Mihály Dobróka

Geophysical Department,
University of Miskolc, Hungary

ROCK STRESS DETERMINATION IN BOREHOLES

WITHOUT OVERCORING

The paper deals with the simultaneous determination of five independent stress
components from the data measured in a single borehole. The proposed technique
consists of cementing a solid cylindrical epoxy probe in which strain gauges have been
embedded into a hole predrilled into the rock mass. Using the rheological equations of
the Poynting-Thomson body analytical solutions are given for the displacement and
strain field of the probe and of the rock. By measuring the time history of the
deformations of the probe one can determine stress components occurring in the rock
continuum without overcoring of the borehole.
1. INTRODUCTION
The stress-relief technique of in situ stress measurements by overcoring is well
known. The first method of this type developed by Leeman (1968) uses rosette strain
gauges directly cemented on the surface of the borehole. In order to overcome the strain
gauge waterproofing problems occurring in wet conditions Rocha and Silverio (1969)
developed the solid inclusion technique, using strain gauges embedded in a cylindrical
epoxy probe. A similar «soft-inclusion» device with an elastic modulus less than that of
the host material was also utilized by Blackwood (1977). In order to reduce the bond
stresses occurring at the interface between the rock and the probe a new technique (the
so-called hollow inclusion) was developed by Rocha et al. (1974) and Worotnicki and
Walton (1976) using a tubular probe with embedded gauges. The complete
displacement and stress field of the probe and that of the rock mass were given by Savin
(1951), Duncan-Fama and Pender (1980).
As a material equation, in all the works mentioned Hooke’s law was used. This results
in the necessity for overcoring. The steps of the rock stress measurements are shown in
Fig. 1. In consequence of the overcoring the rock core expands together with the
cemented probe, producing tensile stresses at the epoxy-rock interface, which may be
sufficently large to break the bond between the rock and the epoxy probe. This is the
case even for the hollow probe, when the rock material is «soft» (Duncan-Fama and
Pender 1980). There is also another disadvantage in connection with overcoring: it can
cause cracks and failures in the investigated rock annulus resulting in erroneous
measurements.
In what follows we propose a new method based on the rheological equation of
the Poynting-Thomson body. The rheological features of the rock continuum result in
time dependent displacements and strains in it after the borehole has been drilled. On
measuring these straings at various times, the in situ rock stresses can be determined
without overcoring the borehole.
2. OPERATION PRINCIPLES
As is well known, in many cases
Hooke’s law gives only an
approximate description of the real
rock continua. In a wide range of
phenomena the rocks may have time-
dependent properties. In order to take
the observed time-dependence into
account (Kurlenya et al. 1973) used
time-dependent moduli in the
computation of the deformations of the
borehole wall. The same method was
applied by Senuk (1973) who
calculated the stresses appearing in a
photo-elastic probe cemented into the
Fig. 1. Steps of measurements borehole.
I – Drilling a borehole to the neighbourhood There are some rock properties (strain
of the point, where the stress is to be retardation and stress relaxation, etc.)
determined; II – Drilling a coaxial hole with a which require that the time-derivatives
smaller diameter; III – Cementing the probe of the material equation be
with embedded strain gauges into the pilot incorporated. The most simple
hole; IV – Overcoring the rock and observing rheological equation describing
the strain induced in the instrument relaxational and retardational processes
is the Poynting-Thomson model
(standard body) where the relationship
between the stress and deformation deviators Tik and Eik is given as a differential
equation

 ∂  ∂
1 + τ Tik = 1 + ϑ 2GEik (1)
 ∂t   ∂t 

while the spherical part of the stress and of the deformation tensors Tik( 0 ) and Eik( 0 )
satisfies a linear equation

Tik( 0 ) = 3KEik( 0 ) (2)

where G is the shear modulus, τ the relaxation time, ϑ the retardation time, K the
modulus of compression. In slow processes, when ϑ << T (T being the characteristic
time of the process) Eqs. 1, 2 can be written as

Tik = 2GEik and Tik( 0 ) = 3KEik( 0 )

while in quick processes, when τ >> T

Tik = 2G * Eik and Tik( 0 ) = 3KEik( 0 )

 ϑ
where G* = G is the dynamic shear modulus. In these two limiting cases the
τ
Poynting-Thomson body can be treated as a linearly elastic one. In the model the
relation ϑ ≥ τ is fulfilled, so in our considerations G* ≥ G .
Since the values of the rheological times, ϑ , τ are of the order of some ten hours, the
drilling of a borehole can be regarded as a quick process. During the drilling process,
1
strains occur Eik (0) = Tik in the host rock, which cannot be measured directly.
2G *
Using Leeman’s method, or cementing a soft hollow probe into the bore-hole, the stress
state of the rock mass can be approximately considered as independent of time: T&ik = 0
(the dot denoted differentiation with respect to the time). By Eq. 1 the strains can be
written as

1   τ 
Eik = Tik 1 − 1 − e −t /ϑ  . (3)
2G   ϑ  
The measurable quantity

1  τ
Eik − Eik (0) = Tik  (1 − (1 − e −t /ϑ )
2G  ϑ

depends on the in situ stresses and – through its time dependence – gives a possibility to
determine the rheological parameter ϑ .
It will be shown later that in real situations, when an elastic probe is used, we can
measure four independent time parameters characteristic of the interaction between the
probe and the rock media, and five independent strains of the probe. By means of these
quantities we can determine five independent elements of the in situ stress tensor and
the material parameters G, K, ϑ , τ as well.
3. ANALYTICAL SOLUTION FOR THE DISPLACEMENT FIELDS
The analytical solution for displacements and strains is required in order to relate the
strains measured in the probe to the stresses occuring in the rock. The results referring
to the linear elasticity are also given in Savin (1951) and Duncan-Fama et al. (1980).
The solutions for the case when the rock material has rheological properties determined
by the Poynting-Thomson model can differ only in their time dependence.
The in-situ rock stresses can be written in cylindrical coordinate system, as

1 ( 0) 1
σ r( 0 ) = (σ x + σ y( 0 ) ) + (σ x( 0 ) + σ y( 0 ) ) cos 2θ + τ xy( 0 ) sin 2θ ;
2 2

1 ( 0) 1
σ θ( 0 ) = (σ x + σ y( 0 ) ) + (σ x( 0 ) + σ y( 0 ) ) cos 2θ + τ xy( 0 ) sin 2θ ;
2 2

1
τ r(θ0 ) = − (σ x( 0 ) − σ y( 0 ) ) sin 2θ + τ xy( 0 ) sin 2θ ;
2

τ rz( 0 ) = τ yz( 0 ) sin θ + τ xz( 0 ) cos θ ;

τ θ( 0z ) = τ yz( 0 ) cos θ + τ xz( 0 ) sin θ ;

σ z( 0 ) = σ z( 0 ) ,

where σ x( 0 ) , σ y( 0 ) , σ z( 0 ) , τ xy( 0 ) ,τ xy( 0 ) , τ yz( 0 ) are the stresses in a Cartesian coordinate system.
While drilling a borehole and cementing a probe to it, there occur changes in the stress
state determined by the equilibrium equations

∂σ r 1 ∂τ rθ σ r − σ θ
+ + = 0;
∂r r ∂θ r

∂ 1 ∂σ θ 2τ rθ
+ + =0;
∂r r ∂θ r

∂τ zr 1 ∂τ rθ τ rz
+ + = 0. ,
∂r r ∂θ r

where σ r , σ θ , τ rθ , τ rz , τ θ z are total stresses in the rock. The material equations can be
written in the form

 νθ  νµθ& 
σ r* + τσ& r* = 2G er + + ϑ  e&r +   ;
 1 − 2ν  1 − 2ν 

 νθ  νµθ& 
σ θ* + τσ& θ* = 2G eθ + + ϑ  e&θ +   ;
 1 − 2ν  1 − 2ν 

 νθ  νµθ& 
σ z* + τσ& z* = 2G ez + + ϑ  e& z +   ;
 1 − 2ν  1 − 2ν 

τ r*θ + τ τ&r*θ = 2G{erθ + ϑe&rθ };

τ rz* + τ τ&rz* = 2G{erz + ϑ e&rz };

τ θ*z + τ τ&θ* z = 2G{eθ z + ϑ e&θ z },

where σ r*, σ θ* , σ z* , τ r*θ ,τ rz* , τ θ*z are stresses appearing in consequence of the borehole
opening, ν is Poisson’s ratio,

∂u 1 ∂υ  ∂w
er = , eθ =  u + , ez = ;
∂r r ∂θ  ∂z
 1  ∂u ∂υ  1 ∂w
erθ =  −υ + r  , eθ z = ;
2r  ∂θ ∂r  2 ∂θ

1 ∂w
erz = , θ = er + eθ + ez
2 ∂r

and u( r, θ , t ), υ ( r, θ , t ), w( r, θ , t ) are the displacements.

The time of the drilling of the borehole is very small in comparison with the rheological
time parameters, so while drilling, the rock can be regarded as a linearly elastic
continuum with the shear modulus G * . So the displacements of the rock are

R  1 ( 0 ) ( 0) R
 * R  R    1 ( 0)
3
 
u =  (σ x + σ y ) + ( χ + 1) −     (σ x − σ y cos 2θ + τ xy sin 2θ   ;
* (0)

2G *  2 r  r  r    2  

R  * R  R    1 ( 0 )
3

υ =
*
( χ − 1) +     (σ x − σ y cos 2θ + τ xy sin 2θ  ;
( 0) ( 0)

2G *  r  r    2 

w* =
R
2G *
{
τ xz( 0 ) cos θ + τ yz( 0 ) sin θ } ,
R
r
χ * = 3 − 4v * .

Similarly assuming that the setting time of the cement for fixing the probe into the
borehole is sufficiently smaller than the relaxation time τ , the initial conditions can be
taken as

u II (0) = u * υ II (0) = υ * w II (0) = w*

for the rock and

u I (0) = u(0) υ I (0) = υ (0) w I (0) = w(0)

for the probe at t = 0 . The boundary conditions – neglecting the deformations of the
thin cement layer between the probe and borehole surface – can be written in the form

u I = u II − u * υ I = υ II − υ * w I = w II − w* ;

σ rI = σ rII τ rIθ = τ rIIθ τ rzI = τ rzII

at r = R , R being the radius of the borehole and

σ rII = σ r( 0 ) + σ r* τ rIIθ = τ r(θ0 ) + σ r*θ τ rzII = τ rz( 0 ) + τ rz* ;

or using the material equations one finds

 vθ II  vθ II    I vθ I  I v1θ I  
σ r( 0 ) + 2G erII + + ϑ e&rII +  = 2G e
1 r + + τ 
 r 1 − 2v   ;
e
& +
 1 − 2v  1 − 2v    1 − 2v1  1 
 { } {
τ (rθ0 ) + 2G e IIrθ + ϑe& IIrθ = 2G 1 e Irθ + τe& Irθ ; }
{ } {
τ (rz0 ) + 2G e IIrz + ϑe& IIrz = 2G 1 e Irz + τe& Irz , }
where G1 , v1 are the elastic parameters of the probe. We seek the solution in a form
similar to that in the Hookean case:

R  r 1 ( 0)  r  r    1 ( 0)
3
 
u = ( χ 1 − 1)a (σ x + σ y ) + b + ( χ 1 − 3)c    (σ x − σ y( 0 ) cos 2θ + τ xy( 0 ) sin 2θ  
I ( 0)

4G1  R2  R  R    2  

R  r
3
r   1 ( 0 ) 
υ =
I
( χ 1 + 3)c  − b   (σ x − σ y cos 2θ + τ xy sin 2θ  ;
( 0) ( 0)

4G1  R R   2 

wI =
R r ( 0)
[
d τ xz cos θ + τ yz( 0 ) sin θ ;
G1 R
]

R  r 1 ( 0 )   R    1 ( 0)  
3
R
u =−
II
 A (σ + σ (0)
) −  ( χ + 1) B + C    (σ − σ ( 0)
cos 2θ + τ ( 0)
sin 2θ  
 r    2
x y x y xy
2G  R 2  r 

R    1 ( 0 )
3
R R 
υ II = ( χ − 1) B − C     (σ x − σ y cos 2θ + τ xy sin 2θ  ;
( 0) ( 0)

2G  r r   2 

w II =
R R ( 0)
G r
[
D τ xz cos θ + τ yz( 0 ) sin θ , ]
where a, b, c, d , A, B, C , D fulfil the set of equations

1 + A + ϑA& = a + τa& ;

b&
{ [ b
1 − 4 B + 3C + ϑ 4 B& (1 − κ ) + 3C& = + τ ;
2 2
]}

b  b& 
1 + 2 B + 3C + ϑ ( 2 B& + 3C& ) = − 3c + τ  − 3c&  ;
2 2 

1 − D − ϑD& = d + τd& ;

 G 
( χ 1 − 1)a = −2ε  A + *  ;
 G 
  G
b + ( χ 1 − 3)c = 2ε ( χ + 1) B + C − χ * *  ;
 G 

 G
( χ 1 + 3)c − b = −2ε ( χ − 1) B − C − χ * *  ;
 G 

 G  G 1  τ
d = ε  D − *  with χ 1 = 3 − 4v1 , ε = 1 , κ = (1 + v )1 −  .
 G  G 3  ϑ

Taking into account the initial conditions we find

ϑ1   τ  −ϑ1 
t

A=  − 1 +  1 −
 ϑ   e ;
ϑ   1  

 − 
t
2ε (ϑ1 − τ )
a = a ∞ 1 − e ϑ1  ; a∞ = ; (4)
  ( χ 1 − 1)ϑ
 

t
−
ϑ2 χ *G 1 − εχB *
B = B p + ( B − B p )e
*
; B* = ; Bp = ;
χG * 1 + εχ

 * G 
 −
t
  1 + εχ * 

b = b∞ 1 − e ϑ2  + 3c ; b∞ = 2εχ  G − B ;
*
(5)
   1 + εχ 
   
 

−
t
 −ϑt −
t

C = − B p − ( B − B p )e
* ϑ2 
+ B p X (e − e
3 ϑ2 ;
 
 

ε ϑ  2 ετ 
1+ − 1 − κ + 
χ 1 ϑ2  3 χ 1ϑ 
X = ;
ε ϑ  ετ 
1+ − 1 + 
χ 1 ϑ2  χ 1ϑ 

 G 
 −t − 
t  1 + εχ * * 
2ε G − B*  ;
c = c'  e ϑ2 − e ϑ3  c' =  (6)
  χ 1  1 + εχ 
   
 

ϑ   τ  − ϑ4 
t

D= 4 1 − 1 − e  ;
ϑ   ϑ4  
  τ 
 −
t
 1+ ε G 
d = d ∞ 1 − e 4ϑ  d∞ = ε  ϑ − , (7)
   1+ ε G* 
   
 

where

τ τ
χ 1 − 1 + 2ε 1 − 2κ + εχ
ϑ1 = ϑ ϑ ϑ =ϑ ϑ;
χ 1 − 1 + 2ε 2
1 + εχ

ε τ τ
1+ 1+ε
χ1 ϑ ϑ.
ϑ3 = ϑ ϑ4 = ϑ (8)
ε 1+ε
1+
χ1

These are the required formulae for describing the displacement field.
4. STRAINS IN THE SOLID PROBE
The active element of the probe is an epoxy cylinder of radius ρ with the usual
arrangement of strain gauges (Fig. 2.). The gauges positioned in the θ = 0 plane
measure the radial strains er (0) and er ( ρ ) at r = 0 and r = ρ , respectively. The strain
gauges positioned around the circumference of the active element with φ = 0 at θ = 0 ,
π 5π π   5π 
θ= ,θ= measure the azimuthal strains eθ (0) , eθ   and eθ   , while those
2 4 2  4 
π 5π
with φ = at positions θ = 0 , θ = measure
4 4

e( 0) =
1
2
[ ]
eθ (0) + γ θ z (0) ,

 5π  1   5π   5π 
e  = eθ   + γ θ z  
 4  2  4   4 
giving the value of the
engineering shear strains γ θ z (0)
 5π 
and γ θ z   , respectively.
 4 
The active element of the probe
is cast into a solid epoxy cylinder
of radius R < ρ . The complete
Fig.2. The active element of the probe instrument is about ten times
longer than its strain-gauged
section; this is to satisfy the plane strain assumption used in the mathematical solution.
Simultaneously solving for the equilibrium equations of the probe and of the rock mass,
the deformation of a solid probe cemented into a borehole can be written as

1  1 ( 0 )   r    1 ( 0)
2
 
er =  (σ + σ ( 0)
)( χ − 1) a +  b + ( χ − 3) c    (σ + σ ( 0)
) cos 2θ + τ ( 0)
sin 2θ  
 R    2
x y 1 1 x y xy
2G1  2  

1  1 ( 0 )   r    1 ( 0)
2
 
eθ =  (σ x + σ ( 0)
y )( χ 1 − 1) a −  b − ( χ 1 + 3) c     (σ x − σ y ) cos 2θ + τ xy sin 2θ  
( 0) ( 0)

2G1  2   R    2  

1   r  2   1 ( 0 ) 
γ rθ = 6c  − b   (σ x − σ y ) sin 2θ + τ xy cos 2θ  ;
( 0) ( 0)

G1   R    2 

γ rz =
d ( 0)
G1
[
τ xz cos θ + τ yz( 0 ) sin θ ; ]

γ θz =
d ( 0)
G1
[
τ yz cos θ − τ yz( 0 ) sin θ ; ]

εz = 0.
5. DETERMINATION OF IN SITU STRESSES
By means of the above formulae one can form at various times the following quantities
from the data measured by strain gauges shown in Fig. 2:

2
χ − 3  ρ  1 ( 0)
e r ( ρ ) − e r ( 0) = 1 3c  (σ x − σ y( 0 ) ) ; (9)
2G1 R 2

2( χ 1 − 1)
er ( ρ ) − eθ ( ρ ) − [er ( ρ ) − er (0)] = χ 1 − 1 a 1 (σ x( 0) − σ y( 0) ) ; (10)
χ1 − 3 2G1 2

2
R
4 −  
er ( ρ ) − eθ ( ρ ) +  ρ  [e ( ρ ) − e (0)] = 1 (b − 3c ) 1 (σ ( 0 ) − σ ( 0 ) ) ; (11)
χ1 − 3 r r
2G1 2
x y

d ( 0)
γ θ z ( 0) = τ yz . (12)
G1
The data given by the left-hand-side of Eq. 10 can be fitted by a function which
is similar to 4 in its time dependence. In this numerical procedure the value of parameter
ϑ1 can be determined. By means of Eqs 9, 11 and 12 together with 5, 6 and 7 the time
parameters ϑ2 , ϑ3 and ϑ4 as well as the time dependent quantities a, b, c, d can be
found in the same way.
 Measuring at a give time t1 the strains eθ and γ θ z at position shown in Fig. 2, the
only unknown quantities appearing in the equations:

4G1eθ (0) = (σ x( 0 ) + σ y( 0 ) ) K1 − (σ x( 0 ) − σ y( 0 ) ) K 2 ;

π 
4G1eθ   = (σ x( 0 ) + σ y( 0 ) ) K1 − (σ x( 0 ) − σ y( 0 ) ) K 2 ;
2

 5π 
4G1eθ   = (σ x( 0 ) + σ y( 0 ) ) K1 − 2τ xy( 0 ) K 2 ;
 4 

G1γ θ z (0) = τ yz( 0 ) K 3 ;

 5π 
G1γ θ z   = −τ xz K 3
( 0)

 4 

with

 − 1 
t

K1 = ( χ 1 − 1)a ∞ 1 − e ϑ1 
;
 
 

 t
 2
 − t1 − 1 
t
 + 3( χ 1 + 1) ρ   e ϑ2 − e ϑ3  c' ;
− 1

K 2 = b∞ 1 − e ϑ2
   R   
  

 t
− 1 

K3 = d ∞ 1 − e ϑ4 
 
 

are the in situ stresses. The determinant of the set of equations

D = −2 K1 K 22 K 3

differs from zero, if t1 > 0 and ε ≠ 0 . Solving the equations we find

G1   π 
σ x( 0 ) = ( K1 + K 2 )eθ (0) + ( K 2 − K1 )eθ  2  ;
K1 K 2   

G1   π 
σ y( 0 ) = ( K1 + K 2 )eθ (0) + ( K 2 − K1 )eθ  2  ;
K1 K 2   

G1   π  K1  5π 
τ xy( 0 ) = eθ  2  + K eθ (0) − eθ  4  .
2K 2    2  
These are the required formulae giving the in situ stress components in terms of strains
measured by the solid epoxy probe cemented into the borehole. The sixth stress
component σ z( 0 ) cannot be measured in this technique, for the probe and the rock mass
around its active part are in a state of plane strain.

REFERENCES

1. Blackwood R.L. 1977: Proc. Int. Symp. of Field Measurement in Rock Mechanics,
Zurich 1, pp. 137-150.
2. Duncan-Fama M.E., Pender M.J. 1980: Int. J. Rock Mech. and Min. Sci. 17, pp.
137-146.
3. Kurlenya M.V., Leonteev A.V., Popov S.N. 1973: Materialy I. Vsesoyuznogo
seminara, chast 1. Pp. 73-77. Novosibirsk
4. Leeman E.R. 1968: Int. J. Rock Mech. and Min. Sci. 5, pp. 31-56.
5. Rocha M., Silverio A. 1969: Geotechnique 19, pp. 116-132.
6. Rocha M., Silverio A., Pedro J.O., Delgado J.S. 1974: Proc. 3rd. Congr. ISRM,
Denver 1. Pp. 464-467.
7. Savin G.N. 1951: Kontsentratsiya napryazhenii okolo otverstii. M.-L. GITTL
8. Senuk D.P. 1973: Materialy 1. Vsesoyuznogo seminara, chast 1, pp. 100-107.
Novosibirsk
9. Worotnicki G., Walton R.J.: Proc. ISRM Symp. on Investigation of stress in rock.
Sydney 1976. pp. 1-8.