Professional Documents
Culture Documents
Aubert/n,
Hassan/& M/tn(eds) 1996Balkerna,
Rotterdam.
ISBN90 5410858X
Modelingof yieldedzoneenlargement
arounda wellbore
Christopher
D. Hawkes
Department
of Geology,
University
ofNewBrunswick,
Fredericton,
N.B.,Canada
Patrick J.McLellan
Advanced
Geotechnology
Inc.,Calgary,
Alb.,Canada
1051
gradient,
i.e. p(r,t),it is possible
to derivethe
10'2reD).Consequently,
porepressure
penetration
following
expression
fortheyielded
zoneradius:
may continue
for severaldays,weeksor months
(])
1-NrPw
+----
1 J,'J
-a'(1-NOR"'-'lrr-"'P(r't)dr=O
beforesteadystateconditions
are achieved.In such
cases,the extentof theyieldedzonewill increase
for
correspondingly
long periodsof time. Use of the
aforementioned
steadystatesolutionto evaluateR
might be overly conservative
if the well is to be
casedlongbeforeR hasreachedits finalvalue(R).
Hence,it is necessary
to solvefor porepressureas a
functionof time. For a poroelasticmedium,the
coupledpore pressurediffusionequationcan be
statedas(RiceandCleary1976):
V2p=]4 6p6it/
690%
k,Me cqt k, cqt'
(2)
whe
1
+ 7o -(1+
="/L
PENETRATION
For permeable
rocks(k > 0.1 mD), the porepressure
gradientwill approach
its steadystateconfiguration
in relativelyshort times (lessthan one hour). For
that volumetricexpansion(or contraction)of the
steadystate pore pressuregradients,equation(1)
porousmatrixwill causea decrease(or increase)in
can be evaluatedanalytically(Wang and Dusseault
pore pressure.It can be shown(Rice and Cleary
1991,McLellanandWang 1994).
1976)that thisterm vanishes
asthe outerboundary
Shales
tendto havelowpermeabilities
(k 10'sto
of the poroelasticregion(ro) tendsto infinityoThus,
The
coupling
term
( )
arises
due
tothefact
O'1'0'3
RADIAL
DISTANCE
yielded
rock
I .... ideal
material
[
I....
typica
mate
I
"---elastic
rock
Figure2: Rheologicalmodelof an idealized,strainweakening
material.
1052
6
o.1
24
1000
2000
3000
2.6
2.2
1.8
1.4
<
n
1
lO
100
1000
10000
n,'
100000
TIME (hours)
Figure3: Sensitivity
of yieldedradiusto residual
Figure4: Sensitivity
of yieldedradiusto residual
angleof friction.
cohesion.
V2p= 67p
LM.
TIME (hours)
(3)
3.1 Uniformpermeabilitysolution
For a rockwhosepermeability
(k), porosity(n) and
matrixcompressibility
(rio) are largelyunalteredby
theyieldingprocess,andassuming
thatdeformation
- porepressure
couplingeffectsare negligible
in the
yielded zone, it is possibleto evaluatethe pore
pressureat any location and time by solving
equation(3). Solutionsto this equationare well
apertures).
In many cases,permeabilityincreasewill be so
drasticthattheporepressure
in theyieldedzonewill
equilibrate with the bottomhole pressure very
rapidly.In this case,with p(r,t) =pw for r, <r <R,
equation(1) reducesto a form which canbe solved
explicitly:
Po N
{(on-a
pw)(1-N,)+
S,+
27
[--(,-1)
Sr- tvpw
+
Nv+l
(4)
p(1-ap) Sr
documented
(e.g. Carslawand Jaeger1959, Crank
1975). Usingthesesolutions,
the pressure
termsin
equation(1) can be evaluatedanalytically
and the
integralcanbe evaluated
by quadratu/'e.
3.2 Enhanced
permeabilitysolution
It is likely that yieldingwill resultin an increasein
permeability
for most low permeability
rocks.The
degreeto which it is increasedwill dependon a
numberof factors,including:
1053
permeability
is so low that evenif it is increased
by
one or two orders of magnitudeupon yielding,
transientpore pressurepenetrationwill persistfor
considerable
lengthsof time.For example,Jouniaux
et al. (1994) measured
an increase
in permeability
by
a factor of approximatelythirty for a 0.1
indurated claystonestressedbeyond its peak
strength.
It would also be expectedthat there shouldbe
somewhat
of anincrease
in porosity(dueto dilation)
upon yielding, as well as an increase in the
2.2
38.3 (0.5)
=37.8(0)
c= 0.7
ii
:
cp= 1
1.6 ......................................
J...................
r........... [
13 1.8
W 1,6
,_1
.J
u.I
_ 1.4
_N 1.2
,_1
.J
500
1000
50
100
150
200
25
1500
TiME (hours)
TIME (hours)
c?rr=R
=0,
9pp
(5g)
/ c*pv
a/c*s
v for
rw
-<r_<
R, (5a) i.e. it is not necessaryto computeporepressuresin
v2P-,ax
the elastic zone. This significantlyreducesthe
V2p,
_k,M,
ft 69p,
for
R_<
r_<
oo,
c?t
accomplished
by Galerkin'smethod,where pore
pressures
areapproximated
by quadraticpolynomials
with boundaryconditions
pArw,O=p.,
(Sc)
p,(R,t)= pe(R,t),
(Sd)
=
p(,O = po,
5rr=R
'
(Se)
(s0
wherethe poroplasticdiffisionparameter
is defined
withanalogyto poroelastic
theory,i.e.
Equations
(5) and(1) arecoupledin sucha way as
to makeanalyticalsolutionimpossible.
In orderto
make numerical evaluation of these equations
computationally
expedient,
it is beneficial
to exploit
the factthat for k,>> ke,fluidflux acrossthe outer
boundaryof the yieldedzoneis very small.Under
1054
2
i 2.6
Rf
1.8
7.8 MPa
2.4
1.6
Cl 2.2
%%%
1.2 .....................................
-i
........................................................
.m
tu 1.8
N
<
a:::
O
200
400
600
800
<
1000
m cr
1.6
TIME(hours)
Cp .......
-..
50
100
150
200
250
o
z
Figure7: Sensitivity
of yieldedradiusto poroelastic
parameter.
TIME (days)
2 15
1.15
110
'
requiredfor R to approachRf is greater,i.e. the
further the impermeableboundaryis from the
wellbore,the slowerthe pressurebuildupin the
yieldedzone.
R is lesssensitive
to thepeakstrength
parameters.
Decreasing
cpfrom 10 to 2 MPa increases
Rf by a
factorof 1.14, and decreasing
p from 50 to 35
increases
Rfby a factorof 1.23.
....
1.7
1.6
Tablel' Sensitivity
analysis
parameters.
0
500
1000
1500
TIME (hours)
2000
Parameter
Value
42
35
4 MPa
2 MPa
0.108 m
4 Sensitivity
Analyses
The sensitivity
of yieldedzonegrowthto thevarious
model parameters was investigatedusing the
numerical solution described above. The values of
the parameters
are aslistedin Table1, exceptwhere
indicatedotherwiseon the figures.
39.8 MPa
37.8 MPa
95 MPa
l
l
10nD
7%
10'3Pa.s
4.5.10']Pa']
4. l Strengthparameters
1055
5 GPa
0.34
4.2 Bottomhole
pressure
Figure5 shows that the initial value of the
yieldedradius
(R)is verysensitive
to thisparameter,
especiallyso at low overbalances.Decreasing
overbalance
from5 to 0 MPa increases
Rt by a factor
1,9
uJ
1.8
1.7
1.6
0.00001
0.001
0.1
10
n-
lOOO
200
400
600
800
1000
TIME (minutes)
TIME (hours)
Figure10: Comparison
of numericalandanalytical
solutions,
uniformpermeability.
of 1.44. It is also evidentfrom Figure5 that the
stabilizing
influenceof a largeoverbalance
is minimal
onceporepressures
in the yieldedzoneapproach
pw.
This suggeststhat, if pore pressurepenetration
parameters
are suchthatRf will be achieved
quickly
relative to the time that the wellbore will remain
4.3 Poroelasticandporoplasticparameters
Rf increases
by a factorof 1.39 as the poroplastic
parameter(a,) increasesfrom 0.7 to 1 (Figure6).
Furthermore,
it is notedthatthe sensitivity
of R
to
a, is most pronouncedin the range 0.9 - 1.
Increasing
the poroelastic
parameter
(a) from0.6 to
1 increases
Rf by a factor of 1.38 (Figure7).
Sensitivity to these parametersis especially
unfortunate
because
theirvaluesareusuallyinferred
or guessed
ratherthanmeasured.
STRENGTH
40
t = 1019 hours
is onlymarginalat largetimes.
The effects of deformationcoupling on pore
39.5
pressure
penetration
canbe observed
in Figure12.
Note thatthe porepressurenearR becomesnearly
1 MPa lessthanpo in the relativelyearlystagesof R
39
38.5
38
37.5
37
i'-"
. 'N
"
0.038:
growth.
"=88-9 .......
:'
47................
7 CONCLUSIONS
..........
.0..:65
.'..._...._.._
5:2
.i
15.?
...............
Solutions
havebeenpresented
whichcanpredict
the rateof growthof the yieldedzonearounda
wellborein an elastic-brittle-plastic
material.
For rock whose permeabilityis unalteredby
36.5
1.2
1.4
1.6
1.8
yielding,
ananalytical
solution
whichneglects
the
coupling
of porepressure
penetration
androck
pressure
penetration.
deformation
hasprovenacceptable.
A numericalsolutionhasbeen developedwhich
estimates
the growth of the yieldedzone for
rockswhosepermeabilities
are increased
by at
leasta factorof five to ten uponyielding.
Theextentof theyieldedzoneisverysensitive
to
the residualstrengthof the rock, and less
sensitive
to the bottomholepressure,poroelastic
andporoplastic
parameters,
andpeakstrength
of
the rock.
6 NUMERICAL
EXAMPLES
fullproblem
asstatedin equations
(Sa)through
(Sf).
sensitivity
analyses
andMonteCarlosimulations
I.e. a largespatialdomainaroundthe wellborewas
in caseswhere materialparametersare poorly
defined.
discretized,
and separatepore pressure
penetration
parameters
werespecified
for theyieldedandelastic Applicationof thesemodelsto suitablefield
zones.Executionof this programwith its coupling
cases
will bepursued
in futurework.
featuresdisabledandpermeability
assumed
equalin
the yieldedand elasticzonesgivesresultswhich
NOMENCLATURE
peakandresidual
cohesion.
Young'smodulus.
(elasticandyieldedrock).
,, q,,,p) permeability
Me,Mp Elasticandyieldedrockporepressure
penetration
parameters.
1057
Cp, Cr
E
n, np
elasticandyieldedrockporosity.
peakandresidualstrength
parameters,
N,,= [l+sin()]/[1-sin(x)].
(in elasticandyielded
p, (p,,pp) porepressure,
rock).
w, o
bottomhole
andnativeporepressure.
r
152.
Kennedy,
T.C. andLindberg,
H.E. 1978.Modeltests
for plasticresponse
of linedtunnels.
ASCEJ. Eng.
radial coordinate.
R, (R,,R) yieldedradius,(initialandfinal).
r,, ro
wellboreandouterdrainage
radii.
rockstrength
Sp,S, peakandresidualapparent
parameters,
S= -2cos(,).cJ[1-sin()].
t, (ts) time,(timeto reach
Biot'sporoelastic
parameter.
poroplastic
parameter.
porousmatrix,mineralgrainandpore
watercompressibility.
volumetric and axial strain.
O'H
tunnellinings.
Proc.3raInt. Congress
on Rock
Mechanics: 1150-1156.
of wellboreinstability:
Applicationof a versatile
poro-elastoplasticmodel. Proc. SPE/ISRM
Eurock'94: 205-214. Rotterdam: Balkema.
peakandresidualfrictionangle.
poroelastic
constant,
= (1-2v)a/J2(1-v)].
mudtiltrateviscosity.
Poisson's ratio.
in situ stressesnormal to wellbore axis.
Papamichos,
E., Vardoulakis,
I. andSulem,J. 1994.
1101.
maximum
andminimum
principal
stress.
Generalized
continuum
models
for
borehole
stabilityanalysis.
Proc. SPE/ISRMEurock'94:3744. Rotterdam: Balkema.
REFERENCES
Eurock '94:
85-92.
Rotterdam:
of
shale
stabilization:
Water
transportthroughshales.SPEFE 9: 129-134.
Bradford, I.D.R. and Cook, J.M. 1994. A semi-
diffusion
solutions for
fluid-saturated
elastic
porousmediawith compressible
constituents.
Rev.
Geophys.SpacePhys.14: 227-241.
Shao,J.F. & Khazraei,R. 1994. Wellborestability
analysisin brittle rockswith continuousdamage
model. Proc.
SPE/ISRM
Eurock'94:
215-222.
Rotterdam: Balkema.
analyticelastoplastic
modelfor wellborestability
L. and Dusseault,M.B.
with applications
to sanding.Proc. SPE/ISRM Thallak, S., Rothenburg,
Eurock'94:347-354. Rotterdam: Balkema.
1991.Analysisof boreholestabilityusingdiscrete
element
models.
Proc.7thInt. Congress
onRock
Mechanics: 813-818. Aachen.
Cheng,H. and Dusseault,M.B. 1993. Deformation van Oort, E. 1994. A novel techniquefor the
and diffusionbehaviourin a solid experiencing investigationof drilling fluid inducedborehole
instabilityin shales.Proc. SPE/ISRM Eurock'94:
damage: a continuousdamage model and its
293-308. Rotterdam: Balkema.
numerical
implementation.
Int. J. RockMech.Min.
$ci. Geomech. Abstr. 30: 1323-1331.
Wang,Y. andDusseault,
M.B. 1991.Boreholeyield
Crank, J. 1975. The mathematicsof diffusion. andhydraulic
fractureinitiationin poorly
Toronto: Oxford.
Hawkes,C.D. 1996.Investigating
theeffectsofpore
pressurepenetrationon the stabilityof a yielded
wellbore. PhD. Thesis. University of New
Brunswick(Canada).
Horsrud, P., Holt, R., Sonstebo,E.F., Svano,G. &
Bostrom, B. 1994. Time dependentborehole
stability: Laboratory studies and numerical
simulation of different mechanisms in shale. Proc.
SPE/ISRM
Eurock'94:
259-266.
Rotterdam:
Balkema.
1058