FREQUENT SURVEYS GUIDE RELIEF WELL TO UNDERGROUND BLOWOUT

Humberto Uzcategui, Javier Cruz
Corpoven SA
Anaco, Venezuela

Fernando Rodrigues
Anadrill-Schlumberger
Caracas

Continuous planning and extensive downhole surveying were vital for a relief well that intersected an underground blowout in Monagas state in northeast Venezuela.

This relief well operation in 1989 was the first known application of anchored, oriented, tubing-conveyed perforating (TCP) used to force communication between two wells.

The wild well, Tejero 2E (T2E), was drilled to 17,003 ft total depth (TD) and was producing with a dual completion string from two zones: the lower Jabillos from 16,210 to 16,258 ft measured depth (MD) and the lower Naricual from 15,660 to 15,700 ft MD.

It was determined that the upper producing horizon was blowing, but there was some uncertainty as to the status of the lower producing horizon.

The Naricual is the principal producing reservoir in the area. The reservoir pressure in the Naricual was 11,670 psi while producing 4,139 b/d of 36.3 gravity oil and 34 MMcfd of gas through a 3/4-in. choke.

The well had been tested and completed and was ready to be put on-line on Mar. 12, 1989. At this time, the 9 5/8-in. casing pressure increased from 2,000 psi to 7,000 psi over 7 1/2 hr. It was decided to bull head 20-ppg mud down the well. During this operation the pressure increased to 7,000 psi. The packing seals in the well head allowed communication between the 9 5/8-in. casing and the 13 3/8-in. casing. This broke down the 13 3/8-in. casing shoe and caused an underground blowout.

A snubbing unit was rigged up and operating on the well on Apr. 3. The intent was to kill the well with any one of the existing production strings or to run a replacement string and kill the well from the bottom, if necessary.

While the snubbing unit operated on the well, two major ground eruptions occurred. The first eruption opened a crater 1,000 ft north of the well and destroyed a local airfield runway. This 300-ft long crater ignited and burned for 2 1/2 hr until Well T-2E was opened to the burning pit. The second eruption occurred when channeling hydrocarbons pressured up the surface sands of an old oil field (Pirital field) located 4 miles north of Well T-2E.

The wells in the Pirital field are shallow and had been declining for years. However, with this phenomenon over 24 of these wells caught fire. To solve this problem in the Pirital field, three vent wells (Well Nos. P-101, P-102, and P-103) were drilled to depths ranging from 3,000 ft to 5,500 ft true vertical depth (TVD) to relieve the pressure in the rejuvenated sands.

After several unsuccessful attempts to control T-2E from the surface with the snubbing unit, the operator planned a relief well that would parallel the 7-in. liner of T-2E for approximately 200 ft at a distance no greater than 2 ft from well center to center.

The relief well plan included the running of a liner and then the orienting and shooting of a large TCP gun. This method would establish complete communication between the wells by perforating both casings and destroying the 2 7/8-in. tubing string.

SURFACE LOCATION

Although a directional well had never been drilled in the area, it was considered the only feasible method of killing the wild well. A national road and a village in the immediate vicinity created some restriction in finding an appropriate location for the rig to drill the relief well.

Safety aspects involving a relief well, such as wind direction, fluid availability, and pumping capacities, were considered extensively before drilling commenced.

Because of safety concerns and the geometry of Well T2E, only the southwest quadrant was available as a surface location for the relief well (Fig. 1). A location almost due west of the T-2E location would minimize any turn that might be required for a trajectory correction after initial detection of the 9 5/8-in. casing in the T-2E.

The surface location for the relief well, Tejero 4 (T-4), was selected at a distance of 2,354 ft at an azimuth of 791'27" from T-2E. The location was constructed in an area of 650,000 sq ft.

WELL PLAN

The detection point of the blowout well had to be a depth at which, in case of a collision, a kick in the relief well could be overcome.

The bottom hole pressures, the reservoir characteristics, and the rig hydraulic capacity were considered to determine mud weight and flow necessary to kill a blowout in both wells. The penetration rate and formation characteristics were also considered because a sidetrack may have been necessary in case the wild well could not be detected on the first attempt.

The only position data available from T-2E were a multishot magnetic survey and an oil-based mud dipmeter log. The multishot films were checked by five people to avoid possible reading errors. The multishot data were considered bad quality even though the bottom hole location from the multishot surveys and the bottom hole location from the dipmeter tool were only 27 ft apart.

Based on these data, the ellipses of uncertainty were computed to have a radius of 37 ft at 13,058 ft TVD (the planned detection point) and 86 ft at 15,368 ft TVD (the planned kill point).

The plan was to drill vertically to 8,000 ft with an 8 3/8-in pilot hole to identify possible gas in surface sands. After this operation, the hole would be opened to 26 in. to 2,000 ft and 17 1/2 in. to 7,500 ft.

At 8,000 ft, a 12 1/4-in. hole would be kicked off with a build rate of 3/100 ft up to 35 inclination in a direction of N 80 E. A lead angle of 7 to the right was chosen because the walk tendencies of the formation were unknown.

Then, with a packed bottom hole assembly (BHA) and polycrystalline diamond compact (PDC) bits, the plan was to drill to 11,500 ft and start the drop off to 8 inclination directing the hole to 25 ft behind the center of the ellipse of uncertainty on T-2E.

WELL DETECTION

The position accuracy of the best survey instruments available is about 2.0 ft per 1,000 ft of survey. Thus, at the planned intersection point of 16,000 ft MD, the survey error could be as much as 32 ft. This survey error is too large to guarantee well bore intersection.

Magnetic ranging tools were used to compensate for this possible survey error, These tools determine the distance and direction from one cased well bore to another well bore. Because the operator did not wish to rely on only one method, both the current injection and magnetic poles search methods were used for ranging.

The current injection technique involves an active analysis (i.e., a signal is transmitted from the logging tool into the surrounding formation) of a magnetic field induced in the pipe located in the wild well.

An alternating current was injected into the formation from an electrode located in a bridle 300 ft above the logging tool. This current then induced an electromagnetic field in the casing of the wild well.

The passive magnetic technique uses a passive analysis (i.e., no signal is transmitted from the logging tool) of magnetic interference caused by pipe located in the wild well. The tool joints, casing collars, and casing shoes are anomalies that create an interference in the earth's magnetic field.

Both of these techniques were used frequently for course control during the drilling of T-4.

KILL PROCEDURE

The procedure to kill T-2E was planned as follows:

Drill relief well T-4 with an 8 3/8-in. hole to within 2 ft center to center of T-2E over a parallel course length of about 200 ft.
Run and cement a 7 5/8-in. liner with a 500-ft overlap in the 9 5/8-in. casing.
Modify the blowout preventer (BOP) stack and connect the pumping units.
Run in the hole with tubing-conveyed perforating (TCP) guns connected to 900 ft of 4 1/2-in. drill pipe with 5-in. drill pipe to the surface.
Orient the guns with an electronic gyro.
Fire the guns.
Monitor the mud in the annulus and the pressure in the drill pipe.
- If there is a small loss of mud, use the centrifugal pumps to keep the hole full, and trip the guns out of the 7 5/8-in. liner.
- If the mud loss in the annulus continues, pump 14.7-ppg mud with the rig pumps until the blowout well is killed.
- If the well begins to flow, close the BOPs and pump into the annulus to kill the well. Monitor the bottom hole pressure through the drill pipe.
- If the level of the mud in the annulus does not drop or the well does not flow, effective communication between the two wells may not have been established. Circulate two bottoms-up to check for gas, and trip out of the hole. The well will need to be perforated again with a different gun orientation.

Three pumping companies participated in the pumping operation.

One company provided six 700-hp pumping units, another provided ten 600-hp units, and a third provided two 450-hp units and one 1,200-hp unit. This gave a total of 12,300 pumping horsepower available.

The pumps were installed in two sets of manifolds. Each of the manifolds was connected to the BOP stack with its own high-pressure line. Two spools were placed in the BOP stack between the lower pipe ram and the blind ram. The two lines from the pumping units, the rig choke manifold, and a pressure gauge were connected to these two spools.

Seven 500-bbl mud tanks and three 700-bbl tanks were located behind the pumping units. Five centrifugal pumps were connected to the tanks to provide intermittent circulation of the mud to keep the barite in solution and to maintain the mud weight.

The mud plant, barite tanks, and electrical generators were set up behind the mud tanks. Water was supplied from a large plastic-lined pond behind the mud plant. An 8-in, pipe connected the rig mud system to the pumping unit mud system. Five additional tanks, located behind the right mud plant, contained 2,500 bbl of mud.

Fig. 2 shows the general equipment layout for the relief well location.

DRILLING OPERATION

The pilot hole was drilled without directional control-the main concern was to investigate the existence of gas. Although a packed BHA was used, the hole built angle to 3 inclination into the northwest quadrant by 5,500 ft (N 78 W).

Fortunately, charged sands were not found. It was assumed that all of the gas escaping from T-2E charged the Pirital field through a fissure fault running in a north-south direction.

The pilot hole was then plugged at two different depths, from 4,900 to 5,400 ft and from 2,000 to 2,900 ft. The 17 1/2-in. hole was started with a packed hole assembly, and single shot surveys were taken every three joints.

SEARCH POINT

The build-up section started at 8,007 ft. A steerable system was run in the hole with a low-speed, high-torque motor with a 3/4 bent housing and a built-in 12 1/8-in., near-bit stabilizer.

A measurement while drilling (MWD) tool was configured for a high rate of data transmission of 3 bits/sec. This sent a tool face measurement every 6 sec for directional and mechanical control of the high-torque motor. By the end of the kickoff and build section, the well attained 35.80 inclination in a direction of N 80.90 E. This interval from 8,007 ft to 9,150 ft was drilled in 67 hr.

Because of the success of the kickoff and angle build, the next section to reach the search point was drilled with a packed-hole rotary assembly with three stabilizers and only five drill collars, including the nonmagnetic collars and the MWD collar. The other drill collars were replaced with 40 joints of heavy-weight drill pipe to avoid mechanical problems.

This section was drilled with a PDC bit; the formation was a very high pressure shale.

When the hole needed to be angled to the right, the PDC bit was replaced with a high-speed journal bit. The purpose was to walk the hole to the right. The hole angled to the right at a rate of 1.30/100 ft while dropping only 3 in 1,600 ft.

At 12,495 ft, the steerable system was run back in the hole to begin the drop-off approach to the detection point. At 13,400 ft, the first magnetic range tool survey showed an inclination of 16.84 at an azimuth of 91.38. Because oilbased mud was used, the first run results of the ranging tool were not good.

Before the use of the passive device, fresh mud had to be circulated in front of the tool to displace any metal cuttings that might have been present. After the third survey, the wild well was located 11 ft away and 90 left of the high side of the relief well.

This placed T-2E on the very outside edge of the previously calculated ellipse of uncertainty.

WELL ALIGNMENT

After detection of T-2E and subsequently with the ellipse of uncertainty reduced from a radius of 37 ft to 5 ft, the well was replanned according to the new coordinates. According to the revised plan, a 113 right-hand correction run placed the relief well bore parallel to the trajectory of T2E at 15,368 ft (Fig. 3).

This is considered a "search and lock" method of detection.

The operator expected that the relief well would follow the same tendencies as the original well (the inclination difference was only 3). The correction run was made in 1,100 ft with an average dogleg severity of 1.63/100 ft. The mud weight was increased to 16.7 ppg during this section.

With well T-4 at 15,500 ft MD and the data from T-2E reviewed again, it was decided that there was no guarantee that the long production string was dead. The final objective was no longer the lower perforations, but instead a zone 200 ft above them. With this new objective, further planning was necessary to ensure the required casing-to-casing separation of 2 ft at the new target depth. The drill plan was adjusted accordingly:

Build up to 12 from 15,500 ft to 15,795 ft.
Run 9 5/8-in. casing.
Build to 14 and turn 10 to the right.
Drill straight for 120 ft.
Correct the well 12 to the left and drop to 9.

This plan would put the relief well exactly behind the wild well. The relief well reached total depth at 16,169 ft MD (15,393 ft TVD), and the distance between the two wells was less than 2 ft over a length of 130 ft. The wells touched each other at 16,102 ft MD. That point was picked as the center for orientation of the TCP guns.

A 7 5/8-in. liner was then run with no problems, and the shoe was set at 16,168 ft.

MUD PROGRAM

From surface to 7,500 ft, the well was drilled with 12-ppg lignosulfonate mud.

From 7,500 ft to 15,794 ft, a 16-ppg oil-based mud system was used. This mud inhibited the sensitive shales and provided lubrication for the pipe in the geometrically complicated well.

From 15,794 ft to TD, the mud system was converted back to a 14.5-ppg lignosulfonate mud. The mud was changed back to lignosulfonate because of its compatibility with the fluid to control and kill T-2E.

CASING PROGRAM

The casing program for the relief well differed from that of the T-2E.

The 30-in. casing was set at a depth of 200 ft as a security measure to prevent shallow gas problems. The 20-in. casing was set at approximately 2,000 ft, effectively separating the upper, slightly overpressured sands from the flowing sands below (those connected to the Pirital field). The 13 3/8-in. casing was set at 7,500 ft, deeper than the normal casing in the area, to ensure isolation of these same sands. The 9 5/8-in. casing was set at about 15,600 ft, in the top of the flowing Naricual sands.

Finally, a 7 5/8-in. liner was set at the bottom of the middle Naricual, thus protecting the hole during the kill operation and allowing the use of 6-in. or 6 1/2-in. oriented TCP guns. A standard 7-in. liner could not accommodate these large guns.

SURVEY PROGRAM

The instruments used in the survey program included the electronic rate gyro, MWD tool, and conventional magnetic single shot. Survey accuracy is extremely important in a relief well operation. Therefore, it was deemed necessary to have two completely independent survey techniques.

The gyro and the MWD tool were selected based on survey accuracy, proven reliability, and the ability to work with high bottom hole temperatures. The MWD tool provided surveys as well as the orientation of the bent housing motor and bent sub while drilling. The gyro provided a check of the MWD surveys.

Because the gyro is more accurate than the MWD tool, the gyro surveys were considered the definitive survey. The MWD tool proved to be accurate to 0.4 for azimuth and 0.1 for inclination, and the gyro proved to be accurate to 0.1 for azimuth and 0.05 for inclination.

The MWD tool was configured to provide high-side tool face updates every 6 sec, which facilitated the precise motor orientation work required by the well plan. An additional feature of the MWD system was that the data from the individual directional sensors were available following each survey. This permitted magnetic interference analysis when drilling in close proximity to the blowout well.

Single shot surveys were used for the drilling of the pilot hole and the 17 1/2-in. hole to a depth of 7,545 ft. A gyro survey was run in the 13 3/8-in. casing from 7,070 ft to the surface. MWD surveys were used while drilling and verified by the gyro at all critical points in the well.

Each gyro run established a new bottom hole location which was then used as a tie-in point for the succeeding MWD surveys. Gyros were run in drill pipe when the heat shield was not used or in open hole when the heat shield was required.

At the 9 5/8-in. casing point, a gyro survey was run all the way back to the 13 3/8-in. casing point. Thereafter, gyros were run more frequently because the close proximity of the blowout well caused some magnetic interference with the MWD surveys.

It was also important to get a survey at the bottom of the well to determine the directional tendency of the previous BHA and the requirements for the next BHA. The MWD survey point was 20-50 ft off bottom, depending on the BHA used.

A final gyro survey was run in the liner. The three gyro surveys run in casing were accepted as the definitive survey set.

The MWD and gyro surveys matched very closely and were well within published accuracy specifications for both tools. The largest discrepancy between the gyro and the MWD surveys occurred at the end of the 90-azimuth portion of the well. The bottom hole location based on MWD surveys was 9 ft to the north of the results from the gyro. This was expected because the well was drilled almost due east with 35 inclination in the northern hemisphere.

For the remainder of the well, the discrepancy between MWD and gyro surveys was less than 5 ft. The MWD tool experienced no failures as a result of the high temperature (265 F.).

KILL OPERATION

Seven 500-bbl mud tanks and three 700-bbl tanks were filled with 14.7-ppg mud. The five additional tanks behind the mud plant contained 500 bbl of 15.1-ppg mud each. The rig's active mud system had 700 bbl of 15.1 -ppg mud. The total available mud on location was 3,200 bbl of 15.1-ppg mud and 5,600 bbl of 14.7-ppg mud.

On the T-2E location there were 1,500 bbl of 11.2-ppg mud with lost circulation material for pumping down the tubing.

On Aug. 28, for the first attempt to control T-2E the TCP guns were positioned across the interval of 16,095-16,115 ft. The two 6-in. OD guns each had fourteen 300-g charges.

These guns were run into the hole on 3 1/2-in. drill pipe to ensure getting inside the liner. The shots were aligned along the gun in three rows with a 5 displacement between each row. The two guns were connected with the center rows displaced 10. After firing, communication was not established between the wells.

On Sept. 7, a second attempt to establish communication used 4 5/8-in. OD guns with a 15 ft spacing between the two guns. The TCP guns had seventy-two 27-g charges. The center row displacement was increased to 33.

Both the first and second attempts were not successful because the accumulated torque in the whole gun assembly, while allowing for initial orientation, did not allow for continued orientation as the guns were fired. The pumping of fluid may have also changed the desired orientation.

For the third attempt, a permanent packer was used with an orienting sub below. The gun assembly with a mule shoe stinger was run into the hole. After the packer was set and its orientation checked, the mule shoe was set inside of the universal borehole orientation device below the packer to anchor the string.

On Sept. 20, the guns were detonated, and immediately the relief well lost circulation.

Once the detonation bar was dropped, the pumping operation began through the drill pipe with 14.2-ppg mud at 1 bbl/min and 750 psi. The pumping lines had been pressured to 500 psi against the well, and the shock when the guns fired was easily noticed.

The drill pipe pressure dropped to 0 psi immediately. The well was observed for 5 min to confirm that communication was established between the wells.

Then, 14.6-ppg mud was pumped into the annulus at 14 bbl/min and 450 psi, and 14.2-ppg mud was pumped down the drill pipe at 5 bbl/min and 240 psi until the drill pipe was full with 286 bbl of mud.

The pumping of mud into the drill pipe was monitored to maintain the bottom hole pressure of 2,200 psi and to avoid gaining from or fracturing the formation.

The pressure on T-2E started to drop 20 min after the pumping started, and 4 hr later T-2E was statically dead. For an additional 7 hr, 15.1-ppg mud was pumped to ensure complete filling of both wells. Table 1 shows the pump rates and corresponding pressures for both wells during the kill operation.

The blowout well was then cemented through the relief well from the killing zone to surface. The relief well was completed and is now a good producer.

Some factors should be considered before starting a relief well operation of this kind:

The availability of data from the wild well is of utmost importance for the search and lock method. The relief well planning depends on the availability and accuracy of the positional data.
The most accurate and manageable surveying and ranging devices on the market should be used. MWD tools should be selected on the basis of their accuracy and their ability to provide fast tool face orientation updates, thus allowing more mechanical control over the steerable motor assembly.
Early in the planning process, a team should be formed which includes representatives from the MWD/directional drilling, gyro surveying, ranging, pumping, drilling, logging, and operating companies.
A well should be planned from "cradle to grave." The kill operation becomes part of the planned well; this may eliminate delays once the well is in position.
Daily or more frequent strategy meetings are vital on these critical operations. Open communication should be encouraged.

ACKNOWLEDGMENT

The authors would like to thank J. Rangel and J. Lake of Corpoven SA, R. Golindano of SAES, P. Slagel and D. Chilton of Anadrill-Schlumberger, D. Teggin of Dowell Schlumberger, R. Waters of Magrange, Gyrodata, and Vector Mechanics.