March 2007

Multiphase and wet gas metering provide a low-capex solution

Combining virtual and physical meters provides wider control of flowline networks for deepwater fields.
Vol. 228 No. 3  

Deepwater Technology

Multiphase and wet gas metering provide a low-capex solution

Combining virtual and physical meters provides wider control of flowline networks for deepwater fields.

�ge Rasmussen, FMC Technologies, Kongsberg, Norway

The benefits of metering the contribution from each well in a field may, in general, be quantified by how efficiently the production system is utilized at any point in time, and ultimately by the overall field recovery factor. Recent studies by Statoil, Hydro and the Norwegian Petroleum Directorate (NPD) show a 15% to 20% difference in recovery factor from subsea to platform wells. 1, 2 Multiphase and wet gas metering play an important role in closing this gap.

More and more offshore field developments, particularly in deep water, are opting for a solution, where a large number of wells are tied back to a limited number of flowlines, since this will provide the lowest capex solution, Fig. 1. To add to the complexity, artificial lift may be provided through a limited number of gas lift lines, or by a limited number of booster or multiphase pumps. All these elements contribute to increasing the number of variables to be optimized simultaneously, which leads to an increase in the overall complexity of the production optimizations. The basis for performing day-to-day production optimization is having an accurate model of the full-flowing network, and accurate measurements of current flowrates and flowing conditions at any point in the flowing network. A combination of physical and virtual metering may provide accurate metering and updated models, in addition to the capability of performing short- and long-term optimization within the same framework.

Fig. 1. Various subsea and platform-based field layouts.

Fig. 1. Various subsea and platform-based field layouts.


As developments move toward deeper water, the previously identified issues become even more pronounced. Due to increasing cost and complexity of operations with increasing depth, the potential benefit increases proportionally, whereas the initial investment in an overall metering and optimization system remains almost unchanged.

Enhanced potential through targeted interventions. As stated above, Statoil, Hydro and NPD have found an average difference in overall recovery factor, between 15% and 20%, for a field completed with subsea wells compared to a field with dry wellheads on a platform. 1, 2 For deepwater developments, it is anticipated that the average recovery factor will decrease as a function of water depth. 3 Obviously, the main reason behind the decrease in recovery factor is the added cost and complexity of performing interventions on subsea wells, factors that are accentuated even further for deepwater developments. Improvements in multiphase or wet gas measurement capabilities will not affect the cost of subsea interventions. These improvements will, however, enable the operator to target the correct wells. Increasing the amount of information available from each well will enable the operators to target the correct wells, and to ensure that the correct intervention type is performed. Together with developing technologies such as light well interventions, quality flowrate measurements may contribute to increasing the recovery factor from subsea and deepwater production systems.

Enhanced potential through production optimization. To add production capacity to existing facilities, or to develop fields at a minimum investment, we are seeing a trend toward more complex field layouts, Fig. 2. A typical manifold layout for a subsea development may be seen as a 4-or-6-slot manifold tied back to the production facility through one or two flowlines. To expand this layout at a minimal cost, it is much cheaper to daisy chain a second manifold on the back of an existing one, than to include a second stand-alone manifold with separate flowlines. For a new field, a two-manifold daisy chain configuration reduces the necessary number of flowlines and risers.

Fig. 2. Increasingly complex subsea layouts are required by fiscal realities.

Fig. 2. Increasingly complex subsea layouts are required by fiscal realities.

Obviously, the flowline and riser sizes will increase, but the flexibility of such a configuration will only be moderately reduced compared to the two-manifold/four-riser option. This solution's drawback is the added complexity of producing through the networks. Taking into consideration the need for artificial lift, either as gas lift or through boosting or multiphase pumping, the complexity of optimizing production increases significantly. The potential benefit from accurate measurements of production from each well increases significantly with the geometry's complexity.

Figure 3 shows an illustration based on production profiles from the Troll B field. The figure shows the difference between best-case flexible routing (blue line), and fixed-well routing (magenta line) over life of field for a four-well manifold tied back to the platform through two flowlines. The curves show a comparison between a fixed routing, and an optimized flexible well-to-line routing. To be able to realize the potential that lies within optimizing field production, it is essential to have accurate knowledge of produced flowrates from each well at all times. Any optimization tool is dependent on an accurate model of the production system. The quality of flow optimizations is linked to the quality of the model, which in turn is dependent on the accuracy of the measured flowrates.

Fig. 3. Optimized vs. fixed routing production..

Fig. 3. Optimized vs. fixed routing production.

Enhanced reservoir modelling potential. Values from Statoil on Gullfaks field show an increase in expected ultimate recovery factor from 46% in 1996 to 61% in 2004. 4 The increase is attributed to several factors, including advances in reservoir modeling. The reservoir models are based on a variety of data from 4D seismic to production logging. However, accurate drainage and injection distribution measurements are a necessary building block to provide accurate reservoir models.


Operators contemplating subsea developments must make a choice between specifying just physical meters or going with a virtual metering option. The advantages and disadvantages of these options are described in the following passages.

Physical meters. All subsea multiphase or wet gas flow meters (MPFM or WGM) face the common challenge, that the property they seek to measure, the volumetric or mass flow of each of the flowing phases, does not lend itself to direct measurement by currently available technology. To measure the volumetric or mass flow directly, the fluids must first be separated, and then measured. As this is not a practical subsea solution, or even practical for most topsides applications, the meters are restricted to measuring primary variables, and companies resort to calculation software to calculate flowrates based on the primary measurements. This requires a model description of the fluid flow, and an algorithm to calculate flowrates based on primary measurements. Figure 4 shows a typical logic diagram for a WGM/MPFM. 5

Fig. 4. WGM/MPFM logic diagram.5

Fig. 4. WGM/MPFM logic diagram.5

The primary sensing elements used range from direct measurements, such as pressure and temperature, to more advanced sensing elements that require calculation and interpretation, such as velocity cross-correlations and Gamma densitometers. Different vendors use different combinations of primary sensing elements that give strengths and weaknesses for different combinations of the three flowing phases. A meter designed to calculate three flowing phases is, as a matter of principle, dependent on a minimum of three primary sensing elements to achieve a mathematically defined problem.

Each primary sensing element is subject to an inherent error or inaccuracy. The total error of the multiphase or wet gas calculations will reflect the sum of the errors of the individual sensing elements, and the error of the applied flow model. A differential pressure measurement across a reduction in flowing area is, together with pressure and temperature sensors, probably the most widely employed primary measurement in WGM and MPFM technology. To give an accurate reading of the overall mass flow from a venturi meter, the flow calculations algorithm must have an accurate description of the density of the each of the flowing phases. If the PVT model used in the flow calculations does not accurately depict the actual flowing properties, the results will contain certain errors. This reflects the greatest challenge to multiphase metering today�the effect of changing fluid properties, and the inability to detect these changes.

To be able to meet their specified accuracy, many meter vendors require special circumstances, such as in-line calibration, or a requirement of only saturated water production at start-up (no free water production). Due to the complexity of many of today's production systems, it may not be possible to achieve reference measurements for meter calibration. In addition, most of the primary sensing elements detect a variable that is dependent on flowing fluids' PVT properties. The calculation software must reflect the correct PVT properties to yield a correct flow split. For most reservoirs, the fluid composition will vary over time. This, in turn, will result in decreased accuracy. Some meters have facilities for providing fluid samples to rectify this problem. This leaves the operator with the challenge of collecting and analyzing samples at a regular interval.

Virtual metering. The term, virtual metering, may be described as the concept of calculating flowrates, based on all available sensory information distributed throughout the production network. Virtual metering may be performed online or offline�the only prerequisite is that all sensory information is gathered within the same timeframe, to facilitate the computational models used. The basic idea is to combine the sensory input with the numerical flow models.

The virtual meter calculates a flow field through the entire production system. The flow field will inherently contain the calculated sensor responses at all sensor locations. The discrepancies between measured and calculated properties are quantified within an object function, and the object function is minimized through an iterative search. The procedure is described in its simplest form in Fig. 5. 6 It shows how the analytically solvable problem of single-phase flow through a venturi can be solved iteratively by imposing an object function.

Fig. 5. The concept of the object function.

Fig. 5. The concept of the object function.

The one-dimensional problem lends itself more easily to illustration than the three-dimensional problem consisting of three flowrates�oil, gas and water�flowing through a production well with multiple sensors. The principle, however, is the same.

Figure 6 shows the operating principles of a virtual meter. All required information is fed into the model, either on an on- or offline basis. The model makes an initial guess of the rates and calculates the equivalent sensor responses. The discrepancy between the measured and calculated system responses is quantified through the object function and evaluated. The procedure is repeated iteratively, until the value of the object function reaches one of a set of cut-off criteria.

Fig. 6. Oil well with standard instrumentation.

Fig. 6. Oil well with standard instrumentation.6

To obtain a well-posed mathematical problem, there must be at least one measurement per unknown, the unknowns being the three flowrates. The previous figures demonstrate the procedure with one and three unknowns. The system can be expanded to a full converging network of wells and lines. Recent software implementations have included the production separator measurements and/or the fiscal export measurements. By increasing the number of available measurements, the calculations will become over-determined, and the virtual meter will become increasingly redundant toward errors and failures in individual sensors.

One of virtual metering's major benefits is the accessibility of data and calculations. Where most physical meters provide the results directly, the virtual meter allows for a higher degree of interpretation and user interaction. Not only the flowrates are reported for each time step, but reporting and trending of the error function also give an indication of the calculations' quality. A typical scenario for a virtual meter is loss or error in one or several sensors. If the remaining number of sensors no longer provides adequate information to obtain a unique solution, the software allows for a reduction of the number of unknowns to obtain a mathematically defined problem. Typically, this may be done by introducing restrictions to the water-to-liquid or gas-to-oil ratios. The restrictions or limitations may be imposed as a function of any variable available to the calculations. The obvious example would be to introduce a fixed, time- or rate-dependent correlation to describe the changes in ratios. A different alternative is to use the knowledge obtained through reservoir simulations to describe the ratios as a function of bottomhole pressure and/or temperature.

A prerequisite to apply restrictions in the ratios is that the software-based solution has the ability to include sporadic or non-continuous measurements as part of the analysis. As opposed to a physical meter, the model used in the virtual meter will continuously be updated toward all available measurements, including sporadic non-continuous information, such as direct well tests or deduction tests. A trend in any production characteristic seen from one well test to the next may be extrapolated into the future to aid the calculations. In the same way that trends and developments may be utilized for predictions, the virtual meter also has the ability to revisit previously calculated time periods. Contrary to a physical WGM or MPFM, the virtual meter contains storage facilities to recalculate all data. Should calculated flowrates be based on a perceived trend or a sensor response that at a later stage is proven erroneous, all results may be recalculated, based on the updated data. As stated in the motivation section above, the primary benefit of accurate metering on a day-to-day basis is the input provided to perform short-term optimization. The long-term benefit of providing accurate information to the reservoir models will be ensured, even if rate calculations are redone at a later stage.

As described for the physical meters, virtual meters are also dependent on an accurate description of fluid properties. Should the actual flowing fluid differ significantly from the compositional description used in the software, this will result in a reduction in the quality of calculations. There is no technology available that can provide continuous or intermittent sampling of fluid properties or composition, subsea. To get an accurate composition, a laboratory test is the only quality solution. The quality of results will also depend on the quality of the analyzed sample. Sampling and analysis of fluid composition and properties form a whole separate field, but a field that has great bearing on the quality of the results obtained from a virtual or physical meter. Due to the virtual meter's ability to continuously report an error function, it will give an indication of the quality of calculations. This may, in turn, be interpreted as an indication of the need for updated fluid characteristics.

Data validation. FMC's virtual meter, FlowManager, uses a number of data validation operations prior to utilizing any data for flow calculations. All imported data are evaluated statistically to ensure that the values are probable. The statistical evaluation employs simple principles, such as maximum and minimum deviation, and rate of change to verify sensor quality. The software also contains features for recognizing certain scenarios, where multiple sensors should detect the same quantity. As an example, a WGM is typically mounted upstream of the production choke. The conditions at this position are essential for production control and optimization, so a typical field will have stand-alone pressure and temperature sensors at about the same location. Continuously comparing the sensor values, and comparing these values to the model responses, gives an indication of both error and a trend in error development (drift) for each of the sensors.

A different example is sensors that only sporadically should show the same value. For the typical FMC subsea tree, the upstream choke sensors are mounted between the master and wing valves. Intermittently, the wing valve will be closed, while the master and choke valves are open. For this scenario, the upstream and downstream choke pressures should be nearly identical. If they are not, this will indicate an error in one or both sensors.

By analyzing the production system, it is probable that a large percentage of sensors may be subject to intermittent verification. By using this information actively, the quality of reported flowrates will increase significantly.


The main challenges for both physical and virtual wet gas and multiphase meters are the quality of individual sensing elements, and the fluid properties that calculations are based on. By combining WGM or MPFM technology with a virtual meter, the system may become over-determined, and to a certain degree redundant toward errors in individual sensors. To calculate flowrates at any point in the production system, a virtual meter depends on having at least three measurements to find a unique solution. The same applies to a physical meter that depends on a minimum of three measurements to calculate three flowing phases. By importing data from the primary sensing elements of the physical meter into the framework of a virtual meter, the number of available measurements will increase.

Any fluid or flowing property that can be both measured and modeled is available to be included in a virtual meter. In this sense, a virtual meter can be used as an addition to, and a substitute for, traditional multiphase meters, if adequate measurements are available. The virtual meter can incorporate individual sensor readings that a multiphase flowmeter consists of, or it may use the calculated rates directly. Readings from the multiphase meters are incorporated with all other measurements within the production system, to give a best fit solution for the entire package. In many fields where multiphase meters are employed, lack of reliability caused by discrepancies between the sum of all meters and the fiscal measurements causes uncertainty in all measurements. The virtual meter is not limited to giving a spot reading at the sampling position. Thus, it may be used as a supervision tool to establish the most likely source of error.

To enable the software to use separator measurements, or the results from a multiphase meter, it is necessary to have a solution that allows for a controlled difference between the measured and calculated quantity. The software contains a weighted object function. Assigning different weights to the different measurements allows the user to accentuate or discard certain measurements. In an over-determined system, this feature gives an operator the possibility to use the system for sensory evaluation.

Additional benefits. Measuring the flowing phases does not in itself provide added value�the value comes from how this information is used in production planning, production optimization and in reservoir modeling. Traditionally, operators have used data from well tests to calibrate models of wells and pipelines for production planning and optimization. Employing a combined solution will give a continuous verification of well model quality, as the sum of all production is verified continuously toward topsides separator measurements. It will also employ all available information to provide, as accurately as possible, a prediction of the flowing phase rates at any point in the production system. Both production planning and optimization will improve, due to the improved accuracy of the employed models.

By employing such a tool, the added benefit of having a full mathematical description of the flowing network may be used to perform a variety of tasks. Online production optimization may be performed on a regular basis by production engineers, or it may be performed as a fully automated, set interval optimization that provides the best possible choke settings for production and gas lift chokes, and routing recommendations. The optimization run by the software is designed around enforcing various system constraints and optimizing toward these constraints. The software also contains facilities for closed-loop choke control. This means that a field may be run in fully automated mode, where choke positions are taken directly from the optimization recommendations.

Employing the complete system may not only improve the field's overall efficiency and recovery factor, it may also limit the manpower needed to run the field, and enable more personnel to work remotely.


Virtual and physical metering are rapidly developing technologies facing many of the same challenges. The main challenge for multiphase flow metering at the current state of technology is the inability to directly detect changes in the compositions of the produced fluids. However, by incorporating the two metering principles in a common framework, the number of available measurements may be sufficient to detect changes in fluid properties, if not to detect changes in the composition directly.

The two technologies also face common challenges related to sensor quality. At the very least, this issue should be dealt with by using continuous statistical surveillance of all sensors' reported values. By evaluating trends over time, and taking advantage of distinct situations where multiple sensors are exposed to similar or identical conditions, the data quality, and subsequently the value of each sensor, will increase significantly.

Incorporating all available data from the production system, including the WGMs/MPFMs, in the framework of a virtual meter paves the way for closing the gap between obtaining relevant data, and turning the information into increased revenue. WO


1 Karstad, P. I., “Subsea IOR challenge and technology solutions,” Statoil presentation at SPE, April 2004.
3 Laherrère, J. H., “Future sources of crude oil supply and quality considerations,”
DRI/McGraw-Hill/French Petroleum Institute conference, June 1997.
4 Hesjedal, A., ”Increased oil recovery on Tampen,” Statoil presentation
to Morgan Stanley, North Sea Conference, May 2005.
5 Cunningham, C., E. Fjøsna and K. Berg, “Combined physical and virtual multiphase metering for redundancy and synergies,” Deep Offshore Technology Conference, November 2004.
6 Rasmussen, Åge, “Field applications of model-based multiphase flow computing,” North Sea Flow Measurement Workshop, October 2004.


Åge Rasmussen is a specialist engineer in Flow Management for FMC Technologies. He holds an M.Sc. degree in Fluid Mechanics from the Norwegian University of Science and Technology (NTNU). Mr. Rasmussen joined FMC's FlowManager department in 2000, and has since been involved with all aspects of the software, from model calibration to software development. He currently works out of FMC's Australia office in Perth, providing flow assurance and field development support for FMC's customers in the Asia Pacific region.



Connect with World Oil
Connect with World Oil, the upstream industry's most trusted source of forecast data, industry trends, and insights into operational and technological advances.