rein-
forcement with the addition of steel and polypropylene
fibres to withstand both structural loads and those result-
ing from fire.

Figure 7 Thrust conditions with assembly defects – Deforma-

tion and stress perpendicular to the direction of thrust

These considerations lead to the conclusion that

reinforcement should be placed on the side opposite that
of the thrust to withstand the stresses resulting from the
ring assembly operations.
Figure 7 Bursting stresses - 3D stress state

The addition of steel fibres distributed in a con-

crete matrix modifies its mechanical properties. It im-
proves the tensile behaviour and counters the progressive
opening of cracks.
The fibres make their contribution once the ma-
trix has begun to crack, by conferring a post-cracking
strength on the composite material which is absent in ma-
trices without fibres.
The strongly degenerative behaviour, typical of a
Figure 6 Stress state due to the concentration of a compressive mono-axial tensile test on concrete, can be significantly
force modified by the addition of fibres, as the percentage vol-
ume is increased. The fibres also improve the tensile be-
haviour of the matrix during cracking, conferring residual
One further aspect analysed during the thrust op-
strength capacity on it.
eration is that of the concentration of the compressive
The use of short discontinuous steel fibres uni-
stress on a reduced area consisting of the thrust distribu-
formly distributed inside the concrete matrix confers
tion plate of the jacks.
greater resistance to cracking on the material and im-
The compressive force in this area spreads into
proves performance in terms of durability as a conse-
the concrete to produce a linear distribution of the
quence.
stresses on the cross section. The compressive stresses
These properties suggested that steel fibre rein-
deviate from the direction parallel to stress as they spread
forced concrete (SFRC) should be used for prefabricated
and this produces high transverse compressive stresses
concrete segments in TBM driven tunnels to partially re-
immediately after the anchor plate and subsequently
place the use of conventional reinforcement. As concerns
transverse tensile stresses.
polypropylene fibres on the other hand, their low resis-
These stresses, termed “bursting stresses” are
tance to fire can be exploited as an advantage to reduce
caused by the diffusion of the concentrated force and are
the problem of spalling in high strength concretes.
attributable to the curve in the trajectory of the compres-
In fact high strength concretes have little perme-
sive stresses. They normally act in the direction in which
ability and therefore the water vapour that develops in-
the load is applied and reach a maximum at a short dis-
side the material at high temperatures cannot escape from
tance from the surface loaded.
the concrete until concrete material itself is expelled by Ansaldo Breda S.p.A. and the Train Consortium to study
the pressure of the water vapour. the issues and to conduct a fire experiment on the body of
When the fibres dissolve, vacuoles are created in a carriage from Line “C” of the Rome Metro.
the matrix of the concrete which allow the water vapour
under pressure to escape through them thereby preventing
the concrete from exploding and shattering.
An experiment was conducted at the Monteli-
bretti school for fire-fighters designed to seek the best so-
lution from the viewpoint of fire resistance and conse-
quently of safety for tunnel users in cases of emergency.
A test tunnel was therefore designed in which different
types of concrete segment were employed, both prefabri-
cated and cast in situ, with and without the addition of
steel and polypropylene fibres in varying amounts.
Figure 9 Partners involved in the experiment
2.1 Fire experiment with the body of a carriage
from Line C of the Rome Metro
The assumptions adopted in the design needed to be
confirmed and the system choices and structural behav-
The main feature of the future Line “C” of the Rome iour needed to be verified with a real fire experiment in
metro is that is an automatic metro with no train drivers which one of the carriages to be used on the line in ser-
on board and with platform doors installed in the stations. vice was burnt, with the creation of an environment as
similar as possible to that of a real tunnel on the metro
line.

Figure 8 A three dimensional model of the body of a carriage

from Line C of the Rome Metro

While on the one hand, the new operating configura-

tion tends to increase the intrinsic safety of the transport
system, on the other hand there are problems concerning Figure 10 Experimental tests – Changes in the strength charac-
knowledge of the consequences of possible fires in tun- teristics of the materials with temperature
nels. In fact the platform doors constitute a true and genu-
ine longitudinal continuity with regard to fluid dynamic
models and they require careful assessment of the input Performing the experiment made it possible to
data, such as the heating power produced, the time taken test and validate on a final basis the simulations carried
for fire to develop, the emissions produced and flashover out at the design stage. This constituted a unique case of
phenomena. national and international value which will provide un-
The experiments programmed inside the test tunnel confutable data not only for Line “C”, but also for all
with a “sacrificial vehicle” allowed sufficient data to be other metropolitan railways, for better formulation of fire
acquired for subsequent simulations to be performed us- prevention and protection strategies and for passenger
ing physical and mathematical models. evacuation strategies.
Passenger safety in the event of a fire on a train in a The site selected to perform the test was the Fire-
metro tunnel was one of the most carefully studied sub- fighters Operational Training School at Montelibretti af-
jects in the design of Line “C” of the Rome Metro. A ter a tunnel 110 metres long was constructed inside the
memorandum of intent was signed between Roma school with the same dimensions and functions as that of
Metropolitane S.r.l., Metro C S.c.p.A, the Fastigi Consor- Line “C”.
tium, Ansaldo Trasporti Sistemi Ferroviari S.p.A.,
 The composition of the rings is shown in the figure
below.

While the general design remained unaltered, the fol-

lowing were constructed:
− 5 rings with completion of the upper half-ring cast in
situ of which 4 with concrete containing steel and
polypropylene fibre additives (type C* solution) and
one ring with concrete containing no additives;
− 1 ring with normal concrete according to the design;
− 1 ring with prefabricated concrete segments con-
structed with the addition of steel fibre reinforcement
according to the design (type A solution);
Figure 11 Test tunnel at the Montelibretti school for fire- − 1 ring with prefabricated concrete segments con-
fighters structed with the addition of steel fibre reinforcement
optimised according to the procedures illustrated in
the calculation report (type B solution);
− 1 ring with prefabricated concrete segments con-
structed with the addition of steel fibre reinforcement
according to the design and with polypropylene fibre
reinforcement (type C solution);
− 1 ring with prefabricated concrete segments con-
structed with the addition of steel fibre reinforcement
optimised according to the procedures illustrated in
the calculation report and with the addition of poly-
propylene fibres (type D solution);
− 4 rings with conventional prefabricated concrete seg-
ments according to the design.
The test was performed by recreating the condi-
tions specified in the design for Line “C” and therefore
the configuration of the site and the materials used were
Figure 12 The constructio n of the test tunnel at the Monteli-
bretti school for fire-fighters maintained as similar as possible to that design.
The objectives achieved by the test were as follows:
− to measure the curve of the heating power emitted by
This tunnel was constructed according to the de- the vehicle, the maximum power developed, the criti-
sign, partly using prefabricated concrete segments (the cal velocity of expansion of the fumes and the fire
same as those used to line the TBM tunnels of line “C”), load of the vehicle;
some of which containing fibre additives, and partly us- − to monitor the parameters of the combustion products
ing segments cast in situ in order to test different lining and of the substances released, the temperatures, the
combinations. visibility following the concentration of the fumes,
etc.;
− to verify the system choices for the design of Line “C”,
with particular attention paid to the safety measures;
− to test the structures of the tunnel following the fire.
The experiment made it possible to acquire all the
elements needed to formulate a specific and innovative
Figure 13 Detail of the rings where the carriage to be sacrificed design for structures in tunnels, because obtained directly
was positioned according to the design from full scale tests.
It is clear from the first results that the addition of
steel fibres can lead to a reduction in the use of conven-
The design specifications for the part of the tunnel tional reinforcement of more than 40% and at the same
(14 segment rings) in which the carriage body to be sacri- time results in greater efficiency at high temperatures, as-
ficed was positioned were as follows: sisted by the introduction of polypropylene fibres.
− 6 rings with completion of the upper half-ring cast in These considerations led to the proposal of a de-
situ; sign for prefabricated concrete segments which involves
− 8 rings constructed using different types of concrete abandoning the use of conventional reinforcement dis-
containing metal and/or polypropylene fibre additives; tributed throughout the volume in favour of prefabricated
− 7 rings starting from the centre protected with structures with steel and polypropylene fibres added to
“fireshield” fire resistant mortar. the entire volume with the addition of an edge in conven-
tionally reinforced concrete, which runs along the outer depth study of the response of a tunnel to fire by using
perimeter where the finite element analysis and experi- mathematical models.
mental results showed that stresses were concentrated. The most effective measures, as demonstrated by
the application of the Italian Risk Analysis Method
(IRAM) on more than 300 km of road and motorway
tunnels, are those given in paragraph 2.3 and those that
follow of Legislative Decree No. 264/06: escape routes
and emergency exits; access for first aid and rescue ser-
vices; parking places; drainage, resistance of structures to
fire, etc. This study on the structural behaviour of prefab-
ricated concrete segments for TBM driven tunnels, per-
formed using finite element method analysis of the stress
state and deformation under different load conditions to
which the prefabricated components are subject during
their normal life falls within this context.
Careful examination of cracking phenomena by means
of the analysis performed here resulted in the generation
Figure 14 Design solution for segments with optimised con- of FEM models that were used to study the probable
ventional + fibre reinforcement causes of the phenomenon from a structural viewpoint
and therefore to seek possible solutions in terms of both
construction measures and the characteristics of the struc-
Verification of this type of solution showed that tural components of the concrete segment linings. It was
optimised conventional plus fibre reinforcement, as illus- clearly found that action needed to be taken in a number
trated in the figure above, produces a configuration of of directions to prevent the manifestation of the cracking
strength to the section capable of withstanding all the encountered by adopting one or more of the following
load combinations generated by the ground in the short approaches:
and long term. − reducing the values of the forces applied;
− adopting specific measures during assembly;
− introducing extra reinforcement additives.
The experiment of a fire in a carriage from Line
“C” of the Rome Metro in a test tunnel at the Monteli-
bretti school for fire-fighters constituted a natural con-
tinuation of that work. Its objective was to furnish data
on the response to fire of the systems and structures of
the future Line “C” of the metro, i.e. those dedicated to
the safety of tunnel users.
These data were used to formulate a specific and
innovative design for structures in tunnels and for prefab-
ricated segments in particular, because they were ob-
tained directly from full scale tests.

Figure 15 Verification of a section of concrete segment with

reinforcement optimised for loads from the ground 4 REFERENCES

Focaracci A. 1999. Gestione progettuale dei preconsolidamenti

At the same time the configuration is also capable III Sessione delle Conferenze Permanenti sull’Alta Veloci-
of preventing possible cracking phenomena due to transi- tà, Bologna, 24 giugno 1999
tory conditions such as handling and assembly with the Focaracci A. 2005. La Gestione progettuale e costruttiva delle
thrust of the TBM. Gallerie, Gallerie e Grandi Opere sotterranee, n 76 Bolo-
gna, agosto 2005 Patron Editore
Focaracci A.,Mattei M. 2007. Le gallerie oggi e nella storia, Le
Strade, n.6, Milano giugno 2007 La Fiaccola srl
3 CONCLUSIONS Focaracci A. 1991 - Interventi conservativi: aspetti riguardanti
l'impresa specializzata. Convegno su il consolidamento del
suolo e delle rocce nelle realizzazioni in sotterraneo, Mila-
A correct engineering approach involves both no - Marzo 1991
Focaracci A. 2007. Utilizzo del sottosuolo in area urbana, Le
developing design solutions with a passenger safety ori- Strade, n.9, Milano settembre 2007 La Fiaccola srl
entation by improving evacuation processes and also in- Focaracci A. 2008. Il preconsolidamento del fronte, Le Strade
3/2008 Milano La Fiaccola srl
Focaracci A. 2008 .La galleria Frena, Le Strade , 12/2008 Mi-
lano La Fiaccola srl
Lunardi P., Bindi R., Focaracci A. 1989. Nouvelles orienta-
tions pour le Project et la construction des tunnels dans des
terrainsmeubles. Etudes et expériences sur le preconfine-
ment de la cavité et la preconsolidation dunoyau au front.
Colloque International "Tunnels et micro-tunnels en ter-
rain meuble - Parigi 7-10 Febbraio1989
Lunardi P., Focaracci A 1998. Mechanical pre-cutting for the
construction of the 21.5 m span arch of the “Baldo degli
Ubaldi” Station on the Rome Underground Atti del Con-
vegno Internazionale su "Tunnels and Metropolises”, São
Paulo, 25-30 aprile 1998
Lunardi P., Focaracci A.1998. Quality Assurance in the De-
sign and Construction of Underground Works Convegno
Internazionale su "Underground Construction in Modern
Infrastructure" - Stockholm 7-9 Giugno 1998
Lunardi P., Focaracci A.. 2000 - Modern Tunnelling in Italy
for High Speed Railway Line, Bauma 2001 Munich, zud to
8 th
EU Directive 2004/54/EC concerning the safety of tunnels in
the trans European road network”
EU Directive 2008/163/EC entitled “safety in rail tunnels in
the conventional and high speed trans European rail net-
work