DC167945 DOI: 10.

2118/167945-PA Date: 10-December-14 Stage: Page: 431 Total Pages: 7

The Use of Modeling To Enhance

the Analysis of Formation-Pressure
Integrity Tests
M.W. Alberty, Hess Corporation, and M.R. McLean, Consultant

Summary while recording the pumped mud volume and the surface pressure
Formation-pressure integrity tests (FPITs) are used to verify the until the target test pressure is achieved. The pumps are then shut
integrity of cement at a casing shoe, measure the stress state of down, and the decay of the pressure is monitored until the pres-
the exposed formations for well planning and operations, and sure stabilizes. These data are then analyzed to determine whether
determine the maximum equivalent circulating density to which a a fracture was introduced, propagated, or closed, and to quantify
shoe can be safely exposed. Critical decisions on operations are the associated pressures.
made directly from the results and include decisions about the There are four reasons to conduct an FPIT:
need for remedial cement operations, maximum mud weights that • To meet local regulatory requirements
can be used to drill the next well section, minimum mud weights • To verify the integrity of cement isolation at the shoe
that can be used to prevent hole collapse, calibration factors for • To measure the stress state of the exposed formation for use
predicting fracture gradients, and the potential need for lost-circu- in well planning and operations
lation-mitigation strategies. • To determine the maximum equivalent circulation density
The interpretation techniques of the result most frequently (ECD) to which the formation at the shoe can safely be
focus on the point at which a fracture first starts, the point at exposed.
which unstable fracture growth begins, or the closure pressure of To determine either the stress state of the exposed formation or
the fracture when pumping ceases. However, the early pressure- the maximum ECD to which the formation at the shoe can be
buildup behavior is often overlooked and can provide much safely exposed, it is necessary to pressure up the formation below
insight on the integrity of cement, the point of the start of a frac- the shoe until a fracture is induced. Fig. 2 displays an idealized
ture, the permeability of the formation being tested, the need for pressure response recorded while conducting a test. The leakoff
cement remediation, and the potential to increase fracture resist- pressure, labeled LOP on the figure, is the point at which a frac-
ance by use of wellbore-strengthening techniques. ture first starts at the wellbore interface. If pressure is increased
This paper presents a model for predicting early pressure- above the LOP, stable fracture growth may occur until the peak or
buildup behavior, discusses how the model can be used to breakdown pressure (labeled FBP on the figure) is achieved, at
improve the interpretation of FPITs significantly, and provides which time unstable fracture growth occurs away from the well
examples of the application in select wells. many meters into the formation to a region that was undisturbed
by the presence of the well. If pumping continues, the pressure
Introduction should stabilize to the fracture-propagation pressure (FPP) as la-
beled in the figure. When pumping is finally halted, the pressure
Formation-pressure integrity test (FPIT) is a generic name immediately drops as the friction pressure is removed from the
assigned to the range of pressure tests that are conducted immedi- well and the instantaneous shut-in pressure (labeled ISIP in the
ately after drilling out a casing or liner shoe and, on occasion, figure) is observed. If well pressure is greater than the pressure
later when additional formation has been exposed while drilling
required to hold the fracture open, the fracture will continue to
ahead. This generic name applies to the different types of tests
grow until the pressure falls below the fracture-closure pressure
that might be conducted including those commonly called jug
and the fracture then closes. Pressure may continue to decline af-
tests, leakoff tests, formation-integrity tests, breakdown tests,
ter the fracture closes if permeability is exposed in the well and as
extended leakoff tests, and repeat extended leakoff tests. Opinions
mud and formation temperatures equilibrate. However, a compari-
vary across the industry about which type of test is best to con-
duct, how to interpret those tests, how to report them, what consti- son of surface and downhole data suggests that the pressure-
tutes a good test, and the behavior of the tests under nontypical decline downhole is often not as marked as that seen at the
conditions. surface. Some of the surface-pressure drop may be a result of
There are alternative methods to configure the rig up for these mud-gelling effects.
tests. The most generalized rig up is shown in Fig. 1. The basic The fracture-closure pressure can be the most valuable piece
procedure would consist of first running a casing-integrity test of information from the test used to determine the stress state of
(CIT) before drilling out the cement at the shoe to establish the in- the formation beyond the influence of the well. However, many
tegrity of the casing and/or liner. After the CIT, the cement pres- operators resist pumping the amount of mud required to achieve
ent inside the casing and the rathole is drilled out and, typically, 3 FPP out of concern that the fracture may grow a significant dis-
m of fresh formation is drilled. The mud is conditioned through tance and encounter lower-stressed rock that may limit the maxi-
adequate circulation and the drill bit pulled back into the casing to mum mud weight that could be used to drill the next hole interval.
minimize any risk of sticking the downhole assembly while con- The leakoff point (LOP on the figure) is also an important
ducting the test. The cement unit is connected to the drillpipe and/ pressure to identify because it establishes the point at which a
or annulus, and circulation is established. The Hydril-type annular fracture starts (or at least at which its growth is visible from the
or pipe rams are closed around the drillpipe, and mud is pumped measurements) and can be used, in part, to establish the maximum
down the drillpipe and/or the chokeline into the closed system safe ECD. When combined with a repeat of the test, the LOP can
also be used to estimate the tensile strength of the formation. The
LOP is most frequently described as the point at which the trend
Copyright V
C 2014 Society of Petroleum Engineers
of the pressure increase deviates from linearity (Postler 1997;
This paper (SPE 167945) was accepted for presentation at the 2014 IADC/SPE Drilling Addis et al. 1998; Van Oort and Vargo 2008).
Conference and Exhibition, Fort Worth, Texas, USA, 4–6 March 2014, and revised for
publication. Original manuscript received for review 14 March 2014. Revised manuscript
In a significant number of tests, the pressure vs. volume trend
received for review 5 June 2014. Paper peer approved 4 September 2014. before the start of a fracture is nonlinear. This can make

December 2014 SPE Drilling & Completion 431

ID: jaganm Time: 18:07 I Path: S:/3B2/DC##/Vol00000/140032/APPFile/SA-DC##140032

 DC167945 DOI: 10.2118/167945-PA Date: 10-December-14 Stage: Page: 432 Total Pages: 7

Idealized Example Test FPIT #1

FPIT #1 - Shut-in
Alternate 350 FPIT #2
FBP FPIT #2 - Shut-in
Flow FPP
300
Path
LOP
250 ISIP

Pressure (psi)
Cement Unit
Valves 200
Mud Pit 150
BOP
100
Valve
Pressure
Gauge 50
Pumping Time Pumping Time
0
0 2 4 6 8 10 12 14
Volume (bbls)/Time (min)

Fig. 2—An idealized example test.

Casing Shoe
New Formation A pressure vs. volume slope of the casing-intensity test (CIT)
is a direct measure of the system compliance before drilling out
Fig. 1—Typical FPIT rig up. the cement and the casing shoe and exposing new formation. That
compliance is a function of the casing expansivity, which is itself
identifying the LOP difficult to determine accurately. Modeling also a function of the cement and rock compressibility and the
can be used to analyze the shape of the trend before fracture ini- fluid within uncemented annuli; but the dominant factor effecting
tiation, which can then greatly aid in understanding the behavior the system compliance in a CIT is the compressibility of the mud.
of the test (and in the accurate determination of the fracture-initia- Rather than break down the CIT slope into the various compo-
tion pressure). The results of the modeling can also be used to nents, it is more convenient to work in terms of an apparent mud
understand the properties of the exposed formation and the poten- compressibility (cm ) as defined by the CIT pressure vs. volume
tial for the presence of a channel behind the casing. slope. Depending on the well construction at the time of the test,
the actual mud compressibility (as an average for the temperature
and pressure conditions of the whole mud column) can be
The Model
expected to be 5 to 15% lower than cm . For all practical purposes,
The behavior of the pressure vs. volume curve (or compliance of the apparent mud compressibility backed out from the CIT will
the system) before the start of the fracture is governed by a num- apply to the formation-pressure integrity tests (FPITs) also,
ber of primary factors: assuming that the basic mud properties have not changed signifi-
• Mud compressibility cantly between the two tests. Adjustment from the CIT conditions
• Casing expansivity to the FPIT conditions needs to take into account the additional
• Cement compressibility mud volume resulting from
• Formation compressibility • Drilling the cement out to the casing shoe
• Air-in-lines • Drilling out the rathole
• Permeability losses • The newly exposed hole (typically only 3 m)
• Channel volume The apparent mud compressibility is then applied to this new
system volume to obtain a first estimate of the expected pressure
1200 vs. volume slope for the FPIT.
ΔVp Fig. 3 shows a typical casing test (maroon curve) and the fit of
a linear regression to the data (red curve). CIT pressure vs. vol-
ume responses should be linear. Apparent nonlinearity may occur
1000 as a result of variations in the pump rate or when leaks in the sys-
tem are present. The effects of air trapped in the lines may also be
observed, but those effects diminish quickly in the early stages of
the test and are usually unnoticeable over much of the CIT plot in
800 which the pressure applied is often well more than a 1,000 psi.
The apparent mud compressibility can then be determined with
Eq. 1, where Vsys_cit is the volume of the system at the time of the
Pressure (psi)

CIT:
600
cm ¼ 1=ðVsys cit  DP=DVp Þ: . . . . . . . . . . . . . . . . . . . ð1Þ
ΔP
The expected pressure vs. volume slope for the FPIT is com-
puted from rearranging Eq. 1, where Vsys_fpit is the total system
400
volume for the FPIT, allowing for the change in volume during
the test resulting from the formation compression (DVfm) below
the shoe. This includes both the rathole and newly exposed forma-
tion. Thus, for the FPIT, we have
200
Casing-Test Data DP ¼ ðDVp  DVfm Þ=ðVsys fpit  cm Þ: . . . . . . . . . . . . . .ð2Þ
Casing-Test Data Shut-in
Modeled Response
Assuming the formation behaves elastically, the increase in
0 hole volume from formation compression, DVfm, can be computed
0 1 2 3 4 5 6 7
from
Volume (bbls)/Pseudo Time (bbls/min.min)
DVfm ¼ DP  p  HS2  LOH =ð4GÞ; . . . . . . . . . . . . . . ð3Þ
Fig. 3—Analysis of the casing test.

432 December 2014 SPE Drilling & Completion

ID: jaganm Time: 18:07 I Path: S:/3B2/DC##/Vol00000/140032/APPFile/SA-DC##140032

 DC167945 DOI: 10.2118/167945-PA Date: 10-December-14 Stage: Page: 433 Total Pages: 7

3500 proportional to the difference between well pressure and forma-

tion pore pressure (DPwf) and inversely proportional to the natu-
ral logarithm of the ratio of the effective-permeability radius or
distance to undisturbed formation pressure (re) to the wellbore ra-
3000
dius (rw). The volume lost to each volume increment can then be
found by multiplying the rate of loss by the time interval (DT)
associated with the volume increment:
2500
DVk ¼ 0:00708  DT  Ff  ðDPwf Þ=½1440  lnðre =rw Þ
                   ð5Þ
Pressure (psi)

2000
Ff ¼ K  h=l: . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð6Þ

The actual FPIT slope is often compared in the first instance

1500 with the CIT slope. When there is not a good match between these
slopes, further investigation can be carried out by considering the
formation compressibility, air-in-lines, permeability losses, and
potential channel-volume properties. These properties can be
1000
adjusted to optimize the fit to the observed data and conclusions
drawn about the amount of air trapped in the lines, the presence of
exposed permeability, and the potential for an exposed channel.
500 CSG Test Fig. 4 displays the expected generalized modeled response to
Air in Line (2 bbls)
the individual components of the model. The casing test (brown
Exposed Soft Formation
Low-Permeability Formation
curve) reflects the compression of the mud and the expansion of
High-Permeability Formation the casing and cement as a function of the applied pressure. The
0 response should be linear. The slope will decrease as the volume
0 2 4 6 8 10 12 14
of the system increases, as additional new formations are exposed,
Volume (bbls)/Pseudo Time (bbls/min.min) or if the compressibility of the fluid increases (e.g., from a mud
change).
Fig. 4—Example modeled responses. When air is introduced into the system (usually when the con-
nections are made between the cement unit and the drillpipe), and
where G is the shear modulus of elasticity for the formation, HS is not circulated out, the response (blue curve) will in the beginning
the hole size of the section exposed below the shoe, and LOH is be nearly flat and then quickly curve upward to nearly parallel to
the length of the open hole exposed at the time of the test. the casing test as the applied well pressure increases. As pressure
In most circumstances, the expansivity of the hole section is applied, the volume of air decreases. As mud is pumped, the air
below the shoe has a negligible influence on the pressure vs. vol- is pushed down the hole, subjecting it to both the pressure of the
ume slope except for “soft” formations that may be encountered column above and the applied surface pressure. After subjected to
at relatively shallow depths below the mudline (e.g., 1/2-m shoe sufficient pressure, the volume of the air becomes so small that its
tests). It is important to note that, in such cases, the deformation is effect on the slope of the curve is negligible as further pressure is
unlikely to be entirely elastic. One should recognize that Eq. 3 applied.
may not always be adequate. When any formation is exposed to the test, the formation will
In the case of both air in the lines and permeability losses, the compress in response to the applied pressure that will cause the
pressure vs. volume slope becomes nonlinear, and the “model” slope of the purple curve to decrease. The softer the formation
needs to be “run” in small timesteps to accurately compute the (lower modulus) and the greater the exposed formation surface
pressure vs. volume curve. Allowing for these effects, Eq. 2 now area, the more the slope of the line will decrease. In most cases,
becomes the amount of change in wellbore volume will be very small. This
response is also linear.
dP ¼ ðdVp  dVfm  dVk Þ=½ðVsys fpit  Va Þ  cm þ Va  ca ; If a permeable formation is exposed during the test, either
below or behind the casing, fluid can be lost into the formation af-
                   ð4Þ ter the well pressure is greater than the formation pore pressure.
The rate of loss will be a function of the permeability of the for-
where, at each small increment (d), the air volume (Va) and air
mation, the amount of exposed permeable surface area, and the
compressibility (ca) need to be recomputed and the permeability
difference between the formation and well pressure. As the differ-
losses (dVk) calculated. For modeling purposes, the air compressi-
ence between the well and formation pressure increases, the rate
bility at each step can be adequately computed with Boyle’s law.
of loss will increase, causing the slope of the line (light blue and
Thus, ca ¼ 1/Pa where, in this case, Pa is the pressure of the air
green lines) to curve toward a lower slope. In high-permeability
bubble as it is pushed down the well with each volume increment
formations and high differential pressures, it is possible for the
and must be absolute pressure, not gauge pressure.
slope to approach horizontal slope without introducing a fracture.
In the case of permeability losses, the computational effort is
The model allows all these effects to be taken into account
compounded in which the pump rate is neither continuous nor
simultaneously. There are other effects that may occur from time
constant. In the simpler case of constant, continuous pumping,
to time (e.g., movement of drillpipe, variations in pump speed,
increments in volume pumped and time are directly proportional.
pump-efficiency issues, and variations in fluid properties) that
The permeability losses to the formation (dVk) are described in
may impact the early curve behavior that the model does not
Eq. 5 with a basic radial-flow model (Smith 1994). More-complex
address. The use of the model can help identify those issues.
permeability models can be used, but the assumption of radial
steady-state flow is adequate for this application, given other
uncertainties such as permeability. Because the true permeability Application of the Model
(K), the length of flow surface area (h), and the viscosity (l) are The basic workflow to model the pressure vs. volume response of
not usually well-characterized, all three can be combined into a a formation-pressure integrity test (FPIT) should take the follow-
single flow factor (Ff) that can be adjusted to fit the observed data. ing general steps:
If it is important to break out the components of the flow factor, (1) Calculate the apparent mud compressibility from the cas-
then Eq. 5 can be used to accomplish that. The loss rate is directly ing test.

December 2014 SPE Drilling & Completion 433

ID: jaganm Time: 18:07 I Path: S:/3B2/DC##/Vol00000/140032/APPFile/SA-DC##140032

 DC167945 DOI: 10.2118/167945-PA Date: 10-December-14 Stage: Page: 434 Total Pages: 7

(2) Calculate the apparent mud response by use of the ex- properties can be cancelled by an opposing error in the assessment
posed volume at the time of the openhole test. of one of the other linear responses. However, the response to the
(3) Calculate the formation-compressibility response. presence of air trapped in the lines or to the presence of perme-
(4) Adjust the amount of air in the lines to fit the early upward ability is nonlinear, and the modeling response is, therefore,
curvature. unique.
(5) Adjust the formation-permeability parameters to fit the During the rig-up process, air can easily be trapped in surface
downward curvature. lines. The compressibility of air is inversely proportional to the
(6) Iterate with air, channel size, and permeability parameters (absolute) pressure. Air is highly compressible compared with ei-
until a best fit to the measured data is found. ther water or synthetic mud. As mud is pumped into the system,
The casing-test pressure vs. volume trend should form a the pressure increases, and the air is pushed down the well,
straight line. Irregularities commonly exist at the low-pressure increasing further the air pressure and decreasing proportionally
readings as the pumping unit operator brings the unit up to speed the air volume. The effect is that air trapped in lines starts off
and adjusts the rate to the desired target rate. These fluctuations in nearly flat and then curves upward rapidly. The effect typically
rate produce variations in the friction pressure that are most pro- becomes indistinguishable by the time 100 psi is applied. Because
nounced when observed with surface gauges, but are significantly the characteristic shape of the air-in-lines is so short-lived, it is
lower when observed with downhole gauges. A linear trend can quite easy to fit by simply adding air to the model. The typical
be determined from the stable higher pressure readings and can be quantity encountered, when present, is usually on the order of a
used in Eq. 1 to calculate the apparent mud compressibility. This few tenths of a barrel.
relationship also requires knowledge of the volume of the mud The presence of permeability results in the loss of mud to the
exposed to the casing test that is made up of the volume of the formation, which creates the downward curvature of the pressure
casing above the cement plug, the volume of the chokeline/kill vs. volume curve mentioned previously. This is the only attribute
line (if used), the volume of the surface plumbing between the that creates downward curvature, so it is uniquely matched by use
well and the cement unit, and the volume and displacement of any of the permeability attributes. By adjusting the flow factor (Ff)
drillpipe that may have been used during the test. and the pore pressure used to calculate the difference between
If the compressibility of the mud that is used during the open- well pressure and formation pore pressure (DPwf), one can opti-
hole test has been significantly altered from that of the mud that mize the fit of the downward curvature of the modeled data to the
was used during the casing test, then an alternative source for the observed data downward pressure. When the quality and sampling
mud compressibility will be required. Most of the major mud ven- rate of the measured data are high, the sensitivity of the model to
dors are able to model the compressibility of their mud products pore pressure can sometimes be good enough to derive an esti-
for a given well configuration. If this is impossible, values from mate of the formation pore pressure.
past experiences can be used. The authors have found that the typ- When trying to fit a data set, it is usually best to begin with the
ical compressibility of water-based mud is on the order of 1.8 to assumption that no channel or permeability is present. Include the
2.5  106 psi1, and synthetic mud is on the order of 3.0 to expected volume of the system, mud compressibility for the cas-
4.0  106 psi1. ing test, and expected formation compressibility from petrophysi-
The pressure vs. volume response to the mud in the well at the cal logs to establish the basic expected response. Then, one
time the FPIT is conducted can be modeled with Eq. 2. This cal- should adjust the air-in-lines to fit any upward curvature observed
culation will also require knowledge of the total volume of the in the early data. Next, the pore pressure and flow factor should
system when the openhole test is conducted, which will have be adjusted to fit the downward curvature. If a significant gap
changed from the volume present when the CIT was conducted. remains between the model data and the observed data, one
This volume can be estimated from an assessment of the casing should add channel volume. One should iterate fine adjustments
volume, openhole volume below the shoe, drillpipe displacement with the air-in-lines, flow factor, formation pore pressure, and
below the blowout preventer (BOP), the drillpipe capacity above channel length until a best match is achieved. When the pressure
the BOP, the chokeline/kill-line volume (if used), surface plumb- is reached at which a fracture is introduced, it creates a higher
ing volume between the drillpipe and the cement unit, and any loss rate as the fracture fills with mud. This introduces a change in
potential channel behind the casing exposed at the time of the the slope of the actual measured data. No adjustment to the model
test. After these inputs are known, the pressure change for each is possible that will mimic the effect of a fracture being intro-
increment in the pumped mud volume can be calculated, and a duced. The actual data will depart to a higher volume compared
comparison to the measured response can be made. If the volume with the model when the fracture starts.
of the system present during the CIT is less than the volume of The presence of a channel can affect the observed pressure vs.
the connected system present during the openhole test, then the volume behavior in a number of ways:
slope of the openhole test will be lower than the slope of the cas- • The additional volume resulting from the presence of the
ing test. If the volume of the system present during the casing test channel will cause the slope of the curve to decrease.
is greater than the volume of the connected system present in the • The increase in exposed formation surface area along the
openhole test, the slope of the openhole test will be higher than path of the channel will cause the slope to decrease further.
the slope of the casing test. This latter case can occur when signif- • The exposure of any additional permeable surface area will
icant casing volume is displaced by the presence of drillpipe dur- cause downward curvature.
ing the openhole test. • The exposure of shallower, potentially lower-pressured per-
When a significant amount of soft-formation surface area is meable formations that will cause more curvature downward.
exposed during the test, these formations can compress, increas- • The presence of contrasting-viscosity fluids in the channel
ing the volume of the system and decreasing the slope of the pres- will cause variations in the loss rate to permeable forma-
sure vs. volume curve. In most cases, the effect of formation tions, and these may change further as the well fluid displa-
compressibility will be negligible. However, if very soft forma- ces the fluid in the channel during the test.
tions are present, low-compressibility mud is being used, and a • Shallower formations will be exposed that may likely have a
large surface area is exposed because of long openhole sections lower fracture gradient.
below the shoe (or a large connected channel behind the casing), As a result of these potential conditions, it can sometimes be
then the impact could be noticeable. Formation compressibility difficult to fit the model to a test when a significant channel is
can often be calculated from petrophysical log data or from corre- present.
lations with effective stress (Mavko et al. 1998). One can infer two important conclusions from the comparison
The impact of the mud compressibility, changes in hole vol- of the model response to the actual data:
ume, and the exposure of a compressible formation are all linear • The point at which the fracture starts will be where the slope
and, therefore, nonunique. Errors in the evaluation of one of these of the actual data departs from the modeled data.

434 December 2014 SPE Drilling & Completion

ID: jaganm Time: 18:07 I Path: S:/3B2/DC##/Vol00000/140032/APPFile/SA-DC##140032

 DC167945 DOI: 10.2118/167945-PA Date: 10-December-14 Stage: Page: 435 Total Pages: 7

1800 1200

1600
1000

1400

800
1200

Pressure (psi)
Pressure (psi)

1000
600

800

400
600

400 Modified Casing-Test Data

200
PWD Test
Modeled Response
Casing-Test Data
Pump Rate (bbls/hr)
200 FPIT 1
Modeled Response
0
0 2 4 6 8 10 12
0
0 2 4 6 8 10 12 14 16 Volume (bbls)/Pseudo Time (bbls/min.min)
Volume (bbls)/Pseudo Time (bbls/min.min)
Fig. 6—Example with permeability present.
Fig. 5—Example of a classic leakoff test.
test. The APWD digital data were merged with the cement-unit
• When conducting tests in clastic sediments, the presence of digital data with the timestamps from each so that the observed
permeability infers that any measured fracture gradient is pressure from the APWD is tied to the measured pumped volume
likely to be in sand because sands are generally subject to a from the cement unit. The data are sampled in 10-second inter-
lower minimum Earth stress than shale, hence the success of vals. The test was run with synthetic mud in the well at a pump
the fracture-stimulation industry. rate of 0.3 bpm.
If permeability is not expected to be present in the open hole There is no linear characteristic to the pressure vs. volume
exposed below the shoe, the presence of permeability in the curve; it has a downward curvature from the very beginning of
response of the pressure vs. volume curve suggests that there may the test. The well pressure is greater than the pore pressure before
be a channel present that connects permeable formations behind any pressure is applied to the well. Therefore, losses are occurring
the casing to the test volume. This channel may identify the need before the test begins. The calculated loss rate on the basis of the
for remedial action to seal the channel. When this occurs, it may early time data is 5.4 bbl/hr. As pressure is increased, the loss rate
be possible to diagnose further the results by spotting a lost-circu- increases because of the higher differential pressure between the
lation pill at the bottom of the well and rerunning the test. If the borehole and the formation. When the fracture is started, the loss
results display an immediate impact of the presence of the pill, rate to the permeability has climbed to 7.66 bbl/hr.
the permeable formation is likely to be in open hole. If the test The optimum model fit to the measured data is found with
shows a delayed response to the pill, the permeability is likely 0.06 bbl of air present and an average mobility of 5.23 md/cp
behind the shoe, and the delay is associated with the displacement more than the 6.7 m of exposed formation below the shoe. One
of the pill up the channel to the permeable formation. can identify the fracture-initiation pressure as the point at which
the measured data depart from the modeled data (at 626 psi).
Fig. 7 shows a sequence of three leakoff tests run with a
Examples water-based mud. The data were hand recorded by the wellsite
Fig. 5 is a classic leakoff version of the formation-pressure integ- supervisor in 1/2-bbl increments, and the pump rate was 1/2 bpm.
rity test (FPIT) in which the test stopped as soon as departure The first test (blue curve) never reached breakdown pressure. The
from linearity was recognized. The slope of the test is very close test was immediately repeated with the same fluid in the well (ma-
to the slope of the casing test. There is a very slight upward curva- roon curve). The model was run (overlying the right-most red
ture at the very beginning of the test that indicates a very small curve) and found a good fit between the model buildup behavior
amount of air present. The best fit used 0.55 bbl of air to fit that and the recorded data. Deviation from the model would imply that
slight upward curvature. No permeability is required to model the a fracture was induced at 680-psi surface pressure.
response. The data were hand recorded by the wellsite supervisor The shoe was squeezed, and then the test was run a third time.
in 1/2-bbl increments at a pump rate of 1/2 bbl/min (bpm). The This test shows very little permeability present, with much closer
inaccuracies with the hand recording are likely the root cause of conformance to the casing-integrity test and a much higher leak-
the slight waviness of the measured data. This test was run with off pressure before the fracture is induced.
synthetic mud. Departure from linearity would be interpreted as Attempts to model the first curve were unable to obtain a good
being between 1,450 and 1,510 psi. A more-accurate determina- fit. It is believed that the difficulty of modeling is the result of a
tion cannot be made without more finely measured data. different fluid in the channel from that in the well. The model was
Fig. 6 is an example of a test run in the presence of a highly unable to obtain a good fit because the formation was exposed to
permeable formation exposed at the time of the test. The data are two different-viscosity fluids as new fluid displaced the fluids
the digital data taken directly from the annular-pressure-while- from the cement job. By the time the second test was run, most of
drilling (APWD) sensor in the measurement-while-drilling tool the annulus was displaced to the drilling mud, and a single fluid
that is downhole a few meters above the shoe at the time of the remained, giving a good fit to the data.

December 2014 SPE Drilling & Completion 435

ID: jaganm Time: 18:07 I Path: S:/3B2/DC##/Vol00000/140032/APPFile/SA-DC##140032

 DC167945 DOI: 10.2118/167945-PA Date: 10-December-14 Stage: Page: 436 Total Pages: 7

1400 700

Repeat FPIT Modeled Response

Modeled Response After Squeeze
Casing-Test Data
1200 Repeat FPIT (before squeeze)
600
Casing-Test Data
#1st FPIT (before squeeze)
#1st FPIT (before squeeze) Shut-in
Repeat FPIT (before squeeze)
1000 500
Repeat FPIT (before squeeze) Shut-in
FPIT 3 (after squeeze)

Pressure (psi)
Pressure (psi)

800 400

600 300

400 200

Casing-Test Data

200 FPIT Before LCM Pill

100
FPIT Before LCM Pill Shut-in
FPIT After LCM Pill
FPIT After LCM Pill Shut-in

0 0
0 10 20 30 40 50 60 70 0 5 10 15 20 25
Volume (bbls)/Pseudo Time (bbls/min.min)
Volume (bbls)/Pseudo Time (bbls/min.min)
Fig. 8—Tests with and without LCM.
Fig. 7—Tests with a channel present.

This test has a number of attributes that seem to indicate that a ids pass by it during the pumping of the first test [the second test
channel was present. The first attribute is the large volume of mud models well with one fluid (a single viscosity)]. The third attribute
required to reach maximum pressure on the first two tests (tests is the very high permeability required to fit the data.
stopped at 50 bbl pumped). Note that, after the squeeze, it took One proactive way to determine whether the permeability is
only 6 bbl to achieve the same pressure. The second attribute is below the casing shoe or in a channel behind casing is to run the
the poor repeatability between the first and second tests that sug- test at the beginning with the drilling mud in a conventional man-
gests that the permeable formation saw two different-viscosity flu- ner and then repeat the test with a lost-circulation-material (LCM)
pill spotted in the openhole below the shoe. The LCM should
have a contrasting higher viscosity to affect the loss rate into the
formation and should contain wellbore-strengthening or stress-
700 cage (Alberty et al. 2004; Aston et al. 2004) material to increase
the fracture resistance. If the pressure vs. volume curve immedi-
ately responds to the presence of the pill, the permeability is
below the shoe. If the response to the presence of the pill is
600
delayed, then the pill needs to be displaced up a channel behind
the casing to affect the losses into the sand.
Fig. 8 shows a case in which this was performed on a well.
500 The blue curve was run with a water-based mud. The data were
hand recorded by the wellsite supervisor in 1/2-bbl increments,
and the test was run at 1/2 bpm. The first test identified the pres-
ence of permeability through the characteristic downward curva-
Pressure (psi)

400 ture. The rig needed to know if a remedial cement job was
required to eliminate the possibility of a channel behind the
casing.
A 50-bbl LCM stress-cage pill containing 30 lbm/bbl of 50-
300
mm D50 calcium carbonate and 20 lbm/bbl of 150-mm D50 cal-
cium carbonate was spotted in the open hole, and the second test
(the maroon curve) was recorded. The response to the LCM pill
200 was immediate, and the conclusion was made that the exposed
permeability was in open hole below the shoe and a remedial
cement squeeze was unwarranted.
Fig. 9 displays the model fit to the second test run with the
100 Casing-Test Data LCM pill. There is a good fit of the model to the pressure vs. vol-
FPIT After LCM Pill ume data, with no air trapped in lines and 32.5 md/cp more than
Modeled Response the 2.7 m of open hole below the shoe. The model results further
0 support the conclusion that the sand is below the shoe in the
0 2 4 6 8 10 12 14 16 18 respect that there is no need to add volume to the system that
Volume (bbls)/Pseudo Time (bbls/min.min) would occur if a channel was present and the value for the perme-
ability required is consistent with expectations for permeable lith-
Fig. 9—Model fit to the LCM case. ologies at this depth and in this area.

436 December 2014 SPE Drilling & Completion

ID: jaganm Time: 18:07 I Path: S:/3B2/DC##/Vol00000/140032/APPFile/SA-DC##140032

 DC167945 DOI: 10.2118/167945-PA Date: 10-December-14 Stage: Page: 437 Total Pages: 7

Conclusions Vsys ¼ volume of the system exposed to the pressure changes

fpit
A model for analyzing formation-pressure-integrity-test (FPIT) during the FPIT
data were presented that one can use to predict the prefracture l ¼ fluid viscosity
behavior of formation-integrity tests. This model incorporates the
effects of mud compressibility, casing expansion, cement com- Acknowledgments
pressibility, air-in-lines, channel volume, formation compressibil-
ity, and permeability losses. The model simplifies the need to The authors wish to thank Hess Corporation for allowing the pub-
treat mud compressibility, casing expansion, and cement com- lication and presentation of this material.
pressibility separately by solving for an apparent mud compressi-
bility from the analysis of the casing test run before drilling out References
the shoe. Addis, M.A., Hanssen, T.H., Yassir, N. et al. 1998. A Comparison of
Examples were shown in which the model was compared with Leak-Off Test and Extended Leak-Off Test Data for Stress Estimation.
a range of field examples, including cases with channels and per- Presented at the 1998 SPE/ISRM Eurock Conference, Trondheim,
meable beds present. A method was demonstrated that one can Norway, 8–10 July. SPE-47235-MS. http://dx.doi.org/10.2118/47235-
use to determine whether permeable beds, recognized in the anal- MS.
ysis of the data, are below the shoe or behind the shoe. Alberty, Mark W. and McLean, Michael R. 2004. A Physical Model for
Understanding the presence of permeable beds and their posi- StressCages. Presented at the 2004 SPE Annual Technical Conference
tion in the well can be useful in making a decision about the need and Exhibition, Houston, Texas, 26–29 September. SPE-90493-MS.
for remedial cement operations and for using FPIT data in cali- http://dx.doi.org/10.2118/90493-MS.
brating fracture-gradient trends. Aston, M.S., Alberty, M.W., McLean, M.R. et al. 2004. Drilling Fluids for
Wellbore Strengthening. Presented at the 2004 SPE/IADC Drilling
Nomenclature Conference, Dallas, Texas, 20–22 February. SPE-87130-MS. http://
dx.doi.org/10.2118/87130-MS.
ca ¼ air-in-lines compressibility
Mavko, Gary, Mukerji, Tapan, and Dvorkin, Jack. 1998. The Rock Physics
cm ¼ apparent mud compressibility
Ff ¼ flow factor (Kh/l) Handbook, Cambridge University Press.
Postler, D.P. 1997. Pressure Integrity Test Interpretation. Presented at the
G ¼ formation shear modulus
h ¼ length of exposed permeable surface, part or all of 1997 SPE/IADC Drilling Conference, Amsterdam, The Netherlands,
4–6 March. SPE-37589-MS. http://dx.doi.org/10.2118/37589-MS.
which may be behind casing connected by a channel
Smith, James T. 1994. Petroleum Reservoir Engineering, Vol. 1, Course
HS ¼ hole diameter of the hole section below the casing
shoe during a formation-pressure integrity test (FPIT) Textbook.
Van Oort, Eric and Vargo, Richard. 2008. A Comparison of Leak-Off Test
K ¼ permeability
and Extended Leak-Off Test Data for Stress Estimation. Presented at the
LOH ¼ length of openhole below the shoe during an FPIT
P ¼ pressure applied during the casing-integrity test (CIT) 2008 SPE/IADC Drilling Conference, Amsterdam, The Netherlands,
2–4 March. SPE-105193-MS. http://dx.doi.org/10.2118/105193-MS.
or FPIT (can be surface or downhole)
Pa ¼ absolute pressure of the air-in-lines
re ¼ effective permeability radius (distance to undisturbed Mark Alberty is a Global Adviser with the Hess Corporation
formation pressure) where he is the Hess pore and fracture pressure Technical
Authority. His expertise includes petrophysics, formation pres-
rw ¼ wellbore radius
sure integrity testing, wellbore strengthening, shallow water
T ¼ time flow, and lost-circulation prevention and remediation. Alb-
Va ¼ volume of air trapped in the system erty’s 41 years of experience includes time with Schlumberger,
Vk ¼ permeability losses Gearhart Industries, and BP.
Vp ¼ volume pumped
Mike McLean is an independent pore pressure and geome-
DPw-f ¼ pressure difference between borehole and formation
chanics consultant who is based in the north of England. His
fluid expertise centers on pore-pressure and fracture-gradient
D Vfm ¼ change in total volume caused by formation com- assessment and prediction, but also includes drilling geome-
pressibility chanics, well integrity, shallow water flow, and seismic rock
Vsys cit ¼ volume of the system exposed to the pressure changes properties. McLean’s 30 years of experience was mostly with
during the CIT BP Exploration.

December 2014 SPE Drilling & Completion 437

ID: jaganm Time: 18:08 I Path: S:/3B2/DC##/Vol00000/140032/APPFile/SA-DC##140032