empiricism, theory, and problem solving in rock engineering

EMPIRICISM, THEORY, AND PROBLEM SOLVING IN ROCK ENGINEERING Nick Barton (www.nickbarton.com)

6th Leopold Mϋller Lecture, with additions

GSL – HONG KONG October 2013

Beaumont TBM Tunnel, 1880 : wedge-failure, stress-failure, tidal influence. Three photos separated by 150 m .

1

WHY THE ‘OVER-BREAK’/ POTENTIAL INSTABILITY? Because of adverse Jn, Jr, Ja (JRC, JCS, ϕr), Jw, SRF ……and dip/dip direction, gravity, density. 2

The origin and numerous applications of these parameters (Jn, Jr, Ja, JRC, JCS, ϕr, Jw, SRF, Q) will be (part of) the subject of this lecture. 3

ACTUAL EMPIRICAL BEHAVIOUR or ASSUMED BEHAVIOUR ?? Empiricism: a posteriori (= behaviour based on experience ) is better than a priori (= ‘behaviour’ (?) based on assumptions). There are too many a priori assumptions (clothed in some amazing algebra) which are used by many of us these days….e.g. GSI/Phase 2? 4

PART 1 5

A DISCONTINUOUS (and idealized) EXPERIMENTAL/EMPIRICAL START IN 1966 at Imperial (‘Empirical’) College, London

6

DST on 200 artificial tension fractures in a variety of brittle model materials (Barton, 1971)

7

NOTE LACK OF ACTUAL COHESION UNLESS STEPPED (“secondary”) FRACTURES ARE TESTED

8

(------ = no decimal places) τ = σn . tan [ 20. log( UCS/σn ) + 30º ]

9

2D JOINTED “ROCK-MASS” Tension-fracture models used for ‘rock slope’ studies (at Imperial College) 1968-1969. ‘Nuclear power plant’ rock cavern investigations (50m) (at NGI) 1977-1978 (pre-UDEC)

10

2D ‘rock mass’ research conducted in the laboratory Physical (1977) models (this colour) follow here which pre-date UDEC:  Artificial, but some useful lessons  They are physical, not conceptual  They are ‘a posteriori’, not ‘a priori’ ! 11

BIAXIAL LOADING Scale-effect investigations 250, 1000, or 4000 blocks.

“Always” got rotational failures with small blocks! 12

BIAXIAL LOADING TESTS WITH HIGHLY ANISOTROPIC STRESS APPLICATION (as under a big rock slide?)

Shearing by rotation of individual blocks, following local ‘kinking’ within the mass?

13

APPROPOS: LARGE DEBRIS and ROCK SLIDES

(Front cover: eds.

J.Clague, D. Stead)

14

FRANK SLIDE (Wikipedia)

15

TRAVEL DISTANCE VARIABILITY • SAY 0.5 to 1 km TRAVEL DISTANCE EXPECTED • WITH ‘AIR-CUSHION’ (Chinese research) 2km • REALITY (sometimes) is > 20 km • SLIDE MASSES TOO HOT FOR RESCUE PARTIES WHY?: Rotational friction, block crushing, extreme heating, ground water converted to steam, ‘steamcushioned slides’ due to ‘gas’ pressure? (V2/V1 = 1,400:1) 16

SUCCESSIVE HALVING OF THE BLOCK SIZE – HAS DRAMATIC ROTATIONAL (degree-of-freedom) EFFECTS, ALSO WITH UDEC-MC. (helps to explain the drama of fault zones: worse with clay and water) Shen, B. & Barton, N. 1997. The disturbed zone around tunnels in jointed rock masses.

17

ROCK JOINTS : THEY (ALSO) SHOWED NON-LINEAR SHEAR STRENGTH (and no cohesion!)

18

130 joint samples. Roughness measurement and tilt test ( Barton and Choubey, 1977)

19

TILT TEST ‘THEORY’

where α° = arctan (τ/σn) (Barton and Bandis, 1990) 20

130 rock-joint samples (Barton and Choubey 1977) Three curved peak shear strength envelopes and no cohesion! 1.Maximum strength with JRC = 16.9 2. Mean parameters JRC = 8.9 JCS = 92 MPa φr = 28°

3. Minimum strength with φr = 26º 21

22

Note: the original tension fracture-based equation (1971) was: τ = σn tan [ 20.log(UCS/σn )+30º]

JRC

JCS

φb (now φr)

TO THOSE WHO HAVE PERFORMED PH.D.’s AND ARE SELLING SOFTWARE – PLEASE

NOTE it is φr since 1977 ! 23

VISUAL MATCHING OF ROUGHNESS – for JRC… USEFUL BUT HAS LIMITATIONS (Barton and Choubey, 1977)

24

EXAMPLE of ROUGHNESS CONTRAST – BACK-CALCULATED FROM DST (L = 400 mm: Nevada Test Site welded tuff)

JRC = 1

JRC = 16

25

NOTE: above JRCJCS strength criterion was developed from tilt and push test correlation with DST (not from analysing roughness profiles!) 26

JCS > UCS (?)

JCS < UCS 27

SCALE EFFECTS FOR INDIVIDUAL JOINTS

28

Tilt tests repeated at different scale -

there is almost no damage. Note: JRC1 < JRC2 (Barton and Choubey, 1977) 29

Bandis 1980 Ph.D.

Ahead-of-their-time scale-effect investigations. One set of many joint replica tests.

30

The angular components of peak shear strength, with asperity strength (SA), and peak dilation angle (dn ) each included. (Barton, 1971)

31

Asperity component SA means that JRC (or φr) cannot be back-calculated by subtracting dilation (dn) from peak strength. Φr or Φb would then be dangerously too high (e.g. 40°) as in some earlier Hong Kong work. (JRC would also be incorrect).

32

SCALE-EFFECTS REDUCTION OF of JRC and JCS with block-size Ln > L 0

(Bandis, Dearman, Barton, 1981)

33

JRCmobilized defined

(also with dimensionless displacement) Barton, 1978, 1982

34

Then possible to predict/model shear stressdisplacement and dilationdisplacement behaviour. (Barton, 1982, with scaling from Bandis et al. 1981).

Note (double) scale effect on shear stiffness (Ks), because it is strongly scale-and-stress-dependent. (Ks usually < 1 MPa/mm, 0.1 MPa/mm if large blocks) 35

Well-jointed wedge. Remains in place because of the higher shear strength of the smaller component blocks ?

Larger blocks defining wedge (failure at much shallower angle of dip)

37

Before leaving shear strength envelopes: When a rock mass fails: 1st, 2nd, 3rd (and 5th) envelopes are mobilized at different strains - not like H-B / GSI estimation (Barton, 1976, 1999, 2006)

38

INTO THE FIELD !! CHARACTERIZATION OF JOINTING, DEFORMABILITY, AT MAJOR DAM SITES • IRAN: KARUN IV 230 m • IRAN: BAKHTIARY 325 m • CHINA: BAIHETAN 283 m (2 x 8,000 MW) 39

“You need to hire a rock-climbing-engineering-geology group to characterise the major joint planes that define the two major wedges that your company are worried about”

40

Some kilos lighter, and not telling his wife the reason, Iranian colleague M.Zargari is profiling major-joint MJ-67, Karun IV Dam, Iran

41

Schmidt rebound (R) on intact rock (> r on joint plane) Karun IV Dam site canyon, Iran

The a/L method for roughness is used when JRC is TOO LARGE

43

Joint Roughness Coefficient (JRC) JRC=0.5 JRC=1 JRC=2 JRC=3 JRC=4 JRC=5 JRC=6 JRC=8 JRC=10 JRC=12 JRC=16 JRC=20 F6a-ZA-B1 F6a-ZA-B2 F6a-ZA-B3 F6a-ZA-B4 F6a-ZA-B7 F6a-ZA-B8 F6a-ZA-B11 F6a-ZA-B13 JRC=22 JRC=24 JRC=26 JRC=28 JRC=30

1000.00

Amplitude of Asperities (mm)

100.00

10.00

For the very rough bedding plane, had to use “a/L” method

1.00

Mean JRCn = 11 0.10 0.1

1

10

(for 2m block size)

Length of Profile (m)

44

COLUMNAR JOINTING AT A MAJOR DAM SITE IN CHINA (Baihetan, 283m, 2 x 8000 MW)

45

BAIHETAN DAM and POWER GENERATION: 2 X 8,000 MW HydroChina/ECIDI

47

RECORDING OF ROUGHNESS FOR JRC ESTIMATION, 70 to 140 m into the canyon walls. (May need stripping to this competent-rock depth)

RECORDING OF SCHMIDTHAMMER REBOUND FOR JCS ESTIMATION 48

APPLICATION OF JRC and JCS to ROCK MASS DEFORMABILITY and to MODELLING with UDEC-BB (see columnar basalt behaviour)

49

Stress-closure and scale-effect shear tests. (Bandis et al. 1981 and 1983 )

The N, S components in rock mass load-deformation mechanisms. (Barton, 1986)

There is plate-load / block-test evidence for these three P-Δ type-curves.

50

UDEC-BB simulations (Chryssanthakis, NGI)

EMPHASISES WHY DISCONTINUUM ANALYSES GIVE MORE EXCITEMENT/INTEREST/ VALUE/REALISM than analyses without joints!

51

PART 2 52

Q-SYSTEM  SINCE 1974 Q HAS ACTUALLY BECOME “A SYSTEM”,

SINCE THERE ARE NOW SEVERAL COMPONENTS.  (50 rock types in first 210 cases, 1250 case records)  Q rockmass classification, Q-histograms  Q for ‘single shell’ NMT support (B + Sfr, RRS)  QC (= Q x UCS/100) for correlating with VP and EMASS  Q as part of QTBM for TBM prognosis  QSLOPE for selecting safe slope angles (in progress)  Qc split into CC and FC (if ‘continuum’ modelling)

‘53

EXAMPLE OF SLOPE ANGLE MATCHED TO GEOTECHNICAL PROPERTIES (or to local Q-slope = 0.1, 1.0, 10). (Panama expansion project, 2011. PCA photo)

Why/how was Q developed? Because of a question to NGI in 1973: “Why are Norwegian underground power houses showing such a variety of deformations”? (from Norwegian State Power Board/ STATKRAFT)

Question passed to NB. Answer given after 6 months of Q-system development! VARIABLES: Rock mass quality, support type/quantity, span/height, depth, stress. 212 case records used. B, S(mr), B+S(mr), CCA. 55

VARIABLE WORLD NEEDS BROAD-REACH CHARACTERIZATION METHOD 56

VARIABLE WORLD CANNOT ALWAYS BE COMPUTER MODELLED – BUT 57 IT CAN BE CHARACTERIZED

Strength contrast, modulus contrast, constructability contrast (15 years/1 year)! 0.001→1000, or 5→95, or F7→F1 ???

58

A GLIMSE OF NMT (SINGLE-SHELL TUNNELLING)

for which the Q-system was actually designed

59

Grimstad and Barton, 1993

(Norwegian conference), Barton and Grimstad, 1994

(Austrian conference).

The Q-system is most strongly associated with ‘single shell’ solutions : (B+Sfr + water control) (= NMT= Norwegian Method of Tunnelling) in mostly better rock, costs about 1/5 x ‘double-shell’ NATM, e.g. 20,000 US $/m compared to 60 100,000 US $/m (Costs from many countries).

RRS is a flexible (until bolted)

‘lattice’ girder.

3D effect because of S(fr) arches.

61

QUESTION: SHALLOW METRO or DEEP METRO? • MIXED-FACE OR ROCK? • 5 m PER WEEK OR 20 m PER WEEK?

• COST DIFFERENCE MAY BE 5:1 (at least)

62

RELATIVE COST FOR TUNNEL EXCAVATION AND SUPPORT (Barton, Roald, Buen, 2001) ………potential benefits of pre-grouting, especially if Q ≈ 0.1

Pre-injection screen 30-70 holes, 20-30m long, 0.5-1.0 m c/c (Hognestad and Frogner, 2005….. and Garshol (ICE/HK, 2010)

BÆRUM TUNNEL RAIL TUNNEL. Pre-injection in progress. Truck mounted

grout storage and mixing. Note up to 70 holes for dry 110m2 tunnel in shales and limestones. Packered injection ‘lances’ are chained for safety. 24-30 hours CYCLE time. Cost of finished NMT tunnel: < 25,000/m

Q-HISTOGRAM METHOD of logging

66

Q - VALUES: Q (typical min)= Q (typical max)= Q (mean value)= Q (most frequent)= B L O C K

(RQD 10 100 71 85

/ / / / /

Jn) 20.0 2.0 9.6 9.0

V. POOR

50

* * * * *

(Jr 1.0 2.0 1.5 1.5

POOR

/ / / / /

Ja) 5.0 1.0 2.8 2.0

* * * * *

(Jw 0.50 1.00 0.69 0.66

FAIR

/ SRF) = Q / 5.0 = 0.010 / 1.0 = 100.0 / 1.5 = 1.78 / 1.0 = 4.68

GOOD

EXC

40

RQD %

30

Core pieces >= 10 cm

20

10 00

10

S I Z E S

80

20

EARTH

30

FOUR

40

50

60

THREE

70

80

TWO

90

100

ONE

NONE

THE RESULT OF Q-HISTOGRAM LOGGING OF SIX CORES AT A PLANNED METRO PROJECT IN HONG KONG. DEVIATED HOLES.

Jn

60

Number of joint sets

40 20 00 20

T A N

(fr)

150

15

12

FILLS

9

6

4

3

PLANAR

2

1

UNDULATING

0,5 DISC.

Jr

100

Joint roughness - least

50

00

and T A N

(fp)

1

0,5

1

1,5

1,5

THICK FILLS

80

2

THIN FILLS

3

4

COATED UNFILLED HEA

60

Ja

40

Joint alteration - least

20 00 20

A C T I V E

13

12

8

6

5

12

EXC. INFLOWS

150

8

6

4

4

HIGH PRESSURE

3

2

WET

1

0,75

DRY

100

Jw

50

Joint water pressure

00

0.05

S T R E S S

10

200

0.1

SQUEEZE

0.2

SWELL

0.33 FAULTS

0.5

0.66

1

STRESS / STRENGTH

150

SRF

100

Stress reduction factor

50 00 20

15

10

5

20

15

10

5

10 7.5

5

2.5 400 200 100 50

20

10

5

2

0.5

1

2.5

67

How do the Q-parameter histograms change, as depth is increased in the same rock type?

CHARACTER OF SAPROLITE AND SOIL

LOGGED CHARACTER OF NEAR-SURFACE SANDSTONES

LOGGED CHARACTER OF DEEPER SANDSTONES

Q - VALUES: Q (typical min)= Q (typical max)= Q (mean value)= Q (most frequent)= B L O C K

(RQD 10 100 50 45

/ / / / /

Jn) 20.0 3.0 10.9 15.0

V. POOR

800

* * * * *

(Jr 1.0 3.0 1.6 1.5

POOR

/ / / / /

Ja) 12.0 1.0 3.5 1.0

* * * * *

(Jw 0.20 1.00 0.68 0.66

FAIR

/ / / / /

SRF) 20.0 0.5 5.4 2.0

GOOD

= Q = 0.000 = 200.0 = 0.26 = 1.49

EXC

RQD %

600

Core pieces >= 10 cm

400 200 00 10

S I Z E S

1000

EARTH

20

30

FOUR

40

50

60

THREE

70

80

TWO

90

100

ONE

NONE

800

Jn

600

Number of joint sets

400 200 00 20

T A N (fr)

800

15

12

FILLS

9

6

4

3

PLANAR

2

1

UNDULATING

0,5 DISC.

600

Jr

400

Joint roughness - least favourable

200 00

and T A N (fp)

1

0,5

1

1,5

1,5

THICK FILLS

1000

2

THIN FILLS

3

COATED

4

UNFILLED HEA

800

J & K rail-link, Kashmir. Here 12m/2 years.

Ja

600

Joint alteration - least favourable

400 200 00 20

A C T I V E

13

12

10

6

5

12

EXC. INFLOWS

2000

8

6

4

4

HIGH PRESSURE

3

2

WET

1

0,75

DRY

Jw

1500

Joint water pressure

1000 500 00 0.05

S T R E S S

8

800

0.1

SQUEEZE

0.2

SWELL

0.33

FAULTS

0.5

0.66

1

STRESS / STRENGTH

600

SRF

400

Stress reduction factor

200 00 20

15

10

5

20

15

10

5

10

7.5

5

2.5 400 200 100 50

20

10

5

2

0.5

1

2.5

72

Class 2: Q = 10 to 40 Histograma de RQD

Histograma de Jn

Normal

Normal

Media Desv .Est. N

1000

3000

88.69 12.16 6239

Media Desv .Est. N

2500

2.628 0.8636 6124

Frecuencia

Frecuencia

800 600 400

1500 1000

200 0

2000

500

0

10

20

30

40

50 RQD

60

70

80

90

0

100

Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 2, Q =10 - 40

0.5 1.0

2.0

3.0

4.0

Jn

9.0

Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 2, Q =10 - 40

Histograma de Ja

Histograma de Jr

Normal

Normal

3500

Media Desv .Est. N

3000

2500

1.705 0.4958 6239

2000 1500 1000

Media Desv .Est. N

2000

Frecuencia

2500

Frecuencia

6.0

1500 1000 500

500 0

0 0.5

1.0

1.5

2.0

3.0

4.0

Jr Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 2, Q =10 - 40

5 0 0.7 1.0

0 2. 0

0 3.0

0 4. 0

0 6. 0

Ja Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 2, Q =10 - 40

2.001 0.6967 6239

Class 3: Q = 4 to 10 Histograma de RQD

Histograma de Jn

Normal

Normal

Media Desv .Est. N

500

8000

75.80 16.90 10891

7000

3.930 1.605

N

10485

6000

Frecuencia

400

Frecuencia

Media Desv .Est.

300 200

5000 4000 3000 2000

100 0

1000 0 0

10

20

30

40

50 RQD

60

70

80

90

5 0 0. 1.

100

0 2.

0 3.

0 4.

0 6.

0 9.

.0 12

.0 15

Jn

Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 3, Q = 4 - 10

Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 3, Q = 4 - 10

Histograma de Ja

Histograma de Jr

Normal

Normal

Media Desv .Est. N

5000

2500

1.539 0.3551 10891

Media Desv .Est. N

2000

Frecuencia

Frecuencia

4000 3000 2000

1000 500

1000 0

1500

0 0.5

1.0

1.5

2.0 Jr

3.0

4.0

Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 3, Q = 4 - 10

75 00 0. 1.

00 2.

00 3.

00 4.

00 6.

00 8.

Ja Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 3, Q = 4 - 10

2.581 0.8259 10891

Class 4: Q = 1 to 4 Histograma de Jn

Histograma de RQD

Normal

Normal

Media Desv .Est. N

400

9000

56.09 19.30 17915

Media Desv .Est. N

8000

6.543 2.614 16187

7000

Frecuencia

Frecuencia

300

200

6000 5000 4000 3000 2000

100

1000

0

0

0

10

20

30

40

50 RQD

60

70

80

90

5 0 0 0 0 0. 1. 2. 3. 4.

100

0 6.

0 9.

.0 12

.0 15

.0 20

Jn Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 4, Q = 1 - 4

Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 4, Q = 1 - 4

Histograma de Ja

Histograma de Jr

Normal

Normal

7000

Media Desv .Est. N

6000

1.486 0.3047 17915

Media Desv .Est. N

2500 2000

Frecuencia

Frecuencia

5000 4000 3000

1500 1000

2000

500 1000 0

0 0.5

1.0

1.5

2.0 Jr

3.0

4.0

Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 4, Q = 1 - 4

75 00 0. 1.

00 2.

00 3.

00 4.

00 6.

00 8.

Ja Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 4, Q = 1 - 4

3.223 1.189 17915

Class 5: Q = 0.1 to 1 Histograma de Jn

Histograma de RQD

Normal

Normal

Media Desv .Est. N

900 800

28.27 15.61 11515

700

9.583 3.200 10380

3000

600

Frecuencia

Frecuencia

Media Desv .Est. N

4000

500 400

2000

300 1000

200 100 0

0

0

10

20

30

40

50 RQD

60

70

80

90

5 0 0 0 0. 1. 2. 3.

100

0 4.

0 6.

0 9.

15

.0

.0 20

Jn Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 5, Q = 0.1 - 1

Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 5, Q = 0.1 - 1

Histograma de Ja

Histograma de Jr

Normal

Normal

3000

Media Desv .Est. N

2500

900

1.374 0.2628 11515

800

Media Desv .Est.

4.628 2.019

N

11515

700

2000

Frecuencia

Frecuencia

.0 12

1500 1000

600 500 400 300 200

500 0

100 0

0.5

1.0

1.5

2.0 Jr

3.0

4.0

Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 5, Q = 0.1 - 1

5 0 0.71.0

2.0

0

0 3.0

0 4.0

0 6. 0

0 8.0

0 .0 10

.0 12

0

Ja Proyecto: Estadística variables de Barton.MPJ; Hoja de trabajo: Class 5, Q = 0.1 - 1

VELOCITY-MODULUSPERMEABILITY-Q-VALUE CHALLENGES, AT BAKHTIARY DAM SITE, IRAN (325 m) 77

How to characterize voids? Velocity-moduluspermeability-Q-value correlation difficulties. 78

Upper diversion tunnel: top heading

79

Q - VALUES: Q (typical min)= Q (typical max)= Q (mean value)= Q (most frequent)= B L O C K

(RQD 10 100 73 80

/ / / / /

Jn) 15.0 2.0 6.0 4.0

V. POOR

100

* * * * *

(Jr 0.5 4.0 2.0 2.0

POOR

/ / / / /

Ja) 6.0 0.8 1.6 1.0

* * * * *

(Jw 0.66 1.00 0.99 1.00

FAIR

/ SRF) = Q / 5.0 = 0.007 / 1.0 = 266.7 / 1.1 = 13.74 / 1.0 = 40.00

GOOD

EXC

80

RQD %

60

Core pieces >= 10 cm

40

20 00

10

S I Z E S

120 100 80 60 40 20 00

20

EARTH

(fr)

150

40

50

60

THREE

70

80

TWO

90

100

ONE

NONE

Jn Number of joint sets

20

T A N

30

FOUR

In diversion tunnel Qm.f. = 40 Qmean = 14

15

12

FILLS

9

6

4

3

PLANAR

2

1

UNDULATING

0,5

Next steps:

DISC.

Jr

100

Joint roughness - least

50 00

and T A N

(fp)

1

0,5

1

1,5

1,5

THICK FILLS

200

2

THIN FILLS

3

COATED

4

1. Convert Q to Qc

UNFILLED HEA

150

Ja

100

Joint alteration - least

50

(UCS?)

00

20

A C T I V E

13

400

• 1.

12

8

6

5

12

8

6

4

4

HIGH PRESSURE

3

2

WET

1

2. Convert to Vp 3. Convert to Emass

0,75

DRY

300

Jw

200

Joint water pressure

100 00

0.05

S T R E S S

10

EXC. INFLOWS

300

0.1

SQUEEZE

0.2

SWELL

0.33

FAULTS

0.5

0.66

1

STRESS / STRENGTH

SRF

200

Stress reduction factor

100 00 20

15

10

5

20

15

10

5

10 7.5

5

2.5 400 200 100 50

20

10

5

2

Rev.

BAKHTIARY DAM HEPP UPPER DIVERSION TUN

0.5

1

2.5

Report No.

NB&A #3 Borehole No. :

Drawn by

Figure No.

6 Date

80

81

TBM prognosis and comparing with drill-and-blast

QTBM 82

145 cases, ≈ 1000 km, open-gripper trends (Barton, 2000).

83

SEVERAL PAGES OF WORLD RECORDS BY TBM – ASSEMBLED BY ROBBINS, GIVE THE FOLLOWING RESULTS WHEN COMBINED Assume 24 hrs/day, 168 hrs/week, 720 hrs/month

8 4

85

86

WHY TBM DELAYS IN FAULT ZONES ? “Theo-empirical” reasons Lack of belief gets paid for!

87

‘THEO – EMPIRICAL’ REASONS WHY FAULT ZONES ARE SO DIFFICULT FOR TBM – THREE BASIC EQUATIONS: 1. AR = PR x U (all TBM must follow this) 2. U = Tm (due to the decelerating advance rate with time) 3. T = L / AR (obviously time for length L must be proportional to 1/AR)

4. T = (L / PR) (1 / (1+m) (from #1, #2 and #3) 5. This is VERY important for TBM……since (-)m is strongly related to Q-values …..in FAULT ZONES. 6. It is important because very negative (-)m values make 1/(1+m) TOO BIG – giving ‘huge delays’ (T in months or years) 88

BUT…Q CAN BE IMPROVED BY PRE-GROUTING ! (IMPROVE –m.....to less negative value)

89

RAIL TUNNEL PROGNOSIS – OSLO-SKI / FOLLO-BANEN

90

91

CENTRAL Q-VALUES AND QTBM VALUES ARE BEST FOR GOOD TBM PROGRESS. TAIL-DISTRIBUTIONS ’BETTER’ WITH D+B ! Note records for drill-andblast: 176m/one face in 168 hours (7x24) week. Whole project 104 m/week average – IN COALMEASURE ROCKS needing B+S(fr). LNS Norway

PHYSICAL (2D) MODELS of ROCK CAVERNS, AS FORE-RUNNER TO UDEC-BB FLEXIBILITY (1977-1978)

93

94

THE ‘jointed’ NATURE OF THE MODELLING MATERIAL (Post- ‘seismic’ loading result, following 0.2 to 0.5 g)

95

Physical and FEM modelling (Barton and Hansteen, 1979) suggested possible ‘heave’ resulting from largecavern construction near the surface…….. ……….depended on joint pattern and horizontal stress level in the physical models.

96

GJØVIK CAVERN INCREASE OF LARGEST CAVERN SPAN BY ALMOST 2 x (Note also the three deep caverns in a Norwegian road tunnel) 97

Gjøvik Olympic cavern represented a big jump…….in span and confidence! (Figure from Sharp, 1996: UK Nirex study)

BLUE: Lærdal Tunnel (three lorry-turning and ‘wake-up-driver’ caverns in 24.5 km long tunnel)

LÆRDAL TUNNEL lorry-turning caverns (three of them) 30 m span, depths 1,000 to 1,400 m (Photo G.Lotsberg)

99

Gjøvik cavern : an ’extension’ of 1974 Q-system data base. (Qmin, Qmean, and Qmax values of 1, 12, 30 logged in the cavern) RQD = 60-90%, UCS = 90 MPa was typical.

100

101

GJØVIK CAVERN JOINT-GEOMETRY ASSUMPTIONS input data, boundary stresses Barton, N., By, T.L., Chryssanthakis, P., Tunbridge, L., Kristiansen, J., Løset, F., Bhasin, R.K., Westerdahl, H. & Vik, G. 1994. Predicted and measured performance of the 62m span Norwegian Olympic Ice Hockey Cavern at Gjøvik. Int. J. Rock Mech, Min. Sci. & Geomech. Abstr. 31:6: 617-641. Pergamon.

102

TOP HEADING TOO WIDE TO OBSERVE FROM ONE LOCATION

103

104

The final modelled 7 to 9 mm (downwards directed) deformations matched the unknown (to be measured) result almost perfectly. (UDEC-BB modelling by Chryssanthakis, NGI)

105

DEFORMATION RECORDS FROM MPBX AND LEVELLING Δ = 7 to 8 mm was typical. Construction period: week 24 to week 52, following arrival of access tunnels (top and bottom). BxHxL = 62 x 24 x 90 = 140,000 m3

Typical NMT B + S(fr) DRAINED

106

CONTINUUM (??) or DISCONTINUUM MODELLING 107

Borehole stability studies at NGI Continuum becomes a discontinuum!

Drilling into σ1 > σ2 >σ3 loaded cubes 0.5 x 0.5 x 0.5 m

of model sandstone

108

Jinping II (D+B) – ISRM News Journal Physical model – bored under stress (NGI) Jinping II (TBM) – ISRM workshop (NB)

Log-spiral shear modes in weaker rock types 109

8

-8

-6

-4

-2

X Axis (m) 0

2

4

6

8

-8

-7

-6

-5

-4

-3

-2

-1

X Axis (m) 0

1

2

3

4

5

6

3

4

5

6

Elastic fracture

10 10

7

Open fracture Slipping fracture Fracture with Water

6

8

Fracom Ltd

5

6

Date: 22/12/2004 11:00:09

4

3

4 2

2 Y Axis (m)

Y Axis (m)

0

1

0

-1

-2 -2

-4

-3

-4

-6 -5

-8

-6

-7

-8

-6

-4

-10 10

Three FRACOD models showing fracturing development. Baotang Shen, 2004 -2

0 X Axis (m)

2

4

6

8

-8

-8

-7

-6

-5

-4

-3

-2

-1

0 X Axis (m)

1

2

11 0

Cundall and Cundall………but the choice is clear! (NGI modelling by Lise Backer)

111

NEED for CHANGE CONVENTIONAL continuum modelling methods are suspect. Poor simulation with Mohr Coulomb or Hoek and Brown strength criteria. ( Hajiabdolmajid, Martin and Kaiser, 2000 “Modelling brittle failure”, NARMS.)

So why performed by so many consultants? 112

Degrade cohesion, mobilize friction: excellent match. ( Hajiabdolmajid, Martin and Kaiser, 2000 “Modelling brittle failure”, NARMS.)

11 3

NOW AN ALTERNATIVE WAY TO ESTIMATE ‘c’ and ‘φ’ FOR ROCK MASSES (but still need to degrade c at small strain, and mobilize φ at larger strain) 114

CC and FC from Qc = Q x σc /100 : Qc = RQD/Jn x Jr/Ja x Jw /SRF x σc /100) CC = cohesive strength ( the component of the rock mass requiring shotcrete)

FC = frictional strength ( the component of the rock mass requiring bolting). Cut Qc into two halves →’c’ and ‘φ’

c RQD 1 CC    Jn SRF 100  Jr  FC  tan   Jw   Ja  1

115

GSI-based algebra for ‘c’ and ‘φ’ contrasted with

Q-based ‘empiricism’ Note: shotcrete needed when low CC, bolting needed when low FC.

Table of Q-parameters with declining quality (resembling weathering) (Barton, 2002). RQD Jn 100 90 60 30 10

Jr

Ja

Jw

SRF

2 2 9 1 12 1.5 15 1 20 1

1 1 2 4 6

1 1 0.66 0.66 0.5

1 1 1 2.5 5

Q

c

Qc

FC° CC MPa Vp km/s Emass GPa

100 100 100 63° 10 100 10 45° 2.5 50 1.2 26° 0.13 33 0.04 9° 0.008 10 0.0008 5°

50 10 2.5 0.26 0.01

5.5 4.5 3.6 2.1 0.4

46 22 10.7 3.5 0.9

Four rock masses with successively reducing character: more joints, more weathering, lower UCS, more clay. Low CC –shotcrete preferred

Low FC – bolting preferred 45

117

HIGH CC

HIGH FC

Q ≈ 100/6 x 4/0.75 x 1/5

(SRF = 5 due to near-surface)

MINING AND CONSTRUCTION MAGAZINE Picture of the mountain-side rock sculpture of the Indian chief ‘Crazy Horse’, in North Dakota, USA.

CC ≈ 100/6 x 1/5 x 150/100 ≈ 5 MPa FC ≈ tan-1 (4/0.75 x 1/1) ≈ 79º 118

FLAC 3D ‘c + σn tan φ’ (left) ‘c then σn tan φ’ (below) (Barton and Suneet Pandey, 2011)

‘New’ approaches:  c then tan φ (not new, but rare!)  Comparing modelled and measured displacements with pre-installed MPBX.  Back-calculating Q from empirical Δ equations, as well as logged Q.

119

‘C then σn tanφ’ (as used in Barton and Pandey, 2011)

120

121

NMT tunnel (= single-shell) through pre-injected (10 MPa pressure) shales, limestones.

122