Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Summary
This paper discusses the theory and application of a pressure-pulse
testing technique to determine the fracture closure pressure (Pc) in
soft formations. This technique is intended for use only after a data
fracture has been pumped and is closed. Implementing the pulse
technique post-closure rather than while the fracture is still open
allows it to be used as an independent determination of Pc without
diminishing the accuracy of the fluid leakoff coefficient (CL) determined from the data-fracture. Previous pulse-testing techniques1
require pumping into a fracture while it is still open. This practice
reduces the accuracy of CL by altering the time necessary for the
fracture to close (tc). Furthermore, previous pulse-testing techniques yield only a possible range for Pc, whereas the post-closure
pulse technique yields a better-defined, singular value of Pc.
The accuracy of post-closure pulse testing is validated with
several field examples from a soft formation on Alaskas North
Slope by comparing a definitive Pc determined using pressure
decline analysis with Pc determined from post-closure pulse testing.
Introduction
Determining an accurate Pc is central to the correct design and
execution of a fracture stimulation. This is especially true in highpermeability, soft rock applications in which obtaining a tip
screenout (TSO) is necessary to generate sufficient fracture width
and to maximize fracture conductivity. Obtaining a TSO at the
correct time is heavily dependent on an accurate estimate of CL,
which, in turn, is based on an accurate determination of Pc. Several
techniques are available for determining an onsite Pc before executing a fracture job. The most common are constant-rate flowback testing, pressure decline analysis, and pulse testing. Pulse
testing, as described by Wright et al.,1 is commonly used in conjunction with other methods to determine Pc. In Wrights method,
pulse testing is used only as a means of bracketing a range of
possible fracture closure pressures. Other methods must then be
relied upon to determine the actual value of Pc within that range.
This paper goes beyond Wrights initial concept to demonstrate
that post-closure pulse testing can be used in soft formations to
actually determine a singular value for Pc (not just a range of
possibilities), resulting in a more accurate determination of CL.
Post-closure pulse testing entails nothing more than pumping a
small (1 to 2 bbl) volume of low-viscosity fluid (water or linear
gel) into a closed, pre-existing fracture at a relatively low rate (5
BPM). The pressure at shutdown is then evaluated to determine Pc.
It is argued that pumping into a rock and evaluating the pressures
at shutdown provides nothing more than an initial shut-in pressure
(ISIP), in which the ISIP is simply the sum of the net pressure in
the fracture (Pnet) and Pc. Thus, Pc is indeterminate from such a
test. However, if the test is performed under conditions such that
Pnet is very small, the ISIP becomes a very close estimate of Pc.
This is the basis for the post-closure pulse-testing technique presented herein.
Post-closure pulse testing requires that a moderate size datafracture (usually 100 bbl) be pumped into the formation before
Theory
Overview. The first key to successfully applying post-closure
pulse testing is to generate an initial fracture with larger dimensions than would be possible with just the small volume of a single
pulse. This is done on the North Slope by pumping a 100- to
200-bbl data-fracture. Generating such a large fracture before
pumping pulses accomplishes several things. First, the pulse fluid
is able to rapidly disperse by traveling along the pre-existing fracture network. Traveling down a pre-existing fracture allows the
pulse fluid to move unimpeded by tip effects, such as fracture
toughness (KIc) and rock tensile strength (To). Because To and KIc
are equal to zero under such conditions, the pulse fluid disperses
down the pre-existing fracture length more rapidly than would
otherwise be possible. Second, dispersing the pulse fluid in a larger
area reduces the average fracture width ( wf), which, in turn, minimizes Pnet. Third, dispersing the pulse fluid in a larger area increases fracture compliance (c), which also minimizes Pnet. Finally, because there are no tip effects (To and KIc) for a fluid pulse
traveling down an existing fracture, linear deformation equations
can be used to describe the Pnet of the fluid pulse accurately.
The second key to successfully applying post-closure pulse
testing is to minimize both near-wellbore fracture tortuosity and
multiple fractures. Fracture tortuosity and multiple fracture generation cause a significant pressure drop at the wellbore, resulting
in erratic pulse ISIPs. This has been the case with every postclosure pulse test tried without first taking corrective action to
minimize these two problems. In each case, once these problems
were addressed and near-wellbore friction pressure was reduced,
post-closure pulses yielded very consistent ISIPs.
Fracture tortuosity and multiple fracture generation are both
minimized by pumping slugs of low-density proppant slurry into
the perforations that eventually are to be fracture stimulated. This
technique was first presented by Cleary et al.2 The rationale for
using low-density proppant slugs is two-fold. First, the proppant
abrades tight, tortuous paths, resulting in reduced friction pressure.
Second, the proppant slugs help minimize the generation of multiple fractures by plugging off smaller fractures. In doing so, fluid
flow to more dominant fractures is enhanced, thereby resulting in
fewer, wider fractures and, consequently, reduced friction pressure.
The third key to successfully applying post-closure pulse testing is to use water or a low-viscosity linear polymer gel as a pulse
230
www.petroman.ir
16rf
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2)
3E
However, the pressure gradient along a fracture is not zero. Therefore, the net-pressure varies from some minimum at the fracture tip
to a maximum at the wellbore (Pnet). Pnet and Pnet are related by
^
Pnet
, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (3)
s
Pnet =
in which s a multiplying parameter defined at pump shutdown.5 For radial and Geertsma-de Klerk (GdK) geometries, this
multiplying parameter is
wf
1 bbl
, . . . . . . . . . . . . . . . . . . . . . . . . . . . . (7)
12 5.615 ft3
which reduces to
Vpulse =
rf2 wf
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (8)
21.45
Solving Eq. 8 for rf and wf and substituting both results into Eq. 6
finally yields the equations linking radial fracture geometry to Pnet
and Vpulse.
wf =
PnetV 2pulse
and rf =
3,438.6
2/3
, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (9)
341,489.2Vpulse
Pnet
1/3
. . . . . . . . . . . . . . . . . . . . . . . . . . (10)
231
www.petroman.ir
Fig. 2Pressure and pump rate curves for data fracture and
pulse.
232
www.petroman.ir
Fig. 5Pressure and pump rate curves for scour stage and
pulses.
Fig. 6Pressure vs. t plot.
233
www.petroman.ir
Fig. 9Pressure and pump rate curves for data fracture and
pulses.
Fig. 10Pressure vs. t plot.
References
1. Wright, C.A. et al.: Robust Technique for Real-Time Closure Stress
Determination, SPEPF (August 1996) 150.
2. Cleary, M.P. et al.: Field Implementation of Proppant Slugs To Avoid
Premature Screen-Out of Hydraulic Fractures With Adequate Proppant
Concentration, paper SPE 25892 presented at the 1993 SPE Rocky
Mountain Regional/Low Permeability Reservoirs Symposium, Denver,
1214 April.
3. Upchurch, E.R.: Flexible, Real-Time Tip-Screenout Fracturing Techniques as Applied in California and Alaska, paper SPE 54629 presented
at the 1999 SPE Western Regional Meeting, Anchorage, 2628 May.
4. Recent Advances in Hydraulic Fracturing, J.L. Gidley et al. (eds.),
Society of Petroleum Engineers, Richardson, Texas (1989) 300303.
5. Nolte, K.G.: Determination of Fracture Parameters From Fracturing Pressure Decline, paper SPE 8341 presented at the 1979 SPE
Annual Fall Technical Conference and Exhibition, Las Vegas, Nevada,
2326 September.
hf
c=
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A1)
2E
Substituting Eqs. 1 and A-1 into Eq. 3 yields
Pnet =
wfE
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-2)
6shf
s =
2n + 2
, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-3)
2n + 3 + a
s = 0.8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-4)
If a pulse of fluid volume Vpulse is introduced into a PKN fracture,
Vpulse = hf2Lf
wf
1 bbl
, . . . . . . . . . . . . . . . . . . . . . . . . . (A-5)
12 5.615 ft3
which reduces to
wf =
33.69Vpulse
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A6)
hf Lf
2.234VpulseE
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A7)
hf2Lf
If we now assume Pnet5 psi and Vpulse2 bbl, Eq. A-7 can be
restated as
hf2Lf = 0.894E. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-8)
Eq. A-8 provides a reasonable guideline for the minimum PKN
fracture dimensions a data-fracture volume must create so that a
2-bbl pulse will generate a Pnet no greater than 5 psi.
Analysis of GdK Geometry. A guideline for applying post-closure
pulse testing when GdK fracture geometry is expected is as follows.
The compliance of a GdK fracture is
c=
Lf
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-9)
2E
234
www.petroman.ir
Fig. 13Pressure and pump rate curves for data fracture and
pulses.
wf E
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (A-10)
6s Lf
235
www.petroman.ir