Casing-conveyed perforating improves tight gas sand economics

The middle and lower Beluga sands have normal pressure in a 1,700-ft section characterized by stacked pay with highly variable pay quality. Permeability in these 5-30 ft thick fluvial sandstones ranges from 0.01 to 3 md. Mineralogy is complex, including a high percentage of clays, volcanics, coals, and fines (Table 1). Metamorphic rock fragments form the framework of the Beluga sandstone, which is bounded by discontinuous shale, siltstone, and coal beds.

Formation gas is lean, 0.56 specific gravity, and has no condensate. The produced formation water contains about 5,000 mg/l. total dissolved solids.

Through 2001, the middle and lower Beluga formation have produced 51.6 bcf of the estimated 260 bcf of original gas in place (OGIP).

The characterization study determined that conventional completion techniques could not develop the Beluga resources optimally. One problem with a conventional approach is "cherry picking" the best sands for stimulation, which leaves much of the pay unstimulated. Unstimulated sands contribute little to a well's productivity.

The study concluded that all middle and lower Beluga pay sands would benefit from fracture stimulation. It, however, was not economically feasible to complete conventionally and stimulate all pay intervals because development costs increase dramatically if more pay is stimulated, especially in an operating environment with limited services and infrastructure.

Conventional completion techniques also prohibit evaluation of low-quality Beluga sands to determine whether they could be commercially developed and added to the reserve base. These low-quality sands contain a potential 70 bcf of recoverable gas, if they could be commercially developed.

The Beluga sands evaluation team received the task of finding a way commercially to develop this potential resource at a cost of $0.45/Mcf.

Casing-conveyed perforating

The casing-conveyed perforating system improves effective stimulation of productive intervals by allowing individual zone stimulation in a rapid, cost-effective manner. The few wells previously completed by this method have not encountered significant completion problems. This gave the project team enough confidence to recommend this new approach.

In this system, the casing contains integral isolation devices, perforating guns external to the casing, and methods to fire the guns and actuate the isolation devices remotely. Fig. 1 illustrates a wellbore in which a second interval is being perforated and the isolation valve actuated.

The perforating guns shoot into and through the casing as well as into the formation (Fig. 2). One usually fires the guns via external hydraulic control lines.

The isolation devices, such as flapper valves, are compatible with conventional primary cementing operations, as well as subsequent hydraulic fracturing. The flapper valves actuate after an interval is perforated, thus isolating lower intervals during hydraulic fracturing.

One can remove the frangible isolation devices with a slickline or coiled tubing. Also the system can incorporate easily and economically other technologies such as continuous bottomhole pressure measurement and downhole chemical injection.

For this system, the term "module" or "perforating module" refers to a unit assembly that includes the isolation valve, perforating gun, and related hardware. These modules are picked up and run as a single unit (Fig. 3).

In the basic casing-conveyed perforating system, one fires the perforating guns hydraulically, using incrementally higher pressures to fire the uphole guns. For example, 2,000-psi surface would fire the lowermost gun, with 3,000 psi being used for the next gun.

This completion system provides a means to stimulate multiple low-quality pay intervals cost effectively and to evaluate the results as part of a broader formation evaluation. Other benefits include:

• Significantly reduced total completion time and accelerated first production.
• Less bypassed pay and improved stimulation quality in a stacked-pay environment.
• Monobore well designs that help prevent liquid loading and facilitate rigless well repairs.
• Lower fracturing-fluid volume requirements because of smaller tubulars and the displacement fluid for one stimulation stage becoming the pad fluid for the next stage.
• Less fracturing horsepower because only one interval is stimulating at a time.
• Improved safety, well control, and environmental operations because the equipment is remotely actuated without having to convey equipment inside the casing.
• Lower total completion costs because of reduced tubular requirements, and associated services.

Drilling, cementing

The initial program in 2001 completed two wells, KBU 42-7 and KBU 24-6, with the casing-conveyed perforating system. Table 2 provides some basic information on the wells completed with casing-conveyed perforating modules, including the two Kenai wells in Alaska.

The perforating system in the Kenai wells used 3 1/2-in. casing, with about a maximum 6 1/2-in. diameter for any external device. These wells have an 81/2-in. borehole, although the system was designed to be run inside a 77/8-in. hole. The 31/2-in. casing has sufficient flexibility that the team did not consider wellbore deviation and dogleg severity as a problem.

The jobs included hole conditioning that followed typical local practice without any special procedures prior to running the perforating modules. Each Beluga sand completion is at about 7,500 ft measured depth (MD). In the first well, KBU 42-7, the 15 modules and all related hardware and external lines were run in about 27 hr, while in KBU 24-6 it took 24 hr to run 12 modules. This time improvement also included 3 hr of unrelated downtime.

The running procedure included simply banding the external 1/4-in. hydraulic lines to the outside of the casing that had specially designed oversized casing couplings with slots to protect the lines. An armored openhole electric logging cable, larger in diameter than the control line and run beside the hydraulic line provided additional protection.

In both wells, the running time included time spent filling the casing and circulating the drilling mud while running in hole.

Running time improved between wells because of pre-assembling the pup-joint assemblies between modules prior to bringing them to location and increased crew familiarity with the operations.

Fig. 4 shows the final wellbore diagram for Well KBU 42-7, including the pressures required for actuating the guns.

These completions included two hydraulic actuation lines, each firing about one-half of the modules. The two lines allowed firing pressures between modules to be significantly different and minimized the maximum surface pressure for firing the top module.

The jobs included 11-ft long guns, regardless of the actual strata thicknesses. In all cases, extra modules at the wellsite ensured that all prospective intervals could be accommodated. Openhole logging determined the final module number and perforating depth.

Module preparation entailed first arming all of the ballistics for perforating. The modules then remained on the pipe rack until the rig was ready for the specified module to be installed into the casing string.

Gamma ray (GR) and casing collar log (CCL) run through the casing verified the module depth relative to the zone of interest. Depth then was adjusted up or down as needed.

Openhole logs always determined the perforating intervals, although for module placement a discrepancy usually exists between the openhole logging measurements and the casing strap. The Kenai well jobs dealt with this discrepancy by having the middle of the bottomhole assembly (BHA) on depth. This allowed any discrepancy to be distributed in two directions.

The completion procedure requires cementing the casing to provide stability and protection for the string along with zonal isolation across productive intervals. The close vertical placement of perforating modules creates a special set of cementing challenges.

The individual modules included a 23/8-in. diameter perforating gun external to the 31/2 casing, thus creating an eccentric wellbore. Therefore, these jobs required special centralization equipment and displacement strategies because of the large and irregular dimension of the securing hardware, such as gun guides and y-blocks.

The special eccentric rigid centralizers placed above and below each perforating module have integral spiral vanes. This equipment centralizes each module and the spiral vanes create turbulence that improves fluid-displacement efficiency.

Cementing operations included increased spacer volume to maximize hole cleaning, and a high quality, "gastight" cement system. The cement slurry had no free water, low fluid loss, excellent compressive strength development, and low pumping rheology for enhancing the bonding across the eccentric producing section of the annulus.

To assure a consistent cement density, the cementing procedure included batch mixing and pumping at simulator optimized flow rates all spacers and cement.

Real-time surface readout of a permanent downhole pressure and temperature sensor allowed one to monitor pumping operations and to obtain slurry heat-of-hydration values.

Device actuation

The perforating modules in both Kenai wells fired successfully. A small hand pump pressurized the hydraulic lines to the pre-set pressuring limit, during the fracture stimulation operations, with the integral ceramic flapper activating simultaneous with the perforating event. This flapper isolates the borehole below the new perforation set, thus allowing the zone to be conventionally stimulated.

After stimulation and production flow back, the work included running coiled tubing to remove all ceramic isolation devices. The devices are frangible and can be removed with a slickline, but coiled tubing generally is used because of the large number of intervals and the potential for depositing proppant on top of the flapper valves.

In the Kenai wells, the coiled tubing had a BHA consisting of a 27/8-in. motor and mill. Future planned applications will eliminate the motor and mill and break the ceramic flappers by setting down weight with a blunt nose, coiled-tubing wash nozzle.

Radioactive tracers and coiled-tubing cleanouts indicated that the integral flapper valves functioned successfully during fracturing, although the team noted some difficulties with proppant cross flow from below a flapper valve into a newly perforated zone. The team is working on modifications to address this problem.

Job success requires positive shot detection-verification. Four-shot detection or verification methods can be employed (see box for description), although acceptable results require more than one method to be used.

Future installations likely will include a geophone attached to the casing at a considerable distance from the surface. The team feels that this will improve shot detection by eliminating surface interference.

Another effort underway ties the separate shot detection methods into a "black box" that will record and synchronize the observations made at each station.

Formation evaluation

The first well, KBU 42-7, had a more robust data-collection scheme than subsequent wells (Table 3). During fracture stimulation operations, the work included minifrac tests and G-function analysis on most intervals. Step-rate tests evaluated near-wellbore friction. Chemical tracers in the fracturing fluids evaluate flow-back response and fluid recovery from individual intervals.

As shown in Table 3, the second well, KBU 24-6, had reduced the openhole logging suite to a mud log, magnetic resonance log, resistivity, and neutron density. The team expects this combination to be the standard openhole logging suite for future Beluga sand wells. KBU 24-6 continued to include diagnostic tools such as radioactive tracer surveys, permanent downhole pressure sensors, and production profiles.

The formation evaluation tools provided valuable information that helped the team optimize development of the resource. But, once the information was obtained, the tools will not be needed in future wells. For instance, the team developed a useful correlation between openhole formation pressures and gas response on the mud log, enabling elimination of future openhole formation-pressure measurements.

Production profiles will be used to quantify the contribution of low-quality sands for reserves booking purposes, but these also will be eliminated in future jobs after quantifying the reserves.

The permanent downhole pressure sensors proved valuable for quantifying downhole fracture pressure, shot detection while perforating, diagnosing problems during fracturing, and aiding flow control during well unloading.

The radioactive tracers confirmed the mechanical integrity of the perforating devices and ceramic flapper valves. The tracers also reinforced the belief that fracture height growth in the Beluga formation is minimal, leading to the conclusion that modules can be spaced closer than the original 100 ft.

The chemical tracers provided only marginal information for evaluating frac-fluid recovery or productivity of individual intervals because of the poor recovery efficiency of the frac fluids and problems with the chemical tracers.

Fracturing fluid selection

As previously discussed, the team considered accurate formation evaluation to be crucial for optimizing development of the Beluga reserves. For instance, well KBU 42-7 was fraced and flowed back in several events, rather than fracturing all modules at one time.

Previous studies indicated that the poor track record for fracturing the Beluga sand was largely due to poor frac conductivity. These jobs used guar borate as well as hydrocarbon-based fracturing fluids.

To optimize the frac fluids, the team had whole core sections and core plugs from the Beluga formation analyzed for rock-pore properties to characterize framework mineralogy (Table 1), clay fraction, cements, porosity type, fluid sensitivity, and rock mechanical properties. The data identified potential completion problems.

Laboratory analyses shows the Beluga formation to be relatively immature with some authigenic precipitation, such as quartz overgrowths or carbonate cement, and porosity losses held to a minimum. It is a moderately-sorted sandstone cemented by clays with a blocky form of chlorite coating most of the framework grains with a thin layer of clay.

The core study characterized reservoir quality as good to poor. Reservoir quality appears to be a function of grain size and, to a much lesser extent, clay fraction content. Coarser grained sections yielded higher permeabilities.

The analysis immersed the formation chips in 4%, 6%, and 10% KCl brines to evaluate the fluid sensitivities further. Immersion in 4% KCl resulted in partial to total disaggregation in the coarser grained, weakly cemented samples while the finer grained samples showed no visible immersion effects. The increased induration, however, may have masked clay swelling alterations.

The coarser grained samples were then immersed in 6% and 10% KCl, and the disaggregation process continued at a slower rate. The tests of the coarser grained samples in a low-viscosity refined mineral oil (Isopar) showed no visible alteration.

The laboratory investigation characterizes the poorly consolidated Beluga sandstone water sensitivity as coming from two primary forces: clay-expansion and clay dispersion of the intergranular cement (chlorite clays). Therefore, aqueous-based fluids were not desirable.

The team selected a gelled-oil (diesel) frac fluid that provides superior rheological stability and proppant transport properties, exhibits low tubular friction pressures, and contains a time-controlled breaker system.

Extensive dynamic and static laboratory testing confirmed the frac fluid compatibility with the formation, quantified the system rheological performance, and obtained the required data for 3D fracture simulation modeling.

Prior work with downhole sensors provided much needed insight into the actual pumping temperatures of frac fluids and reservoir thermal effects after shutdown. Typical industry practice includes estimating frac-fluid break times using static bottomhole temperature.

Sensor data, however, showed that actual temperatures observed during pumping (the effective fluid temperature in the wellbore) and after shutdown (thermal increase during flow back and production) were much lower than static. This led the team to use 100° F. instead of the static reservoir temperature of 133° F. for generation the frac-fluid breaker schedule.

The fluids were continuously mixed on-the-fly with a breaker schedule tailored to deliver rheological stability during pumping and degradation (break) at 100° F. Physical observation and subsequent analysis of flow-back samples confirmed the frac-fluid break.

Fracturing operations

Much prestimulation discussion centered on determining the best displacement strategy. Conventional wisdom is that the proppant-laden fluid should be under displaced to prevent flushing the proppant away from the perforations. With casing-conveyed perforating, however, there is less than 1-bbl capacity between the interval that is being stimulated and the interval that will be stimulated next.

Under displacing the proppant-laden fluid could leave proppant in the wellbore adjacent to the new zone being perforated, potentially plugging perforations or interfering with flapper operation.

In well KBU 42-7, the displacement discussion was complicated further by the desire to conduct minifracs on each interval with ungelled diesel. The project team believed the leakoff of thin fluids would provide a better evaluation of formation properties than viscous gelled fluids. But the displacement fluid had to be straight diesel because the flush of one frac stimulation becomes the prepad of the next stimulation.

This introduced the possibility of viscous fingering through the frac fluid, thereby inadequately displacing the proppant-laden fluid and leaving proppant strung out in the wellbore.

The team finally decided to flush the proppant-laden fluid with a small volume of clean gelled fluid to minimize the effects of viscous fingering. This was followed by straight diesel. Under displacement of the proppant-laden fluid was reduced to zero, but the team did not elect to over-displace the proppant.

The results of this strategy were mixed. The thin fluids provided excellent minifrac and G-function data. The G-function analyses typically indicated no pressure-dependent leakoff, closure pressures that were consistent with prefrac expectations, some height-growth recession, and minimal near-wellbore friction.

Fig. 5 is typical of the leakoff character observed.

This strategy also provided outstanding correlation of permeability derived from the G-function analysis and log-derived permeability from the magnetic resonance log (Table 4). But it also resulted in considerable problems associated with residual proppant left in the wellbore after displacement.

Problems included plugged perforation with proppant prior to initiating a fracture, occasional problems with high near-wellbore friction on initial pump-in, and cross flow from recently fractured intervals.

A different strategy in KBU 24-6 displaced the precise volume to the top perforation with gelled diesel instead of a straight diesel flush.

The change was in part due to the mechanical troubles experienced in KBU 24-6 as well as the elimination of minifracs for the data gathering. This gelled diesel for displacing proppant-laden fluid worked better in terms of preventing plugged perforation, although it did not eliminate the cross flow problem from recently fractured intervals.

One downside of gelled diesel is that it became much more difficult to detect firing of the guns.

In KBU 42-7, the frac jobs were done at three separate times over 3 months. After each fracturing event, the well produced for a time to gauge stimulation results.

The successful stimulations in KBU 42-7 encouraged the development team to optimize fracturing procedures. For instance, KBU 24-6 was fraced in a single event during 5 days, reducing the costs and complications. Well flow back was initiated only after the last frac was displaced and as soon as possible after rigging down the tree saver.

In 2002, completion time further improved for the third Kenai well, KBU 44-6, which included 16 perforation modules and involved 15 fracs in 17 hr.

Because the Beluga formation has normal pressure, the diesel flush leaves a slight formation underbalance that occasionally is adequate to allow the well to flow. Often jobs require coil tubing for jetting the well to initiate production. So far, no attempt has been made to include nitrogen or carbon dioxide in the frac fluids to assist in flow back.

Recovery of the diesel-based frac fluid has been surprisingly poor, about 25% in both wells. Core tests support the use of hydrocarbon-based frac fluids over water-based.

Chemical tracer work in KBU 42-7 showed that the last interval stimulated (the interval with the least leakoff time) has the highest diesel recovery, an estimated 80%.

Economic assessment

The casing-conveyed perforating system in the Beluga sands improved the life-of-well economics and ultimate gas recovery.

A conventional completion consisting of 7-in. casing, 3 1/2-in. tubing and production packer could be expected to develop reserves at a cost of $0.65/Mcf. These completions would have a typical initial 2-3 MMcfd production rate, but much of the pay remained unstimulated and did not contribute to well productivity.

Other previous completion techniques tried to determine if the low-quality Beluga sands could be commercially developed. In 2000, a dual-string tubingless completion in the Beluga sands devoted one string to "research" of these low-quality sands while the second string developed higher quality pay.

While not a complete failure, the dual-string completion was regarded as not being commercially, mechanically, or from an evaluation standpoint adequate.

In contrast, the experience with the two 2001 casing perforating wells in the Kenai gas field has been mostly positive. The work increased middle and lower Beluga sand reserves and reduced development cost to an average $0.45/Mcf. Initial production rates for each well are about 5 MMcfd with possible 7-8 MMcfd/well deliverability.

Production logs verified that the low-quality sands substantially contributed to the production. These observations and the previous reservoir characterization work have lowered the minimum cutoff for effective porosity pay criteria for the Beluga sands.

It is important to recognize that casing-perforating techniques increased the upfront well cost not only because of the perforating hardware, but also because of stimulation costs.

Stimulations that conventionally may be spread out over 20 years within a single wellbore are all being done at the outset of the well's life. But even though the initial costs are higher, the life-of-well economics improved significantly.

This shift in thinking from a focus on initial well cost to total well economics has been one of the major hurdles in promoting casing-perforating technology.

Acknowledgments

The authors thank the management of Marathon Oil Co., BJ Services Co., and the Expro Group for their support, and also thank the people who contributed to development of this technology.

The authors

Gary Eller is a senior production engineer for Marathon Oil Co. in Anchorage, Alas. He primarily has worked in production engineering assignments and also has held drilling and reservoir engineering positions. Eller holds BS and MS degrees in petroleum engineering from Texas A&M University.

Jay Garner works for BJ Services Co. in Anchorage, Alas. He has held various engineering, operations, and management positions with BJ Services. He majored in geology at Colorado State University.

Kevin George is engineering manager for the tubing-conveyed perforating product line for the Expro Group in Burleson, Tex. He has worked in both the operational and design aspects of well completions for about 27 years.

Phil Snider is a senior technical consultant for Marathon Oil Co.'s drilling and completion technology group in Houston. He works on developing completion technology. Snider has a petroleum engineering degree from the University of Wyoming.

Bibliography

1. Dillenbeck, Robert L., and Cooper, Dwayne, "Successful Optimization and Application of Primary Cement Designs Enables Annular Placement of Casing Perforating Guns for Multiple Zone Completions," Paper No. SPE 64526, SPE Asia Pacific Oil and Gas Conference and Exhibition, Brisbane, Oct. 16-18, 2000.

2. BJ Services Technology Center Report No. 01-01-0015.

Based on a presentation to the SPE Western Regional/AAPG Pacific Section joint meeting, Anchorage, Alas., May 20-22, 2002.

Shot detection systems:

Four shot detection systems that are available include:

1. Acoustic listening device that includes a recorder attached to the wellhead. One monitors the recorder with a headset during the perforating process.
   This system has provided reliable shot detection when excess surface noise or interruptions can be eliminated but will likely never be a stand-alone system.

2. A geophone attached to a transmitter that provides a signal to a laptop computer via a receiver. One can monitor the signal in real time during the perforating sequence. This system has limited field testing but preliminary results are encouraging.
   One can manipulate the software associated with this system to filter out much of the surface interruptions that plague the acoustic listening device.

3. Time domain reflectometry (TDR) that measures the resistance in the conductor that is run alongside each perforating module. This conductor terminates at the surface, and the TDR surface instrumentation can be plugged in during the perforating sequence.
   The conductor shortens and is severed by a shaped charge each time a perforating module detonates. As each perforating module actuates, the surface TDR records a shift in the resistance of the cable. Preliminary field testing has been encouraging.

4. Downhole permanent downhole pressure sensors that provide real-time monitoring of bottomhole treating pressure and double as a shot detection system. One monitors the surface readout in the frac van and in many cases the data provides additional reinforcement of the perforating event. One can program the surface display to record pressures in 3-sec intervals.
   Field testing in tight formations shows that the sampling rate needs to be faster than 3 sec to reliably detect the pressure change associated with the perforating event.