Brittle Failure Kaiser

Practical Implications of Brittle Failure in Tunnelling - Kaiser 2010

Peter K. KaiserPresident/CEO Centre for Excellence in Mining Innovation

Chair for Rock Mechanics and Ground Control

Laurentian University

Practical Implication of Brittle Failure on

Hard Rock Tunnelling Construction

Tunnels in Granite - Tneles en granito

Universitat Politchnica de Catalunya

Barcelona - Spain


Acknowledgements

Collaborators: Cai, Diederichs, Hajibdolmajid, Martin, McCreath,

Contractors: MATRANS, TAT, Herrenknecht AG, ...

Mining companies: Vale INCO, Goldcorp, Rio Tinto,

Science Council: NSERC

and many more

Practical Implication of Brittle Failure on

Hard Rock Tunnelling Construction


Toronto

Experiences

from major mining and

tunnelling operations


Mining at depthLessons learned

under stress rock is less forgiving

must learn from costly mistakes

and learn to design smart !


Objective

Hoek/Kaiser/Bawden 1995

Review lessons learned

Interpret observed rock failure processes

Explain factors affecting constructability

to identify opportunities for improvements

support design

rock excavation techniques

ground control measures

to reduce construction problems minimize gap between designer and

contractor


Primary rock mechanics challengewhen tunnelling in massive to moderately jointed rock

Anticipating the actual rock behaviour

Brittle or spalling failure

spalling often dominates over shear failure

Geo-engineering for constructability

fractured rock is often difficult to control


Geological Model !

Rock Mass Model !

Rock Behaviour Model ?

Site characterization

?


Focus on

massive to

moderately

jointed

rock

Modes of tunnel instability


1st Challenge - anticipate failure mode Observe

Interpret Understand

Funka-Bedretto - CH

Trondheim - No

Ltschberg - CH

El Teniente - Chile

URL-CanadaPiora - CH

Mt.Terri - CH

Spalling behaviour must be anticipated in almost rocks !


2nd Challenge - anticipate the extent Observe

Interpret

Stress field

Depth of Failure

Understand

Extrapolate

Stress Level ( ) (or Depth/UCS)


Spalling

leads to

near-excavation

strength

reduction

Appropriate failure criteria to model near-wall behaviour

(Kaiser et al. 2000)


From field observations: micro-

seismicity to visual observations

Field rock mass strength

X


Intact rock too! Revisited data courtesy Hoek (1961)

0.0

1.0

2.0

3.0

4.0

5.0

6.0

0.0 0.1 0.2 0.3 0.4 0.5

s3 /ucs

s1/u

cs

Failure criteria to model

intact rock is actually s-shaped

with full or apparent cohesion mobilization only at

high confinement

ductile

brittle

UCS(I)/10

Apparent UCS(II)

UCS(I)


5,15,25,35,45,50,55,65,69,70

Influence of heterogeneity on propagating path of wing crack in an unconfined sample (simulated with RFPA2D)

WHY? - Griffith crack simulation


Why?Heterogeneity causes tensile stresses

s3 = 2.5 MPa

400

300

200

100

0

-10 0 10 20 30 40

PFC Samples:

Local tension due to

heterogeneity

Yield

Initiation

s1

s3

Courtesy Diederichs 2000


Internal tension causes spalling

Crack

propagation

length

Propagation Spalling

Griffith, Hoek, and many others


Stress ratio= 10 to 20

% tensile stresses spalling limit

400

300

200

100

00 10 20 30 40 50

Yield

Initiation

Area in Tension: 10% 1% 0.1%

s1

s3

Courtesy Diederichs 2000


Spalling

leads to

near-excavation

strength

reduction

Appropriate failure criteria to model near-wall behaviour

(Kaiser et al. 2000)


s3 = 12MPa

Where? near excavations in low s3 range

Ko = 0.75 Ko = 1 Ko = 1.33


Quartzite as an analogue of a rock mass

20

0

100

200

300

400

500

600

700

800

0 10 20 30 40 50 60

s1

[MP

a]

s3 [MPa]

a

~a/2


CoV =

15% to 45%

Probability of yield (100 0% failed elements) and deviatoric stress (s1-s3) contours

x = shear

o = tensionKaiser 2010 Eurock


Spalling limit or s1/s3 - ratio

0

1

2

3

4

5

6

0

5

10

15

0 5 10 15 20

No

rmal

ized

cra

ck le

ngt

h

Stre

ss r

atio

1/

3

Distance along edge of yield zone [m]

Mean stress ratio+1sd-1sdNormalized crack length


Yield actually means deep spalling

Spalling not

shear yield


Stress issues even at shallow depth

Summary from detailed measurements

Martin 1999


under stress - same geology behaves differently -

When getting rock behaviour wrong numerical models and design are likely wrong and construction is often difficulties

Tender documents, tend to

emphasise description of geology, rock and rock mass, and

underemphasise description of the anticipated rock behaviour and spalling is not anticipated (?)


Implications for tunnels

Implications of behaviour misinterpretation illustrated on case example

Stand-up time issues delays $$


Implications of gettingrock mass rock mass behaviour wrong

For example

reduction in advance rate [m/d]

Planned

Actual

Raveling rock behind the open TBM split the advance and support cycle


Stress-driven rock mass degradationoften dominates

(b) (c)


Stand-up time reduction

3-9

mos

6-24

hrs

3-9

mos

6-24

hrs

Bieniawski 1987

Construction tools

GSI


Progressive failure process produces blocky, unravelling ground with rock mass bulking

Volumeincrease inside df

blocky ground = unravelling rockmass broken by stress


Anticipating behaviour at depth

Fracture propagation from stress raisers at corners (e.g. incl. tunnel face)

Massive rock unravels

Instabilities of highly stresses tunnel face


Anticipate TBM face behaviour

now can also anticipated spalling at tunnel face

Observe

Interpret

Understand

Extrapolate to depth

Face behaviour

Increasing stress or depth


Spalling at tunnel face Predictions and

observations


Unravelling tunnel face What is seen in roof is to

be expected at face !

Unravelling of face before wall ! massive

broken


Rock support of brittle failing ground with rock mass bulking

Volume

increase

inside df

blocky ground = unravelling rockmass broken by stress


Challenge anticipate rock mass bulking (geometric volume increase)al. 1996) showing >30% for unconfined floors and 1 to 10% depending on support type (Error!

Reference source not found..b).

0

20

40

60

80

100

120

0 0.5 1 1.5 2

Vertical displacement (mm)

Ver

tical

stre

ss (M

Pa)

-10123456789101112131415

Dila

tion

(%)

= 40s3 = 1 MPa

0

1

2

3

4

5

6

7

8

9

10

0.01 0.1 1 10

Confinement [MPa]

Dila

tion

or B

ulkin

g Fa

ctor

BF

[%]

ELFEN model

Light support

Yielding support

Strong support

(a) (b) Field measurements

Simulation

with ELFEN

al. 1996) showing >30% for unconfined floors and 1 to 10% depending on support type (Error!

Reference source not found..b).

0

20

40

60

80

100

120

0 0.5 1 1.5 2

Vertical displacement (mm)

Ver

tical

stre

ss (M

Pa)

-10123456789101112131415

Dila

tion

(%)

= 40s3 = 1 MPa

0

1

2

3

4

5

6

7

8

9

10

0.01 0.1 1 10

Confinement [MPa]

Dila

tion

or B

ulkin

g Fa

ctor

BF

[%]

ELFEN model

Light support

Yielding support

Strong support

(a) (b)

Courtesy Cai 2006

BF = 0 to 10%

Uni-directional


Yield and s3 pattern near tunnel

Circular tunnel at

2000m

and

Ko=0.51.5m

a = 5m

R = 6.5m

R/a = 1.3

Yield

Bulking


0

2

4

6

8

10

12

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Distance from Springline [m]

Dilati

on

or

BF

[%

]

Plastic strain, far from faceCombined

Radial strain (%) control

with dense bolting and retention (shotcrete)

s3


Lessons learned ...When dealing with brittle failure in tunnelling ...

observe, interpret, and understand

adjust design and construction procedures to match ground behaviour

Stressed ground is less forgiving

stress breaks even good ground

good ground becomes poor ground

massive, brittle rock disintegrates cohesionless ground

Quantum shift in constructability


Lessons learned ...When highly stressed brittle rock fails

by spalling .. not shear

degradation cannot be prevented

conventional failure criteria mislead designers s-shaped

spalling process affects both tunnel walls roof and face

select excavation and support techniques appropriate for broken rock

No ravelling, raining rock and flying arches!


Acknowledgements

Collaborators: Cai, Diederichs, Hajibdolmajid, Martin, McCreath,

Contractors: MATRANS, TAT, Herrenknecht AG, ...

Mining companies: Vale INCO, Goldcorp, Rio Tinto,

Science Council: NSERC

Practical Implication of Brittle Failure on Hard

Rock Tunnelling Construction

www.miningexcellence.ca

Documents

Brittle Failure Kaiser