Paper 7.

First Ever Complete Evaluation of a

Multiphase Flow Meter in SAGD and
Demonstration of the Performance
Against Conventional Equipment

Bruno Pinguet
Schlumberger

Fernando Gaviria
SUNCOR Energy

Laurie Kemp
SUNCOR Energy

John Graham
SUNCOR Energy

Cal Coulder
SUNCOR Energy

Carlos Damas
Schlumberger

Karima Ben Relem

Schlumberger
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

First Ever Complete Evaluation of a Multiphase Flow Meter in

SAGD and Demonstration of the Performance Against
Conventional Equipment
Bruno PINGUET, Schlumberger
Fernando GAVIRIA, Laurie KEMP, John GRAHAM and Cal COULDER, SUNCOR Energy
Carlos DAMAS and Karima BEN RELEM, Schlumberger

1 INTRODUCTION

Suncor and Schlumberger have been working over several years to establish the
performance of multiphase flow meters and determine the accuracy and reliability of flow-
testing equipment for thermal heavy oil production, the first test was carried out in 2007. The
test marked the first successful application of multiphase flow metering in Canadian thermal
recovery operations.

The initial introduction leads to selection of a well for its dynamic conditions and production
history factors such as flowrate, line pressure, line temperature, and fluid properties and
allowed the connection of the MPFM (the Vx) in series with an existing test separator.

The multiphase initial objectives were to measure the well’s oil, water, steam, and gas
flowrates, understand the well’s SAGD dynamics, and the impact of changes in ESP
frequency to experience different production conditions, including line pressure, flowrates,
phase fractions and ratios, i.e., Water/Liquid Ratio (WLR). Finally, to study the Vx’s sensitivity
and performance against a valid reference by defining the results for stability, dynamic
response, repeatability and reproducibility. This leads to several papers presented in different
conferences (Ref [1-3], and in the meantime to work closer to define what should be the
optimal test of a cluster to have a full understanding of the technology and the potential for
permanent installation.

In 2009, a second and more advanced field trial was initiated. A pad was selected as the
location for this field test due to the large number of good producing wells and ideal set up of
the conventional equipment. The MPFM was connected in series with the conventional test
separator and the test was divided into three stages to fully evaluate the safety, reliability,
accuracy, and advantages of the MPFM over the conventional test separator equipment.
During the first stage, 9 wells were tested during the months of October and December 2009,
ensuring that the MPFM was challenged against the full range of operating conditions of the
SAGD wells. During two months of extensive well testing and fluid sampling, no safety
incidents were encountered and the MPFM reported well production rates continuously and
without any performance degradation even during the extreme cold ambient temperatures
recorded at the well site.
To demonstrate the reproducibility of the flow measurements using the MPFM, a second
stage was completed, this time shortening the flow period from 12 hours to 6 hours and then
be able to demonstrate the capability to do more well tests per month per well with the same
quality data. Then, with the Vx meter only, the data of the 12 hours testing were reprocessed
for 4 hours showing a very good consistency in the results with the 12 hours test.
Results, from the second stage of the test confirmed that the data is identical for a 12 hours
test versus the 6-hour test versus the 4 hours test and has very good repeatability and prove
the robustness and independence of the flow regime. These results seem to reveal that
testing for shorter period of time and more frequently gives more information and is likely
preferred to be able to optimize production.
A third phase more dedicated to optimization potential and the impact of changing chokes,
steam injected rates and other numerous parameters was done but it will be only briefly
mentioned in this paper.

1
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

2 PROBLEM AND PROPOSED SOLUTION

The SAGD process presents one of the most challenging environments for flow rate metering
in the industry. Produced fluids are often characterized by unstable flow regimes, high
temperatures, emulsified foamy oil, as well as the presence of H2S and abrasive sand
particles. As such, the capture of good data is extremely difficult.

Timely and accurate well production data is required for the validation process of the reservoir
model and for confirmation of production predictions. Additionally, the oil company could
potentially utilize this valuable information to adjust the operation for optimum performance.
Currently, the industry attempts to meet this need using “test separators”, which employ the
basic principle of gravity of the various phases that produce through the wellhead. Additional
standard equipment, such as a water-cut meter, Coriolis, and orifice plate technologies are
then employed to obtain the measurements.

However, it is well known that there are many issues, which compromise the accuracy and
utility of the data retrieved by this conventional equipment. First, as oil and water density
contrast is very low, the gravitation process is not effective and do not allow for proper
separation of the gas and liquid phases. Second, there are issues regarding the proper
measurement of the mix of steam vs. natural gas, of oil vs. water, and with known errors from
gas entrainment within the liquid lines. Lastly, because many wellheads share a single test
separator, flow measurements from each individual well can only be periodically obtained.
Some additional backpressure may also affect the estimated well production.
These factors lead to a limited understanding of each well’s performance and a negligible
capability for optimization. Thus, it is desired to find a technology that could replace the
separator system with a real time and accurate metering system.

The primary goal of this collaboration is to develop production techniques that will improve the
productivity and recovery factors of viscous oil sand reservoirs in Canada, starting with
Suncor’s Firebag Asset. Firebag is located 140 kilometers northeast of Fort McMurray,
Alberta, Canada (Figure 1). The first barrel of bitumen produced from this operation was
processed in January 2004.

Figure 1: Field location

2
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

With a viscosity of millions of centipoises [cP], the oil is not producible through ordinary cold
production practices. Rather, the bitumen is extracted via an advanced and complicated
recovery technique known to the industry as SAGD (Steam Assisted, Gravity Driven).

From this basis, a collaborative project plan was developed with the mission to bring
optimized workflows and new technologies to Firebag that will achieve a reduction in Steam-
to-Oil (SOR) of at least 0.50 while reducing the environmental footprint. The project was
planned to complete three phases, each with a deliverable during a calendar year:

 Phase One: Enable Firebag to have real-time accurate production flow data

 Phase Two: Working with this data, enable the operation with advanced
monitoring, automation, and optimization systems

 Phase Three: Once optimization techniques are available, add the capability
for active/adaptive completion technologies (e.g. steam injection control)

Figure 2: Field Operation and Equipment Proposed

Only after completion of the phase#1 successfully, the project could advance to Phase 2 of
the aforementioned mission to significantly reduce Firebag's SOR and environmental impact.
To answer all different issues rose above and if this is possible, the solution will be based on
a system without any separation. This paper presents the fieldwork that was conducted during
Q4, 2009 and completes the deliverable for Phase One.

It is proposed to replace the centralized test separator system with a Multiphase Flow Meter
(MPFM), shared either by many wells or ideally at each wellhead for full real time capability.
Such a set up could provide a step change in performance and a reduction in surface
footprint. In addition, continuous well flow measurements could be used to optimize the
artificial lift performance, maximize oil production, and improve the energy balance. If this
present a step change in SAGD conditions, Vx technology has been already accepted in
conventional field development. Approval for allocation in North Sea has been done (Norway
through NPD, UK with DTI) or Gulf of Mexico (USA with MMS).

Schlumberger has successfully developed an advanced MPFM, based on the "Vx technology"
and available as mobile unit (PhaseTester used for this test) and permanent unit

3
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

PhaseWatcher that may provide an ideal solution. Originally designed for challenging deep-
water markets, this multiphase flowmeter has been used in hundreds of applications and has
proven to be reliable and accurate. Over the past few years, this device has been introduced
to the SAGD/Thermal environment via several field trials in Western Canada and Mexico.
Because of these learning, many improvements (e.g. increasing the temperature rating) have
since been commercialized.

3 PROPOSED SOLUTION

An overview on the basic operational principles of MPFM devices was given in 2008 at the
World Heavy Oil Congress [2] and is partially reproduced below for convenience. At a
minimum, all MPFM devices will output the oil, water, and gas flow rates at line conditions.
These outputs are derived using either a physical or an empirical model. However, industry
reports production at standard conditions and therefore the flow rates at line conditions need
to be converted to standard condition using PVT correlations or more properly named “fluid
behaviour model”. Capturing this flow behaviour can be done by simply looking at few macro
parameters: Volume factors (bo, bg, bw), Gas dissolved in Oil (Rst), Gas Dissolved in Water
(Rwst), and the water phase condensation (rgwmp).

Figure 3: Field Operation and Equipment Proposed

Of course the knowledge of the density of each fluid at line conditions will be necessary and
can be obtained based on the above parameters and the knowledge of the dead water and oil
and specific gravity of the gas.

In the SAGD environment, it is a combination of a few simple and basic measurements done
at the well site or through some laboratory analysis, or information available through a PVT
report, which will be used to tune the EOS (usually build for a pad or a flied following the
homogeneity of the reservoir). The combination of the accuracy of the representation of the
mass or volume converted from line to standard conditions with the intrinsic accuracy of the
multiphase flow meter (specific to each technology and then multiphase manufacturer)
provides the global accuracy of the measurement of live effluent. The physical principle and
model has been done by a review systematically of each implication to work in SAGD. This is
the result of 10 years of focus and understanding in details of the Vx technology.

4
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

The Vx multiphase flowmeter (figure 4) contains three main modules:

 An advanced multiphase flow model of the fluid mechanics,

 A robust measurement of the total mass flow rate based on the well-known
Venturi design, and differential pressure concept

 A robust fraction meter measurement based on established nuclear

technology, to produce in real-time fraction measurement of oil. Water and
gas without any separation and being not sensitive to the phase distribution
(emulsion, foam, droplets flow, water contents...).

A detailed explanation of the specific data flow processing routine and various assumptions
made within the Vx technology can be found in the referenced documents [2]. However, a
summary is listed below for convenience:

Figure 4: Field Operation and Equipment Proposed

Two basic measurements are made in the Venturi section:

 Differential pressure across the Venturi (∆P), and

 A multi-energy gamma-ray fraction meter, which provides fraction of each

constituent of the mixture (αo, αw, αg).

Additionally, there are sensors that measure process fluid pressure at the Venturi throat, fluid
temperature upstream of the Venturi section, and ambient temperature.

It is essential here to mention some fundamentals of the fluids mechanics to be able to have
an accurate multiphase flow meter in general. First, the flow is turbulent or chaotic and then it
is necessary to have a system working at a frequency much higher than several 10th of Hz.
Second, the different fraction measurement needs to be correlated; this means that from a
recording time point of view all measurements should be done at the same time. Third,
because the flow is turbulent then the timing is important and because the fluid is a live fluid
then the acquisition should be done at the same space. Fourth, this recording should be done
at the same pressure and temperature. With an example is easy to understand, if for example
the water holdup was done in one place at a certain pressure and temperature or done

5
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

somewhere else at different pressure and temperature then the condensed water will be
different and then the overall calculation of the water content could be either overestimated or
underestimated,

The Vx technology is built on a few physical hypotheses: No slippage inside the liquid phase,
a semi-empirical model of the gas-liquid slippage based on extensive research and
experiences made at low pressures (from 1 to 30 bar) with immiscible fluid, and finally a
model of a “shape factor” for multiphase environment.

Figure 5: Field Operation and Equipment Proposed

Similar to the common dataflow for all MPFM devices, the specific Vx schematic is shown
Figure 5. As already mentioned the volumetric flow rates at standard conditions are the main
outputs and computed from the flow rates at line conditions using either a PVT software
package or a correlation (as it is currently done for the separator).

Figure 6: Field Operation and Equipment Proposed

6
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

The Vx technology is the simplest combination of sensor and technology available on the
market today with only 4 sensors and none of them in direct contact with the fluid avoiding the
high temperature stress.

An additional challenge for SAGD operations is to monitor at surface: Steam (i.e. Vapour of
water), Natural Gas evolved from the Oil, Bitumen, and then Water (Liquid form). This means
there are four phases versus the traditional three at line conditions. Therefore, it is necessary
to use additional information to obtain this fourth phase measurement. The Vx device can
provide four phase measurements and split the fractions of Steam and Natural Gas by using
a dedicated Equation Of State (EOS) package based on the knowledge of the fluid properties.
This allows keeping the multiphase flow meter design as simple as possible without additional
sensors, cost, and maintenance program.

The Vx technology has a unique application within SAGD operations due to the use of a
specific heating device, which allows the entire nuclear-based instrument to be kept at a
specific temperature, and can handle any fluctuation of the fluid temperature passing through
with minimum impact on the accuracy of the fraction measurement.

Figure 7: Field Operation and Equipment Proposed

Extensive testing has demonstrated that under the harsh conditions (no wind, ambient
temperature at 25 degC, and flow temperature at 232 DegC) the sensor will not exceed 80-90
DegC in the worst conditions when the electronic has been designed to handle more than 100
degC in normal operation. In addition, The structural design of the Vx Multiphase Flow meters
complies with the reference codes and standards of Canada or International.

The key points of the Vx Technology are summarized below:

 No Tuning Factor due to a proper understanding of the fluid mechanics

 No Separation by construction
 No Flow Calibration the measurement being based on physics principle
 No issues with Foams or Emulsions due to nucleonic technique of
measurement
 Improve water detection and accurate BSW monitoring (either oil or water
continuous phase)
 Continuous rate measurements
 No moving parts
 Different sizes: 29mm, 52mm, 88mm (100 – 150,000 bpd)

7
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

 Due to the Venturi principle used in the Vx technology an extremely low back
pressure
 Compact solution (<300 kg, 2 feet tall)
 No Sensor in direct contact with the fluid
 Full sampling capabilities (optional)

The Figure 12 presents a comparison in footprint between the conventional Test Separator,
the Vx PhaseTester (periodic applications) and the Vx PhaseWatcher (permanent
installations).

Figure 8: Field Operation and Equipment Proposed

4 MOTIVATION FOR A SECOND TRIAL Vx TEST

As described above, the Vx technology was used for the first time in SAGD in 2007 at the
Suncor Firebag Project. This represented the first introduction of the technology to SAGD
environments and was published in different conferences around the world in 2008 (Ref. [1-
3]).

The decision to initiate a second Vx field trial with Suncor and finalize the introduction of Vx
for SAGD was a result of the completion of several new improvements of the test separator
and a tighter focus on the metrological performance comparison versus 2007. Additionally,
the first Vx Trial in 2007 was specific to a single well in the SAGD pad. The second trial is
covering a much wider range of wells and operating conditions in order to prove the accuracy
and reliability of the Vx flowmeter in all conditions encountered during a SAGD project
operation.

Schlumberger has been working with another operator throughout 2008. During this particular
project, extensive testing was done and both parties were able to challenge the Vx
Technology and conventional equipment to review how a multiphase flow meter could be
becoming a reference measurement in SAGD operations (Ref [4]).

8
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

The main improvements after the 2007 Vx trial are summarized below:

 Design and implementation of a sampling equipment to collect Gas-Steam

samples at line conditions and be capable to measure the steam and natural
gas ratio
 Design and implementation of a sampling equipment to collect liquid samples
at line conditions to be used as reference WLR measurement due to lack of
robust conventional sensor for SAGD conditions
 A comprehensive review of the fluid properties behavior, development of the
full PVT to handle either 3 or 4 phase in SAGD at the well site (Oil, Natural
Gas, Vapor of water, Water)
 A complete review with Engineering center of the future improvement to do in
th
the commercial software to be SAGD compliant (fluid behaviour and 4
phase compliant)
 Starting the process of an entire review by DNV of the equipment for SAGD
operation as a permanent multiphase flow meter
 Improvement of the Vx WLR measurement at line conditions through a new
algorithmic process
 A proper understanding and a written procedure to handle dead sample to
provide the Reference
 An update of the EOS generator (PVT Pro) for SAGD field and taking into
account the steam ratio
 Validation of the Vx measurement at higher temperature then 190 DegC
 Validation of the Vx measurement in environment with water and high H2S
content in SAGD
 Demonstration of the good accuracy in PCP conditions
 Review of the performance in details of the multiphase flow meter (flow rates)
versus conventional equipment and definition of the strong and weaker point
of the multiphase solution versus conventional equipment (Ref. [4]).
 Use in unstable conditions to capture the flow dynamic and also in stable
production to see the same performance than the separator
 Use in the large concentration of CO2 (> 9%)
 Sensitivity analysis and what should be the maintenance program to define to
keep the same level of uncertainty in permanent application through the Vx
advisor software.

Based on this significant list of technical improvements, it was agreed by both parties that
performing a second Vx trial at Firebag will bring value to both companies and would help
reaffirm the capabilities of the Vx technology for SAGD conditions.

4.1 Qualifications & test Objectives

In order to qualify the Vx Technology for SAGD operation, a series of field tests were
conducted at Suncor's Firebag asset during Q4, 2009. The Vx Multiphase flow meter was
connected in series with the conventional test separator and the test was divided into three
stages with review of the performance at the end of each stage before progressing to next
stage and described as follows:

Stage #1: Benchmark (Vx vs. Test Separator)

This first stage will be to compare the performance of the Vx MPFM against the currently
utilized test separator.

1. Demonstrate Vx installation procedure and set up process

2. Demonstrate gas and multiphase sampling capability
3. Benchmark Meet or exceed all metrics currently employed by the operator
1. Accuracy & repeatability of all flow measurements (expected to be within 5%)
2. Frequency of test duration (12 hrs or less)
3. Capability to test each well head at least twice per week
4. Capability to derive Gas/oil Ratio (GOR) and Steam/Oil Ratio (SOR)

9
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

5. No failures or reliability issues

6. Safety

Stage #2: Advanced Vx Capabilities

This stage focus on additional capabilities versus current test separator.

1. Evaluate shorter test duration cycles (e.g. 4 hours vs. the current separator 12 hour
cycle)
2. Demonstrate accurate measurements of both natural gas and water vapor

Stage#3: Optimization Techniques

This last stage focuses on gaining an insight into the potential for optimization, which will be
the basis for the next two phases of the project.

1. Evaluate Vx dynamic response capabilities

2. Explore the potential for production optimization

4.2 Installation

The Pad which was selected for this field test due was upgraded lately and has a large
number of good producing wells and ideal set up of the conventional equipment. Wells were
producing either naturally or by artificial lift (ESP pump). The presence of a 10” by-pass
manifold on the pad production line and the capability to easily modify the current installation
by implementing two 8” spools to connect to the Vx meter was a non-negligible factor. The
metrological equipment of this PAD had also been totally updated or reviewed before this test.

For commodity and cost savings, the Vx mobile unit (“PhaseTester”) was used. However, it
should be noted that the permanent unit (“PhaseWatcher”) utilizes the same technology but it
is lighter and more compact. The only difference between these two tools is the mobility. For
the duration of the three-stage program, 9 of 11 wells were tested. The remaining two other
wells were not accessible or not producing at that time.

Figure 9: Field Operation and Equipment Proposed

The Vx meter was connected downstream of the test separator due to the design. Well
effluents exiting from the test separator were recombined before flowing to the MPFM.

10
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

Figure 10: Field Operation and Equipment Proposed

The schematic (Fig. 10) illustrates the basic set up of the test separator and associated
reference meters. As the two streams exit the separator, the gaseous phase is routed through
a vortex meter in order to measure steam flow and is recombined with the fluid flow later. The
liquid flows at the bottom first through water cut meter probe, then through a Coriolis meter
and liquid sampler bypass.

It should be noted that the flowrate measurements accuracy could be affected due to some
limitation associated with several required assumptions (listed below), but also intrinsic errors
cause by various sensors. As conclusion the separator results should not be considered to be
exact (like any others measurements) but the most correct one if the hypothesis and
equipment are fully within their working envelope.

Assumptions or additional data required for the current test separator system:

 Gas line is assumed to be only water vapor (no natural gas) leading to a
production of 2% more of condensed water in average
 Salinity input is required to get water-cut measurement due to the microwave
technology used and this equipment seems to be very sensitive to the salinity
parameter
 No gas is assumed to be passing through the liquid line
 No liquid is assumed to be passing through the gas line
 A Liquid Sampler by-pass is set to verify Water-cut Meter, and in average
through the data obtained it is more than 8% absolute error that can be noted
with in some cases more than 20% absolute deviations.

4.3 Multiphase Meter Settings

The setting of the multiphase flowmeter is based on a few input parameters that are
essentially associated with the nuclear fraction measurement device. The mass attenuation
parameter is an intrinsic parameter of the fluid composition and it is used to calculate each
phase fraction flowing through the venturi throat. It is strongly recommended in SAGD
operation that the mass attenuation determination for the first commissioning is performed in
the base (more control conditions) and then confirmed on site. However, it is possible to get
these intrinsic properties from the fluid composition. The mass attenuation is constant
whatever the temperature and pressure. Moreover, it is not dependent of the long chain of
alkans and then a composition in C5+ or C9+ is good enough.

11
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

Once a liquid sample is retrieved, the oil is separated from the water and processed in order
to obtain the most accurate mass attenuation data for each phase. From this sample, the
specific gas, oil, and water points are utilized to set the meter. Then when the commissioning
(i.e., configuration file downloaded to the meter) is finished, then a map as illustrated below is
available. Each point of the triangle represents 100% of each phase, and then any mix of the
3 phases will be within the triangle defined by the APEXES of oil, water, and gas (Figure xx).

Figure 11: Field Operation and Equipment Proposed

Based on the composition obtained from Suncor (via Hysis) a VxFluidsID (

i.e. fluid behaviour based on the EOS) was generated after recombination of the sample at
reservoir conditions and used as a generic basis of the PVT behavior for the entire pad. It
should be noted that a large concentration of H2S is coming from this type of oil and above
10%. The entire table set of the Vx Fluids IDs associated with the oil, water and gas
properties are not presented in this document. But it should be known that they are
automatically downloaded to the multiphase flow meter and represent the conversion of the
different parameters from line to standard conditions and vice versa. Specific care was taken
to account not only for water in liquid form, but also in vapor form.

5 OVERALL RESULTS

5.1 Dynamic Response Stage #1

During the Stage#1 of the Vx field trial, the meter was operated in “blind mode”, meaning that
three phase flow measurements were successfully delivered in real time prior to the
knowledge of the flowrate reference (i.e. separator test). Then, post-processing of the data
was performed to report 4 phase flow measurements: Oil, water, natural gas and vapor flow
rates. This was not implemented yet in the software version.

The figure 12 demonstrates the dynamic response to the flow of the Vx compared to the
separator. The separator operating pressure is indicated in green, the Vx line pressure in
blue, and line temperature in red. The high accuracy or performance is achieved via an
optimal design of the inline venturi with a blind “T” for flow conditioning and then eliminating
the issue due to change in flow conditions upstream of the meter. Moreover, this conditioning
process allows getting faster thermo equilibrium at any conditions. The separator’s large
vessel dampens the flow profile because of the need to keep the backpressure constant. The

12
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

pressure profile reveals that the separator pressure profile is flat while in reality there is a
pressure fluctuation during the test.

Figure 12: Field Operation and Equipment Proposed

The red line above confirms that wells were tested at an average of 180 °C flowing
temperature or above. This demonstrates the capability of the Vx to operate at SAGD
conditions.
The figure 13 shows a typical well recording versus time. A comparison Separator water cut,
and Vx WLR is shown.

Figure 13: Field Operation and Equipment Proposed

13
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

The results from the Vx measurements indicate that the flow is turbulent and the WLR is
never constant and always fluctuating. This could explain some of the large discrepancy
between the reference water cut and the liquid sampling procedure (used as conventional
system today inside SUNCOR) if unfortunately both comparisons are not synchronized. This
aspect of the flow dynamic is important since it could affect the validity range of the test data
and, therefore, the duration of each test.

The table 1 & 2 below present an overall performance summary for Stage 1 for all nine tested
wells combined.

Table 1 – Measurement from Separator and Vx meter

Qmass Qliquid Density CWE WLR WLR Qmass Qliquid Density WLR CWE CWE
Sep Sep Liquid Sep sep sep* Vx Vx Liquid Vx Vx Vx
[kg/s] [m3/h] Sep Cor [%] [%] [kg/s] [m3/h] Vx [%] [m3/h] [m3/h]
[kg/m3] [m3/h] [kg/m3]

88.2 353.6 896.4 12.8 88.8 77.2 89.6 355.8 905.7 79.5 10.7 10.7

Table 2 – Comparison Separator – Vx Meter

Vx -    Density  CWE  WLR* Steam
Sep Qmass Qliquid Liquid [m3/h] [%] injected
[%] [%] [%] [kg/s]

12Hr 1.59 0.62 1.04 -0.04 2.0 196.3

*Corrected Values by using the proration factor for the entire field
having no access to a proration factor ratio of the pad

The comparison between the Vx and separator WLR measurements shows a larger
discrepancy on average for all the wells using raw data from the reference water cut meter.
Vx reports consistently lower WLR values than the test separator in the blind test mode.
Further examination revealed that the pro-ratio factors of the entire field were indicating an
over-production of water and under-production of oil, i.e. the water cut at the separator was
too high. In conclusion, there is a significant error with the current water cut meter installed
and the technology proposed. The only way to try to do reconciliation is to correct based on
the proration factor which was done and presented in the table 1 & 2.

It should be noted that in parallel 50% of the liquid samples collected have been analyzed in
laboratory, and compared with the Vx measurements. They are well within the specification
with in average 1% absolute error based on the analysis of few wells. It should be noted that
with the same wells, there is large a discrepancy between the manual sampling done at the
separator and the WLR read from the water cut meter, and then confirmation the correction
applied above.

To summarize the main results from Stage#1:

 Continuous flow measurements provided for each well revealed good
dynamic response
 Benchmarked resulted between the Vx meter and the test separator
equipment are similar
 Total liquid mass flow rates are better than 2%
 Liquid densities are within 1%
 Recalculated WLR within 2%
 WLR within 1% versus half of the samples analyzed to demonstrate the WLR
validation form the Vx.
 No safety or reliability issues during Vx meter testing

14
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

5.2 Measurement Reproducibility Stage #2

To demonstrate the reproducibility of the flow measurements using the Vx technology, a

second test was completed, and this time shortening the flow period from 12 hours to 6 hours.
It is important to mention that only, Well A, B and C were the wells that preserved the same
flowing conditions (frequency, tubing choke size, tubing, and casing steam injection rate)
during both tests. This set of data is highlighted inside the Table 3.

Due to the fluctuation of the initial conditions it was then proposed to playback, the data
initially recorded over 12 hours but only during the 4 first hours. This will then show the
stability of the measurement and then the capability to establish the performance of a well on
a shorter period. It can be also interesting to note that the recorded raw data can always be
playback at any time. This could be instrumental for SUNCOR to demonstrate to ERCB the
performance of the well or be able to recompute data if a wrong setting was done on a
multiphase Vx meter, this is a unique feature of the technology.

With the Vx meter only, the data of the 12 hours testing were reprocessed for only 4 hours
showing a good agreement with the original 12 hours test. This supports the idea that the well
tests, if only done periodically, can be significantly shortened from the current separator-
based operation.

Table 3 – Overall performance with 12, 8 and 4 hours recording

Tubing Choke Frequency Tubing Steam Casing Steam Rate GVF

Size Hz Rate m3/hr %
Well
% m3/hr
12hr 6hr 4hr 12hr 6hr 4hr 12hr 6hr 4hr 12hr 6hr 4hr 12hr 6hr 4hr
A 45 45 45 39 39 39 4.0 4.0 4.0 10.0 10.0 10 95 96 96
B 55 55 55 41 41 41 4.0 4.1 4.0 8.0 8.0 8.0 97 97 97
10 10
C 100 0 0 0 6.0 6.0 6.0 34.0 34.0 34.0 98 98 98
0 0
10
D 45 45 56 54 56 4.0 5.0 4.0 35.0 16.0 35.0 78 80 78
0
E 35 40 35 53 50 53 8.0 4.0 8.0 35.3 8.0 34.0 62 68 62

F 45 40 45 55 55 55 8.0 4.0 8.0 25.0 8.0 25.0 21 56 23

G 52 50 52 54 52 54 8.0 4.0 8.0 21.9 7.9 22.2 50 49 48

H 40 40 40 49 49 49 8.0 4.0 8.0 20.0 7.9 19.9 79 77 79

I 65 65 65 57 55 57 4.0 4.0 4.0 7.9 10.9 8.0 78 79 78

Only the wells highlighted in gray kept exactly the same operating conditions during both tests.

As an example, well production measurements for one well is plotted (see Figure 14, 15, and
16) for the 12 hours, 6 hours, and 4 hours tests. The line temperature is presented in red, line
pressure in blue, liquid flow rate in yellow and gas flow rate in light blue.

15
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

Figure 14: Pressure, Temperature, Gas & Liquid flow rate over 12 hours

Figure 15: Pressure, Temperature, Gas & Liquid flow rate over 6 hours

16
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

Figure 16: Pressure, Temperature, Gas & Liquid flow rate over 4 hours

Table 4 presents a comparison summary on a well-by-well basis for the 12-hour, 6-hour, and
4-hour tests.

Table 4 - The Table Pressure, Temperature, and Condensed Water Equivalent for 12, 8
and 4 hours
Vx Pline (kPaa) Vx Tline (DegC) CWE (m3/h)
Well
12hr 6hr 4hr 12hr 6hr 4hr 12hr 6hr 4hr
A 656.4 661.6 647.9 167.4 167.6 167.8 1.21 1.30 1.23
B 804.9 799.1 813.0 178.2 178.2 178.2 2.56 2.56 2.59
C 855.7 851.6 850.4 183.1 183.0 183.2 3.75 3.71 3.75
D 838.2 772.5 838.4 176.6 174.5 176.6 0.80 0.77 0.79
E 857.2 771.9 857.5 175.6 172.8 175.5 0.39 0.47 0.40
F 766.3 667.4 767.0 168.2 164.9 168.2 0.40 0.18 0.40
G 788.7 685.9 762.9 174.1 166.6 174.2 0.26 0.18 0.24
H 764.8 736.8 765.6 172.2 171.7 172.2 0.66 0.59 0.66
I 757.6 736.4 757.6 172.3 171.5 172.3 0.66 0.64 0.67

Table 5 – Liquid Mass Flow rate, WLR comparison for 12, 6 and 4 hours

Liquid Mass (kg/s) WLR

Well
12hr 6hr 4hr 12hr 6hr 4hr
A 4.1 3.8 4.0 59.4 44.3 59
B 5.3 5.2 5.2 77.4 71.8 71.8
C 5.4 5.4 5.4 80.7 77.3 81.7
D 12.8 11.8 13 84.2 79.7 83.5
E 13.6 13.2 13.5 84.3 82.9 81.4
F 9.7 10.0 9.6 85.7 84.6 85.1
G 16.3 13.2 16.3 79.7 82.4 79.6
H 11.0 11.2 11 80.6 81.3 81.3
I 11.4 10.9 11 83.9 78.6 83.5

17
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

WLR measurement performance decreases with a GVF increasing, when the GVF average
over 10 seconds is above 96-98%, a cut off value has been applied in this case (i.e. No
calculation of WLR measurement if the value is above of the cut-off) in order to reduce the
noise associated with the measurement. It should be noted that a sluggy behavior will be
more favorable and will guarantee a better performance even if in average the GVF was
above 96-99%.

The presented data above confirms that the measurement are identical for a 12 hours test
versus the 6 hour test and even so the conditions may not have been entirely identical there
is a good repeatability over 4 hours test. These results reveal that testing for shorter period of
time and more frequently is providing the same level of confidence in the measurement, and
then it is possible to do more frequent tests and then get a better performance of the
production of each well over time.

5.3 Gas/Oil Ratio Reporting

The natural gas and vapor inside the flow stream could be obtained with an equation of state
developed specifically for SAGD operation for SUNCOR. IA summary of the results for the
vapor phase and SOR results (average of the 12hrs test) is shown below.

Table 6 – Total Gas, Natural Gas and GOR measurements

Total Gas Natural Gas

GOR
Well Flowrate Flowrate
[Sm3/Sm3]
[Sm3/h] [Sm3/h]
A 1496.0 39.4 7.3
B 3428.8 15.0 1.6
C 5118.2 26.6 2.5
D 1111.4 1.3 0.3
E 587.4 14.9 1.8
F 93.43 2.1 0.9
G 441.4 2.92 0.2
H 912.7 6.9 1.1
I 900.4 6.5 1.2

5.4 Sampling Verification

The table below presents a comparison of WLR measurements from Vx, Separator, Vx
Sample, and Manual Samples for each of the four sampled wells.

Table 7 – Comparison Vx measurement, sampling at the Vx, Separator Measurement

and sampling at the separator

Vx Vx Separator Manual
Well GVF
Sample Real-Time Real-time Sampling
A 85 69 93 56 96
B 73 77 80 72 98
C 77 80 79 72 80
D 74 79 80 80 81

It is important to mention that well B was a natural lifted well and this is the primary reason of
an average GVF around 98%. This well was also very challenging to test. The Vx real-time
measurement reported in this table was calculated over the same 15 minutes period of time
when the Vx and Separator samples were taken except for the cases above 96%.
For the wells with a GVF below 96%, the results show a good match of the Vx real time WLR
measurements with the reference samples.

18
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

5.5 Test Conditions

One of the important reliability considerations was to ensure that the Vx flow meter could
operate at the extreme cold conditions during the winter season without any performance
issues. During the second month of the Vx trial at Firebag ambient temperatures close to -
40C were happening. The figure 32 shows a thermometer recording an ambient temperature
(without wind chill factor) of -37.1 C at the well site.

Figure 17: Field conditions in Canada in December

The meter itself has been used most of the time with or without hot air supply and working in
the range of -15 to +15 DegC without showing any change in the performance. Data
Recording beyond -20 DegC (ambient temperature at the meter) has been also done in
exceptional cases and have demonstrated the robustness of the Vx technology in winter
conditions, the issue being if working to too low temperature being to plug the small liner very
quickly.

The Figure 18 shows additional views of the test conditions at Suncor Firebag. The truck was
equipped with a logging cabine and was self autonomous (electricity, warning system).

Figure 18: Field Operation Deployment

19
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

It will be recommended to leave the meter like in other cold country either in container (Alaska
or Kazakhstan Design). Due to the high temperature of the flow, the ambient temperature
inside the box will be kept at least in the range of -15 degC to 15 degC in the worst case and
then this will not need any additional heating as demonstrated in the qualification test.

4.6 Summary

As demonstrated during this trial, the Vx technology provided better accuracy compared to
the test separator data and allowed for more frequent well testing and then better well
performance monitoring. The flow dynamics could be better monitored and understood in real
time with fast acquisition measurements instead of the conventional test separator methods.
An overall summary of the performance of the Vx meter versus the separator and the short
duration test is shown below in the table below. This represents the average of the entire pad
(9 of the 11 wells):

Table 8 – Overall production of the 9 wells for separator and Vx with 12 and 6 hours
recording
Density CWE Density
Sep & Qmass Qliquid WLR Qmass Qliquid WLR CWE CWE
Liquid Sep Liquid
Vx Sep Sep sep* Vx Vx Vx Vx Vx
Sep Cor Vx
[kg/s] [m3/h] [%] [kg/s] [m3/h] [%] [m3/h] [m3/h]
[kg/m3] [m3/h] [kg/m3]
12Hr 88.2 353.6 896.4 12.8 88.8 89.6 355.8 905.7 79.5 10.7 88.2
6Hr 82.6 331.8 898.1 12.9 86.4 85.3 337.9 907.9 75.9 10.4 82.6

Table 9 – Overall comparison of the 9 wells for separator and Vx with 12 and 6 hours
recording
Δ
Δ Δ Δ Δ Steam
Vx - Density
Qmass Qliquid CWE WLR injected
Sep Liquid
[%] [%] [m3/h] [%] [kg/s]
[%]
12Hr 1.59 0.62 1.04 -0.04 2.0 196.3
6Hr 3.27 1.84 1.08 -0.05 -1.3 113.3

Table 10 - Overall production of the 9 wells for Vx with 12 and 4 hours recording
Density
Qmass Qliquid Density CWE WLR Qmass Qliquid CWE WLR
Vx12 Liquid
Vx12 Vx12 Liquid Vx Vx12 Vx12 Vx4 Vx4 Vx4 Vx4
Vx4 Vx4
[kg/s] [m3/h] [kg/m3] [m3/h] [%] [kg/s] [m3/h] [m3/h] [%]
[kg/m3]
Total 89.3 355.7 905.7 10.7 79.5 89.5 353.9 912.6 10.7 78.5

Table 11 - Overall comparison of the 9 wells for separator and Vx with 12 and 6 hours
recording
Δ Qmass Δ Qliquid Δ Density Liquid Δ CWE Δ WLR Qgas D Qgas GVF
Vx12 Vx4
[%] [%] [%] [m3/h] [%] [m3/h] [m3/h] [%]
-0.11 -0.51 0.76 0.04 -1 284.7 2.8 76.5

There are comparable results measured by Vx meter and by test separator equipment:

 Total liquid mass flow rates are better than 2%

 Liquid densities are within 1%
 WLR Vx versus WLR separator recalculated is within 2%

20
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

6 OPTIMIZATION

SAGD well production is composed of and controlled by many man-made variables that are
not found in conventional production operations. These range from initial steam injection – for
steam chamber creation and oil heating, to ramp up – with gas lift injection, steam lift, gas lift,
electro-submersible pumps (ESP) / progressive cavity pumps (PCP) lift, to blow down and
also to break through.

In artificial lift systems, a variety or mechanical and systemic components can limit
optimization of system usage. For example, artificial lift components may be blocked,
damaged, sized improperly, or operated at less-than-optimal rates.
Maintaining the ESPs within a good envelope of operation is critical to optimization. A real-
time optimization SAGD framework could be used to detect certain specific mechanical or
systemic problems at the pump, identify the cause of underperformance, and take preventive
measures that can increase the ESP run-life and decrease production downtimes.
ESPs are working in a harsh environment and it is advisable to reduce as much as possible
the mechanical stress and/or large flow rate variations. ESP performance is altered by
frequency changes due to changing production conditions; it is important to try to work in the
best operating range, which leads to have the proper sizing. Optimization is also linked to
efficiency, which is not only performance, but also efficient use of power.

With Vx technology, it is possible to perform real-time production optimization by adjusting the

frequency control of the pump based on the Vx response. The Vx technology can deliver real-
time production information since no stabilization of flow periods is required. The Vx is also
not dependant on flow regimes and its high-speed acquisition frequency provide results even
when the well production is unstable, giving the full picture of the flow dynamics when the
frequency of the ESP is changed for optimization purposes.

When available, real-time flow measurements obtained from the Vx can be applied against a
reservoir simulation/well model. Discrepancies between the measured data and the model
can be used to determine factors contributing to sub-optimal SAGD performance and make
decisions in a timely manner.

The protocol for Stage #3 was to test the West side of Pad. The pump running frequencies
was to adjust from the initial operating frequency in 2 Hz intervals or so every three hours in
order to capture a representative range of performance of the pump and well system.

The main objective of this stage of the test was to investigate the response of the Vx
measurements to changes in the pump frequencies steam injection rates and chokes. If real-
time measurements can be obtained, this opens the door for a future implementation of a
SAGD optimization method that can result in increased productivity, less energy consumption
and increased ESP run life.

Five wells were selected for the optimization trial, this stage of the trial consisted of a 12
hours baseline test at a fixed frequency and tubing pressure set point followed by another 12-
hour test with a frequency change every four hours. A minimum of three frequency changes
were made per well. Each frequency change was only made after the well had stabilized, and
only if there was room on the pump curve to make the next frequency change. We show you
a typical raw data curve.

21
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

Figure 19: Concept of optimization of each well (data recorded vs. Time)

7 SUMMARY

This paper presents the results of an extensive evaluation on the use of Multiphase
Flowmeters (MPFM) based on the Vx Technology to monitor production flow rates in SAGD
wells at the Suncor Firebag Project.

The collaboration between Schlumberger and SUNCOR to test and qualify the multiphase
technology started in 2007, when a single-well field trial was carried out at the Firebag project
facilities. This was the first introduction of the technology to SAGD environments. In 2009, a
second and more advanced field trial was initiated. A pad was selected due to the large
number of good producing wells and ideal set up of the conventional equipment.
9 wells from Pad 103 were tested during the months of October and December 2009,
ensuring that the MPFM was challenged against the full range of operating conditions of the
SAGD wells. During two months of extensive well testing and fluid sampling, no safety
incidents were encountered and the MPFM reported well production rates continuously and
without any performance degradation even during the extreme ambient temperatures
recorded at the well site.
Flow measurements between the MPFM and the Test Separator for the same stable flow
periods revealed consistent results between both equipments. Total liquid mass flow rates are
within +/- 2% and liquid densities within +/- 1% on average.
The comparison between the Vx and separator Water-Liquid Ratio (WLR) measurements
showed a discrepancy for all the wells comparing the raw data. The Vx reports consistently
lower WLR values than the test separator. Further examination revealed that the pro-ratio
factors were indicating a fair over-production of water and under-production of oil for the
entire field, i.e. the water cut at the separator was too high. In conclusion, there is a significant
error with the current water-cut meter at the opposite of the Vx Technology.

22
 28th International North Sea Flow Measurement Workshop
26th – 29th October 2010

In order to obtain a more accurate reference for comparison, the information about the pro-
ratio factor for the entire field was used to back-calculate what should have been the WLR
separator-based measurement, resulting in WLR Vx versus the recalculated WLR (from
separator) to be within 2%.If the Vx had been used in this field during the test period (using
the data from the Vx) one would have obtained a proration factor (for this pad) between 1.02
and 0.98 for oil and water over the two months (production sold versus what was reported by
Vx).
To demonstrate the reproducibility of the flow measurements using the Vx technology, a
second stage was completed, this time shortening the flow period from 12 hours to 6 hours.
Then, with the Vx meter only, the data of the 12 hours testing were reprocessed for 4 hours
showing a very good consistency in the results with the 12 hours test.The main conclusions of
the test are as follows:

8 CONCLUSIONS

The main conclusions of the test are first that all the initial objectives given for the testing
program were met or exceeded. Second, the Vx multiphase meter managed to demonstrate
all the operating requirements required by Suncor operations. The uncertainties of all the
phases are as per below:

 Uncertainty of within +/- 2% (after correction of the reference measurement

based on a global back allocation factor)
 Uncertainty of minimum +/- 2% on Liquid rate
 Uncertainty of +/- 2% on Total Gas Flowrate (due to wet gas conditions and
assuming that it is only steam)

Third, the Vx technology was capable to capture the entire dynamic of the flow: Start-up,
intermittent and pulsing flow. The meter was used above 180oC in this test. By the order of
accuracy form the output of the meter, the sequence is as follows: Total mass (or volumetric)
flow rate, liquid mass (or volumetric) flow rate, gas mass (or volumetric) flow rate, GVF, GLR,
then WLR, and associated oil and water flow rate.

The Vx meter reveals good dynamic response, repeatability and accuracies of the flow rates
and has shown promise to be able to be used as an optimization tool. This is likely the result
of the design and robustness of the Vx meter, summarized with some key features such as no
Tuning Factor, no Separation, no Flow Calibration, no issues with Foams or Emulsions, no
moving part, no sensors in direct contact with the fluid, and improve water detection and
accurate BSW monitoring.

4 REFERENCES

[1] Performance of Combination of a Venturi and Nuclear Fraction Meter in SAGD

Production Operations. Hompoth, Daniel, et al. Edmonton : s.n., 2008. World Heavy
Oil Congress. 2008-422

[2] Multiphase Metering: Importance of Fluid Property Knowledge in SAGD Operations.

Hompoth, Daniel, Pinguet, Dr. Bruno and Guerra, Elsie. Edmonton : s.n., 2008. World
Heavy Oil Congress. 2008-421

[3] SAGD Production Optimization: Combination of ESP and Multiphase Metering

Bruno G. Pinguet, Elsie Guerra and Colin Drever,Schlumberger. Edmonton : s.n.,
2008. World Heavy Oil Congress. 2008-424

[4] Conventional metering issues for thermal in situ production of extra heavy oils and
improvement of a commercial nucleonic multiphase meter for such applications. V.
Arendo F. Dang and B. Pinguet, PH. Pechard, JL. Roudil. Margarita Island : World
Heavy Oil Congress, 2009. 2009-327

23