BY
S~Y. New triaxial stress gauges have been developed for the determination of the three dimensional
stress field at a point in rock during a single overcoring operation. The strain sensor consists of nine
or more electrical resistance wire strain gauges embedded in a thin-walled epoxy pipe. TWo variants are
used. one for horizontal and down holes and one for up holes. The stress gauges are cemented in an EX size
pilot hole with epoxy adhensive stored in the inner space of the strain sensor pipe. and from which it is
extruded into the space between the gauge and rock by means of a plunger. The gauges permit continuous
observation of strains during overcoring. are fully isolated from drilling water. and operate satisfactorily
at low bond stI'ength with rock. They have been used under a variety of rock conditions to measure stresses
varying from 1.5 to 100 MPa and at temperatures from 15° to 45°C. The theoretical and experimental verifi
cation of the method and the in situ test procedures are presented.
1 INTRODUCTION 2 NOTATION
In 1972 development was vegun of a rock stress radial poSition of strain gauges in the
measurement capability based on parallel application EX size borehole
of two different borehole overcoring methods. the a radius of the EX size borehole
USBM borehole deformeter (Ref. 1) and the LNEC tri c:c axial strain on the surface of the EX size
axial stress tensor gauge (Ref. 2). The latter was borehole
chosen in preference to the CSIR triaxial stress circumferential strain on the surface of
gauge (Ref. 3) because it provided full protection the EX size borehole
of electrical circuitry from drilling water. and strain on the surface of the EX size bore
allowed strain observations to be made during the hole at an angle of :t45° to the borehole
progress of overcoring. axis
y6z shear strain component
Kl.K2,K3'~ = strain correction factors
A number of LNEC gauges were built and tested
OZ.Oy,Oiil = normal stress components in "local'
in the laboratory with a strain gauged area 50 mm
coordinate system
long and a total length of about 100 mm to permit
their use in more closely jointed rock than was Try. TYiil' T= .. shear stress components in ' local'
possible with the original LNEC gauges . An epoxy coordinate system
cement developed for cementing the gauges into pilot Er Young's modulus of rock
boreholes gave bond strength to wet rock (modulus v Poisson's ratio of rock
Young's modulus of epoxy resin
of rupture) of over 7 MPa. As in the original
Young's modulus of rock determined from
design the gauges were for use in down holes with
the cement stored in plastic bags attached to the biaxial tests of cores with hollow inclu
front end of the gauge. sion gauges
0 1 . 02. 0 3 principal stresses
Bl.B2. B3 bearing. east of north
A mathematical analysis carried out using El,E2,E3 Up from hor'izon. positive
Muschelishvilli's method of complex potentials (Ref.
4) showed however. that because of the high Poisson's
ratio of the plastic from which the gauges were made, 3 DESIGN AND CONSTRUCTION
they would behave as rigid inclusions when cemented
into boreholes and would often require a bond (a) Prototype
strength in excess of that of available cements.
This was confirmed by field tests. Fig. 1 shows a prototype gauge developed in
1973 for use in both horizontal and down holes. An
epoxy resin shell with strain gauges was mounted on
In view of this. a 'hollow inclusion gauge' was a steel pipe which served as a cement container
developed in which the solid cylinder used in the with 1 mm air gap between the plastic and steel'
LNEC cell was replaced by a short thin-walled epoxy pipes. All electrical connections were potted with
resin pipe in which three strain gauge rosettes were a cable emerging through the steel rear end.
encapsulated. The latter measure strains on the
surface of the pilot hole during overcoring as in a After filling the container with cement, a
CSIR triaxial cell. This design reduced the requirec steel plunger was inserted into the front end for
strength of bond to the rock to usually less than 1 about 2 cm and held in place by lead shear pins; a
MFa; an alternative method of storing the cement wa~; wooden spacer rod was inserted into the front end
also developed. which allowed the use of the gauges of the plunger to position the gauge at a required
in holes of any direction. distance from the, end of the EX hole. e.g. to avoid
1
WOODEN RESIN SIIE LL WITH STRAIN GAUGES CABLE
OUTLET HOLES fOIl EPOXY LEAD SHEAR PIN POTTING COMPOUND Uti
a joint in the rock. The gauge was then pushed into The recovery of gauges was difficult and time
the end of the EX hole, shearing the lead pins and consuming. The design was therefore modified in
extruding the cement through the holes in the plunger. 1974-75 to reduce the cost of manufacture. The
The cement was forced into the annulus between the steel plunger and pipe were replaced by a hollow
plastic pipe and rock, and held in position by neo acrylic resin plunger. sliding directly into the
prene rubber seals. (In an up-hole version of the epoxy resin pipe and the whole body of the gauge,
gauge the plunger was made solid, and the cement including the rear end, was made of epoxy plastic
was extruded through the openings at the rear of using extrusion moulding techniques. The external
the steel pipe between the plastic shell and the diameter of the strain sensor pipe was increased to
rubber seals.) 36 mm and the strain gauges were placed closer to
the pipe's surface so that when cemented into the
The three 45°/90° strain gauge rosettes were
borehole in the rock they were at an average dis
spaced at 120° along the circumference as shown in
tance of 1. 5 mm from the rock surface (rsg ~ 17.5 nun,
Fig. 2. There were three circumferential and two
wi th radius of borehole a = 19.0 mm). The design of
axial gauges and four gauges at ±45° to the axis of
the horizontal and down-hole stress gauge is shown
the hole. The layout was chosen to give approxima
in Fig. 3 and that for up-hOle gauge in Fig. 4.
tely evely balanced sensitivity to different compo
nents of the stress tensor and to provide for some A breakable resistance wire is stretched across
duplication of gauges so that in the event of a the cement reservoir near its rear end. The wire is
malfunction of one gauge of any of these three connected to a battery and a warning light which is
groups, it would still be possible to obtain the extinguished when the plunger is pushed fully in,
full stress tensor with reasonable accuracy. thus breaking the resistance wire. In the up-hole
type the plunger can be activated by a pull wire
The strain gauges were 10 mm long; from a mathe passing through the rear of the gauge as an a~ter
matical analysis of the relationship between length native to pushing the body of the gauge over the
and sensitivity, 10 mm was considered to be the most plunger.
suitable length for EX holes; it was expected that
satisfactory results would be obtained with a rock (c) Fabrication of Gauges
grain size of up to 4 mm.
A thin-walled pipe 32 mm i.d., 35 nun o.d. is
cast from Araldite 0 using an RTV rubber mould. The
The strain gauge circuit was of three wires
strain gauge rosettes (T.M.L. PRS-IO-ll) are then
quarter-bridge type and utilized the internal dummy
glued to the pipe using an epoxy strain gauge cement.
of the strain indicator. .
Very thin copper strip is stuck to the pipe in the
appropriate poSitions to act as lead wires for the
(b) Modification of Prototype
strain gauges. The breakable resistance wire is
soldered in position. Connections are then made to
The init.i al intention was to recover the stress
the separate conductors of 12 core, 14 x 0076 cable.
gauges and re-use then for stress monitoring or to
Finally, the strain gauges and lead wires are cover
recover the steel pipes and end pieces and replace
ed with a thin coating of Araldite D to make the
the epoxy resin shell and make up a new gauge.
pipe 36nuno.d. At the same time a 22 mmdia. instal
lation end piece is moultled over the cable.
The above moulding technique is made possible
by the use of RTV rubber moulds. These moulds
x
enable dimensions to be maintained, while still
allOWing easy release of the cast part. TwO types
of moulds are required - one to produce the inner
pipe and one to produce the outer surface/instal
lation end piece. The rubber moulds themselves are
y
produced by moulding arOund masters turned from
aluminium. The rubber moulds deteriorate with use,
but experience has shown that they are capable of
producing six to twelve components.
Completion of the gauge requires the position
ing and fixing of the neoprene rubber seals and the
manufacture of the plunger which is produced from
A B c 32 mm o. d. acrylic tubing with a 2 mm wall thickness.
9 30' l70' ISO' The gauges are then given an oven cure fOr 24 hours
f3 O' 90' 45' 45' 90' 135' O' 90' 45' at 55°C to complete the polymerisation of plastic
and ensure the stability of mechanical properties.
Tests of the completed strain sensor pipes gave a
Fig. 2. Orientation of strain gauges in the epoxy Young's modulus of 2.8 - 3.5 GPa and Poisson's
pipe. ratio between 0.35 and 0.40.
2
,/
/
OUTLET HOLES fOR EPOXY CEMUT EPOXY CEMEIIT
/
I
SET INOICATING SHEAR WIRE EPOXY CEMENT ---OUTlET HOLES fOR EPOXY CEMENT )Z••
4 THEORETICAL BASIS AND EXPERIMENTAL VERIFICATION assumed to be the same as that of plastic).
3
vations. If, as usual, the Young's modulus of rock
Uy=U-r 6'89HPa o G'z=6 ' 89HPa
...,
o
ay=O
is determined in a biaxial test of the rock core with
the stress gauge, use can be made of an option in
the program for adjusting the stress-strain coeffi
cients for the axial Ox and shear stresses, TZX and
Tyx to compensate for the 'smaller effect of the 'gap'
between the strain gauges and rock on the axial and
shear strains and stresses as compared with its
effect on the circumferential strains due to stresses
0Y' 0z and T yz • The equations 1-4 are modified by
d~viding both sides of the equation by the average
of Kl and K2 factors
y~--~r--+--- ~- ~-+---+---~--~~~y
t2 (0=19I11III1
--Strain at position rsg=17'SIIIIII If, for some reason, e.g. the core is too short
z and thus determination of modulus in the biaxial test
is impossible, another option permits computation of
stresses in accordance with eqs. 1 to 4 using the
Fig. 5. Circumferential strain on the surface of the "true" modulus of rock as determined by other means.
EX size borehole compared with the circumfe
rential strain at rsg = 17.5 nun, EI' = 69 GPa, In practice, these options give very similar
VI' = 0.25, Epl = 3.5 GPa, Vpl = 0.4. results (usually up to 5\ higher), to those obtained
using the effective modulus of rock from biaxial
tests without K factor corrections.
(bl Computation of Rock Stresses from Measured
Strains
4
TABLE I I
Test resula using 0.20 -0.11 20.00 0.09 -0.44 -0.89 20.0 154· 87· 0. 1 78· -0.7· -0.2 168· -3·
K faclm o~on;
Et=71.0 Pa Std. error 0.40 0. 22 0 . 36 0,17 0.14 0.22
Theoretical,
Infinitely loug 0 23.44 23.44 0 0 0 23.44 270" 23.44 270" 0 0 180" 0"
cylinder
Test .esula Uling -1.57 22.78 23.21 -0.32 -0.04 0.31 23.24 262 0 -84 0 22.8 269 0 5.7 -1,6 1790 0.7 0
leff =65.4 GPa and
adjuRing K faCka Std. error 1.07 0.48 O. &l 0.38 0.32 0.47
of the biaxial pressure jacket (Ref. 1) while the Table II shows the six stress components and the
output of the gauge was monitored (Fig. 7). principal stresses computed from the test results at
the applied pressure of 23.4 MPa. The X direction
Values for the elastic modulus and Poisson's is coincident wi th the axis of the EX hole. The
ratio of the aluminium calculated from the trJee modulus of 65.4 GPa was used, and the stress-strain
circumferential and two axial strain gauges were coefficients at the axial a~ and shear stresses Tzx
E = 65.4 MPa and v = 0.36. The high value of the and Txy adjusted as discussed above.
Poisson's ratio could possibly be attributed to an
axial tensile stress which develops in cores pres (ii) verification of the method in situ
surised over a part of their length only. The
Young's modulus value of 65.4 GPa was about S% lower Verification of the method in situ was obtained
than the "true" value of 71 MPa for the material, by carrying out parallel tests using CSIRO and USBM
due to the gap between the strain gauges and the borehole deformeter gauges and by assessing and com
surface of the EX hole. paring the results obtained in the same area by
different CSIRO gauges.
-§~.--------------------------------.
jg
o
0 -
z ...
;;;:
a:
.... 0
"'0
~pt\
z: :Eo
C
...'"
<It
2
~
<.>
%
~
~
g
~ 8"'"
-Vi
c:
.!c: !8':----~
.... ....0 S
. 10
BIAXIAL
0 5 10 15 20 25 30 35
STRESS (HPa) ....
Fig. 6. Test 1: Strain as sensed by gauge v. uniaxial Fig. 7. Test 2: Strain as sensed by gauge v. biaxial
pressure applied to aluminium block pressure applied to aluminium cylinder
5
TABLE III The following composition was found satisfactory -
STRESS COMFONENTS AND STANDARD ERRORS ASSOCIATED WITH
Araldite resin LC230/Hardener LC230/Araldite 0
PRINCIPAL STRESSES SHOWN IN FIGS. 8 and 9 resin/Hardener HY956 in proportions by weight
100:60:68 :1 2 .
Stress Components (MPa)
Method an Ow av Tnw Twv Tvn
This mix gives a bond strength to wet rock of
Hollow inclusion gauge 03 22 45 17 5 -11 -12 8 - 20 MFa; has a pot life of about 25 min. at 25°C,
05 29 47 14 5 -11 -3
reducing to 10 mins. at 40 o C; extrudes through
06 31 38 8 11 -9 -2
03, OS, 06 combined 29 43 16 7 -9 -7 thin sections without being excessively fluid, and
USBM 23 58 14 -8 -6 -2 polymerises within 12 hours.
USBM and hollow inclusion 21 41 15 3 -8 -4
gaUl_ 03, OS, 06 combined
(a) Adhesives
About 15 different epoxy and polyester adhesive
.ystems were tested. The best results for bond
strength to wet rock, viscosity, pot life and the
ability to polymerise in thin sections, were obtained
by combining the filled epoxy resin, Araldite LC230 Fig. 9. Comparison of principal stresses obtained
(which cures at room temperature) with Araldite 0 from gauges 03,05 & 06 combined (SI) , USBM
resin (manufactured by CIBA-Geigy Australia Ltd.). tests in holes A, B & C (US) and the
results from the three gauges and USBM
tests combined (US & SI) •
w4---+--+--+-~~-+-+~
Fig. 8. Comparison of principal stresses obtained Fig. 10 . Comparison of principal stresses obtained
from gauges 03,05 & 06. Gauges 05 & 06 from gauges 08,11,14 & 16. Gauges 08,11 &
installed in hole B, gauge 03 in hole C. 16 installed in hole A, gauge 14 in hole B.
6
To facil i tate h andling in the field, the two (c) Manpowe r and time
monomer resins, LC230 and Araldite D a re premixed
in the laboratory , as are a lso both hardeners, and Four men are required on-site to install and
transported in two sealed cont aine rs. When needed, overcore a gauge; a driller and assistant plus a
the resins are poured into the hardene rs and two-man crew to prepare, install and r ead the gauge,
stirred vigorously for 3 - 5 minutes before pouring and to carry out the biaxial test. Experience has
into the cement chamber of the stress gauge. shown that in good rock, it is possible to install
and overcore the USBM borehole deformeter in the
(b) Installation Procedure and Equipment first half of a shift, then install a CSlRO gauge in
the second half. The CSlRO gauge can then be over
The installation procedure and equipment are cored the following morning and another USBM test
for the most part the same as for the USBM borehole carried out and/or a second CSIRO cell installed
deformeter, and in most respects follow USBM towards the end of that shift. As another example,
practice for the latter gauge (Ref. 1). with the help of a competent drilling crew, six
CSIRO gauges were installed (in two holes) and over
AlSO mm dia. hole is drilled to the location cored in five, eight-hour shifts. All installations
at which it is desired to measure stresses, using , were successful . No USBM gauges were used at this
thin walled core barrels of length 600 mm. A site.
38 mm dia. pilot hole (usually 450 mm long) is
drilled on through the test position. The gauge Parallel use of USBM and hollow inclusion gauges
is installed normally at 200 to 300 mm from the in the same hole permits twice as much information
kerf of the 150 mm dia. hole in an area free of to be obtained with less than 50% increase in time
jointing. and cost compared with one method only.
7
TABLE IV
1975
Dolpb1n Mine, King
Wand, Tumania
1975 Mt. lIa Mine,
Queendand
Mt. I'a Mine,
Queendand
75 level
Development drive
CroWD pillar of cut-
and-fill nope
19 level
Development drive
75
75
495
975
Pyaxan gamet _
akarn & marble
Argilllte
Argilllte
1_
CIOlely jointed
Severe diacing due
to blgb
Moderate.!y jointed
16
16
30
48
3
5
100
28
3
29 23
2