INTRODUCTION
Currently, the efforts to overcome depleted crude oil production are focusing
on heavy crude oil production by using Enhanced Oil Recovery (EOR) technology
and exploring new oilreservoirs at more severe locations i.e., deep water fields
(MartinezPalou et al., 2011). EOR technique
proposes enhanced recovery methods to effectively lift up higher fractions of
crude oil from the reservoirs. Exploring new oilreservoirs at deep water field
is challenging due to low seabed temperature which can be as low as 4°C
(Ribeiro et al., 1997), high pressure (>10,000
psi shutin pressure), high temperature (>350°F) and complex fluid with
flow assurance difficulties (Olson, 2008). At present,
one of common EOR technique is to reduce heavy oil viscosity by forming Oil
in Water (O/W) emulsion. Both EOR and deep water explorations, amongst others,
are known to have flow assurance problems, including hydrate formation, wax
deposition and crude oil emulsions. To solve these problems, rheological studies
of the crude oil are required in order for the facilities to be designed and
installed appropriately.
For instance, W/O emulsion viscosity measurement done by Ronningsen
(1995) for North Sea Crude Oils showed that viscosity of the emulsion varies
significantly according to the fluid systems. MeriemBenziane
et al. (2012) measured rheological properties of 4 light crude oil
types and the O/W emulsions under a rheometer using a cone and plate geometry.
The results from continuous flow measurements over a shear rate range of 0130
sec^{1} (at 20°C) showed that the viscosities for all the crude
oil types linearly increase with the shear rate showing shear thickening behaviour
and quasiNewtonian rheological behaviour. In addition, from the experimental
results, it was concluded that water contents and the nature of crude oils strongly
affects the rheological and physical stability of the emulsions. Rheological
measurement of a heavy crude oil from Husky EnergyLloydminster, Canada using
a cone and plate geometry was conducted by Hasan et al.
(2010) at a temperature range from 2575°C. The results show that over
the shear rate range, NonNewtonian shear thinning behaviour was observed, contradicting
to the observations documented by (Barnes, 2000). However,
this result is consistent with that of (Ghannam et al.,
2012; Aomari et al., 1998). Another rheological
study on heavy crude oil under the rheometer by ramping the shear rate from
low to high and back to the low shear rate showed that the heavy crude oil exhibit
thixotropic behaviour (Ghannam et al., 2012).
The degree of thixotropy is higher as the temperature is increased. (Aomari
et al., 1998) conducted a strain sweep measurement at a frequency
1 Hz over the strain range of 10^{2}10^{3}% to determine the
strength of dispersant structure of the fluid, storage (G’) and loss (G”)
modulus. Frequency sweep measurement was then performed to determine cross over
point at which the G’ value overlap the G” value. The authors found
that the inverse points or critical fractions above which the emulsion is reversed
from W/O to O/W emulsion were 0.84 and 0.88 for the two light crude oil samples.
Frequency sweep measurements between a frequency of 1 and 40 Hz under various
temperatures (2020°C) were conducted by Pierre et
al. (2004). Pseudo plastic behaviour was observed for the heavy crude
oil used without any yield stress. A yield stress measurement was done by Ghannam
et al. (2012) using a stress ramp method at various temperature (25,
45 and 65°C). The stress was applied to the sample from 0.094 Pa and the
shear stress against shear strain was plotted. The yield stress was the stress
at which the sample starts flowing (shear rate > 0). In other work, stress
sweep of 10010,000 mPa (at frequency 1 Hz) was done using parallel plate geometry
(35 mm) with gap of 2 mm to determine viscoelastic region of crude oil emulsion
(Fingas and Fieldhouse, 2009). All rheological properties
were measured within this region. Geometry diameter and gap utilized were 40
mm and 600 μm, respectively (Pierre et al.,
2004).
Based on existing literature available, various rheological measurements have
been conducted to assess similar rheological parameter i.e., G’, G”,
linear viscoelastic region, yield stress, viscosity, thixotropy, etc. However,
no standard procedure has been proposed to correctly set the parameters in order
for the rheological measurements to be conducted. These procedures are very
important in order for reliable results to be obtained (Ghannam
et al., 2012; Pierre et al., 2004),
(Fingas and Fieldhouse, 2009). Moreover, rheological
measurement procedures may be specific for each fluid system as indicated in
the work done by Ronningsen (1995). Hence, different
rheological measurement procedures may be required for each fluid system.
This study aims to propose the procedures to determine the setting parameters
for rheological measurements of waxy crude oil.
METHODOLOGY
A waxy crude oil from South East Asia region is utilized in this study. Controlled stress rheometer ARG2 was used for all rheological measurements using a 40 mm cross hatched plate geometry to minimize apparent wall slip phenomena. A solvent trap was used to minimize evaporation of the light end components from the sample for all the rheological measurements and maintain the stability of the sample composition. The measurements procedures in this work are as described below.
Nonnewtonian region, WAT and WDT: Two rheological measurements were
first conducted to determine the gelled temperature, Newtonian and NonNewtonian
regions of the sample as well as the Wax Appearance Temperature (WAT) and the
Wax Disappearance Temperature (WDT). The sample was sheared at a constant rate
of 0.1 and 100 sec^{1} from an initial temperature (T_{i})
of 45°C, a temperature well above the anticipated WAT, to a low temperature
(T_{o}) of 15°C, at constant cooling rate of 1°C min^{1}.
Once it reached T_{o}, the sample was heated back to T_{i} at
same rate. The test was known as a thermal cycle test (Marchesini
et al., 2012).
Gap determination and linear viscoelastic (LVE) region: Measurement
gap is an important parameter in any rheological measurements such that continuum
assumption is valid within the gap. The rule of thumb dictates that the measurement
gap is to be 10 times the particle size in order for gapindependent results
to be obtained (Barnes, 2000). To determine the correct
gap for the rheological measurement, the sample was subjected to a thermal cycle
test (Marchesini et al., 2012) under a shear
rate of 100 sec^{1} at 4 different gaps (300, 500, 700 and 900 μm)
within the temperature range of 4515°C. After thermal cycle test, the sample
is subjected to a strain sweep from 0.01200% at a frequency of 10 rad sec^{1}
at the gap defined previously to determine the LVE region.
Inertia dominated region: Inertia dominated region is a region where the inertia of the equipment strongly affects the rheological measurement. It can be observed as an artificial increasing of viscosity or storage modulus. Misinterpretation and inaccurate analysis could result without indentifying the inertiadominated data.
Tas the inertia torque formula shown below for controlled stress (CMT) rheometer should be greater than 10:
It can be seen from the formula that the minimum viscosity would result in a maximum torque ratio (other parameters are kept constant) indicating that the minimum viscosity would give a maximum torque limit above which will be the inertia dominated region. Alternatively, inertia dominated region can also be determined by observing delta and raw phase angle of the sample during measurement following their correlation as below:
The measurement will not be affected at all by system inertia when the inertia
correction factor is equal to 1 or when the sample fully dominates the measurement.
When system inertia start to take part in the measurement, the contribution
of the sample will be less and thus, the inertia correction factor will be <1.
A frequency sweep test was conducted to determine the region. The sample was subjected to a frequency sweep of 0.0120 rad sec at the temperature below (15°C) and above (45°C) the WAT of the crude oil sample; i.e., when the sample exhibits Newtonian and nonNewtonian behavior. In addition, the strain should be set within the LVE region of the sample such that the elastic and loss moduli are independent of the stress and strain applied and are only functions of frequency.
Response time of the sample: Another important parameter to be considered
for the rheological measurements especially for waxy crude oil is the response
time of the sample. It is the time required by the sample to regain stability
under the given conditions (Mazzeo, 2012). If equilibrium
time allowed is shorter than the response time of the sample, the measurement
results would be highly influenced by the shear history of the sample.
To achieve the purpose, a time sweep test was suggested by Mazzeo
(2012) to determine the response time of the sample. The test is conducted
at a frequency 10 rad sec^{1} and a constant strain of 0.5%, i.e.,
within the LVE region and at two measurement temperatures: 45°C (Newtonian
region) and 15°C (Nonnewtonian region) for 15 min.
RESULTS AND DISCUSSIONS
Rheological measurements have been performed using the procedures described in the previous sections and the results are explained below.
NonNewtonian region, WAT and WDT: Two thermal cycle test were first conducted to determine the Newtonian and NonNewtonian regions as well as WAT and WDT. These results are presented in Fig. 1 and 2.
The results showed that at low shear rate, only Nonnewtonian region was observed. However, it was obvious that there were sudden viscosity changes on both cooling and heating curves. These changes were observed at temperature below 20 and above 30°C, respectively. The noisy result at temperature above 30°C was believed due the low shear rate and sample’s viscosity so that the measurement becomes very sensitive to any external disturbance.
At high shear rate, typical Arrhenius temperature dependent of Newtonian fluid were observed at approximately temperature above 20 and 30°C for the cooling and heating curves, respectively while nonnewtonian regions were observed below these temperatures. The two temperatures were known as WAT (the cooling curve) and WDT (the heating curve), respectively. Below WAT, the sample would have higher viscosity value under low shear rate than with high shear rate. It is because high shear rate promotes finer wax crystal formation than low shear rate such that it reduces the viscosity. From these experiments, it was clear that the measurement temperature range covers both Newtonian and nonnewtonian regions of the sample. In addition, higher applied shear rate would significantly reduce the noise (less sensitive to external disturbance) that was shown as nonNewtonianlike behaviour (Fig. 1).
Gap determination and linear viscoelastic (LVE) region: From the thermal
cycle test, as shown in Fig. 3, the viscosity data varies
according to the measurement gaps. WAT and WDT for all the measurements were
found to be around 22 and 34°C (good agreement with previous results), respectively
indicating that the WAT and WDT are independent of the gap setting.

Fig. 1: 
Thermal cycle test result at the shear rate of 0.1 sec^{1} 

Fig. 2: 
Thermal cycle test result at the shear rate of 100 sec^{1} 

Fig. 3: 
Thermal cycle test at 4 different gaps 
In Newtonian region, the viscosity value at the gap of 300 μm was the
lowest. At the gap settings of 700 and 900 μm, the viscosity values were
in agreement. Within the nonnewtonian region, the viscosity values for all
the four gaps were in agreement to each other except for the gap setting of
300 μm when the sample is subjected to heating where the viscosity is significantly
lower.
For both Newtonian and nonNewtonian range (above and below WAT) the viscosity values at the gap settings of 700 and 900 μm were very close to each other indicating that the gap of 700 μm is the minimum gap to ensure gap independent results. From this result, independent gap result can be achieved at gap 700 μm.
However, it should be noted that this gap is valid under the applied shear
rate of 100 sec^{1} and temperature cooling/heating rate of 1°C/min.
Different combinations of shear rate and cooling/heating rate may produce different
results (Marchesini et al., 2012).
In this measurement, viscosity difference on Newtonian region (Arrhenius sector)
during heating and cooling processes were very small (negligible) compare to
the work done by Marchesini et al. (2012) for
all the measurement gap settings. It indicated that there was no significant
irreversible mechanical degradation on the sample and hence, rheological measurements
can be done on the same sample in series.
Strain sweep measurement in Fig. 4 showed that linear G’ values exist at strain 0.12% and thus, this region is the LVE region of the sample. The intersection between G’ and G” appears at strain above 10% followed by dramatic increased on the raw phase angle. It indicated that the sample was completely broken down.
Inertia dominated region: In order to determine inertia dominated region,
a Newtonian fluid was subjected to a frequency sweep measurement as described
in the previous section at gap of 700 m as determine previously.

Fig. 4: 
Strain sweep measurement at non Newtonian region (15°C) 

Fig. 5: 
Frequency sweep measurement at temperature above WAT (Newtonian
region) 
For a Newtonian fluid, a constant viscosity value is anticipated over the applied
frequency range or otherwise the result is highly influenced by the system inertia.
The elastic modulus should also be zero.
Results from the frequency sweep measurements on the sample at both Newtonian (above WAT) and nonNewtonian regions (below WAT) are presented in Fig. 5 and 6, respectively. Within the Newtonian region, it can be seen that delta values start to deviate from raw phase angle at 0.05 rad sec^{1} and become significant at 0.3 rad sec^{1} onwards. As indicate by Eq. 2, the results implied that the system inertia start to dominate the measurement at frequency above 0.3 rad sec^{1} and hence, this region should be avoided. The raw phase angle was found to be above 150° for almost all the measurement range due to the Newtonian behavior.
Within the nonnewtonian region, almost all raw phase angles were below 150°C
indicating that the sample was not broken down. A linear G’ value was observed
at frequency from 0.040.2 rad sec^{1} indicating that the sample was
at its viscoelastic region. Decreasing viscosity trend displays shear thinning
behavior of the sample.

Fig. 6: 
Frequency sweep measurement at temperature below WAT (nonNewtonian
region) 

Fig. 7: 
Time sweep result 
Delta values were found to start deviate from the raw phase angle at frequency
2 rad sec^{1} indicating that system inertia took part in the measurement.
Significant deviation was observed starting at frequency 10 rad sec^{1}
above which is inertia dominated region.
Response time of the sample: A time sweep was applied to the waxy crude
sample to determine the time response at which the sample regains its structure
(Mazzeo, 2012). This set is crucial in order to determine
the correct equilibrium time on the conditioning step of the measurement. The
result is presented in Fig. 7. The G’ and G” is
taken as an indicator of the stability of the crude sample with an equilibrium
value indicating that the sample is stable.
It can be observed that at gelled condition (15°C) even with 900 sec of
time duration, the sample was found to still be unstable. However, it is not
practical to wait until the sample to become completely stable. Hence, it is
reasonable to consider the response time at the point at which slope of the
gradual curve changes is significantly decreased. In this case, the point was
at approximately 360 sec. It should be noted that the sample was subjected to
a temperature ramp from 4515°C under an imposed shear rate of 100 sec^{1}
before being subjected to the time sweep measurement. Hence, it can be concluded
that the sample required an equilibrium time as long as 360 sec after the temperature
ramp step before any other measurements can be taken.
Within the Newtonian region at 45°C, the raw phase angle was all above 150° and highly scattered indicating that the data is inertially dominated and hence not be presented.
CONCLUSION
A series of rheological measurements have been done for a waxy crude oil from South East Asia region. From thermal cycle test, the results showed that the WAT and WDT of the crude oil are 22 and 34°C, respectively and are independent of the gap setting. The appropriate gap for any rheological measurements to be conducted on this sample is 700 μm. The thermal cycle test also provide an evident that the sample has no irreversible mechanical degradation due to the temperature ramping and applied shear rate, hence, the sample can be subjected to a series of rheological measurement. Strain sweep measurement result display LVE region of the sample at strain 0.12%. Frequency sweep measurements showed that rheological measurement on the sample within Newtonian region should be conducted within the frequency range of 0.010.3 rad sec^{1}. At temperatures below WAT (nonNewtonian region), the measurements should be performed within the frequency range of 0.0110 rad sec^{1} for minimum inertia effects. Decreasing viscosity trend during frequency sweep measurement represent shear thinning behavior of the sample. Time sweep measurements confirmed that the crude oil sample needs an equilibrium time as long as 360 sec in order to regain a stable structure. The systematic protocols proposed in this study provide critical parameter values that are important in order for a reliable rheological measurements to be conducted. The protocols may be applied to any other crude oil rheological measurements as preliminary measurement.
ACKNOWLEDGMENTS
The authors would like to thank Universiti Teknologi PETRONAS for the support on the work conducted.