WELL CONTROL / INTERVENTION

Well control operations on a multiwell platform blowout

Two of six wells on a platform offshore India were successfully killed by simultaneous relief well drilling and intersection. Part 1 describes planning leading up to relief well drilling

Mark Mazzella and David Strickland, Cudd Well Control, a division of Cudd Pressure Control, Inc., Houston

 blowout and subsequent fire occurred off the coast of India in the Arabian Sea on a production platform during routine workover operations. A jackup rig had been employed to perform the remedial workover. During this operation on a well, flow was observed from a second well’s 20-in. surface casing x 13-3/8-in. intermediate casing annulus; a third well was experiencing flow from the 30-in. drive pipe x 20-in. intermediate casing annulus. And a fourth well exhibited pressure on each of its casing strings.

Attempts were made to kill the problem wells with seawater and mud. During these attempts, flow from the wells increased, requiring evacuation of the platform and rig. The gas flow ignited soon after the evacuation.

This article will discuss the well-control operations associated with the B-121 Platform well-control incident, and planning that rejected surface intervention in favor of relief well drilling. Part 2 will detail relief well drilling from two jackup rigs.

Introduction / Summary
The National Oil Company of India informed Cudd Well Control on March 11, 1999, of the incident. Well-control personnel reviewed the information and started to prepare preliminary well-control programs. On March 12, two senior well-control specialists were mobilized to assess the incident and discovered a production platform engulfed in flame. Four multiservice vessels (MSVs) were deluging the platform and rig with seawater, successfully preventing the fire from spreading to the rig, Fig. 1.

Fig. 1. MSVs and Sagar Ratna jackup at B-121 platform location.

Surface intervention analysis identified certain hazards associated with moving the jackup from the platform. While the MSVs’ deluge had prevented the fire from spreading to the rig, the seawater damaged some of the rig’s control systems, requiring drilling package skidding, jacking and deck crane systems to be repaired to allow the rig to be demobilized. Well-control personnel developed a plan to repair these systems in a safe / efficient manner, and the rig was moved to a safe location on March 20, eight days after the loss of well control.

Well-control personnel and Oil and Natural Gas Corp. (ONGC) officials developed procedures for surface intervention of the platform wells. During this phase, a support vessel capable of providing a stable work platform, sufficient firefighting support, adequate deck space and living quarters would be required. The existing jackup was determined to be the best vessel to conduct surface intervention operations from.

After detailed plans were developed for modifying the rig, the decision was made to forego surface intervention due to logistical constraints. An estimate of one month for the conversion was predicted – this, plus time estimated for actual surface intervention projects would be much greater than the time required for drilling relief wells.

During initial phases of the well-control incident, gas bubbles and discolored water were seen around the platform, evidence of some gas flow at the seafloor. Before any surface or subsurface intervention effort was attempted, a seabed survey was required around the platform, and a shallow-gas hazard survey was performed before spotting the relief wells. This showed presence of shallow gas at L1, L2 and L3 levels. Previous drilling in the area had not encountered any gas at these levels.

The drilling of relief wells was identified as a control option from the earliest moments of the project. Even if surface intervention were to be performed, relief wells would be drilled as the secondary plan. Planning for the wells was initiated within the first week of the program. Members of the operator’s drilling staff and well-control personnel initiated the planning. After a directional company was selected, their personnel were added to the project team.

Two wells were identified as the source driving the blowout and fire. Two relief wells were drilled and successfully intersected the two problem wells, and both blowout wells were abandoned with kill-weight mud and cement in each wellbore, according to national guidelines. The relief well project was completed 87 days after the loss of well control.

P&A operations for the blowout wells and the B-121 platform were scheduled to be performed following the monsoon season. This work was initiated in 1st Quarter 2000. The P&A project has provided some insights on the well-control incident and the well-kill programs – some of the findings are discussed later.

Well History Before Control Loss
The B-121 platform is a 4-pile, 6-well-slot platform located in the Arabian Sea off the West Coast of India. The platform is situated in the South of Bombay High field in about 250 ft of water and was used to develop LV and LVI limestone formations in two different fault blocks. The wells were drilled by the Sedco Forex Trident II jackup.

Drilling commenced on January 8, 1997. Four wells were completed in the LV or LVI formations, one was drilled through LV and LVI formations, but was determined wet and not completed. Only 20-in. conductor was set in the sixth well. All drilling / completion operations were completed by January 1998.

The jackup Sagar Ratna was mobilized to the platform on February 16, 1999, for workover. Several wells were experiencing annular pressures, and some subsurface safety valves (SSSVs) were malfunctioning.

The B-121 B well was the first to be worked over. It was killed with 13.4-ppg mud utilizing a coiled tubing unit, and tubing was pulled. On February 27, during operations on Well B, a gas leak was observed on Well F, from three points on the 20-in. casing. The leaks were at 90° spacing from one another, at points where bracing had been welded to the 20-in. casing.

A water spray was applied on the gas leak to lower the risk of ignition. A decision was made to complete operations on Well B, then move to Well F. Prior to completing Well B work, Well F was killed with 11.5-ppg mud. This was accomplished by circulating mud down the tubing and up the tubing x production casing annulus. The communication between tubing and casing on Well F was believed to be near the depth of the packer.

After circulating 11.5-ppg mud into the tubing and production casing of Well F, 120 bbl of 11.5-ppg mud was bullheaded down the B-section (9-5/8 in. x 13-3/8 in.) of Well F’s wellhead. At this point, flow from Well F’s 20-in. casing was stopped. Work on Well B continued and Well F was monitored continuously. The Well B workover was completed on March 10.

On the morning of March 10, a gushing of foam-type fluid was observed from the open-ended annulus of the 20-in. surface casing x 30-in. drive pipe of Well D. Water was pumped into the annular space, but the liquid level continued to fall. Pressure recordings were taken on all of the wells’ casing strings, see Table 1. Pressure observed on Well C was of concern and the decision was made to target this well as the priority workover. The first attempt to kill Well C was initiated on March 10, with 500 bbl of seawater bullheaded down the tubing.

This operation dropped tubing pressure to 900 psi and the production casing to 1,300 psi. During this operation, intermittent mud splashing was observed through the holes in the 20-in. casing of Well F. Since Well C was not killed with seawater, the decision was made to bullhead 12.5-ppg mud into the well.

At 9:15 hr on March 11, a significant release gushing mud and gas was observed through the open-ended annulus of the 20-in. x 30-in. annulus of Well D. After this was observed, the decision was made to evacuate the platform and Sagar Ratna. All tree and C-section valves of the wells were closed before evacuation to the two fire fighting MSVs.

Cudd Well Control (CWC) was contacted on March 11 and informed of the well-control incident. Personnel were placed on standby and travel arrangements were secured. The team reviewed the information prepared by ONGC and began to develop a preliminary well-control program. On the night of March 11, with the well-control personnel in transit, the gas flow ignited. Since appropriate steps had been taken by ONGC, there were no injuries.

Well Control Preparations
Two senior well-control specialists were mobilized from Houston, to Mumbai, India. On the morning of March 14, the initial assessment trip to the B-121 location was performed. Four MSVs were positioned around the platform, providing adequate fire suppression to prevent the fire from spreading to the rig, Fig. 1. A discoloration was noticed around and down current from the platform, possibly from seabed disturbances below the platform.

After landing on the lead MSV, water monitors were repositioned to enable a better view of the platform’s wellbay area. An assessment of the platform and rig determined a large fireball present under the structure’s top deck, with individual fires on Wells D and F. Condition of the other wells could not be determined.

The rig had experienced little to no fire-related damage; however, all of the watertight doors had not been secured and severe flooding of the lower rig compartments had occurred. The major compartments affected by the flooding were the SCR room, port-side jacking leg, starboard-side jacking leg, and pump, mud and store rooms.

Project organization, safety considerations. ONGC management ensured that all resources required were available. During the surface intervention program, meetings were held daily at ONGC’s command center to discuss the previous 24 hr’s operation, and future operations. Representatives from ONGC, the well-control company and critical service providers were present.

Additional CWC personnel were mobilized to the location at this time, including two well-control engineers, an additional senior well-control specialist and one well-control specialist. This selection was based on expertise in different areas of operations required to develop the extensive B-121 program.

Personnel safety is the number one priority during a well-control incident. Safety precautions followed during the initial surface intervention are outlined here:

• Number of personnel allowed on the jackup rig limited to capacity of the egress helicopter
• Cudd personnel would control fire suppression equipment located on the MSVs
• Work on the jackup to be performed during daylight hours only, and
• Two Cudd specialists to remain on location 24 hr/day.

Demobilization of jackup rig. The initial goal of the surface intervention program was safe / timely removal of the rig. This was performed in a systematic process with defined objectives and goals for each day’s work. Two teams worked concurrently on the rig and some of the work was performed simultaneously. Major goals / objectives were:

• Pump and siphon water from lower rig compartments
• Prepare emergency generator system for skidding operations
• Repair electrical equipment for jacking operations, and
• Repair crane systems.

The rig skidding was carried out on March 17; and it was moved off the platform on March 20.

Monitoring B-121 Platform
While demobilizing the rig, the condition of the platform was continuously monitored. Two CWC personnel remained on location aboard MSVs during the night to monitor conditions and oversee the water monitor positions. Gas bubbles and discolored (muddy) water were constantly observed under and around the platform. Gas flow from the seabed around the platform was a hazard that had to be addressed prior to developing a more intensive surface intervention program.

After the rig removal, condition of the wells could be assessed more thoroughly, Figs. 2, 3. Gas flow and fire could be seen from Wells D, C and F. A large gasified water and sediment flow was coming out of Well E’s 20-in. conductor. Water could be seen coming from the 30-in. conductor of all six wells. The 30-in. conductors of Wells E, D, C and A had subsided 3 to 10 ft.

Fig. 2. B-121 platform after Sagar Ratna jackup removed.	Fig. 3. Closeup of B-121 wellbay area, with large fluid flow from Well E.

During the night of March 20, a large explosion occurred on the platform. The following morning, a significant change in the platform’s condition was observed. Well D had a large vertical flow component. The resulting fire was about 60 ft in the vertical direction. Closer evaluation confirmed that the tree of Well D was not present. Flow from Well E had stopped, as well as flow from the 30-in. drive pipe of all wells. Gas flow and fire could be confirmed from two wells, D and C. In addition, no more gas bubbling or discoloration of water around the platform was observed.

Well D was flowing through what was remaining of the tubing spool, and from the flange between B and C wellhead sections. The wellhead appeared to have been raised above its original height. A noticeable gap was also present between A and B sections of the wellhead, but no flow was observed exiting this point.

Well C was flowing gas from between B and C wellhead sections. A large gap was present between these flanges also, and the top flange was at about a 10° angle. A large gap was also present between B and A sections of the wellhead. No gas flow could be confirmed from this connection. On March 25, continuous platform observation was halted. One MSV remained in the area to control access to the platform area.

Surface Intervention Planning / Rejection
After the jackup demobilization, surface intervention of the wells was thoroughly investigated. Condition of the platform would make this a difficult task, and major concerns of such a program were:

• Two wells were confirmed to be on fire. These fires had large horizontal components.
• Wells D and C could not be shut-in (after capping) without the other four wells flowing.
• The top deck of the platform had been badly damaged and would have to be removed prior to well intervention.
• Observance of discolored water and bubbles around the platform caused some concern for stability of the seabed under and around the platform.

Surface intervention efforts were to be initiated after fabrication of a suitable work platform could be completed. Major steps of the planned surface intervention were:

• Modify original jackup rig for supporting surface intervention.
• Cut off and remove top deck to allow access to well bay.
• Cut off wellheads of Wells C and D to allow vertical flow.
• Cap and divert each of the wells.
• Independently control each well from surface.

In view of these requirements, the surface intervention program was halted on March 25, based on condition of the wells and logistics involved with the jackup rig modification.

Relief Well Planning
Planning of the relief well(s) for the platform was initiated within the first week of the well-control incident. After initial assessment, relief wells were determined to have the greatest success rate. With gas / water flow from each of the six wells’ 20-in. surface casing x 30-in. drive pipe, the probability of a successful surface intervention program was small. If the wells could be capped, condition of the casing strings and flow from outside the surface casing made shutting-in the wells impossible. Thus the wells would require control from the bottom.

A relief well program would be successful in killing the blowout wells. The uncontrolled flow from most of the wells had been sourced from only one or two of the wells. Relief well(s) would kill the problem wells from the bottom, near the gas flow source. This would stop the uncontrolled flow from the other wells being driven by these wells. If a charged subsurface area existed below the platform, this pressure would bleed off slowly from the other wells.

Identification of relief well targets. Of the six well slots, four were producers at the time of the incident. Although Wells A and E were flowing from their 20-in. surface casing x 30-in. drive pipe annuli, the source of their gas was one of the other wellbores. The workover on Well B had just been completed and kill-weight fluid was still present in the wellbore. In addition, the tubing string and SSSV in Well B had been replaced and was considered still competent. This data supported the belief that Well B was not one of the problem wells.

This left three possible source wells, C, D and F. ONGC representatives at the time of the workover felt confident that the 12.5-ppg mud circulated down the production tubing and up the production casing annulus had killed Well F during the workover. If Well F was still dead, this left C and D as the wells to be targeted by relief wells.

During initial relief well planning stages, a flow change was observed at the platform. The tree section of Well D’s wellhead was cut-off due to erosional velocity and heat of the exiting gas and fire. This allowed the well to relieve itself to the atmosphere and produce at a higher rate. Immediately after Well D’s wellhead blew off, gas / water production from all of the 20-in. surface casing x 30-in. drive pipe annuli stopped. And reducing backpressure on Well D stopped the undergound flow of gas around the platform. Aerial surveys after this occurrence showed gas flow and fire from only Wells C and D.

Shallow seismic gas hazard survey. Prior to final spotting of relief drilling locations, a shallow gas hazard seismic survey was performed and analyzed. The survey was performed in three phases to expedite required results. The initial phase was to shoot the shallow seismic and digital survey over the proposed relief well locations. The second phase was to shoot the data as close to the platform as safely possible. The third phase would be to shoot the remainder of the grid, if required.

The first phase of the seismic survey was initiated on April 2, and shooting around the relief well sites was completed on April 6 – processing was performed concurrently. Preliminary analysis showed presence of gas at depths that corresponded to the LI, LII and LIII layers in the field. Gas had not been previously reported in these formations. No gas hazards were located under the relief well locations or along their directional plans.

Part 2 will detail relief well kill designs, actual drilling operations to intersect and kill Wells C and D, and well abandonment plans. 

Acknowledgment

The authors extend their appreciation to Oil and Natural Gas Corp. and Cudd Pressure Control, Inc., for permission to publish this paper. Numerous other personnel, international governmental agencies and service companies contributed to the success of this project, unfortunately there are too many to list. This article was prepared from a paper presented by the authors at the IADC International Well Control Conference, Dec. 6, 2000, Houston, Texas.

Part 1