Characterization of Fault Zones

Workshop "Characterization of Fault Zones"
EXTENDED ABSTRACTS
th
09 of October 2013
Salzburg Congress, Austria
sterreichische
Gesellschaft fr
Geomechanik
A-5020 Salzburg
Innsbrucker Bundesstr. 67
T +43 (0)662 87 55 19
F +43 (0)662 88 67 48
E [email protected]
www.oegg.at
sterreichische Gesellschaft fr Geomechanik

Austrian Society for Geomechanics
Workshops 2013
October 9 2013, Salzburg, Austria
th
Characterization of Fault Zones
CONTENT
Characterization and Classification of Fault Zones
F. Fasching & R. Vanek
Ingenieurgeologische Charakterisierung von Strungszonen im Innsbrucker Quarzphyllit

A. Tchterle, H. Schierl & C. Reinhold
Results of Laboratory Tests on BIM-rocks

T. Pilgerstorfer
13
Triaxial Tests on Cataclasites

A. Goricki & E. Pimentel
22
Rckrechnung von Gebirgsparametern in Strzonen

machbare Aussagen, Unsicherheit und Grenzen
P. Schubert
28
Tunnelling through a Major Fault: Considerable Overburden, Squeezing and Flowing Ground;
At-Face Selection of Support
U. Glawe
33
A Fault Zone Management for Deep Seated Tunnels

U. Burger, T. Marcher, E. Saurer & L. Soldo
51
Numerical Analysis of Fault Zones Coming Closer to a Solution

M. Brandtner
60
Design of Ductile Supports

N. Radoni & W. Schubert
64

Workshops 2013
th
Characterization and Classification of Fault Zones

a
F. Fasching & R. Vanek

3G Gruppe Geotechnik Graz ZT-GmbH, Graz, Austria
1. INTRODUCTION
The characterisation of brittle fault zones is one of the essential interdisciplinary tasks in the fields of
engineering geology and geotechnics. The prediction of the rock mass model and detailed knowledge about
the mechanical properties of various fault rocks are of particular significance. The complexity of fault
zones and their relevance for constructions make it necessary to implement a project-specific characterisation of the rock mass.
All relevant geological, hydrogeological and geotechnical conditions should be described. In addition, the
construction and economic background of each project has to be taken into consideration in order to be able
to deliver practical information for the designer team. Considering this, a clear path from investigation,
through the development of the geological model to the geotechnical design and the construction is
essential.
The talk summarizes the significant characteristics of fault zones and presents their practical
implementation in a project-specific characterisation. Additional, short descriptions of different types of
fault rocks and fault zones are given. The intention is to illustrate the engineering geological
characterisation of fault rocks and fault zones in order to provide the data needed for geotechnical and
construction design.
2. SUMMARY OF THE TALK

The considerations presented, can be summarised by the following statements:
Fault zones represent complex geological structures that are composed of various rocks with
different properties.
Figure 1: Longitudinal section of a tunnel drive in a fault zone; two cross sections.
R. Vanek & F. Fasching
Each fault zone is unique, due to a wide range of different influencing factors, such as:
modes of tectonic displacement

stress field
temperature-level
duration of displacement (age of fault)
amount and rate of displacement (tectonic activity)
parent rock mass (lithology, mineral content, structure ...)
groundwater conditions
The interaction of those factors causes the special properties of each fault zone, especially:
size and scale of the fault zone (thickness, persistence)

internal structure of the fault zone
rock / rock mass strength
permeability properties
rock mass behaviour
Taking into account all the specific properties of fault rocks, an unambiguous treatment as either
soil material or solid rock is not possible. This gives fault rocks a certain special status.
Figure 2: Zones of all kinds of fault rocks with soil and hardrock properties.
Characterization of Fault Zones | Characterization and Classification of Fault Zones
For an appropriate characterisation of fault zones, detailed structural geological knowledge about
their origin and development and the resulting internal structures are essential.
Figure 3: Evolution of a fault zone causing changes in properties.

The complexity and variety of fault rocks demands an unmistakeable and clear use of terms and
descriptions.
cemented
Hard-rock properties
Mylonite
Tectonic Breccia
Pseudotachylite
Soft-rock and / or soil properties
cohesive
noncohesive
controlling* particle sizes:
Fault rock
without
Blocks
fine-clastic
(clay, silt )
not cemented - loose
*decisive for
mechanical
behavior
grain size
distribution:
wide spread,
size of blocks > gravel
Fault rock
with
Blocks
(Block diameter:
> 63mm)
cohesive
fault rock
(Cataclasite)
coarse-clastic
(sand, gravel)
noncohesive
fault rock
spread of particle sizes:

clay to gravel
spread of praticle sizes:

sand to gravel
cohesive
noncohesive
cohesive fault rock

with blocks
noncohesivefault rock
with blocks
Type
Example:
cohesive fault
cohesive
rock, blocky,
fc75
noncohesive
decisive*
grain size
(Cl to Gr)
fine-clastic (fc)
(clay, silt)
< 25%
25 - 75%
> 75%
fc25
fc50
fc75
coarse-clastic (cc)
(sand, gravel)
cc25
cc50
cc75
Block proportion (> Gr)
cohesive BiM-Rock
grain size distribution:
intermittend
staged,
size of blocks >>
matrix (BiM-Rock)
Type
Example:
cohesive
BiM-Rock,
fc50
cohesive
noncohesive
decisive*
grain size
of matrix
fine-clastic (fc)
(clay, silt)
coarse-clastic (cc)
(sand, gravel)
Example:
noncohesive
fault rock
blocky, cc50
noncohesive BiM-Rock
Block proportion (Block >> Matrix)
< 25%
25 - 75%
fc25
fc50
> 75%
fc75
cc25
cc50
cc75
Example:
noncohesive
BiM-Rock, cc75
Figure 4: Specific classification system of fault rocks (F. Fasching & R. Vanek, 2010).
The engineering geological and geomechanical characterisation of fault zones is essential for an
adequate design and successful construction of underground structures in a faulted rock mass.
ENGINEERING GEOLOGICAL CHARACTERISATION OF FAULT ZONES 1

description
geological environment
fault
components
shear zones /
fault rocks
shear bodies /
preserved rock
portions 2
internal architecture
rock mass
parameters
rock mass
additional. engin.
geol. information 3
parameters
parent rock and geological setting
fault rocks, particle size distribution,

compaction, mineralogy, permeability,
friction angle, cohesion, Young's

modulus,
rock properties, mineralogy, properties of

discontinuity network, wear properties,
strength parameters of rock, shear

parameters of discontinuities, wear
indices,
thickness of shear zones, shape and size of

size of shear bodies, block-matrixshear bodies, volumetric proportion of shear
ration, ...
bodies, anisotropy,
description of strength und deformability

properties, comment on anisotropy of rock
mass properties
rock mass strength, friction angle,

cohesion, Young's modulus,
Poisson's ratio,
permeability, grouting properties, swelling

potential, wear properties,
(as required)
according to "ground types" following the "Guideline for the Geotechnical Design for Underground Structures" (GG, 2008)
the properties of shear bodies usually can be characterised analogously to ground types of hard rock mass
according to project requirements
Figure 5: Main aspects of a project-specific characterisation of a fault zone.
The transfer of the engineering geological model of a fault zone into a realistic geomechanical
model is a great challenge in geotechnical design. An adequate implementation requires intensive
interdisciplinary collaboration between the engineering geologists and geotechnical engineers.
3. DISCUSSION
In the course of the engineering geological characterisation of fault zones, certain simplifications have to be
made concerning internal architecture and rock mass properties. The requirements for a simplified model
are based on one hand on the incomplete knowledge of the condition of the fault zone, as can be derived
from the geological site investigation. On the other hand, the geomechanical design will need adaptations
due to limited performance and system capacity of numerical models. That requires simplifications to the
spatial-geometrical conditions, the rock mass parameters and the constitutive laws.
In order to ensure that no essential geotechnical criteria are lost through such simplifications, detailed
engineering geological knowledge about the ground is desirable. This requires careful and systematic
investigations and sampling of - if possible - all components of the fault zone (different fault rocks and
shear bodies). The results of field and laboratory tests should therefore be subjected to thorough analysis,
where every single result is being assigned to a specific fault material. This detailed knowledge coupled
with a model of all geomechanical behaviour/failure mechanisms is essential as a precondition for the
adequate determination of rock mass parameters.
The quality of both, the engineering geological model and the geomechanical model finally determine the
reliability of prediction for the construction project.
References
[1]
Brodie, K., Fettes, D., Harte, B. & Schmid, R. 2007. Structural terms including fault rock terms.
Recommendations by IUGS Subcommission on the Systematics of Metamorphic Rocks (Web
version of 01.02.2007).
[2]
Brosch, F.J., Kurz, W. & Klima, K. 2006. Definition and Characterization of Faults and Fault Rocks.
Felsbau 24 (5), pp.13-20, Essen: VGE.
[3]
Caine, J.S., Evans, J.P., & Forster, C.B. 1996. Fault zone architecture and permeability structure.
Geology Vol. 24, pp. 1025-1028.
[4]
Eisbacher, G.H. 1996. Einfhrung in die Tektonik. Stuttgart: Enke.
[5]
Fasching, F. & Vanek, R. 2011. Engineering geological characterisation of fault rocks and fault
zones, Geomechanics and Tunnelling Vol. 4 (3), pp. 181194, Berlin: Ernst & Sohn.
[6]
Heitzman, P. 1985. Kakirite, Kataklasite, Mylonite Zur Nomenklatur der Metamorphite mit
Verformungsgefge. Eclogae geol. Helv. Vol. 78 (2), pp. 273-286.
[7]
Hobbs, B.E., Means, W.D. & Williams, P.F. An Outline of Structural Geology. New York: John
Wiley & Sons.
[8]
Mandl, G. 2000. Faulting in Brittle Rocks. Berlin / Heidelberg: Springer.
[9]
Meschede, M. 1996. Methoden der Strukturgeologie. Stuttgart: Enke.
[10]
Naylor, M.A., Mandl, G. & Sijpesteijn, C.H.K. 1986. Fault geometries in basement-induced wrench
faulting under different initial stress states. Journal of Structural Geology Vol. 8 (7).
[11]
GG Fachsektion "Felsmechanik und Felsbau". 2010. Arbeitskreis Geomechanik tiefliegender

Tunnel. sterreichische Gesellschaft fr Geomechanik. Unpublizierte Arbeitsunterlagen.
[12]
Riedmller, G., Brosch, F.J., Klima, K. & Medley, E.W. 2001. Engineering Geological
Characterization of Brittle Faults and Classification of Fault Rocks. Felsbau 19 (4), pp. 13-18,
Essen: VGE.
[13]
Schmid, S.M. & Handy, M.R. 1991. Towards a Genetic Classification of Fault rocks: Geological
Usage and Tectonophysical Implications. Controversies in Modern Geology, Chapter 16, London:
Academic Press.
[14]
Tchalenko, J.S. 1970. Similarities between shear zones of different magnitudes. Geological
Association of America Bulletin Vol. 81, pp. 16251640.
[15]
Twiss, R.J. & Moores, E.M. 1992. Structural Geology. New York: Freeman & Co.
[16]
Wibberley, C.A.J., et.al. 2008. The Internal Structure of Fault Zones. Geological Society Special
Publication, NO. 299, London: The Geological Society.
Authors
Dipl.-Ing. Florian Fasching
Mag. Robert Vanek
3G Geotechnik Gruppe Graz ZT-GmbH

Triester Strae 478 a
8055 Graz-Seiersberg
Austria
3G Geotechnik Gruppe Graz ZT-GmbH

Triester Strae 478 a
Austria
[email protected]
[email protected]

Workshops 2013
th
Ingenieurgeologische Charakterisierung von Strungszonen

im Innsbrucker Quarzphyllit
a
A. Tchterle , H. Schierl & C. Reinhold

Brenner Basistunnel BBT SE, Innsbruck, Austria
Der Brenner Basistunnel (BBT) wird zwischen Tulfes bei Innsbruck (A) und Franzensfeste (I) errichtet
(64 km). Teil des Gesamtprojekts ist ein durchgehender Erkundungsstollen unterhalb und mittig zwischen
den beiden Haupttunnelrhren. Wesentliche Erkundungsziele sind die Lokalisierung, Bestimmung der
Raumlage und geomechanische Charakterisierung von Strungszonen in den zu durchrternden Einheiten.
1. GEOLOGISCHE BERSICHT
Der BBT durchrtert von N nach S den ostalpinen Innsbrucker Quarzphyllit, die emporgewlbten
penninischen Einheiten des Tauernfensters nahe an dessen westlichen Rand, ein schmales ostalpines
Segment (OA) im S des Tauernfensters, die Periadriatische Strungszone und anschlieend den sdalpinen
Brixner Granit (Abbildung 1). Die penninischen Einheiten des Tauernfensters lassen sich grob untergliedern in Obere und Untere Schieferhlle und die Zentralgneise mit dem Alten Dach. Die Trasse des
gesamten BBT zwischen Innsbruck und Franzensfeste verluft zu ca. 75% in Gesteinen mit einer meist
stark ausgeprgten Schieferung, bestehend aus Phylliten und Schiefern der ostalpinen Einheiten sowie der
Oberen und Unteren Schieferhlle. Die restlichen 25% verteilen sich auf kaum bis deutlich geschieferte
Orthogneise und auf ungeschieferten Granit. Bisher wurden im N berwiegend Tunnelbauwerke im
Innsbrucker Quarzphyllit aufgefahren (IQP), wobei hier in erster Linie der Erkundungsstollen mit einer
derzeitigen Lnge von ca. 4,8 km zu erwhnen ist.
Abbildung 1: Geologischer bersichtslngsschnitt des BBT zwischen Innsbruck und Franzensfeste.
2. SPRDE STRUNGSGESTEINE UND GEBIRGE UNTER STRUNGSEINFLUSS NOMENKLATUR FR DIE CHARAKTERISIERUNG VON
TYPISCHEN STRUNGSGESTEINEN IM BBT-PROJEKTRAUM
Eine Strung wird als sprd bezeichnet, wenn es im Zuge der Deformation zu bruchhaften Verformungsprozessen im Gestein gekommen ist. Diese Prozesse knnen sich entweder auf einzelne isolierte Bruchflchen konzentrieren oder ganze Gesteinsabschnitte penetrativ erfassen und zur Bildung sogenannter
Strungsgesteine fhren. Zu beachten ist hier, dass diese Beschreibung mastabsabhngig ist und immer
im Zusammenhang mit der Gre des betrachteten Gebirgsausschnittes zu sehen ist. Unabhngig davon ist
mit der Bruchbildung generell eine Herabsetzung der Festigkeit des Materials verbunden, wobei es durch
nachfolgende Zementationsprozesse zu einer Ausheilung und zumindest teilweisen Wiederverfestigung
der Bruchzone kommen kann. Aufgrund der Verminderung der Materialfestigkeit sind sprde Strungen
jedenfalls von besonderer bautechnischer Relevanz im Tunnelbau.
Zur Beschreibung der Strungsgesteine knnen verschiedene Klassifikationssysteme aus der Fachliteratur
herangezogen werden [1, 2, 3, 4]. Zudem wurde im Zuge der intensivierten geologischen Vorerkundungen
fr den BBT zwischen 1999 und 2006 bereits besonderes Augenmerk auf die Charakterisierung der im
Projektgebiet auftretenden Strungen gelegt. Die verschiedenen Strungsgesteine wurden neben weiteren
geotechnisch relevanten Strukturen beschrieben, fotographisch dokumentiert und in einem Klassifikationsschema untergliedert [5], und die im Projektgebiet auftretenden Strungen wurden anhand von ObertageAufschlssen im Detail beschrieben [6]. Es zeigte sich im Zuge der fortlaufenden Dokumentation im
Erkundungsstollen allerdings zunehmend, dass mithilfe der bestehenden Modelle die Kategorisierung der
aus geschieferten Ausgangsgesteinen gebildeten Strungsgesteine nicht in befriedigendem Mae mglich war.
Deshalb wurde aufbauend auf den Grundlagen aus der Fachliteratur (insbesondere [3]) und den Ergebnissen
aus der Vorerkundung ein projektspezifisches Klassifikationsschema fr Strungsgesteine zusammengestellt (Tabelle 1 & 2). Es hat sich zudem als ntzlich erwiesen, auch das unter Strungseinfluss stehende
Nebengebirge hinsichtlich des Beeinflussungsgrades durch die Strung in Form eines einfachen Schemas
zu untergliedern (Tabelle 3). Alle Definitionen haben dabei folgende Bedingungen zu erfllen:
gute Anwendbarkeit in den fr den gesamten Projektraum zu erwartenden Gesteinen

gute Anwendbarkeit im Zuge der Vortriebsdokumentation
grtmgliche Eingliederung des Klassifikationsschemas aus der Vorerkundung [5, 6]
1
kein Widerspruch zu allgemein anerkannten Definitionen der Struktur- und Ingenieurgeologie
Tabelle 1: Lockergesteinsartige Strungsgesteine.
Insbesondere dieser Punkt beherbergt das meiste Diskussionspotential. Grnde dafr sind bestehende Unterschiede in den etablierten
Definitionen, die einerseits traditionell begrndet sein knnen [2] und andererseits auch auf die unterschiedliche Herangehensweise an
die Thematik verschiedener Fachbereiche zurckgefhrt werden knnen.
Characterization of Fault Zones | Ingenieurgeologische Charakterisierung von Strungszonen
Tabelle 2: Festgesteinsartige Strungsgesteine.
Tabelle 3: Nebengebirge unter Strungseinfluss.
3. GEOLOGISCH-GEOTECHNISCHE VERHLTNISSE IN DEN

VORTRIEBEN IM INNSBRUCKER QUARZPHYLLIT
Im Februar 2010 starteten die Vortriebsarbeiten im Erkundungsstollenabschnitt Innsbruck-Ahrental in der
Sillschlucht bei Innsbruck. Noch im selben Jahr begannen die Arbeiten beim Zufahrtstunnel Ahrental, und
im Mai 2012 erfolgte der Anschlag des Fensterstollen Ampass als Zugangsstollen fr den geplanten
Rettungsstollen Tulfes. Im Sommer 2013 konnten alle Vortriebsarbeiten erfolgreich fertiggestellt werden.
Insgesamt wurden dabei ber 9 km Stollen und Tunnel im Innsbrucker Quarzphyllit hergestellt.
Der Erkundungsstollen Innsbruck-Ahrental (4,8 km) wurde konventionell vorgetrieben und weist eine
Querschnittsflche von ca. 30 m2 auf. Die maximale berlagerung betrgt ca. 900 m. Der Ausbruch erfolgte
praktisch vollflchig, in den greren, alle 250 m angeordneten Ausweichnischen wurde der Querschnitt in
Kalotte und Strosse unterteilt. Aufgrund der Distanz zum Wipptal ist der Einfluss der Wipptalstrung bzw.
die Tektonisierung des Gebirges hier deutlich weniger stark ausgeprgt als das Gebirge im Zufahrtstunnel
Ahrental. Die Hauptkluftscharen fallen hier ebenfalls steil ost- und westwrts ein, aufgrund der annhernd
Nord-Sd verlaufenden Tunnelrichtung aber nicht direkt aus der Ortsbrust, sondern seitlich zu den Ulmen
hin. ber weite Strecken war das Gebirgsverhalten nur durch kleinvolumige, gefgebedingte Nachflle
geprgt. In drei Abschnitten wurden durch schleifend bis achsparallel bzw. spitzwinkelig streichende
Strungen schwach druckhafte Verhltnisse generiert. Nachstehend werden exemplarisch zwei
geotechnisch relevante Strungszonen im Erkundungsstollen Innsbruck-Ahrental beschrieben:
Strungszone ESI-f2088:
Von Stollenmeter ca. 1850 m bis 2100 m begleitete ein Lngsstrungssystem mit einer Mchtigkeit der
Hauptstrung von ca. 3 m den Tunnelquerschnitt am linken bzw. stlichen Ausbruchsrand.
Diese mit ca. 60 nach ENE einfallende Strung lste in Folge lang anhaltende, asymmetrische Kriechbewegungen mit gegenber der Setzungskomponente erhhten horizontalen Querverschiebungen bis max.
160 mm Absolutverschiebung aus. Als Folge traten Schden am Ausbau (Risse und Abplatzungen der
Spritzbetonschale) auf, welche Sanierungsmanahmen und den nachtrglichen Einbau eines Sohlgewlbes
erforderlich machten.
Strungszone ESI-f4158:
Eine weitere Strungszone mit schwach druckhaftem Gebirgsverhalten trat bei Stollenmeter 4140-4180 auf.
Diesmal war ein westfallendes, annhernd spitzwinkelig streichendes Strungssystem Auslser dafr. Die
Hauptstrung ist hier ca. 5 m mchtig bestehend aus dm-mchtigen Scherbahnen mit Fault gouge neben
veruschelten und mrbfesten Quarzphyllit (geschieferter Protokataklasit bis Kataklasit). Je nach Position
des Strungssystems wurden die grten Verschiebungsbetrge beim Hereinstreichen und beim Herausstreichen der Strung in bzw. aus dem Querschnitt beobachtet. Auch hier traten Schden am Ausbau mit
Rissbildungen, Ausknicken der Gittertrger und Aufbiegen der Ankerplatten auf, welche im Nachgang
saniert werden mussten.
Bei beiden Abschnitten war an der Ortsbrust selbst kein unmittelbares Standfestigkeitsproblem mit
verstrkten Nachfllen gegeben. Die augenscheinliche Gebirgsverschlechterung war nicht gravierend. Auch
aus den gemessenen Erstverschiebungen lie sich nicht sofort ein druckhaftes Gebirgsverhalten ableiten.
Aus den gemachten Erfahrungen wurde aber fr den weiteren Vortrieb eine Verstrkung des Ausbaus
beschlossen, sobald an der Ortsbrust erste Indizien fr ungnstig streichende, auch nur geringmchtige
Strungssysteme auftraten. Neben einem dichteren Radialankerungsschema und je nach Position der
Strung unterschiedlichen, zumeist asymmetrischen Ankerlngen waren auch. Verformungsschlitze in der
Spritzbetonschale und eine baubetriebliche Abstimmung fr einen raschen Sohlschluss vorgesehen.
4. ERMITTLUNG DER GEOMECHANISCHEN KENNWERTE DER

STRUNGSZONEN
Fr die geologisch-geotechnische Prognose ist neben Lage, Orientierung, Abmessungen und Gebirgsverhalten der Strungszonen auch die Angabe zutreffender geomechanischer Kennwerte, mit denen die
Strungszonen beschrieben werden knnen, entscheidend. blicherweise wird dabei so vorgegangen, dass
Gesteins- bzw. Bodenproben entnommen werden, diese im Labor bezglich ihrer mechanischen
Eigenschaften untersucht werden und aus diesen Ergebnissen schlussendlich auf die Kennwerte der
Strungszone geschlossen wird.
Bei den im Innsbrucker Quarzphyllit im Erkundungsstollen angetroffenen bestimmenden Strungsgesteinen
handelte es sich im Wesentlichen um geschieferte Protokataklasite (fprc). Solche Gesteine knnen als
Festgestein mit geringen bis sehr geringen Festigkeiten beschrieben werden (UCS < 15 MPa). Gesteinsproben in Form von Handstcken aus dem Vortrieb lassen sich mit einem leichten Hammerschlag oder
sogar schon mit der Hand brechen. Aufgrund dieser geringen Festigkeiten war es derzeit noch nicht
mglich berhaupt Kerne fr einaxiale oder triaxiale Druckversuche zu gewinnen.
Aufgrund der im Labor derzeit noch nicht ermittelbaren mechanischen Eigenschaften der Protokataklasite,
musste eine andere Mglichkeit zur Abschtzung der geomechanischen Eigenschaften der Strungszonen
im Innsbrucker Quarzphyllit gefunden werden. Durch den vorlaufenden Erkundungsstollen besteht die
Mglichkeit der Rckrechnung der im Erkundungsstollen Innsbruck-Ahrental gemessenen Verformungen
und somit der Abschtzung der Gebirgskennwerte der Strungszonen. Auf Basis des geologischen Modells
der Strungszonen, welches auf den Erfahrungen beim Auffahren des Erkundungsstollens basiert, werden
die Verhltnisse im Bereich der Strungszonen mit verschiedenen Berechnungsverfahren nachvollzogen.
Dafr kommen sowohl einfache Berechnungen mit der Kennlinie, als auch komplexe 2- und 3dimensionale numerische Berechnungen zur Anwendung.
Das prinzipielle Vorgehen sieht in drei Schritten die Erstellung des geologischen Modells, die darauf
folgende Erstellung des geotechnischen Modells und schlussendlich die Erstellung des
Berechnungsmodells der Strungszone vor (Abbildung 2). Aus den geologischen Daten welche aus dem
Vortrieb des Erkundungs-stollens gewonnen wurden (Ortsbrustdokumentation, Tunnelbnder) wird im
ersten Schritt ein geologisches Modell der Strungszone erarbeitet. Im folgenden Schritt wird daraus ein
vereinfachtes geotechnisches Modell erarbeitet, welches unter anderem die Aufteilung in Core Zone (CZ)
und Damage Zone (DZ) umfasst. Dieses geotechnische Modell der Strungszone wird dann lagemig in
das Berechnungsmodell integriert. Weiterhin werden die Vortriebshistorie sowie die verwendeten
Sttzmittel detailliert ausgewertet und in das Berechnungsmodell integriert.
Tabelle 4 zeigt beispielhaft die Kennwerte fr zwei Strungszonen im Innsbrucker Quarzphyllit die bereits
im Erkundungsstollen aufgefahren und deren Kennwerte ber Rckrechnungen abgeschtzt wurden. Das
Verhalten beim Vortrieb dieser beiden Strungszonen ist unter Punkt 3 beschrieben. Aus den angegebenen
Core Zone Kennwerten wird unter anderem die Tiefenabhngigkeit der Parameter Elastizittsmodul und
Kohsion ersichtlich. Beide Strungszonen sind dem gleichen regionalgeologischen Strungssystem
(Wipptal-System) zuzuordnen und zeigen auch annhernd einen vergleichbaren geologischen Ausbau.
Lediglich die Mchtigkeiten und berlagerungshhen sind unterschiedlich. Die Rckrechnungen haben
10
eine Zunahme von E-Modul und Kohsion mit steigender berlagerung ergeben. Dieser Sachverhalt ist
insbesondere fr Lockergestein hinreichend bekannt.
Abbildung 2: Beispiel fr die berfhrung des geologischen Modells in ein Berechnungsmodell.

Tabelle 4: Geomechanische Kennwerte ausgewhlter Strungszonen die im Erkundungsstollen
Innsbruck-Ahrental aufgefahren wurden.
Bezeichnung
Kurzbeschreibung
Abmessungen
(DZ: Hangend /
Liegend der CZ)
berlagerung
Parameter
Core Zone
Parameter
Damage Zone
ESI-f2088
CZ: fprc 90%, fg 10%

DZ: jrm 70%, fprc 20%,
fg 10% (GSI 25-30)
CZ: 3 m
DZ: 5,5 m / 5,5 m
300 m
E = 400 MPa
c = 0,15 MPa
E = 1000 MPa
c = 0,45 MPa
= 25
= 26
CZ: fprc 90%, fg 10%

DZ: jrm 80%, fprc 10%,
fg 10% (GSI 30-35)
CZ: 5,5 m
DZ: 10,5 m / 1 m
E = 600 MPa
c = 0,30 MPa
E = 1000 MPa
c = 0,50 MPa
= 26
= 28
ESI-f4158
690 m
5. RESMEE
Fr die Planung und den Bau der Hauptrhren des BBT ist die mglichst genaue Charakterisierung von
potentiell den Vortrieb und insbesondere auch die Wahl des Vortriebsverfahrens (konventionellmaschinell) beeinflussenden Strungszonen von mageblicher Bedeutung. Bei der geomechanischen
Prognose fr tiefliegende Tunnelbauwerke ohne Erkundungsstollen werden die Eigenschaften von
Strungszonen im Normalfall aus Oberflchenkartierungen und Tiefbohrungen abgeleitet. Dabei knnen
sich bedeutende Schwankungsbreiten in den prognostizierten geologisch-geotechnischen Daten ergeben. Im
geologisch-geomechanischen Modell des BBT aus der Vorerkundung (basierend auf Bohrungen und Oberflchenkartierungen) bezog sich beispielsweise die Angabe der Strungszonen auf eine Anzahl je
geotechnischem Homogenbereich mit globalen Gebirgsparametern fr die Strungszonen. Fr das
geologisch-geomechanische Modell der Hauptrhren kann nun basierend auf den Erkenntnissen aus dem
11
Erkundungsstollen eine diskrete Angabe der Einzelstrung mit jeweils bezogenen Gebirgsparametern
erfolgen. Ziel des vorauseilenden Erkundungsstollens beim BBT ist unter anderem, die notwendigen Daten
der Strungszonen prziser und mit geringeren Schwankungsbreiten fr die Haupttunnel prognostizieren zu
knnen. Damit soll das geologisch-geomechanische Risiko beim Bau der Hautrhren erheblich reduziert
werden.
References
[1]
Brodie, K., Fettes, D., Harte, B. & Schmid, R. 2007. Structural terms including fault rock terms.
Recommendations by IUGS Subcommission on the Systematics of Metamorphic Rocks (Web
version of 01.02.2007).
[2]
Fasching, F. & Vanek, R. 2011. Engineering geological characterisation of fault rocks and fault
zones / Ingenieurgeologische Charakterisierung von Strungsgesteinen und Strungszonen.
Geomechanics and Tunnelling 4 (3), pp. 181-194, Berlin: Ernst & Sohn.
[3]
Heitzmann, P. 1985. Kakirite, Kataklasite, Mylonite - Zur Nomenklatur der Metamorphite mit
Verformungsgefgen. Eclogae geol. Helv. 78 (2), pp. 273-286.
[4]
Riedmller, G., Brosch, F.J., Klima, K. & Medley, E.W. 2001. Engineering Geological
Characterization of Brittle Faults and Classification of Fault Rocks. Felsbau 19 (4), pp. 13-18,
Essen: VGE.
[5]
BBT EWIV. 2001. Geologie, Erkundung, Strukturtektonik, Schlubericht. Unverffentl. Bericht von
Brandner, R., Decker, K., Ortner, H., Reiter, F., Bistacchi, A. & Massironi, M.
[6]
BBT SE. 2005. Charakterisierung von Strungszonen. Unverffentl. Bericht von Brandner, R., Dal
Piaz G.V., Bistacchi, A., Decker, K. & Reiter, F.
Authors
Dr. Andreas Tchterle
Mag. Heimo Schierl
Brenner Basistunnel BBT SE

Amraser Str. 8
6020 Innsbruck
Austria

Amraser Str. 8
6020 Innsbruck
Austria
[email protected]
[email protected]
Dr.-Ing. Chris Reinhold

Amraser Str. 8
6020 Innsbruck
Austria
[email protected]
12
steerreichische Gesellschaft
G
f
r Geomechannik
Austtrian Society for
f Geomecha
anics
Worksh
hops 2013
October 9 22013, Salzburrg, Austria
th
Resu
ults of Lab
boratory Tests
T
on BIIM-rocks
a
T. Pilgerstorrfer
Graz Unniversity of Tecchnology, Insttitute for Rockk Mechanics and
a Tunnellingg, Graz, Austrria
Wheen dealing witth tunnels in difficult

d
grounnd conditions,, e.g. fault zon
nes, the know
wledge of the rock
r
mass
paraameters is of utmost
u
imporrtance to ensuure a safe and
d economical tunnel designn. Sections wiith a high
conttent of cataclaasites form the most challeenging stretch
hes during tun
nneling, and a proper geom
mechanical
charracterization iss imperative. However,
H
invvestigating thee overall-propeerties is currenntly a challen
nging task,
origiinating from difficulties
d
in sample acquissition, samplee preparation and
a laboratoryy testing.
In orrder to gain innsight into th
he overall mecchanical propeerties of catacclasite-like maaterial and to study the
basicc mechanical behavior of BIM-rocks
B
an extensive lab
boratory program was carrieed out. Artificcial blockin-m
matrix rocks were
w
fabricated
d for both dirrect shear testts and large oedometer
o
testts. For the direct shear
testss models withh six differentt block orientaations, related
d to the shearr direction, annd three differrent block
propportions weree created. The large oeddometer tests feature samples with three differeent block
arranngements, witth regard to th
he loading dirrection. A straaight forward evaluation m
method is pressented and
the rresults are disscussed in dettail, highlightiing the effect of block orieentation and bblock proportiion on the
sheaar behavior, shhear strength and
a deformatioon behavior of
o BIM-rocks.
1. INTRODU
UCTION
Tecttonic faults are usually com
mposed of lenss-shaped, relattively competeent rock blockks surrounded
d by finely
grainned cataclastic material [1, 2]. Hence, thheir propertiess are highly an
nisotropic andd depend on the
t degree
of thhe regularity of
o the block orrientation, thee total volumeetric amount of the competeent lenses as well
w as the
propperties of the matrix.
m
To study the priinciple mechaanical properrties of fault material an extensive labboratory prog
gram was
n-matrix rockks. Tunneling in weak rock mass is oftenn accompanied
d by large
condducted on artificial block-in
displacements, heence an inten
nded issue waas to investigaate the behav
vior of BIM-rrocks exposed
d to large
straiins. Lindquistt [3] provided
d a comprehennsive study of triaxial testss on artificiall BIM-rocks. However,
beinng a characteristic of triaxiaal tests large strains are diifficult to achieve and the investigation of failure
surfaaces is rarely possible. Heence it was deecided to exaamine the mecchanical behaavior of artificcial BIMrockks in direct shhear tests. Another importaant issue is th
he knowledge about the strress dependen
ncy of the
defoormation properties, especiaally for TBM -advances in weak rock masses. The am
mount of displacements
whicch are expecteed and the risk
k for a shield-T
TBM getting stuck should be
b known a prriori.
2. ARTIFICIIAL BLOC
CK-IN-MAT
TRIX ROC
CKS
2.1. Matrixx material
A sppecial issue waas finding a prroper matrix m
material, meetting following
g goals:
A relativvely short curin

ng time, to ennsure prefabriccation of samp
ples within a ttight timetablee, and
Limited, controllable strength
s
and ddeformation properties
p
in order to ensuree high contrasst between
matrix annd block properties.
Afteer performing several tests, including uniiaxial and triaxial compresssion tests, visuual inspection
n of curing
behaavior (onset of
o cracking du
ue to shrinkagge) it turned out
o that a mix
xture of finelyy crushed rock
k material
(limestone), cemeent and water meets the inttended properrties best. A crushed rock/ccement ratio of
o 5:1 was
od workability
y. This obviouus high ratio is required
usedd, the water/ceement ratio waas set to 1.8, aallowing a goo
due to the fact thaat the mixture contains alm
most no coarse aggregates. The
T propertiess of the matrix
x material,
gainned by triaxiaal tests, are a uniaxial com
mpressive strrength of app
proximately 3 MPa and a Youngs
moddulus of 4000 MPa,
M respectiively.
C
Characterization of Fault Zo
ones
13
T. Pillgerstorfer
2.2. Block material

The blocks were made of cem
ment-stone annd feature an
n ellipsoidal shape, being slightly tapeered in ydirecction (see Tabble 1, block sketch
s
at botttom left). Thee length is 95/100 mm (z-ddirection), thee width is
40/445 mm (in x-ddirection) and
d the depth is 15 mm (in y-direction),
y
reespectively. A Portland cem
ment type
CEM
M II 42.5 wass used, the water/cement rratio was set to
t 0.4. The blocks were prrefabricated in silicone
moldds. Triaxial teests were perfo
ormed on the block materiaal to gain the strength and ddeformation properties,
p
yieldding a uniaxiaal compressivee strength of 337 MPa after seven
s
days curing time, a frriction angle of
o 30 and
a Yooungs Modullus of about 16,000 MPa. T
The strength raatio between matrix
m
materiaal and block material
m
is
abouut 12, and the deformation ratio
r
amounts to the factor 4.
4
3. DIRECT SHEAR
S
TE
ESTS ON B
BIM-ROCK
KS
3.1. Samplle preparattion
The samples weree prefabricated
d in two quaddratic steel sheear boxes. Thee boxes have a height of 18 cm and a
lenggth/width of 200 cm. The gap
p between thee boxes is 2 cm
m, which was sealed with ppolystyrene baars during
sampple preparatioon. First the lo
ower box wass filled with matrix
m
materiaal. After a cuuring time of about one
hourr the blocks were
w placed in the matrix in the intended orientation an
nd block propoortion. With th
he help of
stenccils it was ennsured that thee blocks remaained in the desired
d
positio
on. After the matrix had stiffened
s
a
littlee, and thus thhe blocks would stay in pplace, the upp
per box was filled
f
up withh matrix mateerial. The
sampples were left to cure for seeven days.
3.2. Orienttation of bllocks and vvolumetric block prop

portion
It w
was tried to cover
c
a broad
d range of poossible block orientations with regard tto the shear direction.
Speccimen with foour different block orientaations, each with
w three diffferent volumeetric block prroportions
weree fabricated, along with pu
ure matrix annd pure block
k samples (Taable 1). In orrder to accoun
nt for the
natuural scatter off strength prop
perties and thhe variability in sample preeparation quaality three testts of each
blocck orientation//block proporttion combinatiion were perfo
formed.
Tab
ble 1: Compilaation of perfo
ormed shear tests.
B
Block
prooportion
0%
25%
25%
25%
25%
25%
25%
25%
50%
50%
50%
50%
50%
50%
50%
50%
Orrientation
Symbol
pure matrix
90
90
60
30
30
0
0
90
90
60
60
30
30
0
0
n/a
|||
| | | par.
///
///
\\\
---- long.
---- trans.
|||
| | | par.
///
\\\
///
\\\
---- long.
---- trans.
No. of
tests
3
3
1
3
2
1
2
1
3
1
2
1
2
1
2
1
Block
B
No. of
Orrientation
Symbol
pro
oportion
tests
3
100%
pu
ure block
n/a
75%
90
|||
3
75%
90
| | | par.
1
75%
60
///
2
75%
60
\\\
1
75%
30
///
2
75%
30
\\\
1
75%
0
---- long.
2
75%
0
---- trans.
1
legeend:
shear direction
|||
block
ks vertically alligned, shear in
i
y-direection
| | | par.
p
block
ks vertically alligned, shear in
i
x-direection
///
block
ks inclined in ppos. y-directio
on
block
ks inclined agaainst pos. y-diirection
\\\
---- long. block
ks horizontallyy aligned, sheaar in
z-direection
---- trans. block
ks horizontallyy aligned,
shear in x-directionn
C
ones | Results of Laboratory
y Tests on BIM
M-rocks
14
T. Pillgerstorfer
3.3. Test procedure

p
and
a evaluattion
The tests were peerformed under constant noormal stiffnesss conditions. Due to the ffact that intacct samples
weree tested, speccial attention was paid to the determin
nation of mob
bilization of cohesion, dillation and
friction over the shear
s
displaceement. It is sttated by severral authors thaat there is no permanent co
ohesion in
rockks, at least at relatively low
w confinemennt, where the cohesional sttrength compponent is grad
dually lost
whenn the rock is strained beyond its peak sttrength [4, 5]. The evaluatio
on method in principle is based
b
on a
coheesion-weakeniing friction
nal strengthen
ening (CWFS
S) model, meeaning that th
the mobilized
d strength
com
mponents (coheesive, frictionaal and dilationnal) are strain dependent.
3.4. Resultts
Figuure 1 shows the
t maximum
m shear strengtth, which is normalized
n
by
y the matrix shear strength, versus
blocck proportion. The results depict
d
a concluusive picture. The shear strrength almost linearly increeases with
increeasing block proportion.
p
Th
he samples wiith 30 and 90
0 block orienttation feature nearly the sam
me factors
of sttrength increaase (25% - faactor of 1.6, 50% - factor of 2.2, 75% - factor of 22.8). Up to 50% block
propportion the samples with 60 block orieentation yieldeed the highest contributionn to the shearr strength,
whille a decrease of
o shear resistance for 75% block proporttion was observed.
Figgure 1: Max. shear strengtth (normalizeed by matrix strength) verrsus block prroportion for 30
bloock orientatioon (top left), 60
6 block orieentation (top
p right) and 90 block orien
entation (bottom left).
C
y Tests on BIM
M-rocks
15
T. Pillgerstorfer
A clloser inspectioon of the sheear surfaces aafter the tests backs the ab
bove mentioneed findings (F
Figure 2).
Sam
mples with 90 block orien
ntation experiience during the shearing process a suddden, brittle failure of
bloccks, which ressults in smootth cracks and thus providin
ng low valuess for dilatancyy. The failure plane for
the 330 block oriientation samp
ples features also a smootth surface. However, in thhose cases pu
ure sliding
alonng blocks takees place, whiich yields rellatively high values for diilatancy. The 60 block orientation
o
arranngement yieldds a contrary picture. At thhe onset of th
he shearing prrocess slight sliding along blocks is
present, which leaads to grinding
g and crushingg of the block
ks with ongoin
ng shear displaacement. Thiss results in
an inncreased shearr surface roug
ghness, which further yieldss a higher conttribution to shhear resistancee.
Figu
ure 2: Failuree surfaces aftter the test sh
hown for 30 block proporrtion (top leftt and right), 90 block
prop
portion (bottoom left) and 60 block prooportion (botttom right).
The effect of blocck inclination with regard tto the shear direction
d
(bloccks inclined toowards or agaainst shear
direcction) is nicelly shown in Figure
F
3. Deppicted therein
n is the accum
mulated shear energy over the shear
displacement, whiich is defined by equation 11.
( )
( )
(1)
The shear energy is normalized

d by the accuumulated sheaar energy deveelopment of thhe pure matriix sample.
Exem
mplarily show
wn are the results of the 330 and 60 block
b
orientattion samples ((Figure 3). Itt becomes
eviddent that sampples with bloccks inclined aagainst shear direction
d
featu
ure a pronounnced ductile behaviour,
b
whicch is represennted by a steeady, mostly linear increaase of shear energy
e
develoopment over the shear
displacement. A contrary
c
picturre is given forr samples with
h blocks being
g inclined tow
wards shear dirrection. A
mucch more brittlee behavior is observed, feeaturing a peaak shear stren
ngth value afte
ter relatively low shear
displacement, folllowed by a co
ontinuous decrrease of shear resistance over the remainiing shear disp
placement.
Thiss particular feeature holds true
t
for all teested samples,, hence this behavior
b
is noot influenced by block
propportion.
C
y Tests on BIM
M-rocks
16
T. Pillgerstorfer
Fiigure 3: Norm
malized accum
mulated sheaar Energy verrsus shear dissplacement, shhown for 30 block
prroportion (lefft) and 60 blo
ock proportioon (right).
4. LARGE OEDOMET
O
TER TESTS
S ON BIM-ROCKS
Largge oedometer tests were perrformed on arrtificial block--in-matrix rocks, in order too study the inffluence of
blocck orientation and block pro
oportion on thhe overall defformation prop
perties and too gain knowledge about
the sstress dependeency of modulli.
Currrent laboratoryy techniques for
f the determ
mination of Yo
oungs modulli or constrainned moduli aree afflicted
withh restrictions. Standard oedo
ometer tests, aas characteristtically used fo
or soil testing are not adequ
uate due to
the small sample size and relaatively low sttress levels. Hence,
H
those conditions doo not reflect the stress
condditions found at tunnel leveel and standardd oedometer tests
t
do not alllow obtainingg reliable data about the
stresss dependencyy of moduli att high depths.. In order to circumvent
c
th
hese shortcomi
mings a large oedometer
o
test apparatus waas developed, accounting ffor the issues mentioned above.
a
A detaailed descriptiion of the
largee oedometer teest apparatus is given in [6]].
4.1. Samplle preparattion

The artificial blocck-in-matrix saamples were ffabricated in custom-tailore
c
ed shutterings . All specimen
n featured
a diaameter of 300 mm. The inittial height wass between 128
8 and 156 mm
m. Three differrent block arraangements
in thhree differentt block propo
ortions, as weell as a pure matrix and pure
p
block sam
mple were teested. The
speccified block arrrangements were
w (Figure 44):
Blocks horizontally
h
aliigned (loadingg in y-directio
on of blocks)
Blocks vertically
v
align
ned (loading inn z-direction of
o blocks)
Blocks tw
wisted (loadin
ng in x-directioon of blocks)
C
y Tests on BIM
M-rocks
17
T. Pillgerstorfer
Figgure 4: Diffeerent types off block arran

ngement for large oedometer tests; bllocks horizon
ntally
aliigned (left), blocks
b
vertica
ally aligned (m
middle), block
ks twisted (riight).
4.2. Test procedure

p
and
a data evvaluation
m load of 300
00 kN was utiilized as loadiing device. Th
he vertical
A coompression test machine with a maximum
loadd acting on thee specimen iss measured byy a load cell with
w a nomin
nal load of 20000 kN. A floating ring
conffiguration wass used for the tests, allowinng the determ
mination of fricction forces bbetween oedom
meter ring
surfaace and speciimen. The meeasurement off friction forcees was accom
mplished by thhree load cells. For the
deterrmination of vertical
v
displaacements threee inductive dissplacement traansducers werre used.
The samples were repeatedly loaded
l
and unnloaded. Each
h loading and
d unloading cyycle was kept constant
untill displacemennt rates became almost zero . After each lo
oading-unload
ding loop the lload was increased to a
highher level, whicch was in most cases the douuble amount of
o the previou
us level.
The starting pointts of each load
ding and unlooading loop were
w identified
d and the axiall stress and ap
ppropriate
axial strain were determined. Semilogarithhmic stress-strain diagramss were plotteed and the co
onstrained
moddulus (Es), alsso known as oedometer
o
moodulus, was obtained
o
as seecant moduluss for each loaading and
unlooading loop. Based
B
on Hook
kes law the Y
Youngs Modu
ulus was calcu
ulated by usinng a Poissions ration
of 0..30, based on the results of triaxial tests oon the matrix.
4.3. Resultts
The unloading-mooduli for horizzontally aligneed blocks featture about ten to approximaately thirty tim
mes higher
valuues than the looading-modulii. A slight harrdening behav
vior (with incrreasing stress level) can bee observed
in thhe unloading loops, being more pronounnced for samp
ples with high
her block prooportions. Thee ratios of
unlooading modulii to reloading moduli
m
show values of about 1.3 to 2.0.
The samples withh vertically aligned
a
blockss show a mu
uch higher rattio of loadingg moduli to unloading
u
modduli (~ 20 to 50),
5 compared
d to the speciimen with horizontally alig
gned blocks. O
On the other hand, the
harddening behavior (increase of
o unloading moduli with increasing strress level) is much less prronounced
thann in the horizontally aligneed block arranngement. Thee ratio of unlloading moduuli to reloadin
ng moduli
featuures approxim
mately the sam
me values as thhe samples witth horizontally
y aligned bloccks.
C
y Tests on BIM
M-rocks
18
T. Pillgerstorfer
The ratios of unlooading modulii to loading m

moduli for the twisted
t
block arrangement ffeature factorrs of about
15 tto 30, being in
i the range between the horizontally and vertically
y aligned blocck specimen. Again, a
harddening behaviior is less prronounced, annd is only ob
bserved for higher
h
block proportions. The ratio
betw
ween unloadinng and reloadiing moduli yieelded factors of approximaately two. Thee values for th
he loading
modduli are lower than those of the horizontallly aligned blo
ock samples.
The ratio of loadiing moduli off pure block annd pure matrix samples is between
b
8 (foor low stress states)
s
and
60 ((for high stresss states), hen
nce it can be observed thatt the stress deependency off moduli is much
m
more
pronnounced for thhe pure block
k material. Hoowever, it can
nnot be stated that this is a general behaavior, that
hardder materials feature
f
a mucch higher streess dependenccy of deformaability. It is tthe case that the stress
depeendency and its occurrencce is much m
more a questiion of the cu
urrent degree of porosity, which is
goveerned by the applied stresss level. Furtthermore, imp
posing stress levels exceedding the stren
ngth of a
mateerial yields plastic deformaations, hence eevaluation of deformation moduli
m
consissts of elastic and
a plastic
straiins, which deccrease the appearance of strress dependentt deformation behavior.
A quuite clear pictture is given if the individdual data poin
nts of evaluated moduli aree plotted in a Youngs
moddulus versus axial
a
stress diagram, allowiing the observ
vation of streess dependentt deformation behavior.
The results are deepicted for eaach block arraangement in a separate diag
gram, given iin Figure 5. In
n order to
deterrmine an emppirical relation
nship governinng the develop
pment of the deformation
d
beehavior with increasing
i
stresss level a matthematical relaationship fittin
ing the discrette data points was requiredd. It was foun
nd that the
expoonential functiion shown in equation 2 is aable to yield an
a almost perffect fit.
y e
(2)
Introoducing the Youngs

Y
Modu
ulus E and the axial stress v equation 2 rewrites to (Eqqn. 3):
(3)
E e
A clloser inspectioon of the function parametters and , describing th

he shape of thhe exponential function
curvve, showed thaat they are deependent on thhe volumetricc block proportion. Moreovver, it turned out that
and nearly featuure a linear dep
pendency on tthe block prop
portion. This was
w achieved by slight mod
difications
of thhe best-fit curvves for each block
b
proportioon sample, ho
owever, still yielding regresssion coefficieents above
0.9. This approachh was deemed
d to be valid, since one is dealing
d
with a given mateerial featuring
g a natural
unction param
meters and and their dependence
d
onn the volumetric block
scattter in its propperties. The fu
propportion for thee different block arrangemeents are shown
n in Table 2. The
T great advaantage of this approach
lies in its simpliciity: Merely th
he knowledge of the volumetric block proportion is neeeded as inputt, in order
to caalculate the deevelopment off the overall sttiffness for a desired
d
stress range.
Tab
ble 2: Function parameterss and for different blo
ock arrangem
ments.
Blocck arrangemen
nt
fu
function
paarameter
horizontaally aligned
verttically aligned
d
twisted
for VB
BP < 50
370
for VB
BP 50
4.8 VBP
V
0.0000
032 VBP
0.032 VBP
5 VBP
235
5
11 VBP
58
5.4
0.025 VBP
5.7
5
0.15 VBP
P
8.3
VBP
P volumetriic block proportion in [%]]
C
y Tests on BIM
M-rocks
19
T. Pillgerstorfer
If onne compares the

t diagrams shown in Figgure 5 it can be
b nicely seen
n that the sam
mples with ho
orizontally
alignned blocks feaature values fo
or the Youngs modulus wh
hich are about the double aamount of the vertically
alignned or twistedd block samples. Moreoveer, for the verrtically aligned
d and twistedd block arrang
gements a
lineaar increase off stiffness witth increasing block proporrtion can be observed,
o
whiile for the ho
orizontally
alignned block asseembly this is only the casee for block pro
oportion up to
o 42%. Abovee 42% block proportion
p
stiffn
fness is dispropportionately in
ncreasing.
uated Youngs moduli vs. aaxial stress and fitting fun

nction for horrizontally alig
gned
Figgure 5: Evalu
bloock arrangem
ment (top left)), vertically aaligned blockss (top right) and
a twisted bblock assemblly
(boottom left).
5. CONCLUSION
The performed diirect shear and large oedom
meter tests on
n BIM-rocks provided
p
valuuable informattion about
the mechanical behavior
b
of fault
f
material and yielded a deeper inssight on the factors and properties
p
influuencing the strength
s
and deformation behavior. Fu
urther research will focus on an exten
nsion and
quanntification of the
t findings to
o a broad rangge of the indiividual input properties,
p
likke pure matrix
x and pure
blocck strength andd deformability. However, still putting emphasis
e
on meeting
m
the reequirements off a simple
appllication, basedd on easily deeterminable g eological and
d geotechnicall information,, to allow a proper
p
and
sounnd characterization for faultt rocks.
C
y Tests on BIM
M-rocks
20
T. Pillgerstorfer
Refeerences
[1]
Medley, E.
E 2001. Orderrly Characteriization of Chaaotic Francisccan Melange. Felsbau Rockk and Soil
Engineerinng Vol. 12 (4), pp. 20-33, E
Essen: VGE.
[2]
Riedmllerr, G., Brosch, F.J., Kllima, K. & Medley, E.

E 2001. Enngineering Geological
G
Characterizzation of Brrittle Faults aand Classification of Fau
ult Rocks. Feelsbau Rock and Soil
Engineerinng Vol. 19 (4), pp. 13-18, E
Essen: VGE.
[3]
Lindquist, E.S. 1994. The

T strength and deformaation properties of mlangge. Ph.D. Disssertation,
Universityy of California at Berkeley, U
USA.
[4]
Hajiabdolm
majid, V., Kaiiser, P.K. & M
Martin, C.D. 2002.
2
Modellin
ng brittle failuure of rock. In
nt. J. Rock
Mech. Minn. Sci. 39 (6), pp.
p 731-741.
[5]
Scholfield,, A.N. 1998. Dont use the C word

d Mohr Coulomb
C
erroor correction. Ground
Engineerinng, pp. 29-32.
[6]
Wieser, P., Pilgerstorferr, T. & Schubbert, W. 2012.. Developmen

nt of a Large O
Oedometer Teest for the
Determinaation of Stresss Dependent Moduli on Fault Rocks. In: Proc. E
EUROCK 201
12 - Rock
Engineerinng & Techno
ology for Suustainable Un
nderground Construction.
C
28 - 30 May
M
2012,
Stockholm
m, Sweden.
Auth
hor
Dipl.-Ing. Thhomas Pilgersto
orfer
Graz Univerrsity of Technology
Institute for Rock Mechanics and Tunnelliing
Rechbauersttrae 12
8010 Graz
Austria
[email protected]
C
y Tests on BIM
M-rocks
21

Workshops 2013
th
Triaxial Tests on Cataclasites

a
a
b

Gruppe Geotechnik Graz ZT GmbH, Graz, Austria
ETH Zrich, Institute for Geotechnical Engineering, Zrich, Switzerland
Cataclasites are highly fractured and crushed fault zone materials with generally poor mechanical
properties. They origin from tectonic activities with very high stress levels and in case of deep tunnels the
cataclasites are excavated in depth with high in situ stress. To investigate the behaviour of the low strength
cataclasites in high stress ground conditions triaxial tests with stress levels comparable to the in situ stress
are performed. The focus is set on the evaluation of the mechanical properties for the geotechnical tunnel
design of the fault zone sections. The testing procedure and the results are described in this extended
abstract.
1. TESTING MATERIAL
The ground material tested in the triaxial test consists of highly crushed and faulted phyllites and schists
and was explored during the ground investigations for the Semmering Base Tunnel project. The fault zone
material is highly heterogeneous with generally poor geotechnical properties while the described triaxial
tests focus on the very weak portions of the fault zones which are predicted with thicknesses of several
meters to tens of meters. Figure 1 shows the typical testing material.
Special care was taken during drilling procedure and handling of the samples to reduce the disturbance
before testing to a reasonable minimum. This includes drilling with triple core barrel (including core liner),
immediate vacuum sealing and protection with core liner and air bubble film. The result was a very high
quality of core samples even in heavily sheared rock mass and fault zones.
Figure 1: Examples of typical testing material; cataclasites from phyllite and schist. The material can
easily be broken by hand.
22
2. SPECIMEN PREPARATION
Depending on the degree of shearing, cataclasites can show different structures. For example, they can be
described as bands of sheared rock or hard particles embedded in a relatively soft matrix. Therefore, the
specimen preparation is very demanding, since cutting the sample with a conventional water flushed
diamond disc saw leads to very irregular cutting surfaces. For this reason, the samples are cut on an
electronically controlled air flushed diamond band saw. The saw consists of an endless metal band which
has diamonds both on its cutting edge and on the lateral sides (for polishing while cutting). The sample is
fastened with a prismatic adapter on a table which moves towards the rotating band saw. This equipment
allows vibration less cutting of the hard inclusions without disturbing the matrix (Figure 2). After cutting
one end of the specimen a parallel cut of the other end can be done by rotating the prismatic adapter by
180. During cutting, the samples are protected against drying by covering them with plastic foil. Since the
end surfaces are directly polished, no further treatment is necessary (Figure 3). After cutting, the specimen
is weighed, photographed and measured.
Figure 2: Diamond band saw.
Figure 3: Top view of sample after cutting.
3. TESTING EQUIPMENT
Triaxial tests on cylindrical specimens are the most frequently used laboratory tests to determine the
strength and deformability of geomaterials. The standard equipment varies based on the type of material to
be tested. For rock, high loads and pressures are required, whereas the pore water pressure is not controlled.
On the other hand, low loads and pressures are applied for soils (about one to two orders of magnitudes
lower than for rock), while controlling the pore pressure.
Since the samples of the cataclasites were recovered at depths of more than 350 m, high loads and pressures
must be applied in order to test the specimen at confining pressures similar to the in-situ stresses. Due to the
fact that most pores of the fault material build an interconnected system, the results need to be interpreted
according to the principle of effective stresses. Therefore, the rock testing equipment must be improved in
order to consider the effect of the pore pressure during the execution of the tests. The first experiences on
triaxial testing of fault rock specimens under controlled pore pressure were gained at the ETH Zurich with
the kakirites of the Gotthard Base Tunnel during the exploratory campaign [1]. As a consequence of the
experiences of further investigations on weak rocks from other tunnelling projects, such as kakirites (from
the Gotthard Base Tunnel during the construction stage and from the Ceneri Base Tunnel, Switzerland),
graphitic phyllites (Visp tunnel, Switzerland, and Seich Sou tunnel, Greece) and breccias from the Straits of
Gibraltar, the equipment was replaced and improved.
The used equipment consists of an electromechanical load frame with a nominal load of 600 kN (2 in
Figure 4) and was designed for applying deformation rates down to 0.1 m / min. In the triaxial cell
specimens with dimensions up to 104 mm diameter and 210 mm height can be tested (1 in Figure 4 and
Figure 5). The cell can withstand confining pressures up to 220 bars, which is generated with an electromechanical actuator with a nominal pressure of 250 bars (4 in Figure 4). The pore pressures at both ends of
the specimen are controlled with two electromechanical actuators (5 and 6 in Figure 4) with a nominal
water pressure of 30 bars. The axial load is measured with a high precision load cell (3 in Figure 4).
Characterization of Fault Zones | Triaxial Tests on Cataclasites
23
Furthermore, and in order to increase the accuracy of the measurements, the deformations of the specimen
are measured close to the specimen (1 in Figure 5), i.e. inside the cell. The axial deformation is measured
as the change of the distance between two magnets placed at both ends of the specimen (3 in Figure 4) with
a non-contact position sensor based on the magnetostrictive principle (2 in Figure 5). The radial
deformation is measured contactless with a linear magnetic encoder fixed on a chain extensometer (4 in
Figure 5) and redundantly by measuring the change of the oil volume in the cell (4 in Figure 5). The latter
is done by measuring the position of the piston of the actuator with a further linear magnetic encoder. Both
ends of the specimen are connected with the respective pore pressure devices (5 and 6 in Figure 4 and 5).
These allow applying a hydraulic gradient prior to the deviatoric loading of the specimen, in order to
increase the degree of saturation. Watering the specimen enables the determination of its permeability by
measuring the amount of water pressed in the specimen over time. After watering backpressure for
saturating the specimen can be applied.
Figure 4: Testing equipment.
Figure 5: Triaxial cell without hollow cylinder.
4. TRIAXIAL TEST
4.1. Testing procedure
The consolidated drained triaxial tests (CD) are performed as multi stage tests with three load stages with
confining stresses in the range of the in-situ stresses and a maximum of approx. 20 MPa. The typical testing
procedure consists of the following main phases.
Hydrostatic loading up to 1 MPa.

Water flow through the sample with hydraulic potential at both sides to increase saturation and
measurement of the permeability
Slow increase of the hydrostatic loading and consolidation at a hydrostatic stress level of 5 MPa.
Slow increase of the axial stress (1) until failure.
Second and third loading stage with consolidation at 10 MPa and 15 MPa.
Additionally two unloading/reloading cycles during each load stage.
Additionally, for some samples variations of the testing procedure are done as described below.
Increase of the consolidation pressure (from 5 to 10 MPa) before the triaxial test (representing
approx. 400 m of overburden and a k0 of 1.0).
Variation of the hydrostatic stress level of the different load stages (from 1 to 18 MPa).
Hydrostatic unloading/reloading cycle to measure the compression (bulk) modulus.
Testing of the specimen according to a special stress path by reduction of the axial stress (1)
starting from the assumed in-situ stress condition until failure of the specimen.
24
A full saturation of the specimen cannot be reached. But systematic investigations with kakirites from
Gotthard Base Tunnel [1] show, that partially saturated material behaves almost similar in case of an
interconnected pore system, which is gained by water flow through the specimen before consolidation. The
collected test data are presented by stress-strain diagrams (deviator stress vs. axial strain) and the relation
between volumetric strain v and axial strain 1. In this paper the result of all tests are discussed while the
result plots of only one selected sample is presented; Figure 6 shows the results for the load stage
3 = 15 MPa.
Figure 6: Results of the CD triaxial test for one load stage with confining stress of 15 MPa.
Additionally to the triaxial test geotechnical index tests such as density, water content, porosity and
saturation are performed after the test. The grain size distribution is used to classify the fault material
according to the specified ground types.
4.2. Results - Shear parameters

The tests show friction angles of approx. 24 to 28. These results are comparable with CU triaxial tests
done with typical low stress level of 3 < 0.7 MPa.
The tests show cohesions within a range of approx. 0.1 to 0.8 MPa. A correlation between cohesion and
stress level can be seen where cohesion increases with higher stress levels. Furthermore the values are
significantly higher than the cohesion values gained from CU triaxial tests with low stress level, which are
in a range of 0.10 to 0.14 MPa for similar material. The main reason for this difference can also be related
to the variation of the stress level. It can be stated that the CD tests show an increase of the cohesion with
the increase of the stress level while the Mohr Coulomb failure criterion assumes a linear development of
the failure envelope.
4.3. Results - Deformation parameters

The deformation behaviour is evaluated based on the stress-stain relations from the CD triaxial test results.
For better interpretation of the test results the deviator stresses and the axial strains are normalised by the
confining stress 3. Additionally the axial stains are set to zero after each load stage. Figure 7 shows the
results of one multi stage triaxial test. It can be seen, that the specimen behaves differently in the initial load
stage, which is related to the contraction of the material due to the generally low consolidation stress.
It can be seen that the deformation parameters depend significantly on the stress level and the loading
history. With increase of the confining stress also the E-modulus increases. Furthermore the
unloading/reloading modulus increases with the confining stress and is approx. two to three times higher
than the E-modulus during initial loading. The results can be summarized with E-modulus value of approx.
500 MPa for a stress level of 5 MPa and approx. 4000 MPa for a stress level of 15 MPa.
It is important to notice, that CU triaxial tests with low stress level and similar cataclasites showed for
example an E-modulus of 70 MPa for a confining stresses of 0.5 MPa. This underlines the importance to
test the ground material with a stress level of approx. the in situ stress.
25
Figure 7: Stress-strain diagram (left) and normalized diagram (right) for the evaluation of the
deformation behaviour.
4.4. Results Permeability

The permeability of the cataclasites is calculated form the measured flow through the sample, which is done
before the consolidation of the specimen as described in the triaxial testing procedure above. The
permeability is in a range of approx. 10-10 to 810-9 m/s with a mean value of approx. 610-10 m/s.
4.5. Hardening soil parameters (HSS)

The test results were additionally evaluated according to the Hardening Soil Model with Small Strain
Stiffness (HSS), which considers the non-linear behaviour during initial loading, the different behaviour
during initial loading and unloading/reloading, the stress dependency of the stiffness and the behaviour due
to very small strains (Small Strain Stiffness).
The stress dependent secant modulus E50 is evaluated from the lab tests with an assumed Rf value of 0.9 for
all stress levels. For example, the E50 values of one triaxial test is 450 / 2400 / 4200 MPa for 3 stresses of
5 / 10 / 15 MPa. The stress dependent unloading/reloading modulus Eur is evaluated from two
unloading/reloading cycles performed at each load stage. For example, the Eur values of one triaxial test
2.0 / 3.9 / 5.2 MPa for 3 stresses of 5 / 10 / 15 MPa. Figure 8 shows an example for one load stage with 3
= 10 MPa and an example for the unloading/reloading cycle. By evaluating all triaxial tests both, E50 and
Eur show an almost linear dependency with 3. The inclination of the linearity corresponds to a factor m of
approx. 0.8 to 1,0 for E50. For very small strains E0ref is estimated with 3-4 E50ref and the shear strain 0.7 is
estimated with 1.510-4.
Asymptote
qa
qf
Eur,1
E50
0,5qf
Eur,2
E50~2400 MPa
Figure 8: Example for the evaluation of E50 (left) and Eur (right) for one load stage.
26
4.6. Numerical back analysis of the CD triaxial test

The CD triaxial test is back analysed with the finite element code PLAXIS 2D applying the HSS model.
The FE model is set up with two rotational symmetry axes and is analysed until occurrence of failure in
each load stage.
The back analysis shows a stress-strain diagram with different behaviour of the initial load stage with low
confining stress (5 MPa) as observed in the physical test (see Figure 9). The results of the volumetric strains
are also quite similar to the lab test results.
1)
2)
1)
1)
1)
1) failure, qf ~ 1.4-1.5 q/3
2)
2)
2)
2) path after failure
Figure 9: Results of the numerical back analysis with stress-strain diagram (left) and normalized
stress-strain diagram (right).
5. CONCLUSION
The triaxial testing of cataclasites with stress levels comparable to the in situ stresses and the control of the
pore water pressure leads to reasonable material properties, which can directly be used for geotechnical
design of underground structures. With the use of sophistic testing equipment and special testing
procedures shear parameters as well as stress depending deformation parameters and permeability can be
determined for the usually very critical fault zone section of deep tunnels.
References
[1]
Vogelhuber, M. 2007. Der Einfluss des Porenwasserdrucks auf das mechanische Verhalten kakiritisierter Gesteine. Dissertation Nr. 17079, Institut fr Geotechnik an der ETH, Zrich.
[2]
Daller, J., Atzl, G. & Blmel, M. 2010. Bestimmung von Gesteinskennwerten an Strungsmaterial.
Festschrift zum 60. Geburtstag von Wulf Schubert. Institut fr Felsmechanik und Tunnelbau, TU
Graz.
[3]
Wehnert, M. 2006. Ein Beitrag zur drainierten und undrainierten Analyse in der Geotechnik.
Mitteilungen des Institutes fr Geotechnik, Universitt Stuttgart, Vol. 53.
Authors
Dipl.-Ing. Dr. Andreas Goricki
Dipl.-Ing. Dr. Erich Pimentel
3G Gruppe Geotechnik Graz ZT GmbH

Triester Strasse 478
Austria
ETH Zrich
Institute for Geotechnical Engineering
8093 Zrich
Switzerland
[email protected]
[email protected]
27

Workshops 2013
th
Rckrechnung von Gebirgsparametern in Strzonen

machbare Aussagen, Unsicherheit und Grenzen
a
P. Schubert
iC consulenten Ziviltechniker GesmbH, Bro Salzburg Bergheim, Austria
Das Thema der Rckrechnung von Gebirgsparametern wurde vom Verfasser anlsslich der Auffahrung des
Erkundungstunnels Paierdorf in der Zone der Lavanttaler Hauptstrung aufgegriffen und mit einfachen
Mitteln (Gebirgskennlinie) zu einem befriedigenden Ergebnis gebracht [1]. Die damals gewhlte
Vorgangsweise soll hier nochmals dargestellt werden, ebenso wie ein Verweis auf weitere Untersuchungen
durch andere Personen zu dieser Strungszone.
Eine Verallgemeinerung der hier gewhlten Vorgangsweise stt vor allem auf 2 Hauptprobleme: a) kurze,
inhomogene Strungszonen, wo ein starker 3D Effekt der primren und sekundren Spannungsverhltnisse
eintritt, und b) die Wahl der Vorentspannung, welche bei stark druckhaften Gebirgsverhltnissen auch
wesentlich von der Steifigkeit des Ausbaus abhngt. In beiden Fllen wrde eine Rckrechnung
aufwndigere 3D Berechnungen erfordern.
1. LAVANTTALER HAUPTSTRUNGSZONE DES EKT PAIERDORF

Der Erkundungstunnel besteht aus der Kalotte des spteren Hauptbauwerkes und wurde in der Lavanttaler
Hauptstrung berwiegend mit temporrer Kalottensohle ausgefhrt.
Abbildung 1 zeigt eine typisches Sttzmittelschema mit Deformationselementen reprsentativ fr den
Strungsabschnitt ab Station 1180:
Stauchelemente
Abbildung 1: Regelquerschnitt im Abschnitt TM 1170 bis TM 1350 mit Deformationselementen.

Im Zuge des Tunnelvortriebs wurden umfangreiche geotechnische Messungen durchgefhrt. Neben der
blichen Aufbereitung der Messergebnisse mittels rumlichen Verschiebungsentwicklungen durch
Zustandslinien wurden vektororientierte Auswertungen zur Prognose von Steifigkeitskontrasten herangezogen. Diese Methode liefert Hinweise ber das Gebirgsverhalten vor der Ortsbrust und ermglicht
zeitnahes Reagieren zur Anpassung der Sttzmittel whrend des Vortriebes.
28
P. Schubert
Als weitere Entscheidungshilfe fr den Einsatz von Deformationselementen wurde die Rckrechnung der
Spritzbetonauslastung herangezogen. Dabei wurden ab einer Grenzstauchung von ca. 8 bzw. einer
Auslastung von ca. 80% der Spritzbetonsicherung Deformationselemente zur Anwendung gebracht.
Die Setzungen in der Kalotte zeigen max. ca. 200 mm in den Bereichen mit nachgiebigem Ausbausystem
(Abbildung 2). Gegenber der Erwartung aus der Planungsphase bedeutet dies eine wesentliche Erkenntnis,
da wesentlich hhere Verschiebungswerte erwartet wurden.
2 Verformungsschlitze im Kmpfer
Station [m]
Abbildung 2: Firstsetzung im Abschnitt der Hauptstrung und Sicherungsmittel im Bereich der

Ortsbrust.
Fr die Rckrechnung der Gebirgskennwerte wurde jener Abschnitt der Hauptstrungszone ausgewhlt,
welcher mit den Deformationselementen ausgebaut wurde, das ist zwischen TM 1180 und TM 1330. In
diesem Abschnitt betrug die Firstsetzung im Mittel ca. 150 mm. Die Ermittlung der Gebirgskennwerte
erfolgte mittels Rckrechnung mit dem in der Folge beschriebenen Kennlinienverfahren.
Die Vorverschiebungen in der Natur wurden mit Hilfe der Medatenauswertesoftware GeoFit abgeschtzt.
Das analytische Modell, welches hinter dem Curve-fitting des GeoFit steht, inkludiert auch die
Verschiebungen vor der Ortsbrust, welche natrlich nicht durch die geotechnischen Messungen erfasst
werden. Der Ausbauwiderstand, welcher durch die Spritzbetonschale aktiviert wurde, konnte in diesem Fall
aus der Charakteristik der Deformationselemente abgeleitet werden. Mit der Annahme dass die Verzahnung
zwischen Spritzbeton und Gebirge in der Strung gering ist, wurde mit der aufnehmbaren Normalkraft der
Deformationselemente mittels der Kesselformel ein Radialdruck errechnet. Die Komponente der Ankerung
wurde durch die Annherung bercksichtigt, wonach die Tragkraft eines Ankers bezogen auf das Ankerraster bei voller Ausnutzung etwa dem entsprechenden Ausbauwiderstand entspricht.
Das Kennlinienverfahren darf an dieser Stelle als bekannt vorausgesetzt werden. Wegen der einfacheren
Handhabung und Mglichkeit der Adaptierung auf die Bedrfnisse der Rckrechnung wurde eine Excel
Version verwendet. Das Kennlinienverfahren beruht auf einem zweidimensionalen Modell mit ebenem
Dehnungszustand und elasto-plastischem Materialgesetz, in diesem Fall Mohr-Coulomb. Die dritte
Dimension wird durch die Wahl einer Vorverschiebung bercksichtigt, nach welcher der Ausbauwiderstand
aktiviert wird.
Die Vorgangsweise der Rckrechnung ist wie folgt (Abbildung 3): Fr einen gewhlten Satz von
Gebirgskennwerten E, c und ergibt sich eine Gebirgskennlinie und einer maximalen Verschiebung des
Ausbruchsrandes im ungesttzten Zustand umax. Dem entspricht auch ein relativer plastischer Radius
Pr = rp/ro. Gem Hoek et.al [2] leitet sich von Pr ein Verhltnis uo/umax ab. Bei einem bestimmten umax
ergibt sich also eine Vorverschiebung uo.
Characterization of Fault Zones | Rckrechnung von Gebirgsparametern in Strzonen
29
P. Schubert
support pressure [MPa] and rp/ro
3,00
2,50
2,00
Pr
1,50
0
0,55
u0
1
e 0.15 Pr
u max 3
1,00
0,50
U0
0,00
0
Um
100
200
Umax
pa
300
400
500
600
700
radial displacement [mm]
Abbildung 3: Wesentliche Komponenten des Kennlinienverfahrens und der Methodik der

Rckrechnung.
Aus den Messdaten sind die gemessenen Verschiebungen, die Abschtzung der Vorverschiebung und die
Hhe des Ausbauwiderstandes wie oben ausgefhrt - bekannt. Das analytische Modell muss nun unter
Bercksichtigung der Schritte 1 und 2 und dem aus den Messungen ermittelten Verhltnis von uo/um den
Gleichgewichtszustand ergeben. Fr den charakteristischen Tunnelabschnitt zwischen TM 1200 und TM
1300 wurden folgende Eingabegren fr die Rckrechnung gewhlt:
Tabelle 1: Grunddaten fr die Rckrechnung aus der Beobachtung whrend des Baus.
Charakteristische Firstsetzung [mm]
150
Ausbauwiderstand der Spritzbetonschale [MPa]
0,22
Ausbauwiderstand durch die Ankerung [MPa]
0,24
Summe aktivierter Ausbauwiderstand [MPa]
0,46
Verhltnis Vorverschiebung zu gemessener Verschiebung [-]
ca. 150 : 150
Mit der angefhrten Methodik gibt es in der Regel eine relativ groe Bandbreite von mglichen
Kombinationen zwischen E, C und . Die Kombinationen werden eingeschrnkt durch das Verhltnis
zwischen Vorverschiebung und gemessener Verschiebung (in diesem Fall etwa 1:1) und den plausiblen
Bereich fr den Reibungswinkel (Abbildung 4). Zu letzterem dienten eine Analyse der Kornverteilung und
die Bercksichtigung von Blockanteilen.
1,60
Back Analysis for
E=150MPa
1,40
Back Analysis for

E=200MPa
Cohesion [MPa]
1,20
Back Analysis for

E=250MPa
1,00
0,80
0,60
0,40
Plausible
Values
Back Analysis for
E=350MPa
0.75 : 1
u0/umes
0,20
0,00
15
20
1: 1
u0/umes
25
30
35
Friction Angle []
Abbildung 4: Mgliche Parameterkombinationen und Bereich mit plausiblen Werten.
30
P. Schubert
Die fr die Autoren plausibelste Kombination ist in Tabelle 2 angefhrt.

Tabelle 2: plausible Gebirgskennwerte (aus Rckrechnung).
[]
28
c [MN/m2]
0,27
E [MN/m2]
200
[kN/m3]
23
Das gesamte Verschiebungsverhalten reprsentiert in diesem Abschnitt offenbar eine Mischung aus den
unterschiedlichen mechanischen Komponenten der Strungskernzone. Einzelne Blcke oder auch
Kataklasitzonen haben sich in dem Bauschnitte mit Deformationselementen nicht sehr ausgeprgt in den
Verschiebungen niedergeschlagen. Deshalb ist es sinnvoll, fr weitere Projektierungsschritte Mischkennwerte fr eine tektonische Melange anzunehmen. Eine Einzelbetrachtung der Komponenten feinkrniger,
grobkrniger Kataklasit, Zerrttungszone und Block ist offenbar hier nicht praxisrelevant.
Die Rckrechnungsergebnisse mit der Kennlinienmethode wurden mit einem 2D Flac Modell berprft.
Dabei wurde abweichend vom Kennlinienverfahren die tatschliche Form des Tunnels gewhlt und die
Anker explizit modelliert. Das Materialgesetz Mohr-Coulomb, der primre Spannungszustand und das
Ausma der Vorentspannung wurde analog zum Kennlinienverfahren gewhlt. Dabei konnte gezeigt
werden, dass die Ergebnisse sowohl hinsichtlich Spannungs-Verschiebungsverhalten, Ausbaubelastung und
plastischer Zone weitgehend identisch sind.
2. ERGEBNISSE AUS DEM VERSUCHSSTOLLEN

Die geomechanischen Eigenschaften der Lavanttaler Hauptstrungszone wurden in einem Versuchsstollen
weiter untersucht. Darber haben Pilgerstorfer et al. [3] ausfhrlich berichtet. Es ist anzunehmen, dass der
Versuchsstollen innerhalb der starken Inhomogenitt der Strungszone in einer relativ steifen Zone lag.
Hier wurden Verschiebungsmessungen an der Schale, Dehnungsmessungen im Spritzbeton, Ketteninklinometer ber der Firste, Ortsbrustextensometer, Doppellastplattenversuche und in-situ Scherversuche
an der Kontaktflche Spritzbeton Gebirge durchgefhrt. Die integrale Interpretation der Daten fhrte zu
folgenden Kennwertempfehlungen (Tabelle 3):
Tabelle 3: empfohlene Gebirgskennwerte aus Versuchsstollen abgeleitet.
Station
[]
c
[MPa]
E
[MPa]
TM VS 12,79 m
27
0,30
1500 2100
TM VS 15,36 m
27
0,20
800 2500
TM VS 19,29 m
30
0,25
1000 1800
Man sieht, dass die Einschtzung der Scherfestigkeit (Reibung und Kohsion) sehr hnlich den
Rckrechnungswerten sind, der E-Modul jedoch deutlich hher. Wie stark dieser Befund durch die lokalen
Verhltnisse beeinflusst ist, ist meines Erachtens nicht eindeutig zu sagen. Jedenfalls scheint aber besttigt,
dass tiefliegende Strungszonen durch die hohe Verspannung der berlagerung eine sehr hohe Steifigkeit
aufweisen knnen. Der Einfluss der berlagerungsspannung auf die Gebirgssteifigkeit muss deshalb bei
der Festlegung von Kennwerten unbedingt beachtet werden.
31
P. Schubert
3. PROBLEM VORENTSPANNUNG
Die hier gezeigte, offenbar durchaus erfolgreiche Rckrechnung bezieht ihre Vorverschiebungswerte aus
den 3D Modellrechnungen von Hoek et al. [2]. Diese Rechnungen wurden allerdings ohne Ausbau
gemacht. Cantieni und Anagnostou [4] haben uns allerdings drastisch vorgefhrt, dass die Vorentspannung
bei stark druckhaften Verhltnissen und steifem Ausbau erheblich anders zu bewerten ist. Ein steifer
Ausbau reduziert dann die vorauseilende Entspannung ganz erheblich und erhht damit die Lasten des
Ausbaus.
In unserem Fall ist der Ausbau durch die Deformationselemente sehr weich, wodurch dieser Effekt stark
reduziert wird. Andernfalls scheint aber auch aus diesem Aspekt eine 3D Simulation unausweichlich.
4. PROBLEM 3D EFFEKTE
Die oben angefhrte Rckrechnung fr die Lavanttaler Hauptstrungszone beruht auf einem ebenen 2D
Modell. Das ist bei der 450 Meter langen Strungszone berechtigt. Bei kurzen Strungszonen treten jedoch
erhebliche Randeffekte durch die benachbarten kompetenteren Gebirgsbereiche auf, wodurch die
Spannungs- und Verschiebungsprognose eines ebenen Modells zunehmend falsch wird. Zum Beispiel hat
Graziani [5] gezeigt, dass bei den in seiner Studie gewhlten Parametern ab einer Strungslnge von etwa
50 m die Randeffekte weitgehend verschwinden, und unter 20 m Strungslnge sehr erheblich werden
(Lasten sinken auf unter 50%). Eine Verallgemeinerung ist hier sicher nicht zulssig und kurze Strungen
wrden jedenfalls ein 3D Modell fr eine Rckrechnung erfordern.
5. SCHLUSSFOLGERUNG
Es ist hinreichend erwiesen, dass Tunnelstatik mit vielen Fragezeichen behaftet ist. Dasselbe gilt natrlich
fr Rckrechnungen jeder Art. Der Verfasser ist jedoch berzeugt, dass die einfache Methode der
Rckrechnung mittels Kennlinie am ebenen Modell im Vergleich zur Ableitung von Kennwerten aus
Bohrkernen und Laborversuchen zu erheblich wirklichkeitsnheren Ergebnissen fhrt und deshalb in jedem
Fall lohnend ist. Bei komplexen geometrischen Verhltnissen ist allerdings eine aufwndige 3D
Modellierung nicht zu vermeiden.
Referenzen
[1]
Schubert, P., Hlzl, H., Sellner, P. & Fasching, F. 2010. Geomechanical knowledge gained from the
Paierdorf investigation tunnel in the section through the Lavanttal main fault zone, Geomechanics
and Tunnelling 3 (2), pp. 163-173, Berlin: Ernst & Sohn.
[2]
Hoek, E. 1999. Support of very weak rocks associated with faults and shear zones. In: International
Symposium on Rock Support and Reinforcement Practice in Mining. Kalgoorlie, Australia.
[3]
Pilgerstorfer, T., Radoncic, N., Moritz, B. & Goricki, A. 2011. A Evaluation and interpretation of
monitoring data in the test adit EKT Paierdorf. Geomechanics and Tunnelling 4 (5), pp. 423-434,
Berlin: Ernst & Sohn.
[4]
Cantieni, L. & Anagnostou, G. 2009. The Effect of the Stress Path on Squeezing Behavior in
Tunneling. Rock Mech Rock Eng 42, pp. 289-318, Wien: Springer.
[5]
Graziani, A., Capata, A. & Romualdi, P. 2007. Analysis of Rock-TBM-Lining Interaction in Squeezing
Rock, Felsbaumagazin 25 (2007), Vol. 6, pp. 22-31, Essen: VGE.
Autor
Dipl.-Ing. Dr.mont. Peter Schubert
iC consulenten Ziviltechniker GesmbH
Zollhausweg 1
5101 Bergheim
Austria
[email protected]
32

Workshops 2013
th
Tunnelling through a Major Fault: Considerable Overburden,

Squeezing and Flowing Ground; At-Face Selection of Support
a
U. Glawe
Freelance Consultant in Geomechanics & Rock Engineering
Unexpected occurence of large faults in tunnelling is common in hydropower and infrastructure projects
located in remote areas. It requires adequate geological and geomechanical characterisation of the fault and
of the encountered materials. Tailoring appropriately the available engineering tools to the prevailing
ground conditions is essential for doing well in the construction of the tunnel. This case study directly
addresses such circumstances, where the top heading of a tunnel was driven through an unpredicted fault
and the obtainable construction utensils were modified to deal with the extraordinarily difficult ground
conditions, successfully. The fault is getting characterised by its geometry, the fault structure, the
geological units in the fault as well as by the hydrogeological conditions. Implemented adjustments of the
support and of auxiliary measures to tunnel through squeezing and flowing ground of the fault are
addressed and typical monitoring results are presented.
1. INTRODUCTION AND PROJECT SETTINGS

This contribution to the workshop refers to the construction of the top heading of a 1.175 km long tailrace
tunnel of a hydropower project in tropical SE-Asia. The project is located on the Malaysian Peninsula
150 km south of the Thai/Malay border 50 km west of the coastline.
After tendering and commencement of the initial construction works (river diversion) additional geological
investigations at the tailrace were conducted. They revealed a zone of deteriorated rock at the tunnel level at
a depth of 100 m. The detail design incorporated these adverse ground conditions and provided relevant
assistance to tunnel through the anticipated conditions within the contractual outline of the awarded
project.
The tunnel has an excavation diameter of 9.70 m (for the heaviest support type). After 400 m in reasonably
good rock the top heading entered the zone of deteriorated rock. The following 100 m revealed severe
ground conditions. In addition groundwater was encountered, which led to erosion and squeezing and
flowing ground in the central zone of this tunnelling section. During construction this 100 m wide zone was
identified as a fault of very large scale with outer damage zones and a central fault core.
The contractual available ground reinforcement and support had to be modified to address the prevailing
conditions: there were only basic and conventional methods available that were adjusted to meet the
requirements to build the tunnel, safely.
2. REGIONAL GEOLOGY AND ENCOUNTERED ROCK TYPES

The project region is geologically characterised by low seismic activity and by N-S trending large scale
folds of Permo-Triassic Age.
In the project area the following lithological units are prevailing (Figure 1):
sedimentary rock (low grade metamorphic); graphitic shale, siltstone, fine-grained sandstone
volcanic rock (andesite)
interlocking strata (sedimentary/volcanic), which was defined (thickness, mechanical
characteristics) during construction of the diversion works
granite (4 km away, bee-line distance)
The area is covered by dense jungle and exhibits a hilly topography (elevation 150 m to 400 m asl.). The
depth of weathering depends on rock type and other factors, but reaches for the volcanic rock typically
33
U. Glawe
down to 40 m below ground level at a maximum, where the boundary of moderately to slightly weathered
rock is located. The actual soil thickness (= ISRM weathering grade V and VI) is around 20 m in general.
Figure 1: Schematic sketch showing the project arrangement and the tunnel with respect to the
major geological structures and encountered lithological units.
3. FAULTS AND DISCONTINUITY PATTERN

During previous stages of the project (feasibility study to tender design) focus was predominantly on
identifying faults at the actual dam site. A few faults of minor extend (several tens of metres) with small
off-set were identified through evaluation of aerial photos and field investigations. There was no indication
of the existence of a large scale fault anywhere in the project area.
Beside bedding joints in the sedimentary and the interlocking sequences a discontinuity pattern typical for a
rock mass folded at large-scale is prevailing with well developed extensional joints (ac-jointing) and
conjugated shear jointing (hk0-jointing).
4. TUNNEL GEOLOGY AND EXCAVATION CLASSES

4.1 Tender Stage
For tendering the geological model shown in Figure 2 was used. It describes in sufficient detail the general
ground conditions known at that stage of the project. Four excavation classes were specified for different
rock classes (Q values), which included shotcrete and rock bolts. Two heavier classes were assumed to be
necessary for the portal section and for crossing potential minor faults. Hence, typical support and
reinforcement used for such conditions - steel sets, canopy tubes and face nails were added in the design
and the head & bench method was suggested for excavation.
Figure 2: Sketch showing the geological model for the tunnel at tender design stage. (Ch. = chainage
or tunnel meter).
Characterization of Fault Zones | Tunnelling through a Major Fault
34
U. Glawe
The construction schedule required that the 1.175 km long tunnel will be driven towards the powerhouse
complex from downstream and down-hill at a gradient of 1%. At the downstream portal the tunnel will
discharge into an exiting reservoir. The full supply level of this reservoir is close to the tunnel roof at the
portal. The respective water level is indicated in Figure 3.
4.2 Detail Design Stage

During detail design further geological investigations were conducted along the alignment with focus on
the portal excavation. Furthermore a 120 m deep inclined borehole was drilled (HTS5 in Figure 3) to better
characterise the expected contact of the volcanic rock with the sedimentary rock.
Figure 3: Sketch showing the geological model during detail design prior to construction. FSL stands
for Full Supply Level of the existing downstream reservoir. The dashed red line represents the
assumed rock line. The direction of the tunnel drive is indicated with the black arrow.
Surprisingly, down to an elevation of 10 m below the tunnel invert the entire drilling showed soil like
material with a few small blocks of volcanic rock. The soil in the drilling was visually and from its
estimated grain size distribution similar to residual soil produced by weathering of the volcanic rock.
However, deep-seated weathering was soon excluded from potential sources as the soil was found down to
a depth of more than 100 m and the surface of slightly weathered rock is located generally not deeper than
40 m below ground level. The surfaces of the few blocks in the drilling block showed coatings and signs of
cauterisation. Hence, hydro-thermal alteration along a minor fault or similar genesis was considered to be a
reasonable geological explanation.
Figure 4: Drill cores obtained from borehole HTS-5 (see Figure 3) from the tunnel level with a tripletube core barrel. The borehole penetrated the tunnel at around Ch. 750 (compare Fig. 5 and Table 1
for actual geology at Ch. 750).
35
U. Glawe
As the location and geomechanical characteristics of the contact sedimentary/volcanic rock was still
unidentified, in-depth geological mapping was performed at the morphological ridge along the tunnel
alignment. Quartz pebbles in the residual soil were used as indicator of sedimentary strata. Based on the
results of mapping another borehole was drilled (HTS44 in Figure 3). It revealed that the contact is not
sharp, but it is characterised by interlocking of the individual strata. This interlocking was irrelevant for the
design and for the construction of the tunnel as both rock types are of good quality when fresh.
A newly developed geological model was set up as presented in Figure 3. It shows a zone of deteriorated
rock, soil with rock blocks from Ch. 727 to Ch. 809. A rock line was introduced that separates the soil
type material from the material with rock properties. At that stage of the project it was still unknown that
the tunnel would cross a wide major fault at a cover thickness of approximately 100 m.
The designed tunnel support for the zone of deteriorated rock is characterised by the following
components:
Steel sets HB 200 x 200 x 49.9 kg/m, grade 275, 1 m c/c; on top heading arch outer rib with
footing (steel plate) and inner rib with connecting plate, only
Canopy tubes for forepole umbrella,114.3 x 6.3, grade 275, 0.5 m c/c, 9 m, 3 m overlap, outward
directed at 6.5 to the tunnel centre line
Fibreglass face dowels, 12 m long, 1.5 m c/c staggered, 3 m overlap (= 30% of excavation diameter)
250 mm shotcrete SFR 40, crown and side walls
50 mm SFR 40 at the face
Central exploratory drill hole

In addition to these measures there were provisions for cement grouting and drainage measures provided in
the contract. It is pointed out that the listed support and auxiliary measures were basically the only available
engineering tools to use in tunnelling through the squeezing and flowing ground in the fault core.
Fortunately, the designer has addressed uncertainties in the tunnel geology through permission of directives
of the engineer at the spot as some of the above listed engineering measures had to be significantly
modified, reinforced and tailored to engineering requirements to deal with the prevailing ground conditions.
This is further detailed in chapter 6.
In early 2012 the tunnel was driven from the portal (Ch. 1+176) towards the powerhouse complex. It
entered the damage zone of the fault at Ch. 828.
5. GEOLOGICAL CHARACTERISTICS OF THE FAULT

5.1 Fault Geometry
There were no morphological signs of the fault and of its orientation prior to the construction of the tunnel.
Even now, after driving the tunnel through the fault, the exact orientation of the fault is not entirely clear.
There are several geological indications, such as lithological boundaries found in horizontal probe holes
(drilled at Ch. 765 and Ch. 741, see Figure 5), or the discontinuity pattern in the host rock next to the fault,
which point to a steep dip angle of the fault and an obtuse angle between the tunnel axis and the fault plane.
Further indication referring to the fault orientation is a sharp contact in the fault damage zone at Ch. 721,
which is sub-vertically dipping and orientated perpendicular to the tunnel axis. Based on these observations
it is considered that the fault is dipping at steep angle and is striking in N-S direction (= orthogonal to the
tunnel axis). This statement refers to the location where the tunnel passed through the 100 m fault. The
extension of the fault away from the tunnel is unknown.
One may question why such major fault could not be identified by the geologist prior to driving the tunnel
into the fault. A possible explanation is: both soil types, the completely weathered/residual volcanic
material (ISRM weathering grade V and VI) and the fault material exhibit very similar geological
characteristics. This refers in particular to inspected drill cores as well as to the grain-size distribution (high
fine content with some sand, gravel and blocks). Such resemblance indicates similar proneness to erosion.
This fact and the 20 m thick soil cover of the volcanic material adjacent to the fault does not allow the
development of topographic features typical for large faults in moderate or arid climate.
36
U. Glawe
5.2 Fault Structure and Fault Geological Units

The structure of a fault is shown schematically in Figure 5 and the actual geology exhibited in the tunnel is
detailed in Table 1.
During tunnelling both damage zones were identified and a 39 m wide central fault core could be
designated. Most of the boundaries between zones were clear, but there is also a 12 m wide transition zone
between damage zone and fault core at one side of the fault core.
From Ch. 735 to Ch. 724 strong marble blocks were encountered in the matrix material of the fault core
(light blue blocks in Figure 5). This was entirely unexpected as there are no limestone or marble outcrops
known in the nearby area (for at least several kilometres away from the tunnel). The metamorphic marble
blocks are embedded in a soil matrix that showed no signs of lithification or metamorphism. The block
surfaces showed karstic features. The origin and the genesis of the marble blocks remain mysterious.
Figure 5: The upper block diagram shows the schematised structure of a fault (block diagram
modified after [1]). The actual fault geology and thickness of individual zones observed in the tunnel
is presented in the geological profile below including discharge into the tunnel. The numbers in the
geological profile refer to those used in Table 1. The red bars in top of the profile indicate the extent
of deteriorated rock predicted for detail design. The green lines indicate schematically the probe
drillings ahead of the respective tunnel face.
37
U. Glawe
Table 1: Fault geology, fault zoning and respective tunnel support. The numbers of the zones refers
to the numbers in Figure 5.
#
Chainage
Geology
Zone in Fault
Support
(schematic)
To 828
Andesite (Q > 50)
Host Rock
Spot Bolting
828 - 817
Blocky andesite with randomly shaped large

angular blocks (edge length up to 3 m). No
direct block/block contacts, but blocks are
separated by thin (< 3 cm) yellowish finegrained soil material.
Damage Zone
Ribs &
Shotcrete
817 - 799
As from Ch. 817 to 799 with long (< 1.5 m) and

thick (< 0.3 m) lenses of fine material between
blocks
Damage Zone
Canopy Tubes
& Ribs &
Shotcrete &
Face Support
799 - 775
Andesite (2 < Q < 50)
Large Block in
Damage Zone
Syst. bolting &

Shotcrete
775 - 763
Gradual transition from the andesite (as at Ch.

799 - 775) to fault core material (as at Ch. 763 745).
Transition
Damage Zone /
Fault Core
Ribs &
Shotcrete
763 745
Yellowish-brownish soil material (see Figure

6): intermittent grain size distribution, rounded,
rounded/angular and angular andesite blocks
(block edges always rounded) with a max.
diameter of 0.5 m embedded in soft to medium
stiff fine-grained matrix (see Table 2). Volumetric block content < 25%. Unit weight of blocks
ranging from 18 kN/m3 to 26 kN/m3 indicating
(hydro-thermal?) alteration. Cohesive BiM-rock
fc 25 (after [1]).
Fault Core
Canopy Tubes
& Ribs &
Shotcrete &
Face Support&
Auxiliary
Measures
745 - 724
Brownish-blackish soil material (Figure 7) with

physical characteristics similar to the fault core
material described above (Ch. 763 to 745), but
softer. Locally internal layering in the matrix
material and slickensides (Figure 8).
Volumetric block content << 25%, blocks are
more angular than in previous section. Cohesive
BiM-rock fc 25.
Rounded marble blocks with a diameter of up to
2 m encountered from Ch. 735 to 724.
724 711
Transition from fault core material to damage

zone material (here fractured andesite) from Ch.
724 721. Sub-vertical persistent few centimetres thick fault at Ch. 721. Rapid decrease of
jointing from Ch. 721 to Ch. 711 where Q > 10.
Damage Zone
Initially Ribs
and Shotcrete,
then Shotcrete
and Syst.
Bolting
from 711
onward
Andesite (Q > 10, generally >> 50)
Host Rock
Bolting, Spot
Bolting
38
U. Glawe
Table 2: Grain size distribution and natural moisture content (NMC) determined for the brownishblackish and the yellowish-brownish matrix material.
Brownish-blackish Matrix
Sieve Size [mm]
Yellow-brownish Matrix
Passing
88%
98%
0.212
77%
86%
0.075
75%
71%
NMC
28%
30%
Figure 6: Tunnel face at Ch. 760 prior to shotcreting and placing of steel set. Cohesive yellowishbrownish matrix material and dark-grey andesite blocks (BiM-rock fc25).
Figure 7: Lump of brownish-blackish undisturbed matrix soil of the fault core (Ch. 730). The
material feels similar to damp compacted peat. For grain size distribution and NMC see Table 2.
39
U. Glawe
Figure 8: Brownish-blackish matrix material with slickensides (left) and layering (right). The
slickensides were presumably generated by deformations ahead of the face (Ch. 735) and not by
tectonic processes. The layering may indicate previous internal tectonic deformation of shear lenses
in the fault core.
5.3 Groundwater and Soil Sweating

The groundwater table observed in borehole HTS5 and other drill holes is indicated in Figure 3. Maximum
discharge into the tunnel occurred at the outer boundaries of the individual damage zones: once the tunnel
entered the damage zone at Ch. 828 local discharge of approximately 3 l/sec. was observed (Figure 5).
From Ch. 817 to Ch. 735 discharge occurred locally, generally from the tunnel side walls. In addition,
where the fine-grained matrix was exposed, sweating was observed. This superficial pore water pressure
release was presumably caused by the increased tangential stresses in the face or side walls, unconfined
conditions and due to the high natural moisture content of around 30% for both matrix types. The beads of
water that developed soon after exposure changed rapidly into a thin mud coating of the exposed soil
surfaces and led to serious problems during shotcrete application.
Water discharge increased significantly from Ch. 745 onward when the tunnel approached the other
boundary fault core / damage zone. Over the last 12 metres in the fault core, discharge was entering the
tunnel along a few distinct flow paths in the tunnel face with a total of about 5 l/sec.. Some of the few flow
channels discharged dark-brown muddy suspension (dissolved brownish-blackish matrix material), which
indicated the development of erosional cavities ahead of the face.
6. TUNNELLING THROUGH THE FAULT

6.1 Tunnelling from the Damage Zone into the Fault Core
Due to reasonable stability of the face and limited water discharge in Zone 5 of the fault (see Figure 5 and
Table 1), the top heading was driven full face with all available support and auxiliary measures described
in section 4.2 with a round length of 1 m.
The tunnel drive encountered no worth mentioning problems until it reached Ch. 757, where a combined
face-crown collapse occurred (Collapse I in Figure 5). As with most tunnel collapses there are several
reasons for such failures [2], adverse ground conditions were in this case one of them.
A more than 8 m high chimney-like cavity was back-filled in three stages with a total of more than 80 m3 of
light-weight concrete and some cubic metres of mortar. Light-weight concrete (or foam concrete) was
selected as back-fill material due to its good workability and excavatability not due to its light weight,
only. The design mix of
160 kg sand, 650 kg OPC, 150 l water, 6.5 kg superplasticiser and 11 kg poly-beads per m3
40
U. Glawe
resulted in a unit weight close to 10 kN/m3 and a 28 days cube strength of 6 MPa. This strength allows that
the material can be excavated easily with a standard excavator or even with a pick with difficulty.
After re-establishment of the tunnel a crown failure (= Collapse II in Figure 5) occurred at the same
location as Collapse I (Ch. 757). This happened just before the next rib could be placed underneath the
newly installed canopy tubes. As with the previous collapse, adverse ground conditions were one of the
reasons of the failure. The situation was re-established after only a few days: more than 40 m3 of lightweight concrete and some mortar were pumped into the collapse cavity.
6.2 Tunnelling in the Fault Core

After the second re-establishment, the tunnel was progressively driven through the fault core. The tunnel
instabilities experienced called for adjustments of the support and change of the excavation system. The
heaviest support type was already in place and there was only the option to utilise available tools, which
required on-the-spot modification as described in the following sections, and step-by-step implementation.
The tunnel advanced from Ch. 757 to Ch. 745 with the support and auxiliary measures detailed below at a
rate of less than 0.2 m/day.
6.2.1
Additional fibreglass dowels
To increase face stability in the tunnel the overlap of the face dowels was increased to more than 50% of
the tunnel diameter. This was in line with experience reported by others, e.g. [3] that such overlap is
required to significantly increase face stability in cohesive ground. Increasing the overlap to more than 50%
of the tunnel diameter (= more than 5 m) means for this case effectively reducing the spacing of the 12 m
long dowels by approximately 50% to 0.75 m c/c.
6.2.2
Partial and Sequential Excavation
The excavation was subdivided into three sections and a face buttress was developed. Further along the
tunnel the face was subdivided into five and more excavation and shotcrete sectors. This allowed better
control the face (stability) and eventually created a thick shotcrete shell.
6.2.3
Drainage Canopy
A de-watering shield consisting of 9 m long perforated PVC-pipes wrapped with geotextile and drilled inbetween individual canopy tubes at steeper angle than the tubes was installed to drain water away from the
soil ahead of the face. However, as the ground water followed distinct flow channels in the cohesive soil of
the fault core, very few of the tubes were productive and therefore this measure was no longer implemented
in the tunnel drive.
6.2.4
Temporary Invert Liner
The design provided provisions for a thin temporary shotcrete lining of the invert in the top heading. As the
ground was so soft and could not withstand mechanical impact such as excavator tracks and for other
practical reasons, the shotcrete liner was replaced by a 200 mm mesh reinforced concrete slab.
6.2.5
Change of Excavation Techniques
To separate individual excavation sections in the face, line drilling through the shotcrete was implemented
with a drilling bit of small diameter (2 inches). To avoid overbreak and disturbance of the face during
excavation, the excavator was replaced by a pneumatic percussion drilling rig with a 150 mm drilling bit.
The so easily excavated material was later removed with an excavator.
6.2.6
Improvement of Steel Set Foundations with Cement Grouting
Bearing capacity of the steel set foundations was considered to be crucial for distributing forces from the
stiff mechanical unit forepoles - steel sets - shotcrete vault into the ground. Softening of the foundations
would seriously call into question the entire engineering approach in this regard. Pressure cement grouting
was conducted from the face into the up-coming foundation material, but soon abolished due to the required
effort and questionable success to very locally achieve ground improvement.
41
U. Glawe
6.2.7
Improvement of Steel Set Foundations through Grouted Rib Footings and Ring Drains
Softening of the material in the small excavated pits for the steel set footing due to accumulation of water
from sweating and from local water ways was addressed by the following approach: the readily
excavated pits were cleaned shortly before shotcreting. The gap between ribs and the ground was filled with
aggregates and shortly after with grout to create a load bearing connection between the steel set and the soil
foundation material (Figure 9).
A wound roll of geotextile was fixed onto the walls with nails to prevent water flowing from the walls and
from the tunnel face into the pit. The geotextile roll was extended further into the tunnel profile with a
ribbed PVC-pipe (Figure 9) and simply covered with shotcrete. This simple drain worked well prior and
after shotcreting.
Figure 9: Details of rib footings prior to shotcrete application. The footing of the outer ribs
(highlighted with the dashed white line) stands on aggregate and a wooden formwork was installed
before the rib footing pit is filled with grout. A ring drain of wound geotextile is nailed onto the soil to
prevent water from the walls entering the excavation pit of the footing. A steel plate is welded on the
inner rib in preparation for the installation of micro-piles.
6.2.8
Provisions for Micro-pile Foundations for Steel Sets
In the soft ground micro-piles can act as auxiliary foundation of ribs. Hence, there were an option. To
connect possible micro-piles to the ribs, rectangular steel plates (940 mm x 370 mm x 20 mm) were welded
onto the inner flange of the inner rib (Figure 9).
6.2.9
Strengthening of Vault through dense Steel Sets and increased Shotcrete Thickness
Steel rib spacing was gradually decreased from 1.0 m to 0.5 m and the thickness of shotcrete could be
increased only by increasing the excavation diameter to assure the shotcrete does not penetrate the specified
B-Line (here defined as outer boundary of reinforced concrete liner).
42
U. Glawe
6.2.10 Modification of Canopy Tubes for Pressure Grouting

The canopy tubes were initially designed as un-perforated steel tubes filled with grout followed by annulus
grouting. To modify the tubes for pressure grouting a simple system was implemented as shown in
Figure 10.
Figure 10: Modified canopy tube for pressure grouting. The non-perforated and perforated sections
are divided by a sleeve of mild steel of the diameter of the borehole and welded onto the tube. In a
first stage the annulus of the none-perforated section was filled with grout at low pressure (red =
return pipe). In a second stage the steel tube was filled with grout until return occurred through the
red pipe. This pipe was then closed and pressure grouting was conducted.
6.3 Tunnelling in Squeezing and Flowing Ground of the Fault Core

Due to the low advance rate, which was caused by the necessity to apply required support and implement
auxiliary measures as described above, it could not be avoided that the soil was getting longer in contact
with water and got increasingly soft. From Ch. 745 onward discharge was observed in the face. Erosion of
the soil material ahead of the face was indicated by brown soil suspension flowing into the tunnel.
Once the tunnel entered Zone 7 of the fault at Ch. 745 (see Figure 5), which comprises material that is even
softer than the previous material, squeezing phenomenon were evident and the situation in the tunnel got
increasingly serious. The main concern was possible sudden and brittle failure of the entire shotcrete-ribvault due to bearing capacity failure of the rib foundations.
Figure 11: Spalling of shotcrete at the tunnel perimeter in-between canopy tubes. Radial cracking of
the shotcrete at the face and buttress further demonstrated ongoing instability in the face.
43
U. Glawe
Spalling of shotcrete as well as radial cracking of the shotcrete in the face and face buttress indicated that
the canopy tubes were failing ahead of the face and getting bent down due to squeezing ground (Figure 11).
The tunnel face was getting pushed into the tunnel. Despite all efforts, eventually at Ch. 741 the face gave
way and face collapse of a few cubic metres occurred (Figure 12).
Figure 12: Collapse III at Ch. 741. This face collapse of a few cubic metres only, occurred after
observation of significant squeezing of the tunnel face and the adjacent roof and the side-walls. The
blocks in front of the debris are blocks of matrix soil including smaller rock blocks that were
separated from the face. Note the bent-down canopy tubes in the crown and the discharge of muddy
suspension from the face. The latter was creating large erosional cavities with a total volume of
several tens of cubic metres ahead of the face.
After establishing reasonably safe conditions in the tunnel, three exploratory bore holes were drilled which
indicated far better ground conditions approximately 20 m ahead of the face (for location of drillings and
orientation see Figure 5).
During drilling of the canopy tubes large cavities generated by soil erosion ahead of the face were
discovered. They required filling with more than 40 m3 of light-weight concrete. In addition cement
grouting was conducted along the tunnel perimeter prior to placing the next canopy tubes. At Ch. 741 more
than 30 tons of cement was pumped into the perimeter before the works advanced.
It was evident that the so far implemented measures were insufficient for tunnelling in such poor and
unpredictable ground. To deal with these conditions further support and auxiliary measures were
implemented step by step when the tunnel progressed.
6.3.1
Perimeter and Face Grouting
Perimeter and face grouting was considered necessary for the upcoming 20 m in adverse ground as an
auxiliary measure to contribute to stabilize the face: 9 m long holes were drilled in-between canopy tubes
and grouted at a low water cement ratio with a maximum pressure of 10 bars. The holes were left open and
unsupported prior to grouting; only a 1.5 m steel tube was inserted and the hole was sealed.
44
U. Glawe
Grouting pressure and w/c ratio were consistently adjusted. Grout takes varied significantly, from nil in
some holes to several hundreds of kilograms of cement per hole. In some of the holes that had high grout
takes, grouting was stopped when no pressure changes but continuous takes were recorded. Grouting was
started along the perimeter to radially confine the soil ahead of the face. This was followed by
systematically grouting of the face towards the centreline.
A few grout holes penetrated distinct water paths. These holes were used as drain holes after a 9 m long
geotextile-wrapped PVC-pipe was inserted (where possible).
6.3.2
Reinforcing Canopy Tubes and Canopy Grouting at significant Discharge
The canopy tubes drill holes intercepted distinct groundwater flow paths with increasing discharge and the
previously introduced tube pressure grouting system (shown in Figure 10) became ineffective. The canopy
tubes were reinforced with three centrally placed 26 mm steel bars of grade 460 and modified as shown in
Figure 13.
Figure 13 a and b: Modified canopy tube for grouting in ground with high water discharge. The tube
is closed at both ends. After installation of the tube the annulus is closed at the face. This is followed
by grouting the canopy tube. This procedure guarantees that the entire tube is filled with grout and
no grout is washed away. Together with the three centrally placed steel rebar it can act as stiff and
strong cantilever. In a second stage the annulus is (pressure) grouted.
45
U. Glawe
6.3.3
Dense and Double Forepole Umbrella
The spacing of canopy tubes was further reduced to 0.4 m and where the space allowed, a second row of
canopy tubes was installed. The latter procedure was only possible at Ch. 741 (Figure 14).
Figure 14: Upper section of the face at Ch. 741 with double forepole umbrella on the right side of the
roof and side wall. Note that there is a third row of tubes of the previously installed canopy tube set
above the steel set which reaches 3 m ahead of the last installed steel set. The tubes were installed
with protruding ends to minimise the excavation required for the installation of the next steel set. The
yellow circles indicate optical targets placed strategically onto the last steel rib and in the tunnel face
and buttress. This allowed monitoring displacements during canopy installation and during the
excavation.
6.3.4
Reinforced Micro-Piles as Steel Set Foundations
As previously noted, bearing capacity failure of the foundation soil beneath rib footings was a very serious
concern of the engineer in charge. The results of tunnel monitoring indicated continuous and differential
settlements of the shotcrete-rib vault (Figure 15 and Figure 17). Radial cracking of the shotcrete between
steel ribs at Ch. 746 occurred.
Section 6.2.8 describes the preparation of the ribs to connect them to micro-piles and hence immediate
installation was possible. Material available at the site (= canopy tubes) was used as micro-piles: 6 m long
steel tubes, 114.3 x 6.3, grade 275, one pile at each side of a rib (= four piles per steel rib). The principle for
fixing the piles to the ribs is detailed in Fig. 16. It leads to immediate mobilisation of the piles at any
settlements of the ribs.
Due to restriction in space (drilling rig and location of foot plate of the outer rib) the micro-piles could only
be drilled at a maximum inclination of about 30 to the horizontal. Hence, they are acting primarily through
bending rather than skin friction and end bearing. Therefore, the piles were reinforced with a centrally
spaced steel bar of 26 mm diameter (grade 460).
Initially it was not clear whether the micro-piles could substantially support the steel set foundations and
take the loads transferred from the reinforced forepole umbrella onto the ribs. Tunnel monitoring was an
essential tool to observe functioning of the micro-piles and to identify their role in tunnel stabilisation. The
monitoring results presented in Figure 17 indicate that the micro-piles required more than 20 mm of
displacements until they were significantly contributing to stabilise the tunnel, which is considered to be
small bearing in mind the poor ground conditions.
46
U. Glawe
Figure 15 a and b: Displacement monitoring results for sections at Ch. 738 and 745 for a period of 27
days in early 2013 (Feb 23 until March 22, 2013). The results indicate differential settlements
between the two sections as well as differential settlements within the section at Ch. 738.
47
U. Glawe
Figure 16: Construction principle for fixing the micro-pile steel tubes onto the steel sets leading to
immediate mobilisation of the piles (frontal view): a) A steel plate (940 mm x 370 mm x 20 mm) is
welded onto the inner rib. U-shaped openings upside down are cut into the plate; the yellow lines
represent welded seams. b) Holes are drilled for the piles (into the drawing plane and downward).
c) Holes are filled with grout (red) and the micro-piles with a centralized 26 mm steel rebar are
inserted. d) Small steel plates of 20 mm thickness with upside down U shaped openings are welded
onto the larger plate and connected to the steel tube (yellow lines are welding seams).
Figure 17: Results of monitoring the crown settlements at Ch. 745 for the period Dec 07, 2012 until
April 05, 2013. After the collapse at Ch. 741 the micro-piles were soon installed at Ch. 756 to Ch. 745
and settlements continued. It took 23 mm and 11 mm of settlements, respectively, after the piles were
installed to stabilise the tunnel (= cease of settlements).
48
U. Glawe
6.4 Tunnelling out of the Fault Core into better Ground

The anticipated ground ahead of the face (Ch. 741 to Ch. 724) consisted of
Soft matrix with few small andesite blocks and completely saturated soil ( = mud)
Few strong marble blocks (c >> 150 MPa) with a diameter of up to 2 m (Ch. 735 to Ch. 724)
Water in distinct flow paths
Light-weight concrete and grout
Unexpected cavities of unknown volume and location
Hence, all necessary support and auxiliary measures were put in place and the tunnel advanced at low rate.
Figure 18: Canopy tube grouting at Ch.735 after pneumatic (dry) drilling and tube installation. The
photograph shows the release of liquefied in-situ mud of the brownish-blackish matrix material from
the holes, which is flowing down the face.
Figure 19: The geological conditions got rapidly better once the damage zone was reached. Fractured
andesite and lenses or bands of clayey-silty soil at Ch. 722 are exhibited in the tunnel face. The lenses
of soil are marked with dashed yellow lines. Some of the lenses were connected with long shear
fractures (dashed white lines).
49
U. Glawe
The material was at some locations so soft that it was flowing into the tunnel when the face was opened or
during drilling of holes for canopy tubes (Figure 18). Surprisingly, out of the 40 m3 of light-weight concrete
and 30 tons of cement pumped into the face at Ch. 741, only a few cubic metres were found in the mucked
material.
As shotcrete shell settlements and squeezing were still ongoing extreme care was taken. At one location one
could visually observe opening of radial cracks in the face. The situation was controlled by immediately
placing a stabilising buttress/ramp of rock fill against the face and spraying additional 250 mm SFR 40 onto
the face, starting from the crown perimeter and continuing downward.
Simultaneously with the cease of the rib settlements (Figure 17) the last canopy tube set in the fault core
was installed at Ch. 729. These canopy tubes reached into better ground (= damage zone) and squeezing
stopped promptly. From Chainage 724 onward the ground conditions improved rapidly (Figure 19).
Break-through of the top heading of this tunnel was in mid July 2013 and it was in schedule of the project.
Finally it is noted that despite the three collapses and frequent and unavoidable very dangerous working
conditions at the face and in the tunnel no one was injured and no equipment was lost in this operation.
7. FINAL REMARKS
The top heading of a 9.7 m excavation diameter tunnel has been driven successfully through a 40 m wide
core of a major fault at 100 m overburden thickness. The material consisted predominantly of soft soil and
water was encountered leading to squeezing and flowing ground conditions. This required modification of
the initially designed support and auxiliary measures.
The geology of the fault is considered to be reasonably well understood and characterised, but there are
many open geological questions. These inquiries refer to the genesis of the fault, the materials in the fault
(e.g. the origin of the marble blocks) and several other geological issues.
The adjustment of the measures to deal with the prevailing ground conditions was essentially based upon
engineering and experienced-based judgement. The approach selected by the authorised engineer in
charge avoided lengthy academic discussions, but crucial decisions were made pragmatically and promptly
at the tunnel face, and they were instantly directed.
This case study is contributed due to its high relevancy to the subject of the workshop. It is envisaged to
publish this data as part of a more comprehensive paper that will include reference to and
acknowledgement of all participatory stakeholders.
References
[1]
Fasching, F. & Vanek, R. 2011. Engineering geological characterization of faults. Geomechanics

and Tunnelling 4 (3), pp. 181-194, Berlin: Ernst & Sohn.
[2]
Seidenfuss, T. 2006. Collapses in Tunnelling. Master Thesis, Technical College Stuttgart, Stuttgart.
[3]
Cassani, G. 2013. Excavation of tunnels in structurally complex formations. Geomechanics and

Tunnelling 6 (3), pp. 197-214, Berlin: Ernst & Sohn.
Author
Ulrich Glawe
PhD. Eng Geol.
MSc. Eng Rock Mech., DIC.,
MSc. Eng Geol.
[email protected]
LinkedIn: Ulrich Glawe
50

Workshops 2013
th
A Fault Zone Management for Deep Seated Tunnels

a
U. Burger , T. Marcher , E. Saurer & L. Soldo

Galleria di Base del Brennero Brenner Basistunnel BBT SE, Innsbruck, Austria
b
ILF Consulting Engineers Ltd., Rum/Innsbruck, Austria

c
GEODATA Engineering S.p.A., Milan, Italy
1. INTRODUCTION
Geological uncertainties and the ensuing risks in the construction of long tunnels at great depth has been
described in the ITA Report no. 4 - Long tunnels at great depth (ITA) [1]: . the deeper the tunnel, the
larger the uncertainties; the higher the probability of encountering adverse or unforeseen conditions for
tunnelling, the greater the effort and the cost for site investigations to reduce the uncertainties.
Among the identified hazard sources, faults play a dominant role, due to their squeezing potential, swelling
and creep, possible inflow of water (and/or gases) and debris, or eventual displacements along active shear
zones.
Tunnelling through fault zones may therefore lead to critical events for both the construction process and
the safety of the personnel. In order to minimise the risk, to be prepared in the case of an event or to
overcome fault zones efficiently, a Fault Zone Management Plan has been developed. This Fault Zone
Management Plan provides a systematic process developed for the cross-border BBT (Brenner Base
Tunnel) railway tunnel at great depth between Italy and Austria (Bergmeister [2]).
Information from other base tunnels, the existing BBT geomechanical project and the knowledge from
currently driven BBT exploratory and main tunnels will be taken into account in order to optimise the fault
zone management process.
2. SPECIFICS FOR A TUNNEL AT GREAT DEPTH

2.1. General
Examining the design projects of tunnels at great depth, the experience gained from the first BBT
exploration tunnels crossing different fault zones and from international tunnel projects served to define
several specifics for tunnels at great depth. These specifics have to be considered in the fault zone
management plan.
Experience in the BBT project include the following:
Perpendicular main fault in granite crossed by the Aica-Mules exploration tunnel with a
instationary inflow of 160 l/s (Perello et al. [3]).
A fault zone in granite striking sharply to the tunnel axis of the Aica-Mules exploration tunnel: due
to the deformation a four month stop of TBM followed (Barla et al. [4]).
A fault zone striking perpendicular to the tunnel axis of the shallow tunnel Saxen: the tunnel
crossed a principal nappe fault with approx. 20 l/s instationary inflow.
Main fault zones in the phyllite, striking parallel to the exploration tunnel Innsbruck-Ahrental and
flat lying fault zones as crossed by the access tunnel Ahrental with approximately 15 l/s
instationary inflow.
51
2.2. Specifics for tunnels at great depth derived from design projects and
experiences in fault zone crossing
Specifics for tunnels at great depth, which have to be considered in the Fault Zone Management Plan:
Regional character of tunnel projects at great depth:
Due to the depth of the tunnels, the involved volume of the rock mass is large; this means that
regional knowledge of the geology and hydrogeology is required.
Long tunnels at great depth usually cross multiple tectonic units, therefore tunnels at great depth
usually cross regional fault zone systems. Additional literature on this topic see e.g. Damiano et al.
[8] and Eusebio et al. [9].
Due to the regional character of the fault zone systems, similar faults may be crossed several times
and even by different lots (e.g. NE-SW and N-S striking faults in the BBT project).
Events may have a regional impact, larger than the actual construction lot limits.
Complex shear zones:
Deformation along shear zones leads to the characteristics of the rock mass, including the fabrics
and mineral assemblages.
Major shear zones that cross the crust down into the upper mantle show both brittle and ductile
sectors. Brittle fault related rock masses, such as breccias or gouges, are obviously related to most
of the critical conditions for tunnels, both during excavation and operation.
Complex project:
Tunnels at great depth have more construction lots.

Tunnels at great depth have huge project teams, often situated in different places and even in
different nations.
2.3. Risk management

The Risk Management paper published by the International Tunnelling Association (ITA) [5] is widely
considered as a guideline. The AFTES Recommendations [6] follows this conceptual approach.
The insurance and re-insurance groups are very actively promoting the use of Risk Management at all
stages of a project in order to minimize insurance losses. An International Code of Practice, (german
version: Richtlinien zum Risikomanagement von Tunnelprojekten) which follows the ITA guidelines
closely, has been published by the International Tunnelling Insurance Group (ITIG) [7].
In simple terms, the risk management approach consists in identifying and listing the potential hazards
associated with the tunnelling activities, assigning a probability of occurrence to each hazard, and
allocating an index of severity to the consequence. The next steps involves a definition of the measures to
reduce the probability of occurrence of an event and to reduce the severity of the consequence (the so called
mitigation measures).
The analysis process prosecutes reassessing the remaining risk level after the application of the mitigation
measures, obtaining an updated risk level, a "residual risk level". This "residual risk level" should be
examined for acceptance (and then shared among the parties involved in the project) considering the
"global cost" necessary for reducing or completely eliminate the source of risk.
The adoption of a RMP enhances the conceptual framework given by the observational method
establishing a previously planned, sound and rational, framework for the design changes during
construction that finally creates a favourable environment for cost or time savings and avoid unnecessary
claims.
The proposed BBT approach for the risk management related with fault zones gives a comprehensive map
of actions that shall be followed to avoid accidents together with the necessary countermeasures in presence
of unfavourable conditions. But also it introduces a procedure for their continuous updating during
construction, based on the registered evidences and feedbacks implementing appropriate amelioration if
necessary.
Characterization of Fault Zones | A Fault Zone Management for Deep Seated Tunnels
52
3. FAULT ZONE MANAGEMENT - WORK PROCESS

3.1. Flow chart
The general structure of fault zone management in terms of working steps, necessary decisions and
additional input is illustrated in the BBT Fault Zone Management Flow Chart (see Figure 1).
Figure 1: Fault Zone Management Plan flow chart with main packages of the working process in
the case of an event.
53
As can be seen from the flow chart, the Fault Zone Management Plan can be applied during the
construction of the exploratory tunnel (see left black line on the flow chart) and the main tunnels (see right
black line on the flowchart). The major steps (boxes) are described in the following chapters.
3.2. Localisation and preliminary fault zone model

On account to the large size of the project team it becomes necessary to provide a first characterization of
the fault zone. The first fault zone model has to be distributed to all members of the project. Therefore a
simple template has to be filled out by the site geologist, the site geotechnical engineer and the
responsible person for the construction lot.
Main contents of the template are:
Localization of the fault zone in relation to the tunnel

Preliminary geological fault zone model including graphic illustration
Decision whether or not the fault is situated in a zone with applicable authority requirements
Decision about necessary additional investigations from a geological and / or geotechnical point of
view
Definition of the event class (minor, critical, major)
Activation of required members
3.3. Investigation
If the first characterisation shows that an investigation programme is recommended for preparing the final
geological and geotechnical model, a detailed investigation programme including the aim of the
investigations has to be drawn up. The purpose of the investigation programme is to characterise in detail
the fault zone with respect to:
Localisation, orientation and geometry

Material
Hydraulics
A tool box with investigation methods and the related investigation interests is included in the fault zone
management plan.
3.4. Geological, hydrogeological and geotechnical model

The geological / hydrogeological model shall contain at least all parameters included in the template of the
fault zone catalogue (see Figure 2).
54
Figure 2: Template for the fault zone characterisation. The upper 2 sheets include a description,
illustration and characterisation of the geology and geotechnics of the fault zone. The lower 2 sheets
of the template contain information on the deformation behaviour in the case of available data from
exploration tunnels. The example shows the information for a fault zone inside the Quartz phyllite
striking parallel to the Innsbruck-Ahrental exploration tunnel and dipping east.
55
3.5. Hazard identification and definition of countermeasures

With the information obtained from the geological/hydrogeological and geotechnical model, the dominant
hazard has to be identified and defined (hazard identification).
A table shall be used for the definition of the hazards (type) and their description (in columns). An excerpt
of such a table is provided in the tool box (see Table 1). In the column labelled indicators, parameters
which could give an indication of the type of hazard are provided.
Table 1: Definition of possible hazards (excerpt).
For the design, a distinction is made between preventive countermeasures and those applicable in the case
of the occurrence of an event.
Preventive countermeasures
Event-related countermeasures and equipment
A list with fault zone phenomena and proposed countermeasures from experience made during tunnel
construction under comparable boundary conditions and corresponding references is given in a tool box
(see Table 2).
The definition of possible hazards and preventive countermeasures are based on experience and on
literature, e.g. Amberg [10], Bonzanigo et al. [11], Daller et al. [12], Fellner et al. [13], Ferrari et al. [14],
Rthlisberger et al. [15], Sausbriber et al. [16], Stadelmann et al. [17], Weh et al. [18], Wildbolz [19],
Ziegler [19].
56
Table 2: Fault zone phenomena and proposed countermeasures (excerpt). Cases A to D are referred
to the knowledge of a fault zone: Case A = assumed fault zone; Case B = fault zone localized with
investigations ahead of the tunnel; Case C = close to the tunnel lying fault zone localized by
interpretation of monitoring data; Case D = intercepted fault zone.
3.6. Application of countermeasures and monitoring

The designed countermeasures are applied in order to overcome the construction difficulties in the fault
zone. Main steps are:
The countermeasures have to be applied according to the design.

During the application, the behaviour of the fault zone has to be observed and monitored
intensively in order to be able to validate the design.
A report on the application of countermeasures and the excavation has to be prepared (e.g. as-built
reports from BA/DL).
4. CONCLUSIONS
The proposed BBT fault zone management plan provides a comprehensive action plan that shall be
followed in the case of an event. The plan includes different tool boxes such as proposals for required
countermeasures in presence of unfavourable fault zone conditions during the excavation process.
The presence of a permanent team of specialists, specifically working within the framework of the fault
zone management approach, facilitates transferring the continuously identified design and construction
best procedures and solutions into the subsequent design and tunnel construction packages.
The advantages derived from this ongoing learning, test, apply and transfer sequence will provide a
wealth of experience to the whole project team.
Experience in the next BBT construction lots may provide the opportunity to validate and verify the action
plan and to improve and adapt it.
57
References
ITA WG 17. 2010. Long Tunnels at great depth. http://www.ita-aites.org/en/wg-
[1]
committees/committees/itatech/downloads/content/22/working-group-17-longtunnels-at-great-depth. (last access 08/2013).
[2]
Bergmeister, K. 2011. Brenner Basistunnel der Tunnel kommt. pp. 263, Lana: Tappeiner.
[3]
Perello, P., Baietto, A., Burger, U. & Skuk, S. 2013. Excavation of the Aica-Mules pilot tunnel for
the Brenner base tunnel: experience gained on water inflows in tunnels in granitic massifs. Rock
Mechanics and Rock Engineering. In press.
[4]
Barla, G., Ceriani, S., Fasanella, M., Lombardi, A., Malucelli, G., Martinotti, G., Oliva, F., Perello, P.,
Pizzarotti, E.M., Polazzo, F., Rabagliati, U., Skuk, S. & Zurlo, R. 2010. Problemi di stabilit al fronte
durante lo scavo del cunicolo esplorativo Aica - Mules della Galleria di base del Brennero. In: Barla,
G. & Barla, M., Proc. XIII ciclo di conferenze di meccanica e ingegneria delle rocce MIR 2010,
Torino.
[5]
ITA Working Group No. 2. 2004. Guidelines for tunnelling risk management. Tunnelling and
Underground Space Technology 19, pp. 217-237.
[6]
AFTES Recommendation GT32.R2A1. 2012. Recommendations on the characterisation of

geological, hydrogeological and geotechnical uncertainties and risks.
[7]
International Tunnelling Insurance Group (ITIG). 2006. Richtlinien zum Risikomanagement von
Tunnelprojekten (A Code of Practice for Risk Management of Tunnel Works).
[8]
Damiano, A., Venturini, G., Alzate, M., Mancari, G., Dematteis, A., Soldo, L. & Vendramini, M.
2011. The Reference Geological Model for the Corredor Bioceanico Aconcagua project. In: Congrs
International AFTES: Lyon.
[9]
Eusebio, A., Soldo, L., Pelizza, S. & Pettinau, D. 2005. Preliminary geological studies for the design
of Mercantour base tunnel. AFTES Tunnelling for a sustainable Europe: Chambery.
[10]
Amberg, F. 2006. Geologie, Vortriebsmethoden und Bauhilfsmassnahmen in der Multifunktionsstelle Faido. Geologie und Geotechnik der Basistunnels, pp. 225-237, Zrich: Lw Hochschulverlag.
[11]
Bonzanigo, L. & Oppizzi, P. 2006. Low Angle fault zones and TBM excavation in Bodio section of
Gotthard Base Tunnel. Geologie und Geotechnik der Basistunnels, pp. 156-184, Zrich: Lw
Hochschulverlag.
[12]
Daller, J., Atzl, G. & Weigl, J. 2011. The new Semmering base tunnel tunnel design in the fault
zone. Geomechanics and Tunnelling 4 (3), pp.237-253, Berlin: Ernst & Sohn.
[13]
Fellner, D., Bachmann, M. & Bnsch, W. 2007. DS-TBM over its Limits / Schwierigkeiten mit
Zuckerkrnigem Dolomit beim Bau des Shanggongtunnels in Sdwestchina. In: Internationales
Symposium BBT: Innsbruck.
[14]
Ferrari, A., Fabbri, D. & Sidler, M. 2007. Bewltigung von unerwarteten Strzonen beim TBMVortrieb im Gotthard-Basistunnel Baulos Bodio. In: Internationales Symposium BBT, pp. 56-62:
Innsbruck.
[15]
Rthlisberger, B., Stadelmann, R. & Wei, Z.Q. 2007. Bewltigung der Herausforderungen im
Bauabschnitt Faido des Gotthard Basistunnels. In. International Symposium BBT: Innsbruck.
[16]
Sausbriber, T. & Brandner, R. 2001. The relevance of brittle fault zones in tunnel construction.
Lower Inn Valley Feeder Line North of the Brenner Base Tunnel, Tyrol, Austria. Mitt. Osterr.Geol.
Ges. 94, pp. 157-172.
[17]
Stadelmann, R., Rehbock-Sander, M. & Rausch, M. 2007. Crossing of fault zones in the MFS Faido
by using the observational method. Underground Works under Special Conditions. Romana,
Perucho, Olalla (eds). London: Taylor & Francis.
[18]
Weh, M. & Bertholet, F. 2011. TBM advance in the main adit to the Nant de Drance pumped storage
power plant. Geomechanics and Tunnelling 4 (5), pp. 284-591, Berlin: Ernst & Sohn.
[19]
Wildbolz, A. 2006. Befreiung der Tunnelbohrmaschine Gabi II in Amsteg. Schweizer Baujournal.
[20]
Ziegler, H.-J. 2011. 150 km Vortrieb durch die Alpen Ingenieurgeologische Erfahrungen. In: Swiss
Tunnel Congress 2011.
58
Authors
Mag. MSc. Ulrich Burger
Dipl.-Ing. Dr. Ing. Thomas Marcher

Amraser Strasse 8
6020 Innsbruck
Austria
ILF Consulting Engineers Ltd.

Feldkreuzstrasse 3
6063 Rum/Innsbruck
Austria
[email protected]
[email protected]
Dipl. Bau-Ing. Dr. sc. ETH Erich Saurer
Geol. MSc Ph.D. Luca Soldo
ILF Consulting Engineers Ltd.

Feldkreuzstrasse 3
6063 Rum/Innsbruck
Austria
GEODATA Engineering S.p.A.

Via Leopardi 1
20123 Milan
Italy
[email protected]
[email protected]
59

Workshops 2013
th
Numerical Analysis of Fault Zones - Coming Closer to a Solution

a
M. Brandtner
IGT Ziviltechniker GmbH, Salzburg, Austria
Numerische Modellierung von Strungszonen - Eine Annherung: In der GG Richtlinie fr die

geotechnische Planung von Untertagebauten [1] ist ein wichtiger Entwurfsschritt die berprfung, ob das
ermittelte Systemverhalten mit den Anforderungen bereinstimmt. Fr heterogene Bereiche bzw. Strungszonen reichen analytische Methoden wie z.B. das Kennlinienverfahren bzw. numerische 2D Modelle wegen
der unzutreffenden Randbedingungen bezglich rumlichen Spannungszustand und Porenwasserdruck im
Falle eines Bergwassers nicht mehr aus. In manchen Fllen sind axialsymmetrische Modelle noch
vertretbar. Die Inkaufnahme etwaiger Ungenauigkeit muss in jedem Fall gesondert beurteilt werden.
Bezglich der Modellierung des primren Spannungszustandes und des Porenwasserdrucks knnen diese
Modelle in den meisten Fllen eine 3D Anylyse nicht ersetzen.
According to the Guidline for the Geotechnical Design of Underground Structures published by the
Austrian Society for Geomechanics [1], the assessment of the system behaviour is an essential step in the
design process. For heterogeneous ground conditions and fault zones methods with closed solutions based
on the Ground Reaction Curve (GRC) and numerical 2D analyses are inapplicable due to unrealistic
boundary conditions concerning the stress state and pore water pressure in case of groundwater. In rare
cases axisymmetric models can do the job. The acceptance of this mismatch has to be judged in each
individual case seperately. However in terms of initial stress conditions and pore water pressure these
models cannnot fully replace 3D models.
1. CHARACTERISTICS OF THE 3D MODEL

In most cases the normal vector of the fault zones orientation plane is not parallel to the tunnel axis and
therefore an axisymmetric model cannot reflect the real situation. For the sake of reduced model size a
symmetry plane can be utilized if the fault zone strikes perpendicular to the tunnel axis and the initial stress
state is homogeneous (Figure 1), a full 3D model has to be chosen otherwise.
Figure 1: Typical 3D model for fault zone with transition zones adjacent to unfaulted regions.
60
M. Brandtner
The determination of suitable model boundaries is essential. One should bear in mind that each plane on the
boundary is a symmetry plane anyway and the influence to the region where the tunnel is driven through
the fault zone should not be influenced by model boundaries.
2. MATERIAL MODEL FOR ROCK MASS

Since the area of interest is rather in the zone of the intersection between tunnel and fault zone, the Mohr
Coulomb criterion in unfaulted regions is sufficient. This might have some disadvantages concerning the
undrained behaviour of the rock mass, but in order to reduce computational costs it can be accepted. In the
faulted region an advanced model has to be employed for several reasons:
Different stiffness for unloading and reloading (excavation)
Realistic stress path in combination with pore water pressure (strength)
Double hardening constitutive model (shear and compression hardening)
Stress dependent stiffness (yield surface expands with increase of plastic straining)
The Hardening Soil Model was originally developed for soils, but for cataclastic rocks it seems to be
suitable as well.
Figure 2: Determination of parameters from laboratory tests.

The comparison of stress strain relations gained from laboratory tests (Figure 2 left side) with calibrated
parameters for the constitutive model (Figure 2 right side) shows a good compliance and justifies the
deployment of the HSS-Model. The consideration of groundwater needs a special treatment but different to
soil mechanics. The thesis by Wehnert [2] gives an excellent overview concerning drained vs. undrained
analysis, but deals only with soils, where the bulk modulus of the skeleton material (Ks) is much higher
than that one of bulk material itself (Ks >> K). In soil mechanics it may be accepted that saturated ground is
assumed to be incompressible, but deep rock tunnels are exposed to a higher stress level. Water has a bulk
modulus of approx. 2.3 GPa and cannot be treated as incompressible anymore. A starting point to find an
adequate theoretical basis for the treatment of water under such conditions is the theory of porous rocks. It
should be emphasized that the debates on the applicability of the principle of effective stresses for rocks
seem to be a never ending story, but it is accepted that Terzaghis approach has to be adapted for a rock
mass [3] [4]. A different approach, which was applied successfully in several pressure tunnels, is provided
by Innerhofer [5]. In addition attention should be paid to the contact zone between fault and unfaulted rock
mass. Whilst the material in the fault zone may have very low permeability, the adjacent rock mass can
even be nearly dry. In [6] some situations are sketched.
Characterization of Fault Zones | Numerical Analysis of Fault Zones
61
M. Brandtner
Figure 3: Contact between porous rocks according to [6].
3. STRUCTURAL ELEMENTS
The shotcrete lining in reality behaves highly non linear, but for the sake of simplicity on this scale of
models a pragmatic approach has been chosen. An average stiffness of 5 GPa regardless of age or time
dependent creep effects is accounted for the support resistance of the primary lining, which is modelled
with linear shell elements. Rock bolts have been completely omitted. The resistance is taken into account
by increase of the cohesion. Special attention was paid to the effects of damping elements. The dissipation
of displacement energy results from LSC elements (lining stress controllers) with a given stress strain
relation. These elements are modelled with nonlinear beams showing a characteristic working curve
depicted in the diagram below.
Figure 4: Working curve for damping elements.

After a shortening of approx. 50% of the total element length (0.4m) absorption of energy is no longer
possible and the axial load is transferred directly to the primary lining.
4. RESULTS
If dealing with the development of pore pressure the analysis has to be made in real time scale and post
processing the results is another challenging task. One should bear in mind that the job is not done when the
computation is finished, most of the times the effort for looking through and visualizing the results is
underestimated. Exemplarily some results are show in the following figures.
62
M. Brandtner
Figure 5: Absolute displacements after the tunnel has been driven through the fault zone.
Figure 6: Axial forces in the damping elements (Lining Stress Controllers).

The elapsed computation time for typical models containing features described above is between one day
and half a week. It is obvious that performing parameter studies ends up in excessive computational costs.
However we should keep in mind that such analyses have the objective to assess the system behaviour and
cannot be seen as a detailed stress analysis.
References
[1]
Austrian Society for Geomechanics. 2010. Guideline for the Geotechnical Design of Underground
Structures with Conventional Excavation.
[2]
Wehnert, M. 2006. Ein Beitrag Zur Drainierten Und Undrainierten Analyse in der Geotechnik. Mitt.
Inst. fr Geotechnik.
[3]
Nur, A. & Byerlee, J.D. 1971. An Exact Effective Stress Law for Elastic Deformation of Rock with
Fluids. Journal of Geophysical Research Vol. 76, pp. 64146419.
[4]
De Boer, R. & Lade, P.V. 1997. The Concept of Effective Stress for Soil, Concrete and Rock.
Gotechnique 47, pp. 6178.
[5]
Innerhofer, G. 2008. Action of Force on Rock Mass by Crack Water Pressure. Geomechanik und
Tunnelbau Vol. 1, pp. 583589.
[6]
Mandl, G. 1988. Mechanics of Tectonic Faulting. Amsterdam: Elsevier.
Author
Ing. Markus Brandtner
IGT Ziviltechniker GmbH
Mauracherstrae 9
5020 Salzburg
Austria
[email protected]
63
Workshops 2013
th
Design of Ductile Supports

a
W. Schubert , N. Radoni
Graz University of Technology, Institute for Rock Mechanics and Tunnelling, Graz, Austria
b
Geoconsult ZT GmbH, GTU Koralmtunnel, Wals bei Salzburg, Austria
1. INTRODUCTION
The combination of high stresses and poor ground quality frequently leads to large displacements,
unsustainable by conventional tunnel supports. Ideas for ductile supports have been around for many
decades, but systematic application has started on a larger scale only in the nineteen nineties. Over the past
decades, various ductile elements have been proposed and sometimes successfully applied. The successful
applications showed to increase safety, as well as considerably reduce displacements. While most systems
fulfilled their role as deforming element, costs and operational feasibility, as well as their interaction with
shotcrete and deforming ground are seldom discussed. This paper focuses on important issues in design and
implementation of ductile elements in combination with shotcrete linings.
2. REQUIREMENTS
The requirements on the ductile support are straightforward and based on common imperatives: structural
safety and operability and cost-oriented considerations. They can be summarised as follows:
1.
The overall structural stability, safety and the minimal clearance profile must be granted in all
situations by the applied support system.
2.
Conventional materials and construction methods should be used, and the applied concept should
not be different from the other support concepts applied in the tunnel.
3.
The load-displacement relationship of the entire support system, and therefore of the yielding
elements, should be easily adaptable in accordance to the observed system behaviour.
4.
The support resistance should be maximized, in order to reduce the displacement magnitude
fluctuations in heterogeneous ground and reduce ground disintegration in the vicinity of the
excavation. In addition, the required overcut of the profile is minimised.
5.
The load-displacement relationship of the yielding elements should be adjusted to the timedependent shotcrete strength development and the overall displacement characteristic of the
ground. Generally said, low initial stiffness (in order to prevent overstressing of the young
shotcrete) followed by a smooth and steady ductile region are favourable. The load-displacement
relationship should not feature strong oscillations, in order to prevent the entire system jumping
into the next valley and producing avoidable fluctuations in final displacements (Figure 1).
64
Figure 1. Unfavourably oscillating (a) and favourable (b) load-displacement relationship, with regard
to the overall support resistance and final displacement prediction.
3. BASIC EFFECT OF DUCTILE LININGS

The application of open deformation gaps or installation of ductile elements allow the otherwise too stiff
shotcrete lining to deform in such degree that an equilibrium with the relaxing ground can be attained. As
the experience has proven time and again, the usual approach, common for conventional structural analysis,
of increasing the capacity of the structure until the loads can be sustained, leads to additional stiffness
increase and attraction of additional loads. Simply put: the effort of devising a support which would be able
to withstand the ground pressure is a race which is either impossible to win or very costly.
In case open gaps are used, the contribution of the lining is very limited and support mobilisation of the
lining is governed by shear bond between the ground and the shotcrete and shear resistance of the rock
bolts. Hence, system behaviour is very sensitive to changes in the ground quality, and the displacement
field of the tunnel support can strongly change within a few meters of advance. This leads to additional
stress-redistribution processes triggered in the ground (since stiffer rock mass portions attract additional
stresses and are prone to brittle failure), and lining cracks due to incompatible displacement differences
between two adjacent areas. Additional disadvantages are the occurrence of long-lasting displacements
(since the support pressure is low) and presence of a kinematically free lining system in case of sudden
failure (Schubert & Riedmller 1995).
The application of yielding elements (or open gaps) results in certain kinematical peculiarities which have
to be accounted for. The absorption of tangential strains by the yielding elements inevitably results in
relative displacements (slip) developing between the shotcrete and the ground, and additional shear loading
of the installed rock bolts. The occurring slip and its consequences have a significant influence on the axial
forces in shotcrete segments and the overall support mobilisation. In addition, if only the top heading is
excavated, the axial loading of the shotcrete is usually unsustainable by the ground beneath the lining feet,
and they penetrate into the ground considerably. This leads to the generation of two additional yielding
elements and translational movement of the entire support (Figure 2).
Characterization of Fault Zones | Design of Ductile Supports
65
Figure 2. Large deformations and their influence on the lining incorporating yielding elements;
Top: undeformed system state; Bottom: deformed system state.
4. DESIGN
4.1. Boundaries of conventional support
Many criteria for fast determining whether a ductile support must be used or not have been published over
the years, some based on the idea of critical radial strain (Hoek 2001, Aydan et al. 1993) or on the
definition of a critical ground classification number (Singh et al. 1992, Goel et al. 1995). However, all of
the proposed methods fail to recognise that the ground forms an integral part of the ground-support system,
and that the yielding support is simply required when the conventional one is not able to economically and
safely withstand the loads. Simply put: the application boundaries for a ductile support are defined by the
capacity of conventional support methods being applied at the time, and should be changed in accordance
to the progress of material technology.
Radoncic (2011) conducted a Monte-Carlo simulation with this inverse reasoning in mind: first the
maximum capacity of different conventional support concepts (immediate ring closure, top heading, top
heading with temporary invert) has been determined, and then the ability of the chosen concept to attain
equilibrium with the randomly chosen set of ground parameters has been checked. The results show a clear
trend, featuring a clear boundary defined by overburden and radial strain, calculated for the case of
unsupported ground (Figure 3).
66
Figure 3. Results of Monte-Carlo sampling, showing a clear boundary between lining failure and
equilibirum.
The systematic analysis allows the drawing of the pre-design chart depicted in Figure 4 and the definition
of the relationship (Equation 1).
Figure 4. Boundaries of support failure for a friction angle of 15.

The relationship includes friction angle as an parameter since the study has shown that the boundaries move
up the ability of the support to attain an equilibrium with the ground rises with the increasing friction
angle.
H crit
2

X

( H 0 H tan ) 75 1
X 0
*
Equation 1
67
The corresponding parameters for different support concepts are given in Table 1.
Table 1. Parameters defining the critical overburden Hcrit.
Full-face excavation
Top heading
Top heading w. Invert
X [-]
0 [-]
H0 [m]
H*[m]
0.062
0.062
0.030
0.035
0.045
0.030
100
100
75
680
680
375
4.2. Pre-design
Radoncic et al. (2009) have presented a method based on extensions to the convergence confinement
method, allowing incorporation of the influences of the ground conditions, advance rate, time-dependant
shotcrete strength and overall support layout into the analysis.
Support pressure [MPa]
0.1
0.2
0.3
0.4
Radial displacement [m]
5
4
3
2
1
0
0.1
0.1
0.2
0.3
0.4
0.5
Radial displacement [m]
0.2
0.3
0.4
0.5
10
15
20
25
5
4
3
2
1
0
0.6
0.6
Distance to face [m]
Distance to face [m]
0.6
0.5
5
10
6
Time [d]
6
10
12
10
12
15
20
25
30
30
35
35
Figure 5. Determination of the shotcrete thrust capacity (blue line: maximal mobilised shotcrete
thrust dotted line: total support resistance, green line: available shotcrete thrust capacity).
Radoncic (2011) added additional improvements, incorporating the influence of shear bond between ground
and shotcrete and dowel action of the rock bolts (Figure 6).
68
Krb,shear,7 Krb,normal,7
Fshear,7
u7
Lseg
8
9
u,tot
KYE
Figure 6. Applied discretisation of a shotcrete half-segment, with rock bolts acting like non-linear
springs (both with regard to axial deformation and to tangential shear).
The applied discretisation scheme allows the determination of the shotcrete thrust distribution over the
entire length of the shotcrete segment (Figure 7), while still being directly combined with thedesign graph
presented in Figure 5.
Figure 7. Influence of rock bolting density and friction angle at the rock mass lining contact surface
on the thrust distribution in shotcrete (UCS equals 1 MPa, Yielding element thrust of 900 kN, tunnel
radius of 5 meters).
69
4.3. Design and fine tuning

The internal anisotropy of the tectonic faults causes a pronounced mechanical anisotropy. This has a strong
effect on the system behaviour, influencing the final displacements, displacement development and the
overall contrast in the radial displacements. The internal shear bands act as natural planes of movement and
suppress the re-distribution of shear stresses. Hence, in case of an advance perpendicular to the fault
orientation, the displacement field tends to be homogenous and the displacements tend to be low, however
strong initial displacements can be expected. In the opposite case (parallel advance), the highest radial
displacements occur perpendicularly to the internal structure, and the displacement development is longlasting. Such effects should be accounted for in the final stage, and yielding element layouts should be
prepared for the most likely cases.
4.4. Construction and monitoring

It indisputable that in case tunnelling through highly heterogeneous faults, the observational approach
represents the only feasible design method. Hence, the design has to be constantly updated based on the
observed system behaviour. In this case, the issues of required over-excavation and shotcrete utilisation /
support resistance represent the key issues. Due to already discussed kinematic peculiarities, the shotcrete
utilisation and section forces cannot be back-calculated from the absolute displacement monitoring, since
the tangential displacements and strains of the shotcrete shell (at the yielding elements position) feature a
discontinuity of an unknown magnitude. Hence, it is strongly recommended to install additional
displacement monitoring points at both sides of the yielding elements, allowing direct and straightforward
determination of the support resistance. Strain gauges can used for gaining additional information, allowing
direct determination of the section forces in shotcrete and supplementing the observations with absolute
displacement monitoring.
5. CONCLUSIONS
Following recommendations and conclusions are deemed generally applicable in case of designing and
constructing a tunnel with integrated yielding elements:
1.
Although appearing as a trivial conclusion, the roles of proper investigation, ground

characterisation and ground behaviour determination cannot be overstressed. Both the
investigation programme and the lab test programme have to be custom-tailored to the envisioned
analysis means and identified failure mechanism. This renders the design, especially in the early
project stages, an iterative process: after all possible mechanisms of stress re-distribution and
displacement development have been identified; additional effort has to be invested into
determining reliable parameters for the best-suited calculation model.
2.
Proper determination of ground behaviour is an excellent starting point for the basic support
design: after the influences of structure, rheological behaviour and water have been determined,
the basic layout of lining should follow the anticipated deformation pattern. The layout of the
yielding elements within the cross section does NOT have to be symmetrical, but to orientate itself
on the kinematics of a ductile lining subjected to unsymmetrical deformation field.
3.
The yielding element load-displacement characteristic must comply to the time-dependent

development of shotcrete capacity and the displacement development characteristic of the system.
Displacement monitoring (especially if including the respective element shortening) should be
frequently and systematically conducted and evaluated by competent personnel. A great room for
optimisation with regard to overexcavation dimensions, associated support measures and risk
minimization is present when the measurement data are evaluated on time and the right
conclusions are drawn.
4.
The chosen excavation sequence has to closely follow the envisioned support mobilisation.
Advancing a top heading with yielding elements of extreme capacity without taking care about the
proper abutment of the top heading feet is basically meaningless. The systematic rock bolting,
temporary invert installation and/or elephant feet have a limited effect on the top heading
resistance mobilisation. While the temporary invert causes adverse support geometry, the elephant
feet and rock bolting are intrinsically bound to the ground properties and can have a strongly
varying effect. Above a certain level of thrust required in the lining, a short bench and early ring
closure should be envisioned, however this does not immediately imply that full-face excavation
70
should be used. It is associated with several operational problems: use of heavy machinery, low
accessibility of the crown and shoulders, a great amount of immediate displacement (and energy
release) and heavy face support are the usual consequence.
5.
The rock bolt pattern has to adhere to the chosen yielding element layout. Long and stiff rock bolts
should be either installed far from the yielding elements, or other means of preventing the
combined thrust and shear loading should be used. The general rock bolting concept should
combine frequent short rock bolts (with the goals of homogenizing the rock mass and increasing
its ductility in the vicinity of the excavation, where the strains are highest) and several strong and
long rock bolts for the overall stability (thus anchoring the shotcrete segments and the loosened
weight of the rock mass).
References
[1]
Schubert, W. & Riedmller, G. 1995. Geotechnische Nachlese eines Verbruches Erkenntnisse und
Impulse. Proc. 10th Christian Veder Kolloqium, Semprich et al. (eds.). Graz.
[2]
Hoek, E. 2001. Big tunnels in bad rock. 36th Karl Terzaghi Lecture, ASCE Journal of Geotechnical
and Geoenvironmental Engineering. ASCE. Seattle.
[3]
Goel, R.K., Jethwa, J.L. & Paithakan, A.G. 1995. Tunnelling through the young Himalayas a case
history of the Maneri Uttarkashi power tunnel. Engineering Geology 39. pp. 31-44.
[4]
Radoncic, N., Moritz, A.B. & Schubert, W. 2009. Ductile lining design. Geomechanics and
tunnelling 2 (5). pp. 561577.
[5]
Radoncic, N. 2011. Tunnel design and prediction of system behaviour in weak ground. PhD Thesis,
Institute for Rock Mechanics and Tunnelling, Graz University of Technology.
[6]
Aydan, ., Akagi, T. & Kawamoto, T. 1993. The squeezing potential of rock around tunnels: theory
and prediction. Rock Mechanics and Rock Engineering 2. pp. 137163.
Authors
Dipl.-Ing. Dr.techn.
Nedim Radoni
Geoconsult ZT-GmbH
GTU Koralmtunnel
Hlzlstrae 5
5071 Wals bei Salzburg
Austria
[email protected]
O.Univ.-Prof. Dipl.-Ing. Dr.mont.

Wulf Schubert
Graz University of Technology
Institute for Rock Mechanics and Tunnelling
Rechbauerstrae 12
8010 Graz
Austria
[email protected]
71

Workshops 2013
th
Proceedings Workshop "Characterization of Fault Zones"

th
Publisher: Austrian Society for Geomechanics

5020 Salzburg
Austria
Editor: O.Univ.-Prof. Dipl.-Ing. Dr.mont. Wulf Schubert
Dipl.-Ing. Alexander Kluckner
Dipl.-Ing. Thomas Pilgerstorfer
Print: Graz University of Technology, print and copy center
2013 Copyright and all rights therein are retained by the authors and the publisher.
In-house publisher 2013
st
1 edition: 160 copies
72
Workshop "Characterization of Fault Zones"
EXTENDED ABSTRACTS
th
09 of October 2013
Salzburg Congress, Austria
sterreichische
Gesellschaft fr
Geomechanik
A-5020 Salzburg
T +43 (0)662 87 55 19
F +43 (0)662 88 67 48
E [email protected]
www.oegg.at

Characterization of Fault Zones

Uploaded by

Copyright:

Available Formats

Characterization of Fault Zones

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Characterization of Fault Zones

Uploaded by

Copyright:

Available Formats

Workshop "Characterization of Fault Zones"

sterreichische Gesellschaft fr Geomechanik