You are on page 1of 10

SPE 116460

Field Testing of an Automated Well Tester in Extra-Heavy and Diluted Oil


Application
Rafael Bastardo, Asher Imam, Weatherford; José Scoglio, Petrozuata

Copyright 2008, Society of Petroleum Engineers

This paper was prepared for presentation at the 2008 SPE Annual Technical Conference and Exhibition held in Denver, Colorado, USA, 21–24 September 2008.

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

In the past, gas-liquid cylindrical cyclone (GLCC) technology was chosen mainly for low viscosity applications, where liquid
and gas phases could be completely separated from the mixture and measured independently. Due to its reduced footprint,
investment cost, and simplicity, this turned out to be an attractive solution for oil industry applications. However, the practice
of not using GLCC technology when handling heavy oils became a model. This model shift happened because of GLCC’s
short residence time, poor separation efficiency, high gas carry-under (GCU), and limitations of single-phase flow-metering
instruments.
The main improvement made to this technology, was finding an efficient way to separate a majority of gas from heavy-oil
mixture and finding a combination of metering technologies able to handle certain amounts of GCU without affecting liquid-
rate accuracy.
The evaluation program was conceived to test the behavior of a partial separation system (PSS) when handling extra-
heavy and diluted oil. A total of 42 tests were performed on 16 wells from 3 well pads.
The diluent injection rate was determined to be one of the main parameters impacting the PSS performance and liquid
metering accuracy. It was observed that the variation in the diluent percentage impacted the liquid metering by over- or
under-estimating the total liquid measurement. High viscosities and foamy mixture effects were countered by making minor
adjustments to the GLCC level control and instrument configurations.
Based on the success criteria, the PSS evaluation was qualified as successful. The system proved to be ±10% accurate on
total liquid-volume measurement with a confidence interval of 90%. Regarding repeatability, a maximum deviation of
1 to 2% was shown under normal well operating conditions.
For water-cut measurements, the objective was to measure within ±5% of the actual value without performing individual
well calibrations. Testing of fluctuations in extra-heavy oil properties and concentration of diluent, as well as field conditions,
proved that a ±5% accuracy for the water-cut measurements could be achieved if individual well calibrations were
performed.
Throughout the entire duration of the project, the PSS, with its flexibility and ease of operation, maintenance and
diagnostics, proved to be a solution for extra-heavy and diluted oil applications in the Orinoco Belt of Venezuela.

Introduction
The Petrozuata multiphase flow metering (MPFM) evaluation program took place in San Diego de Cabrutica, Venezuela.
The MPFM program was put in place to evaluate the performance of partial separation systems (PSS) in an extra-heavy and
diluted oil application.
The PSS used in the Petrozuata MPFM evaluation program is a GLCC based technology. It is a simple, compact, low-
weight and low-cost separator, and is designed to separate and measure liquid and gas phases of a process stream,
individually. This is achieved by introducing the multiphase flow into the vessel at a calculated angle to induce the optimum
centrifugal rotation necessary to separate gas from liquid. Both gas and liquid are then individually measured and recombined
back into the system (Ref. 1, 2)..
The PSS utilizes level, flow and pressure control to optimize the separation and measurement procedure; thereby keeping
GCU and liquid carry over (LCO) to a minimum. The infrared (Red Eye® 2G) water-cut meter provided the system with an
2 SPE 116460

accurate water-cut measurement over a full range (0 to 100%) of oil-water concentration in the liquid stream. All of this
combined, PSS provided a complete method of process management to the customer (see Fig 1).

Fig. 1. Partial Separation System Layout.

The PSS used in this evaluation had a 10-in. diameter vertical separator with an 8-in. tangential inlet. Gas measurement
was carried out by dual gas legs (1 in. and 2 in.) with Coriolis and Vortex meters. The 2-in. liquid leg used a Coriolis meter
and infrared water-cut meter to measure the liquid mass rate and water cut, respectively. Liquid level in the GLCC was
measured using a differential pressure transmitter. Pneumatically actuated control valves on the gas and liquid legs were used
to maintain the level of the GLCC.
The evaluation took place between March 16th and May 28th, 2007. During that time, 42 tests were completed on 16
wells from 3 well pads.
Field Test Setup
The following section details the steps required to implement the MPFM evaluation program. It covers a detailed explanation
of the field operation, the criteria against which the PSS was evaluated and details of the in-field test setup.
Field Operation. Fig 2 illustrates the layout of a typical well pad in Petrozuata. Each well pad contained a series of wells
that were tied into a multiport selector valve; which was used to select the individual well to be tested. Those wells not under
tests were diverted to general production.

TANK 1
DILUENT

WELL 1 TANK 2

WELL 2
TANK 3

WELL 3
DILUENT LINE

WELL 4 PRODUCTION LINE


MULTIPORT VALVE

WELL 5
TEST LINE

WELL 6

CHOKE
VALVE

PSS

Fig. 2. Test Layout

A separate skid was used to control the injection of diluent at the head of each well. Diluent was added to the process to
help reduce wellhead pressure and to raise the API gravity of the oil to help increase flow rates. Diluent rates varied from
SPE 116460 3

well to well depending on the behavior of the downhole/surface equipment and wellhead pressure. During normal operating
conditions of the wells, the percentage of dilution was kept between 15 and 25%.
In-field Setup. The actual field conditions found were as follows:
• Liquid Rate: 100–2,500 BPD
• Water-cut: 0–80%
• Gas Rate: 1,000– 600,000 SCFD
• Temperature: 90–150°F
• Pressure: 90–250 PSI
• API: 12-17 (blend); 8.3-9.5 (extra heavy)

In order for the field tests to be successful, the MPFM evaluation program was divided in two phases:
1. The preliminary testing phase was designed to prove that the PSS was capable of handling the actual conditions found,
even though the system was not originally designed for this field. Wells were tested to demonstrate that the infrared
water-cut meter, liquid-flow meter and level-control strategy could handle a wide range of flow conditions, API, GVF,
water-cut and flow rates.

2. Once preliminary testing was complete, a total of 42 tests were performed during the formal testing phase to validate the
accuracy and repeatability of the PSS.
• Well pad A, 4 wells, 8 tests total.
• Well pad B, 4 wells, 10 tests total.
• Well pad C, 8 wells, 24 tests total.

The total liquid volume measured by the flow computer was validated against a series of Frac tanks. A choke valve
installed on the outlet of the PSS was used to ensure that the pressure on the system when flowing to the tanks was similar to
pressure when flowing to general production. PVT data calculations show that the actual shrinkage factor was between 0.98
and 0.99.
Water-cut measurements were averaged by the flow computer and the results were compared against one or two hand
samples taken during testing.
Since gas measurement was not a priority at this stage of the project, results were not validated against any reference
meter.
Objectives of the Testing. Establish, with field tests, whether the equipment reaches the repeatability and accuracy
indicated by the manufacturer. The requirements were:
• Accuracy. ±10%, i.e., the average of all “deltas” between both measurements, tank and equipment, should be less than
±10%. Confidence interval of 90%.
• Repeatability. ±2%, i.e., between measurements from the same well and operating conditions, the data gathered from the
equipment should have a “delta” no greater than ±2%.

Test Unit. As mentioned previously, the PSS used in this evaluation was not originally designed for this particular
application. It was agreed to test the system in this particular oil field to evaluate the overall capability of a PSS to efficiently
separate gas from extra-heavy and diluted oil. The PSS was designed using a the methodology developed by The University
of Tulsa (Ref. 3).
Design specification of the test unit as follows:
• Built to ANSI B31.3 piping code
• ANSI 600 system
• Design pressure = 1,020 psig at 200°F
• Design Temperature = 200°F
• GLCC O.D. = 10 in.
• GLCC Inlet = 8 in.
• Inlet/outlet connections = flanges 4 in. 600# RF
• High pressure hoses = 3 in. 1000 psig (one end 4 in. 600# RF; other end 3 in. 300# RF)
• Skid estimated dry weight = 5,900 lbs.
• Overall dimension (H × W × L) = 16.5 × 7.5 × 5 ft.

The test unit working window design is presented in Table 1. The PSS used for this field test was originally designed to
handle fluid viscosities up to 10 Cp; however, field test were performed with API gravities from 11 to 19 and viscosities over
150 Cp. The difference between design and actual field conditions affect the total accuracy of the PSS.
4 SPE 116460

Table 1. Test unit working window

Parameter Units Min Max


Liquid Flow Rate BPD 314 3,774
Gas Flow Rate (Leg A) SCFD 100,000 1,060,000
Gas Flow Rate (Leg B) SCFD 7,000 100,000
Average Water Cut % 0% 100%
Operating Pressure PSIG 29 200
Operating Temperature DEG F 100° 250°
Viscosity Cp 10
Liquid Accuracy % ± 5%
Gas Accuracy % ± 3%
Water Cut Accuracy % ± 3%

Process & Instrumentation


The main focus of the evaluation was to determine whether PSS technology was capable of efficiently separating and
independently measuring gas and liquid flow rates before recombining them back into the process.
Three fundamental components of the PSS were evaluated during this field test: GLCC technology, level control logic
and instrumentation.
Gas-Liquid Cylindrical Cyclone Technology. GLCC is the core technology of the PSS. Fluid is introduced into the
system at a calculated angle to induce the optimal centrifugal rotation necessary to separate gas from liquid. It is important to
note that the separation of gas from liquid inside the GLCC does not have to be perfect. One main advantage of PSS is that
when it is combined with the right instrumentation, it prevents at least small amounts of GCU from affecting the overall
accuracy of the system.
The level inside of the GLCC is measured using a differential pressure transmitter. It was maintained by pneumatically
actuated control valves on both the gas and liquid legs. Fig 3 shows the main parts of the PSS.

6
1. Level Transmitter
2. Liquid Flow Meter
1
3. Infrared Water Cut Meter
4
4. Liquid Control Valve
3 5. Gas Flow Meter
6. Gas Control Valve
2

Fig. 3. Partial Separation System

Level Control Logic. Level control logic is a critical component of a PSS. The proprietary control logic allows for
simultaneous control of both the gas and liquid control valves to maintain a specified level inside the GLCC; known as the
“level set point”.
The level set point can vary on a well-to-well basis depending on fluid properties and flow rates of each well. An ideal
level to maintain in the GLCC is critical to finding a balance in minimizing GCU and LCO in the system.
Instrumentation. The right combination of instruments is vital to the overall success of a PSS. Separation inside the
GLCC is not always ideal, therefore GCU and LCO can adversely effect the accuracy of metering instruments.
Gas. The test unit used in this application had two gas legs: a 1 in. and 2 in.. A Coriolis and vortex shedding meters were
used on both legs to measure gas flow rate. Since this system was designed as a test unit, both metering technologies were
employed to compare their performances at different flow conditions. Although gas measurement was not validated against a
reference, it was still important to monitor what was occurring in the gas legs during testing. Sight glasses were installed on
each leg to provide the operator with a visual inspection point for LCO. Analyzing data from the Coriolis mass meter also
gave us an idea of whether or not there was any kind of liquid in the gas leg.
Liquid. Liquid measurement was carried out on a 2 in. liquid leg with a Coriolis and infrared water-cut meter. The
Coriolis meter measures a liquid mass rate using two oscillating tubes while the infrared water meter (Weatherford Red Eye®
SPE 116460 5

2G) uses a patented optical sensor technology to accurately measure the full range (0 to 100%) of oil and water concentration
in the liquid stream.
The main advantage of using a Coriolis meter is that gas carry under and foamy mixture have minimal effect on liquid
mass rate readings. However, they do drastically affect density readings, making the volume flow rate from the Coriolis
meter unreliable. It was therefore, necessary to find alternative ways of calculating liquid volume rate.
The ability of the infrared meter to calculate an accurate water-cut percentage with up to 15% GCU, makes it an ideal
collaborator to the Coriolis meter. Using accurate water-cut readings, the flow computer is able to calculate an instantaneous
mixture density that is not affected by GCU up to 15%. In order for infrared meters to read accurately, well-specific
calibrations must be performed. However, in this application, where oil properties (density and infrared absorption) change
with diluent rate and flow pattern, it was the objective of this field test to evaluate the behavior of the combination Coriolis-
infrared meters when using well-specific or field average calibration factors. For this reason, an average fixed oil density of
61 lb/ft3 was used to correct the actual density measured from the coriolis, and diluted-dry oil samples from well pad A were
used to calibrate the infrared water-cut meter.
The fixed density value was calculated by averaging the density of oils between 15 and 25% dilution. Using a fixed
density was also helpful in that the rate of diluent injection remained constant, but the flow rate of each well did not, making
it difficult to obtain accurate real-time density values. On the other hand, not using well-specifics for well pads B and C
resulted in larger water-cut inaccuracies than what was to be expected.

Findings
Impact of Diluent. During standard operating conditions at the Petrozuata field the diluent percentage was maintained
between 15 and 25%. The amount of injected diluent played a critical role because it helped reduce the pressure drop in pipe
lines, WHP (well head pressure) and to raise the API of the blend oil to help increase the production rate. Therefore, during
the preliminary testing phase, different diluent rates were tested to see what effect it had on fluid properties and the overall
performance of the PSS.
It was observed that when only a small percentage of diluent was added to a well (below the minimum 15%), the
combination of heavy oil and diluent caused the liquid to become a foamy mixture leading to an overestimation of total liquid
volume. This was discovered after analyzing real-time data recorded from the flow computer. When the liquid control valve
was fully closed to raise the level of liquid inside the GLCC, the breakout of the entrained gas in this foamy mixture resulted
in the Coriolis meter reading its movement through the liquid as mass flow. These mass readings were added to the
accumulated liquid-flow counter leading to a liquid volume overestimation. This phenomena can be observed in Fig 4.
300
Liquid Mass Flow [lb/min]

250

200

150

100

50
0

Instantaneous Liquid Mass Rate Lb/min


100
Liquid Valve Position [%]

90
80
70
60
50
40
30
20
10
0
Liquid Valve Position %
Time

Fig. 4. Mass readings when dilution is below 15%.


6 SPE 116460

500

Liquid Mass Flow [lb/min]


400

300

200

100

0
Instantaneous Liquid Mass Rate Lb/min
100
Liquid Valve Position [%]

90
80
70
60
50
40
30
20
10
0
Liquid Valve Position %
Time

Fig. 5. Mass readings under normal conditions.

In order to compensate for these false mass readings, the mass rate low-flow cut off value was raised from 6 lb/min to 12
lb/min. It was noticed that when the liquid control valve opened, the mass flow rate readings of each well were above 12
lb/min therefore there was no risk of ignoring any real flow values. Under normal conditions the PSS successfully separated
the majority of gas from the mixture of heavy oil and the diluent using a precise level control set up. With minimal GCU, the
combination Coriolis and infrared water-cut meter were able to measure the liquid-flow rate. Fig 5 illustrates how the control
logic functioned during this condition.
Impact of adding a large percentage of diluent (above 25%) was also tested to see its effects on liquid metering. In this
case, total liquid volume production was found to be underestimated. This phenomenon was mainly observed on low-
production wells, where the flow rate of diluent was kept steady while the well flow rates fluctuated constantly. Although the
mass measurement from Coriolis was very accurate, the fixed density (used to reduce the impact of bled oil density
fluctuations when handling gas) was over estimated leading to an underestimation of the liquid volume rate calculation. It
was recommended for low production wells to keep the dilution rate below 25%.
The last case tested was one in which there was no diluent injection into the well. Though well testing is never done
without diluent due to the pressure drop in the test line, it was important to see what effects extra-heavy oil had on the overall
performance of a PSS. Analysis of the Coriolis meter and liquid control valve data Fig 6 shows that although the liquid valve
was open the entire time, the liquid mass rate from the Coriolis meter read zero.
When the gas control valve was fully closed, the gas build-up inside the GLCC forced the liquid (extra-heavy oil) down
through the leg at higher rates. Once the gas valve opened, the relief of back pressure inside the GLCC caused the liquid rate
to slow down below the low-flow cut-off value of the Coriolis meter resulting in zero readings. In order to alleviate this issue,
the low-flow cut-off value would need to be lowered to almost zero to measure these low-flow rates. No corrective actions
were taken regarding this scenario because, as mentioned earlier, well testing was always done with the addition of diluent.
The data confirmed that PSS performed accurately during the normal operating conditions of 15 to 25% dilution.
Variation of diluent percentage outside the normal range did have an impact on the total liquid measurement but its effect
was countered by making minor adjustments to the GLCC level and low-flow cutoff settings.
Findings regarding the impact of diluent are summarized in Table 2.
SPE 116460 7

100
90

Liquid Mass Flow [lb/min]


80
70
60
50
40
30
20
10
0
Instantaneous Liquid Mass Rate Lb/min
100
Liquid/Gas Valve Position [%]

90
80
70
60
50
40
30
20
10
0
Liquid Valve Position %
Time
Gas Valve Position (A) %

Fig. 6. Mass readings and valves position (extra-heavy oil).

Table 2. Summary of findings (impact of diluent).

Findings Observed Effects Solution


High Viscosity: - Total liquid volume underestimation. - Increase dilution over 15%
- Extra-heavy oil - Low velocity in liquid leg, below cutoff
- Low dilution (<15%) value of the instrument.
Foamy oil: - False mass rate readings when liquid - Increase dilution over 20%.
- Diluted oil leg is closed. - Increase liquid cutoff value.
- Total liquid volume overestimation. - Increase GLCC level set point.
- Increase of GCU
High API: - Total liquid volume underestimation. - Reduce dilution within the
- Low flowing wells and high - Waste of diluent. recommended range (15-25%).
dilution(>25%).

Control Logic. In a PSS, control logic is critical to the operation and successful separation of gas from liquid inside a
GLCC. Without a proper control logic in place; GCU and LCO can be detrimental to the overall success of the system. As
discussed above, GCU and LCO can result in an over or under-estimation of total liquid volume. That is why it was
important to fine tune the control logic for the wells being tested; especially since this particular PSS was not designed for
this field.
With the addition of diluent, LCO was a concern when the crude became a foamy mixture resulting from high amounts of
entrained gas in the liquid. Due to the foamy nature of this mixture, liquid level in the GLCC rose higher, making it easier for
the free gas coming in from the inlet to carry this lighter liquid to the top of the GLCC and into the gas leg causing high
LCO. Sight glasses on the gas legs were helpful in visually observing this phenomenon. This was countered by lowering the
set point, allowing the gas and liquid mixture to spin below the inlet arm providing better separation in the vertical separator
and letting the free gas, and a majority of the entrained gas to escape from the mixture carrying only minimum amount of
liquid to the top.
Lowering the set point too low however, started resulting in GCU through the liquid leg. To counteract this, the level set
point of the GLCC had to be slightly raised to allow for a longer residence time of the liquid inside the GLCC. This permitted
the entrained gas in the foamy mixture to break out and release into the gas leg of the system. Even though minimal amounts
of entrained gas still traveled through the liquid leg, the instrumentation used in this system was capable of maintaining
accurate readings. After a few attempts of fine tuning the level in the GLCC, an appropriate set point was determined where
both LCO and GCU were at a minimum, and overall system accuracy was at a maximum.

Results and Discussions


During the formal testing of the PSS in Petrozuata, water-cut, total liquid volume production and total gas volume production
were calculated and recorded by the flow computer. PSS’s performance regarding these three vital parameters will now be
discussed in detail.
8 SPE 116460

Water-cut. The infrared water-cut meter is a well-specific device. The meter must be calibrated with dry oil samples
from each well and then a combination of wavelengths must be manually selected to give precise water-cut readings.
Therefore, it’s necessary to calibrate the meter with dry oil samples from all the wells on which the PSS is tested to obtain
high water cut accuracy. However, it was requested that the meter should only be calibrated using oil samples from well pad
A for all tests. This was done to evaluate the performance of the meter with a generic set of calibration values. Results from
the forty-two tests are shown in Fig 7.

25.00%
Calibrated Red Eye® Generic Calibrations No. Wells: 16
Well Pad A Well Pad B & C API: 11-19
Temperature: 98-110 F
20.00% Pressure: 100-170 psig
WATER-CUT % DIFFERENCE

15.00%

10.00%

5.00%

0.00%
0 5 10 15 20 25 30 35 40 45

-5.00%

-10.00%
REMMS Water-cut Difference
-15.00%
TEST NUMBER

Fig. 7. Water-cut measurement results.

When using well-specific calibrations, the infrared water-cut meter performed within the ±5% accuracy for all but one
test on well pad A. The larger errors can be attributed to the reference comparisons being used. While the flow computer
accumulated a total average water-cut over the entire time of the test, reference comparisons were averaged from samples
taken at one or two occasions during the whole test. Using only one or two hand samples to validate results is not reflective
of an overall average of a well; water-cut percentages fluctuate constantly so it is necessary to take numerous samples
periodically, in order to validate results.
For well pads B and C, the same calibration values from well pad A were used. During this round of tests, there were a
greater number of results that fell outside of the ±5% accuracy. This is mostly attributed to the use of non-specific oil
calibrations. Infrared meter calibrations can vary significantly from well to well depending on the oil’s absorbance
characteristics. Having the right calibration values along with choosing the right combination of wavelengths can improve
results drastically.
Liquid Volume. The main focus of this evaluation was to test how well a PSS was at measuring total liquid volume
production. Coriolis mass meters were chosen for liquid-flow measurement mainly because low amounts of GCU has little
impact on mass-rate accuracy, making them a good fit for the PSS where small amounts of entrained gas was occasionally
present in the liquid flow. Results from the forty two tests can be seen in Fig 8.
SPE 116460 9

15.00%

TOTAL LIQ. PRODUCTION % DIFFERENCE


TEST 2
10.00%

5.00%

0.00%

-5.00%

-10.00%
TEST 28

-15.00%

-20.00% No. Wells: 16


TEST 3 API: 11-19
Temperature: 98-110 F
-25.00% Pressure: 100-170 psig
TEST 30

-30.00%
0 5 10 15 20 25 30 35 40 45
TEST NUMBER
REMMS Total Volume Difference

Fig. 8. Total Liquid Volume Results

Overall, results show that a majority of the tests were within the ±10% accuracy. There were only four tests that fell
outside this range (Tests 2, 3, 28 and 30). During test 2 and 3, the control logic was still being fine tuned to optimize the
performance of the test unit. Once a correct configuration of GLCC level was discovered, the readings from the Coriolis
meter improved drastically resulting in total volume accuracies within ±10%.
Test 28 and 30 were performed on the same well. Significant slugging was witnessed on this well during these tests,
resulting in a considerable amount of GCU which negatively impacted the liquid volume readings. Under normal
circumstances PSS can generally handle heavy slugging by adjusting the level in the GLCC, in this situation, however, the
test unit was not designed for the large amounts of instantaneous flow, so it was not possible to raise the level any further.
Regarding repeatability, the largest difference between tests from the same well, at the same flow conditions (pressure,
temperature and diluent rate) was only 1% in total volume error, well within the ±2% requested by Petrozuata.
These results proved that a PSS is capable of handling the extra-heavy and diluted oil found in this field. It is important to
keep in mind that Frac tanks were used in this case as a reference to compare the results.

Conclusion
The Petrozuata MPM evaluation program was conceived to test the performance of the PSS in a heavy oil application. After
executing a comprehensive test plan and analyzing the results that were subsequently obtained, the testing of the PSS was
categorized as “successful” based on the predetermined success criteria of overall volumetric flow accuracy of +10% and
performance repeatability of ±2%.
During normal operating conditions, when the diluent rates were within a 15 to 25% range, the PSS was successful in
finding an efficient way of separating a majority of the gas from the mixture of heavy oil and diluent, due to a precise level
control set up. The Coriolis and infrared water-cut meter proved to be the best combination for liquid-flow metering in a
heavy oil applications, because of their capability of handling certain amounts of entrained gas without losing accuracy.
Injected diluent rate was determined to be one of the main parameters impacting the liquid metering performance.
Therefore, we extensively tested a wide range of diluent percentages, including those outside the normal operating range. It
was observed that when the diluent rate was low, the combination of heavy oil and diluent caused the liquid to become a
foamy mixture leading to an overestimation of total liquid volume. At a high diluent percentage the total liquid volume
production was found to be underestimated. Since large amounts of diluent decreases the density of oil, using a fixed density
caused the total liquid volume calculations to be lower than they actually were. Using live density values instead of a fixed
number would have been helpful in accurately calculating the liquid volume. When no diluent was added, and in the absence
of any backpressure, the flow rate of the extra-heavy crude through the liquid leg was lower than the Coriolis low-flow cut
off, leading to the underestimation of the total volumetric flow.
When using well-specific calibrations, the infrared (Red Eye® 2G) performed within the ±5% accuracy on all tests except
one. That error was attributed to the improper reference reading. Global calibration values were used during the testing of
well pads B and C, this resulted in a greater number of tests falling outside of the ±5% accuracy limit. The conclusion is that
individual well calibrations must be performed for the water-cut meter, to meet their ±5% accuracy specification.
10 SPE 116460

The Petrozuata MPFM evaluation program was designed to test each of their performances in a heavy oil application. It
was determined that the collaboration of these three led to an efficient separation of gas and liquid in the GLCC by using
precise level control, with minimum GCU. The Coriolis-infrared water-cut meter combination was then able to accurately
meter liquid flow rate with some entrained gas.
Throughout the entire duration of the project, the three fundamental components of the PSS - GLCC based separation
technology, level control logic and Coriolis-infrared meters combination for liquid metering were evaluated under severe
flowing conditions. The Petrozuata MPFM evaluation program was designed to test each of their performances in a heavy oil
application. It was determined that the collaboration of these three led to an efficient separation of gas/liquid mixture (by
using a precise level control of GLCC), accurately measuring the liquid volume and water cut.
The Weatherford PSS’s open architecture (not a “black box”), flow-metering technology, flexibility, ease of operation,
maintenance and diagnostics proved to be a good solution for the extra-heavy and diluted oil applications.

Acknowledgements
The authors would like to thank Petrozuata and Weatherford International for the permission to publish this paper. A special
acknowledgement should be offered to Petrozuata field personnel, operators, production engineers and management team for
all their cooperation and help. Without their support this work would never have taken place.

References
1. A Review of Gas-Liquid Cylindrical Cyclone Technology. By G.Kouba and O.Shoham. Presented at the Production Separation
Systems International Conference. April 1996.
2. Two-Phase Flow Modeling. By Ovadia Shoham. The University of Tulsa. April 2000.
3. Aspect Ratio Modeling and Design Procedure for GLCC Compact Separators. By L.Gomez, R.Mohan O.Shoham, J.Marrelli and
G.Kouba. Presented in the Journal of Energy Resources Technology. March 1999.

You might also like