Вы находитесь на странице: 1из 16

SPE 115478

Field Validation Processes for Multiphase Wet Gas Surface Well Testing
Solutions: Example from the Yamburgskoe Arctic Gas Condensate Field,
Russia
B. Theuveny, Y.Shumakov, A Zhandin, Schlumberger and I.Zinchenko, Gazprom

Copyright 2008, Society of Petroleum Engineers

This paper was prepared for presentation at the 2008 SPE Russian Oil & Gas Technical Conference and Exhibition held in Moscow, Russia, 28–30 October 2008.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
Surface welltesting of Gas-Condensate with multiphase flowmeters is still considered a challenge for production metering.
Traditional means of well testing have been deployed for years and used consistently for reservoir and production
management. However, it can be difficult to compare data sets obtained with different measurement devices.

Multiphase flow meters have been proved for multiphase production metering by many operation companies worldwide.
However, in artic environmental conditions like those of Yamburgskoe gas-condensate field, with low ambient temperature
and production rate regulation restrictions, this process had to be revalidated and the operational capabilities confirmed with
all of the logistical challenge of this environment. A number of recommendations to prevent and mitigate the impact of the
hydrate and document major benefits of multiphase well testing are based on the accumulated operational experience from the
operations of various multiphase flowmeters in the area.

Most of the gas condensate wells in the Yamburgskoe gas-condensate field are flowing with liquid unloading issues in a slug
flow regime that can be observed and monitored accurately with multiphase flow meters. The importance of slug flow regime
identification relates to the selection of stable flow intervals for surface sampling and further recombination factor
determination. This information is also extremely important for well completion design. There is good agreement between the
slug frequencies obtained with nodal simulations and the actual dynamic measurements from the meter – an impossible task
for conventional measurements based on separator or batch processing of the liquid.

Surface well testing to control changes of reservoir parameters, PVT composition, and production back-allocation performance
is one of the key parts of gas-condensate field development in Russia. A number of multiphase welltesting operations have
been successfully performed recently that show acceptable performance capability and the benefits of this methodology
compared to the parallel production measurement with a traditional gas separator. This novel comparison and qualification
process of multiphase flow meters also provides a better understanding of the aptitude of conventional means to collect rate in
the case of wet gas wells in Northern Siberia.

Introduction
Many of the benefits of using in-line Dual Energy Gamma Venturi multiphase flow meters for accurate rate measurements on
gas wells have been published ([1], [2], [3]). Their utilization in oil wells has been known for a number of years, but their
deployments in gas and condensate wells is rather recent (2005) and has significantly changed the existing approach to
multiphase well testing and production back allocation in Northern Siberia of Russia.
The biggest part of gas reserves in Russia are concentrated in fields with arctic environmental conditions (Northern Siberia,
Arctic Shelf, Okhotskoe Sea). Yamburgskoe gas, gas-condensate field is one of these fields, which have been selected for the
field evaluation campaign of multiphase flow meter production capabilities (see Fig. 1). The Yamburgskoe field is one of the
mature fields in that area, with different geological objects of development and an extended range of well operating conditions.
To perform a comprehensive field evaluation of multiphase technology, three typical wells, representing different groups of
production conditions were selected. These three wells were selected describing the range of expectable variations of fluid and
2 SPE 115478

production environments such as well head operation pressure, GOR, condensate and water contained in production fluid.
Despite the fact that traditionally a wide range of operation conditions typically are extremely difficult to measure using one
type and size of production flow meter, the performance obtained with the multiphase flowmeter is deemed adequate to
provide valuable information.
All the wells in Yamburgskoe field are producing commingled
from different formations, all with reservoir pressure below
dew point and different PVT properties. These facts did not
allow existing and normalized PVT properties to acquire during
the tests performed earlier for conversion of rate measurements
from line to standard conditions.

Validation methodology
These three representative wells were in production prior to the
tests, and no preliminary shut-in period for a significant amount
of time was planned, prior to the tests, in order to reduce the
impact on the production flow and fluids regimes to be
analyzed. The sequence of events during these test was quite
similar to traditional so-called “hydrodynamic test” (i.e.
deliverability test) performed by the operator company on all
the production well stock. During tests, the well flowed under
the wellhead pressure that is conventional for in-line production
regime, then – through various choke diameters under and
below the pressure on base production regime. The average
production time of the wells on each regime did not exceed 24
hours according to the test program. The objectives of the tests
were: production back calculation, determine apparent
reservoir pressure of multiple formations and obtain PVT
property of reservoir fluid.

Additionally to increase the evaluation range of multiphase


flow meter operation conditions, two choke manifolds in
surface layouts were used. The changes of the well production Figure 1. Location of Yamburgskoe field
regime during the test was performed using one of the choke
manifolds putting multiphase meter to operate under upstream or downstream pressure on each regime of the well.

Any multiphase flowmeters require PVT information and that aspect is no different from any other types of flow measurement
devices such as Coriolis, Ultrasonic, and Microwave multiphase or single-phase measurements from a separator. All of these
devices are measuring some type of flow rate at line (operating) conditions and therefore must use some PVT information to
convert the measured flow rates to standard conditions. A large part of the total uncertainty on the measurement of the rates
obtained from the Dual-Energy Gamma Venturi Multiphase Flowmeter (MPFM) comes from underlying uncertainties in the
PVT data of produced fluid. It is therefore important to operate the MPFM using most accurate PVT properties of the fluid
best determined from samples obtained in the flow at the well site in each of the wells. This is now possible using a set of
recently developed techniques. During the test of the well, representative surface samples were collected, using a newly
developed multiphase sampling device add-on to the now-traditional multiphase flow meter. One of the main benefits of this
approach is to rapidly provide an accurate and tailored data set of fluid property parameters at the well site, using an in-situ
measurement rather than sending samples to a PVT Laboratory taking days or weeks, a long turnaround time mostly caused by
the challenge of the logistics in the Northern Siberia operating area.

The full metrological performance of a multiphase flowmeter is the combination of three types of factors:
1. The inherent uncertainty linked to the measurement methodology
2. The uncertainty propagated from the individual measurement uncertainty
3. The uncertainty linked to the lack of knowledge or poor description of the well fluids that are metered.

These factors must be evaluated in order to obtain a reliable description of the overall flow rate uncertainty. A number of
challenges are associated with this approach. The methodology used to optimize the overall metrology performance of the
multiphase flow meter specifically in this type of fluids and logistical environment are demonstrated in Afanasyev et al [4].
SPE 115478 3

It is an established fact that the actual metrological performance of the underlying technology of any flow meter can only be
demonstrated during flow loop tests. Many tests of Venturi-dual energy flow meter, on different flow loop facilities,
successfully established an uncertainty model within last 10 years, proving the metrological performance of the meter over a
large range of operations. In 2005, the model of computation and methodology of measurement of the Dual Energy-Gamma
Venturi multiphase flowmeter was enhanced to include the ability to perform a reasonable measurement of liquid and accurate
measurement of gas rates in extremely high GVF ranges (95-99%). Such conditions are of interest in gas wells, gas lifted
wells. The necessity to always revert to the metrological performance of a well test device and the propagation of errors is
essential during any actual field operations condition, as they must be performed in absence of any reference flow rate
measurements, which are also usually very difficult. This stand-alone methodology is very important to ensure that the quality
of the diagnostics that can be deduced from the acquired well test information is consistent with some quantifiable
metrological data and thus enables a proper prioritization of the consequent recommendations.

Initial trial tests against conventional measurements were used to get a better feeling for the actual expectable metrological
performance and accuracy of the multiphase flowmeter measurement. This was achieved using a conventional centrifugal gas
separator. The purpose of this process was to confirm the uncertainty estimations based on performed flow loop tests that may
have been changed dramatically in actual field operation conditions due to condensate drop out and the lean mixture of fluids
that are normally encountered during the well testing operations (97-99.5% GVF). It was soon realised that the multiphase
flowmeter would document and quantify the relative instability levels of the natural flow behavior of the wells.

These types of comparison tests are always difficult to conduct and a number of best practices have been developed and are
presented in other papers. However, the paper will focus on some of the specific comparison challenges that have been
observed in the case of Yamburgazdobicha operations. One of the main concerns is always the quantification of the actual
separation efficiency, the underlying hypothesis that is required to ensure accurate single-phase measurements. In order to
quantify the amount of condensate carry-over to the gas line of the test separator additional independent flow rate
measurement at the outlet of the separator were deployed in one of the wells.

The gas phase was measured in traditional orifice based methodologies on the gas outlet line. The liquid phase flow rate
separator measurements were derived from the time required to fill up the separator tank between two valves located at
different levels and direct measurement of associated gas using orifice plate at tip of the flare. The operation time of such type
of separator is fully dependant from the time required to fill up the volume of the tank. This batch process was repeated a
number of times on each of the flow choke settings. The average representative flowrates from the separator is then reported.

It was not part of the exercise to determine the actual metrological performance of the separator based measurement process.
Some insights however about the efficiency of the separations during the batch processes have been established and are
presented later in the paper.

The direct verification methodology involved a comparison of the following measurements:


• Total mass rate
• Liquid mass rate of condensate (stable condensate, in the Russian nomenclature)
• Comparison of the fluids composition
• Gas volumetric rate at standard conditions
• Water content
• Condensate to gas ratio (mass per volume)

It is important to note the importance of the fluid properties and to develop a proper understanding of difference of reporting
between the Russian standards and the western practices. Of particular attention, one needs to differentiate properly the stable
liquid in standard conditions from the C5+ stable fraction in that liquid, which is one of the standards, used in Russia Gazprom
operations for example. The challenge of establishing and agreeing upon the proper nomenclature before is important. That
point is one of the main lessons learnt on this process.

General well test operating conditions


The Yamburgazdobbycha fields presently produce significantly below their original dew point and produce a leaner effluent
than initially, due to a large condensate bank extending throughout the reservoir. In this study, the average condensate to gas
ratio was 79.5 g/Sm3, as traditionally reported by the Russian well test standards. This compares to 15.4 STB/MMSCF in oil
field units and 97 to 99.5% Gas Volume Fraction. The operating temperature varied from 0 deg C to 23 deg C and the
measurement pressure from 40 to 110 bars. Water liquid ratio ranged from 0% to 40 % (water in the liquid phase at
measurement conditions – not including the water vapor in the gas phase).
4 SPE 115478

From a multiphase measurement perspective, the main challenge is the high Gas Volume Fraction (GVF) that is experienced.

From a traditional well test perspective, the issue lies mostly in the separation efficiency and the characterization of the
amount of carry over, if any, in the gas outlet of the separation device.

The measurement methodology used traditionally by Yamburgazdobycha consists in using a vertical centrifugal separator. The
flow metering provided consists of a differential pressure measurement on the gas line and volumetric batch measurement on
the condensate.

Actual field data and their validation challenges


Early work about the deployment of Dual-Gamma Venturi multiphase flowmeters in the Yaburggazdobycha operating area has
been presented by Theuveny, Zinchenko and Shumakov ([5], [6]). During the production metering campaign, all the wells
demonstrated some level of unstable production behaviour, which could be retrospectively inferred from small fluctuations of
pressure and temperature. These small fluctuations related to surprisingly large variations of flows at surface as reported by the
multiphase flowmeter. A number of conjectures as to these instabilities were developed. The observed temperature behaviour
of the wells can be divided into two types: slow trends, related with contrast of ambient temperature during the night and
daytime and fast changes related with changes of liquid contains in the flowing fluid. The first effect is also compounded by
the slowness of the transfer of heat between the upper part of the completion in contact with cold formation or permafrost and
the fluid in the tubing. These temperature trends were a consequence of the short shut-in period necessary to perform some
pressure gradient operations with slick line prior to the beginning of the test or in between chokes. Changes of the liquid
content due to slugging in the completion affect significantly the Joule-Thompson effect at the choke and consequently
downstream temperature behaviour. It shall also be noted that the present rates observed in the range of 100,000 Sm3/day to
150,000 Sm3/day are now large in consideration to the tubing capacity optimized for the initial well capacity about 10 years or
more ago.

During production, metering of tested wells hydrate formation issue on surface facilities was expected. Hydrate formation
issue is a well-known problem of Yamburgskoe area. Regardless of continued methanol injection during the tests, except a
period of the sampling operations to mitigate the expected hydrate formation, coating and plugging problems were one of the
biggest challenges during the metering campaign. Due to low potential production of the wells in most of the cases, there was
no chance to move well operation point in to hydrate-free area and the crew had to be contended with this problem.

The hydrate formation issues not only related with safety aspects of the operations such as plugging in the surface facilities,
but also expectedly affects the performance of the MPFM. The hydrate formation is making significant impact on production
measurement using Dual-Energy Venturi multiphase flow meter and already been presented in Theuveny at al [2]. The three-
phase flow meter interprets by default the presence of hydrates as a large amount of moving fresh water. This fact was used for
early hydrate warning during the metering campaign. The early hydrate detection in combination with hydrate curve formation
is a powerful tool to control the adequacy of injecting hydrate inhibitor. However, the hydrate formation was observed only by
the MPFM during operations downstream of the choke – a fact confirmed with the utilization of the conventional separator
suffering also from plugging and hydrating issue. As a consequence the MPFM was operated upstream of the choke when
possible to avoid the problem and we recommend to use this configuration in the future.
SPE 115478 5

Figure 2. Hydrate formation curve with actual operation point of Figure 3. Hydrate formation observed by
MPFM for one of the tested well MPFM
Figure 2 shows the operation region of the multiphase flowmeter as a function of pressure and temperature. It is easy to see
that the regimes to the left of the hydrate formation curve will be strongly affected by hydrate formation. Figure 3 shows the
operational triangle effects on the nuclear measurements of the multiphase flowmeter. The 45-Hz spectral analysis of the dual-
energy gamma provides clearly a visual warning of the presence of hydrates. The operating point is pulled rapidly (in a matter
of seconds) towards the fresh water point (at the bottom left) in this plot of attenuations at high energy versus low energy
attenuation of the mixture.

The separation device traditionally used also is affected by the presence of hydrates. The issue is mitigated in two ways:
injection of methanol during the non sampling period, utilization of batch heating of the body of the separator during the
sampling period to ensure that clean samples would be collected – as will be discussed later, the change of temperature is not
so challenging from a sample representativity perspective as the “stable” C5+ fraction is mostly considered.

Slug Flow Regimes


An important part of the validation methodology relies on the understanding of the difference of measurement content
especially in the liquid part. The batch process used offers some benefits in terms of integration of a quantities of liquid, the
sampling simplicity and ensures that even tiny amount of liquid could be measured, but on the other hand does not provide an
easy understanding of the stability of the condensate / water flow – an important factor for the full analysis of the well
performance and of the representativity of the samples collected.

Most of the gas-condensate wells of Yamburgskoe field, as mentioned earlier, are producing in a slug flow regime. The Dual
energy gamma – Venturi multiphase flowmeter clearly captures the dynamic nature of the flow because of the high speed
measurement phase resolution of the undisturbed multiphase flow. Well production modeling was run to support our
comments regarding observed well flow regime within selected certain period of the well production when flow rates were
measured using separator and multiphase flow meter. The simulation results of flow map on wellhead shows that the well
operates under slug flow regime.

Original completion design with larger rates guaranteed a rather stable flow initially, but with depletion and massive
condensate banking, the well productivity has somewhat been reduced and therefore gas vertical velocities tends to be too low
to enable stable transport of the condensate – a very challenging feat in any case.
6 SPE 115478

Slug Values
PI-SS (Severe Slugging Group) 37.74
Mean Slug Volume (m3) 0.07
Mean Slug Length (m) 14.67
Mean Slug Frequency (1/min) 0.23
1 in 10 Slug Volume (m3) 0.13
1 in 10 Slug Length (m) 27.82
1 in 10 Slug Frequency (1/min) 0.02
1 in 100 Slug Volume (m3) 0.21
1 in 100 Slug Length (m) 47.03
1 in 100 Slug Frequency (1/min) 0.002
1 in 1000 Slug Volume (m3) 0.31
1 in 1000 Slug Length (m) 68.42
1 in 1000 Slug Frequency (1/min) 2.3E-04

Figure 4. Flow regime map over selected interval of well production and statistical distribution of the slugs to be expected

The calculation of slug volumes and slug frequencies using a probabilistic approach through vertical flow performance
simulations, take into account the actual well string configuration demonstrated variation of slug frequency with mean value
equal to 0.23 1/min. These calculation results are in good agreement with actual, observed flow behaviour by the MPFM
demonstrated in the (Figure 5). The accurate determination of slug value parameters is extremely important for accurate design
of slug catcher in surface facilities infrastructure. Use of conventional flow meters with averaging data in time by the principal
of measurements makes it smoother and doesn’t give information about actual behaviour of the wells.

Figure 5. Liquid Rate fluctuation on example of one of the well production regime

It is notable to see how little the gas flow rate seems undisturbed by the presence of liquid slugs – this fact is probably the
result of the relatively small mass contribution of the liquid to the total flow and the very small amount of relative volume
occupied at line conditions by the liquid.
SPE 115478 7

MPFM Metrological performances – Consistency of MPFM measurements


between various operating conditions
The wells selected for the comparison analysis of the production measurements would not be in any manner and place be
considered as the easiest to test – in particular considering the combination of environmental impacts, logistical and
operational issues and finally the proper well behaviours. However, this approach was selected to ensure that a wide range of
operating conditions would be met in that analysis/validation and would be representative of wells even for future larger
deployments of the multiphase well testing in the fields. The production of the wells during the tests was not fully stabilized,
which complicates the comparisons of flow rates more complicated. Flow rate measurements were performed upstream and
downstream of the choke with the MPFM, the test separator was down stream of the choke and in one case a second MPFM
was deployed on the gas outlet of the separator. Theses instabilities were well identified by the MPFM, with the excellent
ability of multiphase flow meter to measure all the transient flow behaviour. This was used for a selection of stable production
periods, avoiding the effect of hydrate formation and filtering out high fluctuation of temperature and pressure due to slugging
of the wells.

Flow loop tests performed under similar conditions as operation of MPFM in Yamburgskoe field leads to an expected
uncertainty of 2% for the gas rate assuming there are no associated uncertainties resulting from the determination of fluid
property. The expected uncertainty of liquid rate measurement is around 15% at such high GVF.

However, the actual well production behaviour is much more complicated, and contains features that are difficult to simulate
in flow loops tests – thus to the importance of the field validation of the MPFM. The discrepancy of liquid and gas rate
measurements at 15% (Figure 6) cannot be simply considered as the actual metrological performances of the flow meter in
such conditions – the balance of uncertainty of the reference, the slow stabilization effects, the lack of simultaneity between
the some of the measurements – all build up further in the error budget of the comparison.

Table 1. Comparison of production metering results using MPFM upstream and downstream of the choke
Upstream rate measurements Downstream rate measurements
Choke Operation Operation
Well Condensate, Water, Gas, Condensate, Water, Gas,
size, mm pressure, pressure,
sm3/day sm3/day Msm3/day sm3/day sm3/day Msm3/day
bar bar
1  9.53  97.8  10.7  1  133.1  32.1  12.3  1.1  144.9 
1  11.11  93.3  16.7  1.9  192.7  40.2  16.2  1.8  176.1 
   1  12.7  90.9  20.4  2.8  223.1  52.3  20.7  2.9  229.4 
   2  9.53  86.2  6.9  7.8  113.2  29.9  14.7  17.2  87.6 
   2  11.11  79  15  16.2  121.1  38.2  10.3  14.2  139 
   2  12.7  77.4  9.8  15.2  145.3  50.8  15.8  18.3  155.7 
8 SPE 115478

Figure 6. Cross plot comparison of production metering results upstream and downstream of the choke with the multiphase
flowmeter.

Despite all of these factors, the dual-energy gamma Venturi multiphase flow meter provides a consistent agreement between
actual production measurements upstream and downstream the choke in various operational conditions.

Separator production rate measurements


Along with continuous production rate measurements using the MPFM, during a number of selected production intervals
avoiding hydrates, separator production measurements were performed. As was mentioned earlier the duration of these
production measurement is limited by volume of the tank (liquid volume is limiter to 160 litres – about one barrel). These
short production rate measurements can be affected due to sluggy or unstable production of the wells. A statistical analysis
approach was applied, to see the probability of a given condensate to gas ratio value, which could be measured by
unconventional separator. The actual fluctuation of the rate observed by multiphase flow meter was used as initial parameters
for modeling (Fig. 7). Calculated results in Table 2 show that probability range of condensate-gas ratio that could be measured
with a unconventional separator will be confined between 42,5 g/m3 and 75,2 g/m3, averaging 62,9 g/m3.
SPE 115478 9

Figure 7. Fluctuation of the rates measured by MPFM, applied for probability calculation of CGR.

Table 2. Probability value of CGR which


could be measured by unconventional Cum ulative probability distribution of CGR
separator under slug flow regime 100
CGR, Measurement 90
Probability
g/m3 time, min. 80
Cumulative probability

100% 42.30 7.1 70


distribution, %

90% 42.56 7.5 60


80% 43.32 8.0 50
70% 45.54 8.7 40
30
60% 48.92 9.6
20
50% 53.24 10.5
10
40% 58.25 11.5
0
30% 64.11 12.3
40 50 60 70 80
20% 69.83 12.9
10% 75.25 13.1 Condensate Gas Ratio, g/Sm 3

0% 78.79 13.2
Figure 8. Cumulative probability distribution of Condensate Gas Ratio
10 SPE 115478

The following table shows the differences between the traditional testing means and multiphase well test results.

Table 3. Mass comparison of production metering results obtained during the validation process

Mass flow rate measured by Phase Tester Vx Mass flow rate measured by Separator
Choke, Total Mass
Well N
mm Difference, %
Gas mass Liquid mass Total mass Gas mass Liquid mass Total mass
rate, t/day rate, t/day rate, t/day rate, t/day rate, t/day rate, t/day

1  9.53  108.8  10.0  118.9  100.3  11.4  111.7  6.4 

1  11.11  132.2  13.6  145.8  127.3  14.2  141.4  3.1 

2  9.53  64.1  28.2  92.3  73.0  10.1  83.1  11.1 

2  11.11  88.6  27.5  116.1  91.8  22.6  114.5  1.4 

3  11  124.4  5.8  130.2  114.5  9.2  123.7  5.3 

3  12.7  165.2  8.9  174.1  152.3  11.5  163.8  6.3 

Despite of the short periods of production measurements using the separator good agreements of mass rates was achieved
between separator and multiphase flow meter. The well instabilities were probably the main cause of discrepancy of liquid
rates and gas rates. Overall all the comparison of the measurements is within the expectations. The mass comparison is close
except in one instance discussed later. One shall also not be too concerned at the seemly large difference in reported liquid
rates. As seems in most of the instances, the instabilities of the flow, caused by a combination of liquid loading and the
production way below dew point creates quite a challenge in comparison of the rates.

The measurements performed by separator three times over the flow of the well 2 under 9.53 mm choke demonstrated the
average gas rate 98.9 thousand sm3/day and 11.3 sm3/day of liquid rate respectively. The actual measurement of the rates by
multiphase flow meter installed below the separator and operated at the same conditions are 87.6 thousand m3/day of gas and
31.9 sm3/day of liquid respectively. The discrepancy of 11% of total mass rate between separator and multiphase flow meter
measurements on one of the instances for well 2 may have found an explanation in the analysis of the densities measured by
second multiphase flow meter installed in separator gas outline. The comparison of direct measurements of specific gravity of
flow mixture over the time with gas specific gravity at operation conditions demonstrated insignificant liquid carry over into
the gas line from separator. That may have resulted to overestimation of gas production rate and underestimation of liquid rate
by separator.

An interesting experiment was carried out to verify the efficiency of the separator and has been illustrated by locating a second
multiphase flowmeter at the outlet of the separator, on the gas line. During the period of flow with the separator online, it is
clearly possible to see that the gas at the outlet is almost dry, confirming the relative efficiency of the separator (around 90 %
efficient in terms of stripping the liquid from the gas). The following figure illustrates one of these periods. The mixture
density reported by the multiphase flowmeter matches closely the theoretical dry gas density. The instantaneous variations are
caused by small changes of pressure and temperature in the flow line. This also demonstrates clearly the capability of the dual
energy Gamma Venturi multiphase flowmeter to differentiate accurately between a 99% mixture of gas and liquid and dry gas
(100% gas). This figure also demonstrates the sluggy nature of the flow, with instantaneous variations of the mixture density
when the separator is not active (the beginning and the end of the plot). These variations match appreciatively the predicted
slug frequency caused by the lack of velocity in the tubing. These slugs are certainly challenges to the traditional separator, but
clearly seem to be caught by the centrifugal design used by Yamburggazdobicha. The ability to actually measure the total
volume of those slugs should then be properly achieved by the separator. One of the issue remains the time synchronization
with the multiphase flowmeter measurements remains the dependence on the very manual procedure used. This batch process
of measuring the liquid rate by closing the liquid drain valve and letting the level raise to a know level – a secondary outlet, is
performed every 15 minutes to 60 minutes, depending on the liquid flowrate and may carry some uncertainty.
SPE 115478 11

Figure 9. Inlet separator rate measurement of multiphase flow meter installed upstream of the separator and density
measurements by second multiphase flow meter in gas outline of the separator.

An important area of discussion was also raised on the proper set-up of the multiphase flowmeter. In the present mode of
operation, the nuclear attenuation of the liquid are performed through in-situ measurement performed at ambient conditions of
pressure and temperature. It has been suggested that some slight difference would be observed if the in-situ was performed
under pressure, to avoid the liberation of some light components out of the condensate. This point is quite valid in theory,
however, one shall also note that the propagation of the uncertainty on the liquid rate in the mode of operation of the dual-
energy gamma Venturi multiphase flowmeter is mostly a function of the gas attenuation and in-situ density. This fact may be
seen as surprising as it differs significantly from the separator, but can be resolved easily. The multiphase flowmeter
measurements are performed with no prior separation of the fluids. Therefore, the presence of an overwhelming phase
(definitively our case here, with 97% to 99.5% of gas) controls the overall uncertainty and propagation of errors from the fluid
property set-up.

A typical uncertainty analysis is provided in the following (Fig. 10). In this evaluation, the impact of setting up the meter with
the low and high energy attenuation determined at ambient condition rather than at line conditions is evaluated. The changes
are rather small in nature on the liquid and gas rates computed by the meter. One also notice that the most important factor is
the change of attenuation between the composition of the gas at line condition and the gas at ambient conditions. This also
justifies why in-situ determination of the gas attenuation is performed at the well site under the conditions of the flow, in order
to reduce the uncertainty on the liquid measurement. Regardless a pressurized in-situ attenuation cell has been made available
in order to reduce the impact of the determination of condensate attenuations on the measurement.
12 SPE 115478

Figure. 10. Impact of uncertainty due to slight change of composition between in-situ condensate at pressure and temperature
of the flow and ambient determination of the attenuations.

Conclusions and Recommendations


• The deployment of multiphase flowmeters in the Yamburgskoe Arctic Gas-Condensate field is operationally possible.
• The installation of Dual-Energy Gamma – Venturi multiphase flowmeters either on permanent basis or on mobile
basis is confirmed.
• The centrifugal separator should be equipped with some digital monitoring capability and a sight glass to monitor the
liquid level would also enhance the repeatability of the process
• The numbers of tests performed recently in Yamburgskoe Arctic Gas-Condensate field have clearly demonstrated the
ability of Dual Energy Gamma Venturi multiphase flow meters to perform high resolution and transient dynamic gas,
condensate and water rate measurements. This fact was additionally confirmed by agreement between the multiphase
flow meter measurements and the gas separator.
• Reasonable agreements in production measurements were obtained by multiphase flow meter at different operational
conditions and the separator in selected stable flow conditions.
• The biggest issue during production metering campaign related to the environmental conditions and particular hydrate
formations. Hydrate formation issue do not allow the utilization of conventional separator over the full period of the
test. To perform the rapid rates a measurement, avoiding any impact of hydrate formation, unconventional separator
is applied over many years of field development. The actual metrological performances and limited application of
short production rate measurements of well flow rate producing in slug flow regime is demonstrated.
• Hydrate formation issues make significant negative effect on performances of the test including production
measurement using multiphase flow meter. The ability of earlier hydrate identification by Venturi-Dual energy flow
meter was used to select interval for surface sampling and production measurement by separator. In the same time
multiphase flow meter Vx has unique possibility of multiphase production measurement in hydrate free area –
upstream of the choke, which was demonstrated during the tests.
• The accurate measurement of transient flow rate and identification of slug value parameters are extremely important
for well flow modeling, well completion and surface facilities designs. Observation of transient flow behaviour is also
important for selection of stable production intervals.
SPE 115478 13

• A number of operational benefits of using dual-energy gamma Venturi multiphase flowmeters have been identified:
o Dynamic flow measurement, evaluation and quantification of transient effects from the wells
o Ease of logistics
o Ability for continuous monitoring
o Quantification of slug frequencies and volumes / length
o Capability to identify early the quick accumulation of hydrates
• A comparison methodology was established allowing an easier analysis of the data. It consists in ensuring a proper
filtering of the unstable flow periods and avoiding comparison during the injection of methanol.

Acknowledgments
The authors are grateful to the Yamburggazdobycha for granting permission for publication of this paper. Furthermore the
operational and logistical support and contribution of the Yamburggazdobycha Gazprom well testing crew under the
leadership of N.M. Lebenkov. Thanks to Schlumberger in Russia A. Zhandin and V. Aanasyiev who operate the multiphase
flow meter in challenging environmental. Special thanks to G.Conort and V.Pivovarov for facilitate and communication
between Schlumberger and Yamburggazdobycha Gazprom.

References
[1] 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 Technical Conference and Exhibition, Dallas, Oct. 1-4.

[2] Theuveny B.G. and Mehdizaed P.: “Multiphase Flowmeters for Well and Fiscal Allocations”, Paper presented at the SPE
Western Regional / AAPG Pacific Section joint meeting, 20-22 May 2002, Anchorage, Alaska.

[3] Jayawardane S. and Theuveny B.: “PVT Sampling wuth Multiphase Flowmeters”, SPE 77405, San Antonio, Sep. 30 –
Oct. 2. 2002, Texas

[4] V. Afanasyev, B. Theuveny, A. Zhandin, O. Kulyatin, V. Bastos, P. Guieze, S. Jayawardane, S. Romashkin, “Sampling
with Multiphase Flowmeter in Northern Siberia - Condensate Field Experience and Sensitivities”, Paper SPE 115622,
presented at the 2008 SPE Russian Oil & Gas Technical Conference and Exhibition held in Moscow, Russia, 28–30 October
2008.

[5] Theuveny, B., Zinchenko,I.A., Shumakov, Y., "Testing Gas Condensate Wells in Northern Siberia with Multiphase
Flowmeters", Paper SPE 110873 presented at the 2007 ATCE in Los Angeles, USA, Nov 2007.

[6] SPE 105362, "The Identification of Condensate Banking With Multiphase Flowmeters—A Case Study", B.C. Theuveny,
P.D. Maizeret, N.S. Hopman, and S. Perez, Paper presented at the 2007 MEOS conference in Bahrain, March, 2007.

Si Metric Conversion Factor


cp × 1.0* E-03 = Pa·s
bar × 1.013 25* E+05 = Pa
psi × 6.894 757 E+00 = kPa
bbl × 1.589 873 E-01 = m³
B/D × 6.624 471 E-03 = m³/h
cuft × 2.831 685 E-02 = m³
cf/d × 1.179 869 E-03 = m³/h
lb/ft³ × 1.601 846 E+01 = kg/m³

* Conversion factor is exact


14 SPE 115478

Figures

Figure A-1. Equipment layout


SPE 115478 15

Figure A-2. Production history of the test for well number 1

Figure A-3. Production history of the test for well number 2


16 SPE 115478

Figure A-4. Production history of the test for well number 3

Вам также может понравиться