Scott Gaisford
Multi-fluid Inc.
Golden, Colo.
John Amdal
Saga Petroleum As
Sandvika, Norway
Hans Berentsen
Den Norske Stats Oljeselskap As
Stavanger, Norway
An encouraging multiphase meter test may lead to improved economics for developing marginal offshore oil and gas fields.
The test was sponsored by Saga Petroleum AS and carried out by Den norske stats oljeselskap AS (Statoil) on the Gullfaks B platform, Norwegian sector of the North Sea.
Saga and Statoil have been working on multiphase problems for many years. The multiphase meter was developed and supplied by Multi-Fluid Inc. and Multi-Fluid International A/S.
BACKGROUND
Companies worldwide are looking for new production methods for offshore oil fields. Traditional methods have been developed for larger fields where economies of scale justify the capital investments.
But in many areas, undeveloped smaller fields cannot bear the cost of dedicated production facilities.
Multiphase transportation to existing production facilities is widely held to be the key to producing these smaller fields.
BENEFITS
New multiphase technology can extend the distance over which unseparated oil, water, and gas streams can be transported, from a limit of several kilometers today to perhaps 200 km in the future.
Offshore, such multiphase transportation can make marginal fields economical and extend the life of platform production facilities.
Likewise onshore, long distance multiphase transportation to existing facilities can improve the economics of remote fields.
While the benefits of multiphase transportation methods are obvious, development of this technology has been difficult, although research in multiphase problems has been well financed over the last 10-15 years.
Some successes are being achieved in multiphase pumping and multiphase transport modeling, but hydrate formation and corrosion phenomena are not well understood in multiphase flow lines.
Conversely, multiphase metering is advancing toward practical applications.
COST SAVINGS
For efficient production, a well's production is sampled through a test separator usually once a month. Onshore the costs are low, but for a remote subsea field the costs are high.
A major cost factor is the installation of a test line through which production of individual wells can be routed to a topside test separator.
Another significant problem is allocation. Quite often, owners of topsides facilities are not the same as owners of the marginal field.
Consequently, the production must be measured separately. This may require placing on the topside facility an additional test or production separator dedicated to the marginal field. Costs can be high for preparing the space to accommodate the additional equipment.
Multiphase meters capable of continuously measuring the production rates of oil, water, and gas coming from wells or flowing in pipes can eliminate the test line, the dedicated separator, and the need for extra space.
By installing the meter on each subsea well head, well production and reservoir behavior is known at all times. Neither the test line nor the test separator is needed.
A meter placed on the main production line from the subsea field eliminates the dedicated separator for allocation purposes.
Many efforts are under way to develop and prove multiphase meters. In Norway alone, there are four significant independent efforts.
MULTIPHASE METER
The meter being tested by Saga and Statoil is a 4 in., ANSI Class 600 meter. It is supplied by Multi-Fluid and called the LP meter.
The full-bore instrument is 66.7 cm (26.3 in.) in length and has no moving parts. The meter spool piece was fabricated in 316L stainless steel and weighs about 100 kg (220 lb). Some of the important meter specifications are listed in Table 1.
Unlike some multiphase meters under development, this meter measures the complete production stream without having to separate partially or completely the gas from the liquid.
Fig. 1 shows a 3-in. LP meter spool piece, the associated electronics in an explosion-proof enclosure, and the PC interface system used to configure the system at installation. The spool piece is described in Fig. 2.
MEASUREMENTS
The complete multiphase meter has two separate meters.
One is the composition meter for measuring the instantaneous volume or mass fractions of oil, water, and gas in the sensor. The other is a velocity meter for determining the speed of the mixture through the sensor.
An instantaneous volume or mass production rate for each component is calculated by combining the outputs from the two meters.
The composition meter measures the raw well stream's dielectric properties (permittivity and conductivity) and the density. The dielectric properties are more sensitive to the water, while the density is more sensitive to the gas.
The dielectric measurements are made with a new and patented microwave technique that allows accurate, stable, rapid, full-bore electrical measurements of the process fluid.
The density is measured with a conventional gamma-ray densitometer mounted directly on the sensor spool piece.
The velocity meter determines flow rate using cross-correlation techniques. More specifically, the velocity meter makes very rapid microwave dielectric measurements in two cross-sectional measurement sections. The axial spacing between these sections is known.
Statistical analysis of the signals from each of the measurement sections determines an average time for the material to flow from the first measurement section to the second. By knowing the transit time and the spacing between the measurement sections, the velocity can be determined.
LIMITATIONS
The multiphase meter tested by Saga and Statoil has some limitations.
It is unable to measure production streams where the water is the continuous phase of the liquid. In practice, this limits the usage of the meter to applications having water/oil ratios not much greater than 1.
The meter is also designed to operate with bubble-flow conditions where the velocity exceeds 2 m/sec. Consequently, the manufacturer recommends that the meter be installed immediately downstream of a static mixer with flow directed vertically upward.
In this configuration, the mixer can break up most annular or stratified flow conditions for the short distance necessary to perform the measurements.
The meter is intrusive. Although it has virtually no pressure drop, it cannot be pigged.
TEST SETUP
The meter was installed on the Gullfaks B platform in a 4-in. bypass loop connected to the 8-in. test line running from the test manifold to the test separator. Twelve different wells were available for testing. The water/oil ratios vary from near zero to more than 3. Gas/oil ratios at standard conditions are typically 50-70 cu m/cu m (281-393 scf/bbl). Six different wells were tested.
At test conditions, the temperature and pressure were about 65 C. and 76 bar (149 F. and 1,100 psi). The flow velocities on the tested wells varied from about 3 m/sec up to about 9 m/sec (about 10-30 fps).
To avoid slug or plug flow conditions, the 8-in. main production line leading into the test separator was reduced to 4 in. in front of the meter. Therefore, the flow regime for these tests was uniformly bubble flow in spite of most wells producing about 40% gas by volume.
The test separator itself is a very large three-phase separator about 15 m (49 ft) in length and 3.3 m (11.5 ft) in diameter. Because of its size, the separator is very efficient at separating oil, water, and gas phases.
Typically, the residual water content of the oil line is less that 0.5% and the oil in the water line is less than 0.1%. Consequently, the reference measurements from the test separator are more accurate than would normally be expected.
The instrumentation on the test separator consists of an orifice meter on the gas line and turbine meters on the oil and water lines (Fig. 3).
The accuracy of the liquid reference meters is estimated to be within +/- 2% of reading and the gas reference within +/- 5%.
Test separator readings at 10 and 30 min intervals were totaled to obtain the production rates for oil, water, and gas. From these rates, average oil, water, and gas volume fractions and flow velocities were calculated for each test period. These were compared to the meter readings.
CALIBRATION
The meter was calibrated in about 10 min. First, the gamma densitometer was calibrated in air at atmospheric conditions prior to pressurizing the bypass loop.
Second, the estimated oil and gas densities at process conditions were keyed into the instrument. The values used were 815 and 53 kg/cu m, respectively.
Finally, the approximate density of the water (980 kg/cu m) and its conductivity (60 millisiemens/cm) were keyed into the meter. This completed the calibration process.
No specific on site dielectric or density measurements of the constituent oil, water, and gas components were necessary. Moreover, no attempt was made to differentiate between the oil, water, and gas in the different wells tested. The same values were used throughout.
The simple calibration process is one of the attractive features of the meter.
Results from the meter were logged every 10 sec, although its actual data output rate was more than once per second. The values were integrated for 10 and 30 min intervals timed to match the measurement intervals from the test separator. These data sets were compared to determine meter performance.
TEST RESULTS
The meter was tested during a 3-week period in December 1992. The second week of tests had to be scrapped because the test separator was out of commission while one of the wells was being treated.
During the first week, a static mixer was placed upstream of the meter. After this, the mixer was removed with no apparent effect on the results.
COMPOSITION METER
The composition meter performed exceptionally well during the test.
Table 2 shows the results for two test periods. The meter and test separator data are in percent by volume for each component (oil, water, and gas) at process conditions.
In addition, the volume percent hydrocarbon is also displayed. This is equal to the sum of the oil and gas. This value is considered more valuable by many field engineers for high pressure and temperature fluids.
The percent error represents the difference in absolute percentage between the meter and the test separator readings.
During the first test period, the meter was on three wells. One of the wells was producing about 15% by volume water (0.33 water/oil ratio). The other two wells produced very little water.
For wells producing less than 1% water, the water leg of the test separator was closed. Therefore, the amount of water in the mixture for Wells B10 and B20 was not measured with the test separator, but was known to be less than 1%.
The composition meter was accurate to within +/- 3% for all components. The figures for percent hydrocarbon and percent water were within +/- 1 %. While these results are quite good, they are more than the +/- 2% specified.
At the conclusion of the first test period, the manufacturer found two small errors in the measurement algorithms. These were corrected before the second test period. The mixer was also removed.
The results for the four wells in the second test period are even better. Well B11 had a water continuous phase; therefore, the meter did not function. For the remainder of the wells, the meter was essentially within +/- 1% for all components. This is considerably better than specified.
Well B20 was tested during both periods and illustrates the improvement in performance attributed to the elimination of the errors in the algorithms.
While Table 2 shows the performance of the composition meter over long periods, it does not illustrate the relative quality of the meter's performance for "instantaneous" measurements with respect to the test separator.
The test separator required about 30 min to stabilize after the test well was changed. The composition meter, on the other hand, reacted instantly to the input and demonstrated that the new well needed from 5-6 min to stabilize.
Moreover, the composition meter readings changed very little during the course of a well test. They were generally stable to better than 0.1% over time spans of 20 sec.
Fig. 4 shows the composition meter readings and the measured values from the test separator at 10 min intervals (the shortest available test period with the test separator) for Well B21. It became quite clear during the test that the composition meter gave very accurate, reliable, and fast trend information about the production of individual wells.
The manufacturer's specification of +/- 2% accuracy seems supported by the results.
VELOCITY METER
The results for the velocity meter are shown in Table 3. In general, the results are inferior to the composition meter, but are still encouraging.
The data in Table 3 show that the velocity meter is reading consistently high in all tests and the results are essentially unchanged by the presence of the static mixer.
In all tests but that of Well B11, the meter seems to be reading consistently high by 10-12%. A simple reduction of the readings by 10% would put the velocity meter within its specified accuracy of +/- 5%. The reason for this offset has not been determined.
As with the composition meter, the velocity meter proved to be more reliable in the short term than the test separator. Fig. 5 shows the velocities measured on Well B21 by the velocity meter over 10 min intervals as compared to the test separator. Overall, the measurement trend is the same for each.
PRODUCTION RATES
When the results of the composition and velocity meters are combined to calculate production rates in terms of produced volume per unit time for each component, the results were consistently high by about 10%, +/- 3%. The error is directly attributable to the high velocity readings.
Target accuracy for the first generation of multiphase meters is +/- 10% for allocation purposes. This meter appears capable of achieving this target or exceeding it if the velocity meter can be improved.
FUTURE PLANS
More tests will be carried out in 1993 at Gullfaks, and other tests are planned. Among the more interesting investigations will be to test the meter as a mass flow device. When operated as a mass flow meter, the investigators believe that the meter will be more accurate and less sensitive to calibration parameters.
In addition, it should be easier to convert measured mass data at process conditions to volume or mass information at standard conditions. To perform these tests, mass flow meters will be added to the test separator at Gullfaks B.
Multi-Fluid, together with Saga, Statoil, and the other companies named subsequently agree that this first meter is application limited. Many applications for multiphase meters will require the capability of measuring higher water contents.
Therefore, Statoil and Saga--together with British Petroleum Norway A.S., Elf Aquitaine Norge A.S., Phillips Petroleum Co. Norway, and Total Norge A.S.--are sponsoring Multi-Fluid to develop a second multiphase meter capable of measuring 0-100% water.
This second multiphase meter is based on very similar technical principles as used by the first meter. It will be ready for field trials later this year.
In spite of the limitations, this first meter marks a significant step forward in multiphase measurement technology.
It is becoming increasingly clear that no single instrument will be capable of measuring all of the different flow regimes with the accuracy and reliability required by users. Instead, different meters will serve the different applications.
The tested meter should be well-suited to measure new field production where water content is relatively low and test-separator liquid output where free-gas breakout is a problem.
Copyright 1993 Oil & Gas Journal. All Rights Reserved.
