SPE 134330

Refracs – Why Do They Work, And Why Do They Fail In 100 Published Field
Studies?
M.C. Vincent, SPE, consultant to CARBO Ceramics, Inc.

Copyright 2010, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Annual Technical Conference and Exhibition held in Florence, Italy, 19–22 September 2010.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
A database has been compiled and analyzed, summarizing more than 100 field studies in which restimulation treatments
(hydraulic refracs) have been performed, along with the production results. Field results demonstrate that refrac success can
be attributed to many mechanisms, including:
− Enlarged fracture geometry, enhancing reservoir contact
− Improved pay coverage through increased fracture height in vertical wells
− More thorough lateral coverage in horizontal wells or initiation of more transverse fractures
− Increased fracture conductivity compared to initial frac
− Restoration of fracture conductivity lost due to embedment, cyclic stress, proppant degradation, gel damage, scale,
asphaltene precipitation, fines plugging, etc.
− Increased conductivity in previously unpropped or inadequately propped portions of fracture
− Improved production profile in well; preferentially stimulating lower permeability intervals [reservoir management]
− Use of more suitable fracturing fluids
− Re-energizing or re-inflating natural fissures
− Reorientation due to stress field alterations, leading to contact of “new” rock

Although less frequently published, unsuccessful restimulation treatments are also common. Documented concerns
illustrated in this paper include:
− Low pressured, depleted wells (especially gas wells) posing challenges with recovery of fracturing fluids
− Low pressured or fault-isolated wells with limited reserves
− Wells in which diagnostics indicate effective initial fractures and drainage to reservoir boundaries
− Wells with undesirable existing perforations, or uncertain mechanical integrity of tubing, casing, or cement

This paper will explore the common problems that lead to unsatisfactory stimulation, or initial treatments that fail over time.
Guidelines for evaluating refrac candidates and improving initial treatments will be reviewed. The paper summarizes
restimulation attempts in oil and gas wells in sandstone, carbonate, shale and coal formations. This organized summary of
field results and references will provide significant value to readers evaluating or designing restimulation treatments.

Introduction
A compilation of published literature can provide a database of analog field examples to guide operators in design of refrac
treatments. Since existing production models often fail to accurately predict the production achieved from refracs, field
examples will provide real-world calibration of expectations and better estimation of economic potential when recompleting
and restimulating wells. A database containing 143 field studies in which restimulation treatments have been performed has
been analyzed. Field results demonstrate that refrac success can be attributed to numerous mechanisms which will be
discussed in this paper. This paper will highlight field examples of various mechanisms, and provide some guidance in the
selection of optimal candidates for restimulation, as well as an overview of the questions operators should explore to design
the proper treatment for their field. This paper will examine the common problems that lead to unsatisfactory stimulation, or
2 SPE 134330

stimulations that fail over time. The major concerns will be outlined and an overview of the diagnostic techniques available
for use will be given.

Sallee and Rugg (1953) estimated that initial fracture treatments were successful in 75% of attempts industry-wide, although
their review of 2000 initial stimulations in Oklahoma and Texas indicated that more than 85% were successful. The results of
two sequential frac stimulations performed in the first year of production on a well in the Strawn formation are shown in
Figure 1.

Figure 1 – As early as 1953, restimulation treatments were shown to increase production rates and pay back investments
within 60 days [Sallee and Rugg, 1953].

Garland et al (1957) studied 24 wells in the Strawn formation that had been stimulated with nitroglycerin five to ten years
prior to a hydraulic fracture treatment. Propped fractures were beneficial in all categories of wells, but better results were
observed in previously untreated wells. Eleven Strawn wells had been treated with sequential propped fractures (refracs),
with up to 55 elapsed months between fracs. While the refracs did not typically increase production by the same percentage
as the initial fracture, the rate of production decline was typically lower after a second treatment. The results from a well
receiving three sequential propped fracs is shown in Figure 2.
Oil Production, bbls/month

Initial Re-
Frac Frac
Tri-
Frac

Figure 2 – Three sequential propped fracture treatments increased monthly oil production from 200 to over 1000 BPM,
despite four years of depletion [Garland et al, 1957].
SPE 134330 3

It is interesting to note that as of 1970, approximately 35% of the 500,000 frac jobs performed by the industry had been
restimulation treatments [Howard and Fast, 1970; Coulter and Menzie, 1973]. However, by 1996, it was estimated that only
2-3% of current stimulation activity was directed towards restimulation [GRI, 1996]. Why? Did the industry drastically
improve the design, implementation, and durability of initial frac treatments? Did technological improvements cease? Did
the economic parameters change? Is the industry no longer sufficiently staffed to analyze refrac opportunities? Are refracs
not as effective as initially believed? Did restimulation simply fall from vogue and become a lost art?

Where successful, restimulation treatments have an undeniable allure. The ability to increase production rate and/or reserves-
without the expense of drilling a new well, with minimal environmental impact and minimal regulatory permitting
requirements- can make restimulation treatments among the most quickly implemented procedures which deliver outstanding
return on investment. In some fields, refracs are highly predictable and consistently economic. In other fields, significant
engineering effort is required to identify appropriate candidates. Inadequate characterization of the initial fracture, poor
knowledge of reservoir pressure, vague candidate selection criteria, and less experience with refrac design can cause
restimulation campaigns to entail more uncertainty than does traditional infill drilling.

The Appendix of this paper compiles 143 actual field examples documenting the success and failure of treatments to assist in
calibrating expectations for refrac performance. The published Appendix represents only a few columns from the compiled
database that is used when designing stimulation treatments for clients. Extensive history-matching of field examples has
provided numerous insights into the performance of initial treatments, but many of those details are omitted from this paper.
This paper concludes with discussions of successful strategies and procedures to identify, design, and implement
restimulation treatments.

The author recognizes a significant literature bias. The petroleum industry tends to publish success stories more frequently
than it discloses failures. When a refrac is attempted, but is unsuccessful, that frequently terminates the restimulation
campaign. There is little incentive for publication of that failure, as there is little statistical significance to the solitary
example, and few accolades to gain. Therefore, the literature becomes enriched with successful examples while having less
representation of failed efforts. This does not mean the documented examples are invalid, but merely that they may not
faithfully represent the full spectrum of outcomes and the correct proportions of success/failure. While profitable
restimulation opportunities exist in most formations, careful review and design is usually merited before embarking on a
restimulation program. Refrac evaluation demands a more comprehensive understanding of the geology, initial treatment
performance, and outcomes achieved with other treatment designs in the candidate field and analogous fields. For this
reason, some successful operators assign this discrete task to an employee or consultant who can focus on restimulation
without the distraction of competing priorities, or without undue allegiance to previously engrained design methodologies.

It is hoped that a review of the mechanisms related to refrac success and failure, coupled with a discussion of the diagnostics,
techniques, and expertise necessary to optimize restimulation treatments will assist readers in the selection of candidates and
encourage more effective recovery of oil and gas by improvement of initial and remedial stimulation treatments.

Restimulation Field Examples Which Illustrate Different Mechanisms

In the vast majority of published restimulation treatments, more than one variable were altered between the initial and
restimulation treatments. Commonly, the total proppant mass, concentration, and quality are upgraded simultaneously –
often making it difficult to identify what portion of the increased productivity should be attributed to superior coverage of the
pay interval, improved fracture length, or superior fracture conductivity and/or proppant durability. It appears that multiple
benefits may have been achieved in almost every published example. Nonetheless, specific field examples have been
highlighted from the Appendix to emphasize the benefits of restimulation achieved by the following mechanisms:
- Enlarged fracture geometry, enhancing reservoir contact
- Restoration of fracture conductivity lost due to embedment, cyclic stress, proppant degradation, gel damage,
scale, asphaltene precipitation, fines plugging, etc.
- Improved fracture conductivity compared to initial frac
- Improved frac fluid selection which addresses a frac fluid –formation incompatibility
- Refrac success due to superior containment within the target formation
- Re-energizing or re-inflating natural fissures
- Reorientation due to stress field alterations, leading to contact of “new” rock

Field Example – Improved Vertical Coverage of Pay -- “Targeting Bypassed Pay”

There are numerous examples demonstrating the challenge in effectively contacting desired pay intervals with fracture
treatments. Fractures may not initiate in all desired intervals - geologic understanding may evolve, or economic criteria may
change - necessitating the need to restimulate bypassed pay.
4 SPE 134330

Evidence of bypassed pay

In stacked tight gas sands it is often impractical to mechanically isolate and individually fracture stimulate each potentially
productive interval. For instance, Pinedale Anticline wells may encounter 50 discrete sand intervals within 5500 feet of gross
section [Huckabee, 2005]. However, it is common to treat Pinedale wells with 15 to 20 mechanically isolated fracture stages,
relying on some degree of perforation-friction controlled diversion (limited entry) for intra-stage diversion. Although a
review of the productivity of 446 Pinedale stages [Vincent, 2007] confirmed that this strategy consistently resulted in
productive treatments, it was not determined how much incremental benefit could be achieved with additional stages in this
field. In the Piceance Basin of western Colorado, three to five frac stages are often utilized to target an average of 26 sand
intervals. However, an analysis of thirteen Piceance Basin production logs indicated that 28% of the targeted sands failed to
produce measurable gas rates [Esphahanian, 1997]. Detailed work by Craig & Odegard (2008) prove that limited-entry
treatments in the Piceance ineffectively stimulate some layers, and that production logs alone are insufficient to identify
layers meriting restimulation. In the Jonah field of Wyoming, each well may encounter 30 to 40 gas sands, however
production logs suggest approximately 35% to 40% of the intervals fail to meaningfully contribute to production [Eberhard,
2003]. Although production logging alone cannot determine whether the failure was due to poor reservoir quality or poor
fracture diversion, additional diagnostics such as detectable proppant logging, tilt-meter, microseismic mapping, and
selective isolation and retreatment have provided additional insight. An analysis of 136 fracture treatments incorporating
radioactive tracer in the Almond, Cotton Valley, Delaware, and Red Fork formations demonstrated that in nearly 40% of the
wells at least one targeted interval failed to receive any of the desired treatment [Fisher, 1995].

Restimulation of bypassed pay

In the Hugoton Field, Hecker et al (1995) published an excellent field example of designing restimulation treatments to target
bypassed pay. Small quantities of radioactive particles were blended into the fracturing treatments of 150 gas wells to
identify unstimulated portions within the 200-ft thick Chase Group. Although the research provided significant
improvements in the perforating and fracturing strategies of future wells, several refrac opportunities were identified with
poor radioactive tracer coverage in desired intervals. As shown in Figure 3, only minor amounts of tracer appeared to enter
the upper perforations of a well, poorly stimulating three zones at 2550-2660 ft depth. The well quickly declined from an
initial production rate of 440 mcfd to 215 mcfd. A retrievable bridge plug was set at 2680 ft, and a restimulation treatment
was pumped with a unique isotope into the upper intervals. Initial production following the refrac improved to 450 mcfd
over the first four months, increasing to 510 mcfd one year later when surface flowing pressure was reduced.

Gas production rate

Tracer after initial frac Tracer after refrac

Figure 3 – Tracer logging indicated the refrac (middle image) provided better distribution of proppant into upper intervals
poorly stimulated during initial treatment (left). Following the restimulation treatment, production increased from 215 mcfd to
approximately 500 mcfd (right image) [Figures adapted from Hecker, 1995]

Without the diagnostic tracer, it may not have been possible to determine which intervals were poorly stimulated, and the
operator may have attempted to restimulate the entire well, requiring a larger, more expensive treatment with little assurance
of effectively treating the upper three intervals. In addition to identifying target restimulations, the tracer logs promoted a
SPE 134330 5

better understanding of fracture containment, fracture staging, and perforating strategies. When bypassed pay intervals are
located between existing perforations, coiled tubing with isolation packers or seal assemblies can be used to selectively
isolate and restimulate desired intervals [Marsh, 2000] or casing liners may be preferred [Barba, 2009]. In some locations,
due to environmental or safety concerns, the use of radioactive tracer is discouraged or forbidden. One proppant supplier can
coat proppants with a resin containing taggant that can be made temporarily radioactive during logging to avoid handling
radioactive materials at surface. Another ceramic manufacturer provides proppant in which an entirely non-radioactive
material is permanently incorporated in each proppant pellet that can be detected with standard neutron logs. In addition to
addressing environmental, safety and tracer segregation concerns, these innovations allow the proppant location to be
identified at any time in the future, compared to traditional radioactive tracers which degrade with half-life ranging from 60-
85 days. This may allow identification of bypassed pay intervals many years following the original treatment.

Field Example – Improved Coverage of Horizontal Wellbore with Refrac

With increased drilling of horizontal wellbores have come additional challenges with cost-effective stage isolation to achieve
stimulation of all desired intervals along the lateral. In numerous reservoirs developed with horizontal wells, restimulation
treatments have been proven to be cost-effective and can increase both production rates and ultimate recovery of reserves. In
the Elm Coulee field of Montana, Lantz et al (2007) demonstrated that restimulation of cemented laterals could achieve
production rates exceeding the peak production from the initial fracture treatment. The Estimated Ultimate Recovery (EUR)
of sixteen refractured wellbores was increased by over 1,300,000 barrels. In this development, the operator intended to drill
wellbores in the direction of the maximum principal stress to achieve longitudinal fractures, although evidence of transverse
growth was observed in some wells. Radioactive tracers were utilized to determine whether portions of the wellbore were
unstimulated by the original treatment. As shown in Figure 4, the initial treatment in one well failed to contact the entire
wellbore. Additional perforations were added and a refrac was performed utilizing 10 ppg proppant slugs and ball sealers to
improve diversion. Significantly better coverage and entry of proppant into previously unstimulated areas is apparent in the
upper track of Figure 4.

▼Tracer from refrac

Tracer from original frac▲

Figure 4 – Tracer logs in horizontal well indicate coverage was increased by refrac. [Figure adapted from Lantz, 2007]

The effect of depletion was observed in the treating records, as breakdown pressures averaged 669 psi lower than initial
treatments. However, net pressure generated during the refracs was approximately 50% greater than during the initial fracs.
Pressure data frequently suggested that refracs were diverted into higher stress (undrained) areas of the reservoir. Post-
refracture production often achieved production similar to initial rates, despite pressure depletion from producing tens of
thousands of barrels of oil. Post-refrac, the average gas-oil-ratio (GOR) declined from 915 scf/bbl to near-original 520
scf/bbl, again indicating the refracture treatments had contacted virgin reservoir not in communication with the initial
treatments. Figure 5 shows the production from the same well depicted in Figure 4.
6 SPE 134330

Figure 5 – Production rate increased and GOR dropped following the refrac in May 2004, corroborating pressure data
suggesting the refrac contacted previously undrained portions of the reservoir. [Figure from Lantz, 2007]

Field Example – Benefit of Increased Conductivity in Refracs

The Kuparuk Field on the North Slope of Alaska may be the most definitive demonstration of the benefit of improved
fracture conductivity in restimulation treatments, with results from more than 200 refracs reported [Pospisil, 1992]. The
Kuparuk “A Sand” is a continuous, blanket reservoir with predictable well productivity. Although the overlying “C Sand”
had a relatively mature waterflood, only on rare occasions would a frac initiated in the A Sand increase the watercut –
demonstrating that most fractures remained contained in the target A Sand. Therefore, the benefit of refracturing cannot be
attributed to height growth or contacting additional pay in the vertical extent; instead, it is reasonable to infer that the entire
20-30 feet of fully-perforated pay was penetrated by all fracturing attempts. Refracturing was never observed to result in
premature breakthrough of the line-drive waterflood, suggesting that no dramatic reorientation of fracture azimuth was
achieved. Previous efforts to inject up to 1,000,000 pounds of sand at low concentration provided no incremental benefit
beyond short fracs with similar low conductivity, demonstrating that improved reservoir contact was not beneficial when
constrained by a low conductivity fracture. Despite low closure stresses of 3200 psi, ceramic proppants were eventually
incorporated – yielding a study in which proppant degradation should be minimal. For these and other reasons, it is believed
that the primary benefit of refracturing in the Kuparuk field can be attributed to increases in fracture conductivity.
As shown in Figure 6a, natural frac sand was initially considered to be adequate due to low expected stresses of ~3200 psi.
However, many wells received multiple refrac treatments in the same interval, with downhole pressure buildup testing
between each treatment. Optimization involved switching to Light Weight Ceramic proppants (LWC), then to progressively
larger proppant sizes, higher concentrations and reduced pad size. A systematic reduction of silica flour, 100-mesh sand and
other damaging fluid loss additives also served to improve retained fracture conductivity. Figure 6a shows the success of
nearly 200 refracture stimulations, demonstrating that it was possible to double the productivity of the previously stimulated
wells by placing progressively higher conductivity fractures. In this figure, no attempt was made to adjust for changing
reservoir pressure. Stimulation treatments were initially focused on producing wells, and the increased productivity outran
the existing waterflood injection support. As the reservoir pressure dropped, the Gas-Oil-Ratio increased, requiring many of
the restimulated wells to be shut in during the summer months due to facility gas handling constraints. The reduced reservoir
pressure and increased shut in time is largely responsible for the decline of the well rates in Figure 6a. Reservoir pressure
recovered from a combination of shut-in production and stimulation of injection wells. This improved pressure support
provided much of the “rebound” shown in Phase IV wells’ rates.
SPE 134330 7

70 4000
Phase IV - 12/18 LWC (52 wells)
Phase III - 16/20 LWC (97 wells) Original Fracture (20/40 Sand)
3500

Production from Fracture (bfpd)

60
(6a) Phase II - 20/40 LWC (27 wells)
Phase I - 20/40 Sand (9 wells)
(6b) Phase I refrac (20/40 Sand)
3000 Phase III refrac (16/20 LWC)
50
BOPD/ Foot of Pay

2500
40
First
2000 Refrac
Incremental
30 Oil Exceeds
1500 1,000,000
barrels
20
1000 Incremental
Oil exceeds
650,000
10 500 barrels
Second
Refrac

0 0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 May-84 May-86 May-88 May-90 May-92 May-94 May-96 May-98 May-00

Time Since Refrac (months) Date

Figure 6 – The results from 185 refracturing attempts showed increasingly higher productivity possible with improved
fracture conductivity (a). Initial production rate from Kuparuk Well 2F-08 was increased from 500 bopd to over 3500 bopd by
more aggressive refracturing (b). Figures adapted from SPE 24857 with updated production data.

Figure 6b shows the production response of a single Kuparuk well. Although injection support was inadequate to sustain the
high production rates, it is clear that each refracture treatment was fantastically economic. These production gains cannot be
explained using traditional modeling assumptions; however, models which incorporate non-Darcy, multiphase flow, cyclic
stress, and other realistic damage factors can match the observed production response achieved with improved fracture
conductivity.
Field Examples – Proppant Durability Affecting Refrac Success
Although poorly recognized by most frac engineers, there are ample laboratory data demonstrating that the flow capacity of
proppant packs degrade over time in various laboratory conditions. In Figure 7a, McDaniel (1986) tested sand, resin coated
sand (RCS) and three ceramic proppants at 8,000 psi confining stress. Within 15 days, the sand lost 80% of the flow
capacity, RCS lost 55% and ceramics lost 25-30%. McDaniel showed similar degradation regardless of whether proppant
was tested in brine or in dry nitrogen gas. McDaniel recommended that engineers adjust published data - assuming sand
would sustain 7-8% of the short term conductivity, while RCS would sustain 17% and ceramics would retain ~50%. Testing
by Cobb (1986) in a Teflon sleeve designed to eliminate cell corrosion (Figure 7b) demonstrated that ceramic proppants
confined at 10,000 psi stress lost ~20% of their conductivity over 70 days while sand confined at 5000 psi lost more than
30%.

1
Permeability Ratio

0.8

0.6

0.4 IDC at 10,000 psi (69 MPa)

LWC at 10,000 psi (69 MPa)
0.2 Sand at 5000 psi (35 MPa)

0
0 15 30 45 60 75
(7a) (7b) Days at Constant Stress

Figure 7– Extended duration tests routinely show continued degradation to the flow capacity of proppants. Tested at 8500 psi
stress in a conductivity cell (7a), sand and RCS showed more severe degradation over time compared to three ceramics
[McDaniel, 1986]. In tests designed to eliminate all cell corrosion (7b), ceramics in 2% KCl at 200ºF exhibited 20% loss in
conductivity while sand lost 30% despite being subjected to merely half the closure stress. [Cobb, 1986]
8 SPE 134330

Hahn (1986) showed that, when tested at 8,500 psi in a non-corrodible sleeve, two ceramics lost between 20% and 50%
of the flow capacity over 50 days while white frac sand and resin coated sand lost ~90% of their flow capacity within 7 days
(Figure 8a). Montgomery (1984) showed that two sand samples held between steel plates at 5,000 psi confining stress lost
~50% of their flow capacity over a nine month period (Figure 8b).

10000
100

% Original Conductivity
Conductivity (md-ft)

80 20/40 Sand at 75F

10/20 Sand at 250F
60
1000
IDC - Intermediate Density Ceramic
40
Proflow - Precursor to LWC
RCS - Resin Coated Sand
20
Ottawa Sand
100
0
0 10 20 30 40 50
Days at Constant Stress, 8500 psi 0 30 60 90 120 150 180 210 240 270
Days at Constant Stress, 5000 psi
(8a) (8b)

Figure 8 – Hahn [Hahn, 1986] showed two ceramics to lose 20-50% of the initial conductivity within 40 days, while sand and RCS
lost ~90% of the flow capacity within one week (8a). Montgomery [Montgomery, 1984] held two sand samples in a conductivity cell
for 9 months, documenting over 50% reduction in flow capacity (8b).

More recent testing in modern cells flowing deoxygenated, silica-saturated 2% KCl brine and inert nitrogen gas further
reinforce these conclusions. Handren (2007) presented a more descriptive evaluation of the non-Darcy effects measured
throughout a 40 day test with Economy Light Weight Ceramic (ELWC), RCS, and Uncoated Sand. While all proppants
degraded over time, the ceramics were more stable. Initial beta factors were ~2.5 times lower (superior) with ceramic, but
after 40 days they were 3-6 times superior. Despite the relatively benign conditions of 6,000 psi closure stress at 250º F, it is
clear that proppants experience significant degradation in the test cell. Unfortunately, no commercial simulators have
incorporated proppant degradation into production models and few engineers design fractures to accommodate degradation
over time. It is possible that proppant degradation is partially responsible for steep production decline and subsequent
success of refracture treatments. The author is not aware of a single field trial specifically evaluating whether more durable
proppants will avoid or delay the need for restimulation. However, in the Bakken reservoir of North Dakota and Montana,
several studies can be combined to make a comparison of refrac attempts in horizontal wells:
− In cemented Bakken wells initially stimulated with sand, refrac attempts were economically successful in essentially
100% of wells (new perforations were added). [Lantz, 2007]
− In Bakken wells with uncemented liners initially fractured with sand, refrac attempts were economically successful
in 87% of completions (retreating existing perforations) [Eberhard, 2008]
− In Bakken wells initially treated with ceramic proppant, refrac attempts are economically merited in less than 50%
of cases. Successful instances appear to correlate with wells in which radioactive tracer suggested poor coverage
during initial treatment. [Besler, 2007, 2008]
While the dataset is imperfect, these results suggest that more durable ceramic proppants have delayed and/or reduced the
need to restimulate Bakken wells. There are a large number of additional field results which demonstrate superior EUR has
been achieved with superior proppants [Vincent, 2009]; however, few studies specifically investigate proppant durability.
There are numerous studies in the Appendix in which sand or glass beads were used in initial treatments, followed by
successful restimulation with stronger ceramic proppant. However, it does not appear that increased productivity can be
confidently attributed solely to proppant durability instead of an overall increase in fracture conductivity.

Field Example – “Replacing” Degraded Proppant

The success of a great number of the refrac examples in the Appendix owe a portion of their success to restoring conductivity
to fractures that were damaged due to proppant embedment, degradation, fines plugging, asphaltene precipitation, and other
phenomena that result in compaction and/or loss of permeability within the proppant pack. One field example is the Rangely
oil field of western Colorado, in which barium sulfate scale and asphaltene precipitation is believed to degrade the fractures.
Bagzis (1989) documented the evolution of 1700 refracture treatments in the 5 mD Weber sandstone from 1947 through
1989, ranging from unpropped oil squeezes to more aggressive fracs carrying high concentrations of proppant. In this field,
most wells have received three or four restimulations but still remain viable restimulation candidates. An early Rangely Field
SPE 134330 9

refrac performed in 1953 increased production from 30 bopd to 142 bopd and 58 bwpd, declining to 96 bopd after 15 months.
Extensive comparisons indicate that lower sand concentrations resulted in steeper production declines, attributed to crushing
and embedment of the frac sand. According to Bagzis, “in general, refrac successes have been achieved by using higher
sand concentrations…” During 1987-1989, when treatments specified 8 to 10 ppg concentrations, refracs were successful in
83% of wells, superior to 69% to 77% success rates previously achieved when using lower proppant concentrations. Hejl
(1992) noted that further increases in the designed half-length or sand concentrations (above 10-14 lb/gal) “had no noticeable
change on the success rate” and had not resulted in premature breakthrough of the CO2 flood. Although restoration of
fracture conductivity lost due to scale precipitation was a primary goal of these treatments, many refracs were successful in
the Rangely field in wells that had no indication of scale formation. Bagzis noted that particularly good refracturing
candidates include wells in which the previous stimulation met one of more of the following conditions:
− Fracs with premature screenouts, failing to place a sizeable percentage of designed treatment
− Fracs overflushed or staged with diverters
− Jobs that did not cover the entire pay zone due to pump rate limitations or mechanical limitations
− Fracs using low sand concentration (1 to 4 ppg)
− Small frac lengths relative to well spacing
− Settled pack jobs that did not reach an equilibrium frac height covering the pay
While it appears that the benefit of refracturing was likely achieved by more than one mechanism, the extensive history
documented at the Rangely Field indicates continued degradation of fracture conductivity and demonstrates economic value
in restimulation with higher conductivity fractures.

Field Examples – Different Fracturing Fluids

Some reservoirs can be effectively stimulated with a variety of fracturing fluids while other formations appear to be more
sensitive to specific fluid formulation. Smith and Murphy (1974) tested seven different fracturing fluids including water,
alcohol, crosslinked guars, emulsions, gelled methanol, gelled acids, and gelled condensate in the Canyon Sand. While each
fluid had specific advantages, the authors determined that avoiding formation damage could not economically be a primary
factor in selecting the frac fluid. A well initially treated with a 30,000 gallon waterfrac was retreated with 320,000 lb sand
using an emulsified fluid, but achieved insignificant improvement in production rates, clearly an economic failure.

Leshchyshyn et al (1999), however, showed that economic restimulation of wells believed to be damaged by initial water-
based treatments can be achieved. Two wells initially completed in the Viking and Glauconite formations of Canada were
believed to be undersaturated with water and failed to respond to initial water-based frac treatments. The wells suffered from
early screenouts and poor load recovery and were considered for abandonment. Refrac treatments with oil-based fluids and
larger proppant volumes resulted in excellent load recovery and sustained production rates. If initial damage was indeed
caused by aqueous phase trapping, these field examples suggest that damage can be reversed or bypassed by restimulation.

In the Olmos formation in Texas, initial frac treatments with low proppant concentrations in water-based fluids provided
disappointing results. As documented by Pauls et al (1985), gelled hydrocarbon fluids carrying extremely high proppant
concentrations reaching 17 ppg were much more successful. One well initially treated with 3 ppg proppant in gelled
concentrate was refractured with 12 ppg proppant in gelled diesel, achieving “remarkable” increases in oil production.

In gas storage fields, the region near the wellbore is commonly desiccated due to years of injection and withdrawal of
pipeline-quality (dehydrated) natural gas. Nonetheless, deliverability loss is significant in gas storage fields, and tens of
millions of dollars are spent annually to recover lost deliverability [Brown et al, 2003]. Restimulation attempts (matrix acid
injections, acid fracs, or hydraulic propped fractures) re-introduce liquids which may defer the benefits of restimulations by 6
to 48 months (mean of 16 months). These wells may provide a unique opportunity to investigate the effect of near-wellbore
saturations on deliverability, but analyses have been plagued by poor record-keeping and generally low-tech approaches to
frac optimization.

Cipolla (2005) evaluated the restimulation of a well in the Barnett Shale that was initially drilled to achieve a longitudinal
fracture. The initial treatment using crosslinked gel resulted in microseismic activity predominantly confined near the
wellbore axis (Fig 9a). However, a subsequent refrac using slickwater was observed to induce microseismic activity over a
larger reservoir area (Fig 9b) and provided significant increase in gas production rate (Fig 9c). Although the breadth of
microseismic events may be due to a combination of less viscous fluid and the stress induced by the initial treatment, it has
been shown that reduced fluid viscosity generally results in more complex fracture geometry [Cipolla et al, 2008].
10 SPE 134330

(9a) (9b)
(9c)
Figure 9 – Restimulation of Barnett well with slickwater is believed to contribute to increased fracture network complexity and
superior gas production [Cipolla, 2005].

Field Example – Refrac Success due to Superior Containment

Gil et al (2007) described how some fractures may undesirably grow out of zone due to relatively small stress barriers,
wasting proppant out of zone and creating insufficient fracture length. Subsequent production can reduce the stress in the
productive layers, allowing refracs to be more contained in the desired intervals.

Field Example – Conflicting Interpretations or Multiple Mechanisms affecting Refrac Success

Wells in the Denver-Julesburg (DJ) Basin have been frequently restimulated with refracs, tri-fracs, quads, and even the
colorfully named “cinco-de-fraco”. The mixed success of 29 restimulation treatments in the Muddy J formation were
reported as early as 1979 [Parrot, 1979], indicating that larger treatments were often more productive. Restimulation
treatments in the Codell formation began in the mid 1990’s and eight recent papers have attributed restimulation success to a
variety of different mechanisms. The Codell formation is a gas/condensate reservoir at ~7500 ft. depth and has been
penetrated by thousands of wells. Typical GOR can reach 15,000 scf/bbl, introducing significant multiphase flow
complications, with the added potential of condensate banking in the reservoir and/or fracture. Between 1997 and 2001, the
industry restimulated 1500 Codell wells with three early interpretations:
− Emrich et al (2001) emphasized the role of fracturing fluid composition and injection rates in refrac success. Codell
restimulations with a reduced polymer CMG (carboxymethyl guar) fluid provided approximately 200% rate of
return (ROR), compared to 66% with previous fluids and selection criteria, and typically increased each well’s
recoverable reserves by 35,000 boe.
− Shaefer and Lytle (2001) also focused on fracturing fluids, reviewing the variety of fluid systems historically
utilized in the Codell. Comparing 479 recent restimulation treatments, fracs utilizing lower polymer CMG with
cleaner base water typically increased pre-refrac rates by 500% and produced favorably to almost all offsets
restimulated with high polymer hydroxypropyl guar (HPG). Attempts to determine effective frac lengths via
production data analyses were compromised due to multiphase flow and condensate banking in the fracture –
making the more productive CMG frac appear 1/5 the length of an offset HPG frac, despite significantly better
productivity.
− Meanwhile, Sencenbaugh et al (2001) published the results of an early neural net analysis, which was inconclusive
at defining factors critical to success in recent restimulation attempts. Although the restimulation of 750 wells
resulted in economic rate of return (ROR) exceeding 100% and attractive reserve additions at costs below $4.23/boe,
the characteristics of superior candidates were not apparent. Sencenbaugh noted that although current well spacing
was 40 acres, simulation and production observations suggested most wells appear to be draining only 15 to 30
acres. Refrac success was largely attributed to extension of fracture half-lengths during restimulation. Refracturing
was claimed to intersect essentially “virgin” reservoir pressure and often achieved rates comparable to the original
treatment. Restimulation was found to be the preferable choice to infill drilling, due to reduced capital costs and
surface disturbance.
The following year, Asadi et al (2002) attempted to examine the flowback and cleanup of fracs in the Codell using unique
chemical tracers in each stage. The concentration of these water-soluble tracers in the flowback fluid indicated poor recovery
of cross-linked stages, ranging from 4% to 16%, attributed to excessive crosslinker and stabilizer coupled with conservative
breaker concentrations. By 2004, the number of Codell refracs exceeded 4000 [Miller, 2004] and results from 1000 recent
 SPE 134330 11

refracs suggested the viscosity profile of the fracturing fluids may correlate to increased refrac success. When fluid
viscosities were too high, undesirable height growth was achieved. With low fluid viscosity, poor sand transport caused
bridging and elevated net pressures, again suspected to result in undesirable growth out of zone. Interestingly, optimization
of fluid viscosities reduced the frequency of fractures breaking into offset wells, which was attributed to a significant change
in fracture azimuth. Restimulation pressures and subsequent production rates were similar to or greater than initial
treatments, interpreted as the refrac contacting undrained portions of the reservoir. However, another operator [Pagano,
2006] reached a different conclusion as to the mechanism that resulted in “unconditional” production improvement with
sequential restimulations in the Codell, as shown in Figure 10.

Figure 10 – Production after initial treatment, refrac and “tri-stim” of Codell [Pagano, 2006].

Pagano posited that instead of considering the Codell as a tight sandstone, it should be envisioned as a shaly siltstone with
minimal to no effective porosity. His core and field testing suggested that the Codell has a dual-porosity network with
significant pseuodomicroporosity possessing significant interconnectedness under virgin overpressured conditions. However,
after the initial pore pressure (~0.62 psi/ft gradient) is depleted, the interconnectedness collapses, causing production declines
that are further exaggerated due to condensate banking and increased viscosity of dead oil. Pagano claimed that refracs
reengage the closed and potentially damaged microporosity, thereby renewing productivity. Pagano’s review of GOR trends
and character of produced fluids indicated that refracs do not contact new, virgin rock, but instead reengage previously
drained reservoir. This conclusion was challenged by subsequent tiltmeter data [Wolhart, 2007] indicating that initial
fracture azimuths were N66ºE in the Codell, while significant and highly variable reorientation was observed during
restimulation treatments. Interestingly, there was no obvious correlation between the degree of refracture reorientation and
subsequent production. While Wolhart recognized that increased fracture length and improved conductivity may also play a
role in refracture success, reorientation was documented to be a significant component in the Codell. Cramer (2008)
provided some different interpretations and several practical suggestions regarding restimulation of Codell wells. Cramer
interpreted the failure of low viscosity fluids in the Codell to be attributed to downward fracture growth and proppant settling
below the pay interval – necessitating high viscosity fluids for proppant suspension, yet low gel loadings required for
cleanup. Often, following a 3000 barrel refrac treatment, many Codell wells will recover only 50-100 barrels of fluid and
then quit flowing. Shutting those wells for 30-60 days following the treatment often allows pressure to build and perhaps
fluid to dissipate such that commercial production rates can be sustained. When initial fracture treatments intersect offset
Codell wells, Cramer notes that simultaneously refracturing both offset wellbores may be one of the most effective ways to
place an effective stimulation. Cramer also recognized that asymmetrical drainage induced by condensate banking in the
Codell can provide an excellent opportunity for refracturing, especially when some degree of reorientation is achieved.
Despite the industry experience with over 4000 successful refracs in the Codell, there remains significant disagreement as to
the predominant mechanisms and optimal restimulation design. Data gathering in the DJ Basin is often hindered because
many wells are completed in multiple stages in the Codell, Niobrara, and J-Sand intervals which are commingled with
infrequent production logs to identify production from each set of perforations. Additionally, most wells are not individually
metered, instead flowing to a common manifold with production allocated mathematically to each well instead of directly
measured. Despite these engineering challenges, significant refrac programs are economic and pursued by most DJ Basin
operators.
12 SPE 134330

Field Example – Restimulating Unpropped or Inadequately Propped Fractures

There is ample evidence demonstrating that most unpropped or inadequately propped portions of fractures will collapse over
time, necessitating restimulation to recover productivity.
Restimulation of unpropped fractures: Many formations can sustain some degree of stimulation following an injection
treatment without proppant [Palisch 2008]. However, documented comparisons indicate that, in most reservoir types,
superior production and recovery can be achieved when proppant is placed within the fracture [Howard, 1970; Ghauri, 1980;
Lambert, 1980; Palmer, 1992; MacDonald, 2002; Harper, 1985; Britt, 2006; Coulter, 2004; Gottschling, 1985; Rieb, 2005;
Rieb, 2006; Crow, 1977; Rushing, 2003; Wong, 2000; Tuffs, 2009]. Since propped treatments sustain greater production, it
can be concluded that most unpropped fractures close over time – validated by laboratory studies documenting collapse of
unpropped fractures and collapse of unpropped arches above proppant banks [Stim-Lab, 2007-2009].
Restimulation following initial treatments with inadequate proppant conveyed in viscous fluids: It is important to
recognize that most fracture treatments inject diffuse slurries of proppant. For example, 1 pound of sand added to 1 gallon of
gelled water represents a very diffuse slurry – approximately 5% solid and 95% fluid. Following injection, the fracture may
slowly close upon the few suspended proppant grains, resulting in significant reductions in fracture width and the potential
for stranded or discontinuous proppant beds [Vincent, 2009]. There are many cases shown in the Appendix in which wells
were restimulated with higher concentrations of proppant, achieving dramatically improved production rates - which may be
attributed to improved continuity of the proppant pack [in addition to improved conductivity achieved with superior fracture
widths].
In the Cotton Valley, many wells were initially fractured with large quantities of frac sand (150,000 to 800,000 lb) at
modest concentrations. Numerous refracs were documented [Hunter, 1986] with proppant concentrations reaching 10 ppg,
often achieving 3-4 fold increases in production and 1 BCF of added reserves per well.
In the AWP Olmos Field [Saucier, 1988], proppant concentrations reached the “unprecedented” levels of 15 lb/gal, as the
operator observed superior sustained production (reduced decline rates) in wells with higher proppant concentrations. An oil
well initially treated with 6 ppg sand had declined to 80 bopd and was subsequently restimulated with 15 ppg sand,
increasing production to over 140 bopd. Data compiled from 17 refrac attempts in the Olmos [Conway, 1985] are shown in
Figure 11, demonstrating that rates were more stable following refracs with high proppant concentration compared to initial
fracs utilizing low concentrations.

250
Average Well 1 Well 2 Well 3 Well 4 Well 5
Well 6 Well 7 Well 8 Well 9 Well 10 Well 11
Well 12 Well 13 Well 14 Well 15 Well 16 Well 17

200
More stable
production from
refracs at 14 ppg
Steep production
declines following 150
initial fracs
BOPD

100

50

1 month after initial 6 months after 12 months after 1 month after 6 months after 12 months after
frac initial frac initial frac Refrac Refrac Refrac

Initial fracs, low prop concentration Re-fracs, high prop concentration

Figure 11 – Initial fracs utilized low proppant concentrations, often demonstrating steep production declines. Refracs targeting
14+ ppg provided superior sustained production [Adapted from Conway, 1985].

In the McAllen Ranch [Saucier, 1988] a gas well initially fractured with sand was refractured with intermediate density
ceramic, increasing production from 250 mcfd to 1600 mcfd. Later work in the McAllen Ranch Field [Durrani, 1994]
corroborated these results when a well initially fractured with <5 ppg sand was subsequently refractured with 8 ppg ceramic,
increasing production from 150 mcfd to more than 1000 mcfd.
SPE 134330 13

Restimulation of initial slickwater treatments/low viscosity treatments: Clearly, in many slickwater treatments, the vast
majority of the created fracture volume is not effectively propped. Even operators concluding that gas production from the
Barnett is correlated with the proppant mass [Coulter et al, 2004] may introduce merely 300,000 lbs of sand in 600,000 to
1,500,000 gallons of water – approximately 1% to 4% sand by volume and 96% to 99% water. Since leakoff is modest in
low permeability reservoirs, most injected fluids are believed to remain in the fracture during the treatment, and it is clear that
the entire created fracture network cannot be effectively propped. Some portion of the eventual decline in productivity
should be attributed to progressive collapse of inadequately propped portions of the fracture. Analyses of production from 41
Barnett wells demonstrate that the fracture networks do not sustain high relative conductivity throughout the mapped extent
[Cipolla, 2009]. Significant improvements in production would be expected if fracture conductivity and continuity can be
increased.
Restimulation treatments are frequently economic in shale gas reservoirs following slickwater treatments, with several
examples in the Appendix. There is uncertainty as to what portion of the success should be attributed to reorientation of the
second treatment, allowing contact with previously undrained rock, or whether much of the result is from simply replacing
degraded proppant, dilating the unpropped but collapsed fissures, and providing new but temporary access to the fracture
network before the wellbore connection progressively collapses again.

Field Example – Enlarged Fracture Geometry and Improved Reservoir Contact

In the 1950’s, it was common to complete the Mesaverde formation by detonating a charge of nitroglycerin in an openhole
section of the wellbore. Hower and Decker (1992) provided production performance analyses and geological evidence
suggesting that these stimulations failed to open all the productive sand lenses in the heterogeneous pay zone. Wells
exhibiting linear p/z trends [Fetkovich, 1990] were believed to be producing from single intervals and were identified as
restimulation candidates. Despite more than 30 years of production, 22 wells were sidetracked or redrilled with cemented
casing and stimulated with propped fractures. Tremendous incremental reserves exceeding 2 BCF per well were added, even
when the bottomhole locations were within a short distance from original wellbore. One well was abandoned due to drilling
challenges; however, 22 wells were outstanding economic successes with reserves added for approximately $0.20/mcf.

Field Example – Reorientation of Refracs

Fractures tend to propagate in the direction of the maximum compressional stress, as this orientation allows them to open
(generate width) against the smallest principal stress. It has long been recognized that the stress field surrounding a well can
be altered as a result of fluid production, injection, or creation of nearby fractures. Warpinski and Branagan (1989)
recognized that stress reorientation could provide an opportunity to achieve a favorable fracture orientation. The first field
demonstrations of fracture reorientation were provided by Wright & Conant (1994) and Wright & Stewart (1994) when
tiltmeter mapping of refracture treatments proved that dip and azimuth of treatments were affected by production-induced
fluid pressure gradients or potentially by shear failure of the formation. The tendency of fractures to grow toward high stress
(to avoid opening against high stress) may induce refracture treatments to become “reserve seeking missiles” as they may
reorient toward higher stress, undrained regions of the reservoir. On the other hand, in waterfloods, restimulation of
producers may cause undesirable growth of fractures towards injection wells, resulting in increased watercut and decreased
sweep efficiency.

It is important to note that reorientation can be induced by two distinct mechanisms – stress created by the opening of an
adjacent fracture, or pore pressure alteration due to previous production or injection periods. Roussel & Sharma (2010) used
a 3d model to examine the extent of stress perturbation induced by adjacent fractures and provided guidelines as to the degree
of interference between closely spaced fractures in horizontal wells. The paper provides a mathematical spacing to maximize
reservoir contact by reducing interference and achieving mostly parallel transverse fractures. The paper noted cases in which
complete stress reversal would be observed, resulting in longitudinal fracture growth along a wellbore oriented to achieve
transverse fracs in the initial stress field. The authors viewed this outcome as undesirable in terms of increased reservoir
contact, and did not explore the potential benefits in wellbore connectivity of inducing a subsequent longitudinal fracture that
could “link up” existing transverse fractures and provide improved connection with the wellbore. Roussel & Sharma (2009)
examined the stress perturbations induced by production or injection and how poroelastic stress alteration will affect
orientation of refracture treatments.

Examples of aziumthal reorientation of refracs have been documented in the Lost Hills Diatomite [Wright & Stewart, 1994],
[Wright & Conant, 1995], the Austin Chalk [Wright & Conant, 1994], the DJ Basin [Wolhart, 2007], the Barnett Shale
[Siebrits et al, 2000], the Van oilfield [Wright & Conant, 1995] the Ansai oilfield [Li, 2006], the Daqing oilfield [Wang,
2007] and [Liu 2008], the Xin Zhan field [Yao et al, 2007] and in the Mounds cuttings disposal study [Moschovidis, 2000].
Diversion and/or reorientation of subsequent treatments in a horizontal well have been shown by Leonard (2009), Dunek et al
(2009) and Waters et al (2009). Some evidence suggesting reorientation in the Black Warrior CBM was summarized by
Palmer (1993). Surjaatmada (2007) demonstrated that a refracture performed within 30-60 minutes following a previous
treatment can induce diversion and reorientation, and Li (2006), Wang (2007), and Yao (2007) showed that use of paraffin
14 SPE 134330

balls or other temporary diverters in conjunction with temporary shut-ins can influence refrac reorientation.

Field Example – Production Data Analyses Indicating Need to Refrac

Shaefer (2006) conducted multiple types of analyses to evaluate the production history of four wells stimulated with
hydraulic fractures in the 2nd Frontier formation in Wyoming. The closure stress was estimated at 5500 psi, and 20/40
Ottawa sand had been selected for the initial completions. However, Reciprocal Productivity Index (RPI), Agarwal-
Gringarten, and Pseudo Steady State (PSS) analyses all showed poor apparent frac lengths. All four wells were restimulated
with economy light weight ceramic or intermediate density ceramic. As shown in Figure 12a, these treatments increased
well production above initial production rates, despite evidence that the pore pressure gradient had been depleted to 0.40
psi/ft from initial gradient of 0.52 psi/ft during the first two years of production. Subsequent analyses showed the effective
fracture lengths were dramatically increased compared to initial sand completions, despite smaller injected masses of
proppant. Two wells that had been producing for 17 years (Fig 12b) and were anticipated to have much greater depletion
still showed dramatic increases in production following restimulation with ceramic proppant.

(12a) (12b)

Figure 12 – Four wells were initially fractured with 20/40 Ottawa sand. Two wells restimulated with ceramic proppant after
two years of production showed 5-fold production increases and >20,000 BOE incremental gas in first 2 years (Fig 12a).
Two wells restimulated with ceramic proppant following 17 years of production provided 4-fold production increases and
>14,000 BOE incremental gas in first 20 months (Fig 12b).

Field Examples – Refracs even though Production Data Analyses Suggested Initial Fracs were Sufficient
In some fields, production trends or pressure transient testing can indicate whether the well will benefit from refracturing. In
the Norge Marchand Unit of Oklahoma, initial fracs typically specified 66,000 lbs of 20/40 or 10/20 sand at concentrations
up to 3 ppg [Branch, 1991]. Pressure buildups (PBUs) were gathered on 36 wells, and of 8 wells selected based on buildup
interpretation, 50% success was achieved in restimulating. Two wells were refractured despite NOT being recommended
based on PBU results with neither being successful. Five wells were refractured without any PBU testing with a 40%
success rate. In general, wells with superior permeability and high skin (determined via pressure transient analyses) were the
best candidates. Good wells made good refrac candidates; selecting low performing wells drastically hindered economic
success of refracs.
However, in other fields, refracs have been successful even when the well test indicated the presence of a deeply
penetrating, highly conductive frac [Elbel and Mack, 1993]. It has been argued that these results may be attributed to refrac
reorientation or frac extension into regions of higher pore pressure.

Field Examples – Restimulating Shallow, High Perm Coal Bed Methane (CBM)
It is likely that the combination of damage mechanisms in soft coal formations may be unique from other examples
previously discussed. Stimulation of CBM is particularly challenging for the following reasons:
• Proppant in soft coal formations is particularly susceptible to embedment and coal spalling into proppant pack
• Mobile coal fines may progressively plug proppant packs, especially those with irregular proppant size/shape
• High fracture conductivity may be necessary to efficiently dewater the coal and accommodate multiphase flow
• Gas desorption is driven by Fickian diffusion, a process much more sensitive to pressure than Darcy flow.
Therefore, proppant packs with minimal pressure losses may have much larger benefit in coal than expected.

Numerous authors have recognized that fracture conductivity can be a primary concern in CBM stimulation. Jeu et al (1988)
SPE 134330 15

noted that “high fracture conductivity is paramount” and that larger mesh proppants are less prone to coal fines plugging and
thereby allow superior sustained production. McDaniel (1990) noted that “very high fracture conductivity is needed to
ensure rapid dewatering” of low pressured CBM reservoirs, which has made 12/20 sand a popular material for many
operators. Palmer (1992) noted that high fracture conductivity “is more important than heretofore recognized” in CBM
completions. Lehman et al (1998) noted that ultimate gas recovery from CBM depends on maintaining fracture conductivity.
Rodvelt et al (2001) analyzed 900 CBM wells in Virginia and noted that production problems were caused by inadequate
fracture conductivity. Since CBM desorption is driven by Fickian diffusion, production rates are much more pressure-
dependent than typical gas wells [Lehman (1998) and McCabe (1999)] and reducing unnecessary pressure losses is critical.

Clawson (2003) showed that gas production rates from the Raton Basin CBM were doubled when surface facility
modifications (wellhead compression, field boosters, looping gathering lines) allowed the flowing pressure to be lowered. In
many cases, reducing the flowing pressure by 30 psi made all the difference, converting water wells into productive methane
wells. If reducing the pressure at surface is this effective, reducing the pressure losses in the fracture should provide similar
or larger benefits, as desorption could be initiated throughout a larger portion of the contacted reservoir. In the Black Warrior
Basin, Palmer (1992) showed that increased proppant concentrations generally resulted in more effective cumulative gas
recovery, as shown in Figure 13a. In the Helper Field in Northeast Utah, Stutz et al (2002) documented that gas production
rates could be increased 15-fold with larger, more conductive fracture treatments with higher sand concentrations (Figure
13b). Other analyses by Stutz show a dramatic increase in production with larger frac treatments. However, in neither study
could it be proved that the increased production was solely attributed to fracture conductivity, as the treatments were believed
to increase the fracture length and height as well.
2001: Re-frac, 330,000
lb 16/30 sand
20
Cum MMCF/Years Produced/Ft of Pay

1999: Initial frac,

60,000 lb 16/30 sand

15-fold increase

10

0
0 10 20 30 40 50
Pounds of Sand/Ft of Pay (thousands)

(13a) (13b)
Figure 13 – In the Black Warrior Basin of Alabama, Palmer (1992) showed cumulative well production corresponded to
proppant concentration (13a). In the Helper Field of Utah, Stutz et al (2002) showed that restimulation of a well with 6 times
the frac sand provided a 15-fold increase in CBM production (13b). Figures adapted from Palmer (1992) and Stutz (2002).

One CBM trial was designed to specifically investigate the effect of fracture conductivity upon methane production in the
San Juan Basin refracs. The operator initiated a 20-well trial in 1999, with the intent to refracture 10 wells with the “typical”
20/40 sand design, and 10 wells with 20/40 ceramic proppant. Note that the Fruitland coal lies at ~3500 ft depth, and
crushing of the frac sand was not a concern. Rather, the operator selected man-made proppant for the uniform size and
spherical shape as it was believed to be more resistant to compaction, fines plugging, and would provide reduced inertial
pressure losses. Other than the proppant substitution, all other treatment parameters were to be held constant between wells
refractured with sand and ceramic. The first two wells were restimulated with ceramic (Figure 14) and incremental
production was reported to lie within the top 10% of all refracture treatments attempted to date, despite two previous
stimulation treatments on each well and a decade of production. Following this initial success, the remaining eight wells
were scheduled for treatment.
16 SPE 134330

3500 500 2000 180

Gas May 1999 Frac: Gas
450 1800 160
Water Water May 1999
3000
300,000 lb 20/40 1600 Initial Frac in Frac:
400 140
Initial Frac in LWC 1989:
1989: May 1995 Frac:

Water Rate, BWPD

2500 1400 Oct 1994 Frac: 258,000 lb

Water Rate, BWPD

350

Gas Rate, MCFD

120
Gas Rate, MCFD

304,000 lbs 20/40 LWC

48,000 lb 40/70 5,000 lb 100 mesh 1200 sand 12,000 lb 40/70 +
+ 24,000 lb 20/40 300
2000 sand + 466,000 82,000 lb 12/20 Sand 100
lb 12/20 sand Sand 1000
250
80
1500 800
200
60
600
1000 150
40
400
100
500 200 20
50
0 0
0 0 Jan-90 Jan-91 Jan-92 Jan-93 Jan-94 Jan-95 Jan-96 Jan-97 Jan-98 Jan-99 Jan-00 Jan-01
Jan-90 Jan-91 Jan-92 Jan-93 Jan-94 Jan-95 Jan-96 Jan-97 Jan-98 Jan-99 Jan-00 Jan-01

Figure 14 – In the San Juan Basin, Vastar stimulated two CBM wells with ceramic proppant to investigate the performance
of spherical, uniform proppant compared to sand. Well names are Southern Ute 12-2; 32-9 and Southern Ute 18-2; 32-8.
Figures adapted from Vincent (2002).

However, the property was sold in March, 2000 and the acquiring company had no interest in continuing a trial using
“strong” ceramic proppant in a shallow coal formation in which the closure stress was less than 3500 psi. The acquiring
company discontinued the ceramic trial. In 2006, the results of subsequent refracs by this operator in the San Juan basin were
summarized (JPT Online, 2006). From 1998 to 2006, 138 refracture treatments were performed, many incorporating surface
modification agents in an effort to mitigate migration of coal fines. In 2001, the average production improvement was 130
mscfd. In 2004, due to improved candidate selection and improved treatment design, the average production improvement
increased to 350 mscfd. Although two wells certainly don’t provide a statistically valid trial, the refracs designed to increase
fracture conductivity in Figure 14 provided 1400 and 800 mcfd increases, substantially higher than the reported average in
the area.

The desorption of methane from coal and subsequent movement through complex fracture networks is not well described in
most production models. To date, the author has not been able to reach a defendable conclusion whether production
impairment should be attributed to residual gel damage, fines plugging, fracture complexity, or hindered desorption due to
large pressure losses in fractures. However, most field data indicate that improved fracture conductivity appears to be more
beneficial in CBM reservoirs than generally acknowledged, and that refracs are frequently beneficial.

Field Example – Restimulation of Shallow Oil Wells with Higher Concentration of Larger Frac Sand
There are at least 50 examples summarized by Vincent (2009) in which productivity of oil and gas reservoirs shallower than
5000 ft depth have been improved with increased fracture conductivity. Therefore it appears that elevated closure stress on
proppant is not the sole cause of fracture insufficiency. Fleming (1992) reported the results from 32 fracture restimulation
attempts in the North Westbrook Unit in West Texas. The Middle Clearfork formation is an anhydritic dolomite at 3000 ft
depth. Initial fracs conducted between 1954 and 1960 placed up to 1 lb/gal sand using lease crude as the fracturing fluid.
Between 1960 and 1975, gelled water was used to place sand at concentrations up to 2 lb/gal. In 1976, crosslinked gelled
water was introduced with continual increases in proppant concentration through 1991. Based on observed advantages with
higher proppant concentrations, sand concentrations as high as 14 lb/gal were successfully implemented in the North
Westbrook Unit in 1991, with excellent success as shown in Figure 15.
SPE 134330 17

100 1000

of all wells refractured since Dec 1989

of Well NWU 7021
Oil Rate, bopd

Oil Rate, bopd

(15a)
10 100
(15b)

1 10
Jan- Jan- Jan- Jan- Jan- Jan- Jan- Jan- Jan- Jan- Jan- Jan- Jan- Jan-
86 87 88 89 90 91 92 86 87 88 89 90 91 92

Figure 15 – Refrac with 14 ppg sand in January, 1991 resulted in 20-fold increase in production for Well NWU 7021 (15a).
Refracs of 32 wells between Dec 1989 and early 1991 paid out in less than a year, generating over 200% rate of return on
investment (15b). Plots adapted from Fleming (1992).

Improved oil production was achieved with high concentrations of 12/20 Brady sand. Even greater production gains were
documented with 8/16 Brady sand, but operational challenges forced a return to 12/20 sized particles. As shown in Figure
15, oil production increased by 5-fold to 30-fold, with average improvement of 32 bopd/well. Restimulation treatments paid
out in less than a year, exceeding 200% rate of return on investment. Insufficient information on fracture geometry was
available to uniquely match observed production; however, the production gains after restimulation indicate that the pressure
losses in original fractures were much larger than would be predicted for simple, planar fracs with Darcy flow.

Field Example – Variety of Outcomes Restimulating Tight Gas Wells

Despite the numerous examples of increased production rate when superior proppant is utilized, it is important to recognize
that conductivity is not a silver bullet that overcomes all evils. Ennis (1989) provided a detailed account of seven tight gas
wells completed in the Morrow formation. Pressure transient testing on similar wells stimulated with a resin coated sand
suggested poor effective lengths and low fracture conductivities. As shown in Figure 16a, efforts to restimulate with
improved fracture conductivity provided production increases ranging between zero and 2700 mcfd, averaging 600 mcfd
incremental sustained production. Although the overall program was economic, a better frac did not make Well #1 more
productive. It was a “duster” following the initial frac, and better proppant did not create gas. Clearly, refrac success
depends on much more than fracture conductivity.
18 SPE 134330

Well Initial Initial Refrac Post Frac Benefit

Frac Rate Design Stabilized (mscfd) 2500
24,000 gal + 32,000 gal +
1 10,000 lb 0 50,000 lb 0 0

Stabilized Rate (MSCFD)

glass beads 20/40 ISP
2000
25,000 gal + 80,000 gal +
2 46,000 lb
20/40 sand
125 115,000 lb
20/40 ISP
400 275
60,000 of 1500
16,090 gal +
110,000 lb
3 10,941 lb
bauxite
300 20/40 ISP 3000 to 1100 800
(equip)
100,000 gal +
60,000 of 1000
120,000 lb
4 132,000 lb
sand
150 20/40 ISP 1100 to 600 450
(equip)
10,000 gal + 75,000 gal + 500
5 10,000 lb
glass beads
400 120,000 lb
20/40 ISP
2500 to 1500 1100

18,000 gal + 87,000 gal +

0
6 20,000 lb
20/40 RCS
500 98,000 lb
20/40 ISP
2400 to 2000 1500
Pre Frac 10,000 gal 80,000 gal + 75,000 gal +
80,000 gal + 3% acid + 100,000 lb 120,000 lb
40,000 gal +
110,000 lb
7 74,000 lb
sand
300 20/40 ISP 750 to 400 100 10,000 lb
glass beads
20/40 sand 20/40 ISP
(out of zone)

(16a) (16b)
Figure 16 – Seven Morrow wells in Oklahoma provided a range of outcomes when restimulated with increased fracture
conductivity (16a). Production from Well 5 after three stimulation treatments (16b). Data from Ennis (1989).

The history of Well #5 (Figure16b) provides an interesting case study in fracture optimization. The well was productive on
initial completion, which is a notable achievement in many reservoirs. A small acid job and fracture with 10,000 lbm of glass
beads increased the production to 600 mcfd. After some period of time, a much larger refrac using sand doubled well
production to 1300 mcfd. It was not stated how much additional time passed before a third stimulation treatment was
performed, utilizing a greater mass of stronger proppant in less total fluid. Production increased to an impressive 2400 mcfd.
Several questions may be asked. Which treatment was successful? Arguably, everyone who touched this well was
successful. The driller left the well productive – and each stimulation treatment increased production. Which treatment was
optimal? This is a difficult question, as no costs were provided, and it is unclear what production would have been achieved
were any step omitted. However, the preferred answer is that none of these treatments is possibly optimal. There is no
evidence that a 5th column could not be added to Figure16b, if a clever person found a better way to stimulate the well. With
the nearly unlimited number of treatment parameters, it is improbable that the ideal combination has been found. There is no
indication from the available data that superior designs are out of reach, or that previous efforts have even reached a point of
diminishing returns.

Field Example – Successful Restimulation of Underperforming Wells

As will be shown later, many field studies suggest that the most economic candidates for restimulation are the best wells, not
the poorest wells. However, that does not preclude successful restimulation of poorly producing wells. Crow (1977)
documented the results of three sequential stimulation attempts on a <0.5 mD oil well in the Salem Formation of Indiana. As
shown in Figure 17, the initial propped frac using gelled acid, and the subsequent treatment with HCl could not achieve
sustained production rates above 3 bopd longer than one month. However, Crow’s strategy to place high sand concentrations
(up to 14 ppg of 10-20 sand) to pack the fracture, combined with forced closure, strategic use of 100-mesh sand, and lower
injection rates provided initial production rates exceeding 750 bopd and sustaining 90 bopd four months later.
SPE 134330 19

3 months later
700
Well #1, Salem Formation, 30,000 gelled water
600 Gibson County, Indiana 20,000 lb 100 mesh
in pad
500 58,000 lb 20/40
5,000 lb 10/20
400
BOPD

Initial Frac: Acidize 8 months

later
8500 gal gelled
300 5% HCl 1500 gal 14% Initial rate exceeding
HCl 700 bopd, sustaining
31,500 lb Sand 90 bopd 120 days
200 8 bopd, declining
Declined from 10 later
to 3 bopd in 7
to 3 bopd within
100 days
30 days

0
Initial Frac Acidized Flush Rate after 120 days post
Refrac refrac

Figure 17 – Despite two unsuccessful attempts to stimulate oil production, a third treatment utilizing higher
concentrations of larger diameter proppant provided sustained productivity. Data from Crow (1977).

Analysis and Interpretation of the Field Examples:

Taken in isolation, it is impossible to “rank” the mechanisms in relative importance. Through extensive review, the author
has developed interpretations of the factors that influence success and failure in restimulation attempts:
− Touching the pay is Job #1. Clearly, many laminated reservoirs cannot be effectively drained unless the fracture
contacted all hydrocarbon-bearing intervals. In many reservoirs, the vertical permeability is less than 0.1% the
horizontal perm [kv/kh <0.001]. Refracturing bypassed pay is frequently a technical success; economic success will
depend on required wellwork to effectively identify, isolate, and retreat intervals.
− Achieved fracture conductivity is generally inadequate. While it is challenging to analyze restimulation treatments due
to multiple mechanisms contributing to production increases, there are other definitive field results to consider. In more
than 200 field studies, increased frac conductivity provided excellent improvement in well productivity [Vincent, 2009].
In the vast majority of fields with successful restimulation treatments, higher quality proppants and/or higher proppant
concentrations have been introduced in the refrac, which undoubtedly contributes to success.
− Unpropped and poorly propped portions of fractures should be expected to collapse. Almost all stimulation treatments
show greater sustained production when proppant is introduced into fractures, validating laboratory studies suggesting
that unpropped fissures should be expected to collapse over time. It is important to recognize that 1 ppa proppant
represents a very diffuse slurry – approximately 5% solid and 95% fluid. Clearly, in many slickwater treatments, the
vast majority of the created fracture extent cannot be effectively propped. Analyses of production from 41 Barnett
wells demonstrates that the fracture networks do not appear to provide high relative conductivity [Cipolla, 2009], and
significant improvements in production would be expected if fracture conductivity and continuity can be increased.
− Laboratory testing shows that all proppants degrade, but that ceramic proppants are much more durable than sand-based
proppants. It is clear that some portion of refrac success is due to restoration of conductivity to degraded proppant
packs.
− Horizontal wells stimulated with transverse fractures will always suffer from inadequate near-wellbore conductivity
[Shah, 2010]. Restimulation can frequently benefit production by introducing superior proppant near-wellbore, or by
increasing wellbore-to-fracture intersection if new perforations are added.
− Horizontal wells stimulated with longitudinal fractures would not be expected to require additional fracture conductivity
[Shah, 2010]. Restimulation success in longitudinal fractures appears to be related to inducing fractures in previously
unstimulated portions of the lateral.
− Significant reorientation has been documented in some reservoirs, but this phenomenon may occur in many more to a
less dramatic and perhaps undetectable extent. Reorientation is most likely to occur in reservoirs with modest initial
contrast in horizontal stresses. Stress reorientation may be induced by the creation of a nearby propped fracture or by
production/injection altering the pore pressures. Fracture mapping, observation of pressure trends while pumping, and
evaluation of reservoir pressure or GOR has indicated we can “touch new rock” with refracs in many reservoirs.
20 SPE 134330

Examples of Economic Failures in Attempted Refracs

Restimulation efforts may be unsuccessful for many different reasons, including:
• Poor reservoir quality
• Limited reserves in restricted drainage area
• Low reservoir pressure prohibiting cleanup
• Selection of undesirable candidates
• Poor mechanical integrity or undesirable perforations
• Formation damage induced by refracturing fluids
• Unknown reasons

Field Examples – Failed Refracs due to Very Low Reservoir Quality

Operators are understandably reluctant to touch the best wells and instead often focus first refrac efforts on underperforming
wells. However, as will be discussed later, this is the wrong approach for many reservoirs. A review of 32 refracs in the
tight “J sand” in the Wattenberg Field was reported by Parrot and Long (1979). In an area with permeability ~0.005 mD,
refracturing of old wells with larger proppant volumes increased productivity by six-fold, paying out treatment costs within
one year. However, in areas with lower permeability (<0.005 mD), insufficient production was obtained to pay out refracs
with gas prices below $1.50/mcf, and it was unclear that the wells would even recover the condensate used to formulate the
polyemulsion refrac fluid. It should also be noted that refrac technology and economics subsequently evolved and successful
restimulation throughout the Wattenberg are described in the Field Example – Conflicting Interpretations or Multiple
Mechanisms affecting Refrac Success section.
There are several reports of unsuccessful attempts to restimulate non-productive wells. As described above, Ennis (1989)
reported an unsuccessful refrac of a Morrow formation, with the well remaining non-productive post refrac. In New Zealand,
Green et al (2006) reported the results of three restimulation treatments in a water-sensitive formation. Although excellent
productivity gains were obtained with 12/18 LWC and diesel based fluids, attempts to restimulate a non-productive well
failed, and initial well productivity was highly correlated with the quality of candidate for both initial and restimulation
success. Despite evidence that some refracs contacted “new” reservoir rock, according to Green, “poor wells make poor
candidates for fracturing and also poor candidates for refracturing.”
Field Examples – Poor Refrac Results in Tight Gas
The Gas Research Institute surveyed operators in 25 different plays, and found operators perceived restimulation of tight gas
wells to carry high risk with marginal economic returns [GRI, 1996]. In general, operators were unenthusiastic about further
research into restimulation of gas wells. Pagano (2006) claimed that the industry’s success rate restimulating tight gas wells
was “less than 1 in 10” although the Codell formation responds “unconditionally” to restimulation. GRI later funded a
scoping study and subsidized restimulation research in several tight gas wells [GRI, 1996; Reeves et al, 1999]. Unfortunately,
identifying ideal candidates and optimized refrac designs proved elusive. Although one well in the Green River Basin was
improved by ~50% with restimulation, one was damaged by 50%, and a third well was improving, but after five months had
not returned to the production rate prior to the refrac. Additional evaluation of the candidates with a neural network
[Mohaghegh et al, 2000] was able to rank the successful refrac treatment above two unsuccessful candidates, but was not
structured to explain why two refrac attempts actually hindered production from two tight gas wells.
Shelley (1999) used a neural network to study refrac candidates in the Red Oak Field of Oklahoma. Historically, about 25%
of refracs in this field fell below economic limits, and about 20% exhibited significantly improved economic performance.
After training the neural network with the results from 17 recent refracs, the tool could match recompletion production of the
training dataset supplemented with eight additional refracs with an average error less than 5%. The 102 remaining wells
were evaluated to identify the best refrac candidates, with 30% of the remaining wells identified as good economic
candidates. In general, superior production was achieved with increased amounts of N2 foam, 16/30 proppant, and increased
treatment volumes. Poor candidates generally included older wells (lower pressured) on the periphery of the field, wells
initially treated with larger volumes of fluid, and wells with low production rates following the initial stimulation.

Field Example – Refrac Failure due to Limited Drainage Area or Gas in Place
Benedict and Miskimins (2009) reported the results from a reservoir simulation to investigate the effect of refrac reorientation
in lenticular reservoirs. Twenty-six scenarios with various combinations of fracture length, permeability anisotropy, and
reservoir extent were considered for an 18-ft thick interval with a reservoir permeability of 0.08 mD depicting a Piceance
Basin interval. If the only benefit of refracturing is to achieve reorientation, the 10-year cumulative production from wells in
a 40-acre drainage area can be increased by 15 to 22% by refracturing. However, production from wells draining a mere 10-
acre region could only be increased by 1 to 9% by creating a refrac perpendicular to the initial fracture. If a well is enclosed
in a small reservoir compartment of limited pore volume, the benefit of refrac reorientation alone is often insufficient to
justify restimulation.
SPE 134330 21

Field Examples – Failed Refracs due to Pressure Depletion and Correspondingly Poor Cleanup of Refrac Fluid
Restimulation of the Almond formation in the Wamsutter Field has resulted in “huge financial successes and several notable
failures” according to Cramer (1995). In the most productive portions of the field, the Almond is comprised of high-
permeability bar sands, which can provide excellent productivity, but deplete more thoroughly than lenticular, stratified,
lower quality stringers dominating much of the region. Therefore, the success of restimulation treatments in the bar sands is
highly dependent on the remaining reservoir pressure. Cramer reviewed nine recent restimulation treatments, noting that the
only failures were two wells that failed to energize the frac fluid with CO2 or N2, perhaps indicating the gaseous phase
reduced formation damage and/or improved cleanup and recovery of the water-based fracturing fluid.

Field Examples – Failed Refracs due to Candidate Selection

There are an interesting series of studies performed in the Vicksburg Formation. Holditch and Ely (1973) documented the
benefit of increased fracture conductivity in the Vicksburg, with EUR increases exceeding 1.7 BCF/well obtained with higher
sand concentrations. Tucker (1979) reported pressure transient analyses designed to identify which well should be
restimulated with bauxite proppant to counteract excessive crushing of frac sand. Unfortunately, the cost of the newly-
invented bauxite was 23-times greater than the previously used sand, so the refrac treatment size was reduced and welltesting
was conducted to confirm proppant durability. Following restimulation, extensive welltesting was performed to demonstrate
that fracture integrity was maintained with the stronger proppant and not with the original sand. This success was
instrumental in justifying the use of bauxite in a new Vicksburg well, with “superior flow” and “encouraging results” even
prior to the well being subjected to full closure stress. Despite this compelling evidence that initial fracs were insufficient,
refrac success has not been universal in the Vicksburg. Flores and Indriati (2009) documented that the economic success
frequency of refracs in the Vicksburg had declined to 45%, with finding costs increasing to $1.30/mcf with traditional
candidate identification and treatment design. Instead, Flores focused on high quality areas of the reservoir, and developed a
ranking system based on type of proppant initially utilized, physical condition of wellbore, shape of decline curve, and
historic water production. Refracs with a lightweight ceramic were designed for high potential areas with wells that met
specific criteria. Economic success was achieved in 80% of new refrac candidates, with finding costs reduced to $0.70 per
mcf of added reserves.

Field Examples – Failed Refracs due to Poor Mechanical Condition of Wellbore.

Restimulation treatments can be economic even when wells must be entirely redrilled or sidetracked due to undesirable initial
completions [Hower and Decker, 1992]. However, corroded tubulars, undesirable perforations, and lack of pressure integrity
can drastically increase workover expenses and can prohibit restimulation. There are numerous unpublished examples in
which wells failed to pass mechanical integrity tests, requiring more extensive wellwork and, in extreme cases, abandonment
of wells following pressure testing or mechanical failure during restimulation treatments.

Field Examples – Failed Refracs due to Formation Damage

In the Steinle Ranch oil field, Kohlhaas (1982) demonstrated that initial completions using ceramic proppants provided far
superior and more sustained production than wells stimulated with sand or glass beads, with a 2300% return on incremental
investment. This prompted restimulation attempts of four wells initially treated with sand or glass beads. However, only two
of the four refrac treatments resulted in an increase in recoverable reserves, and one refrac actually damaged the well. The
author noted that these wells were particularly sensitive to formation damage from the fracturing fluids. Despite the failed
restimulation attempts, the two successful refracs added 62,000 bbls of additional reserves, providing a 210% return on
investment for the entire restimulation program.

Field Examples – Refrac Surveys Indicating Refrac Failures

In 1995, Spears and Associates compiled a market survey of restimulation activity and success [GRI, 1996]. At that time, the
industry appeared to be unenthusiastic about the potential for restimulation of tight gas wells, and the refrac market was
estimated to be between 2 and 3% of all stimulation activity, or 450-550 refracs annually. Twelve years later, a smaller
restimulation survey primarily targeting operators and consultants in the Rocky Mountains summarized the success rate and
candidate selection criteria [Craig, 2010]. The 18 respondents had completed at least 240 refracs in 2006. Refrac success
was variable, with approximately 1/3 of the respondents reporting production increases exceeding 100%, with another 1/3
reporting that wells failed to respond to restimulation. When asked about candidate selection criteria, fully 61% of the
respondents said that well production trends were a primary criterion, with “intuition” the next most frequent technique at
39%. Fewer than 25% of respondents utilized fracture diagnostics or pressure transient testing and no respondent used neural
networks to identify refrac candidates. The problems that were identified for remediation included “short effective fracture
lengths” (56%), “bypassed pay” (56%), “loss of conductivity” (22%), and less than 20% “other,” “phase trapping,” or “don’t
know.” The most common barriers to a successful restimulation program were identified as “time required for candidate
selection,” “diagnostic methods,” and “risk.” Clearly, if superior candidate selection techniques or superior refrac design
could eliminate the failed refrac attempts, significant economic opportunities for restimulation are available.
22 SPE 134330

Comments on Candidate Selection Procedures

Identifying the most lucrative candidates for restimulation is a challenging task. It is highly probable that a single procedure
or flowchart will fail to yield the most promising candidates in all fields, as the level of available data and beneficial
restimulation mechanisms are likely unique in many fields. For example, in the Kuparuk oilfield, extensive pressure and
production data are available, with multiple pressure transient tests conducted on most producing wells. Restimulation has
been highly effective and has reliably increased production rates when the retreatment was successfully designed and
implemented to increase fracture conductivity. In contrast, multiple wells in the DJ Basin flow into a common manifold and
individual well rates are mathematically allocated in lieu of measurement. Few production logs have been conducted and
actual production from the Codell, J-Sand, and Niobrara intervals are typically unknown. The mechanisms for benefit from
restimulation in the DJ Basin are in debate, and many operators restimulate specific layers or all intervals with little feedback
as to the efficacy of individual restimulation treatments, merely relying on statistics indicating that the overall restimulation
program is cost-effective.
Although the author firmly believes that candidate selection procedures should be customized to the particular field, there are
several general comments that may be of benefit.

Candidate Selection based solely on Unsatisfactory Production

Operators are understandably reluctant to conduct wellwork on the best producers in a field and have often focused
restimulation efforts on wells that underperform compared to offsets or expectations. This strategy has been successful in
some fields, such as Cooperstown in Pennsylvania [MacDonald et al, 2005]. Seventeen wells were identified which
exhibited at least 50% poorer productivity than offset wells. Restimulation of the underperforming wells typically resulted in
approximately 80 MMCF incremental reserves per well with net present values of ~$80,000 after recovering refrac
investments.

However, many operators have provided contrary reports, indicating that restimulation of the lowest performing wells will
frequently provide the least incremental recovery. Random selection of candidates has been shown to provide superior
results to systematic selection of underperforming wells in some fields. Numerous authors have concluded that the best
wells will often make the most economic refrac candidates, including Parrot, 1979; Hunter, 1986, Branch, 1991; Reese,
1994; GRI, 1996; Shelley, 1999; Reeves, 1999&2000; Ely, 2000; Husen, 2003; Green, 2006; and Flores, 2009. The fact that
the most prolific wells can usually be improved with a refrac suggests that initial treatments are rarely sufficient or optimized.
If the best wells can be improved, it is probable that lower performing wells can also benefit from a restimulation treatment,
but would require a greater percent improvement to achieve similar economic returns.

Candidate Selection based solely on Production Trends

Crowell and Jennings (1978) recognized that restimulating wells solely based on unsatisfactory performance often resulted in
marginal success. Instead, detailed pressure transient testing was suggested to compare “effective” fracture length and
conductivity to values anticipated from propagation modeling. Reeves (2000) similarly concluded that simple production
data comparisons are not effective at identifying high-potential restimulation candidates in many fields, and recommended
virtual intelligence techniques.

Candidate Selection with Artificial Neural Networks

In some fields, virtual intelligence or neural networks appear to improve candidate identification. Reeves (2000) was able to
train virtual intelligence systems incorporating neural networks, genetic algorithms, and fuzzy logic to select candidates that
agreed within 80% of those selected with more detailed reservoir modeling. In one of the most successful field examples,
Shelley (1999) demonstrated in the Red Oak Field that neural networks could be trained to match refrac production with
average errors less than 5%. In the DJ Basin, there was less success identifying the parameters associated with successful
stimulations [Mohaghegh, 2002] and significant effort was invested to identify and cull approximately 50% of records
considered to be contaminated or unreliable [Popa, 2003]. The operator eventually adopted a statistical algorithm based on
hydrocarbon pore volume, cumulative recovery percentage, initial perforating strategy, EUR, GOR, and relative EUR
compared to best offset. Using these criteria, one operator claims only 3% of DJ refracturing treatments are economic
failures, and primarily associated with undesirable communication with overlying formations or with offset wells [Dozier,
2003]. Although automated candidate selection routines have shown promise in some specific studies, a more universal
procedure has largely eluded researchers, likely indicating that beneficial restimulation mechanisms are unique in different
fields. In most fields, there are insufficient numbers of historical restimulation attempts to train neural networks, and
application of rules derived from seemingly analogous fields should be made only with careful engineering judgment.
SPE 134330 23

Candidate Selection with Well Testing

A primary criterion for candidate selection should be the remaining reservoir pressure, as this relates directly to the
recoverable reserves and effectiveness of initial treatment at depleting the reservoir. Although numerous studies indicate the
potential to “frac into” virgin reservoir pressure during a restimulation treatment, evidence of remaining reservoir pressure
prior to the refrac increases the likelihood of economic success. In addition to determining reservoir pressure, well testing
may identify the presence or absence of a conductive fracture of adequate effective length from the initial completion, as
previously discussed. To identify bypassed pay in multilayer completions, Craig (2006) developed pressure transient
solutions to evaluate refracture candidates. This involved short injections above frac gradient into prospective layers,
analyzing both pre- and post-closure behavior. Variable storage behavior can identify pre-existing fractures and characterize
damage to the initial fracture.

Diagnostic Techniques to Evaluate Initial Treatment prior to Refrac

Table 1 summarizes some common techniques used to characterize candidates for restimulation.

Table 1
Mechanism to Diagnostic Technique Comments
Investigate
Video Camera Evidence of perf erosion
Temperature log Run immediately post-frac, can indicate where
cooler frac fluid entered

Infer where proppant exited Production log (spinner survey) Cannot distinguish between poor frac and poor
perforations reservoir quality

Radioactive tracer ProTechnics –half-life of isotope+env restrictions

Proppant with induced radioactive signature Hexion – reduced logging speed
Non-radioactive detectable proppant CARBO Ceramics – neutron log
Predictive modeling Propagation models –uncertain stress boundary
data

Fracture height Measurement Temperature data describes where fluid entered,

not proof of proppant entry
Detectable proppant yields an estimate of
minimum fracture height
PTA (drawdown, buildup – pressure transient) Non-unique estimate of effective fracture
conductivity and length

PDA (analyze flowing history of well) Can indicate performance over time
Fracture effectiveness
Shape of decline curve Flat production is not signature of highly
stimulated well

Diagnostic injections May be able to identify restim candidates

Determination of remaining PTA, PDA, Long term shut in, material balance with
reservoir pressure known/assumed HCPV
Laboratory testing, periodic welltesting, field
Fracture degradation
experimentation with remedial treatments
Core studies, field experimentation with different
Fluid sensitivity
completion and fracturing fluids
Production forecasting using realistic conductivity
Improved conductivity estimates. Field validation with different proppant
quality and concentration
Determination of principle stresses (borehole
ellipticity, breakouts, anelastic strain recovery,
velocity anisotropy, dipole sonic, formation imaging,
DSCA, AE, DRA, PSHA)
Stress reorientation
Mathematical prediction of reorientation
If significant reservoir pressure remains, this
establishes the minimum reserves, even if the
refrac does not contact new rock
Recognize that required restimulation frequency
may depend on proppant quality and
concentration (frac width)
Although many cores appear to be very sensitive
to frac fluids, field data often suggests formation
damage may be tolerated in exchange for highly
conductive fractures. Frac complexity generally
increases with reduced fluid viscosity, with
jeopardized proppant transport
Models generally underestimate the importance
of conductivity due to complex flow regimes and
imperfectly planar fracs
When stress contrast is small, some degree of
reorientation and/or diversion is likely due to
pore pressure alteration or creation of propped
frac widths.

Tiltmeter or microseismic mapping More definitive documentation of far-field

frac/refrac geometry
24 SPE 134330

Discussion of Candidate Selection Criteria:

− Well Productivity: This can be a good indicator of success; highly productive wells are frequently the most economic
restimulation candidates
− Pressure Transient Analyses: Wells with high skin and high remaining pressure are excellent candidates for
restimulation.
− Cumulative production to date: This is often a poor indicator of success, as the best wells are typically superior refrac
candidates. However, high cumulative production conversely correlates to depleted zones.
− Well age: Age is a poor criterion presuming mechanical wellbore condition is acceptable. In a shallow gas field in
Alberta, refracs often exceeded initial production rates, despite significant gas depletion during 30 years of
production [Gutor, 2003]. If mechanical integrity of well has been jeopardized with age, refracs are economically
hindered.
− Reservoir pressure: When significant reservoir pressure can be verified, this is a strong indicator of success.
However, absence of ability to document good reservoir pressure is not a reliable predictor of failure, as it appears
restimulation treatments can “frac into” higher pressured layers or regions in many reservoirs.
− Poor conductivity in initial frac: Low conductivity in initial treatments is a strong indicator of success when
restimulating with high conductivity refracs.
− Low proppant concentrations placed in crosslinked fluids: This is a strong indicator of success when refracturing
with increased proppant concentrations, as better vertical and lateral continuity within proppant pack can be
achieved.
− Frac sand or resin coated sand used in high stress or high temperature applications: This is a strong indicator of
success when restimulated with more durable ceramic proppants.
− Shape of decline curve: A great deal of information can be extracted from analyses of historical production;
however, a thorough description is beyond the scope of this paper. In general, long term flat production is
frequently indicative of an understimulated well. Steep production declines can be attributed to numerous
phenomena ranging from proppant degradation to limited drainage area. Interpretations are generally non-unique
without additional data or engineering judgment.
− Existing perforated intervals: Incorrect perforation design is an excellent indication of refrac success. When
insufficient perforations were placed to provide reasonable connection to productive intervals, refracs are frequently
merited following reperforation. Identifying and stimulating bypassed pay is frequently a technical success. When
excessive perforations were placed such that sufficient injection into all perforations was not assured, refracs are
frequently merited; however, more expensive workover or isolation procedures may restrict the economic viability
of restimulation attempts.
− Number of previous stimulation treatments: The number of previous treatments is a poor indication of refrac success.
Some wells have been successfully refractured as many as five times and remain economic candidates for future
restimulation.

Conclusions
1. Restimulation has been very successful in a wide variety of reservoir types and conditions – demonstrating that
unrecognized potential exists in many reservoirs.
2. Refracs may be justified due to:
o Inadequate initial fracture design
o Flawed execution of initial treatment
o Improved fracturing technology and materials
o Improved reservoir knowledge
o Increase in hydrocarbon price, justifying different exploitation approaches
o Changes in reservoir stress providing new opportunities (superior containment in partially depressured
intervals or refrac reorientation due to pore pressure changes or induced stress)
3. Although there are many exceptions in the literature, poorer candidates for restimulation include:
o Wells with questionable mechanical integrity (requiring expensive workovers)
o Fringe type wells with low or inadequate reservoir quality
o Wells that are non-productive following a correctly implemented initial fracture treatment
o Wells with thorough depletion in all layers - minimal remaining reserves
4. In most cases, conventional models fail to accurately predict the production rates achieved with fracture
restimulation, indicating that traditional modeling fails to properly describe reservoir/hydraulic fracture dynamics.
5. There are many documented examples in which restimulation treatments have achieved production rates higher than
initial production from the original completion, despite significant pressure depletion.
6. Significant engineering effort is required to identify superior refrac candidates and design optimal retreatments. For
this reason, many operators overlook refrac opportunities in lieu of more routine drilling and development.
SPE 134330 25

7. A fresh perspective and new approach may achieve restimulation success when previous attempts have failed.
8. In many reservoirs, refracs can more cost-effectively add reserves compared to incremental drilling, with fewer
environmental impacts as the surface footprint of operations remains unchanged.
9. The best wells frequently make the best restimulation candidates.
10. Significant opportunities are available to increase production from most reservoirs. In more than 200 field studies
[Vincent, 2009], large increases in production could be obtained by redesign of the fracture treatments. It is highly
improbable that we’ve stumbled upon the optimal frac design for any reservoir, and therefore, production can be
improved by restimulating most wells. However, careful economic analyses are required in many cases to identify
which treatments are cost effective and how to optimize the economic potential.

Nomenclature
bbl barrel
BCF billion cubic feet
boe, BOE barrels of oil equivalent
bopd barrels of oil per day
bwpd barrels of water per day
BPM barrel per month
CBM coalbed methane
CMG carboxymethylate guar
CO2 carbon dioxide
ELWC economy light weight ceramic, ASG ~ 2.7
EUR estimated ultimate recovery (recoverable reserves)
ft feet
GOR gas-oil ratio
HCl hydrochloric (acid)
HPG hydroxylpropyl guar
IDC intermediate density ceramic, aka ISP, ASG ~3.3
ISB, HSB intermediate/high strength bauxite – unusual abbreviations used in Fig 7
lbs, lbm pounds, mass
lb/sq ft pounds of proppant per square foot of fracture area
LWC lightweight ceramic, ASG ~2.7
KCl potassium chloride, a salt
mcfd thousand standard cubic feet per day
MMCF million cubic feet
mD, md milliDarcies, a measure of permeability
mD-ft milliDarcy-feet, a measure of conductivity
N2 nitrogen
PBU pressure build up (a pressure transient test)
ppa pounds of proppant added per gallon of clean fracturing fluid
ppg pounds of proppant in one gallon of slurry, also lb/gal
PSS pseudo steady state
RCS resin coated sand
ROR rate of return, an economic measure of return on investment
RPI reciprocal productivity index, a form of production data analyses
scf/bbl standard cubic feet per barrel of liquid

Acknowledgements
The author would like to acknowledge the extensive support of StrataGen Engineering and CARBO Ceramics for supporting
this compilation. In particular, Cindy Cortez has spent numerous hours re-reading nearly 150 papers and compiling an
extensive database of reservoir properties, completion details, and production results that are used when designing
restimulation treatments. Additionally, the author recognizes the work of several hundred authors who have shared their
restimulation results with the industry. Summarizing these critical field results in the Appendix is in no way meant to demote
the importance of their important contributions, but was necessary to keep the body of the paper to a somewhat manageable
length.
26 SPE 134330

References noted in paper text (additional references in Appendix)

Asadi, M. et al, 2002: Monitoring Fracturing Fluid Flowback with Chemical Tracers: A Field Case Study. SPE 77750 presented at the
2002 Annual Technical Conference, San Antonio TX, Sept 29-Oct 2.
Bagzis, J.M., 1989: Refracturing pays off in the Rangely field. World Oil Nov 1989 39-44.
Barba, R.E. 2009: A Novel Approach to Identifying Refracturing Candidates and Executing Refracture Treatments in Multiple Zone
Reservoirs. SPE 125008 presented at the SPE Annual Technical Meeting, New Orleans, Oct 4-7.
Benedict, D.S. and Miskimins, J.L. 2009: Analysis of Reserve Recovery Potential from Hydraulic Fracture Reorientation in Tight Gas
Lenticular Reservoirs. SPE 119355 presented at the 2009 Hydraulic Fracturing Technology Conference, The Woodlands, TX, Jan 19-21.
Besler, M.R. et al, 2007: Improving Well Productivity and Profitability in the Bakken—A Summary of Our Experiences Drilling,
Stimulating, and Operating Horizontal Wells. Paper SPE 110679 presented at the 2007 Annual Technical Conference, Anaheim, Nov.
11-14.
Besler, M.R. 2008: Re-fracturing: Is Too Little Prevention Requiring the Cure? Presented at the 2008 Williston Basin Petroleum
Conference, Minot ND, April 27-29.
Branch, G.A. and Drennan, K.M. 1991: Refracture Stimulations in the Norge Marchand Unit: A Case Study. SPE paper 21642 presented at
the Production Operations Symposium, Oklahoma City, April 7-9.
Britt, L.K., et al 2006: Water-Fracs: We Do Need Proppant After All. SPE paper 102227 presented at the 2006 Annual Technical
Conference, San Antonio, Sept 24-27.
Brown, K.G. et al 2003: Assessment of Remediation Treatments in Underground Gas Storage Wells. SPE paper 84393 presented at the
Annual Technical Conference, Denver, Oct 5-8.
Cipolla, C.L. 2005: The Truth about Hydraulic Fracturing – It’s More Complicated than We Would Like to Admit. Distinguished Lecture
2005-2006, SPE 108817.
Cipolla, C.L. et al. 2008: The Relationship between Fracture Complexity, Reservoir Properties, and Fracture Treatment Design. SPE paper
115769 presented at the SPE Annual Technical Conference, Denver, Sep. 21-24.
Cipolla, C.L., Lolon, E., and Dzubin, B., 2009: Evaluating Stimulation Effectiveness in Unconventional Gas Reservoirs. SPE paper
124843 presented at the 2009 Annual Technical Conference, New Orleans, Oct 4-7.
Clawson, T. 2003: Production Enhancement in a Low Pressure CBM Reservoir – Apache Canyon Area. Presented at the IPAMS
Conference, Denver, April 16, 2003.
Cobb, S.L. and Farrell, J.J. 1986: Evaluation of Long-Term Proppant Stability. SPE paper 14133 presented at the International Meeting on
Petroleum Engineering, Beijing, Mar 17-20.
Conway, M.W. et al 1985: Expanding Recoverable Reserves Through Refracturing. SPE paper 14376 presented at the Annual Technical
Conference, Las Vegas, NV, Sept. 22-25.
Coulter, G.R. and Menzie, D.E. 1973: The Design of Refrac Treatments for Restimulation of Subsurface Formations. SPE paper 4400
presented at the SPE/AIME Rocky Mountain Regional Meeting, Casper, Wy, May 15-16.
Coulter, G. R. et al, 2004: Water Fracs and Sand Quantity: A Barnett Shale Example. SPE paper 90891 presented at the 2004 Annual
Technical Conference, Houston, Sept. 26-29.
Craig, D.P. 2006: Analytical Modeling of a Fracture-Injection/Falloff Sequence and the Development of a Refracture Candidate Diagnostic
Test. Dissertation, Texas A&M, May 2006.
Craig, D.P. and Odegard, C.E. 2008: Identifying Bypassed and Ineffectively Stimulated Layers in a Well with Commingled Production
from Multiple Layers: Mesaverde Case History. SPE 114777 presented at the Unconventional Reservoirs Conference, Keystone, CO,
Feb 10-12.
Craig, D.P. 2010: Shale Pressure Diagnostics: Pre-frac Reservoir Evaluation Post-frac Stimulation Effectiveness. Presentation at StrataGen
Shale Completion Strategies Workshop, April 22, 2010. Discussion at refrac.net/refracnetblog/
Cramer, D.D. 1995: The Evolution of Hydraulic Fracturing in the Almond Formation. Paper SPE 30480 presented at the Annual Technical
Conference, Dallas, Oct 22-25.
Cramer, D.D. 2008: Stimulating Unconventional Reservoirs: Lessons Learned, Successful Practices, Areas for Improvement. SPE paper
114172 presented at the Unconventional Reservoirs Conference, Keystone, CO Feb 10-12.
Crow, W.L., 1977: Mini Massive Frac – A Stimulation Technique for Low-Porosity, Low-Permeability Zones. SPE paper 6627 presented
at the 1977 AIME Eastern Regional Meeting, Pittsburgh, Oct 27-28.
Crowell, R.F. and Jennings, A.R. Jr. 1978: A Diagnostic Technique for Re-stimulation Candidate Selection. SPE paper 7556 presented at
the Annual Fall Conference, Houston, Oct 1-3.
Dozier, G. et al. 2003: Refracturing Works. Oilfield Review Autumn 2003, p 38-53.
Dunek, K.L. et al, 2009: Far-Field Volumetric Distribution of Fracturing Fluids Away from an Uncemented Horizontal Liner in the Bakken
Formation. SPE paper 115826 presented at the Rocky Mountain Technology Conference, Denver, April 14-16.
Durrani, A.J et al, 1994: The Rejuvenation of 30-Year Old McAllen Ranch Field – An Application of Cross-Functional Team
Management. SPE paper 24872, JPT, Dec 1994.
Eberhard, M. and Mullen, M. 2003:The Effect of Completion Methodologies on Production in the Jonah Field. SPE 84959, SPEPF (August
2003) 145.
Eberhard, M. 2008: Review of Current Bakken Practices, MT and ND. Presented at the 2008 Williston Basin Petroleum Conference,
Minot ND, April 27-29.
Elbel, J.L. and Mack, M.G. 1993: Refracturing: Observations and Theories. SPE 25464 presented at the Production Operations
Symposium, Oklahoma City, OK, Mar 21-23.
Ely, J.W. et al, 2000: Restimulation Program Finds Success in Enhancing Recoverable Reserves. SPE 63241 presented at the 2000 Annual
Technical Conference, Dallas, Oct 1-4.
SPE 134330 27

Emrich, C. et al, 2001: Codell Restimulations Evolve to 200% Rate of Return. SPE 67211 presented at the 2001 Production Operations
Symposium, Oklahoma City, Mar 24-27.
Ennis, B.K. 1989: Case History of Restimulations in Western Oklahoma. SPE 18861 presented at the 1989 SPE Production Operations
Symposium, Oklahoma City, Mar. 13-14.
Esphahanian, C. and Storhaug, D.G 1997: A Statistical Approach to Pay Identification in the Tight, Fractured, Heterogeneous Reservoirs of
the Piceance Basin, paper SPE 38366 presented at the 1997 Rocky Mountain Regional Meeting, Casper, Wyoming, 18-21 May.
Fetkovich, M.J. et al. 1990: Depletion Performance of Layered Reservoirs Without Crossflow. SPEFE, Sept, 1990, 311-318.
Fisher, K., Robinson, B.M., and Voneiff, G.W., 1995: A Comprehensive Study of the Analysis and Economic Benefit of Radioactive
Tracer Engineered Stimulation Procedures. SPE paper 30794 presented at the 1995 SPE Annual Technical Conference and Exhibition,
Dallas, Texas, 22-25 October.
Flores, D.Q. and Indriati, S. 2009: Finding Additional Resources in a Mature Field: Refracturing Engineering Workflow for a Successful
Campaign in Vicksburg Basin, South Texas. SPE abstract 125260 and presentation at 2009 Tight Gas Completions Conference, San
Antonio, TX June 15-17.
Fleming, M.E. et al. 1992: Successful Refracturing in the North Westbrook Unit. Paper SPE 24011 presented at the 1992 Permian Basin
Oil and Gas Recovery Conference, Midland, March 18-20.
Garland, T.M et al. 1957: Effects of Hydraulic Fracturing Upon Oil Recovery from the Strawn and Cisco Formations in North Texas.
USBM, RI 5371.
Gil, I.R. et al 2007: “Super Height Growth” – Strange Surface Pressures and Refrac Potential. SPE 106083 presented at the Hydraulic
Fracturing Technology Conference, College Station, TX, Jan 29-31.
Ghauri, W.K. 1980: Results of Well Stimulation by Hydraulic Fracturing and High Rate Oil Backflushing. SPE 1382 JPT June 1980.
Gottschling, J.C. et al 1985: Nitrogen Gas and Sand: A New Technique for Stimulation of Devonian Shale. SPE paper 12313, JPT May
1985.
Green, D., et al 2006: Hydraulic Fracturing of Miocene and Oligocene Sandstones in the Taranaki Basin, New Zealand. SPE 101121
presented at the Asia Pacific Oil & Gas Conference, Adelaide, AU, Sept 11-13.
GRI Gas Research Institute, 1996: Assessment of Technological Barriers and Potential Benefits of Restimulation R&D for Natural Gas
Wells, Final Report. GRI 96/0267, July 1996.
Gutor, C. et al 2003: New Life for Old Wells – A Case study of the Effects of Re-Stimulating Gas Wells using Fracturing Through Coiled
Tubing and Snubbing Techniques. SPE 81730 presented at the SPE/ICoTA Coiled Tubing Conference, Houston, Apr 8-9.
Hahn, G. 1986: How Long will it Prop? Drilling, the Wellsite Publication. Vol. 47, No. 6, Issue 596, April 1986.
Handren, P. and Palisch, T. 2007: Successful Hybrid Slickwater Fracture Design Evolution. Paper 110451 presented at the 2007 Annual
Technical Conference, Anaheim, Nov 11-14.
Harper, T. R., et al 1985: Fracturing Without Proppant. SPE paper 13858 presented at the 1985 Low Permeability Gas Reservoirs
Symposium, Denver, March 19-22.
Hecker, M.T., Houston, M.E, and Dumas, J.D. 1995: Improved Completion Designs in the Hugoton Field Utilizing Multiple Gamma
Emitting Tracers. SPE paper 30651 presented at the 1995 Annual Technical Conference, Dallas, 22-25 Oct.
Hejl, K.A. 1992: High-Rate Refracturing: Optimization and Performance in a CO2 Flood. SPE paper 24346 presented at the 1992 Rocky
Mountain Regional Meeting, Casper, May 18-21.
Holditch, S.A. and Ely J, 1972: Successful Deep Well Stimulation Utilizing High Proppant Concentration. SPE paper 4118 presented at
the 1972 SPE Annual Fall Meeting, San Antonio, October 8-11.
Howard, G.C. and Fast, C.R. 1970: Monograph 2 – Hydraulic Fracturing. AIME, Millet the Printer, Dallas, TX (1970).
Hower, T.L. and Decker, M.K. 1992: Identifying Recompletion Candidates in Stratified Gas Reservoirs. SPE 24307 presented at the Mid-
Continent Gas Symposium, Amarillo, TX, April 13-14.
Huckabee, P. et al 2005: Field Results: Effect of Proppant Strength and Sieve Distribution Upon Well Productivity. SPE paper 96559
presented at the SPE Annual Technical Conference, Dallas, Oct 9-12.
Hunter, J.C. 1986: A Case History of Refracs in the Oak Hill (Cotton Valley) Field. SPE paper 14655 presented at the East Texas Regional
Meeting, Tyler, TX, April 21-22.
Husen, A.A. et al. 2003: Hydraulic Fracturing the Best Producer - A Myth? SPE 81543 presented at the 2003 SPE Middle East Show and
Conference, Bahrain, April 5-8.
Jeu, S.J. et al 1988: Exploitation of Deeply Buried Coalbed Methane Using Different Hydraulic Fracturing Techniques in the Piceance
Basin, Colorado, and San Juan Basin, New Mexico. SPE paper 18253 presented at the 63rd Annual Technical Conference, Houston, Oct
2-5.
JPT Online, March 2006: Techbits: Refracturing in Low-Permeability Reservoirs. SPE editorial summary of presentation by Devin Brown
entitled “Refracturing in the Fruitland Coal” presented at ATW, San Antonio.
Kohlhaas, C.A. 1982: Field Study – Steinle Ranch, an Intermediate Depth Oil Field, Shows Significant Benefit From Bauxite Proppants.
SPE 10875 presented at the 1982 Rocky Mountain Regional Meeting, Billings, MT May 19-21.
Lambert, S.W., Trevits, M.A. 1980: The Feasibility of No-proppant Stimulation to Enhance Removal of Methane from the Mary Lee
Coalbed. US DOE Report of Investigation, April, 1980.
Lantz, T, et al, 2007: Refracture Treatments Proving Successful in Horizontal Bakken Wells: Richland County, Montana. SPE paper
108117 presented at the 2007 Rocky Mountain Oil and Gas Symposium, Denver April 16-18.
Lehman, L.V. et al. 1998: Desorption Enhancement in Fracture-Stimulated Coalbed Methane Wells. SPE paper 51063 presented at the
2998 Eastern Regional Conference, Pittsburgh, Nov 9-11.
Leonard, D. 2009: Refracturing of Barnett. Presentation at PNR Energy Forum, Brookhaven College, May 28, 2009.
Leshchyshyn, T. et al. 1999: Technical and Economic Rationalization of Oil Refracture Programs on Wells Previously Stimulated with
Water-Based Fracture Fluids. Petroleum Society paper 99-60 presented at the CSPG/Petroleum Society Joint Convention, Calgary,
Alberta, June 14-18.
28 SPE 134330

Li, P., Song, Z. and Wu, Z. 2006: Study on Reorientation Mechanism of Refracturing in Ordos Basin - A Case Study: Chang 6 Formation,
Yanchang Group, Triassic System in Wangyao Section of Ansai Oil Field. SPE paper 104260 presented at the 2006 International Oil &
Gas Conference , Beijing, December 5-7.
Liu, H., et al, 2008: Evaluation of Refracture Reorientation in Both Laboratory and Field Scales. SPE 112445 presented at the 2008
International Symposium on Formation Damage Control, Lafayette, February 13-15.
MacDonald, R.J. et al. 2002: Comparing Production Responses from Devonian Shale and Berea Wells Stimulated with Nitrogen Foam and
Proppant vs. Nitrogen Only, Pike Co., Kentucky. SPE paper 77464 presented at the 2002 Annual Technical Conference, Sep 29 – Oct 2.
MacDonald, R.J. et al. 2005: A Rapid and Efficient Method to Identify Underperforming Stripper Gas and Oil Wells. SPE 95511 presented
at the Annual Technical Conference, Dallas, Oct 9-12.
Marsh, J. et al. 2000: Economic Fracturing of Bypassed Pay: A Direct Comparison of Conventional and Coiled Tubing Placement
Techniques. SPE 60313 presented at the Rocky Mountain Regional/Low Perm Reservoirs Symposium, Denver, Mar 12-15.
McCabe, M.A. et al. 1999: Investigation of a New Fracturing Fluid and Conductivity Enhancement Technology on Coalbed Methane
Production. SPE paper 52193 presented at the 1999 Mid-Continent Operations Symposium, Oklahoma City, Mar 28-31.
McDaniel, B.W. 1986: Conductivity Testing of Proppants at High Temperature and Stress. SPE Paper 15067 presented at the 56th
California Regional Meeting, Oakland, April 2-4.
McDaniel, B.W. 1990: Hydraulic Fracturing Techniques Used for Stimulation of Coalbed Methane Wells. SPE paper 21292 presented at
the 1990 Eastern Regional Meeting, Columbus OH, Oct 31-Nov 2.
Miller, J.D. et al. 2004: Fracturing Fluid Viscosity Profiles Determine Well Productivity in a Remedially Stimulated Tight Gas Reservoir.
SPE paper 90194 presented at the Annual Technical Conference, Houston, Sep. 26-29.
Mohaghegh, S. et al. 2002: Identification of Successful Practices in Hydraulic Fracturing Using Intelligent Data Mining Tool; Application
to the Codell Formation in the DJ-Basin. SPE 77597 presented at the Annual Technical Conference, San Antonio, TX, Sep 29-Oct 2.
Montgomery, C.T. and Steanson, R.E. 1984: Proppant Selection – The Key to Successful Fracture Stimulation. SPE paper 12616 presented
at the Deep Drilling and Production Symposium, Amarillo, TX April 1-3.
Moschovidis, Z. et al, 2000: The Mounds Drill-Cuttings Injection Field Experiment: Final Results and Conclusions. SPE paper 59115
presented at the IADC/SPE Drilling Conference, New Orleans, Feb 23-35.
Pagano, T.A. 2006: Rock Properties, Deliverability Mechanism influence Codell Restimulation. OGJ July 10, 2006. 44-47.
Palisch, T. et al. 2008: Slickwater Fracturing – Food for Thought. SPE paper 115766 presented at the Annual Technical Conference,
Denver, Sept 21-24.
Palmer, I.D. 1992: Review of Coalbed Methane Well Stimulation. SPE 22395 presented at the 1992 International Meeting on Petroleum
Engineering, Beijing, China, March 24-27.
Palmer, I.D. 1993: Induced Stresses Due to Propped Hydraulic Fracture in Coalbed Methane Wells. SPE 25861 presented at the Rocky
Mountain Regional Low Perm Symposium, Denver, Apr 12-14.
Parrot, D.I. and Long, M.G. 1979: A Case History of Massive Hydraulic Refracturing in the Tight Muddy “J” Formation. SPE paper 7936
presented at the Symposium on Low-Permeability Gas Reservoirs, Denver, May 20-22.
Pauls, R.W. et al 1985: Successful Stimulation of the Olmos Formation Using Oil-Base Fluids and High-Proppant Concentrations. SPE
13817 presented at the Production Operations Symposium, Oklahoma City, Mar 10-12.
Popa, A.S. et al 2003: Identification of Contaminated Data in Hydraulic Fracturing Databases: Application to the Codell Formation in the
DJ Basin. SPE 83446 presented at the SPE Western Regional/AAPG Pacific Section Meeting, Long Beach, CA, May 19-24.
Pospisil, G. et al 1992: Results of a Large-Scale Refracture Stimulation Program, Kuparuk River Unit, Alaska. SPE paper 24857 presented
at the 1992 SPE Annual Technical Conference and Exhibition, Washington, D.C., Oct. 4-7.
Reese, J.L. et al 1994: Selecting Economic Refracturing Candidates. SPE paper 28490 presented at the Annual Technical Conference, New
Orleans, Sept 25-28.
Reeves, S.R. et al 1999: Restimulation of Tight Gas Sand Wells in the Rocky Mountain Region. SPE paper 55627 presented at the Rocky
Mountain Regional Meeting, Gillette, WY, May 15-18.
Reeves, S.R. et al 2000: Benchmarking of Restimulation Candidate Selection Techniques in Layered, Tight Gas Sand Formations Using
Reservoir Simulation. SPE 63096 presented at the Annual Technical Conference, Dallas, Oct 1-4.
Rieb, B.A. and Leshchyshyn, T. T. 2005: Production Success of Proppant Stimulation on Horseshoe Canyon Coalbed Methane. SPE paper
96864 presented at the 2005 Annual Technical Conference, Dallas, Oct 9-12.
Rieb, B. 2006: Proppants Raise the Curve for Coalbed Methane Production. World Oil, Aug 2006.
Roussel, N.P. and Sharma, M.M. 2010: Quantifying Transient Effects in Altered-Stress Refracturing of Vertical Wells. SPE 119522
presented at the 2009 Hydraulic Fracturing Technology Conference, The Woodlands, TX, Jan 19-21.
Roussel, N.P. and Sharma, M.M. 2010: Optimizing Fracture Spacing and Sequencing in Horizontal Well Fracturing. SPE 127986 presented
at the 2010 International Symposium Formation Damage, Lafayette, LA, Feb 10-12.
Rodvelt, G. et al. 2001: Multiseam Coal Stimulation Using Coiled-Tubing Fracturing and a Unique Bottomhole Packer Assembly. SPE
paper 72380 presented at the 2001 Eastern Regional Meeting, Canton, OH, Oct. 17-19.
Rushing, J.A. 2003: Evaluation of a Hybrid Water-Frac Stimulation Technology in the Bossier Tight Gas Sand Play. SPE 84394 presented
at the Annual Technical Conference, Denver, Oct. 5-8.
Sallee, W.L and Rugg, F.E. 1953: Artificial Formation Fracturing in Southern Oklahoma and North Central Texas. Bulletin, AAPG (Nov
1953) 37, No 11, 2539-2550.
Saucier, R.J. 1988: A Field Development Program in Hydraulic Fracture Stimulation. SPE 18257 presented at the Annual Technical
Conference, Houston, Oct 2-5.
Sencenbaugh, R.N, et al. 2001: Restimulating Tight Gas Sand: Case Study of the Codell Formation. SPE paper 71045 presented at the
Rocky Mountain Petroleum Technology Conference, Keystone, CO, May 21-23.
Shaefer, M. 2006: Awakening an Old Field – A Case Study of a Refracturing Program in the Greater Green River Basin. SPE paper 101026
presented at the Annual Technical Conference, San Antonio, Texas, Sept 24-27.
SPE 134330 29

Shaefer, M.T. and Lytle, D.M. 2001: Fracturing Fluid Evolution Plays a Major Role in Codell Refracturing Success. SPE paper 71044
presented at the Rocky Mountain Petroleum Technology Conference, Keystone, CO, May 21-23.
Shah, S.N. et al. 2010: Fracture Orientation and Proppant Selection for Optimizing Production in Horizontal Wells. SPE paper 128612
presented at the SPE Oil and Gas India Conference, Mumbai, Jan 20-22.
Shelley, R.F. 1999: Artificial Neural Networks Identify Restimulation Candidates in the Red Oak Field. SPE paper 52190 presented at the
Mid-Continent Operations Symposium, Oklahoma City, Mar 28-31.
Siebrits, E. et al. 2000: Refracture Reorientation Enhances Gas Production in Barnett Shale Tight Gas Wells. SPE 63030 presented at the
Annual Technical Conference, Dallas, Oct 1-4.
Smith, C.F. and Murphy, D.L. 1974: Field Evaluation of Gas Well Stimulation Methods – Sutton County, Texas. SPE 4800 presented at the
Permian Basin Oil Recovery Conference, Midland, TX, Mar 11-12.
Stim-Lab Proppant Conductivity and Fluid Damage Consortia, 2007-2009. Duncan, OK.
Stutz, H.L. et al. 2002: Calibrating Coal Bed Methane Fracture Geometry in the Helper Utah Field Using Treatment Well Tiltmeters. SPE
paper 77443 presented at the 2002 Annual Technical Conference, San Antonio, Sep 29-Oct 2.
Surjaatmadja, J. 2007: The Mythical Second Fracture and Its Optimal Placement for Maximizing Production. SPE 106046 presented at the
2007 Europec/EAGE Annual Conference, London, June 11-14.
Tucker, R.L. 1979: Practical Pressure Analysis in Evaluation of Proppant Selection for the Low-Permeability, Highly Geopressured
Reservoirs of the McAllen Ranch (Vicksburg) Field. SPE paper 7925 presented at the Symposium on Low Permeability Gas Reservoirs,
May 20-22.
Tuffs, B. 2009: Mannville CBM in Central Alberta. CSUG 11th Annual Unconventional Gas Conference, Nov 18-20.
Vincent, M.C. 2002: Proving It – A Review of 80 Published Field Studies Demonstrating the Importance of Increased Fracture
Conductivity. SPE paper 77675 presented at the 2002 Annual Technical Conference, San Antonio, Sep 29-Oct 2.
Vincent, M.C. et al 2007: Field Trial Design and Analyses of Production Data from a Tight Gas Reservoir: Detailed Production
Comparisons form the Pinedale Anticline. SPE paper 106151 presented at the 2007 Hydraulic Fracturing Technology Conference,
College Station, TX Jan 29-31.
Vincent, M.C. 2009: Examining our Assumptions – Have Oversimplifications Jeopardized Our Ability to Design Optimal Fracture
Treatments? Paper SPE 119143 presented at the 2009 Hydraulic Fracturing Technology Conference, The Woodlands, Jan 19-21.
Wang, S.G., et al, 2007: Case Studies of Propped Refracture Reorientation in the Daqing Oil Field. SPE 106140 presented at the 2007
Hydraulic Fracturing Technology Conference, College Station, January 29-31.
Warpinski, N.R. and Branagan, P.T.. 1989: Altered Stress Fracturing. JPT Sept 1989 990-997.
Waters, G. et al. 2009: Utilization of Real Time Microseismic Monitoring and Hydraulic Fracture Diversion Technology in the Completion
of Barnett Shale Horizontal Wells. OTC 20268 presented at the Offshore Technology Conference, Houston May 4-7.
Wolhart, S. et al. 2007: Surface Tiltmeter Mapping Shows Hydraulic Fracture Reorientation in the Codell Formation, Wattenberg Field,
Colorado. SPE paper 110034 presented at the Annual Technical Conference, Anaheim, Nov. 11-14.
Wong, S.W. et al. 2000: Fresh Water Injection Stimulation in a Deep Tight Oil Reservoir. SPE paper 62618 presented at the 2000
SPE/AAPG Western Regional Meeting, Long Beach CA June 19-23.
Wright, C.A. and Conant, R.A. 1994: Reorientation of Propped Refracture Treatments. SPE paper 28078 presented at the 1994 SPE/ISRM
Rock Mechanics Conference, Delft, The Netherlands, Aug 29-31.
Wright, C.A., Conant, R.A. et al. 1995: Hydraulic Fracture Orientation and Production/Injection Induced Reservoir Stress Changes in
Diatomite Waterfloods. SPE paper 29625 presented at the Western Reional Meeting Mar 8-10.
Wright, C.A. and Conant, R.A. 1995: Hydraulic Fracture Reorientation in Primary and Secondary Recovery from Low-Permeability
Reservoirs. SPE paper 30484 presented at the Annual Technical Conference, Oct 22-25.
Wright, C.A. and Stewart, D.W. et al. 1994: Reorientation of Propped Refracture Treatments in the Lost Hills Field. SPE paper 27896
presented at the 1994 Western Regional Meeting, Long Beach, CA, Mar. 23-25.
Yao, F, et al. 2007: Reorientation Refracturing Case Study. SPE paper 106595 presented at the Production and Operations Symposium,
Oklahoma City, Mar 31 – Apr 3.
 Region Paper Author - Reference Companies Refrac Findings Reason for Refrac Success or Failure Location Type

Sallee, W.L. and Rugg, F.E.: "Artificial
Formation Fracturing in Southern Oklahoma
USA-Oklahoma
AAPG and North-Central Texas". Bulletin of the
and Texas
AAPG, Vol. 37, No.11 (November 1953). pp
2539-2550
Of 2,000 initial frac treatments studied, >85% were economically successful. Treatment sizes ranged from 1,000 gallons of fluid with 750 lb of sand to
Exploitation of the reservoir will be more economical when fundamental geological 100,000 gallons of fluid with 100,000 lb of sand. There is some indication that Southern Oklahoma
Dowell Inc. aspects are thoroughly understood. Many dry holes have been converted into larger treatments with higher sand concentration result in flatter production and North Central oil
commercial producers and marginal leases changed into valuable properties. One decline curves. Refracs often repaid investment in 60 days and restored Texas
of the first commercial fractures increased daily production to 200 bbl from 25 bbl. production to initial peak rates and lessened decline.

Fracturing has substantially increased reserves in North Texas. Within a 2-yr

Garland, T.M,, et al: "Effects of Hydraulic
Report of Fracturing upon Oil Recovery from the
USA-Texas Investigation Strawn and Cisco Formations in North
5371 Texas". Report of Investigation 5371, US
Dept. of Interior, Bureau of Mines, 1957
The age and condition of a well is a primary consideration. Results indicated
interval, 1,278,000 barrels of oil was recovered due to fracture treatments that
that larger treatments will generally result in a larger increase. Better results
would not have been recovered by primary means. Fracture treatments resulted
could be obtained from thick formations by isolating and treating the interval
in increased production that did not decline to pretreatment level until
Bureau of Mines zones separately. Wells previously treated with nitroglycerin did not respond North Central Texas oil
approximately 30 months after treatment. Treatments of old wells in which the
as well as those that had not been shot. Production increase from refracs
reservoir energy was depleted resulted in substantial increases in the rate of oil
usually will not be as large as that caused by the first. However, the decline
production. However, even greater benefits were obtained from wells that were
rate is typically reduced following restimulation.
treated before the reservoir energy had declined appreciably.

Ghauri, W.K.: "Results of Well Stimulation by
Hydraulic Fracturing and High Rate Oil
USA-California 1382 Backflush," paper SPE 1382 presented at the
1959 Fall Meeting of the LA Basin Section,
Oct. 22-23.
Along Whittier fault of Los Angeles basin (low perm and shaly) hydraulic fracturing
methods of hydrofrac and sandoil frac have been tried extensively along with the
stimulation technique of "high rate oil backflush" - injecting large volumes of
Shell Oil Co.
formation crude at high rates into production zone. Several treatments individually
have yielded no additional oil, but well stimulation by these methods over-all has
been highly successful. 53 jobs in 45 wells studied.
Wells were first hydrofraced, then sandoil frac, then high rate oil backflush. All
methods were successful overall, but Sandoil frac yielded higher additional oil
Southern California,
recoveries (> 62,000 bbl) than the other two methods. High rate oil backflush oil
Los Angeles Basin
was least productive. Although both propped and unpropped injections could
increase rate, propped fracs/refracs yielded superior sustained rate and EUR.

In the Strawn formation, the initial fracturing treatments used less than 1,000 gal of Wells that responded to the first fracturing job typically respond to re-
fluid with 1,000 lb of sand. Second and third treatments ranged from 1,500 to treatment, often with production rates equal to or greater than after the original
15,000 gallons of fluid with 1,200 to 30,000 lb of sand. Before the first treatment stimulation. Successful re-treatment is due to 1) extension of existing fracture
Howard, G.C., Fast, C.D., Hydraulic
SPE Editors for SPE the production averaged 199 bbl/month. After refracture they averaged 711 system, 2) re-opening of previous fractures, 3) washing of fracture faces, 4)
USA-Texas Fracturing. SPE Monograph 2, 1970. Chapter various oil and gas
Monograph monograph bbl/month. In gas storage wells, gas production following treatment has been replenishing of embedded proppants, and 5) opening new fractures in
10 "Hydraulic Fracturing".
several-fold greater that that obtained upon initial completions. In low permeability previously unfractured areas. Previously unstimulated wells responded better
areas, fracture treatments were so successful in increasing oil production that to propped fractures than wells that had been previously stimulated with
operators are drilling areas previously skipped. nitroglycerin.

Success attributed to development of viscous water-based fluid, allowing

Holditch, S.A. and Ely J.: "Successful Deep
Well Stimulation Utilizing High Proppant
USA-Texas 4118 Concentration," paper SPE 4118. Presented
at the 1972 SPE Annual Fall Meeting, San
Antonio, October 8-11.
This paper presents the results obtained with a viscous water-based fluid
Shell Oil Co. and
developed for fracturing deep wells where adequate flow capacity is difficult to
Halliburton
maintain. High closure stresses embed or crush the proppant, necessitating the
(consultant)
use of higher proppant concentrations/wider fracs to accommodate this damage.
placement of sand at 5-6 ppg instead of 2-3 ppg used in initial treatments.
Production rates were increased 3 to 4 fold, and EUR increased by ~1.7 BCF
for wells receiving hydraulic fracs after several years of production. Long-term
Texas gas
improvement is clear when compared to previous low-viscosity treatments. In
the hard dolomitic limestone of Devonian, Wilcox, and Hosston formations,
improvement was also noted.

Two methods of redistributing old proppant system are discussed: 1) Placing a

quantity of diverting gel on top of the existing proppant bed, allowing it to set,
As of 1973, > 500,000 frac treatments performed, more than 35% refracs. Refrac then pumping through the old proppant at a low rate until the flow rate can be
Coulter, G.R. and Menzie, D.E.: "The Design candidates include: low flow capacity proppant used initially, proppant degraded, increased sufficiently to remove the old proppant from the wellbore area, then
Halliburton,
of Refrac Treatments for Restimulation of insufficient length or height, or frac face damage. Primary considerations would be placing the new proppant system in the fracture. 2) Proppant redistribution by
USA 4400 University of increasing fracture width. By using highly viscous, proppant suspending
Subsurface Formations," paper SPE 4400. remaining reservoir energy and recoverable reserves. Lab tests conducted to
Oklahoma fracturing fluid at a sufficient rate, the existing fracture width would be
1973 determine a method of removing or replacing the old proppant system in the
fracture. Two techniques and their economic considerations are discussed here. increased, allowing the old proppant system to decrease in height, allowing
simultaneous displacement of the old proppant and placement of a new
proppant.
A well initially treated with a 30,000 gallon waterfrac was retreated with
Attempts to stimulate marginal gas producers in Sutton County, Texas, have 320,000 lb sand using an emulsified fluid, but achieved insignificant
provided an opportunity to evaluate several types of stimulation fluids on a improvement in production rates. The Massive Canyon and Holman Sand
Smith, C.F. and Murphy, D.L..: "Field
comparative basis. Seven fluids studied here include water base, oil base, and formation layers are different and require completely different well analyses.
Evaluation of Gas Well Stimulation Methods - Dowell Div. of Texas, Sutton
USA-Texas 4800 acid base fluids. A thorough study of the production history of wells following This paper discusses the test of emulsion frac and the modification of two gas
Sutton County, Texas," paper SPE 4800, Dow Chemical County
treatment demonstrates the effectiveness of the different fluids on a long term aqueous systems for special application in the Canyon. These systems are
1974
basis. Conclusions drawn from this study provide a guide to selection of fracture designed primarily to give maximum frac penetration in staged treatments at
fluids for tight gas sands. minimum cost per gallon of frac fluid, better clean-up properties, better sand
transport and, if possible, reduced formation damage.
Ohio, Holmes
Results of restimulation work performed in the Silurian Clinton Sand Formation, Refrac benefit was correlated to size of initial treatment. Wells initially County
east central Ohio were determined using decline curve evaluation. A total of fractured with small treatments provided the largest increase after Ohio, Muskingum
Greene, J.: "Study of Refracturing Results in
Oxford Oil seventy-one oil wells were restimulated with hydraulic fractures between 1972 and restimulation. It is also apparent that these restimulation reserves are the County
USA-Ohio 6361 the Clinton Sand Formation," paper SPE oil
Company 1975. Of these wells, seven had to be plugged and abandoned due to poor most economic reserves that can be established. Due to heavy salt water Ohio, Muskingum
6361, 1976
mechanical integrity. The average increased production due to restimulation was production, it is believed that large volumes of injected fresh water are County
4,000 bbls per well. beneficial to remove salt deposits from the fracture and formation. Ohio, Hocking
County

SPE 134330 Appendix Page 1 of 16 Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference
Crow, W.L.: “Mini Massive Frac - A
Stimulation Technique for Low Porosity, Low
USA-Indiana 6627 Permeability Zones," paper SPE 6627
presented at the 1977 Eastern Regional
Meeting, Pittsburgh, Oct 27-28.
Crowell, R.F. and Jennings, A.R.: "A
Diagnostic Technique for Re-stimulation
USA 7556 Candidate Selection," paper SPE 7556
presented at the 1978 ATCE, Houston, Oct 1-
3.
Kundert, D.P, and Smink, D.E.: “Improved
Stimulation of the Escondido Sandstone,"
USA-Texas 7912 paper SPE 7912 presented at the 1979 SPE
Symposium on Low-Permeability Gas
Reservoirs, Denver, May 20-22.
Tucker. R.L.: “Practical Pressure Analysis in
Evaluation of Proppant Selection for the Low-
Permeability, Highly Geopressured
USA-Texas 7925 Reservoirs of the McAllen Ranch (Vicksburg)
Field," paper SPE 7925 presented at the
Symposium on Low-Permeability Gas
Reservoirs, Denver, May 20-22.
Parrot, D. and Long, M.: “A Case History of
Massive Hydraulic Refracturing in the Tight
USA-Colorado 7936 Muddy 'J' Formation," SPE 7936 presented at
1979 SPE Symposium on Low-Permeability
Gas Reservoirs, Denver, May 20-22.
Kohlhaas, C.A. : “Field Study - Steinle Ranch,
an Intermediate Depth Oil Field, Shows
Significant Benefit from Bauxite Proppants" of Mines, Inexco,
USA-Wyoming 10875
paper SPE 10875 presented at the 1982
Rocky Mountain Regional Meeting, Billings,
MT, May 19-21.
Cramer, D.D.: "An Analysis of Post-
Stimulation Production Response in the
USA-North
12922 Madison: Elk Area, ND," paper SPE 12922
Dakota
presented at the 1984 Rocky Mountain
Regional Meeting, Casper, WY, May 21-23.
Pauls, R.W. et al. : “Successful Stimulation of
the Olmos Formation Using Oil-Base Fluids
and High-Proppant Concentrations" paper Halliburton, Royal
USA-Texas 13817
SPE 13817 presented at the 1985 SPE
Production Operations Symposium,
Oklahoma City, OK, Mar. 10-12.
Harper, T.R. et al.: "Fracturing Without
Proppant," paper SPE 13858 presented at
England 13858
the SPE/DOE 1985 Low Permeability Gas
Reservoirs, Denver, CO, May 19-22.
Conway, M.W. et al. : “Expanding
Recoverable Reserves Through Refracturing"
USA-Texas 14376 paper SPE 14376 presented at the 1985
Annual Technical Conference and Exhibition,
Las Vegas, NV, Sep 22-25.
Hunter, J.C.: “A Case History of Refracs in
the Oak Hill (Cotton Valley) Field," paper SPE
USA-Colorado 14655 14655 presented at the 1986 SPE East
Texas Regional Meeting, Tyler, TX, Apr 21-
22.
SPE 134330 Appendix Page 2 of 16
Companies Refrac Findings Reason for Refrac Success or Failure Location Type
Use of the Mini-Massive Frac has provided productivity increases many times
greater than conventional treatments in low-porosity, low-permeability
formations. This technique combines several existing techniques: Fine sand to
In the Salem Formation, many different treatments were attempted with acid,
control fluid loss into stress cracks formed or intersected; injection rates Indiana, Gibson
water, oil, and viscous gel fracturing. The best well was selected for restimulation
Dowell, Dow designed to control fracture height; extremely high sand concentrations County, Salem oil
with high sand concentrations up to 8 ppg, reducing payout time from 10 years to
produce packed factures which have better conductivity; larger size proppant Formation
35 days.
used to tail in at the end of the treatment can prevent proppant return to well-
bore during quick release of pressure.
Amoco This method involves the use of type curves, computer analyses of the original
This paper presents a method to determine the present condition of refracturing
Production, treatment, and a logical comparison of the results. The information necessary
candidates in low permeability gas formations and allows the operator to decide if
Halliburton to make a good assessment is discussed. Refrac using 875,000 lb sand
further stimulation is warranted.
Services provided significant increase in productivity
Low viscosity frac fluid used in original fracs may have resulted in short
Initial fracs using small sand volumes in low viscosity fluid doubled production, but fractures with low bed height. A polymer emulsion system was chosen for low
Halliburton, SW Texas, Webb
rapidly declined to pre-frac rate. Five wells were selected for refracturing with fluid loss and improved proppant transport, and to reduce clay swelling. It was
Consultant and County, Mesquite
higher viscosity fluids, increased proppant concentration, larger proppant diameter determined that high concentrations of sand would be necessary to provide the gas
Freemont Energy Field, Escondido
and greater proppant mass. Average production increase due to the refracs was needed flow capacity. A comparison of the decline curves indicates that a
Corporation Formation
620% and recoverable reserves were doubled. substantial increase in recoverable gas can be attributed to the improved
facture design.
Steep declines in production in this highly geopressured, low permeability field
A series of pressure transient tests showed fractures with sand were deteriorating require massive hydraulic stimulation for commercial production. PTA
South Texas,
Forest Oil over time. Use of a stronger proppant in restimulation treatments and in new indicated loss of length and conductivity in wells treated with glass beads or gas
Vicksburg Formation
completions resulted in superior, sustained performance. sand. Newly introduced bauxite proppant was tested in 1978 and early PTA
and production suggested sustained performance.
Initial fracs with 150,000 lbs sand provided low initial rates and rapid decline. On Success in Area A due to massive hydraulic refrac treatments with 300,000 gal
new wells, production increased with treatments as large as 1.25 million lbs. polyemulsion fluid and up to 735,000 lb of proppant. Areas B & C were
Wattenberg Field,
Amoco Refracs of wells in the better part of the field were found to payout in as little as 4 unsatisfactory after initial frac, probably due to lower permeability. Refracs in gas
DJ Basin
months with a gas price of $1.37/mcf. Large treatments appeared more areas with perm < 0.005 mD will likely not realize payout at gas prices below
successful. $1.50/mcf.
Initial stimulation treatments using sand or glass beads in the mid-1970s provided Success due to using appropriate proppant for conditions. Superior production
Colorado School disappointing production response. Significantly higher production rates were was achieved when water-based fluids were replaced with gelled oil, and sand Converse County,
obtained using bauxite with a return on incremental investment of 23 to 1. This and glass beads were replaced with sintered bauxite proppant. Other variables WY, Steinle Ranch
oil
Western prompted one operator to refrac 4 wells with bauxite. Two were failures, but two included well selection - wells which were good producers had higher rates of Unit, Muddy
Company successes added 62,000 bbls of reserves with 210% return on investment for return - and natural fractures - poor producers might have originally been formation
entire program. stimulated in a part of the formation with no natural fractures.
Breakdown acid systems have been used to bypass formation damage
Wells achieving low oil production after acid stimulation often responded favorably causing low productivity. Wells retreated with a gelled acid system have
The Western to propped refracs. Tail-ins with 12/20 sand and intermediate density proppant shown production rates more consistent with predictions. Proppant fracturing North Dakota,
oil
Company reduced proppant flowback, improved near-wellbore conductivity, and improved provided 1.5 to 9.1 fold production increases. Post-fracture decline rates Madison Formation
production compared to treatments only utilizing 20/40 sand. dropped to 25% from 54%.
Development of a new method of gelling hydrocarbon based fluids has allowed
Initial Olmos fracs with water based fluids and low proppant concentrations were the successful refracturing of deep, hot, low-compressive strength, sensitive
not very successful, with short apparent fracture lengths and rapid decline. Well formation. This enables very high concentrations of proppant to be pumped
Olmos formation,
#5: Large refrac at 12 ppg provided 13-fold increase in oil rate, and 13-fold over a long period of time at relatively high frac temperatures. oil
Oil and Gas Corp Texas
increase in 180 day cumulative production. Challenges included high bottom hole A review of production histories on wells indicates that this technique has
temperature and depths approaching 10,000 ft. substantially increased production rates, which remain more stable over time.
The results of this study indicate that conductive fractures can be formed
The results of fracture stimulation treatments implemented with and without
without proppant to effectively increase well productivity in the short run, and Egmanton and
British Petroleum propping agents in two reservoirs in England are compared. The fields studied are
suggest that the longer the fracture, the greater the effect. Refrac attempts Bothamsall fields, oil
Company not notably naturally fractured. Similar decline rates were noted for propped and
with proppant at 5-7 bpm provided benefit, in some cases significantly central England
unpropped fractures, however propped refractures were more effective.
increasing the rate of unpropped fractures.
This paper shows refrac results from three fields including the Olmos formation of
the Tilden oil field. Initial treatments in the Tilden utilized low sand concentrations. Successful refrac treatments used all available information and technology to
Olmos formation,
Refracs were designed to place 14+ ppg proppant. The results from 17 refracs are determine economic feasibility (sufficient reserves) and assess the possibility
Tilden field + an
Halliburton shown to provide sustained production increases in many wells by placing higher that a more effective fracture can be created. Success depends on accurately oil
undisclosed shallow
conductivity fractures. Similar results in an undisclosed South Texas field at 3000 assessing the results of the initial treatment(s), establishing appropriate
S. Texas oil field
ft depth placed 12/20 sand in refracs up to 10 ppg and documented 83% success fracture design, and developing techniques to achieve them.
rate with restimulation.
Initial treatments utilized <500,000 lb sand. Over time, designs specified more
Pipeline curtailments allowed shut-in time for pressure buildup without
sand with less frac fluid. This paper describes 7 refracs designed for 2 million lbs
intentional deferral of production. Some natural fracture leakoff may have Taylor Sand, Oak
reaching 10 ppg concentration. All 7 refracs were successful, average benefit 1030
Graham caused problems in some of the wells - natural fractures should be accounted Hill Field, Cotton
mcfd/well, FOI=3.6, Incremental reserves estimated at 1 bcf/well by operator, GRI gas
Resources for in simulations. Some re-fracs would not be cost effective during times of Valley, Rusk County
later determined 1.88 bcf/well. Wells with highest pre-frac kh responded best to
low gas prices, some performed well. Guidelines were established defining TX
increases in proppant mass and concentration. Post-frac buildup suggested 250 ft
good candidates for re-frac.
with conductivity of only 300 md-ft.
Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference
Niemeyer, B.L and Reinart, M.R.: "Hydraulic
Fracturing of a Moderate Permeability
USA-Alaska 15507 Reservoir, Kuparuk River Unit," paper SPE
15507 presented at the 1986 ATCE, New
Orleans, LA, October 5-8
Barby, B.G. and Barbee, W.C.: "Ultra-High
USA-Colorado,
Conductivity Fracture Stimulations: A Case
Kansas,
16222 History," paper SPE 16222 presented at the
Oklahoma,
1987 Production Operations Symposium,
Texas
Oklahoma City, Mar 8-10.
Branagan, P.T., et al : "Case History of
Hydraulic Fracture Performance in the
Naturally Fractured Paludal Zone: The
USA-Colorado 16397 Transitory Effects of Damage," paper SPE
16397 presented at the 1987 Low
Permeability Reservoirs Symposium, Denver,
May 18-19.
Griffith, P.J, et al:"Optimization of Fracture
Stimulation within the North Ward Estes
USA-Texas 17281 Field", paper 17281 presented at the 1988
Permian Basin Conference, Midland, Mar 10-
11.
Saucier, R.J.: "A Field Development Program
in Hydraulic Fracture Stimulation" paper
USA-Texas 18257
18257 presented at the 1988 ATCE,
Houston, October 2-5.
World Oil (Nov. Bagzis, J.M.: “Refracturing Pays Off in
USA-Colorado
1989) Rangely Field,” World Oil (Nov. 1989) 39-44.
Ely, J.W. et al : "Successful Proppant
USA- Fracturing of Ultradeep Sour Gas Reservoir,"
19768
Mississippi paper SPE 19768 presented at 1989 ATCE,
San Antonio, October 9-11.
Snow, S.E. and Brownlee, M.H.: "Practical
and Theoretical Aspects of Well Testing in
Norway 19776 the Ekofisk Area Chalk Fields," paper SPE
19776 presented at the 1999 ATCE, San
Antonio, October 8-11.
Pearson, C.M. et al. : “Optimal Fracture
Stimulation of a Moderate-Permeability
Reservoir—Kuparuk River Unit, Alaska,”
USA-Alaska 20707
paper SPE 20707 presented at the 1992 SPE
Annual Technical Conference and Exhibition,
New Orleans, Louisiana, Sept. 23-26.
SPE 134330 Appendix Page 3 of 16
Companies Refrac Findings Reason for Refrac Success or Failure Location Type
This paper examines the early history of refracs of the Kuparuk River Unit program, Design was evolved by continual interpretation of field results and related data
and shows the evolution of treatments to reach the desired fracture flow efficiency. from previous stimulations - including pressure data, temperature logs and 3-
Treatment design started with 20/40 mesh sand in gelled diesel, evolving to smaller dimensional simulation. Changes in fluid, proppants and schedule were made
treatments with higher 12/20 sand concentrations and gelled water. The third until success was achieved. More than 300 frac stimulations were made, 188
Kuparuk River,
Arco Alaska phase reduced sand concentration and went back to gelled diesel. Screenouts used the 20/40 mesh sand in gelled diesel that became the standard design. oil
Alaska
were a problem. The final treatment design used 20/40 sand in increased Increased job success rate and uniform designs have allowed up to four jobs
concentrations, in gelled diesel, in a reduced job size. Average post-fracture oil to be conducted by one crew in a single day, creating cost reduction as well as
rate was 755 BOPD, up from 227 BOPD. Screenouts were negligible, and costs desired flow efficiency. Subsequent papers 20707 and 24857 documented
were reduced. See SPE 20707 & 24857 additional improvement in refrac design.
In the Morrow formation, large diameter frac sands were inadequate at 4500 psi Fracture treating the production wells resulted in significantly improved
closure stress. An 8/12 ISP ceramic proppant was successfully placed at over 8 production. Large, high strength proppants improved conductivity and
Barby Energy
ppg. Proppant fracturing increased well productivity between 2.4 and 4.2 fold from productivity and are better in small concentrations than smaller sand in higher Four State area
Corp, Edinger, oil
previous acid stimulations. When bottomhole pressures are further reduced, concentrations. Acidizing was of marginal benefit sandstone. If fracture covered by Morrow.
Inc.
solution gas evolves and liquid production is reduced, validating laboratory studies treating after acid treatment, consideration should be given to over-flushing the
regarding pressure losses within multiphase systems. wellbore, as proppant flowback was problematic.
A well completed in a naturally fractured zone was initially productive without
The damage mechanisms (liquid or frac/fluid polymer damage to natural
stimulation. An unpropped fracturing experiment reduced production capacity by
Sandia National fractures) were theorized as being time-dependent. Interval was isolated for Colorado, Rulison
about 20%. A subsequent propped refrac further reduced deliverability by 20%.
Laboratories, 18 months and re-tested. Production was then superior, potentially due to fluid Field, Piceance gas
The continued decrease in production following each fracturing experiment was
CER Corp., DOE imbibition or dehydration. This paper details the well testing, analytical Basin``
attributed to decreasing the flow capacity of the natural fractures that were
evaluations and reservoir modeling throughout the experiments and re-entry.
intersected by the hydraulic fracture, and residual fluid or gel damage.
Production modeling indicated that high sand concentrations up to 12-16 ppg
Production declines are frequently offset by refracturing, and it is not uncommon
are required to optimize production from this field. However, concern with Texas, North Ward
Chevron for a well to have been refractured several times. Of 39 recent refracs, 29 were oil
sand production caused a recommendation for 9 ppg. Previous refracs added Estes Field
economically successful (74%).
18,000 bbls of reserves per attempt.
A Fracture Development program was instituted to improve hydraulic fractured well
productivity through an increased understanding of frac behavior and improved Cooperation between operating company, service company and R&D, and
design and optimization techniques. Fracture fluid type was changed from water to utilization of modern testing methods resulted in improved understanding of
Shell Western oil to water based fluids. Minifrac testing was initiated, field quality control frac behavior. Refinement of frac fluid choice and proppant type in higher South Texas, AWP
oil
E&P procedures were instituted, proppant type was changed and proppant concentrations led to higher productivity while cutting costs. Due to QC Olmos
concentrations reached ever higher levels. Improved well performance, increased program, frac jobs are pumped largely problem free, and frac performance has
technical capabilities, enhanced R&D model development, improved service improved.
company products and service, and reduced costs were documented.
Bagzis details the evolution of frac design from 100-lb hydrafracs in 1947 to high
Refrac candidate selection is detailed.
conductivity fracs with 8 to 10 lb/gal concentrations and ceramic proppant tail-ins.
Current successes can be attributed to: Elimination of use of diverters to stage
The Weber formation ranges in depth from 5500 to 6500 ft, with average
fracs; placement of higher proppant concentrations; better gels and better leak- Colorado, Rangely
Chevron permeability of 5 to 10 md. By 1989, >1700 frac jobs had been performed on 891 oil
off control. Field
wells, with many re-fraced 3-4 times with proppant ranging from ground walnut
Recommendations: Perform additional tests with ceramic tail-in; continue
hulls, to glass beads, to natural frac sand. “In general, refrac successes have been
improvement in gel quality; determine optimum length.
achieved by using higher sand concentrations.”
The success of this program was accomplished through planning, engineering,
Stimulation in the deep, high-H2S, high temp, high pressure Johns field presented and the utilization of state-of-the-art technology, using the best equipment and
challenges to safety and technology. Previous moderate volume, low proppant materials available. Bauxite and high temperature stabilized zirc fluids were Rankin County,
Tomlinson concentration treatments had not been satisfactory. Four wells were restimulated selected for the high temperature environment. Onsite evaluation of all Mississippi,
Interests, with >500,000 lbs of bauxite as concentrations reaching 8-9 ppg. All of the fracturing fluid combined with intense quality control procedures assured that Smackover gas
Republic Refining refracturing treatments have been financial successes and productivity in the quality fluids were mixed and pumped, and that problems could be corrected formation, Johns
Jones field was increased by more than 300%. Current production is about 650 immediately in the field, as screenout could have resulted in loss of the well. field.
Mmcf/mo up from 180 Mmcf/mo. Mini-frac treatments were conducted on each well not only to evaluate leakoff,
but to evaluate the equipment and the capabilities of the service company.
Reductions in stimulation effectiveness occur gradually and periodic
An efficient testing program for this harder, more complex chalk formation has restimulations can be beneficial. Pressure tests indicate that kh is dependent
Phillips Petroleum been developed using the equivalent fracture system concept, to accurately on stimulation efficiency, particularly on diversion effectiveness. Pressure test Ekofisk Area, North
gas & oil
Co., Norway forecast production, monitor well performance, evaluate the need for restimulation, results have been successfully used for optimizing stimulation designs and for Sea
and optimized completion and stimulation designs. matching fracture growth models. Periodic pressure tests have been effective
in defining the condition of the stimulation and the potential for restimulation.
This paper describes the evolution from small treatments pumped with 20/40 sand
Stimulation of moderate-perm formations required different approach.
to a variety of treatment sizes with larger LWC proppant.
Overriding design criterion is fracture conductivity, not fracture length.
These design changes were shown to more than double the post-frac productivity
Significant steps made in optimizing the treatment design include removing the North Slope, Alaska,
ARCO of the wells. The authors state: “The permeability of the proppant pack is one of oil
100-mesh sand, limiting silica flour to the pad stage, using higher-conductivity Kuparuk River Unit
the most important factors governing the final conductivity of the
16/20 and 12/18 mesh LDC proppant, increasing the treatment size to 60k to
fracture...evaluation of a proppant needs to be made in terms of cost per millidarcy,
150k lbm, and changing frac fluids to more viscous water-based gels.
not cost per pound.” 140 wells in project. See SPE 24857
Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference
Branch, G.A. and Drennan, K.M.: “Refracture
Stimulations in the Norge Marchand Unit: A
USA-Oklahoma 21642 Case Study,” paper SPE 21642 presented at
the 1991 SPE Production Operations
Symposium, Oklahoma City, Apr 7-9.
Olson, K.E.: “A Case Study of Hydraulically
Refractured Wells in the Devonian Formation,
Crane County, Texas," paper SPE 22834
USA-Texas 22834
presented at the 1991 Annual Technical
Conference and Exhibition, Dallas, TX, Oct 6-
9.
Krasey, R. and Jackson, M.: “Revitalization of
Canada,
JCPT 92-09-08 the Pembina," paper JCPT 92-09-08 , JCPT
Alberta
(Nov 1992) 65-72.
Fleming, M.E.: “Successful Refracturing in
the North Westbrook Unit,” paper SPE 24011
USA-Texas 24011 presented at the 1992 SPE Permian Basin
Oil and Gas Recovery Conference, Midland,
Texas, Mar. 18-20.
Hower, T.L., and Decker, M.K.: “Identifying
USA-Colorado Recompletion Candidates in Stratified Gas
and New 24307 Reservoirs," paper SPE 24307 presented at
Mexico the 1992 SPE Mid-Continent Gas
Symposium, Amarillo, TX April 13-14.
Hejl, K.A.: “High-Rate Refracturing:
Optimization and Performance in a CO2
USA-Colorado 24346 Flood,” paper SPE 24346 presented at the
1992 SPE Rocky Mountain Regional Meeting,
Casper, Wyoming, May 18-21
Pospisil, G. et al. : “Results of a Large-Scale
Refracture Stimulation Program, Kuparuk
River Unit, Alaska,” paper SPE 24857
USA-Alaska 24857
presented at the 1992 SPE Annual Technical
Conference and Exhibition, Washington,
D.C., Oct. 4-7.
Durrani, A.J. et al. : “Rejuvenation of 30-Year-
Old McAllen Ranch Field - An Application of
USA-Texas 24872
Cross-Functional Team Management," paper
SPE 24872, JPT (Jan 1994) 1065-1072.
Elbel, J.L. and Mack, M.G.: "Refracturing:
Observations and Theories," paper SPE
USA 25464 25464, presented at the 1993 Western
Regional Meeting, Long Beach, CA. March
23-25
Palmer, I.D.; "Induced Stresses Due to
Propped Hydraulic Fracture in Coalbed
Methane Wells," paper SPE 25861 presented Amoco Production
USA 25861
at the 1993 Rocky Mountain Regional Low
Permeability Symposium, Denver, April 12-
14.
SPE 134330 Appendix Page 4 of 16
Companies Refrac Findings Reason for Refrac Success or Failure Location Type
Although only 6 of 15 refracs were successful, entire program generated 64%
Initial completions placed ~66,000 lbs of sand reaching 3 ppg. None of the initial ROR. Successful refractures resulted in average increase of 50 BOPD and 98
Grady City, OK,
fracs screened out. Refracs with larger diameter ISP increased production 50 MBO additional reserves per well. Good wells made best candidates.
Norge Marchand
Oryx Energy bopd and added 98,000 bbl additional reserves per well. Six premature screenouts Intermediate strength proppant used to obtain higher fracture conductivity. oil
Unit, Northwest
(attempting to place 10/20 or 16/20 in refracs). Well testing showed ability to Successful refracs had perm >.14 md and positive skin more than .39.
Chickasha Field
identify superior refrac candidates, which would further improve economics. Probability of screen-out reduced if 20/40 ISP is pumped in early stages with
16/20 or 10/20 tail. PTA helpful in identifying refrac candidates.
Original fracs placed 2 ppg 20/40 sand. PBUs on 12 wells indicated no
appreciable frac contribution. 3 wells refractured with up to 75,000 lbs 20/40 ISP at
Candidate wells should be reperforated in desired interval. Hydraulic isolation
concentrations up to 5 ppg. Well #1: Refrac placed 23,000 of 75,000 lb design.
of desired interval was beneficial. Appropriate proppant and high
Well doubled in production, eventually stabilizing at 40 bopd with 14 month payout.
Mobil concentrations were specified to achieve needed conductivity. High injection Devonian Field, oil
Well #2: Reperfed, refraced with 71,000 lb ISP, generating 8-fold increase in oil
rates with viscous gels were used to maximize fracture width when adequate
rate, with 3 month payout. Well #3 refractured with 90,000 lb ISP. This refrac
barriers were present to contain height growth.
undesirably communicated with Upper Devonian, but still doubled production rate,
generating an 18 month payout.
Occidental performed 83 fracture treatments in the mature Pembina field. Over
Despite more than 45 years of field development, only 19 of 83 refracs were
time, frac size has increased from 25 tons in gelled crude to 50-70 tonnes of 20/40
unsuccessful. Detailed results shown for all 83 attempts. No change in WOR
Canadian sand in polyemulsion fluids. The average refrac payout was one year, with Pembina field,
suggests minimal or no reorientation of refracs, yet clear, documented oil
Occidental incremental reserves per job of 2130 m3. 64 of 83 refracs were considered Cardium formation
increases in recoverable reserves. Infill drilling was successful 82% of time,
successful, with no measurable increase in WOR. Restimulation of existing wells
compared to 77% success rate with refracturing.
had stronger economics than infill drilling.
This paper traces the evolution of stimulation in the Westbrook Field, and
In the Westbrook Field of central Texas, Fleming reports the restimulation of 32 oil
discusses the treatment design that came out of the 1989 field study. The
wells. The Middle Clearfork formation is an oil-bearing dolomite at 3000 ft depth. Texas, Westbrook
Chevron components discussed include proppant selection, proppant concentration, oil
Larger sand (12/20 Brady) and aggressive ramps up to 14 ppg allowed twice the Field
frac fluid selection, quality control and the forced closure technique.
incremental rates of refracs performed with lower concentrations of 20/40 sand.
ROR using these criteria is >200% with payout of less than one year.
In the San Juan Basin, the Mesaverde formation was originally stimulated with Original nitroglycerin shots failed to contact all productive intervals. Use of p/z Mesaverde Group,
nitroglycerin in the 1950's. In the late 80's, Prima recompleted wells, sidetracked, plots helped to identify wells producing from a single interval. Recompletion San Juan Basin,
Intera Bergeson, redrilled and hydraulically refractured wells. 14 openhole wells sidetracked and and restimulation of shot wells is cost effective way to open up all layers, Blanco Mesaverde
gas
Prima Oil and Gas refraced, with 100% success rates. Incremental rates of 609 mcfd/well achieved, recovering reserves that would have been missed. . Conventional propped Field of New Mexico,
with an average FOI of 25. Operator estimated 2 bcf incremental reserves per fracturing techniques are more effective approaches to stimulate multi-layered and Ignacio Blanco
well, with only $383,000 investment. wells. Field of Colorado
In the Rangely Field in Colorado, five decades of fracture optimization can be High sand concentrations and increased injection rates beneficial. Money
traced. A 1992 update by Hejl showed the results of 198 recent refracture saved by: High pump rates to perform single stage treatments. Fluids and
treatments, with current designs specifying pumping rates exceeding 80 bbl/min additives optimized to reduce costs & contamination. Casing or 4-1/2 in.
Colorado, Rangely
Chevron and 8 to 12 lb/gal sand. Pads have been reduced and FLA eliminated to maximize tubing to reduce hydraulic horsepower cost. Procedure optimized by oil
Field
conductivity. The average incremental production from 85 wells restimulated in eliminating acidizing, perforation break down and streamlined pumping
1990 was 52 bopd during the first year, and over 1.66 MMBO incremental from procedures. Treatment schedule altered to increase lengths and proppant
March 1990 to January 1992 concentrations w/o increasing volumes.
One of the most convincing demonstrations of the benefit of fracture conductivity is
Increase in productivity from changes in treatment design, increased job size,
in the Kuparuk Field on the North Slope of Alaska. By 1992, more than 880
different proppant types, proppant mesh sizes and fluid types, oriented and North Slope, Alaska,
ARCO stimulation treatments, including 200 refracs had allowed ARCO to further improve oil
aligned perforations. Most changes focused upon increasing fracture Kuparuk River Unit
production. Optimization involved switching to LWC, then to progressively larger
conductivity and achieving superior connection with wellbore.
proppant sizes, higher concentrations and reduced pad size
Gas production has been increased by 250% in the McAllen Ranch Field. 75 fracs Original frac used < 5 ppg sand, while refrac specified 8 ppg IDC. With
or refracs performed. This study highlights a well refractured in 1991, resulting in stacked pay, operator began to complete and individually stimulate all stages
Shell Western McAllen Ranch
production increasing from 150 mcfd to 1000 mcfd, a 6.7 fold-of-increase. GRI for commingled completion instead of leaving some intervals for future gas
E&P Field, Vicksburg
later estimated 1.7 BCF reserves added by this treatment, which demonstrates the development. Cross-functional team able to generate innovative approaches,
redevelopment opportunities in mature gas fields. cut time and costs.
A review of a number of refracturing treatments was made to determine what Obvious refrac candidates include wells with inadequate initial length or
insight could be obtained from their treatment pressure histories. The effect of the conductivity. However, several wells were successfully refractured despite
Dowell previous proppant and the stress changes due to pore pressure depletions were of PTA suggesting presence of deeply penetrating, highly conductive initial fracs.
Schlumberger particular interest. Comparing pressure histories of the treatments led to some Refrac reorientation may be induced by pore pressure gradients in the
observations and theories on the role played by reservoir pore pressure and formation. The optimal time to restimulate for reorientation will depend on well
reorientation in refracturing treatments. properties.
Possible benefits of stress-induced reorientation include circumventing the gel
Because of high net pressures in coalbed formations, stresses induced by propped damage caused by the original fracture, or propagating a fracture across the
fractures can be substantial. If the original propped fracture has altered the dominant natural fracture system to enhance production. In another case, a various gas
Research
stresses sufficiently, a refracture may initiate orthogonal to the original fracture. sacrificial frac treatment can change an unfavorable stress contrast into a
favorable one.
Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference
Schraufnagel, R.A. et al. : “Restimulation
Techniques to Improve Fracture Conductivity GRI, Taurus
USA-Alabama 26198 and Overcome Damage," paper SPE 26198 Exploration, Stim- paid out investment in 180 days, and added 383 MMcf reserves at $.10/mcf. Well
presented at the 1993 SPE Gas Technology
Symposium, Calgary, Alberta, June 28-30.
Kuuskraa V.A. et al. : “Economic and
Technical Rationale for Remediating Advanced
Inefficiently Producing Eastern Gas Shale Resources
USA-Michigan 26894 and Coalbed Methane Wells," paper SPE International, GRI,
26894, presented at the 1993 SPE Eastern
Regional Conference, Pittsburgh, PA, Nov. 2- Exploration
4.
Wright, C.A. et al.: "Reorientation of Propped
Refracture Treatments in the Lost Hills Field,"
USA-California 27896 paper SPE 27896, presented at the 1994
Western Regional Meeting, Long Beach, CA,
March 23-25.
Hailey, R.G. and McKaughan, C.A.: “Case
Study: Isolation and Restimulation of Granite
Wash Zone in Mendota, NW Field Using
USA-Texas 27933 Inflatable Packer Frac Tools" paper SPE
27933 presented at the 1994 SPE Mid-
Continent Gas Symposium Amarillo, TX, May
22-24.
Reese, J.L. et al.: "Selecting Economic
Refracturing Candidates," paper SPE 28490, Amoco Production conducted in wells in permeable reservoirs require increased fracture conductivity
USA 28490
presented at the 1994 ATCE, New Orleans, Company
LA, September 25-28.
Hopkins, C.W., et al : "Screening
Restimulation Candidates in the Antrim
USA-Michigan 29172 Shale," paper SPE 29172 presented at the
1994 Eastern Regional Conference,
Charleston, November 8-10.
McMillan, D.J., and Suffron, R.D.: "Innovative
Re-fracturing Program Produces Economic
and Technical Success in Foothills Cardium
Canada,
CIM 95-44 Formation," paper CIM 95-44 presented at Service, Amerada
Alberta
the 1995 Annual Technical Meeting of the
Petroleum Society of CIM, Banff, Alberta,
May 14-17.
Kiburz, D.E et al.: “A Methodology to
Optimized Fracturing Treatments in a Tight
Gas Field of Northwest Louisiana,” paper
USA-Louisiana 29554
SPE 29554 presented at the 1995 Rocky
Mountain Regional Symposium, Denver,
Colorado, Mar. 20-22.
SPE 134330 Appendix Page 5 of 16
Companies Refrac Findings Reason for Refrac Success or Failure Location Type
This paper describes the success of restimulation of CBM wells. Well P3 squeeze-
Cement can be used to abandon undesirable hydraulic fractures prior to
cemented, refraced with N2 foam and 186,000 lbs sand. Increased production Rock Creek site,
restimulation. Alternative gel polymer minimized permeability damage during
Birmingham, gas
restimulation. Foamed frac fluid may provide more effective CBM restimulation
Lab P4 refraced with 117,000 lbs sand without cement squeeze. P4 payout in 500 Alabama
compared to original water and sand treatment.
days, adding 48 MMscf incremental reserves at $.63/mscf.
This paper summarizes results in Antrim Shale, Michigan. Well testing was
Successful remediation requires: simple, low cost diagnostics for separating
conducted in an 8-well pattern to select a single well for restimulation. This well Michigan Basin,
poor completions from poor geology; good data to understand cause of
was restimulated in two stages designed for merely 10,000 lb proppant. Antrim Shale, gas
completion problem and to develop good remediation plan; diligent
Taurus Incremental rate of 348 mcfd (FOI=6.6). Incremental reserves of 0.64 Bcf later Bagley East Project
implementation; and post-remediation to assess what works.
estimated by GRI, with the restimulation cost only $15,000.
Four new wells with modern initial fracture treatments were refractured in the same
intervals with treatments similar to the original treatments. All wells responded with The most striking conclusion was that refracture treatments can reorient and
post refracture production equal to or slightly less than IP following the original propagate along a significantly different fracture plane than original fracture
Lost Hills treatments. On the average these wells showed production increases of 50 BOPD treatments of the exact same intervals. Possible mechanisms for this
Petroleum over the previous daily rates of approximately 10 BOPD. This paper summarizes phenomena are discussed. This phenomena allows refracture treatments to Lost Hills field,
oil
Engineering, the investigation into possible reasons for apparent "virgin" production response reach untapped reserves, thereby increasing production. However, in Bakersfield, CA
Chevron resulting from refractures. Several potential mechanisms were considered. Stress waterflood applications, the induced refracture planes may no longer be
change induced reorientation of the refractures along a different fracture plane parallel, leading to premature breakthrough, insufficient sweep, and poor
than the original fracture was considered most likely. recovery efficiency. Mitigation of these problems could be complicated.
Information on "pre 1987" fracs vs. "post 1987" fracs is also given.
Original fracs were ineffectively stimulated with limited entry designs. Infill wells
treated and tested each zone individually, with enhanced rates. One refrac
candidate was initially fractured with 510,000 lb 20/40 sand. One of 6 intervals Very low gas recovery for wells originally stimulated by limited entry method.
Seagull Midcon, was isolated with inflatable packers and refractured with 142,900 lbs 20/40 sand. Upper intervals my have taken majority of original treatment; lower intervals Mendota (Granite
gas
Baker Oil Tools Post-refrac production from single “G” interval alone peaked at 1070 mcfd and inadequately stimulated. Excellent results were obtained by treating and Wash) Field
stabilized at 650 mcfd. Of the 120 mcfd initial (six zones) assumed only 20 mcfd testing each zone individually with frac liner.
from “G” interval. GRI later estimated over 1 BCF incremental reserves added with
this restimulation treatment.
Conclusions include:
A literature survey shows that refracture stimulation treatments in tight formations - Refracs in permeable formations should be designed to increase fracture
require increased fracture length and that refracture stimulation treatments conductivities.
-Refracs in low perm formations should be designed to increase fracture
multiple
to be commercially successful. Refracturing in depleted reservoirs may not be length.
commercially viable. This paper evaluates refracturing and provides guidelines -Refracs in depleted low perm formations are probably not economically
with respect to stimulation design and commercial viability. viable.
- Refracs in paritally depleted permeable reservoirs are potentially viable.
The Gas Research Institute is involved in a detailed research project to advance
Gas Research
technology in the Antrim. Objectives include increasing the effectiveness of As a result of this study, many wells were identified for restimulation. Because
Institute, Terra
fracture stimulations, developing methods for screening restimulation candidates, of sand flow back problems or over flushing resulting in poorly propped
Energy, Ward
identifying optimal recompletion candidates and procedures, and improving the fractures at the wellbore, or inadequate stimulation, 500 to 1000 Antrim wells
Lake, Muskegon
understanding of the Antrim reservoir. This paper focuses on methods for could be candidates for restimulation. The methodology for screening Michigan Basin,
Development,
screening restimulation candidates. candidates is described. The annulus injection tests developed to determine if Antrim Shale, gas
Oilfield
A novel injection test unit and procedures to help identify the best restimulation a hydraulic fracture exists at the wellbore can be run for $500-600, in less than Bagley East Project
Investments,
candidates in a cost-effective manner, in a reasonable amount of time has been 2 hours. The methodology can be applied to all wells that might benefit from
Wolverine
developed. The test is designed to determine if there is an effective hydraulic restimulation, provided modifications are made to the test design for specific
Environmental
fracture connected to the wellbore. 50 tests have been run, to date, identifying formation/well types.
Production
potential restimulation candidates in about 1/3 of the cases.
Amerada study in Western Foothills documents evolution of fracture designs. In
1960’s – 20 tonnes and concentrations up to 600 kg/m3. In 1990 – acidizing
(marginal results + scaling). In 1992 – Intense re-fracturing program, with
Nowsco Well Success due to better candidate selection and treatment design. New hybrid
concentrations increased 3.3 fold; tonnage was increased 100%. 17 refrac
frac fluid allowed proppant concentrations to be increased while decreasing Cardium, foothills oil and gas
treatments performed (2 on gas wells). All wells show increased production,
Hess Canada frac fluid to proppant ratio.
ranging from 62% to 1043%. Average three month production rates up 2.5 fold.
Total incremental oil production estimated at 165,210 m3. NPV of project to date is
$3.3 million.
Using modern tools and proper data, optimized fracture stimulation treatment
can be designed, including fluid volumes, proppant volumes and proppant
Refrac program began in 1990 to restimulate wells previously fractured with low
types. Production up 72%, Costs down 14%. Louisiana, Gray
Marathon, proppant concentrations. Using large diameter uncoated proppant tail-in increased
Increase in productivity attributed to use of high-temp encapsulated breaker to Sand Reservoir,
Schlumberger conductivity without increasing proppant flowback. Increasing proppant gas
enhance fracture conductivity; larger mesh proppant as tail-in further increased Smackover
Dowell concentration to 5-7 ppa, with more aggressive ramp and increasing breaker
conductivity while lowering job costs; final proppant concentration increased Formation
improved well productivity.
from 5 to 7 ppa. Comparing performance between wells important to measure
effects of design changes.
Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference Companies Refrac Findings Reason for Refrac Success or Failure Location Type

Waterflooding has been extensively employed to increase total recovery, and to Waterflood injection in diatomite reservoirs has been observed to cause infill
Wright, C.A. et al: "Hydraulic Fracture
Pinnacle mitigate the potentially catastrophic effects of reservoir compaction and the hydraulic fracture azimuths to rotate by more than 60 degrees from the original
Orientation and Production/Injection Induced
Technologies, resulting surface subsidence. The low permeability of the diatomite reservoirs, fracture orientation, and to cause horizontal fractures where previously were Belridge and Lost
Reservoir Stress Changes in Diatomite
USA-California 29625 Area Energy, however, results in generation of significant reservoir pressure gradients which vertical fractures. Waterflood design can be greatly compromised by the Hills reservoirs, oil
Waterfloods," paper SPE 29625 presented at
Crutcher-Tufts result in significant changes in the local reservoir stress field. This in turn causes assumption of a constant hydraulic fracture azimuth. Measurement of tilt at California
the 1995 Western Regional Meeting,
Production reorientation of hydraulic fractures on infill wells and contributes to wellbore casing many locations, together with the far-field nature of tiltmeter mapping, allows
Bakersfield, March 8-10.
failure. determination of fracture orientation.

Cramer, D.D.: "The Evolution of Hydraulic
Fracturing in the Almond Formation," paper
USA-Wyoming 30480
SPE 30480 presented at the 1995 ATCE,
Dallas, October 22-25.
Hydraulic fracturing treatment design has evolved over the years. More than 600
wells were evaluated for reservoir, production and treatment characteristics in the
Almond formation. An effective fracture stimulation design combines proppant
BJ Services scheduling of the late 70s with fluid and gel-breaker systems of today. An overview
of the evolution is given, with problems and successes discussed. Refrac success
highly tied to remaining reservoir pressure. Depleted intervals appear to benefit
from N2 or CO2 energized fluids.
Highly productive wells have been successfully treated with crosslinked gelled
water treatments utilizing moderate concentrations of 20/40 sand. Remedial
Washakie Basin and
polymer-specific enzyme treatments have reversed fracture/formation damage
Wamsutter Arch,
caused by previously-injected fluids. Fracture overpressuring damage caused gas
Rock Springs,
by near-wellbore packing of the fracture has in some cases been reversed by
Wyoming
CO2/solvent treatments. Pressure-drained Almond intervals can be damaged
by all-water-based treatments and respond well to energized retreatments.

Wright, C.A. and Conant, R.A.: "Hydraulic Operational changes can mitigate at least some adverse effects. Recognizing
Reorientation in low-perm reservoirs changes due to production activities over the
Fracture Reorientation in Primary and that putting new infill wells on production can alter fracture reorientation, a
life of the field. Fluid production results in depletion which alters the stress state,
Secondary Recovery from Low-Permeability strategy was put in place to drill and fracture all new infill wells before any were Lost Hills field, San
Pinnacle but not symmetrically or universally. A method is presented that allows indirect
USA-California 30484 Reservoirs," paper SPE 30484 presented at put on production, which reduced the degree of reorientation. Larger pipe was Joaquin Valley, oil
Technologies, Inc measurement of the cumulative effects. Capitalizing on refrac reorientation could
the presented at the 1995 Annual Technical used in new wells in order to facilitate future refrac treatments. Strategies for California
significantly increase field production without drilling new wells, an economic
Conference and Exhibition, Dallas, Texas, both primary and secondary recovery should consider the impact of
advantage.
October 22-25. reorientation so that it can be either exploited or mitigated.

Mobil Oil Spectral gamma ray technology used in 150 gas wells in the Hugoton field The information from the tracers revealed unstimulated pay zones, and was
Hecker, M.T., et al : "Improved Completion
Corporation, provided important information that was used to identify unstimulated pay zones used to continually modify completions. Through limited stress barriers and
USA-Kansas, Designs in the Hugoton Field Utilizing Multiple Hugoton Field,
Western Atlas and to continuously improve the design of fracture stimulations. The spectral logs permeability variations, perforation schemes were successfully modified to
Oklahoma, 30651 Gamma Emitting Tracers," paper SPE 30651 Kansas, Oklahoma, gas
Logging Services were used to evaluate diversion techniques and fracture containment. Perforating improve fracture containment and proppant placement. Bypassed pay was
Texas presented at the 1995 ATCE, Dallas, October Texas
and Technics strategies were altered based on this knowledge, minimizing costs while identified by lack of radioactive tracers, yielding both candidates for
22-25.
International maximizing ultimate gas recovery. restimulation and improved perforating strategies.

Radioactive tracer technology was used to analyze 98 wells with 136 fracture
treatments in four formations. Results indicate that in nearly 40% of the Almond formation,
Radioactive tracers can be used to reliably measure fracture height, so that
ProTechnics completions one or more zones did not receive any or all of the designed Wyoming, Cotton
Fisher, K., et al: "A Comprehensive Study of treatments can be modified. Unstimulated perforation sets can be identified
International, S.A. treatment. In this study, production and revenue forecasts for the actual treatment, Valley formation,
the Analysis and Economic Benefits of and the problem solved via refracturing or alteration of future perforating
Holditch & the optimal treatment, and other possible outcomes was generated. It was found East Texas,
USA 30794 Radioactive Tracer Engineered Stimulation strategies. Understimulated pay intervals can be corrected. Tracers are not gas
Associates, that if problems are correctable in only a small percentage of completions, the Delaware formation,
Procedures," paper 30794 presented at the profitable on every well, but when averaged over a large number of wells,
Meridian, UPRC, economic benefits outweigh the costs due to significant production increases in a New Mexico, Red
1995 ATCE, Dallas, October 22-25. tracer technology can increase field reserves by up to 10%, increasing profits.
Enron Pennzoil small percentage of the wells. Guidelines are discussed for designing and Fork formation,
Benefit-to-cost ratios ranged from 9:1 to 12:1.
analyzing radioactive tracer programs that can maximize productivity and overall Oklahoma
profit.

A decrease in productivity due to increasing skin damage near the wellbore was Success was attributed to quality control and real-time evaluation of mini-frac
Davidson, B.M., et al: "Stimulation Program in noted. Several types of treatments were performed. Acid treatments, xylene and fracturing data. Initial fracs were superior with higher proppant
Empresa
High Permeability Oil Sands - Case Study," treatments, clay stabilizer treatments, and high pressure gas fracturing had concentrations, shorter fracs, and increased net pressure reached during TSO
Colombiana de
South America- paper 39050 presented at the 1997 Latin disappointing results. Hydraulic fracture treatments provided the best results and designs. Treatments were optimized on location based on observed data Putumayo Basin,
39050 Petroleos oil
Colombia American and Caribbean Petroleum justified the economics of implementing a major stimulation program. 17 achieving dramatic results. Modest benefit shown with refrac. Using the Southern Colombia
Ecopetrol, Neo-
Engineering Conference, Rio de Janeiro, RJ, treatments on 10 wells have been performed, and the productivity of the wells has correct fluid, proppant, proppant volume and proppant concentrations each
Energy Partners
Brazil, August 30 - September 3. increased from 614 bbl/d to 1461 bbl/d. Water cut was not increased by these contributed to increased production. Fracture modeling, history matching and
treatments. 3-D modeling were used to design for higher conductivity fractures.
Fairchild Jr., N.R., and Wood, D.D.:
Fairchild and Wood documented the benefit of refracturing gas storage wells in Real-time 3d fracture modeling and reservoir simulation improved proppant
“Optimization of Clinton Restimulation
Ohio. Early treatments utilized 5000 to 45,000 lbs of 20/40 sand, with 25% pad and placement and fracture conductivity. Modern technologies were used for well
Program in Stark-Summit Gas Storage Field: Ohio, Stark-Summit
USA-Ohio 39220 East Ohio Gas 2-3 lb/gal concentration. A refrac accepted 40,000 lb of 20/40 sand with selection, and test wells were used to evaluate application of technologies to gas storage
A Field Cast Study,” paper SPE 39220 Field
concentrations up to 8 lb/gal. Flow testing showed post-frac deliverability doubled, optimize treatment design and proppant concentrations. Many wells
presented at the 1997 SPE Eastern Regional
and was over 187% higher than any previous treatment. restimulated, now experiencing proppant flowback.
Meeting, Lexington, Kentucky, Oct. 22-24
Amoco
Moschovidis, Z.A., et al: "The Mounds Drill
Production, Exxon This paper describes an extensive field experiment to improve understanding of The intention of the experiment was to find an ecologically sound way to
Cuttings Experiment: Determining Placement
Production the mechanics and modeling of the processes involved in the injection of drill dispose of drilling waste, while creating fractures near and far field from the
USA-Oklahoma 48987 of Drill Cuttings by Hydraulic Fracturing Mounds, Oklahoma oil
Research, ARCO cuttings. Mapping and coring definitively prove reorientation can occur with wellbore, and to verify that fracture models can be calibrated to predict fracture
Injection," paper SPE 48987 presented at the
E&P Technology, subsequent injections. behavior. Conclusive evidence of reorientation.
1998 ATCE, New Orleans, September 27-30.
BP Exploration
Wang, F., Ding, Y., and Yong, L.: “A Study of In Northwestern China, four low perm oil wells were refractured with higher
Success is due to treatment designed for low-permeability reservoirs. This
Refracturing in Low Permeability Reservoirs,” Langfang conductivity proppants. Field “S” has an average permeability of ~ 7 md, at a depth
paper discusses what to look for in refracture candidates, proppant and fluid China, northwest,
China 50912 paper SPE 50912 presented at the 1998 SPE Research of 9800 ft. Well testing indicated the degradation of the initial fracture conductivity oil
selection, and tools to use to effectively evaluate initial fracture behavior for Field "S"
International Oil and Gas Conference and Institute, with time. Refracs performed with a higher permeability ceramic proppant
designing new treatment.
Exhibition, Beijing, China, Nov. 2-6. substantially increased production rates.

SPE 134330 Appendix Page 6 of 16 Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference
Leshchyshyn, T., et al: "Technical and
Economic Rationalization of Oil Refracture
Canada, Petroleum Programs on Wells Previously Stimulated
Alberta Society 99-60 with Water-based fracture Fluids," presented
at the 1999 CSPG and Petroleum Society
Joint Convention, Calgary, June 14-18.
Shelley, R.F.: "Artificial Neural Networks
Identify Restimulation Candidates in the Red
USA-Oklahoma 52190 Oak Field," paper SPE 552190 presented at
the 1999 Mid-Continent Operations
Symposium, Oklahoma City, March 28-31.
International, Gas
Reeves, S.R., et al: "Restimulation of Tight
Gas Sand Wells in the Rocky Mountain
USA 55627 Region," paper SPE 55627 presented at the
1999 Rocky Mountain Regional Meeting,
Gillette, May 15-18.
Reeves, S.R., et all: "Restimulation for Tight International, GRI,
Gas Sand Wells," paper SPE 96482
USA 56482
presented at the 1999 ATCE, Houston,
October 3-6.
Mohaghegh, S. et al: "Performance Drivers in
Restimulation of Gas Storage Wells," paper
USA-Ohio 57453 SPE 57453 presented at the 1999 Eastern
Regional Meeting, Charleston, October 21-
22.
Aguilar-Razo, R.: “Propped Fracturing in Gas
Carbonate Formations in Mexico," paper SPE
Mexico 58987 58987 presented at the 2000 International
Petroleum Conference, Villahermosa,
Mexico, Feb. 1-3.
Exxon, Branagan
& Assoc, Sandia,
Pinnacle, OGCI,
Moschovidis, Z.A., et al: "The Mounds Drill
Advantek, Arco,
Cuttings Experiment: Final Results and
Baroid, Chevron,
USA-Oklahoma 59115 Conclusions," paper SPE 59115 presented at
GRI, Halliburton,
the 2000 IADC/SPE Drilling Conference, New
Orleans, February 23-25.
West Virginia
Mohaghegh, S., et al.: "Development of an
Intelligent Systems Approach for
USA 59767 Restimulation Candidate Selection," paper
59767 presented at the 2000 SPE/CERI Gas
International, and
Technology Symposium, Calgary, April 3-5.
SPE 134330 Appendix Page 7 of 16
Companies Refrac Findings Reason for Refrac Success or Failure Location Type
More proppant was placed with the oil-based refracture, increasing propped
While water-based fractures are more economical, they may not work in
fracture length and flowback load recovery to make the stimulations
undersaturated reservoirs as load fluids can cause reservoir damage. Wells from
Fracmaster, successful. Casing fracture programs were designed to avoid proppant
each formation were fractured with a cross-linked water-based fluid. Both wells
Crestar Energy, concentration/frac-width problems. Resin coated proppants were used, as Glauconite formation
failed to respond, had very low load fluid returns, and were considered for oil and gas
Schlumberger ordinary 20/40 sand cannot withstand crushing of greater depths. Injection and Viking formation
abandonment. A re-interpretation of well logs suggested oil-based re-fracture
Oilfield Services rates were reduced to provide reduced fracture height, more fracture length
programs. The wells were not permanently damaged, and responded successfully
with higher fracture conductivities, producing higher production rates. 3D
to the oil-based re-fracture programs designed and placed in both wells.
simulators enabled better design.
This paper documents the development of an Artificial Neural Network that Using the ANN methods described in this paper, operators can screen well
predicted Red Oak post-recompletion production with an average error of less than candidates before performing more detailed engineering analyses. This
5%. Reservoir simulators, 3D design models and various pressure and log- method is appropriate when an experience base exists. The ANN is not a
analysis techniques require detailed well history, reservoir information and stand-alone tool, but can be economical and expedite a well-evaluation
Red Oak field,
Halliburton geological understanding. These methods can be time-consuming and expensive, process. The number of labor hours required to complete the ANN procedures
Latimer and Le Flore gas
Energy Services and often records are incomplete or unavailable. ANN collects data from previous are minimal compared to hours required for conventional geologic and
counties, Oklahoma
recompletions then develops evaluation criteria that can be used on the remaining engineering evaluation. In Red Oak, 30% of wells were good refrac
wells in the field. This paper focuses on determining whether data from candidates. Poor candidates included older, lower pressured wells on
recompletions can be used with ANN to identify wells with the best opportunity for periphery of field, wells initially treated with larger volumes of fluid, and wells
economic gain through recompletion. with low production rates following the initial frac.
Advanced
Resources GRI believes that candidate selection methodology is a technology
advancement which will provide the greatest industry benefit. Poor candidate
In this study, a series of sequential analyses would be used to reduce the total well
Research selection procedures seem to be the reason that many restimulation projects
population to the ~15% that are actually high-potential refrac candidates. The first
Institute, Ely & are unsuccessful. While refinements and improvements in each analytic level Green River Basin,
level should be rapid screening and low in cost, followed by incrementally more gas
Associates, were made, the inconsistencies in selection remained, exemplifying industry- Piceance Basin
detailed, labor-intensive levels. Then, well performance must be separated into
Holditch & wide difficulty of candidate selection. What is clear is that restimulation
reservoir quality and completion efficiency components. (Reference paper 29172)
Associates, potential does exist, but that some degree of site-specific knowledge and
StimLab, West methodology customization is required.
Virginia U.
Advanced
A multi-process candidate selection methodology has been developed to identify
Resources
viable restimulation candidates. It consists of production comparisons, engineering
Production type-curve matching and production performance comparison
based performance assessments, and pattern recognition technology. Also
Holditch Reservoir attempt to identify underperforming wells. Pattern Recognition Technology Green River Basin,
incorporated are individual well reviews, economic analysis, and a new short-term gas
Technologies, was incorporated to better indentify activities that may have led to an inefficient Piceance Basin
field test for candidate verification. Lab studies have also identified new
Stim-Lab, Ely & completion.
procedures for effective clean-up of unbroken gel in propped an natural fractures.
Assoc., West
(Reference paper 55627)
Virginia Univ.
Parameters are interdependent. There is no "one" most important factor. It is
best to build accurate and representative model, then query on a case by case
The Clinton sand gas storage wells are on an annual restimulation program.
basis. Statistical analysis indentified sand volume to be most important
Several wells have been refractured multiple times, while some have only been
West Virginia criterion. Backward elimination showed that average rate is most important and
fractured once in 30 years. A study has been conducted to identify the Clinton formation,
University, CNG emphasized the role of the service company in the success of fracture jobs. gas storage
performance drivers in each series of frac jobs, which could be an important Northeast Ohio
Transmission When first, second and third fracs were analyzed separately, it was revealed
economic factor in the design of new jobs. The study combined neural network
that as the number of fractures performed on a particular well increases, the
and fuzzy logic tools.
importance of sand volume decreases and influence of service company
increases.
Premature screenout. A net increase of 0.2 MMCFD was achieved, but not
Pemex clarified that 12/18 LWC reaching 8 ppg was used to restimulate a
enough to recover the investment in a short term. This paper shows that
previously acidized completion. The well was not perforated optimally for a
combined analysis gives a better picture - screenout could have been avoided Mexico, Mata gas
Pemex propped treatment, and high net pressures were interpreted as 9 multiple fractures,
by employing step down test. Use of real data is important. Perf program Pionche field, condensate
causing an early screenout. Despite these complications, the well produced at 1.6
should be designed for fracturing. Modified frac fluid systems should be
mmcfd, higher than previous matrix and acid frac treatments.
considered.
PCM Technical,
Summary of the experiment to improve understanding of mechanics and modeling
BP/Amoco, Multiple fractures were created spanning an azimuth range of 20-33 degrees.
of downhole injection of drill cuttings. (Reference SPE 48987) Over 20 intermittent
Intermittent injections were effective in limiting fracture propagation.
cuttings-slurry injections were conducted into each of two disposal formations while
Repeatable and stable injection pressure behavior was observed.
imaging the created fractures with surface and downhole tiltmeters and downhole Mounds, Oklahoma
Conventional fracture simulations estimated the dimensions of the disposal
accelerometers, verifying the imaged fracture geometry with comprehensive
Hughes- domain with acceptable accuracy for waste disposal operations. Reorientation
deviated-well, and coring and logging programs through the hydraulically fractured
Christensen, of subsequent injections provide large implications for restimulation treatments.
intervals.
Mobil, MSD,
Schlumberger,
Shell, UPRC,
A/OC
Artificial neural networks, genetic algorithms and fuzzy logic are the facets of virtual Virtual intelligence has been modified from its original implementation to better
University,
intelligence techniques used in this study to develop a methodology for capture complexity in the candidate selection process. This methodology is
Advanced
restimulation candidate selection in tight gas sands. This paper discusses how now capable of selecting candidate wells that show improvement after Rocky Mountain gas
Resources
virtual intelligence works, and how the three main facets work together to solve the restimulation, although it cannot at this time estimate the potential incremental
problems presented. production. This paper gives examples of how this was used in field studies.
GRI
Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference
Wong, S.W., et al: "Fresh Water Injection
Stimulation in a Deep Tight Oil Reservoir,"
Oman 62618 paper SPE 62618 presented at the 2000
SPE/AAPG Western Regional Meeting, Long
Beach, June 19-23.
Siebrits, E., et al : "Refracture Reorientation
Enhances Gas Production in Barnett Shale
USA-Texas 63030 Tight Gas Wells," paper SPE 63030
presented at the 2000 ATCE, Dallas, October
1-4.
Jeffrey, R.G. and Settari, A.: "Hydraulic
Fracture Growth through Offset Pressure-
Australia 63031 Monitoring Wells and Boreholes," paper SPE
Taurus Reservoir
63031 presented at the 2000 ATCE, Dallas,
October 1-4.
Ely, J.W., et al: "Restimulation Program Finds
Success in Enhancing Recoverable
USA 63241
Reserves," paper SPE 63241 presented at
the 2000 ATCE, Dallas, October 1-4.
Marquardt, M.B., et al., "Production Gains
from Re-Fracturing Treatments in Hassi
Algeria 65186 Messaoud, Algeria," paper SPE 65186
presented at the 2000 European Petroleum
Conference, Paris, October 24-25.
Guoynes, J. et al. : “Optimizing Deliverability
in Five Gas-Storage Reservoirs—Case
USA-Illinois 65636 Studies,” paper SPE 65636 presented at the
2000 SPE Eastern Regional Mtg,
Morgantown, VA, Oct. 17-19
Ammer, J.R., and Sames, G.P.: “Advances
for Improved Storage: Deliverability
US DOE, Nat'l
Enhancement and New Storage Facilities,”
USA-Ohio 65638
paper SPE 65638 presented at the 2000 SPE
Technology Lab
Eastern Regional Meeting, Morgantown,
Virginia, Oct. 17-19.
Emrich, E., et al : "Codell Restimulation
Evolve to 200% Rate of Return," paper SPE
Patina Oil & Gas,
USA-Colorado 67211 67211 presented at the 2001 Production and
Operations Symposium, Oklahoma City,
March 24-27.
Shaefer, M.T. and Lytle, D.M.: "Fracturing
Fluid Evolution Plays a Major Role in Codell
Refracturing Success," paper SPE 71044
USA-Colorado 71044
presented at the 2001 Rocky Mountain
Petroleum Technology Conference,
Keystone, May 21-23.
Sencenbaugh, R.N. et al. : “Restimulating
Tight Gas Sand: Case Study of the Codell
Formation," paper SPE 71045 presented at HS Resources, BJ
USA-Colorado 71045
the 2001 SPE Rocky Mountain Petroleum
Technology Conference, Keystone, CO, May
21-23.
SPE 134330 Appendix Page 8 of 16
Companies Refrac Findings Reason for Refrac Success or Failure Location Type
A water squeeze trial was conducted under the belief that removing the salt
Petroleum
surrounding the silicilyte formation would increase permeability. This produced a PTA, decline analyses, and chemical evaluation of salt concentrations in
Development
temporary production increase, later determined to be propagation of unpropped flowback fluid assisted in understanding phenomena of multiple injection Athel formation,
Oman, Pinnacle oil
fractures instead of the intended matrix injection. However, the lack of proppant cycles of fresh water. Based on the transient collapse, it appears fractures South Oman
Technologies
allowed the fracture to heal, and benefit was short-lived. Propped fractures in this should be propped for sustained performance in this formation.
Delft
formation were shown to increase productivity by a factor of 7.
This paper discusses reorientation of restimulation treatments. A number of wells
Mitchell Energy Refracture reorientation penetrates untapped sections of the reservoir to
in the naturally fractured Barnett Shale were successfully refractured. Tilt mapping
Corporation, increase production. Reorientation may allow more cost-effective reservoir Barnett Shale gas
showed significant reorientation, and production data demonstrate dramatic
Schlumberger contact than closely spaced infill wells.
productivity improvement.
Several hydraulic fracture treatments are described, including data not usually
available. Conclusions include: Treatment over old mines shows that the free
CSIRO Treatments were designed to investigate fracture growth in coal and rock and to surface strongly affects the fracture growth and must be accounted for in
Petroleum, induce the rock to cave into a mine opening. Fractures that grew through adjacent modeling. A minifrac and a re-fracture of the same interval the next day New South Wales,
monitored boreholes provided direct measurements of fracture growth rate or produced nearly identical pressure response at the injection well, and at a Australia
Solutions pressure in the fracture away from the injection point. monitoring well located 50 m away. Growth rates were also similar. Tensile
stresses and ahead of the leading edge of the fracture produced noticeable
pressure reductions at the monitor points as the fracture approached.
Gas Research Institute initiated a restimulation study. One of the focuses of the GRI believes many restimulations fail due to poor candidate selection. They
project is candidate selection. GRI believes this is where the greatest industry also argue that often good producers are the best candidate, even though that
benefit resides. They describe the multi-level methodology they field-tested and the seems counterintuitive. Before re-stimulating an underproducer, "why" it Green River Basin,
Gas Research
results of the tests. They detail how these procedures can be used cost-effectively. underproduces should be understood, as it may not be a good candidate. Piceance Basin and gas
Institute
Cotton Valley candidates were underperforming, refrac'ed with CO2 and 80,000 to Proper design and execution are critical, as is recovery of the stimulation fluid. Cotton Valley
170,000 lbs sand compared to 1-4 million pounds in initial fracs. Modest benefit in Limits placed on design and execution from mechanical conditions can result
3 wells. in failure.
Nine refracs were performed in the Hassi Messaoud field between 1996 and 2000
The success of refracs resulted from improved field practices of treating open-
resulting in three types of responses. Wells with large production gains were
hole and slotted liner completions, and from the use of state-of-the-art
successful due to the amount and type of proppant placed, correcting the
fracturing equipment and engineering tools.
inadequate conductivity from the original treatment. Wells with modest refrac
Sonatrach, The lessons learned give new insight to candidate selection for this field.
benefit were often in portions of the field with asphaltene damage, and success Hassi Messaoud
Halliburton Limited conductivity of the original fracture and proppant pack damage in the oil
was achievable despite premature screenouts, as benefit was attributed to the near- field, Algeria
Energy Services near wellbore area are more likely to show improvement through refracture
wellbore placement of fresh undamaged proppant to bypass the damage. Marginal
with 12/18 bauxite and doubling fracture width. It is difficult to create longer
refrac success was achieved in wells with relatively low permeabilities in which frac
effective frac lengths, and refracs targeted at this goal were of marginal
extension was attempted with a refrac. One example showed that refrac parting
benefit.
pressure was elevated higher than initial treatment.
Success attributed to custom treatments based on accurate identification of
Seventy-five wells in five gas storage reservoirs were diagnosed and treated, with
damage mechanisms, using new testing methods and candidate selection
Halliburton, evaluations made at 1- and 2- year intervals. In the Cook’s Mills field in central Illinois, Cook's Mills
methods. gas storage
Kinder Morgan Illinois, wells re-stimulated with higher conductivity fractures achieved 70 to 80% Field
In this case, hydroblasted to clean wellbore, then fraced with SMA frac sand
injection/withdrawal increases.
and low polymer fluids, improving performance immediately by 70-80%.
Technologies tested: liquid CO2 with proppant, propellant, tip-screenout and
Ammer and Sames discuss results in the Stark-Summit/Chippewa gas storage field extreme over-balanced fracturing. Liquid CO2 and proppant fracture
in East Ohio. 29 fracture treatments were performed as part of the project, testing treatments led to a 7-fold increase in deliverability; tip-screenout fracture
Energy Eastern Ohio gas storage
four new and novel fracture stimulation technologies. Wells with higher non-Darcy treatments led to an increase of nearly 50 MMcfd in deliverability; Well
flow components responded better to re-stimulation. selection criteria was improved; wells with higher non-Darcy flow components
responded better to stimulation.
Success of this program is due to combination of stringent well selection
The authors report on results of the most recent 200 refractured Codell wells. criteria, high fluid quality control guidelines and effective operational field
Using a reduced CMG polymer fluid system has yielded incremental monthly practices. A diagnostic algorithm improved candidate selection and contributed Wattenberg Field,
production results of approximately 80% of the original initial production. to improved performance. The combination of the CMG fluid system and DJ Basin, Codell gas
BJ Services
The evolution and experimentation of treatment design (pump rates, fluid and sand source water control contributed to improved performance. The rate of return Formation
volumes, and modern technology improvements) since the 70s is discussed. increased from 66% to 208% based primarily on fracturing fluid and selection
criteria.
A lower percentage of proppant entered the Codell when it was completed
comingled with the Niobrara. Designing a treatment just for the Codell led to
Average oil and gas rates have been increased by greater than 500% by
increased production. Wattenberg Field,
BJ Services, HS restimulating over 750 Codell wells. The evolution of applied fracturing fluid
Treatments pumped with the CMG based fluid show increased production Denver-Juleseburg oil and gas
Resources technology has played a major role in the successes in the Codell refrac program.
compared to HPG treatments. Lower polymer loadings and cleaner base basin, Colorado
Different fluids are discussed.
water lead to economic success. CMG requires less polymer loadings, causing
less formation damage, and has superior rheological properties.
Between 1997 and 2001, HS Resources refractured 750 Codell wells, generating a
Improvement of frac fluids and placement techniques. Designs changed
project ROR of 100%. Although well spacing is 40 acres, it appears wells drain 15 Wattenberg Field,
toward larger sand concentrations and higher pump rates. Cost cutting gas-
to 30 acres. Refracturing is a preferable choice to infill drilling to recover reserves. DJ Basin, Codell
Services techniques such as limited entry completions. Successful for underperforming condensate
Added 20 MMBOE reserves (26.6 MBOE/well) by refracturing. Daily production Formation
wells as well as good producers.
increased 10 bopd and 148 mcfd on average.
Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference Companies Refrac Findings Reason for Refrac Success or Failure Location Type

Most failures are due to wellbore problems, fracturing materials and

equipment, fracture mechanics or reservoir properties. Proppant
Jennings, A.R.: "When Fracturing Doesn't When the limits of existing technology are reached, producing disappointing concentrations have increased, and stronger proppants have become
Work," paper SPE 71657 presented at the Enhanced Well fracture treatments, evolutionary improvements have been made in the industry. A available. Monitoring and measuring advances have lead to more realistic
Global 71657
2001 ATCE, New Orleans, September 30- Stimulation historical overview is given, common reasons for failures are described, and modeling. Fracturing fluids have evolved to fill specific needs. Equipment has
October 3. suggestions are offered to provide further improvements. been designed to keep up with size and pressure needs. Formation flow
capacity and reservoir pressure have been identified as important properties
influencing fracturing success.

Attempting to identify the most important performance driver in such a complex

Mohaghegh, S., et al,: "Performance Drivers
USA-Ohio 74715 in Restimulation of Gas-Storage Wells,"
paper SPE 74715 revised from SPE 57453
system may be an oversimplification. Parameters are interdependent. It is best
An update on the Clinton sand annual restimulation program, established in the to build an accurate model and query on a case-by-case basis to identify
West Virginia
late 1960s and analyzed in SPE 57453 in 1999. Linear regression analysis, fuzzy drivers for each design. For Clinton sand, neural networks identified average Clinton formation,
University, CNG gas storage
curves and neural networks are applied to the problem of identifying influential rate and the role of the service company as most important, and revealed that Northeast Ohio
Transmission
parameters to design optimum fractures and refracs. as the number of fractures performed on a particular well increases, the
importance of sand volume decreases and the influence of the service
company increases.

Successful stimulations are engineered for individual well conditions. The

Barree, R.D, et al : "A Practical Guide to
Hydraulic Fracture Diagnostic Technologies,"
USA 77442 paper SPE 77442 presented at the 2002
STCE in San Antonio, September 29 to
October 2.
Barree & Various diagnostics are discussed and compared, with strengths and limitations authors state that each diagnostic technology has strengths and weaknesses,
Associates, explained. Near-wellbore diagnostics include radioactive tracers and temperature and that often more than one technology needs to be applied to adequately
Pinnacle logging. Far-field diagnostics include tiltmapping and microseismic fracture describe the fracture parameters needed to design individual treatment. One
Technologies, mapping. These methods can be used together to get a more complete view of Oklahoma example using RA tracer indicated bypassed pay. A bridge plug
ProTechnics actual well conditions, to formulate a more complete treatment design plan. allowed restimulation of upper perfs allowing 75% production improvement and
0.5 to 0.75 BCF added reserves.

Stutz, H.L. et al: "Calibrating Coal Bed
Methane Fracture Geometry in the Helper
Utah Field Using Treatment Well Tiltmeters,"
USA-Utah 77443
paper SPE 77443 presented at the 2002
Annual Technical Conference and Exhibition,
San Antonio, Sept 29-Oct 2.
Vincent, M. C., “Proving It – A Review of 80
Published Field Studies Demonstrating the
USA-Colorado
Importance of Increased Fracture
and New 77675
Conductivity”, paper SPE 77675 presented at
Mexico representing over
the 2002 Annual Technical Conference and
Exhibition, San Antonio Sep 29-Oct 2.
Jin, L.. et al .: “Large-Scale Hydraulic
Fracturing in a Frontier Area in China," paper
China 77680 SPE 77680 presented at the 2002 Annual
Technical Conference, San Antonio, Sept 29-
Oct 2.
Asadi, M. et al: "Monitoring Fracturing Fluid
Flowback with Chemical Tracers: A Field
USA-Colorado 77750 Case Study," paper SPE 77750 presented at Patina Oil & Gas
the 2002 Annual Technical Conference and
Exhibition, San Antonio, Sep 29-Oct 2.
Sauer, P.W. et al: "Re-Fracturing:
Evaluation, Design, and Implementation of a
Petroleum Corp., in 1999 with 17,000 lbs curable RCS. In 2002, a caustic soak was pumped in an
Chester Oil Well in SW Kansas," paper SPE
USA-Kansas 80916
80916 presented at the 2003 Production and
Operations Symposium, Oklahoma City,
Technologies
March 23-25.
Husen, A.A. et al .: “Hydraulic Fracturing the
Best Producer - A Myth?," paper SPE 81543
Egypt 81543
presented at the 2003 SPE Middle East
Show and Conference, Bahrain, April 5-8.
Gutor, C. et al : “New Life for Old Wells - A
Case Study of the Effects of Re-Stimulating
Gas Wells Using Fracturing Through Coiled
Canada,
81730 Tubing and Snubbing Techniques" paper
Alberta
SPE 81730 presented at the 2003
SPE/ICoTA Coiled Tubing Conference,
Houston, TX Apr. 8-9.
In the Helper Field, restimulation of coal bed methane (CBM) wells proved
Anadarko,
successful. Figure 6 shows a 15-fold increase in gas rate following a refrac.
Pinnacle
Figure 4 shows dramatically higher productivity of wells corresponding to high
Technologies
proppant volumes.
CARBO
Two coal bed methane wells in the San Juan Basin were successfully restimulated
Ceramics, and improved by increasing fracture conductivity. This summary demonstrates that
with ceramic proppant, providing tremendous increases in gas production and
authors well productivity and profitability can be frequently be improved by redesign of
accelerated dewatering. Although the closure stress was minimal, the low beta
hydraulic fractures, despite the failure of many existing models to predict those
factor of the uniform spherical proppant provided substantial benefits.
70 companies
Burlington
Initial fracs in this 0.05 md gas sand did not produce adequate rates to justify field
Resources, Yuma
development. Refracs have provided 4-15 times higher production rates, using up
Exploration &
to 700,000 lbs of higher quality ceramic proppant at concentrations up to 8 ppg.
Production
This paper presents a novel technique - Chemical Frac Tracers - whereby
fracturing fluid flowback can be effectively monitored in a complex multi-zone
ProTechnics,
fracture system. A family of environmentally safe chemical frac tracers, with
unique characteristics was developed so that flowback of each separate stage
Corp.
could be calculated, and so flowback efficiency of each stage can be known, as
well as the flowback efficiency of the well.
Anadarko A well was initially fractured with 173,000 lbs 16/30 Ottawa sand, then restimulated
Integrated attempt to dissolve resin, followed by a tri-frac containing 52,000 lbs ELWC.
Petroleum Pressure transient testing, treatment pressure analyses and Reciprocal Productivity
Index comparisons indicate a successful treatment.
Although operators are reluctant to restimulate the best wells, good wells are often
Schlumberger,
excellent candidates. This operator selected a good well, and designed an
GUPCO (Suez
aggressive TSO treatment incorporating up to 12 ppa 16/20 LWC. Production
Petroleum
doubled to more than 3000 bopd and paid out the frac treatment in less than 30
Company)
days.
Fifteen shallow gas wells producing from the Medicine Hat and Milk River
Enerplus
formations were selected for restimulation. Wells were originally completed in the
Resources Fund,
1970’s. Despite significant pressure depletion over 30 years, refracs performed
Schlumberger
well, often exceeding initial production rates.
Tiltmeter instruments are available which can measure fracture height and
width in real time from within the treatment wellbore, eliminating the need for a
monitor well. The results in the Helper Field were fed into a 3D fracture
simulator. The results indicate that adequate height is being created to cover Utah, near Price.
gas, CBM
large pay section through a single, small perforated interval. This reduces Ferron Coals
stimulation and completion costs as compared to the multi-stage technique.
The limited interval perforating technique also minimizes the number of
fractures and reduces fracture complexity.
In over 80 field studies, in all types of reservoirs, well productivity was
various
benefits.
Previous refracture failures were attributed to several causes, including each
discipline performing its work without feedback from others, lack of adequate
equipment, and environmental and governmental challenges. China, Sichuan
Engineering challenges were identified and plans were implemented to Basin. Bajiaochang gas
overcome them. Modern practices, an integrated, multidisciplinary team, and gas field, XX4 group
using proper equipment and experienced personnel are cited as the reasons
for the success of the program.
Flowback analysis of each frac fluid stage can help in better designing frac
fluid additives such as gel stabilizer, oxidizer breaker, and enzyme breaker. It Wattenberg Field,
also helps to distinguish fracture-tip from near-wellbore flowback. Analysis of DJ Basin, Codell gas
Wattenberg flowback indicated poor cleanup in refracs, attributed to excessive Formation
crosslinker and stabilizer.
In the failed stimulations, the volume of fluid was believed to be insufficient to
create the desired fracture dimensions, and the proppant concentrations and
size were wrong for the Chester formation. Poor lateral proppant distribution
Chester formation,
was believed to be a contributing problem. Before the third, successful, oil
Seward Co., Kansas
stimulation, efforts were made to flush the old proppant from the near-wellbore
area. Fluid rheology, proppant type, and treatment schedule were altered to
achieve a more desirable fracture.
Key to success is integration of all data, transforming data into knowledge and
Western Desert,
using it to make appropriate practical decisions. People, Technology and oil
Egypt
Process.
Outlined ways to consider & select wells that are candidates for re-frac. Well
preparation prior to stimulation is essential to reduce overall cost. Both coiled Southern Alberta.
tubing conveyed and snubbing conveyed fracture stimulation of re-entry Medicine Hat, Milk gas
shallow gas wells resulted in significant increase in production, but coiled River formations
tubing was more time efficient and cost effective.

SPE 134330 Appendix Page 9 of 16 Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference Companies Refrac Findings Reason for Refrac Success or Failure Location Type

Popa, A.S., et al,: "Identification of

Contaminated Data in Hydraulic Fracturing
Databases: Application to the Codell
West Virginia
Global 83446 Formation in the DJ Basin," paper SPE
University
83446 presented at the 2003 Western
Regional/AAPG Pacific Section Joint
Meeting, Long Beach, May 19-24.
Not all records are completely accurate. Errors in stored data can be the result of
A neuro-cluster system was developed to ID contaminated data, using a neural
improper or incomplete data collection, errors in data entry, lack of proper
network. This methodology can be extended to any type of database for the ID Codell formation,
interpretation, etc. This paper presents a method for identification of contaminated
of contaminated data. In this dataset, approximately 50% of data was culled. Wattenberg Field, gas
data, separating data into good data, slightly contaminated data, or bad data. This
Removal of suspect data will enable best practices identification, and better Colorado
method was applied to an existing database, enabling successful fracturing
treatment design.
interpretation.

381 stimulations studied. Initial treatment effectiveness does not necessarily

Brown, K.G. et al. : “Assessment of
Remediation Treatments in Underground Gas
Storage Wells" paper SPE 84393 presented
USA 84393
at the 2003 SPE Annual Technical
Conference and Exhibition, Denver, CO, Oct
5-8.
The North American gas storage industry uses over 400 reservoirs and 15,000
wells. Tens of millions dollars are spent annually to recover losses of deliverability.
Schlumberger
Hydraulic fracturing and refracturing were shown to be popular and effective
methods to increase deliverability.
represent the optimal approach, if it fails to consider longevity of the treatment.
In carbonate reservoirs, acidizing and fracturing appear to be most effective
and longest lasting treatments currently employed, with fracturing most
Nationwide gas storage
optimal. In sandstones, combination acidizing-hydroblasting-and-perforating
treatment, fracturing, and refracturing are most effective currently. Opportunity
exists to accelerate benefits of stimulation by eliminating or reducing cleanup
effects, and by learning how to reduce underlying cause of damage.

Nor-Azlan N. et al .: “Massive Hydraulic

Fracturing - A Case History in Western
Siberia, Russia," paper SPE 84916 presented
Siberia 84916
at the 2003 SPE International Improved Oil
Recovery Conference, Kuala Lumpur, Oct 20-
21.
Burnstad, R.G., et al.: "A Case Study of a
JOB Pertamina-
Mature Field Redevelopment Using Propped
Talisman Tanjung
Indonesia 88604 Hydraulic Fracturing," paper SPE 88604
presented at the 2004 Asia Pacific Oil and
Gas Conference, Perth, October 18-20.
Initial frac increased production and yielded a negative skin. However, an Massive Hydraulic Fracturing strategy used successfully in Russia. Treatment
Western Siberia,
aggressive refrac with 216 tons 16/30 ceramic was conducted, building net sizes increased to 150 tons or more to effectively stimulate the entire
Schlumberger, Vyngayakhinskoe
pressure to increase fracture width. Skin was reduced to -4.75, and oil production productive intervals and reach desired proppant concentrations. Production oil
Sibneft Field, BP12
increased from 60 tons/day with the initial conservative frac to 150 tons/day as a rates often doubled with larger treatments. Analysis approach described to
formation.
result of larger, more aggressive treatments. select candidates and design treatment.
Although prolific fields are often not considered candidates for restimulation, a Using Nodal analysis, frac treatments were redesigned assisted by mini frac
significant increase in oil productivity from a mature field was attained with propped interpretation from downhole gauges. Optimization allowed treatment sizes to
fractures. Early mis-steps with poor designs and sand flowback caused few fracs be reduced, with costs reduced by 50% despite goals to achieve very high Tanjung Raya Field,
to be attempted between 1960 and 2000. New tip screenout designs reaching 12 concentrations. The use of resin coated sand solved the problem of proppant Barito basin, oil
Ltd and BJ
to 16 ppg of RCS solved conductivity and proppant flowback problems. Fracs in flowback. Candidate selection processes, minifrac and step rate tests, the use Kallimantan
Services
mature reservoir enhanced recovery, better utilized vertical wells as a low cost, of bottom hole gauges, and techniques to recognize and mitigate tortuosity
effective alternative to horizontal wells, and allowed wider well spacing. were key.

Significant improvement to well productivity resulted from a systematic

More than 4000 Codell wells have been remedially stimulated with massive
Miller, J.D., "Fracturing Fluid Viscosity modification of the viscosity profile of the frac fluid. The parameters in the
hydraulic fracturing treatments, with ~1,000 studied for this paper. A majority of
Profiles Determine Well Productivity in a study that had the strongest influence on well productivity were: time to 500 Codell formation,
Patina Oil & Gas the refracturing treatments in this study have yielded higher oil and gas rates than
USA-Colorado 90194 Remedially Stimulated Tight Gas Reservoir," centipoise, peak viscosity, and percentage viscosity decline. Optimizing the Wattenberg Field, oil and gas
and BJ Services achieved following the original treatment. There is evidence that the fracturing-fluid
paper SPE 90194 presented at the 2004 FVP improved proppant placement as evidenced by the ability to control height Colorado
viscosity profile (FVP) has a multi-faceted impact on well productivity that is more
ATCE, Houston, September 28-29. growth and eliminating screen-outs, and possibly due to a significant change in
than just being able to place the proppant into the formation.
fracture azimuth, as evidenced by the elimination of offset well breakthroughs.

The "completion efficiency" concept provides a means to determine whether

Barba, R.E. and Shook, R.A., "Post Frac
Evaluation of Multiple Zone Fracture
USA-Texas 90483 Treatments Using the 'Completion Efficiency' Services, Cimarex frac analysis. The concept of completion efficiency is proposed and described.
Concept," paper SPE 90483 presented at the
2004 ATCE, Houston, September 28-29.
Wiley, C. et al: "Improved Horizontal Well
Stimulations in the Bakken Formation,
USA-Montana 90697 Williston Basin, Montana," paper SPE 90679
presented at the 2004 ATCE, Houston,
September 28-29
sub-optimal production is a function of low pressure, low kh, or an ineffective
fracture treatment. The study results strongly suggest that fracture treating
Simulation suggests that when multiple zones are perforated in a single frac stage
more than one perforated interval in a single fracture stage may result in only
multiple fractures of varying lengths may be created, with the target propped frac
one of the perforated intervals being effectively stimulated. Long perforated
Integrated Energy length not being obtained in each sand. Verification is difficult with traditional post
intervals in wells with >2 degrees deviation should be avoided. Excessive
SE Texas
drawdown pressures should be avoided. The study suggests that if the zone
Corporation Completion efficiency involves estimating the productivity of a zone with a minimum
is to be killed with fluid prior to fracing subsequent stages, extended flowback
acceptable frac geometry and comparing it to actual production.
periods between stages do not improve completion efficiency. In addition to a
post-frac analysis tool to help determine best practices, the productivity models
developed for the process can be useful in determining the optimum
perforation and staging strategy for new wells.
Lyco Energy
A review of the approach taken to evaluate the most effective way to complete and Four main changes in treatment design resulted in a 50% production
Corporation,
fracture laterals. RA tracer show the improvements in proppant distribution along improvement, with reduced well costs: noncemented liners and longer laterals; Williston Basin,
Sleeping Giant
the wellbore caused by changes in fracture treatment design, and its impact on staged treatment and diversion mechanisms (ball sealers); cleaner fracturing Richland County, oil
LLC, Barree &
well productivity. With low stress contrast, refracturing was identified as potential fluid with a more aggressive breaker package; use of more proppant per foot MT
Assoc., and
strategy. (see SPE 108117) of lateral. Sets stage for 108117 refrac paper.
Halliburton

MacDonald, R.J et al : "A Rapid and Efficient
Method to Identify Underperforming Stripper
USA-
95511 Gas and Oil Wells," paper SPE 95511
Pennsylvania
presented at the 2005 ATCE, Dallas, October
9-12
Dept. of Energy, A method has been developed to identify underperforming stripper gas and oil Of 2,200 wells, 339 were identified as underperformers, with 168 economically
Great Lake wells based on comparing various production indicators. An important outcome of viable candidates. 8 wells identified with this methodology were restimulated
Crawford, Venango,
Energy Company, this methodology is that it highlights substandard wells, which may only need low- successfully with modern treatments. Incremental gas is estimated at 615
and Warren oil & gas
Belden & Blake, cost efforts to improve their productivity. For example: wells stimulated with MMscf. Success attributed to recognizing underperforming, viable candidate
Counties, Pa
and Schlumberger nitrogen and little or zero proppant may be restimulated with greater quantities of wells quickly and efficiently. Costs are contained by focusing attention on wells
DCS proppant, enabling production of significant levels of additional gas. that can be made profitable, spending zero time on wells that cannot.

Craig, D.P and Blasingame, T.A.: "A New

Refracture-Candidate Diagnostic Test
Determines Reservoir Properties and
USA 96785
Identifies Existing Conductive or Damaged
Fractures," paper SPE 96785 presented at
the 2005 ATCE, Dallas, October 9-12.
Oil & gas wells often contain potentially productive layers bypassed either
intentionally or inadvertently during an original completion. Refracturing programs
Halliburton, Texas
are sometimes unsuccessful because the programs focus on commingled well
A&M
performance and well restimulation without investigating individual layer properties
and the refracturing potential of individual layers.
A new refracture-candidate diagnostic test is presented that requires a brief
isolated-layer injection at a pressure exceeding the fracture initiation and
propagation pressure followed by an extended shut-in period with the pressure
falloff recorded. This test can be used to find bypassed layers and to help
formulate the proper treatment to successfully refracture the well.

SPE 134330 Appendix Page 10 of 16 Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference
Craig, D.P.: "Analytical Modeling of a
Fracture-Injection/Falloff Sequence and the
Dissertation,
Global Development of a Refracture-Candidate
TAMU
Diagnostic Test" Dissertation submitted to
Texas A&M, May 2006.
Pagano, T.A.: "Rock Properties, Deliverability
USA-Colorado OGJ 2006 Mechanism Influence Codell Restimulation." Kerr-McGee Corp.
Oil & Gas Journal (July 10, 2006) pp 44-47
E&P (Nov. East, L. and Craig, D. "Refracs reverse
Global
2006) production dips," E&P (Nov. 2006) pp 45-47
Shaefer, M.: “Awakening an Old Field - A
Case Study of a Refracturing Program in the
Greater Green River Basin," paper SPE
USA-Wyoming 101026
101026 presented at the 2006 Annual
Technical Conference, San Antonio, Sept. 24-
27.
Green, D. et al : "Hydraulic Fracturing of
Miocene and Oligocene Sandstones in the
Taranaki Basin, New Zealand," paper SPE
New Zealand 101121
101121 presented at the 2006 Asia Pacific
Oil & Gas Conference, Adelaide, September
11-13.
Dedurin, A.V et al: "Designing Hydraulic
Fractures in Russian Oil and Gas Fields to
Accommodate Non-Darcy and Multiphase
TNK-BP, SIAM,
Russia 101821 Flow - Theory and Field Examples," paper
CARBO Ceramics
SPE 101821 presented at the 2006 SPE
Russian Oil and Gas Technical Conference,
Moscow, October 3-4.
Moore, L.P and Ramakrishnan, H.:
"Restimulation: Candidate Selection
USA, China 102681 Methodologies and Treatment Optimization,"
paper SPE 102681 presented at the 2006
ATCE, San Antonio, September 24-27.
Ketter, A.A., et al: "A Field Study Optimizing
Completion Strategies for Fracture Initiation
USA-Texas 103232 in Barnett Shale Horizontal Wells, " paper
SPE 103232 presented at the 2006 ATCE,
San Antonio, September 24-27.
Li, P., Song, Z. and Wu, Z.: "Study on
Reorientation Mechanism of Refracturing in
Ordos Basin - A Case Study: Chang 6
Formation, Yanchang Group, Triassic System
China 104260
in Wangyao Section of Ansai Oil Field," paper
SPE 104260 presented at the 2006
International Oil & Gas Conference , Beijing,
December 5-7.
JPT Online, March 2006: Techbits:
Refracturing in Low-Permeability Reservoirs.
USA - Colorado
JPT Online SPE editorial summary of presentation by
and New
March 2006 Devin Brown entitled “Refracturing in the
Mexico
Fruitland Coal” presented at ATW, San
Antonio.
SPE 134330 Appendix Page 11 of 16
Companies Refrac Findings Reason for Refrac Success or Failure Location Type
To identify bypassed pay in multilayer completions, Craig developed pressure
transient solutions to evaluate refracture candidates, with short injections above Significant evidence of bypassed pay is presented in dissertation. Diagnostic
Texas A&M
frac gradient into prospective layers, analyzing both pre- and post-closure technique can successfully identify intervals that were poorly stimulated and
University
behavior. Variable storage behavior can identify pre-existing fractures and have adequate remaining pressure for future refracturing.
characterize damage to the initial fracture.
This study argued that the Codell would respond "unconditionally" to repeated The pseudo microporous reservoir's deliverability depends on the number of
Codell formation,
refracture treatments, often restoring productivity to virgin conditions. Contrary to open and effective microporosity features that network through the matrix. The gas +
Wattenberg Field,
other papers, this study concluded that refracs do not contact new rock in the microporosity will close upon depletion, and a refrac can re-engage the closed condensate
Colorado
Codell, but rather reengage previously drained reservoir. and potentially damaged microporosity, renewing deliverability.
Enhanced performance can be obtained through multiple interval stimulation
This article discusses why multi-layer wells can require re-stimulation, the mistakes
technologies and new refrac-candidate diagnostics. Past practices or mistakes
Halliburton that have been made in the past, the technologies that are producing good results,
that result in poor production often can be remedied by properly designed
and how to indentify good candidates for restimulation.
refracture treatments.
Four gas wells initially stimulated with Ottawa sand exhibited disappointing fracture Multiple diagnostics used to identify poor fracture performance. Fluid system
length and conductivity based on production analyses. Wells were restimulated and proppant selection affect the success of stimulation treatments. Fracture
BJ Services, Wyoming, GGRB,
with ceramic proppant, resulting in greater apparent frac lengths and production Impulse Tests (FIT) can give valuable insight into the type of proppants
BERCO Fontenelle Field, gas
rates. Gas increased 250% to 1500% with 10-month payout, not considering NPV necessary as well as the type of fluid system that should be used. In these
Resources 2nd Frontier.
of increased reserves. Despite 17 years of depletion, refracs restored production wells, crosslinked gel fluid carried more proppant farther, and superior
to near-original rates. proppants enhanced conductivity.
Hydraulic fracture stimulation and restimulation of Kauri and Tariki sands can be In an area where hydraulic fracturing was rare, poorly planned fracture
successful and economically achieved. Wells with little or no pre-fracture treatments produced mixed results. Water-based fluids contributed to poor
production will generally have little or no post-fracture production - good wells performance, as was near wellbore tortuosity. Productive treatments were
Swift Energy, BJ Taranski, New
make good fracture candidates. In good candidates, in general, larger treatments created using mini-frac and step rate tests to garner quality information. In oil & gas
Services Zealand
yielded superior production. Larger proppant grain size increased production. some successful refracs, the minifrac showed differing friction losses, closure
Proppant flowback was alleviated by use of deformable particles added to the last stresses, and leakoff compared to initial treatment. Proper fluid and proppant
50%. Oil-based fluids outperformed water-based fluids. selection and better treatment design produced successful stimulations.
When non-Darcy and multiphase flow effects are minimized, a corresponding
production increase is realized. The authors note that while three ceramic Achimovskoya
In the Orenburg region, initial fracs used 16/30 ISC. Refracs with higher
proppants by three manufacturers are considered interchangeable by some, formation, Tomsk,
concentrations of a larger diameter 12/18 ISC provided a typical production
tests show that similarly sized products can have significantly different Western Siberia and gas & oil
increase of 73% or 37 tonnes per day incremental oil. Despite 15% depletion in
strengths and conductivity. Higher proppant concentrations are shown to yield Orenburg, Volga-
reservoir pressure, refracs provided large benefits.
an extra 27% of NPV after 1 year. Stimulation treatments were designed to Urals
improve fracture width, and incorporate highly permeable proppant.
For successful restimulations, proper candidate well selection and proper
treatment design are key. Areas with restimulation potential have common
Production data alone may not offer an effective restimulation candidate selection traits, which will help to identify candidates that will respond to resimulation.
methodology. Other parameters such as remaining reservoir energy, recoverable First pass evaluation using moving domain analysis or neural network type
Schlumberger various gas & oil
reserves, and favorable response to original fracture jobs are important. Case techniques using existing data can help prioritize wells for in-depth analysis.
histories in Barnett, Vicksburg, Codell, Red Oak, and China are discussed. Treatment parameters such as total fluid volume, fluid type, and proppant
volume and proppant type have a major impact on the outcome of any
restimulation.
A study of 256 horizontal Barnett shale wells was conducted to identify the causes Image logs can be used to prevent inefficient fracture initiation by selectively
of near-wellbore issues and to recommend an optimized completion strategy to locating perforations. Reducing the number of clusters per stage minimizes
minimize problems, increase stimulation coverage, and decrease unplanned stress interference. Proper perforation orientation is discussed. Perforation Fort Worth basin,
Schlumberger,
completion expenses. Using optimized strategies has reduced fracture initiation length should be no greater than four wellbore diameters. Using ASC and/or Denton, Wise and gas
Devon Energy
difficulties by 74%. Barnett should not be considered to provide a homogeneous shorter perforation clusters reduces horsepower required to stimulate the wells. Tarrant counties, TX
stress environment throughout lateral length, with significant implications as to Adding 100 mesh sand to the pad stage controls leakoff. Initiating fractures
fracture diversion and refrac potential. with crosslinked gel provides sufficient width for placement of proppant.
Changquing
Although mathematical models predicted minimal reorientation due to stress field We can design fracture treatments and implement successful fracture
Petroleum
change, oriented perfs, TSO and wax diverters were expected to reach criteria for reorientation to tap into lateral "undrained regions" and enlarge the drainage
Exploration
reorientation. With sufficient net pressure, potential to open orthogonal natural area to enhance oil production using Tip Screen Out (TSO) with Temporary Ordos basin,
Bureau, Institute oil & gas
fissures may be met. A well producing 1.45 tonne/day of oil improved to 3.97 t/d Blocking Treatment (TBT). For refracture treatments, with the initial hydraulic northeast China
of Applied
following restimulation. Tiltmeter mapping did indicate reorientation, which can be fracture temporarily blocked with wax beads, multi-fractures and new fractures
Mathematics and
undesirable in a waterflood. can be achieved through oriented perforations.
Mechanics
From 1998 to 2006, 138 refracture treatments were performed, many incorporating
Although candidate selection and use of SMA to control coal fines was touted Fruitland CBM, San
surface modification agents in an effort to mitigate migration of coal fines. In 2001,
as evidence of superior refrac design, substantially higher production Juan Basin, coal-bed
BJ, BP the average production improvement was 130 mscfd. In 2004, due to improved
increases (1400 and 800 mcfd increases) were previously documented with Colorado and New methane
candidate selection and improved treatment design, the average production
refracs in the Fruitland Coal (SPE 77675) Mexico
improvement increased to 350 mscfd.
Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference
Malik, K., Mohaghegh, S. and Gaskari, R.:
"An Intelligent Portfolio-Management
Approach to Gas Storage Field Deliverability
Maintenance and Enhancement: Part One -
Global 104571
Database Development and Model Building,"
paper SPE 104571 presented at the 2006
Eastern Regional Meeting, Canton, October
11-13.
Fairhurst, D.L., et al,: "Fracturing Previously
Bypassed Highly Laminated Tight Gas
Sands: A Production Optimization Case
USA - Texas 105681 Study in South Texas," paper SPE 105681
presented at the 2007 Hydraulic Fracturing
Technology Conference, College Station,
January 29-31.
Li, P.: "Theoretical Study on Reorientation
Global 105724 Mechanism of Hydraulic Fractures," paper
SPE 105724.
Weng, X. and Siebrits, E.: "Effect of
Production-Induced Stress Field on
Refracture Propagation and Pressure
Global 106043
Response," paper SPE 106043 presented at
the 2007 Hydraulic Fracturing Technology
Conference, College Station, January 29-31.
Surjaatmadja, J.: "The Mythical Second
Fracture and Its Optimal Placement for
Global 106046 Maximizing Production, " paper SPE 106046
presented at the 2007 Europec/EAGE Annual
Conference, London, June 11-14.
Gil, I.R., et al,: "Super Height Growth -
Strange Surface Pressures and Refrac
USA-Texas 106083 Potential," paper SPE 106083 presented at
the 2007 Hydraulic Fracturing Technology
Conference, College Station, January 29-31.
Wang, S.G., et al,: "Case Studies of Propped
Refracture Reorientation in the Daqing Oil
China 106140 Field," paper SPE 106140 presented at the
2007 Hydraulic Fracturing Technology
Conference, College Station, January 29-31.
Ingram, S.R., et al,: "Remediation of
Production Loss Die to Proppant Flowback
Through Coiled-Tubing Intervention," paper Halliburton, XTO
Global 106532
SPE 106532 presented at the 2007
Production and Operations Symposium,
Oklahoma City, March 31-April 3.
Yao, F. et al: "Reorientation Refracturing
Case Study," paper SPE 106595 presented
China 106595 at the 2007 Production and Operations
Symposium held in Oklahoma City, March 31
- April 3.
Phillips, Z.D. et al: "A Case Study in the
Bakken Formation: Changes to Hydraulic
Fracture Stimulation Treatments Result in Hess Corporation,
USA-North Improved Oil Production and Reduced
108045
Dakota Treatment Costs," paper SPE 108045
presented at the 2007 Rocky Mountain Oil & Bass Enterprises
Gas Technology Symposium, Denver, April
16-18.
SPE 134330 Appendix Page 12 of 16
Companies Refrac Findings Reason for Refrac Success or Failure Location Type
Re-stimulation candidate well selection can be economical with new software being
developed, which uses well deliverability parameters as well as change in skin.
West Virginia The new methodology uses different trained neural networks and is very flexible to Better, more economical candidate selection should allow superior refrac
gas storage
University design genetic algorithms based on different parameters, so that it can be used performance.
successfully in differing fields. This software combines data editing, well-test
analysis and artificial intelligence in one software package.
Image logs show a more complex reservoir environment than conventional
logging tools. Can be used to increase the accuracy of log based volumetric
Recent advances in logging technology can discern gas bearing permeable sands
Schlumberger, evaluation. A pseudo three-dimensional fracture design model is used in
from clay dispersed in silt and sand for accurate productivity indication. Production LaSalle County,
Escondido conjunction with the production model to optimize the stimulation treatment gas
data from newer wells with the high-resolution analysis were compared to those South Texas
Resources and job design parameters, identifying additional pay and developing
with standard analysis, allowing identification of bypassed pay opportunities.
previously bypassed resources effectively and economically, adding to the
reserve base.
Tip screen out fracturing can build large net pressures. When combined with
Fracture reorientation is not only achieved with geo-stress alteration due to
Shanghai diversion or blocking treatments, refracs in the Ansai oilfield indicated
production/injection or fracture placement. If proper fracturing technology or
University reorientation. With more accurate rock mechanic and geo-stress parameters
methods are adopted, fracture reorientation can be induced.
for specific wells and layers, this theoretical analysis will become more reliable.
Unlike the initial fracture treatment in an unperturbed reservoir where in-situ stress Once the in-plane fracture starts to propagate along the initial fracture path, it
is uniform, a refracture treatment creates a hydraulic fracture(s) that propagates in competes against the orthogonal fracture for fluid. To predict the propagation
a non-uniform stress field due to the induced stress changes. While refracs may and the corresponding treating pressure for the two intersecting fractures, a
Schlumberger initiate in new locations/directions, the propagation pressure of the orthogonal dual-fracture model that can handle stress variation along the fracture path is
fracture should increase resulting in reopening of the initial fracture. If true, a much needed. A simple PKN-type 2D model is developed in this study and is
larger fluid volume than the initial fracture treatment may be required for the presented. Refracs need to be focused to either improve conductivity or
refracture to penetrate beyond the initial frac length. extend frac length.
Multiple, consecutive fractures can be created in a way that enables them to reach By creating more fractures, it is more likely that productive pockets will be
formations in a manner not reached before while using conventional means. intersected. To consistently intersect highly permeable pockets, the preferred
Hydrojet and immediate refrac can take advantage of temporary stress alteration maximum direction must be modified. A second fracture is quickly initiated
Halliburton and induce a refrac into a new plane. With 30-60 minutes between stages, frac 2 after the first, to take advantage of the stress modification created from the
was nearly horizontal, and proppant strength was insufficient to support higher first fracture, allowing the second fracture to reach more productive rock not
overburden. Microseismic data indicated the 2nd and 3rd fracs "tried" to initiate accessible to the first. In this way we harness a natural phenomenon to our
perpendicular to preceding fractures. advantage.
A well that exhibits a pressure "bump" after injecting only water or linear gel is
In presence of sand/shale stress contrast, low viscosity frac fluids provide the
almost certainly a candidate for a re-frac, since super height growth probably exists
best frac length and containment results. Under low stress gradient regime,
NSI Technologies, (height growth upwards with simultaneous closure at bottom and no further length
and in absence of sand/shale stress contrast, low viscosity fluids are most Wilcox and
Devon Energy development). How to recognize and capitalize on this situation is discussed. Case
efficient choice. For steep stress gradient, provided no sand/shale stress Vicksburg fields, gas
Corp., Chevron, histories are presented. Analysis to find minimum degree of depletion needed for
differences exist, moderate viscosity fluids are best alternative. Very high South Texas
BP effective fracture containment is described. These results were used to select
viscosity fluids can be used to increase fracture width at the expense of
candidates for re-frac, which were successful. Some wells showed dramatic and
penetration.
sustained production increase, while others had more modest gains.
Chinese Academy
of Sciences The paraffin balls worked to temporarily plug original fractures and encourage
The paper examines cases of re-fracture treatments using a new technique
Guangzhou Inst. new reoriented fracturing. Surface tiltmapping verified fracture reorientation
whereby paraffin balls are pumped half way through the propped treatment, Daqing field,
Of Geochemistry, and the presence of both vertical and horizontal fractures. Volume percent for
followed by a brief shut in to allow fractures to close and subsequently to initiate Heilongjiang oil
Daqing Oilfield the horizontal fractures was 26%-67%. The well with the least post-fracture
new or reoriented fractures. The wells in this study indicate tiltmeter reorientation Province, China
Co. Ltd., Pinnacle reorientation showed the least post-fracture production increase. Reorientation
and show production increases from 117% to 288%.
Technologies, can add significant economics for mature field redevelopment.
CARBO Ceramics
A combination of coiled tubing, pressure pulsing tool, and low-viscosity
The treatment chemicals are place precisely into propped fractures to form a
consolidating agent (liquid curable resin) provides a reliable and economical
consolidated, highly permeable pack that can withstand high drawdown,
remedial treatment to overcome proppant flowback problems, without mechanical various
Energy enabling higher rate production. The objective is to treat the proppant near the
isolation, to eliminate subsequent cleanouts and improve the operator's return on
wellbore, not the entire proppant pack.
investment by saving costs.
China U. of
This paper discusses stress field prediction software, and the process of
Petroleum, By reorienting during refractures, we can decrease water production and Xinzhan Oilfield,
reorienting fractures. Monitoring results show that reorientation did occur, oil
RIPED, Daqing increase oil production, improving oilfield recovery. China
validating the software.
Oilfield Co.
To optimize treatment design an understanding of the created fracture
geometry and intersection with the wellbore is necessary. If fractures are
In the Bakken, refracs have been successful with a variety of fluids, proppant longitudinal with excellent wellbore intersection, proppant selection is less
Pinnacle types, and isolation techniques. In this paper, slickwater refracs of horizontal wells crucial. If fractures are transverse, proppant characteristics are key. Williston Basin,
oil
Technologies, previously stimulated with ceramic proppant suspended in crosslinked gel provided Best results (30% production increase with 40% cost decrease) were found North Dakota
increases in 30-day production. with lower-viscosity fluids in higher volume; lower proppant concentrations;
with an increase in treatment rate. It was noted that the one well fraced in the
new method with ceramic proppant was the best producer.
Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference
Lantz, T., et al: "Refracture Treatments
Proving Successful in Horizontal Bakken
Wells: Richland County, Montana," paper
USA-Montana 108117
SPE 108117 presented at the 2007 Rocky
Mountain Oil & Gas Technology Symposium,
Denver, April 16-18.
Cipolla, C.L.: "The Truth about Hydraulic
Fracturing-It's More Complicated Than We
Global 108817
Would Like to Admit." SPE 108817.
Distinguished Lecturer Series 2005-2006
Wolhart, S., et al: "Surface Tiltmeter Mapping
Shows Hydraulic Fracture Reorientation in
USA-Colorado 110034 the Codell Formation, Wattenberg Field,
Colorado," paper SPE 110034 presented at
the 2007 ATCE, Anaheim, November 11-14.
Quirein, J.A., et al: "Microseismic Monitoring Energy Services,
of a Restimulation Treatment to a Permian
USA-Texas 110333 Basin San Andres Dolomite Horizontal Well,"
paper SPE 110333 presented at the 2007 Consultants, OXY
ATCE, Anaheim, November 11-14.
Besler, Monte, R., et al: "Improving Well
Productivity and Profitability in the Bakken - A
USA-
summary of Our Experiences Drilling,
N, Dakota, 110679
Stimulating, and Operating Horizontal Wells,"
Montana
paper SPE 110679 presented at the 2007
ATCE, Anaheim, November 11-14.
Leonard, R., et al: "Barnett Shale
Completions: A Method for Assessing New
ProTechnics and
USA-Texas 110809 Completions Strategies," paper SPE 110809
Range Resources
presented at the 2007 ATCE, Anaheim,
November 11-14.
Edwards, J.T., and Crafton, J.W.: "CO2
Restimulation in the Michigan Basin: A
USA-Michigan 111179 Second Chance for the Prairie Du Chien"
paper 111179 presented at the 2007 Eastern
Regional Meeting, Lexington, October 17-19.
Myers, R., et al,: "The Oriskany Sandstone of
Guernsey County, Ohio: A Study of High
Surface Treating Pressure Response During
USA-Ohio 111210
Hydraulic Fracturing," paper SPE 111210
presented at the 2007 Eastern Regional
Meeting, Lexington, October 17-19.
Cadwallader, R.: "New technology,
FWBO&G (Nov. innovations continue to improve fracturing
USA-Texas
2008) process." Ft. Worth Basin Oil & Gas
(November 2008) pp 4-9
"StimMORE Service Increases EUR in
USA-Texas SLB 08-ST-054 Barnett Shale Well by 0.25 Bcf"
www.slb.com/reservoircontact
SPE 134330 Appendix Page 13 of 16
Companies Refrac Findings Reason for Refrac Success or Failure Location Type
Sixteen horizontal wells originally treated with ~300,000 lb RCS were restimulated, Success attributed to effectively stimulating longitudinal intervals missed in
ProTechnics,
with perforations added in areas apparently not contacted with the original original treatments. Benefit is through placing more perforations and in the
Enerplus Spring Lake/Elm
treatments. Treating pressures and GORs indicated contact of previously use of diversion techniques. Performance analysis and offset comparison are
Resources, Coulee, Richland oil
undrained rock, and post refrac rates often exceeded IP. Refracs added 1.3 of increased importance in candidate selection, although tracer logs remain a
Sleeping Giant County, Montana
MMBO additional reserves. One of 16 refracs screened out, and became best primary tool in well evaluation. Elapsed time or cumulative production did not
LLC, Halliburton
producer. correlate with success. Reserves added at <$5/bbl.
An overview of fracture complexity is given, and a brief introduction to fracture Prudent application of fracture diagnostic technologies reduces uncertainty.
diagnostic technologies. Limitations of fracture modeling analyses and the need Utilizing multiple diagnostic technologies is many times the key to success in
Pinnacle
for calibrated fracture models is discussed. Case histories including refracturing complex environments. By identifying fracture complexities one can various
Technologies
are presented to illustrate fracture complexity and some solutions that overcame understand well performance, and enhance design strategies by selecting
those problems. proper fluid, proppant, stages, and well placement.
Successful refracturing of Codell wells in Wattenberg Field has been attributed to Refracture reorientation allows existing wells to contact new reservoir
Noble Energy,
several reasons including reorientation of fractures. Tiltmeter mapping shows increasing gas recovery per well. Improved completion and stimulation Wattenberg Field,
Anadarko gas
significant reorientation, but production data shows no obvious correlation between practices, fracture fluids and treatment designs lead to success. The potential Colorado
Petroleum
degree of reorientation and production. exists for a third fracture for additional reserve recovery.
Halliburton
Few microseismic events were observed during restimulation of a 3000 ft openhole Operationally, orientation and velocity calibration from ball drops is West Welch Unit,
Applied lateral, perhaps due to previous stimulation, excessive open area for fracturing, or economically superior to using stringshots. Data reprocessed with a calibrated Midland Basin,
Seismology oil
attenuation of signal by highly fractured dolomite. Tools could observe ball layered velocity model show a more defined pattern of transverse and complex Dawson County,
activated sliding sleeves. fracture growth. Texas
USA, University of
Toronto
In transverse fracs all efforts to improve fluid flow via wider fractures, cleaner
Optimizing fracture treatment designs for horizontal wells requires an estimation of
fluids, better proppants or superior implementation provide increased rate. In
Hohn the fracture geometry. Definitive fracture mapping is important. Longitudinal and
completions using ceramic proppant, the authors have not found it necessary
Engineering, transverse oriented fractures require very different designs. Correctly designed Williston Basin,
to restimulate many wells, as there is minimal proppant degradation or
Nance Petroleum, multiple transverse fracs exhibit superior reservoir contact and increased North Dakota and oil
flowback. Ceramic proppant provides substantially higher production with
Pinnacle productivity and recovery, but require improved conductivity due to limited Montana
smaller treatments, and produces higher EUR.
Technologies intersection between wellbore and fracture.
The authors found superior production from wellbores oriented NW-SE,
allowing creation of multiple transverse fracs.
Radioactive particles and chemical fluid tracers were used on four sets of
Use of particulate tracers in the proppant and liquid chemical tracers in the frac
simultaneously-fractured wells to identify key completion parameters. Larger
fluid allow a superior understanding of: 1) quality of lateral isolation (cement or
proppant volumes and more effective horizontal stage isolation appear to be
packer effectiveness) 2) effectiveness of perforating strategy 3) offset well Fort Worth Basin,
advantageous. Simultaneously-fracturing wells could enhance the degree and
interference (frac fluid and/or tracer in offset wellbores or at surface). Despite RR Commission oil
uniformity of interwell fracturing and expand the fracture network area. The
significant interwell communication between 500 to 1000 ft offsetting Newark, East, Texas
simultaneously-fractured Barnett wells appear to be yielding favorable post- wellbores, simulfracs appear to be yielding favorable production, which may
completion well performance. Discussion of re-entry candidates using have implications to refracs.
appropriately phased perforations.
Initial fracs were successful, but production had dropped. Earlier attempts to refrac Advantages of CO2 treatment include increased fluid recovery, hydrostatic
with XLG or N2 foam provided mixed results. Four refracs with CO2 were head increase due to fluid density, and solubility with aqueous fluids for
BJ Services, performed, increasing well NPV by 50%. The CO2 stimulations showed increased reduced surface tension, while avoiding many problems seen in this field with Prairie Du Chien,
Performance effective fracture length, and increased pressure/drainable area. It is also likely that cross linked fluid or nitrogen foam treatments. The major issues are product Michigan Basin, gas
Sciences additional pay was connected, as initial treatments probably failed to connect up to availability, equipment capabilities, pricing and operational safety concerns. Michigan
200 feet of vertical pay. Production rates increased in all four wells and recoverable Refracs appeared to either add pay or contact new rock due to improved
reserves increased in three wells. reservoir pressure or drainage area post-refrac.
Possible explanations for the pressure behavior include: slurry dehydration
The Oriskany sandstone is one of the toughest sands to drill and fracture stimulate,
BJ Services, combined with narrower than normal fracture apertures causing bridging in the
and many treatments at the Guernsey Storage Field are actually refracs. This
Columbia Gas fracture system and/or the minimum horizontal in-situ stress gradient is high for
paper explores the high surface pressure responses seen during both original frac Guernsey Co., Ohio gas storage
Transmission reasons not yet known. Fracture diagnostics to evaluate the near wellbore and
jobs and refracturing treatments, for the purpose of improving future stimulation
Corp. far-field stress fields, such as tracers and microseismic may prove useful in
designs. End of job ISIP gradients can exceed 1.6 psi/ft.
solving the fracture geometry puzzle, but refracs are adding value.
New technologies allow selective multi-zone operations through the production
Operators constantly search for ways to squeeze the most out of their investments. string. Simul-fracing can improve production because fracturing two wells
Innovation comes from challenging the difficulties of fracing and pursuing better close together increases the pressure on each bore, which can alter fracture
efficiency. Newer, multi-stage mechanical devices allow rapid completion. growth. Some operators use fiber optic to monitor wellbore temperatures
Technology continues to improve refracturing tools and techniques. One (DTS) to collect real-time data on fluid diversion in wellbore. Surfactant Barnett Shale gas
constant challenge in horizontal wells is effective isolation to allow restimulation of polymers extend in the presence of water, which impedes water flow, and
desired intervals. Diverting agents including fibers, salt, and starch, plus temporary contract in hydrocarbons, which encourages gas flow. Scale inhibitors block
shut downs have been attempted. the formation of barium sulfate buildup. Hybrid jobs incorporating both gelled
and slickwater fluids may combine the advantages of each fluid system.
Underproducing sections can be identified. Real-time mapping evaluated the
Microseismic monitoring shows large increase in microseismic activity in Barnett
use of diversion stages to induce fracture initiation. Although much of the
Schlumberger refrac compared to initial treatment. Daily production rates increased threefold, Barnett Shale gas
refrac appeared to restimulate the original fracture, approximately 25%
and EUR was improved by 0.25 BCF.
improvement in lateral coverage was achieved.
Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference
Liu, H., et al: "Evaluation of Refracture
Reorientation in Both Laboratory and Field
China 112445 Scales," paper SPE 112445 presented at the
2008 International Symposium on Formation
Damage Control, Lafayette, February 13-15.
Surjaatmadja, J.: "Placing Two Fractures
Consecutively in Close Proximity to
Significantly Increase Revenue -to-Cost
USA 114600
Ratio," paper SPE 114600 presented at the
2008 Asia Pacific Oil & Gas Conference,
Perth, October 20-22.
Craig, D.P. and Odegard, C.E.: "Identifying
Bypassed and Ineffectively Stimulated Layers
in a Well with Commingled Production from
USA-Colorado 114777 Multiple Layers: Mesaverde Case History,"
paper SPE 114777 presented at the 2008
Unconventional Reservoirs Conference,
Keystone, February 10-12.
Lolon, E.P., et al,: "Integrated Fracture and
Production Modeling Study in the Lower
USA-Louisiana 115467 Cotton Valley Sands, Northern Louisiana,"
paper SPE 115467 presented at the 2008
ATCE, Denver, September 21-24.
Palisch, T.T., et al: "Slickwater Fracturing -
Food for Thought," paper SPE 115766
Global 115766
presented at the 2008 ATCE, Colorado,
September 21-24.
Cipolla, C.L., et al: "Hydraulic Fracture
Complexity: Diagnosis, Remediation, and
Global 115771 Exploitation," paper SPE 115771 presented
at the 2008 Asia Pacific Oil & Gas
Conference, Perth, October 20-22.
Waters, G., et al,: Utilization of Real Time
Microseismic Monitoring and Hydraulic
Fracture Diversion Technology in the
USA-Texas OTC 20268 Completion of Barnett Shale Horizontal
Wells," paper OTC 20268 presented at the
2009 Offshore Technology Conference,
Houston, May 4-7.
Sparkman, D., et al: "Real-Time Monitoring
USA 'Steers' Fractures." The American Oil & Gas
Reporter (December 2009) pp95-99
Dunek, K.L., et al: "Far-Field Volumetric
Distribution of Fracturing Fluids Away From
an Uncemented Horizontal Liner in the
USA-Montana 115826 Bakken Formation," paper SPE 115826
presented at the 2009 Rocky Mountain
Petroleum Technology Conference, Denver,
April 14-15.
SPE 134330 Appendix Page 14 of 16
Companies Refrac Findings Reason for Refrac Success or Failure Location Type
Lab studies indicate multiple or T-shaped fractures, and further confirm the
Refracturing has proved to be the most effective means of increasing oil recovery impact of cement quality and fracture fluid properties on fracture reorientation.
PetroChina from the Daqing oilfield. In shallower intervals, tiltmeter mapping suggests both Refracs would reopen previous fissures unless differential stress >3MPa,
Daqing Oilfield, horizontal and vertical fractures were created during restimulation. In deeper unless temporary plugging agents were incorporated. Paraffin balls were
Daqing Field, China oil
Pinnacle intervals, refracs appear predominantly vertical. Reorientation has been noted for injected followed by a shut-in to allow fractures to close before restarting
Technologies some refracs, which must be considered in waterflooded fields. All four wells in injection. The tightest well may not have been shut in long enough for
this study showed post refracture production increases between 117% to 288%. complete closure, and minimal reorientation and smallest production benefit
were observed.
Fractures caused by stimulation will likely propagate away from high-permeability
Repressurization of the wellbore may preferentially reopen the first fracture.
areas, but achieve significant production improvement because of the large flow-
Therefore, initiating a second fracture must be done before plastic creep from
path increase created. Placing two fractures consecutively in close proximity to
Halliburton the first treatment dissipates, with time delays of less than 30 minutes. simulation
each other may offer a tremendous benefit by taking advantage of the temporary
Simulations show that production from the second fracture could reach 4-5
stress modification caused by the opening of the first fracture so that the second
times the production of the first fracture.
fracture will extend into the highly permeable areas.
Reservoir It is difficult to effectively stimulate all layers in the Piceance Mesaverde. High- Isolated-layer refracture-candidate diagnostics confirmed microseismic maps
Development angle slant hole wells and single-entry-point fracture treatments were not effective. indicating bypassed layers. Production logs are not reliably used without other
Piceance Basin,
Consulting, Conductivity of the connection between stages varies widely. pressure diagnostics or multilayer modeling to identify refracture-candidate gas
Western Colorado
Williams A successful evaluation method could lead to a better plan to develop existing layers. Nitrogen fracture-injection/falloff tests are effective in evaluating
Production RMT reservoirs that are producing below potential and significantly increase the EUR. refracture-candidate layers.
Wells in this study show evidence of damaged hydraulic fractures. Potential Recommended procedures include: Keep wells on small surface choke for as
Pinnacle
damage mechanisms include high production drawdown and large effective long as possible to delay potential damage. Pump more, smaller, fracture
Technologies, J- Lower Cotton Valley,
stresses on the proppant in conjunction with high temperature, fracture complexity, stages and stimulate major sand separately to avoid inadequate proppant
W Operating Bienville Parish, gas
dipping fractures, and formation collapse. Re-fracs in these wells have mixed coverage which will increase risk of fracture conductivity damage and poor well
Company, EOG Louisiana
results. Possible causes and solutions are discussed. Exact form and location of performance. "Partial" flowbacks between stages may lead to fracture face
Resources
damage is not clear at this point. Suspected causes are listed. damage; use either long-term flowbacks or no inter-stage flowbacks.
CARBO Although proppantless injections are beneficial in some specific formations, in most Benefits of slickwater treatments include reduced gel damage, lower costs,
Ceramics, Insight published examples, refracs carrying proppant will improve production and higher fracture network complexity, improved height containment,
Consulting, longevity of the stimulation. Paper discusses differing design principles with environmental advantages. Concerns include reduced transport of proppant, multiple gas
Denbury crosslinked and slickwater treatments. Field data indicates that increasing the large amounts of water required, potential for settled banks failing to cover
Resources conductivity in slickwater fracs economically increases production. the desired vertical pay interval.
Reviews Barnett slickwater refrac following crosslinked gel treatment. Larger
Using techniques to diagnose fracture complexity leads to the selection of
fracture network (greater stimulated reservoir volume) correlated with superior
appropriate remediation or exploitation measures.
production. In Canyon Sands, prior pressure depletion appeared to alter frac
Fracture initiation and proppant slug strategies can be used to control fracture
Pinnacle growth and productivity. In North Sea, efforts to reduce near wellbore complexity
complexity. Well placement may affect fracture complexity and infill drilling gas
Technologies by displacing viscous gel to perfs prior to injection and use of 100 mesh sand
strategies. In some reservoirs, simply understanding the nature of the fracture
improved subsequent proppant placement. Describes role of fracture complexity in
complexity is the key to success, allowing proper stimulation design using a
treatment design. This paper focuses on techniques to diagnose fracture
calibrated fracture model.
complexity and the selection of appropriate remediation or exploitation measures.
Similar example in SPE 119636- refrac in Barnett used additional perforations,
larger proppant volume, variable injection rates, fibers, intermittent shutdown, and Real-time microseismic monitoring allows for real-time adjustments and
real time microseismic acquisition to induce frac diversion. Pumping pressures and corrections to the stimulation. Degradable diverting materials allow the fracture
Schlumberger, Fort Worth basin,
fracture mapping did indicate diversion. Production following restimulations was to be diverted into untouched reservoir, and also away from geohazards. The gas
Devon Energy Barnett shale, Texas
400% greater, and has been sustained over a long period of time, resulting in a recovery factor is increased dramatically when the hydraulic fractures more
significant increase in EUR. These techniques were proven in initial completions densely contact the reservoir.
and recompletions, and should work in other low perm reservoirs.
Microseismic monitoring has proven to be an effective way to see whether By precisely planting the seismic monitoring equipment, rich data sets can be
fractures propagate according to plan, and to identify untreated reservoir sections acquired. Processing the data in real time, allows "on the fly" decisions to
that can be targeted on subsequent refracs. Coupled with diversion technology, improve fracture quality and reservoir contact. New techniques have been
fractures can be "steered" to contact specific reservoir volumes, allowing enhanced by microseismic monitoring, including "zipper-frac" which creates a
Devon Energy, Barnett Shale, North-
numerous, closely spaced fractures for improved productivity. In the Barnett Shale, complementary stress field around the well just stimulated, preventing the well gas
Schlumberger Central Texas
these methods were used to prevent fractures from propagating into the wet being stimulated from intersecting previous fractures, resulting in thorough
Ellenberger. Production wells have been restimulated using these technologies to reservoir contact with minimal interwell contamination. Degradable diversion
perforate between the previously treated zones, resulting in a 400% improvement media allows temporary bridging and also prevents proppant from "slumping"
in production. More detail can be found in OTC 20268. to the bottom of the fracture which ensures uniform distribution.
Surface tiltmeter mapping recorded the growth of fractures from a Bakken Reduced fluid viscosity, higher pump rate, and stress shadowing associated
horizontal well during initial crosslinked stages, showing transverse, longitudinal, with the higher total volume are considered causes for substantially more total
Marathon Oil, and horizontal components developing over time. Six weeks later the well was reservoir treatment in second treatment over first. Changes in design resulted Williston Basin,
oil
Pinnacle refrac'ed with slickwater, with increased development of transverse fractures near in cost savings, greatly reduced mechanical risk, and accelerated cycle times. Montana
the heel. Knowledge of fracture growth allows superior fracture design and Open hole waterfracs performed similarly or better than previous cross-linked
identification of refrac opportunities. treatments with mechanical diversion.
Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
 Region Paper Author - Reference
Benedict, D.S. and Miskimins, J.L., "Analysis
of Reserve Recovery Potential from Hydraulic
Fracture Reorientation in Tight Gas Lenticular
Global 119355
Reservoirs," paper SPE 119355 presented at
the 2009 Hydraulic Fracturing Technology
Conference, The Woodlands, January 19-21.
Roussel, N.P. and Sharma, M.M.,
"Quantifying Transient Effects in Altered-
Stress Refracturing of Vertical Wells," paper
Global 119522
SPE 119522 presented at the 2009 Hydraulic
Fracturing Technology Conference, The
Woodlands, January 19-21.
Potapenko, D.I., et al : "Barnett Shale
Refracture Stimulations Using a Novel
Diversion Technique," paper SPE 119636
USA-Texas 119636
presented at the 2009 Hydraulic Fracturing
Technology Conference, The Woodlands,
January 19-21.
Rajappa, B., et al: "Paradigm Change:
Proppant Fracture Treatments in the Lower
Green River Formation of the Altamont-
USA-Utah 123526 Bluebell Field, Utah," paper SPE 123526
presented at the 2009 Rocky Mountain
Petroleum Technology Conference, Denver,
April 14-16.
Pournik, M., et al,: "Acid Re-fracturing: is it a
Good Practice?" paper SPE 124874
USA 124874
presented at the 2009 ATCE, New Orleans,
October 4-7.
Barba, R.E., "A Novel Approach to Identifying
Refracturing Candidates and Executing
Integrated Energy
USA 125008 Refracture Treatments in Multiple Zone
Reservoirs," presented at the 2009 ATCE,
New Orleans, October 4-7.
Flores, D.Q. et al : "Finding Additional
Resources in a Mature Field: Refracturing
Engineering Workflow for a Successful
USA-Texas 125260 Campaign in Vicksburg Basin, South Texas,"
SPE abstract 125260 presented at the 2009
Tight Gas Completions Conference, San
Antonio, June 15-17.
Bundy, T.E., "Chittim Ranch Redevelopment
Case Study," SPE abstract 125277
USA-Texas 125277
presented at the 2009 Tight Gas Completions
Conference, San Antonio, June 15-17.
Vargas, W.O., "Rejuvenating "Marginal"
Permian Basin Assets: Deploying New Tools
in Mature Fields to Unveil Hidden Potential,"
USA-Texas 125279
SPE abstract 125279 presented at the 2009
Tight Gas Completions Conference, San
Antonio, June 15-17.
Rodvelt, G.D., et al,: "Case History:
Recompletions in a Virginia Coalbed-
Methane Field Yield Additional Gas,: paper
USA-Virginia 125458
SPE 125458 presented at the 2009 Eastern
Regional Meeting, Charleston, September 23-
25.
SPE 134330 Appendix Page 15 of 16
Companies Refrac Findings Reason for Refrac Success or Failure Location Type
The aim of refracturing has been to repair or replace an ineffective initial fracture
treatment to increase production rates and reserves. Reorientation theory adds
Based on simulations, refracturing to achieve reorientation is often not
Aspen Operating, another reason - to tap unstimulated pay.
economic if reservoir drainage area is as small as 10 acres. In larger lenses
Colorado School This study concludes that: reorientation can be studied using a reservoir simulator; stimulation gas
covering 40 acres, reorientation can increase 10-year cumulative production by
of Mines incremental gains were observed that were consistent with possibility of
15 to 22% in lenticular tight gas sands.
reorientation; it is not economical under typical conditions in most tight gas
lenticular reservoirs due to limited volumes.
Fracture reorientation can be induced by production or injection prior to Conditions leading to orthogonal secondary fractures have been analyzed to
restimulation. A model is described that can be applied to compute the optimal establish some rules of thumb. Effects of bounding layers and permeability
University of
time for refracturing four different types of reservoirs: shale gas, tight gas, heterogeneity and anisotropy on stress reorientation are also discussed. The
Texas
conventional oil reservoirs and conventional gas reservoirs to best achieve desired conclusions of this study are useful for the design of refracturing operations
reorientation. and candidate well selection.
To increase the productivity of existing wells and book additional reserves at
A degradable diverter generates a temporary plug within the existing fracture
reduced cost, vertical wells have been restimulated with demonstrable success.
system allowing fracture propagation in new sections of the reservoir, while
This provides opportunity to enhance refracture treatment by targeting bypassed
maintaining original fractures. This does not require the existence of
and ineffectively stimulated zones. Limited success restimulating horizontal wells Newark, East field,
Schlumberger perforations or perforation tunnels, which are likely to be eroded after pumping gas
has been achieved with current stimulation techniques. A new fracture diversion Texas
proppant. Real-time fracture monitoring is key, to enable operators to make
technique is discussed that creates a temporary bridge within the active fracture
informed decisions that influence fracture geometry, increase lateral coverage,
network to cause treatment redirection to understimulated intervals. Data available
and improve gas recovery. See also OTC 20268.
for one well showed that this application improved EUR by 20%.
Previous acid stims were attempted. If initial propped frac screened out, rarely
would a refrac be pumped to completion, even with increased pad and low
BJ Services, El Production data from 20 hydraulic recompletions is compared. Propped fracs were
proppant concentration. Attributed to pre-existing fracture complexity
Paso Exploration more expensive than acid jobs, but provided higher EUR and typically yield higher oil & gas
generated by initial frac. Adding 100 mesh sand to the pad stage improved
& Production ROR.
placement. 3D modeling helps in designing an efficient proppant fracture
treatment and better segmenting of individual fracture stages.
A set of four different experiments were conducted using different acid
composition, leak-off conditions, and closure stress loading to determine the effect
While the mechanisms are not identical between acid fracs and propped
Texas A&M of these parameters on acid re-fracturing. All the experiments indicate that re-
hydraulic fracs, there are some interesting parallels with sequential injections Middle East
University acidizing results in more etching with rougher profile and greater wormhole activity
above parting pressure.
than the first acidizing. While the first acidizing weakens the fracture face, the
majority of experiments indicate stronger fracture faces after the re-acidizing.
In one Wilcox/Vicksburg analysis, pay was bypassed due to excessive
Effective restimulation of multizone wells involves determining underlying cause for
"shotgun" perforating. Long perforated stages, stages with partial flowback
poor production, identifying the specific zones with remaining reserves, determining
Services followed by shut-in, and wells with drawdown exceeding 50% of reservoir
reservoir pressure, and selectively stimulating the target zones.
pressure appeared to have poorer completion efficiencies.
To develop a methodology to identify re-stimulation candidates in a short time with A point-based ranking system was developed. Indicators used were: Initial
Anadarko limited engineers and limited budget, case histories from prior re-stimulation productivity to long-term performance relationship; Physical condition of the
Vicksburg Basin,
Petroleum, treatments were evaluated to find trends and success indicators. Using this current wellbore; The type of proppant used in the original stimulation; Historic gas
South Texas
Schlumberger improved methodology, success was increased to 75%, and costs per MSCF gas water production; Shape of the historic decline trend. Procedure outlined to
were reduced by 40%. identify and implement refracs in multilayered wells.
Integrated teams, using new technologies and a new well-manufacturing
The Chittim Field was discovered in the mid-fifties and was steeply declining.
process, are currently drilling 6 wells per week using a coiled tubing drilling rig,
Studies determined that there was substantial remaining resource, but
costing 20% less than the 2006 pilot program. Chittim Field,
redevelopment would require multistage fracturing. High waterflood injection
ConocoPhillips Success attributed to managing the work flow (focus on efficiency), identifying Maverick County, oil
pressures appeared to initiate horizontal fractures. Out-of-zone injection partially
problems with original water flood, effective organization, optimum technology, Texas
caused by poor cement jobs. Recompletions shown adding 25-50 bopd,
measuring performance, and stimulating (and planning) in groups versus
increasing field rate to 2100 bopd.
single well.
Modern multidisciplinary teams are able to reexamine reservoirs with state-of-the-
Advantages over 1940s-50s: Multidisciplinary teams, multi-stage hydraulic
art tools and processes to evaluate the Permian Basin with new eyes. 70 boepd
fractures, ultra-deep-penetration guns and newly-deployed industry equipment. Permian Basin,
ConocoPhillips was predicted, actual production was 100 boepd in 15 wells drilled with new oil
We now have more efficient analysis and monitoring to identify potential and Texas
technology. Many mature fields could be revisited in the same way. Current
machinery/processes to get to it.
technology can revitalize "marginal" leases to become prime assets.
Wells initially targeted 15 to 25 CBM seams with 4 to 6 frac stages. 20/40 brown Recompletion candidate selection requires integrated review of well files, logs,
sand was used followed by 12/20 sand tail-ins. Wells with production of sand or and production data for candidates and direct offsets. When stage-fracturing
CNX Gas coal fines targeted for restimulation. The re-frac wells were treated with foam fracs multiple seams, the two thickest seams' thickness should make up at least Oakwood field,
coal-bed
Company, and the SMA Chemical to improve conductivity and control fines. Scale inhibitor 70% of the perforated thickness in a stage. Refracturing problematic Buchanan County,
methane
Halliburton was placed in an overflush stage. Fluid returns indicated little or no fines, and sand/fines completions with SMA-treated proppant improved production and Virginia
producing rates have continued to climb. Refracs added reserves at $0.61/mcf reversed decline. Remediation with RSS treatments reversed decline of coal-
and were the most effective form of remediation attempted. fines producers.
Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)
Region Paper Author - Reference Companies Refrac Findings Reason for Refrac Success or Failure Location Type

Flow capacity of the fractures can be significantly increased if larger diameter,

University of Current challenge remains to improve effectiveness and longevity of the fracturing more spherical, and stronger proppants can be introduced into the fracture.
Shah, S.N., et al: "Fracture Orientation and
Oklahoma, Insight treatments in horizontal wells. Data indicate proppant degradation may largely Wells treated with more aggressive standards achieved increased productivity.
Proppant Selection for Optimizing Production
Consulting, contribute to the need for restimulation. Successful design of transverse fractures High quality ceramic proppants last longer, often justifying expense. Decline USA, Siberia, North
Global 128612 in Horizontal Wells," paper SPE 128612 various
StrataGen requires increased focus on fracture conductivity. Fracture orientation and rates after several years should be included in economic analysis of proppant Sea
presented at the 2010 Oil and Gas India
Engineering, proppant selection are discussed in detail. Design and orientation of the wellbore, selection. Current fracture models do not accurately describe realistic fracture
Conference, Mumbai, January 20-22.
CARBO Ceramics perforations, and induced fractures should be considered simultaneously. flow capacity and complex geometry, so fracture optimization and refrac
potential should be confirmed with field production data.

Diagnostics are necessary in a resource play dominated by well spacing issues,

Surveys indicated 1/3 of restimulated wells show no improvement, but
drilling efficiency, stimulation scheduling, and IP rates. Available Pre-frac and Post-
Craig, D.P.: "Shale Pressure Diagnostics: Pre- Reservoir selection criteria are often oversimplified, including production trends and
Presentation, frac diagnostics, and their importance, are described. Observations about preferred gas
Global frac Reservoir Evaluation, Post-frac Development "intuition". More sophisticated techniques to inject into candidate intervals, various
April 2010 treatments for various formations are discussed. New pressure type curve discussions
Stimulation Effectiveness." April 2010 Consulting, carefully evaluating pre- and post- closure behavior can lead to superior
matching can be used when conventional analysis methods fail. Closure response
selection of refrac candidates.
can be interpreted with type curve methods.

SPE 134330 Appendix Page 16 of 16 Refrac Field Examples in Chronological Sequence, (printed version shows only a few columns from compiled database)