Diagnostics & Kill Planning

The first step in blowout diagnostics is to gather as much archived data as possible about the well, field, reservoir, fluids, etc.  The next step is to gather data from the initiating event through escalation to the blowout incident, then finally gather data from the aftermath at the site after the blowout.  The diagnostics and kill planning team consisted of a coordinator (John Wright), two hydraulic kill specialists (one from Well Flow Dynamics) and support drilling, reservoir and production engineers.

Initiating Events were as follows:

Initial Kill Plan. The design of the first kill attempt was performed with the use of sophisticated two-phase hydraulic kill software. Two models were used, both designed to first simulate the blowout flow and then to evaluate hydraulic kill sensitivities on factors such as: hole diameter, flow path, oil rate, GOR, kill mud density, HHP requirements, injection pressure and kill rates.

Unfortunately little direct diagnostic data was available to the technical support group for evaluating the first kill (e.g. oil rate, GOR, flowpath, flowing bottom hole pressure, etc.). Estimates were made based on visual observations and offset reservoir and production data. Controllers were placed on the model for mud density, hydraulic horsepower and injection pressure of 15 ppg, 4000 HHP and 7000 psi respectfully. The mud density of 15 ppg was the heaviest weight that could be reasonably held in suspension with the given mixing system, 4000 HHP was the estimated effective horsepower of all the cementing trucks in country, 7000 psi was the highest pressure utilizing a 20% safety factor for burst on the existing tubulars.

The maximum estimated achievable pump rate was between 30 and 35 bpm. Simulations estimated that a 25 bpm pump rate of 15 ppg mud was capable of killing flowrates of 50 Mbopd and 60 MMscf/d of gas on the assumed flow path with a volume of less than 1000 bbls. If the maximum pump rate could be achieved an oil rate of 90 Mbopd and 140 MMscf/d of gas could potentially be killed. The kill rate was much more sensitive to increasing gas rate than to oil. Invert gunk recipes were successfully pilot tested but sufficient quantities could not be procured in time for kill attempt 1.  The team felt there was a 50/50 chance of success.

Initial Kill Attempt. A tremendous effort was made by the Control Team to prepare for the first kill attempt which occurred only five days after the initial action plan was put in place. 3000 bbls of 15 ppg mud  followed by 2200 bbls of 11.5 ppg followed by 2000 bbls of 10 ppg mud followed by 10,000+ bbls of produced water were pumped at maximum pump rates of  31 bpm. Pump pressures rose quickly and then leveled off after the first 1500 bbls of 15 ppg mud was pumped. No additional significant pressure rise was seen during the remainder of the mud displacement. The flow out of the crater never diminished significantly at any stage of the pumping operation.

Analysis of the pump schedule indicated there was high uncertainty as to whether the well was killed dynamically for a short time but due to high loss rates was unable to transition to a static kill or the well was in two phase flow and the gas rates were too high to be killed at 31 bpm. This uncertainty was due to the inability to measure bottom hole pressure directly while pumping and the large variation in surface pumping pressure due to small variations in flowrate and viscosity of the fluids. The high pressure kill equipment produced 60% of its rated HHP.

Kill Attempt 2 Action Plan.  It would take ten days to prepare for the next kill attempt as it required bringing in equipment from outside the country.  The action plan for the task forces were: 

Task Force 1:  Continue diagnostics to define blowout problem.  Plan next kill attempt.  Evaluate invert gunk as a reactive kill fluid.
Task Force 2:  Run PVT log, remove rig, re-attach diverters, install dual inlet pump-in tree.  Click here to see rig move.
Task Force 3:  Re-configure mud plant to handle invert gunk and larger mud volumes and rate. Build 2000 bbls of 10.1 ppg Invert Gunk, 3000 bbls of 15 ppg, 4000 bbls of 10 ppg, 100 tonnes of LCM, 1500 bbls of cement.
Task Force 4:  Procure and test new 12,000 HHP high pressure pumping plant
Task Force 5:  Initiate relief well plan, prepare surface locations for spud
Task Force 6:  Continue preparation of crude oil containment and recovery system, construct fire safe lagoon, site security, safety and civil works.  Click here to see oil lake.  Click here to see plan view of site.

Additional Diagnostics.  A pressure/temperature log was run in the blowout drillpipe on May 12th. It indicated the flowing pressure at the perforations was 2415 psi at a temperature of 240°F. The exit point of the flow was clearly indicated at 105 m directly below the shoe of the 18-5/8" surface casing at a temperature of 180°F. There were no other temperature anomalies that would indicate a significant down hole choke in the annulus.

This flowing bottom hole pressure was much higher than estimated by the preliminary diagnostics. Since there appeared to be no restrictions the flow rates had to be considerably higher than originally predicted. The hydraulic models were used to attempt to match flow rate with the log data. This could be achieved, however, by increasing the oil rate while keeping the gas rate constant, increasing the gas rate or increasing both. The maximum gas rate that could be killed through the drillpipe under the assumed conditions with 15 ppg mud was estimated at 160 MMscf/d at a pump rate of 40 bpm. The oil rate was insensitive and could vary from 40 to 100 Mbopd with little effect on the kill rate.

The oil rate was estimated to be between 40 and 60 Mbopd based on oil evacuation measurements and evaporation calculations. The gas rate was more difficult to estimate by visual observations due to the diameter of the crater.

Invert Gunk.  Due to the high oil rate of this blowout invert gunk was pilot tested to assess whether this plugging agent might assist the dynamic kill. The active ingredient of this fluid is amine treated clay of the type used to make typical oil base muds. This clay will hydrate when it is mixed with oil but not when mixed with water. It works in the opposite (invert) manner as a normal gunk pill. The amine clay is suspended in fresh water with pumpable characteristics similar to normal drilling mud (e.g., 10.1 ppg, pV = 15, yp = 30).

Large volumes of amine clay are required to mix invert gunk with a typical mixture ranging from 200 to 300 ppb. Pilot tests indicated if the invert gunk is mixed with oil in concentration from 0.5/1 to 4/1 (gunk to oil), the clay will hydrate and form a thick viscous paste with a consistency ranging from peanut butter to a solid oily mass of clay with little excess water in 1 - 2 minutes. This reaction ratio matched well with the estimated pump rates of 40 bpm and oil flow rates of 40 to 60 Mbopd. Enough amine clay was sourced in Europe to build ˜ 1750 bbls of invert gunk at 220 ppb in the given time period.

Kill Plan # 2.  The second kill attempt would be controlled by an injection pressure of 7000 psi (80% of burst) on the existing tubulars and 40 bpm rate. The horsepower plant would be derated to 60% based on kill attempt # 1 requiring a total of 12,000 HHP. The additional required high pressure pumps were air freighted into the country from Europe within seven days.

It was still not practical to increase the mud weight higher than 15 ppg with available mixing equipment and it could not be exchanged in the ten day time period planned for the second kill attempt. An additional 2000 bbls of active tanks would be required however to mix and store the invert gunk until the kill attempt.

Invert gunk would be the primary kill fluid and would be pumped first at a maximum rate of 40 bpm. If the kill was successful it would be displaced with water and then cement. If the invert gunk was not successful a dynamic kill attempt would follow with 3000 bbls of 15 ppg mud at maximum rate. This would be followed by 4000 bbls of 10 ppg mud and then water if necessary. If the kill was successful, enough cement was available to pump 1500 bbls of 15.8 ppg. LCM material was added to the mud in volumes of 5 to 20 ppb.

The first kill attempt was limited by existing in country pumping equipment, the second would be limited by the maximum rate 15 ppg mud could be pumped down the drill pipe.

Results Kill Attempt # 2.  The pumping sequence was: 1750 bbls of invert gunk at 41 bpm, followed by, 3000 bbls of 15 ppg mud at 40 bpm, followed by 4000 bbls of 10 ppg mud at 40 bpm, followed by 9.6 ppg brine at 40 bpm. 88 tons of LCM were mixed with the various mud mixtures. The invert gunk built pump pressure to 5800 psi and leveled off, a rise of 1300 psi after 1200 bbls of gunk was pumped. The crater showed a substantial reduction in flow but did not stop completely. The 15 ppg mud was started as described in the decision tree but the pressure did not hold as steady as the gunk. The pressure fell slowly during the pumping of the 10 ppg and 9.6 ppg brine.

It appeared that the invert gunk may have been close to stopping the flow when the pumps lost suction at the switch over to 15 ppg mud. This is based on observations at the crater and pressure rise from the start of pumping and a 2400 psi flowing pressure at the perforations. It also appeared the 15 ppg mud may have initially stopped the flow and was in a dynamic unstable loss-flow situation. It was even more difficult, however, on this kill attempt (than attempt #1) to calculate bottom hole pressure rise due to the uncertainty in pump rate and the pressure drop due to friction at 40+ bpm pump rates. Given the fact that similar conclusions were drawn at the end of kill attempt #1, the most likely scenario as of this date is the kill attempts have been unsuccessful either due to severe losses after the well achieves an initial dynamic kill or the flow rates of gas were higher than 160 MMscf/d.

Kill Attempt 2 Summary

During the first two kill attempts pressure gauges were run in offset wells producing from the same reservoir to monitor pressure.  The high gas rate was confirmed several days later after evaluating the pressure log data. A mass balance calculation based on daily pressure drop data put the gas flow rate in the range of 500 MMscf/d.

It was not possible to hydraulically kill this flowrate using the existing tubulars.  The only alternatives were relief wells, which would require a minimum of sixty days after spud or other surface attempts involving snubbing or exotic reactive fluids which would require greater than ten days to evaluate and prepare.

Evaluation of Further Surface Control Options

The Source Control Team had determined the safest method with the highest probability of success for controlling the blowout at this point was with relief wells. Additional surface control methods were evaluated however. The criteria for evaluating those options were as follows:

1. Safety: Any operation that has a high or unknown risk of injuring the intervention team should not be attempted.
2. Deterioration of current well integrity: Any operation that has a high or unknown risk of causing further well deterioration (e.g. destabilizing crater, damage to wellhead or BOP, damage to drillpipe), should not be attempted.
3. Jeopardize relief well success: Any operation that would jeopardize the relief well success (e.g. removing drillstring electromagnetic target, losing ability to pump down drillpipe of blowing well), should not be attempted.
4. Any operation that does not have a high probability of success (e.g. 70% or higher) should not be attempted.

Heavy mud (18 to 21+ ppg):   Procedure: Pump 2000 bbls of 9.6 ppg water, followed by 3000 bbls of 18 - 21 ppg mud at maximum rate, followed by cement.

Analysis: Fits all kill criteria, however, mud mixing equipment for heavy mud was not in the country and the 7000 psi pressure limit on the drillpipe would be exceeded by the required 70 bpm flowrate now calculated to kill the well with 18 ppg mud.

Massive invert gunk kill:   Procedure: Pump invert gunk at 50+ bpm with an effect minimum volume of 3500 bbls. Follow gunk with cement.

Analysis: Fits all kill criteria, however, low availability of amine clay required 30 + days to deliver to country. Probability of success less than 50% based on the second kill attempt and higher confirmed gas rate.

Massive polymer kill:   Procedure: Design cross linking or linear polymer to start gelling in the annulus based on 240° F flowing temperature. Pump water at maximum rate for 2000+ bbls to slow flow and cool drillpipe, follow with polymer at maximum rate for 4000+ bbls, follow with cement.

Analysis: Would require pilot testing of various polymer mixtures and temperature evaluation of dynamic flow down drillpipe to avoid plugging pipe. Due to uncertainty in an unproven technique the team gave the probability of success at less than 50%.

Sodium silicate and cement (or gunk):  Procedure: Snub 5" drillpipe out of well, snub in 9-5/8" casing (bottom sealed with pump out plug), snub 5" tubing inside 9-5/8" casing, pump-out plug in 9-5/8" and snub 5" outside end of casing. Pump 3000 bbls of 15 ppg mud down 5" tubing at maximum rate, follow the mud with cement at maximum rate, ˜ 30 bpm. Pump sodium silicate down the 9-5/8" x 5" annulus at 15 bpm. Once annulus seal is formed, stop silicate and bullhead cement into reservoir.

Analysis: Chance of success if pipe is successfully placed: 95%, same as relief well. Risks involved: (1) safety of workers in snubbing basket is questionable, (2) inability to remove 5" drillpipe, (3) inability to snub 9-5/8" past hole in casing, (4) inability to get pipe back to 2880 m, (5) could lose relief well target if unsuccessful, (6) lose ability to pump down both relief well and blowout well drillpipe if unsuccessful.

Packer options:   Many different scenarios of snubbing in inflatable or up-setting packers below the hole in the 13-3/8" casing then inflating to stop the flow were evaluated. The well would then be bullheaded dead down the drillpipe followed with cement.

Analysis: Chance of success, unknown. Trying to set a packer in a high velocity flow stream in a 12-1/2" diameter casing will create a tremendous dynamic load on the packer, drillpipe (shut-in pressure is ˜ 2500 psi) and wellhead/BOP. Packer may also get damaged when snubbed past the hole in the casing. Risks: Safety of men in snubbing basket, casing fail below packer causing new crater. If packer is set just below the hole and properly designed, should not lose access to drillpipe if attempt fails. If packer is placed adjacent cemented pipe (˜ 1100 m) then pipe must be snubbed out. In this case you could lose access to drillpipe for ranging and pumping if you could not get down.

The time to mobilize a snubbing unit to country with associated BOP equipment is ˜ 1 week by airfreight. Time to be ready for a kill with option 6.3.1 is ˜ 2 weeks after rig-up for 6.3.2 ˜ 1 week.

Pump junk shot down annulus in attempt to plug hole in 13-3/8": Procedure: Rig up pump in lines to annulus, pump knotted kevlar rope, golf balls, hard rubber, etc. (exact mix to be determined, maximum diameter is 3"). If hole plugs bullhead annulus and pump cement down drillpipe.

Analysis: Hole in casing appears to be big based on flowing surface pressure. Not likely it can be plugged with junk < 3" in diameter. Risk: Considered low, however, some concern on affecting crater movement if partial blockage occurred momentarily and then released.

Same objective as snubbing only use a crane and strip the packer into the well with the annular.

Analysis: Same risks as snubbing with increased risk of dropping pipe with crane and losing it or damaging the BOPs and/or causing ignition.

Final Analysis

After evaluating all the potential surface control options the Source Control Team determined that none met all the criteria set-up to justify additional surface kill attempts.