Well Production Testing Using a Rod Pump Controller

LeMoyne Boyer, Sam Gibbs, Ken Nolen, Andy Cordova and Alan Roberson
Lufkin Automation

Abstract
Well tests are crucial to managing rod pumped wells, and operators struggle to get tests as frequently as
they desire. A production test method that has proven to be accurate and reliable has been developed
using the downhole pump card generated by a rod pump [Link] method will be explained and
data will be presented of field tests showing actual well tests compared to the calculated well test from
the rod pump controller.

© 2006 Lufkin Automation.

For more information [Link] SAM is a trademark of Lufkin Automation.
(LUF617/300/0506)
Introduction tests, by as much as 10% or more. Guidelines for
testing include monthly testing, correction for
Traditional Production Testing recorded run times, periodic calibration of test
separators and meters, and increasing the test
Production testing has been an integral part of
period if possible for best results.2 Good sampling
oil producing operations for many [Link]
procedures are important on high water cut wells.
gathering of this data is required for many reasons.
Some reasons include governmental regulation,
environmental conservation concerns, reserve Diagnostic Methods
estimates, business purposes and well A traditional production test may give an indication
troubleshooting. Business purposes include the of a change in the condition of the well and/or its
allocation of leaseholder royalties and costs. associated equipment, but it does not reveal any
Production specialists often use production tests as specific causes for the change. Diagnostic methods
indicators that a well requires troubleshooting. have been developed over the years to identify and
A rapid decline in production between tests may rectify a [Link] and error methods are still
indicate a mechanical problem such as a rod part, used today, but superior results can be obtained
worn pump, tubing leak, or a bad flowline check with modern dynamometer analyses. A fluid level
valve that needs to be [Link] change may instrument by itself cannot typically determine the
also be due to a change in reservoir conditions cause of an operational problem.
related to secondary recovery operations. Modern diagnostic methods began in the 1960’s with
Accepted methods and equipment for production the development of a method for determining the
testing are well documented.1 Production test downhole pump card from surface dynamometer
methods include manual and automatic tank gauging data by Dr. Gibbs.3,4 Although downhole
of oil, portable trailer-mounted well-testing dynamometer measurements have been performed
equipment that measures oil, water and gas, and since the late 1930’s,5 the expense and time made
permanent facility equipment such as lease direct measurements impractical. As computers and
automatic custody transfer systems (LACT).Various solution methods evolved, it became practical to
types of meters are in common use, for example determine pump cards from easily obtained surface
turbine, positive displacement, orifice, ultrasonic and dynamometer [Link] ready availability of pump
coriolis meters. cards allowed for qualitative determination of
numerous pump problems such as defective pump
A common method of metering with newer valves or barrels, gassy or pounding wells,
automated facility systems is that a well is redirected unanchored tubing and parted [Link] reliability of
from a common flowline header from multiple wells this method has been well established.
and switched “on test” [Link] total flow of
all wells is then metered for custody transfer Quantitative information from pump cards reveals
[Link] test period may be a few hours in even more information about a given well. Fluid load,
[Link] “snapshot” of that well’s production is pump fillage, well friction, pump leakage, liquid and
then assumed to be a normal operating condition at gas throughput (gross and net), pump efficiency,
all times while the well is “off test.” pump intake pressure and oil shrinkage can be
calculated using proper data inputs and correlations.
There are several uncertainties in making a snapshot Calculation of pump leakage from dynamometer
assumption. If the well has any downtime at all, for valve check measurements was described in 1990 by
example due to a rod part, the actual production Nolen and Gibbs.6
during a longer time period will be less than the
reported test. Downtime is often neglected entirely One can observe that all requisite information is
in the reported test. Likewise, if a well was having available (from the dynamometer analysis) to utilize
problems during a test, for example a sanded screen, the downhole pump as an accurate metering device.
the actual production may be more than the The one limitation of the dynamometer analysis is
reported test if the problem was corrected in a that it is also a “snapshot” and may not represent
timely manner. Meters require routine checks and what occurs every stroke of the pumping [Link]
maintenance and may have a problem which is not next logical step to take is to implement diagnostic
discovered for some [Link] actual production methods on a real-time basis at each well site using
can vary significantly from reported production a rod pump controller.

For more information [Link]

page 2

Well Production Test via a well K for better agreement; and, wells G and X
Rod Pump Controller suggest that either the wrong pump diameter was
entered, a bad well test, or some other problem
Rod pump controllers (RPC) have advanced exists because it would require a k-factor greater
significantly in the last 30 years. Many are “surface than 1 to have a good agreement between the test
card controllers” which measure polished rod load and the IP.
and position, and determine pump off by observing
Data for these wells was tracked over a one-year
changes in the surface card [Link] have been
[Link] of these were selected for presentation.
used to approximate production by observing
Figure 3 shows results for Well I. Initially, the
“surface fillage,” see Figure [Link] the addition of
calculated IP was high, as expected since the k-factor
onboard well diagnostic capability they can
was at the 1.00 default. After setting the k-factor,
accomplish many more functions in real [Link]
the IP and well test are in good agreement with
latest RPC well manager (WM) technology from
less than 15% difference for the remainder of the
Lufkin Automation7 incorporates pump card
test period.
technology to more accurately calculate production.
This is discussed in more detail below. Figure 4 shows results for Well R. Again the IP
compares very well with the well test, within 10%.
SAMTM IP method—k-factor It is not certain if the k-factor was adjusted in March
2004, but it is apparent that the IP calculation was
The current method in the SAM WM calculates the adjusted, since the agreement is within 5% after
net stroke of the pump card, determines the stroke this date.
volume, and accumulates the incremental volume of
each stroke during a 24-hour [Link] is referred In Figure 5, the daily IP recorded for a one-month
to as inferred production (IP).The incremental period on a different Permian Basin well can be
volume is calculated from compared with two well tests during this period
(from June 4 to July 3, 2004).The daily IP varies from

V1

 d2 S
n (1) about 68 to 80 BBL during this period with about
4
five days of downtime. If each of the 30 daily IPs are
where d is the pump diameter, Sn is the net stroke. added, a total monthly production of 1824 BBL is
This is illustrated in Figure 1b. IP assumes that the calculated. On the other hand, the two well tests
pump is in good condition, leakage is minimal, the average about 74 BBL. If it is assumed to be constant
tubing is anchored at or near the pump, free gas in during the 30-day period, then the monthly report
the pump is negligible at the time of traveling valve would be 2220 BBL, about 22% [Link] highlights
(TV) opening, and oil shrinkage effects are minimal. the ability of IP to automatically and more
In reality, not all of these assumptions are true; appropriately account for run time.
therefore, this volume calculation will typically be
Table 1 presents results for three wells in California.
greater than that reported from a well test. A gross
These wells are tested through one test unit daily.
adjustment, referred to as a k-factor, is applied to
The test unit is calibrated [Link] metered
bring the IP into agreement with the production test
production was reported on three different days. In
and account for the error in the underlying
each case, the total production from IP was within
[Link] approach is reasonable as long as
2% of the test unit.
conditions are steady. As such, k-factors should
generally be less than 1. A common range is from It is easy for one to say, “Well, I’ll just change the
0.85 to 0.9. If a lower number is required, there may k-factor to exactly match the well test.” Indeed that
be excessive pump leakage or a tubing leak that is a good start, but as time goes on, some of the
should be addressed. underlying assumptions may go awry, especially the
one regarding pump leakage. As time goes on, the
RPCs capable of pump IP calculation were installed
pump will wear, resulting in additional [Link]
on several wells in the Permian Basin. IP was
pump still produces what IP calculates, but now
configured, and k-factors entered based on recent
additional leakage means that less arrives at the
well tests. Figure 2 shows the comparison between
surface. Periodic adjustments may be acceptable, but
the IP and the well test on 26 wells. In general, the
require extra effort to maintain. If continuous IP is
agreement is good. On wells S and U the k-factor
intended to be comparable to intermittent well
could have been lowered to better agree with the
tests, the limiting assumptions need to be addressed.
well test; the k-factor could have been raised on

For more information [Link]

 page 3

SAM Well Test Free Gas and Oil Shrinkage

For reasons stated above, new technology is being Figure 6 shows a pump card with free gas in the
developed to eliminate the underlying assumptions pump at the time of TV [Link] volume of free
in the IP [Link] new patent pending gas after compression may not be small and is a
technology is referred to as SAM well test (SWT). function of the pressure of the gas as it enters the
It is anticipated that this method will be used to pump (pump intake pressure). Sgas@Pa represents the
minimize or even replace traditional well tests and corresponding stroke length. It is necessary to
reduce facility testing infrastructure. At the determine the pump intake pressure,
minimum, it can be used for allocation purposes.
Currently, SWT is in an initial phase of field trials; as Lf
Pi
Pa  (2)
such, this paper will introduce the method, but Ap
extensive field data is not yet available. A discussion
of how the limiting assumptions in IP are eliminated where Pi is the pump intake pressure; Pa is the pump
follows. Figure 6 illustrates the relevant concepts discharge pressure due to hydrostatic head of
from the pump card. oil-gas-water in the tubing and the tubing head
pressure; L f is the fluid load derived from the pump
Pump Leakage card; and Ap is the area of the plunger. SWT
automatically determines L f from the shape of
As mentioned previously, diagnostic methods are the pump card. Figure 8 shows a representative
available for quantitatively determining the quantity downhole card where L f has been calculated.
of pump leakage from measured valve checks.
The user can adjust L f to account for additional
Several techniques have been published elsewhere6
well friction if necessary.
and include the TV load-loss method from a valve
check or from the critical velocity on a real-time Pa is dependent on the amount of free gas and
pump card. SWT currently supports the TV load-loss solution gas metered into the tubing each stroke, as
method from a recorded valve check performed well as the amount of water and oil. Data input
previously; the user can also manually enter a includes water cut and reservoir PVT properties.
“known” leakage. Both are based on a 24-hour run SWT includes an iterative solution algorithm,
time; SWT accounts for a stroke period and commonly referred to as PIP, to determine all
calculates the leakage for a single stroke. Figure 7 outputs such that equation (2) is satisfied. PIP uses
illustrate the TV load-loss method. Sleak in Figure 6 Nolen non-dimensional curves for solution gas and
is the equivalent stroke length due to TV/plunger oil shrinkage as functions of [Link] GLR, Pi,
leakage. P a, and Sgas@Pa are some of the outputs generated
from the algorithm. A detailed description is beyond
Some training is required to properly perform this the scope of this paper.
measurement. It is expected that pump leakage
changes slowly with time, so monthly or quarterly
SWT Discussion
valve checks should be sufficient, and only take a
few minutes to perform. Automated methods for All of the effects above are used to calculate oil,
determining pump leakage are being investigated, water and tubing gas volume produced with each
but are not discussed here. stroke. Status information for SWT includes
calculated production for fluid, oil, water and tubing
Tubing Movement gas; and pump volumetric efficiency, pump intake
pressure and pump fillage, see Figure 9. Also
A simple static Hook’s law model is employed to available are cumulative production since gauge off is
subtract the amount of pump stroke due to tubing shown, as well as yesterday’s production, the
movement for the unanchored portion of the tubing. instantaneous rate, and a projected value based on
It is assumed that the tubing anchor is holding, if current run time. 60-day history plots and data. Data
installed. In Figure 6 St is the stroke length due to entry for SWT is shown in Figure 10. By directly
tubing movement. taking into account pump leakage, tubing movement,
free gas and oil shrinkage, SWT has eliminated the
need for the k-factor utilized in the IP method.
page 4

It should be understood that SWT is not without References

limitations. Certainly, if a tubing, flowline check valve,
or flowline leak develops, the calculated production 1. API Manual of Petroleum Measurement Standards,
will be greater than what is received at the stock various chapters.
tank. Recirculation of chemicals should be taken into
account. Sudden changes in SWT and/or run time 2. Joe D. Clegg, Sucker Rod (Beam) Pumping,
from the historical trend may indicate that such 1st Ed., presented at SWPSC, April 1998,
problems have arisen. Proper water cut must be Lubbock,Texas.
entered by the user and checked periodically. As
discussed above, valve checks must be performed 3. S G Gibbs, Method of Determining Sucker Rod
periodically to determine pump TV/plunger leakage. Performance, US Patent 3,343,409, issued
Sept 26, 1967.
Summary 4. Gibbs, S G and Neely, A B: “Computer Diagnosis
of Down-Hole Conditions in Sucker Rod
In summary, RPCs with advanced IP and SWT Pumping Wells,” JPT, Jan 1966, pp 91–98.
technology are capable of calculating production
volumes by utilizing the pump as a flow [Link] 5. Gilbert, G: “An Oil Well Pump Dynagraph,”
IP method has been proven in the field with API Drilling and Production Practices, 1936,
extensive testing to be comparable quantitatively pp 84–115.
with traditional production test [Link]
improvements incorporated in SWT eliminate the 6. Nolen, Ken and Gibbs, S G: “Quantitative
reliance on the underlying assumptions that exist in Determination of Rod-Pump Leakage with
the IP method. Dynamometer Techniques,” SPE Production
Engineering, August 1990.
SWT provides much more detail than has been
previously possible with respect to an individual 7. Dugan, L and Howard, L: “Beyond Pump-Off
well’s production. It can definitely be used to more Control with Downhole Card Well
intelligently determine allocation. As industry Management,” 49th SWPSC, April 2002,
acceptance is obtained, it may reduce or eliminate Lubbock,Texas, pp 55–62.
traditional production testing and related facilities.
SWT may not replace custody transfer
measurements at point of transfer to the customer,
but it can be used as a cost-effective tool to find and
reduce discrepancies. More field testing will need to
be performed to substantiate the SWT concept, and
industry participation is encouraged.

Table 1. IP results on three California wells

11/4/2004 11/8/2004 11/11/2004

Well #1 378 358 346

SAM IP

Well #2 300 291 288

Well #3 518 568 537

Total 1196 1217 1171

Test Production 1207 1232 1192

% Difference 0.9% 1.2% 1.8%

For more information [Link]

 page 5

Surface Stroke Length Gross Stroke (Sg)

Polished Rod Load

Pump Load
POC
Set Point

Sp Pumped Off
Sp Full Net Stroke (Sn)

a) Surface Stroke Position b) Pump Stroke Position

Figure 1. Net stroke determination in RPC for a) the surface card, and b) the pump card (IP method)

350 1
0.9
300
Daily Production (BBL)

0.8
250 0.7
0.6

K Factor
200
0.5
150 0.4
100 0.3
0.2
50
0.1
0 0
A B C D E F G H I J K L M N O P Q R S T U V WX Y Z
Well Test IP K Factor

Figure 2. Well test comparison to IP for 26 Permian Basin wells

k=1.0 k=0.66
150 100%
Daily Production (BBL)

IP/WT Difference (%)

120 80%
90 60%
60 40%
30 20%
0 0%
7/03 9/03 11/03 1/04 3/04 5/04 7/04 9/04 11/04
SAM IP Well Test % Difference

Figure 3. IP and well test tracking on Well I

page 6

Daily Production (BBL) 245 100%

IP/WT Difference (%)

230 80%
215 60%
200 40%
185 20%
170 0%
10/03 12/03 2/04 4/04 4/04 6/04 8/04 10/04 12/04
SAM IP Well Test % Difference

Figure 4. IP and well test tracking on Well R

100
Daily Production (BBL)

80
60
40
20
0
4-Jun 11-Jun 18-Jun 25-Jun 2-Jun
SAM IP Well Test

Figure 5. Daily IP over one-month period compared with two well tests on a Permian Basin well

For more information [Link]

 page 7

Net Pump Discharge

Liquid Volume

Net Stock Tank Liquid Volume

Figure 6. Pump card illustrating SAM well test concept

Figure 7. Pump Leakage calculation from on-site valve

checks (TV loadloss method)
page 8

Figure 8. SWT pump card showing fluid load lines Figure 9. SWT status screen

Figure 10. SWT configuration screens

For more information [Link]

Sales and Service Centers
Bakersfield
5301 Office Park Drive, Suite 305
Bakersfield, CA 93309
Tel: 661-325-3099
Fax: 661-325-3105

Calgary
Suite 1050
808-4th Avenue S.W.
Calgary, AB, Canada T2P 3E8
Tel: 403-234-7692
Fax: 403-265-6913

Midland
2064 Market Street
Midland,TX 79703
Tel: 432-697-2228
Fax: 432-697-0192

Oklahoma City
2300 S. Prospect
Oklahoma City, OK 73129
Tel: 405-677-0567
Fax: 405-677-7045

Houston (Headquarters)
11375 W. Sam Houston Parkway South Suite 800
Houston,TX 77031
Tel: 281-495-1100
Fax: 281-495-6333

Sales and Service Representatives

Argentina, Belarus, Brazil, Canada, Colombia, Egypt, France, India, Italy, Mexico, Oman, Russia,Trinidad,
Venezuela, U.S.A. (Wyoming, North Dakota, Louisiana, Utah, Michigan), United Kingdom

For more information [Link]