SPE 163275

Laboratory Challenges
of
Sand Production
in
Unconsolidated
Cores
1
2
1
1
2
2

Mohammad
A.J. Ali. , K. Al-Hamad , A. Al-Haddad , S. AlKholosy , , H. Abu Sennah , T. Sanyal , J. Aniel
1
Kuwait Institute for Scientific Research, 2 Kuwait Oil Company

Copyright 2012, Society of Petroleum Engineers
This paper was prepared for presentation at the SPE Kuwait International Petroleum Conference and Exhibition held in Kuwait City, Kuwait, 1012 December 2012.

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 pa per 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 mus t contain conspicuous acknowledgment of SPE copyright.

Abstract
Improved oil recovery for heavy oil reservoirs is becoming a new research study for Kuwaiti reservoirs. There are two
mechanisms for improved oil recovery by thermal methods. The first method is to heat the oil to higher temperatures, and
thereby, decrease its viscosity for improved mobility. The second mechanism is similar to water flooding, in which oil is
displaced to the production wells. While more steam is needed for this method than for the cyclic method, it is typically more
effective at recovering a larger portion of the oil.
Steam injection heats up the oil and reduce its viscosity for better mobility and higher sweep efficiency. During this
process, the velocity of the moving oil increases with lower viscosity oil; and thus, the heated zone around the injection well
will have high velocity. The increase of velocity in an unconsolidated formation is usually accompanied with sand movement
in the reservoir creating a potential problem.
The objective of this study was to understand the effect of flowrate and viscosity on sand production in heavy oil reservoir that
is subjected for thermal recovery process. The results would be useful for designing completion under steam injection where
the viscosity of the oil is expected to change due to thermal operations.
A total of 21 representative core samples were selected from different wells in Kuwait. A reservoir condition core flooding
system was used to flow oil into the core plugs and to examine sand production. Initially, the baseline liquid permeability was
measured with low viscosity oil and low flowrate. Then, the flowrate was increased gradually and monitored to establish the
value for sand movement for each plug sample. At the end of the test, the produced oil containing sand was filtered for sand
content.
The result showed that sand production increased with higher viscosity oil and high flowrate. However, sand
compaction at the injection face of the cores was more significant than sand production. In addition, high confining pressure
contributes to additional sand production. The average critical velocity was estimated ranged from 18 to 257 ft/day for the 0.74
cp oil, 2 to 121 ft/day for the 16 cp oil, and 1 to 26 ft/day for the 684 cp oil.
Background
Sand production can be defined as the production of sand particles dislocated and detached from the reservoir matrix along
with the produced hydrocarbon to the surface. Sand production is a very challenging and a complex problem, which has
troubled the oil industry worldwide. Millions of dollars are spent every year on cleaning out sand from wells, surface facilities,
and wellbore. Sand production from highly unconsolidated formation can occur as soon as the well brought to production.
However, in more consolidated sandstone formation, sand production occurred initially for short periods then was decrease
and eventually stopped.
A study by Musaed et. al. (2005) on unconsolidated cores where he introduced a sand production capability factor. He
postulated that the capability factor is a function of confining stress at any given two points. He observed that confining stress
highly affects sand production at different flowrates. He also noted that sand production from the yielded zones around vertical
wells is higher than that in the horizontal well.
De-Hua et. al. (2007) performed several laboratory studies on sand movement. They claimed that a simple universal model for
predicting critical velocity of sand production in heavy oil reservoirs is unrealistic at this stage. Generally, several factors
affecting sand movement due to fluid velocity such as rock texture pore fluid properties, and interaction between pore fluids
and rock matrix at different temperatures. Their analysis showed that sand could be either oil-wet or water-wet. For water-wet
sand, oil is a detached pore fluid and does not withstand confining pressure. For oil-wet sand, the oil in pore fluid is directly
contacted with the free sand. The amount of sand movement varies with radial distance around the wellbore, i.e. maximum
concentration of sand is near the wellbore radius. Sand movement occurs when the drag forces of the moving fluid exceeds the
body forces holding the detached sand in place.

SPE 163275

Experimental Procedure
Oil Viscosity: An automated, high-pressure and temperature, closed-capillary tube was used to measure oil viscosity. Initially,
the oil sample was filtered with filter paper to remove any sand particles in the original oil sample. Next, the water cut was
measured to be less than 1%, so there was no need to remove the water content. Initially, the viscosity was measured at
ambient temperature of 25 oC and pressure of 30 psi. The temperature was increased while maintaining a constant pressure. A
minimum of 2 hours was required to arrive to a stabilized temperature each time. Table 1 presents the viscosity and density
measurement at different temperatures; only three values are used in this study at 25, 100, and 180 oC to yield viscosity of 684,
16, and 0.74 cp, respectively.
Coreflood Test: A reservoir-condition bench top core flood apparatus was used for this study. The core flood chamber is
made of stainless steel and can handle a confining pressure of 1000 psi at a temperature of 185 oC. The apparatus is connected
to an electronic balance and high flow rate dual piston pump. Pressure transducers are loaded at the core inlet and outlet to
measure differential pressure across the core sample. The coreholder and the oil accumulator were separately wrapped with
heating jackets and temperature control unit. Initially, the baseline liquid permeability was measured with low viscosity oil
and low flowrate.
The initial setup was conducted by placing the core sample inside the core holder and gradually applying a 600 psi
confining pressure before starting the injection. It was observed that while building the confining pressure, sand was produced
at the outlet production line. In some cases, too much sand was produced and the rubber sleeve was ruptured causing
termination of the test. In some samples, the cores were deformed causing a leak from confining fluid into the core and
damaging the test. Ideally, it is preferred to install a backpressure regulator (BPR) in a typical core-flooding test. However, for
this particular study, it was not advised to use it as it would trap sand grains inside the tubes and cause high differential
pressure that would terminate the experiment. Therefore, it was decided not to use backpressure regulator and allow sand
grains to produce from the core sample.
Another approach however was carried out by maintaining a minimum of 200 psi difference between the confining
pressure and injection pressure by manually adjusting the confining pressure during the test. At the start of the test, a one-way
valve was placed at the exit outlet flow line in order to pressurize the pore-pressure. Starting a flow of 1 cc/hr, the pore
pressure was gradually increased to reach to 50 psi, and kept at a constant pore-pressure and at 250 psi confining pressure for
overnight. Then, the outlet valve was slowly open with extreme care to prevent a sudden pressure drop and sand movement.
Once the outlet valve is fully open, then the injection was started at low flowrate and gradually increases it while maintaining a
constant confining pressure. This procedure was also very challenging to conduct, as it was very difficult to maintain a
constant net overburden pressure during varying the flow rates.
Applying and maintaining a constant confining pressure was very challenging. The core plugs were deforming from
their original size, and sand was compacting just by applying confining pressure of 600 psi. Even when confining pressure was
applied very slowly, it required one to two hours for pressure stability. In some cases, the rubber sleeve was damaged, and core
was collapsed due to confining pressure.
Initially, the system and oil were heated to give low viscosity oil injection at constant flow rate until baseline
permeability was established, or sand production was achieved. Later, the temperature was reduced to give higher viscosity oil,
and was injected. Different flowrates were used ranging from 6 to 360 cc/hr, to inject for approximately 30 ml of oil (1 porevolume) for each flowrate. However, these flowrate values were not always obtained. Therefore, the mass balance flowrate
was used instead of the pump setup values. A weighing balance at the core outlet recorded the weight of oil with time, giving a
direct and precise flowrate.
Sand Filtration: Oil samples containing sand coming out from the outlet of the cores were collected in glass beakers. The oil
samples were cooled to room temperature, and the weight of the oil containing sand was measured. Then, a 100 ml of toluene
was added to the oil and filtered with a 0.42 m filter paper. The remaining sand and oil in the beaker were rinsed with toluene
for complete removal of any residue. The weight of the dry filter paper was measured and labelled with the sample name. The
sand with the filter-paper was dried in an oven at 100 oC and cooled to room temperature. The weight of the sand was
calculated by subtracting the weight of the dry sand with filter paper from the weight of dry filter-paper.
Results and Data Analysis
The results from the initial laboratory setup indicated several phenomena in most core plugs. First, sand was produced in all
samples, except for the very low permeability plugs. Most unconsolidated plugs were deformed when applying confining
pressure causing sand production even without fluid displacement (Fig. 1). It appeared from these laboratory experiments that
the major problem was anticipated by sand compaction at the injection face, which causes severe permeability damage, high
upstream pressure, and sand production at each flowrate did not always increase with higher flowrates. In some cases,
wormholes were formed due to high flowrate and compacted sand as seen in Fig. 2. Permeability values measured by the
coreflood test were discarded because they were unrepresentative to the actual field permeability data because of the noncylindrical deformation during the test. The reduction in the cross sectional area (diameter) along the length of the core caused
uneven velocity at single flowrate, (Fig. 3). Therefore, the term critical velocity defined as the velocity when sand is produced

SPE 163275

was very challenging and the term critical flowrate was used instead.
Another phenomenon observed in this study, was that the first flowrate for each plug sample produced high sand
content than the proceeding flowrate values, no matter how low the flowrate values were. The explanation for this was mainly
due to the effect of sand compaction from confining pressure, these initial reading were not considered as critical values.
Estimating Critical Flowrate
Measurement of sand production was very challenging in all core samples. The main challenge was by maintaining a stable
and constant confining pressure without using backpressure regulator for the unconsolidated cores. Sand was produced at the
start of the test just by applying confining pressure. The production of sand was not always increasing with flowrate. As can
be seen in Figs. 4-5, low concentrations of sand were produced at all flowrates, but the production was not always increasing
with flowrate. Ideally, no sand production anticipated at low flowrates. Then, sand is produced at higher flowrates, but if the
flowrate is too high, then sand compaction and bridging occurs. In this study, there was no clear pattern as to sand production
vs. flowrate. Fig. 4 shows a flowrate of 24 cc/hr gives highest sand production using high viscosity oil of 684 cp, but lower
sand production was observed at 41 and 64 cc/hr. The test was terminated after 64 cc/hr because the pore pressure was
exceeding the confining pressure of 600 psi. Similarly, when low viscosity oil was injected of 0.74 cp, there was different
sand production at different flowrates. The maximum sand production was seen at 289 cc/hr, and lower sand production was
observed at 333 and 360 cc/hr. Due to the limitation of this publication, only few samples are presented in this paper to
illustrate the results and discus the challenges, we conclude that in order to improve the quality of the test, a new core must be
used for each flowrate and viscosity. This however would require a large number of core samples and tests, which may not be
feasible.
Therefore, plotting flowrates vs. the cumulative sand production gives a semi linear fit that describes the general
behavior of sand production. Sand production does not always increase monotonically with flowrate. Sand production may
increase or decrease with increasing flowrate. Therefore, the cumulative sand production is used to correlate sand production
with flowrate. The assumption was that first, a correlation must be established between flowrate and the cumulative produced
sand. Once that correlation has been established, it would then be possible to estimate sand production at any given flowrate.
However, if it is required to examine the individual flowrate on sand production, separate plugs must be used for each
flowrate. Figs. 6-8 present examples of correlations between total sand production at different flowrates, viscosities, and rock
type. It appears that oil viscosity did not always correlate with sand production. Although, by theory it is expected to see less
sand production with low viscosity oil, but this was not always true.
Low viscosity oil gave more sand production as shown in Fig. 6. A second observation was that the relationship
between flowrates and the cumulative sand production was a logarithmic fit; similar trends were seen in other samples as
shown in Figs 7-8. Although it appears that the cumulative sand production was increasing with flowrate, but there was no
clear trends between oil viscosity and sand production. For example, Fig. 6 shows that a low viscosity oil of 0.74cp produced
more sand than the higher viscosity oils. Fig. 7 shows systematic sand production; where lowers viscosity oil produce less
sand, until a flowrate of 150cc/hr when sand production sharply increased. Fig. 8 shows that a high viscosity oil of 684cp
produced less sand than at lower viscosity oil. It is believed that the reason for data discrepancy is that there was not enough
sand to be produced after each test. Therefore, the experiments would properly been more representative if a new core was
used for each flowrate and each temperature. However, this procedure will require a huge number of core samples, which may
not always be available. Fig. 9 shows the net total sand production during oil injection with different viscosities. The general
trends seemed at first to show small grains cores produced more sand than large grain size cores, but a sample with an average
grain diameter of 207 m showed less sand production than a core with higher grain size. This may indicate that if the
permeability is low, then high sand production is unlikely to occur. The degree of consolidation also plays a big factor on sand
movement. For example, core samples with similar permeability, grain size distribution, and subjected to similar oil viscosity
and flow rate, but they have different on the degree of un-consolidation, would have different critical flow rate and different
produced sand content.
Predicting the critical flowrate at which highest sand production was a very challenging technique particularly when
using the same core with different flowrates as illustrated in Table 2. Figs. 10-12 show examples of sand production at
different flowrates, viscosities, and average grain diameter. It appears that smaller grain diameter cores produced sand at lower
flowrates than the larger grain diameter cores. Although the results were not consistent for all core plugs, but the general trend
shown in Fig. 13 indicate that the critical flowrate is a function of fluid viscosity and particle diameter. Large size particles
require larger dragging force to move in the pore channel than smaller size particles.
High viscosity oil tends to produce more sand than low viscosity oil. In addition, critical flowrate is more sensitive to
low viscosity oil than at high viscosity oil. At 0.74 cp viscosity oil, the critical flowrate varied from 60 to 850 cc/hr (18 to 258
ft/day) for grain diameter of 277 to 366 m, and at 16 cp. The critical flowrate varied from 6 to 401cc/hr (2 to 121 ft/day) for
grain diameter of 277 to 366 m, and at 684 cp viscosity oil. The critical flowrate varied from 3 to 87cc/hr (1 to 26 ft/day) for
grain diameter of 277 to 366 m. Fig. 13 shows a illustration diagram for the change of velocity along the length of the core
plug inside the coreholder Because the diameters of the core plugs were changing due to sand production and reduction in bulk
volume, it would be incorrect to calculate the velocity from flowrate conversion. If however, it is necessary to report critical
velocity for field application, then uniform core diameter must be assumed. Using Eqs. 1&2, the critical flowrates are
converted from lab unit to field unit in ft /day.

SPE 163275

(1)
(2)

Conclusions
The results obtained from these coreflood experiments give an insight into field application, but it is not conclusive. Dealing
with unconsolidated core plugs in the laboratories is extremely crucial and very challenging. The coreflood experiment is a
tool that is used to better understand the mechanism of sand movement. Therefore, depending on the core flooding experiment
alone in characterizing the reservoir is very dangerous and the validity of the setups remains to be questionable.

Unconsolidated core plugs are very challenging to work with, and extreme care must be taken during coring,
preparation, preservation, and coreflood setup in order to arrive at good quality results.
The unconsolidated plugs were deformed due to confining pressure, which resulted in sand production.
The first flowrate gave high sand production due to confining pressure; and therefore, the first points were omitted.
Sand production from unconsolidated formation could be affected by flowrate, fluid viscosity, and grain diameter.
Large grain diameter has higher critical flowrate than smaller grain diameter.
The degree of unconsolidation plays a big factor on sand production. Similar permeability but different consolidation
gives different behavior of sand production.

References
De-Hua Hon, Qiuliang Yao, Hui-zhu Zha, 2007. Complex Properties of Heavy Oil, Rock Physics Lab, University of Houston,
SEG/San Antonio 2007 Annual Meeting
Musaed N. J. Al-Awad, Abdel-Alim H. El-Sayed, and Saad El-Din. Desouky 1998. Factor affecting sand production, from
unconsolidated sandstone Saudi Oil and Gas Reservoir, J. King Saud Univ. Vol 11,. Eng. Sct (1). PP 151-174.

Table 1. Viscosity and Density Measurements.

Temp.

Viscosity

Density

(oC)
25
60
80
100
140
160
180

(cp)
684
89
36
16
3.25
1.54
0.74

(gm)
0.9622
0.9342
0.9202
0.9062
0.8782
0.8642
0.8502

Table 2. Estimation of Critical Flowrate of Sand Movement.

No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

0.74 cP
(cc/hr)
218
850
154
289
174
60
61
297
2
2
-

16 cP
(cc/hr)
95
248
401
30
364
202
149
48
6
91
178

684 cP
(cc/hr)
56
32
11
87
12
24
3
56
25
-

SPE 163275

Fig. 1: Deformed unconsolidated core due

to applying 600psi confining pressure.

Fig. 3: Deformed core showing squeezed diameter

at the outlet causing uneven velocity.

Fig. 5. Sand production using oil with viscosity

of 0.74 cp.

Fig. 2: Sand compaction at the core inlet

face showing warmholes.

Fig. 4. Sand production using oil with viscosity

of 684 cp.

Fig. 6. Total sand production at different viscosities

with core sample having average grain diameter
306 m.

SPE 163275

Fig. 7. Total sand production at different viscosities

with core sample having average grain diameter
338 m.

Fig. 9. Net total sand production using different

viscosity oil.

Fig. 8. Total sand production at different viscosities

with core sample having average grain diameter
327 m.

Fig. 10. Esimating of critocal flowrate using oil

viscosity of 0.74 cp.

SPE 163275

Fig. 11. Esimating of critocal flowrate using

oil viscosity of 16 cp.

Fig. 12. Esimating of critocal flowrate using

oil viscosity of 684 cp.

Fig. 14. Velocity changing in core sample.

Fig. 13. Correlations between oil viscosity and

critical flowrates.

Acknowledgments: The authors would like to thank Kuwait Oil Company (KOC), heavy oil (HO) team, and research and
technology (R&T) group for their support and providing the oil and core samples. Gratitude is extended to Kuwait Institute
for Scientific Research (KISR) for utilizing their laboratory facilities and technical assistance.