PROGRESSIVE CAVITY PUMPS PROVE MORE EFFICIENT IN MATURE WATERFLOOD TEST
Donny W. Wright
Amoco Production Co.
Odessa, Tex.
Richard L. Adair
Highland Pump Co.
Odessa, Tex.
For mature waterflood wells, field tests in the Permian basin indicate that progressive cavity (PC) pumps provide greater mechanical efficiency and use less electricity than beam and electric submersible pumps (ESP).
The field tests are being conducted in wells that produce 500-1,000 b/d from 3,800-5,000 ft.
The testing started in 1991 and will continue through 1993.
ENVIRONMENT
The maturing of a waterflood creates constant changes that affect artificial lift design. Most often, the lifted liquid volume increases while the percent oil in the produced fluid decreases. This increased expense and decreased return on capital causes many producing wells to become marginally economic.
The need to lift greater fluid volumes more efficiently led to examining alternative artificial lift for economic attractiveness.
Wells evaluated produce high volume (500-1,000 b/d) with beam pumps and low volume (800-1,000 b/d) with ESPs. Depths range from 3,800 to 5,000 ft. All these wells produce from the Grayburg and San Andres horizons.
The typical environment is:
- Water/oil ratio: 90-98%
- Hydrogen sulfide: 820% in the gas phase
- Hydrogen sulfide: 1,000 ppm in the water phase
- Crude gravity: 36-38 API
- Corrosion: moderate to extreme
- Iron sulfide (FeS): moderate to extreme in the fluid.
The focus was on operating at or near pumped-off conditions (net lift equal to pump setting depth).
TEST OBJECTIVES
Rene Moineau presented a doctoral thesis to the University of Paris in the mid-1930s.1 Simply stated, Moineau's principle uses a rotor with a single external helix that is inserted into a stator containing an internal double helix. This arrangement creates a series of cavities (Fig. 1).
When one member is rotated, cavities progress from one end (suction) to the other (discharge) creating a continuous flow.
Positive displacement is created due to the distance between the centerline of the rotor and stator. Fig. 2 illustrates the geometry of this helical gear.
PC pumping systems have successfully pumped shallow low-gravity crude oil and water, but limited testing and documentation has been done with environments and depths proposed for this study.
The first step in the study was to become familiar with technology used in shallow heavy-oil production and determine what was applicable for higher gravity, less viscous fluid.
For this reason, primary and secondary, operational concerns where identified.
PRIMARY CONCERNS
Primary operational concerns were defined as problems that caused catastrophic failure or were detrimental to the use of PC pumping systems under conditions previously stated. These were:
- Adequate pump stages to lift required volume from required depth
- Elastomer compatibility with higher aromatics in crude oil and production chemicals, such as corrosion inhibitors
- Adequate drive-head-bearing performance because of increased weight/load.
SECONDARY CONCERNS
Secondary operational concerns were defined as 11 problems that must be addressed because of noncatastrophic failures" or "problems not detrimental to the use of PC pumping systems" in the above-stated environment. These were:
- Tubing back off because of friction in the pump
- Tubing wear from rods and rod couplings because of increased rotational speed (up to 720 rpm)
- Adequate pump-off control
- Environmental exposure such as stuffing-box performance
- Installation and pulling expense.
TEST DESIGN
The equipment was designed in four phases. Phase 1 included design of a PC pumping system using existing technology, and choosing candidate wells representing the majority of currently operated mature waterflood wells.
In Phase 2, 12 PC pumps were installed and design changes made as necessary, and the primary operational concerns were identified and corrected.
The third phase continued to identify and correct primary operational concerns as well as the secondary operational concerns. Also, electrical and mechanical efficiency was monitored and evaluated.
In the last phase, Phase 4, the system's operation is being optimized over a 2-year period that will determine pump life. Effective pump-off control systems were developed and evaluated. Also, a comparison was made of the electrical and mechanical efficiencies to beam and ESP pumping systems under like conditions.
PHASE 1
Phase 1 determined the PC pumping system design and application criteria.
Efforts were made to capitalize on all existing technology that applied to mature waterfloods. Reference 2 was very useful. Also, a visit to manufacturing and field facilities in Alberta helped evaluate applicable technology.
One criteria for the pump selected was that each pump stage must support 100 psi pressure. At a minimum, an 18-stage pump was required but a 26-stage pump was selected because of more flexibility of rotor/stator fit, compensation for elastomer incompatibilities, and passage of solids.
Although other models are applicable, a model 200/26 stage was selected for comparative reasons. This pump is designed to lift 1.75 bid/rpm.
Tubing anchor catchers (TAC) and modified packers allowed the tubing to be set in tension and prevented tubing back off. For evaluation purposes, some installations were without TACs or modified packers.
The lack of a TAC caused no problems but it is believed that tubing in tension during operations prevents tubing wear.
In all installations, the tubing was 2.875-in. EUE 8RD. One 8-ft tubing sub at the discharge of the pump allowed bottom rod movement. In some installations for strength, heavy wall tubing was added to the bottom and top of the tubing string.
All installations had 1-in. sucker rods except for two that had 7/8-in. rods. Different pony-rod arrangements at the bottom and top of the rod strings offset the harmonics of the rotating strings. This design was based on work done by the Centre For Frontier Engineering Research (C-FER) in Edmonton, Alta.
A variety of rod protection devices were evaluated. Most installations had polyurethane-coated couplings. Two installations included four rods with molded-on rod guides (three per rod) installed at the pump discharge. Spray-metal rod couplings were used in four wells.
After tubing failures occurred, snap-on, lotus-type rod guides were later added to these strings, one per rod.
Combination molded/snap-on rod guides allowed a molded plastic surface to spin through a snap-on polyethylene guide. For wear, these provide a bearing surface of plastic on plastic.
All installations had a standard polished rod, 26-ft long, to allow rotor removal from the stator and allow fluid bypass without the need to remove surface equipment.
Vertical electric drive (VED) units powered the system. Produced volume determined the electric motor's size. All motors were NEMA B.
Initially, standard compression-packed stuffing boxes were used. Later, these were replaced by oil-lubricated seals.
Mechanical drum-type speed limiters were installed initially to constrain the back-spin speed to within sheave limits. Later, some were replaced with hydraulic-type limiters.
In one instance, a small blowout preventer (BOP) was set below the drive head to allow stuffing box repacking without killing the well.
In all cases, electronic controllers allowed programmable kill and restart features as well as constant monitoring of electrical usage. These included a flow line pressure switch and flow line turbine meter that monitored production rate. The controller was set to stop the motor when the rate reached one half of the designed rate for that well.
For monitoring efficiency, in-line kilowatt hour meters were installed on five wells. Where possible for comparative reasons these meters were installed and monitored while the wells were beam pumped.
Fig. 3 illustrates a typical installation. Modifications were made once problems were identified.
Candidate wells were chosen based on production rate, depth, completion horizon, and problems associated with previous pumping designs.
The intent was to choose wells representing the majority of those found in South Permian basin mature waterfloods. Large volume producers with high water/oil ratios were targeted.
The pumping systems in the candidate wells were at or near the limits. Most candidates were known to have beam-pump problems because of iron sulfide (FeS) in the well bore.
Because H2S content is present in most wells, candidate wells had H2S ranging from 8 to 20% in the gas phase and 1,000 ppm in the water phase. Aromatic content of the oil was approximately 8%. Inhibitors containing no aromatics mitigated corrosion.
PHASE 2
In the second phase, the object was to identify and correct the primary operational concerns.
Five of the 12 PC pump installations were evaluated for electrical and mechanical performance (Table 1).
The 26-stage pump was adequate although net lift never exceeded 4,400 ft. An attempt was made to increase production from 1,000 to 1,200 b/d by increasing pump speed from 600 to 720 rpm. Because only a small increase in volume was realized, it was determined that the pump volume begins to peak at 600 rpm, depending on net lift and rotor/stator fit.
There were no problems with the drive head and bearing because of the increased weight loads. The bearings subjected to the highest loading and rotational speed were inspected and found to be in excellent condition.
No problems were identified with aromatic incompatibilities with the high-nitrile elastomer used. Normally used oil-soluble corrosion inhibitors were replaced with a water-soluble inhibitor to avoid potential problems.
Although softening and blistering associated with aromatic attack were not experienced, five stators failed to pass bench testing when pulled. Test failure was due to high torque.
Internal inspection revealed different degrees of surface hardening of the elastomer. The worst case of hardening was only '/8-in. deep, but was enough to cause surface cracking and increased torque beyond allowable limits.
It is suspected that H2S attack is responsible for the stator failures, although other stators exposed to the same environment, for even longer periods of time, continue to perform without a problem.
The sulfur present in the H2S is suspected of creating a post curing of the elastomer. High revolutions per minute may have also contributed to this process. Samples of the failed elastomers are still being analyzed at the time of this writing.
PHASE 3
Phase 3 identified and corrected the secondary operational concerns.
Tubing back off was experienced in one installation. The tubing string had a standard TAC set just below the pump. Analysis indicated that when the TAC released, the tension in the tubing string moved the stator upwards and allowed the rotor to contact the tag bar (bottom portion of the pump used for spacing the rod string).
The back off occurred instantaneously, just after fluid reached the surface. Although the tubing unscrewed, the cause was a standard TAC failure.
In two other wells, standard TAC failures occurred before modified packers were installed. The modification included a positive-locking J setting that prevents turning of the tubing. All TAC installations that followed were of this type.
No appreciable tubing wear was identified where polyurethane coated couplings and polyurethane centralizers were used. Some debonding and breaking of the polyurethane occurred in most wells.
Other urethane blends are currently being tested. Various types of molded and field-applied rod guides are also being tested.
After 4-6 months, tubing wear was identified in wells with spray-metal rod couplings. This wear was most predominant at the bottom 100 ft, with some wear also at the top. The use of spray-metal coated rod couplings has been discontinued.
Molded rod guides such as those used in beam-pumped wells experienced premature wear.
A new type of tag bar for spacing the rotor has been developed. The new design prevents the rotor from sticking if the rotor contacts the tag bar while operating.
Although some pitting of the rotor's chrome plating has occurred, it has not been detrimental.
Correct spacing was difficult in the beginning. Failure to allow enough space for rod stretch because of loading resulted in rotor contact with the tag bar. Increased stretch allowance solved this problem.
Rods have been over torqued and, in one instance, parted because of rotor contact with the tag bar. Proper spacing has eliminated this problem.
Improper alignment has caused polished-rod failure, well head movement, and stuffing-box leakage. These problems have been eliminated by using a dial indicator to assure proper well head equipment alignment with the well.
Because of rotational speed, an oil-lubricated seal has been developed to replace the compression-type stuffing boxes. The seal gland allows, if required, slight movement with the polished rod.
Originally, mechanical drum-type speed limiters restricted reverse rotational speed when the tubing and annulus fluid levels equalized. These were unsuccessful and have been replaced with a newly designed hydraulic speed limiters. The new design allows the release of energy stored in the rod string and prevents reverse rotation speed. Although some efficiency loss has been detected, in each case it has been attributed to a downhole cause.
As shown in Table 1, the average overall system efficiency for PC pumps is 63.4%. This is 23% more efficient than the evaluated beam pumps. Also included in Table 1 are the results from an efficiency study from Reference 3 on wells meeting the same conditions as previously discussed.
Results show that PC pumps are 13% more efficient than beam pumps and 50% more efficient than ESP pumps. It should be noted that all system efficiencies in Table 1 were determined using the same method.
PHASE 4
In the last phase, the pumps are being optimized. Monitoring pump performance before and after installation has been very beneficial.
Graphs such as Fig. 4 should be maintained for future reference when the pumps are pulled. By studying graphs such as Fig. 5, it appears possible to design a specific rotor/stator fit to match the conditions of the well.
Because PC pumping systems remain relatively constant, as opposed to cyclic, system performance is easily monitored. The volume is proportional to rotation speed, and controllers can be used to increase or decrease production rates.
Use of controllers has indicated possible prediction of net lift. Although more evaluation is necessary, PC pumps appears to have significant potential. This study will continue through 1993, after which a conclusive evaluation will be done.
ACKNOWLEDGMENTS
The authors gratefully acknowledge Amoco Production Co.'s South Permian Basin Business Unit and Highland Pump Co. personnel for their assistance in collecting these data.
REFERENCES
- Moineau, R.J.L., "A New Capsulism," Doctoral Thesis presented to the faculty of science of the University of Paris, 1935.
- Lea, J.F., Anderson, P.D., and Anderson, D.G., "Optimization of Progressive Cavity, Pump Systems in the Development of the Clearwater Heavy Oil Reservoir," Paper No. 87-38-03, Petroleum Society of CLM, Calgary, June 7-12, 1987.
- Lea, J.F., and Minissale, J.D., "Efficiency of Artificial Lift Systems," Southwestern Petroleum Short Course, Lubbock, Tex., April 22-23, 1992.
Copyright 1993 Oil & Gas Journal. All Rights Reserved.