NEW FRACTURING TECHNIQUES REDUCE TIGHT GAS SAND COMPLETION PROBLEMS

Oct. 12, 1992
Paul L. Bruce, Jeffrey L. Hunter Pennzoil Exploration & Production Co. Houston Robert D. Kuhlman Halliburton Services Duncan, Okla. Don D. Weinheimer Halliburton Services Houston New fracturing stimulation technology contributed to solving problems in completing tight gas sands in the Carthage Cotton Valley field in Texas. These technologies included improved fluid systems, computer-controlled proppant placement, multiple isotope radioactive logs, mechanical properties logs, and innovative
Paul L. Bruce, Jeffrey L. Hunter
Pennzoil Exploration & Production Co.

Houston

Robert D. Kuhlman
Halliburton Services
Duncan, Okla.
Don D. Weinheimer
Halliburton Services
Houston

New fracturing stimulation technology contributed to solving problems in completing tight gas sands in the Carthage Cotton Valley field in Texas.

These technologies included improved fluid systems, computer-controlled proppant placement, multiple isotope radioactive logs, mechanical properties logs, and innovative casing design.

CARTHAGE FIELD

Drilling activity in the Carthage field commenced on a large scale in 1978 and 1979. At that time, the Natural Gas Policy Act of 1978 (NGPA) first allowed higher gas prices.

Pennzoil Exploration & Production Co. drilled 32 wells during that period. Initial development was on 640 acre spacing (Fig. 1).

The Cotton Valley sands are found mostly between 8,200 and 9,700 ft. As many as 13 different productive sandstone members are recognized and correlated within the interval.

In 1980, these low-permeability sandstones officially were classified as "tight gas sands" by the Federal Energy Regulatory Commission (FERC). This classification qualified the sands for NGPA incentive gas prices.

After the Texas Railroad Commission (RRC) changed the field rules to 320 acre spacing, another round of development drilling began. In 1981 and 1982, Pennzoil drilled and completed 22 infill development wells before the gas market crashed in 1982.

In mid-1988, the RRC further reduced allowable well spacing to 160 acres. Pennzoil responded by initiating a new 74 well infill drilling program. Six drilling rigs were employed from October 1988 until September 1989. Three rigs continued to drill until completion of the project in April 1990.

The project was successful. Drilling operations averaged 30 rig days (545 hr rotating)/well. This time was 15% less than the original projection.

By December 1990, the 74 wells were flowing to the pipeline and had produced a cumulative of 35 bcf. Flow averaged 1,012 Mcfd.

CASING MODIFICATION

Earlier Cotton Valley completions had 5-1/2 in. casing and 2-7/8 in. tubing. A packer was set at approximately 8,200 ft, above all perforations and productive intervals.

Through 1982, completion procedures involved perforating the 5-1/2 in. casing and fracturing through the 2-7/8 in. tubing. After stimulating the perforated zone or group of zones, a sand plug was set to isolate the fractured interval from the next interval to be perforated and fractured.

After the last (uppermost) interval was fractured, a snubbing unit washed out the sand plug. For circulating out the sand, ball sealers, and miscellaneous junk, coiled tubing was not used because of the risk of sticking the coiled tubing in either the 2-2/8 in. tubing or the packer.

The most recent drilling program modified the completion procedure to allow for several improvements.

While fracturing, the 2-7/8 in. tubing prevented injection rates to exceed about 25 bbl/min. This rate was inadequate for successfully fracturing larger intervals, especially where weak barriers allow extensive fracture height growth during treatments.

In the recent completions, this rate limitation was minimized or eliminated by fracturing down the casing and then running the production tubing. Fracturing down the casing lowered friction pressures during treatment and consequently lowered horsepower requirements.

Perforating with 3-3/8 in. casing guns was another advantage to treating down the casing. These larger guns give greater penetration than the previously used tubing guns. The better penetration lowered the pressure to breakdown the formation and further reduced the horsepower needed for the stimulation.

Because coiled tubing now could be used instead of a snubbing unit, the cost of washing out the sand plug after fracturing was reduced.

Because of these advantages, the wells drilled in 1988 and later were completed with 4-1/2 in. casing of sufficient strength to withstand fracturing pressures. The 2-3/8 in. production string was installed after the wells had been fractured and cleaned out.

CROSS FLOW

With multiple pay strings over a large interval, the Cotton Valley sands must be stimulated by fracturing in multiple stages. Sand plugs provide the necessary diversion between fracturing stages.

Throughout the 1979 and 1982 projects, and through part of the current project, problems were experienced with the sand plugs. Since then, significant improvements have been made.

In the first 13 wells of the recently completed 74-well program, 28 sand plugs were set. The plugs were tagged with wire line and pressure tested to 5,000 psi.

Immediately upon perforating for the next fracturing stage, eight of the sand plugs were ruined by cross flow caused by hydraulic communication between the new perforations and the perforations below that were to be isolated by the sand plugs. When this occurred, the ruined sand plugs had to be reset.

Because perforations existed above the reset plugs, the plug location could only be checked by tagging with wire line. The plugs could not be pressure tested.

Radioactive tracer logs from these wells showed that three of the reset plugs did not hold.

Fig. 2 is a radioactive tracer log showing a cross-flow failure. The iridium isotope, used to tag Stage 1 (the lower zone, perforations 9,4589,469 ft), also is seen at the perforations of Stage 2 (the upper zone, perforations 9,118-9,149 ft).

After tagging Stage 2, the scandium isotope was observed throughout the Stage 1 interval. This indicated that the sand plug did not hold. By not holding, the lower zone accepted part of the fracturing fluids and proppant intended for the upper stage.

Multiple isotope radioactive tracer logs revealed this problem for the first time. It is believed that such failures had occurred with some frequency during earlier development activity but were never detected.

After identifying the plug failure problem, the fracturing procedure was modified to include a 6-12 hr flow back period immediately following each fracture treatment. This procedure induced fracture closure and relieved supercharging of the stimulated zone.

Chemical breakers were added to the fracturing fluid system to facilitate immediate flow back of fluids. Flow back periods are monitored closely and controlled by small positive chokes to prevent any significant sand flow from the fracture.

After the modification, cross-flow sand plug problems were reduced to about 1 in 20. Also, reset sand plugs have not failed to divert since the procedure change. This has ensured placing most of the proppant and treating fluids in the right zones.

Well deliverabilities improved after these changes.

Because of the flow back operation, fracture closure was induced more quickly, hence, trapping a maximum concentration of proppant in the fractured zone immediately adjacent to the perforations. This allowed for fracture closure while more proppant was suspended in the fracture and yielded a better distribution of proppant throughout the fracture height.

It was, however, difficult to document the magnitude of tangible improvements in well deliverability because of the number of variables involved in comparing wells. Variables include pay zone thicknesses, porosity, permeability, etc. Also, the zones were fractured with different fluid volumes and injection rates, larger proppant stages, and fracturing fluid systems.

FRACTURE CONTAINMENT

The multiple-isotope gamma ray energy spectroscopy radioactive-tracer log was one of the valuable tools for evaluating and designing optimal fracturing and completion procedures. This tool was invaluable in detecting the sand plug cross-flow problems.

This type of log can determine if radioactivity from the tracer material is being detected from within the casing or in the formation. Radioactivity logs have also been useful in identifying cement bond problems.

Fig. 3 is a radioactive tracer log that shows unexpected vertical height growth from a fracturing treatment. The perforations (9,114-9,121 ft) were intended to have a fracture that extended from 9,090 to 9,140 ft. This log detected the unexpected height growth to 9,010 ft.

Because the casing is hot with radioactivity up to 8,700 ft, conventional gamma ray tracer logs likely would have indicated height growth up to 8,700 ft. This log, however, interpreted most of that radioactivity inside the casing, particularly above 9,000 ft.

With this type of accurate information, appropriate modifications were made to the frac design and perforations for the next fracturing stage in this particular well.

SCREENOUT

This tracer log also helped identify a partial screenout condition.

Limited entry perforating techniques were frequently used to obtain better distribution of the fracturing fluids and proppant across multiple zones. The log could qualitatively evaluate the effectiveness of limited entry perforation programs in the larger fracture intervals which cover multiple zones.

Usually, for fracture designs over multiple zones, the proppant is tagged with two different isotopes. Fig. 4 shows a log that measured scandium and iridium isotope tracer material.

The iridium was used to tag the main part of the proppant slurry, up to 7 ppg. For tagging the remaining 7-8 ppg proppant slurry, scandium was used. As seen from the log, only the upper perforations were receiving fluids during the 7-8 ppg slurry (Fig. 4).

For offset wells, this information reduced the number of perforations placed in the upper part of this interval.

This information also could be useful in evaluating potential gas recoveries and planning future remedial or additional stimulation work.

BARRIERS

Mechanical properties and calculated potential fracture height growth predictions from long-spaced sonic logs and the after-fracturing radioactive tracer logs were combined to identify effective and ineffective barriers to fracture height growth in various areas of the field.

In this field, Pennzoil generally has designed fracturing stimulation treatments for individual zones whenever possible. When barriers between zones were not sufficient to contain a fracture, groups of two or more productive zones were placed together between adequate barriers. The newer wells usually required three or four frac stages.

Fig. 5 shows a well log with a five-stage treatment. The five frac stages all stayed well in zone, and the isotopes were distributed throughout entire intervals, especially in the large fourth and fifth stages.

Some wells had no effective barriers to fracture height growth between zones within the Cotton Valley interval. Those wells were stimulated with large, single-stage frac treatments.

Fig. 6 is a net pressure plot from the first stage of a treatment that was originally designed to cover a 450-ft interval. The net pressure plot, however, shows unexpected height growth as indicated by the -1 slope after 100 min. A log run following the first stage (Fig. 7) indicated that the fracture had grown upward into Zone 1, the target of the second stage.

Based on this information, the second stage was canceled and Zone 1 was perforated. By indicating significant production from Zone 1 perforations, subsequent production logs confirmed the fracture height.

FRACTURE DESIGN IMPROVEMENTS

The original 32 wells were stimulated with massive hydraulic fracturing treatments with the best fluid systems and latest available fracturing technology. Fluid and proppant volumes averaged 313,500 gal/well and 503,375 lb/well. The zones of interest were fractured in two or three stages.

In one of these earlier wells, the maximum volume pumped was 554,000 gal of fluid with 893,000 lb of proppant. The treatment was pumped in four different stages with a maximum proppant concentration of 4.0 ppg.

JOB SIZE INCREASES

In the early 1980s, improved fracturing fluid systems allowed proppant concentrations to reach 7-8 ppg. Pennzoil's second round of development in this field (320 acre spacing) involved treatments that averaged 510,000 gal of frac fluid and 1.131 million lb of proppant/well.

From the previous program, this was a 62% increase in fluid volumes and a 125% increase in proppant volumes.

Increases in fracture height growth (identified from after-fracturing gamma ray logs from these earlier wells) enlarged the total volumes used in the fracturing.

The increase in the ratio of proppant volume to fluid volumes was caused by the attainable higher proppant concentrations and also because higher concentrations were injected in the larger volume jobs.

During the second development project, for a single stage, the maximum volumes injected were 411,000 gal of frac fluid and 1.15 million lb of proppant.

The maximum volumes pumped for a single well were 522,000 gal fluid and 1,305,000 lb proppant in four stages with maximum proppant concentration of 7.0 ppg.

In the recently completed development program (74 wells), treatments averaged 943,000 gal of fluid and 2.91 million lb of proppant/well. This was an increase of 65% in fluid volumes and 157% in proppant volumes over the 1981 to 1982 period.

To date, the largest fracturing stage had a volume of 968,200 gal of frac fluid and 4.13 million lb of proppant. Maximum proppant concentration was 8 ppg.

COMPUTER CONTROL

The ability to place larger proppant volumes at higher concentrations is partly due to the introduction of computerized proppant control systems that allow for new proppant stage ramping techniques.

Table 1 illustrates volumes and rates for various wells in the 1978 to 1990 project.

Because treatments are larger, occasionally stimulation costs are nearly equal to all other well costs. For this reason, optimal fracturing designs significantly impact well costs.

One goal in the stimulation program was to create 1,000-ft frac wings with a minimum proppant bed concentration of 1.0 lb/ft and an average 2.0 lb/ft after fracture closure.

All proppant systems used were crosslinked gels. The early frac designs called for pad volumes of 35-40% comprised of 50 lb of gel/1,000 gal.

The sand-laden fluids usually consisted of 40 lb of gel/1,000 gal. Typically, fluid-loss additives in the pad were various combinations of 100 mesh sand and silica flour. Early in the program, the silica flour was replaced by a liquid (diesel base) fluid-loss additive.

With the cooperation of service companies and the introduction of new technology, stimulation work in the Cotton Valley is now carefully controlled by computer vans that monitor treatment pressures, rates, densities, volumes, and fluid rheological properties.

The system generates net pressure log-log plots in real time and uses fracture design software that permits modification of the treatment design on location.

With improved monitoring and mixing systems, coupled with other technology improvements such as liquid gel concentrate and liquid potassium chloride (KCl) substitutes, Pennzoil reduced the amount of pad and gel used per pound of proppant placed, hence, reducing chemical and fracturing costs.

STAGE DESIGNS

The Taylor sandstone, lowest Cotton Valley sand (designated by Pennzoil as Zone 10), is always fractured separate from the other zones.

For this zone, the fracturing design zone included 17% pad with 40 lb of gel/1,000 gal in the pad and in 50-75% of the sand-laden fluid. The remaining 25-50% of the sand-laden fluid had 35 lb of gel/1,000 gal.

This design yielded a savings of 10% in chemical costs, or about 4-5% savings in total fracturing cost when compared to previous designs.

In most of the wells, the second stage was designated as Zone 9, this sometimes included Zone 8. The fracture design for this interval used a 23-25% pad consisting of 40 lb of gel/1,000 gal. The same gel concentration was used in 50-75% of the proppant-laden fluid. Then 35 lb of gel/1,000 gal was used for the remaining 25-50% of proppant-laden fluid.

This design was gradually modified by reducing the original 34-39% pad volumes consisting of 50 lb of gel/1,000 gal and 40 lb of gel/1,000 gal for the sand-laden fluid. This saved significantly on fracturing costs.

The third stage covered a 250-400 ft interval which usually consisted of part or all of Zones 4 through 8. In the early part of this program, some problems were encountered with increasing treating pressures and screenouts. As a result, the fluid-loss component was modified to include liquid fluid-loss additive (diesel base) in up to 50% of the proppant-laden fluid.

The fluid design for this interval was optimal for a 30% pad consisting of 50 lb of gel/1,000 gal with 40 lb of gel/1,000 gal in 75% of the sand-laden fluid, and 35 lb of gel/1,000 gal in the remaining 25% of the proppant-laden steps.

The fourth stage was in an interval of about 150-250 ft that included Zones B, 1, 2, and 3. For this stage, Pennzoil used a 27-29% pad made up of 40 lb of gel/1,000 gal. This same gel loading was also included in the 75% of the sand-laden fluid. The remaining 25% of the sand-laden fluid consisted of 35 lb of gel/1,000 gal.

The authors believe that these designs are close to the optimum for each interval, but some further changes may be made, especially on Stages 1 and 4.

Lower gel concentrations have been successful in the deeper intervals. This is contrary to what is expected based on the higher pressures and temperatures encountered at those depths. It is likely, however, that this is caused by better barriers to fracture height growth in the lower zones, such as Zone 10. Therefore, these zones may be successfully treated at higher injection rates without concern for fracturing out of zone.

ACKNOWLEDGMENT

The authors thank the management of Pennzoil Exploration & Production Co. and Halliburton Services for permission to publish this article.

Copyright 1992 Oil & Gas Journal. All Rights Reserved.