SPE 139581

A General Correlation for Proppant Settling in VES Fluids

Sahil Malhotra, SPE, Mukul M. Sharma, SPE, The University of Texas at Austin

Copyright 2011, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Hydraulic Fracturing Technology Conference and Exhibition held in The Woodlands, Texas, USA, 24–26 January 2011.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
Polymer-free viscoelastic surfactant-based (VES) fluid systems are used to minimize damage to the proppant pack and to
efficiently transport proppants into fractures. Proper selection of proppants and fracturing fluids for maximum propped
fracture length requires reliable data and correlations for the impact of fracture fluid viscoelasticity and the effect of fracture
walls on proppant settling. Current models and correlations neglect the important influence of fluid elasticity on proppant
transport. This paper presents an experimental study that investigates the impact of fluid elasticity and fracture width and
presents a general correlation for proppant settling in VES fluids.

Proppant settling experiments are performed in shear thinning VES fluids. Experimental data is presented to show that fluid
elasticity plays an important role in controlling the settling rate of proppants. Increasing fluid elasticity can either increase or
decrease the settling velocity depending on the rheological properties of the fluid and the properties of the proppants. A new
experimental correlation is presented to quantify the settling velocity of proppants in VES fluids as a function of the fluid
rheology and proppant size. It is shown that the VES fluids should be designed such that the relaxation time is greater than
the critical relaxation time (Tcrit).

The productivity of fractured wells depends strongly on proper placement of proppants in the fracture. Experimental
data/correlations are presented for the first time to show that the settling velocity of proppants is significantly impacted by the
fracture width and in VES fluids this dependence is different than for the non-elastic fluids. Data is presented to show that
settling velocity is reduced as proppant size becomes comparable to the fracture width. Results show that elasticity reduces
the retardation effect caused by fracture walls. An experimental correlation to quantify the retardation effect due to fracture
walls is presented. Proposed correlations highlight the advantages and limitations of using VES fluids for efficient proppant
transport. These correlations can be directly used in fracture simulators for proppant selection and for the design of fracturing
fluids.

Introduction
Productivity of fractured wells in mainly determined by fracture conductivity and propped fracture length. Both these factors
are dependent on effective proppant transport inside the fracture. It is essential that the fracturing fluids have excellent
proppant carrying abilities to carry the proppant further into the fracture and provide a conductive path for the hydrocarbons
to flow back. The proppant transport inside the fracture is governed by the rheology of the fracturing fluid, properties of the
proppants and the geometry of the fracture. Various types of fracturing fluids have been used for hydraulic fracturing
operations. These range from conventional fracturing fluids which include water-based and polymer-containing fluids,
energized fluids and foams and hydrocarbon-based fluids. Less commonly used fracturing fluids include polymer-free fluids,
methanol-containing fluids, liquid CO2 based fluids and liquefied-petroleum based fluids. A review of unconventional
fracturing fluids and their applications can be found in Gupta (2009).

Surfactant based viscoelastic (VES) fluid systems fall under the category of polymer-free fluids and have been widely used
for hydraulic fracturing operations (Samuel et al. 1997; Mathis et al. 2002; Leitzell 2007). Free of polymers, these fluids
leave very little residue and facilitate rapid flowback. The operational simplicity of these fluids also provides an advantage
over conventional fracturing fluids (Gupta 2009). For VES fluids, fluid elasticity plays an important role in suspending the
proppant particles. It has been pointed out by Asadi et al. (2002) that the zero shear viscosity is an important parameter for
evaluating proppant transport in these fluids. However, there are no reliable models/correlations available which take into
2 SPE 139581

account the important influence of fluid elasticity and fracture width and quantify proppant settling. In this paper, we present
an experimental study that highlights the impact of fluid elasticity and fracture walls on the settling velocity of proppants. We
present new experimental correlations that quantify the settling velocities in VES fluids and highlight the advantages and
limitations of using VES fluids for efficient proppant transport.

Proppant Settling
The settling velocity of single spherical particle in a Newtonian fluid in the creeping flow regime was first derived by Stokes
in 1851 (commonly referred to as the Stokes equation). Subsequent researchers studied the settling at higher Reynolds
numbers and presented expressions to calculate the drag force (Proudman and Pearson 1957; Ockendon and Evans 1972). For
purely viscous non-Newtonian fluids, an effective or apparent viscosity is used in the corrected Stokes equation in order to
calculate the settling velocity of a spherical particle. A wide range of literature is available for the equations for settling
velocities of spherical particles in purely viscous Newtonian and non-Newtonian fluids (Novotny 1977; Clark and Quadir
1981; Roodhart 1985; Acharya 1986; Daneshy 1990; Gadde et al. 2004).

On the other hand, the past work on the determination of settling velocities of proppants in viscoelastic fracturing fluids is not
as complete. Acharya (1988) performed proppant settling experiments in uncrosslinked and crosslinked HPG and
carboxymethyl HPG (CMHPG) gels and concluded that in the creeping flow regime (Re < 2) the viscous parameters dictate
the proppant settling rate and settling is not influenced by fluid elasticity for both uncrosslinked and crosslinked gels. It was
pointed out that at higher Reynolds numbers the settling velocities calculated on the basis of viscous parameters of the fluid
were lower than experimental settling velocities. It was concluded that fluid elasticity exerts considerable influence and
causes the proppants to settle faster at high Reynolds numbers. A correlation was presented to calculate the settling velocity
of proppants in these gels.

Kruijf and Roodhart (1993) showed that the static proppant settling in borate-crosslinked HPG solutions was not controlled
by the viscous modulus, G′′ but by the elastic modulus, G′ of the fluids. In other words the static settling velocity of
proppants was determined by the elastic properties of the gels. Goel et al. (2002) performed proppant settling experiments in
non-crosslinked and crosslinked guar gels. They observed that for non-crosslinked gels there was a critical guar concentration
above which the settling velocity of the proppants decreases considerably. For crosslinked gels it was observed that the solids
transport behavior correlated better with the elastic modulus rather than the viscosity. It was also observed that drag
coefficients in crosslinked gels were dissimilar at three different pHs because of the difference in the crosslinked networks.
To the best of our knowledge no data/results have been presented for proppant settling in surfactant-based viscoelastic (VES)
fluids. As shown in this paper the absence of polymers and the elasticity of these fluids impact the settling velocity of
proppants differently from gel-based fluids.

Effect of Fracture Width on Proppant Settling

The presence of the fracture walls reduces the settling velocities of the proppants. This effect has been quantified in terms of
a wall factor, Fw which is defined as the ratio of settling velocity in the presence of the confining walls to the unbounded
settling velocity in the same fluid. The wall factor (Fw) is defined by:

Settling velocity with confining walls (1)

Fw =
Settling velocity in unbounded fluid

Faxen (1923) used the method of reflections to determine the wall factors as a function of ratio of particle diameter to the
spacing between parallel walls. It was pointed out that for Newtonian fluids in the creeping flow regime, the wall factors
depend only on the ratio of particle diameter to fracture width irrespective of the viscosity of the fluid. It was shown that the
retardation effect of the parallel walls increases with increasing ratio of particle diameter to spacing between walls.

Miyamura et al. (1981) formulated a 19th order polynomial to determine the wall factors for spheres settling between two
parallel plates in Newtonian fluids. Machac and Lecjaks (1995) conducted experiments with purely viscous shear thinning
(power-law) fluids and showed that the retardation effect of the walls decreases with the decreasing flow behavior index, n of
the fluid. In other words increased shear thinning behavior reduces the retardation effect of the walls. They proposed a
correlation to calculate the wall factors in terms of the diameter to wall spacing ratio and the flow behavior index, n given by:

1 (2)
Fw =
⎡⎣1 + k1 ( d p / DE ) + k 2 ( d p / DE ) 2 ⎤⎦

where DE is the effective diameter given by:

SPE 139581 3

dp 1 (3)
=
DE 2(a / d p - 1)

In the above equations dp is the diameter of the spherical particle and a is the distance between the parallel walls. The
constant k1 in Equation (2) was observed to be a function of the flow behavior index given by Equation (4) and k2 was
observed to be a constant equal to 0.49.
k1 = 1.120 - 3.025n + 3.715n2 (4)

Liu and Sharma (2005) conducted a series of experiments with highly viscous Newtonian fluids and linear guar gels and
presented correlations for wall factors as a function of the rheology of the fluids. They showed that the effect of the fracture
walls in reducing the settling velocity of the proppants becomes significant as the ratio of the proppant diameter to the
fracture width increases. It is important to note that all the above mentioned results have been presented for inelastic fluids
and no data is available which shows the effect of elasticity on retardation due to fracture walls.

Experimental Methods
Description of the Fluids
In our experiments proppant settling is studied in a two component VES fluid system (Zhang 2002). The fluid consists of an
anionic surfactant (sodium xylene sulfonate) as one component and a cationic surfactant (N,N,N, trimethyl-1-
octadecamonium chloride) as the second component. The two components are diluted in distilled water and mixed using an
overhead mixer to ensure proper mixing. When the two components are mixed at different concentrations and in different
proportions the surfactant mixture forms worm like micelles that yield a variety of different rheological properties. The
rheological properties of these fluids are discussed in the next section.

Rheological Measurements
Steady shear-viscosity measurements and dynamic oscillatory-shear measurements are made for all the fluid mixtures using
the ARES rheometer by TA Instruments. Figure 1 shows the shear stress (τ) and viscosity (µ) as functions of shear rate ( γ& )
for one of the samples. The fluids exhibit a shear-thinning behavior. The power-law (K, n) model is fitted to the data in the
range of the shear rates encountered by the particles during the settling experiments. The shear rate used is the surface
averaged particle shear rate defined as (2V/dp), where V and dp are the settling velocity and particle diameter respectively.

Figure 1: The steady shear properties of Fluid 5.

The dynamic oscillatory-shear measurements are made in order to quantify the elasticity of the fluids. These measurements
are made over a range of frequencies from 0.1 rad/s to 100 rad/s. Figure 2 shows the elastic modulus, G′ and the viscous
modulus, G′′ as functions of angular frequency, ω. The data is used to calculate the ratio of the viscous modulus to the elastic
modulus (G′′/G′) as a function of the angular frequency as shown in Figure 3. It is observed that the ratio decreases with the
increase in the angular frequency. This data is fitted to a Maxwell model and the relaxation time, T of the fluid is calculated
using the following equation:
G ''(ω ) 1 (5)
=
G '(ω ) ωT
4 SPE 139581

Figure 2: Dynamic oscillatory-shear properties of Fluid 5.

Figure 3: Ratio of storage modulus to elastic modulus fitted with a Maxwell model.

It is observed that the dynamic modulus data fits the above equation very well for all the fluid samples used in the study.
Table 1 shows the rheological properties (K, n, T) of all the fluid mixtures used. The relaxation time is a measure of the
elasticity of the fluid. A fluid with zero relaxation time is referred to as an inelastic fluid and the elasticity of the sample is
larger for a larger relaxation time.

Table 1: Rheological properties of the different fluids used in the experiments.

n
Fluid # K (Pa.s ) n Relaxation time, T (s)

1 0.363 0.484 0.171

2 0.472 0.389 0.389

3 0.336 0.579 0.555

4 0.785 0.771 0.31

5 0.876 0.7395 0.284

6 2.830 0.9805 0.212

7 2.792 0.9755 0.227

8 23.75 0.9645 2.245

SPE 139581 5

Procedure for Measuring Settling Velocities in Unbounded Fluids

Uniform glass spheres are used in the settling experiments. These spheres have smooth surfaces and have diameters ranging
from 1mm to 5mm. The particles are selected such that they are near-perfect spheres and their diameters are measured with a
high resolution microscope. The settling experiments are performed in containers with diameters at least 25 times the
diameter of the particles. This is done to ensure that there is no effect of the confining walls on the settling velocity of the
particles. The containers are filled with the VES fluids and a sphere is immersed in the fluid and allowed to settle.

A meter stick is placed alongside the cell and the settling process is captured using a video camera. The recorded video is
then used to track the position of the particle and measure the terminal settling velocity. We use a software application called
“Tracker 2.0” to get accurate measurements of the settling velocities. The rheological properties of the fluid are measured as
illustrated in the previous section and the density of the fluid is measured using an accurate weighing balance. All
measurements are made at 25°C. At least 3 measurements are made for each reported settling velocity under each unique set
of conditions. In most cases the experiment is repeated 3 to 4 times to ensure reproducibility. Error bars provided in the
experimental results clearly show the reproducibility of the results and the possible variability in the experimental results.

Procedure for Measuring Settling Velocities between Parallel Walls

Two experimental cells made of plexiglass are used for performing the experiments. These cells are constructed such that the
walls are smooth and perfectly parallel to each other. The width between the walls in the two cells is 3.6 mm and 8 mm
respectively. The aspect ratio of the two cells is kept low in order to ensure that there is no effect of the walls orthogonal to
the parallel walls. Figure 4 shows a schematic of the experimental cell. The experimental cells are filled with the viscoelastic
fluid and the particles are immersed in the fluid and allowed to settle between the walls. Particles of diameters varying
between 1 mm to 5 mm are used in this setup. As with the unbounded settling velocity measurements, the settling process is
recorded with a high resolution camera and the video is used to measure the terminal settling velocity. Repeated
measurements are made to ensure reproducibility and to obtain error bars for each measurement.

Figure 4: Schematic of the experimental cell (not to scale).

Results for Settling in Unbounded Fluids

Settling velocities in unbounded viscoelastic fluids are experimentally measured for glass spheres of five different diameters
in all the eight fluid systems mentioned in Table 1. The experimentally measured settling velocity is denoted with the symbol
V∞EL where ‘∞EL’ refers to unconfined viscoelastic fluids. This settling velocity is compared with the settling velocity
(V∞INEL) calculated on the basis of apparent viscosity data based on the power-law parameters. Here‘∞INEL’ refers to
unconfined inelastic fluids. In other words the experimental settling velocity is compared with the settling velocity of the
same spherical particle in an inelastic fluid of the same viscosity.
6 SPE 139581

The values of V∞INEL are calculated using the Equations given below. The following equations have been obtained from
Acharya (1988).

For RePL < 2 (creeping flow regime):

1 (6)
⎡ ( ρ p - ρ f ) g d pn +1 ⎤ n
V ∞ INEL =⎢ ⎥
⎢⎣ 18 K F ( n ) ⎥⎦

where F(n) is the drag correction factor given by

(3 n -3)
⎡ 33 n 5 - 63 n 4 -11n 3 + 97 n 2 + 16 n ⎤ (7)
F (n) = 3 2
⎢ ⎥
⎣ 4 n 2 ( n + 1)( n + 2)(2 n + 1) ⎦

For 2 < RePL < 500

-1/ 2 (8)
⎪⎧ 3ρ f ⎡ 24 F ( n ) f ( n ) ⎤ ⎫⎪
V∞INEL =⎨ ⎢ + 2 f3 ( n ) ⎥ ⎬
⎪⎩ 4( ρ p - ρ f ) gd p ⎣ 4Re PL Re PL ⎦ ⎪⎭

where
f 2 ( n ) = 10.5 n - 3.5 (9)

and f 3 ( n ) = 0.32 n - 0.13 (10)

In the above Equations RePL is the Reynolds number for a sphere falling in a power law fluid given by:

ρ f V ∞ IN E L 2 - n d pn (11)
R e PL =
K

Any deviation of the experimental settling velocity (V∞EL) from the inelastic settling velocity (V∞INEL) is due to the influence
of the elasticity of the fluid. This deviation from the inelastic settling velocity is expressed in terms of velocity ratio, which is
the ratio of the V∞EL to V∞INEL. A ratio greater than 1 suggests an increase in settling velocity due to elasticity and vice versa.
The elastic effects are represented using the Weissenberg number defined as:

2TV∞EL (12)
We ∞EL =
dp

Figure 5 shows the variation of velocity ratio as a function of the diameter for Fluid 2. It is observed that the velocity ratio is
greater than 1 for the three smaller particles and smaller than 1 for the two larger particles. This suggests that the settling
velocity is increased by elasticity for the three smaller particles and it is reduced by elasticity for two larger particles. These
results clearly show that the elasticity of the fluid can increase as well as decrease the settling velocity.

The increase in the settling velocity due to fluid elasticity has been observed earlier (Acharya et al. 1976; Chhabra et al. 1980;
Walters and Tanner 1992; McKinley 2002). At lower Weissenberg numbers there is a virtual elimination of the wake behind
the sphere. As the sphere settles most of the energy is dissipated in the wake and consequently there is less dissipation of
energy in viscoelastic fluids leading to a reduction in drag and an increase in settling velocity. However at higher
Weissenberg numbers the extensional effects in the wake of the sphere become important and the dilatational stresses slow
down the settling (Solomon and Muller 1996; McKinely 2002; Chhabra 2007). We define the relaxation time at which the
effect of the dilatational stresses become important i.e. when the settling velocity is smaller than the settling velocity with no
elasticity as the critical relaxation time (Tcrit).

It is important to highlight that the increase in settling velocities at lower Weissenberg numbers followed by a reduction at
higher Weissenberg numbers has been observed before only for Boger fluids (constant viscosity elastic fluids) and this
behavior has not been reported in shear-thinning viscoelastic fluids.
SPE 139581 7

Figure 5: Velocity ratios for different size particles settling in Fluid 2.

Figure 6 shows the variation of velocity ratio with diameter for Fluid 2, Fluid 3 and Fluid 8. The plot shows that all the
particles settling in Fluid 3 and Fluid 8 experience a decrease in settling velocities (unlike Fluid 2). It can also be seen that the
magnitude of the velocity ratio is different for the same size particles in the three fluids of different rheologies. This clearly
suggests that the increase as well as decrease in the settling velocities is a combined effect of the rheological properties of the
fluid and the size of the spherical particles.

Figure 6: Velocity ratios for different sized particles in different fluids.

Similar experiments were performed with all the other fluid systems shown in Table 1 and the settling velocities were
recorded. These settling velocities are compared with the settling velocities calculated on the basis of the apparent viscosity
data. Based on a dimensional analysis, an attempt is made to correlate the velocity ratio, V∞EL/V∞INEL as a function of the
properties of the fluid and the properties of spherical particles. Using regression analysis, the following correlation is fitted to
the experimental data.
V 1 +We 4.5
n-1.16 (13)
∞EL ∞INEL
=
V∞INEL 1 + We∞INEL5 Re∞INEL 0.12

In the above equation We∞INEL is a dimensionless number relating the relaxation time of the fluid to the diameter of the
particle and settling velocity based on apparent viscosity. It is defined as:
2 TV∞ INEL (14)
We ∞ INEL =
dp
8 SPE 139581

Similarly, Re∞INEL is the Reynolds number calculated using the settling velocity based on apparent viscosity.
ρ f V∞INEL 2- n d p n (15)
Re ∞INEL =
K

The correlation in Equation (13) is applicable in the following range of variables:

0.39 ≤ n ≤ 1.0, 0 ≤ We∞INEL ≤ 22, 6.12 × 10-6 ≤ Re∞INEL ≤ 6.63

Equation (13) is an explicit correlation which can be used to calculate the settling velocity in VES fluids using the properties
of the fluid and the spherical particle. The coefficient of determination of the fit is 0.89. Figure 7 shows a comparison of the
velocity ratio from all the experiments and the correlation (Equation(13)), as a function of the Weissenberg number. It is
observed that the values from the correlation match very well with the experimental results and important trends of increasing
and decreasing velocity ratios are captured by the correlation. It is important to point out here that the dimensionless number
We∞INEL is different from the actual particle Weissenberg number. We use this quantity in order to get an explicit correlation
between the settling velocity and the properties of the fluid and the particles.

Figure 7: Comparison of the experimental velocity ratios with those predicted using the correlation (Equation (13)).

Increase/Decrease of Settling Velocity due to Fluid Elasticity: The Concept of Critical Relaxation Time
It has been highlighted in the previous section that the elasticity of the fluid can increase as well as decrease the settling
velocity of the particles. In this section we explain this increase/decrease using the correlation (Equation(13)) derived from
our experiments. Let us assume a spherical proppant of size 20 mesh and density 2500 kg/m3 settling in a fluid with density
1000 kg/m3, K = 0.4 Pa.sn, n = 0.7 and a relaxation time = 0 s (inelastic). The settling velocity, V∞INEL can be calculated using
Equations (6) through(11). Let us increase the relaxation of the fluid to 0.25 s and calculate the settling velocity, V∞EL using
Equation(13). The calculations show that V∞EL is higher than V∞INEL by a factor of 1.63 which suggests that proppant will
settle 1.63 times faster in the elastic fluid as compared to the inelastic fluid of the same viscosity.

However, if the relaxation time of the fluid is 1.5 s, the settling velocity V∞EL is lower than V∞INEL by a factor of 0.76 which
suggests a 24% reduction in the settling velocity of the proppant. Figure 8 shows the variation of the settling velocity as a
function of the relaxation time of the fluid for the proppant and the fluid properties mentioned above. It can be seen that there
is an increase in the settling velocity with increase in relaxation time which is followed by a reduction in settling velocity at
higher relaxation times. The relaxation time at which the transition from velocity increase to velocity reduction occurs can be
referred to as the critical relaxation time (Tcrit) and it is observed to be a function of the fluid properties as well as the
proppant properties.
SPE 139581 9

Figure 8: Variation of the settling velocity with the relaxation time of the fluid.

Figure 9 shows the variation of the velocity ratio (V∞EL/ V∞INEL) as a function of the relaxation time for fluids with different n
values (K is kept constant at 0.4 Pa.sn). It is observed that for a constant relaxation time, the magnitude of increase/decrease
in the settling velocity in different for fluids with different n values. It is observed that the velocity ratios are highest for the
most shear-thinning fluids. This suggests that shear-thinning effects in viscoelastic fluids result in an increase in the settling
velocity of particles. It is also observed that the relaxation time at which the transition from velocity increase to velocity
decrease occurs i.e. the relaxation time at which the velocity ratio crosses 1, is highest for the most shear-thinning fluid.

Figure 9: Variation of velocity ratio as a function of the relaxation time for fluids with different n values.

Figure 10 shows the variation of the velocity ratio as a function of the relaxation time for fluids with different K values (n is
kept constant at 0.7). It is observed that the velocity ratios are higher for more viscous fluids (higher K values). It is also
observed that the relaxation time at which the transition from velocity increase to velocity decrease occurs is higher in a more
viscous fluid.
10 SPE 139581

Figure 10: Variation of velocity ratio as a function of the relaxation time for fluids with different n values.

Results for Settling in the Presence of Fracture Walls

Settling velocities are experimentally measured for glass spheres settling between parallel walls. Multiple measurements are
taken so as to get reliable averages and error bars. As discussed before, two experimental cells having parallel walls are used.
The spacing between the parallel walls is 3.6 mm and 8 mm respectively. Different diameter particles are used such that data
points are uniformly obtained for the complete range of particle diameter to wall spacing ratio (r) varying from 0 to 1. The
settling velocities are compared with the settling velocities in the same fluid under unbounded conditions and the wall factors
defined by Equation (1) are calculated.

Figure 11 shows the wall factors (Fw) for Fluid 5. Settling for these particles occurs in the creeping flow regime. It can be
seen that the wall factor drops to a value below 0.46 as the ratio of particle diameter to slot with increases to 0.82. This
signifies a 54% reduction in the proppant settling velocity as the fracture width becomes comparable to the proppant
diameter. It is also clear from the plot that, unlike Newtonian fluids, the wall factors are not a function of only the diameter to
wall spacing ratio. Similar experiments are performed for Fluid 1 through Fluid 7 shown in Table 1.

Figure 11: Experimentally measured wall factors as a function of ratio of particle diameter to wall spacing, Fluid 5.

In order to examine the effect of elasticity on the wall effects, the wall factors are plotted as a function of the particle
Weissenberg number, We for a constant value of r. Figure 12 shows the wall factor against We for three different values of r.
It is observed that the wall factors increase with increasing Weissenberg numbers. This suggests that the effect of the
SPE 139581 11

confining walls decreases with increasing elasticity. It is also important to note that for lower values of r the increase of wall
factors with Weissenberg numbers is not very significant. On the other hand, for higher values of r this increase is much
more pronounced. This suggests that as the ratio of particle diameter to wall spacing increases the impact of elasticity on
reducing the wall effects is higher.

An attempt is made to fit the wall factors in terms of the settling velocity under unbounded conditions and the fluid and
particle properties. The following correlation is obtained using non-linear regression analysis:

Fw = (1- r ) p (16)
Here p is given by:
p = 0.44We∞EL-0.19 (17)

Here We∞EL refers to the particle Weissenberg number under unconfined settling conditions as defined in Equation (12).
Equations (16) and (17) are valid in the following range of variables:
0 ≤ r ≤ 1, 0.39 ≤ n ≤ 1, 0.25 ≤ We∞EL ≤ 9.7

This correlation is an explicit function of the properties of the fluid and the particle and the settling velocity in the same fluid
under unbounded conditions. The above correlation fits the wall factors for all the experiments very well and the coefficient
of correlation for the fit is 0.963. Thus the effective settling velocity in the presence of fracture walls can be calculated
explicitly using Equations (16) and (17).

Figure 12: Experimentally measured wall factors against the particle Weissenberg number
for three different fixed values of r.

Effect of Shear Rate inside the Fracture on Proppant Settling in VES Fluids
If we assume a fracture of height h with width w at any location along the length of the fracture and a purely viscous fluid is
injected at rate i, the shear rate inside the fracture at that particular location is given by:

1 (18)
i ⎛ 2n + 1 ⎞ ⎛ 2 y ⎞ n
γ& = ⎜ ⎟⎜ ⎟
hw 2 ⎝ n ⎠ ⎝ w ⎠
12 SPE 139581

Here y is the location between the fracture walls, being 0 halfway between the fracture walls and w/2 at the fracture walls and
n is the flow behavior index. This averaged shear rate over the width of the fracture is given by:

i ⎛ 2n + 1 ⎞ (19)
γ& av = ⎜ ⎟
hw 2 ⎝ n + 1 ⎠

If ww is the fracture width at the wellbore, the width at any location along the length of the fracture can be evaluated using:

1/( 2 n + 2 ) (20)
⎛ x ⎞
w ( x ) = ww ⎜ 1 -
⎜ x ⎟⎟
⎝ f ⎠

where xf if the half-fracture length. For a given injection rate, fracture width at the wellbore and height of the fracture, the
average shear rate inside the fracture as a function of the length can be calculated using Equations (19) and (20). Figure 13
shows the average shear rate as a function of length along the fracture, given an injection rate 10 bbl/min, half fracture length
of 1000 ft, fracture height of 50 ft and fracture widths at the wellbore of 0.1 in and 0.5 in (these widths can be at different
locations along the height of the fracture as shown in the sketch alongside the plot). We do not take into account any leak-off
in this example. It is observed that the shear rate increases especially towards the tip of the fracture.

Figure 13: Variation of shear rate of the fracturing fluid inside the fracture. The sketch on the side shows that the width at the
wellbore varies along the height of the fracture.

For fluid moving along the length of the fracture, the proppant particle experiences this shear rate imposed by fluid motion
along with the shear rate due to particle settling in the vertical direction. To take into account the effect of the shear rate
Novotny (1977) suggested that the effective shear rate on the proppant particle is the vector sum of the shear rate due to
proppant settling (V/dp) and the shear rate imposed by the fluid motion ( γ&av ). The effective shear rate is given by:

2 (21)
⎛V ⎞
γ& = ⎜ ⎟⎟ + γ&av
2
⎜d
⎝ p ⎠

The settling velocity for a proppant settling in an inleastic power-law fluid is thus given by a modified form of Equation(6)
(for Re < 2), given by:
1- n (22)
⎞ ⎡ ⎛ V∞INEL ⎤
2 2
⎛ (ρ p - ρ f ) g d p2 ⎞
⎟⎟ ⎢ ⎜⎜ ⎟⎟ + γ&av ⎥
2
V∞INEL =⎜
⎜
⎝ 18 K F ( n ) ⎠ ⎢⎣ ⎝ d p ⎠ ⎥
⎦
SPE 139581 13

The settling velocities of proppants, in the creeping flow regime, in purely-viscous inelastic power-law fluids along the
length of the fracture can be calculated using Equation (22) by using an iterative procedure.

Let us assume proppant of size 20 mesh with a density of 2500 kg/m3 settling inside a fracture whose width at the wellbore is
0.1in. The fluid has properties of K = 0.5 Pa.sn, n = 0.7, T = 0 s (inelastic) and density of 1000 kg/m3. Figure 14 shows the
settling velocity along the length of the fracture denoted by “Inelastic”. We see that the proppant settling velocity is greater
near the fracture tip due to higher shear rates (and thus lower effective viscosities). If the fluid was elastic with a relaxation
time of 0.05 s, the settling velocities in the elastic fluid can be calculated using Equation (13). The velocity profile in the
elastic fluid is also shown in Figure. It is clearly seen that the settling velocities are higher in the fluid with relaxation time of
0.05 s as compared to the elastic fluid.

Figure 14: Comparison of the settling velocity profile along the length of a fracture
for inelastic and elastic fluids.

However if we increase the relaxation time to 1.0 s we see that the settling velocities of the proppants are lower than the
inelastic fluid throughout the length of the wellbore as shown in Figure 14. This clearly suggests that the relaxation time is a
critical parameter in determining the suspension capabilities for VES fluids.

In the example cases illustrated above, we have used the shear rate and width profiles (Equations (19) and (20)) from the
literature for inelastic power-law fluids. This is an assumption in the above cases because the shear rates and width profiles
induced by fracture propagation using viscoelastic fluids will be different. Our main goal is to illustrate the effect of elasticity
on settling velocities of the proppants.

The retardation effect of the fracture walls on the settling velocity of proppants is important and should be taken into account.
We have already presented a correlation that quantifies the wall factors for VES fluids given by Equation (16). In this section
we illustrate this retardation effect on the settling velocity profile inside the fracture. Figure 15 shows the settling velocities
along the fracture length for the inelastic fluid as well as the fluid with a relaxation time of 1 s. The settling velocities in
presence of the walls for the inelastic fluid are calculated on the basis of the correlations proposed by Machac and Lecjaks
(1995) given by Equations (2) through (4). The settling velocities for the elastic fluid are calculated using Equations (16) and
(17). It can be observed from the figure that the wall retardation effect increases as the distance from the wellbore increases.
This is because the fracture width decreases near the fracture tip and becomes comparable to the proppant diameter. It can
also be observed that the reduction in settling velocities is larger in the inelastic fluid as compared to the elastic fluid. This is
in accordance with our observation from Figure 12 that an increase in elasticity reduces the wall retardation effects.
14 SPE 139581

Figure 15: Influence of the fracture width on the settling velocity profile for an inelastic and elastic fluid.

Design Considerations for VES Fluids

As has been highlighted from the experimental data and examples in the previous sections, the relaxation time of the VES
fluids is an important parameter influencing proppant settling. Clearly for efficient proppant transport the fluids should be
designed so that they can suspend the proppant for a longer time. This can be achieved by not only increasing the viscosity
but also by controlling the relaxation times of these fluids. While designing frac fluids and selecting the proppant sizes, care
should be taken that the velocity ratio expressed by the correlation in Equation (13) is less than 1. Velocity ratios greater than
1 imply an increase in settling velocity caused by fluid elasticity leading to smaller fracture lengths. For a given proppant
diameter the relaxation time must be greater than a certain minimum value, defined in this paper as the critical relaxation
time.

Conclusions
1. Experimental results are presented to show that fluid elasticity can lower or increase the proppant settling velocity of
particles in VES fluids.
2. A new correlation for the settling velocity in unbounded VES fluids in terms of the properties of the fluid and the
proppants is presented. The correlation can be directly used in fracture simulators to predict proppant settling.
3. It is shown that the settling velocity first increases with the relaxation time of the fluid (keeping the other properties of the
fluid constant). This is followed by a decrease at higher relaxation times. The relaxation time at which the velocity ratio
becomes one is defined as the Critical Relaxation Time (Tcrit).
4. For a fixed proppant diameter the VES fluid should be designed such that the relaxation time is greater than the critical
relaxation time. This will ensure that the elasticity enhances proppant suspension inside the fracture.
5. Experimental results are presented to show the wall retardation effects for spherical particles settling in VES fluids
between parallel fracture walls. Data is presented to show that the reduction in the settling velocities due to the presence
of the fracture walls can be very significant and should be accounted for.
6. It is observed that the wall retardation effects are smaller with an increase in fluid elasticity and this effect is more
pronounced at higher ratios of particle diameter to fracture width.
7. A new correlation is presented to calculate the settling velocity between fracture walls as a function of the unbounded
fluid Weissenberg number and the ratio of proppant diameter to fracture width. This correlation can be implemented
directly into fracture simulators to model the effect of fracture walls on settling.

Acknowledgments
The authors are grateful to the companies sponsoring the JIP on Fracturing and Sand Control at the University of Texas at
Austin (Anadarko, BJ Services, BP America, ConocoPhillips, Halliburton, Schlumberger and Shell) and to RPSEA and DOE
for providing the financial support that made this work possible.

Nomenclature
a = Spacing between parallel walls, L, m
DE = Effective diameter, L, m
dp = Particle diameter, L, m
SPE 139581 15

F(n) = Drag correction factor

f2(n) = Dimensionless function dependent on the flow behavior index
f3(n) = Dimensionless function dependent on the flow behavior index
Fw = Wall factor
g = Acceleration due to gravity, L/t2, m/s2
G′ = Elastic modulus, m/Lt2, dyne/cm2
G′′ = Viscous modulus, m/Lt2, dyne/cm2
h = Height of the fracture, L, m
i = Injection rate of fracturing fluid, L3/t, m3/s
K = Flow consistency index, m/Lt2-n, Pa.sn
k1 = Dimensionless coefficient dependent on the flow behavior index
k2 = Dimensionless constant coefficient in the correlation
n = Flow behavior index
p = Dimensionless coefficient dependent on the particle Weissenberg number
r = Ratio of the particle diameter to the fracture spacing
Re = Reynolds number for a sphere settling in a fluid
Re∞INEL = Reynolds number calculated on basis of apparent viscosity for sphere settling in an unbounded fluid
RePL = Reynolds number for a sphere falling in an inelastic power-law fluid
T = Relaxation time of the fluid, t, s
Tcrit = Critical relaxation time, t, s
V = Settling velocity of the proppant inside the fracture, L/t, m/s
V∞EL = Settling velocity in an elastic unbounded fluid, L/t, m/s
V∞INEL = Settling velocity calculated on the basis of apparent viscosity in an unbounded fluid, L/t, m/s
w = Width of the fracture, L, m
We = Particle Weissenberg number based on settling velocity between parallel walls
We∞EL = Weissenberg number of the particle settling in an unbounded elastic fluid
We∞INEL = Dimensionless number relating the relaxation time to particle diameter and unbounded settling velocity, V∞INEL
ww = Fracture width at the wellbore, L, m
x = Distance from the wellbore along the fracture, L, m
xf = Half fracture length, L, m
y = Distance from the centerline of the fracture, L, m
γ&av = Average shear rate at any location along the length of the fracture, 1/t, 1/s
γ& = Shear rate, 1/t, 1/s
ω = Angular frequency, 1/t, rad/s
ρf = Density of the fluid, m/L3, kg/m3
ρp = Density of the particle, m/L3, kg/m3
µ = Viscosity of the fluid, m/Lt, Pa.s
τ = Shear stress, m/Lt2, dyne/cm2

References
Acharya, A., Mashelkar, R.A. and Ulbrecht, J. 1976. Flow of Inelastic and Viscoelastic Fluids past a Sphere: 1. Drag Coefficient in
Creeping and Boundary Layer Flows. Rheol. Acta 15: 454-470.
Acharya, A. R. 1986. Particle Transport in Viscous and Viscoelastic Fracturing Fluids. SPEPE March: 104-110.
Acharya, A. R. 1988. Viscoelasticity of Crosslinked Fracturing Fluids and Proppant Transport. SPEPE November: 483-488.
Asadi, M., Conway M. W. and Barree R. D. 2002. Zero Shear Viscosity Determination of Fracturing Fluid: An Essential Parameter In
Proppant Transport Characterizations. Paper SPE 73755 presented at the SPE International Symposium and Exhibition on Formation
Damage Control, Lafayette, Louisiana, 20-21 February.
Chhabra, R. P., Uhlherr, P. H. T. and Boger, D. V. 1980. The Influence of Fluid Elasticity on the Drag Coefficient for Creeping Flow
Around a Sphere. J. Non-Newtonian Fluid Mech. 6(3-4): 187-199.
Chhabra, R.P. 2007. Bubbles, Drops and Particles in Non-Newtonian Fluids, second edition, Boca Raton, Florida: Taylor & Francis.
Clark, P. E. and Quadir, J. A. 1981. Prop Transport in Hydraulic Fractures: A Critical Review of Particle Settling Velocity Equations.
Paper SPE/DOE 9866 presented at the SPE/DOE Low Permeability Symposium, Denver, Colorado, 27-29 May.
Daneshy, A. 1981. Proppant Transport. In Recent Advances in Hydraulic Fracturing, J. L. Gildley, S. A. Holditch, D. E. Nierodem and R.
W. Veatch JR, Chap. 10, Monograph Volume 12, Henry L. Doherty Series, Richardson, TX: Girst Printing.
Faxen, H. 1923. Der Widerstand gegen die Bewegung einer starren Kugel in einer zähen Flüssigkeit, die zwischen zwei parallelen ebenen
Wänden eingeschlossen ist. Ann. Phys. 68: 89-119.
16 SPE 139581

Gadde P. B., Liu Y., Norman J., Bonnecaze, R. and Sharma, M. M. 2004. Modeling Proppant Settling in Water-Fracs. Paper SPE 89875
presented at the SPE Annual Technical Conference and Exhibition, Houston, Texas, 26-29 September.
Goel, N., Shah, S. N. and Grady, B. P. 2002. Correlating Viscoelastic Measurements of Fracturing Fluid to Particle Suspension and Solids
Transport. J. Pet. Sci. Eng. 35: 59-81.
Gupta, D. V. S. 2009. Unconventional Fracturing Fluids for Tight Gas Reservoirs. Paper SPE 119424 presented at the SPE Hydraulic
Fracturing Tehcnology Conference, Houston, Texas, 19-21 January.
Kruijf, A. S. D., Roodhart, L. P. and Davies, D. R. 1993. Relation Between Chemistry and Flow Mechanics of Borate-Crosslinked
Fracturing Fluids. SPEPF August: 165-170.
Leitzell, J. R. 2007. Viscoelastic Surfactants: A New Horizon in Fracturing Fluids for Pennsylvania. Paper SPE 111182 presented at the
SPE Eastern Regional Meeting, Lexington, Kentucky, 17-19 October.
Liu, Y. and Sharma, M. M. 2005. Effect of Fracture Width and Fluid Rheology on Proppant Settling and Retardation: An Experimental
Study. Paper SPE 96208 presented at the SPE Annual Technical Conference and Exhibition, Dallas, Texas, 9-12 October.
Machac, I. and Lecjaks, Z. 1995. Wall Effect for a Sphere Falling Through a Non-Newtonian Fluid in a Rectangular Duct. Chem. Eng. Sci.
50(1): 143-148.
Mathis, S. P., Pitoni, E., Ripa, G., Ferrara, G., Conte, A. and Ruzic, M. 2002. VES Fluid Allows Minimized Pad Volumes and Viscosity to
Optimize Frac-Pack Geometry: Completion Type Evolution in Barbara Field, Central Adriatic Sea. Paper SPE 78317 presented at the
SPE European Petroleum Conference, Aberdeen, Scotland, 29-31 October.
McKinely, G. H. 2002. Steady and Transient Motion of Spherical Particles in Viscoelastic Liquids. In Transport Processes in Bubbles,
Drops and Particles, ed. D. DeKee and R. P. Chhabra, Chap. 14, second edition, New York: Taylor & Francis.
Miyamura, A., Iwasaki, S. and Ishii, T. 1981. Experimental Wall Correction Factors of Single Solid Spheres in Triangular and Square
Cylinders, and Parallel Plates. Int. J. Multiphase Flow 7: 41-46.
Novotny, E. J. 1977. Proppant Transport. Paper SPE 6813 presented at the Annual Technical Conference and Exhibition of the Society of
Petroleum Engineers of AIME, Denver, Colorado, 9-12 October.
Ockendon, J. R. and Evans, G. A. 1972. The Drag on a Sphere in Low Reynolds Number Flow, J. Aero. Sci. 3: 237.
Proudman, I. and Pearson, J. R. A. 1957. Expansion at Small Reynolds Number for the Flow Past a Sphere and a Circular Cylinder. J.
Fluid Mech. 2: 237.
Roodhart, L. P. 1985. Proppant Settling in Non-Newtonian Fracturing Fluids. Paper SPE 13905 presented at the SPE/DOE Low
Permeability Gas Reservoirs, Denver, Colorado, 19-22 May.
Samuel, M., Card, J. C., Nelson, E. B., Brown, J. E., Vinod, P. S., Temple, H. L., Qu, Q. and Fu, K. 1997. Polymer-Free Fluid for
Hydraulic Fracturing. Paper SPE 38622 presented at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas, 5-8
October.
Solomon, M. J. and Muller, S. J. 1996. Flow Past a Sphere in Polystyrene-based Boger Fluids: The Effect of the Drag Coefficient of Fluid
Extensibility, Solvent Quality and Polymer Molecular Weight. J. Non-Newtonian Fluid Mech. 62: 81-94.
Walters, K. and Tanner, R. I. 1992. The Motion of a Sphere through an Elastic Fluid. In: Transport Processes in Bubbles, Drops and
Particles, ed. R. P. Chhabra. and D. DeKee, Chap. 3, New York: Hemisphere.
Zhang, K. 2002. Fluids for Fracturing Subterranean Formations. U.S. Patent No. 6,468,945.