David D. Cramer
BJ Services Co.
Denver
In the Bakken formation, stimulation treatments are used sparingly in horizontal well completions. But in vertical wells, stimulation is used extensively and successfully.
This concluding part of a two-part series, started in the Apr. 22 issue, shows the stimulation designs that are effective in the Bakken formation.
HYDRAULIC FRACTURE HEIGHT
Post-treatment temperature and radioactive tracer surveys allow estimates to be made of the extent of hydraulic fracture height growth in the Bakken formation. Fig. 1a shows a temperature survey which was run following a crosslinked gelled-water treatment of a vertical well in the Bicentennial field.
The well was perforated in the basal Lodgepole interval. Significant height growth occurred above the False Bakken interval, and lesser growth occurred in the Three Forks zone. A CBL log indicated that cement bonding was excellent through the Lodgepole, Bakken, and Three Forks intervals.
Fig. 1b shows surveying runs that followed a gelled oil treatment of a vertical well in the Buckhorn field. A radioactive tracer (Iridium 192) enabled proppant location to be identified with a gamma ray too[.
The temperature survey was not useful for interpretation above the perforations because the well was flowed while doing the survey. Perforations were placed from the False Bakken interval to the top of the Three Forks zone.
The tracer survey implies that extensive fracture growth occurred in the massive Lodgepole interval. Reduced gamma ray intensity adjacent to the basal Lodgepole zone suggests that fracture width is narrow in this interval.
An amplitude log suggests the presence of gas-cut cement above the Bakken formation and raises the possibility of annular channeling. Yet the inferred extent of fracture height growth is very similar to the observations in the Bicentennial field.
Fig. 1c shows a temperature survey which was run after a gelled-oil treatment in a vertical Antelope field well. The well was perforated in the Sanish sand member of the upper Three Forks formation. Fracture growth occurred in the Lodgepole as indicated by the "hot nose" anomaly. 23
A major cooling anomaly is present in the lower Bakken shale and Three Forks intervals and suggests that significant fluid injection occurred in these zones.
The preceding examples are representative of most surveys conducted after Bakken treatments. These surveys show that the Lodgepole zone is not a barrier to hydraulic fracture height growth and that fracture growth occurs in the upper part of the Three Forks zone.
TREATING PRESSURE
Analysis of bottom hole pressure during and immediately after hydraulic fracture treatments provides insight into various treatment phenomena, such as fracture height containment, minimum horizontal stress (i.e., fracture closure pressure), proppant bridging, lateral fracture extension, and fracturing fluid efficiency.
The following examples show pressure phenomena commonly observed in treatment of Bakken wells.
CLOSURE PRESSURE
Multiple treatments were conducted recently on a vertical well in Dunn County, N.D. The well was perforated in the middle Bakken silt and lower Bakken shale. A breakdown treatment was implemented with 8,400 gal of lease oil, injected at an average rate of 11 bbl/min.
Following this job, the bottom hole pressure was monitored and plotted vs. the square root of time (Fig. 2a). Deviation from the linear trend of this plot suggests a change from linear to radial flow in the formation, denoting closure of the hydraulic fracture. 21
Fracture closure pressure is equivalent to the minimum horizontal stress. It is used in calculations of net fracture pressure and is needed for determination of fracturing fluid efficiency from posttreatment pressure decline analysis. 24 25
A large-scale treatment was subsequently conducted on this well. During the job, 54,000 gal of gelled oil, 155,000 lb of 20/40 mesh sand, and 20,000 lb of 20/40 mesh proppant (IDC) were injected into the formation. A small amount of particulate fluid-loss additive (i.e., silica flour 15 lb/1,000 gal) was added to the first 14,000 gal of fluid.
The job was conducted via open-ended 2-7/8 in. tubing. Bottom hole pressure was calculated by adding the hydrostatic pressure of the static annular fluid column to the monitored backside pressure (i.e., reflected bottom hole pressure). Fig. 2b shows a diagnostic net (or excess) pressure plot, constructed by deducting the previously determined fracture closure pressure from the reflected bottom hole treating pressure. Several pressure trends occurred. Initially, the net pressure declined during the treatment (Points A to B). This behavior suggests the development of radial, vertically unconfined fracture growth. Net pressure increased sharply at Point B and marked an increase in fluid injection rate and the entry of viscous gelled fluid into the fracture. Both elements increased friction pressure in the fracture.
The net pressure trend then resumed declining until Point C, when the pressure leveled off and built slightly to Point D. Some degree of vertical fracture containment, probably in the Three Forks interval, may have occurred during this time.
A negative pressure trend resumed at Point D and coincided with the injection of less viscous gelled oil, supplied from a different storage tank. A sharp increase in net pressure (unit slope trend) starting at Point E was concurrent with the entry of the 5 ppg, 20/40 mesh sand stage into the formation.
The slope of this trend lessened to one third at Point F when a lower viscosity gelled oil from a different storage tank entered the fracture. Fluid injection ended at Point G.
The positive slope of net pressure during the time period between Points E and G indicates that proppant increased the resistance to flow in the fracture. Increased flow resistance could be caused by particle bridging, increased fluid viscosity due to solids addition, 26 or a combination of these elements.
A temperature survey, run 4 hr after the treatment, indicated that fracture growth stayed within the upper 50 ft of the Three Forks formation but extended 200 ft into the Lodgepole zone. A significant change in the rate of cooling occurred adjacent to the basal Lodgepole interval, suggesting that this zone acted as a pinch point.
If this is the case, particle bridging occurred at this point and restricted fluid flow to a portion of the original fracture height. This type of proppant bridging increases the rate of fracture length extension in the pay zone.
Net fracture pressure was suppressed throughout the treatment because of extensive vertical fracture height growth, which decreased the flow velocity in the fracture, and tapered gelled fluid viscosity, which reduced the viscous drag force in the fracture.
FISSURE OPENING PRESSURE
Low net bottom hole pressure inhibits the onset of pressure-sensitive fluid leakoff, which occurs when auxiliary fissures open. 27 Fissure opening can happen when net fracture pressure reaches a critical level, as related in the following equation. 24
PC - (sigma H max - sigma H min)/(1 - 2v) (4)
where:
PC = Fissure opening pressure
sigma H max = Maximum horizontal stress
sigma H min = Minimum horizontal stress
v = Poisson's ratio
An elastic strain relaxation analysis of oriented Bakken core indicates that the difference between minimum and maximum horizontal stresses in the Bakken intervals is roughly 500 psi. 21
Using Equation 4, the critical net pressure for fissure opening is in the 850-1,250 psi range. Net pressure levels during treatment of the studied well did not exceed 650 psi.
FLUID EFFICIENCY
The bottom hole pressure was monitored after the end of injection (shut-in), and pressure decline analysis was conducted to estimate fracturing fluid efficiency (that is, the volume of fluid and proppant present in the fracture at shut-in divided by the total volume of fluid and proppant injected during the treatment).
Diagnostic plots suggested that the fracture closed on the proppant pack 30 min after the end of injection. Using analytical methods developed by Nolte, 25 the calculated fluid efficiency and leakoff coefficient were 39% and 0.00192 ft/(min) 0.5, respectively. 28
The fluid-efficiency value derived for the above treatment is similar to findings in other stimulated Bakken wells. Fracturing fluid leak-off rate is normally moderate to low during Bakken treatments and does not correlate with the productivity of the well.
This suggests that low-permeability rock is usually present at the walls of the induced fracture. It is unlikely that the induced hydraulic fracture intersects many natural fractures in most treatments conducted in the Bakken. This implies that the azimuths of open natural fractures and the hydraulic fracture are very similar.
TREATMENT RESULTS
Stimulation treatments are commonly used on vertical Bakken wells. Treatment techniques and materials have been experimented with in efforts to improve well productivity. Some of the most significant experiences involve the treatment of previously stimulated wells.
In these cases, treatment performances can be more accurately compared because reservoir properties should be constant, at least before significant pore pressure reduction occurs.
WELL BORE EXTENSION
Fig. 3a shows the production history of a vertical well in the Buckhorn field, Federal No. 10-1. An initial frac treatment used 40,000 gal of foamed oil to place 22,000 lb of 20/40 mesh sand into the formation at a maximum concentration of 2 ppg.
Proppant blockage in the fracture or well bore caused the treatment to be ended before the intended job completion. This event, known as a "Screenout," was attributed to the poor Theological properties of foamed oil at high temperature (bottom hole static temperature = 260 F.).
Production rate following the treatment, 32 bo/d at 460 psi FTP (flowing tubing pressure), was disappointing because good hydrocarbon shows were encountered during the drilling of this well.
This well was restimulated 14 months later using 54,000 gal of gelled oil, 47,500 lb of 20/40 mesh sand, and 21,000 lb of 20/40 mesh Interprop (a ceramic proppant) at a maximum concentration of 6 ppg.
Fluid-loss additives (e.g., 100 mesh sand, silica flour) were not used. This job increased production from 28 bo/d at 140 psi FTP to 250 bo/d at 800 psi FTP.
Comparing the two treatments suggests that a small "scour frac," as implemented initially, inadequately connected the well bore to the natural fracture system. Larger volumes of fluid and proppant were necessary to establish effective well bore extension into the reservoir.
It is notable that in the second treatment, inferred fracturing fluid efficiency from post injection pressure decline was high. This suggests that the induced fracture did not cut across many open natural fractures and likely paralleled the natural fracture trend.
BYPASSING DISTURBED ZONE
Fig. 3b shows the production history of the Blacktail Federal No. 1-28, a vertical well located in the Elkhorn Ranch field. An initial treatment used 24,000 gal of low-viscosity gelled oil to place 15,000 lb of 20/40 mesh sand into the formation. Maximum proppant concentration was 1.5 ppg.
This job increased production from 21 bo/d (80 psi FTP) to 63 bo/d (90 psi FTP). The well was restimulated 4-1/2 years later using 64,000 gal of gelled oil, 84,000 lb of 20/40 mesh sand, and 20,000 lb of 20/40 mesh Interprop at a maximum concentration of 8 ppg. Following the treatment, production increased from 32 bo/d (50 psi FTP) to 91 bo/d (700 psi FTP).
Fig. 3c shows the production history of the Blacktail Federal No. 3-19, a nearby vertical well in the Elkhorn Ranch field. It was initially treated with 32,000 gal of low-viscosity gelled oil and 20,000 lb of 20/40 mesh sand. Maximum proppant concentration was 1.5 ppg.
This job increased well production from 20 bo/d (50 psi FTP) to 105 bo/d (200 psi FTP). Compared with the initial treatment of the Blacktail Federal No. 1-28, the use of slightly larger fluid and proppant volumes seemed to improve production results.
The well was restimulated 4-1/2 years later using 61,000 gal of gelled oil and 84,000 lb of 20/40 mesh sand at a maximum concentration of 7 ppg. High-strength ceramic proppant was not used. This treatment increased production from 32 bo/d (50 psi FTP) to 90 bo/d, the same outcome obtained with the remedial treatment of the Blacktail Federal No. 1-28.
Although remedial stimulation of both Elkhorn Ranch wells was successful, the use of more proppant and a high strength proppant tail-in in the Blacktail Federal No. 128 treatment apparently did not improve job results.
This may be due to the anisotropic nature of the Bakken reservoir in this area. It has been hypothesized that the hydraulic fracture drains the low-permeability rock that is transverse to the unidirectional natural fracture trend. In this limiting case, the major benefit derived from fracture treatments is improved well bore communication with the natural fractures.
This is accomplished by extending a fracture past the radius of formation (natural fracture) damage or disturbance and propping the fracture vertically through all discrete productive intervals. Certain critical fluid and proppant volumes are needed to achieve this.
Once the induced fracture bypasses the disturbed zone, increasing the fracture penetration will result in marginal production improvement.
PROPPANT SETTLING
Fig. 3d shows the production history of a vertical well located in the Devils Pass field in Billings County, N.D.
This well was initially treated with 47,000 gal of gelled oil and 34,400 lb of 20/40 mesh sand. Maximum proppant concentration was 3 ppg. Equipment problems caused the job to end prematurely.
The treatment increased production from 20 bo/d (200 psi FTP) to 120 bo/d (1,000 psi FTP). It was felt that the low proppant concentration displaced during the treatment could have impaired well productivity. A remedial treatment, designed to place a high concentration of proppant near the well bore, used 25,000 gal of gelled oil and 29,500 lb of 16/30 mesh Ac-Frac PR, a precured resincoated proppant. Proppant concentration reached 8 ppg.
Production initially declined from the pretreatment rate of 51 bo/d (100 psi FTP) due to a problem recovering incompletely broken fracturing fluid. Three months after the treatment, the production rate peaked at 66 bo/d and stabilized at 53 bo/d.
The marginal improvement from this remedial treatment suggests that proppant and fluid volumes may have a more significant effect on treatment results than surface proppant concentration or proppant permeability. Proppant settles during and after injection until the fracture closes, increasing proppant concentration in the lower portion of the fracture. 6
Proppant settling increases fracture conductivity through the thin Bakken zone in the Devils Pass field, but could be a problem in thicker intervals, where a higher propped height may be needed.
AQUEOUS FLUID
Acid and water-based stimulation fluids have been used with success in the Bakken in several areas of the basin. Fig. 3e shows the production history of a vertical well, USA No. 43X-27A, located in the Squaw Gap field in McKenzie County, N.D.
This well is perforated in the lower Lodgepole and upper Bakken intervals. It was initially stimulated with 5,000 gal of 20% HCl acid and 15,000 gal of emulsified 28% HCl acid. This job increased production from 12 bo/d (swabbing) to 31 bo/d (600 psi FTP).
The well was restimulated several months later with 130,000 gal of crosslinked gelled water and 130,000 lb of 20/40 mesh sand. Maximum sand concentration was 5 ppg. The treatment screened out before attaining the intended completion. Following this job, the production rate increased to 105 bo/d.
The improved results following acidization offered proof that reservoir potential existed in the Lodgepole intervals and enabled the reservoir to be properly evaluated. Load water flowed back slowly but steadily following the remedial treatment.
Improved production following the prop frac treatment could be the result of increased lateral fracture penetration and communication into the upper Bakken shale and lower zones, intervals not affected by the first treatment. The upper shale is almost inert in HCl acid, and the fracture created during the acid treatment probably closed in this zone, isolating lower zones from the perforations.
In 1989, the well stopped flowing because of decreased reservoir pressure. A rod pump installation restored production to roughly 100 bo/d. Reducing the bottom hole pressure by pumping did not impair the ability of the formation or propped fracture to deliver fluid to the well bore. Frac sand provided adequate fracture conductivity in this instance.
Although water as a stimulation fluid has produced good results in the preceding cases, low productivity Bakken wells that are treated with gelled water sometimes have difficulty unloading fluid and seem to be less productive than comparable wells treated with gelled oil. 6 Damaged microfractures/microporosity from water imbibement or gel residues 29 could be the root of this problem.
CERAMIC PROPPANT
Fig. 3f shows the production history of a vertical well in the Bicentennial field. This well was initially perforated in the basal Lodgepole interval and treated with 10,000 gal of crosslinked gelled water, 10,000 gal of emulsified 28% HCl acid, and 30,000 lb of 100 mesh sand.
Production following the job was 115 bo/d at 900 psi FTP (the well was not tested prior to the treatment). It appears evident that the acid soluble Lodgepole interval contains reservoir-quality fracture systems, although the 100 mesh sand may have prevented the hydraulic fracture from completely closing in the Bakken shale.
In 1988, reservoir pressure decreased to the point that the well stopped flowing. Following this occurrence, perforations were added in the Bakken intervals, and a remedial fracture treatment was conducted on the well, using 69,000 gal of gelled oil and 84,000 lb of 20/40 mesh Interprop. Maximum proppant concentration was 5 ppg.
Production peaked at 135 bo/d following the installation of a rod pump. The production rate has declined sharply since the treatment and is probably related to high fluid withdrawal rates in the area. Comparing the history of this well with the Squaw Gap well, it does not seem that high-strength ceramic proppant provided a significant advantage over frac sand.
SCALE REMOVAL
Fig. 3g shows the production history of a vertical well, Harvey Hopkins No. 3-A, in the Antelope field. This well was completed in the lower Bakken Shale and upper Three Forks intervals. It was never stimulated with a prop frac but was treated periodically with fresh water and acid (both HCl and HCl/HF blends) to remove buildups of salt and carbonate scale, respectively.
In 1983, the well was acidized with 3,300 gal of an acetic/HCl acid blend and 3,300 gal of an oil/solvent/surfactant mixture (Bakken-Sol); benzoic acid flakes were also used in an attempt to divert the treatment. The Bakken-Sol mixture was designed to break emulsions formed by the spent acid, reduce the interfacial tension of the load fluid, and help dissolve the diverting material. The acid job increased production from 26 bo/d (flowing intermittently) to 91 bo/d (100 psi FTP).
Slugs of spent and partially spent acid were recovered from the well for up to 1 year after the 1983 treatment.
PLUGGING FORMATION
Fig. 3h shows the production history of a well in the Devils Pass field. An initial stimulation attempt on this well screened out during pad fluid injection when large amounts of fluid-loss additives (i.e., silica flour and 1 00 mesh sand) plugged the narrow hydraulic fracture.
This well was successfully stimulated in December 1982 following a "scour" frac treatment that abraded and washed away a portion of the particulate material. The main job consisted of 50,000 gal of gelled oil and 68,000 lb of 20/40 mesh sand. Maximum proppant concentration exceeded 6 ppg.
This treatment increased production to 94 bo/d (360 psi FTP). In 1987, nitrified 15% HCl acid and solvents were used in an attempt to remove scale and debris that were thought to be present near the perforations. As a result of this job, production decreased from 51 to 12 bo/d.
A possible reason for this treatment failure is plugging of the proppant pack and natural fractures.
The source of plugging material could have been formation fines generated by reaction with acid-soluble rock intervals, debris in the tubing, or the fluid-loss material deposited during the original treatment.
A remedial fracture treatment was attempted in 1990 and screened out shortly after proppant entered the perforations. This example shows that a proppant pack is an efficient filter. Caution must be exercised when acidizing a Bakken well that has been previously treated with proppant.
MASSIVE FRAC
Fig. 3i shows the production history of a well in the Elkhorn Ranch field. This well, Federal No. 19-4, was stimulated with one of the largest jobs ever conducted on a Bakken well.
Material usage included 95,000 gal of crosslinked gelled water and 449,000 lb of 20/40 mesh sand. Maximum proppant concentration was 14 ppg. Following the treatment, production increased from 20 to 79 bo/d.
This result is comparable with nearby producers that had pretreatment rates close to 20 bo/d. Those wells were stimulated with smaller volumes of fluid and proppant, in the 50,000 gal and 80,000 lb range. This case seems to be another example of sharply diminishing returns when the treatment volume exceeds a critical level. Similar experiences have been encountered with large gelled oil treatments.
GEL BREAKERS
Fig. 3j shows the production history of a well in the Buckhorn field. This well was stimulated in late 1986, using a gelled oil system that did not contain any gel breaking additive. Immediately following the job, the well produced at a rate of 23 bo/d.
Poor frac fluid recovery was suspected of hampering the productivity of this well. Eighteen months after the treatment, the production rate of the well suddenly doubled, from 22 to 45 bo/d. This increase was attributed to delayed cleanup of the gelled oil or a residual by-product of the gelled oil.
Accurate proportioning of chemical additives improves the properties of gelled oil. Fig. 4a shows the Theological behavior of a high performance gelled oil system. Fracturing fluids should provide adequate viscosity during the duration of the treatment (e.g., 120 min) and break to a thin consistency within 24 hr without forming fracture plugging gel residue. This type of performance is achieved with the use of 1.2 gal of gel activator/1,000 gal of oil.
All test fluids contained 1.0 gal of GBO-6/1,000 gal of oil, a proprietary liquid gel breaker. Incomplete gel degradation occurs without the use of GBO-6. Because of subtle differences in the composition of lease oils, Theological testing should be conducted with the actual oil to be gelled in the field to optimize the concentrations of gelling surfactant (GO-53), gel activator (XLO-1), and gel breaker.
Rheological properties should be measured with a Fann 50 viscometer and at the same bottom hole temperature of the well to be treated.
This method was used recently in a treatment conducted on a well in the Elkhorn Ranch field. The job consisted of 50,000 gal of Super Allo Frac 10 (with 1.2 gal XLO-1/1,000 gal of oil and 1.0 gal of GBO-6/1,000 gal of oil) 82,000 lb of 20/40 mesh sand and 24,000 lb of 20/40 mesh Carbo-Lite, a low-density ceramic proppant.
Maximum proppant concentration was 9 ppg. Flowback samples were evaluated for Theological properties using a Fann 35 viscometer. As can be seen in Fig. 4b, the recovered fluid was thin, and very little gel structure was evident. The treatment increased production from an unsustained small flow to 90 bo/d.
REFERENCES
- Nolte, K., "A General Analysis of Fracturing Pressure Decline with Application to Three Models," SPEFE, December 1986, pp. 571-83.
- Nolte, K., and Smith, M., "Interpretation of Fracturing Pressures," JPT, September 1981, pp. 1767-75.
- Holditch, S., et al., "The Effect of Viscous Fluid Properties on Excess Friction Pressures Measured During Hydraulic Fracture Treatments," paper No. SPE 18208, 63rd Annual SPE Technical Conference, Houston, Oct. 2-5, 1988.
- Warpinski, N., "Dual Leakoff Behavior in Hydraulic Fracturing of Tight, Lenticular Gas Sands," SPEPE, August 1990, pp. 243-52.
- Cramer, D., "Treating Pressure Analysis in the Bakken Formation," paper No. SPE 21820, SPE Low Permeability Symposium, Denver, Apr. 15-17.
- Branagan, P., et al., "Case History of Hydraulic Fracture Performance in the Naturally Fractured Paludal Zone: The Transitory Effects of Damage," paper No. SPE/DOE 16397, SPE/DOE Low Permeability Reservoirs Symposium, Denver, May 18-19, 1987.
Copyright 1991 Oil & Gas Journal. All Rights Reserved.