Proceedings, 11th FIG Symposium on Deformation Measurements, Santorini, Greece, 2003.

TUNNEL MONITORING DURING THE EXCAVATION PHASE:

3-D KINEMATIC ANALYSIS BASED ON GEODETIC DATA
Villy Kontogianni and Stathis Stiros
Lab. of Geodesy, Dept. of Civil Eng., Patras University, Patras 26500, Greece

Abstract
Tunnel excavation produces a re-distribution of stresses in the rock and a tendency for
closure of the void produced a tendency counteracted by the tunnel support shell. In some
tunnels a very stiff support shell not permitting any ground deformation is selected. In other
tunnels, excavated on the basis of the popular so-called NATM a quite different approach is
adopted: A controlled deformation of the ground is permitted (i.e. limited closure of the rate of
approximately 1%) and this gives the opportunity to the stresses to be partly released and a less
stiff and less-expensive support system to be used. In this case, an accurate, systematic and
continuous monitoring of the tunnel deformation is absolutely necessary during the excavation.
In the last years, the introduction of modern total station instruments provided a simple, low
cost and functional way to record the absolute 3-D displacements of a large number of
controlled points (usually 3-5 points every approximately 10-20m). Such data from a number of
tunnels were used to study: i) the distribution of deformation in the various tunnel sections, ii)
the variation of maximum section closure along the tunnel axis, iii) the time-history of
deformation and iv) the relationship between closures and ring closure delays.

1. Introduction
Tunnel openings tend to close under the geostatic stresses of the surrounding rock mass and,
depending on the ground characteristics, this effect may cause either reduction of the excavated
area or collapse. Thus, a basic requirement for the excavation of a successful tunnel is the
control of section closure or of its instabilities, with the establishment of a detailed monitoring
system to record tunnel deformation during the excavation and indicate the need for additional
measures if convergence exceeds certain safety levels.
Furthermore, modern tunnel excavation is based on the principle of controlled deformation for
permitting partial stresses relaxation and installation of a less stiff and more economic support.
According to NATM (Kovari, 1993) and other techniques (shallow tunnelling, cut and cover
etc.) monitoring is crucial for selecting the proper support system and provide a cost-effective
solution to the requirement for stable tunnel sections.
In the past, deformation control was based on INVAR tapes and wires (Kovari and Amstad,
1993), while the contribution of geodesy to the tunnels construction was mainly limited to the
optimization of the breakthrough accuracy (Kienast, 1995). Tapes provided a high accuracy
(~0.13mm/10m; Dunnicliff, 1993) in deformation measurement but were expensive and timeconsuming and could only measure changes in the distance between selected points fixed on the
tunnel periphery. In the 1980s, however, the introduction of modern electronic total stations
and levelling instruments made surveying very popular for many types of projects, including
underground excavations. In the last years, geodetic techniques are widely used to tunnel
excavations and provide a simple, inexpensive and functional way to control tunnel deformation
during tunnel construction with an accuracy of a few mm which is very satisfactory for all
projects.

EDM
targets
measuring stations
Figure 1: Arrangements of targets and of EDM measuring stations along the tunnel axis.

2. Geodetic monitoring of tunnel deformation

Total stations are placed on stations along the tunnel axis and record the 3-D coordinates of all
targets fixed on the primary tunnel lining (support shell; Fig.1), shortly after the excavation with
an accuracy of a few mm (or a few cm for >5km long tunnels). Geodetic measurements are
taken periodically (usually once per day) until stabilization of the height of all targets is
obtained (usually during a two-month period; Kontogianni et al., 1999).
Coordinates of stations are measured relatively to one or two stations outside the tunnel, on
stable grounds, and all measurements refer to a local coordinates system suitable for the project.
Recording measurements are transferred automatically to the computer at the end of each
measurement epoch.
The large amount of geodetic data permits to estimate the absolute displacements of the targets
and the final deformed profile of the sections. However, a more detailed analysis of the
monitoring data may provide substantial information on the kinematics of the support system
due to the imposed stresses and to the selected excavation procedure.

3. Kinematic analysis based on the geodetic data

The tunnels support shells react to the growing stresses of the surrounding rock mass tending to
reduce the opening and this reaction induces a time-dependent deformation of the various
sections. An attempt to explain the relationship between ground and lining requires the detailed
recording of their interaction and the analysis of the monitoring data in terms of timedependency, tunnelling method adjustments and local conditions. The geodetic observation of
tunnel deformation provides several information concerning the behaviour of the excavated area
and permits the understanding of the kinematics of the support shell. The most important
parameters of shell kinematics are: i) the distribution of deformation in the various tunnel
sections, ii) the variation of maximum convergence along the tunnel axis, iii) the time history of
deformation and iv) the relationship between convergence and ring closure delays

3.1 Distribution of deformation in the various tunnel sections

A main characteristic of the observed deformation is that it does not correspond to a uniform
radial section closure, but the amplitude of radial displacement is variable along the tunnel
periphery. Geodetic monitoring provides absolute displacements for at least 3 points at each

top-heading (i.e. upper part) section (one at the crown and two at the sidewalls) and 2 points at
each bench (lower part) section (one at each side). Thus, the deformed profile of the excavated
area can be obtained. An example is the measured non-uniform closure of the Tymfristos tunnel,

Figure 2: Representative deformed profiles from Tymfristos tunnel, Greece, a long time (1
year) after the excavation. Real scale convergence is indicated by black.
central Greece: at least nine targets were measured at each monitored section both prior to and
after deformation. Displacements of all points were plotted on the tunnel section giving the final
excavated area of the tunnel at several positions along the tunnel axis. Two selected deformed
sections for Tymfristos tunnel, compiled on the basis of the survey data, are shown in Fig.2.

3.2 Variation of maximum convergence along the tunnel axis

In many tunnel cases a swell-type pattern of tunnel closure along axis is observed for areas of
high deformation. In those cases there is evidence that deformation did not gradually develop at
each section, but, on the contrary, it developed at certain sections and then propagated bilaterally along a distance of several tens of meters, affecting neighbour sections. This
mechanism for destabilization of previously stabilized adjacent sections and propagation of
deformation along the tunnel axis has been observed in many cases by plotting maximum
section closure along the tunnel axis. An example is the Bolu tunnel, north Turkey (Fig.3;
Dalgic, 2002), where at locations of poor ground conditions and stoppage of excavations,
extreme deformation occurred and spread to adjacent sections. A similar phenomenon occurred
at the Frejus tunnel, France (Lunardi, 2000), where due to a temporary excavation stoppage,
deformation increased at the tunnel face and propagated at both directions of the tunnel.

3.3 Time-history of deformation

One of the most essential parameters to test the effectiveness of the excavation method and the
selected lining is the support shell deformation versus time. The continuous geodetic monitoring
provide a completive set of data for plotting the deformation / time curves. The typical pattern
of a time-dependent deformation can be represented by a curve of cumulative convergence
which reaches asymptotically the final (stabilization) value after a short (~1 month) time period.
vertical displacement (cm)

distance along tunnel axis (m)

0

500

1000

1500

2000

2500

3000

0
2
4
6
8

Figure 3: Local high convergence at points of high deformation affected adjacent tunnel
segments as well, producing a swell-type deformation along a distance (data from Bolu
tunnel, Turkey).

Vertical Convergence (mm)

500

1.

2.

3.

4.

5.

400
300
200
100
0
0

100

200

300

400

500

Time (days)
Figure 4: Resumed deformation after a stabilization period (between 170 and 240 days) for
one section of the Kallidromo tunnel, Greece
However, some parts of some tunnels follow a different pattern of convergence. This last pattern
reveals that some sections stabilize temporarily, following the typical stabilization curve
described above, but some time after (a few days to a few months) deformation resumes and
stabilizes to higher levels. This is a major threat for the tunnel construction and stability. One
tunnel case of extreme closure, measured in detail by geodetic data is the Kallidromo railway
tunnel (central Greece), in which large time-dependent deformations occurred and made
necessary its re-excavation three successive times (Fig.4).

3.4 Closure in relation to ring closure delays

In most tunnels convergence ceased when the opening was fully supported and consequently
deformation was stabilized to low levels. In some cases, however, the hysterisis between section
excavation and ring closure (i.e. installation of the full support system around the tunnel
periphery) played a significant role to the stability of the tunnel: in case of loose rock masses a
long delay in ring closure may lead to extreme, unexpected deformation cumulating for a long
time, even after the full support of the section.
However, the detailed geodetic recording of the tunnel closure time-history and the
behaviour of the opening to the different stages of the excavation sequence is sometimes a
warning to modify the excavation procedure and avoid large deformation for long times. This
was the case with most sections of the Tymfristos tunnel during its re-excavation phase (Fig.5;
Tsatsanifos et al., 1999). The belayed ring closure stimulated further ground deformation;
extreme closure affected the stability of tunnel sections.

time (days)

vertical displacement (m)

10

20

30

40

0
0,1
0,2
0,3
0,4

Figure 5: Vertical displacement of the crown for one representative section of the Tymfristos
tunnel versus time. Arrow indicates invert (floor of tunnel) excavation. A few days later ring
was closed (i.e. the tunnel periphery at this section was fully supported by steel rings and
concrete) but deformation continues to cumulate.

4. Conclusion
Since the excavation of the first tunnels (Kienast, 1995) it was realized that deformation is a
main factor controlling the failure and cost-effectiveness of underground excavation. In the last
two decades, however, deformation monitoring, mostly based on geodetic methods, became a
fundamental requirement for assessing the stability of underground openings and for
quantifying the risk of unacceptable rock response.
Except for safety control, geodetic monitoring provides a wealth of data describing the 3-D
kinematics of the support shell and the time-history of deformation depending on the excavation
technique parameters (velocity, delays etc.). Such information can be input to back analyses for
improving the geotechnical models of tunnel excavation and optimize the excavation process.
Thus, it may lead to a more safe and economic project.

Acknowledgements
This article is a contribution to the research project PENED 2001 (GSRT). We are indebted to
Attiko Metro S.A. and ERGOSE S.A. for providing unpublished data.

References
Dalgi, S. (2002). Tunnelling in Squeezing Rock, the Bolu Tunnel, Anatolian Motorway,
Turkey. Engineering Geology, 67(1-2), 73-96.
Dunnicliff, J. (1993). Geotechnical Instrumentation for Monitoring Field Performance. John
Wiley & Sons, Inc., U.S.A., 577pp.
Kaiser, P. (1993). Deformation Monitoring for Stability Assessment of Underground
Openings. Compressive Rock Engineering, Bergamon 4, 607-630.
Kienast, H. (1995). Die Wasserleitungdes Eupalinos auf Samos. Deutsches Archaeologisches
Institut, Bonn.
Kontogianni, V., Tesseris, D. and Stiros S., 1999. Efficiency of geodetic data to control tunnel
deformation. Proc. of The 9th FIG International Symposium on Deformation
Measurements. Olsztyn, Poland. 206-214.
Kontogianni, V., and Stiros, S. (2002). Shallow tunnel convergence: predictions and
observations. Engineering Geology, 63 (3-4), 333-345.
Kovari, K. (1993). Erroneous concepts behind NATM. Lecture given at the RabcewiczGeomechanical Colloquium in Salzburg.
Kovari, K., and Amstad, C. (1993). Decision making in tunnelling based on field
measurements. Compressive Rock Engineering, Bergamon 4, 571-606.
Lunardi, P. (2000). Design and constructing tunnels ADECO-RS approach. Tunnels and
Tunnelling International special supplement (http://www.rocksoil.com/).
Tsatsanifos, C. P., Mantziaras, P.M., and Georgiou, D. (1999). Squeezing rock response to
NATM Tunnelling. A case study. Proc. of the Int. Symp. on Geotechnical Aspects of
Underground Construction in Soft Rock. Japan, 167-172.