Hydraulic Fracturing of Coalseams

Prévia do material em texto

June 2007
Chapter 8
Hydraulic Fracturing of Coalseams
8.1 N
The coalb
methane c
dewatered
gas. Hydr
1948, pro
elevating 
Although 
reservoirs
for the coa
• The su
• The co
fractu
the fl
pressu
• Fractu
filter c
• Multip
• High p
• Young
• Induce
mine-
• Horizo
Warri
• Fines 
• Coalse
thick, 
Hydraulic Fracturing of Coalseams 357
eed for Fracturing Coals
ed methane (CBM) industry began after the realization that large
ontents of coals could often be produced profitably if the seams were
 and if a permeable path to the wellbore could be established for the
aulic-fracturing technology, developed in the oil and gas industry after
ved to be the answer in many cases for facilitating dewatering and
gas production rates to economic levels.
hydraulic fracturing had been highly developed for conventional gas
 of low-permeability sands, adjustments to the process were necessary
l because of the following phenomena:
rface of the coal adsorbs chemicals of the fracturing fluid.
al has an extensive natural network of primary, secondary, and tertiary
res that open to accept fluid during hydraulic fracturing but close upon
uid afterwards, introducing damage, fluid loss, fines, and treating
res higher than expected.
ring fluid can leak deep into natural fractures of coal without forming a
ake.
le, complex fractures develop during treatment.
ressures are often required to fracture coal.
’s modulus for coal is much lower than that for conventional rock.
d fractures in some vertical CBM wells may be observed in subsequent
throughs.
ntal fractures occur in very shallow coals, such as the Pratt group in the
or basin.
and rubble result from fracturing brittle coal.
ams to be fractured may be multiple and thin, perhaps only 1 or 2 ft 
requiring a strict economical approach to the operations.
Coalbed Methane: Principles and Practices
358 Hydra
Successful application of fracturing to coalseams has been helped by research
during the 1980s in the Black Warrior basin at the Gas Research Institute’s Rock
Creek site. The research helped reduce the costs and improve the performance of
hydraulically fractured coalseams, serving somewhat as a field laboratory for the
development of the process. Improvements continue, especially in preventing
damage to
8.1.1 Ap
The centra
in place. S
coal gas f
low produ
Vertical, u
Appalachi
less than 
produced 
wells that
smallscale
it was not 
It became
could ben
The effect
results of 
Pennsylva
coal was 
1,000 ft d
maximum
150-ft and
in Fig. 8.1
condition
ulic Fracturing of Coalseams June 2007
 the coal.
palachian Wells Inadequately Stimulated
l and northern Appalachian basins have an estimated 66 Tcf of CBM
everal decades before the CBM process became commercially viable,
rom vertical wells in the Appalachian basins was being produced, but
ction rates from these early wells contrast sharply with current rates.
nstimulated, or inadequately stimulated CBM wells in the northern
an basin completed before 1980 produced methane at modest rates of
140 Mcf/D with most of the wells at 10–30 Mcf/D.1 (Those that
more than 100 Mcf/D had permeabilities greater than 10 md.) Of the
 were hydraulically fractured, the sizes of the hydraulic projects were
. Although production could be sustained for long times at these rates,
economical to produce for pipeline sales.
 apparent in these early wells that the low-permeability formations
efit from fracturing and that the benefit depended upon fracture length.
 of fracture length is indicated from the field data and the simulation
a test well drilled in 1975 into the Pittsburgh seam in Greene County,
nia.1 (The Pittsburgh seam is mined in the area.) Permeability of the
about 1.3 md and gas content 190 scf/ton. The coalseam was about
eep and about 6.5 ft thick. The well was not fractured, and it gave a
 production of 21 Mcf/D. Simulation results of Hunt and Steele for
 250-ft half-length fractures are compared to the unstimulated well data
. The results demonstrate the need for hydraulic fracturing under these
s, which could have yielded 80 Mcf/D with a 250-ft fracture
Coalbed Methane: Principles and Practices
June 2007
half-length. Peak gas production would have occurred several years sooner in
fractured wells. 
Further co
gives adde
production
over the fi
increases.
350 ft co
half-length
Fig. 8.1—
Hydraulic Fracturing of Coalseams 359
mputer simulation by Hunt1 with data from wells in Greene County
d insight into the positive effect of longer fracture half-lengths on gas
 rate over a period of 10 years. Production rates increase dramatically
rst few years from coals of low permeability when fracture half-length
 Production rates from three half-length fractures of 150 ft, 250 ft, and
nverge at 10 years, but at the peak rate after 2 years the 350-ft
 would produce at a rate 66% higher (see Fig. 8.2).
Extent of fracturing effects.1
Coalbed Methane: Principles and Practices
360 Hydra
The benef
the abso
Schraufna
Black Wa
fracture h
permeabi
improvem
range, frac
Fig. 8.2—
ulic Fracturing of Coalseams June 2007
it of the fracture length at infinite fracture conductivity is qualified by
lute permeability of the seam. Simulations by Spafford and
gel2 (Fig. 8.3) are based on reservoir parameters indigenous to the
rrior basin and show 5-year cumulative gas production as a function of
alf-length and as a function of absolute permeability. A range of
lities exists in which longer fractures show marked production
ents, but beyond the high end and the low end of the permeability
ture length becomes unimportant.
Sensitivity to fracture half-length.1
Coalbed Methane: Principles and Practices
June 2007
Fracture 
Therefore
cannot be 
The length
above the
connect t
near-wellb
Fig. 8.3—
Hydraulic Fracturing of Coalseams 361
length assists productivity especially between 0.5 and 6.0 md.
, if the absolute permeability of a prospect is too low, the property
made economical by fracturing.
 becomes inconsequential as permeabilities exceed 10 md. Therefore,
 propitious permeability range, the goal of stimulation may be to
he wellbore with the natural fracture system, circumventing any
ore damage.
Efficacy of fracture length dependent on permeability level.2
Coalbed Methane: Principles and Practices
362 Hydra
8.1.2 Unstimulated Wells in Big Run Field
An interesting case history is the Big Run field in Wetzel County, West Virginia.
Conventional gas was produced from the Big Injun and Gordon sands below the
seam of co
and plugg
above the
water satu
(about 1,0
30 years, a
52 unstim
methane c
rates from
without fr
Fig. 8.4—
ulic Fracturing of Coalseams June 2007
al from 1905 until 1932, at which time the well was to be abandoned
ed. Upon pulling the casing, flow of gas was initiated from the coals
 abandoned sands; nearby mining in the Pittsburgh seam had reduced
rations to a low level. Recompletion of the well in the Pittsburgh seam
70-ft depth) proceeded to produce 200 MMcf of methane over the next
lbeit at a slow rate, without stimulation.1 Other wells were drilled and
ulated wells have produced from the field. After 43 years, 2 Bcf of
umulative production has resulted (see Fig. 8.4). Typical production
 the low-permeability Pittsburgh seam amounted to only 38 Mcf/D
acturing.
Big Run field, unstimulated.1
Coalbed Methane: Principles and Practices
June 2007
8.2 Unique Problems in Fracturing Coals
Most anomalies in fracturing coals result from uncommon values of properties of
the coal reservoir, such as rock mechanical properties and extensive natural
fractures in the coals. As a consequence of these coal reservoir properties,
induced f
altering of
higher tha
path of th
pressures
fracturing
Excessive
the coal. U
productio
particlesc
The organ
Fluid dam
of the coa
from the f
in the intri
Perhaps th
fluids hav
through th
During fr
penetratio
A consequ
rock of s
knowledg
Hydraulic Fracturing of Coalseams 363
ractures are very sensitive to complex in-situ stress profiles and the
 those stresses when drilling and fracturing. Treating pressures may be
n conventional reservoir fracturing. The cleat system influences the
e fracture and may introduce multiple fractures to increase treating
. Rubble generated near the wellbore or fines introduced during
 may contribute to higher treating pressures.
 fines are generated during fracturing because of the friable nature of
nfortunately, the fines continue to be generated during subsequent gas
n to reduce conductivity. Unlike the conventional reservoir, the
an be the size of powder or blocks large enough to plug perforations.
ic composition of the reservoir rock makes it susceptible to damage.
age to the coals occurs by two mechanisms. First, the organic surface
l is especially susceptible to fluid damage by adsorption of chemicals
racturing fluid or drilling fluid. Second, the fluids may become trapped
cate fissure network that constitutes the flow path.
e more pervasive problem is the trapped fluids. Cement and drilling
e been found to permeate surprisingly long distances from the wellbore
e natural cleat system to physically block these conduits of gas flow.
acturing, the imposed pressures open the cleats to allow fluid
n, subsequently trapping the gel upon closure to obstruct gas flow.
ence of the experience gained by the industry in fracturing a reservoir
uch different and complex properties is an advancement in the
e and understanding of fracturing in general.
Coalbed Methane: Principles and Practices
364 Hydra
8.2.1 Fines
Fines contribute to elevated pressures during fracturing.3 Fines are known to
deteriorate fracture conductivity with time, possibly packing into secondary and
tertiary natural fractures to damage permeability.
Some rese
fracturing
fracturing
as the mo
create rub
could pack
cause high
quantitativ
Laborator
load the f
frictional 
fracture. 
average of
Jeffrey d
lb/1,000-g
of fines ge
in the flow
More imp
the wellbo
Injection f
this.
Fines are 
flowing p
hydroxyp
rates in a c
ulic Fracturing of Coalseams June 2007
arch has helped explain qualitatively the contribution of fines to high
 pressures. Several mechanisms are offered.4,5 Fines could load the
 fluid to increase its viscosity and consequently increase pressure drop
re viscous fluid moves through the fracture. Parting of the coal could
ble and fines near the wellbore for a more tortuous flow path. The fines
 in the tips of developing fissures or bridge elsewhere in the fracture to
er treating pressures. A more important question revolves around the
e impact of fines on fracture treating pressures.
y burst-tests verify the generation of fines but in volumes that will not
racturing fluid appreciably. Therefore, there should not be excessive
pressure drops introduced by fines in the flow of the fluid through the
In coal burst-tests in the laboratory by Jeffrey and coworkers,6 an
 0.0144 lb of fines per sq ft of fracture surface area was created.
etermined the increase in apparent viscosity from loading a 40
al noncrosslinked fluid with 120- to 170-mesh coal fines. The volume
nerated in his tests would not significantly increase the pressure drop
 of the fracturing fluids in coals.
ortant effects on treating pressures come from fines concentrating near
re to create high pressure drops in the fluids flowing through them.
alloff tests in CBM wells that reveal high skin factors are indicative of
also created from the attrition of the fracturing fluid, loaded with sand,
ast the coal surface. In a laboratory experiment,6 a 40 lb/1,000-gal
ropyl guar (HPG) gel with 8 lb/gal sand flowing at typical fracturing
oal-simulated fracture generated fines linearly with time (see Fig. 8.5).
Coalbed Methane: Principles and Practices
June 2007
A tortuous fluid path causing high-velocity fluid flow, such as near the wellbore
or through opened butt or tertiary cleats, would contribute to the attrition of fines.
Shear stresses on the coal that move one face of the fracture or cleat relative to
the other face would also be expected to generate fines.
Perforatin
Creek in p
were gen
abrasiven
where the
the fractur
the inorga
Perforatin
entrained 
fracture of
Fi
ne
s,
 
sq
 ft
 x
 1
0-5
Fig. 8.5—
Hydraulic Fracturing of Coalseams 365
g only in the rock partings between seams proved effective at Rock
reventing pump repairs and workovers, primarily because fewer fines
erated.3 Since the fracturing fluid loaded with sand increases in
ess with velocity, most damage occurs in the vicinity of the wellbore
 cross-sectional area of the flow channel is smallest and the velocity of
ing fluid is greatest. In the case of thin, multiple seams, perforating in
nic rock avoids the high attrition of coal fines near the wellbore.
g in an acceptable rock parting may later help remove coal fines
with production fluids by screening those fines in the sand-propped
 the inorganic rock before they concentrate at the wellbore.
40 lb/1,000 gal gel
8 lb/gal sand
Coal: Rock Creek Seam, Utah
Time, hr
0 1 2 3 4 65
0
5
10
15
20
Fines from fluid abrasion laboratory flow tests.6
Coalbed Methane: Principles and Practices
366 Hydra
In many cases, it is desirable to perforate only the coalseams to avoid directing
the hydraulic fracture treatment into a lower-stress sandstone or carbonate. The
operator must then have a remedial process for alleviating damage caused by
fines plugging the sandpack and wellbore area. 
A post-fr
blockage t
fines is th
surface of
formation
8.6 shows
thin carri
helping br
of the frac
restoring c
Fig. 8.6—
ulic Fracturing of Coalseams June 2007
acture service that helps remove wellbore damage and coal fines
hrough a powerful backflush has been developed. The mobility of the
en restricted with a proprietary chemical formulation that makes the
 the coal particle “tacky,” enabling them to stick together and cling to
 features away from the critical flow paths in the proppant pack. Fig.
 how fines “clots” can accumulate near the wellbore in the pack. The
er fluid is pumped under high pressure into the damaged fractures,
eak down the clots of coal fines and displacing them to the outer limits
ture system. The clots are immobilized at the far reaches of the pack,
onductivity to the wellbore. 
Removing and holding fines away from the wellbore.
Coalbed Methane: Principles and Practices
June 2007
This proprietary system (marketed by Halliburton as CoalStim® Service) can also
be formulated to remove polymer damage from fracturing treatments. While the
well is shut in after treatment to allow the chemical process to alter the coal fines’
surface, polymer breakers will have time to dissolve residue to improve pack
conductiv
this treatin
This proc
increase g
8.7 depict
service on
with a pay
Fig. 8.7—
Hydraulic Fracturing of Coalseams 367
ity. Both guar and polyacrylamide polymers have been removed with
g fluid. 
ess has been used in the Rocky Mountain and Appalachian basins to
as production from 17.5% to 25% with payouts of less than 9 days. Fig.
s one operator’s success in using the process. Another operator used the
 a 30-well program, increasing production an average of 66 Mcf/D
out of 32 days.7
Production increase from controlling fines.
Coalbed Methane: Principles and Practices
368 Hydra
Another improvement in fines control is the use of a surface modification agent
(SMA) on the surface of the proppant grains during hydraulic fracturing that
provides several benefits:
• Helps maintain a high well production rate fora longer period of time.
• Enhan
• Reduc
• Helps 
• Adds 
coated
• Stabili
forma
With the a
pack/form
proppant 
production
the beginn
place the 
de-waterin
to provide
SMA was
increase p
history. 
SMA-tre
performan
SMA, pro
months af
ulic Fracturing of Coalseams June 2007
ces the frac fluid cleanup (see Fig 8.6).
es proppant settling to help improve permeability of the proppant pack.
reduce proppant flowback.
surface modification agent (SMA) on-the-fly to help eliminate leftover
 proppant.
zes the proppant pack/formation interface to reduce the intrusion of 
tion material into the proppant pack.
mount of fines generated during a stimulation treatment, a stabilized
ation interface is critical to maintaining conductivity through the
pack (Fig. 8.8). Intrusion of fines into the pack is the major cause of
 decline in a CBM producer. Besides plugging the pack, fines can be
ing point for scale precipitate formation. Using SMA, the operator can
rod pump below the lowest perforations, allowing a more efficient
g of all coals. All CBM projects can benefit from lowering the pumps
 lower backpressure on the coals.
 used in the Fruitland Coal in the San Juan basin8 for an operator to
roduction from no production up to 200 Mcf/day in a re-frac case
Low-gel borate (LGB) fluid was used to place 300,000 lb of
ated proppant in two of three re-fracs confirming the process
ce. LGB was used on all three wells. However, in the two wells using
duction showed a four-fold increase that was being maintained several
ter treatment. Economic value to the operator was $720,000 per year.
Coalbed Methane: Principles and Practices
June 2007
8.2.2 Flu
The organ
of ingredi
the inorg
entrapme
tertiary f
diffusion, 
Molecules
adsorbed 
permeabil
the adsorb
A possibl
matrix sw
3.5-in. dia
basin) we
permeabil
Fig. 8.8—
conducti
Hydraulic Fracturing of Coalseams 369
id Damage
ic surface of coal has the potential of being damaged from adsorption
ents of the fracturing fluid (or drilling fluid) in a manner unlike that of
anic surfaces of conventional reservoirs. Adsorption and physical
nt of polymer molecules in the coal obstructs butt and face cleats,
issures, and micropore openings to restrict methane desorption,
and Darcy flow.
 small enough to enter the micropores, such as CO2, that are strongly
in the micropores cause swelling of the coal matrix with attendant
ity reduction. The degree of swelling is dependent upon the affinity of
ate for the solid surface.
e problem of chemicals in crosslinked gels altering permeability by
elling from adsorption has been investigated by Puri, et al.9 Cores of
meter (from the San Juan basin) and 2.0-in. diameter (from the Warrior
re evaluated in the laboratory by Amoco for polymer damage to
ity. The flow tests were structured to isolate permeability damage from
A stabilized proppant pack/formation interface helps maintain 
vity through the proppant pack.
Coalbed Methane: Principles and Practices
370 Hydra
sorption effects and to minimize extraneous effects of cleats physically bridging
and packing with gel. The gel in the tests had been broken and the fracturing fluid
filtered. It was found that HPG decreased permeability by a factor of 10 in each
of the two coals. In Fig. 8.9, the Fruitland coal exhibits a precipitous decline in
permeability simultaneously with the commencing flow of the fracturing fluid.
After dete
reinstated.
In Fig. 8.1
damage fr
Pe
rm
e
a
bi
lity
,
 
m
d
0.00
0.05
0.10
0.15
0.20
Fig. 8.9—
ulic Fracturing of Coalseams June 2007
rioration of permeability from sorption, permeability could not be
 The damage was mostly irreversible.
0, the higher permeability Warrior basin coal demonstrated a similar
om the broken polymer in the Amoco test.
Reverse H O Flush Started2
Forward H O Flush Started2
Stable H O Permeability
Start of Frac Fluid Flow
2
Time, hrs
0 10 20 30 6040 50 70 1201101009080
Gel damage, San Juan core.9
Coalbed Methane: Principles and Practices
June 2007
It is recog
fissures of
protected 
of gels to 
treating pr
by closure
Mineback
extended 
formation
It should 
successfu
effects fro
substantia
Fig. 8.10
Hydraulic Fracturing of Coalseams 371
nized that the primary and secondary cleat system as well as the tertiary
 coals represent the flow system for future gas production and must be
during the drilling or completion process.10 Besides chemical damage
the organic surface, blockage of the natural fractures can occur as high
essures open fissures for fluid invasion and as the gels become trapped
; filter cakes may not limit fluid invasion as in sandstone formations.
 has revealed unbroken gels in fractures far from the wellbore at
times after treatment. An estimated 25% of the gel remained in the
 in an Oak Grove, Alabama test conducted by Amoco.11 
be emphasized that fracturing with gel fluids has produced many
l wells that are economical and operate with no apparent deleterious
m the fluid. However, gel damage does often occur, and it can be
l.
—Gel damage, Warrior coals.9 
Coalbed Methane: Principles and Practices
372 Hydra
At the Rock Creek test site, remedial treatments of poorly performing wells12
were conducted. The criteria for selecting the wells for corrective action were as
follows. The criteria reflect the probability of the original fracturing fluid
damaging
• Origin
breake
• Fluid 
• Some 
The restim
been used
retarded a
hydroxyet
380 Mcf/D
Fig. 8.11
ulic Fracturing of Coalseams June 2007
 the coal:
al stimulations used guar-based fracturing fluids with an enzyme
r.
returned at high viscosity after fracturing.
wells underachieved in the midst of good performers.
ulation of Well P3 at Rock Creek is a classic example.13 HPG gel had
 originally to fracture the well. Production rates from the well were
t 65 Mcf/D. The well was refractured with nitrogen foam containing
hyl cellulose (HEC). After the remedial treatment, production reached
 (see Fig. 8.11).
—Restimulation with nitrogen foam.13
Coalbed Methane: Principles and Practices
June 2007
8.2.3 Excessive Treating Pressures
A higher pressure than ordinary may be necessary to initiate a fracture in coal.14
With normal expectations of overburden pressure gradient of 1.0–1.2 psi/ft and
of minimum horizontal stress of 0.6–0.8 psi/ft, the pressure to initiate the fracture
should be
create a v
gradient g
fracturing
indicated 
It is evide
normal 1.
Fig. 8.12
Hydraulic Fracturing of Coalseams 373
 approximately 100 psi greater than the minimum horizontal stress to
ertical fracture,5 or no more than a 1 psi/ft gradient. Instead, a fracture
reater than 1.0 psi/ft is often encountered in coals.9 A survey5 of the
 gradients encountered in the Black Warrior basin of Alabama
the distribution as presented in Fig. 8.12.
nt that most of the fracture gradients in the Warrior basin exceed the
0-psi/ft gradient. Note that some pressures exceeded 2.0 psi/ft. The
—High fracture gradients in Warrior basin.5
Coalbed Methane: Principles and Practices
374 Hydra
preponderance of wells were within the 1.0–2.0-psi/ft range. Only about 20% of
the wells exhibited gradients less than 1.0 psi/ft.
The following mechanisms have been postulated to account for the higher than
expected f
1. Bo
ati
ing
the
fra
2. Bu
ne
ch
3. To
oc
pa
ma
un
pa
op
4. A 
Th
flu
5. Fr
to 
6. Ra
The propo
8.13. The 
wellbore f
and multip
ulic Fracturing of Coalseams June 2007
racturing pressures in coal:
rehole instability or perforating causes rubble at the point of fracture initi-
on. Any stress relief of the coals results in breakup of the coal block. Drill-
 the wellbore, perforating, and even fracturing realign stresses surroundingborehole. The unconsolidated coal chips retard initiation of the hydraulic
cture.
rsting of the rock at fracture initiation generates fines that bridge the crack
ar the wellbore. Further from the wellbore, the accumulation of fines and
ips blocks the fracturing fluid front, redirecting the path of the fracture.
rtuous fracture path develops as the path follows cleats, slippage at joints
curs, and horizontal components at the rock interface develop. A tortuous
th may develop at the wellbore if the perforations are not aligned with the
ximum horizontal stress.15 Otherwise, the fracture may propagate radially
til extending in the direction of maximum horizontal stress. The tortuous
th causes greater pressure drops in the fluid, requiring higher pressures to
en the apertures sufficiently for sand traverse.16 
network of fractures, multiple fractures, and parallel fractures develops.
ese have been documented in minethroughs. They tend to divert fracturing
id, necessitating higher pressures to propagate the primary fracture.
acture tip anomalies occur from fines at the tip or fluid lag.17 This is similar
(3), but it occurs at the fracture tip.
ising pore pressures near the wellbore makes the coal subject to failure.
sed mechanisms causing high fracturing pressures are depicted in Fig.
most likely causes of the high fracturing pressures are rubble near the
rom poroelastic effects, tortuous path near the wellbore and beyond,
le fractures.
Coalbed Methane: Principles and Practices
June 2007
Laborator
Jones indi
high press
wellbore s
short dista
Fig. 8.13
Hydraulic Fracturing of Coalseams 375
y and simulator uses of field data by Khodaverdian, McLennan, and
cate that coal fragments in the fracture near the wellbore help cause the
ures.5 The pressures in the fracture as a function of distance from the
how the effect of near-wellbore damage, as pressures drop off rapidly a
nce from the well18 (see Fig. 8.14).
—Mechanisms causing excessive fracturing pressures.4
Coalbed Methane: Principles and Practices
376 Hydra
Fluid leak
that mech
modulus 
increasing
In the cas
parting be
from perf
initiation
wellbore. 
The five p
individual
each mech
M
ax
im
u
m
 In
-fr
a
ct
ur
e 
Pr
es
su
re
,
 
ps
i
2
3
4
5
6,000
Fig. 8.14
ulic Fracturing of Coalseams June 2007
off from the fracturing process increases pore pressure to the extent
anical properties of coal deteriorate near the wellbore. Young’s
decreases and Poisson’s ratio increases in such instances, thereby
 the fines generation and causing the failure of the coal matrix.
e of multiple, thin seams, perforating below the coalseam or in the
tween seams, if the bounding rock is suitable, reduces coal rubbling
orations, fines generation from the bursting of the coal at fracture
, and attrition of fines from the high velocity of the fluid near the
It also may avoid degrading poroelastic effects.
roposed mechanisms presented in Fig. 8.13 may work in consort or
ly. Most have been verified. The amount of the pressure drop due to
anism is unknown in the coal fracturing process.
Minimum Horizontal Stress, Hmin
,000
,000
,000
,000
0 1 2 3 4 5
Distance from Wellbore, in.
—Near-wellbore damage.18 
Coalbed Methane: Principles and Practices
June 2007
8.2.4 Leakoff
Historically, when coalseams were encountered in the hydraulic fracturing of
conventional formations, the coal acted as a barrier to fracture growth because of
fluid leakoff, elastic properties of the coal, and the likelihood of slippage at the
coal-rock
penetrate 
magnified
The follow
• Loss o
• Fractu
• Forma
• Screen
The sever
observatio
of a coal m
In anothe
stimulatio
governme
during fra
unpropped
stair-stepp
In extensi
hydraulic
factor may
hvAb-ran
superpose
coal in th
pressurizi
gels.
Penny and
with 3.5-i
Hydraulic Fracturing of Coalseams 377
 interface. With the advent of the CBM process and the objective to
or stay within the bounds of the coal, the problem of leakoff became
.
ing deleterious effects result from leakoff in coals:
f fluid limits penetration of the fracture.
ring efficiency decreases.
tion damage likely occurs.
out probability increases.
ity of the leakoff problem in coals is substantiated from mineback
ns. For example, cement was observed in a natural fracture in the roof
ine 133 ft from the wellbore at Oak Grove in the Black Warrior basin.
r instance, unbroken gel was spotted in a fracture 7 months after the
n was completed.19 In a third case of eight field treatments in a
nt-sponsored test where fluorescent paint was part of the fluid system
cturing, paint was observed as far as 630 ft from the wellbore in
 face and butt cleats. The paint in some intercepted fractures revealed
ed butt and cleat joints propagating through the coal.20 
ve natural fracture networks of coals, the pressures imposed during
 fracturing open the fissures to compound the leakoff problem. This
 be accentuated in the fairway section of the San Juan basin where the
k coal has an elaborate network of cleats, closely spaced, including
d tertiary cleats from a reoriented stress field. The high-permeability
e fairway is more susceptible to leakoff of fracturing fluids upon
ng, and greater damage to the coals may result from fracturing with
 Conway21 addressed the leakoff problem in laboratory experiments
n. × 2.9-in. mined coal samples taken from the Fruitland formation of
Coalbed Methane: Principles and Practices
378 Hydra
the San Juan basin. Because of the randomness of the cleat system, the
permeabilities of the samples ranged from 1 to 100 md with an average value in
his tests of 40 md. Although 1-md samples were impermeable to all fracturing
fluids, both crosslinked and noncrosslinked HPG fluids moved into the natural
fractures of the 40-md samples unhindered by any filter-cake buildup at modest
driving pr
obstruct le
Although 
is possible
gap and po
and large
intimated 
Fig. 8.15
ulic Fracturing of Coalseams June 2007
essure differentials (see Fig. 8.15). Note that no filter cake develops to
akoff at any pressure. At the higher pressures, loss of fluid increased.
the polymers do not bridge the cleat openings to initiate a filter cake, it
 to do so with the correct proppant size. The proppant may bridge the
lymer build upon it to prevent leakoff. The bulk of the fracturing fluid
r size proppant is then diverted to a primary induced fracture. It is
that multiple fractures might be reduced to a single dominant fracture
—Leakoff in Fruitland cores.21
Coalbed Methane: Principles and Practices
June 2007
and tortuosity of the single fracture reduced by use of proppant slugs.16 Slugs of
100-mesh or 40/70-mesh sand early in the pad could direct the fluid and proppant
to a single fracture.
Sand of 100-mesh in concentrations as low as 2 lb/gal proved effective in
reducing l
cake in the
Note in Fi
added. Im
7
6
5
4
3
2
0
1
0
Le
ak
o
ff
Vo
lu
m
e 
x 
1,
00
0 
m
l
Fig. 8.16
Hydraulic Fracturing of Coalseams 379
eakoff to an insignificant level by facilitating the formation of a filter
 laboratory experiments of Penny and Conway21 (see Fig. 8.16).
g. 8.16 that leakoff progressed unabated until the 100-mesh sand was
mediately, a filter cake formed to eliminate the loss of fluid at the
1 2 3 4 5 6 7 8
No FLA 2 lb/gal 100-mesh
35 lb/1,000 gal
Guar/Borate
Crosslinked
ΔP = 400 psi
40 md coal
Time, min0.5
—Leakoff prevention in Fruitland cores.21
Coalbed Methane: Principles and Practices
380 Hydra
higher 400-psi test condition. The results have implications for reducing fluid
damage to the coals and for creating a single dominant hydraulic fracture.
A leakoff coefficient, Cw, may be calculatedusing Eq. 8.122 to provide an
approximation of how much fluid will leak into the formation, affecting height
and penetr
where
Cw =
m =
A =
For the ca
0.001 ft/m
The fine-m
fissures in
are dilated
Cramer23 
basin to s
effective u
cleats and
mechanism
segregatio
different f
the coarse
ulic Fracturing of Coalseams June 2007
ation of the fracture.
leakoff coefficient, ft/min1/2
slope of fluid-loss curve (filtrate volume/ ), ml/min0.5
cross-sectional area of sample, cm2
se of the 40-md Fruitland samples of Fig. 8.16, Cw is determined to be
in0.5 with the 100-mesh sand in the fluid.
esh sand should be scheduled so that it is present as the cleats and
itially spread apart.10 Injecting the fine mesh later after the apertures
 may compound the problem.
reports the effective use in the field of 40/70-mesh sand in the San Juan
eal cleats and to prevent leakoff. Palmer and Kutas24 also relate an
se of 40/70-mesh sand preceding a coarser 12/20-mesh sand to seal the
 secondary pathways that open when fracturing San Juan coals. The
 was verified when radioactive tracers in the two sands indicated a
n of the two sand sizes in the coal and placement of the two sizes in
ractures. The fine sand went to close secondary and tertiary fissures;
r sand propped the main fracture.
2A
m 0.0328 = Cw (8.1)
 time) (flow
Coalbed Methane: Principles and Practices
June 2007
A 100-mesh sand was used to control leakoff in the U.S. Department of Energy’s
multiwell experiment, resulting in completing the fracturing as designed.10 
Since fluids may enter the cleats and secondary fissures when they are dilated
from treating pressures, later cleanup at reduced pressures may leave gel trapped
to reduce 
possible t
selecting p
Partly bec
in fracturi
8.3 T
For metha
must be a
system an
fissures to
cavity com
There has
fluid—wh
without pr
of the prec
Cost, form
the choice
and it is s
foams wo
considera
crosslinke
Hydraulic Fracturing of Coalseams 381
permeability.10 It becomes important, therefore, to restrict as much as
he growth of complex fractures and fluid loss to them by properly
roppant size and schedule.
ause of better control of leakoff, nitrogen foams are increasingly used
ng coals.
ypes of Fracturing Fluids for Coal
ne production rates to be economical, permeability of the formation
dequate. Permeability of the coalseam depends on the natural fracture
d the connection of the fracture system to the wellbore. Connecting the
 the wellbore must be by hydraulic fracturing or by regionally limited
pletions.
 been uncertainty in the industry on the choice of the proper fracturing
ether to use linear polymer, crosslinked gel, water with proppant, water
oppant, or nitrogen foam. The history of changing popularity of each
eding fluids reflects the uncertainty.
ation damage, proppant placement, and propped fracture length dictate
. Table 8.1 summarizes the general attributes of the fluid selections,
urmised from the tabulation that either crosslinked gels or nitrogen
uld be preferred. Formation damage evolved as an important
tion in selecting a fluid, moving the preferred fluid selection from
d gels toward nitrogen foams.
Coalbed Methane: Principles and Practices
382 Hydra
8.3.1 Cr
In the man
Black Wa
lb per 1,00
Polymer c
and addi
specificati
The wate
propylene
stability. 
galactose 
chain.
Table 8.1—Fracturing Fluid Ratings
Cost Formation Damage
Proppant 
Placement 
Propped
 Length
Wate
Wate
Linea
Cros
Nitro
ulic Fracturing of Coalseams June 2007
osslinked Gels
y CBM wells that have been fractured in the San Juan basin and the
rrior basin, the fracturing fluid most frequently used has been a 30–35
0 gal HPG in 2% KCl water solution crosslinked with the borate ion.25
ontent of the gel is minimized to reduce residual unbroken gel, cost,
tional produced-water treatment requirements to meet BOD
ons.
r-soluble HPG polymer is derived from guar by combining it with
 oxide to achieve a polymer with less residue and higher temperature
The structure of HPG is presented in Fig. 8.17.26 It contains one
unit to two mannose units as the basic repetitive group of the polymer
r w/o proppant Good Good Poor Poor
r w/ proppant Good Good Poor Poor
r gel Fair Poor Fair Fair
slinked gel Fair Poor High High
gen foam High Good Good Good
Coalbed Methane: Principles and Practices
June 2007
Crosslinki
The borat
fluids. It l
O
H
H
HO
CH OR2
Hydraulic Fracturing of Coalseams 383
ng increases viscosity of the fluid with a minimum amount of polymer.
e ion is most commonly used as the crosslinker in CBM fracturing
inks the polymer as shown in Fig. 8.18.26 
O
H
O
H
OH
CH OR2
H
H
OH
OH
O
O
OH RO
H
HH
HH
CH2
O
OH
Fig. 8.17—Structure of HPG polymer.26 
Coalbed Methane: Principles and Practices
384 Hydra
The gel is
making i
borate-cro
temperatu
provide th
Juan basin
transport b
The relati
borate cro
without cr
the natura
absolute te
not dissoc
Fig. 8.18
ulic Fracturing of Coalseams June 2007
 shear thinning but reforms its structure with the borate ion crosslinker,
t easy to work with in the field. Apparent viscosity of the
sslinked gel is high, and it provides excellent proppant transport. At the
res encountered in CBM wells, structures of the gel are stable and thus
e viscosity needed for sand transport.27 Black Warrior basin and San
 temperatures of 105 to 120°F are in ranges that provide good proppant
y fracturing fluids.28
onship of apparent viscosity to temperature for one HPG gel with
sslinker is given in Fig. 8.19.27 Note that the apparent viscosity of HPG
osslinker follows the relationship with temperature of Eq. 8.2, where
l logarithm of the apparent viscosity is linear with the reciprocal of
mperature at temperatures where the polymer molecular structure does
iate. The gel’s apparent viscosity is much higher, but its viscosity
—HPG crosslinked with borate.26
Coalbed Methane: Principles and Practices
June 2007
decreases at the same rate as the polymer solution at temperatures encountered in
CBM wells; the gel viscosity declines with temperature according to Eq. 8.2.
where
µa =
ß, α =
T =
Higher te
abruptly, a
e = /Ta
αβμ (8.2)
Fig. 8
Hydraulic Fracturing of Coalseams 385
apparent viscosity
constants
absolute temperature
mperatures above those encountered in CBM wells break the gel
nd its viscosity declines to that of the base polymer solution.
.19—Apparent viscosity of gelled fracturing fluids.27
Coalbed Methane: Principles and Practices
386 Hydra
Fracturing with gels maximizes the fracture length and increases proppant
loading over longer distances. Good results have been reported in the Warrior
basin as well as the San Juan basin. HPG polymers crosslinked with the borate
ion as 30–35 lb of polymer per thousand gallons of solution are commonly used;
less than 1 11
Two exam
fracture c
12/20-mes
The proc
injected at
In a secon
35 lb/1,00
per minut
sand and 2
When com
possible d
1. Th
$5
W
2. Ch
3. Th
50
4. Br
After rese
coal, serv
LGB syst
polymer. T
desirable 
reducing f
damage m
ulic Fracturing of Coalseams June 2007
0 lb/gal of 20/40-mesh sand is common.
ples of fracturing treatments of coalbeds are as follows. A typical
onducted by Taurus in the Mary Lee group was designed to use
h sand, filtered water, hydroxypropyl guar, and borate ion crosslinker.
ess involved 63,000 gallons of fluid with 145,000 lb of proppant
 40 bbl/min; proppant load was ramped.29
d example, a 4,000-ft well in the San Juan basin was fractured with a
0-gal HPG crosslinked with the borate ion. Fluid was injected at 55 bbl
e, and proppant was injected in two stages: 22,000 lb of 40/70-mesh
10,000lb of 20/40-mesh.28
pared to water as the fracturing fluid, crosslinked polymers have four
isadvantages. 
e cost is higher. For similar jobs, fracturing with a gelled fluid costs 
0,000 while water fracturing costs $28,000 in the Oak Grove field of the 
arrior basin.7 
emicals in the gelled fluid may alter the surface properties of the coal. 
e polymer or gel may plug flow channels. Gel may penetrate into the coal
 ft from the vertical fracture and be trapped upon closure.11 
eakers added to the gel may be inadequate and leave unbroken gel in seams.
arch of fracturing fluids identified the possible damage mechanisms to
ice companies have improved the performance of the crosslink gels.
ems have been optimized to provide high viscosity with 50% less
ypical gel loadings have been reduced to 15–20 lb/Mgal of fluid. It is
to use a high-viscosity fluid that will transport sand efficiently while
luid lost to the coal cleat system. Whole fluid invasion is the primary
echanism when deciding which fluids to use. Shallow coal plays
Coalbed Methane: Principles and Practices
June 2007
generally have low bottomhole pressure. The driving force to produce back fluids
lost into the cleat system may not be present. The addition of nitrogen to the fluid
system can help alleviate fluid loss and provide energy to return treatment fluids.
Regardless of which fluid system is chosen, minimizing contact time with the
coal is the
cleaned a
stimulatio
production
Guar syste
High-per
instances
Cleaner b
by Palmer
with good
Crosslinke
permeabil
Fig. 8.20
Hydraulic Fracturing of Coalseams 387
 best method of reducing damage. It is recommended that wellbores be
nd the well placed on pump within 72 hours of performing the
n treatment. This may mean delaying the stimulation treatment until
 equipment is in place.
ms are preferred over HPG systems to lower the cost of gelled fluids.
formance enzyme breakers have been developed that eliminate
 of unbroken gel even at bottomhole temperatures as low as 55°F.
reaks mean higher regained conductivity (Fig. 8.20). In a survey done
, et al.,30 LGB fluid was the predominant fluid used in the Raton basin
 results. In Appalachia, the use of nitrogen foams predominates.
d foams have been used to provide improved sand transport on higher
ity coals.
—Cleaner gel breaks yield higher regained conductivity.
Coalbed Methane: Principles and Practices
388 Hydra
8.3.1.1 Hydrogen Peroxide 
As discussed in the previous section, polymers can penetrate the cleat system and
cause damage. Even the lower gel-loading systems used today can leave residual
damage. One emerging solution is the use of hydrogen peroxide (H2O2) as a
cleanup a
polyacryla
been an is
Lack of p
H2O2. The
service co
Halliburto
pumping 
placement
have a sa
treatments
In additio
cleat apert
dioxide an
operators 
One draw
inexpensiv
the delive
equipmen
8.3.2 Wa
Water has
with the m
12/20-san
1–1.5 lb/g
more sand
ulic Fracturing of Coalseams June 2007
id. H2O2 is a strong oxidizer capable of dissolving guar and
mide, commonly used products in fracturing. Placement of H2O2 has
sue of concern in the past.
rocess knowledge and understanding of risk have limited the use of
 rapid reaction of H2O2 with steel manifolding and tubulars prevented
mpanies from pumping it; operators did not want the safety liability.
n has designed a process using composite coiled tubing, stainless
equipment, and a chemical stabilization system that allows safe
 of the product in the coal with minimal surface risk. Operators now
fe, remedial, treatment process for removing gel damage from past
. 
n, the reaction of H2O2 on minerals in the coal serves to enhance the
ure, effectively increasing permeability. Reaction products are carbon
d water, both commonly found in coal. This is highly desired by coal
when the target zone is later to be mined.
back of the process could be cost. Proximity of location to an
e supply of H2O2 delivery could make the process economical. Cost of
ry system would best be minimized with a sequence of wells when
t is mobilized.
ter
 been used as the ultimate cheap, nondamaging fracturing fluid but
ajor deficiency of reduced sand transport. Less than 5 lb/gal of a
d has been used. Fracturing with water in coalbeds may pump only
al of sand without screenout; if the water flow rate is increased to carry
, the height of the fracture may grow. Excessive height growth of the
Coalbed Methane: Principles and Practices
June 2007
fracture in sand/water fracturing increases the problem of sand settling from the
water. Propping a limited portion of the fracture is indicated in Fig. 8.21 from a
simulation run by Amoco6 to match the results of fracturing the Black Creek
group in Alabama with water-carrying sand. Possibly, only one-third of the
seams in t
In the Oa
fracture t
sealers to
self-propp
formation
that the wa
than treatm
Fig. 8.21
Hydraulic Fracturing of Coalseams 389
he group were propped by the sand.
k Grove field, Amoco30 evaluated the use of water without sand to
he Pratt, Mary Lee/Blue Creek, and Black Creek seams using ball
 direct fluid flow. The concept is to create fractures that are
ing; slippage of the ragged fracture faces from shear stresses of the
 is supposed to support the fracture upon closure. Amoco concluded
ter fracture treatments with sand gave better gas production in the field
ents with water alone.
—Schematic of proppant distribution in water fracture.11
Coalbed Methane: Principles and Practices
390 Hydra
Without proppant present, coal fragments may help support the fissure. If in-situ
shear stresses cause slippage at the interface during fracturing, the rugosity of the
faces may provide a propped fracture. Some successes with water fracturing in
thin, multiple seams have been seen.
It is possib
less stress
them less 
face cleats
8.3.3 Co
A field st
fracturin
Twenty-th
and 10 wi
Characteri
The tabula
treatment,
coals wer
indicated 
had 20% 
Apparentl
more seam
formation
the water 
ulic Fracturing of Coalseams June 2007
le that water fracturing without sand creates fractures of less width and
 redistribution. These restricted widths may close face cleats parallel to
than wider fractures propped with sand, where closing of the parallel
 would divert gas flow to the less permeable butt cleats.31
mparison of Gel and Water
udy in the Oak Grove field of the Warrior basin compared water
g with gelled-fluid fracturing under controlled conditions.11
ree wells were fractured, 13 with water-soluble crosslinked polymer
th water. The selected wells were interspersed to avoid bias of location.
stics of the water and water-gel treatments are compared in Table 8.2.
tion shows approximately a 50% cost saving from the water-fracturing
 but the gel fluid transported more than twice as much proppant. The
e of good permeability and boreholes were cased and perforated as
in Table 8.3. After 12 months of production, the water-fractured wells
more methane production with less formation water production.
y, although the gel created longer and better propped fractures through
s, the shorter and poorly propped water fractures had negligible
 damage. The tradeoff in this case of a high-permeability coal favored
treatment.
Coalbed Methane: Principles and Practices
June 2007
The comp
fluids in 
fractures,
refractures
Table 8.2—Comparison of Water and Gel Fractures11
Characteristic Water Gel
C
Pr
Fl
N
Pr
C
Ef
P
C
D
Hydraulic Fracturing of Coalseams 391
arison was broadened to include the results from additional fracturing
the San Juan basin as well as the Warrior basin. Sandless water
 water with sand fractures, crosslinked gel fractures, sandless water
, and cavity completions were compared30 (see Table 8.4).hemicals No polymer Borate crosslink,HPG, 30 lb/1,000 gal
oppant <5 lb/gal 12/20
 70,000 lb/zone
10 ppg 12/20,
100,000 lb/zone
ow rate, bbl/min 50 to 60 40
umber of wells 10 Oak Grove 13 Oak Grove
oduction 12 months 12 months
ost, USD $28,000 $50,000 
ficiency, % <20 50 to 80
Table 8.3—Field Properties of Oak Grove Pilot11
Parameter Comments
ermeability 5 to 20 md
ompletions
• Cased and perforated. 
• Individual seams of Black Creek and Mary 
Lee/Blue Creek.
• Perforated, stimulated Black Creek.
• Repeated Mary Lee/Blue Creek.
epth 2,000 ft—Black Creek1,500 ft—Mary Lee/Blue Creek
Coalbed Methane: Principles and Practices
392 Hydra
The result
and a need
San Juan 
formation
A somewh
indicated
fracturing
fracture ap
8.3.4 Fo
Nitrogen 
surfactant
foam, or v
Table 8.4—Comparisons of Stimulation Treatments30 
Basin X Y
Gas 
Production 
Stimulation 
Cost
San Juan
San Juan
Black Wa
(Oak Gro
Black Wa
(Oak Gro
Black Wa
ulic Fracturing of Coalseams June 2007
s indicate a cost savings with the water, formation damage with gels,
 for proppant support of the fracture. A special case is indicated in the
basin where a good permeability and cleat system are sensitive to
 damage.
at similar study by Taurus in the Cedar Cove field of the Warrior basin
 a better performance of the crosslinked polymer than the water
 fluid in the first nine months of production,29 where a long, propped
parently overshadowed formation damage to increase production.
am
foam is a gas-in-water emulsion made stable by the addition of a
 and a viscosifying agent, such as HEC or HPG. The quality of the
olume percentage of nitrogen in the foam, may range from 60–90%.
X/Y X/Y
Cavity Gel 5 to 10 11.0
WFS Gel 2.5 0.5
rrior 
ve) WFS Gel 1.2 to 1.4 0.5
rrior 
ve) WFS SWF 1.9 2.0
rrior SWF
refracture
Gel original 
fracture 2.0 0.25
Coalbed Methane: Principles and Practices
June 2007
Nitrogen foam reduces formation damaging effects of the fracturing fluid for the
following reasons:
• The nitrogen provides energy to clean the fracturing fluid from the formation.
• The foam requires about 70% less water than a gel.32
• HEC i
• Foam 
In addition
enhance m
pressure o
diffusion o
Nitrogen d
adsorbed 
detrimenta
Advantage
• Cleans
• Leave
• Leave
• Inflict
• Enhan
• Provid
• Reduc
The disad
• More 
• More 
• Diffic
A laborato
seam) fro
permeabil
water, vis
The result
fluids,33 a
Hydraulic Fracturing of Coalseams 393
s used at reduced levels and is a less damaging viscosifier.
has better leakoff characteristics.
 to assisting fluid cleanup, the nitrogen released from the foam acts to
ethane desorption and production. The mechanism is to reduce partial
f methane in the coal, thereby creating a concentration gradient for
f methane from the micropores.
oes not cause appreciable swelling of the coal because it is less readily
than the methane. Carbon dioxide, if used in the foam, could induce
l matrix swelling because it is preferentially adsorbed by the coal.
s of nitrogen foam as a fracturing fluid may be summarized as follows:
 up quickly from the induced fracture.
s virtually no unbroken fluid.
s a minimum residue to plug the reservoir.
s minimum damage to coal.
ces CH4 desorption by lowering CH4 partial pressure.
es good proppant transport.
es leakoff.
vantages of a foam fracturing fluid for coals are as follows:
expensive.
difficult quality control.
ult to characterize rheologically.
ry analysis of permeability damage to Warrior basin coal (Blue Creek
m flow contact with a 70% nitrogen foam showed a high recovery of
ity after the test. The continuous phase of the foam was 2% KCl in
cosified with HEC polymer as 30 lb of polymer per 1,000 gal of liquid.
s in Fig. 8.18 illustrate the nondamaging aspects of N2 foam fracturing
s 78% of the permeability had been recovered shortly after foam
Coalbed Methane: Principles and Practices
394 Hydra
treatment, and improvement was continuing at that time. Although more
expensive than HPG, the HEC polymer is less damaging to the formation.34
8.3.5 Pr
Sand pro
economica
Some com
proppant a
matrix of
proppant, 
and (4) tra
Fig. 8.22
ulic Fracturing of Coalseams June 2007
oppant Considerations
ppant has sufficient strength for CBM applications, so it is the
l and practical choice.
mon problems encountered in conventional fracturing involving
re magnified in coalbed fracturing: (1) embedment of proppant into the
 the soft formation, (2) trapping of large volumes of fines by the
(3) leakoff of the sand-bearing fluid into secondary fissures and cleats,
nsport of the proppant through a tortuous path.
—Nondamaging aspects of foam.33
Coalbed Methane: Principles and Practices
June 2007
Because of the soft, elastic properties of coal, proppant embeds in the coal matrix
to reduce conductivity. In doing so, it causes spalling of the fracture face.
Consequently, the coal chips that collect in the sandpack further contribute to the
deterioration of fracture conductivity.25 As described by Eq. 8.3, the initial width
of the pac
sandpack 
where
Weff =
Wi =
ΔWc =
ΔWemb =
ΔWs =
Hardness 
laboratory
from hand
proppant e
from Fig. 
Low-volat
to proppan
Hydraulic Fracturing of Coalseams 395
ked sand in the fracture is decreased to eventually give an effective
width, Weff.
effective sandpack width
initial sandpack width
sandpack compression
sand embedment
sand width loss due to spalling
of coal, the property affecting embedment, is difficult to measure in the
 because of the randomness of fissures and the introduction of fractures
ling of the sample.35 A general indication of the susceptibility to
mbedment as a function of coal rank is given in Fig. 8.23. It is evident
8.23 that the hardness of coal increases rapidly at the anthracite rank.
ile bituminous and medium-volatile bituminous coals are most subject
t embedment.35,36
(8.3)W - W -W -W = W sembcieff ΔΔΔ
Coalbed Methane: Principles and Practices
396 Hydra
Higher lo
Holditch37
1.0 lb/ft2.
120
Vi
ck
e
rs
 M
ic
ro
ha
rd
ne
ss
,
 
kg
/m
m
2
Fig. 8.23
microhar
ulic Fracturing of Coalseams June 2007
adings of the proppant in the fracture will alleviate the problem.
 concludes that the fracture design should be for proppant loadings of
hvCb hvBb hvAb mvb lvb ansa
Coal Rank
0
20
40
60
80
100
—Relative embedment potential of coal ranks as determined by Vickers 
dness.35,36
Coalbed Methane: Principles and Practices
June 2007
Three other problems—fines, leakoff, and tortuous path—might be alleviated by
proper selection of size distribution for proppant and their schedule of
introduction. Radioactive tracers amid 100-mesh, 40/70-mesh, and 12/20-mesh
proppant used in the San Juan basin confirmed24 that the 100-mesh and
40/70-me
situated in
mechanism
tertiary c
fracturing
creates lar
Therefore
process i
prevents b
fracture.
A proper 
coal fines
schedule 
deep pene
flowback
wellbore.3
8.4 In
8.4.1 Ro
The mech
imposed s
or in-situ 
fractures,
reservoirs
Hydraulic Fracturing of Coalseams 397
sh sands become segregated from the 12/20-mesh sand, each size
 a particular part of the induced and natural fracture system.24 The
 is one of the small particles located at the openings of secondary and
leats and obstructing flow into the cleats, thereby forcing more
 fluid to be diverted into the main induced fracture. The diverted flow
ger widths in the main fracture to accommodate the 12/20-mesh sand.
, not only does the finer fraction of proppant reduce leakoff, but in the
ndirectly helps place the larger proppant in the primary fracture,
ridging in the primary fracture, and reduces tortuosity of the primarysize distribution of proppant helps prevent the movement of sand and
 through the proppant bed to the wellbore. Holditch, et al.37 propose a
of the following: 100-mesh sand for secondary fissure blocking and
tration, followed by 40/70-mesh sand to screen coal fines and proppant
, followed by 20/40-mesh sand to reduce flow resistance near the
7
-Situ Conditions
ck Properties
anical properties of the coal determine the reaction of the rock to
tresses of fracturing. Elastic properties determine the effect of imposed
stresses on existing natural fractures or previously created hydraulic
 directly affecting the permeability of the rock system. In coalbed
, rock mechanical properties and related stresses are of great concern.
Coalbed Methane: Principles and Practices
398 Hydra
Young’s modulus is an elastic property of rock defined by Eq. 8.4 that gives a
measure of fractional elongation as a consequence of stress imposed on the rock.
where
Ex =
σx =
εx =
Young’s m
coal, and i
a fracture
Young’s m
Soft, elast
Converse
constricte
Grove fiel
wellbore.
Some rep
microfract
ten contra
substantia
The surro
thickness 
σ x
x = E (8.4)
ulic Fracturing of Coalseams June 2007
Young’s modulus (psi)
stress, x direction (psi)
strain (x direction)
odulus is important in establishing the width of the fracture in the
t plays a minor role in limiting fracture height. Maximum width, w, of
 near the wellbore is inversely proportional to the fourth power of
odulus38 as in the fracturing model of Geertsma and de Klerk.38
ic coal of low Young’s modulus will be conducive to a wide fracture.
ly, hard formations may be adjacent to the coalseam and have a
d flow path in the fracture.32 Minethrough observations in the Oak
d show sand-propped fractures 1.5 to 2.5 in. wide within 10 ft of the
resentative rock properties of coal and its bounding rock from
ure tests are presented in Table 8.5.39-41 The table illustrates a factor of
st in Young’s modulus, E, of coal and adjacent rock, as well as its
lly higher Poisson’s ratio, v.
unding rock will represent a high percentage of the overall formation
in the multiple, thin seams of basins similar to the Black Warrior.
ε x
)
E
1( w 1/4~
Coalbed Methane: Principles and Practices
June 2007
The high modulus of adjacent rock contrasted with the low modulus of coal will
contribute to confining a fracture in the coal, but the confinement from modulus
is secondary to restraints to fracture growth from in-situ stresses.
Data from
modulus i
unchangin
the modu
laboratory
Table 8
G
Hydraulic Fracturing of Coalseams 399
 van Krevelen42 illustrate the effect of coal maturation on Young’s
n Fig. 8.24. For hvAb-rank coal through lvb-rank, Young’s modulus is
g, but beginning with anthracite, the modulus increases rapidly. Again,
lus is affected by fissures in the rock, and it is difficult to make
 measurements that are representative of field conditions.
.5—Contrasting Elastic Properties of Coal and Bounding Rock39-41
Ecoal
(psi)
Ebounding
(psi) νcoal νbounding
290,000
erman Creek 3,481,000 0.35 0.22
300,000
Bowen Basin 2,320,000 0.39 0.23
400,000
Mary Lee 7,000,000 0.35
0.20
Shale
Coalbed Methane: Principles and Practices
400 Hydra
 
Young’s m
the Black
seams) in
average n
Sparks4 to
100,000–5
al.45 show
account f
formation
the labora
Fig. 8.24
ulic Fracturing of Coalseams June 2007
oduli measured44 from core analyses across the Mary Lee zone and
 Creek zone (formations from Black Creek to Mary Lee/Blue Creek
 Alabama are illustrated as a function of the depth in Fig. 8.25. An
on-coal value of E = 2.5 × 106 psi was determined by Palmer and
 exist across the zones. (Typically, Young’s modulus for coal would be
00,000 psi.37) History matching with the simulator by Lambert, et
ed that a value of Young’s modulus of about 1.3 × 106 psi would best
or pressures encountered during the fracturing.45 Fractures in the
 would effectively reduce Young’s modulus so that core evaluations in
tory supply an upper-limit value.37
—Young's modulus of coal.42,43 
Coalbed Methane: Principles and Practices
June 2007
Poisson’s 
of the lat
longitudin
where
v =
ε2 =
ε1 =
Coal
E
( x 10 psi)6
Depth
(ft)
1,000
Fig. 8.25
Hydraulic Fracturing of Coalseams 401
ratio is an elastic property of rock defined by Eq. 8.5 that is a measure
eral expansion as compared to the longitudinal contraction for a
ally imposed load, the ratio of transverse strain to longitudinal strain.46
Poisson’s ratio
strain or fractional lateral expansion
strain or fractional deformation in longitudinal direction
Siltstone/Shale
Sandstone
M
BC
A
B
C
D
E
FG
H
I
1,050
1,100
1,150
1,200
1,250
1,300
1,350
1,400
1,450
4.5
1.4
3.1
1.7
4.8
3.2
—Young's modulus of Black Creek zone.4,45
ε
εν
1
2- = (8.5)
Coalbed Methane: Principles and Practices
402 Hydra
The sign convention establishes expansion as the negative direction. Poisson’s
ratio for the reservoir rock and surrounding rock influences the stress profile, the
reservoir parameter that defines fracture boundary and orientation. It is a factor in
determining fracture width. Poisson’s ratio and Young’s modulus are essential
for fracture model evaluations.
8.4.2 Str
In-situ mi
large diffe
Coal usua
and in the
order of m
subservien
the fractur
pattern, so
restrict fra
For an id
coalseam,
minimum
stress.
ulic Fracturing of Coalseams June 2007
ess
nimum stress differences of strata limit fracture height growth, and
rences in the strata of Young’s modulus limit fracture height growth.
lly has a much smaller Young’s modulus than the surrounding rock,
 case of the Fruitland coal adjacent to the Pictured Cliffs sandstone, an
agnitude less.24 It has been determined that modulus contrasts are
t to in-situ stresses in limiting fracture height growth. The effect is for
e induced in such strata of different modulus to conform to the stress
 that strata of high stress rather than elastic properties of the rock will
cture height growth.
ealized depiction of high-stress areas confining a fracture to the
 consider Fig. 8.26. A vertical fracture propagates perpendicular to the
 horizontal stress and is limited in height by bounding strata of high
Coalbed Methane: Principles and Practices
June 2007
Fracture h
minifrac 
multiwell 
The result
distance o
stressed M
height gr
confined t
growth of
Overburden
Fig. 8.26
Hydraulic Fracturing of Coalseams 403
eight is controlled by in-situ stresses of the formations. As an example,
tests determined stress variations at the Department of Energy’s
experiment site in the lower Mesaverde group of the Piceance basin.47
s showed a large in-situ stress variation of about 2,000 psi over a short
f 100 ft of formation between the Cozzette sandstone and the highly
ancos shale, seen in Fig. 8.27. The stressed shale would limit fracture
owth if the sandstone were to be fractured; the fracture would be
o the Cozzette. A lateral, high-stress area would pinch out the vertical
 the fracture.48 
High stress
confining height
of fracture
Minimum
horizontal
stress
Stress confines
downward
growth
—Fracture height confined by stresses.
Coalbed Methane: Principles and Practices
404 Hydra
Fig. 8.27
ulic Fracturing of Coalseams June 2007
—In-situ stress measurements.47
Coalbed Methane: Principles and Practices
June 2007
Minimum in-situ stress profiles were established from microfracture tests made
at the Rock Creek site of the Warrior basin.45 The profile for depths of
1,000–1,450 ft spanned the Mary Lee/Blue Creek seams at about 1,200 ft to the
deepest Black Creek seam at approximately 1,415 ft. The stress profile ispresented
Warrior b
Fig. 8.28
Hydraulic Fracturing of Coalseams 405
 in Fig. 8.28. Forty miles from Rock Creek at Moundville in the
asin, stress profiles have been found to be similar.
—Stress profile Black Creek zone.45
Coalbed Methane: Principles and Practices
406 Hydra
Note the high stress in the siltstone/shale interbedded with the lower seams of the
Black Creek group. A fracture initiated through perforations in the lower Black
Creek should not grow downward but possibly extend upward into the Mary
Lee/Blue Creek seams. Fig. 8.28 depicts the fracture that spanned the multiple
seam interval.
After the s
fracture p
Mary Lee
to intercep
groups wa
The stress
possibility
properties
stresses m
Another e
occurs in 
sandstone
than the c
Young’s m
A general
8.29 wher
which in t
the case w
than the l
minimum
This is tru
face cleats
The adven
the hydra
horizontal
ulic Fracturing of Coalseams June 2007
tress profile was obtained, fracturing with crosslinked gel resulted in a
ropagating from the perforations at 1,375–1,383 ft upward into the
/Blue Creek seams, and the fracture propagated downward far enough
t the lowermost Black Creek seams. Communication between the coal
s evident.
 profile over an interval of multiple seams shown in Fig. 8.28 raises the
 of lowering costs of completing and making marginally economical
 profitable by fracturing all the seams of one zone in one operation. The
ust limit the fracture to the desired interval.
xample of the effects of stress contrasts of the coal and bounding strata
the northwestern part of the San Juan basin, where Pictured Cliffs
 below the coalseam at about 2,900 ft has a stress value 746 psi less
oal; the fracture grows across the interface into the sand, even though
odulus of the sandstone is an order of magnitude larger.24 
 indication of the orientation that a fracture will take is given in Fig.
e a vertical fracture develops perpendicular to the least principal stress,
his case is the minimum horizontal stress. Similarly, Fig. 8.29 depicts
here a horizontal fracture is possible if the overburden weight is less
ateral stress, as might be the case in a very shallow coalseam. The
 in-situ stress orientation determines the orientation of the fracture.49
e of the general trend of the fracture. Localized trends follow butt and
 in a highly irregular path.
t of CBM operations with minethrough afforded visual observations of
ulic fracture. Consequently, minethroughs gave insight into when a
 or a vertical fracture would occur.
Coalbed Methane: Principles and Practices
June 2007
Horizonta
fractures o
two, orien
A horizon
interface i
than the 
coefficien
interface 
interface w
and type 
determine
OverburdenOverburden
Fig. 8.29
Hydraulic Fracturing of Coalseams 407
l fractures have been observed shallower than about 750 ft; vertical
ccur in the coalseams deeper than 2,000 ft.38 In between either of the
tations or inclined fractures occur.
tal component of the fracture may be created at the coal and roof rock
f the shear strength, τ, of the interface described by Eq. 8.632 is less
tensional stress of the propagating fracture. Therefore, if a low
t of friction of the interface or a low normal stress acting on the
or the product of these two parameters are present, slippage at the
ill occur to terminate the vertical growth of the fracture. The amount
of fill material at the interface and the rugosity of the two faces
 τo and µf. The normal stress decreases at shallower depths.
(a) Vertical Fracture (b) Horizontal Fracture
—Stresses orient fracture in coals.48
Coalbed Methane: Principles and Practices
408 Hydra
where
τ =
τo =
σn =
µf =
The comb
of shear s
unbonded
CBM seam
been ampl
With the 
componen
seam.24 Fr
the roof of
the Germa
If the coa
stresses of
of the two
propagate
Slippage 
within the
friction an
coals may
Stress pro
stress is a
vertical fra
and width
σμττ nfo + = (8.6)
ulic Fracturing of Coalseams June 2007
shear stress at interface to overcome cohesive and friction forces
cohesive shear strength of interface
normal stress
coefficient of friction
ination of normal stress and friction coefficient that gives a low value
tress will be conducive to the horizontal propagation of the fracture at
 interfaces. If the overburden stress is low, as it is at the depth of many
s, the T-shaped fracture is more likely to occur. The T fracture has
y documented in minethroughs.
relationship of increasing normal stress with depth, the horizontal
t of the T is more often found in the roof than in the floor of the
actures of T shape with a horizontal component have been observed at
 coalseams in the San Juan and Warrior basins of the United States and
n Creek mine of Australia.19,39
l and bounding strata at the interface are bonded and the minimum
 the two strata at the interface are similar, the relative elastic properties
 rocks and strength of the interface, τo, determine whether the fracture
s across the boundary.47
also may occur as the fracturing fluid increases macropore pressure
 coal in the natural fracture system. Thus, by decreasing coefficient of
d allowing coal faces to slip relative to each other, permeability of the
 be permanently altered.10
file is the most important parameter for designing fracture heights. The
lso important in determining proppant embedment, horizontal or
ctures, proppant crushing, surface treating pressures, fracture azimuth,
s of the fracture.32 
Coalbed Methane: Principles and Practices
June 2007
8.4.3 Determining Stress Values
Stress profiles of the coal and other rock strata between coal groups may be
obtained by pump-in microfracture tests. Microfractures involve pumping a
small volume of fluid into the formation and measuring the instantaneous shut-in
pressure (I
method is
restricted 
discrete po
However
stresses to
Two impo
the GRI i
microfra
summariz
1. Iso
2. In
3. Br
4. Ex
5. Af
6. Ta
Hydraulic Fracturing of Coalseams 409
SIP), which is close to the value of the minimum horizontal stress. The
 reliable when used in low-permeability rock having less than 1 md of
leakoff.47 Microfracturing provides stress measurements for the few
ints tested. The procedure is relatively expensive and often neglected.
, an increasing emphasis is being placed on importance of in-situ
 CBM production.
rtant series of in-situ, state-of-stress (ISSOS) tests were conducted for
n the Piceance and Warrior basins.50,51 The steps used in their
cture techniques were similar in each basin. The procedure is
ed as follows:
late the test interval of the formation with straddle packers.
ject 10–20 gal of fresh water at 4–6 gal/min.
eak the formation.
tend the fracture at constant pressure for 1 minute.
ter shut-in, monitor the pressure decline.
ke the ISIP as the minimum horizontal stress.
Coalbed Methane: Principles and Practices
410 Hydra
If the comprehensive pump-in tests require unacceptable time and expense, an
estimate of minimum horizontal stress can be made with Hubbert’s equation (Eq.
8.7).
where
σmin =
v =
σE =
pR =
σz =
To profile
ratio, rese
calculated
if external
state. Whe
horizontal
example, W
the lower
external st
In the mos
of σE in t
principal
compressi
straight l
throughou
then supe
pressures 
line repre
forces can
ν (8.7)
ulic Fracturing of Coalseams June 2007
minimum horizontal stress (psi)
Poisson’s ratio
externally generated stress (psi [must be measured])
reservoir pressure (psi)
overburden stress
 the stresses in the coal zone, Poisson’s ratio is needed. With Poisson’s
rvoir pressure, and overburden stress the horizontal stress may be
 according to linear elastic theory. The calculation would be complete
 horizontal stresses were not presentand if the rock were in a relaxed
n tectonic action or nearby mountain ranges have created significant
 stresses, the calculations without external stresses are not accurate. For
arpinski showed that calculated values of stress from the equation on
 Mesaverde group in the Piceance basin, which is subjected to large
resses, did not match well with measured values.47 
t comprehensive evaluation of Eq. 8.7, Sparks detailed the importance
he Cedar Cove field of Alabama.52 Fig. 8.30 presents the minimum
 stress as calculated from Eq. 8.7 without any contributing
onal tectonic forces, where this calculation is presented as the lower
ine. Closure pressures from microfracture tests in the 400 wells
t the field, as an approximation of the minimum principal stress, were
rimposed on the calculated line of Fig. 8.30. Most of the closure
fall above the calculated base line, and their distance above the base
sents the magnitude of tectonic stress, σE. It is evident that tectonic
not be neglected in most of the Cedar Cove field.
σσνσ ERRz + p + )p - )( - 1( = min
Coalbed Methane: Principles and Practices
June 2007
Poisson’s 
test, or it m
analysis o
lower elas
dynamic t
8.5 V
The inters
opportunit
investigat
3,000
2,
1,
Cl
os
ur
e 
Pr
es
su
re
,
 
ps
ia
Fig. 8.30
Hydraulic Fracturing of Coalseams 411
ratio may be determined from cores stressed in the laboratory in a static
ay be determined on undisturbed coal in place in the formation from
f sonic logs as a dynamic test. Unfortunately, static tests result in a
tic constant, as the cleats and fissures of the coal are not affected in the
ests but are in the static tests.
isual Observation of Fractures
ection of hydraulically induced fractures by mines has afforded the first
y to view fracture characteristics. A study by the U.S. Bureau of Mines
ed the fracture characteristics of 22 stimulation treatments that had
000
000
0
0 1,000 2,000 3,000 4,000
True Measured Depth, ft
PTectonic
W/O PTectonic
1 psi/ft
Hubbert and Willis Equation
—Minimum principal stresses at Cedar Cove.52
Coalbed Methane: Principles and Practices
412 Hydra
been mined through. From those investigations, Diamond and Oyler19 reported
the sand-propped fracturing of a 5.6-ft coalseam with a vertical fracture 0.5 in.
wide. A T-shaped fracture formed at the coal/shale interface of the roof, and the
horizontal fracture component was filled with sand (see Fig. 8.31). No horizontal
componen
Fractures 
in Austra
interface, 
elliptical w
5.6 ft
Fig. 8.31
ulic Fracturing of Coalseams June 2007
t occurred at the floor interface.
of T shape were observed in minethroughs at the German Creek mine
lia.39 The horizontal segment of the fracture occurred at the roof
where most of the proppant was deposited. The horizontal fracture was
ith the major axis in the direction of maximum stress.
Coalbed
Underclay
1 2/ -in. wide
Sand filled
Shale
—Minethrough observation of T fracture.19
Coalbed Methane: Principles and Practices
June 2007
Further documentation of the horizontal component of the fracture at the roof
parting of the coal comes from radioactive proppant tracer used in fracturing
Fruitland coals of the San Juan basin.24 The tracers profile horizontal
components of the fracture at the roof of the coal. Furthermore, the horizontal
fracture is
The vertic
in rock el
observatio
induced f
fracture p
induced cr
modulus f
at a high-s
The offse
interface
downhole
Sparks.4 
Black Cr
basin are p
Extensive
by hydra
CBM we
minethro
generated
with 100-
and docum
reported b
Fig. 8.33.
to extend
and to be 
I , 352 f t
maximum
fracture w
Hydraulic Fracturing of Coalseams 413
 found more often at the top of the coal than at the floor.
al fracture is terminated by a high in-situ stress rather than a difference
astic properties. The phenomenon is indicated in the minethrough
ns of Warpinski.49 In his noncoal application, a hydraulic fracture was
rom a horizontal wellbore in a low modulus formation. The induced
ropagated across the interface without a horizontal component, as the
ack moved in a continuous fashion without offset upon entering a high
ormation. However, the fracture terminated in the downward direction
tress peak in the low modulus formation below.
t of a fracture at the coal
 was also observed in the
 telemetry of Palmer and
Their observations in the
eek coals of the Warrior
resented in Fig. 8.32.
 fractures that were induced
ulic fracturing in vertical
lls have been observed in
ughs. A long fracture,
 by a large water treatment
mesh and 20/40-mesh sand
ented by minethrough, is
y Steidl20 and illustrated in
 The fracture was observed
 525 ft from the wellbore
propped with sand at point
 f rom the wel l .2 0 The
 observed width of the
as 0.3 in.
Fig. 8.32—Downhole camera results.4
Coalbed Methane: Principles and Practices
414 Hydra
O
A B G H I
J
K
Fig. 8.33
ulic Fracturing of Coalseams June 2007
N M
L
2-12-2 CDE F
0 100 200 Feet
Legend
N Well surface location
Well bottom location
Observed fracture
Possible fracture
—Minethrough documents long fracture.20
Coalbed Methane: Principles and Practices
June 2007
References 
1Hunt, A.M. and Steele, D.J.: "Coalbed Methane Development in the Northern 
and Central Appalachian Basins—Past, Present and Future," Proc., Coalbed
Methane Symposium, Tuscaloosa, Alabama (May 1991) 127-141.
2Spafford
terly Rev
15-18.
3Spafford
Restricte
loosa, Al
4Palmer, I
hole TV C
3, 270.
5Khodave
Develop
GRI-91-0
6Jeffrey, R
turing to 
Methane
7HO3679
8HO2289
9Puri, R., 
Hydrauli
Alabama
10Warpins
ruary 19
11Palmer,
Gel-Fra
Proc., C
233-242
Hydraulic Fracturing of Coalseams 415
, S.D. and Schraufnagel, R.A.: "Multiple Coal Seams Project," Quar-
iew of Methane from Coal Seams Technology (July 1992) 10, No. 1,
, S.D.: "Stimulating Multiple Coal Seams at Rock Creek with Access 
d to a Single Seam," Proc., Coalbed Methane Symposium, Tusca-
abama (May 1991) 243-246.
.D. and Sparks, D.P.: "Measurement of Induced Fractures by Down-
amera in Black Warrior Basin Coalbeds," JPT (March 1991) 43, No.
rdian, M., McLennan, J.D., and Jones, A.H.: "Spalling and the 
ment of a Hydraulic Fracturing Strategy for Coal," final report,
234 (April 1991) 43.
.G., Hinkel, J.J., Nimerick, K.H., and McLennan, J.: "Hydraulic Frac-
Enhance Production of Methane from Coal Seams," Proc., Coalbed
 Symposium, Tuscaloosa, Alabama (April 1989) 385-394.
, Halliburton Internal Sales Data Sheet.
, Halliburton Internal Sales Data Sheet.
King, G.E., and Palmer, I.D.: "Damage to Coal Permeability During 
c Fracturing," Proc., Coalbed Methane Symposium, Tuscaloosa,
 (May 1991) 247-255.
ki, N.R.: "Hydraulic Fracturing in Tight, Fissured Media," JPT (Feb-
91) 43, No. 2, 146.
 I.D., Fryar, R.T., Tumino, K.A., and Puri, R.: "Comparison Between 
cture and Water-Fracture Stimulation in the Black Warrior Basin,"
oalbed Methane Symposium, Tuscaloosa, Alabama (May 1991)
.
Coalbed Methane: Principles and Practices
416 Hydra
12Spafford, S.: "Re-Stimulation Treatments for Poorly Performing Wells," 
paper presented at the 1992 Eastern Coalbed Methane Forum, Tuscaloosa,
Alabama, 1 September.
13Spafford
Quarter
10, No.
14Bell, G.J
Hydraul
Methan
15Davidso
ysis of 
Fracture
nology S
16Cleary, 
mature 
tration,"
Regiona
Denver,
17Jones, A
egy for 
(March 
18McLenn
Strategy
ogy (Fe
19Diamon
beds an
U.S. Bu
20Steidl, P
Warrior
loosa, A
21Penny, 
Support
Methan
ulic Fracturing of Coalseams June 2007
, S.D. and Schraufnagel, R.A.: "Multiple Coal Seams Project," 
ly Review of Methane from Coal Seams Technology (October 1992)
 2, 17-21.
., Jones, A.H., Morales, R.H., and Schraufnagel,