Fiber optics, advanced technologies complete ERD producer/injector

July 5, 1999
Fiber optics and advanced completion technologies were successfully run in a Wytch Farm extended-reach drilled (ERD) well that is both an oil producer and water injector.

Fiber optics and advanced completion technologies were successfully run in a Wytch Farm extended-reach drilled (ERD) well that is both an oil producer and water injector.

BP Amoco plc (BP) operates the field on behalf of partners Arco British Ltd., Premier Oil plc, ONEPM Ltd., Kerr McGee U.K. Ltd., and Talisman North Sea Ltd. Baker Oil Tools supplied equipment and services for the completion.

The work showed that:

  • This type of completion can be run in one trip.
  • A well can have a retrievable lower completion string and an independently retrievable upper completion string.
  • A completion can include fiber-optic, well-bore temperature monitoring.

Project background

The field extends under the southern part of Poole Harbor on the south coast of England and out under the sea to the east about 6 miles. Early field development tapped reserves from the shallower, Bridport reservoir. The second development phase included the deeper Sherwood reservoir, which extends beneath the English Channel.

To tap Sherwood`s eastern reaches, BP considered production platforms and subsea development. Following months of evaluation and consultation, the initial development plan called for building an artificial island in Poole Bay.

In December 1991, however, the development partners felt confident enough with advances in horizontal drilling techniques and BP`s own extensive research and development programs to abandon the island scheme in favor of horizontal wells.

Development of the eastern half of the Sherwood reservoir started with four wells drilled from the existing F site on the shores of Poole Harbor and continued from a new well site, M, adjacent to the existing site. This marked the third phase of the Wytch Farm development.

Artificial lift with electric submersible pumps (ESPs) and water injection supported existing production. At first, only production wells were drilled into the eastern part of the Sherwood reservoir, and these were completed with ESPs in 95/8-in. casing. Producing intervals were mainly in 51/2-in. cemented and perforated liners.

The third-phase ERD wells generally have between 2.4 and 6 mile horizontal step-outs. Well cost has been relatively high for land operations.

The first ERD wells were all completed as producers. Then, to reduce costs and provide water injection to support the eastern end of the field, BP decided to have one ERD well bore for both oil production and water injection.

Equipment BP owned from an earlier North Sea project was used for the first combined injector/producer, M-10, completed in May 1997. The plan included running the completion in two trips and using coiled tubing between the two runs to deploy cross-flow sleeves.

The completion achieved BP`s goals for production and injection, but the installation took 15 days with 11 coiled tubing runs instead of the planned 2.

M-12 plans

Click here to enlarge image

During 1997, BP earmarked a second well, M-12, as a potential candidate for a combined water injection/oil production completion. The planned well trajectory included a long horizontal section through the oil reservoir that dropped off into the aquifer into which water would be injected (Fig. 1).

The planned upper (ESP) completion had oil being produced up the tubing while water was injected down the annulus. The lower completion, straddling the oil reservoir, was the reverse, with oil being produced up the annulus while water was injected down the tubing. Oil and water flow crossed over between the upper and lower completions.

BP planned to preperforate the producing interval and perforate the injection interval after running the completion string.

The "wish list" included:

  • Independent injection and production operations
  • Retrievable lower completion
  • Independently retrievable upper completion for ESP workovers
  • One-trip installation
  • No dynamic seals
  • No coiled tubing operations
  • Permanently installed well bore temperature monitoring
  • Maintaining the flow meter above the ESP.

Eliminating coiled tubing

Elimination of coiled tubing was relatively straightforward considering the basic completion requirements. The M-10 completion had been run in two stages, with the lower completion having two retrievable sleeves set in the upper production packer. One sleeve allowed the two packers to be hydraulically set at the top and bottom of the lower completion. The second sleeve facilitates cross flow.

The coiled tubing runs were required for swapping the sleeves before running the upper completion.

The M-12 completion plan called for running the completion in one trip, with the ESP above the upper packer. Because this technique blocked access to any sleeve, a tubing-retrievable cross-flow sub was positioned in the tubing string above the top packer.

This position significantly increased flow area because the full casing ID could be used rather than only the packer bore.

Click here to enlarge image

Fig. 2 shows fluid flow through noninterconnecting ports of the cross-flow sub, anchor, and packer.

The flow cross-over point and the upper packer are in the 95/8-in. casing just above the 7-in. liner top. The 41/2-in. lower-completion tubing is in the 7-in. liner.

Click here to enlarge image

With a tubing-mounted cross-flow sub, it was impossible for the hydraulic upper packer to be set with tubing pressure and for the hydraulic lower packer to be set until the upper packer was set and the annulus pressure diverted through the cross-flow sub (Fig. 3).

A seal mandrel run instead of the top packer and stabbed into the liner top would have separated the injection and production flow paths. However, the lower completion needed to remain in the well when pulling the upper completion, and the liner lap had to be protected from high injection pressures.

A dedicated upper packer would retain pressure and provide a fresh seal bore into which the lower completion could be anchored.

Packer selection

A short list of packers began the multistage process for packer selection. Pros and cons were generated. The packers had to set with minimal or no tubing manipulation, which meant running either hydraulically or compressionally set packers.

The upper packer selected was a high-performance, compression-set packer with a set of tie-back seals for the liner top, polished-bore receptacle (PBR). This packer did not require hydraulic setting pressure, although pressure to an hydraulic packer could have been supplied through a control line.

The packer had a larger bore than the standard big-bore hydraulic completion packers. This provided a larger annular flow area and more room for control line clamps.

One drawback was that the set-down weight required for packer pack-off was close to completion`s compression limit. This could have been overcome by packing off with the aid of annulus pressure. This pressure, at the same time, would also set the hydraulic lower packer.

Because of the completion`s complexity, however, the top packer was run on drill pipe, with the upper and lower completions being run together later. The lower completion passed through the upper packer.

Although this prevented a true one-trip completion, running the upper packer on drill pipe offered BP other benefits.

The seal bore and latch profile in the upper packer differed from the standard liner top PBR supplied, but having a true back-up packer required being able to set one packer against the liner top and, if necessary, the back-up packer against the first packer.

This, in turn, required two different seal mandrels. One sealed in the bottom of the 71/2-in. ID, 15-ft liner top PBR. The other sealed in the 73/8-in. ID, 6-ft packer seal bore.

To achieve this, both packers were standard products with 73/8-in. by 6-ft seal mandrels. Also separate, add-on 71/2-in. by 9-ft seal mandrels could be either prefitted in a workshop or fitted on the rig site.

Lower completion

The lower completion had to be retrievable in case of future remedial work. Therefore, a shear-release anchor was used to connect the lower completion to the upper packer, and an hydraulic permanent lower packer was selected with an extended upper seal bore and shear-release anchor.

The primary constraint on the lower packer was the 420 ft of tubing-conveyed perforating (TCP) guns below the packer. Shock loading from the guns could have released a retrievable completion packer.

During the planning stage, BP was not certain about the length and weight of the guns. The shear-release anchor in the lower packer had to carry this unspecified tail pipe load and allow tubing pressure to set the packer before shearing.

An hydraulic actuated, shear-release anchor was selected to make the completion work less difficult. The lower completion design required a sequence of four hydraulic-actuated events, as follows:

  1. Set packer
  2. Release anchor
  3. Open tail-pipe flow path
  4. Fire annulus-operated TCP guns.

Shear-rating had to be carefully selected, particularly because the shear device for opening the tail-pipe flow path had to operate against an atmospheric chamber instead of a straight differential pressure.

An atmospheric-chamber-operated valve was selected because it allowed multiple opening devices to be run together to provide flow area. All these devices could be operated even if they did not open simultaneously.

The lower completion had to be retrievable for future remedial work and the most likely work was for shutting off water in the producing interval. Therefore, a permanent packer at the bottom was unlikely to cause interference.

Recompletion would require another hydraulic packer and valves set above the existing one. For this reason, the lower packer was set about 300 ft beyond the last perforation in the producing interval.

Fiber optics

Because the well could not be conventionally logged, a novel solution was required for determining water influx. BP had studied fiber-optic temperature monitoring with a view to performing field trials. Wytch Farm provided an appropriate place for the test.

BP selected the Sensor Highway fiber-optic system for the M-12 well. The system had been installed in nearly 100 steam flood wells in Canada and California. Wytch Farm was the system`s first time in lower-temperature environments.

The system suited the M-12 completion because the fiber could be installed and retrieved independently of the completion. This allowed temperature logging across the lower completion with an independently retrievable upper completion.

The optical fiber could be installed after running the completion string. All that was needed was a 1/4 in. U-tube conduit (control line) run with the completion across the production interval and back to surface. This required two control lines to be disconnected when pulling the upper completion.

On/off disconnect tool

For this completion, a tool was developed for disconnecting and reconnecting two 1/4-in. control lines between the upper and lower completions. The optical fiber requires that the conduit bore be smooth and without any gaps at the disconnect point. The on/off disconnect tool had to meet additional design requirements such as:

  • Being able to carry (deploy) the 5,500 ft-lower completion containing TCP guns
  • Allowing a straight-pull disconnect with no rotation
  • Providing a large annular bypass area for water injection
  • Accommodating tubing movement expected from a combined producer/injector well.

The 1/4-in. control line connectors had to remain stationary once the optical fiber was installed. This was achieved by making the entire on/off disconnect stationary.

Tubing movement calculations revealed that a stationary on/off disconnect could be achieved without undue tubing stress. The calculations took into consideration:

  • ESP above the on/off disconnect
  • Cooling and reverse ballooning during injection
  • Warming and internal ballooning during production.

To balance the tubing forces, some tensile capacity was required in the on/off disconnect because set-down weight alone was inadequate. A shear-release anchor was incorporated in the on/off disconnect. The shear rating had to be limited to prevent accidental release of the straight-pull, shear-release lower completion.

This limited shear rating could not carry the weight of the lower completion or the load from pressure testing across the on/off disconnect seal mandrel. Therefore, an hydraulically released interlock mechanism was incorporated, with the hydraulic pressure supplied down the 1/4-in. conduit.

The most reliable way to release this system was to do the release and disconnect during the completion phase. This operation was integrated with the space-out requirement for the upper completion that included:

  • Landing the lower completion
  • Releasing the on/off disconnect
  • Pulling back to check the release
  • Spacing out the upper completion
  • Landing off again
  • Latching the shear-release anchor in the on/off disconnect.

Anchoring the on/off

The male/female sealing connectors for the 1/4-in. conduit were necessarily short and had to be fully engaged when running the completion with the interlock holding the two halves together. The anchor could not be fully engaged because of the requirement to disconnect and space out the upper completion.

To solve this problem, the completion included a snap-in/snap-out (SISO), set-down-activated shear-release anchor.

This anchor can be run latched together but will unlatch and relatch repeatedly. It only fully engages the shear mechanism after weight is set down on it.

The weight shears a "commit" shear ring and moves the anchor mandrel down under the latch. The mandrel supports the anchor and prevents it from snapping out again.

The shear ratings selected for the SISO anchor were critical and required a compromise to achieve a workable system.

Too low a commit shear rating could cause the anchor-to prematurely engage when releasing the interlock or snapping in and out while spacing out the completion string.

Conversely, too high a shear rating would make it impossible to engage the shear ring because of compression limitations in the completion string.

The fiber conduit connectors had to compensate for the small movements of the anchor body during engagement. They had to maintain their zero gap without taking any completion compression loads. A spring-loaded alignment and stab sub met these needs.

Other issues

The problem of the alignment and stab of the fiber conduit connectors was also an issue.

The well profile and production string geometry did not allow for any string rotation either from surface or when engaging misaligned connectors. Therefore, one-half of the alignment mechanism was given limited ability to rotate around the tool.

Click here to enlarge image

Fig. 4 shows the on/off disconnect with optical-fiber feed through.

Click here to enlarge image

The fiber conduit passed from the upper completion annulus to the lower completion annulus through the cross-flow sub. This technique avoided complicated packer feed-throughs. The annular space between the packer bore and the lower completion tubing had enough room for positioning the control line and control line clamps (Fig. 5).

A formation saver valve was another important piece of equipment in the lower completion. The valve automatically shuts off when the ESP stops, thus preventing fluid-column backflow into the reservoir. The valve also prevents completion-fluid loss into the reservoir during an upper completion and ESP workover.

The formation saver valve was placed between the on/off disconnect and the cross-flow sub. The on/off disconnect, formation saver valve, cross-flow sub, shear-release anchor, and handling pups were all preassembled and tested.

Control line pig tails were prefitted to minimize control line installation work on the rig.

Friction drag

The long M-12 well profile required a study of friction drag during the completion work. BP had good records of actual hook loads from previous Wytch Farm completions. These loads were overlaid on theoretical drag plots.

The profile of actual vs. theoretical could be matched very closely by varying the friction coefficient in the theoretical calculations. BP showed that the friction coefficient was fairly constant for all Wytch Farm ERD work.

BP`s completion fluid contained a lubricant that kept friction factors at about 0.12. Hook loads for the M-12 well were calculated for various operations using the planned completion weights and dimensions with the planned well profile, and the common friction factor.

Two important considerations when looking at the drag figures are:

1. To pull out the completion in the event the on/off disconnect would not disconnect. This requires a large overpull to shear the shear-release anchor in the top packer. It also requires that the top end of the completion withstand the completion weight, shear load, and associated drag loads without over-stressing the pipe.

2. To know the expected hook loads during the completion, particularly when disconnecting from the lower completion.

Unfortunately, misinterpretation of the plots casued an error for the theoretical hook load after disconnect. This error was not discovered until after disconnecting from the lower completion.

Fortunately, the sequence of events leading up to the disconnect prevented a premature commitment of the SISO anchor in the on/off disconnect.

The ESP shroud hanger had to carry the loads of the tail pipe and pressure testing during completion work. The shroud hanger design and tensile information became available very late in the completion design process and revealed a weak point, that led to a review of the upper shear-release anchor shear value.

This also caused a problem during the tubing and shroud pressure test during completion.

Running the completion

Shortly before running the completion, the hydraulic shear-release anchor selected for the lower packer was found to have a potential problem. It had been selected over a straight-pull shear-release anchor because of the unspecified length and weight of TCP guns it had to deploy.

Like the lower packer, however, it could not be hydraulically actuated until the top shear-release anchor was landed in the upper packer. This would prevent release in the event the lower packer pre-set while running.

A quick investigation of the hydraulic shear-release anchor showed that the latch design and tensile rating was such that, if necessary, it could be separated from the packer without putting undue load on the completion above it.

Thorough post-drilling clean-up was critical because the completion process included running a packer across a long horizontal preperforated interval. Cleaning horizontal well bores is difficult, and BP had to resort to previously untried methods.

Following liner installation, BP discovered a liner lap leak. But because of the selection of a high-performance liner top packer and integral tie-back seal mandrel as the upper completion packer, the liner lap leak could be shut off during the first phase of the completion. However, one-half of the tough Teflon-based seals on the pack-off bushing on the packer running tool were lost when releasing from the packer.

The pack-off bushing was believed to have dropped out of the seal bore onto the latch threads in the packer top, ripping the upward facing seals off the bushing. An extended nose on the pack-off bushing would prevent this from recurring in the future.

In the M-12 well, BP decided to continue running the completion without attempting to clean out the seals and had no problem during the remainder of the installation or subsequent production.

The TCP guns and lower completion were run smoothly. The lower completion consisted of 41/2-in., flush-joint pipe with special slim-line, control-line protectors to maximize the annular flow area for production.

A test fixture was fitted to enable high-pressure testing of the conduit connections without releasing the disconnect. This was done after installing the on/off disconnect sub assembly at the top of the lower completion.

At this point, a pressure-relief valve was fitted to the control line reel and set at 3,500 psi to prevent accidental release of the disconnect while running the completion. The conduit was run with a 3,000 psi shut-in pressure.

A four-joint spacing was placed between the on/off disconnect and the pump shroud to allow for the slight offset of the 75/8-in. shroud caused by the control-line protectors. The 160-ft shroud consisted of five joints of flush-joint pipe.

The shroud hanger was installed in the top of the shroud, and the shroud and tail pipe were pressure tested down to the formation saver valve before the pump was run.

At this point, the combined load of the tail pipe and the test pressure just exceeded the design rating of the shroud-to-hanger connection, deforming the top of the shroud. These items were replaced with back-up parts before running the pump and reducing the test pressure.

After the pump was installed in the shroud, any tubing pressure tests would apply pressure to the outside of the motor lead extension (MLE) and motor pothead. Pressure needed to be bled off slowly to prevent damage to the cable cladding and insulation.

The main production string consisted of about 12,000 ft of 51/2-in. pipe with a 41/2-in. sliding sleeve just above the ESP and a 41/2-in. safety valve nipple at 160 ft below the wellhead.

The ESP included an Exal flow meter with its associated instrumentation cable. Two control lines, one instrument cable, and one power cable were run with the pipe between the ESP and safety valve nipple. An additional control line was run upward from the safety valve.

These lines prevented applying annular pressure to test the top anchor and set the lower packer after landing off in the top packer. The completion had to be spaced out and the tubing hanger made up and landed off before pressure could be applied to the annulus.

An overpull was taken to check anchor engagement after landing off in the top packer. The tubing then was slacked off to obtain a neutral point at the on/off disconnect based on the drag calculations.

Too much compression on the on/off disconnect could cause the shear-release anchor to prematurely engage, releasing the interlock. Too much tension on the tool during disconnect would cause the fiber conduit connectors to part with pressure on them. But the small conduit bore and fluid friction would prevent massive fluid discharge across the connector seals as they parted.

Drag calculations indicated that a hook load of 150,000 lb was needed for the release. After release, the actual pick-up weight was 175,000 lb.

The on/off disconnect was still in tension, despite a hook load of 25,000 lb less than it should have been. This was because the last action before releasing involved taking an overpull on the top anchor and because of the well deviation.

The completion was pulled back, and the space-out calculations indicated that a maximum 35,000 lb set-down load at the on/off disconnect would be needed to limit tubing stresses during production. The on/off disconnect required 25,000 lb set-down to engage the SISO anchor. It had been anticipated that this load would not reach the disconnect because of the well deviation. Therefore, application of annulus pressure after landing the hanger specifically to reverse balloon the tubing and engage the SISO anchor was programmed.

This step was required anyway to set the lower packer, release the lower anchor, and open the valves above the TCP guns.

The annulus pressure was required to engage the SISO anchor. This operation was checked by subsequently releasing the tubing hanger and taking an overpull on the completion.

To enable pump-down of the fiber when required, the optical fiber conduit had to allow pumping 1 l./min with a back pressure of less than 10,000 psi. In the M-12 installation, only 8,000 psi were needed for the circulation rate.

Optical fiber installation

After the completion was finished and the rig moved off, it was time to install the optical fiber. This marked the first time such a long length of fiber (33,000 ft) had been installed in a well.

At first, the fiber could not be pumped into the conduit because of oil in the control line and back pressure that needed to be overcome to pump down the fiber.

Water-based fluid is considered normal for fiber installation. However, a thin oil film on the control line inside wall was causing the fiber to stick when it contacted the wall. To overcome this problem, the control line was flushed with a solvent, then displaced back to water-based fluid.

Back force on the fiber, where it passed through the stuffing box in the injector, was overcome by using a winch system installed inside the pressure housing, downstream of the stuffing box.

The combined solutions enabled the fiber to be installed in the well successfully, so that a full temperature profile could be measured from surface to the lower packer at 16,730 ft, across the producing interval.

The well was initially used for production only and the TCP guns were left unfired. After this initial production period and commissioning of surface water injection pipe work, the annulus was pressure cycled to fire the guns and open up the injection flow path into the aquifer.

Temperature measurement

Click here to enlarge image

Installation of optic fiber allows well bore temperature monitoring at any time, from the well head to the bottom of the producing interval, without intervention or loss of production. Measurements are acquired every 1 m, with the entire acquisition process taking about 15 min. (Fig. 6 at right)

Recordings have been made with the well flowing, both with and without water injection, and during shut-in as the temperature begins to return to the geothermal gradient. Some of the profiles recorded are shown on the plot below.

When producing without injection, the heating from the ESP pump is clearly evident, as is the effect of gas breaking out of solution (with resulting cooling) as the flow moves up the tubing. The effect of injecting water, in addition to producing oil, diminishes the former and masks the latter effect on the temperature profile.

When production commenced a low water cut (about 10%) was also observed. Was this from a water finger found on logs, which had been cemented off, or from another source?

Temperature events along the producing interval correlated with zones expected to flow, but also indicated production occurring behind the unperforated section of casing where the water finger had been identified. Thus the temperature data analysis confirmed the source of water production to be from the water finger leaking behind the casing.

Temperature data taken when the well was shut-in suggests the water finger continued to produce (and cross flow into other zones). This interpretation is supported by the formation pressure tester data which shows the water finger has a higher formation pressure and permeability than the surrounding zones.

Acknowledgments

The authors thank the operator and Wytch Farm co-venturers for permission to publish this article. The views expressed in this article are not necessarily those of the operator or co-venturers.

The Authors

Click here to enlarge image

Russell Miller is a region engineer for Baker Oil Tools in Great Yarmouth, England. He has held a variety of quality assurance, operational, and engineering positions at Baker. Miller holds a degree in mechanical engineering and has U.K. Engineering Council professional engineering qualifications.

Click here to enlarge image

John Davies is currently the production technologist for BP Amoco`s Wytch Farm oilfield, responsible for the design, development, and performance of well completions. He has spent 19 years with BP Exploration in several petroleum engineering related posts. Davies holds degree in mechanical engineering and energy studies from the University College Cardiff.

Click here to enlarge image

George Brown is manager of interpretation development at Sensor Highway, developing interpretation techniques for permanently installed fibre optic measurements. He was previously with BP Exploration and Schlumberger. Brown holds degree in mechanical engineering.