SPE 165057

PCP Production Optimization in Real Time with Surface Controller

Alejandro Romero, Sinopec Exploration & Production Argentina, Alberto Gasparri Lufkin Argentina,
Eugenio Ferrigno, Lufkin Industries

Copyright 2013, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Artificial Lift Conference-Americas held in Cartagena, Colombia, 21-22 May 2013.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract

During 2011 and 2012 a pilot was made to test the possibility of optimizing and protecting the PCP pump with control
systems installed on surface. A tool was searched for that would allow for the detection of changes in the behavior of the
contribution of the wells, which leads the wells to sub produce or to have the PCP pump working with low submergence.
(below the safety levels suggested by the manufacturers). Water injection shutdowns, erratic changes in the inflow, nearby
well shutdowns, among others, are the daily problems of the oilfield. The results of this test are recorded in this document.

Introduction

Caadn Minerales oilfield is in the north of the province of Santa Cruz, operated by Sinopec Exploration & Production
Argentina. The oilfield started in the late 50s. Nowadays, it has around 200 producers, of which 130 use artificial lift with
PCP pumps and the other 70 with mechanical pumping. The production of the area is 3774 bopd and 224 Mscfpd. 100% of
the production is obtained through secondary recovery with 62 injection wells and a gross production of 60Kbfpd, which
represents an average water cut of 94%.

The automation of the wells in the area started in 2008 with the installation of the pumping controllers. Injection controllers
were installed later. And remote terminals (RTUs) were installed in wells with PCP, that report the situation of the variable-
speed drive (VSD) and instruments installed on site. The area has well supervision and control software (SCADA) from
which information is obtained and the situation is monitored.

Real Time Monitoring in wells with PCP

Of the wells with PCP, around 10% have downhole sensors. These pieces of equipment obtain the intake pressure of the
pump (PIP) and temperature. With this information, the fluid level above the pump and the well improvement opportunities
can be estimated. In certain cases, these instruments are connected to the VSD, allowing for the adjustment of speed
according to preset parameters.

The principal reason for not having downhole sensors in the other 90% of the wells is the total cost associated. To evaluate
the cost, the following has to be taken into account: downhole sensor, assembly fittings (support, swage, cable), computer
and its installation (more rig time to perform the task). Although the sensor durability is greater than the time of use of the
PCP, an incorrect handling during the installation of the equipment results in problems with its functioning, normally
associated with damages in the cable that brings the information to surface. Ref [1]- [2]. In wells with high water cut (90%
or more) and relatively low production it is difficult to amortize the downhole sensors.
2 SPE 165057

This fact means that only new wells or those with erratic behavior are candidates to have these devices installed. For the rest
of the wells, another option was searched for that would allow for real time control.

The controller tested uses as a technique the control of the pump based on the production of the well. It has the same
platform as the mechanical pumping controllers, so it was not necessary to buy new equipment and if the well is converted
from one system to another, the same controller can be used (with a change of firmware).

The control method is based on the modification of the revolutions per minute (RPM) of the pump rod in predefined steps
and measurement of the production at this new speed, as an increase of percentage speed in a well with PCP, should produce
a similar percentage of increase in the production rate. Therefore, in an ideal case, the increase percentage of RPM (%) and
the production rate (%) should be established at the same value. For example, if there is a pump of 200 BPD each100 RPM
and 75% of efficiency; at 100 RPM it should obtain (200/100) *100*0. 75 = 150 BPD. If the speed is increased in 10% at
110 RPM and the efficiency is the same, the new production rate is (200/100) * 110 * 0,75 = 165 BPD (10% increase in the
production rate).

In reality, the new production rate will probably be below the expected value, as the fluid level above the pump (FOP) has
decreased. Due to this, the pressure delta through the pump will have to increase and therefore, the slippage through the
pump will increase (decrease of the "efficiency"). Continuing the previous example, an increase of 10% of the speed (110
RPM) caused a production of 159 BPD, which represents an increase of 6% and a drop in efficiency to 72%.

The main concept is that, within the range of correct functioning of the PCP, the increase of percentage of production is
related to the increase of percentage speed. In the presence of an increase in speed, the controller expects a minimum value
of increase of production (Setpoint), if this value is not reached; the controller will decrease the speed. If this situation is
repeated, the controller will gradually decrease the speed until it receives a response to the increase of speed which is above
the Setpoint of required production.

Implementation of the Project

For this pilot, it was decided to use a well with downhole sensor, so as to be able to have a follow up of the behavior of the
well and to have a better comprehension of control decisions. The selected well was CL-3011 and the activities to put the
installation in conditions began in November 2011. Gross production from the well at that time was 534 bfpd, 84% watercut
and net production of 32 bopd.

This well has an API 18-40-1500 pump, at 912 meters depth and 1 rods type D. The well has a downhole sensor with an
operative range of 0-3000 PSI at a depth of 877.8 meters. The variable- speed drive installed is of the 6 poles type, with
60HP, intake voltage 380Vac at 50Hz.

To measure flow, a wedge flowmeter was installed in the production bridge of the well. Also, a RPM sensor was installed to
measure the turning rate of the polished rod. This was achieved through a Hall effect type sensor which detects the passing
of a magnet mounted on the polished rod (magnetic pickup). The well also has a pressure sensor in the production line
connected to the controller. In figure 1, there is a diagram of the assembly of the implemented system.

The controller and instruments were started up in December 2011, and then the SCADA interface development was worked
on, as well as the historical information to collect and the flowmeter behavior was controlled. Once it was observed that the
flowmeter was working precisely and the SCADA environment was ready, on February 20th 2012 the well began to be
controlled based on its production.

Control Configuration

For the startup of the equipment, it is necessary to configure the controller with a series of data such as the characteristics of
the pump, the motor, and the head of the PCP, the type of variable-speed drive, analogical ranges of the intake sensors
(flowmeters, pressure sensors, etc.) and outlet of the VSD control. Once all the sensors are configured, the desired control
mode of the well is configured, where the most relevant variables are the following:

Time for Setting up and Sampling: The control is performed based on two periods of operation: the setting up time and the
sampling time. The first one is after a change of speed, when you wait for the production to normalize on the surface. After
this period, the controller starts measuring the production during the time called sampling time. The controller makes an
average of all the values measured during this period and when finished, a new decision of change of speed is made based on
the measured production. For this test, both times were fixed at 10 minutes.
SPE 165057 3

Maximum and Minimum Operating Speed: Maximum and minimum operative speeds must be defined, which are the
application limits of the algorithm. If the algorithm is at the maximum operational speed, regardless of the behavior of the
production, the RPMs are decreased at the next change of speed. The minimum speed is very important as the controller may
be programmed to decrease speed to the minimum under certain circumstances (for example high line pressure). The
configured values were Min RPM =100 and Max RPM = 320.

Maximum Speed of Acceleration and Deceleration: The maximum increase of speed and the maximum decrease of speed are
also configured. With these parameters the idea is to regulate the speed of increase and decrease in each period. A maximum
increase of 8 RPM and the maximum deceleration of 20 RPM were configured.

Minimum Increase of Speed: There are also parameters of minimum changes of speed, one of them defined in RPM and the
other in % of change. The aim is not to make insignificant changes in velocity compared to the related production. For this
case, the configuration is in 2 RPM and 2% respectively. If the operational speed is low, for example 80 RPM, the % is more
restrictive, as the 2% of 80RPM is 1.6RPM. On the other hand, at high speed for example 200 RPM, the 2% is 4RPM and
therefore, the control of the value is more restrictive. In other words, there is a simple form of regulating the changes over an
extensive range of operation.

Minimum Requirement of Change of Production: It is the decision parameter (Setpoint [%]) of the algorithm. If the
reached production after a sampling period is not greater than this percentage value measured in the previous cycle, the
control makes the decision of decreasing the speed. If on the contrary, the measured value on the last sample is greater at
least in this value than in the previous sample, the decision will be to increase the speed. For this case, it was configured in
1.5%.

Maximum Consecutive Steps of Deceleration: In case of a decrease of speed, the production should drop, which could
produce at the end of the following sampling period, the decision to drop the speed again, and this could happen again and
again until reaching the minimum operation speed configured. To avoid these loops that lead to a continuous decrease, the
quantity of continuous steps decreasing the speed allowed should be configured, before starting on a new intent to reach the
setpoint of production, increasing the speed. For this test, the value was configured in two steps; in other words, after two
decreases of speed, in the third step regardless of the measured production, the speed will be increased at the minimum
allowed increase of speed.

Screen developed in SCADA

In order to have the best control possible, on the company SCADA system, a screen was developed with the most relevant
information of the situation of the well. A copy of the screen is shown in Figure 2. Data of instruments and the most
relevant parameters are shown.

Results

Production

Measured production in tank of the well CL-3011on February 24th, 2012 was 597 bfpd. Considering this value as the
starting point, an increase of gross production on an average of 7% was observed (660bfpd) within the period February-
December 2012, with a maximum increase of 20% (717bfpd 12-June-2012).
Net production did not have significant changes. On February 24th, 2012, 32.89bopd were measured. An average increase of
0.27% (33.71bopd) and a maximum increase of 20%, coinciding with the maximum gross production (39.43bopd 12-June-
2012) were obtained.
In figure 3, the behavior of the well production measured in the tank can be observed.

Early Detection of Events

The equipment helped with the early detection of abnormal operation and shutdowns. Shutdowns due to power cut and belts
were immediately detected. It also gave information of the shutdown duration during maintenance interventions. This made
it easier to calculate deferred production and simplified production personnel tasks of those in charge of the well. It was also
observed that in one opportunity the controller had decreased the speed significantly, which led to a revision of the well
situation. During the analysis, the drop of PIP was observed and the decrease of the revolutions was to maintain
approximately a constant PIP. The reason of the PIP drop was due to the fact that an injection well nearby was out of
operation. Therefore, the controller, by decreasing the speed, kept the PCP permanently within the safe operation zone.
4 SPE 165057

Start-Ups

The methodology applied by the area for start-ups after a power cut, is to send an operator to start the pump up at minimum
speed, and as echo meter measurements are taken, decisions are made to increase speed until the submergence of the well is
stabilized within the desired ranges. In many cases, this requires several days.
In Figure 4, the recovery after a shutdown of an automated well can be observed. The figure starts at 09:05am and it is
observed that the well was working in a normal and stabilized way. At 01:06 pm we have the last sample received. At
08:44pm, the first sample is taken after energizing the well. Speed and production are at zero, while the PIP is at a maximum
of 381PSI. In the next sample of the well, it is already in operation at set start up speed (265RPM). From this moment on, it
can be seen how the controller increases speed based on the fact that the production follows this increases proportionally.
The maximum speed reached also matches the maximum RPM configured at 310RPM.
In this figure, the fact is that in only 12 hours and with no assistance of the operator, the well is already stabilized, operating
within the same ranges that it had before the power cut. This has been one of the most important points of the application of
this technology.

Optimization of the Production

It can be seen that once reached the speed of optimum operation, the controller continues searching for opportunities to
optimize production. Figure 4 shows how, once the well is stabilized, the controller increases the RPM in the minimum
value, always looking for greater production.
In figure 5, the search for greater production can be seen as a continuous wave of ups and downs. It is important to underline
that the changes in speed of the figure are the minimum configured, around 6 RPM in this case, so as not to expose the PCP
to sudden decreases of submergence. These changes are so small that they are not observable to the operator of the well,
unless the operator observes them on the screen of the equipment or on the SCADA system from the office.

Operations

Automation allows the optimization of field personnel tasks (operators, maintenance, etc.) as well as personnel responsible
for the operation (analysts, technicians, engineers). In SPE paper 95046 Application of Intelligent Well Management System
to Optimize Field Performance in Golfo San Jorge Basin Argentina [4] benefits obtained are detailed and can be
summarized as:
Optimization of times associated with well visit routines.
Minimization of well visits
Efficiency in the definition of visit programs, determined by the behavior and tendencies remotely observed.
Improvement in the times to determine the root cause of problems, avoiding the need to locally investigate events which
are pre-established remotely.

The optimization associated with the management of the production information is also mentioned. The following are
examples:
Collection of data associated with reports of deferred production (Chapter IV API) through their automatic creation
based on the information collected from the field.
Consolidation of the performance information of the lift systems, with the possibility of verifying in real time the
operation, as a tool of assistance to production operations and engineering.
Minimization of time for adaptation and analysis of well response in secondary recovery projects and new high
declination projects.

Other Points found

Even though the controller has a K factor for flow adjustment, it was detected that this did not work with any of the firmware
versions tested. There was also a difference to what was presented in the controller manual and the results obtained, for the
parameter maximum consecutive steps for deceleration, as the value to configure in the controller is N-1times the real
decrease steps. That is to say, if three steps are configured, in practice, it only decreases two steps and in the third it tries to
increase speed, while the manual indicates that it would decrease the three steps.
SPE 165057 5

Conclusions

From the results, the following conclusions were obtained:

Rapid recovery of the optimum level was observed after the changes produced as a consequence of shutdowns due
to power cuts, changes in secondary recovery contribution, situation of productivity of nearby wells.
As a consequence of having the well flow measurement at the wellhead, changes in production and evaluation of
behavior not only of the well but also of the response of the secondary recovery injection pattern were rapidly
detected.
Well measurement tests at the wellhead were made with magnetic and wedge flow meters, deciding on the latter
option as the primary element of measurement.
The controller looks for a systematic way to obtain greater production, always within the safe parameters
configured.
Although this equipment does not inform the PIP or the submergence, it was observed that submergence is kept
relatively constant and if there is any change in level, it directly affects the speed of operation.
Desktop Information. The project was monitored from production and maintenance personnel desks, which allowed
for an early detection of changes in well behavior.
Routine of field personnel. Visit routines could be optimized through a detailed report of well behavior that could
be obtained automatically. It improves the times for the resolution of problems, as there is a history of the well.

Improvements and Progress with the Project

In August 2012 the second controller in well CL-1446 was installed, and in December 2012, the third device was installed in
well CL-3020. These wells do not have downhole sensors; the results at the moment are similar to the ones obtained
presented in this paper.

There are certain functions that were not tested in well CL-3011 but that are being worked on in the new wells that are related
to torque. Through analogical outputs from the VSD, information can be obtained about the instantaneous torque.
Functional ranges of torque can be defined in the controller, so it can selectively make decisions of operation speed, in
relation to the relevant changes of the torque.

Another function related to torque information is the one called Sand Blowout and it is applied for the removal of sand
trapped in the PCP. To make use of this function, it is necessary to define a threshold torque, so when it goes over this value,
the controller decreases the speed during a period of time, allowing for an increase in pump submergence. After this time,
the controller will increase speed and, by passing great amounts of fluid, the sand can be removed.

It is also desirable to include communication between records of the VSD and the controller, so that all of the VSD records
can be published in SCADA via the controller.

With the results obtained from the evaluated wells, a plan for the installation of 25 devices is currently being developed. The
goal would be to finish 2013 with at least 60 devices operating. There are also plans to take this technology to other areas of
the company in the region. Candidate wells are being studied to install the controllers.

Thanks

We thank the support of Sinopec E&P Argentina, the production personnel of Caadn Minerales area for the help to be able
to carry out this pilot project, and the management for allowing us to publish this paper.

References

1.- Eck J., Ewherido U., Mohammed J., Ogunlowo R., Ford J., Fry L., Hiron S., Osugo L., Simonian S., Oyewole T.,
Veneruso T.: Monitoreo en el fondo del pozo su evolucin, Oilfield Review, Schlumberger, Spring 2000.
2.- Van Gisbergen S.J.C.H.M., Vandeweijer A.A.H.: OTC10945, Reliability Analysis of Permanent Downhole Monitoring
System resume from JPT magazine, October 1999.
3.- Moyano H., Dalle Fiore R., Mazzola R., Ponce C., Ferrigno E., SPE 95046 Application of Intelligent Well Management
System to Optimize Field Performance in Golfo San Jorge Basin Argentina
6 SPE 165057
Figures
Fig. 1 Assembly Diagram
Fig. 2 SCADA Screen
Glossary
Velocidad: Speed
Torque: Torque
Pres. Linea: Line Pressure
Pres. Casing: Casing Pressure
Temp fdo: Downhole temperature
SP pres. g/lt: Pressure SP
Pres fdo calc: Downhole Pressure
calculation
Estado SAM: SAM State
Tiempo en estado: Time in state
Velocidad salida: Outpunt Speed
Caudal: Flow rate
Vol. hoy: Today Volume
Vol. ayer: Yesterday Volume
Arranque: Start-up
Retardo: Delay
Alerta: Alert
Rampas: Ramps
Tiempo: Time
Parada: Shutdown
Tiempo max.: Maximum time
Modo operac.: Operation Mode
Tiempo p/control: Time for Control
Tiempo Muestreo: Sampling Time
Cambio Min Prod: Minimum Change of
Production
RPM % min cambio: Minimum change
in % of RPM
RPM max escala: Maximum scale
RPM min escala: Minimum scale
Max aceleracin: Maximum acceleration
Max desacelerac.: Maximum deceleration
Min. Cambio Vel.: Minimum change in
speed
Max. Cambios Consec.: Maximum
consecutive changes
Tensin Salida: Output Voltage
Corriente: Current
SPE 165057 7

Fig. 3 CL-3011 Production Control against Tank

8 SPE 165057

Fig. 4 Recovery after well shutdown

SPE 165057 9

Fig 5 Continuous Search for Production