SPE

SPE 24005

Automatic Well Testing Using a Two-Phase Separator

Gary Fransen, Agar Corp.
SPE Member

Copyright 1992, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the 1992 SPE Permian Basin Oil and Gas Recovery Conference held in Midland, Texas. March 18-20, 1992.

This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author@). Contents of the paper,
as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author@). The material, as presented, does not necessarily reflect
any positionof the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publicationreview by Editorial Committees of the Society
of Petroleum Engineers. Permissionto copy is restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain conspicuousacknowledgment
of where and by whom the paper is presented. Write Librarian Manager, SPE, P.O. Box 833836, Richardson, TX 750833836. Telex, 730989 SPEDAL.

Abstract

The traditional methods of testing the production of Well testing consists of isolating the production
oil wells have involved the task of separating the stream of a selected well and analyzing the content of
produced stream into its components; oil, water, and that stream under conditions that approximate those of
gas. However, development of an energy absorption
device for measurement of oillwater mixtures in any normal production as closely as possible. The
ratio has eliminated the need for three-phase production parameters typically measured include the
separation. This technology allows the use of a much following:
simpler, less expensive and more accurate system:
the two-phase well tester. Gas Flow
Gross Production Rate
Introduction Net Oil Production
Net Water Production
The testing of oil wells has always been an important
part of oilfield operations. The results derived from The net oil production figure is generally regarded as
such testing are critical to forecasting revenues, the most important of these measurements,
optimization of recovery methods and efficient particularly in cases where the gas or water flows are
reservoir management. They are further applied to insignificant or very nearly constant. Also, the net
evaluate taxes and to determine the allocation of oil production is commonly the basis for taxes and the
revenues based on lease ownership and, in some primary source of profit to the producing company.
situations, custody transfer. Because this information The primary focus of this discussion will therefore be
directly impacts the profitability of operations, it on the best and most accurate means to determine the
becomes vitally important that the testing methods net oil and water in the produced stream. In that
yield the most accurate results possible. pursuit, we will examine the design and function of
the traditional well testing system (the three-phase
method), and the present methods employed to
References and illustrations at end o f paper. overcome its shortcomings. Secondly, the design and
functional advantages of the two-phase system will be calculate the net amount of oil and water in the
presented as a preferred alternative both in terms of oillemulsion. When the level falls below the "LO"
test accuracy and operating cost. point, "LC 1" closes the dump valve.

Methods of Well Testing When the oil-water interface reaches the "HI" point
set by the water level controller, "LC2" the water
The producing streams from an area are typically level control valve, "WCV" opens. If proper
routed to centralized manifolds where they are separation has taken place, oil-free water is measured
combined into a group or "family" line for transfer to through a separate flow meter as it is dumped. The
production separators (tanks or vessels) and on to water is dumped in batches just like the oillemulsion
further treatment. These manifold settings generally to maintain relatively constant flow rates through the
have a test separation tank or vessel into which each flow meter. This free water is added to the water
individual well can be routed and its production totals from the oillemulsion leg to determine net
evaluated. (See Fig. 1.) water produced by the well. When the oil-water
interface reached the "LO" point, the water dump
Traditional well test methods involve separation of the valve closes.
well stream into three phases: gas, oil (emulsion),
and water. The gas is liberated via flash pressure The pressure regulator "PCV" holds a back pressure
drop in the vessel and is vented overhead to a on the tester by venting gas that has separated from
metering run. This isolates the liquid phase, allowing the liquids. With the separator operating at near
for measurement of its rate and oillwater content. constant pressure, the liquid flow rate during dumps
is held relatively constant increasing liquid
In some instances, this liquid phase is then manually measurement accuracy through the flow meter. The
sampled and tested by centrifuge. In automatic gas flow is calculated and recorded using standard
systems, the liquid phase is allowed to separate into industry methods.
water and emulsion layers, with the emulsion phase
typically being floated or "pushed" over a weir into The raw measurements at the test separator are totals
a separate compartment (oil leg). Each phase is then of oil, gas and water. To convert this into usable test
metered for volume with the emulsion phase being data, the tests are carefully timed and typically
tested either manually or continuously via some type converted to daily production rates.
of percent water measurement device. (See Fig. 2.)
Fundamental Accuracy Problem
The oillemulsion level in the separator is controlled
by the oil level controller, "LC 1" . When the liquid Even in instances where the three-phase system
behind the weir reaches the "HI" point, the oil leg performs exactly according to specification (see below
control valve, "OCV" snaps open, dumping oil. The for factors affecting this), there is a fundamental
oillemulsion discharge is typically metered by a problem associated with its use.
turbine or positive displacement flow meter. Wells
that produce at low rates will require fewer dumps Assume the well feeding into a three-phase separator
per hour than higher producers. Because the oil and (Well 'A') separates into gas, free water, and a 20%
water cannot be completely separated in the well water in oil emulsion. After a specified test time, the
tester, the water content of the oillemulsion is next well (Well 'BY)is switched into the separator.
measured with an oillwater monitor. Traditionally This new well separates into gas, free water, and a
these oillwater monitors have operated on the 10% water in oil emulsion.
principles of capacitance or conductivity. The outputs
from both the flow meter and the oil/water monitor The free water from Well 'B' will drop through the
on the oil leg, are combined in a microprocessor to
emulsion left in the separator from Well 'A', and be and 100% water in the water continuous phase.
measured through the water leg flow meter. Since its
emulsion component is drier than that left in the 2. Oil in the Water Phase - Here again, poor
separator from Well 'A', the emulsion from Well 'B' oiliwater separation results in mixing of the phases.
will initially tend to float on top of that from Well It is assumed that all produced fluid dump out of the
'A'. water leg is 100% water and that hydrocarbon under
carry does not exist. This can be particularly
However, with time the cohesive nature of emulsions troublesome in areas where produced crude gravity
will cause Well 'B's emulsion to mix with the (API) is low. In any case, oil will be dumped with
emulsion from Well 'A', and this mixture will be the water, and with no monitoring device on the
drawn into the oil leg. As a result, the AWT will water leg, the test accuracy will suffer accordingly.
register an oil content for Well 'B' that is actually a This is a particularly common problem and very
mixture from the two wells. troublesome because the operator will rarely be aware
that it is occurring as the streams are recombined and
Since the second emulsion is drier than the first (and fed to the group line.
displacement does not occur), it is impossible to
project a suitable purge time with any accuracy. Improving the Se~aration

Factors affect in^ Net Oil Measurement A great deal of effort and expense has been focused
on various means to achieve the oiliwater separation
The quality of the oillwater separation is a critical process. The most common can be summarized as
factor in the successful (i.e., accurate) application of follows:
three-phase well testers. When this separation is
irregular or incomplete, the interface controller will Heating of the Produced Stream
be less effective, and the test results can suffer both Addition of Emulsion-Breaking Chemicals
in terms of absolute accuracy and repeatability. The Increasing Vessel Residence Time
problems that arise from poor separation are quite
straightforward and can be summarized as follows: 1. Heating of the well stream is a common method
of enhancing separation. The resultant reduction in
1. Excess Water in the Oil Leg - The probes utilized the viscosity and specific gravity of the oil allows for
to monitor the percent water (water cut) of the faster coalescence and settling of water droplets.
emulsion stream typically use capacitance or However, there are several drawbacks to this method:
conductivity as a basis for measurement. These
methods rely on the assumption that the measured a. Significant expense involved in installation and
stream will be non-conductive. When the emulsion is maintenance of heat exchange equipment.
oil-continuous (water droplets suspended and b. Poor control and measurement accuracy resulting
surrounded by oil), this technology can provide from inability of well tester instrumentation to
reasonably accurate results. However, when the compensate for higher temperatures.
emulsion contains a high percentage of water, the c. The BTU cost of energy expended in heating the
fluid exiting the oil leg can contain both water- production streams from many wells to many testers.
continuous emulsion (oil surrounded by water) and
free water. Both of these constituents are highly Based on these factors, adding heat is potentially
conductive, and result in saturation of the oillwater helpful, but rarely cost effective or a functional
probe. This "blinds" the probe to the actual oillwater solution to good separation.
ratio present and results in a significant loss in
measurement accuracy. Traditional capacitance 2. The addition of demulsifying chemicals serves to
probes can not distinguish between 50 % , 70 % , 90 % disrupt the chemical and physical agents that stabilize
the interface between oil and water. This also allows rate at 'which they absorb high frequency
for improvement in water droplet coalescence. electromagnetic energy. With the individual
However, the great expense of such chemicals will absorption rate for each phase known (establishing
add significantly to the operating costs (and therefore absolute limits), any measured rate between those
lift costs) of any production area. Further, the limits will correspond to an oillwater mixture of
efficiency of such chemicals is strongly dependent known content.
upon dosage. It is difficult to maintain the optimum
concentration of chemicals when production rates and With this new technology, the AWT system can be
water cuts vary from well to well. In some cases, the down sized to a much smaller "two-phase" vessel for
overfeed of chemicals will actually reduce the gaslliquid separation. This system would also greatly
oillwater separation process of the emulsion, resulting simplify both the equipment and maintenance needs
in further loss of accuracy. (and associated costs inherent in its operation.

3. Increased residence time for the liquid phases can The Function and Design of the Two-Phase Well
only be accomplished by either altering the operating Tester
level in the vessel (requiring extensive internal work)
or by outright replacement of the vessel with a larger The use of two-phase separators in no way changes
one. The alteration of the level will frequently result the existing configuration of the production manifold
in the entrainment of liquid in the gas phase. The use or test lines. A two-phase separator simply takes the
of large vessels is expensive, and significantly isolated well stream, and functions as a gas
increases the time required to purge the vessel of the eliminator. Further, it requires no internal weir,
previous well's fluid prior to the next test. baffles or oillwater interface controllers. (See Fig.
3.)
In summary, the above-listed methods of improving
the oillwater separation can be helpful, but in no way The total liquid level is controlled by "LC1" which
provide a conclusive or cost effective solution to the opens the level control valve "LCV" when the liquid
problem. builds up the "HI" point and closes the valve when
the level falls to the "LO" point. While the separator
Eliminating the Need for Separation is filling, the oil and water will separate to some
extent. During the dump, the 0 to 100% oil-water
Many of the problems associated with obtaining monitor will see free water, water-continuous
accurate well test results from three-phase AWT's can emulsion and oil-continuous emulsions. Despite the
be attributed to the oillwater separation process. presence of a water-continuous emulsion, the oil-
Therefore, automatic well testing can best be water monitor will accurately measure the oil content
improved by eliminating the need to perform of the fluid throughout the entire dump. The signal
oillwater separation. from the oillwater monitor is combined with that
from the flow meter in a totalizer to calculate net oil,
If the oil and the water are not separated, the AWT net water, and total produced fluid.
system must use an oillwater probe capable of
measuring emulsion from 0-100% in both the oil Gas control and measurement are accomplished in the
continuous and water continuous phases. same manner as in a three-phase separator. Again,
totals of oil, water and gas are multiplied by the test
The probe utilized to accomplish this 0 to 100% time to calculate daily production rates.
measurement, functions on the principle of energy
absorption, thereby overcoming the limitations of The key to the success of this simple and efficient
capacitancelconductivity technology. It can means of testing wells is its accuracy. Simplifying
differentiate between oil and water by measuring the the design, reducing the size of the vessel, and
eliminating the separation process allows for a much royalties from leases sharing common treatment
better purge of production from previous tests. This facilities. Although state law required that the
improvement in purge efficiency eliminates the production of each individual lease be monitored by
ljrocess of feeding a new well into a system tank testing only, an evaluation of the two-phase
containing "old" fluid, a significant source of error in system (with 0 to 100% probe) proved to provide
three-phase testing. equal or better accuracy. The system was approved
for use as a formal exception to the state law.
The Typical Two-Phase Installation
Data from this and other testing has been included
While the two-phase tester can be applied in virtually with this paper. (See Table 1.) This data was
any type of production environment, certain common compiled in testing on a variety of systems handling
applications can best display its benefits. crude from 13" to 36" API and temperatures from
100" to 400" + F. (See Appendix.)
A typical two-phase well tester would be in a mature
oil field undergoing water or steam flood. The Advantages of the Two-Phase Design
conventional three phase tester was designed before
the waterlsteam flood. During three phase operation, Unlike the three-phase system, the accuracy of the
the increased water flow does not allow sufficient two-phase design is strictly a function of the quality
residence time for the oil and water to separate. The of instrumentation utilized for flow and net oillwater
high water content in the separator's "oil" stream measurements. However, the advantages derived
(above 50%) is above the working range of the from its use are not limited to the performance of a
capacitance probe traditionally used to determine 0 to 100% oillwater monitor. It is, in fact, the
water cuts. In addition, oil emulsions are often simplification of the process that yields the greatest
dumped with the free water. These problems lead to improvements.
errors in the calculation of net oil produced.
The two-phase well tester is not dependent on or
This three phase separator (See Fig. 4) has been affected by the quality of the oillwater separation.
converted to a two phase separator by closing off the For this reason, a much smaller vessel can be
oil leg (Point A) and disabling or removing the weir employed with none of the added costs associated
(Point B). The previous oil level control "LC1" is with augmenting the separation process. This in turn
now controlling the liquid dump valve, "LCV". shortens the time required to test any given well
accurately, (particularly low gross wells).
A 0 to 100% oil-water monitor has been installed at
Point C upstream of the flow meter. The flow meter The shorter test time is further improved by a
and the oil-water monitor are connected to the net oil significant reduction in the time required to purge the
meter, which combines the signals to calculate net oil fluid of the previous test from the vessel (a common
and net water totals. The monitor is in a vertical line source of inaccuracy in three-phase testing).
so that the oil-water emulsion is uniform at the cross-
section of the flow stream. A minimum constant In addition to the improvements in accuracy and test
fluid velocity of 5 ftlsec is maintained through the time, the two-phase design reduces operating costs.
oillwater monitor by dumping the liquid in batches. With smaller vessels (requiring no internals for
This also keeps the flow rates within the design range separation) and only half the basic instrumentation
of the flow meter for maximum accuracy. requirements, the cost of new systems drops
significantly. Both new and converted systems
One excellent example of the capabilities of this benefit from a reduction in maintenance costs, as well
system involved a large West Texas producer who as the elimination of the above-mentioned costs of
was faced with the problem of allocating production augmenting the oillwater separation process.
The advantages of the two-phase system are
summarized in the Appendix of this report.

Proving the Well Test

A wide variety of methods are typically applied to

prove the accuracy of well test results. (See Table 2.)
It is interesting to note that most of these methods are
quite prone to error by their very nature. This is
particularly true of the manual and composite
sampling methods.

Obviously it is necessary to verify the results obtained

from any system, and in this pursuit the use of
atmospheric storage or test tanks is strongly
recommended. This provides the best means of
"sampling" the entire volume of fluid tested, and can
usually be readily evaluated given sufficient settling
time.

For those systems where tank testing is not an option,

the alternative method utilized should be conducted
carefully and rigorously, with repetitive testing to
verify repeatability.

Summarv

Major reservoir/production decisions are based on

well test data. The accuracy of these well tests is of
critical importance. Conventional three-phase testers
have been the traditional choice based on the
limitations in the technology of measuring the water
content of the fluid stream. The development of an
accurate 0 to 100% oil water monitor has now
allowed for replacement of the three-phase system
with one not dependent on the separation of the oil
and water phases. The two-phase well tester offers
proven accuracy in a simple design, while greatly
reducing the equipment and maintenance costs
associated with well testing. Either in new systems
or in conversion of existing vessels, it now provides
an efficient solution to accurate determination of net
oil and water production.
 TABLE I

TEST RESULTS UTILIZING A TWO-PHASE SYSTEM

WEST TEXAS*
% OIL

TOTAL % OIL 0 T O 100%

APPENDIX DATE PROWCTION T E S T TANK NET O I L EM
S'Y
I?!

02-08
02-09
SUMMARY OF ADVANTAGES OF WELL TESTING WITH 02-10
2-PHASE V. 3-PHASE SEPARATOR 02-1 1
02-12
02-13
WELL TESTING RATE 2-PHASE 3-PHASE 02-14
02-15
Retention Time: Short. Only need to separate liquid Long:necessary tobmk~rrmlaions. 02-16
from gas. 02-17

Purging Time: Short. Negligible dead volumes. Long, due to dead volumes; varies
with em~laiiontype well flaw W. K E R N COUNTY, CALIFORNIA

04-2 1
ACCURACY 04-22
04-23
Purging: Smsller dead volumes reduce Fluid from previous well can greatly 04-24
error. affect results if not totally purged. 05-05
Emulsions: 05-06
Measures from 0 to 100% water Treats all water phase emulsions as
100% water. 05-08
in oil. 05-09
05-10
Inversions: Does handle density inversions. Can not hmdle density inversions. 05-13
05-13
Inmmmentation: Accurate over full range of E m due to number of instruments.
0 to 100% water cuts.
ALBERTA, CANADA

ECONOMICS "Pro-ration values of monthly plant bitumen [avg. 8" API gravity crude] volumes to well test bitumen
volumes, formerly averaging 1.5 (ranging between 0.3 and 2.5), improved to 1.05 (ranging between 0.9
Number of Instruments: I flow meter, 1 set of level Requim2 flow mkrs, 2 seLF oflevel and 1.3) after the use of net oil monitors was adapted." - From Producer's paper submission for
controls and I control valve. wntrols and 2 control valves. CIMIAOSTRA technical conference.
Vessel Size: Smaller with equal capacity. Large because of Retention Time.
(Specific well test results were not made available.)
Piping. Only I leg. Complex: both water leg md oil leg.
* TESTING PERFORMED FOR THE TEXAS RAILROAD COMMISSION
Maintenance: Much simpler system. Twice the number of instruments.

Wells per Vessel per Day: Many more. Few (see Well Testing Rate).
TABLE 2
ChemicalsIHeating: No emulsion breaking needed. Ennulsionbteakingcsnbeverymutly.
METHODS FOR PROVING WELL TESTERS

* HISTORICAL RECORDS

* RANDOM SAMPLE AND FLOW METER

* COMPOSITE SAMPLE AND FLOW METER

* SECOND WELL TESTER IN SERIES

* ATMOSPHERIC STORAGE TANK

 GAS

w NET OIL METER

A WELL
'STREAM

ylDshvbow
METER
OI L-WATER
TMoNlToR

FIGURE 3. 2-PHASE WELL TESTER

FIGURE I. GATHERING STATION

NET OIL METER

I I 1
GAsT
,---I--+---I--
I I
I I I
I I I
OIL WELL
&REAM
F ~ WI
I
,-----
METER
+----

FLOW
METER

FIGURE 2. 3-PHASE WELL TESTER

FIGURE 4. TYPICAL 2-PHASE TESTER