Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Stage:
Page: 1
Total Pages: 11
Summary
Formation damage caused by overbalanced drilling with waterbased mud (WBM) is inevitable as a result of mud filtrate invading the near-wellbore formation. The invasion radius is critical to
the multiphase flow when the well is put on production. It contributes to the total skin that hinders the hydrocarbon production. Furthermore, the response of the logging tools may be affected as a
result of such invasion, rendering many inaccurate calculations in
formation evaluation. To evaluate the skin caused by mud-filtrate
invasion, it is important to determine the radius of invasion. A
thorough literature review indicated that no practical and reliable
method with solid theoretical basis to quantify formation damage
is available. Former studies assumed that single-phase drilling
fluid displaces reservoir fluid during the invasion. The neglecting
of residual reservoir fluid in the invaded zone will introduce error
to invasion-radius estimation. This work takes the residual reservoir fluid into account; thus, the estimation of invasion radius is
more accurate.
This work proposes a practical model to determine the depth
of mud-filtrate invasion near the wellbore drilled by WBM. The
distribution of mud-filtrate saturation in the near-wellbore region
is also calculated by using drilling-operation parameters, mud-filtration-test data, relative permeability, and drilling time. With the
accurately determined invasion radius and known wellbore radius,
reservoir permeability, and damaged-reservoir permeability, one
can evaluate skin factor more accurately. With the knowledge of
invasion volume and radius, one can design the wellbore-cleanup
procedure appropriately. The proposed model allows engineers to
predict the well performance and to diagnose wellbore problems
by checking any deviation from the predicted production. This
study also can assist with the correction of parameters inferred
from log measurements, thereby reducing the over- and/or underestimation of log-derived parameters used in various formationevaluation calculations.
Introduction
Skin is a dimensionless term to be used in diagnosing well performance. A positive skin often indicates the decrease of well performance caused by factors such as formation damage, limited
entry, and choked effect. It could also be a negative skin of
increasing the well performance, which usually is attributed to
well stimulations of acidizing and hydraulic fracturing. For our
purposes, the skin of formation damage resulting from overbalanced drilling (OBD) is studied. Hawkins (1956) equation for
damage skin is
This paper (SPE 168184) was accepted for presentation at the 2014 SPE International
Symposium and Exhibition on Formation Damage Control, Lafayette, Louisiana, 2628
February, and revised for publication. Original manuscript received for review 14 November
2013. Revised manuscript received for review 18 September 2014. Paper peer approved 1
December 2014.
s
k
rs
1 ln ; . . . . . . . . . . . . . . . . . . . . . . . . . 1
ks
rw
where k reservoir permeability, ks damaged reservoir permeability, rs radius of damaged zone, rw wellbore radius, and
s skin factor.
As shown in Eq. 2, well productivity depends on a number of
rock and fluid properties. A positive damage skin impairs the oil
production. Without sufficiently evaluating the magnitude of-well
skin, one cannot analyze oil production properly:
qo
kkro hDp
; . . . . . . . . . . . . . . . . . . 2
1 re
141:2lo Bo ln s
2 rw
Page: 2
Total Pages: 11
tion distribution in the flushed and transition zones, formationdamage skin is calculated accordingly. Then, one can plan the
well stimulation accordingly and appropriately, if it is necessary
to remove or alleviate formation damage. This model also can
estimate the cleanup period at the onset of production. It allows
engineers to predict the well performance and to diagnose wellbore problems by checking any deviation from the predicted production. In this regard, one can consider the effect of invasion on
the response of the logging tools, to avoid or minimize the overor underestimation of reservoir parameters such as water
saturation.
Drilling Fluid
pm
Stage:
k
h
rw
pe
re
re
pm
kkrf @Apf
; . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
@r
lf
Filtrate
pe
kkro @Apo
. . . . . . . . . . . . . . . . . . . . . . . . . . . 3
lo @r
pm
rw
rw
pe
k
r
re
Fig. 2Radial reservoir system: (a) plan view, (b) lateral view.
2
Drilling Fluid
qt = qo+qf | r+r
qt = qo+qf | r
pm
pe
h
rw
k
r
r
re
qf
kkrf @Apo Pc
: . . . . . . . . . . . . . . . . . . . . . . 6
@r
lf
qf : . . . . . . . . . . . . . . . . . . . . 8
@r
@r
kkrf
Subtracting Eq. 8 from 7, we obtain
lf
@APc 1 lo
qo qf : . . . . . . . . . . . . . . . . . . 9
@r
k kro
krf
To simplify the analysis, we introduce the concepts of the total
filtrate-invasion rate and fractional flow, which are defined as
qt qo qf
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Stage:
qf qf r qf qf rDr Dt
1
ff
Pc
qt lo
@r
ff
: . . . . . . . . . . . . . . . . . . 14
kro lf
1
krf lo
If capillary pressure is negligible, Eq. 14 collapses to
ff
Total Pages: 11
and
ff
Page: 3
1
: . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
kro lf
1
krf lo
where Sf filtrate saturation, qf filtrate density, Dt time interval, t time, Dr radius increment, and / porosity.
Simplifying Eq. 16, we have
qf qf r qf qf rDr Dt
@qf qf
@Sf qf
2prh/
: . . . . . . . . . . . . . . . . . . 19
@r
@t
@qf
@Sf
2rph/
: . . . . . . . . . . . . . . . . . . . . . . . . 20
@r
@t
@ f f qt
@Sf
2rph/
: . . . . . . . . . . . . . . . . . . . . . 21
@r
@t
Stage:
Page: 4
Total Pages: 11
Spurt loss
Qf
Qsp
0.5
dff
from core relative permeabildSf Sf
ity and fluid properties. In case the core relative permeability is
not available, one can use the Corey (1954) relative permeability
model. The total filtrate-invasion volume, Qf tqf-av, is discussed
in the next section. The invasion time t is the period between the
initial penetration of the formation and the setting of cement in
the cased hole.
The total filtrate-invasion volume can be influenced severely
by total invasion time, pressure differential, filtrate viscosities and
reservoir fluid, reservoir permeability and thickness, permeability
and thickness of mudcake, and magnitude of formation damage.
The invasion volume consists of the spurt loss, the invasion upon
the mudcake formation (we call this Stage 1), and the invasion at
constant mudcake thickness (we call this Stage 2). The static and/
or dynamic filtration tests are used to estimate the filtration rate.
The flow of mud filtrate through a mudcake is given by Bourgoyne et al. (1986):
analysis. One can calculate
; . . . . . . . . . . . . . . . . . . . . . . . . 22
qf -av @t
@r
where
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
dt
lf hmc
: . . . . . . . . . . . . . . . . . . . . . 24
qf -av @t
dSf @r
s
p
kmc Dpmc fsc
1 A t; . . . . . . . . . . . . . . . 31
Qf 2
lf
fsm
Qf
qf -av
t
qf -invasion
@Sf
@Sf
dt
dr 0
@t
@r
fsc
1 hmc : . . . . . . . . . . . . . . . . . . . . . . 32
Qf A
fsm
or
@Sf
@Sf dt
: . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
@r
@t dr
Substituting Eq. 26 into Eq. 24 yields
dff
2rph/
dt
dr: . . . . . . . . . . . . . . . . . . . . . . . . . 27
qf -av
dSf
Integrating Eq. 27 yields an equation for filtrate-front position
rf as
s
tqf -av dff
; . . . . . . . . . . . . . . . . . . . . . . . 28
rf
ph/ dSf f
where rf filtrate-front position in a radial reservoir system.
For any filtrate saturation Sf, one can calculate the position by
s
tqf -av dff
rSf
; . . . . . . . . . . . . . . . . . . . . . . 29
ph/ dSf Sf
where rSf position of any filtrate saturation in a radial reservoir
system.
Eq. 28 dictates the deepest filtrate invasion, or the outerboundary of the damaged zone. Eq. 29 gives the filtrate-saturation
distribution from wellbore to the outer boundary of the damaged
zone. It is noted that the damaged radius is rarely beyond 10 ft
(Nunes et al. 2009), considering the limited filtrate volume that
invades into the formation.
Now, we can discuss the parameters required for the calculation in Eq. 29. One can obtain porosity from well logs or core
4
khDpformation
;
k
162:6lf logt log
3:23
0:87s
/lf ct rw2
34
Stage:
Page: 5
Total Pages: 11
The sum of Eqs. 39, 37, and 38 gives the total filtrate volume
invading the formation:
t1
t2
2prw hQsp-lab
Qf
qf -invasion1 dt qf -invasion2 dt;
Alab
t1
40
where t1 time to establish constant mudcake thickness, or time
period of Stage 1, and t2 total invasion time.
One can use Eq. 31 to calculate the time to establish constant
mudcake thickness. One can calculate the filtrate-invasion volume
in Eq. 31 from Eq. 32. The mudcake thickness in Eq. 32 is the difference between bit size and wellbore radius, which is measured
from the caliper log. One should note that the parameters used for
calculation should be measured under downhole conditions, if
possible.
pm pp Dpmc Dpformation
162:6qf -invasion1 lf logt log
v
u 2k Dp
mc
mc p
t
qf -invasion1 lf u
u
t fsc
1 lf
fsm
0:0011268kmc A
k
3:23 0:87s
/lf ct rw2
kh
35
where pp formation pore pressure and qf-invasion1 filtrate-invasion rate at Stage 1, and
A 2prw h: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Calculation Procedure
The calculation procedure is as follows:
1. Calculate spurt loss in wellbore with Eq. 39.
2. Divide the invasion period into several time intervals; the
overbalanced pressure should be constant in each interval.
3. Choose Time Interval 1; assume skin factor 0; calculate
invasion rate by Eq. 37 if mudcake thickness is less than
constant mudcake thickness or by Eq. 38 if mudcake thickness is equal to or greater than constant mudcake thickness.
4. Calculate mudcake thickness by Eq. 33.
5. Repeat Steps 3 and 4 until the calculated mudcake thickness
converges.
6. Calculate invasion volume by Eq. 40; calculate invasion radius by Eq. 28; calculate a new skin factor by Eq. 1.
7. Choose Time Interval 2, with the new skin factor obtained
from Step 6; repeat Steps 3 through 6 until finishing all
time intervals.
8. With the invasion volume obtained in Step 7, calculate the
filtrate-front position and any filtrate-saturation position
through Eqs. 28 and 29.
Case Study
A synthetic case was used to illustrate filtrate invasion in drilpm pp
ling. The input data are shown in Table 1. Table 2 is the relative
v
:
u 2k Dp
permeabilities and calculated parameters vs. filtrate saturation.
mc
mc p
u
!
t
lf u
One can obtain drilling-fluid and filtrate properties from drillingk
t fsc
162:6lf logt log
3:23 0:87s
1 lf
fluid laboratory measurements. One can measure reservoir-rock
/lf ct rw2
fsm
properties, fluid properties, and relative permeability through a
0:0011268kmc A
kh
pressure/volume/temperature test and core analysis. Reservoir
37 pay thickness, wellbore radius, and constant mudcake thickness
are estimated by well-logging interpretation. Overbalanced presSimilarly, combining Eq. 30 with Eq. 33, we have the filtrate- sure and total invasion time are from drilling operations. In case
invasion rate at Stage 2:
of a gas reservoir, one should change the inputs accordingly.
One should replace oil properties with gas properties. One
qf -invasion2
should use gas relative permeability instead of oil relative perpm pp
; meability. One should use gas pseudopressure to replace presk
162:6lf logt log
3:23 0:87s sure. Considering that gas properties are functions of pressure
/lf ct rw2
lf hmc
and temperature, one can use average gas properties as an
approximation.
kh
0:0011268kmc A
The spurt loss in the wellbore is 0.055 bbl. The time to estab 38
lish constant mudcake thickness is approximately 11.5 minutes, or
690 seconds. The cumulative-invasion volume at the time the
where qf-invasion2 filtrate-invasion rate at Stage 2.
Because the permeability of mudcake tends to decrease with mudcake thickness reaches a constant value is 0.04 bbl. The total
increasing pressure, or kmcDpmc a constant (Bourgoyne et al. filtrate-invasion volume is 18.9 bbl at a total invasion time of 100
1986), the spurt loss in the wellbore can be approximately calcu- hours. Therefore, the filtrate-invasion volume at Stage 2 is more
significant than those of spurt loss and Stage 1, as shown in Figs.
lated from static and/or dynamic filtration-test data, which is
5 and 6. The plots of filtrate-invasion rate and mudcake thickness
vs. time are shown in Fig. 5. Fig. 6 shows the cumulative filtrateQsp-lab Qsp-well
Stage:
Page: 6
Total Pages: 11
0.007
1000
Stage 1
0.006
Stage 2
0.005
100
0.004
Filtrate invasion rate (bbl/day)
Mudcake thickness (ft)
0.003
10
0.002
0.001
1
0.0001
0.001
0.01
0.1
0
100
10
Time (hours)
100
1000
Filtrate invasion rate (bbl/day)
10
100
0.1
0.01
10
Stage 1
1
0.0001
0.001
Stage 2
0.0001
0.001
0.01
0.1
10
Fig. 5The plots of filtrate-invasion rate and mudcake thickness vs. time.
100
Time (hours)
Fig. 6The plots of filtrate-invasion rate and cumulative filtrate-invasion volume vs. time.
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.0001
0.001
0.01
0.1
10
Sensitivity Analysis
Sensitivity analysis is conducted to better understand the effects
of crucial parameters on filtrate-penetration radius. Only the controllable parameters are discussed for practical-application
purposes.
0.9
0.8
0.9
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.0001
0.001
0.1
0.01
Page: 7
Total Pages: 11
Stage:
0.8
0.7
0.6
0.5
Pressure difference = 100 psi
0.4
0.3
0.2
0.1
0
0.0001
10
0.001
0.1
0.01
10
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.0001
0.001
0.01
0.1
10
Stage:
Page: 8
Total Pages: 11
0.009
80
0.008
0.007
70
0.006
60
0.005
50
0.004
40
90
0.003
30
0.002
20
0.001
10
0
0.001
0
0.01
0.1
10
100
Fig. 11Plots of filtrate-invasion volume and mudcake thickness vs. time for two scenarios (no additional mudcake buildup and
additional mudcake buildup after running well logging).
logging tools into and out of the hole, and running casing into the
hole. The buildup of mudcake will be offset partially or significantly. In many cases, drillers use (1) a higher mud-circulation
rate with two or more circulations before tripping out and (2)
short tripping in and out to avoid the sticking of the drillpipe and
to prepare a gauge hole for running well-logging tools and/or casing. Under these conditions, the additional mudcake buildup will
be small. Furthermore, the mudcake thickness is obtained from bit
size and well logging, which measures wellbore diameter during
tripping out of the hole. Therefore, the difference is mainly in running casing. Mudcake buildup is depressed while running casing
because of the scratch effects of the casing, casing shoe, and the
stabilizers installed outside of the casing and the movement of the
drilling/completion fluid. It is also very difficult, if not impossible,
to gain mudcake thickness while running casing. Fig. 11 shows
plots of filtration and mudcake thickness vs. time for two scenarios: additional mudcake buildup and no additional mudcake
buildup after running well logging. The difference in total invasion volume is 2.1 bbl, or 3.1% (68.9 bbl for constant mudcake
thickness, 66.8 bbl for mudcake buildup after running well logging), even with a 25% increment in mudcake thickness.
Validation of the Proposed Model With Field Data
The mud and mudcake properties, reservoir-rock and fluid properties, wellbore geometry, and drilling parameters of a well drilled
in the South China Sea were used to validate the model. Waterbased mud was used to drill the well. Overbalanced pressure
ranged from 425 to 481 psi during the penetration of the reservoir
and cementing. The total exposure time of the formation to drilling
fluid was 59 hours. After running the casing, well tests were conducted to test the well productivity, estimate the permeability and
skin factor, obtain the reservoir pressure, and detect the reservoir
boundary. Table 3 shows the inputs for the calculations. Fig. 12
shows the relative permeabilities of oil and filtrate. Drilling-fluid
and filtrate properties were obtained from drilling-fluid analysis at
the wellsite. Reservoir-rock properties, fluid properties, and the
relative permeability were measured through pressure/volume/
temperature analyses of fluid samples taken from well tests and
conventional and special core analyses. Reservoir pay thickness,
wellbore radius, and constant mudcake thickness were estimated
by well-logging interpretation. Overbalanced pressure and total
invasion time were from records in drilling and completion
operations.
The proposed model calculated a filtrate-invasion depth of 4.2
ft (or an invasion radius of 4.2 0.328 4.528 ft). Well-test analysis gave a skin factor of 18.8 and a reservoir permeability of 16
md. Applying the Hawkins (1956) skin-factor formula, we can
calculate the invasion radius, which was 4.8 ft. The absolute error
is 0.272 ft, and the relative error is 5.6%. The difference can
result from reservoir heterogeneity, inaccurate pressure drop during the invasion, inaccurate relative permeabilities, variation of
mud properties along the wellbore, inaccurate mudcake thickness,
and any deviation from the aforementioned assumptions. Therefore, the proposed model gives reasonable results. Fig. 13 shows
overbalanced pressure and the calculated filtrate-invasion depth at
different invasion times.
Relative Permeabilites vs. Filtrate Saturation
1
0.9
0.9
krf
kro
0.8
0.7
0.8
0.7
0.6
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0
0
0
0.2
0.4
0.6
0.8
Table 3The input data for filtrate invasion in a well drilled in the
South China Sea.
8
460
3.5
440
420
400
380
2.5
Overbalanced pressure vs. Invasion time
Filtrate-invasion depth vs. Invasion time
2
1.5
360
340
320
0.5
300
0.0001
0.001
0.01
0.1
10
0
100
Fig. 13Overbalanced pressure and the calculated filtrate-invasion depth vs. invasion time.
Conclusions
The following conclusions have resulted from this work:
The model developed in this study estimates the filtration distribution along the radius from the wellbore wall to the outerboundary of the invasion zone.
The filtrate does not completely displace the original reservoir
fluid away from near-wellbore regions. In other words, the displacement is not a piston-like process. The calculated invasion
is deeper than that of a 100% displacement process.
The filtrate-invasion volume after mudcake thickness become
constant is more significant than those of spurt loss and mudcake buildup. Therefore, major effort should focus on reducing
invasion at constant mudcake thickness.
The reduction of formation permeability caused by invasion
reduces the filtrate-invasion rate. Impairment as a result of damage can limit damage and may even be desirable in the nearwellbore region.
The proposed model cannot be applied to a heterogeneous reservoir. In the case of low-permeability reservoirs, one should
take into account capillary pressure, as in Eq. 14. Ignoring it
may lead to unacceptable results.
Future Work
The capillary pressure plays an important role in affecting
filtrate invasion, especially in shale/tight reservoirs. Future
work should focus on quantification of filtrate invasion with
capillary pressure effects.
We should conduct experiments and develop guidelines to
calibrate the proposed model to each specific well.
Nomenclature
A flow area
Alab filter-press area in filter-press test
Bo oil formation volume factor
ct total compressibility
ff filtrate fraction
fsc the volume fraction of solid in the mudcake
fsm the volume fraction of solid in the mud
h pay thickness
hmc mudcake thickness
k reservoir permeability
kmc mudcake permeability
kro relative permeability to oil
krf relative permeability to filtrate
ks damaged-reservoir permeability
Pc capillary pressure
pe reservoir outer-boundary pressure
pf filtrate pressure
pm drilling-fluid pressure
po oil pressure
pp formation pore pressure
Stage:
qf-invasion
qf-invasion1
qf-invasion2
qo
qt
r
re
rf
rs
rSf
rw
s
Sf
t
t1
t2
lf
lo
Dp
Qf
Qsp-lab
Qsp-well
qf
qf-av
Dpformation
Dpmc
Dt
Dr
qf
/
Page: 9
Total Pages: 11
Acknowledgments
The authors are grateful to the Petroleum Engineering Department
at the University of North Dakota. This research is supported in
part by the US Department of Energy under award number DEFC26-08NT0005643 and the North Dakota Experimental Program
to Stimulate Competitive Research (EPSCoR) Program under
award number EPS-0814442.
References
Al-Anazi, H.A., Bataweel, M.A., and Al-Ansari, A.A. 2009. Formation
Damage Induced by Formate Drilling Fluids in Gas-Bearing Reservoirs: Lab and Field Studies. Presented at the SPE/IADC Drilling Conference and Exhibition, Amsterdam, The Netherlands, 1719 March,
SPE-119445-MS. http://dx.doi.org/10.2118/119445-MS.
Altunbay, M., Sy, R., and Martain, R. 2003. Formation Damage Assessment and Remedial Economics From Integration of NMR and Resistivity Log Data. Presented at the SPE Annual Technical Conference
and Exhibition, Denver, Colorado, 58 October. SPE-84384-MS.
http://dx.doi.org/10.2118/84384-MS.
Amanullah, M. 2003. A Novel Method of Assessment of Spurt and Filtrate
Related Formation Damage Potential of Drilling and Drilling-in Fluids. Presented at the SPE Asia Pacific Oil and Gas Conference and Exhibition, Jakarta, Indonesia, 911 September. SPE-80484-MS. http://
dx.doi.org/10.2118/80484-MS.
Bourgoyne, T.A., Chenevert, E.M., Millheim, K.K. et al. 1986. Applied
Drilling Engineering, Vol. 2. Richardson, Texas: SPE Textbook
Series.
Brownson, G.W. and Peden, J.M. 1980. A Systematic Approach to Assessing Permeability Impairment Due to Drilling Fluids. SPE-9727-MS.
Byrne, M.T., Spark, I.S.C., Patey, I.T.M. et al. 2000. A Laboratory Drilling Mud Overbalance Formation Damage Study Utilising Cryogenic
SEM Techniques. Presented at the SPE International Symposium on
Formation Damage Control, Lafayette, Louisiana, 2324 February.
SPE-58738-MS. http://dx.doi.org/10.2118/58738-MS.
Corey A.T. 1954. The Interrelation Between Gas and Oil Relative Permeabilities. Prod. Monthly 19 (1): 3841.
Stage:
Page: 10
Total Pages: 11
Stage:
Page: 11
Total Pages: 11
Peng Pei is a research engineer at the Institute of Energy Studies, University of North Dakota. He holds a PhD degree in geological engineering and an MS degree in mechanical
engineering, both from the University of North Dakota. Pei also
holds a BS degree in mechanical engineering from North
China Electrical Power University. His research area focuses on
energy-related rock mechanics.
11