Professional Documents
Culture Documents
Keynote Lecture
Historic buildings
Street level Soft clay First sand layer Second sand layer Stiff clay 6.5m diameter tunnels
Adv anc e
Tu nn
el
TUNNEL
Pre-convergence
In TBM tunnels the fraction varies significantly (< 70%) depending on the method
Face extrusion
Conclusion : In non-TBM tunnels, control of pre-convergence (face extrusion) is critical in urban tunnelling
Control of pre-convergence is contrary to the basic NATM principle of mobilising rockmass strength by deformation This NATM principle is mainly applicable in mountain tunnels
Mountain tunnels : Stability is critical Deformation not critical (usually desirable) Urban tunnels : Deformation critical : to be minimised Stability is ensured by controlling deformation
Support load
(uR D)
Face pre-treatment
Urban tunnelling methods : TBM tunnelling Control of pre-convergence by face pressure and ground conditioning in closed-face machines
Slurry shield
p
EPB shield
screw conveyor
Urban tunnelling methods : TBM tunnelling Control of pre-convergence by the size of cutter-head openings in open face machines
Athens Metro (1998) Urban tunnelling methods : NATM tunnelling (North of Alps) Control of pre-convergence by multi-drifting (uR D)
Urban tunnelling methods : NATM tunnelling (North of Alps) Control of pre-convergence by multi-drifting (uR D)
1 1
1 1
Excavation reduces 3 to zero causing face instability. Forepoling : The presence of a stiff beam reduces the major (vertical) stress (1) on the face
arch
Athens Metro : Monastiraki Station (18m wide span) micro-tunnel pipe arch (bicycle chain)
1. Face protection methods : Reduction of 1 ahead of tunnel face 1.3 Vertical nails (or piles) from ground surface
ATHENS METRO
3 =
n Fy P = A ( FS F ) A
2
FSo =
(1 ) N s
1 = (1-) po
Ns =
2 po
cm
po = geostatic stress
Factor of safety with FG-nails :
FS = FSo +
1 3 2 tan 45 + (1 ) 2 p o
FSo =
(1 ) N s
1 = (1-) po
Ns =
FS = FSo +
2 po
cm
2 c tan 45 + (1 ) 2 p o
Athens Metro Athens Metro : Ground improvement ahead of TBM (via a pilot tunnel) using fiber-glass anchors and TAM grouting
Continuum models
Intact rock strength controls response
Discrete models
Structural features control response
Continuum models
Rockmass strength controls response
Typical numerical analysis using computer programs : UNWEDGE (for tunnels) SWEDGE (for slopes)
2-D analysis of tunnel face stability: UDEC Results Kamata & Mashimo (2003)
Modelling stages are direct : 1. Geostatic (initial conditions) 2. Installation of face support 3. Advancement of the excavation (one step) 4. Installation of side support 5. REPEAT steps 34 until new face support 6. Install face support ..
However : Input preparation and output presentation is often complicated Analysis is time consuming Improved accuracy may be incompatible with the level of knowledge of ground conditions
Use of 3-D FE/FD models for face pre-treatment : Modelling face treatment Constitutive model (E-sensitive analyses) Knowledge of input ground parameters (E)
Purely elastic response for Ms > 4 uy,max = maximum extrusion (at tunnel face)
u y ,max D
unlined tunnel
Spyropoulos, 2005
= 0.0004 M s2.25
Ms = E 1000 H 0.90 D 0.10
1.00 1.20 1.40
0.10
Maximum extrusion uy,max (at tunnel face) as a function of the controlling ground parameter Ms. Extrusion is not influenced by the installation of shotcrete lining (thickness t) behind the face (distance L) correlation uy,max & Ms is useful E
Purely elastic response for Ms > 4 uz,max = crown settlement (at tunnel face)
Spyropoulos, 2005
unlined tunnel
Ms =
0.10
Crown settlement uz,max (at tunnel face) as a function of the controlling ground parameter Ms. Crown settlement is strongly influenced by the installation of shotcrete lining (thickness t) behind the face (distance L). Crown settlement cannot be used to assess the value of Ms ahead of the tunnel face
uy,max = maximum extrusion (at tunnel face) uy = extrusion at distance (x) from tunnel face
Spyropoulos, 2005
uy u y ,max
= (1 + e
x 2 x R R
Extrusion uy as a function of the distance from tunnel face. Since the value of uy,max is related to Ms correlation uy & Ms (for any x/R) is useful
Spyropoulos, 2005
n = number of FG-nails F = mean axial force in FG-nails A = tunnel section area H = vertical overburden
f |G =
Ms =
nF A H
Spyropoulos, 2005
without FG-nails unsupported without FG-nails supported with shotcrete fG = 4000
H = vertical overburden
f |G =
nF A H
Ms =
Crest settlement is only slightly reduced by installing FG-nails (and any reduction is masked by the shotcrete liner)
Spyropoulos, 2005
2.5 without forepoles
fF1 =1
fF1 = I / S (cm3), I = moment of inertia of a forepole tube S = axial distance between forepoles L = length of forepole overlap D = tunnel diameter
1.5
1.0
Ms =
fF1 =100
fF1 =50
0.5
0.8
fF1 =1
fF1 = I / S (cm3), I = moment of inertia of a forepole tube S = axial distance between forepoles L = length of forepole overlap D = tunnel diameter
uz(x=0) / D (%)
0.6
0.4
Ms =
fF1 =100
0.2
Spyropoulos, 2005
0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6
Ms
0.7
fF1 < 20
Zone 2 supported
Zone 1 unsupported
()
()
()
= deconfinement ratio
= 1
=1
p p = po (1 ) po
()
()
()
p = (1 ) po
po = geostatic stress (isotropic)
(1 2 ) (1 ) E= Eo ( ) + 1 2
Eo = ground E-modulus
Example :
Example :
=0.70 p = 30% po
=0.70 E = 10% Eo
Advantage : Good in anisotropic fields
p/po
0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10
Values of / for = 0.30 0.533 0.400 0.300 0.222 0.160 0.109 0.067 0.031
= 1 - p/po
E (1 2 ) (1 ) = (1 2 ) + Eo
uR(p)
uR(x)
2-D model Calculation method : 3-D model : uR = uR(x) 2-D model : uR = uR (p) or uR = uR () Thus :
= (x)
Determination of the deconfinement ratio () along the tunnel axis FLAC-3D : Spyropoulos, 2005
-4 0 -0.1 u R / u R,max (crest settlement) -0.2 -0.3 -0.4 -0.5 -0.6 -0.7 -0.8 -0.9 -1 -3 -2 -1 0 1
= f ; Ms
2 3
x R
x/R 4
FLAC3D
uR 0.37 x 1 exp 2 . 2 = + M s u R R
1.2
Ms =
TUNNEL
Conclusions
1. Ground deformations are critical 2. Estimates of ground deformations require 3-D numerical analyses ( + ground model + ground properties) 3. Relevant ground properties (mainly E) can be obtained by measurement of face extrusion & numerical back-analyses (or use of the normalised graphs) 4. For many tunnel designers, 3-D analyses may seem too sophisticated : Methods exist to analyse the problem in 2-D using the deconfinement method () Normalised graphs are available to estimate () in tunnels without / with face pre-treatment