Design of Bolo Tunnels - Parte-1

Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

102.2-R-133 / 12.

37

Other main consequences evidenced by the damage survey where:


1. Permanent increment of axial loads in lining that has been observed at monitored sec-
tions following to the Düzce event.
2. Important damages at shotcrete and temporary support were observed in soft ground
stretches, where poor mechanic characteristics soils were encountered.
3. Among similar ground stretches, cracks at invert were recorded where invert was not
reinforced.
Is worth here noting that the attained earthquake greatly exceeded the design earthquake in
terms of PGA (0.81g against 0.4g). Also, following to the Izmit and Düzce events the mo-
torway owner (Turkish General Directorate for Highways) required a revision of the design
of the entire motorway to deal with a design seismic event of PGA of 0.81 g.

4. ANALYSES PERFORMED FOR BOLU TUNNEL


A two phases analysis has been performed for the Bolu tunnel seismic upgrade:
1. A “seismic screening” basing on simplified assumptions and elastic model,
2. A detailed 2-D analysis by Finite Difference Method considering the real behavior of
the soil and shape of the structure.

4.1 Seismic screening


A “seismic screening” has been performed to evaluate critical sections in which more de-
tailed analyses needed to be performed. The procedure, detailed in ref. 3, consists in apply-
ing an estimated seismic shear strain1 to the soil surrounding the tunnel, and to deduce,
moving from closed form solutions (i.e. hole in elastic half space) the thrusts, bending mo-
ments and shear forces induced in the lining. This method aims to estimate the maximum
seismic load undergone by the tunnel lining.

Approximations driven by such an approach are mainly:


1. Shear strain in the subsoil only estimated, here the PGV has been estimated to 0.8 m/s
by the statistical correlations for stiff soils (Shear wave velocity between 200 and 750
m/s).
2. The soil is considered elastic, (i.e. eventual rupture effects cannot be evaluated)
3. Layering of the soil is not fully taken into account
4. Tunnel shape is not modeled exactly (the tunnel cross section is assumed circular to
adopt closed forms solutions).

1
A wide literature is available about the methodology to evaluate the seismic shear induced
in the subsoil by a seismic event. A rough estimation is the ratio between shear wave veloc-
ity Cs and peak particle velocity at tunnel depth:
Vs G
γ where Cs  Vs is the peak ground velocity at tunnel depth
Cs ρ
ρ is the density of the soil
G is the shear modulus of the soil
A better evaluation of the induced maximum shear strains is attained by performing a linear
equivalent analysis on a resonant column modeling the soil stratigraphy.

-5-
102.2-R-133 / 12.37

The original tunnel design followed the NATM principles, proceeding with shotcrete, rock
bolts and light steel ribs. While excavating, following to major tunneling difficulties, new
heavier cross sections needed to be added to the original design.
The excavation was proceeding from both portals, western (Asarsuyu) and eastern (El-
malik). On 12.11.1999, stretches of about 1670 m and 730 m were already excavated re-
spectively from Asarsuyu and Elmalik portals. A stretch of about 960 m was still to exca-
vate and a reconnaissance gallery was excavated on the right tube (Elmalik face) whose
face was only about 380 m from the opposite face.

3. DAMAGES DESCRIPTION
The area suffered the two major Turkish earthquakes of 1999 (see figure 4 for a seismic
map of the area). Kocaeli (Izmit) Earthquake struck on 17.08.1999, with an Mw 7.4. The
epicenter was located about 150 km west of the site. Close to tunnel site a PGA of
0.25÷0.35 g was recorded. No major damages were suffered by the tunnel.

Figure 4 – Seismic map of Turkey close to Bolu area

Düzce Earthquake struck on 12.11.1999, the magnitude (Mw 7.2) was lower than Izmit one,
but the epicenter was only 20 km west of the site. PGA and PGV of 0.81 g and 0.66 m/s
were experienced at site; the surface rupture extended to 3 km from western portal. The
tunnel suffered major damages and collapsed in a stretch of almost 400 m (figure 3).
The heavier damages have been recorded at Elmalik (Ankara) side where the above men-
tioned total collapse occurred in both tubes.
Also at Asarsuyu side damages to lining and invert uplift (in non reinforced stretches) were
observed. The excavation, where a major fault gauge was being crossed, was proceeding by
advancing two 5 m diameter pilot tunnels at benches (OPTION 4). These last were badly
affected by the earthquake: the ribs buckled and invert heaved, requiring heavy re-profiling
and reparations.

-4-
102.2-R-133 / 12.37

The induced shear strain at the tunnel level depends on a huge amount of factors (i.e.: layer-
ing of the medium, shear modulus and density of the layers, damping ratio, inclination of
layers boundaries, etc…). In particular, in a layered medium reflection and refraction of
seismic waves can induce concentration of shear strains at certain levels. Therefore the
evaluation of the shear strain induced by an earthquake at a certain depth cannot neglect the
layering of the soil stratigraphy.

2. PROJECT DESCRIPTION
A portion of the Trans European Motorway (TEM) network, the Turkish motorway linking
Ankara to Istanbul is currently under construction. The crossing of the Bolu Mountain is
included in the 120 km section between Gümüşova and Gerede. The last 25 km stretch, the
most challenging, runs parallel to the North Anatolian Fault. This stretch, under construc-
tion, includes several viaducts and the Bolu tunnel.
The Bolu tunnel is a 3360 m long twin bored three lanes tunnel.
The average excavated radius is 8 m, corresponding to an x-section area between 190 and
260 m2. The width of the ground pillar between tubes varies from about 28 m at portals to
almost 48 m. Depths are up to 250 m; on 86% of the total length, overburdens are higher
than 100 m and on 48% higher than 150 m.

In Figure 3 the Bolu tunnel longitudinal profile is sketched and the lithology encountered is
indicated. A wide range of soils is represented, mainly highly tectonised series of mud-
stones, siltstones and limestones, fault gouge clays. The consistence of the soil varies from
competent rock requiring blasting to very weak clayey zones where heavy advancement
problems were encountered.

Figure 3 – Bolu Tunnel - longitudinal profile

-3-
102.2-R-133 / 12.37

Seismic Loads to Structures due to Ground Shaking


Surface structures Underground structures
The soil surface vibrates and the struc- The seismic waves propagating in the
Phenomenon
ture oscillates causing inertial forces medium induce drifts in the structure.
Load introduced by seismic forces Load introduced by seismic strains
Loads
(proportional to accelerations) (proportional to velocities)
In common problems is possible to ne- Not possible to neglect soil-structure
Interaction
glect soil-structure interaction interaction
In common problems linear analyses
Analysis Highly non-linear approach
are possible
Table 1 – Differences in Seismic Analyses for underground and surface structures

While longitudinal loads (figures 1a and 1b) are generally not critical for overall stability
(mainly radial cracks are induced and can be partly absorbed by radial joints in lining), the
ovaling of the cross section is critical, and, affecting the integrity of the cross section, can in-
duce the failure of the lining.

Tension

Compression

Figure a – Axial Compression Figure b – Axial Bending Figure c – X-Sect. Ovaling


Figure 1 - Seismic load on tunnel linings (Ground Shaking)

The ovaling of the cross section is induced by shear waves propagating within the soil from
the vibrating bedrock toward the surface. The phenomenon can be sketched as in Figure 2.

Figure 2 – Drifts induced by seismic shear.

-2-
102.2-R-133 / 12.37

SEISMIC DESIGN OF UNDERGROUND STRUCTURES


THE BOLU TUNNEL

Walter AMBERG
Chief of the Technical and Scientific Department

Marco RUSSO
Engineer of the Technical and Scientific Department

Lombardi SA Engineering Limited


Via Simen 19
CH – 6648 Minusio
Switzerland

ABSTRACT: Traditionally low seismic risk is associated to tunnels and buried structures.
Nonetheless collapses or heavy damages in underground structures have been reported fol-
lowing to Kobe and Loma Prieta earthquakes. More recently, following to Düzce earthquake,
in Turkey, an almost 400 m long stretch under construction of the Bolu tunnel collapsed con-
firming that, in soft soils or close to important seismic sources, a seismic analysis of tunnels is
advisable. Here the procedure followed for seismic upgrade of the design of the Bolu tunnel is
exposed together with some main results.

1. GROUND SHAKING INDUCED LOADS FOR UNDERGROUND STRUCTURES


Main causes of seismic risk for underground structures can be liquefaction, active faults
crossings, seismically induced landslides and ground shaking intensity. Damages induced at
Bolu tunnels by the 12th November 1999 Düzce earthquake where mainly due to the ele-
vated intensity of ground shaking.
Below the design procedure followed for Bolu Tunnel against ground shaking loads is
summarized together with some preliminary outstanding results.
Concerning loads induced by ground shaking, some differences exist between above and
below the ground structures: while surface structures are loaded mainly by inertial forces on
the basis of structural amplification, in underground structures, due to the high grade of
constraint, very low inertial forces are experienced. In table 1 the major conceptual differ-
ences in handling the two phenomena are resumed.
Seismic loads due to pure soil shaking in underground structures are thus induced by the
relative displacements caused in the medium by the seismic wave propagation.
The possible loads induced by seismic waves on tunnels are due to:
− Seismic waves propagating in the longitudinal direction of the tunnel (see Figure 1a and b).
− Seismic waves hitting the tunnel in the plane of the cross section (see Figure 1c).

-1-

You might also like