Electric submersible pump (ESP) and auto gas lift completions have overcome such production engineering challenges in the Stag oil field as high gas fraction, continuous slugs with a short frequency, large volumes of sand, rapid onset of water production and rapid reservoir pressure depletion.

The Stag oil field presented one of the most challenging environments in which a field can be developed using ESPs. When reservoir pressure depleted more rapidly than expected, a high gas fraction and continuous slugs with a short frequency occurred. Shortly thereafter, water production began, followed by large volumes of sand. A newly improved completion method allows natural flow via the tubing or the annulus, ESP lift via the tubing augmented by auto gas lift through the annulus, or conventional compressor-driven gas lift via the annulus. Measures were taken to overcome poor ESP run life caused by sand-laden slug flow. The improved completion has played a key role in raising field performance above the target level following poor performance in early field life.
The Stag oil field is in less than 151ft (46m) of water, 40 miles (65km) northwest of Dampier, Western Australia. It is expected that more than 40 million bbl of 19° API oil will be produced during the next 15 years. The reservoir, about 2,230ft (680m) below sea level and 66ft (20m) thick, is a poorly consolidated, Lower Cretaceous, M.australis glauconitic sandstone. Reservoir temperature is 125°F, and oil viscosity at reservoir conditions is 7 centipoise (cP). Well productivities are in the range of 5 to 30 STB/d per pound pressure.
The field development plan anticipated using five ESP-lifted horizontal production wells, with horizontal sections up to 3,000ft (915m) long. The wells were each expected to produce 5,000 to 6,000 STB/d initially. Three water injection wells also were planned. Eight oil-producing wells and two water-injection wells (Figure 1) are on a central production facility (CPF) exporting crude to a storage tanker via a buoyant mooring system. Two subsea water injection wells are tied back to the CPF.
After production began in May 1998, it became clear the field had a larger than expected gas cap. The producing gas-oil ratios (GORs) - now declining from an initial value in excess of 2,000 scf/STB - were significantly higher than expected when the initial completion designs were developed. As the gas was produced, and because of a poorer than expected response from the edge drive aquifer, the reservoir pressure fell more rapidly than originally expected. After 1 year of production, the average reservoir pressure had fallen from about 1,050 psi to 700 psi. Consequently, free gas fractions above 80% at the pump suction caused gas locking at the ESP. These high-gas fractions and horizontal section trajectories also caused terrain slugging in the wells. After the onset of water production in one of the wells, significant sand production began and then stabilized at 0.1% by volume - 870lb (395kg)/1,000 bbl - during several months.
Initial design
The initial completion design incorporated a shrouded ESP system in the wells' horizontal sections, with no gas venting facility. Special gas handling stages were installed in the pumps to precondition the gassy crude.
A mechanically actuated fluid-loss valve, designed to isolate the reservoir during workovers, was incorporated in the tailpipe, below a deep-set, retrievable packer. Chemical injection valves, for demulsifier and scale inhibitor dosing, also were included.
Vent packer
In an attempt to address the poor performance of the wells, conventional vent packer, high-gas ESP completions were installed. These incorporated a vortex rotary gas separator upstream of the pump to separate a portion of the free gas and exhaust it via the annulus. The free-gas fraction is reduced entering the special gas-handling and pump sections.
Vented packers with surface-controlled, subsurface safety valves (SCSSSVs) were included above the pump to allow isolation of the annulus flow path. A sliding sleeve was incorporated below the vent packer to facilitate natural flow while bypassing the pump. The tubing-retrievable, surface-controlled, subsurface safety valve (TRSCSSSV) was retained in the design about 100ft below the seabed.
Well performance
The vent packer completions increased production 15% to 20% compared to production with the original completion design. The potential of the wells was not being met, because the ESP's operation continued to be constrained.
Multiphase flow also occurred under auto gas lift via the annulus. A large pressure drop occurred across the vent packer SCSSSVs and in the 2in. annulus flowline at the surface. The gas slugs were not vented efficiently, which resulted in intermittent pump gas lock.
ESP, auto gas lift completion
The objective of redesigning the completion was to debottleneck the annulus flow path. When the GOR declines and natural flow is not possible, ESP lift is initiated. The debottlenecked annulus allows the vortex gas separator to operate more efficiently. This allows auto gas lift to occur, which produces liquid in the annulus and the tubing.
The debottlenecking was achieved by eliminating the vent packer and its associated SCSSSV. The shallow-set TRSCSSSV also was eliminated. In their place, a TRSCSSSV was installed below the ESP system. Placing the valve below the pump allows isolation of the entire system of tubing and annulus (Figure 2).
The wellhead spools, which have 2in. by 2in. annulus side outlets, will be replaced. New spools will have a single, 4in. side outlet that will minimize pressure drop at the wellhead and limit the potential for erosive damage. Annulus flowlines (4in.) also will be installed.
The ESP was relocated farther up the hole from the horizontal section in the deviated section. This allows more effective natural separation of gas and oil before entry into the ESP and more effective venting of gas slugs past the ESP.
The combined system is flexible and can produce naturally via the tubing, without the requirement of an open sliding sleeve. This is done via an auto flow sub (AFS) incorporated above the pump. This device is a flapper valve, which allows tubing-to-annulus communication. When the pump is being operated, tubing pressure holds the valve closed. When the pump stops, the AFS
valve opens, allowing natural flow to bypass the pump.
The AFS also prevents the settlement of solids onto the uppermost pump stages following a shut-in. The solids fall through the side port into the annulus, rather than settling onto the pump. Prior to the incorporation of the AFS, a sand trap valve was used above the pump to prevent settlement.
Conventional surface-compressor-driven gas lift may also be used. A casing-flow mandrel is installed above the AFS, which allows gas to be injected via the tubing so that gas-lift production occurs via the annulus. This approach is preferred to gas injection via the annulus and gas lift via the tubing, because production potential is up to 40% greater.
ESP sizing
When auto gas lift occurs, a nontraditional ESP sizing approach must be adopted, allowing the split of liquid production between the ESP and the annulus to be accounted. Software was developed to iterate tubing and annulus performance until the equilibrium match condition is achieved. In this way, the three liquid streams, below the pump and in the tubing and annulus, are represented. Due to the way in which GOR, watercut and reservoir pressure varies in Stag wells over well life, the range tends to be wider than in a more conventional application. A larger volume pump was selected, one with a flat head curve over a wide range.
The size of the gas separator was increased to allow improved performance at the higher in-situ flow rates.
ESP, auto gas lift performance
Combined ESP and auto gas lift completions have been installed in six of the eight production wells.
In one well, a definitive incremental production rate of 1,700 STB/d (60%) was achieved. The other wells are new or have had other remediation measures applied (additional perforations). Their behavior demonstrates that incremental production of 50% to 100% is the direct result of the improved completion design. ESP run life also has been improved by more than 100%.
Well A
Recompletion in Well A resulted in a 60% increase in total liquid production from 4,400 STB/d (35% water cut) to 7,000 STB/d. Intermittent gas locking in the pump was eliminated, and auto gas lift was initiated in the annulus. Separate well testing of the tubing and annulus indicated 40% of liquid production was via the annulus.
After a few weeks of production, the water cut in the well increased to about 75%. A second annulus line was installed and resulted in a 2,000 STB/d increase in the total liquid production rate. The rate subsequently stabilized at close to 11,000 STB/d. Despite the increase in water cut to 75%, the new completion allows the well to produce about the same amount of oil as it did when the water cut was 35%.
Well B
Prior to recompletion, Well B was naturally flowing through a completion without an ESP. After recompletion, the free gas fraction at the pump suction was about 90% due to low reservoir pressure. A production rate of 1,800 STB/d at 0% water cut was achieved by ESP and auto gas lift. The well potential under natural flow was higher, and it was possible to flow the well naturally via the AFS at about 2,500 STB/d.
As the reservoir pressure and GOR declined, the threshold condition was reached, and natural flow was not possible. ESP lift then was invoked at 2,400 STB/d. The free gas fraction at the pump suction had declined to 80%. ESP lift had not been possible in other wells in the field with vent packer completions when the free gas fractions had been as high as 80%. It can be concluded that the incremental production delivered by the combined ESP and auto gas lift system was 2,400 STB/d.
Well C
Productivity was poor in Well C, so perforations were added at the time of recompletion. The well was subsequently comparable to Well B, with a total liquid production rate of 3,300 STB/d. The pump suction pressure was the same as for Well B, though the free gas fraction at the pump intake was slightly lower at about 70%.
Wells D and E
These wells were new when completed with the improved completion design. They are more productive than most other wells in the field, and reservoir pressure is higher. Shortly after startup, they each were producing 6,500 to 7,500 STB/d.
Completion reliability
Following the increase in water production from 35% to 75% in Well A, significant sand production commenced. This stabilized at about 0.1% (870lb/1,000 bbl) during a period of several months. With the onset of water production, a reduction in capillary bonding and apparent cohesion between the sand grains occurs, with resultant sand production. It is expected this will be typical of all wells in the field following the onset of water production.
Remedial sand control of the wells was considered. This option was rejected because of the significant reduction in productivity that would probably be experienced if screens were installed inside the preperforated liners.
ESP system reliability
The ESP installed in Well A failed following 3 weeks of production. During this period, some 200,000lb of sand was produced.
Pump failure was initiated by sand-induced radial wear of the zirconia bearings in the pump and vortex rotary gas separator.
To improve abrasion resistance, zirconia has been replaced with silicon carbide bearings and bushings. The number of bearings has increased to one per stage in order to distribute the load more effectively.
Harder Type 2 Ni resist (30% Nickel) materials have been specified for diffusers and impellers to improve abrasion resistance.
The number of pump stages has been increased to enable operation at a lower frequency than possible with the previous ESP specification.
Mixed-flow-type pumps were adopted to gain more tolerance of abrasive solids than the radial pumps.
Solids transport prediction
As downhole sand control is not being adopted, and in view of the large quantities of sand that have been produced, it is important produced sand does not accumulate in the wellbore.
To investigate the movement of sand out of the wellbore, solids transport in horizontal, inclined and vertical tubing and annular multiphase flow was modeled. A multiphase-flow well performance model was built and history-matched. Sensitivities were run to predict mixture velocity and fluid property profiles during the range of possible production rates.
The threshold mixture velocities were calculated along the length of the wellbore. The risk of solids deposition in each section of the well was determined by comparing the predicted threshold velocity profiles with the predicted mixture velocity profiles.
The comparison indicated that, at the operating conditions in Well A , 95% of produced solids are being effectively transported from the 7in. liner into the 95/8in. casing section. Some settlement of the larger solids is probably occurring in the casing section. Deposition should be self-limiting, since the threshold velocity for solids lift is close to the mixture velocity. Above the pump, solids deposition is not expected in the tubing. In the annulus, where auto gas lift occurs, some solids deposition will likely occur. Unlike the deposition in the horizontal 95/8in. section, this will not be self-limiting. The well may be susceptible to annular blockage by sand bridging.
Forward artificial lift strategy
Notwithstanding the steps taken to improve ESP reliability, an environment in which high gas and sand are flowing under slugging conditions is fundamentally hostile to ESPs. The risk of failure remains high.
Gas lift is widely regarded as the most effective artificial lift method where lift gas is readily available and wells with high production potential produce sand.
Gas lift performance was determined for all wells and compared with ESP performance.
Gas lift deliverability was found to be poorer than ESP.
The minimum ESP run life to make gas lift uncompetitive was determined at between 2 and 6 months. Because this ESP run life is considered achievable, gas lift is not being considered, other than to ensure sufficient completion flexibility.
Conclusions
A combined ESP and auto gas lift completion design has been developed for wells experiencing a hostile operating environment. The design enabled production rates up to 60% higher than could be achieved with a conventional vent packer design.
The abrasion resistance of the ESP system also has been upgraded. Run life has improved significantly. In the well where the ESP failed after 3 weeks, the improved system has been operating for 9 months at press time.
Acknowledgements
The authors thank the Stag joint venture partners - Santos Ltd. and Globex Far East Ltd. - for permission to publish this paper. All views and conclusions presented are opinions of the authors and not necessarily views of the aforementioned companies.
The authors thank Chris Farrow, Heiko Morgenroth and Klaus Scherwitzel, all of Helix Well Technologies Ltd., for their work on reviewing the likely sand failure mechanism, multiphase solids lift simulation and gas lift performance simulation.
The authors thank Gordon Kappelhoff and his colleagues in Schlumberger Reda Production Systems.