Well Control And Intervention

CO₂ blowouts: An emerging problem

After thirty years of injecting CO₂ for secondary recovery, corrosion and pressure change properties of CO₂, and applications in aging wells, are creating conditions conducive to increasing blowouts during workovers / recompletions

Les Skinner, PE, Well Control Engineering Manager, Cudd Well Control, Houston

Blowouts on gas producers containing high concentrations of CO₂ have occurred for years in the industry. Recently, however, there has been an increased frequency of CO₂ blowouts in injection projects requiring intervention by well control companies. One company has responded to five such blowouts in the last three years, whereas the entire well-control industry only responded to two injection-related CO₂ blowouts in the previous 27 years since injection began.

This article discusses several physical and chemical properties of CO_2; blowout characteristics, unusual hazards, corrosion, preventive measures and methods for controlling wells in injection projects. Summaries of three of the five recent blowouts provide insight that will, hopefully, prevent the upswing in blowouts from becoming a long-term trend.

INTRODUCTION

Gas wells with high CO₂ concentrations have been drilled throughout the world for several years. Some of these have suffered blowouts during drilling / production operations including well-control problems on CO₂ source wells in New Mexico, Colorado and Wyoming. Producing wells with even moderate CO₂ concentrations have also suffered corrosion-related problems. This so-called "sweet corrosion" has been well documented in the literature. Recently, blowouts in CO₂ tertiary recovery floods have become problematic.

Continued injection has resulted in higher reservoir pressures in most projects, with CO₂ now at or above its critical point. The extreme expansion of this fluid when surface pressure control is lost, and the resulting violence of the blowouts, is astounding. Blowouts have complications that other blowouts may not exhibit, due to the characteristics of supercritical CO₂. Fighting a CO₂ blowout requires a change in thinking to properly manage these unusual conditions, although most intervention techniques follow classical procedures.

Whether this increased frequency of blowouts represents a trend is unknown. However, recent experience tends to verify that blowouts in CO₂ floods may, in fact, be an emerging problem.

CO₂ PROPERTIES

CO₂ is a stable oxide of carbon composed of two oxygen atoms, each double-bonded to a single carbon atom. CO₂ is soluble in most crudes. When dissolved, it swells crude oil, slightly increases in-place volume and reduces viscosity. It can displace oil miscibly or immiscibly, depending on oil composition and reservoir pressure.

Due to its weakly bi-polar nature, CO₂ is highly soluble in water, with which it reacts to form carbonic acid. Failures from CO₂-related corrosion can cause the loss of well control. CO₂ is not classified as a toxic material, i.e., it is not a poison. But breathing can cause severe negative health effects. CO₂ is more dense than air, with a specific gravity of 1.55 (Air = 1.0). It can collect in high concentrations in low areas such as depressions, pits and cellars.

Pressure vs. phase changes. Of greatest significance is the tremendous expansion of supercritical CO₂ when pressure containment is lost. Fig. 1 is a phase diagram showing its critical point at 1,071 psia and 88°F. Above this pressure and temperature, there is no distinction between liquid and vapor phases, and even small pressure drops can produce large volume increases, and vice versa. Minimum miscibility pressure for most CO₂/crude systems exceeds critical pressure; and reservoir temperatures are greater than the critical temperature for most, if not all, injection projects. Thus, most floods in which injection has been underway for several years contain CO₂ at conditions above the critical point.

Fig. 1. Phase diagram showing CO2 critical point.

When pressure containment is lost, two processes occur simultaneously. First, the CO₂ (and a fraction of miscible products) converts from a supercritical "fluid" to a vapor, with significant expansion. This vapor continues to expand with decreasing confining pressure as it moves up the wellbore. Flow velocities increase accordingly. Any mud or other fluid in the well is expelled quickly, leaving little hydrostatic pressure to resist reservoir influx. The result is that more supercritical CO₂ flows into the wellbore, expanding as it does.

The flowrate eventually stabilizes as an equilibrium is established between backpressure caused by fluid friction from the blowout and the pressure drop across the formation face. Often, the flowrate is controlled by the opening through which the plume escapes at the surface. Flow through small openings (holes in casing, leaks around pipe rams or in the wellhead, etc.) can reach sonic velocity, limiting flowrate and, consequentially, CO₂ influx from reservoir to wellbore.

This flow behavior is almost explosive in its violence. And this is usually not expected by field workers. Often, only a small volume of supercritical “liquid” CO₂ in the wellbore is enough to trigger the process, causing the well to blow out in a matter of seconds. Reaction time is minimal and some equipment, particularly manual BOPs and stab-in safety valves, cannot be installed and closed fast enough to avoid complete liquid expulsion from the well and total loss of pressure control.

Effects of expansion cooling. The second effect is rapid cooling of wellbore and fluid streams due to expansion. Once the CO₂ stream falls below the triple point temperature and pressure of -63°F and 76 psi, respectively, solid dry ice particles can form quickly. Several special problems result from this unique phase behavior: 1) high flowrates complicate surface intervention work and expose workers to gas moving at high velocities; 2) CO₂ and produced fluids from hydrates that can collect in BOPs, the wellhead and other surface equipment; 3) the cold CO₂ condenses water in the atmosphere, resulting in reduced visibility in the white “cloud” around the wellbore, Fig. 2; and 4) free oil and condensed miscible fluids swept out of the near-wellbore area can collect on the surface, creating a ground-fire hazard, Fig. 3.

Fig. 2. Vapor cloud from water in the air condensed by cold CO2 reduces visibility near wellbore, hindering hand-signal communications.

Fig. 3. Injection well where oil swept out of the wellbore by CO2 blowout collected in pools on location.

Further, dry ice formation often results in pea- to marble-size projectiles expelled at very high velocities, sufficient to injure workers. Fig. 4 shows a large accumulation of ice around the front of the pumping unit skid and on the gearbox. The white area behind the pumping unit is a 1 to 2-in.-deep dry ice accumulation.

Fig. 4. Accumulation of dry ice and hydrates on the pump unit skid and gear box, plus 1 to 2 in. accumulation on the ground.

Corrosion. “Sweet” corrosion on metal goods results from formation of mild corrosives in CO₂-water reactions. While not as rapid as “sour” corrosion caused by H₂S or strong acid solutions it, over time, is just as insidious. Some wells in CO₂ floods were drilled in the 1940s and 1950s, and cumulative corrosion is now becoming problematic.

CO₂-related corrosion is generally attributed to carbonic acid. Only a small fraction of the total CO₂ volume dissolved in water reacts. The remainder of the gas remains in solution to supply a continuous CO₂ source. The corrosion is localized, likely the result of small galvanic cells formed in specific areas. Other chemical reactions can also create scales that protect one area, while a nearby area is exposed to the acid.

Many reservoirs in which CO₂ is injected also produce corrosive H₂S and high chloride waters. Few older wells are equipped with corrosion-resistant casing and wellhead components, although some have been equipped in recent years with improved-metallurgy tubing / packers.

Two recent blowouts were the result of surface casing failure caused by corrosion. In both cases, a hole in the production casing allowed CO₂ to internally corrode the surface casing, resulting in pipe failure and loss of well control.

Continued corrosion is expected on wells and tubular goods due to increasing CO₂ partial pressure in most wells. The ratio of natural gas to recovered CO₂ should decrease with time, unless the natural gas is re-injected, an unlikely scenario for commercial reasons.

Trends in corrosion-related problems are difficult to assess and predict. However, one operator reported that most holes in casing strings formerly occurred deep in the wellbore (+/- 6,000 ft) as a result of external corrosion. Recently, wells in this project are experiencing holes in the casing 200 to 300 ft from surface due to internal corrosion related to CO₂ production. Also, in the past, a rig was needed occasionally to repair casing leaks. Now, two rigs are continually used for cement squeeze jobs, an indication that more corrosion pits are now penetrating the pipe.

PRE-BLOWOUT EVENTS

Four of the five recent blowouts occurred during remedial work. Workovers are routinely required on wells in CO₂ floods, including cleanouts, conformance improvement jobs, stimulations, conversions and paraffin / asphaltene removal. Repairs to downhole tubulars, sensing equipment and production equipment are also required frequently.

Generally, the wells are killed with cut brine or low-density muds by bullheading down the tubing or annulus. Some wells die for a time after being bled down at surface. Once the well is dead, the head is nippled down and a BOP is installed. The BOP stack is often a 3,000-psi-rated manual with pipe and blind rams. Rarely is an annular preventer installed, due to height limitations under the rig floor.

Experience shows that much of this service rig blowout prevention equipment (BOPE) has aged and has not been maintained properly. The same ram blocks and stab-in safety valves are frequently used for several years without reconditioning. In many instances, the BOPs are not pressure-tested after installation, so there is often no sure way to confirm whether ram packers will seal and control the well.

Also, there is limited information available on the effect of CO₂ on ram packers (some is available on its effects on other sealing components such as O-rings, however). Blistering due to rapid decompression of elastomers permeated by CO₂ has been observed. Some reservoirs are now pressured to the point that the stack and other BOPE component ratings would be exceeded by shut-in surface pressures if a CO₂ blowout occurs.

Crew training may need adjusting for working in CO₂ injection projects. Some have worked in production operations, but they have no experience with the highly expansive nature of supercritical CO₂, and the need to shut-in the well as quickly as possible after a flow is detected.

CASE STUDIES

The following are brief case studies of three of the five CO₂ blowouts requiring well-control specialist intervention over the past three years:

Blowout No. 1. This well, in a miscible, West Texas CO₂ displacement project, was an injector being serviced to replace corroded tubing joints and packer. The tapered 2 7/8 x 2 3/8-in. tubing was being pulled, and only a few joints of 2 3/8 in. and the packer were left in the hole. Air slips were chained to the top of the dual manual BOP.

The well began to flow unexpectedly, and the crew closed the manual BOPs, dressed with 2 7/8-in. ram blocks. An early report indicated at least one tubing joint was ejected and hung in the derrick. The well blew out within 30 sec and the crew evacuated. Fig. 4 shows the well upon arrival of the control team.

The air slips had apparently opened at some point allowing the tubing to drop into the BOP. The tubing was hanging on the partially closed pipe rams and the air slips had cocked sideways, spreading the plume horizontally around the wellhead; visibility was poor. The pipe rams were opened to drop the tubing and packer. An attempt was made to close the blind rams, but they were frozen in place. It was not possible to confirm whether the tubing had fallen downhole, due to ice buildup in the BOP. Fluid could not be pumped into the frozen wellhead. A hot-oil truck thawed the wellhead and BOP, and 242 bbl of water were pumped.

The next morning, the hot oiler again thawed pump lines, tubing head and BOP. The pump began injecting water down the annulus at about 0.5 bpd. Control specialists confirmed the tubing had dropped and the BOP stack was clear. Pump rate was increased to help load the well. Then blind rams were worked to break them free, and they were closed.

High-rate CO₂ flow from the well had apparently damaged the ram packers and the BOP leaked badly, indicating it could fail at any time. The pump rate was increased to 20 bpd and the well was killed. The BOP was stripped off and a new stack was nippled up. The dropped tubing and packer were fished, and workover operations proceeded without further problems.

Blowout No. 2. This well was an active CO₂ injector, being converted to reservoir pressure monitoring. Plans were to squeeze the top of a 4 1/2-in. liner and run new tubing with sensors and a packer. The well was killed, injection tubing was pulled and the old packer was removed.

A cement retainer was started in the hole on the old tubing. With the retainer at 6,300 ft, a pickup joint was made up on the injection tubing and run. Pipe rams were closed on the tubing, air slips were set and a stab-in safety valve was installed. Blind rams were left open.

Next morning, the crew found CO₂ blowing from the BOP. Swept-out-oil had collected on location in pools and puddles. It was unlikely that the gas plume would ignite because of its high CO₂ concentration, but oil on the ground was a serious fire hazard, so a foam blanket was applied. Then, the specialists approached the blowing well and confirmed that all flow was coming out the top of the BOP – it appeared that the pipe rams had failed. A line was laid and 177 bbl of brine were pumped down the annulus without affecting flow. The rig could not be started because of the fire hazard, so a winch truck was backed in to raise the blocks.

The tubing was raised, and a saddle was installed to hot tap the tubing with a 0.5-in. bit. About 300 bbl of brine was pumped down at 4 1/2 bpm to kill the blowout, but the well continued flowing. Pipe rams were closed on the tubing; flow stopped; and the well was finally killed by bullheading fluid down the annulus.

The pickup joint was backed off and laid down and was found to be flattened slightly on one side – an area only about 3/4-in. wide and about 1/16-in. deep, the length of the joint. It appeared that the joint had been pulled through a partially closed blind ram and the entire CO₂ flow had exited the well between the flat spot and the closed pipe ram. When the rams were closed on the undamaged joint, the flow stopped.

Blowout No. 3. This well was also an injector in a miscible CO₂ flood that required a workover to clean out fill. The well was killed and the packer was released. Injection tubing was pulled and stood back.

A small-diameter “stinger” made from 1 1/2-in. tubing was screwed onto the bottom of a joint of tubing, as had been used to clean out other wells. The crew elected not to install an annular preventer or change pipe rams before running the stinger.

Blind rams were opened and the crew lowered the stinger. Suddenly, the well began to flow. Pipe rams were closed, but they would not seal around the small-diameter stinger. An attempt was made to lower the stinger and tubing joint, but flow uplift would not let the tubing go down. The crew apparently attempted to drop the tubing but, instead of falling, the stinger bent and the joint fell over. Oil reached the surface a few seconds later and the crew evacuated Fig. 5. Oil collected on the location, and dirt berm was pushed around the site to contain it.

Fig. 5. Blowout No. 3 with oil on wellhead equipment and location, and bent stinger joint extending from BOP.

A single-jet abrasive cutter was rigged up on a boom and a line was run to a pump truck. Gelled fluid and an abrasive were mixed and pumped through the jet and the boom was telescoped to the correct position to cut the stinger off just above the BOP. The cut was made and the stinger fell into the hole, and blind rams were closed stopping the flow. The well was killed by bullheading brine down the casing, and the stinger was fished after the annular preventer was rigged up.

FUTURE OUTLOOK

CO₂ injection is expected to increase. Several new floods have just begun in Cogdell Reef (Canyon) field, North Hobbs (San Andres) Unit, the Boyd Unit in the Slaughter (San Andres) field and others. CO₂ injection is expected to increase in other projects now that better supplies are available.

Continuing corrosion is likely to be a problem in these projects as wellbore age and economic pressure precludes drilling replacement wells. Repairs and stimulation workovers will continue, with multiple opportunities for well-control problems.

Several sequestering projects have been proposed and are being studied to remove CO₂ from the atmosphere, especially CO₂ produced by fossil-fuel-fired power plants. Much of this work will involve personnel not familiar with the intricacies of CO₂ injection, use of reconditioned, abandoned wells with inappropriate metallurgy for this service, and pressures that may exceed limits of existing equipment.

Continued exploration and development with high CO₂ concentrations also provides the potential for well-control problems. With improved separation technology particularly in molecular sieves, many reserves that were formerly non-commercial will be developed. This includes areas such as the Norphlet in the U.S. Deep South, the Natuna Sea in Indonesia and the South China Sea, e.g., offsets to the Vicky well.

PREVENTIVE MEASURES

Several proactive procedures can be used to reduce probabilities of a CO₂ blowout and mitigate adverse effects if one should occur. These include: wellbore integrity surveys on existing wells; improved BOPE maintenance; and installation of additional BOPE on suspect wells. Improved crew awareness and well control training, and aggressive blowout contingency planning and emergency response training for operator personnel are essential.

Every BOP stack used by service rigs in CO₂ projects, whether manually or hydraulically-operated, should be performance and shell-tested at least annually. There is ample reason to recommend pressure testing blind rams, pipe rams and other BOPE components such as stab-in safety valves on each well when the stacks are nippled up. This BOPE should be tested to its rated pressures to ensure the stack can hold not only expected pressure from the well, but imposed pressures required during well kill operations, such as bullheading.

Annular BOPs are rarely used on well servicing units. At least two of the five blowouts mentioned above could have been prevented, or at least mitigated, by installation of an annular preventer.

The installation of profile nipples above packers in tubing strings would allow plugs or back-pressure valves to be set before running or pulling tubing. This would result in the crew pulling a wet string, but flow from the well would not come up the tubing requiring crews to install stab-in safety valves.

The importance of proper crew training and repetitive drills cannot be overstated. Many crew members have little or no well control training. Experienced crew members may have had no training in CO₂ blowout control. The use of specialists familiar with CO₂ blowouts to train and perform simulated incident drills may be justified.

In conclusion, response measures to safely handle CO₂ blowouts require an awareness that these blowouts are not the same as gas-well or oil-well blowouts. With proper training and planning, and a firm knowledge about the peculiarities of supercritical CO₂, the industry can respond to these events and, hopefully, prevent the recent increase in CO₂ blowouts from becoming an epidemic.  

ACKNOWLEDGMENT

The author appreciates the permission and cooperation of Cudd Well Control in preparation of this article. Thanks are also extended for information provided by the well-control specialists and engineers who worked on these blowouts, including Steve Burrow, Mark Mazella, Eddie Goodman, Gabe Gibson and Chuck Roberts.

BIBLIOGRAPHY

Weeter, R. F. and Halstead, L. N.: “Production of CO₂ From a Reservoir - A New Concept” paper SPE 10283, Journal of Petroleum Technology (September 1982), pp. 2144-2148.

Lynch, R. D., McBride, E. J., Perkins, T. K.  and Wiley, M. E.: “Dynamic Kill of an Uncontrolled CO₂ Well,” paper SPE 11378, JPT (July, 1985), pp. 1267 - 1275.

Norman, C.: “CO₂ for EOR is Plentiful but Tied to Oil Price,” Oil & Gas Journal (February 7, 1994).

Newton, L. E., Jr. and McClay, R. A.: “Corrosion and Operational Problems, CO2 Project, Sacroc Unit,” paper SPE 6391, presented at the 1977 SPE Permian Basin Oil and Gas Recovery Conference, Midland, TX, 10-11 March, 1977.

Gair, D. J. and Moulds, T. P.: “Tubular Corrosion in the West Sole Gas Field,” paper SPE 11879, SPE Production Engineering (May, 1988), pp. 147-152.

Gunaltun, Y.: “Carbon Dioxide Corrosion in Oil Wells,” paper SPE 21330, presented at the SPE Middle East Oil Show, Bahrain, 16-19 November, 1991.

Palacios, C. A. and Chaudary, V.: “Corrosion Control in the Oil and Gas Industry Using Nodal Analysis and Two-Phase Flow Modeling Techniques,” paper SPE 36127, presented at the Fourth Latin American and Caribbean Petroleum Engineering Conference, Port-of-Spain, Trinidad & Tobago, 23-26 April, 1996.

Sienko, M. J. and Plane, R. A.: Chemistry, 3^rd Edition, McGraw-Hill Book Company, p. 87.

Farshad, R. R., Garber, J. D. and Polaki, V.: “Comprehensive Model for Predicting Corrosion Rates in Gas Wells Containing CO₂,” paper SPE 65070, SPE Production & Facilities, 15 (3) (August, 2000) p. 186.

Shakhashiri, B. Z.: “Carbon Dioxide,” (handout) Chemistry 104, University of Wisconsin-Madison (March 18, 2002).

Ormiston, R. M. and Luce, M. C.: “Surface Processing of Carbon Dioxide in EOR Projects,” paper SPE 15916, JPT (August, 1986), pp. 823-828.

Holm, L. W.: “Evolution of the Carbon Dioxide Flooding Process,” paper SPE 17134, JPT (November, 1987), pp. 1337-1342.

Crolet, J. L.: “Acid Corrosion in Wells (CO₂, H₂S): Metallurgical Aspects,” paper SPE 10045, JPT (August, 1983), pp. 1553-1558.

Stone, P. C., Steinberg, B. G. and Goodson, J. E.: “Completion Design for Waterfloods and CO₂ Floods,” paper SPE 15006, SPE Production Engineering (November, 1989), pp. 365-370.

Personal interview, Kerry Shoemake, Kinder Morgan (October 1, 2002)

Moritis, G.: “Future of EOR & IOR: new Companies, infrastructure, projects reshape landscape for CO₂ EOR in US,” Oil & Gas Journal, (May 14, 2001)

Moritis, G.: “California steam EOR produces less; other EOR continues,” Oil & Gas Journal, (April 15, 2002)

Schempf, J.: “Future Looks Good for CO2,” Hart’s E&P, (June, 2001)

Schempf, J.: “CO2 injection grows in Gulf states,” Hart’s E&P, (September, 2001)

THE AUTHOR
	Les Skinner, PE, Well Control Engineering Manager for Cudd Well Control, a division of Cudd Pressure Control, earned a BS in chemical engineering from Texas Tech University in 1972. He has 20 years' oilfield experience with major / independent operators as a drilling / production engineer, plus 11 years with well-control companies, covering 14 U.S. states and 13 international countries. He has designed / supervised drilling / completion programs for deep and horizontal wells. He has worked on several major blowouts, including those in Kuwait during the Gulf War. Mr. Skinner is a licensed professional engineer in Texas. He has published several technical papers and holds three patents. He is a member of SPE, TSPE, NSPE and AIChE.