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

Geological Prediction of Weak Dyke during construction of a Rock Storage

Cavern
G. KANNAN1, C. P. CHAKRAVARTHY1, C. KUMAR1, A. NANDA1 & S. K. JAIN2
1
Sub Surface Projects Division, Engineers India Ltd., New Delhi, 110066, India.
2
Associate Prof., Dept. of Civil Engg, Jaypee University of information Technology, Solan, HP, 173234, India
e-mail: gopi.kannan@eil.co.in

ABSTRACT: This Paper deals with a mafic fractured dyke of about 30m thick and more than 500m length
encountered during construction of a large underground storage cavern in India. The geological and
geotechnical problems associated with this dyke and the challenges encountered during the investigation and
construction stage are presented in this paper. The fractured dyke is in general perpendicular to the cavern
alignment but it offsets suddenly about 50m from the original orientation and encountered in one of the caverns
in parallel orientation for about 50m and dips steeply towards the wall of the cavern and possesses stability
problems in the cavern wall. The probe hole drilled inside the cavern wall shows that these dykes are extended
up to a distance of 10 m behind the cavern wall. Borehole televiewer (BHTV) was also conducted in the water
curtain boreholes drilled above the cavern to study the geometry of the shifted dyke in the intermediate cavern
pillar. Unwedge and FEM studies were conducted to ascertain stability of cavern from wedge sliding as well as
wall stability. Special supports were recommended based on the analyses and investigations. Now the cavern is
fully excavated in this zone and the extensometer readings show the cavern condition is stable.
1 INTRODUCTION
The rock cavern project under study is a strategic crude oil storage project under construction in west coast
of India near Mangalore. The project consists of four U shaped caverns of 30m height, 20m width and about
700m length (See Fig.1). Each U shaped cavern consists of inlet and outlet shaft at the open ends for crude oil
movement through pipelines. The cavern is seated about 60m below the ground and is excavated through two
access tunnels; one for each 2 U shaped caverns. The caverns are placed about 100m inter-distance. Each leg of
U shaped cavern is having about 30m width of rock pillar and above each U shaped cavern, water curtain
system is developed to contain oil and gas using Hydro-geological confinement.

Fig. 1 Layout of the project

Water curtain system consists of a water curtain tunnel parallel to and 20 m above the cavern and a series of
water curtain boreholes drilled perpendicular to the tunnel while extending cover of about 20m beyond both
sides of U shaped caverns. These boreholes are pressurized with water during water curtain tunnel construction
to prevent lowering of water table during cavern excavation below. As a matter of fact, the water curtain
systems should work properly atleast 50 m ahead of cavern excavation to prevent de-saturation in rocks near
cavern which will be difficult to re-saturate once de-saturated leaving air traps and air bubbles risky for stored
oil and gas to leak in future.
Since the water curtain tunnel is excavated before cavern excavation, this acts as a pilot tunnel exposing the
actual geology to be encountered during cavern excavation. As the caverns are 30m height, they are excavated
in four stages; top gallery and 3 benches each of about 7-8m in height.
INDOROCK 2013: Fourth Indian Rock Conference 29 – 31 May 2013

2 GEOLOGY OF THE PROJECT

The project is seated in Peninsular Gneissic rock complex of Karnataka. The regional geology is
characterized by heterogeneous mixture of different types of granites intrusive into the schistose rocks after the
latter were folded, crumpled. They include granites, granodiorites, gneissic granites and banded or composite
gneisses, the granitic constituents of which show distinct signs of intrusion. The gneissic types are probably due
to the intensive granitisation of older schistose rocks and show streaky and contorted bands of which are
granitoids to porphyritic and other granulitic intrusive. They include composite gneisses, migmatites, granitised
older crystalline rocks and true granites with their aplitic and quartz vein systems.
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.
The project area has undergone extensive weathering and has lateritic soil at the top followed by weathered
and fresh bed rock. The bedrock is slightly to moderately weathered with occasional occurrence of shear seams,
dykes, fracture zones and clay coated joints.
There are three major discontinuities observed in the project area; two sub-verticals and one sub-horizontal.
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. Major tectonic and geomorphic features are also aligned parallel to these
discontinuities. Mafic dykes are also found roughly oriented parallel to these discontinuities indicating these
intrusions are both syn-tectonic as well as post tectonic. Sub-horizontal joints are oriented east to N60E and
dipping about 5 - 15 degrees both ways.
Four different types of mafic dykes are found intruded in the bed rock in the project area (See Fig.2). Two
of them TD1 & TD2 are thin (0.5 to 1m) and highly permeable. Their contact is slightly weathered with
occurrence of 5 -10 mm soft gougy material. These dykes are fresh & hard and moderately fractured. Third dyke
TD3 found was about 3m thick and the contact is fresh & tight and is totally dry. This dyke intersects the other
dykes present and its thickness varies and found branched at these intersections.
The fourth kind of dyke namely TD5 found was about 30m thick and would intersect all the caverns
perpendicularly. The dyke rock is hard and fractured and at contact with parent rock, it is fresh and tight in
general but at some places, sandy silt particles can be found in the contact. Very little amount of water is
intersected in this region. Fracture zone (FR4) is found on either side of this dyke with clayey silt infilling of
about 5-10mm. Dyke + fracture zone consists of about 50 - 60 m thick and would make it a very critical zone
during construction.

Fig.2 Geological model of the project

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

3 CONSTRUCTION OF CAVERN THROUGH THICK MAFIC DYKE

During investigation stage, this 30m thick fractured mafic dyke TD5 was encountered in a corehole and was
studied in detail. Its orientation, thickness, geological, geotechnical and hydro-geological properties were well
understood. It was classified as class IV rock as per Q system (Barton et al, 1974). The layout of storage cavern
was chosen (along with all other relevant geotechnical parameters) such that this dyke cut across almost
perpendicularly all the storage caverns.
Accordingly, rock supports were designed for this 30 m thick mafic dyke zone after detailed numerical
analysis & unwedge analysis. The construction sequence for the upper water curtain tunnel was planned with
probe holes ahead of excavating this dyke. The dyke zone was classified as class IV rock after excavation. As
the dyke intersected water curtain tunnels (as expected) obliquely (almost perpendicularly) and with very little
ground water seepage, it was not considered a big challenge geotechnically. Rock support of 100mm shortcrete
with 5m rockbolts and 1.5 – 1.75m spacing centre to centre was applied to stabilize this dyke zone.
As a part of construction of water curtain above storage cavern, water curtain boreholes were drilled
perpendicular to water curtain tunnel. In this dyke zone, these holes were drilled parallel & in to the dykes.
Some of these boreholes were imaged by BHTV (Borehole televiewer) as a part of construction strategy to
understand the complete geology beyond the tunnel and above the storage caverns. These were extrapolated to
predict the geology that would encounter in cavern heading. A geological model was developed at cavern
heading level and the design of caverns was updated (A. Nanda, 2012). Intersection of these dykes in the
caverns along with other critical geological locations are classified as Geological hot spots and marked in the
geological model (See fig.3) and issued for construction and planning purposes.

Fig.3 Geological model showing 30m thick dyke as geological hotspots of the project

During cavern heading excavation, sudden change in the position of the dyke TD5 was observed in one of
the storage caverns (See fig.4). The dyke offsets about 50m from the original orientation and encountered in the
shifted position in one of the caverns. But their orientation was same. The shifting could have taken place in the
rock pillar between the two adjacent storage caverns. To confirm this, 3 water curtain boreholes drilled in this
rock pillar just above the observed shifted dyke was examined with BHTV. It revealed that the shifting in dyke
does occur in the rock pillar and the orientation of dykes, which were split in to numerous dyke swarms, were
found to be both perpendicular as well as parallel to the cavern alignment. The parallel orientation of the dyke
swarms along with its steep dip towards north wall in one of the caverns was found to be critical. The fracture
zone FR4 found along either direction of the TD5 dyke seems unchanged in orientation and got overlapped with
parallel alignment of dyke in this zone. This zone in north wall of cavern was physically found to be fractured
and crushed with silty clay infilling and was classified as rock class V where as in south wall of the cavern, no
shifting in dyke was found and the dyke remains perpendicular to the cavern wall and is classified as class IV.
INDOROCK 2013: Fourth Indian Rock Conference 29 – 31 May 2013

A close study on the northwall geology and BHTV studies suggest that the offset of the dyke is ductile and
at the offset, it locally splits into number of small dykes (dyke swarms) and some of them runs parallel to the
cavern orientation for about 50m and some of them runs perpendicular to the cavern same as original dyke but
overlapped with similarly oriented fracture zone. Thus the overall geology in this area are quite complex (See
Fig.4). Also the shifting in dyke could have caused by dyke TD3 which is intersecting this TD5 dyke at this
same location. Another possibility could be that this dyke was emplaced on pre-existing fracture system and
during emplacement dyke found its own way along the weaker zone explaining the ductile mechanism observed.
Thus geologically the phenomenon is not still well understood whether it is brittle or ductile and which intrusive
dyke is younger one and the fracture zone associated (FR4) is syn-tectonic or post tectonic could be best left to
the structural geology experts.
To ascertain the extent of branched out dyke passing parallel to the cavern, 4 nos. of probe holes of about 15
m length were driven. The probe hole result shows that these dykes are extended upto a distance of 10 m behind
the dyke. However, the persistence, thickness and orientation of each branched dyke rock were still unknown. In
addition, these branched dyke rocks were highly irregular and discontinuous.

Fig. 4 Shifted dyke & interpretation from BHTV studies

4 CAVERN WALL STABILITY STUDIES & GEOTECHNICAL MONITORING

4.1 Initial Design
Design of cavern during the initial stage consists essentially of two parts. The one is based on the empirical
rock mass classification schemes proposed by Barton (Q-system) which is checked and modified by Unwedge
analysis (using statistical joint data) to analyze the stability of possible wedges in the jointed rock mass & the
other through finite element method to analyze the extent of rock deformation into the tunnels and cavern due to
distressing of the rock mass resulting from the excavation.
Intact rock properties, in-situ stresses and joint orientations were taken from the geotechnical investigation
and surface mapping performed during the site investigation stage. The details of the structural discontinuities,
dyke etc considered for unwedge analysis are shown in Table 1.

Table 1 Joint sets considered for Wedge analysis

Joint set number Dip/Dip Direction
1(FR4 zone & TD5 dyke) 85/70
2 (major set) 85/20
3 (random) 15/0
1(FR4 zone & TD5 dyke changed) 85/160
)orientation)

At crown, a wedge formed with maximum apex height 4.0 m (Refer table 2), suggests that bolt length
should be at least 1 m more than the wedge apex height. Therefore minimum length of the bolt should be 5.0 m.
Summary of the recommended rock support for the caverns are given in Table 3.
INDOROCK 2013: Fourth Indian Rock Conference 29 – 31 May 2013

Table 2 Wedge analysis with dyke orientation obtained during initial investigations
Joint set Location of wedge Apex Unsupported factor of FOS with
Combination Height safety Support

Roof 2.35 0.073 9.6

J1-J2-J3
Wall 3.65 0.073 11.8

Table 3 Recommended Rock Support for cavern

Cavern 20 X 25.4–31 m (W X H)
Q-value 1<Q<4 Q<1
Support Type IV (Poor) V (V.Poor)
Crown Support
Bolt length 5m 5m
Shotcrete thickness 75 mm 150 mm
Bolt spacing 1.5 m 1.5 m
Sidewall Support
Bolt length 5m 5m

4.2 Design verification during excavation

With the shifted orientation of dyke in north wall of cavern, stability of wedge analysis has been performed.
This wedge analysis results shows that due to shift in orientation, the wall wedge weight and apex height has
been increased. However, the required factor of safety of wedges could be achieved with the recommended
initial class V rock support (see table 4).

Table 4 Wedge analysis with Shifted dyke orientation in northern wall of cavern
Joint set Combination Location of Apex Unsupported factor FOS with
wedge Height of safety Support
Roof 1 0 Stable
J2-J3-J4
Wall 4.5 0.5 2.7

As the rock mass around this zone is highly fractured, closely jointed and of complex geology, numerical
analysis using Phase2 software was also carried out to access the north wall stability with assumed rock support.
The analysis models were developed from respective surrounding rock mass and geometry of mafic dyke of
about 10m thick placed on the north wall of cavern with the properties of dyke varying from poor to very poor.
Results of the FEM analysis shows that Stress level observed around the cavern are moderate and stable.
Yielded zone varied between 2-3m and is restricted close to central part of north wall of cavern. Limited
yielding of rock bolts was noted. Maximum displacement observed during excavation of bench 2 was estimated
between 4-8 mm. Overall displacement is relatively small and acceptable. Strength factor observed between the
pillars is around 2 and the shear stress is very close to zero. This state ensures the stability of pillar between two
caverns.
The results of the analyses both from unwedge & FEM indicates that there is no cause for concern for
wedge sliding as well as wall stability. However, due to uncertainties based on the fact that the branched dykes
were highly irregular and their persistence, orientation, etc were still unknown, one row of 12 m rock bolt (25
φ) were recommended at both bench 2 and 3 level, which will cross over the dyke zone of about 10m. In
addition shotcrete thickness of 100mm was proposed on north wall. The revised recommended rock supports at
weak dyke zone are shown in Table 5.
To check the integrity of applied recommended systematic rock bolt, pull out test with a load 1.5 times the
design load i.e. 20 MT was performed on 5 rock bolts of 5 m in length on north side of wall. The results of the
pull out test shows a maximum deformation of 4.3 mm at ultimate load and none of the rock bolts were failed at
this load.
INDOROCK 2013: Fourth Indian Rock Conference 29 – 31 May 2013

4.3 Geotechnical Monitoring

The readings taken from two borehole extensometers installed from water curtain tunnel located above the
cavern at this zone was observed during the heading excavation. These extensometers are anchored at 4 points.
The result from extensometer shows a deformation of about 1mm at crown, which stipulate the stable condition
as there is no significant deformation/movement of the rockmass at the crown region during and after the
excavation.
In addition, before start of bench 2 excavation, one extensometer (3-point anchor) each on north & south
wall were installed in this zone to observe the deformation in wall. A deformation of about 3 mm is observed
which is well within the acceptable limit.

Table 5: Recommended Additional Rock Support

Support Type Q<1 (initial) Q<1 (Additional) North Wall

Rock Bolt 5 m @ 1.5m c/c 1 row of 12m each in bench 2 & 3

Shotcrete 150 mm 100 mm

5 CONCLUSIONS
During cavern heading excavation, a sudden change in the position of Mafic fractured dyke was observed. The
dyke offsets about 50m from the original orientation and encountered in one of the caverns in shifted position.
Shifting could have taken place in the rock pillar between the two adjacent storage caverns where the dyke runs
parallel to the cavern alignment. A Borehole image processing studies (BIPS) were conducted in three boreholes
drilled in the water curtain tunnel just above the observed shifted dyke to confirm this phenomena and also 4
nos. of probe holes of about 15 m length were driven to ascertained the extend of dyke. The probe hole result
shows that these dykes are extended upto a distance of 10 m behind the cavern wall. These branched dyke rocks
were highly irregular and discontinuous and it locally splits into number of small dykes (dyke swarms) and all
of them runs parallel to the cavern orientation for about 50m. As the rock mass around this zone is highly
fractured, closely jointed and of complex geology, wall stability of the cavern in this zone was re-ascertained
using both wedge and numerical analysis. Results of the analyses both from unwedge & FEM indicates that
there is no cause for concern for wedge sliding as well as wall stability. However, as the branched dykes were
highly irregular and their persistence, orientation, etc were still unknown, one row of 12 m rock bolt (25 φ)
were recommended on north wall at both bench 2 and 3 level. The displacement plot of extensometer both from
wall as well as from water curtain tunnel confirms stable conditions.

REFERENCES
1. Barton. N., Lien. R, Lunde, J., 1974, Engineering classification of rock masses for the design of tunnel
support, Rock Mechanics, Vol.6, n°4, pp 189-236.
2. Nanda A., Design and construction of storage caverns, Proceedings of Indian Geotechnical conference,
Dec.13-15, 2012, Delhi.