IPTC 11573

An Innovative Multiphase Sampling Solution at the Well Site to Improve Multiphase

Flow Measurements and Phase Behavior Characterization

Downloaded from http://onepetro.org/IPTCONF/proceedings-pdf/07IPTC/All-07IPTC/IPTC-11573-MS/1802024/iptc-11573-ms.pdf/1 by Petrobras, Tamires Soares on 16 April 2025

F. Hollaender, SPE, J.J. Zhang, B. Pinguet, SPE, V. Bastos, SPE, E. Delvaux, SPE; Schlumberger Testing Services

Copyright 2007, International Petroleum Technology Conference

Overview
This paper was prepared for presentation at the International Petroleum Technology Multiphase flow meters (MFM) have been accepted now by
Conference held in Dubai, U.A.E., 4–6 December 2007.
the oil industry (Ref. [11], [12], [26], [28]). All MFM as other
This paper was selected for presentation by an IPTC Programme Committee following review
of information contained in an abstract submitted by the author(s). Contents of the paper, as
devices (Coriolis, Ultrasonic, Separators…) by construction
presented, have not been reviewed by the International Petroleum Technology Conference are measuring flow rates at line conditions and therefore the
and are subject to correction by the author(s). The material, as presented, does not necessarily
reflect any position of the International Petroleum Technology Conference, its officers, or use of PVT data to convert the flow rates at standard
members. Papers presented at IPTC are subject to publication review by Sponsor Society
Committees of IPTC. Electronic reproduction, distribution, or storage of any part of this paper
conditions is necessary (Ref. [3]). Currently, this is not
for commercial purposes without the written consent of the International Petroleum Technology addressed properly by most multiphase suppliers.
Conference 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 Schlumberger has been a leader to bring this into the picture
acknowledgment of where and by whom the paper was presented. Write Librarian, IPTC, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.
via the synergy of several divisions within Schlumberger
being each expert in knowledge of phase behavior and
Abstract sampling techniques. A generic software solution to address
Representative reservoir fluid sampling and characterization the needs of our customers has been proposed from Heavy Oil
has become increasingly important over the years. With to Condensate (Ref. [2], [3], [28]).
exploration, appraisal and development activities moving into The way to handle this need is very simple and until now
marginal fields and more challenging environments, accurate only two approaches were possible:
fluid characterization becomes more critical. This can be said
for the formation tester, DST and multiphase sampling and 1. using published correlations to estimate oil, water and
fluid characterization environments with the most challenging gas properties. This simple approach leads to
area in recent years arguably being the multiphase acceptable metrological results up to 1500 psia. The
environment. Multiphase flow meters have been accepted for upper limit can be extended to 5000 psia with great
several years now by the industry. Their use in permanent or care (Ref [2], [3], [6]). 1
well testing applications has been growing rapidly. In many 2. using experimental and modelled data from an
cases, multiphase flow meters have replaced the separator for Equation of State (EOS) software package using data
flow rate evaluation, but some fundamental needs from from a PVT report. This requires a competent phase
operators were not addressed properly, such as the ability to behaviour specialist to generate and control the
collect representative samples for phase-behavior generated fluid data (Ref [2], [3]).
characterization. Moreover, metering accuracies has been
questionable in many cases (at very high GVF or in wet gas A MFM could be located anywhere from the well head
conditions, high pressure or /and high temperature). flowing temperature and pressure up to separator conditions or
This paper focuses on the Multiphase Active Sampling lower. Figure 1 illustrates this.
Device Service (MASS), a fluid sampling and analysis service
that can be provided with the Vx* multiphase metering As said previously, MFM's request PVT information for
technology with the objective of collecting representative conversion, this is unavoidable, and has been similarly
samples, isolating and analyzing each fluid phase, and required to obtain flow rates at standard conditions with a
providing data from the analysis as input to the Vx* separator or any other devices. Figure 2 illustrates a generic
acquisition software data to obtain more accurate flow rates. flow path from line to standard conditions for any MFM and
The collection of phase representative samples also opens the any fluid types (i.e. from Heavy Oil to Gas Condensate).
opportunity for a full recombination PVT study to be On the left in Figure 2 are the data coming from flow
performed using the improved recombination ratio at line meters, be there MFM, separators or other devices. It is based
conditions from the multiphase flow meter. This dedicated on the physical need of some parameters to do the calculation
multiphase fluid sampling and analysis system, combined with 1
Vx technology provides a comprehensive and accurate set of At higher pressures and temperatures the fluid properties
fluid properties for improved flow-rate measurement accuracy correlations developed for separators are outside their range of
compared to conventional systems. application. In addition, the accuracy of the PVT inputs for
conversion from line to standard conditions and vice-versa is
more critical and can seriously affect the overall performance
of the meter if care is not taken in their measurement.
2 IPTC 11573

at standard conditions. These measured data have to be biggest effect on the global uncertainty of a meter such as in
converted to standard conditions taking into account the gas high pressure or wet gas conditions (Ref [4])3.
and liquid dissolved in the various phases as well as the
volumetric variations associated with changing pressure and The global uncertainties of a MFM encompass these two
temperature conditions. distinctive sources of errors related to fluid properties and
In summary, three sets of data are required to calculate deployed technology for each technology4.
flow rates at standard conditions: densities, volumetric
conversion factors from line conditions to standard conditions Fluid Property Measurement and Modelling
(bo, bw, bg) and Solution Ratio (Rst, Rwst, rgmp). These techniques

Downloaded from http://onepetro.org/IPTCONF/proceedings-pdf/07IPTC/All-07IPTC/IPTC-11573-MS/1802024/iptc-11573-ms.pdf/1 by Petrobras, Tamires Soares on 16 April 2025

values can be obtained from a classical PVT laboratory (Ref Several options are available to generate fluid property inputs
[2]). The liquid viscosity at line conditions is also a pertinent which can generally be described as:
parameter for Heavy Oil2. There are, therefore, a total of 10
parameters required to obtain a full PVT profile for any flow 1. Black Oil Correlations (BOC): This is the basic
metering device and in any type of fluid. approach using BOC to estimate oil, water and gas
properties from simple stock tank oil and gas
Intrinsic Error and Global Error measurements. However, in many cases the fluid
All MFM’s and any other metering device such as Coriolis, behaviour may differ significantly from these
separators… require some knowledge of PVT to allow flow correlations, then it is preferable to use data collected
rates calculation of the different phases at standard conditions. either experimentally or simulated using a PVT EOS
Meanwhile, each multiphase flow meter is also dependent on simulator.
several intrinsic features, measurements and models. Some of 2. Wellsite Fluid Property Measurement (WSPM): A
the most common one are listed here-after: dedicated wellsite measurement of the required fluid
properties, such as can be produced with the PVT
1. Hardware: Venturi throat diameter, Venturi inlet Express†, although requiring some expertise can
diameter, Distance between pressure ports, Distance deliver an accuracy equivalent to PVT laboratory
between elements of cross correlation, Pipe section, measurements or better in some applications. In
Acquisition frequency, Electronic stability, pipe addition to measuring the relevant fluid properties for
diameter or distance between sensors (Cross MFM input a full PVT report can be produced within
correlation)... a day from a representative recombined sample. The
2. Physical Measurement Type: Permittivity, need for wellsite fluid property measurement
Conductivity, Optical or Nuclear measurement for the becomes a particular issue at higher MFM operating
different fluids and mixture at different range of pressures and temperatures and with increasing
temperature and pressure… volatility of the produced fluids.
3. Measurement Accuracy and stability for the above 3. Equation of State (EOS): The required fluid property
measurements at high frequency to be able to capture inputs can be generated using an EOS tuned to data
the chaotic nature of flow. from a full PVT report in a software package such as
4. Validity of the Interpretation Engine: for different PVT-ProTM. The PVT report is often generated at
hypotheses regarding the flow rate (i.e. flow regime, the exploration and appraisal stage from well test or
mixing conditions, slugs or not…) or about viscosity, formation tester samples. Specialist PVT Expertise is
inversion phase, slippage and other physical or required to both conduct the EOS simulation and to
mathematical assumptions.... quality control the PVT data prior to its use.
Although sometimes a complex task, EOS tuning
Overall, there are more than 20 parameters, which could provides the essential information necessary to run
significantly affect any multiphase measurement. Some can be the meter in any type of fluid including volatile oils
adequately quality controlled during the construction, but and gas condensate. The turnaround time is short
others, such as the validity of models and some physical once the PVT data is validated.
measurements, are more complex to understand and may lead 4. Laboratory Measurement (LM): PVT Laboratory
to error. This is what we have called the intrinsic measurement measurements on samples obtained from the MFM
performance. That is only a first part of the overall will necessarily require a longer turnaround time than
performance of the MFM. all the other techniques but in most cases offer the
most accurate measurements. This may be a
Beyond the intrinsic error that can in itself be significant, requirement when using an MFM in a production
characterizing the produced fluid properties at standard
conditions is crucial to the successful deployment and 3
A review of the performance of different black oil
application of multiphase flow meters. The fluid properties correlations has already been completed (Reference [2], [3],
input accuracy is rarely challenged, even if this may be the [6], [13], [18]), and demonstrated that their application is quite
limited.
2 4
In this case, the dependency is primarily which is in this case There has been little published on this element of the
depending primarily on the WLR, temperature and dissolved propagation error attributed to the fluid properties input alone
gas. (Reference [1], [2], [3]).
IPTC 11573 3

allocation application where the cost and delay in reconditioning in laboratory was necessary. This was usually
producing the fluid property inputs is offset by the seen or noticed when a PVT report was given where a
need for superior accuracy. recalculated GOR is obtained based on the sample analysis
results. The advantage of the separator sample was to be in
Back to the fundamental of the life of the well, flow rate theory monophasic. The differences in pressure and
measurements are critical inputs at various stages from temperature are not too critical when dealing with black oil
Exploration and Appraisal to Production Monitoring via and a small variation is not in general inducing a large transfer
Development and Clean Up. This is used in the primary stage of elemental component such as C1, C2, CO2… from the gas
to define the future performance of the reservoir and give key to the oil phase or the opposite. The disadvantage of the

Downloaded from http://onepetro.org/IPTCONF/proceedings-pdf/07IPTC/All-07IPTC/IPTC-11573-MS/1802024/iptc-11573-ms.pdf/1 by Petrobras, Tamires Soares on 16 April 2025

information about the future potential of the reservoir. In the separator sample is also the fact that it is not possible to take
cleanup phase this is used to ensure that the well has been samples at high pressure and in the case of large flow rate of
conditioned properly for future production, then during the one phase against the other it can often induce have some
production monitoring period this is used for tax purposes but carry-over or carry-under if the separator design is not optimal
also for tuning the reservoir model with the latest information for the flowing conditions, leading to imperfect separation.
coming from downhole gauges combined with the surface
flow rates. Key indicators include GOR and water cut Schlumberger was one of the first to focus on the
tracking, as well as flow stability, especially in the case of development of a sampling solution for the multiphase
lifted wells. For all those applications, a well tuned MFM with environment. The commercial PhaseTester† tool coming out in
high frequency outputs can provide more accurate, 1999 had sampling capabilities from the start as shown on
representative data than devices lacking dynamic response Figure 4.
capabilities and using approximate fluid properties. This needs It can be noticed from this design (Figure 5) that the gas
of accurate flow rate and sampling information is summarized sampling was upstream of the Venturi and the liquid sampling
in the classical drawing shown in Figure 3. 3 axes are port was downstream therefore as in the separator both sample
represented, one showing the importance of the flow rate, the were not taken at the same conditions and could require some
other one that related to sampling and the final one significant reconditioning in laboratory before being able to do
considering fluid properties. It can be seen clearly with the a proper analysis. This was discussed in reference [28]and
different point presented on the figure. demonstrates that in the case of volatile oil a reconditioning
was compulsory.
Fluid properties and first sampling are a subject which
have not been discussed for years in details. It is commonly The Multiphase Active Sampling Service
taken today at a separator when this cannot be done downhole. Due to the wide range of operating conditions and flow
The trend to take it at the surface is essentially motivated by regimes that a multiphase flowmeter can face (Ref. [1], [4],
the risk associated with the possibility to leave the tool [5], [7], [8], [9], [10], [14], [16], [17], [19], [20], [21], [22],
offshore or about deffered production due to downhole [23], [24], [26]), there is no possibility of measuring
intervention. However, taking a sample at a separator assumes specifically phase densities in-line. The ultimate solution is
thermodynamic equilibrium, which means that the both phase therefore fluid sampling and analysis at the wellsite because it
oil and gas are at the same pressure and temperature and in will allow direct measurements of the required MFM input
equilibrium one with the other. Thermodynamic equilibrium parameters rather than relying on an estimated value from a
between the different phases for the given reservoir fluid are correlation or EOS model. It will also provide an opportunity
critical for any sampling operation. This is possible when the to take samples for a recombination PVT study using the
pressure and temperature remain constant for a period of time. improved MFM GOR at line conditions.
During this equilibrium there should be: Due to the complexity of multiphase flow regimes it will
1. No molecular composition change in any phase when never be possible to sample only one phase at a time but there
P and T remain constant is always a position in the flow line where one phase is
2. No mass transfer between phases when P and T predominantly present. The sampling technique, therefore,
remain constant needs to allow the selection of the optimum sampling position
This equilibrium is generally independent of the sample for the required phase and subsequently segregate and enrich
volume and phase fractions. There have been intense the sample with that phase at isobaric and isothermal
discussions about the dynamics of establishing thermodynamic conditions with reference to the meter flowing pressure and
equilibrium between phases in multiphase conditions. Some temperature. It should also allow all three phases to be
authors consider that equilibrium is reached after a few collected separately but at the same point in the meter so that
milliseconds while others consider that it is never fully the samples are all collected at the same conditions and can be
achieved. It is, however, generally accepted that the considered to be in equilibrium.
equilibrium is reached a few feet after a choke, change in pipe
size or pressure-loss generating flowline assembly causing The Multiphase Active Sampling Service (MASS) has
high frictional or turbulent losses. In a separator, the residence been developed first as a dedicated add-on to the Vx
time of the fluid in the vessel is generally considered long multiphase flowmeter providing a unique and integrated
enough (typically 1 minute) to approach the true equilibrium. multiphase solution to improve flow rate measurement. The
In fact, this is difficult to achieve and the pressure and complete installation of the MASS setup on a PhaseTester Vx
temperature of both phases are often different and a is shown on Figure 6.
4 IPTC 11573

The Multiphase Active Sampling Hardware consists of phase depending on its quality (gas, water, oil, liquid) and on
three main elements (Figure 7): the flow regimes (mist, bubble, slug, churn flows, etc.).
1. A multi-probe sampling device that retrofits to the
liquid sampling port on the Vx PhaseTester. The purpose of the MASS is to first collect representative
2. A wellsite fluid property measurement package phase concentrated fluids (oil or condensate, gas & water) at
which allows the direct measurement of the key fluid line conditions for onsite measurement of the Vx fluid
property inputs at line and standard conditions for property input parameters to improve the meter accuracy.
any type of multiphase flow meter. Secondly it is to collect the same representative phase
3. A dedicated data acquisition software to receive the concentrated fluids at line conditions for recombination and a

Downloaded from http://onepetro.org/IPTCONF/proceedings-pdf/07IPTC/All-07IPTC/IPTC-11573-MS/1802024/iptc-11573-ms.pdf/1 by Petrobras, Tamires Soares on 16 April 2025

directly measured fluid property inputs as an subsequent full PVT analysis at the PVT laboratory or at the
alternative to the standard correlation available with well site with the PVT Express service where a rapid
the Vx multiphase meters. turnaround is required.

The MASS addresses the sampling and thermodynamic An Optical Phase Detector (OPD) probe is used to sense
equilibrium issues with the following features: the type of fluid entering or leaving the sample chamber. It is a
• The multi-probe sampling device is inserted through key element of the sampling apparatus, as it allows monitoring
the liquid sampling trap and into the PhaseTester what is entering into the sample chamber from a given probe
stream. This has the advantages of good mixing due and what is expelled during the enriching process. When the
to the venturi that is located before the sampling required phase has been segregated the sample is transferred to
probes. Additionally the samples are taken at the a flash apparatus for the measurement of the fluid properties.
same point, which ensures that the pressure and The properties are measured at a single pressure and
temperature is constant and the samples refer to the temperature point and are then used to tune a mapping PVT
same thermodynamic equilibrium. model in the MFM acquisition software to compensate for
• There is no modification to do to get a retrofit for the variations in line pressure and temperature. The samples can
permanent part. also be used to for a PVT recombination study, taking
• The multi-probe sampling device has several probes advantage of the more accurate MFM GOR measurement that
in the flow stream arranged axially and facing both results from the improved fluid property characterization.
upstream and downstream. This allows the selection
of a sampling probe which will sample predominantly Flow Loop Validation
the required phase. The sampling principle was tested and validated in a three-
• The ability to enrich any desired phase during the phase reference flow-loop. A numerical simulation was
sampling process. If a specific phase exists in a very completed prior to the tests and the results show a good
small fraction inside the flow the Multiphase Active correspondence. Probe #1 that is positioned in the flow
Sampling Device can actively enrich this phase so a direction takes advantage of the difference in inertia between
large enough volume will be captured for all the gas and liquid and samples predominantly gas. Due to
required measurements. centrifugal forces acting in the elbow preceding the sampling
• The pressure and temperature is maintained by zone, liquid is flowing predominantly at the top of the pipe
heating and insulation during the sampling, and therefore probe #4 is the better choice for liquid sampling
segregation and enrichment process to ensure there is (i.e. very low GVF in the sample chamber). Other probes are
no mass transfer between the phases. there to allow for sampling in any unexpected conditions.
• The ability to verify that the right phase has been This is illustrated on Figure 10, showing the gas fraction
captured or transferred utilizing an optical phase captured in the sample chamber depending on the probe used
detection technique. and the flowing GVF.
Another interest was to verify a past study on the
The benefits of the Multiphase Active Sampling Service possibility that the WLR sample could be representative of
are: what is flowing in multiphase conditions5. In Figure 11 the
sampling technique is able to provide a WLR check at line
• Rapid turnaround of accurate Vx fluid property inputs
conditions. These tests were performed in the flow loop up to
at the wellsite rather than requiring sending samples
99.5% GVF being only limited by the accuracy of the
to a PVT Laboratory taking days or weeks.
reference flow loop at higher GVF. This shows an extremely
• Improved flow rate measurements computed with the
good consistency between WLR values measured using the
more accurate Vx fluid property inputs that result
optical probe from the MASS with input flow-loop WLR,
from direct measurement versus correlations or EOS
especially when using samples taken using the probe favoring
modeling.
liquid fractions, and this even at very low liquid fractions
• Provision of samples for a recombination PVT study
present in the flowline. In other word, the results indicate it is
utilizing the improved Vx GOR measurement.
possible to collect samples with a representative WLR when
only 0.5% of the fluid is flowing through the main pipe.
The multi-probe sampling device, in Figure 8 and Figure 9,
allows sampling at different positions and directions in the
flow line. This feature allows the capture of a predominant 5
.i.e. there is no significant slippage between oil and water
IPTC 11573 5

Field Trial Results Case B is a well with a wetter gas with a GVF close to
Following tests performed in flow loops and under well- 95%. Results are shown in Table 3 and Table 4. In this case it
controlled conditions, several multiphase active sampling kits can be seen that the oil flow rate at line conditions is
were sent to various worldwide locations in order to validate underestimated by the black oil correlation and there is a
the concept and data quality in a wide range of environments. reasonable match between the EOS and S&M. The gas flow
2 cases will be presented below. rates at line conditions remain consistent.

First series of field trials Second series of field trials

Having demonstrated that a representative liquid and gas An introduction test was done in by the end of 2006, this

Downloaded from http://onepetro.org/IPTCONF/proceedings-pdf/07IPTC/All-07IPTC/IPTC-11573-MS/1802024/iptc-11573-ms.pdf/1 by Petrobras, Tamires Soares on 16 April 2025

sample could be taken at the same pressure and temperature was a gas condensate well with an expected CGR in the range
from the same sampling point in a flow loop the next step was of 30 bbl/MMscf. This was a saturated reservoir (for the past
to field test the concept with live fluids at the wellsite on a Vx 20+ years). 3 chokes were tested in comparison with test
well test operation. Samples were taken on four gas separator. The flowing pressure was between 1600-1800psi
condensate wells and the Multiphase Active Sampling Service with a gas flowrate from 22 to 40 MMscf/d.
Measurements (S&M) were compared with the results from a This was an extensive campaign made of 24 hour flow
the Black Oil Correlation model (BOM), PVT Laboratory periods. 3 sets of samples of gas and condensate were taken
measurements with the PVT ExpressTM (PVT-XP) and a for each flow period from the separator, MASS, and the
tuned equation of state (PVT-Pro). The EOS was tuned only conventional Vx sampling port. With some spare samples
on the dead oil density and bubble point at line conditions taken to verify the effect of the temperature, these lead to more
including the composition. than 22 samples acquired in total. All those samples were
In Figure 12, comparisons were made of the calculation of analyzed in a PVT laboratory where the same analysis
liquid density at line conditions with the various methods. The procedures were used independently of the source of the
deviations are reported versus the measurement made with the samples.
MASS. Without going into too many details, the first comparison
A reasonable match was found between all techniques but of the various samples was performed via some basic
there was generally a superior match between the direct measurements on the samples: shrinkage, flash GOR and
measurements at the wellsite and the laboratory. pycnometer density on the condensate samples as well as
A similar comparison was made (Figure 13 and Figure 14) pycnometer density measurement on gas samples. The
for oil shrinkage (bo) and the GOR of the flashed oil (Rst). differences between the samples where generally lower than
Both comparisons demonstrate the clear benefit of 10% for the flash GOR, 2% for the pycnometer densities and
measurement over prediction. There is an excellent 4% for the shrinkage measurements. Those uncertainties can
consistency in both cases between wellsite and laboratory be considered as very reasonable knowing that the analyses
measurement for the four different fluids at different well were performed on small sample volumes to allow for repeat
conditions whereas the relative error for bo by prediction was analyses.
seen to be as high as 21.7% using a black oil correlation and
similarly a relative error of up to 38% for Rst was observed A preliminary study was the gas composition, this was
with the same technique. This is not surprising given the simplified for the sake of concisiveness (Table 5). It can be
known limits of the published correlations [2]. noticed that the overall composition is very consistent between
the separator and MASS samples and considering as well the
Having demonstrated the superiority of measurement micro gas chromatograph analysis integrated in the MASS kit
versus prediction of the fluid properties the next step was to made at the well site. The sample taken from the gas sampling
investigate the impact on the MFM flow rates at both line and port from the Vx is clearly different with a larger heavy end
standard conditions. The following tables show the effect of composition (i.e C7+) and also in C4-C6. This shows clearly
the fluid properties on calculated flow rates using real data that the gas taken is wetter and in other words contaminated
post-processed with the BOM, EOS model and S&M fluid with some condensate. This analysis yielded similar results for
property inputs. Again this information here below is not the different gas sample during different flow periods.
multiphase technology dependent, it is purely the error The same study was done with the liquid sample taken
introduced by doing the conversion from line to standard from the separator and from the MASS and the liquid
conditions. sampling port. The results are shown in Table 6.
Case A is a well with a very high GVF of over 99% and The comparison between separator and MASS samples
results are shown in Table 1 and Table 2. leads to the similar compositions with very little differences. It
In this case, it can be seen that the influence on the is clear that in the case of the Vx sampling port, the
calculated GVF and the gas flow rates at both line and composition is slightly lighter, hinting at the capture of free
standard conditions is minimal but there is some improvement gas along with the liquid. Not all comparisons have been
in the oil flow rate at line conditions with the wellsite shown here for the sake of concisiveness, but results are
measurement technique. At this very high GVF the well similar for all flow-periods in terms of trends, showing a very
stream is very close to being single phase which would explain good agreement between separator sample and MASS sample
the minimal errors associated with one method versus another. compositions, and a tendency for contamination of
conventional samples.
6 IPTC 11573

Reusing the different information here above it is possible

with the Vx technology to make a post-processing of the same Discussion
data with different fluid properties. Here below is presented Multiphase technology is perceived to be complex and this
the result of this work. may lead to difficulty in deploying multiphase flow meters
Before looking at the resulting flow-rates, the first check successfully in the field. We have observed that well selected
point was to consider the differences between the fluid field technicians and engineers trained with very specific
properties used depending on the calculation mode considered, objectives can develop a full understanding of the challenges
be it black oil correlations (BOC), EOS-derived properties, associated with multiphase metering and can successfully
measurements obtained from the MASS or laboratory operate multiphase flow meters. The focus on PVT training

Downloaded from http://onepetro.org/IPTCONF/proceedings-pdf/07IPTC/All-07IPTC/IPTC-11573-MS/1802024/iptc-11573-ms.pdf/1 by Petrobras, Tamires Soares on 16 April 2025

measurements. and the ability to collect representative samples is essential.
Table 7 shows a comparison of values of the oil shrinkage, This has led us to the conclusion that the Multiphase Active
oil density and gas density used by the different models and Sampling is best delivered as a service rather than a product
compared with laboratory measurements performed on ASD sold with the Vx multiphase flow meters.
samples for the second flow period. While the conformance
between on-site measurements (MASS) and laboratory Multiphase flow meters can be operated at pressures well
measurements is excellent for this case, results are typically above a separator. This environment usually results in better
within 2% for the other flow periods. measurements of rates at line conditions since the Gas Volume
Fraction is lower. In gas environments this effect is even more
The differences between the results obtained depending on acute with the higher level of liquid drop out observed for
the fluid propreties model used and with test separator results retrograde condensates. Under these circumstances, the usual
are presented in Table 8 and Table 9. This shows that, while PVT properties model developed for the separator can reach
the sensitivity related to the gas rate is limited, that associated its limits. It is unavoidable to have an understanding of the
to the liquid rate is quite significant and clearly highlights the PVT and the knowledge about the error propagation on the
importance of fluid properties over the final results. multiphase flow meter performance to ensure it is fit for
More striking are the differences between results obtained purpose.
from the test separator as compared to MFM results. The
liquid rates are significantly lower than those obtained from The dual-energy gamma ray and Venturi (i.e. Vx
the PhaseTester Vx, independently of the fluid properties technology) has been tested by many operating companies in
model used. This can be traced back to 3 main causes: their fields over the recent years. In addition to testing wells
• Separator results do not consider the amount of periodically, monitoring clean-up operations and evaluating
condensate dropping out of the gas phase after the reservoir performance in combination with production
separation pressure, that accounts for more than 10% logging, Vx Technology is also used for optimizing gas lift,
of the total production sampling fluids and measuring accurate flow rates, pressure
• The shrinkage value used for separator calculations and temperature. The main benefits of Vx Meters are: a stand-
were obtained from a shrinkage tester operating at a alone meter, portable and no moving parts. There is no need
lower temperature than the actual separator or for flow calibration by cross reference to another metering
flowing temperature, leading to an under-estimation system or any other calibration. The measurement is
of the shrinkage of 12 to 15% as compared to independent of the flow regime, e.g. slug flow, foaming. It has
laboratory analysis results, thus leading to an a wide operating envelope with the capability of sampling
additional under-estimation of the condensate rate from fluid sampling ports. In addition, it offers real-time data
• Some liquid carry-over was suspected during this management.
test, leading to a slight over-estimation of the gas rate
but a significant under-estimation of the liquid rate The meter acquires production rate, cumulative volume
Even though the differences appear significant, it is worth and operating condition data (including pressure and
pointing out that considering the total mass rate obtained from temperature) in real time without requiring separation of fluids
the various series of results, differences are inferior to 3% with and capturing the entire dynamics of the well response thanks
respect to the MASS results, showing consistency between the to a fast acquisition. There is continuous advancement in the
various metering technologies. The discrepancy in individual current Vx Technology leading to a better precision, and
phase rates between the separator and MFM can be assigned to extension of working conditions on the entire GVF range from
incorrect and incomplete fluid properties input as well as to 0 to 100% through two innovative models and in terms of
imperfect separation in this case where the density contrast temperature and pressure range (i.e. 302ºF/5,000 psia up to
between the gaseous and liquid phase is not too significant. 392ºF/10,000 psia)

The main benefits for operators are improved safety, data

These results consider the second flow period but quality, logistics, and longevity. It had been demonstrated that
differences are somewhat similar for the third flow period, the playback facility is a key element in this type of
with differences in liquid rate between the various post- conditions.
processed results being slightly higher, especially when
considering the results obtained from correlations. The robust physics behind the dual energy gamma ray
venturi multiphase flowmeter enables a strong and predictable
IPTC 11573 7

performance of the meter in specific operating conditions. It is understanding of the performance of the MFM for the life of
also possible to perform the error budget on the computed the well and the ability to manage the accuracy within a
rates and quantify accurately the sensitivities to the input narrow boundary.
parameters of the flow meter. The Multiphase Active
Sampling Service has been optimized for this specific type of The Multiphase Active Sampling Service is a unique
meter. solution to improve the overall accuracy of flow rate
measurement by reducing the uncertainty associated with the
Conclusions fluid properties input parameters both with variations in
The Multiphase Active Sampling service has been field tested pressure and temperature.

Downloaded from http://onepetro.org/IPTCONF/proceedings-pdf/07IPTC/All-07IPTC/IPTC-11573-MS/1802024/iptc-11573-ms.pdf/1 by Petrobras, Tamires Soares on 16 April 2025

to demonstrate the quality of the measurement versus
conventional PVT Laboratory measurements as well as the The Vx Technology enhance the PhaseTester and
impact on the MFM flow rates. The fluid properties PhaseWatcher to become the first meter capable to measure
measurements were typically within 1% accuracy between the accurately 3 phases (oil, water, gas) from 0 to (at least) 99.8%
well site and the equivalent laboratory measurement. The GVF and from 0 to 100% WLR. This is the first meter capable
improvement against the standard model in this type of to measure any phase without any online separation.
conditions was better than 10% to 20% in the measured oil, Moreover, the Vx Technology provides now an entire
water and gas phase flow rates for challenging metering multiphase solution, which addresses not only the metering
conditions. The direct measurements from the MASS but also the sampling and the knowledge of the fluid
hardware therefore provided a far superior Vx flow rate output properties at the well site.
at standard conditions when compared with both Correlation
and EOS modeling where a single-point calibration is References
acceptable. The multiphase sampling technique may find a [1] THEUVENY B.C., SEGERAL G. and PINGUET B.:
limit at ultra high GVF's but this has yet to be determined with “Multiphase Flow Meters in Well Testing Applications” paper
sampling having been successfully performed at GVF higher SPE 71475 at the 2001 Annual Technical Conference and
than 99.8%. Exhibition, New Orleans, Sep. 30th- Oct. 3rd.
[2] PINGUET B., GUIEZE P., DELVAUX E.: “Criticality of the
PVT Model in Multiphase Flow Meters to ensure accurate
Phase behaviour input accuracy requirements are rarely volumetric flow rate reporting” paper Multiphase Pumping &
challenged, even where this may have a major effect on the Technologies Conference in Abu Dhabi, February 22nd-25th
global uncertainty higher pressure, condensate... The industry 2004.
reports its production at standard conditions and therefore, the [3] PINGUET B.G, HADDAD N., BIRKETT P.G.:“Fluid
flow rates at line conditions need to be converted to standard Properties on the main path for MFM and WGM Accuracy: An
condition using PVT model in one way or another. The mass Analytical approach”, 4th South East Asia Conference, Kuala
and volume flow rate accuracy converted from line to standard Lumpur Conference, March 8th-11th 2005
condition combined with the intrinsic accuracy of the [4] HOPMAN N., PINGUET B.G, PEREZ S, GUIEZE P.,
VANDENBERG S, BOURGEOIS AM: “Field Experience in
multiphase flow meter technology provides the global Gas Well Testing: The benefit of the Vx Technology from 0 to
accuracy of the MFM. 100% GVF”, 5th South East Asia Conference, Kuala Lumpur
Conference, March 9th-12th 2006
The global uncertainties, therefore, hide two distinctive [5] PINGUET B.G, GUIEZE P, HOPMAN N, : “Field Experience
sources of errors, which have the potential to mislead the in gas well testing from 0 to 100% Gas Volume Fraction” at
community in determining the real performance of one meter Rio2006 Oil&Gas Conference in September, 11-14 2006, in Rio
against another. Flow loop tests are most of the time used by de Janeiro
oil companies to determine the performance of an MFM, [6] PINGUET B.G, DESTARAC P: “Importance of the fluid
however generalising the results of a comparative flow loop properties and predictable measurement in Multiphase flow
metering to ensure accurate reporting in high water cut
tests or a specific successful field experience to make conditions and/or high CO2 concentration” at 1st International
decisions about an application elsewhere where the fluid Jornadas sobre Medicin de Gas, Petrleo y Derivados South
properties are substantially different could lead to America Instituto Argentino de Petroleo y Gas
disappointment in the selected MFM’s performance. [7] ECONOMIDES M. J., HILL D.A. and EHLIG-ECONOMIDES
It is therefore valuable to be able to quantify the global C.: Petroleum Production Systems, Upper Saddle River, NJ
uncertainty of a meter as a function of the separate (1994) 523-550.
uncertainties in the prime measurement and from fluid [8] FRESHMAN R., LEKIC H.O.: “Artificial Lift for High-Volume
properties inputs for any given application. Production”, Oilfield Review (Spring 1999): 49-63.
[9] MUS E.A., TOSKEY E.D., BASCOUL S.J.F., BARBER E.C.:
“Development Well Testing Enhancement Using a Multiphase
Few suppliers, based on the available publications, are Flow Meter”, paper SPE 77769 presented at the 2002 Annual
capable today of quantifying and demonstrating the entire Technical Conference and Exhibition, San Antonio, Sep. 29th-
performance of their metrology and even fewer have an in- Oct. 2nd.
depth understanding of fluid properties. Through the unique [10] RETNANTO A., AZIM A.: “Monitoring Well Performance
specification of the Vx Technology it is possible to identify Using Multiphase Flow Meter” SPE 68718 presented at the 2001
the global error and the intrinsic performance of the Asia Pacific Oil and Gas Conference and Exhibition, Jakarta,
PhaseTester or PhaseWatcher. Having a measurement system Indonesia, Apr. 17-19.
simple that is predictable offers the advantage of a better
8 IPTC 11573

[11] FALCONE G., HEWITT G.F., ALIMONTI C. and Logging and Formation Evaluation, Natural Gas Engineering,
HARRISON, B.: “Multiphase Flow Metering: Current Trends and Petroleum Reservoir Engineering).
and Future Developments” paper SPE 71474 presented at the [28] JAYAWARDANE S. and THEUVENY B. C. “PVT Sampling
2001 Annual Technical Conference and Exhibition, New with Multiphase Flowmeters”, SPE 77405, San Antonio, Sep.
Orleans, Sep. 30th- Oct. 3rd. 30-Oct. 2, 2002, Texas.
[12] ATKINSON D.I., BERARD M., and SEGERAL G.:
“Qualification of a Nonintrusive Multiphase Flow Meter in
Viscous Flows” paper SPE 63118 presented at the 2000 Annual Tables
Technical Conference and Exhibition, Dallas, Oct. 1-4.
S&M BOC EOS
[13] ATKINSON D.I., PINGUET B., SEGERAL G. and
QvolOil@lc bpd 543 539 445

Downloaded from http://onepetro.org/IPTCONF/proceedings-pdf/07IPTC/All-07IPTC/IPTC-11573-MS/1802024/iptc-11573-ms.pdf/1 by Petrobras, Tamires Soares on 16 April 2025

THEUVENY B.C.: “Field Implications of Uncertainties in
Multiphase Flow Measurements” paper SPE 77403 presented at QvolOil@sc bpd 385 397 342
the 2002 Annual Technical Conference and Exhibition, San QvolWater@lc bpd 11 11 12.4
Antonio, Sep. 29th- Oct. 2nd. QvolWater@sc bpd 0 0 0
[14] BAKER C. B.: “Flow Measurement Handbook: Industrial QvolGas@lc MMcfd 0.302 0.310 0.311
Designs, operating principles, performance and applications”, QvolGas@sc MMcfd
Cambridge, United Kingdom (2000) 4-9.
21.3 20.5 21.3
[15] FREUND J. E.: “Modern Elementary Statistics”, Prentice-Hall, GVF % 99.0 99.0 99.2
Inc., Englewood Cliffs, NJ 1979. WLR % 3.8* 3.8 3.8
[16] COUPUT J.P., PROUVOST H., COQUIL M., LEPORCHER E. Table 1: Case A, comparison of reprocessed MFM flow-rates, GVF
and DYKESTEEN E.: "Implementation of multiphase metering and WLR
on unmanned wellhead platform", paper OTC13220 presented at
the 2001 Offshore technology Conference in Houston, April S&M-BOCEOS-BOC
30th-May 3rd, 2001. Oil Relative Error @lc % 0.8 -17.3
[17] RETNANTO A.: "Production Optimization using Multiphase Oil Absolute Error @lc bpd 4.5 -93.0
Well Testing: A Case Study from East Kalimantan, Indonesia", Oil Relative Error @sc % -4.3 -14.9
paper SPE 71556 presented at the 2001 Annual Technical
Conference and Exhibition, New Orleans, Sep. 30th- Oct. 3rd, Oil Absolute Error @sc bpd -17 -60
2001. Gas Relative Error @lc % -2.8 0.2
[18] THEUVENY B., SEGERAL G., MOSKNES P.O.: "Detection Gas Relative Error @sc % 3.6 3.8
and Identification of Scales Using Dual energy / Venturi Subsea Error WLR % 0 0
or Topside Multiphase Flow meters", paper OTC 13152 Error GVF % 0.04 0.17
presented at the 2001 Offshore technology Conference in Table 2: Case A, relative and absolute differences between flow-
Houston, April 30th-May 3rd, 2001. rate, WLR and GVF results obtained using different fluid
[19] KONTHA I.N.H., WEIMER B., RETNANTO A., AZIM A., properties modesl with reference to correlations
MARTINON D.: "Monitoring Well Performance using
Multiphase Flow meter", paper SPE 68718 presented at the SPE S&M BOM EOS
Asia Pacific Oil and Gas conference and Exhibition, Jakarta, 17- QvolOil@lc bpd 2134 1181 1736
19 April 2001.
[20] HADY A.A., "The Use of a Multiphase Flow Meter to optimize
QvolOil@sc bpd 1342 997 1326
Gas Lift Well Operations”. QvolWater@lc bpd 0 0 0
[21] MUS E. A., TOSKEY E. D., and BASCOUL S.J.F., "Added QvolWater@sc bpd 0 0 0
Value of a Multiphase Flow Meter in Exploration Well Testing”, QvolGas@lc MMcfd 0.240 0.246 0.243
paper OTC 13146 presented at the 2001 Offshore Technology QvolGas@sc MMcfd 18.9 20.7 19.6
Conference held in Houston, April 30th-May 3rd, 2001.
GVF % 95.3 97.4 96.2
[22] TURNA E.P., PINGUET B.G., KOSHY T., KHOORI A. and
BEKKOUSHA AEK.: "Multiphase Flow meters and production WLR % 0 0 0
Logs Diagnose Well Response in an Onshore ADCO Field, Abu Table 3: Case B, comparison of reprocessed MFM flow-rates, GVF
Dhabi” paper SPE 81534 presented at the SPE 13th Middle east and WLR
Oil Show & Conference held in Bahrain 5-8 April 2003.
[23] PINGUET B.G., BARRETO W. : “Multiphase Flow experience S&M-BOM PVT Pro-BOM
in Brazil: An artificial lift Focus”, South East Asia Conference, Oil Relative Error @lc % 81 47
Singapore, March 9th-11th. Oil Absolute Error @lc 952 555
[24] PINGUET B.G, TURNA E.P and BARRETO W. : “Field bpd
Experience in Gas Lift Conditions with a Venturi and Nuclear Oil Relative Error @sc % 35 33
fraction Meter Combination“, Multiphase Pumping and Oil Absolute Error @sc bpd 346 329
Technologies, Abu Dhabi Conference, February 22nd-25th
[25] MEHDIZAED P.: "Multiphase Meters" Hart's Petroleum Gas Relative Error @lc % -2.0 -1.2
Engineer International, May 98, p63-70. Gas Relative Error @sc % -8.6 -5.6
[26] THEUVENY B. G. and MEHDIZAED P.: ”Multiphase Error WLR % 0 0
Flowmeters for Well and Fiscal Allocations”, Paper presented at Error GVF % -2.1 -1.3
the SPE Western Regional / AAPG Pacific Section joint Table 4: Case B, relative and absolute differences between flow-
meeting, 20-22 May 2002, Anchorage, Alaska. rate, WLR and GVF results obtained using different fluid
[27] BRADLEY, Petroleum Engineering Handbook, SPE (SPE properties modesl with reference to correlations
Letter & Computer Symbols Standard for Economics, Well
IPTC 11573 9

Separator ASD Vx MGC

Non-HC 6.11 5.87 6.04 6.36
C1 76.46 76.46 75.19 76.12
C2-C3 14.84 14.89 15.28 14.86
C4-C6 2.52 2.66 3.24 2.61 Figures
C7+ 0.06 0.11 0.23 0.04
MW (g/mol) 21.529 21.593 22.135 21.567
Gross Ht Value
(MJ/m3) 42.87 43.20 44.09 -

Downloaded from http://onepetro.org/IPTCONF/proceedings-pdf/07IPTC/All-07IPTC/IPTC-11573-MS/1802024/iptc-11573-ms.pdf/1 by Petrobras, Tamires Soares on 16 April 2025

Table 5: Comparison of gas samples compositions from different
sources

Separator ASD Vx

Component Flashed Gas Flashed Oil Recombined Flashed Gas Flashed Oil Recombined Flashed Gas Flashed Oil Recombined
mole % mole % mole % mole % mole % mole % mole % mole % mole %
Non-HC 7.167 0.000 4.327 7.603 0.000 4.672 6.314 0.000 3.953
C1 51.187 0.000 30.901 51.443 0.000 31.610 56.145 0.000 35.153
C2-C4 38.392 8.002 26.348 37.637 7.299 25.941 34.383 7.295 24.255
C5-C8 3.252 46.914 20.556 3.316 45.465 19.566 3.156 48.881 20.253
C9-C20 0.001 44.609 17.680 0.001 46.850 18.063 0.001 43.520 16.273
C21+ 0.000 0.475 0.188 0.000 0.386 0.149 0.000 0.304 0.114
MW 28.96 117.67 64.12 28.86 119.06 63.64 27.72 116.55 60.93 Figure 1: Schematic of a typical multiphase meter set-up
Table 6: Comparison of liquid samples compositions from
different sources

bo rho_o rho_g
[-] [kg/m3] [kg/m3]
BOC 0.525 558.5 118.7
EOS 0.504 558.2 118.9
MASS 0.559 571.5 121.5
Lab 0.561 572.7 121.0
Relative error
[%] [%] [%]
BOC-Lab -6.4% -2.5% -1.9%
EOS-Lab -10.2% -2.5% -1.7%
6
MASS -0.3% -0.2% 0.4% Figure 2: Flow path from line to standard conditions
Table 7: Comparison of fluid properties used by different
interpretation models against laboratory measurements Sampling
5
BOC EOS MASS Separator
4
QGas @LC MMcf/d 0.245 0.245 0.247 -
3
Well Types
QGas @SC MMscf/d 31.31 31.69 31.41 31.80
2 Exploration & Appraisal
QLiquid @LC bbl/d 1750 1734 1438 -
QLiquid @SC stb/d 934 962 885 430 1 Development
WLR % 1.6 1.5 1.5 - 0 Production
Water Cut % 2.8 2.5 2.2 2.4
GVF % 94.2 94.2 94.9 -
CGR stb/MMscf 29.0 29.6 27.6 13.2 Fluid processing
Table 8: Comparison of reprocessed MFM flow-rates and ratios Metering and disposal
along with test separator results Figure 3: Importance of sampling, metering and fluid processing

BOC-MASS EOS-MASS Sep-Mass

QGas @LC % Relative -0.7% -0.7% - 6
qgmp is the gas volumetric flow rate at MFM conditions.
QGas @SC % Relative -0.3% 0.9% 1.2%
QLiquid @LC % Relative 21.7% 20.6% -
This gas splits into two phases at standard conditions which
QLiquid @SC % Relative 5.5% 8.6% -51.4%
are qggsc (gas flow rate) and qgosc (oil flow rate due to the
WLR % Absolute 0.1% 0.0% - liquid dropping out of the gas phase at lower pressure and
Water Cut % Absolute 0.5% 0.3% 0.2% temperature). qomp, is the oil at MFM conditions. This oil
GVF % Absolute -0.7% -0.7% - splits into two phases at standard conditions which are qoosc
CGR % Relative 5.3% 7.4% -52.1% (oil flow rate) and qogsc (gas flow rate), the gas evolved from
Table 9: Relative differences between flow rates and ratios the oil at standard conditions. qwmp will follow a similar path
obtained using different fluid properties model and separator to the oil splitting into two phases, qwwsc and qwgsc. qsmp
results with reference to MASS inputs refers to a possible solid phase. The sum of the different
outputs gives the total volumetric flow rates of gas, oil and
water at standard conditions.
10 IPTC 11573

Downloaded from http://onepetro.org/IPTCONF/proceedings-pdf/07IPTC/All-07IPTC/IPTC-11573-MS/1802024/iptc-11573-ms.pdf/1 by Petrobras, Tamires Soares on 16 April 2025

Figure 4: PhaseTester Vx physical setup
Figure 7: Overall view of the multiphase active sampling service

Figure 8: MASS connected to the PhaseTester Liquid port

Figure 5: Liquid and gas sample ports of the PhaseTester Vx (left

and right respectively)

Figure 9: Multi-probe sampling device

Figure 6: PhaseTester Vx Fitted with the MASS

Figure 10: Sample chamber gas fraction versus flow-loop

GVF
IPTC 11573 11

Rst S&M
BOM
PVT Xp

263 - 3.0 %
300
PVT Pro

271

240 - 3.6 %

- 10.4 %
249
250

- 27.7 %

223
- 30.5 %
168 - 38.0 %

196
200

173
Rst (m3/m3)

+ 17..4 %
+ 22.4 %
150

96 + 6.6 %

+ 1.1 %
- 3.3 %
- 3.6 %

Downloaded from http://onepetro.org/IPTCONF/proceedings-pdf/07IPTC/All-07IPTC/IPTC-11573-MS/1802024/iptc-11573-ms.pdf/1 by Petrobras, Tamires Soares on 16 April 2025

110

108

93
100

92
90

89
86
50

0
IA111 IA113 OT143 OT142
Well

Figure 14: Comparison of the GOR of the flashed oil (Rst) in 4

Figure 11: Sampled WLR versus reference flow-loop wells with the different techniques

Oil phase density at LC

800
S&M
670 - 0.6 %
660 + 0.9 %

658 + 0.7 %

655 + 0.2 %

659 - 2.4 %

BOM *
Mark of Schlumberger and Framo Engineering
646 - 4.2 %

PVT Xp
592 + 12.8 %
675

700 PVT Pro

†
579 + 10.3 %

585 + 3.4 %
653

Mark of Schlumberger
571 + 0.9 %
559 + 6.5 %

545 - 3.7 %

600
566
525

500
density (kg/m3)

400

300

200

100

0
IA111 IA113 OT143 OT142
Well

Figure 12: Comparison of the liquid density for the different

techniques on four wells

Oil Phase Shrinkage (bo)

+ 2.8 %

+ 0.5 %
+ 0.5 %

0.9
- 2.1 %
- 7.4 %
- 5.0 %

S&M
BOM
0.770
0.767

0.766
0.750

0.750

0.8 PVT Xp
0.746

PVT Pro
+ 21.7 %
0.709

0.709

+ 13.7 %
+ 18.6 %

0.7
+ 5.9 %
+ 2.0 %
+ 2.2 %
0.595

0.580

0.580

0.6
0.540
0.520
0.510
0.500
0.489

0.5
bo (m3/m3)

0.4

0.3

0.2

0.1

0
IA111 IA113 OT143 OT142
Well

Figure 13: Comparison of the oil shrinkage (bo) in 4 wells with the
different techniques