See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/282909412

Design and construction of a sewer tunnel in difficult site conditions

Conference Paper · May 2011

DOI: 10.1201/b12748-22

CITATIONS READS

0 759

4 authors, including:

P. Croce
Università degli studi di Cassino e del Lazio Meridionale
58 PUBLICATIONS   642 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

geomechanical analysis on point cloud from laser scanners View project

All content following this page was uploaded by P. Croce on 17 October 2015.

The user has requested enhancement of the downloaded file.

Design and construction of a sewer tunnel in difficult site conditions
P. Croce
University of Cassino, Italy
S. Di Maio, G. Speciale, L. Cassibba
Sering Ingegneria s.r.l., Italy

ABSTRACT: The paper deals with a sewer tunnel, recently excavated in the city of Palermo. The tunnel lay-
out is close to existing urban facilities and residential buildings. The geotechnical profile is quite variable and
the soil cover ranges between 7 and 9 m. The groundwater level is located at mid height between the invert
arch and the vault of the tunnel. Excavation support was granted by means of the canopy technique, employ-
ing micropiles and jet grouting. A detailed monitoring program was also organized and tunnel construction
was followed step by step allowing for proper design changes, according to the subsoil conditions met during
excavation. The case history is reported by describing the tunnel layout, the geotechnical profile and the
monitoring plan. Recorded settlements of the existing structures are analyzed, considering the influence of the
geotechnical properties of the subsoil as well as the peculiar construction sequence of tunnelling.

1 INTRODUCTION technique can also be customized in order to account

for variable soil properties, by making proper use of
Tunnel construction in urban areas is generally a the different soil reinforcing techniques. Design fea-
very difficult task for several typical reasons that tures can also be modified in real time to meet un-
may be listed as follows: foreseen situations, such as unexpected soil condi-
i. the presence of weak soils (including made tions, undetected underground facilities or excessive
land), characterized by low shear strength, surface settlements, as it frequently happens in urban
which may compromise the stability of the areas. In such cases, it may be appropriate to apply
underground excavation the observational method (Nicholson et al., 1999;
ii. the occurrence of seeping water, coming Patel et al., 2007; Croce, 2010) in which the design
from local aquifers or water lines leakage, is reviewed during construction as allowed by the
which may induce soil piping along the cav- Eurocode EC7.
ity In the following, the typical features of the can-
iii. the excavation process which generates soil opy technique are first recalled, considering the
deformations that may induce excessive set- available soil reinforcing methods and their imple-
tlements of adjacent or overlying buildings mentation in the tunneling procedure. This technique
iv. the crossing of other underground facilities was successfully employed in order to excavate a
which may interfere with digging activities sewer tunnel, built in difficult site conditions in the
It is well known that tunneling through weak city of Palermo.
soils may result in face and vault instability, unless The case history is then reported in some detail
proper soil support is provided during excavation. by describing the tunnel layout, the geotechnical
This goal can be reached either by using a TBM or profile, the design features and the monitoring plan.
by installing a set of reinforcing elements, forming a Finally, recorded settlements of the existing struc-
sort of canopy along the cavity contour and behind tures are analyzed, considering the influence of the
its face. geotechnical properties of the subsoil as well as the
However, the use of a TBM becomes technically peculiar construction sequence of tunneling.
and economically feasible only for long tunnels with
ample curvature radii, circular cross section and con-
stant diameter. 2 CANOPY TECHNIQUE
On the contrary, the canopy technique can be
readily used for short tunnels, non circular cross sec- The canopy technique is based on the use of soil
tions and abrupt changes of direction. The canopy reinforcing methods which provide stability of the
   6

1
5 

3 4  7
2

Figure 3. Plan view. 1) Tunnel layout 2) Park 3) Retaining wall

4) Subway Ticket Office 5) Pedestrian underpass 6) Subway
line 7) Buildings.

In principle, each reinforcement technique is best

Figure 1. Schematic drawings of the canopy technique: A – re- suited for particular soil types and seepage condi-
inforcement by jet grouting columns , B – reinforcement by tions. Jet grouting treatments are generally preferred
steel micropiles and fibreglass bars or tubes. for sandy soils, providing larger columns which can
be set aside forming a sort of supporting arch (Fig.
1A). The latter can provide also waterproofing,
when properly dimensioned, if pore water pressures
are not too high.
However, for fine grained materials jet grouting
is not very effective and so for clayey soils it is usu-
ally more convenient to choose micropiles (Fig. 1B).
In particular, steel micropiles are used for the tunnel
contour while fibreglass bars or tubes are employed
for the face reinforcement, since they can be easily
truncated during excavation. Finally, in difficult
cases or doubtful soil conditions, it may be useful to
combine jet grouting and micropiles in order to form
a sort of reinforced jet column.
Whatever method of soil reinforcement is chosen,
the canopy technique is characterized by a construc-
Figure 2. Tunnelling sequence by the canopy technique: A – tion sequence which proceeds by subsequent spans,
reinforcement of tunnel contour by jet grouting and/or steel
micropiles, B – excavation, C – excavation completed, D –
as depicted in Fig. 2. The span length is usually
face reinforcement by jet grouting and/or fiberglass bars or comprised between 6 and 10 m. For each span there
tubes (optional). are two main construction phases: treatment along
the tunnel contour (Fig. 2A) and soil digging (Fig.
tunnel contour during excavation. Such canopy is 2B). After the excavation is completed (Fig. 2C) it is
obtained by means of steel micropiles and/or jet possible to consolidate the tunnel face (Fig. 2D)
when this treatment is needed.
grouting columns which are installed in advance
with respect to excavation. Moreover, for very weak
soils, the stability of the tunnel face may be granted 3 TUNNEL DESIGN
by diffused reinforcement of the soil behind the face,
which can be pursued either by jet grouting treat- The sewer tunnel, which is still under construction
ments or by the insertion of fibreglass elements (bars in the city of Palermo, is almost 5 km long and is
or tubes). characterized by frequent changes of cross section
During excavation, the face reinforcing compo- and abrupt variations of direction, in order to meet
nents are removed together with the soil, while the peculiar hydraulic requirements and to conform to
contour elements are progressively supported by set- the urban features of the city. The tunnel stretch
ting up steel ribs and shotcrete in order to complete considered in the present paper is about 350 m long
the provisional lining of the tunnel. The final lining, and is located downtown Palermo under a busy
made of reinforced concrete, can then be installed at street, named “Corso Re Ruggero” (Fig. 3). This
a later and more convenient time. street is bordered on one side by a subway line
which was previously built by cut and cover.
 clay silt sand gravel

% weigth
alluvium

Figure 5. Grain size distribution of the alluvial soils.

exception of one single edifice founded on rein-

forced concrete piles.
In the second half of the tunnel stretch the street
is bordered by an old masonry retaining wall, stand-
ing over a city park (Fig. 4b). Another relevant
structure is the subway ticket office, which is con-
nected to the subway line through a pedestrian un-
derpass (Fig. 4c). This particular situation is met in
the middle point of the tunnel stretch.
For each span the cross section of the excavation
ranges between the following values: width 5.10-
5.40 m, height 5.40-6.00 m. The soil cover, taken as
the distance between ground level and tunnel crown,
is comprised between 7 m and 9 m.
The tunnel intersects different geologic forma-
tions. At the bottom, just below the invert arch, there
is a continuous thick stratum of dense silty sands.
On top of this stratum there are several layers of cal-
carenitic rocks, characterized by a very variable de-
gree of cementation. The overall thickness of the
calcarenitic layers ranges between 3 and 11m. More-
over, the continuity of the calcarenitic formation is
interrupted by alluvial sediments deposited by the
ancient river Kemonia which bordered the city of
Palermo many centuries ago. The grain size distribu-
tion of these alluvial soils is reported on Fig.5.
Finally, on top of the natural soils, there are rele-
vant layers of made land i.e. materials of anthropic
origin (bricks, pottery, fragments, transported soil,
etc.). The groundwater level is located at the depth
of 10m, practically at mid height between the invert
arch and the vault of the tunnel.
Figure 4. Typical cross sections: a) Micropiles Canopy b) Jet Several continuous borings were performed, at
grouting Canopy c) Steel frame. (M made land; C calcarenites; the design stage, providing the elements for drawing
A alluvial soils; S silty sands; W.L. water level ) the geotechnical profile along the tunnel. However,
due to the remarkable lithological variability of the
Therefore, on this side, the tunnel runs along a sheet subsoil, it was not possible to detect the contact be-
pile wall (Fig.4). On the opposite side of the tunnel, tween the alluvial materials and the calcarenitic for-
three different conditions are met. Along the first mation.
half of the tunnel stretch, there are some residential Moreover only few undisturbed samples of cal-
buildings (Fig. 4a) of various dimensions and char- carenite were retrieved, since these materials are
acteristics, such as masonry and reinforced concrete. very fragile and are thus subjected to extensive
They are all placed on shallow foundations, with the breakage during mechanical boring. However, it was
Figure 6. Tunnel Construction: a) Geotechnical Profile; b) Final settlements; c) Canopy Types (M made land; C calcarenites; A al-
luvial soils; S silty sands)

well known that the degree of cementation of the where excavation was essentially supported by the
calcarenites is very variable, generally low, and pre-existing reinforced concrete slab and piles (see
some times negligible due to their peculiar petro- Fig 4c).
graphical features (Canzoneri et al., 2002).
Design was thus accomplished by dimensioning
three types of canopies, and each type of canopy was 4 CONSTRUCTION AND MONITORING
associated to a typical stratigraphical sequence
(Fig.6). In particular, steel micropiles were pre- Tunnel excavation was recently completed, under
scribed for supporting the vault in the calcarenitic “Corso Re Ruggero”, but the final concrete lining
formation, considering that the cover was mainly has not yet been cast in place at the time of writing.
composed by made land (canopy type 1.a). The mi- During excavation a detailed monitoring program
cropiles were extended down along the tunnel pil- was carried on and the tunnelling process was fol-
lars, where it was expected to intercept the dense lowed step by step for each construction span, hav-
silty sands (canopy type 1.b). Jet grouting was pre- ing an average length of 9m.
scribed instead for the alluvial soils of river Kemo- This observational procedure allowed for proper
nia (canopy type 2). Horizontal drains were pre- design implementation of the design canopies, ac-
scribed for all the tunnel spans. A special steel frame cording to the subsoil conditions met during excava-
was devised for crossing the pedestrian underpass, tion. In particular, the following monitoring activi-
ties were carried on: ‐30 ‐20 ‐10 0
y [m]
10 20 30 40
i. lithological observation of the excavation 0
face and evaluation of water drainage a
ii. sub-horizontal borings to check subsoil con- 20

S [mm]
ditions for subsequent spans 40
iii. deformation measurements of the provisional
lining by means of convergence bolts 60
iv. deformation measurements of the steel ribs
80
by strain gauges
v. direct stress measurements on the steel ribs 100 Span 23 Span 24 Span 25 Span 26 Span 27 Span 28 Span 29
by means of load cells y [m]
‐30 ‐20 ‐10 0 10 20 30 40
vi. topographical observations of the ground sur-
0
face and of the existing buildings
b
vii. inclinometric, assestimetric and piezometric
20
measurements from vertical borings

S [mm]
All the above measurements provided useful con-
40
tribution for checking the design and for prescribing
the most appropriate solutions for each excavation
60
span, according to the principles of the Observa-
tional Method. However, in practice, lithological
80
and topographical observations proved to be the
most valuable means for construction control (Leta
et al., 2007). A typical stratigraphical condition ob- 100 Span 23 Span 24 Span 25 Span 26 Span 27 Span 28 Span 29
served on site is shown by a picture taken at the tun-
nel face (Fig. 7). Figure 8. Surface Settlements (S) vs. Distance (y) from excava-
tion face as recorded at bench mark 8.3 (under retaining wall).
a) Settlement progress b) Data interpolation by cumulate prob-
ability curve.

The final settlements, measured after tunnelling

completion, reached quite different values (Fig. 5b),
corresponding to the various subsoil conditions (Fig.
5a) and canopy types (Fig. 5c).
Buildings settlements were always less than 10
mm and did not cause any problem to the existing
structures. This behaviour was related to the over all
stiffness of the calcarenitic layers. On the other
hand, the settlements of the retaining wall reached a
maximum value of 100 mm causing relevant tilting
towards the park. Such larger settlements were asso-
ciated to the lower stiffness of the alluvial soils of
river Kemonia and to a higher percentage of ground
volume loss.
Figure 7. Lithological observation of excavation face: M made The same data were later analysed in order to in-
land; C1 slightly cemented calcarenite; C2 weakly cemented vestigate the effects of tunnelling in more detail. For
calcarenite; C3 cemented calcarenite. such purpose the settlements recorded at the bench
marks have been plotted versus distance y from the
Topographical data were retrieved and processed in tunnel face at the time of recording. Such distance
real time, allowing for continuous checking of sur- was taken along the horizontal, with reference to the
face settlements. However, measurements of bench orthogonal projection of the bench mark on the tun-
marks placed on the road surface were affected by nel axis.
traffic disturbance which induced excessive scatter- Fig. 8a shows the settlements measured by the
ing. On the contrary, continuous settlement data bench mark which experienced the maximum final
were recorded from the bench marks placed on the settlement. This bench mark, labelled by number
existing buildings and on the masonry retaining wall 8.3, was placed on the previously mentioned ma-
located along the park. These data provided the most sonry retaining wall located between the street and
useful mean for monitoring the near by structures. the park. The plot points out that settlement is pro-
duced not only by excavation of the span located
 y [m] Table 1. Settlements data.
‐60 ‐40 ‐20 0 20 40 60 80 Bench Smax H k Soil Type Canopy Type Existing k Smax
Mark mm m Structures mm
0 1,2 3,5 11,7 4,3
Bench 
a Marks 2,1 3,5 11,7 3,9
3.1 2,2 3,9 11,7 4,5
0,2 3.2 2,3 9,4 11,2 2,9
3.3 1.a
S / Smax

3,1 10,6 11,2 2,3

Calcarenite
3.4 Buildings 2,7 7,7
0,4 3.5 3,2 9,9 11,2 2,2
8.1 3,3 9,7 11,2 2,0
8.2 3,4 9,9 10,6 1,4
0,6 8.3 3,5 9,4 9,6 2,2
8.4 1.b
4,1 6,9 9,6 1,3
8.5
4,2 2,6 10,0 0,8
0,8 8.6
7.2 4,3 2,2 9,9 1,2 Steel frame Underpass 1,0 2,5
4,4 2,6 9,9 1,1
7,2 70,3 9,7 1,1
1
7,3 67,6 9,8 1,0
y [m] 8,1 71,3 9,7 0,9
‐60 ‐40 ‐20 0 20 40 60 80 8,2 83,4 9,7 0,8

Alluvium
8,3 95,9 9,7 0,8
0 2 0,9 64,0
8,4 99,2 9,7 0,9
b k values 8,5 68,2 9,8 0,8 Retaining
0,2 2,7 8,6 39,2 9,8 0,9 wall
2,2 8,7 26,3 9,8 0,8
1,0 8,8 19,0 9,8 0,9
S / Smax 

0,4 0,9  8,9 15,5 9,9 2,8

Calc.
8,10 11,0 10,0 1,4
1.a 2,2 12,9
8,11 13,0 10,2 1,0
0,6 8,12 12,2 10,3 3,6

0,8
k
0 1 2 3 4 5
1
0

Figure 9. Settlements ratios for all bench marks. a) Recorded 20

data b) Average settlement ratios (k=2.7 - building settlements
Smax [mm]

on calcarenite; k=2.2 - retaining wall settlements on cal- 40

carenite; k=1.0 - under pass settlements on steel frame; k=0.9 -
retaining wall settlements on alluvial soils).
60 1 (buildings  / calcarenite)
2 (underpass  / calcarenite)
directly under the check point but also by excavation
80 3 (wall / alluvium)
of the preceding and of the following spans. It can
4 (wall / calcarenite)
also be appreciated that, when excavation is ap-
100
proaching the bench mark (i.e. negative distance on
plot) the settlement proceeds not only during exca-
vation but also when the tunnel face is immobile Fig. 10 Maximum settlements and k values for different com-
binations of structures and subsoil characteristics.
(dotted lines between excavation spans). A similar
behaviour was observed for all bench marks.
However, if these discontinuities are disregarded, Clearly, the results are scattered because of the vari-
the measured settlements can still be interpolated by able conditions encountered along the tunnel stretch.
a Gauss cumulate probability curve (Fig. 8b) as first However these data can be further analysed by con-
suggested by Peck (1969) and later confirmed by sidering four different combinations of structures
several authors (e.g. O’Really and New, 1982; and subsoil characteristics, listed as they were met
Nyren et al. 2001). Moreover, if the settlements S during tunnelling:
recorded at each bench mark are scaled with respect 1. buildings on calcarenite (canopy 1.a/1.b)
to their maximum value Smax, then all the data can be 2. underpass on calcarenite (steel frame)
grouped in a single plot as reported on Fig. 9a. 3. retaining wall on alluvial soils (canopy 2)
The parameter k was estimated for each bench 4. retaining wall on calcarenite (canopy 1.a)
mark according to the well known relation For each of the above groups, average values of
Smax and k are reported in Table 1 and average set-
y − y2
S 1 tlement trends are drawn in Fig. 9b. Moreover Smax
∫e
2
= 2( kH )
dy values are plotted versus k values in Fig. 10, in order
S max kH 2π −∞
(1) to provide a general picture of the observed settle-
ments.
The k values, obtained by numerical best fitting By comparing data groups 3 and 4, the influence
of the recorded data, are reported on Table 1 to- of subsoil properties can be singled out. In fact, in
gether with the maximum settlement Smax and the these two cases, the influence of the masonry wall
depth H of the tunnel centre line. (see Fig. 4b) is equal and probably very limited due
to its poor rigidity. It is thus confirmed that the Clearly, more observations are needed in order to
amount of settlement is mostly related to the subsoil gain a better understanding of movements induced
stiffness, higher for calcarenites and lower for allu- by canopy technique tunnelling In fact only few
vial soils, even though relevant data scattering is ob- cases on this topic have been published so far (e.g.
served particularly for the alluvial soils. Some scat- Croce et al. 2004, Russo & Modoni, 2006). Some
tering is also observed for buildings settlements on new settlement data gained by other case histories of
calcarenite (data group 1) due to their variable struc- canopy tunnelling are going to be reported in the
tural and foundation features (Standing, 2006; van near future (Croce et al., 2011).
Tol, 2006). However, in the average, buildings set-
tlements are smaller than those of the masonry wall
on calcarenite. This difference can be attributed to REFERENCES
their higher structural rigidity and also to their larger
distance from the tunnel axis (see Fig. 4a). The un- Canzoneri V., Giammarinaro M.S., Gugliuzza G., & Val-
derpass settlements are even smaller due to the very lone P. 2002. “Area urbana palermitana: analisi delle
high rigidity of this piled structure (see Fig. 4c). proprietà geotecniche dei terreni supportata da tecno-
Finally from Fig. 9b it can be seen that the set- logia GIS”. Atti del XXI Convegno Nazionale di Geo-
tlement curves are steeper for the alluvial soils with tecnica, L’Aquila 2002.
respect to the calcarenitic formation. It also appears Croce, P. 2010. Application of the Observational Method
that the higher rigidity of the buildings, compared to in Urban Tunnelling. 14th Danube-European Confer-
that of the masonry wall, provides a more elongated ence on Geotechnical Engineering. 2 – 4 June, Brati-
shape to the length-settlement curves. slava, Slovak Republic.
Croce P., Modoni G. & Russo G. 2004. Jet-grouting per-
formance in tunnelling. Geosupport 2004. J.P.Turner
5 CONCLUSIONS
& P.W. Mayne eds. ASCE Geotechnical Special Pub-
The project of a shallow sewer tunnel in difficult site lication No. 124. pp. 910-922.
conditions has been successfully accomplished by Croce P., Russo G. & Spacagna R.L. 2011. Settlements
means of the canopy technique, making proper use induced by tunneling in clay shales. XV European
of different soil improvement methods: steel and fi- Conference on Soil Mechanics and Geotechnical En-
breglass micropiles, jet grouting, horizontal drains. gineering. Athens. September 12-15.
In particular, design was carried on by dimensioning Leta M., Barba D., Canzoneri V., & Giammarinaro M.S.
three different types of canopies, each one associ- 2007. Public works: an important resource to upgrade
ated to a typical stratigraphical sequence. territorial knowledge. VI Forum Geoitalia. Rimini
During excavation a detailed monitoring program 2007.
was organized and the tunnelling process was fol- Patel D., Nicholson D., Huybrechts N., & Maertens J.
lowed step by step for each construction span. This 2007. The observational Method in Geotechnics. Pro-
observational procedure allowed for proper imple- ceedings of XIV European Conference on Soil Me-
mentation of the design canopies according to the chanics and Geotechnical Engineering: Geotechnical
subsoil conditions met during excavation. engineering in urban environments, Madrid. vol. 2.
Settlement recording was the most effective Nicholson D., Tse C., & Penny C. 1999. Observational
means for monitoring the near by structures in real Method in ground engineering – principles and appli-
time and provided also useful data for investigating cations. Report 185. CIRIA. London.
the effects of tunnelling afterwards. It was found that Nyren R.J., Standing J.R., & Burland J.B. 2001. Surface
the final settlements measured after tunnel comple- displacements at St James’s Park greenfield reference
tion reached very different values, corresponding to site above twin tunnels through the London Clay.
the various subsoil conditions and canopy types. Case studies from construction of the Jubilee Line
It was also observed that settlements are pro- Extension. London. Thomas Telford. Pp. 387 – 400.
duced not only by excavation of the span located di- O’Really M.P., & New B.M. 1982. Settlements above
rectly under the check point but also by excavation tunnels in the United Kingdom – their magnitude and
of the preceding and of the following spans. In addi- prediction. Tunnelling ’82. London. IMM, pp.173-
tion it was appreciated that settlements proceed not 181.
only during excavation but also when the tunnel face Peck, R.B. 1969. Deep excavation and tunneling in soft
is immobile. However, by careful processing of all ground, Proceeedings of VII ICSMFE, Mexico City.
the settlement data, it was observed that the relation State of the art volume. pp. 225-290.
between settlement ratio and distance from the tun- Russo G., & Modoni G. 2006. Monitoring results of a
nel face still follows, with acceptable approximation, tunnel excavation in an urban area. Geotechnical As-
the typical shape of a Gauss cumulate probability pects of Underground Construction in Soft Ground.
curve as repeatedly observed for TBM tunnelling. Bakker et al. eds. Taylor and Francis. London.
Standing J.R. 2006. Monitoring ground and structural re-
sponse to underground construction works. Geotech-
nical Aspects of Underground Construction in Soft
Ground. Bakker et al. eds. Taylor and Francis. Lon-
don. pp. 67-79.
van Tol A.F. 2006. The effects of tunnelling on existing
structures. Geotechnical Aspects of Underground
Construction in Soft Ground. Bakker et al. eds. Taylor
and Francis. London. pp. 31-41
Viggiani G., & Soccodato F.M. 2004. Predicting tunnel-
ling induced displacements and associated damage to
structure. Rivista Italiana di Geotecnica XXXVIII, n.4,
pp.11-25.

View publication stats