See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/283802002

Sand screen selection

Article · September 2015

CITATIONS READS
0 6,143

6 authors, including:

Jamie S. Andrews Joseph Ayoub

Equinor 70 PUBLICATIONS   885 CITATIONS   
46 PUBLICATIONS   946 CITATIONS   
SEE PROFILE
SEE PROFILE

Mehmet Parlar Somnath Mondal

Schlumberger Limited Shell Oil USA
117 PUBLICATIONS   702 CITATIONS    19 PUBLICATIONS   132 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Negotiations Conflict Managment and Business Development Research View project

Pressure Transient Analysis of Diagnostic Fracture Injection Test View project

All content following this page was uploaded by Somnath Mondal on 02 March 2016.

The user has requested enhancement of the downloaded file.

 Sand Screen Selection

Many formations produce sand that may hinder production or damage completion
and surface equipment. For decades, the industry has chosen sand control screens
to address this threat based on traditional practices. Research suggests a new
methodology that uses numerical simulation for selecting screen size and type may
improve outcomes.

Jamie Stuart Andrews Adopting methods first used in water wells, early In time, the oil and gas industry developed
Statoil 20th century oil and gas operators concerned sand retention methods that incorporated
Stavanger, Norway with potential sand production from unconsoli- screens, resin- or plastic-coated particles and
dated formations completed wells using pipe that gravel packs. Some companies have, in recent
Joseph A. Ayoub had slotted or round openings. The openings, years, sought to distinguish between sand man-
Rajesh A. Chanpura placed across the production interval, were sized agement and sand retention, in which the former
Mehmet Parlar
to prevent sand from entering the wellbore while uses techniques such as orientation of the well-
Sugar Land, Texas, USA
minimally constricting fluid flow. bore and perforations, monitoring and control of
Somnath Mondal
Shell International E&P
Houston, Texas

Mukul M. Sharma
The University of Texas at Austin
Austin, Texas

Oilfield Review 27, no. 2 (September 2015).

Copyright © 2015 Schlumberger.
1. Tronvoll J, Dusseault MB, Sanfilippo F and Santarelli FJ:
“The Tools of Sand Management,” paper SPE 71673,
presented at the SPE Annual Technical Conference and
Exhibition, New Orleans, September 30–October 3, 2001.
2. Coberly CJ: “Selection of Screen Openings for
Unconsolidated Sands,” Drilling and Production Practice
(January 1, 1937): 189–201.
The notation refers to the percent of sand particles by
mass within the formation that are larger than that value.
A d10 designation means 10% of the sand particles in a
formation are larger than that value; thus, 10% of the
sand particles in a formation are larger than a d10 sand
particle size.
3. Coberly, reference 2.

22 Oilfield Review
well pressures, fluid rates and sand influx to limit
sand production.1 Sand retention, or sand con-
trol, refers to the use of screens and other tools to
reduce the risks of sand production without
restricting oil and gas productivity.
Early sand control efforts centered on the
assumption that choosing the optimal sand
screen was based on a relationship between
screen opening and a single point in grain size
distributions. Experiments performed under
ideal prepack test conditions using spheres of a
single diameter led early researcher C.J. Coberly
to conclude that negligible particle production
occurs through rectangular slots of widths that
are twice the particle diameter or through circu-
lar openings that have diameters three times that
particle diameter (Figure 1).2
In formation sand samples, particles have a
size distribution, which forced Coberly to pick a
characteristic diameter, d, within that size distri-
bution based on physical experiments using for-
mation sand samples. Sizing the slot width to
twice d10 (2d10) to allow negligible transient sand
production is known as the Coberly rule. In
response to Coberly’s work, H.D. Wilson wrote
that for sand samples from the US Gulf Coast, for
example, proper retention of sand required sizing
the slots to no larger than d10.3 Industry experts
have concluded that the differences in those con-
clusions are related to what constitutes a negli-
gible amount of produced sand and to the attempt
to characterize the entire particle size distribu-
tion using a single parameter.
Other aspects of selecting a slot or screen size
based on traditional practices involve taking rep-
resentative sand samples and characterizing
those samples. Most representative samples are
obtained through conventional cores retrieved
from known depths.
To characterize formations, laboratory tech-
nicians determine the particle size distributions
(PSDs), typically by sieve or laser analysis or
both. In recent years, the use of laser particle
size analysis (LPSA) has become common in
some companies because such analysis can bet-
ter provide the details of the finer portion of the
particle size distribution than can sieve analysis.
In addition, laser analysis is less labor intensive
than sieve analysis and thus typically lower in
cost, which allows operators to economically ana-
lyze many samples.
Using the most representative sample avail-
able, engineers typically determine proper
screen openings based on the coarsest 10% of a
particle size distribution, or d10. Screens that
have slot widths determined by this process are
designed to allow some amount of sand to pass
September 2015
0.35
0.30
gh
ou
th r
s
se
s
pa
0.25
nd
Sa
0.20
Slot width, in.
g es
rid
n db
0.15 Sa
0.10
0.05
0.00
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
Screen opening, in.
Figure 1. Traditional method for sizing screens. Tests performed in the 1930s,
using sands of a single grain size on screens that had rectangular slots,
resulted in curves that are nearly linear functions. A stable particle bridge
formed across slots whose width was about twice that of the grain diameter
(red), and all sand passed through the screen that had a slot size about three
times the grain diameter (blue). (Adapted from Coberly, reference 2).
while the coarsest particles are retained by size be performed to determine the optimal choice for
exclusion or bridging. In the process of reten- the target formation. A case history from offshore
tion, fine particles are retained by the pore West Africa demonstrates the potential for the
space of the coarse grains and even finer parti- methods discussed.
cles retained between the pore space of the fine
particles; this process repeats until sand pro- How Choices Are Made
duction ceases. Before the drill bit breaks ground, operators
This article describes the process by which must make various decisions that will impact how
engineers match optimal wire wrap and metal the completion is finally configured. Engineers
mesh stand-alone screen (SAS) size and type to must then decide whether to case, cement and
target formations in openhole completions. In perforate the production interval or to use an
addition, this article discusses a technique that openhole completion.
allows engineers to use the entire sand size dis- Openhole completions, typically less costly
tribution when selecting a screen and to quickly than cased hole completions, may be completed
narrow the range of screen sizes and types to using gravel packs or stand-alone screens if the
optimize sand control. This process often results formation is expected to produce sand. Stand-
in sand control decisions more suited to the well alone screen types include wire wrap screens
at hand than is possible using past practices that (WWSs) and metal mesh screens (MMSs). To cre-
use only one design parameter, such as d10, and ate a WWS, manufacturers wrap wire around a
reduces the number of laboratory tests that must perforated base pipe. The wire is either placed
23
 around the pipe during manufacturing or manu-
factured as an individual jacket that is later
welded to a base pipe. Mesh screens include one
or more layers of woven stainless steel or mesh
wire wrapped around a base pipe. The mesh,
which acts as a filter, is covered by a protective
shroud (Figure 2). Although uncommon, opera-
tors have included shrouds on WWSs in side-
tracked wells that have challenging casing exits.
Even when widespread agreement exists that
Wire wrap screen Protective Metal mesh SASs are appropriate, recommendations for
shroud screen
screen type and opening size often vary widely.
Early efforts at screen sizing were based on a
single point (d10) on the PSD and some amount of
sand production that was assumed to be accept-
able, as described earlier.4
In the 1990s, a mathematical model was
developed to optimize sizing of slots in sand con-
Perforated trol devices. This model was based on a fractal
base pipe description of the entire PSD given in terms of
the number of particles rather than particle
mass.5 A series of laboratory tests were performed
to establish a database of wire wrap screen
behavior results using sands from the North Sea
and the Haltenbanken Area offshore Norway.
From these experiments and the number-based
particle size distributions, four slot widths were
defined for each sand type tested: d22, d2, d+
and d++ (Figure 3). The designation d22 was the
largest slot size at which severe plugging
occurred and d++ was the smallest slot size at
Figure 2. Wire wrap and metal mesh screens. Both wire wrap screens (left)
and metal mesh screens (right) are constructed around a perforated base which continuous sand production occurred. The
pipe. Wire wrap screens include a screen that can be slipped over the base d2 and d+ slot widths were defined as the small-
pipe and welded into place. The metal mesh screens, made of woven metal est hole size that did not allow plugging and the
layers that may include sintered metal, are placed between the base pipe and largest slot size that did not allow continuous
the perforated protective shroud.
sand production, respectively.6 The ideal slot size
was stipulated to be between d2 and d+.
Completion engineers often use these crite-
ria to constrain screen size options before per-
Will always Will not forming sand retention tests (SRTs) in the
plug plug
laboratory to determine a final screen size. Two
types of SRTs are available: slurry tests and pre-
pack tests. Slurry tests are designed to replicate
4. Coberly, reference 2.
5. Markestad P, Christie O, Espendal Aa and Rørvik O:
May plug Optimum zone May produce “Selection of Screen Width to Prevent Plugging and
sand Sand Production,” paper SPE 31087, presented at the
SPE Formation Damage Control Symposium, Lafayette,
Louisiana, USA, February 14–15, 1996.
6. Markestad et al, reference 5.
7. Chanpura RA, Hodge RM, Andrews JS, Toffanin EP,
Will not Will always Moen T and Parlar M: “A Review of Screen Selection
produce sand produce sand for Standalone Applications and a New Methodology,”
SPE Drilling & Completion 26, no. 1 (March 2011): 84–95.
d– – d– d+ d++ 8. Chanpura et al, reference 7.
9. Mondal S, Sharma MM, Chanpura RA, Parlar M and
Figure 3. Slot width ranges for sand screen design. Mathematical modeling Ayoub JA: “Numerical Simulations of Sand-Screen
and laboratory results led scientists to define four slot widths for each target Performance in Standalone Applications,” SPE Drilling
sand based on sand grain diameter (d). The lower and upper limits of width & Completion 26, no. 4 (December 2011): 472–483.
sizes are defined by d22 and d++. The optimum size range that will neither
plug nor produce sand is bounded by d2 and d+ (green).

24 Oilfield Review
gradual failure of the rock surrounding the Sand-fluid inlet
borehole (Figure 4). During slurry tests, a low-
concentration slurry is pumped at a constant
rate to form a sandpack around the screen. The
mechanism of sand retention, therefore, is dic-
Pressure
tated only by particle size exclusion. monitoring
To perform prepack tests, which represent points
complete hole collapse, technicians place a Screen coupon
sandpack on the screen and pump clean solids-
free liquid through the pack. Because a sand-
pack is already in place, sand retention during a
prepack test is achieved through both size
exclusion and bridging.
Recent research has shown that current SRT Outlet for sample capture

setup and interpretation methods tend to favor

one screen type or other. The traditional criteria
used to choose between a gravel pack or an SAS
Sand-fluid inlet
are overly conservative and often lead analysts to
opt for a gravel pack. Numerous experiments
indicate that, contrary to accepted wisdom, Piston
screen plugging is rarely a problem in clean sand
formations; when plugging is a threat as a result Retaining mesh
of other factors such as contaminated fluids, the Pressure
risk can be mitigated through proper hole prepa- monitoring
points Sandpack
ration procedures.7
To address the variability and inconsistency
inherent in screen selection and to better under- Weave sample

stand the physics of sand control, scientists

recently used a numerical simulation approach
to evaluate sand screen performance. The effort
was part of a larger plan to produce a systematic
screen selection process.
Screen sizing practices that relied on
accepted standards were based on PSDs that
did not use the results of sand retention tests.
Despite the limitations of these standards, which
are based on a few parameters of the formation
sand size distribution and implicit assumptions
about acceptable levels of sand production, most
experts continue to use such standards not only
to narrow screen size options but also to perform
SRTs to confirm final screen selection.
In general, three results from SRTs are of
interest: sand production correlated to the
screen’s sand retention efficiency, pressure
development correlated to screen plugging ten-
dency and size distribution of produced particles
with which to evaluate the risk of screen erosion.
However, because it has now been established
that screen plugging is rarely a problem in clean
formation sand of any PSD, the main criteria for
screen selection become transient sand produc-
tion and PSD of produced particles. Engineers
can determine both criteria using models devel-
oped in the last five years for specific screen and
PSD combinations without having to conduct
actual SRTs.8
Outlet for sample capture
Figure 4. Two types of sand retention tests. Slurry tests (top) are designed
to simulate gradual failure of the formation surrounding the borehole.
Technicians pump a low-concentration sand slurry through a screen coupon,
then measure the weight of solids produced through the screen and the
pressure buildup across the screen versus the amount of sand contacting the
screen. Laboratories design prepack tests (bottom) to simulate a complete
borehole collapse by placing a sand sample directly on the screen. A liquid is
then flowed through the sand and screen. Technicians then create a confining
stress on the sample that forces the sand into full contact with the screen. The
test measures the amount of sand that passes through the screen—measured
by weight—and the pressure drop across the screen.
Model Alternative understand the science of sand retention. The
A team from academia and industry reviewed study used numerical SRT simulations that
recent screen testing advancements, interpreta- matched experimental data in an effort to aid the
tion and modeling for SAS applications. Based on team in understanding and relating PSD-screen
its findings, the team has proposed a screen combinations and to correlating sand production
selection method based on laboratory test–veri- with formation sand PSD until sand production
fied numerical and analytical models. stops or is limited to fines.9
The primary purpose of this method is to elim- The team first studied WWSs, which have a
inate or reduce the number of physical SRTs that simpler geometry than that of other screen types,
must be performed when selecting a screen size and performed simulations using the discrete
and type for a given application and to better element method (DEM). This numerical model
describes mechanical behaviors, such as mass,

September 2015 25
Simulation Box Simulation End to form stable particle bridges, whereas the most
critical parameter affecting the number of sand
particles produced is the ratio of the slot width to
particle diameter. Similarly, high fluid viscosities
and low pressure gradients facilitate particle
bridging; increased fluid pressure increases par-
ticle production when pressure gradients are up
to about 2.3 MPa/m [100 psi/ft]. At higher gradi-
ents, however, there is no such dependence.
When the results from the DEM model were
plotted, the team observed a power-law relation.
This relationship was confirmed by plotting the
experimental data, which revealed excellent
agreement and consistent trends between model
and experimental results. Based on this newly
established relation, the team developed the
Figure 5. Simulations of sand retention tests using the discrete element method (DEM). Using the Mondal-Sharma (M-S) method, which uses the
DEM, scientists track information such as mass, velocity, force and momentum about each particle
number and size of the produced solids to esti-
within the computational domain, or simulation box (left). Researchers used the DEM and a molecular
dynamics simulator to model performance in a prepack experiment by generating a packing of mate the mass of sand produced (Figure 6).
polydisperse granular spheres (multicolored balls) over a wire wrap screen geometry (blue layer) and When comparing the estimated mass of sand pro-
then flowing liquid through the pack. The individual size and number of particles per size were obtained duced using the M-S method with the mass of
from the measured particle size distribution of the formation sand sample used for the corresponding
sand produced in experiments, a good match was
experiment. Discrete element method simulations were then used to calculate the mass of sand
produced per unit area of screen for various screen sizes and particle size distributions. Near the found. The M-S method, which uses DEM simula-
end of the polydisperse simulation, which required 24 hours on a 48-processor network cluster, sand tion results to develop a simple correlation, can
particles (right, green, purple, brown, blue and white) bridge across the screen openings (pink). be used to estimate the mass of sand produced
without performing DEM simulations for every
possible sand and screen combination.11
velocity, force and angular momentum, of assem- mass of sand produced per unit area of screen for The research team next extended the applica-
blies of spheres (Figure 5).10 The study simulated various screen sizes and PSDs. tion of the M-S method to include plain square
prepack experiments by first generating a pack- To accurately represent the physics of the mesh (PSM) screens, achieving much the same
ing of polydisperse granular spheres over a WWS problem, the model was tested and validated outcomes. Some conclusions from WWS and PSM
geometry and then flowing a fluid through the using a range of various parameters. The team simulations included the following:
pack. The research team could then compute the found that friction and shear forces are necessary • Simulations are able to estimate the mass of
sand produced for a given PSD and screen size.
• Simulations results strongly agree with those
100 from carefully controlled prepack experiments.
90 • Simulations show that the mass of sand pro-
duced per unit screen area and for unit open
80
flow area is larger for single layer PSMs than
Cumulative % by weight

70
for slot geometry of the same rating and corre-
60
sponding standard open flow area.
50 • Simulations show that the ratio of wire thick-
40 ness to opening size seems to be a key factor
30 contributing to the increased mass of sand pro-
20 duction from single layer PSMs.12
10 Researchers then turned their attention to
0
analytical solutions and Monte Carlo simulations
1,000 100 10 to predict sand production through WWSs and
Particle size, microns PSM screens under slurry test conditions. Their
Figure 7. Particle size distribution (PSD) of retained and produced formation results showed that the analytical solution and
sand. The PSD of the first layer of sand retained on the screen (red) has the the numerical simulation were in excellent
expected distribution of the sand particles of widths greater than the slot agreement. The team showed that its proposed
size. The PSD of the second layer (green) is approaching that of the formation
methods were able to estimate both mass and
sand (blue). Because the sizes of the particles retained on the second layer
are dictated by the pore sizes of the first layer, the retained particles will size distribution of the produced solid in a slurry-
eventually be of the same PSD and permeability as the formation sand. type SRT, taking into account the full PSD of for-
(Adapted from Chanpura et al, reference 13.) mation sand. Simulations also showed that, with
the exception of a mobile fines problem, sand

26 Oilfield Review
 Number of particles produced/m2 for 10% open flow area
1 × 1012

1 × 1011

1 × 1010
y = A5x – b 5
R 2 = 0.8998
1 × 109

1 × 108

1 × 107

1 × 106
0.01 0.1 1
Dp /W

100

80
PSD A PSD B
Cumulative by weight,%

60

40

20

0
D5A

D4A
D3A
D2A

D1A

1,000 100 10 1
Diameter, microns

Figure 6. Determining mass of sand produced and entire formation particle size distribution. The
Mondal-Sharma (M-S) method uses a correlation between the number of particles of diameter Dp
produced through a screen slot opening of width, W. The number of particles of each diameter
produced through the screen are counted and plotted against Dp /W from every simulation (top). In
this case, formation PSDs A and B were distributed into five bin sizes each (bottom, dashed lines)
to generate the number-based size distributions (D1A to D5A and D1B to D5B) used to populate the
simulation box (bottom). (Adapted from Mondal et al, reference 9.)

production becomes negligible once the slot tion based on an acceptable level of sand produc-
opening has been covered by particles larger tion. Final screen selection may then be
than the opening (Figure 7). confirmed through a slurry-type SRT.13 Results
As in the case of modeling prepack-type SRTs, showed that more than 90% of the total sand pro-
the proposed methods can be used to estimate duction by mass occurs during the formation of
sand production in slurry-type SRTs for various the first layer of particles on the screen and that
screen sizes, thereby enabling screen size selec- the PSD of the retained sand approaches that of
10. Cundall PA and Strack ODL: “A Discrete Numerical 12. Mondal S, Sharma MM, Hodge RM, Chanpura RA,
Model for Granular Assemblies,” Géotechnique 29, no. 1 Parlar M and Ayoub JA: “A New Method for the Design
(March 1, 1979): 47–65. and Selection of Premium/Woven Sand Screens,”
11. Mondal et al, reference 9. SPE Drilling & Completion 27, no. 3 (September 2012):
406–415.
the formation sand after a few layers of sand
accumulate on the screen.
Results also revealed that the mass of sand
produced during the formation of the first layer
of particles on the screen is independent of the
shape of the PSD for grains smaller than the
aperture-pore size and is governed by the shape
13. Chanpura RA, Fidan S, Mondal S, Andrews JS, Martin F,
Hodge RM, Ayoub JA, Parlar M and Sharma MM:
“Advancements in Screen Testing, Interpretation and
Modeling for Standalone Screen Applications,” paper
SPE 143731, presented at the SPE European Formation
Damage Conference, Noordwijk, The Netherlands,
June 7–10, 2011.

September 2015 27
 0.8
Mass of sand produced/screen area, kg/m2
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
175-1 175-2 175-3
Formation sample
2.5
Mass of sand produced/OFA, kg/m2
2.0
1.5
1.0
0.5
0.0
175-1 175-2 175-3
Formation sample
Figure 8. Comparison of sand production through plain square mesh (PSM)
and wire wrap screens (WWSs). The mass of sand produced for seven PSD
values through a 175-micron single-layer PSM is greater than that produced
through a 175-micron WWS per unit screen area (top) and per unit open flow
area (OFA) (bottom). (Adapted from Chanpura et al, reference 15.)
of the PSD of grains greater than the aperture-
pore size. In addition, researchers found that
sand production through the filter layer of a PSM
screen of a given pore size is greater than that of
a WWS of the same slot size (Figure 8).14
The Mythology of Screen Selection
The team’s work has cast doubt on, or added
qualifications to, numerous widely held industry
beliefs about WWSs and PSMs. These axioms,
upon which many traditional screen selection
methodologies for SASs have been based, include
the contention that formation sand plugs screens.
However, research has shown that following
SRTs, when only trapped particles remained on
the screens, final screen permeability was in the
range of 5% to 100% of original screen permea-
bility; the final value, then, of even the low-per-
meability SAS screens, which have an original
screen permeability of about 300 D, would be a
minimum 15 D. The screen permeability is thus
28
Single-layer PSM WWS
175-4 175-5 175-6 175-7
Single-layer PSM WWS
175-4 175-5 175-6 175-7
significantly higher than most formations and
thus too great to cause plugging; plugging is
commonly quantified by a pressure differential
created across the screen. Instead, plugging
more likely occurs as the result of poorly condi-
tioned mud or filtercake mixed with formation
sand, mixed coarse and fine formation sands
from a variety of zones or clay and shale mixed
with formation sand.15
PSD and PoSD
When SRTs are performed in the laboratory using
formation sand, the sand PSD is often not needed.
However, PSD is required if there is a large
spread in formation PSD along the well, or if the
SRT is performed using a sample that was gener-
ated based on specified PSD or if a model is used
to estimate sand production for a given sand
PSD–screen combination. Particle size distribu-
tion of formation sand is typically determined
through dry sieve analysis or laser particle size
analysis (LPSA).16
Dry sieve analyses determine PSD through a
mechanical separation of particles by filtering
them from top to bottom through a series of pro-
gressively finer sieves. The measured weight of
the sand captured in each sieve is used to calcu-
late cumulative percentage mass of each, which
is then plotted against sieve size on a semiloga-
rithmic scale.
Laser particle size analyses determine PSD by
measuring how light is scattered as a laser beam
is passed through a sand sample. The angle of
scatter is inversely proportional to the particle
size.17 To ensure that the sand samples are deliv-
ered to the measurement device in the correct
concentration and in a stable state, LPSA is per-
formed on samples whose dispersion is controlled
by dry or, when necessary, fluid dispersants.
Sand control experts have long used dry
sieve and LPSA nearly indiscriminately, and
persistent differences in the results obtained
from the two methods have been well docu-
mented. Recent research indicates these incon-
sistencies may be caused by the aspherical
shape of the particles, sampling practices for
LPSA, fluids used and various light blocking lev-
els used in the LPSA. Based on these observa-
tions, PSD determined by dry sieve analysis is
recommended for both slurry-type SRT testing
and sand production prediction using the above
14. Chanpura RA, Mondal S, Andrews JS, Mathisen A-M,
Ayoub JA, Parlar M and Sharma MM: “New Analytical
and Statistical Approach for Estimating and Analyzing
Sand Production Through Plain Square-Mesh Screens
During a Sand-Retention Test,” SPE Drilling &
Completion 28, no. 2 (June 2013): 135–147.
15. This research revealed that numerous assumptions
regarding sand production and screen characteristics
were unfounded. For more on the team’s discussion of
traditional assumptions: Chanpura RA, Mondal S,
Sharma MM, Andrews JS, Mathisen A-M, Martin F,
Marpaung F, Ayoub JA and Parlar M: “Unraveling the
Myths Associated with Selecting Standalone Screens
and a New Methodology for Sand-Control Applications,”
SPE Drilling & Completion 28, no. 3 (September 2013):
227–236.
16. Zhang K, Chanpura RA, Mondal S, Wu C-H, Sharma MM,
Ayoub JA and Parlar M: “Particle Size Distribution
Measurement Techniques and Their Relevance or
Irrelevance to Sand Control Design,” paper SPE 168152,
presented at the SPE International Symposium and
Exhibition on Formation Damage Control, Lafayette,
Louisiana, February 26–28, 2014.
17. Ballard T and Beare S: “Particle Size Analysis for Sand
Control Applications,” paper SPE 165119, presented at
the SPE European Formation Damage Conference and
Exhibition, Noordwijk, The Netherlands, June 5–7, 2013.
18. Zhang et al, reference 16.
19. Mondal S, Wu C-H, Sharma MM, Chanpura RA, Parlar M
and Ayoub JA: “Characterizing, Designing, and
Selecting Metal Mesh Screens for Standalone Screen
Applications,” paper SPE 170935, presented at the
SPE Annual Technical Conference and Exhibition,
Amsterdam, October 27–29, 2014.
20. Agunloye E and Utunedi E: “Optimizing Sand Control
Design Using Sand Screen Retention Testing,” paper
SPE 172488, presented at the SPE Nigeria Annual
International Conference and Exhibition, Lagos, Nigeria,
August 5–7, 2014.
Oilfield Review
Figure 9. High-resolution microCT scans of PSM. A 3D PSM screen image (left) can be reconstructed from a microCT scan using a commercially available
computer-aided design format that is able to preserve and reproduce minute detail (center and right). (Adapted from Mondal et al, reference 19.)
models. However, errors or differences attrib-
uted to particle shape differences may still
occur.18 These differences may be minimized by
characterizing the particle shape and aspect.
Recent investigations of mesh screens have
highlighted the need to account for screen com-
plexity from layered design when modeling sand
production. Using microcomputed tomography
(microCT) images, researchers constructed 3D
images of two metal mesh screen types: PSM and
plain Dutch weave (PDW) (Figure 9). These 3D
images of virtual screens were validated by com-
parison with the microCT images.
The team then conducted DEM simulations
that were validated by experiments of prepack
SRTs through multilayer PSMs and PDWs.
Analyses of microCT scan meshes indicated that
mesh screen layers overlap significantly and thus
impact retention efficiency. The group developed
a method to calculate the retention pore size dis-
tribution (PoSD) and effective pore size for a
given overlap of PSM samples. The calculated
PoSD can be used in the analytical model to
improve sand production prediction in a slurry-
type SRT.
As a consequence of this work, the perfor-
mance of nominal size MMSs can be simulated
using any reservoir sand size distribution. To
date, because the team has been able to charac-
terize PSMs, operators are able to evaluate a
large number of PSMs in a short time and thus
reduce the number of SRTs that must be run to
choose the optimal screen size for a given reser-
voir.19 In time, this work will be expanded to
include additional screen types.
September 2015
View publication stats
By the Numbers
Engineers use SRTs to choose the optimal screen
from a range of screens selected based on a rela-
tionship between screen openings and grain
sizes. Although SRT results can be impacted sig-
nificantly by relatively small changes to test con-
ditions, when performed properly, the SRT is
widely considered a reliable method for finalizing
screen choice. The drawback to this process,
however, lies in the dubious traditional practices
used to narrow the range of screen choices and in
misinterpretation of pressure developments in
standard SRT experiments. This process often
forces operators to choose to perform many time-
consuming SRTs before qualifying a screen as
optimal for long horizontal sections that have
varying sand PSD.
By replacing traditional methods with numer-
ical and analytical models, operators may reduce
and eventually eliminate the dependence on
SRTs. In addition, because traditional screen
selection methodology tends to be conservative, a
software-based approach may allow operators to
opt for SASs over gravel packs, which are typi-
cally more expensive.
When working offshore West Africa required
sand control for a nonuniform unconsolidated
formation, a major operator based its screen
selection process on traditional d10 preselection
criteria and on SRTs for finalizing its selection.
The completions team also compared the results
of the laboratory tests to numerical models.
The targeted reservoir is the second sand in
the offshore field; wells in the first sand of the
field were completed using sand control devices
selected based solely on traditional methods.
However, the first formation produced is made up
of highly uniform, well-sorted reservoir sands
that have very low levels of fines content. By con-
trast, the targeted sand in the second reservoir is
much less uniform, poorly sorted and has higher
fines content. In the face of these adverse sand
control indicators, the operator opted to perform
as rigorous a selection process as possible and to
check selections based on traditional and SRT
methods against those using simulations and
mathematical models.
In comparing results, the operator concluded
that selections based on the results of SRTs and
those based on the mathematical models
matched closely. The operator added that
although models require laboratory data for
proper calibration, they held significant potential
for aiding screen size selection without the need
for continued laboratory testing when applied in
regions for which extensive SRT data existed.20
The quantity of and interaction between the
variables that engineers must consider in choos-
ing a sand control strategy can be daunting. For
decades, engineers have relied on the experience
of their predecessors to help them sort the data
and arrive at decisions. Today, however, because
of the growth of computing power and capacity,
operators may avail themselves of more accurate
and less compromising methods for sand control
selection. Based on physics and mathematics,
these new methods promise not only a quicker,
less costly path through the selection process,
but one that provides engineers with the cer-
tainty that they have chosen an optimal sand con-
trol strategy for any given formation. —RvF
29