You are on page 1of 6

INDOROCK 2013: Fourth Indian Rock Conference 29 – 31 May 2013

Design Practice for Large Underground Structures


C. KUMAR1, A. MANDAL2 & A. NANDA1
1
Sub Surface Projects Division, Engineers India Ltd., New Delhi, 110066, India.
2
Associate Professor, Deptt. of Civil Engg, VNIT, Nagpur, 440010,India
e-mail: chandan@eil.co.in

ABSTRACT: Design of cavern during the basic engineering design (BED) was carried out by collating the
inputs from geotechnical investigation and surface mapping performed during the initial site investigation stage.
However, during excavation, geological hotspots or locations of particular geotechnical interest are identified
along the cavern alignment based on the geological information obtained from the excavation of cavern at
various stages, including core drilling and deformation of excavation surface measured from monitoring targets.
One of the geological hotspot, which consists of mafic dyke intruded in the parent rock, is discussed here.
Stability of cavern is further re-checked at this location owing to an assessment of geotechnical features
encountered. The paper outlines the design aspects of underground rock caverns based on updated structural
geological maps, staged excavation, & verification by monitoring.
1 INTRODUCTION
1.1 Project Description
Rock cavern taken for the analysis are generically of U-shaped in Plan & D shape in cross section and are
kept parallel to each other. Orientation (trend) of the underground cavern is kept at N60°E. The underground
facilities; See Figure 1, essentially consist of access tunnel, main cavern (30 m height, 20m width and about
700m length) and water curtain tunnels running parallel to and 20m above each U shaped cavern with a series of
water curtain boreholes drilled perpendicular to it. The cavern gallery roof is seated at about 60m below the
ground surface with each leg having a separation distance of 30m between them. The project consists of two U
shaped caverns placed at an inter distance of about 100m.
Water curtain tunnel, which is excavated before the cavern excavation, also acts as a pilot tunnel exposing
the actual geology to be encountered during cavern excavation. The excavation of cavern is carried out in four
stages i.e. top gallery and three benches each of about 7-8m in height.

Fig. 1 Layout of project

1.2 Geological & Geotechnical description


The project area comprises of banded and granitic gneisses, migmatites along with true intrusive granites
and few mafic intrusions. The banded gneisses consist of white bands of quartz-feldspar (felsic bands)
alternating with dark bands containing hornblende, biotite and minor accessory minerals (mafic bands). The
granites are porphyritic to granular with typical quartz vein system and at times with intrusive properties. The
mafic intrusive are in the form of doleritic dykes of varying thickness in the above parent rock.
INDOROCK 2013: Fourth Indian Rock Conference 29 – 31 May 2013

All the major discontinuities along with different types of dykes etc encountered in the project area are
shown in Figure 2.

Fig. 2 Geological model of the project

Details of the structural discontinuities within the rock mass as determined from site investigation mapping
reveals three major discontinuities (sub-vertical) and one sub-horizontal as shown in table 1. The sub-vertical
discontinuities are persistent and oriented in almost north-south and east-west direction with a dip of about 80-85
degrees both sides. Sub-horizontal joints are oriented East to N60°E and dipping about 5 - 15 degrees both ways.

Table 1 Joint sets considered for Wedge analysis


No Joint set number Dip Direction Dip (°)
1. Joint set 1 (TD5) 250/70 85
2 Joint set 2 100 85
3 Joint set 3 (TD1) 200/20 85
4 Horizontal 150/330 15
)or
ien
Geotechnical parameters of rock mass, in-situ stresses and joint orientation etc have been derived from the
tati
results of geotechnical investigation, surface mapping and laboratory tests performed during the site
on The rock mass has been further divided into 5 classes of very poor, poor, fair, good & very
investigation stage.
good as per Barton’s
) Rock tunnelling Quality Index; Q, Barton (1974) System criteria.
2 INITIAL DESIGN OF CAVERN
The design of cavern essentially consists of two parts, Mandal et al. (2011):
(i) Wedge Analysis (using statistical joint data): To analyze the stability of possible wedge in the jointed rock
mass formed due to the tunnel excavation. For the stability of possible sidewalls wedges, a factor of safety
(FOS) with 1.5 and for the roof, a FOS of 2.0 is considered.
In the analysis, Mohr-Coulomb joint strength criterion was used and the values of friction angle and cohesion
are obtained from the results of a study by Barton (1974) for the shear strength of filled discontinuities for
granites and gneissic type rock.
(ii) FEM analysis : To analyze the extent of rock deformation into the cavern due to distressing of rock mass
resulting from the excavation. Generalised Hoek & Brown criterion is used to assign the rock mass properties.
Criterion for the maximum movements is limited with 0.5% of the maximum span of an excavation under
normal loading conditions.

Results from the wedge analyses shows that rock bolt in combination with shotcrete is essential in increasing
the factor of safety to an acceptable level. The maximum apex height of wedge formed at crown is about 4.0 m.
Therefore, minimum length of recommended rock bolt is 5.0 m i.e. at least 1 m more than the maximum wedge
apex height.
INDOROCK 2013: Fourth Indian Rock Conference 29 – 31 May 2013

Results of the FEM analysis shows that no global failure occurs even in the case of poor and very poor rock
mass condition. The proposed rock support was found to be adequate and sufficient. Summary of recommended
rock support for the cavern are given in Table 2.

Table 2 Recommended rock support


Q-value Q>40 10<Q<40 4<Q<10 1<Q<4 Q<1
Support/Rock Type I-Very Good II-Good III-Fair IV-Poor V-Very Poor
Crown Support
Bolt length 5m 5m 5m 5m 5m
Shotcrete 50mm 50mm 75mm 100mm 150mm
Bolt spacing, m(c/c) 2.5 2.0 1.75 1.5 1.5
Sidewall Support
Bolt length 5m 5m 5m 5m 5m
Shotcrete 50mm 50mm 75mm 100mm 150mm
Bolt spacing, m(c/c) SB 2.5 2.0 1.75 1.50
SB: Spot Bolting

3 DESIGN VERIFICATION DURING CONSTRUCTION


3.1 Identification of critical area
The basic design needs to be verified and modified during the excavation stage based on the updated
geological condition encountered, especially in areas where stability could be critical, Nanda (2012).
A series of water curtain boreholes perpendicular to the water curtain tunnel, which runs parallel and is 20m
above main cavern tunnel were drilled. Some of these boreholes were logged or imaged by BHTV (Borehole
televiewer) to understand the complete geology beyond the tunnel and above the storage caverns. A geological
model was developed at cavern heading level based on the extrapolation from water curtain borehole geology.
Also the geology of the entire project is further being updated after each round of cavern excavation i.e. heading,
first bench and second bench. Based on the updated geology, some critical locations which are classifies as a
geological hot spot are identified and marked in the geological map as shown in Figure 3.

Fig. 3 Geological model showing geological hotspots of project

Out of those identified critical locations, analysis of one geological hotspot stretched in a span of about 50m,
and is marked in circle, is discussed here. The dyke TD1 that is intruded in the parent rock is thin i.e. about 1m
and highly permeable. Their contact is slightly weathered with occurrence of 5 -10 mm soft gouge material.
These dykes are fresh & hard and moderately fractured. The dyke was found to be highly persistent and this was
accounted for in the analysis.
INDOROCK 2013: Fourth Indian Rock Conference 29 – 31 May 2013

3.2 Geological discontinuities


To verify the quality of rock mass and to detect major geological features at specific location, two nos. of
core drilling from water curtain tunnels above the cavern units was carried out. The borehole has been drilled
horizontally up to a length of 52m with a slope of 1:10.
From drilling core log which was placed just over the geological hotspot, the bedrock found was dominantly
light to dark grey coloured medium to coarse-grained granite gneiss with intrusion of black fine-grained mafic
dyke TD1, intersected by shear seams; fresh to slightly weathered; very strong to very weak. RQD in the
bedrock varied from 30% ; found in highly fractured basic dyke (fine-grained basic rock) & fractured rock near
to dyke zones to 100%. The joints were planar and smooth to rough. From other drilling core log, which was
kept just away from the geological hotspot, the bedrock was found to be in a very good condition. The RQD in
bedrock varied from 91% to 100%. Results of drill core log suggest that effect of TD1 dyke is constrained and
the extent of geological hot spot, which is taken as about 50m, is adequate.
Also at the geological hotspot, joint information from heading and first bench excavation in the sidewall of
cavern was studied in detail. The joint sets analysis was done by plotting the stereographs of joint set data for the
sidewall region of concern using the program DIPs. The rock mass is characterized by prominent one plus
random to three plus random joint set, which are rough (or irregular) and planar to undulating with unaltered
joint wall to softening clay mineral fillings. In general, the rock mass is characterized by dry/damp ground water
conditions. Major joint sets for both north & south wall are given in Table 3 & 4.
Table 3 Structural features on north wall
Discontinuity Dip Direction Dip (°) Persistence (m)
Joint 1 180 10 1-10
Joint 2 105 75 3->10
Joint 3/TD1 025 70 >10
Joint /Random 260 70 1-10
Joint/Random 135 45 1-10
Table 4 Structural features on south wall
Discontinuity Dip Direction Dip (°) Persistence (m)
Joint 1 335 10 3-10
Joint 2 125 85 3->10
Joint 3/TD1 030 75 >10
Joint /Random 260 80 3->10
3.3 Analysis & design updation
With respect to the updated geotechnical information, cavern design is re-checked by performing a block
analysis using the program Unwedge to establish if the current recommended rock support in that region of
sidewall had sufficient capacity to keep the wedges in place with a suitable safety margin. A safety factor of 1.5
was adopted as requirement for the limit equilibrium analyses of the wedges. For satisfying the required FOS
level, a typical rock support as per the Q system, additional rock bolt or minimum rock support is recommended.
A combination analysis was performed to sort through all possible joint combinations to select the joints,
which define the most critical wedges. The wedge formations, which have the potential for plane sliding with the
realistic shaped are, discussed in the results. Persistence of about 10m of each joint set is considered for all joint
set combinations. However, the persistence for TD1was varied from 10 to 20 m to observe the effect on the size
of wedge and factor of safety thereof.
Mohr-Coulomb joint strength criterion was used in the analysis. Angle of friction was taken as 32º for shear
seams and TD1 assuming the surface of joints to be rough (or irregular), planar with unaltered joint wall, with
softening clay mineral coatings fillings. For other joints, angle of friction was taken as 40°. Cohesion component
of all joint sets was ignored. Results of the analysis for both north and south wall are given in Table 5 & 6.

Table 5 Results of Unwedge analysis & size of critical wedge formed on north wall
Wedge Information Joint Combination
J1,J2 & J3 J1,J3 & J5 J3,J4 & J5
FOS without support stable 0.8 0.8
Wedge volume (m3) 0.33 12.537 67.44
Wedge weight (MN) 0.009 0.35 1.88
Apex Height (m) 0.5 2.85 4.5
Excavation face area (m2) 2.82 13.87 44.95
Min. Rock support type As per Q system Class-II Class-III
INDOROCK 2013: Fourth Indian Rock Conference 29 – 31 May 2013

Table 6 Results of Unwedge analysis & size of critical wedge formed on south wall
Wedge Information Joint Combination
J1,J3 & J4 J1,J2& J3 J1,J2 & J4
FOS without support 0.8 No wedge is formed stable
3
Wedge volume (m ) 24.94 - 13.36
Wedge weight (MN) 0.70 - 0.37
Apex Height (m) 4.67 - 2.09
2
Excavation face area (m ) 16.0 - 19.16
Min. Rock support type Class-II As per Q-System As per Q-System

Results of the analysis in north wall indicate that a large wedge ; see Figure 4 (a), of about 4.5 m height,
weight of about 1.88 MN and wedge volume of about 67.44 m3 is formed with joint set combinations ( J3,J4 &
J5).Here persistence of mafic dyke TD1 has a considerable role in the formation of wedge size. Apex height of
wedge varies from 3.5 m to 4.5 m when the persistence of TD1 varies from 10m to 12.5m. In the present case,
when the persistence of other joint set is limited to 10m, maximum apex height of wedge formed is restricted to
4.5m even after the persistence of TD1 extent larger than 12.5m. Based on above, a minimum rock support of
rock support type Class III i.e. rock bolt 5m @ 2.0m c/c and 75 mm shotcrete is recommended. For joint set
combinations (J1, J3 & J5), a minimum rock support of rock support type Class II is required. For all other joint
set combination, minimum rock support as per Q system is adequate.
Results of the analysis in south wall indicate that in general a minimum rock support as per Q-system is
adequate except for a joint set combinations (J1, J3 & J4), wherein a minimum rock support type class-II i.e.
rock bolt of 5m @ 2.5m c/c and 50 mm shotcrete is required. Apex height of formed wedge is about 4.6m ; see
Figure 4 (b), however, the wedge volume and weight are less as compared to the wedge formed in north wall.
Persistence of TD1 has no considerable role in wedge height and it is restricted with the overall persistence of
other joint set of 10m.
Based on the results, it was therefore recommended that before taking up next bench excavation, an
additional rock support to be install in the present bench (if required) in the given segment subject to the given
joint set combination as shown in table 5 & 6.

Fig. 4(a) Wedges formed on north wall Fig. 4(b) Wedges formed on south wall
Fig. 4 Critical wedges formed
3.4 Geotechnical Monitoring
Monitoring array stations within the caverns are typically arranged at every 25m interval with five optical
targets being placed within the top heading and two targets on either side in each bench of excavation. In
addition, some extensometers at geologically critical areas are installed in the drilled boreholes of water curtain
tunnel located above the cavern. These extensometers are anchored at 3 points to measure vertical displacement
of each point.
INDOROCK 2013: Fourth Indian Rock Conference 29 – 31 May 2013

At one of the monitoring array station falling in the present geological hotspot segment, displacement
recorded at pilot and side slashing was about 6-8mm, which was greater than or equal to trigger value in top
heading. The trigger value is set at 80% of the estimated design displacement. However, the displacement plot
shows stabilizing trend and seems stable. It was further advised to monitor the present array station once every
day before taking up any further excavation. It was also recommended to inspect physically at this location for
any distress in rock mass in terms of cracks in the shotcrete and increase in seepage etc. Displacements recorded
on both the north & sidewall were limited to 3-4mm and showed stable condition.
Result from the extensometer shows a deformation of about 3-4mm at crown, which stipulate a stable
condition as there is no significant deformation/movement of the rock mass observed at the crown region.
Stability of cavern was further ascertained from the monitoring results after bench-2 excavation, which
shows stable condition with the above installed support.
4 CONCLUSIONS
Following conclusions can be drawn from the present study to develop best practice to tackle the enormous
uncertainties involved in construction of underground large cavern.
Stability of large rock cavern excavation should be ensured by an observation approach i.e. the original design
assumptions are to be updated by means of the results from monitoring during construction, to allow for an
optimized design and not to be limited to a one off design exercise, which is performed at the basic design stage.
In the cavern, which is discussed in this paper, some geological hotspot or locations of particular geotechnical
interest are identified along the cavern alignment based on the updated geological information collated from the
excavation at various stages. Quality of rock mass and major geological features at specific location has been
derived from core drill log performed from water curtain galleries just above the main cavern as well as joint
information collated from heading and first bench excavation in the sidewall of cavern. With respect to the
updated geotechnical information, cavern design is re-checked by performing a block analysis using the program
Unwedge. A minimum rock support based on the class type is recommended for both walls. Stability of cavern is
also being ascertained from the displacement behaviour obtained from geotechnical monitoring scheme. Here,
the displacement plot both from optical targets and from extensometer shows stabilizing trend, however, regular
monitoring of displacement behaviour has been suggested in the specific geological hotspot area.

REFERENCES
Barton N., Lien R. and Lunde J. (1974), “Engineering classification of rock masses for the design of tunnel
support, Rock Mechanics”, Vol.6, no. 4, pp. 189-236.
Mandal A., C. P. Chakravarthy, R. Rath, A. Nanda and A. Usmani (2011), “Analysis and Design Approach of
Large Storage Caverns”, (In Press), Int. Journal of Geomechanics, ASCE. doi:10.1061/(ASCE)GM.1943-
5622.0000180
Nanda A., Design and construction of storage caverns, Proceedings of Indian Geotechnical conference, Dec.13-
15, 2012, Delhi.

You might also like