CORIOLIS-BASED NET OIL COMPUTERS GAIN ACCEPTANCE AT THE WELLHEAD

K.T. Liu
KTL & Associates
Cerritos, Calif.
G. E. Kouba
Chevron Petroleum Technology Co.
La Habra Calif.

Full-range, water-cut capability and a nonintrusive nature enable the Coriolis-based net oil computer (NOC) to meter produced fluids directly from the wellhead.

Several hundred NOCs are currently installed at individual wellheads on low-GOR wells. A majority of the wells have wellhead pressures greater than the bubble point pressure of the crude oil. Because no entrained gas is present in the liquid stream, a test separator is unnecessary.

SYSTEM BENEFITS

The Coriolis-based NOCs are cost effective, especially in high water-cut applications. Field experience shows the units are rugged, reliable, and accurate when properly installed and operated.

The NOC system continues to evolve with improvements in implementation procedures and techniques to enhance operation and expand application opportunities.

Increasing numbers of Coriolis-based NOCs are used simultaneously to measure crude and water production from individual wells or leases. In this NOC system, the Coriolis force flowmeter (CFF) serves as both water-cut analyzer and flowmeter.

Compared to other conventional measurement systems, this technology offers the following advantages:

Measures full-range water cut
Reduces repair frequency and test cycle time. This lessens operating and maintenance costs
Lowers overall capital costs.

Since introduction in 1988, more than 1,200 Coriolis-based NOCs have been installed for commercial operation worldwide. About 700 NOCs are in tandem with test separators. The remaining NOCs are at the wellhead and monitor well production on a real time, continuous basis.

The Coriolis-based NOC system feasibility has been demonstrated and described elsewhere.1 The inherent simplicity of NOC systems, however, has in some instances led to overlooking fundamental principles required for successful applications.

APPLICATIONS

Measurement of crude oil produced from individual wells or leases is important, often essential in oil field operations. Accurate and timely information ensures prompt operating decisions, equitable royalty distribution to interest owners, and efficient reservoir management. But measurement accuracy is often in doubt.

Problems may stem from several possible shortcomings of conventional well testing systems, such as inefficient separation, poor sampling, rangeability of flowmeters, failure to maintain instrument calibration, and lack of appropriate measurement technology.

Conventional oil and water flow measurement from a producing well or a group of wells involves a two-phase or three-phase separator to remove produced gas. Free water and the emulsion flow rates are measured with respective flowmeters.

Emulsion stream water cut is determined by an on-line sampler, a grab sample from the emulsion flow line, or an on-line capacitance-type water cut analyzer. Combined signals from the flowmeters and water cut analyzer then determine oil and water production.

One major measurement error is attributed to water cut determination, especially when water cut is relatively high. High water cut can occur in the following situations:

Maturing reservoir
Small test separator for high production wells
Two-phase separator
Tight emulsion that prevents separation of free water.

On-line and grab samples may not represent total flow because crude and water usually are not uniformly distributed in the pipe, and water cut will generally fluctuate with time. Consequently, systems that rely on sampling techniques for water cut determination can suffer significant measurement errors.

Analyzers based on capacitance measurement principles are reasonably accurate only when water cut is relatively low and the oil/water emulsion is oil-continuous (i.e., water droplets in oil-continuous phase). Because these analyzers depend on dielectric properties of the emulsion, erroneous measurements occur at high water cuts when the emulsion becomes water continuous (i.e., oil droplets in the water-continuous phase).

No distinctive transition region exists in which an oil continuous emulsion changes completely to a water-continuous emulsion. In most cases, the highest water cut of the oil continuous emulsion occurs at about 3560%.

Recently, several new water cut analyzers have been developed to measure full water-cut streams. These include those based on microwave and radio frequency techniques. In the authors' opinion, however, none of these analyzers has yet fully demonstrated its claimed capabilities.

Another difficulty pertains to flowmeter reliability. Conventional well-testing systems typically employ a positive displacement meter or a turbine meter to measure various liquid streams.

Because these flowmeters have rotating parts, they are susceptible to wear and clogging caused by sand or paraffins in the production stream. Consequently, these flowmeters frequently need repair and recalibration, and the measurement becomes unreliable in the interim.

The Coriolis NOC technology was developed to alleviate problems with high water-cut measurement and flowmeter wear.

CORIOLIS-BASED NOC

For net-oil measurement applications, the Coriolis force flowmeter (CFF) simultaneously serves as a flow-meter and a density meter.

In brief, a typical CFF shown in Fig. 1 has two identical tubes that are vibrated in opposition at their natural frequency by an electromagnetic drive mechanism. Because of the Coriolis effect, the fluid flowing through the vibrating tubes creates an asymmetric distortion between the inlet and outlet legs.

The distortion magnitude, measured by two position detectors placed on opposite tube legs, is directly proportional to the mass flow rate.

Besides mass flow rate, the fluid density can also be measured by the change in vibrating frequency of the meter tubes. A resistance-type temperature sensor continuously monitors the meter tube temperature for various signal processing purposes.

Many CFFs provide mass flow and density, measurement accuracies of 0.15% and 0.0005 g/cc, respectively. Details of the operating principle of CFF 2 3 are described elsewhere.23

From the basic outputs of a CFF, water cut, net oil volume, and net water volume are calculated with the following equations:

Xe = (De-Do)/(Do-Do) (1)

Net oil volume = (Me,/De) X

(1-Xw) X Ctl,o (2)

Net water volume =

(Me/De) x Xw x Ctl,w (3)

where:

Xe = Water cut in emulsion, volume fraction

De = Emulsion density measured by the Coriolis meter

Do = Known, predetermined density of dry crude oil

Dw = Known, predetermined density of produced water

Me = Total emulsion mass measured by the Coriolis meter

Ctl,o = Crude-oil volume correction factor that adjusts measured crude volume to standard temperature.

Ctl,w = Produced-water volume correction factor that adjusts measured water volume to standard temperature.

Because water cut is based on density difference between oil and water, accurate measurements can be obtained over the full range of 0-10017( water cut, regardless of whether the emulsion is oil-continuous or water-continuous.

Also, because the Coriolis meter is a nonintrusive instrument and has no moving parts within the flow path, it requires significantly less maintenance than conventional positive displacement and turbine meters.

High rangeability of the Coriolis flowmeter (up to 80 to 1 turn-down ratio) allows the NOC to accommodate a wide range of flow rates from different wells. Also, its internal temperature sensor corrects net oil volume to standard temperature.

PERFORMANCE CHARACTERISTICS

Under normal operations, NOC accuracy is primarily governed by the accuracy of input dry oil and produced water densities, and the accuracy of the emulsion density measured by the Coriolis meter. Typical performance characteristics in terms of water cut and net oil determinations are illustrated in Fig. 2.

Note that these performance curves are generated on the basis that the uncertainties of dry oil density (Do), produced water density (Dw), and measured emulsion density (De) are all on the order of 0.5 kg/cu M. Field data show that these uncertainty levels are achievable with proper calibration.

Fig. 2 shows that the measurement uncertainties for both water cut and net oil decrease as the API gravity of the crude oil increases. it is intuitive that the lighter the crude oil, the larger the density difference between the oil and water, and therefore the better the measurement accuracy.

Fig. 2a also shows that, for a given API gravity of crude oil, the uncertainty of the water cut measurement is relatively insensitive to the water cut level. With a 26 API oil, the water cut uncertainty ranges from 0.5% to 0.6% over the entire water cut range.

The corresponding uncertainty in net oil volume, however, depends on both the crude's API gravity and also on the instantaneous water cut of the emulsion. As shown in Fig. 2b, the uncertainty is relatively small in the low water cut range and is only slightly higher in the mid water cut range, up to 80%.

As the water cut increases further, however, the uncertainty increases sharply. Mathematics explain this behavior.

In Table 1, the net oil volume uncertainties resulting from inaccurate water density and mixture density are proportional to the inverse of (1 - Xw). This implies that the overall net oil uncertainty will increase exponentially at very high water cuts.

Again with 26 API crude, Fig. 2b shows that the net oil uncertainty, increases from 0.6% to about 3% between 0 and 80% water cut. It then increases to about 6% at 90% water cut and escalates to about 12% at 95% water cut.

It should be noted that water-cut analyzers will also exhibit the same behavior with high water-cut streams.

To obtain the best performance from the Coriolis-based NOC, it is imperative that:

Dry oil and produced-water densities at operating pressure must be predetermined accurately.
Dry oil and produced-water densities must be consistent over time.
Emulsion streams must contain no free gas
Direct measurement of extremely high water-cut streams should be avoided or minimized.

ACCURATE DENSITIES

At metering conditions, crude and produced-water densities from individual wells or leases must be predetermined and programmed into the NOC. The "live oil" density rather than the "dead oil" density must be used as the input parameter.

Live oil in this context refers to the crude that is saturated with solution gas at the separator pressure and temperature. Reduction of liquid pressure to stock-tank pressure will cause the live oil to lose its solution gas or light-end components and become a dead oil.

Depending on GOR and separator pressure, the live and dead oil densities can be very different. Water cuts will read too low and net oil volume too high if the dead-oil density is used.

For example, for a well producing a 26 API oil at 50% water cut, if the dead-oil density is 10 kg/cu m higher than the live oil density, water cut would read about 4% too low and net oil volume would read about 8% too high as a result of wrong crude-oil-density input. Table 1 lists ways to estimate the sensitivity of various parameters on water cut and net oil measurements.

Three methods are commonly being used to determine the live oil density accurately.

The first involves sampling liquid, with a pressurized bottle, near or at the separator. The sample is processed in the laboratory to remove free water while maintaining the sample slightly above the separator pressure to prevent the solution gas from flashing off. The dry crude oil is then measured with a high-precision laboratory density meter, capable of measurement at elevated pressures.

The second method uses a Coriolis meter as a density measurement device. It involves blocking off the liquid inlet and outlet of the test separator for an extended time to allow free water and the emulsion to separate in the separator. The emulsion, which may still contain water, is then flowed through the Coriolis meter, and density and temperature readings are recorded.

Meanwhile, an emulsion sample is withdrawn from the flow line and its water content determined. The dry crude-oil density can then be computed from the density, temperature, and water content of the emulsion.

The third method uses a sampling and measuring assembly (Fig. 3).

Typical operating procedure involves withdrawing a liquid sample into the small cylindrical sample container and allowing the sample to stand an extended period while the sample container is connected to the separator gas line.

Similar to the second method, the produced water and crude densities can be determined by flowing water and dry emulsion through the density meter. The main advantages of this approach are relative cost, portability, and local processing of the samples so that oil and water densities can be quickly updated with relative ease.

GAS EFFECT

Many Coriolis meters tolerate an amount of entrained gas. A small amount of gas may not affect mass measurement accuracy. Depending on meter design and size, the gas tolerance limits can range from several percent to 10-20% by Volume, especially when gas bubbles are well dispersed in the liquid stream.

The entrained gas, however, will certainly affect the water cut determination of the NOC. Entrained gas will cause a decrease in density reading, which is misinterpreted as a decrease in water cut.

For example, a 0.5% by volume of entrained gas will underestimate water cut by as much as 4.3% for a 26 API crude.

There are two ways in which entrained gas may be present at the Coriolis NOC:

Gas carry-under with the liquid stream
Gas evolving out of solution in the crude oil.

Gas carry-under may be caused by high liquid viscosity, improper separator operation, or poor separator design. Problems with separator operation and design are beyond the scope of this discussion but must be corrected for proper operation of the NOC.

A pressure reduction may cause gas to flash out of solution as crude flows from the separator to the NOC. The amount of evolved gas depends on crude properties, operating temperature and pressure, and pressure drop.

Most test separators are designed to remove practically all entrained gas from the liquid. In some production situations, however, extremely tight emulsions can occur in the production line and test separator. Tight emulsions are characterized by unusually high viscosities, sometimes many times more than the crude oil viscosity.

Separating the entrained gas from this emulsion is not efficient even with a large test separator. Demulsifier chemicals that break the tight emulsion and facilitate free gas removal, however, have been found to be effective and practical.

One way to minimize or prevent the solution gas from flashing off is to install the Coriolis meter below the separator. In a properly designed system, the static head gain effectively offsets the dynamic pressure losses in the flow line. This results in a higher pressure at the meter than at the separator, thus preventing solution gas from flashing. The design criterion is expressed as:

_Ps _Pp + _Pm

where:

Ps = Liquid static head measured from the liquid level in the separator to the flowmeter

Pp = Dynamic pressure loss in flow line and fittings from the test separator to the flowmeter inlet

Pm = Pressure drop across the flowmeter

The frictional pressure loss (PP) can be minimized by installing the Coriolis sensor as close to the test separator as practical and using larger-diameter connecting pipes. Piping elements, such as tees, elbows, and reducing unions, should also be minimized between separator and meter. Sampling ports, static mixer, meter proving connections, dump valve, back pressure regulator, or other flow-restricting devices should be installed downstream of the flowmeter.

If a cut-off valve must be installed between the separator and the flowmeter, a full-port valve should be considered.

To minimize the pressure drop across the flowmeter, (Pm), it may be necessary to use a larger flowmeter than normally needed. Because of the high rangeability of the Coriolis flowmeter, flow measurement accuracy can still be achieved even at very low flow rates.

The minimum vertical height between the Coriolis meter and the liquid level in the test separator can now be estimated from:

Hmin = 2.3 x (Pp + Pm)/Do

where:

Hmin is in ft

Pp and Pm are in psi

Do is in g/cc

To illustrate this design, consider an application in which a 3-in. Coriolis flowmeter is installed at the separator outlet. Assuming a maximum fluid flow rate of 6,000 b/d and fluid viscosity of 1 cp, the pressure drop across the Coriolis flow meter (Pm) is calculated to be 1.1 psi.

It is further assumed that the total equivalent length of the pressure drops through the 3-in. diameter connecting pipe and other piping elements (elbows, valve, etc.) is 20 ft.

The piping pressure (Pp) drop is calculated to be 0.7 psi. Therefore, the required static pressure head (Ps) must not be less than 1.8 psi. For a 26 API oil (Do = 0.8975 g/cc), the Coriolis flowmeter must be installed at least 4.6 ft below the liquid level in the separator to prevent the solution gas from flashing off in the meter.

Another method to prevent gas flashing raises the liquid pressure above the separator pressure by installing a charge pump upstream of the flowmeter. Although a limited approach, it effectively boosts pressure and has been successful in several installations.

FLUID DENSITIES

The NOC requires that oil and water densities in the product stream not vary significantly with time. For an individual well, this is generally the case.

The possibility of varying oil and water densities exists for wells with multizone completion and for wells undergoing initial stages of enhanced oil recovery (EOR). Again, the potential effect needs to be evaluated prior to installing the NOC.

In theory, the NOC may be unsuitable for measuring commingled production. Changing production rates from individual wells may cause the oil and water densities in the commingled stream to vary.

In practice, however, no appreciable density variations have been reported, to date, in all applications of this type. Nevertheless, it is advisable that the potential effect of density variations be evaluated against the expected performance.

When the NOC is used in commingled applications, more frequent update and adjustment of oil and water densities may be required.

HIGH WATER CUT

To reduce measurement uncertainty with very high water cut (e.g., 90% or higher), it is imperative to keep instantaneous water cut as low as practical when the emulsion passes through the meter. One effective method uses a three-phase separator to remove a portion of the free water from the bulk production stream, thereby keeping the water cut in the emulsion stream at a lower level.

If a two-phase separator is used, a snap-acting control mechanism has been found to be very effective. This control mechanism allows some free water to settle at the bottom of the separator during the standing cycle.

At the beginning of the dumping cycle, clean produced water (i.e., 100% water cut) flows through the meter. Then an oil/water emulsion stream follows that has a significantly lower water cut than the overall value.

This operating scheme also is ideal for determining or updating the produced water density. Because only clean produced water flows through the meter at the beginning of the dumping cycle, the density reading from the Coriolis meter can be used as the produced water density.

WELLHEAD APPLICATION

For wellhead applications, the typical crude oil GORs are less than about 100 scf/bbl (17.8 cu m/cu m), and the bubble point pressure ranges from about 100 to 250 psig (791 to 1,825 kPa). The gravity of the crude oil varies from 22 to 40 API with water cut ranging from 0% to as high as 95%.

These NOCs are installed on flowing wells, and wells equipped with electrical submersible pumps and beam pumps. Well production rates range from about 2,000 to 20,000 b/d (318-3,180 cu m/day) of gross fluid.

The applications generally require 3-in. or larger meters. Fig. 4 shows a typical installation. Because production wells are normally located in remote areas, these NOCs are typically linked to a central computed by means of radio telemetry.

To prevent solution gas from breaking out from the crude oil, it is critically important to ensure that the fluid pressure at the flowmeter is at least several psi above the bubble point pressure of the crude oil. Bubble point pressures can vary from well to well.

In addition to laboratory PVT analysis and flash calculations, using the installed NOC to determine the bubble point pressure has been found to be easy and practical. The method involves varying the well head pressure and observing the density indicated on the NOC. The bubble point corresponds to the pressure where the density (or the indicated water cut) reading on the NOC decreases abruptly.

Major users of this type of application include PDO (Petroleum Development Oman), OXY Colombia, CPI (Caltex Pacific Indonesia), and Saudi Aramco.

A wellhead NOC enables wells to be monitored on a continuous basis. Significant savings on capital investment and operating expenses have been demonstrated because individual flow lines and conventional well test facilities are eliminated.3

When wellhead pressures are maintained below the bubble point pressure, a certain amount of gas will be present in the production stream. The use of a Coriolis-based NOC in this situation requires some form of gas separation. Several users have successfully taken the approach of attaching the NOC downstream of a Simple and low cost in-line gas eliminator.

Fig. 5 shows a gas eliminator assembly that handles a production stream containing a small amount of free gas (up to about 3-5% by volume). The gas and liquid are separated in a vertical gas elimination pipe, and the liquid stream is measured with the NOC and recombined with the gas.

Currently, Chevron Corp. is developing a similar system aimed at handling a full range of multiphase flow streams. Fig. 6 shows the Chevron Multiphase Metering Loop (CMML) concept.

Separation in the gas eliminator is performed by a cylindrical cyclones The performance of the gas/liquid cyclone is further enhanced by initial separation in the downward-sloping tangential inlet pipe. After separation, the gas flow rate is measured by any appropriate conventional gas meter, and the oil and water rates are measured by the Coriolis NOC.

Initial laboratory and field prototypes have demonstrated that this system is capable of handling much higher GORs than a simple vertical separator.

REFERENCES

1. Liu, K.T., and Revus, D.E,, "Net-oil computer improves water-cut determination," OGJ, Dec. 19, 1988.

2. Spitzer, D.W., editor, Flow Measurement: Practical Guides for Measurement and Control, Instrument Society of America Publication, 1991.

3. Liptak, B.G., editor, Flow Measurement, Chilton, 1993.

4. Taylor, R.D.H., and Nuttall, R.C.H. "On-Line Well Monitoring and Its Application in a South Oman Oilfield," SPE Paper No. 25169, SPE Middle East Oil Technical Conference & Exhibition, Bahrain, Apr. 3-6 1993.