Multilateral, underbalanced well drilled in mature Brazil field

April 7, 2003
A multilateral well drilled in northeastern Brazil was highly successful in a mature field that had already been under waterflood and steam injection.

A multilateral well drilled in northeastern Brazil was highly successful in a mature field that had already been under waterflood and steam injection.

The seven-leg well was drilled underbalanced with semishort-radius build rates in Carmopolis field about 35 miles northeast of Aracaju, Brazil.

The field has multiple pay zones with about 2.2 billion bbl oil in place, and to date, has produced about 390 million bbl of oil. During the field's 20 years, waterflood and steam injection have been used in various parts, with production now reaching about 19,000 bo/d and 115,000 bw/d.

The project

Click here to enlarge image

The field has complex geology and faulting. And, with the 1,450 wells already in the field on narrow spacingmake it difficult to find suitable locations to drill (Fig. 1). The producing formations have a low permeability (50-1,000 md) and a relatively heavy crude (20-21° API). Vertical wells are no longer economic in many parts of the field.

The operator chose multilateral technology to obtain needed reservoir exposure in all of the producing zones.

Well profile

Most of the vertical wells in the field are equipped with sucker rods and beam pumps.

Semishort radius (intermediate radius) drilling was considered to reduce drilling costs, minimize potential drilling problems, and maximize investment rate of return (ROR). The tools for intermediate radius drilling are generally more conventional (not articulated) but can still provide build rates of up to 60°/30 m in a 43/4-in. hole.

Multilateral system

Multilateral systems come in varying types and degrees of complexity. During selection of a multilateral system for any well, it is always best to use the simplest solution that will still meet all well-design requirements.

Because isolation between zones or near the junction was not required, the well was to be commingled during production.

On this particular well, the junction could be placed in reasonably competent sandstone. This allowed use of stacked windows (i.e., two to three windows spaced 2 m apart within the same joint of casing), which meant the junctions could all be placed closer to the reservoir to maximize drawdown during production

Underbalanced drilling

The well was drilled and completed underbalanced to prevent formation damage during drilling.

And, due to the depletion of the reservoir, underbalanced drilling was a sure way to minimize formation damage. In addition, the type of oil play involved, with the absence of gas in the reservoir, permitted a relatively simple underbalanced drilling system.

The team

A multidisciplinary team was to merge multilateral drilling, underbalanced drilling, and intermediate-radius drilling into one project. It consisted of geologists, reservoir engineers, drilling engineers, mud engineers, underbalanced-drilling engineers, directional drillers, casing and cementing specialists, fishing specialists, and health, safety, and environmental professionals.

The team worked through all issues regarding the merger of the technologies in order to minimize problems at the rig site. The team worked and reworked the drilling program before the rig arrived on location.

And the team constantly addressed safety and environmental issues throughout the project. On this well, it was particularly important because of the underbalanced drilling.

Main bore profile

The main bore well plan called for a conventional, low-inclination, directional well. Although the main bore could have been vertical, it was directional to simplify tool orientation while drilling out the windows.

With an 8° inclination at the junction depth, there would be no need for a gyro or steering tool to drill out the windows and drill away from the main bore casing. Therefore, with 8° inclination at the windows, conventional slimhole MWD tools could be used to drill out and drill ahead in each lateral.

Lateral profiles

Click here to enlarge image

The junctions were placed within the CPS-1 sandstone, a low porosity, low-permeability zone that was not expected to contribute to the well's productivity (Fig. 2). The number of productive zones in Carmopolis field required the drilling of at least four laterals to ensure each zone could be successfully depleted. Therefore, the well was planned with two stacked window joints, with two premilled windows in each joint.

Window placement was to be established after running wireline logs in the main bore.

The two lower laterals would be targeted to land in the CPS-3A at 90° inclination, and about 38-40 m below the window. They would then be drilled horizontally for 130 m.

The formation was dipping up and away from the wellbore, which meant the well would pass through the CPS-3A conglomerate, cross a thin shale, and then enter the CPS-3B conglomerate.

The two upper laterals would be targeted to land in the CPS-2, drill horizontally through the CPS-2, through another thin shale, and then enter the CPS-3A conglomerate.

Since the well profiles would be cutting through thin shale sections, the team decided to run slotted liners and drop them off in the openhole sections of the horizontal legs.

This would ensure that the shale sections would not slough and hinder production of the lateral legs. The tops of the drop liners were also equipped with reentry funnels to facilitate access for future coiled-tubing cleanouts.

Drill pipe fatigue

Intermediate radius curves with build rates of 45°/30 m, required the drillstring to be subjected to high cyclical loading while operating in rotation. The number of alternating stress cycles that each drill pipe joint would be exposed to needed to be controlled. In order to monitor and control, however, one must know how the drill pipe was used in the past.

Since this information is all but impossible to obtain for a used string, the team purchased for the project 60 joints of new 27/8 in., 10.40 lb/ft, S-135 grade drill pipe with 27/8-in. SL-H90 connections.

During the project, each joint of drill pipe subjected to cyclical loading was carefully identified and tagged with a serial number. While drilling the horizontal section of each lateral, the drill pipe was monitored for the total number of revolutions each made as it was being rotated through the high dogleg section of the well profile.

The drillstring was rearranged for every lateral so that each joint of drill pipe did not experience excessive numbers of cycles.

Main bore drilling

The team conventionally drilled the main bore pilot well (overbalanced), and took special care with regards to the section with the window joints.

The primary objective of the main bore was to obtain pay-zone information for the placement of the premilled windows, as well as provide a good place for the artificial lift equipment.

The well was drilled near vertical to minimize rod and tubing wear during the life of the well.

The well profile was built with an inclination of 8º at the anticipated depth where the windows would be installed.

Wireline logs indicated that the upper and lower double-window joints could be placed in the CPS-1 sandstone as planned. With the placement of the windows, the laterals could also be landed as per plan, using the same dogleg on the build-up curves for all four legs (around 45º/30 m).

This scheme meant that while drilling the laterals, the same motor settings and drilling parameters could be kept from one leg to the next, making each one of the build sections a repetition of the previous one.

Casing, cement

The 7-in. casing string needed an orifice for allowing communication between the interior of the 7-in. casing and the 7 in. x 95/8-in. casing annulus so that nitrogen (N2) could be used to achieve an underbalanced condition.

Since the well was very shallow, the orifice needed to be placed as deep as possible in the 7-in. casing string, while still staying inside the 95/8-in. casing shoe. A 60-m overlap was planned.

It was imperative that the orifice remain open after the cement job. The orifice had to remain open during the job while a good cement job across the window joints and the 60 m overlap was obtained.

The windows were placed about 40 m above the planned landing point for each lateral, making it possible to land each lateral with approximately 45º/30 m doglegs.

The windows were aligned as follows:

  • First window pointed at N45W.
  • Second window pointed at N45E.
  • Third window pointed at N90W.
  • Fourth window pointed at N90E.

The team drilled out the first two windows and drilled straight ahead on the correct azimuth for the remainder of the lateral. The third and fourth windows, however, required a slight turn of 30° to put them on the correct azimuth of N60W and N60E, respectively.

A gyroscope and a magnetic single shot read the window orientations. The casing was oriented simply by putting a little right-hand torque into the casing string and reciprocating until the applied torque was transmitted to the bottom.

Since the well was going to be drilled underbalanced, the N2 orifice had to stay open. Maintaining an open N2 orifice required a higher level of control on the cement job.

The casing was run with the orifice plate already open, and the cement job was performed with an inner cementing string and a retainer.

With a production head installed on the 7-in. casing, water was circulated down the 23/8 in. x 7-in. annulus, through the open N2 orifice, and back up the 7 in. x 95/8-in. annulus while the cement was being pumped into place through the 23/8-in. cementing string.

The added benefit of conducting the cement job in this manner was that now there was no need to pump a casing wiper plug past two window joints with four windows. That was one of the risks of running premilled windows.

Lateral No. 1 (lowest window)

The first step was to retrieve the two internal sleeves used to protect the windows during cement jobs (one internal sleeve for each double-window joint). The whipstock was set and locked into the latch coupling installed immediately below the lower window joint.

There is no need for a gyro to orient the whipstocks to each window. When it is properly set and tested in the latch coupling, it is already aligned to the first window and on the correct depth.

The windows were drilled in order of deepest first and shallowest last. Drilling the laterals in this sequenceprevented problems with drilling debris falling into windows which were already opened up. Open windows were always below the drilling whipstock.

The whipstock was assembled with double-latch assemblies so that the whipstock would not need to be pulled back to surface after Lateral No. 1 was completed. It was simply removed from Window No. 1 and set in Window No. 2 on the same trip.

Two bottomhole assemblies drilled each leg. The first drilled out the composite wrap covering the window opening, and it was also used to drill the 45°/30 m build section of the lateral. The only problem was loss of measurement-while-drilling (MWD) signal when the tools encountered greater than 55°/30 m doglegs.

Simply continuing to drill ahead and getting the tool through the high dogleg area (usually only a few meters) corrected the problem. As soon as it was clear, it would start working again.

The second bottomhole assembly was set to give an 8°/30 m build rate and was used to navigate and drill the horizontal section.

The remaining 120 m of available horizontal section was distributed between the two zones assigned for each leg. The total displacement allowed for these wells was about 160 m due to a fault that extends from northeast to northwest.

The build section of the lateral consumed about 40 m of horizontal displacement, leaving only about 120 m of horizontal section to navigate within the targeted zones.

This first leg landed horizontally in the CPS-3A and then extended to the CPS-3B by crossing a thin shale which separates both zones.

Click here to enlarge image

Ultimately, 184 m of open hole was drilled. Table 1 shows the resulting reservoir exposure.

There was 104 m of reservoir exposure, but just 36 m of reservoir exposure in the targeted zones (CPS-3A and CPS 3B

A 27/8-in. perforated liner with a re-entry collet (guide funnel) was run and dropped off 1 m below the bottom of the window to ensure well integrity over the life of the well. The reentry collet at the top of the liner will allow future coiled tubing access if required.

During the 2 days of drilling the first leg, the well produced about 70 bbl of oil. This first leg was already producing more oil per day than most of the vertical wells in the field.

Lateral No. 2

The second leg was drilled the same as the first. The drilling of the curve followed the same pattern as the first leg and the problems with the super slim MWD tools were treated with much more confidence and faster, since everyone already knew what was happening.

There was a total of 131 m of reservoir exposure and 104.5 m of reservoir exposure in the targeted zones (CPS-3A and CPS 3B). That was much better than the first leg.

The remaining laterals were targeted to stay within a single zone since the formation dip was different than expected, and drilling through the shale layer separating the two zones of interest consumed more open hole than originally expected.

The produced oil during the drilling of the second leg was 163 bbl, and the contaminated mud discharged reached 320 bbl, showing, as anticipated, an increase in production rate with incremental length of reservoir exposure.

Lateral Nos. 3-6

Two openhole sidetracks were added to these two zones because the desired reservoir exposure was not achieved within the CPS-3A and 3B zones.

The first openhole sidetrack came from the third window, and the second openhole sidetrack came from the fourth window.

Therefore, the third and fourth windows each had two producing legs.

This meant that the multilateral well would now have six producing legs. The legs drilled as an openhole sidetrack were lined with perforated dropped liners, and the sequence of drilling allowed access to the lower zone.

The remaining laterals were drilled and the resulting reservoir exposure is shown in Table 1. With increasing reservoir exposure, the well increased its rate of production.

The well was drilled and completed within the expected timeframe.

After the well was completed, the remaining drill pipe life was calculated and each joint tagged with the information for future use.

The maximum fatigue life used on any single joint of drill pipe was 20.1% of the useable life. The average fatigue life used on the drillstring was only 11.9%.

Underbalanced drilling

The well was completed with a 31/2-in. tubing string, a sucker-rod pump, sucker rods, and pump jack.

The pump was placed at 679 m, about 4 m below the lowest window. This maximized the drawdown on all the legs.

Offset vertical wells in the field produce an average of 30 bo/d with varying degrees of water cut. The best unstimulated vertical producer in the field produces 62 bo/d and 13 bw/d (~20% water cut).

There is one near horizontal, high-angle well in the field which produced for about 1 year at 120 bo/d with 105 bw/d. Six months later, the production rate slowly declined to 62 bo/d and 155 bw/d.

Some of the vertical wells have been fractured. Of those wells, one of the best produced 215 bo/d with 62 bw/d. After 6 months, the production rate had slowly declined to 107 bo/d and 20 bw/d.

The average fractured well stabilizes at about 95 bo/d after 1 year. Injecting water in the surrounding wells maintains this production rate.

Without pressure maintenance, the production rate would continue to decline rapidly.

The multilateral well initially produced 235 bo/d and 50 bw/d. After 6 months, the production rate has slowly declined to 138 bo/d and 50 bw/d.

Although the multilateral well produced at higher rates than any other well in the field, it exhibited the same decline characteristics as the stimulated wells in the short-term.

First impression of the production data reveals that fracturing the existing vertical wells will maximize return on investment (ROI) when compared to all other technologies. However, the production rate of the multilateral well should remain higher than the stimulated wells.

More time is required, however, to verify this. If the multilateral well continues to outperform the fractured wells in the long-term, then there may be some additional value in incremental reserves recovered that have not yet been considered.

The authors

Dean Lee is a multilateral-special projects engineer for Halliburton's Sperry-Sun product service line. He supplies engineering and project management services to clients in South America. He started his career as a diamond driller in 1984. With 19 years of industry experience, Lee has held various positions, including a drilling and completions engineer in Canada and the US. Prior to joining Halliburton, Lee spent 2 years at Chevron's technology center. He holds a BS in mechanical engineering for the University of Saskatoon, Saskatchewan, and is a member of SPE Brazil.

Fernando José Brandão is coordinator of the engineering department for Halliburton's Sperry-Sun product service line in Brazil. With more than 20 years of worldwide industry experience, Brandão has held positions as a directional driller and a drilling specialist. During the past 10 years he has worked on special projects, including the Carmópolis multilateral project and ERDs.

Click here to enlarge image

Gabriel Paulo Gutierrez Sotomayor has been a senior petroleum engineer for Petrobras in Macae, Brazil, since 1983. He has held various positions as a drilling engineer in Brazil. He recently has been involved in all aspects of multilateral projects, including well engineer, design, and execution. He holds a BSc degree in eletronic engineering from the Rio de Janeiro State University and MSc degree in petroleum engineering from the State Campinas University, Brazil. ([email protected])

Click here to enlarge image

Humberto de Lucena Lira is a senior petroleum engineer for Petrobras in Brazil. He joined the company in 1983 and has held various positions as a drilling engineer in Brazil. His responsibilities include all aspects of well engineering, design, and execution. He holds a BSc degree in mechanical engineering from the Paraiba Federal University and MSc degree in petroleum engineering from the State Campinas University, Brazil. ([email protected])

Click here to enlarge image

Paulo Silva Filho is a Petrobras geologist responsible for work in Carmopolis field. He has 21 years of experience in onshore and offshore reservoir. ([email protected])

Other benefits from merging technologies

Normally, multilateral wells are drilled first and later the production rate is established after the entire well is completed. When that is done, each leg's production contribution is unknown, unless a well test is done on each independently.

Also, it is unknown if additional legs drilled really give incremental value to the project.

Perhaps one of the biggest benefits from merging multilateral technology with underbalanced technology on this particular well was the production information obtained while drilling the well. Although actual contributions from each leg could still not be determined with precise accuracy, the relative amount of contribution to production was clearly evident.

The table below shows the amount of oil produced while drilling each phase of the multilateral well. Although the production data while drilling Legs 3 and 4 and Legs 5 and 6 were placed together, the same production trend exists.

Click here to enlarge image

Because the drawdown was virtually the same throughout, these data would indicate that the production rate was directly proportional to the amount of reservoir exposure. It further indicates that this well could have benefited from even more reservoir exposure and longer horizontal legs; however, due to faulting in the area this was not possible.