Larry H. Flak
Wright Boots & Coots
Houston
Various types of reactive materials, or gunk, can react directly with produced fluids (oil, condensate, or brine) or with an additionally injected fluid to form a plug to kill blowout wells or shut off large flow paths. Several recent blowouts were successfully controlled with reactive plugs; other conventional methods would have been more difficult operationally and cost more.
Several plug mixtures are available on the market and can be made to suit the type of application and any particular environmental concerns. With proper planning and application, reactive plugs should be considered as a prime well control method when injection into the blowout flow path is available. This method of blowout control can save significant time and expense.
Most reactive plugs were initially developed for lost circulation applications, but they also have excellent application in blowout control, particularly for underground blowouts. The first reactive squeeze materials, developed in the 1940s and 1950s and still in use today, consisted of fast-setting cement or CaC12 brines mixed with sodium silicate based muds. The reaction of sodium silicate with any Portland cement or CaC12 brine results in nearly instantaneous solidification.
Extremely fast-Setting partially dehydrated gypsum/Portland cements were first used in the 1950s by Halliburton. In 1961, W.C. Goins and others at Gulf Oil Corp. developed gunk: bentonite-diesel oil (BDO) and bentonite-cement-diesel oil (BCDO).1 Later, invert gunk (amine-treated clay in water) that reacted with oil was developed by both Goins and Mobil Oil. Then, Halliburton and Goins developed guargum-based gunk which reacts better with brines. Today, polymer blocks developed from complex cross-linked polymers are also used in plugging applications.
There are two basic types of reactive plugs:
- Hydration
- Chemical
Application of these reactive plugs in blowout control involves blocking the blowout flow path or greatly increasing the viscosity of the kill fluid introduced into the flow path (with resultant high dynamic pressure drops at relatively lower pump rates).
It is surprising that with a 30-year history of gunk use, the excellent suitability of gunk to many applications in blowout control remains unrealized. Suggestions of gunk or other reactive plugs as a blowout control option have often met unfounded resistance. Reactive plugs generally have only one major disadvantage - they must not be reacted except where solidification is desired.
Table 1 (3666 bytes) lists many reactive plugs, their formulation, application, benefits, and disadvantages. The reactive plug of widest application, lowest cost and greatest material availability is BCDO. This gunk is easily mixed in modern cement batch mixers and can be stored indefinitely, assuming there is good agitation to prevent settling. In one operation, a major broached blowout was plugged with 500 bbl of BCDO reacted with 16.0-ppg Portland cement.
INJECTION FLOW PATHS
Two basic methods of using reactive fluids are available, depending on the flow path: injection down a single flow path to react with produced fluids and injection through two independent flow paths to a target plugging depth (any combination of drill pipe, tubing, annular, and relief well flow paths).
Controlling the location and timing of the reaction is the key to using any reactive mixture. Missed steps in the plugging procedure can result in reactions in downhole paths or surface lines and will greatly complicate control.
SINGLE FLOW PATH
In blowouts producing large amounts of water or oil, reactive materials can be injected into the flow path to react with these produced fluids to form a plug:
- Invert gunk (275 lb/bbl amine-treated clay in water) can plug off oil flows
- BCDO (200 lb/bbl bentonite and 150 lb/bbl cement in diesel) can plug off brackish water flows
- Guar gum gunk (300 lb/bbl guargum in diesel) can plug off brine flows.
The greatest difficulty with this method of reactive plug use involves determining the blowout flow rate so that the correct mix rate is achieved for a good plug. Too little reactant (for example, less than 1:1 gunk to water ratio) mixed into too much flow will not accomplish plugging, but it may help achieve a dynamic kill through greater combined fluid viscosity. Too much reactant (for example, greater than 8:1 gunk to water ratio) will provide no plugging and little additional frictional pressure drop.
The flow rates can be determined through the following:
- Hydraulic modeling to estimate downhole flow rates'
- Flow tests to sample produced fluids
- Production logs to help estimate flow rates .3
The proper reactant mixture and appropriate pump rates can be selected, if produced fluid chemistry and blowout flow rates are known. Volume requirements can only be made on hole capacity and reaction time. Pilot tests are used with produced fluids to help define reaction time. In blowout control operations, rarely is less than 95 bbl of reactant initially mixed (based on near capacity of most cement batch mixers).
TWO FLOW PATHS
Planned and controlled mixing of reaction materials underground can be accomplished if there are two independent flow paths to convey the two materials to a mixing depth that makes the plug at a good point.
Two independent flow paths can be provided by combinations of drill pipe and annulus or casing and casing annulus (Fig. 1)(36365 bytes). Two independent flow paths can also be provided by snubbing pipe or coiled tubing (Fig. 2)(27685 bytes). Another option for two independent well paths is a relief well to intercept a blowout (Fig. 3).
PROCEDURE OVERVIEW
A gunk application typically starts with relatively high ratios of gunk to water (typically 4:1), and the "gunk rate" is slowed as squeeze pressure is seen. A constant injection rate of water (mud or cement) helps to monitor squeeze pressure development. As surface pressure is known and rate is constant, then any pressure increase on the water side of injection is directly related to increases in bottom hole pressure. Bottom hole pressure can be accurately determined with knowledge of frictional pressure drop at this constant rate.
Other reactant materials can also be controlled by adding less of one component or the other and adjusting pump rates. Gunk reaction is nearly instantaneous at ratios of 1:1. Starting operations with this gunk rate can produce a too-short plug for good plug strength as reservoir pressure builds up.
Gunk plugs are not high strength plugs unless reacted with cement. Cement and sodium silicate plugs are very hard. There is no thermal limit for BCDO gunk (the application in Fig. 2 (27685 bytes) had a bottom hole temperature greater than 300 F.). Other reactant mixtures must be pilot tested. Early squeeze off can always be a problem because excess reactant cannot be reversed out. If required, a spacer can be displaced to the reversing point to allow reversing. Alternatively, excess material can sometimes be bullheaded out or staged out as pipe is tripped out.
USING CEMENTS
The use of cement to react with gunk or sodium silicate requires simple chemistries that are fairly retarded, highly dispersed, and only modified with silica flour at high temperatures. Cement is mixed "on the fly" with modern recirculating cement mixers. Fluid loss control is not used because fast cement dehydration is desired.
Typical cements are 16.0 ppg Class H with 0.5% dispersant and a lignosulfonate retarder for an 8-hr pumping time. Reaction with gunk or silicate will quickly set up the cement.
For stronger plugs, higher density cements are used with BDO or BCDO gunk. When cement is used, it is very important that displacement occurs after the blowout flow path is squeezed off. Thus, the application of cement with gunk or silicate is for more difficult blowouts (such as a broach to surface, only one available kill attempt, large diameter flow path, or very short plug length).
APPLICATION STEPS
The use of reactant plug mixtures in blowout control requires the following steps:
- Properly isolate and clean out all storage tanks and lines.
- Use independent pumping plants and pump-in lines.
- Use cement batch mixers to blend gunk or invert gunk, and store large volumes in either multiple blenders or mud tanks with good agitation.
- Use at least 10 bbl of spacer ahead of gunk.
- Pilot test the reaction at surface at various mix ratios with actual reactant materials and fluids. (Caution: Never pump a material without knowledge of its non-reacted rheology, reaction speed, and final plug consistency.)
- Add viscosifier if settling is a problem (amine clay in diesel mixtures or prehydrated bentonite in water-based mixtures).
- Determine how cleanup will be handled with diesel-based mixtures.
- Add strong water-wetting agents to speed up reaction of oil-based gunk.
- If desired, add lost circulation materials and barite to gunk, at the expense of bentonite, cement, guar gum, or polymer.
- Never attempt to reverse out excess gunk.
ENVIRONMENTAL CONCERNS
In instances where environmental concerns limit the application of diesel, then sodium silicate is used with CaC12 brines or cement. Also, a gunk made with nontoxic oil is available (Table 1)(36666 bytes).
Other available gunks are acid soluble (98%) and less damaging to reservoirs. BDO, BCDO, invert gunk, and sodium silicate plugs are not acid soluble and can be very damaging.
CASE HISTORIES
The following case histories describe recent reactive plug applications.
HORIZONTAL WELL
Onshore Texas in the Austin chalk horizontal drilling trend, an operator stuck a horizontal drilling assembly in the lateral during underbalanced drilling (Fig. 1)(36365 bytes). The drill pipe was backed off just above the bottom hole assembly. After the back off, pressure was discovered behind the 7%-in. protective casing on the 10/4-in. surface casing.
Surface leaks in wellheads and knowledge that the annular cement top was below the surface pipe shoe depth indicated that an underground blowout was under way with possible surface blowout risk. Wellhead leaks were controlled with injected materials, and a temperature log was run in the drill pipe to determine possible underground flow. Temperature anomalies were found at 300 ft and 2,500 ft.
Fluid flow occurred from the horizontal well bore up the 4-in. drill pipe x 7 5/8-in. casing annulus and out: parted casing (based on collar log response) into the 7 5/8-in. x 10 3/4-in. annulus and then down to the 10 3/4-in. shoe at 2,500 ft. Flow then turned out immediately below the shoe into a sandstone.
The following plan was made to control underground flow
- Displace the drill pipe with freshwater for squeeze pressure monitoring.
- Displace the 4-in. drill pipe x 7%-in. casing annulus with 13-ppg freshwater mud, and establish a constant 2-bbl/min rate.
- Lead the BCDO gunk with 10 bbl of diesel down the 7 5/8-in. x 10 3/4-in. annulus.
- Follow the diesel with BCDO gunk down the 7%-in. x 10 3/4-in. annulus to squeeze off the annular flow path. Start the squeeze at 8 bbl/min (4:1 ratio of gunk to mud), and slow the rate as squeeze pressure occurs.
- When squeeze pressure occurs, either bullhead mud for kill down the 4-in.. drill pipe x 7%-in. annulus, or circulate mud down the 4-in. drill pipe to surface for kill, depending on the 1,500-psi pressure limit set on the 103/4-in. casing.
Before application, the initial drill pipe pressure was psi, indicating a flowing hot tom hole pressure less than water gradient. Initial injection of mud down the 4-in drill pipe x 7 5/8-in. annulus did not affect surface pres sure on the drill pipe, indicating that pressure was con trolled only by injection pres sure at the formation sand face. The initial surface pres sure on the 10 3/4-in. casing was 1,000 psi.
After gunk contacted water-based mud at the shallow casing part, the drill pipe pressure and mud injection pressure quickly rose. The gunk rate was slowed to 4 bbl/min. Displacement with diesel was started after only 35 bbl of gunk were pumped.
Annular squeeze off occurred after approximately 25 bbl of gunk were displaced below the parted casing. Gunk injection was shut down to prevent placing gunk into the 4-in. x 7 5/8-in. annulus. The pressure rise was dramatic after the annulus was squeezed off to the 1,500-psi surface pressure maximum set prior to the job.
The well was opened through the choke, and drill pipe was displaced with 13-ppg mud. A conventional circulated kill was accomplished without exceeding the 1,500-psi surface pressure limit. The well was dead and on a vacuum after circulation; the total control effort was less than 24 hr.
The operator needed to repair wellheads and parted casing after the kill operation, thus requiring removal of the blowout preventers. A plug had to be set below the damaged casing to allow a safe repair. Additional gunk was then used to make a 1,000-ft plug, up from the end of the 7 5/8-in. casing shoe.
BCDO gunk was pumped out of the drill pipe and mixed with mud pumped in the annulus at a ratio of 2:1 gunk to mud. The plug was tested, and the drill pipe was safely tripped out. The blowout preventers and casing spool were removed, the parted casing was pulled, and the casing was repaired.
The well was sidetracked at the top of the lost tools and completed as a horizontal oil producer.
HPHT WELL
Offshore Louisiana, a high-pressure, high-temperature (HPHT) producing well had an underground blowout (Fig. 2). The well was shut in with 11,300-psi tubing pressure just prior to catastrophic failure of the production tubing, production casing, and protective casing. Production logs indicated flow out of tubing and production casing holes at 14,850 ft and 10,000 ft up to sands located just below the 16-in. surface casing set at 4,469 ft. A dynamic kill attempt failed because of continued mud losses, disabling static control. Therefore, gunk was considered with a strategy using snubbed-in 1 1/4-in. tubing to provide two independent flow paths. The downhole flow rate exceeded 15 MMscfd of gas and 1,000 bo/d with a flowing bottom hole pressure of more than 10,000 psi.
Three BDO and BCDO squeezes were used to control underground flow. The final squeeze was with BCDO gunk. Because of limited tubular clearance, great depth, and high density kill fluids, injection rates were very limited. The 18.5-ppg water-based mud was pumped down the 1 1/4-in. x 2 7/8-in. annulus at 0.26 bbl/min (2:1 gunk to mud ratio) successfully squeezing off the annular blowout flow path. A total of 90 bbl of BCDO gunk were used.
This operation is believed to be the first application of snubbed-in tubing (to convey gunk to a point just below a tubing hole) and the outer annulus (used to convey the kill mud to the hole for the reaction). Had this gunk method failed, alternative control methods would have been extremely costly and difficult, given the well depth, productivity, and completion geometry.
BROACHED GAS BLOWOUT
Onshore Argentina, an operator had a surface blowout and fire during drilling operations (Fig. 3)(31089 bytes). An underground flow up a steeply dipping sandstone formation to a surface outcrop about 2,000 ft away complicated blowout control. The combined flow rate exceeded 200 MMscfd of gas.
The rig was removed, the fire extinguished, the blowout capped, and the flow diverted. Drill pipe (4 1/2-in.) was snubbed to bottom to attempt a circulated kill. Initial kill attempts failed because of high blowout flow rates, large hole diameter, and supercharging of shallow formation.
A relief well was spudded to intercept the blowout well bore directly, given the high flow rate and complexity of continued underground flow. The relief well intercepted the blowout within a salt bed cap rock just above the reservoir. A method was needed to block the flow path between the highly permeable reservoir and the shallow supercharged sand.
The initial step of blowout control required slowing down the blowout with weighted salt mud (to limit further blowout hole washout) by pumping down the relief well annulus. The relief well drill pipe was displaced with water to monitor bottom hole pressure. Water was injected down the blowout well annulus to increase back pressure and raise the flowing bottom hole pressure as much as possible.
Interestingly, a fracture gradient of I psi/ft was found in the shallow supercharged sandstone. After 3,000 bbl of salt mud were pumped at up to 80 bbl/min with little increase in flowing bottom hole pressure, pumping was slowed as sa.It-saturated cement was pumped down the relief well drill pipe and annulus at rates increasing to 60 bbl/min. A 40% molar sodium silicate solution was simultaneously pumped at 20 bbl/min down the drill pipe in the blowout well bore to react with the salt cement for a nearly instantaneous flash set. Approximately 500 bbl of cement were pumped before the blowout annulus was squeezed off, killing the blowout flow. Excess cement was circulated out of the relief well bore and the kill was monitored.
REFERENCES
1. Goins, W.C., et al., "Method and composition for scaling lost circulation in wells," U.S. Patent 2990016, June 27, 1961.
2. Smestad, P., Rygg, O.B., and Wright, J.W., "Blowout control: Response, intervention and management-Hydraulics modeling," World Oil, April 1994, pp. 75-80.
3. Slungaard, C., and Smestad, P., "Noise and temperature, logging used to determine underground flow path," Society of Professional Well Log Analysts European Symposium, Budapest, October 1490.
4. Flak, L.H., and Gloger, D., "Blowout control: Response, intervention and management-Case History", World Oil, July 1994, pp. 53-59.
