Sei sulla pagina 1di 11

New Approach For In-Line Production Testing for Mature Oil Fields Using

Clamp-on SONAR Flow Metering System

Giovanni Morra, Sara Scagliotti (Eni E&P) Dhia Naeem, Ali Al Alwan (South Oil Company Iraq) Siddesh Sridhar, Ahmed
Hussein (Expro Meters)


Monitoring produced oil, gas and water rates from individual wells plays an important role in reservoir management and
production optimization. While beneficial, obtaining timely and accurate wellhead measurements can be challenging due to a
range of factors.

This paper describes a cost-effective and convenient approach to production surveillance of black oil wells using clamp-on
flow meters (Sonar), integrated with a PVT and multiphase flow engine to calculate the properties of the produced fluids, and
the individual phase flow rates. The PVT engine calculates the gas and liquid properties of the produced fluids, at the pressure
and temperature conditions measured where the Sonar flow meter is clamped-on. The Sonar flow meter provides a direct
measurement of the mixture flow velocity within the flow line. The mixture flow velocity is then interpreted in terms of actual
gas and liquid flow rates. Once the gas and liquid flow rates are determined at actual conditions, the oil, water, and gas flow
rates are reported at standard conditions based on the PVT data calculated by the PVT engine.
The new approach was deployed in a mature field to test 6 different wells against existing conventional test separators as the
reference. The tests showed that Sonar meters could be installed either upstream or downstream of the choke manifold thus
allowing more flexibility in terms of field installation. Some of the wells were tested at fixed flowing conditions while others
were tested at multiple choke settings. All data is found to be within acceptable limits. The results show that, overall, the new
Sonar-based approach could perform within ±15% or better from the reference, depending on the accuracy of the input PVT
data (including water cut) and the flowing conditions.

The new clamp-on approach requires about 90 minutes for installation and commissioning which allows the possibility to
perform multi-rate testing of the wells in one day. Therefore, the Sonar clamp-on methodology offers the opportunity to
increase the well rate measurement frequency at a field-wide level thus allowing a field production back-allocation with
consequent benefits in field/production management.


Accurate measurement of net oil rate from individual wells is a critical component in effective oil field management,
influencing production optimization strategies and allocation issues. The main equipment currently used for this purpose in the
field is either the CTS (Conventional Test Separators) or the MPFM (conventional Multi Phase Flow Meter). The separators in
the field DGS (De-Gassing Station), which were originally designed as test separators, are currently being used as Production
Separators (PS).The field under investigation is characterized by both carbonate and sandstone reservoirs on production since
more than 60 years. A total of 240 wells have been drilled, of which more than 100 (40 in natural flow and 60 completed with
ESP) are on production and 40 are water injectors. Assuming that each producing well should be tested at least twice a year,
more than 200 well tests need to be performed per year. Considering the fact that each test takes an average of 3 days, the
consequences are quite significant:

a. Multiple well test packages would be required to accomplish the desired testing frequency, which may not be
available and may be cost prohibitive.
b. The amount of gas and oil flared to the atmosphere may exceed environmentally accepted levels, unless the existing
wellhead connections and well test equipment are modified.
c. The production losses/deferral associated with traditional well testing methodologies may be significant.
d. Shutting in wells to rig up test equipment can negatively impact the long-term performance of the well.
e. Higher HSE risk associated with the increased number of well tests.

The solution of re-injecting the produced fluids into the flow line has been already considered for the new well site facilities.
To achieve this, it is necessary to have two gate valves in the production flow line, one to connect the well with the CTS and
the other one for re-injection. The MPFM offers a flare-less alternative, but it is usable only where the two gate valve system
has already been implemented at the well site. In order to address some of the shortcomings described above, it was decided to
evaluate the capabilities of Sonar metering as an alternative for well testing.

To evaluate the feasibility of using the clamp-on Sonar as an alternative for well testing, it was initially decied to do a
preliminary trial on two wells. The tests were carried out on wells A-01 and A-02 in conjunction with ongoing production
testing activities. The CTS was used as a reference to check the Sonar meter accuracy. At the conclusion of the preliminary
trial, it was deemed that the technology was suitable for the field, providing sufficient accuracy and easy handling in terms of
field deployment and installation. It was consequentially decided to deploy Sonar metering on a larger scale in a field-wide
well test campaign.The field-wide campaign comprised the testing of 94 wells (including 3 water injection wells) over a period
of 108 days. The Sonar results from the individual producing wells were aggregated and compared to the total production rates
as measured at the Degassing Stations. The aggregate Sonar production data was in reasonable agreement with the actual
production rates as measured at the DGS and to hydraulic well models expectations. In order to address some differences seen
between the Sonar test results and the expected flow rates, 4 wells were retested against the CTS at the conclusion of the Sonar
field-wide campaign. Again, the results showed that the Sonar meters performed within expectations.

This paper presents the results of the Sonar testing, including detailed comparisons between the Sonar and CTS oil, gas and
water rates for the six wells where reference data is available.

a. Production Allocation / Well Testing

Worldwide, the majority of oil & gas production allocation is achieved using separator-based measurements. Test
and production separators are generally used to allocate production from individual wells and fields. Most
separator-based measurement methodologies rely on the assumption of complete separation of the gas and liquid
phases prior to measurement.

In practice, however, complete separation is often difficult, if not impractical, to achieve, the accuracy of separator-
based measurements could be limited by flow instrumentation-related issues.
On the flow measurement side, the test separator functionality may be limited by the quality of the incoming fluid
stream, which many times requires further processing, such as de-sanding (or solids removal), in order to protect
both the separator and the ancillary instrumentation. In addition, sanding and/or solids may cause the separator to be
by-passed for certain periods of time, with no flow measurements available during the by-pass period. Another
potential issue is incorrect selection and sizing of flow meters on the liquid and gas legs of the CTS (for example,
wrong orifice plate size) which may yield inaccurate measurement results.

The size and the cost associated with mobilizing and operating CTS packages in the field may also limit the number
of wells to be tested. Therefore, for field management purposes, various schemes and models are often developed to
infer the production rates for each well. These models are used to predict individual well performances and in turn
to make selection of intervention and work-over methods based on incomplete data with regard to actual well
performances. To reduce the uncertainties of the models and the consequent risk associated to the selection of the
interventions, it is mandatory to strengthen them with measured data. This data has been provided efficiently by
Sonar Clamp-on meter within the required timeframe.

b. Overview of Technology

Clamp-on Sonar-based flow meters utilize Sonar processing techniques to determine the speed at which naturally
occurring, coherent flow structures convect past an array of sensors clamped-on to the outside of the pipe.

Figure 1 illustrates the naturally occurring, self-generated, coherent structures present within turbulent pipe flows.
Naturally occurring, self-generating, turbulent eddies are superimposed over the time-averaged velocity profiles.
These coherent structures contain fluctuations with magnitudes on the order of 10 percent of the mean flow velocity
and are carried along with the mean flow. These eddies remain coherent for many pipe diameters and convect at, or
near, the volumetrically-averaged flow rate in the pipe.

Figure 1 : Sonar-based Flow Meters with Coherent Structures within Pipe Flows

In Sonar array processing, the spatial/temporal frequency content of sound fields are often displayed using “K-ω"
plots. K-ω plots are presented as surface plots in which the power of a sound field is allocated to bins corresponding
to specific spatial wave numbers and temporal frequencies. On a k-ω plot, the power associated with coherent
structures convecting along with the flow is distributed along “the convective ridge”. The slope of this ridge
indicates the speed of the turbulent eddies. Thus, identifying the slope of the convective ridge determines
volumetric flow rate.

Slope – 34 ft/sec



Figure 2: K-w plot from a Clamp-on Sonar meter on a Gas Well

Figure 2 shows an example of a k-ω plot generated from the diagnostic output of a Sonar flow meter clamped on to
a 4-inch, schedule 80 pipe with black oil flowing at 30 barg. As shown, the k-ω plot exhibits a well-defined
convective ridge. The slope of the convective ridge was determined by the SONAR flow meter to be 34 ft/sec.

Figure 3: Flow regime operating range – Sonar Technology

Sonar technology capabilities include:

 Bulk flow velocity (mixture flow) measurement over (almost) the entire GVF range
 Multiphase flow interpretation from velocity measurements and PVT data
 Specific algorithm/model developed for each particular range (as outlined above – the depicted values are
to be used only as guidelines)

c. New Approach to Field-Wide Well Testing

The approach utilizes a Sonar flow meter to clamp-on to well head piping to measure the mixture volumetric flow
rate at actual conditions. The flow measurement is then combined with measurements of the process pressure and
temperature and user-supplied compositional information to determine individual phase flow rates. Several samples
are taken of the fluids at line conditions and a base sediment and water (BS&W) test is made to determine water cut
(WC). This process is shown schematically in Figure 4.

Pressure &
Temperature SONAR Velocity

Multiphase Flow Qgas
& Black Oil Model


Watercut (WC)
Figure 4: Schematic of Sonar based Production Surveillance Methodology

Compositional analysis and equation of state PVT modeling of black oils can be quite complex. Unlike gas and gas
condensate wells for which the hydrocarbon composition is made up primarily of short length hydrocarbons (C7+
>~10%), black oils typically contain a large fraction of very long hydrocarbon chains rendering detailed
compositional characterization and subsequent EoS PVT modeling quite challenging. Historically, the oil and gas
industry has modeled the PVT properties of produced oils using “black oil” models. These models utilize readily
measurable physical characteristics of produced hydrocarbons (such as Gas to Oil Ratio (GOR), WC, and bubble
point pressure) as input to empirical correlations to predict the PVT properties of the black oil over a range of
pressure and temperatures.

The Sonar approach to oil well production surveillance utilizes “black oil” models for oil wells to interpret the
output of the Sonar meter using flow line pressure, temperature and well bore composition in terms of flow rates of
gas, oil and water at standard conditions. The Sonar meter measures mixture velocity at line pressure and
temperature. The flow computer then calculates the liquid volume fraction at actual conditions within the process
flow line using a Black oil model and using input WC. This provides a measure of the oil rate at actual conditions.
A shrinkage factor calculated by the black oil model is applied to report oil rate at standard conditions. The gas and
water rates associated with the produced oil are then inferred from the measured oil rate.

The Sonar meter is typically clamped on the pipe in about one hour, allowing for a relatively quick installation &
commissioning. Within meter installation guidelines, the Sonar meter can be installed on the pipe anywhere at the
wellhead, either upstream or downstream of the choke. The clamp-on non-intrusive nature of the meter approach
eliminates any pressure drop and production loss, and minimizes the HSE risks associated with invasive well
testing methods. The duration of the test may vary depending on the stability of the flow conditions, but in general

an assumption of one test per day is reasonable. Multi-rate tests are also possible on an intra-day basis, provided the
flow is stabilized at each rate. This enables to test the wells at various choke settings and/or ESP frequencies with
the goal of determining the settings resulting in optimum production.

d. Initial Sonar trial – 2 wells

During the initial trial, the clamp-on Sonar meter was installed on the pipe connecting the CTS to the X-Mas tree.
For the first well, the Sonar was installed upstream of the choke manifold (Figure 5).

Figure 5: Well A-01 Sonar Setup

The flow was monitored for 2 hours when the well was flowing at a choke setting of 48/64th. The test measurements
are displayed in Figure 6, while the results are plotted in Figure 7.

Well A‐01 Sonar Measurements

30 300

25 250
Pressure (psig) and Temperature (degF)
Sonar Velocity (Vsonar) (ft/sec)

20 200

15 150

10 100

5 50
Well Head Pressure

Well Head Temperature

0 0

Figure 6: Well A-01 Sonar Measurements


Well A‐01 Sonar vs Test Separator

2500 2.0


2000 1.6

Test Sep Oil Rate 1.4

Gas Rate (mmscfd)

Sonar Oil Rate

Oil Rate (STB/d)

1500 1.2
Test Sep Gas Rate
No Test Sep Inaccurate Test Sep Gas Sonar Gas Rate
Gas Flow Flow Measurement 1.0

1000 0.8


500 0.4


0 0.0

Figure 7: Well A-01 Sonar vs. Test Separator

The test revealed certain limitations associated with the CTS methodology. First, the flow stream could not be
diverted through the test separator at the beginning of the well test due to the high amount of solids entrained in the
flow stream, which is a typical condition for new or re-worked wells. Therefore, the CTS flow instrumentation
could not be used to measure the flow rates for that period of time.
The second issue is related to the separator inertia, which means that the separator takes time to reach the designed
level at which both the liquid and the gas measurements are deemed valid.
The third issue is related to the separator gas measurement. Since the typical flow measurement of the gas leg is
performed using the orifice meter, the proper sizing/selection of the orifice plate and/or dP sensor becomes very
important. Choosing a plate with a smaller beta ratio may limit the flow measurement since the dP generated across
the plate may exceed the dP sensor capacity.
The Sonar meter, due to its clamp-on design and relatively wide flow range, overcomes some of the limitations
described above and offers the possibility of monitoring the flow over a wide range of conditions.
For convenience, the averages associated with the above measurements and results are summarized in Table 1.

Sonar meter Test Separator Difference (%)

Duration Choke
(min) (64ths) Qgas Qoil Qgas Qoil
Qgas Qoil
(MMscfd) (stb/d) (MMscfd) (stb/d)
A-01 142 48 0.71 2152 0.67 2106 7 -2
Table 1: Well A-01 Sonar vs. Test Separator

For the second well, the Sonar meter was clamped on to the connecting pipe between the test separator and the
choke manifold, as depicted in Figure 8.

Figure 8: Well A-02 Sonar Setup

The Sonar flow was monitored during the majority of the test period, at all three different choke settings of 24/64th ,
48/64th and 64/64th. The measurements are shown in Figure 9, and the results are plotted in Figure 10.

Well A‐02 Sonar Measurements

70 250


Pressure (psig) and Temperature (degF)
Sonar Velocity (VSonar) (ft/sec)



Pressure 50
10 Temperature

0 0

Figure 9: Well A-02 Sonar Measurements


Well A‐02 Sonar Rates

5000 5

Sonar Oil rate
4500 4.5

4000 Sonar Gas Rate 4

3500 3.5

Gas Rate (mmscfd)

3000 3

OIL Rate (bpd)

2500 2.5

2000 2

1500 1.5

1000 1

500 0.5

0 0

Figure 10: Well A-02 Sonar Oil and Gas Flow Rates

For convenience, the averages associated with the above measurements and results are summarized in Table 2.

Clamp–on Sonar CTS Difference (%)

Duration Choke
Well Qgas Qoil Qgas Qoil
(min) (64ths) Qgas Qoil
(MMscfd) (stb/d) (MMscfd) (stb/d)

120 24 0.347 527 0.57 885 39 40

A-02 120 48 1.366 2127 1.29 2201 -6 3
75 64 1.585 2486 1.48 2533 -7 2
Table 2: Well A-02 Sonar vs. Test Separator

Both the gas and oil rates are well within a 10% error band for the higher choke settings. The low choke setting
(24/64”) test point showed a 40% difference for both oil and gas rates between the Sonar and the test separator.
Possible explanations for this deviation include incorrect separator settings (the records showed unstable parameters
throughout the flow duration) or the low flow limit of clamp-on Sonar at the low pressure/velocity conditions this
well. The Sonar results from both well tests were consistent and considered satisfactory and sufficient to move
forward with field-wide implementation.

e. Implementation of field-wide Sonar well testing

Field-wide production testing using the Sonar meter commenced after the trial tests. There were 94 wells tested in
the first testing campaign. These included 91 oil wells (naturally flowing and ESP lifted) and 3 water injection
wells. Sonar testing was also performed at the gas outlet of the 1st Stage Separator in the de-gassing stations to
obtain volumetric gas flow rates. Reservoir PVT data was provided and used as an input for the black-oil models to
interpret multiphase flow rates.
The simplicity and the flexibility of the clamp-on Sonar system allowed for a relatively quick turnaround in terms
of training.
The individual wells flow rate measurement gave the opportunity to promptly take necessary actions in order to
optimize the whole process stream for the field and to identify eventual problems in the wells. As a consequence the
production downtime was reduced and the production optimization workflow is strengthened.
All the well optimization models were tuned using the Sonar well test data, and particular benefit were identified on
ESP completed wells, due to the presence of real time down-hole pressure and temperature sensors. Sonar testing
resulted in updating of the hydraulic well models and real-time ESP frequency optimization at the well site.

A very small percentage of tests where discarded, due to the flow at or below Sonar measurement capabilities. For
these wells, a change in piping layout was recommended to allow for installation of a smaller Sonar meter size.

f. Validation – Comparisons with the test separator – 4 wells

At the end of the first field-wide testing campaign, 4 wells were tested in conjunction with test separators for
validation purposes. Three wells were tested at single choke settings and stable flow conditions, while a fourth well
was tested at multiple choke settings.
The Sonar meter was clamped upstream of the temporary choke manifold (part of the test separator package) for all
tests. The multiple-choke setting well test was chosen for illustration purposes (See Figure 11).

Figure 11: Well A-03 Sonar Setup

The test measurements are displayed in Figure 12, while the results are plotted in Figure 13.

Well A‐03 Sonar Measurements

40 400

35 350

30 300
Pressure (psig) and Temperature (degF)

25 250
Sonar Velocity (VSonar) (ft/sec)


20 200

15 150

10 100

5 Pressure 50


0 0

Figure 12: Well A-03 Sonar Measurements


Well A‐03 Sonar vs. Test Separator

3.00 Choke-64/64" 3000

2.50 2500

2.00 2000

Qgas (mmscfd)

Qoil (STB/d)
1.50 1500

1.00 1000

Sonar Gas Rate

0.50 500
Test Sep. Gas Rate
Sonar Oil Rate
Test Sep. Oil Rate
0.00 0

Figure 13: Well A-03 Sonar vs. Test Separator

As observed during the initial evaluation trial against the CTS, the clamp-on Sonar system tracked the separator
measurements relatively closely. Also, the Sonar proved once again to be a more responsive system compared to the
CTS at the beginning of a well test after the well shut down.
The data from all 4 tests is summarized in Table 3.

Clamp–on Sonar CTS Difference (%)

Duration Choke
Well Qgas Qoil Qgas Qoil
(hrs) (64ths) Qgas Qoil
(MMscfd) (stb/d) (MMscfd) (stb/d)

10.5 24 0.486 994 0.467 1152 4 -14

A-03 12 48 0.998 2041 0.929 2106 7 -3
2.7 64 1.351 2762 1.229 2805 10 -2
A-04 6 64 1.900 2752 1.686 2621 13 5
A-05 5 64 2.491 3194 2.280 3105 9 3
A-06 2.7 48 6.794 4249 5.291 3417 28 24
Table 3: Validation Results - Sonar vs. Test Separator

The validation tests again proved that the new clamp-on Sonar approach for well testing is reliable and yields a
reasonable measurement uncertainty over a wide range of choke settings and flow conditions. The validation testing
results were consistent with the measurements recorded during the initial evaluation tests, giving the confidence to
proceed with subsequent field-wide well testing campaigns using the clamp-on Sonar flow meter.


Both the individual and the field-wide well test evaluations of the clamp-on Sonar system were successfully completed without
any safety incidents. The results were overall considered to be within the accepted uncertainty range.The following
conclusions were derived from the Sonar well testing campaign:

a. The clamp-on Sonar system measured three phase well production with accuracy comparable to conventional
well testing methods such as CTS and MPFM. For the wells where the flow was at or below Sonar measurement
capabilities, a change in piping layout was recommended to allow for installation of a smaller Sonar meter size.
b. The Sonar deployment provided more individual well production data than could have been collected by
conventional means over the same time period, thereby contributing significantly to the improvement in
production allocation and facilities management. Periodic individual well testing resulted in better production
allocation and optimum use of production separators and other equipment further downstream and contributed to
decision making for work-over prioritization.
c. Due to its portability and ease of installation, the Sonar system allowed a large number of wells to be tested in a
short period of time without any production interruption/loss, proving to be a very cost effective way to do field-
wide well testing. The simplicity of the Sonar system allowed and timely and effective training
d. The new approach, due to the amount of data provided over such a short period of time, helped to improve field
management, i.e.:

1. On a Natural Flow multilayer production well Sonar meter allowed to diagnose a leaking in the
annulus packer
2. On ESP lifted wells, the Sonar meter allowed for real-time ESP optimization at the well site, tests
at multiple ESP drive frequencies helped optimize ESP frequency for target production.
3. On ESP lifted wells, the Sonar meter allowed to diagnose blocked circulation valve in the annulus
flow line.

CTS = Conventional Test Separator
DGS = De-Gassing Station
dP = Differential Pressure
ESP = Electric Submersible Pump
GOR = Gas to Oil Ratio
HSE = Health Safety and Environmental
MPFM = Multi-Phase Flow Meter
PS = Production Separator
PVT = Pressure, Volume, Temperature
WT = Well Test
WH = Well Head
WC = Watercut

The authors gratefully acknowledge the management of South Oil Company of Iraq (SOC), ENI S.p.a. and Expro Group for
permission to publish this work and acknowledge the efforts of numerous colleagues who have contributed to the
development and evaluation of this technology.

1. Kragas, T. K., et al., “Downhole, Fiber Optic Multiphase Flow Meter: Field Installation” Presented at the Society
of Petroleum Engineers Annual Conference, SPE Paper #77654, September 2002.
2. Gysling, D.L., Loose, D.H, "SONAR-Based Volumetric Flow And Entrained Air Measurement For Pulp And
Paper Applications", TAPPI Spring Technical Conference, Paper 58-1, Chicago, IL, May 2003.
3. Gysling, D.L., Lu, M., Wen, T., “Clamp-on Two Phase Measurement of Gas Condensate Wells Using Integrated
Equation of State Compositional Models”, 28th International North Sea Flow Measurement Workshop, 26th –
29th October 2010, St. Andrews, Scotland, UK