Douglas J. Wilson, Mark F. Barrilleaux
Exxon Co. U.S.A.
New Orleans
Prepacking and waterpacking techniques were successful in gravel packing the long, highly deviated completion intervals (up to 445 ft at angles approaching 70) in the wells on the Lena guyed tower, located in the Gulf of Mexico.
During 8 rig-years of operations, these techniques were used for both single and multizone completions.
While some operators consider this technology antiquated, field experience at Lena demonstrates that water packing results in high-performance completions.
LENA DEVELOPMENT
In 1982, Exxon Co. U.S.A. installed the Lena guyed tower in 1,000 ft of water, approximately 50 miles southeast of Grand Isle, La. The platform, placed in Mississippi Canyon Block 280, serves Blocks 280, 281, and 360.
Development drilling from 1983 to 1989 resulted in 93 completions in 56 well bores.
The completed intervals are characterized by finegrain, unconsolidated sands that are finely interbedded with shale. These Pliocene reservoirs, which have limited areal extent, dip at 20 to 40 from a piercement salt dome. Radial faults, dissecting the reservoir, add further complexity to the production strategy.
Field characteristics and development strategies determined that aggressive Sturn wells were needed to penetrate stacked targets, and "build and hold" wells should be drilled to reach objectives located up to 11,500 ft horizontally from the platform.
Well bore deviations averaged 45, with some approaching 80. Because of the high incident angles at reservoir contacts and thick sand packages, gravel packs were frequently installed across long completion intervals, some exceeding 400 ft.
Lena well depth ranged from 7,400 to 15,300 ft measured and 7,000 to 12,000 ft true vertical.
Well design typically included:
- 7-in. casing for single completions
- 75/8-in. casing for dual completions
- 95/8-in. production casing with a 7-in. liner for wells requiring protective casing due to high torque and drag or differential sticking conditions.
Each well was completed with the drilling rig immediately after production casing was Set.
DESIGN BASIS
At first, the completion designs were based on Exxon's previous operating experience and gravel pack modeling performed in the mid-1970s.1 Subsequent research 2 conducted with a more elaborate model reinforced the design basis and permitted a more detailed investigation of gravel placement.
This research was conducted with a 7 in. x 23/8-in. gravel pack completion model that was 25-ft long and contained up to 12 perforations/ft.
The model was fabricated from clear plastic so that visual observation of the gravel packing process could be made. The assembly could be rotated on its stand. The rotation allows simulating well bore deviations up to 11O from vertical.
More than 200 full-scale gravel pack tests were performed with varying well deviations, pump rates, gravel concentrations, and fluid viscosities. Carriers used for the gravel placement testing were either water or fluids with viscosity controlled by HEC (hydroxyethyl cellulose).
Based on these studies, numerous conclusions about gravel placement were incorporated into completion designs at Lena.
The primary conclusion was that water was a more universal transport fluid for gravel placement than were viscous fluids, particularly in long, highly deviated intervals. For this application, higher pump rates along with lower gravel concentrations and large wash-pipe diameters yielded the best results.
While gravel placement with viscous fluids has attractive operational advantages over gravel placement with water, tests with viscous transport fluids indicated that:
- The packs were commonly nonuniform and tended to settle with time. At deviations greater than 60, settling caused voids along the top of the pack.
- Areas around the coupling and blank sections of the screen were poorly packed.
- Formation stratification and high fluid leak could cause excessive node buildup on the perforations. The buildup could interfere with gravel placement.
Because of the combination of long completion intervals and high well angles, completion designs were standardized with water (completion brine) as the gravel pack carrier fluid.
The designs also specified prepacking the perforations with a viscous carrier prior to installing the screens. The prepack ensured complete gravel packing of the perforation tunnels.
PLACEMENT DESIGNS
The completion combinations included singles (one zone), single selectives and duals (two zones), and dual selectives (three zones).
Fig. 1 illustrates a typical dual-selective gravel pack. Multizone completions were designed to isolate independent intervals and to allow simultaneous or selective production from separate reservoirs.
A distinct advantage of the water pack over the slurry pack was the ability successfully to gravel pack without breaking long intervals into two or more short sections. As a result, the additional time, equipment, and complexity associated with multiple zones was only applied when reservoir isolation was desired.
Gravel pack design for a given interval within a completion did not vary significantly for different configurations (i.e., single vs. dual). The generic design, as illustrated in Fig. 2, was basically repeated as subsequent gravel packs were stacked.
In cases where clearances between zones were tight, the length of the gravel pack equipment above the screen and, if necessary, the length of the gravel reserve were shortened.
Each gravel pack screen was sized for a minimum of 1/2-in. radial clearance with the ID of the casing. Where possible, smaller screens were used to allow for additional gravel volume.
A dual completion inside 7-5/8-in. casing would typically have a 27/8-in. screen across the lower zone and a 41/2-in. screen across the upper zone (for concentric innertubing string clearance).
Each completion was installed with 6-gauge screen and 40/60-mesh gravel.
Continuous screen was run from approximately 10 to 20 ft below the bottom perforation to 80-100 ft above the top perforation.
Because water packing did not require a telltale screen, no blank pipe was run across the gravel placement interval. This reduced the probability of voids.
The additional footage above the perforations was designed to provide a gravel reserve in the event of pack settlement or gravel loss during gas injection. At greater than 60 (the approximate angle of repose for gravel), the reserve gravel was unlikely to resettle.
Even at high angles, however, the small cost of additional screen and gravel was inexpensive insurance against premature failure.
Prepacked screen, of the same dimension as the standard screen, was run opposite the top 40 ft of perforations. This provided a second line of defense in the event that excessive gravel loss consumed the reserve above the perforated interval.
Wash pipe, with flush-joint connections, was chosen to satisfy a 0.8 ratio of wash pipe OD to screen-tubing ID. The gravel pack model demonstrated that this ratio ensured desirable flow characteristics during gravel placement.
After gravel-packing in the inevitable presence of well bore trash, this ratio still left adequate clearance for removal of the wash pipe from inside the screens.
DRILLING AND COMPLETION
Well bore characteristics impacting completion complexity (angle, production casing size, etc.) were factored into the initial well design. Through to the objective formations, use of modified "build and hold" and S-turn directional designs increased drilling complexity, but reduced well deviation.
This trade-off facilitated completion operations, particularly when the well angle was brought less than 60.
Research data and operational experience indicated that 60 was a threshold angle for the following reasons:
- The gravel pack model demonstrated that at greater than 60, additional measures had to be taken to achieve effective gravel placement (these measures being incorporated into the Lena completion design).
- Cased-hole wire line operations were difficult, often requiring pumping tools downhole through the work string.
- Hole cleaning was difficult because of low-side settling, even at high circulating rates.
- Use of mechanical indicating devices, which relied on hook-load monitoring at the surface, was often difficult due to high drag in deep, high-angle wells. When high well bore deviation was unavoidable, completion design and field practices largely compensated for the increased complexity.
The importance of having experienced personnel at the well site during the completion of these wells, however, cannot be understated.
WELL BORE PREPARATION
The completion work string consisted of a dedicated string of standard 31/2-in. drill pipe. Although completion recommendations called for "pickling" the pipe to remove dope and scale, this was infrequently performed in the field.
Light doping of the drill pipe pin ends was carefully observed to minimize introduction of contaminants inside the work string. Swapout from drilling fluid to clear brine was preceded by a casing scraper run and with a seawater flush.
Care was taken throughout the completion process to ensure that the circulation system was clean and the fluid filtered.
PERFORATING
Tubing-conveyed, under-balanced perforating at Lena was comparable to that of most operators. Perforating guns with big-hole charges were used for gravel packing efficiency. Gun body sizes were based on wash-pipe dimensions to ensure that the guns could be washed over if necessary.
Low-side phasing was used for wells with hole angles greater than 45. This prevented excessive entry of unconsolidated sand into the well bore.
It is important to note that low-side perforating was not justified by concerns over the ability to fill the high-side perforations and annulus with sand. Model tests demonstrated that this was not a problem when water packing was performed.
The guns were run below a packer with a pressure firing head and crossover port to allow underbalanced perforating. Zones were generally perforated 500 psi underbalanced and allowed to flow 10-20 bbl (depending on interval length) to clean debris from the perforations.
Either a gas-cushion or a light-weight completion fluid was used in the work string to establish the underbalance. Guns were positioned with a sump packer and a snap latch, or by wire line logs.
PREPACKING
Prepacking was designed to fill the perforation tunnels with gravel prior to installing the gravel-pack screens. The gravel was displaced downhole through the work string in a viscous slurry and squeezed into the perforations until sand-out pressures were observed.
Viscosified completion fluid acted as the carrier to keep the sand in suspension until it reached the perforations.
Prepacking operations were performed with an open-ended work string. The well bore was cleaned thoroughly of gun debris. Then, to select an appropriate prepack rate, injection pressures were established at 1-2 bbl/min.
When possible, the initial injection pressures were at least 1,000 psi below the estimated formation fracture pressure. Prepacking at rates exceeding the formation fracture pressure forces the prepack gravel into the formation beyond the casing, leaving the perforations unpacked.
Prepack gravel volume was based on a ratio of 30 lb/ft of net perforations, plus enough additional volume completely to fill, across the gross interval, the casing with slurry.
The gravel was carried in a completion fluid viscosified by HEC. The viscous carrier was sheared and filtered, mixed with sand at 200 lb/bbl in vertical blenders, and then pumped downhole with a standard skid-mounted pumping unit.
Sand-out was reached at 1,000 psi greater than initial injection pressure. If this pressure did not occur, an additional slurry volume was calculated and the process repeated.
Some particularly tight sands could only be prepacked with a bradenhead squeeze method. (The slurry was displaced across the perforation interval and then squeezed into the formation.)
Cleaner sands would exhibit a more textbook response. In this case, injection pressure slowly increased as prepacking progressed and flow was diverted to deeper and less permeable perforations.
Well bore cleanup after prepacking was critical. In some instances, residual gravel from prepacking prevented the proper location of the screen assembly. At angles greater than 60, gravel settled to the low side of the casing and the drill pipe during circulation, even with viscous fluid.
It was found that constant rotation of the drillstring during reverse circulation was necessary for adequate well bore cleanup. The same process was effective for removing gun debris after perforating.
GRAVEL PACKING
The purpose of the waterpacking process was to fill the annular space between the screen and casing uniformly with gravel. The gravel packing assembly was run on the bottom of the completion work string.
Completion brine transported the gravel downhole and returned via the annulus. After sand-out, circulation was reversed through the pack to disturb any bridges that might have formed. The pack was then repressured to ensure proper sand placement.
The basic components of the water-pack assembly, as shown in Fig. 2, were similar to that for a slurry-pack, except that the service tool allowed reverse circulation through the work string.
With the gravel-pack assembly in position, circulating rates were established to determine baseline circulation pressures and to verify that no restrictions existed.
Surface equipment consisted of a high-pressure, dual-pot gravel injector and flow manifold. These were downstream of the pumping unit. The gravel injector was designed to feed gravel gradually into the completion fluid flow stream.
With the dual-pot configuration, the flow could be diverted to one pot while the other pot was refilled with sand.
More advanced gravel-pumping systems have since been developed which deliver more uniform gravel concentrations than pot systems.
Gravel was continuously circulated downhole in a 0.5 ppg concentration at a rate of 2 bbl/min. This pumping continued until gravel completely filled the annular volume across the screen and provided 5 ft of excess gravel above the screen.
As the initial volume was displaced down to the screens, intermittent batches of gravel were injected into the flow stream until sand-out occurred. This eliminated waiting for a full drill pipe displacement if the initial volume was insufficient.
Throughout the process, the pump gauges were monitored for pressure increases, and return fluid was sampled for traces of sand. When sand-out occurred, indicated by a 1,000-psi pressure increase over the baseline, the pack was "fluffed" by reversing a few barrels slowly through the gravel pack.
Fluffing was done to disturb any bridges that might have formed during initial gravel placement. The pack was then repressured with forward circulation to top off the pack. This process was repeated until fluffing produced no change in pack-off pressure.
Field experience indicated that this procedure was an efficient and less troublesome way of installing watercarried gravel packs.
The gravel placement process lasted from 2 to 10 hr, depending on interval length, Tight sands were packed with sand volumes slightly in excess of the calculated volume, whereas unconsolidated formations required as much as 50% excess gravel volume, possibly to fill void areas behind the casing.
It was not unusual to fluff and repressure the pack twice or more before no change in sand-out pressure was experienced. This indicated that the reverse flow cycle had some effect.
Pack-off pressures were not difficult to read without a telltale screen. Circulating pressures slowly increased as gravel placement progressed, followed by a final abrupt increase in pressure.
No logging was required to ensure the pack was suc-cessful.
PRODUCTION RESULTS
Because virtually all of the gravel packs in this field were installed using similar gravel-packing techniques, the effects of different techniques cannot be obtained. But these data can be compared by the reader with performances of wells in other fields, completed with different techniques.
From Lena's production results one can analyze the effect of different completions.
Two of the results are:
- Installing two (or more) packs in a single well bore saves considerable money over drilling separate wells.
- Installing multiple tubing strings and producing two zones at once increases current revenue.
However, the incremental risks due to the complex geometry are difficult to quantify. Will a gravel pack in a dual or other complex completion produce as well as a single cased hole pack? Is it more likely to fail prematurely? Will the resulting workover be riskier?
COMPLETION TYPES
Design parameters, such as interval length, screen size, and general completion complexity, vary widely. Therefore, the effects of these parameters can readily be evaluated. Having a data base of almost 100 gravel packs in the same field lends statistical validity to the conclusions.
In total, 93 gravel packs have been installed in 56 well bores. Eighty-three of the packs are either still producing or have depleted their reservoirs. Only five packs have failed, and only three of these failed before producing at least 500,000 bbl of oil. Three packs would not flow and two have not yet been placed on production.
Gravel pack performance has been very good despite the fact that many of the completions have been subjected to "one well turn around" gas injection/oil production cycling.
Twenty-nine of the well bores contain single packs, 19 have dual packs, and 7 have single selectives. There are 3 dual selectives and 1 twin selective (three packs and one tubing string).
This totals 59 well bores; 3 are counted twice due to workovers performed.
To evaluate the success of these completion strategies, maximum sustained rate and cumulative oil production were tabulated for each gravel pack. Net and gross perforation intervals were recorded so that rate and cumulative production could be normalized on a per-foot basis.
Additional information was gathered on the three packs that failed prematurely.
RATES AND VOLUMES
Table 1 lists some basic completion interval and performance statistics. It should be noted that many of the completions are still on production.
No attempt has been made to extrapolate performance into the future. Thus, cumulative production totals are conservatively stated.
There is a very wide range of productivities represented in the field. This makes it difficult to isolate the causes for the variation.
One major parameter that must be considered is the difference in formation properties between wells. This effect will not be completely resolved here.
Except for an acknowledgment that formation property variation skews the comparison among completion types, the variation will be assumed evenly distributed among the data.
The effect of completion type (singles vs. duals vs. single selectives, etc.) on performance is evaluated first. Figs. 3 and 4 show average cumulative production per foot and average maximum sustained rate per foot for each completion type on a total well bore basis.
On a per-string basis, the single wells outperformed the other completion types.
On a total well bore basis, the performance of the multiple configurations is comparable to the single wells' performance but not quite as good. This does not mean, however, that the installation of the multiple completions was a poor decision.
The decision to install a multiple completion can be based on a number of factors, such as accelerating production or reducing field life to reduce expenses.
The savings obtained by completing multiple zones all in one workover or completion effort is also important.
Pre-investing relatively minor amounts of money to install second or third packs in wells has allowed Exxon to complete zones economically with less reserves. These zones, on their own, could not justify the cost of completion or recompletion.
Smaller zones have less oil produced at lower rates because of low permeability or limited areal extent. These zones are profitable, however, and would not be tapped at all if not for multiple completion strategies.
This is a part of the reason for the discrepancy between production performance of singles and other completion types.
SCREEN CLEARANCE
Figs. 5 and 6 show average cumulative production per foot and average maximum sustained rate per foot for completions with various screen and casing combinations.
The question to be answered is whether annular clearance between the screen and the casing has any effect on well performance.
Casing-size selection will be affected by the desire for large annular clearances to allow higher productivity. This is especially true in multiple-pack well bores where upper zones must have large bores.
To increase annular clearances, casing sizes would increase which would cause drilling costs to rise accordingly.
As Figs. 5 and 6 show, gravel packs in a 27/8 in. screen x 7 in. casing outperformed all other combinations. Because there are only three 21,'8 in. x 5 in. packs, those data should be viewed with caution.
Previously, it was suggested that single wells outperformed other configurations. Therefore, it is reasonable to ask if the 27/s in. x 7 in. packs performed well because they were in single wells, or because they are somehow mechanically superior to the other configurations.
Forty-three of the packs are 27/8 in. x 7 in. About half of these are in single wells and half in multiple configurations.
Sorting these 43 wells by completion type shows that the cumulative production per foot and rate per foot from single wells is twice as much as from the dual and selective completions. The higher production rate is still unexplained.
Two factors could be influencing these results. Operationally, something is being done on the multiple wells that is causing them to produce less effectively, or completions are being made in less attractive zones.
As previously stated, completing smaller zones is one of the reasons for putting multiple completions in wells.
On the other hand, the lower zones of duals or selective wells are open longer to a well bore containing kill-weight fluid. Therefore, these lower zones are subject to more fluid invasion and contamination.
Both factors probably play a part in the productivity disparity.
Only single completions and lower zones of multiples can be made with the 21/8 in. x 7 in. design. Upper zones requiring an inner isolation string must be designed with a larger screen.
Forty of the packs were built with 4-in. screens in 7-in. casing or 41/2-in. screens in 75/8-in. casing.
With an outside screen diameter of 4.62 in., the radial clearance between the 4-in. screen and 7-in. casing is as little as 0.66 in., depending on casing weight. With such tight clearance, pack failure due to premature bridging and pack-off during gravel placement must be considered.
Two of the three premature failures were in 4 in. x 7 in. packs. However, there were extenuating circumstances. The other failure was in a 27/8 in. x 7 in. pack.
INTERVAL LENGTH
Figs. 7 and 8 show the relationship between net perforated interval and cumulative production per foot and between net perforation interval and maximum sustained rate per foot.
Although the data are quite scattered, there is clearly a general decline in rate and cumulative production on a per-foot basis as interval length increases.
This could be caused by shortcomings in any of the operational considerations. For instance, the underbalance perforating flow-back volume is based on well bore capacity rather than perforation interval.
Typical flow-back volume is 10-20 bbl. This volume might be insufficient for a 200-400 ft zone.
A 500-psi underbalance might be insufficient. But there are safety concerns that must be satisfied. These concerns limit the acceptable flow-back and underbalance levels.
In cases where zones are acidized, the diversion may be insufficient to stimulate all of the perforations.
Flowmeters detected that in several long-interval wells the lower permeability intervals were not producing.
A general reduction in rate across the entire pack would suggest a basic problem with gravel packing technique. However, this does not seem to be the case.
Work is required on diversion techniques to acidize the lower permeability zones effectively so that they can contribute to production.
It is recognized that some of the longer interval completions contain a higher percentage of poor quality sand. The poor quality portions might simply be incapable of significantly contributing to production, even when formation damage is not present.
HOLE ANGLE
Hole angles through the completion intervals ranged from zero (vertical) to 66. They were evenly dispersed through the range of angles. Twenty-five zones had angles greater than 50. The angle through the completion interval showed no correlation to flow rate or to cumulative production.
Hole angle effects were examined on the three prematurely failed packs. The failures were in wells with angles of 0, 20, and 37 from vertical. Clearly, high angle was not a problem.
FAILURES
In Well A-21 S the selective zone had an 88 ft gross (54 ft net) perforated interval at a 20 angle.
The screen was 4 in. in 7-in. casing.
The zone was perforated underbalanced, prepacked, and water-packed. The gravel pack was probably intact until opened for production by punching holes in the inner string across the perforated intervals.
Although low power hole-punch charges were used, screen damage from the concussion of detonation was probably responsible for the pack failure. After this occurrence, the procedure was changed to perforating above the screen interval across blank pipe.
In later wells, using sliding sleeves eliminated the need for perforating inner strings.
Well A-26, a single completion, was one of the few gas wells at Lena. The well was completed across 82 gross ft of perforations (44 ft net) at 37.
The screen was 27/8 in. in 7-in. casing.
The zone was perforated underbalanced, prepacked, and water-packed. The well flowed gas for 2 years at various rates before sanding up.
No special cause for the failure has been determined.
Well A-31 was a single selective well completed across 376 gross ft of perforations (122 ft net) in a vertical hole. The screen was 4 in. in 7-in. casing.
The zone was perforated underbalanced, prepacked, and water-packed. The well sanded up within hours of being lifted in.
A unique feature of this well was that a landing nipple in the middle of the screen assembly contained a 100-ft long blank section. The section was inserted for the purpose of shutting off water production in the future.
A void in the pack outside this blank section might have caused the failure.
WORKOVERS
Opponents of multiple-completion options have always pointed out that difficulties are likely when repairing a well with multiple open zones, or with complex tightclearance equipment stacked in the well.
In fact, several difficulties have been encountered on the workovers that have been performed.
It has been difficult to kill multiple wells when the zones have had different bottom hole pressures. The long intervals and high permeabilities have compounded this problem.
Experience in washing over tight-clearance packs, specifically 4-in. screens in 7-in. casing, has been mixed. In two attempts, one was successful. The failed attempt may have been caused by slightly collapsed casing.
Workover difficulty must be considered in the decision to use multiple-completion strategies.
Doing it right the first time has eliminated the need for most remedial workovers.
Where workovers have been required, extensive preplanning to mitigate the problems has been successful. Economic analysis of the various completion alternatives has proven that the selective and dual completions were the proper choice.
REFERENCES
- Gruesbeck, C.. Salathiel, W.M., and Echols, E.E., "Design of Gravel Packs in Deviated Wellbores," JPT, January 1979, pp. 109-115.
- Penberthy, W.L. Jr., and Echols, E.E., "Gravel Placement in Wells," unpublished Exxon Production Research Co. report, 1989.
Copyright 1991 Oil & Gas Journal. All Rights Reserved.