Buyon Guo HF Dan Acid Matrix Acidizing

Guo, Boyun / Computer Assited Petroleum Production Engg 0750682701_chap16 Final Proof page 243 21.12.
2006 2:30pm
16 Matrix Acidizing
Contents
16.1 Introduction 16/244
16.2 Acid–Rock Interaction 16/244
16.3 Sandstone Acidizing Design 16/244
16.4 Carbonate Acidizing Design 16/247
Summary 16/248
References 16/248
Problems 16/249
Guo, Boyun / Computer Assited Petroleum Production Engg 0750682701_chap16 Final Proof page 244 21.12.2006 2:30pm
16/244 PRODUCTION ENHANCEMENT
(1)(100:1)
16.1 Introduction b15 ¼ (0:15)
(2)(36:5)
Matrix acidizing is also called acid matrix treatment. It is a
technique to stimulate wells for improving well inflow per- ¼ 0:21 lbm CaCO3 =lbm 15 wt% HCl solution:
formance. In the treatment, acid solution is injected into the
The dissolving power on a volume basis is called volumet-
formation to dissolve some of the minerals to recover per-
ric dissolving power and is related to the gravimetric dis-
meability of sandstones (removing skin) or increase per-
solving power through material densities:
meability of carbonates near the wellbore. After a brief
r
introduction to acid–rock interaction, this chapter focuses X ¼b a, (16:2)
on important issues on sandstone acidizing design and rm
carbonate acidizing design. More in-depth information where
can be found from Economides and Nolte (2000).
X ¼ volumetric dissolving power of acid solution,
ft3 mineral=ft3 solution
ra ¼ density of acid, lbm =ft3
16.2 Acid–Rock Interaction
rm ¼ density of mineral, lbm =ft3
Minerals that are present in sandstone pores include mont-
morillonite (bentonite), kaolinite, calcite, dolomite, sider-
ite, quartz, albite (sodium feldspar), orthoclase, and 16.2.3 Reaction Kinetics
others. These minerals can be either from invasion of The acid–mineral reaction takes place slowly in the rock
external fluid during drilling, cementing, and well comple- matrix being acidized. The reaction rate can be evaluated
tion or from host materials that exist in the naturally experimentally and described by kinetics models. Research
occurring rock formations. The most commonly used work in this area has been presented by many investigators
acids for dissolving these minerals are hydrochloric acid including Fogler et al. (1976), Lund et al. (1973, 1975), Hill
(HCl) and hydrofluoric acid (HF). et al. (1981), Kline and Fogler (1981), and Schechter (1992).
Generally, the reaction rate is affected by the characteristics
of mineral, properties of acid, reservoir temperature, and
16.2.1 Primary Chemical Reactions rates of acid transport to the mineral surface and removal of
Silicate minerals such as clays and feldspars in sandstone product from the surface. Detailed discussion of reaction
pores are normally removed using mixtures of HF and kinetics is beyond the scope of this book.
HCl, whereas carbonate minerals are usually attacked
with HCl. The chemical reactions are summarized in
Table 16.1. The amount of acid required to dissolve a
16.3 Sandstone Acidizing Design
given amount of mineral is determined by the stoichiom-
etry of the chemical reaction. For example, the simple The purpose of sandstone acidizing is to remove the dam-
reaction between HCl and CaCO3 requires that 2 mol of age to the sandstone near the wellbore that occurred dur-
HCl is needed to dissolve 1 mol of CaCO3 . ing drilling and well completion processes. The acid
treatment is only necessary when it is sure that formation
damage is significant to affect well productivity. A major
16.2.2 Dissolving Power of Acids formation damage is usually indicated by a large positive
A more convenient way to express reaction stoichiometry skin factor derived from pressure transit test analysis in a
is the dissolving power. The dissolving power on a mass flow regime of early time (see Chapter 15).
basis is called gravimetric dissolving power and is defined as
nm MWm 16.3.1 Selection of Acid
b ¼ Ca , (16:1)
na MWa The acid type and acid concentration in acid solution used
where in acidizing is selected on the basis of minerals in the
formation and field experience. For sandstones, the typical
b ¼ gravimetric dissolving power of acid treatments usually consist of a mixture of 3 wt% HF and
solution, lbm mineral=lbm solution 12 wt% HCl, preceded by a 15 wt% HCl preflush. McLeod
Ca ¼ weight fraction of acid in the acid solution (1984) presented a guideline to the selection of acid on the
nm ¼ stoichiometry number of mineral basis of extensive field experience. His recommendations
na ¼ stoichiometry number of acid for sandstone treatments are shown in Table 16.2.
MWm = molecular weight of mineral McLeod’s recommendation should serve only as a starting
MWa ¼ molecular weight of acid. point. When many wells are treated in a particular forma-
tion, it is worthwhile to conduct laboratory tests of the
For the reaction between 15 wt% HCl solution responses of cores to different acid strengths. Figure 16.1
and CaCO3 , Ca ¼ 0:15, nm ¼ 1, na ¼ 2, MWm ¼ 100:1, shows typical acid–response curves.
and MWa ¼ 36:5. Thus,
Table 16.1 Primary Chemical Reactions in Acid Treatments

Montmorillonite (Bentonite)-HF/HCl: Al4 Si8 O20 (OH)4 þ 40HF þ 4Hþ $ 4AlFþ
2 þ 8SiF4 þ 24H2 O
Kaolinite-HF/HCl: Al4 Si8 O10 (OH)8 þ 40HF þ 4Hþ $ 4AlFþ
2 þ 8SiF4 þ 18H2 O
Albite-HF/HCl: NaAlSi3 O8 þ 14HF þ 2H $ Na þ AlFþ
þ þ
2 þ 3SiF4 þ 8H2 O
Orthoclase-HF/HCl: KAlSi3 O8 þ 14HF þ 2Hþ $ Kþ þ AlFþ 2 þ 3SiF4 þ 8H2 O
Quartz-HF/HCl: SiO2 þ 4HF $ SiF4 þ 2H2 O
SiF4 þ 2HF $ H2 SiF6
Calcite-HCl: CaCO3 þ 2HCl ! CaCl2 þ CO2 þ H2 O
Dolomite-HCl: CaMg(CO3 )2 þ 4HCl ! CaCl2 þ MgCl2 þ 2CO2 þ 2H2 O
Siderite-HCl: FeCO3 þ 2HCl ! FeCl2 þ CO2 þ H2 O
MATRIX ACIDIZING 16/245
16.3.2 Acid Volume Requirement and

The acid volume should be high enough to remove near-
Vm ¼ p r2a r2w ð1 fÞCm , (16:4)
wellbore formation damage and low enough to reduce cost
of treatment. Selection of an optimum acid volume is VP ¼ p r2a r2w f, (16:5)
complicated by the competing effects. The volume of acid
needed depends strongly on the depth of the damaged where
zone, which is seldom known. Also, the acid will never ra ¼ radius of acid treatment, ft
be distributed equally to all parts of the damaged forma- rw ¼ radius of wellbore, ft
tion. The efficiency of acid treatment and, therefore, acid f ¼ porosity, fraction
volume also depends on acid injection rate. To ensure that Cm ¼ mineral content, volume fraction.
an adequate amount of acid contacts most of the damaged
formation, a larger amount of acid is necessary. Example Problem 16.1 A sandstone with a porosity of
The acid preflush volume is usually determined on the 0.2 containing 10 v% calcite (CaCO3 ) is to be acidized with
basis of void volume calculations. The required minimum HF/HCl mixture solution. A preflush of 15 wt% HCl
acid volume is expressed as solution is to be injected ahead of the mixture to dissolve
the carbonate minerals and establish a low pH
Vm
Va ¼ þ VP þ Vm , (16:3) environment. If the HCl preflush is to remove all
X carbonates in a region within 1 ft beyond a 0.328-ft
where radius wellbore before the HF/HCl stage enters the
formation, what minimum preflush volume is required in
Va ¼ the required minimum acid volume, ft3 terms of gallon per foot of pay zone?
Vm ¼ volume of minerals to be removed, ft3
VP ¼ initial pore volume, ft3
Table 16.2 Recommended Acid Type and Strength for Sandstone Acidizing
HCl Solubility > 20% Use HCl Only
High-perm sand ( k > 100 md)

High quartz (80%), low clay ( <5%) 10% HCl-3% HFa
High feldspar ( >20%) 13.5% HCl-1.5% HFa
High clay ( >10%) 6.5% HCl-1% HFb
High iron chlorite clay 3% HCl-0.5% HFb
Low-perm sand ( k < 10 md)

Low clay ( <5%) 6% HCl-1.5% HFc
High chlorite 3% HCl-0.5% HFd
a
Preflush with 15% HCl.
b
Preflush with sequestered 5% HCl.
c
Preflush with 7.5% HCl or 10% acetic acid.
d
Preflush with 5% acetic acid.
300
8 wt% HF
250
4 wt% HF
Percent of Original Permeability
2 wt% HF
200
150
100
50
Berea
sandstone
80 ⬚F-100 psi
0.1 1 10 100 1,000
Pore Volumes of Acid
Figure 16.1 Typical acid response curves (Smith and Hendrickson, 1965).
Solution lution and reaction product precipitation. Fortunately,

research results have shown that acidizing efficiency is
Volume of CaCO3 to be removed:
relatively insensitive to acid injection rate and that the
Vm ¼ p r2a r2w ð1 fÞCm highest rate possible yields the best results. McLeod
(1984) recommends relatively low injection rates based
¼ p 1:3282 0:3282 ð1 0:2Þ(0:1)
on the observation that acid contact time with the forma-
¼ 0:42 ft3 CaCO3 =ft pay zone tion of 2–4 hours appears to give good results. da Motta
(1993) shows that with shallow damage, acid injection rate
Initial pore volume: has little effect on the residual skin after 100 gal/ft of

VP ¼ p r2a r2w f injection rate; and with deeper damage, the higher the
injection rate, the lower the residual skin. Paccaloni et al.
¼ p 1:3282 0:3282 (0:2) ¼ 1:05 ft3 =ft pay zone
(1988) and Paccaloni and Tambini (1990) also report high
Gravimetric dissolving power of the 15 wt% HCl solution: success rates in numerous field treatments using the high-
y m MWm est injection rates possible.
b ¼ Ca There is always an upper limit on the acid injection rate
ya MWa that is imposed by formation breakdown (fracture) pres-
(1)(100:1) sure pbd . Assuming pseudo–steady-state flow, the max-
¼ (0:15)
(2)(36:5) imum injection rate limited by the breakdown pressure is
¼ 0:21 lbm CaCO3 =lbm 15 wt% HCl solution expressed as

Volumetric dissolving power of the 15 wt% HCl solution: 4:917 106 kh pbd p Dpsf
qi, max ¼ , (16:6)
r ma ln 0:472r
rw þ S
e
X ¼b a
rm
where
(1:07)(62:4)
¼ (0:21)
(169) qi ¼ maximum injection rate, bbl/min
¼ 0:082 ft3 CaCO3 =ft3 15 wt% HCl solution k ¼ permeability of undamaged formation, md
h ¼ thickness of pay zone to be treated, ft
The required minimum HCl volume pbd ¼ formation breakdown pressure, psia
p ¼ reservoir pressure, psia
Vm
Va ¼ þ VP þ Vm Dpsf ¼ safety margin, 200 to 500 psi
X ma ¼ viscosity of acid solution, cp
0:42 re ¼ drainage radius, ft
¼ þ 1:05 þ 0:42
0:082 rw ¼ wellbore radius, ft
¼ 6:48 ft3 15 wt% HCl solution=ft pay zone S ¼ skin factor, ft.
¼ (6:48)(7:48) The acid injection rate can also be limited by surface
¼ 48 gal 15 wt% HCl solution=ft pay zone injection pressure at the pump available to the treatment.
This effect is described in the next section.
The acid volume requirement for the main stage in a mud
acid treatment depends on mineralogy and acid type and
strength. Economides and Nolte (2000) provide a listing of 16.3.4 Acid Injection Pressure
typical stage sequences and volumes for sandstone acidizing In most acid treatment operations, only the surface tubing
treatments. For HCl acid, the volume requirement increases pressure is monitored. It is necessary to predict the surface
from 50 to 200 gal/ft pay zone with HCl solubility of HF injection pressure at the design stage for pump selection.
changing from less than 5% to 20%. For HF acid, the volume The surface tubing pressure is related to the bottom-hole
requirement is in the range of 75–100 gal/ft pay zone with flowing pressure by
3.0–13.5% HCl and 0.5–3.0% HF depending on mineralogy. psi ¼ pwf Dph þ Dpf , (16:7)
Numerous efforts have been made to develop a rigorous
method for calculating the minimum required acid volume where
in the past 2 decades. The most commonly used method is psi ¼ surface injection pressure, psia
the two-mineral model (Hekim et al., 1982; Hill et al., 1981; pwf ¼ flowing bottom-hole pressure, psia
Taha et al., 1989). This model requires a numerical tech- Dph ¼ hydrostatic pressure drop, psia
nique to obtain a general solution. Schechter (1992) pre- Dpf ¼ frictional pressure drop, psia.
sented an approximate solution that is valid for Damkohler
number being greater than 10. This solution approximates The second and the third term in the right-hand side of
the HF fast-reacting mineral front as a sharp front. Readers Eq. (16.7) can be calculated using Eq. (11.93). However, to
are referred to Schechter (1992) for more information. avert the procedure of friction factor determination,
Because mud acid treatments do not dissolve much of the the following approximation may be used for the frictional
formation minerals but dissolve the materials clogging the pressure drop calculation (Economides and Nolte, 2000):
pore throats, Economides and Nolte (2000) suggest taking
the initial pour volume (Eq. [16.5]) within the radius of treat- 518r0:79 q1:79 m0:207
Dpf ¼ L, (16:8)
ment as the minimum required acid volume for the main stage 1,000D4:79
of acidizing treatment. Additional acid volume should be where
considered for the losses in the injection tubing string. r ¼ density of fluid, g=cm3
q ¼ injection rate, bbl/min
m ¼ fluid viscosity, cp
16.3.3 Acid Injection Rate D ¼ tubing diameter, in.
Acid injection rate should be selected on the basis of L ¼ tubing length, ft.
mineral dissolution and removal and depth of damaged
zone. Selecting an optimum injection rate is a difficult Equation (16.8) is relatively accurate for estimating fric-
process because the damaged zone is seldom known with tional pressures for newtonian fluids at flow rates less than
any accuracy and the competing effects of mineral disso- 9 bbl/min.
Example Problem 16.2 A 60-ft thick, 50-md sandstone

pay zone at a depth of 9,500 ft is to be acidized with an
acid solution having a specific gravity of 1.07 and a
viscosity of 1.5 cp down a 2-in. inside diameter (ID) coil
tubing. The formation fracture gradient is 0.7 psi/ft. The
wellbore radius is 0.328 ft. Assuming a reservoir pressure
of 4,000 psia, drainage area radius of 1,000 ft, and a skin
factor of 15, calculate
(a) the maximum acid injection rate using safety margin
300 psi.
(b) the maximum expected surface injection pressure at
the maximum injection rate.
Solution
(a) The maximum acid injection rate:

4:917 106 kh pbd p Dpsf
qi, max ¼
ma ln 0:472r
rw þ S
e
4:917 106 (50)(60)ð(0:7)(9,500) 4,000 300Þ

¼
(1:5) ln 0:472(1,000) þ 15
(0:328) Figure 16.2 Wormholes created by acid dissolution of
¼ 1:04 bbl=min limestone (Hoefner and Fogler, 1988; courtesy AIChE).
(b) The maximum expected surface injection pressure:
pwf ¼ pbd Dpsf ¼ (0:7)(9,500) 300 ¼ 6,350 psia 16.4.2 Acidizing Parameters
Acidizing parameters include acid volume, injection rate,
Dph ¼ (0:433)(1:07)(9,500) ¼ 4,401 psi and injection pressure. The acid volume can be calculated
518r0:79 q1:79 m0:207 with two methods: (1) Daccord’s wormhole propagation
Dpf ¼ L model and (2) the volumetric model, on the basis of desired
1,000D4:79
penetration of wormholes. The former is optimistic, whereas
518(1:07) (1:04)1:79 (1:5)0:207
0:79
the latter is more realistic (Economides et al., 1994).
¼ (9,500)
1,000(2)4:79 Based on the wormhole propagation model presented by
¼ 218 psi Daccord et al. (1989), the required acid volume per unit
thickness of formation can be estimated using the follow-
psi ¼ pwf Dph þ Dpf ing equation:
¼ 6,350 4,401 þ 218 ¼ 2,167 psia 1=3 d
pfD2=3 qh rwhf
Vh ¼ (16:9)
bNAc
16.4 Carbonate Acidizing Design where

The purpose of carbonate acidizing is not to remove Vh ¼ required acid volume per unit thickness
the damage to the formation near the wellbore, but to of formation, m3 =m
create wormholes through which oil or gas will flow after f ¼ porosity, fraction
stimulation. Figure 16.2 shows wormholes created by acid D ¼ molecular diffusion coefficient, m2 =s
dissolution of limestone in a laboratory (Hoefner and qh ¼ injection rate per unit thickness of
Fogler, 1988). formation, m3 =sec-m
Carbonate acidizing is a more difficult process to pre- rwh ¼ desired radius of wormhole penetration, m
dict than sandstone acidizing because the physics is much df ¼ 1:6, fractal dimension
more complex. Because the surface reaction rates are very b ¼ 105 105 in SI units
high and mass transfer often plays the role of limiting NAc ¼ acid capillary number, dimensionless,
step locally, highly nonuniform dissolution patterns are
usually created. The structure of the wormholes depends where the acid capillary number is defined as
on many factors including flow geometry, injection rate, fbg a
NAc ¼ , (16:10)
reaction kinetics, and mass transfer rates. Acidizing de- (1 f)gm
sign relies on mathematical models calibrated by labora-
tory data.
Table 16.3 Recommended Acid Type and

16.4.1 Selection of Acid Strength for Carbonate Acidizing
HCl is the most widely used acid for carbonate matrix
acidizing. Weak acids are suggested for perforating fluid Perforating fluid: 5% acetic acid
and perforation cleanup, and strong acids are recom- Damaged perforations: 9% formic acid
mended for other treatments. Table 16.3 lists recom- 10% acetic acid
mended acid type and strength for carbonate acidizing 15% HCl
(McLeod, 1984). Deep wellbore damage: 15% HCl
All theoretical models of wormhole propagation predict 28% HCl
deeper penetration for higher acid strengths, so a high Emulsified HCl
concentration of acid is always preferable.
where The designed acid volume and injection rate should be

adjusted based on the real-time monitoring of pressure
g a ¼ acid specific gravity, water ¼ 1:0 during the treatment.
g m ¼ mineral specific gravity, water ¼ 1:0.
Based on the volumetric model, the required acid volume
per unit thickness of formation can be estimated using the Summary
following equation: This chapter briefly presents chemistry of matrix acidizing
2
and a guideline to acidizing design for both sandstone and
Vh ¼ pf rwh rw2 ðPV Þbt , (16:11)
carbonate formations. More in-depth materials can be
where (PV )bt is the number of pore volumes of acid found in McLeod (1984), Economides et al. (1994), and
injected at the time of wormhole breakthrough at the end Economides and Nolte (2000).
of the core. Apparently, the volumetric model requires
data from laboratory tests.
References
Example Problem 16.3 A 28 wt% HCl is needed to daccord, g., touboul, e., and lenormand, r. Carbonate
propagate wormholes 3 ft from a 0.328-ft radius wellbore acidizing: toward a quantitative model of the worm-
in a limestone formation (specific gravity 2.71) with a holing phenomenon. SPEPE Feb. 1989:63–68.
porosity of 0.15. The designed injection rate is 0.1 bbl/ da motta, e.p. Matrix Acidizing of Horizontal Wells,
min-ft, the diffusion coefficient is 109 m2 =sec, and the
Ph.d. Dissertation. Austin: University of Texas at
density of the 28% HCl is 1:14 g=cm3 . In linear core
floods, 1.5 pore volume is needed for wormhole Austin, 1993.
breakthrough at the end of the core. Calculate the acid economides, m.j., hill, a.d., and ehlig-economides, c.
volume requirement using (a) Daccord’s model and (b) the Petroleum Production Systems. Englewood Cliffs, NJ:
volumetric model. Prentice Hall, 1994.
economides, m.j. and nolte, k.g. Reservoir Stimulation,
3rd edition. New York: John Wiley & Sons, 2000.
Solution fogler, h.s., lund, k., and mccune, c.c. Predicting the
flow and reaction of HCl/HF mixtures in porous
(a) Daccord’s model: sandstone cores. SPEJ Oct. 1976, Trans. AIME,
1976;234:248–260.
y m MWm (1)(100:1) hekim, y., fogler, h.s., and mccune, c.c. The radial
b ¼ Ca ¼ (0:28)
ya MWa (2)(36:5) movement of permeability fronts and multiple reaction
¼ 0:3836 lbm CaCO3 =lbm 28 wt% HCl solution: zones in porous media. SPEJ Feb. 1982:99–107.
fbga (0:15)(0:3836)(1:14) hill, a.d. and galloway, p.j. Laboratory and theoretical
NAc ¼ ¼ ¼ 0:0285 modeling of diverting agent behavior. JPT June
(1 f)g m (1 0:15)(2:71)
1984:1157–1163.
qh ¼ 0:1 bbl= min -ft ¼ 8:69 104 m3 =sec-m
hill, a.d., lindsay, d.m., silberberg, i.h., and schechter,
rwh ¼ 0:328 þ 3 ¼ 3:328 ft ¼ 1:01 m r.s. Theoretical and experimental studies of sandstone
1=3 d
pfD2=3 qh rwhf acidizing. SPEJ Feb. 1981;21:30–42.
Vh ¼ hoefner, m.l. and fogler, h.s. Pore evolution and channel
bNAc
p(0:15)(109 )2=3 (8:69 104 )1=3 (1:01)1:6 formation during flow and reaction in porous media.
¼ AIChE J. Jan. 1988;34:45–54.
(1:5 105 )(0:0285)
lund, k., fogler, h.s., and mccune, c.c. Acidization I: the
¼ 0:107 m3 =m ¼ 8:6 gal=ft dissolution of dolomite in hydrochloric acid. Chem.
Eng. Sci. 1973;28:691.
(b) Volumetric model:
lund, k., fogler, h.s., mccune, c.c., and ault, j.w.
2
Acidization II: the dissolution of calcite in hydro-
Vh ¼ pf rwh rw2 ðPV Þbt
chloric acid. Chem. Eng. Sci. 1975;30:825.
¼ p(0:15)(3:3282 0:3282 )(1:5) mcleod, h.o., jr. Matrix acidizing. JPT 1984;36:2055–
¼ 7:75 ft3 =ft ¼ 58 gal=ft: 2069.
paccaloni, g. and tambini, m. Advances in matrix stimu-
This example shows that the Daccord model gives opti-
mistic results and the volumetric model gives more realistic lation technology. JPT 1993;45:256–263.
results. paccaloni, g., tambini, m., and galoppini, m. Key factors
The maximum injection rate and pressure for carbon- for enhanced results of matrix stimulation treatment.
ate acidizing can be calculated the same way as that for Presented at the SPE Formation Damage Control
sandstone acidizing. Models of wormhole propagation Symposium held in Bakersfield, California on Febru-
predict that wormhole velocity increases with injection ary 8–9, 1988. SPE Paper 17154.
rate to the power of 1⁄2 to 1. Therefore, the maximum schechter, r.s. Oil Well Stimulation. Englewood Cliffs,
injection rate is preferable. However, this approach may NJ: Prentice Hall, 1992.
require more acid volume. If the acid volume is con-
smith, c.f., and hendrickson, a.r. Hydrofluoric acid
strained, a slower injection rate may be preferable. If a
sufficient acid volume is available, the maximum injection stimulation of sandstone reservoirs. JPT Feb. 1965,
rate is recommended for limestone formations. However, Trans. AIME 1965;234:215–222.
a lower injection rate may be preferable for dolomites. taha, r., hill, a.d., and sepehrnoori, k. Sandstone acid-
This allows the temperature of the acid entering the for- izing design with a generalized model. SPEPE Feb.
mation to increase, and thus, the reaction rate increases. 1989:49–55.
Problems (b) the maximum expected surface injection pressure

at the maximum injection rate.
16.1 For the reaction between 20 wt% HCl solution and
calcite, calculate the gravimetric and volumetric dis- 16.6 A 40-ft thick, 20-md sandstone pay zone at a depth
solving power of the acid solution. of 8,000 ft is to be acidized with an acid solution
16.2 For the reaction between 20 wt% HCl solution and having a specific gravity of 1.07 and a viscosity of
dolomite, calculate the gravimetric and volumetric 1.5 cp down a 2-in. ID coil tubing. The formation
dissolving power of the acid solution. fracture gradient is 0.65 psi/ft. The wellbore radius is
16.3 A sandstone with a porosity of 0.18 containing 8 v% 0.328 ft. Assuming a reservoir pressure of 3,500 psia,
calcite is to be acidized with HF/HCl mixture solu- drainage area radius of 1,200 ft, and skin factor of 15,
tion. A preflush of 15 wt% HCl solution is to be calculate
injected ahead of the mixture to dissolve the carbon-
ate minerals and establish a low-pH environment. If (a) the maximum acid injection rate using a safety
the HCl preflush is to remove all carbonates in a margin of 400 psi.
region within 1.5 ft beyond a 0.328-ft-radius wellbore (b) the maximum expected surface injection pressure
before the HF/HCl stage enters the formation, what at the maximum injection rate.
minimum preflush volume is required in terms of
gallon per foot of pay zone? 16.7 A 20 wt% HCl is needed to propagate wormholes
16.4 A sandstone with a porosity of 0.15 containing 12 v% 2 ft from a 0.328-ft radius wellbore in a limestone
dolomite is to be acidized with HF/HCl mixture so- formation (specific gravity 2.71) with a porosity of
lution. A preflush of 15 wt% HCl solution is to be 0.12. The designed injection rate is 0.12 bbl/min-ft,
injected ahead of the mixture to dissolve the carbon- the diffusion coefficient is 109 m2 =sec, and the den-
ate minerals and establish a low-pH environment. If sity of the 20% HCl is 1:11 g=cm3 . In linear core
the HCl preflush is to remove all carbonates in a floods, 1.2 pore volume is needed for wormhole
region within 1.2 feet beyond a 0.328-ft-radius well- breakthrough at the end of the core. Calculate the
bore before the HF/HCl stage enters the formation, acid volume requirement using (a) Daccord’s model
what minimum preflush volume is required in terms and (b) the volumetric model.
of gallon per foot of pay zone? 16.8 A 25 wt% HCl is needed to propagate wormholes
16.5 A 30-ft thick, 40-md sandstone pay zone at a depth 3 ft from a 0.328-ft radius wellbore in a dolomite
of 9,000 ft is to be acidized with an acid solution having formation (specific gravity 2.87) with a porosity of
a specific gravity of 1.07 and a viscosity of 1.2 cp down 0.16. The designed injection rate is 0.15 bbl/min-ft,
a 2-in. ID coil tubing. The formation fracture gradient the diffusion coefficient is 109 m2 =sec, and the den-
is 0.7 psi/ft. The wellbore radius is 0.328 ft. Assuming sity of the 25% HCl is 1:15 g=cm3 . In linear core
a reservoir pressure of 4,000 psia, drainage area floods, 4 pore volumes is needed for wormhole
radius of 1,500 ft and skin factor of 10, calculate breakthrough at the end of the core. Calculate the
acid volume requirement using (a) Daccord’s model
(a) the maximum acid injection rate using safety and (b) the volumetric model.
margin 200 psi.
Guo, Boyun / Computer Assited Petroleum Production Engg 0750682701_chap17 Final Proof page 251 3.1.2007 9:19pm Compositor Name: SJoearun
17 Hydraulic
Fracturing
Contents
17.1 Introduction 17/252
17.2 Formation Fracturing Pressure 17/252
17.3 Fracture Geometry 17/254
17.4 Productivity of Fractured Wells 17/256
17.5 Hydraulic Fracturing Design 17/258
17.6 Post-Frac Evaluation 17/262
Summary 17/264
References 17/264
Problems 17/265
17.1 Introduction reduce the injection rate requirement, a low leaf-off frac-
turing fluid is essential. Also, to prop the fracture, the sand/
Hydraulic fracturing is a well-stimulation technique that is
proppant should have a compressive strength that is high
most suitable to wells in low- and moderate-permeability
reservoirs that do not provide commercial production enough to resist the stress from the formation.
rates even though formation damages are removed by This chapter concisely describes hydraulic fracturing
acidizing treatments. treatments. For detailed information on this subject, see
Hydraulic fracturing jobs are carried out at well sites Economides and Nolte (2000). This chapter focuses on the
following topics:
using heavy equipment including truck-mounted pumps,
blenders, fluid tanks, and proppant tanks. Figure 17.1
illustrates a simplified equipment layout in hydraulic frac- . Formation fracturing pressure
turing treatments of oil and gas wells. A hydraulic fractur- . Fracture geometry
ing job is divided into two stages: the pad stage and the . Productivity of fractured wells
slurry stage (Fig. 17.2). In the pad stage, fracturing fluid . Hydraulic fracturing design
only is injected into the well to break down the formation . Post-frac evaluation
and create a pad. The pad is created because the fracturing
fluid injection rate is higher than the flow rate at which the
fluid can escape into the formation. After the pad grows to a 17.2 Formation Fracturing Pressure
desirable size, the slurry stage is started. During the slurry Formation fracturing pressure is also called breakdown
stage, the fracturing fluid is mixed with sand/proppant in a pressure. It is one of the key parameters used in hydraulic
blender and the mixture is injected into the pad/fracture. fracturing design. The magnitude of the parameter de-
After filling the fracture with sand/proppant, the fracturing pends on formation depth and properties. Estimation of
job is over and the pump is shut down. Apparently, to the parameter value begins with in situ stress analysis.
Fracturing Proppant
fluid
Blender
Pumper
Figure 17.1 Schematic to show the equipment layout in hydraulic fracturing treatments of oil and gas wells.
Pad stage Slurry stage
Open Closed Create fracture Open Fill fracture with

Open
Open Inject fluid only sand/proppant
Open
Inject frac
fluid/proppant mixture
Closed Closed
Closed
Figure 17.2 A schematic to show the procedure of hydraulic fracturing treatments of oil and gas wells.
HYDRAULIC FRACTURING 17/253
Gas
Oil
Water
ptf
pwt
p
pe
Figure 17.3 Overburden formation of a hydrocarbon reservoir.
Consider a reservoir rock at depth H as shown in Because of the tectonic effect, the magnitude of the hori-
Fig. 17.3. The in situ stress caused by the weight of the zontal stress may vary with direction. The maximum hori-
overburden formation in the vertical direction is expressed zontal stress may be sh, max ¼ sh, min þ stect , where stect is
as called tectonic stress.
rH Based on a failure criterion, Terzaghi presented the
sv ¼ , (17:1) following expression for the breakdown pressure:
144
where pbd ¼ 3sh, min sh, max þ T0 pp , (17:5)
sv ¼ overburden stress, psi where T0 is the tensile strength of the rock.
r ¼ the average density of overburden
formation, lb=ft3 Example Problem 17.1 A sandstone at a depth of 10,000 ft
H ¼ depth, ft. has a Poison’s ratio of 0.25 and a poro-elastic constant of
0.72. The average density of the overburden formation is
The overburden stress is carried by both the rock grains 165 lb=ft3 . The pore pressure gradient in the sandstone is
and the fluid within the pore space between the grains. The 0.38 psi/ft. Assuming a tectonic stress of 2,000 psi and a
contact stress between grains is called effective stress tensile strength of the sandstone of 1,000 psi, predict the
(Fig. 17.4): breakdown pressure for the sandstone.
s0v ¼ sv app , (17:2)
where
s0v ¼ effective vertical stress, psi
a ¼ Biot’s poro-elastic constant,
approximately 0.7
pp ¼ pore pressure, psi.
The effective horizontal stress is expressed as
n
s0h ¼ s0 , (17:3)
1n v
where n is Poison’s ratio. The total horizontal stress is
expressed as
Figure 17.4 Concept of effective stress between
sh ¼ s0h þ app : (17:4) grains.

1
Solution mqi ð1 nÞR 4
ww ¼ 2:56 , (17:6)
Overburden stress: E
rH (165)(10,000) where
sv ¼ ¼ ¼ 11,500 psi ww ¼ fracture width at wellbore, in.
144 144
m ¼ fluid viscosity, cp
Pore pressure: qi ¼ pumping rate, bpm
pp ¼ (0:38)(10,000) ¼ 3,800 psi R ¼ the radius of the fracture, ft
E ¼ Young’s modulus, psi.
The effective vertical stress:
s0v ¼ sv app ¼ 11,500 (0:72)(3,800) ¼ 8,800 psi Assuming the fracture width drops linearly in the radial
direction, the average fracture width may be expressed as
The effective horizontal stress: 1
n 0:25 mqi ð1 nÞR 4
s0h ¼ s0 ¼ ð8,800Þ ¼ 2,900 psi w ¼ 0:85 : (17:7)
1 n v 1 0:25 E
The minimum horizontal stress:
17.3.2 The KGD Model
sh, min ¼ s0h þ app ¼ 2,900 þ (0:72)(3,800) ¼ 5,700 psi Assuming that a fixed-height vertical fracture is propagated
The maximum horizontal stress: in a well-confined pay zone (i.e., the stresses in the layers
above and below the pay zone are large enough to prevent
sh, max ¼ sh, min þ stect ¼ 5,700 þ 2,000 ¼ 7,700 psi fracture growth out of the pay zone), Khristianovich and
Zheltov (1955) presented a fracture model as shown in
Breakdown pressure: Fig. 17.5. The model assumes that the width of the crack
pbd ¼ 3sh, min sh, max þ T0 pp at any distance from the well is independent of vertical
¼ 3(5,700) 7,700 þ 1,000 3,800 ¼ 6,600 psi position, which is a reasonable approximation for a frac-
ture with height much greater than its length. Their solution
included the fracture mechanics aspects of the fracture tip.
They assumed that the flow rate in the fracture was con-
17.3 Fracture Geometry stant, and that the pressure in the fracture could be approxi-
It is still controversial about whether a single fracture or mated by a constant pressure in the majority of the fracture
multiple fractures are created in a hydraulic fracturing job. body, except for a small region near the tip with no fluid
Whereas both cases have been evidenced based on the penetration, and hence, no fluid pressure. This concept of
information collected from tiltmeters and microseismic fluid lag has remained an element of the mechanics of the
data, it is commonly accepted that each individual fracture fracture tip. Geertsma and de Klerk (1969) gave a much
is sheet-like. However, the shape of the fracture varies as simpler solution to the same problem. The solution is now
predicted by different models. referred to as the KGD model. The average width of the
KGD fracture is expressed as
17.3.1 Radial Fracture Model " #1=4
A simple radial (penny-shaped) crack/fracture was first qi mð1 nÞx2f p
presented by Sneddon and Elliot (1946). This occurs w ¼ 0:29 , (17:8)
Ghf 4
when there are no barriers constraining height growth or
when a horizontal fracture is created. Geertsma and de where
Klerk (1969) presented a radial fracture model showing w ¼ average width, in.
that the fracture width at wellbore is given by qi ¼ pumping rate, bpm
Area of highest
flow resistance
xf
w(x,t)
ux
x l
tica
e llip e
r
w(o,t) ate ctu
r o xim of fra
p
Ap hape
rw s
hf
Figure 17.5 The KGD fracture geometry.

m ¼ fluid viscosity, cp models is the requirement to specify the fracture height or

G ¼ E=2(1 þ n), shear modulus, psia to assume that a radial fracture will develop. It is not
hf ¼ fracture height, ft always obvious from data such as logs where, or whether,
the fracture will be contained. In addition, the fracture
17.3.3 The PKN model height will usually vary from the well to the tip of the
Perkins and Kern (1961) also derived a solution for a fixed- fracture, as the pressure varies.
height vertical fracture as illustrated in Fig. 17.6. Nordgren There are two major types of pseudo–three-dimensional
(1972) added leakoff and storage within the fracture (due to (P3D) models: lumped and cell based. In the lumped (or
increasing width) to the Perkins and Kern model, deriving elliptical) models, the fracture shape is assumed to consist
what is now known as the PKN model. The average width of of two half-ellipses joined at the center. The horizontal
the PKN fracture is expressed as length and wellbore vertical tip extensions are calculated

qi mð1 nÞxf 1=4 p at each time-step, and the assumed shape is made to match
w ¼ 0:3 g , (17:9) these positions. Fluid flow is assumed to occur along
G 4
streamlines from the perforations to the edge of the ellipse,
where g 0:75. It is important to emphasize that even for with the shape of the streamlines derived from simple
contained fractures, the PKN solution is only valid when analytical solutions. In cell-based models, the fracture
the fracture length is at least three times the height. shape is not prescribed. The fracture is treated as a series
The three models discussed in this section all assume of connected cells, which are linked only via the fluid flow
that the fracture is planar, that is, fracture propagates in a from cell to cell. The height at any cross-section is calcu-
particular direction (perpendicular to the minimum stress), lated from the pressure in that cell, and fluid flow in the
fluid flow is one-dimensional along the length (or radius) vertical direction is generally approximated.
of the fracture, and leakoff behavior is governed by a Lumped models were first introduced by Cleary (1980),
simple expression derived from filtration theory. The and numerous papers have since been presented on their
rock in which the fracture propagates is assumed to be a use (e.g., Cleary et al., 1994). As stated in the 1980 paper,
continuous, homogeneous, isotropic linear elastic solid, ‘‘The heart of the formulae can be extracted very simply by
and the fracture is considered to be of fixed height (PKN a non-dimensionalization of the governing equations; the
and KGD) or completely confined in a given layer (radial). remainder just involves a good physics-mathematical
The KGD and PKN models assume respectively that the choice of the undetermined coefficients.’’ The lumped
fracture height is large or small relative to length, while the models implicitly require the assumption of a self-similar
radial model assumes a circular shape. Since these models fracture shape (i.e., one that is the same as time evolves,
were developed, numerous extensions have been made, except for length scale). The shape is generally assumed to
which have relaxed these assumptions. consist of two half-ellipses of equal lateral extent, but with
different vertical extent.
In cell-based P3D models, the fracture length is discre-
17.3.4 Three-Dimensional and Pseudo-3D Models tized into cells along the length of the fracture. Because
The planar 2D models discussed in the previous section only one direction is discretized and fluid flow is assumed
are deviated with significant simplifying assumptions. to be essentially horizontal along the length of the fracture,
Although their accuracies are limited, they are useful for the model can be solved much more easily than planar 3D
understanding the growth of hydraulic fractures. The models. Although these models allow the calculation of
power of modern computer allows routine treatment fracture height growth, the assumptions make them pri-
designs to be made with more complex models, which are marily suitable for reasonably contained fractures, with
solved numerically. The biggest limitation of the simple length much greater than height.
Figure 17.6 The PKN fracture geometry.

Planar 3D models: The geometry of a hydraulic fracture is Table 17.1 summarizes main features of fracture models
defined by its width and the shape of its periphery (i.e., height in different categories. Commercial packages are listed in
at any distance from the well and length). The width distri- Table 17.2.
bution and the overall shape change as the treatment is
pumped, and during closure. They depend on the pressure
distribution, which itself is determined by the pressure gra-
dients caused by the fluid flow within the fracture. The 17.4 Productivity of Fractured Wells
relation between pressure gradient and flow rate is very Hydraulically created fractures gather fluids from reser-
sensitive to fracture width, resulting in a tightly coupled voir matrix and provide channels for the fluid to flow into
calculation. Although the mechanics of these processes can wellbores. Apparently, the productivity of fractured wells
be described separately, this close coupling complicates the depends on two steps: (1) receiving fluids from formation
solution of any fracture model. The nonlinear relation be- and (2) transporting the received fluid to the wellbore.
tween width and pressure and the complexity of a moving- Usually one of the steps is a limiting step that controls
boundary problem further complicate numerical solutions. the well-production rate. The efficiency of the first step
Clifton and Abou-Sayed (1979) reported the first numerical depends on fracture dimension (length and height), and
implementation of a planar model. The solution starts with a the efficiency of the second step depends on fracture per-
small fracture, initiated at the perforations, divided into a meability. The relative importance of each of the steps can
number of equal elements (typically 16 squares). The ele- be analyzed using the concept of fracture conductivity
ments then distort to fit the evolving shape. The elements defined as (Argawal et al., 1979; Cinco-Ley and Sama-
can develop large aspect ratios and very small angles, which niego, 1981):
are not well handled by the numerical schemes typically kf w
used to solve the model. Barree (1983) developed a model FCD ¼ , (17:10)
kxf
that does not show grid distortion. The layered reservoir is
divided into a grid of equal-size rectangular elements, over where
the entire region that the fracture may cover. FCD ¼ fracture conductivity, dimensionless
Simulators based on such models are much more com- kf ¼ fracture permeability, md
putationally demanding than P3D-based simulators, be- w ¼ fracture width, ft
cause they solve the fully 2D fluid-flow equations and xf ¼ fracture half-length, ft.
couple this solution rigorously to the elastic-deformation
equations. The elasticity equations are also solved more
rigorously, using a 3D solution rather than 2D slices.
Computational power and numerical methods have im- Table 17.1 Features of Fracture Geometry Models
proved to the point that these models are starting to be A. 2D models
used for routine designs. They should be used whenever a Constant height
significant portion of the fracture volume is outside the Plain strain/stress
zone where the fracture initiates or where there is signifi- Homogeneous stress/elastic properties
cant vertical fluid flow. Such cases typically arise when the Engineering oriented: quick look
stress in the layers around the pay zone is similar to or Limited computing requirements
lower than that within the pay. B. Pseudo-3D (2D 2D) models
Regardless of which type of model is used to calculate the Limited height growth
fracture geometry, limited data are available on typical Planar frac properties of layers/adjacent zones
treatments to validate the model used. On commercial State of stress
treatments, the pressure history during the treatment is Specialized field application
usually the only data available to validate the model. Even Moderate computer requirements
in these cases, the quality of the data is questionable if the C. Fully 3D models
bottom-hole pressure must be inferred from the surface Three-dimensional propagation
pressure. The bottom-hole pressure is also not sufficient Nonideal geometry/growth regimes
to uniquely determine the fracture geometry in the absence Research orientated
of other information, such as that derived from tiltmeters Large database and computer requirements
and microseismic data. If a simulator incorporates the Calibration of similar smaller models in conjunction
correct model, it should match both treating pressure and with laboratory experiments
fracture geometry.
Table 17.2 Summary of Some Commercial Fracturing Models

Software name Model type Company Owner
PROP Classic 2D Halliburton

Chevron 2D Classic 2D ChevronTexaco
CONOCO 2D Classic 2D CONOCO
Shell 2D Classic 2D Shell
TerraFrac Planar 3D Terra Tek ARCO
HYRAC 3D Planar 3D Lehigh U. S.H. Advani
GOHFER Planar 3D Marathon R. Barree
STIMPLAN Pseudo–3D ‘‘cell’’ NSI Technologies M. Smith
ENERFRAC Pseudo–3D ‘‘cell’’ Shell
TRIFRAC Pseudo–3D ‘‘cell’’ S.A. Holditch & Association
FracCADE Pseudo–3D ‘‘cell’’ Schlumberger EAD sugar-land
PRACPRO Pseudo–3D ‘‘parametric’’ RES, Inc. GTI
PRACPROPT Pseudo–3D ‘‘parametric’’ Pinnacle Technologies GTI
MFRAC-III Pseudo–3D ‘‘parametric’’ Meyer & Associates Bruce Meyer
Fracanal Pseudo–3D ‘‘parametric’’ Simtech A. Settari
In the situations in which the fracture dimension is much The previous analyses reveal that low-permeability res-
less than the drainage area of the well, the long-term ervoirs, leading to high-conductivity fractures, would
productivity of the fractured well can be estimated assum- benefit greatly from fracture length, whereas high-perme-
ing pseudo-radial flow in the reservoir. Then the inflow ability reservoirs, naturally leading to low-conductivity
equation can be written as fractures, require good fracture permeability and width.
Valko et al. (1997) converted the data in Fig. 17.7 into
kh pe pwf
q¼ , (17:11) the following correlation:
141:2Bm ln rrwe þ Sf
xf 1:65 0:328u þ 0:116u2
sf þ ln ¼ (17:15)
where Sf is the equivalent skin factor. The fold of increase rw 1 þ 0:180u þ 0:064u2 þ 0:05u3
can be expressed as where
r
J ln e u ¼ ln (FCD ) (17:16)
¼ re rw , (17:12)
Jo ln rw þ Sf
where Example Problem 17.2 A gas reservoir has a permeability
J ¼ productivity of fractured well, stb/day-psi of 1 md. A vertical well of 0.328-ft radius draws the
Jo ¼ productivity of nonfractured well, reservoir from the center of an area of 160 acres. If the
stb/day-psi. well is hydraulically fractured to create a 2,000-ft long,
0.12-in. wide fracture of 200,000 md permeability around
The effective skin factor Sf can be determined based on the center of the drainage area, what would be the fold of
fracture conductivity and Fig. 17.7. increase in well productivity?
It is seen from Fig. 17.7 that the parameter
Sf þ ln xf =rw approaches a constant value in the range Solution Radius of the drainage area:
of FCD > 100, that is, rffiffiffiffi rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
which gives A (43,560)(160)
re ¼ ¼ ¼ 1,490 ft
Sf 0:7 ln xf =rw , (17:13) p p
meaning that the equivalent skin factor of fractured wells Fracture conductivity:
depends only on fracture length for high-conductivity frac- kf w (200,000)(0:12=12)
tures, not fracture permeability and width. This is the FCD ¼ ¼ ¼2
kxf (1)(2,000=2)
situation in which the first step is the limiting step. On
the other Figure 17.7 reads
hand,
Fig. 17.7 indicates that the parameter
Sf þ ln xf =rw declines linearly with log (FCD ) in the Sf þ ln xf =rw 1:2,
range of FCD < 1, that is,
which gives
kf w
Sf 1:52 þ 2:31 logðrw Þ 1:545 log Sf 1:2 ln xf =rw ¼ 1:2 lnð1,000=0:328Þ ¼ 6:82:
k

0:765 log xf : (17:14) The fold of increase is
Comparing the coefficients of the last two terms in this J ln rre ln 1,490
¼ re w ¼ 1,4900:328 ¼ 5:27:
relation indicates that the equivalent skin factor of frac- Jo ln rw þ Sf ln 0:328 6:82
tured well is more sensitive to the fracture permeability
and width than to fracture length for low-conductivity In the situations in which the fracture dimension is com-
fractures. This is the situation in which the second step is parable to the drainage area of the well, significant error
the limiting step. may result from using Eq. (17.12), which was derived based
Figure 17.7 Relationship between fracture conductivity and equivalent skin factor
(Cinco-Ley and Samaniego, 1981).
on radial flow. In these cases, the long-term productivity of . Specifications of fracturing fluid and proppant
the well may be estimated assuming bilinear flow in the . Fluid volume and proppant weight requirements
reservoir. Pressure distribution in a linear flow reservoir . Fluid injection schedule and proppant mixing schedule
and a linear flow in a finite conductivity fracture is illus- . Predicted injection pressure profile
trated in Fig. 17.8. An analytical solution for estimating
fold of increase in well productivity was presented by Guo
and Schechter (1999) as follows: 17.5.1 Selection of Fracturing Fluid
Fracturing fluid plays a vital role in hydraulic fracture treat-
J 0:72 ln rrwe 34 þ So
, ment because it controls the efficiencies of carrying proppant
¼ (17:17)
Jo ðze pffifficffi þ SÞ 1pffi
1pffi and filling in the fracture pad. Fluid loss is a major fracture
c xf 2xf c
1e design variable characterized by a fluid-loss coefficient CL
where c ¼ ze2k and a spurt-loss coefficient Sp . Spurt loss occurs only for
wkf and ze are distance between the fracture
and the boundary of the drainage area. wall-building fluids and only until the filter cake is estab-
lished. Fluid loss into the formation is a more steady process
than spurt loss. It occurs after the filter cake is developed.
17.5 Hydraulic Fracturing Design Excessive fluid loss prevents fracture propagation because of
Hydraulic fracturing designs are performed on the basis of insufficient fluid volume accumulation in the fracture.
parametric studies to maximize net present values (NPVs) Therefore, a fracture fluid with the lowest possible value of
of the fractured wells. A hydraulic fracturing design fluid-loss (leak-off) coefficient CL should be selected.
should follow the following procedure: The second major variable is fluid viscosity. It affects
transporting, suspending, and deposition of proppants, as
1. Select a fracturing fluid
well as back-flowing after treatment. The viscosity should
2. Select a proppant
be controlled in a range suitable for the treatment. A fluid
3. Determine the maximum allowable treatment pressure
viscosity being too high can result in excessive injection
4. Select a fracture propagation model
pressure during the treatment.
5. Select treatment size (fracture length and proppant
However, other considerations may also be major for
concentration)
particular cases. They are compatibility with reservoir
6. Perform production forecast analyses
fluids and rock, compatibility with other materials (e.g.,
7. Perform NPV analysis
resin-coated proppant), compatibility with operating
A complete design must include the following components pressure and temperature, and safety and environmental
to direct field operations: concerns.
2,000
Pressure(psi)
0
1,340
1,260
20
1,180
1,100
180
1,020
340
940
860
500
780
perpendicular to
700
660
the fracture(ft.)
in the direction
620
820
540
460
Distance
980 Distance in
380
fracture direction(ft.)
300
1,140
220
140
1,300
60
0
Figure 17.8 Relationship between fracture conductivity and equivalent skin factor.

17.5.2 Selection of Proppant n rH
s0h ¼ app
Proppant must be selected on the basis of in situ stress 1 n 144
conditions. Major concerns are compressive strength and
0:25 (165)(10,000)
the effect of stress on proppant permeability. For a vertical ¼ (0:7)(2500) ¼ 3,236 psi
1 0:25 144
fracture, the compressive strength of the proppant should
be greater than the effective horizontal stress. In general, Therefore, the minimum required proppant compressive
bigger proppant yields better permeability, but proppant strength is 3,236 psi. Figure 17.9 indicates that the pack of
size must be checked against proppant admittance criteria the intermediate-strength proppants will have a perme-
through the perforations and inside the fracture. Figure ability of about kf ¼ 500 darcies.
17.9 shows permeabilities of various types of proppants
under fracture closure stress. 17.5.3 The maximum Treatment Pressure
The maximum treatment pressure is expected to occur
Example Problem 17.3 For the following situation, esti- when the formation is broken down. The bottom-hole
mate the minimum required compressive strength of 20/ pressure is equal to the formation breakdown pressure
40 proppant. If intermediate-strength proppant is used, pbd and the expected surface pressure can be calculated by
estimate the permeability of the proppant pack: psi ¼ pbd Dph þ Dpf , (17:18)
Formation depth: 10,000 ft
Overburden density: 165 lbm =ft3 where
Poison’s ratio: 0.25 psi ¼ surface injection pressure, psia
Biot constant: 0.7 pbd ¼ formation breakdown pressure, psia
Reservoir pressure: 6,500 psi Dph ¼ hydrostatic pressure drop, psia
Production drawdown: 2,000 and 4,000 psi Dpf ¼ frictional pressure drop, psia.
Solution The second and the third term in the right-hand side of Eq.
(17.18) can be calculated using Eq. (11.93) (see Chapter
The initial effective horizontal stress: 11). However, to avert the procedure of friction factor

n rH determination, the following approximation may be used
s0h ¼ app for the frictional pressure drop calculation (Economides
1 n 144
and Nolte, 2000):
0:25 (165)(10,000)
¼ (0:7)(6500) ¼ 2,303 psi 518r0:79 q1:79 m0:207
1 0:25 144 Dpf ¼ L, (17:19)
1,000D4:79
The effective horizontal stress under 2,000-psi pressure
drawdown: where
r ¼ density of fluid, g=cm3
n rH q ¼ injection rate, bbl/min
s0h ¼ app
1 n 144 m ¼ fluid viscosity, cp
D ¼ tubing diameter, in.
0:25 (165)(10,000)
¼ (0:7)(4500) ¼ 2,770 psi L ¼ tubing length, ft.
1 0:25 144
The effective horizontal stress under 4,000-psi pressure Equation (17.19) is relatively accurate for estimating fric-
drawdown: tional pressures for newtonian fluids at low flow rates.
Figure 17.9 Effect of fracture closure stress on proppant pack permeability

(Economides and Nolte, 2000).
Example Problem 17.4 For Example Problem 17.1, h

rp ¼ (17:25)
predict the maximum expected surface injection pressure hf
using the following additional data: Af ¼ 2xf hf (17:26)
Specific gravity of fracturing fluid: 1.2
Viscosity of fracturing fluid: 20 cp Vfrac
Tubing inner diameter: 3.0 in. h¼ (17:27)
Vinj
Fluid injection rate: 10 bpm
1h
Solution Vpad ¼ Vinj (17:28)
1þh
Hydrostatic pressure drop: Since KL depends on fluid efficiency h, which is not
Dph ¼ (0:433)(1:2)(10,000) ¼ 5,196 psi known in the beginning, a numerical iteration procedure
Frictional pressure drop: is required. The procedure is illustrated in Fig. 17.10.
518r0:79 q1:79 m0:207 3. Generate proppant concentration schedule using:
Dpf ¼ L
1,000D4:79
t tpad «
518(1:2)0:79 (10)1:79 (20)0:207 cp (t) ¼ cf , (17:29)
¼ (10,000) ¼ 3,555 psi tinj tpad
1,000(3)4:79 where cf is the final concentration in ppg. The proppant
Expected surface pressure: concentration in pound per gallon of added fluid (ppga) is
psi ¼ pbd Dph þ Dpf ¼ 6,600 5,196 þ 3,555 expressed as
¼ 4,959 psia cp
c0p ¼ (17:30)
1 cp =rp
17.5.4 Selection of Fracture Model
An appropriate fracture propagation model is selected for the and
formation characteristics and pressure behavior on the basis of 1h
in situ stresses and laboratory tests. Generally, the model «¼ : (17:31)
should be selected to match the level of complexity required 1þh
for the specific application, quality and quantity of data, allo-
cated time to perform a design, and desired level of output. 4. Predict propped fracture width using
Modeling with a planar 3D model can be time consuming, Cp
whereas the results from a 2D model can be simplistic. w¼ , (17:32)
Pseudo-3D models provide a compromise and are most often 1 fp rp
used in the industry. However, 2D models are still attractive where
in situations in which the reservoir conditions are simple and
well understood. For instance, to simulate a short fracture to be Mp
Cp ¼ (17:33)
created inathick sandstone,theKGDmodelmay bebeneficial. 2xf hf
To simulate a long fracture to be created in a sandstone tightly Mp ¼ cp (Vinj Vpad ) (17:34)
bonded by strong overlaying and underlaying shales, the PKN
model is more appropriate. To simulate frac-packing in a thick cf
cp ¼ (17:35)
sandstone, the radial fracture model may be adequate. It is 1þ«
always important to consider the availability and quality of
input data in model selection: garbage-in garbage-out (GIGO). Example Problem 17.5 The following data are given for a
17.5.5 Selection of Treatment Size hydraulic fracturing treatment design:
Treatment size is primarily defined by the fracture length. Pay zone thickness: 70 ft
Fluid and proppant volumes are controlled by fracture length, Young’s modulus of rock: 3 106 psi
injection rate, and leak-off properties. A general statement can Poison’s ratio: 0.25
be made that the greater the propped fracture length and Fluid viscosity: 1.5 cp
greater the proppant volume, the greater the production rate Leak-off coefficient: 0:002 ft= min1=2
of the fractured well. Limiting effects are imposed by technical Proppant density: 165 lb=ft3
and economical factors such as available pumping rate and Proppant porosity: 0.4
costs of fluid and proppant. Within these constraints, the Fracture half-length: 1,000 ft
optimum scale of treatment should be ideally determined
based on the maximum NPV. This section demonstrates how Assume a KL value
to design treatment size using the KGD fracture model for
qiti = Afw + 2KLCLAf rp ti
simplicity. Calculation procedure is summarized as follows:
1. Assume a fracture half-length xf and injection rate qi ,
calculate the average fracture width w
using a selected
fracture model. ti 1 8
KL = h + p(1-h)
2. Based on material balance, solve injection fluid volume 2 3
Vinj from the following equation:
V inj = qiti
Vinj ¼ Vfrac þ VLeakoff , (17:20)
where V frac
V frac = Afw h=
Vinj ¼ qi ti (17:21) V inj
Vfrac ¼ Af w
(17:22) 1−h
V pad = Vinj
pffiffiffi 1+h
VLeakoff ¼ 2KL CL Af rp ti (17:23)

1 8 Figure 17.10 Iteration procedure for injection time
KL ¼ h þ p(1 h) (17:24)
2 3 calculation.
Fracture height: 100 ft Pad volume:

Fluid injection rate: 40 bpm 1 h 1 0:3875
Final proppant concentration: 3 ppg «¼ ¼ ¼ 0:44
1 þ h 1 þ 0:3875
Assuming KGD fracture, estimate Vpad ¼ Vinj « ¼ (6:26 104 )ð0:44Þ ¼ 2:76 104 gal
a. Fluid volume requirement It will take 17 min to pump the pad volume at an injection
b. Proppant mixing schedule rate of 40 bpm.
c. Proppant weight requirement b. Proppant mixing schedule:
d. Propped fracture width
t 17 0:44
Solution cp (t) ¼ (3)
37 17
gives proppant concentration schedule shown in Table
a. Fluid volume requirements: 17.3. Slurry concentration schedule is plotted in Fig. 17.11.
The average fracture width: c. Proppant weight requirement:
" #1=4
qi m(1 n)x2f p cf 3
¼ 0:29
w cp ¼ ¼ ¼ 2:08 ppg
Ghf 4 1 þ « 1 þ 0:44
2 31=4 Mp ¼ cp (Vinj Vpad ) ¼ (2:08)(6:26 104 2:76 104 )
(40)(1:5)(1 0:25)(1,000)2 ¼ 72,910 lb
¼ 0:294 (3106 )
5 ¼ 0:195 in:
4
2(1þ0:25) (70) d. Propped fracture width:
Fracture area:
Mp 72,910
Af ¼ 2xf hf ¼ 2(1,000)(100) ¼ 2 105 ft2 Cp ¼ ¼ ¼ 0:3645 lb=ft3
2xf hf 2(1,000)(100)
Fluid volume based on volume balance: Cp 0:3645
pffiffiffi w¼ ¼ ¼ 0:00368 ft ¼ 0:04 in:
þ 2KL CL Af rp ti :
qi ti ¼ Af w (1 fp )rp (1 0:4)(165)
Assuming KL ¼ 1:5,

0:195
(40)(5:615)ti ¼ (2 105 ) þ 2(1:5)(2 103 ) 17.5.6 Production forecast and NPV Analyses
12 The hydraulic fracturing design is finalized on the basis of

70 pffiffiffi production forecast and NPV analyses. The information
(2 105 ) ti
100
gives ti ¼ 37 min. Table 17.3 Calculated Slurry Concentration
Check KL value: t (min) cp (ppg)
Vinj ¼ qi t ¼ (40)(42)(37) ¼ 6:26 104 gal
0 0
0:195
Vfrac ¼ Af w ¼ (2 105 ) (7:48) ¼ 2:43 104 gal 17 0.00
12 20 1.30
Vfrac 2:43 104 23 1.77
h ¼ ¼ ¼ 0:3875 26 2.11
Vinj 6:26 104 29 2.40
32 2.64
1 3 1 3
KL ¼ h þ p(1 h) ¼ (0:3875) þ p(1 0:3875) 35 2.86
2 8 2 8
37 3.00
¼ 1:48 OK
3.5
3.0
Slurry Concentration (ppg)
2.5
2.0
1.5
1.0
0.5
0.0
0 5 10 15 20 25 30 35 40
Injection Time (min)
Figure 17.11 Calculated slurry concentration.
of the selected fracture half-length xf and the calculated but this operation is quite time consuming and is not
fracture width w, together with formation permeability the goal of the exercise. Perfect matches are sometimes
(k) and fracture permeability (kf ), can be used to predict proposed by manually changing the number of fractures
the dimensionless fracture conductivity FCD with Eq. during the propagation. Unfortunately, there is no inde-
(17.10). The equivalent skin factor Sf can be estimated pendent source that can be used to correlate a variation of
based on Fig. 17.7. Then the productivity index of the the number of fractures. The option of multiple fractures
fractured well can be calculated using Eq. (17.11). Produc- is not available to all simulators. Nevertheless, much pres-
tion forecast can be performed using the method presented sure adjustment can be obtained by changing parameters
in Chapter 7. controlling the near-wellbore effect. Example parameters
Comparison of the production forecast for the fractured are the number of perforations, the relative erosion rate
well and the predicted production decline for the unstimu- of perforation with proppant, and the characteristics
lated well allows for calculations of the annual incremental of fracture tortuosity. These parameters have a major
cumulative production for year n for an oil well: impact on the bottom-hole response but have nothing
to do with the net pressure to be matched for fracture
DNp,n ¼ Npf ,n Npnf,n , (17:36)
geometry estimate.
where
Matching the Net Pressure during Calibration Treat-
DNp,n ¼ predicted annual incremental cumulative
ment and the Pad. The calibration treatment match is part
production for year n
of the set of analysis performed on-site for redesign of the
Npf ,n ¼ forecasted annual cumulative production
injection schedule. This match should be reviewed before
of fractured well for year n
proceeding with the analysis of the main treatment itself.
Npnf,n ¼ predicted annual cumulative production
Consistency between the parameters obtained from both
of nonfractured well for year n.
matches should be maintained and deviation recognized.
If Eq. (17.36) is used for a gas well, the notations DNp,n ,
The first part of the treatment-match process focusing
Npf ,n , and Npnf,n should be replaced by DNp,n , Npf ,n , and Npnf,n ,
on the pad is identical to a match performed on the cali-
respectively.
bration treatment. The shut-in net pressure obtained from
The annual incremental revenue above the one that the a minifrac (calibration treatment decline) gives the magni-
unstimulated well would deliver is expressed as tude of the net pressure. The pad net pressure history (and
DRn ¼ ð$ÞDNp,n , (17:37) low prop concentration in the first few stages) is adjusted
by changing either the compliance or the tip pressure. The
where ($) is oil price. The present value of the future Nolte–Smith Plot (Nolte and Smith, 1981) provides indi-
revenue is then cation of the degree of confinement of the fracture.
X m
DRn A positive slope is an indication of confinement, a negative
NPVR ¼ , (17:38) slope an indication of height growth, and a zero slope an
ð1 þ iÞn
n¼1 indication of toughness-dominated short fracture or mod-
where m is the remaining life of the well in years and i is the erate height growth.
discount rate. The NPV of the hydraulic fracture project is Using 2D Models. In general, when the fracture is
NPV ¼ NPVR cost: (17:39) confined (PKN model) and viscous dominated, we either
decrease the height of the zone or increase the Young’s
The cost should include the expenses for fracturing fluid,
modulus to obtain higher net pressure (compliance is
proppant, pumping, and the fixed cost for the treatment
h=E). For a radial fracture (KGD model), we adjust the
job. To predict the pumping cost, the required hydraulic
tip pressure effect to achieve net pressure match. If the
horsepower needs to be calculated by
fractured formation is a clean sand section and the fracture
qi psi is confined or with moderate height growth, the fracture
HHP ¼ : (17:40)
40:8 height should be fixed to the pay zone. In a layered forma-
tion/dirty sandstone, the fracture height could be adjusted
because any of the intercalated layers may or may not have
17.6 Post-Frac Evaluation been broken down. The fracture could still be confined, but
the height cannot a priori be set as easily as in the case of a
Post-frac evaluation can be performed by pressure match-
clean sand zone section. Unconsolidated sands show low
ing, pressure transient data analysis, and other techniques
Young’s modulus ( 5 105 psi), this should not be
including pumping radioactive materials stages and run-
changed to match the pressure. A low Young modulus
ning tracer logs, running production logging tools, and value often gives insufficient order of magnitude of net
conducting back-pressure and performing Nodal analysis.
pressure because the viscous force is not the dominating
factor. The best way to adjust a fracture elastic model to
17.6.1 Pressure Matching match the behavior of a loosely consolidated sand is to
Pressure matching with a computer software is the first increase the ‘‘apparent toughness’’ that controls the tip
step to evaluate the fracturing job. It is understood that the effect propagating pressure.
more refined the design model is, the more optional Using Pseudo-3D Models. Height constraint is adjusted
parameters we have available for pressure matching and by increasing the stress difference between the pay-zone
the more possible solutions we will get. The importance of and the bounding layer. Stiffness can be increased with an
capturing the main trend with the simplest model possible increase of the Young modulus of all the layers that are
can only be beneficial. Attention should be paid to those fractured or to some extent by adding a small shale layer
critical issues in pressure matching such as fracture con- with high stress in the middle of the zone (pinch-point
finement. Therefore, all the lumped pseudo-3D models effect). Very few commercial fracturing simulators actually
developed for processing speed of pressure-matching ap- use a layer description of the modulus. All of the lumped
plications are widely used. 3D models use an average value. Tip effect can also be
The final result of the net pressure-matching process adjusted by changing toughness (Meyer et al., 1990). For
should ideally be an exact superposition of the simulation some simulators, the users have no direct control of this
on the pumping record. A perfect match is obtainable by effect, as an apparent toughness is recalculated from the
adjusting controlling parameter of a fracture simulator, rock toughness and fluid-lag effect.
Simulating controlled height growth with a pseudo-3D Post-propped Frac Decline. The simulator-generated
model can be tricky. Height growth is characterized by pressure decline is affected by the model of extension
a slower rate of pressure increase than in the case of a recession that is implemented and by the amount of sur-
confined fracture. To capture the big picture, a simplifica- face area that still have leakoff when the simulator cells are
tion to a three-layer model can help by reducing the num- packed with proppant. It is very unlikely that the simula-
ber of possible inputs. Pressure-matching slow height tor matches any of those extreme cases. The lumped solu-
growth of a fracture is tedious and lengthy. In the first tion used in FracProPT does a good job of matching
phase, we should adjust the magnitude of the simulated pressure decline. The analysis methodology was indeed
net pressure. The match can be considered excellent if developed around pressure matching the time to closure.
the difference between the recorded pressure and the The time to closure always relates to the efficiency of the
simulated pressure is less than 15% over the length of fluid regardless of models (Nolte and Smith, 1981).
the pad.
The pressure matching can be performed using data
from real-time measurements (Wright et al., 1996; Burton 17.6.2 Pressure Buildup Test Analysis
et al., 2002). Computer simulation of fracturing operations Fracture and reservoir parameters can be estimated using
with recorded job parameters can yield the following frac- data from pressure transient well tests (Cinco-Ley and
ture dimensions: Samaniego, 1981; Lee and Holditch, 1981). In the pressure
transient well-test analysis, the log-log plot of pressure
. Fracture height derivative versus time is called a diagnostic plot. Special
. Fracture half-length slope values of the derivative curve usually are used
. Fracture width for identification of reservoir and boundary models. The
transient behavior of a well with a finite-conductivity
A typical pressure matching with a pseudo-3D fracturing
fracture includes several flow periods. Initially, there is a
model is shown in Fig. 17.12 (Burton et al., 2002).
fracture linear flow characterized by a half-slope straight
Efficiency and Leakoff. The first estimate of effi- line; after a transition flow period, the system may or
ciency and leakoff is obtained from the calibration treat- may not exhibit a bilinear flow period, indicated by a
ment decline analysis. The calibration treatment provides one-fourth–slope straight line. As time increases, a for-
a direct measurement of the efficiency using the graphical mation linear flow period might develop. Eventually,
G-plot analysis and the 3⁄4 rules or by using time to closure the system reaches a pseudo-radial flow period if the drain-
with a fracturing simulator. Then calibration with a model age area is significantly larger than the fracture dimension
that estimates the geometry of the fracture provides the (Fig. 17.13).
corresponding leakoff coefficient (Meyer and Jacot, 2000). During the fracture linear flow period, most of the
This leakoff coefficient determination is model dependent. fluid entering the wellbore comes from the expansion
Propped Fracture Geometry. Once we have obtained of the system within the fracture. The behavior in the
both a reasonable net pressure match, we have an estimate period occurs at very small amounts of time, normally
of length and height. We can then directly calculate the a few seconds for the fractures created during frac-packing
average width expressed in mass/area of the propped frac- operations. Thus, the data in this period, even if not
ture from mass balance. The propped geometry given by distorted by wellbore storage effect, are still not of prac-
any simulator after closure should not be any different. tical use.
4,000 30
28
26
3,500 24
22
20
Slurry Rate(bbl/min)
18
Prop Conc(PPA)
3,000 BHP (Job data)

16
BHP(psi)
BHP (PropFRAC)
14
&
Slurry Rate (bbl/min)

Prop Conc (PPA) 12
2,500
10
8
6
2,000 4
2
0
1,500 −2
0 5 10 15 20 25 30 35 40
Treatment Time(min)
Figure 17.12 Bottom-hole pressure match with three-dimensional fracturing model PropFRAC.
Bilinear flow:
Fracture
linear flow:
Pseudo-radial flow
Reservoir linear flow
Figure 17.13 Four flow regimes that can occur in hydraulically fractured reservoirs.
The bilinear flow regime means two linear flows occur 17.6.3 Other evaluation techniques
simultaneously. One flow is a linear flow within the frac- In addition to the pressure-matching and pressure buildup
ture and the other is a linear flow in the formation toward data analyses, other techniques can be used to verify the
the fracture. Bilinear flow analysis gives an estimate of fracture profile created during a fracpack operation. These
fracture length and conductivity. A calculated pressure techniques include (1) pumping radioactive materials in
distribution during a bilinear flow is illustrated in the proppant stages and running tracer logs to verify the
Fig. 17.2 (Guo and Schechter, 1999). fracture heights, (2) running production logging tools to
The formation linear flow toward the fracture occurs determine the production profiles, and (3) conducting
after the bilinear flow. Linear flow analysis yields an esti- back-pressure and performing Nodal analysis to verify
mate of formation permeability in the direction perpen- the well deliverability.
dicular to the fracture face. If the test time is long enough
and there is no boundary effect, a system pseudo-radial
flow will eventually occur. Pseudo-radial flow analysis Summary
provides an estimate of formation permeability in the This chapter presents a brief description of hydraulic frac-
radial direction. The reader is referred to Chapter 15 for turing treatments covering formation fracturing pressure,
analysis and interpretation of pressure transient data. fracture geometry, productivity of fractured wells,
It is important to note that by no means does the pressure- hydraulic fracturing design, and post-frac evaluation.
match procedure and the pressure transient data analysis
More in-depth discussions can be found from Economides
give details of the fracture geometry such as fracture width
et al. (1994) and Economides and Nolte (2000).
near the wellbore, which frequently dominates the post-
treatment well performance. The fracture width near the
wellbore can be significantly lower than that in the region
away from the wellbore. This can occur because of a number References
of mishaps. Overdisplacement of proppant leads to the frac- argawal, r.g., carter, r.d., and pollock, c.b.
ture unsupported near the wellbore, resulting in fracture Evaluation and prediction of performance of low-
closure. Fluid backflow reduces the amount of proppant permeability gas wells stimulated by massive hydraulic
near the wellbore, which results in less fracture width sup- fracturing. JPT March 1979, Trans. AIME
ported. If the proppant grains do not have compressive 1979;267:362–372.
strength to withstand the stress concentration in the near-
wellbore region, they will be crushed during fracture closure, barree, r.d. A practical numerical simulator for three
resulting in tight fracture near the wellbore. The reduced dimensional fracture propagation in heterogeneous
fracture width near the wellbore affects well productivity media. Proceedings of the Reservoir Simulation Sym-
because of the fracture choking effect. Post-treatment flow posium, San Francisco, CA, 403-411 Nov. 1983. SPE
tests should be run to verify well performance. 12273.
The effect of near-wellbore fracture geometry on post- burton, r.c., davis, e.r., hodge, r.m., stomp, r.j., palthe,
treatment well production is of special significance in p.w., and saldungaray, p. Innovative completion
deviated and horizontal wells (Chen and Economides, design and well performance evaluation for effective
1999). This is because a fracture from an arbitrarily
Frac-packing of long intervals: a case study from the
oriented well ‘‘cuts’’ the wellbore at an angle, thereby
West Natuna Sea, Indonesia. Presented at the SPE
limiting the communication between the wellbore and the
reservoir. This feature of fluid entry to the wellbore itself International Petroleum Conference and Exhibition
causes the fracture-choking effect, even though the near- held 10–12 February 2002, in Villahermosa, Mexico.
wellbore fracture is perfectly propped. Certainly, a hori- Paper SPE 74351.
zontal well in the longitudinal to the fracture direction and chen, z. and economides, m.j. Effect of near-wellbore
using 180-degree perforation phasing that can be oriented fracture geometry on fracture execution and post-treat-
will eliminate the problem. However, to align the horizon- ment well production of deviated and horizontal wells.
tal wellbore in the longitudinal to the fracture direction, SPE Prod. Facilities August 1999.
the horizontal wellbore has to be drilled in the direction cinco-ley, h. and samaniego, f. Transient pressure
parallel to the maximum horizontal stress direction. The
orientation of the stress can be obtained by running tests in analysis for fractured wells. J. Petroleum Technol.
a vertical pilot hole of the horizontal well. Special log September 1981.
imaging (e.g., FMI and FMS) can be run in combination cleary, m.p. Comprehensive design formulae for
with an injection test at small-rate MDT or large-scale hydraulic fracturing. Presented at the SPE Annual
minifrac to fracture the formation and read directly the Technology Conference held in Dallas, Texas, Septem-
image in the wellbore after the fracture has been created. ber 1980. SPE 9259.
cleary, m.p., coyle, r.s., teng, e.y., cipolla, c.l., 162 lb=ft3 . The pore–pressure gradient in the sand-
meehan, d.n., massaras, l.v., and wright, t.b. stone is 0.36 psi/ft. Assuming a tectonic stress of
Major new developments in hydraulic fracturing, with 1,000 psi and a tensile strength of the sandstone of
documented reductions in job costs and increases in 800 psi, predict the breakdown pressure for the sand-
stone.
normalized production. Presented at the 69th Annual 17.2 A carbonate at a depth of 12,000 ft has a Poison’s
Technical Conference and Exhibition of the Society of ratio of 0.3 and a poro-elastic constant of 0.75. The
Petroleum Engineers, held in New Orleans, Louisiana, average density of the overburden formation is
25–28 September 1994. SPE 28565. 178 lb=ft3 . The pore–pressure gradient in the sand-
clifton, r.j. and abou-sayed, a.s. On the computation of stone is 0.35 psi/ft. Assuming a tectonic stress of
the three-dimensional geometry of hydraulic fractures. 2,000 psi and a tensile strength of the sandstone of
Presented at the SPE/DOE Low Perm. Gas Res. Sympo- 1,500 psi, predict the breakdown pressure for the
sium, held in Denver, Colorado, May 1979. SPE 7943. sandstone.
17.3 A gas reservoir has a permeability of 5 md. A vertical
economides, m.j., hill, a.d., and ehlig-economides, c.
well of 0.328-ft radius draws the reservoir from the
Petroleum Production Systems, Upper Saddle River,
center of an area of 320 acres. If the well is hydraul-
New Jersey, Prentice Hall PTR, 1994. ically fractured to create a 2,000-ft long, 0.15-in. wide
economides, m.j. and nolte, k.g. Reservoir Stimulation, fracture of 200,000-md permeability around the cen-
3rd edition. New York: John Wiley & Sons, 2000. ter of the drainage area, what would be the fold of
geertsma, j. and de klerk, f. A rapid method of predict- increase in well productivity?
ing width and extent of hydraulic induced fractures. 17.4 A reservoir has a permeability of 100 md. A vertical
J. Petroleum Technol. Dec. 1969;21:1571–1581. well of 0.328-ft radius draws the reservoir from the
guo, b. and schechter, d.s. A simple and rigorous IPR center of an area of 160 acres. If the well is hydraul-
equation for vertical and horizontal wells intersecting ically fractured to create a 2,800-ft long, 0.12-in. wide
fracture of 250,000-md permeability around the cen-
long fractures. J. Can. Petroleum Technol. July 1999. ter of the drainage area, what would be the fold of
khristianovich, s.a. and zheltov, y.p. Formation of increase in well productivity?
vertical fractures by means of highly viscous liquid. 17.5 For the following situation, estimate the minimum
In: Proceedings of the SPE Fourth World Petroleum required compressive strength of 20/40 proppant. If
Congress held in Rome, Section II. 1955, pp. 579–586. high-strength proppant is used, estimate the perme-
lee, w.j. and holditch, s.a. Fracture evaluation with ability of the proppant pack:
pressure transient testing in low-permeability gas res-
ervoirs. J. Petroleum Technol. September 1981. Formation depth: 12,000 ft
Overburden density: 165 lbm =ft3
meyer, b.r., cooper, g.d., and nelson, s.g. Real-time 3-D
Poison’s ratio: 0.25
hydraulic fracturing simulation: theory and field case
Biot constant: 0.72
studies. Presented at the 65th Annual Technical Con- Reservoir pressure: 6,800 psi
ference and Exhibition of the Society of Petroleum Production drawdown: 3,000 psi
Engineers, held in New Orleans, Louisiana, 23–26 Sep-
tember 1990. Paper SPE 20658. 17.6 For the Problem 17.5, predict the maximum expected
meyer, b.r. and jacot, r.h. Implementation of fracture surface injection pressure using the following addi-
tional data:
calibration equations for pressure dependent leakoff.
Presented at the 2000 SPE/AAPG Western Regional Specific gravity of fracturing fluid: 1.1
Meeting, held in Long Beach, California, 19–23 June Viscosity of fracturing fluid: 10 cp
2000. Paper SPE 62545. Tubing inner diameter: 3.0 in.
nolte, k.g. and smith, m.b. Interpretation of fracturing Fluid injection rate: 20 bpm
pressures. J. Petroleum Technol. September 1981.
nordgren, r.p. Propagation of vertical hydraulic fracture. 17.7 The following data are given for a hydraulic fractur-
ing treatment design:
SPEJ Aug. 1972:306–314.
perkins, t.k. and kern, l.r. Width of Hydraulic Fracture. Pay zone thickness: 50 ft
J. Petroleum Technol. Sept. 1961:937–949. Young’s modulus of rock: 4 106 psi
sneddon, i.n. and elliott, a.a. The opening of a griffith Poison’s ratio: 0.25
crack under internal pressure. Quart. Appl. Math. Fluid viscosity: 1.25 cp
1946;IV:262. Leakoff coefficient: 0:003 ft= min1=2
valko, p., oligney, r.e., economides, m.j. High permeability Proppant density: 185 lb=ft3
fracturing of gas wells. Gas TIPS. October 1997;3:31–40. Proppant porosity: 0.4
Fracture half length: 1,200 ft
wright, c.a., weijers, l., germani, g.a., maclvor, k.h.,
Fracture height: 70 ft
wilson, m.k., and whitman, b.a. Fracture treatment
Fluid injection rate: 35 bpm
design and evaluation in the Pakenham field: a real- Final proppant concentration: 5 ppg
data approach. Presented at the SPE Annual Technical
Conference and Exhibition, held in Denver, Colorado, Assuming KGD fracture, estimate
6–9 October 1996. Paper SPE 36471. a. Fluid volume requirement
b. Proppant mixing schedule
Problems c. Proppant weight requirement
d. Propped fracture width
17.1 A sandstone at a depth of 8,000 ft has a Poison’s
ratio of 0.275 and a poro-elastic constant of 0.70. 17.8 Predict the productivity index of the fractured well
The average density of the overburden formation is described in Problem 17.7.

Buyon Guo HF Dan Acid Matrix Acidizing

Încărcat de

Informații document

Descriere originală:

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Buyon Guo HF Dan Acid Matrix Acidizing

Încărcat de

Drepturi de autor:

Formate disponibile

Guo, Boyun / Computer Assited Petroleum Production Engg 0750682701_chap16 Final Proof page 243 21.12.

16/244 PRODUCTION ENHANCEMENT

Table 16.1 Primary Chemical Reactions in Acid Treatments

MATRIX ACIDIZING 16/245

16.3.2 Acid Volume Requirement and

High-perm sand ( k > 100 md)

Low-perm sand ( k < 10 md)

16/246 PRODUCTION ENHANCEMENT

Solution lution and reaction product precipitation. Fortunately,

MATRIX ACIDIZING 16/247

Example Problem 16.2 A 60-ft thick, 50-md sandstone

(a) The maximum acid injection rate:

4:917 106 (50)(60)ð(0:7)(9,500) 4,000 300Þ

(b) The maximum expected surface injection pressure:

16.4 Carbonate Acidizing Design where

Table 16.3 Recommended Acid Type and

16/248 PRODUCTION ENHANCEMENT

where The designed acid volume and injection rate should be

MATRIX ACIDIZING 16/249

Problems (b) the maximum expected surface injection pressure

17/252 PRODUCTION ENHANCEMENT

Pad stage Slurry stage

Open Closed Create fracture Open Fill fracture with

HYDRAULIC FRACTURING 17/253

Figure 17.3 Overburden formation of a hydrocarbon reservoir.

17/254 PRODUCTION ENHANCEMENT

Figure 17.5 The KGD fracture geometry.

HYDRAULIC FRACTURING 17/255

m ¼ fluid viscosity, cp models is the requirement to specify the fracture height or

Figure 17.6 The PKN fracture geometry.

17/256 PRODUCTION ENHANCEMENT

Table 17.2 Summary of Some Commercial Fracturing Models

PROP Classic 2D Halliburton

HYDRAULIC FRACTURING 17/257

17/258 PRODUCTION ENHANCEMENT

HYDRAULIC FRACTURING 17/259

Figure 17.9 Effect of fracture closure stress on proppant pack permeability

17/260 PRODUCTION ENHANCEMENT

Example Problem 17.4 For Example Problem 17.1, h

HYDRAULIC FRACTURING 17/261

Fracture height: 100 ft Pad volume:

17/262 PRODUCTION ENHANCEMENT

HYDRAULIC FRACTURING 17/263

3,000 BHP (Job data)

Slurry Rate (bbl/min)

17/264 PRODUCTION ENHANCEMENT

HYDRAULIC FRACTURING 17/265

S-ar putea să vă placă și