You are on page 1of 12

SPE 95287

Measurement of Proppant Transport of Frac Fluids


P.C. Harris, R.G. Morgan, and S.J. Heath, Halliburton
Copyright 2005, Society of Petroleum Engineers

This paper was prepared for presentation at the 2005 SPE Annual Technical Conference and
Exhibition held in Dallas, Texas, U.S.A., 9 12 October 2005.

This paper was selected for presentation by an SPE Program Committee following review of
information contained in a proposal submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to a proposal of not more than 300
words; illustrations may not be copied. The proposal must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract
The object of hydraulic fracturing is to produce a propped
fracture extending from the wellbore. Extensive time and
effort are expended in measuring rheological properties of
crosslinked fluids (without proppant) under simulated
downhole temperature and shear-history conditions.
Normally, it is assumed that if a fluid meets certain minimum
viscosity conditions, it will transport proppant successfully.
Measurements of actual proppant transport under dynamic
simulated downhole conditions have been attempted and
described, but such measurements have been very
cumbersome and require significant equipment and expense to
execute. Such measurements are well beyond the scope of
routine QA/QC analyses.
A proppant viscometer recently constructed can measure
fracturing fluids containing propping agents across a wide
range of concentrations. The device was designed to work as a
conventional Fann Model 50-type viscometer. This unique
viscometer has been used to measure typical fracturing fluids
containing realistic concentrations of proppants, at
temperatures and times up to several hours, representative of
actual fracturing treatments. The measurements show regions
of elastic transport typical of viscoelastic fracturing fluids
where proppant is transported efficiently, usually followed by
regions of purely viscous transport where proppant slowly
settles.
The advantage of the new proppant viscometer is that all
components of a fracturing fluid, including proppant, can be
tested. Shear-history effects of proppants on frac fluids are
usually unknown or ignored, but such effects were observed in
this work. Breakers were usually added to the fluids, showing
reasonable times when the crosslinked fluid no longer
transported proppant efficiently and proppant began to settle.
This paper shows how different types of fracturing fluids
can support proppant based on their chemical type, i.e. metal
and borate crosslinked fluids, linear gel fluids, and surfactant
gel fluids. Proppant concentrations are also considered. The
physical characteristics of the proppant viscometer are also
addressed.

Introduction
Most hydraulic-fracturing treatments use a gelled fluid to
create a subterranean fracture and partially fill the fracture
with propping agents. When the fracture closes and the fluid is
recovered, a conductive channel into the reservoir remains.
Proppant transport is a function of (1) wellbore and fracture
geometry; (2) volumetric rate; (3) proppant size, concentration
and specific density; and (4) carrier-fluid rheology.
Instruments such as Fann Model 50 viscometers are available
for measuring viscosity at high temperatures and pressures,
but elasticity is much more elusive to measure. Additionally,
most viscometers, such as the Model 50, are designed only to
handle the clean fluid systems, e.g. without proppant. By
default, we generally assume that higher viscosities will do a
better job of transporting proppant, as well as generating the
desired fracture geometry. There have been several attempts
1-9

to characterize transport properties of fluids using modified
bench-scale viscometers or through large slot or pipeline
apparatus, but those efforts are very expensive and are not
performed using routine quality-assurance measures.
Several rheological properties directly impact a frac fluids
performance: (1) apparent viscosity, (2) yield stress, (3)
dynamic viscosity, (4) rheomalaxis (irreversible thixotropy),
(5) viscoelasticity (for example G, G), and (6) the related
issue of turbulent-drag reduction. In laboratory research,
sample volume is often very limited, thus necessitating
rheological testing and evaluation of small quantities. Also,
most bench-top rheometers use batch mode, that is, small
samples are placed in a testing chamber as opposed to flow-
through testing, as is the case for pipe viscometers. This
presents the challenge of simultaneously:
1. Imparting viscometric shear history that simulates the
wellbore travel path.
2. Not exceeding the proper mechanical energy input;
the bench-top batch process should impart about the
same amount of integrated work as the wellbore path.
3. Maintaining satisfactory thermal balance, e.g. being
sure not to create localized hot spots in the bench-
top process because of its batch mode of operation.

In the case of most polymer-based fracturing fluids, the
capability to transport proppant is directly related to their
rheological equations of state (RES). Extent of crosslinking,
breaking, shear history, and volume-average shear rate are
2 SPE 95287
major factors affecting a fluids RES.
10-11
As an illustration,
Table 1 is a summary showing how fracture height and width
affect volume-average shear rates for a Newtonian fluid.
Where viscous drag dominates, as in the classical case of a
two-winged vertical fracture with parallel-plate geometry, the
challenge in proppant transport is to ensure that vertical
settling time is much greater than horizontal travel time.
Sufficient vertical settling time allows the particle to reach a
maximum horizontal distance, thus avoiding a duning effect.
Table 2 is a simple illustration of the ratio of horizontal
velocity to vertical settling velocity of 20/40 sand being
carried by a 500 cp Newtonian fluid. Table 2 illustrates the
sensitivity of proppant transport to fracture geometry. The
shaded areas were selected as examples in which the
horizontal transport time was at least 50 times greater than the
vertical settling times.
In the case of crosslinked gels, the elastic forces are
designed to dominate, preventing any substantial viscous
settling during the fracturing and placement of proppant.
Morris
12
presents the case that a minimum value of G
(oscillatory elastic modulus) of 10 to 12 Pas is sufficient for
most 20/40 frac sands. The complex viscoelastic nature of
crosslink fracturing fluids presents a dilemma for the
fracturing rheologist. Conventional rheometers are designed
for measuring viscoelastic properties through well-controlled
oscillatory deformations that are small, non-destructive, and
within the linear elastic range. However, the actual fracturing
process involves large amounts of shear strain of multiple
orders of magnitude, well beyond the linear elastic range. As
the breaker reaction begins to dominate, the transport
mechanism shifts from elastic to viscous, leading to settling
caused by the low viscosity of the broken fluid system.
The Fann Model 50 viscometer was designed for
characterizing fracturing gels under simulated downhole
temperature-time conditions. However, the Model 50 and most
other bench-top viscometers/rheometers are not adequately
equipped to handle proppant-laden fluids. In the case of
concentric cylinders, the centrifugal forces tend to stratify the
particles, thus resulting in nonrelevant data. In cone-plate and
plate-plate viscometers, the small gaps necessary to provide
torque sensitivities result in particle jamming. Additionally,
the large density differences between most proppants and
conventional fracturing fluids result in settling during testing,
thus producing unreliable results.
Many researchers have modeled and/or measured the
effects of fluid rheology and particle size, shape, and density
on suspension attributes in either single-particle systems or at
dilute concentrations, in which particle-to-particle interactions
are neglected. Others have attempted to include the impact of
particle-to-particle collisions.
13-15

Morgan et al.
16
presented a bench-top device and model for
predicting the apparent viscosity of proppant-laden fluids for
Newtonian and Power-law fluids with a wide range of particle
sizes and concentrations. Their work was based on a
viscometer uniquely designed to keep highly concentrated
dense particles suspended in fluids while measuring volume-
averaged shear stresses and shear rates. The work was
limited to atmospheric temperature and pressure and did not
include crosslinked gels. Table 3 summarizes the experimental
limits of the work. The impact of particle concentration
followed Huggins theory, that is, that apparent viscosity of the
bulk slurry was a function of particle concentration, as shown
in Fig. 1.
The proppant viscometer described here measures in the
highly concentrated range with 7.5 power, according to Fig. 1.
We have designed and tested an apparatus capable of routine
measurement of gelled fracturing fluids containing wide range
of proppant particles under downhole temperatures, ranging
from linear to fully crosslinked, to broken gels. By measuring
various types of fracturing fluids, it is possible to classify the
transport properties of each fluid type. By adding breaker
chemicals to fluid, it is possible to determine the length of
time a fluid maintains its transport capability, or whether the
proppant will settle quickly.

Experimental
Base gel fluid was mixed in a blender with polymer, buffers,
surfactant, and clay control in tap water. A 500-mL sample of
base gel fluid and 725 g of 20/40 mesh proppant (145 g/100
mL) were stirred at moderately high speed with an overhead
stirrer. Reactive additives, i.e., crosslinker and breaker, were
added while stirring. The slurry was quickly transferred to the
cup and placed on the proppant viscometer. The transferring
operation required about 1 minute. The cup was rotated at 25
RPM, with an estimated volume-average shear rate of 6 sec-1.
Torque was monitored while the sample was heated at
5F/min. Volume-average viscosity (VAV) and temperature
data were plotted vs. time.

Description of Proppant Viscometer
The proppant viscometer
17
is based on a Fann Model 50-type
viscometer with modified bob and cup. The proppant
viscometer has an inner sensing surface composed of flags, in
place of the concentric cylinder design of the Model 50
viscometer (see Fig. 2). To aid mixing within the cup,
protrusions were welded to the inner walls of the cup. Proper
clearances were provided between the flags and the wall
extensions to help prevent proppant from sticking between the
surfaces. As the flags on the moving cup sweep by the static
flags on the central sensing shaft, a cyclic torque is measured.
The oscillating force contains both viscous and elastic
components.
The proppant viscometer has complicated geometric
surfaces, making calculation of shear rate difficult. However,
laboratory personnel approximated volume-average shear rate
by studying known fluids. The resistance to movement of a
slurry increases as proppant concentration increases.
Placement of the mixing and sensing surfaces toward the
bottom of the cup allows greater sensitivity to concentrating of
proppant as it settles. An increase in torque indicates that
proppant is settling.

Theory
Viscometry studies by Steffe and Morgan
18-20
were used to
develop the following equations for estimating volume-
average shear rates, shear stresses, and viscosities for the
device described in Fig. 2.

) (
1
RPM k
vol
= & ...................................(1)
SPE 95287 3
) (
2
Torque k
vol
= ............................... (2)

=
vol
vol
vol

&
........................................ (3)
where:
vol
& is the volume-average shear rate within the novel
viscometer;
vol
is the volume-average shear stress; and
vol

is the volume-average viscosity (VAV) within the modified
Model 50 viscometer. The value of is unique to the
geometry of the device and slightly related to the sample
volume, while the value of is unique to the torque-
measuring device. Newtonian viscosity standards and Power-
law fluids were used to quantify the values of and for
the fracturing fluid/proppant systems used in this study.
1
k
2
k
1
k
2
k

Proppant Viscometer Response for Linear Gel Fluids
The volume-average viscosity (VAV) is a product of fluid
viscosity, particle concentration and cup RPM. When the sand
concentration was held constant at 12 lb/gal and non-
crosslinked polymer concentration varied, two separate effects
were observed in the curves, as shown in Fig. 3a. The most
noticeable effect is that the slope of the curve decreased as
polymer concentration increased. Once the elastic transport
forces no longer dominate, the particles begin to settle at a rate
dependent upon their size, density, concentration, and the
viscosity of the fluid medium. The slope of the VAV vs. time
curves are defined as rate of viscous settling (RVS) values,
which are plotted in Fig. 3b. Proppant settled very rapidly for
the 40-lb/Mgal base-gel fluid at 75F and 6 sec-1; whereas
settling may not have been complete in the 100-lb/Mgal fluid
until about 4 hours. The slope of these lines is useful to
describe purely viscous settling behavior in more complex
fluid systems, such as those with crosslinker and breakers.
The second effect is that higher polymer concentrations
have a higher torque at zero time, where all the proppant is
evenly distributed throughout the fluid. This zero-time
intercept value is a function of base-gel viscosity and particle
concentration. Surfactant gel fluids behaved similarly to non-
crosslinked polymer fluids. Micelles and vesicles that
comprise surfactant gels are analogous to polymer gels of high
molecular weight, whose viscous behavior depends solely on
hydrodynamic volume interference and limited entanglement,
rather than upon crosslinking.

Proppant Viscometer Response for Crosslinked Gels
Crosslinked gels respond very differently from linear-base
gels. Their lower RVS slope indicates that even low polymer
concentrations of crosslinked fluid may transport better than
high polymer concentrations of noncrosslinked fluids. This
fact indicates a fundamental difference between the fluids.
Noncrosslinked fluids support particles on the basis of viscous
forces; whereas crosslinked fluids are tied together in a
network structure. The network provides elasticity to the fluid
structure.
The initial decline in the proppant viscometer response for
crosslinked slurry (lowest curve of Fig. 4) is caused by
temperature thinning of the supporting fluid. Once
temperature is reached (upper curve), the VAV maintains a
stable reading, decreasing only slightly as the breaker works to
reduce the polymer network. This stable, low-VAV response
is the "elastic transport region," where virtually perfect
proppant transport occurs. Eventually, the breaker acts to
sever sufficient network connections and allows particles to
fall, resulting in the "onset of sand settling." Settling usually
begins at a low rate, which may increase as the breaker
continues to decrease the degree of networking in the polymer
fluid. This response is the "viscous settling region," where
settling approximates that of a non-crosslinked polymer
solution. The slope in this region may be compared to RVS
slopes of Fig. 3 as an indicator of the viscous state of the fluid.
Another curve may be superimposed on the VAV curve of
Fig. 4. The dashed line (middle curve) represents a typical
Fann Model 50 viscometer response of the same sample
without sand. This curve shows a decline in viscosity because
of the effects of chemically breaking the polymer. By itself,
this curve shows no indication of whether the fluid will
transport sand. By coupling the Model 50 curve with the sand-
transport curve, it is possible to correlate viscosity with the
time the onset of sand settling occurs. Fluids may then be
compared to determine the minimum viscosity at which they
will give proppant transport. However, both fluids must have
similar shear histories during the crosslinking reaction for this
procedure to be valid.
A clear plastic cup was constructed for the proppant
viscometer to allow visualization of proppant settlement to
coincide with torque measurements. The clear cup reinforced
the VAV measurements to give greater confidence in the
technique. Fig. 5 shows a low-temperature, metal-crosslinked
fluid with an enzyme breaker. Photographs of the clear cup
containing the slurry are superimposed on the plot of VAV at
corresponding times. The arrows point to the top of the
proppant bed. As the proppant settled from breaker activity
over time, the VAV value increased, thus validating similar
measurements at higher pressure and temperature in the
stainless steel cup.
Fig. 6a shows a concentration series of low-polymer
metal- crosslinked fluids containing oxidizer breaker at 250F.
The initial low-viscosity region parallel to the time axis is the
elastic transport region, where virtually no settling occurs.
Proppant is supported by a network structure having crosslink
bonds between polymer strands. As the viscometer cup rotates
at 25 RPM, the network stretches for about sec and relaxes
for sec alternately, producing torque oscillations. For the
metal-crosslinked elastic fluid, the oscillations have small
amplitude. Breaker actively reduces the network structure
within the elastic region until proppant is no longer supported
perfectly and the onset of sand settling is reached. Following
that point, we encounter viscous settling, where the increasing
slope indicates that proppant concentration is increasing at the
bottom of the cup. The slopes of the VAV vs. time curves in
Fig. 6a were computed and are plotted in Fig. 6b to illustrate
the sensitivity of this novel proppant viscometer to
characterize both the elastic proppant transport regime and
also the rate at which proppant settles during breaking. For the
viscous settling region, the amplitude of the torque oscillations
is greater and increases as proppant settles out. Oxidizer
4 SPE 95287
breaker was a practical addition to these tests, because without
breaker, the onset of sand settling would be much longer.
The capability of the metal-crosslinked fluid to transport
proppant elastically (judged by constant VAV value) was not
improved by higher polymer concentration up to the onset of
sand settling. The higher polymer concentration did increase
the time to arrive at the onset of sand settling. Undoubtedly
the fluids with higher polymer concentration had more
molecular interaction or entanglement, and longer time was
required for the breaker to reduce those network interactions
to the point of releasing sand for settlement.
Fig. 7a shows the effect of breaker concentration on a
metal-crosslinked fluid at 250F. With no oxidizing breaker
present, the fluid continues to support proppant elastically for
an extended period. When breaker is added, transport
capability is relatively sensitive to breaker concentration,
indicating that accuracy of metering the breaker is important
to good proppant placement. Once again, the RVS were
computed and are shown in Fig. 7b. Note the sensitivity of
the proppant viscometer to detect changes in dynamic settling
of the proppant.
Another breaking mechanism for metal-crosslinked fluids
is acid hydrolysis. Fig. 8 shows a fluid that can be crosslinked
from pH 5 to 10. The fluid can be readily stabilized at neutral
and alkaline pH values, but acid hydrolysis will eventually
break down the polymer network. The fluid character changes
from elastic to viscous transport, allowing proppant to settle.
The appearance of proppant transport plots for borate-
crosslinked fluids differed significantly from plots for metal-
crosslinked fluids (Fig. 9). Short-term oscillations for these
fluids at 100F were more intense than for metal-crosslinked
fluids, indicating greater elasticity. The intensity of the
oscillations makes the curves more difficult to interpret.
Approximate VAV values for proppant transport (22,500 cP)
and proppant settling (37,500 cP) are indicated by dashed
lines. Below the lower line, proppant is transported elastically.
Above the upper line, proppant is being deposited. Between
the two lines, proppant settles in a viscous mode. Very low
polymer concentrations, i.e. 15 lb/Mgal, showed elastic
transport for a very short period of time, then the proppant
settled very quickly. At 20 lb/Mgal, the fluid maintained
elastic transport for about 2 hours before rapid settling
occurred. The 25-lb/Mgal fluid without breaker transported
proppant for almost 3 hours before a rapid increase could be
detected; the transition from elastic to viscous transport is not
very obvious.

Discussion
Data for fracturing fluids in the proppant viscometer illustrate
three major fluid classes having differences in capability to
transport proppant. The first class of fluids does not have a
network structure and does not transport proppant perfectly.
These fluids include Newtonian fluids (oil and water), and
non-Newtonian, non-crosslinked polymer fluids, oil gels, and
surfactant gels. This first class transports by viscosity alone
and settling occurs during transport, depending upon flow rate
and viscosity. It may be possible to transport proppant
efficiently with combined rate and viscosity if the settling time
is not too long, but such a combination would likely result in
very high pumping pressures.
A second class of fluids is the metal-crosslinked group.
These fluids exhibit viscoelasticity. When polymer
concentration and crosslink density are high enough, proppant
is transported perfectly, without settling, usually at a low
pumping pressure. Once the perfect transport threshold is
achieved, adding more polymer or crosslinker to increase
viscosity does not improve transport, but may increase
pumping pressure. Some breaker mechanism is required to
initiate disruption of the network structure and allow settling
to begin.
A third class of fluids is the borate-crosslinked group.
Borate fluids have a transient network structure. The state of
the crosslink network is dependent upon shear rate as well as
fluid composition. When borate fluids are static long enough
to become relaxed, settling of proppant may occur in the
viscous settling mode. However, when some small shear rate
is imposed on the fluid, elasticity increases and the fluid acts
to transport perfectly. Some minimum degree of shear must be
maintained on borate fluids to maintain the elastic network
condition.
Foamed fluids were also measured on the proppant
viscometer. Foams made with noncrosslinked polymers
exhibited viscous settling, but at a slower rate than the neat
liquid phase alone. The retarded settling of 70% quality foams
indicates the presence of some fluid structure, although not
necessarily the same as a crosslinked polymer network. It may
be that the retardation is an indication of a yield point in the
foam. One would expect that crosslinked foams would
transport at least as well as the external liquid-phase gel.
The proppant viscometer shows that borate-crosslinked
and metal-crosslinked fluids are very different in structure and
in capability to transport proppant. It also indicates that
measurements of viscosity, such as in a couette viscometer, do
not forecast the capability of a fluid to transport proppant. We
found that low-polymer metal crosslinked fluids can transport
proppant perfectly for long periods of time, even though their
couette viscosity readings are not very high. On the other
hand, low-polymer borate fluids produce very impressive
couette viscosity values, often 10 times as high as the metal-
crosslinked fluids, but their proppant transport may not be as
good.

Conclusions
A modified Fann Model 50-type viscometer has produced
measurements indicating the capability of a fracturing fluid to
transport proppants. The proppant viscometer can further
distinguish between fluid types having either viscous or elastic
components. Proppant is included as a component of the fluid,
so chemical effects from the presence of proppant material can
be observed.
Three main classes of fracturing fluids were described
according to transport capability:
Purely viscous fluids. These fluids include
Newtonian fluids (oil and water), non-Newtonian
non-crosslinked polymer fluids, oil gels, and
surfactant gels. This class transports by viscosity
alone and settling occurs during transport according
to flow rate and viscosity. In these cases, it is
important to ensure that the ratio of the fracture-
SPE 95287 5
transport velocity is much greater than the vertical-
settling velocity.
Metal-crosslinked viscoelastic fluids. Metal-
crosslinked fluids have a permanent network
structure. When polymer concentration and crosslink
density are high enough, proppant is transported
efficiently, without settling,
Borate-crosslinked viscoelastic fluids. Borate-
crosslinked fluids have a transient network structure.
The state of the crosslink network is dependent upon
shear rate as well as fluid composition. Some
minimum degree of shear must be maintained on
borate fluids to maintain the elastic network
condition.

Acknowledgements
The authors would like to acknowledge the support of
management of Halliburton for the opportunity to present this
work. Special thanks are also due to Harold Walters, Billy
Slabaugh, Johnny Johnson, and David Barrick for their
contributions to constructing the apparatus and implementing
the experimental work.

References
1. Biot, M.A. and Medlin, W.L.: Theory of Sand
Transport in Thin Fluids, paper SPE 14468
presented at the 1985 SPE Annual Technical
Conference and Exhibition, Las Vegas, NV, 22-26
September.
2. Prudhomme, R.K.: Rheology of Fracturing Fluid
Slurries: Topical Report, Report No. NTIS PB94-
159183, January 1992, Gas Research Institute.
3. Fan, Y. and Holditch, S.A.: Testing Crosslinked
Fluids on the Fann 50 Viscometer Using The
Volumetric Average Shear Rate, paper SPE 26895
presented at the 1993 SPE Eastern Regional Meeting,
Pittsburgh, PA, 2-4 November.
4. Penny, G.S., Conway, M.W., and Schraufnagel, R.A.:
The Evaluation of Proppant Transport and Cleanup
of Foamed Fluids Used in Hydraulic Fracturing of
Shallow, Water-Sensitive Reservoirs, paper SPE
26923 presented at the 1993 SPE Eastern Regional
Meeting, Pittsburgh, PA, 2-4 November.
5. De Kruijf, A., Davies, D.R., and Fokker, P.A.:
Novel Rheological Equipment for the
Characterization of Hydraulic Fracturing Fluids,
paper SPE 27398 presented at the 1994 SPE
Formation Damage Control International
Symposium, Lafayette, LA, 9-10 February.
6. Goel, N., Willingham, J.D., Shah, S.N., and Lord,
D.L.: A Comparative Study of Borate-Crosslinked
Gel Rheology Using Laboratory and Field Scale
Fracturing Simulations, paper SPE 38618 presented
at the 1997 SPE Annual Technical Conference and
Exhibition, San Antonio, TX, 5-8 October.
7. Thesing, A.: New Device for Rheology
Measurements of Proppant-Laden Fluids with the
Fann 50 Viscometer, paper SPE 58759 presented at
the 2000 SPE International Symposium on Formation
Damage Control, Lafayette, LA, 23-24 February.
8. Tremblay, B., De Rocco, M., Ridley, R.K., Scott, K.,
Singh, S., Browne, D., Lukos, B., and ONeil, B.: A
Device and Method of Determining the Rheological
Quality of Gelled Fracturing Fluids, paper CIM
2000-58 presented at the 2000 Annual CIM
Petroleum Society Technical Meeting, Calgary,
Canada, 4-8 June.
9. Asadi, M., Conway, M.W., and Barree, R.D.: Zero
Shear Viscosity Determination of Fracturing Fluids:
An Essential Parameter In Proppant Transport
Characterizations, paper SPE 73755 presented at the
2002 International Symposium and Exhibition on
Formation Damage Control, Lafayette, LA, 20-21
February.
10. Harris, P.C. and Walters, H.G.: Real-Time Control
of Low-Polymer Fracturing Fluids, paper SPE
63238 presented at the 2000 SPE Annual Technical
Conference and Exhibition, Dallas, TX, 1-4 October.
11. Walters, H.G., Morgan, R.G., and Harris, P.C.:
Kinetic Rheology of Hydraulic Fracturing Fluids,
paper SPE 71660 presented at the 2001 SPE Annual
Technical Conference and Exhibition, New Orleans,
LA, 30 September-3 October.
12. Morris, J.F.: Personal communication, City College
of New York (2004).
13. Morris, J.F. and Brady, J.F.: Pressure-driven Flow
of a Suspension: Buoyancy Effects, Int. J.
Multiphase Flow, Vol 24: 105-130, 1998.
14. Asmolov, E.S.: The Inertial Life on a Spherical
Particle in a Plane Poiseuille Flow at Large Channel
Reynolds Number, J. Fluid Mechanics, Vol 381: 63-
87, 1999.
15. Cherukat, P. McLaughlin, J.B. and Dandy, D.S.: A
Computational Study of the Inertial Lift On a Sphere
in a Linear Shear Flow Field, Int. J. Mulltiphase
Flow, Vol. 25: 15-33, 1999.
16. Morgan, R.G., Lord, D., Kannan, R. and Srinivasa,
A.R.: Rheology of Highly Concentrated Particle-
Liquid Systems, presented at the 74th Annual
Meeting-Society of Rheology, Minneapolis, MN 13-
17 October, 2002.
17. Walters, et.al.: Testing Device and Method for
Viscosified Fluid Containing Particulate Material,
U.S. Patent 6,782,735, Aug 31, 2004.
18. Steffe, J.F. and Ford, E.W.: Rheological Techniques
to Evaluate the Shelf-Stability of Starch-Thickened,
Strained Apricots, J. Texture Studies. 16:179-192.
(1985).
19. Steffe, J.F., Castell-Perez, E.M. and Zabik, M.E.:
Rapid Testing Method for Characterizing the
Rheological Behavior of Gelatinizing Corn Starch
Slurries, Cereal Chemistry 66:65-68. (1989).
20. Steff, J.F.: Rheological Methods in Food Process
Engineering, Freeman Press, 2807 Still Valley Dr.,
East Lansing, MI. ISBN 0-9632036-0-6, (1992).

6 SPE 95287
Table 1Example Effects of Fracture Geometry on Vol.-Avg. Shear Rate
W
f
, in. hf = 10
ft
20 ft 30 ft 40 ft 50 ft 60 ft
0.1 4,436 2,218 1,479 1,109 887 444
0.2 1,109 555 370 277 222 111
0.3 493 246 164 123 99 49
0.4 277 139 92 69 55 28
0.5 177 89 59 44 35 18
0.6 123 62 41 31 25 12
0.8 69 35 23 17 14 7
1.0 44 22 15 11 9 4
Assumptions:
Wellbore flow rate = 35 bbl/min
Average leakoff rate = 50%
hf = average vertical height of fracture, ft
wf = average width of fracture, in.


Table 2Ratio of Horizontal to Vertical Velocity in Fracture
W
f
, in. hf=10
ft
20 ft 30 ft 50 ft
0.1 230 115 77 46
0.2 115 58 38 23
0.3 77 38 26 15
0.4 58 29 19 12
0.5 46 23 15 9
Assumptions:
Wellbore flow rate = 35 bbl/min
Average leakoff rate = 50%
Particle size, microns = 500 (20/40 sand)
Newtonian viscosity, cP = 500


Table 3Range of Conditions for Computing Viscosity of Fracturing Slurries
16
Particle size 250 to 1,020 microns
Particle density 1.13 to 3.6 gm/cc
Sphericity 0.7 to 0.95
Power-law consistency
coefficients
950 to 17,000 mPa-
sec
n
Shear thinning index 0.2 to 1.0
Volume fractions of particles 0.0 to 0.55


SPE 95287 7


Fig. 1Effect of particle concentration on fluid viscosity.



Fig. 2Diagram of proppant viscometer.
17

8 SPE 95287


Fig. 3aProppant settling in noncrosslinked base-gel fluids.





Fig. 3bRate of viscous settling slopes computed from Fig. 3a.

SPE 95287 9


Fig. 4Response of a viscoelastic crosslinked slurry with breaker.





Fig. 5Transport visualization with metal crosslink fluid with enzyme breaker.

10 SPE 95287


Fig. 6aEffect of polymer concentration in metal-crosslinked fluid.





Fig. 6bRate of viscous settling as affected by gel loading.

SPE 95287 11


Fig. 7aEffect of breaker concentration on metal-crosslinked fluid.





Fig. 7bRate of viscous settling as affected by breaker content.

12 SPE 95287


Fig. 8Effect of acid hydrolysis on crosslinked polymer.





Fig. 9Transport with borate crossliked fluids at 100F.