TUNNEL SUPPORT SELECTION

FROM Q-CLASSIFICATION and

SUPPORT ELEMENT PROPERTIES

NB #3
 ORIGINAL DATA BASE
 UNSUPPORTED TUNNEL CHARACTERISTICS
 NMT CONCEPTS OF TUNNEL SUPPORT
 UPDATING Q-SUPPORT WITH S(fr) SHOTCRETE
 STRENGTH-DEFORMATION CHARACTER OF S(fr)
 STRENGTH-DEFORMATION CHARACTER OF BOLTING
 WATER CONTROL WITH FREE-STANDING LINERS
 RRS FOR TUNNELLING IN BAD GROUND
 ‘NOMINALLY UNLINED’ HEADRACE TUNNELS
 NMT (SINGLE-SHELL) COST ESTIMATION
1
212 case records, mostly B+S(mr) (from Barton, Lien and Lunde, 1974)
2
SUPPORT PRESSURE GUIDELINES (FROM 1974) - used for checking bolt and
anchor capacity needs 3
APPROXIMATE GUIDE-LINES FOR TEMPORARY SUPPORT
AND FOR (PERMANENT) WALL SUPPORT

4
5
6
UPDATED SUPPORT CHART PRINCIPLES, FOR 1989 TO 1993 INCORPORATION
OF NEW B+ S(fr) CASE RECORDS. Grimstad, 1989
7
8
In 1992 (Barton et al. 1992) coined the phrase NMT –
the Norwegian Method of Tunnelling – to differentiate
it from NATM, which was soon to get so bad publicity
from high-profile collapses around the world
(Heathrow, Munich, São Paulo).

There was also widely published criticisms of NATM

from a prominent Swiss engineer (Prof. Kovari)
concerning the questionable scientific/engineering
basis for NATM.

9
10
11
12
Some details concerning NMT. Tunnels are dry, drained, and PC-element
cladded if required for road or rail use . (‘Pigging’ = scaling)

13
Some of the key differences between NMT and (the
sometimes misused term) NATM that were more obvious
10 years ago than today, following technology
improvements. More NATM tunnel operations use S(fr)
today than 10 years ago.

14
After Grimstad and Barton, 1993 (Norway) and Barton and Grimstad, 1994 (Austria).
15
UPDATED ESR RATINGS
(TUNNEL-USE-and-SAFETY rating for adjusting EQUIVALENT SPAN)
(Barton and Grimstad, 1994) 16
Grimstad’s new case records 17
18
19
20
21
22
Cycle-time recordings (drilling the next round, blasting, waiting for gasses
to clear, mucking, logging, bolting-if needed, shotcreting-if needed) as a
function of Q-value. (Grimstad, NGI: pers com.1998).
23
SOME IMPROVED TECHNOLOGY ASPECTS OF NMT

1. relevant shotcrete technology and

equipment (no steel arches!)

2. relevant bolting technology (corrosion

protected)

3. relevant water control (pre-injection and the

free-standing liner)

24
B+S (better still B+Sfr) gives by far the best tunnel-stabilizing result
according to 5 years of deformation monitoring at an experimental
tunnel in mudstone.

25
Ward et al. 1983 (left), and Barton 1994. (right).
 THE POTENTIAL (ACTUAL)
INEFFECTIVENESS OF
STEEL SETS, UNTIL (IF?)
THEY ARE PUT IN
CONTACT WITH THE
TUNNEL ARCH.

ONLY THE SPILING

BOLTS DRILLED
INTO THE FACE ARE
GOING TO BE
EFFECTIVE, WHEN
THE NEXT ADVANCE
IS MADE.
26
Relevant shotcrete technology and equipment

Road-licensed, diesel/electric, high-output robot trucks, which can

serve several tunnel faces.

Each are capable of 20 to 25 m3/hour on-the-tunnel-wall shotcreting

with S(fr).

Rebound 4 to 6 % (contractor guarantee) with right concrete design.

Air/water jet cleaning > 15 minutes before each 1 hour of S(fr)!

27
28
 Typical S(fr) mix design for C45 to C55 (MPa) shotcrete.
Note operator location close to nozzle, where rebounds of 4 to 6%
(and almost dust-free air) make quality control very easy.

29
 Large-scale testing of S(fr) by Robocon in the mid-eighties. Fracture
energy (area under load-deformation curves) was 60 to 80 times that of
unreinforced shotcrete, depending on fibre dosage 40 or 60 kg/m3.

30
31
32
SLAB TEST LOAD-DEFORMATION BEHAVIOUR WITH S(fr)
Torsteinsen and Kompen, 1983.

33
Conceptual comparison of S(mr) and S(fr)

M. Vandevall (Bekært, Belgium)

34
CONTRACTOR ‘PAINTING’ TO AVOID PROBLEMS (??)

35
ONE OF THE FIVE MAIN PROBLEMS, same tunnel.

36
The advantages of S(fr) compared to S(mr). There is today the additional
advantage of alkali-free accelerator, allowing thick layers of S(fr) to be
built up rapidly, without the previous loss of long-term strength when 37
using ‘too much’ accelerator.
LOAD-DERORMATION COMPARISON OF S (mr) and S (fr)

38
POLYPROPOLENE FIBRE….gaining ground!

39
40
Comparison of load (= kraft) versus deformation behaviour for various
Dramix-steel, and Barchip-polypropylene fiber-reinforced shotcretes.
NFF lecture, Grimstad, 2010.
 Note energy absorption classes E1000J, E700J (and remainder E500J). In
boxes: dimensions of RRS (rib reinforced shotcrete arches) with thickness
(cm), number of reinforcing bars, and c/c spacing of the ribs.
From Grimstad et al., 2002.
(See Fig. 3.16 for original correct bolt lengths: small errors in present figure).

42
(From Grimstad, 2010. NFF lecture.)
43
EXAMPLES OF POOR PRACTICE USING
THIN MESH, HAND-HELD EQUIPMENT

(one from NATM advertisement !)

44
POOR PRACTICE….INCOMPLETE SHOTCRETING
MAKE THE MESH APPEAR STRONG !
(Shadow effect, and later corrosion likely)

45
Wrong distance, poor-practice shotcreting.
A strange choice for an NATM advertisement of….S(mr) ?
AMV (Andersons Mekaniske Verksted – Andersons
Mechanical Workshop) road-licensed, deisel/electric
shotcrete robot: 20-25 m3/hr shotcreting capacity.
 Relevant bolting technology (corrosion protected)

• Because NMT uses S(fr)+B as the final support of tunnels and caverns
(Barton et al. 1992, Barton and Grimstad, 1994), it is important that also
the bolts are of good quality, with suitable long-life corrosion
protection.

• The widespread use of NMT principles in Norway for the last 25 years
(35 years if S(mr) is included) has meant excellent development of
corrosion protected bolts in this country.

• The CT-bolt, manufactured by Ørsta Stål, incorporates a simple end-

anchoring (wedge-lock) for easy installation and tensioning, followed
by double-annulus grouting using a PVC-sleeve.

• With the layers: galvanising, Combi-coat (epoxy paint), grout, PVC-

sleeve, grout : it has five layers of initial corrosion protection, and four
are left if/when the outer layer of grout is cracked due to joint
deformation. (This is the usual start of corrosion for conventional 49
bolts).
 The CT-bolt with PVC sleeve (many meters length in practice).
Maximum load capacities are 33 and 30 tons in tension and shear,
50
respectively, for the 20mm diameter bolt (22mm with thread).
 The double-annulus grouting (shown in red). Shotcreting can be
performed after end-anchoring and before grouting of the bolt, if desired,
using a tube extension. See www.CTbolt.com for good 3D animation of
51
process.
Grouting of the CT
bolt can be
performed after
shotcreting, if
desired.
An over-cored CT bolt showing crack (joint) penetration to outer
layer of grout – the usual commencement of corrosion
for a conventional bolt. There remain four layers of corrosion
protection even with the joint/crack. 53
54
55
 Some (slightly exaggerated) potential problems with un-sleeved
conventional rock bolts. (Blue is lost grout due to water flow, black is
void, DD is good, BB less so). 56
Direct-shear tests of bolted
joints. BJURSTRÖM 1974

57
An unstable wedge
in the cavern wall.

The force diagram

demonstrates ‘closure’
using the minimum bolting
force B when inclined at
the approximate
available/mobilized
friction angle…….

in relation to
the critically inclined
discontinuity.
The optimal bolt
installation angle
(and resulting
capacity) will vary
with the shear
displacement that
has already
occurred.

The critical
discontinuity/joint
might be at post-
peak strength.

59
 Relevant water control

• hydrostatic liner and membrane

• free-standing liner
• pre-injection

There are several solutions to the water problem,

and the different solutions tend to have
widely different prices.

60
An example of one of the most expensive tunnelling solutions, like
conventional NATM, with B+S(mr) for primary support, and final cast concrete
liner with membrane and drainage fleece, for secondary support. This high-
speed rail tunnel through jointed chalk in Southern England, had final (year
2000) costs of US$ 128M /3.2 km, or $ 40,000 per metre. This was three to
four times higher than a typical NMT tunnel, with similar Q-value rock, using
B+S(fr) as permanent rock support, and a PC-element+membrane liner, for a
61
drained-but-dry solution.
NMT concepts in diagram form
Note: stage No. 6 must precede stage No.
2 if stability /stand-up time is very poor.

Note the PC-element free-standing but

bolted liner (= # 8).

This has an outer continuous membrane

lying over it, if required due to continued
water inflow or drips – e.g. if high
pressure pre-grouting had not completely
controlled the water.

#7 Permanent B+S(fr) in single shell

This is the accepted, and economic

single-shell support and reinforcement.

#8 PC-element free-standing liner

This is completed at rates up to 1000m

per month, with suitable mounting 62
machinery.
An example of a PC-element final liner,
placed after cleaning of muck and fill in
the invert. Membrane and frost
insulation (sandwich) shown.
An example of PC-element mounting
for a two-lane road tunnel. Note the
primary B+S(fr) permanent support,
and the mostly dry surface of the
shotcrete,. 63
An example of the primary rock
reinforcement and support.

This consists of B+S(fr) in the

arch and upper walls of a heavily
jointed (and therefore heavily
over-breaking) rock mass.

Photographed while checking the

shotcrete for any signs of
‘druminess’.

This was followed by a free-

standing liner with outer
membrane for this two-lanes
motorway tunnel.

64
 9 tons
ARCH
elements
being lifted
up ready for
fixing
(bolting) in
220 km/hr
double-track
rail tunnel.

NMT (B+Sfr)
then PC-
element liner
with outer
membrane
PC-elements stacked at rail tunnel portal ready for mounting with outer
membrane sheet – at rates of 900m/month (after learning curve)

66
 THE PRE-INJECTION OPTION
FOR WHEN INFLOW AND
GROUNDWATER DRAWDOWN
ARE UNACCEPTABLE

(see LECTURE on PRE-GROUTING)

67
RRS (rib-reinforced shotcrete arches)

FOR VERY BAD ROCK CONDITIONS

(e.g. Q < 0.1)

68
 RRS philosophy
• When Q-values are below approx. 0.1 (i.e. extremely poor),
it can be expected that there will be the possibility of large
over-break, low stand-up time, and significant early
deformations.

• The use of steel sets should be avoided in such situations,

due to the actual relatively larger rock-block loosening that
they allow, unless followed immediately by bolting or
shotcrete, or both.

• It is for this category of problems that RRS (or rib-

reinforced shotcrete) has been developed.

• This is a much more effective measure than steel arches or

lattice girders when conditions are very bad, because it
provides a more rapid and much stiffer support than these
two ‘solutions’. 69
70
RRS or steel-reinforcing-bar reinforced shotcrete arches, for the next-to-worst
categories of rock mass, e.g. 0.01 < Q < 0.1. 1= first layer of general S(fr) –
accelerated with non-alkali additive, 2 = build-up local, smooth but not necessarily
circular arch (or arches) of non-alkali accelerated S(fr), 3 = drill bolt holes at
e.g.1m centres round arch, and install end-anchored bolts with pre-fabricated,
welded cross-bars. 4 = attach (wire and weld) 4-6x16mm reinforcing bar ‘steel-
arches’ to each bolt-head cross-bar (pre-fabricate in bundles, for easier
attachment. (Note: these bars can be bent into overbreak zone, therefore requiring
less shotcrete volumes than with e.g. stiff lattice girder), 5 = spray over reinforcing
bars with shotcrete, to complete arch and provide foundation for: 6 = bolts and
washer, tensioned (bolt thread pre-protected with plastic caps. 7 = Spray over bolt
heads to complete RRS arch.
71
 RRS
is a
flexible
(until
bolted)
‘lattice’
girder.

3D
effect
becaus
e of
S(fr)
arches.

72
AN IMPORTANT CONTRIBUTION
OF THE Q-SYSTEM:
‘NOMINALLY UNLINED’
HEADRACE TUNNELS FOR
HYDROPOWER
(economic, rapidly constructed)

73
Hydropower tunnels in
Norway. 3,500 km total
length.

From the 1960’s:

nominally ‘unlined’.
(Post 1974, often Q-
based, = NMT: S(fr)+B
where needed).

‘Rocks in the
turbines’? Never?
Why? Parabolic flow
velocity.
Thidemann and Bruland, 1992, Broch 2011.
Two options: concrete
lined, or larger DRILL-
AND-BLASTED section
with S(fr) + B where
needed.
 USE FRACTURE-FLOW ANALOGY TO FLOW
ALONG
ROUGH-WALLED HYDROPOWER TUNNEL

• THE SO-CALLED ‘ROCK-TRAP’ IS ACTUALLY ONLY

SLOWLY FILLED WITH SILT, SAND, AND MAYBE
SMALL GRAVEL PARTICLES, AFTER e.g. 10 TO 15
YEARS OF OPERATION BEFORE EMPTYING……NO
‘ROCKS IN THE TURBINES’
Louis and
Wittke,
1969.

‘BED-LOAD’
TRANSPORT
NOT EASY
WITH ZERO
VELOCITY.
With occasional HEADRACE TUNNEL collapse/blockage/break-
through trauma, large blocks can be transported, due to ultra-
high velocities. (Sources: NB and confidential)

Debris reach 400m from collapse * Coarse material upstream * Stones collected from 4km
 CONTENTS OF A ‘ROCK-
TRAP’. THE STONES ARE
PUMICE THAT CAME
FLOATING.
NO SIGNIFICANT SIZE OF BED
TRANSPORT….DUE TO
VELOCITY non-ISOTROPY
(9 km TBM + 1 km D+B)
 HJULSTRÖM DIAGRAM
from 1937 Ph.D. river studies.

Does not apply to rough

particles along rough walls.

(Does apply to rounded

particles/stones in rivers)

REALITY IN A
‘NOMINALLY UNLINED’
HEADRACE’ TUNNEL:
1.5, 2.5 AND EVEN 3.5
m/s WATER FLOW
VELOCITIES DO NOT
MOVE PARTICLES EVEN
AS BIG AS
10X10X10 mm.
 FINALLY:
COST ESTIMATION
IN
RELATION TO
THE
Q-VALUE

82
 Historic cost
differences:
concrete
lined/ ‘nominally
unlined’

• Lining where not

needed is 4 to 5 x
more costly:
NATM / NMT
These cost-versus-Q curves were developed using Norwegian prices for
bolting, shotcrete, RRS and concrete (plus transport prices for tunnel mucking),
dating from 2000. (Approximate doubling of price in last 10 years).
This Q-value statistic was estimated from Q-histogram logging of
about 800 m of core from evenly spaced boreholes drilled along
a future motorway ring-road tunnel. The estimated number of
meters of the given Q-classes are shown.
 Multiplication of meters
frequency by cost/meter
for given rock classes.

Note that pre-injection

could potentially ‘transfer’
much of the low-Q-high-
cost area towards the right.

(See lecture on pre-

injection)