Вы находитесь на странице: 1из 6

Australian Geothermal Energy Conference 2011

287
Wellbore Temperature and Thermo-elastic Stress Analyses During
Drilling or Stimulation
Bisheng Wu
*
, Bailin Wu, Xi Zhang and Rob Jeffrey
*
Bisheng.wu@csiro.au
CSIRO Earth Science and Resource Engineering
Ian Wark Lab, Clayton, VIC 3168, Australia

In this paper, stress analyses around the bottom
of a vertical wellbore during drilling were carried
out using a fully coupled thermo-elastic wellbore
model. To account for temperature evolution
during drilling, a fluid circulation model was
developed to track the fluid temperature in drill
string and annulus. The combination of these two
models provides the information required for
wellbore stability analyses. In particular, numerical
methods that were implemented in the models
solving the respective problems are given. A
special formula for the heat transfer coefficients
between fluids and surrounding solid materials,
which is dependent on the flow behaviour and
material properties, is chosen. The predicted
temperature change is incorporated into the
wellbore stability model to calculate the stress
changes. The results show that the drilling fluid
circulation rates have a considerable effect on the
borehole bottom-hole temperature; the choice of
the model for calculating the heat transfer
coefficient has a great effect on the prediction of
the wellbore temperature. In addition, the thermal
effect on the near-wellbore stresses is important.
The cooling effect can not only reduce the
compressive radial and hoop stresses, but also
the vertical stresses. This stress reduction can
have an impact on hydraulic fracture initiation in
the rock formation.
Keywords: Wellbore temperature prediction, fluid
circulation, analytical solutions, stress change,
drilling
Introduction
Nowadays it is common to drill high temperature
and high pressure wells deep into the earths
crust, for instance, to develop enhanced
geothermal systems (EGS). As the depth of the
wellbore increases, it becomes increasingly
difficult to maintain the stability of the well due to
the high in-situ stresses encountered at depth. In
addition to variation of drilling fluid pressure along
the well, the wellbore temperature can change
due to the circulation of the drilling mud and the
heat exchange with the host rock. These changes
in pressure and temperature will affect
significantly the thermo-elastic stress distributions
around the wellbore. Under certain
circumstances, rock failure can take place
surrounding the wellbore, which can lead to
wellbore collapse. It is therefore of great
importance to accurately evaluate the bottom-hole
temperature and the coupled thermo-elastic effect
on wellbore stability during drilling.
A great deal of research on the temperature
distribution due to drilling fluid circulation in a well
has been carried out over the last several
decades (Raymond 1969; Holmes & Swift 1970;
Sump and Williams 1973 and Fomin et al. 2005).
For example, Raymond (1969) used an explicit
finite difference method to numerically solve this
problem, but he did not provide a stability and
convergence analysis of the proposed numerical
method. Fomin et al. (2005) studied the borehole
temperature during drilling in a fractured rock
formation, but gave only the steady state solutions
to the formulated problem. In our recent work (Wu
et al. 2010), the problem similar to the cases
studied in these two papers is re-visited with an
objective of obtaining the fully-coupled transient
solution. In particular, the plane-strain condition is
used to simplify the problem as most models do.
The analytical solutions are obtained in Laplace
space and then the Stehfest inverse Laplace
transformation is applied, which has been shown
to give accurate results as a function of time.
This paper describes a solution method to
incorporate the temperature prediction model for
circulating fluid inside the well with the stress
prediction model for thermo-elastic solids. A cold
fluid (or drilling fluid) is circulated into a tubing
string and back up the annulus between the
tubing and the borehole wall. The heat exchange
between fluids and solids will generate a time-
dependent temperature profile along the well. The
corresponding temperature change and rock
deformation will be calculated by the proposed
model.
Problem formulation
In Fig.1, a system for circulating fluids in the
wellbore is illustrated. At the beginning (t=0), the
fluid and rock temperatures are in equilibrium with
the linearly varying temperature profile with depth
shown on the right of Fig. 1. We refer to this
temperature distribution as static. At t>0, the fluid
with a given surface temperature T
in
*
, is pumped
into the drill string at an average fluid mass rate
G
d
, The corresponding volumetric flux is equal to
G
d
/
l
in which
l
is the liquid density. When the
fluid mass reaches the bottom of the tubing string,
it will return to the surface along the annulus.
Although the fluid mass rate can, in general, be
variable along the annulus, it is assumed to be
http://www.ga.gov.au/image_cache/GA20085.pdf
Australian Geothermal Energy Conference 2011
288
constant for this analysis. The return mass flow
rate is denoted as G
a
, which can be less than the
mass injection rate after considering fluid loss.
The tubing wall thickness is denoted as
0
, which
cannot be zero because, otherwise, the fluid
temperature in the tube and in the annulus will be
equal at all times.

Figure 1 Injection system considered by the model.
The heat transfer along the well is a complicated
process, which includes the heat advection in the
tubing string, the heat exchange between the fluid
and the interior or outside surfaces of the tubing
string, and the heat exchange with the formation
rock. Here the temperatures of the fluid inside the
annulus, the tubing and the rock are denoted as
T
a
*
, T
d
*
and T
r
*
, with the first two of these
independent of the radial position, r
*
. The
temperature on the borehole wall is denoted as
T
w
*
. The initial static temperature distribution for
the rock is assumed to be linear with depth as
shown in Fig. 1.
Note that the z
*
position coordinate is pointing
down so that z
*
is zero at the surface and is
positive and increases with depth in the model.
On the other hand, the fluid pressure at the
bottom of the borehole can be assessed by the
fluid density and the height as
l
gH in which g is
the gravitational acceleration. For the current
version of the model, we ignore the details of
viscous friction in the drill string and annulus.
We have provided a plane-strain model for stress
analysis on the cross-section of a vertical wellbore
(Wu et al. 2010). Basically the model provides the
fully coupled solution to the stress, pore pressure
and temperature in the thermo-elastic rock around
the wellbore. A linear thermo-poro-elastic
constitutive law is used. When the pressure and
temperature are prescribed on the well wall and
the initial conditions are stationary, the solutions
can be obtained analytically in Laplace space.
The reader is referred to our previous papers (Wu
et al. 2010) for more details. In the present paper,
we only use the thermo-elastic results.
In next section, we will provide the governing
equations and initial and boundary conditions for
the fluid circulation part only.
Governing equations and boundary
conditions for fluid circulation model
The temperature of the fluid in the tubing is
determined by the rate of heat convection down
along the tubing and the rate of heat exchange
between the tubing and the annulus. The
temperature of the fluid in the annulus is
determined by the rate of heat convection up
along the annulus, the rate of heat exchange
between the tubing and the annulus, and the rate
of heat exchange between the fluid and the
surrounding rock. According to Raymond (1969),
the convective heat transfer equations for fluids
inside the tubing string and the annulus are as
follows
* *
* *
0 * *
* *
* * * *
0 * *
2 ( ) ,
2 ( ) 2 ( ) ,
ad d d
l l d d d a l d l
ad w a a
l l a a d a w i w a l a l
T T
c A v r h T T A c
z t
T T
c A v r h T T r h T T A c
z t
t
t t
c c
+ =
c c
c c
+ + =
c c

(1)
The physical parameters in the above equations
are assumed to be independent of temperature
and pressure. The parameters definition and
typical values for drilling fluid and granite are
listed in Table 1.
For the surrounding rock formation, because the
temperature gradient in the radial direction is
much greater than that in the vertical direction in
the near wellbore region, the derivative of the
formation temperature with respect to z
*
can be
ignored (Raymond 1969). Then the governing
equation for heat transfer follows the classical 2D
heat conduction equation
*
*
* * * *
1
r r r
r r
T k T
r
t c r r r
c c c | |
=
|
c c c
\ .
. (2)
At the borehole wall, the heat flux from the rock
formation into the annulus satisfies Fouriers law.
That is, the heat exchange between the rock
formation and the annulus fluid is expressed as
*
* *
*
2 ( ) 2 at ,
r
w r r a w r w
T
r h T T r k r r
r
t t
c
= =
c
(3)
On the other hand, heat transfer between the
fluids in the tubing and in the annulus, has been
considered in Eq. (1).In this equation we use the
combined heat transfer coefficient h
ad
to include
the effect of the tubing thickness.
To solve the present problem, the boundary
conditions are needed at the surface and the well
bottom. These boundary conditions are given as
* * * * *
* * * * * * * *
( , ) , when 0,
( , ) ( , ), when ,
d in
d a
T z t T z
T z t T z t z H
= =
= =
(4)
and the initial condition which is that at t=0, the
formation temperature is given as
* * * * *
0 0 0
( , ,0) ,
r
T z r T A z B = = +

(5)

Under static condition, well will
reach a thermal equilibrium --
radius of investigation = depth of
invasion
Australian Geothermal Energy Conference 2011
289
where A
0
and B
0
are the constants which
determine the initial static formation temperature
as a function of depth.
In summary, the physical parameters defined
above are given below
0
0.8 0.43 0.25
0 0
0.8 0.43 0.25
0 0
0.8 0.43 0.25
0 0
0
1 1 1
,
0.021(Re ) (Pr ) (Pr / Pr ) / (2( ) ),
0.021(Re ) (Pr ) (Pr / Pr ) / (2 ),
0.021(Re ) (Pr ) (Pr / Pr ) / (2( ) ),
2( )
Re , Re
ad
a d d
a a l l d l w
d d l l d l
w a l l w l w
l a w
a
h h h k
h k r r
h k r
h k r r
v r r
o
o
o
o

= + +
=
=
=

=
0
2 2 2
0 0
2
,
Pr , Pr , Pr ,
, , ,
, , ( ), .
l d
d
d l w
d d l l r r
l d r
l d r
l l d d r r
a d
a d a w d
a l d l
v r
d d d
k k k
d d d
c c c
G G
v v A r r A r
A A




t t

=
= = =
= = =
= = = =

It is clear from the above definition that the heat
transfer between fluids and the surrounding solids
is dependent on the Reynolds number and the
Prandtl number, in addition to the well and drill
string radii (Fomin et al., 2005). However, these
empirical correlations are only valid when the
Reynolds number is larger than 10,000. In the
model, for Reynolds number less than 10,000, we
use an approximation method. If Reynolds
number is less than 2000, the fluid flow is laminar
and the heat transfer coefficient is 3.66 W/(m
2
K)
and the borehole wall heat flux tends to become
stable soon after circulation begins. If Reynolds
number is within the range 2000 to 10,000, the
heat transfer coefficient can be obtained by an
interpolation method.
Methods for solution
Here we only provide the solution method for
temperature changes along the well. The reader
is referred to our previous work (Wu et al. 2010)
for the solution method for stress analysis of the
wellbore. The main approach for solving the fluid
circulation problem is to use Laplace
transformation to convert the partial differential
equations (PDEs) into ordinary differential
equations (ODEs). Then the analytical solutions
are obtained by solving the ODEs based on the
transformed boundary and initial conditions. The
numerical values are calculated by using the
Stehfest inverse Laplace transformation. The
details for numerical inversion can be found in
(Wu et al. 2011).
To validate the numerical inversion method, the
results from the present numerical model have
been compared to the experimental
measurements and to simple cases such as no
drill string present and an infinitely long wellbore.
A good agreement is found.
Table 1 Parameters for the calculation
Parameter Value
Pipe internal radius r
0
(m) 0.0462
Pipe thickness
0
(m) 0.01
Wellbore radius r
w
(m) 0.1
Wellbore length H
0
(m) 4132
Injection rate G
d
(Kg/s) 13
Pump out rate G
a
(Kg/s) 13
Injection temperature T
in
*
(

C) 27
Temperature gradient A
0

(

C/m) 0.0467
Surface temperature B
0
(

C) 27
Initial temperature T
0
*
(

C) A
0
z
*
+B
0
Fluid specific heat c
l
(J /(kgK)) 4200
Rock specific heat c
r
(J /(kgK)) 790
Pipe specific heat c
d
(J /(kgK)) 460
Fluid thermal conductivity k
l
(W/(mK)) 0.68
Rock ther. conductivity k
r
(W/(mK)) 2.2
Pipe ther. conductivity k
d
(W/(mK)) 50
Fluid mass density
l
(Kg/m
3
) 900
Rock mass density
r
(Kg/m
3
) 2700
Pipe mass density
d
(Kg/m
3
) 7800
Fluid viscosity (Pas)
0.0004
Numerical results
Bottom-hole and outlet temperature
First of all, let us consider the temperature
changes along the well. Figs. 2 and 3 display the
time dependent response of bottom-hole and
surface outlet temperatures, respectively for
several mass flow injection rates and fluid mass
densities. Based on the given parameters, the
rock temperature at the wellbore wall and annulus
fluid temperature are identical along the well for
these cases. It is clear from these figures that with
increasing the injection flow rate, the bottom-hole
temperature will drop more quickly to a lower
level. The temperature continues to decrease
after 1000 hours, but at an ever slower rate.
Therefore, a significant portion of the cooling
occurs in the first 100 hours or less and higher
rates are required to produce significant cooling
effect on wellbore temperature.
The outlet temperature can be easily measured in
the field. Such a measurement would provide
useful data for model verification presented here.
It is found that after about 2 hours, the outlet
temperature is nearly stable. Increasing the
injection rate results in an increase in the outlet
temperature at the annulus. This is largely due to
more heat is carried away from the deep portion
of the wellbore.
In addition, changing the fluid densities does not
have a great effect on the bottom-hole and outlet
temperatures as shown Fig. 2(b) and Fig. 3(b). It
should be noted that the mass injection rate is
used in the model.
Australian Geothermal Energy Conference 2011
290
Tubing and annulus temperature
Figure 4 shows the temperature distribution along
the well at a specific time when G
d
= G
a
=13 kg/s.
The temperature continuity between the tube and
annulus fluid temperature at the bottom of the well
is satisfied as expected. The red line is the
reference line, giving the initial rock static
formation temperature. The two curves for the
annulus fluid temperature and borehole wall
temperature are identical. This means the heat
loss from the formation to the annulus is
extremely small. On the other hand, the heat loss
across the steel pipe is clear although the
difference is not significant after 150 hrs injection.
Figures 5 and 6 show the temperature
distributions in the tubing and in the annulus at
different times. The drop of bottom-hole
temperature with time elapsed is clear. Of course,
the trends for two temperature curves are the
same. Significant temperature drop occurs at
larger depth. After 200hrs circulation, the bottom-
hole temperature can be reduced by 50 degrees.
The cooling will definitely change the local stress
distribution due to formation shrinkage.

Time (hours)
0 200 400 600 800 1000 1200 1400
T
e
m
p
e
r
a
t
u
r
e

(
0

C
)
140
160
180
200
220
G
d
=
G
a
=1 Kg/m
3
G
d
=
G
a
=7 Kg/m
3
G
d
=
G
a
=Kg/m
3
G
d
=
G
a
==19 Kg/m
3

(a)
Time (hours)
0 10 20 30 40 50
T
e
m
p
e
r
a
t
u
r
e

(
0

C
)
170
180
190
200
210
220
l
=900 kg/m
3

l
=1300 kg/m
3

l
=1700 kg/m
3

(b)
Figure 2 Bottom-hole temperature responses when
Tin
*
=B0=27 C. (a) Varying only the injection rates with the
fluid density l=900 kg/m
3
, (b) varying only the fluid densities
with the injection rate Gd=Ga=13 kg/s.
Time (hours)
0 20 40 60 80 100 120 140
T
e
m
p
e
r
a
t
u
r
e

(
0

C
)
26
28
30
32
34
G
d
=
G
a
=1 Kg/m
3
G
d
=
G
a
=7 Kg/m
3
G
d
=
G
a
=13 Kg/m
3
G
d
=
G
a
=19 Kg/m
3

(a)
Time (hours)
0 20 40 60 80 100 120 140
T
e
m
p
e
r
a
t
u
r
e

(
0

C
)
26
28
30
32
34

l
=900 kg/m
3

l
=1300 kg/m
3

l
=1700 kg/m
3

(b)
Figure 3 Outlet temperature responses when Tin
*
=B0=27 C.
(a) Varying only the injection rates with the fluid density
l=900 kg/m
3
, (b) varying only the fluid densities with the
injection rate Gd=Ga=13 kg/s.
Temperature (
0
C)
0 50 100 150 200
D
e
p
t
h

(
1
-
z
/
H
0
)
0.0
0.2
0.4
0.6
0.8
1.0
Tubing
Annulus
Borehole wall
Formation

Figure 4 Temperature profiles in the drill string, annulus and
borehole wall after 150 hours of circulation time when
Tin
*
=B0=27 C and Gd= Ga=13 kg/s.
Australian Geothermal Energy Conference 2011
291
Temperature (
0
C)
0 50 100 150 200
D
e
p
t
h

(
1
-
z
*
/
H
0
)
0.0
0.2
0.4
0.6
0.8
1.0
2 hours
20 hours
200 hours
1000 hours
Formation

Figure 5 Temperature in the drill string as a function of time
when Tin
*
=B0=27 C and Gd= Ga=13 kg/s.
Temperature (
0
C)
0 50 100 150 200
D
e
p
t
h

(
1
-
z
*
/
H
0
)
0.0
0.2
0.4
0.6
0.8
1.0
2 hours
20 hours
200 hours
1000 hours
Formation

Figure 6 Temperature in the annulus as a function of time
when Tin
*
=B0=27 C and Gd= Ga=13 kg/s.
Temperature in the rock
Figure 7 shows the temperature contour around
the circular well. It can be seen that the near-well
region at the bottom of the well is greatly
impacted by the circulation. The low temperature
area near the bottom of the well can extend into
the formation to a depth of one wellbore radius
after 150 hrs. This cooled zone will induce
thermo-elastic stresses that may be useful in
helping initiate fractures and to enhance the
permeability of certain natural fracture sets during
stimulation.
Thermal effects on the stress changes
The temperature changes from the fluid
circulation model are introduced in the stress
analysis model as a boundary condition. A typical
example is demonstrated here to show the
thermal effect on the stress change. For values of
the parameters for the thermo-elastic rock mass,
see Wu et al. (2010).
In the case where the wellbore bottom hole
temperature is reduced to 140 C, the radial and
hoop stresses along the
max
directions (=0) are
displayed in Fig. 8 for different elapsed times. In
the real cases, the temperature drop at the bottom
is reduced progressively, and then the stress drop
caused by accumulated heat exchange would be
larger than the predictions given here. Therefore,
the stress drop provided here is conservative.
In particular, from Fig. 8, the radial stress can be
reduced up to 5 MPa and the hoop stress to 8
MPa at the location of twice the wellbore radius
from the wellbore centre. This stress drop would
provide another boundary condition for hydraulic
fracture growth from the wellbore wall. The near-
well fracture growth can be investigated
accordingly.
Figure 7 Temperature contours in the surrounding rock at
t=150 hours when Tin
*
=B0=27 C and Gd= Ga==13 kg/s.
r (m)
0.2 0.4 0.6 0.8 1.0
o
r
r
(
M
P
a
)
-160
-140
-120
-100
-80
t=10 s, THM
t=16.7 mins, THM
t=2.77 hrs, THM
t=3.47 days, THM
Elastic
r (m)
0.2 0.4 0.6 0.8 1.0
o
u
u
(
M
P
a
)
-130
-120
-110
-100
-90
t=10 s, THM
t=16.7 mins, THM
t=2.77 hrs, THM
t=3.47 days, THM
Elastic
Figure 8 Radial and hoop stresses in the max direction (=0)
for Tw*=140 C (negative stress is taken as compression) (a
top, b bottom).
Australian Geothermal Energy Conference 2011
292
In addition, the thermal effects have a
considerable effect on the vertical stress as
shown in Fig.9. The stress reduction in the region
adjacent to the wellbore wall along the
max

direction can reach a maximum of 20 MPa (less
compressive) when the wellbore temperature is
reduced to 140 C. The lower the vertical stress,
the more likely the creation of horizontal tensile
fractures.
Conclusions
In this paper, the solutions for the wellbore
temperature changes during fluid circulation are
obtained and are introduced as a boundary
condition for near-well stress analyses. Then the
stress changes due to the cooling effects are
investigated so that the wellbore stability caused
by temperature variations during drilling or
stimulation can be considered. Some conclusions
are drawn as follows:
1. The borehole bottom temperature change is
sensitive to the injection rate used, especially
at the early time. Higher rates can produce
significantly more cooling effect.
2. The surface outlet temperature will reach a
steady state after several hours. The outlet
temperature can be easily measured in the
field and such data would allow an inversion
method to be used to obtain heat transfer
coefficients.
3. More importantly, the cooling can cause
significant change in thermo-elastic stresses
around the wellbore. The reduction in near-
well stresses can facilitate hydraulic fracturing.
References
Raymond, L.R., 1969, Temperature distribution in
a circulating drilling fluid, SPE-AIME.
Fomin, S., Hashida, T., Chugunov V. and
Kuznetsov, A.V., 2005, A borehole temperature
during drilling in a fractured rock formation,
International Journal of Heat and Mass Transfer
48, 385394.
Sump, G.D. and Williams, B.B., 1973, Prediction
of wellbore temperature during mud circulation
and cementing operations, Journal of Engineering
for Industry, 1083-1092.
Holmes, C.S. and Swift, S.C., 1970, Calculation of
circulating mud temperatures, Journal of
Petroleum Technology, 670-674.
Wu, B., Zhang, X. and Jeffrey, R. G., 2010. A
Thermo-poro-elastic analysis of stress fields
around a borehole. Paper ARMA 10-442
presented at the 44
th
US Rock Mechanics
Symposium and 5
th
US-Canada Rock Mechanics
Symposium held in Salt Lake City, UT June 27-
30, 2010.




Figure 9 Vertical stress contours at different wellbore temperatures. Left Tw*=220 C at (a) 3 seconds and (b) 3.47 days. Right
Tw*=140 C at (a) 3 seconds and (b) 3.47 days.

Вам также может понравиться