Advances in Dynamometer Technology

O. Lynn Rowlan, Echometer Company

Abstract
Dynamometer testing of sucker rod lifted wells is performed routinely in a safe and efficient
manner throughout the world. Advances in dynamometer and computer technology have led to
more accurate measurement of force and position at the surface and calculation of loads along
the rod string and at the pump. Using an accelerometer to accurately detect sudden small
movements of the polished rod gives a detailed picture of the changing load conditions
throughout the pumping cycle. Improvement of the polished rod transducer calibration
technique has resulted in acquiring dynamometer load measurements equal to the accuracy of
precision strain gage load cells. This paper describes advances in dynamometer technology that
allow the operator to better monitor and analyze a sucker rod lift system.
Introduction
Recording load and position data on sucker rod lifted wells with a dynamometer transducer has
been performed in the oil field for years. In the 1950s the popular Johnson-Fagg dynamometer
determined the polished rod load by compressing calibrated steel rings mounted on each side of
the polished rod. The surface load and position is scribed onto a rectangular wax coated card as
the dynamometer moved with the polished rod. Two disadvantages of the Johnson-Fagg
dynamometer were 1) the recorded position was distorted1 and 2) the installation of this
dynamometer between the polished rod clamp and carrier bar was difficult and time consuming.
A big advantage the Leutert dynamometer holds over the Johnson-Fagg dynamometer is that it is
quickly and easily mounted on the well between the carrier bar and polished rod clamp by using
a permanently mounted spool and two specially shaped spacers. The integral hydraulic pump
inside the Leutert dynamometer extends two pistons that support the load causing the hydraulic
pressure in the Leutert body to change as the well load is applied to the dynamometer. The
calibrated springs in the registration unit convert the hydraulic pressure into load. A string
attached to the wellhead turns the registration unit in proportion to the polished rod position,
thereby scribing the load and position onto the wax dynamometer card for each stroke. The
calibrated springs weaken due to use and over time the spring constant gradually changes
affecting the accuracy of the loads. Inaccurate load measurement of up to 40% is a documented
problem when using the Leutert, which the inaccuracies are caused primarily by the mechanics
of the load registration unit2 and load hysteresis due to drag of the pistons seals against the
cylinder walls. Another disadvantage for both of these wax dynamometer card systems is that
the load and position traces have to be tediously digitized by hand, before any detailed analysis
can be performed.
Current portable dynamometer technology uses high performance digital data acquisition
systems to record dynamometer data for analysis of the sucker rod lift pumping system. The
data acquisition system consists of a laptop computer, an analog to digital converter, and a load
cell with accelerometer. The load cell uses a strain gauge circuit to accurately measure the load
on the polished rod. The load cell can be of the horseshoe type, which is temporarily positioned

on the polished rod between the carrier bar and the polished rod clamp, or the donut type, which
is permanently mounted. Alternately, a polished rod transducer of a special design that is easily
clamped directly onto the polished rod is used to measure load indirectly by sensing the polished
rod deformation. The signals from the load cell and accelerometer are connected to the (A/D)
converter for conditioning and digitizing at a user-specified sampling frequency. The digital data
is then routed to the computer where the signals can be processed, displayed, monitored and
recorded. Numerical integration of the acceleration signal is used to calculate the velocity and
the position of the polished rod versus time. This polished rod load and position data is typically
processed to plot a surface dynamometer card and calculate a pump dynamometer card.
The surface and pump dynamometer plots are used commonly to analyze the operation of a
sucker rod pumping system. The surface dynamometer card is the plot of measured polished rod
load at the various positions throughout a complete stroke. The pump dynamometer card is a
plot of the calculated load at various positions of pump-plunger stroke and represents the load
the pump-plunger applies to the bottom of the rod string. Identifying how the pump is
performing and analyzing downhole problems is one of the primary uses of the pump
dynamometer plot.
Surface dynamometer cards sometimes do not allow complete performance diagnostics of the
sucker rod lift system. Measured surface dynamometer cards are valuable for diagnosing rod,
structural, and torque loads on the unit and prime mover. When attempting to diagnose
downhole pump problems, a visual inspection of the surface dynamometer card is often not
sufficient to determine what conditions exist at the pump. Some diagnostics can be done through
practical experiences where certain downhole problems are associated with certain surface
dynamometer card shapes. In shallow to medium depth wells, interpretation of the surface
dynamometer card may be reasonably effective in diagnosing pump performance. In deeper
wells, however, the complex dynamics of the lift system usually prevents diagnosing pump
performance from surface dynamometer cards.
W.E. Gilbert of Shell developed a mechanically operated downhole dynamometer and he became
one of the early pioneers in pump dynamometer card interpretation. The downhole
dynamometer provided both an accurate representation of pump-plunger loading and a valuable
insight into pump mechanics. In 1936 Gilbert published a classic paper on the interpretation of
actual pump dynamometer cards3. Glen Albert4 developed an electronic downhole dynamometer
that is used to acquire rod loading and position over a pre-programmed time interval. Both of
these downhole dynamometers could be installed at a particular location in the rod string and the
dynamometer data was recorded while the well was operating but could only be retrieved by
pulling the rods.
During the 1960s, S.G. Gibbs determined it was mathematically feasible to "wave down" the
conditions acquired from a surface dynamometer and calculate a precise and reliable downhole
dynamometer card.(need reference here) The ability to use a computer5 to calculate the
downhole card has become the basis of modern pump card diagnostic analysis
Polished Rod Position from Accelerometer
Using an accelerometer to accurately detect sudden small movements of the polished rod results
in a detailed picture of the polished rod motion throughout a pumping cycle. The accelerometer

is a solid-state device and does not require routine maintenance, as does the string type
potentiometer. The (A/D) converter continuously acquires "real-time" acceleration data at user
selectable time intervals of 15, 30, 60, 120, 240, 480, or greater samples per second. Polished
rod velocity, Fig. 1, (inches per second) is the integral of acceleration with respect to time, and
polished rod stroke length, Fig. 2, (inches) is the integral of velocity with respect to time.
Accurate polished rod position and velocity data can be determined that describes the motion
during a pumping cycle. Velocity of the polished rod in an upward direction is considered
positive (+), while velocity in a downward direction is considered negative (-). The time axis of
Fig. 1 and Fig. 2 is shown in seconds and represents the elapsed time during one stroke. For a
small duration of time, there are two points at which the polished rod comes to a stop: 1) the
beginning of the upstroke at zero time (or the end of the downstroke), 2) the top of the upstroke
(or beginning of the downstroke). The resolution of the polished rod position determined using
the accelerometer is dependent on the sampling rate and can be almost continuous, while in
comparison the position determined by using string box potentiometers attached to the carrier bar
or by inclinometers attached to the walking beam tends to have poor resolution at the areas in the
stroke where the polished rod comes to a temporary stop. Fig. 3 shows actual acquired polished
rod position at the beginning of the stroke from a beam mounted inclinometer, the red curve is
the smoothing of the data required to prevent false load spikes from appearing on the pump
dynamometer card when the position data is processed using the wave equation.
Sudden changes in the speed of the pumping system is sometimes caused by abnormal load
changes being applied to the rod string, pumping unit, and or prime mover and are clearly
reflected in the acceleration signal. When observing the acceleration data in real time the
operator will often hear a sound at the time when these abnormal load changes occur and he can
relate the sound he hears to certain spikes in the acceleration data. These noises are reflected in
the acceleration versus time display and can be used by the operator to detect a potential problem
and identify its source in relation to the stroke. A fluid pound is a sudden change in load on the
rods and is an example of a downhole event that frequently is seen as an acceleration spike at the
polished rod.
Polished Rod and Plunger Position
The polished rod position and plunger position versus time plots aid in the understanding of the
movement of the pump plunger relative to the polished rod position. Rods stretch when they
pick up the differential fluid load on the plunger during the upstroke. Rods shorten when
releasing or transferring the fluid load from the rods back to the tubing during the beginning of
the down stroke. Unanchored tubing stretches during the transfer of fluid load from the rods to
the tubing during the early part of the down stroke and shortens when the fluid load is transferred
from the tubing to the rods.
For one complete stroke the positions of both the polished rod and the pump plunger are shown
by the two curves in Fig. 4. This figure is an excellent example of the effect of rod stretch on
plunger motion for a well with anchored tubing. During the 8.37 seconds duration of one stroke,
the polished rod position curve has a near-sinusoidal shape with a maximum amplitude, S, of 168
inches, while the magnitude of the plunger stroke, Sp, is 153.6 inches. Starting at the beginning
of the upstroke, point A, the polished rod moves upward but the pump plunger remains
stationary while the rods gradually stretch, until the fluid load, force from the differential
pressure across the plunger, is transferred from the tubing (standing valve) to the stretched rods

(traveling valve). At point B, the rods have stretched enough to balance the fluid load on the
plunger and the plunger begins to move upward. The plunger does not move relative to the
anchored tubing for the time period from point A to point B and the plunger has near-zero
velocity shown in Fig. 5 during this time period. During the rest of the upstroke, the rod stretch
remains approximately constant and the plunger and polished rod move in sync. Point C is at
the beginning of the polished rod down stroke, the plunger again remains stationary while the rod
stretch decreases. The fluid load at point C is beginning to be transferred from the rods and
plunger back to the closed standing valve supported by the anchored tubing. At point D, where
the stretch of the rods becomes zero, the plunger is completely unloaded and again both plunger
and polished rod move together during the down stroke. This process is repeated for each
additional stroke. A horizontal line for plunger position from points A to B and from points C to
D is an indication that the traveling and standing valve are not leaking, the tubing anchor is
holding, and very little fluid slippage through the clearance between the plunger and barrel is
occurring.
Example Pump Card
Fig. 6 is the example pump card that corresponds to Fig. 4 and Fig. 5. The maximum plunger
travel, MPT, is the maximum length of the plunger movement with respect to the pump barrel
during one complete stroke. The fluid load, Fo, is the force caused by differential pressure acting
on the pump plunger and acts on the traveling valve on the upstroke and is transferred to the
standing valve on the down stroke. The differential pressure is the difference between the
pressure due to the tubing fluids and the pressure in the pump barrel. The pressure in the pump
barrel during the upstroke is related to the pressure at the pump intake minus the pressure drop
through the standing valve. The pressure at the pump intake is equal to the annulus pressure at
the depth of the standing valve minus the pressure drop through the dip tube, gas anchor (if
present) or screen that may be attached below the pump. In most cases the pressure drop through
the assemblies attached to the pump are considered to be small and are ignored. When fluid
entry into the pump in partially blocked, a long dip tube is attached to the pump intake, or the
viscosity of the fluid is high; then this pressure drop should not be ignored, because the
additional pressure affects the fluid load that is applied to the pump. The magnitude of the fluid
load is equal to the pump discharge pressure minus the pump intake pressure multiplied by the
plunger area. From points B to C the rods carry the fluid load, when the traveling valve is
closed. From points D to A the tubing carries the fluid load, when the standing valve is closed.
The effective plunger travel, EPT, is the length of the plunger travel when the full fluid load is
acting on the standing valve.
The successive steps in the pump operation are:
1. At the start of the upstroke (point A), the traveling valve and standing valve are both
closed.
2. From point A to point B, the fluid load is fully carried by the tubing prior to point A and
is gradually transferred to the rods at point B. The load transfers as the rods stretch to
pick up the fluid load (Fo). If the tubing is anchored, the plunger does not move relative
to the tubing. The pressure in the pump barrel (Pb)decreases from the static tubing
pressure (Pt) to the pump intake pressure (Pint) and any free gas in the space between
valves expands..

3. The standing valve begins to open at B, allowing fluid to enter the pump when the
pressure in the pump drops below the pump intake pressure (Pint).
4. From point B to C, the rods carry the fluid load while well fluids are drawn into the
pump.
5. At C, the standing valve closes as the plunger starts down, and the traveling valve
remains closed until the pressure inside the pump is slightly greater than the pump
discharge pressure (Pd) that is equal to the static tubing pressure (Pt).
6. From C to D, gas in the pump (if present) is compressed as the plunger moves down to
increase pressure on the fluid from the intake pressure (Pint) to the static pressure in the
tubing. As the fluid in the pump barrel is compressed, then the fluid load is gradually
transferred from the rods to the tubing. The plunger does not move, if the pump barrel is
full of an incompressible fluid, until the rod stretch is recovered.
7. At D, the pump barrel pressure (Pb) equals the static tubing pressure (Pt), and the
traveling valve opens.
8. From D to A, the fluid in the pump is displaced through the traveling valve into the
tubing as the closed standing valve holds the fluid load on the tubing.
During the upstroke the fluid load, FoUp, applied to the rod string is due to differential pressure
acting on the plunger and is equal to the pump discharge pressure minus the pump intake
pressure time the area of the pump plunger. For low viscosity fluids the friction force between
barrel and plunger is usually small and the pressure drop across the standing valve is usually
small so that the barrel pressure is close to the pump intake pressure. On the downstroke the
fluid load the pump applies to the rods is near zero, because the pressure inside the pump barrel
is usually almost equal to the pump discharge pressure. The friction force between barrel and
plunger and the force due to fluids moving through traveling valve are usually small. For
viscous fluids the friction and pressure drop through the valve may become significant. In
addition to FoMax, it is useful to show two additional reference lines on the pump card as shown
in Fig. 7:
1) On the downstroke, FoDn represents the average pump card loads. Generally this line
should be near zero load.
2) On the upstroke, FoUp represents the average pump card loads. When the well is
produced at is maximum potential with low pump intake pressure, then the fluid load on
the upstroke should be near the reference line of FoMax.
FoMax is equal to the pump discharge pressure times the plunger area, this is the maximum
fluid load required to lift the fluid to the surface, with pump intake pressure near zero and the
well providing no help in lifting the fluid to the surface.
Polished Rod Transducer
The polished rod transducer, PRT, is a very popular and versatile dynamometer that can be
quickly installed in less than 30 seconds simply by clamping it to the polished rod below the
carrier bar. The PRT is used to gather load and position data that allows the calculation and
determination of a surface dynamometer card, a pump card and traveling and standing valve tests
when using the portable Well Analyzer and Total Well Management6 software. The
dynamometer data is sufficiently accurate for most pumping unit analysis in a very safe manner
with a minimum of effort.

The position of the plunger in the pump barrel is not changed by the installation of the PRT, as
occurs in some horseshoe dynamometer installations. For this reason, the polished rod
transducer analysis may be more representative of actual well performance than an analysis
using a horseshoe transducer that raises the rods and plunger in the pump. The PRT7 contains
load measuring semi-conductor strain gauges and an accurate accelerometer. It measures the
change in diameter of the polished rod and converts the change in diameter to the change in load
on the polished rod. The extremely small diameter changes of the polished rod are detected
using sensitive solid-state strain gauges. The strain in the radial direction resulting from a stress
in the axial direction is converted into changes in axial load on the polished rod using a
generalized form of Hooke's law, z z/ E, for homogeneous isotropic materials and rz.
Since the Poisson's ratio ( for steel is about 0.3, the radial strain (r) is about 30% of the axial
strain ( z). The PRT is approximately 3 times more sensitive than the normal strain gauge load
cell that measures axial strain. The acceleration data is twice integrated as discussed in a
previous section to determine the polished rod position. The change in load for a pump stroke
and the calculated positions from the acceleration data are used to generate a surface
dynamometer card. The software calculates a pump card using the wave equation solution for
rod loads and displacements, the acquired relative load values and the polished rod position data.
The relative loads of the surface and downhole cards are calibrated by software using the
principle that the average mechanical load the pump applies to the rod string on the downstroke
when the traveling valve is open should equal zero and thus the pump card should rest on the
zero load line. The surface card, which contains both positive and negative relative load values
is adjusted, by the same offset that is utilized to set the average of the down stroke pump card
loads to zero, resulting in the display of absolute load values. Software performs these
calculations and plots the calibrated surface and pump cards.
If the proper amount of friction along the rod string is NOT accounted for by the wave equation
(based on average damping factors) when waving down to the resultant pump card then the
surface dynamometer card loads may be offset by a value representing an unaccounted
friction. This unaccounted friction shifts the surface loads upward by approximately 1/2 of the
value of the unaccounted friction, because similar amounts of friction usually occur on both the
up stroke and the down stroke. Paraffin, a crooked hole, or viscous oils can cause unaccounted
friction. Shown in Fig. 7 is an example of a pump card determined from data acquired using an
accurate calibrated load cell; due to unexpected friction from accumulated paraffin along the rod
string the pump card has approximately 850 lbs of unaccounted friction on both the up stroke and
the down stroke. If the PRT had been used to acquire the dynamometer data on this well, then
the surface loads would be shifted upward by 850 pounds. A horseshoe transducer should be
used in special cases where the drag or damping factor on the sucker rods cannot be estimated
with reasonable accuracy.
Modifications in the software have also improved the accuracy of surface loads determined when
using a PRT on wells that are tagging on the down stroke. Acquiring dynamometer data on wells
with a tag on the down stroke is often impossible when installing a horseshoe transducer between
the pumping unit carrier bar and the permanent polished rod clamp, because the horseshoe load
cell lifts the pump up 3 inches, re-spacing the pump, and the tag usually goes away. Fig. 8
shows how setting the average of the down stroke pump card loads to zero improves accuracy of
surface loads on wells with a tag. In the past the minimum pump card load was set on zero,

which resulted in an upward shift of the surface loads equal to the absolute value of the negative
tag. The calculated peak polished rod load without the proper adjustment for the tag is 23,843
pounds, which is too high by the 2,914 pounds of negative load being applied by the tag. The
surface PRT dynamometer cards peak polished rod load is a more accurate 21,041 pounds when
the pump card load offset due to tagging is adjusted properly.
Fig. 9 compares the of polished rod and pump dynamometer cards data acquired using
simultaneously the PRT and a horseshoe load cell on the same well. The shapes of the
dynamometer cards are the same and the downhole pump fillage of 55% was calculated to be the
same for both systems. The average deviation is only 50 lbs when comparing the measured
loads of the two systems, while absolute maximum deviation is 250 lbs. The peak polished rod
load, PPRL, determined using the PRT is within 171 lbs of PPRL of 13,012 lbs determined by
using the horseshoe load cell. The minimum polished rod load, MPRL, determined using the
PRT is within 60 lbs of the MPRL of 5,885 lbs determined by using the horseshoe load cell. The
difference in load acquired using the PRT when compared to the load acquired with a horse load
cell is small and the loads are more accurate than would normally be expected when comparing
two load cells.
Horseshoe Load Cells
Horseshoe transducers are used to very accurately measures polished rod load using 12 strain
gauges, which are mounted on three supporting members. Offloading or side loading due to the
carrier bar being tilted does not affect the accuracy of the load measurement due to the averaging
effect of the multiple gages. The transducers are manufactured with instrumentation grade
stainless steel and incorporate a high accuracy accelerometer from which the software computes
the velocity and position of the polished rod. The horseshoe transducer is calibrated to yield an
overall accuracy of 0.5% of range or better. The 4" horseshoe transducer is rated at 30,000 Lb.
and the 5" horseshoe transducer is rated at 50,000 Lb.
A horseshoe transducer is installed between the pumping unit carrier bar and the permanent
polished rod clamp. To install the 4" horseshoe transducer, a temporary polished rod clamp is
positioned on the polished rod about 4 inches above the stuffing box when the polished rod is
stopped at the bottom of the stroke. A temporary knock-off block is located on the stuffing box
as the polished rod with the temporarily installed clamp is on the down stroke. Also when on the
down stroke, the motor is turned off the momentum of the system causes the polished rod to
continue downward until the temporarily installed polished rod clamp comes in contact with the
knock-off assembly. The pumping unit brake is set when the polished rod is at the bottom of the
stroke. This causes the permanently installed polished rod clamp that normally rests on the
carrier bar to be several inches above the carrier bar. The 4" O.D. transducer is 3" high, and it is
positioned into the free space between the carrier bar and the permanent polished rod clamp.
The pumping unit brake is then released, causing the load to be transferred from the knock-off
block to the carrier bar. Then the knock-off block is removed and the motor is started again.
An alternative system uses a hydraulic lift horseshoe dynamometer that requires permanent
installation of an inexpensive spacer spool over the polished rod positioned between the polished
rod clamp and the pumping unit carrier bar. This 5 OD horseshoe load cell dynamometer
system is designed to facilitate safe and quick installation of the load cell, and to eliminate

measurement errors caused by changes in pump spacing resulting from temporary installation of
a horseshoe transducer between the carrier bar and the permanent polished rod clamp. When a
dynamometer test is desired, the horseshoe transducer and a hydraulic lift assembly are easily
inserted into the spool. The hydraulic lift is actuated using a small portable hydraulic pump,
which transfers the polished rod load from the spacer onto the 5" horseshoe load cell. After a
inch thick spacer plate is inserted between the hydraulic lift and the horseshoe transducer, the
hydraulic liquid is bled back into the pump, and the polished rod load is lowered onto the
horseshoe transducer, spacer plate, and the hydraulic jack. The entire polished rod load is
supported on the horseshoe transducer. Once the pressure in the pump is released, then the hose
and fitting can be disconnected from the hydraulic jack. After the dynamometer test, the
horseshoe transducer, hydraulic jack and spacer are removed in reverse manner leaving the spool
on the well for future tests when desired.
The main advantage of the 5" horseshoe transducer hydraulic jack assembly is that the polished
rod is raised less than 1/4" during installation of the load cell. Thus, the pump plunger continues
in the same operating position in the pump when the dynamometer test is performed. Also, the
installation of the 5" horseshoe transducer (using the hydraulic jack) is easier and safer than
installation of a horseshoe transducer that requires separating the pumping unit carrier bar from
the permanent polished rod clamp a sufficient distance (3 inches) so that the horseshoe
transducer can be inserted between the carrier bar and the clamp. The disadvantage of the
horseshoe transducer hydraulic jack assembly is that a spool must be placed (in general
permanently) over the polished rod between the pumping unit carrier bar and the polished rod
clamp requiring a spool for each well.
Modified Leutert Dynamometer
With a conventional Leutert system the surface dynamometer load and position data are traced
on a waxy paper card. The dynamometer trace on the card is generally not digitized and is used
only for qualitative interpretation of pump operation. Modifying the Leutert dynamometer unit
to communicate with an A/D device and then processing the dynamometer data with software,
allows the operator to perform a very detailed analysis of the total sucker rod lift system. The
main concern with using a hydraulic system to measure varying loads is inaccuracy of the
acquired load data, caused by hysteresis due to excessive friction in the registration unit and in
the cup piston seals. Hysteresis is defined as a lag effect when the forces acting on a body are
reversed. When load is applied to the Leutert dynamometer and then released to a specific load,
the measured load is higher than the actual load. Hysteresis will result in erroneous load
readings, which affect rod stress, gearbox loading and other calculations. The replacement of the
registration unit with electronic sensors results in a reduction of a large amount of hystheresis.
The registration unit has a small moving piston that reacts to pressure changes within the Leutert.
This moving piston is thought to contribute to the majority of the hystheresis effects. There is
also a spring and linkages, which transfer the load differences to a stylus, which traces on a waxcoated card attached to a drum. The registration unit not only contributes to the hystheresis, but
also requires the highest amount of maintenance. By replacing the registration unit with a
pressure transducer and accelerometer, then the load data is more accurate, the acquired load
values are repeatable and maintenance problems are reduced. Hystheresis for the modified
Leutert is usually reduced to within 0.5 and 1.5% and is almost equal to the precision of a
calibrated strain gauge load cell.

Instead of using horseshoe a load cell or PRT for acquiring dynamometer data, some operating
companies have decided to modify existing Leutert dynamometers based on the following
factors:
1) Current operator experience dictates using the same type of portable system.
2) Leutert dynamometer is easier to install on the polished rod, than the 4 horseshoe type
load cell that requires the rods to be stacked off on the stuffing box. If stacking off is
not performed correctly, this practice can cause damage to the stuffing box and/or
polished rod and at times can be dangerous if the pumping unit break is not operating
properly.
3) The Leutert dynamometer is used by the operating company, so the wells already have
Leutert spacers installed, precluding the associated cost of purchasing the spool and
washers for a 5 load cell.
4) The Leutert dynamometer only raises the rods l/16 to l/8 of an inch, resulting in data
representative of actual operating conditions.
Permanently Mounted Donut Load Cells
The load cell is cylindrical (donut) shaped and usually has strain gauge rosettes located at equal
distances around cylindrical body. The donut load cell is permanently mounted over the
polished rod between the carrier bar and the polished rod clamp. Position of the walking beam is
often determined through some type transducers where the voltage signal is proportional to the
angle swept through by the walking beam. Generally these installations are used in conjunction
with pump off controllers or as field end-elements in SCADA systems. The most accurate
dynamometer measurements are obtained using a calibrated strain gauge load cell, which directly
measures the load on the polished rod.
Proper analysis of dynamometer data depends on accurately measured loads and the polished rod
loads should be acquired using a calibrated load cell. Often the load measured with a
permanently mounted, donut type load cell can be in error due to:
1) Damage to the load cell,
2) An error in the calibration of the load cell,
3) A load cell that is not centrally loaded at the top and bottom,
4) The load cells offset for zero load has changed or is unknown and the well loads must be
stacked off of the load cell to determine the zero offset, or
5) An error in the calibration of the controller incorrectly converting the mV/V output from the
load cell into pounds of load.
In order to verify the accuracy of the measurements and the operation of the pumping system, the
permanently mounted donut load cells can be used to measure load data in conjunction with
portable dynamometer data acquisition systems. An external string box or position transducer is
sometimes used to determine position, but concerns about accuracy of position data still exist as
previously discussed. A special accelerometer transducer, similar in size to the PRT but
containing only the accelerometer function, can be used to determine position. This special
accelerometer transducer is used to replace any string box or inclinometer that could be used
with any type of load cell. The special accelerometer transducer can be quickly installed and is
simply clamped to the polished rod below the carrier bar. The special accelerometer is small and
can be easily installed on a well to obtain acceleration data along with load data from the
separate load cell.

Dynamometer Data
The previous sections in this paper describe advances in dynamometer technology, which allow
the operator to better monitor and analyze a sucker rod lift system. Acquiring accurate
dynamometer data representative of the wells normal operating conditions requires the operator
to take care when he acquires data. A sucker rod pump operated in a continuous mode is
constantly lifting fluids to the surface, while maintaining steady state conditions with a constant
producing bottom hole pressure and continuous inflow into the wellbore. The pumping unit
must be stopped in order to attach a dynamometer transducer, for safety purposes the operator
should disconnect the power and set the break. The normal steady state operating conditions are
disturbed during the time the well is shut down. It is recommended the pumping unit operate for
a sufficient time after being shut down until the well again operates at stabilizes conditions, then
the acquisition of dynamometer data will be representative of the conditions of how the well
normally produced. Other times the problem needing analysis may be of an intermittent nature
and the operator must acquire data when the problem occurs, this may sometimes require the
operator to collect data for extended periods of time.
Analysis of the collected dynamometer data shows the loads and horsepower requirements of the
surface dynamometer card and the pump card. A traveling valve and standing valve test can be
performed. A comparison of the measured load to the calculated buoyant rod weight is an
excellent check that the well's rod data are entered correctly. The traveling valve and standing
valve tests allow the calculation of pump intake pressure, pump leakage, traveling valve and
plunger performance and standing valve leakage performance. Gearbox loading is calculated by
software using the polished rod load and position data. The counterweight moment must be
calculated using the known properties of the cranks, counterweights and counterweight positions
or the counterweight moment can be determined by measurement of the counter balance effect
using the accurate load transducer. Gearbox loadings and a permissible load diagram are
calculated and displayed. Dynamometer data processed with software allows the analysis of
polished rod power requirement, pumping unit beam loadings, rod loadings, pump power
requirements, and pump performance. These types of analysis allow the operator to effectively
monitor and analyze the sucker rod lift system.
Conclusion
Technology derived from computer software and improvements in transducers is continually
evolving. New technology is improving the ability of the operator to better understand and
analyze the operation of the sucker rod lift equipment. The advent of portable computers and
software has made analysis and acquisition of dynamometer data easier, faster, and more
accurate. Use of a portable system permits further in-depth analysis of the sucker rod pumping
system at the well site. Polished rod position data determined by using an accelerometer is far
superior to the position data determined from other devices. Accelerometers can be used to
accurately determine position in conjunction with any type of load cell. Polished rod loads
determined using a polished rod transducer are usually accurate to with-in the expected 0.5 %
error of a calibrated load cell.
References
1.

Slonneger, J. C.,Dynagraph Analysis of Sucker Rod Pumping, Gulf Publishing, 1961., pp.65-69

2.
3.
4.
5.
6.
7.

Garrett, M., Rowlan, O. L., McCoy, J.N., and Podio, A.L., Improved Field Measurements Aid in Sucker Rod Lift
Analysis presented at the Southwestern Petroleum Short Course, Lubbock, Texas, 1996.
Gilbert, W.E.: An Oil Well Pump Dynagraph, API Drilling and Production Practices, 1936, pp.94-115.
Albert, G.: Downhole Dynagraph Measurement A Thesis for MS at University of Oklahoma, 1986
Gibbs, S. G. and Neely, A. B.: "Computer Diagnosis of Down-Hole Conditions in Sucker Rod Pumping Wells,"JPT (Jan.
1966) 91-98
Podio, A.L., McCoy, J.N., etc.: Total Well Management II, SPE 67273, SPE Production and Operations Symposium
Oklahoma City, Oklahoma, 2427 March 2001
McCoy, J.N., Jennings, J. W., Podio, A.L., A Polished Rod Transducer for Quick and Easy Dynagraphs, SWPSC,
Lubbock, TX, 1992

Nomenclature
MPT

= Maximum Length of the Plunger Movement

Fo

= Fluid Load Acting on the Plunger

EPT = Effective Plunger Travel

= radial strain
= Poisson's Ratio
= axial stress
= Young's Modulus

FoUp = Fluid Load on Up Stroke

Fig. 1 Polished Rod Velocity is the Integration of Acceleration with Respect to Time.

Fig. 2 Polished Rod Position is the Integration of Velocity with Respect to Time.

Fig. 3 Polished Rod Position from Inclinometer Requires Smoothing of Error.

Fig. 4 Polished Rod and Plunger Position Versus Elapsed Time for Anchored Tubing.

Fig. 5 Polished Rod and Plunger Velocity Versus Elapsed Time for Anchored Tubing.

Fig. 6 Example Pump Card w/ Load (k-lbs) and Position (inches)

Fig. 7 Pump Card with Unaccounted Friction from Paraffin

Fig. 8 Surface PRT Dynamometer Adjusted for Load Offset due to Tagging on the Down Stroke

Uncorrected

Corrected

Fig. 9 Comparison of Polished Rod and Pump Dynamometer Cards for PRT and Load Cells