SPE-178961-MS

Theoretical and Experimental Study on Optimal Injection Rates in

Carbonate Acidizing
K. Dong, D. Zhu, and A. D. Hill, Texas A&M University

Copyright 2016, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE International Conference & Exhibition on Formation Damage Control held in Lafayette, Louisiana, USA, 24 –26
February 2016.

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

Abstract
Optimal acid injection rate is critical information for carbonate matrix acidizing design. This rate is
currently obtained through fitting acidizing coreflood experimental results. A model is needed to predict
optimal acid injection rates for various reservoir conditions.
A wormhole forms when larger pores grow in cross-sectional area at a rate that greatly exceeds the
growth rate of smaller pores due to surface reaction. This happens when the pore growth follows a
particular mechanism, which is discussed in this paper. We have developed a model to predict wormhole
growth behavior. The model uses the mode size in a pore size distribution - the pore size that appears most
frequently in the distribution, to predict the growth of the pore. By controlling the acid velocity inside of
it, we can make this particular pore grow much faster than other smaller pores, thus reaching the most
favorable condition for wormholing. This also results in a balance between overall acid/rock reaction and
acid flow. With the introduction of a porous medium model, the acid velocity in the mode-size pore is
upscaled to the interstitial velocity at the wormhole tip. This interstitial velocity at the wormhole tip
controls the wormhole propagation. The optimal acid injection rates are then calculated based on
semi-empirical flow correlations for different flow geometries.
The optimal injection rate depends on the rock lithology, acid concentration, temperature, and rock
pore size distribution. All of these factors are accounted for in this model. The model can predict the
optimal rates of acidizing coreflood experiments correctly, as compared with our acidizing coreflood
experimental results. In addition, based on our model, it is also found that at optimal conditions, the
wormhole propagation velocity is linearly proportional to the acid diffusion coefficient for a diffusion
limited reaction. This is proven both experimentally and theoretically in this study. Since there is no flow
geometry constraint while developing this model, it can be applied to field scales. Applications are
presented in this paper.

Introduction
Experimental results have shown that the best wormholing efficiency can be obtained at a certain acid
injection rate. We call this the optimal conditions. Extensive work has been done to study this particular
condition, both experimentally and theoretically.
2 SPE-178961-MS

Acidizing coreflood experiments are an important method to determine the optimal conditions for a
specific acid/carbonate system. The experimental setup usually consists of a syringe pump, acid/brine
accumulators, a coreholder and a backpressure regulator. Each experiment is carried out with a specific
acid injection rate, and the corresponding breakthrough acid volume is measured. Repeated experiments
with different acid injection rates are needed to find the optimal condition.
Through acidizing coreflood experiments, several investigators have studied factors that affect the
optimal conditions. These factors include reservoir temperature, HCl concentration, and carbonate rock
types. In general, increasing temperature can make both the optimal acid flux and the corresponding
optimal breakthrough pore volumes become larger. Higher acid concentration results in larger optimal
acid flux and lower optimal breakthrough pore volumes. The optimal breakthrough pore volumes for
dolomite is larger than that of limestone.
Other than experiments, different models for predicting optimal conditions have also been developed.
Wang et al. (1993) developed a model using the largest pores naturally existing in a rock as a parameter
for wormhole growth. A transition pore area was defined, that was used to distinguish the growth
mechanisms of small pores and large pores. If the area of a pore is larger than this transition pore area,
this pore grows rapidly and a wormhole can form. To some extent, the model can predict the optimal acid
injection rate for linear coreflood acidizing experiments. However, the surface reaction rate is used to
represent the overall reaction rate in the model, with diffusion rate being ignored. This leads to an
incorrect prediction of temperature effect.
Gong and El-Rabaa (1999) published a correlation to calculate the optimal conditions using flow/
reaction dimensionless numbers. Fundamentally, this model is based on the model described by Daccord,
G. et al. (1989). They calculated the derivatives of the diffusion limited relationship and made it equal to
zero. The idea is that the minimum point of a curve has a zero derivative. The optimal injection rate
calculated from this model is orders of magnitude less than the experimental results.
Panga et al. (2005) studied different conditions of dissolution patterns based on the ratio of transverse
to axial length scales. He showed that when the transverse length scale and the axial length scale are of
the same order, the optimal condition happens. Furthermore, he studied the optimal conditions for
kinetically controlled reaction and mass transport controlled reaction separately. However, experimental
and theoretical studies have shown that a kinetically controlled reaction produce uniform dissolution but
not wormholing dissolution.
Fredd and Fogler (1999) studied the effects of transport and reaction on wormholing process using a
wide range of reactive fluids. The optimal conditions were identified for low concentration HCl, EDTA,
CDTA, DTPA and HAc respectively with Indiana limestone. They defined the Damkohler number as the
ratio between the overall acid reaction rate at the wormhole wall and the acid flow rate in the wormhole.
They found that when the Damkohler number equals to 0.29, the optimal condition can be achieved for
all the reactive fluids they studied. However, in order to have this Damkohler number available, a
pre-existing wormhole diameter and length need to be identified.
With more experimental data available oday, more physics can be unveiled for wormhole propagation.
The development of a reliable model is possible.
Pore Growth
When acid is introduced into a rock, it reacts with the minerals and changes its pore structure. The macro
properties like porosity and permeability of the rock change accordingly. In order to describe this
phenomenon, Schechter and Gidley (1969) studied the changes of pore structure and pore size distribution
due to surface reaction. They set up a porous medium model with pores represented by capillaries
distributed randomly. Pore enlargement is described by a pore growth function. The details about the
porous medium model are illustrated in Appendix A. They concluded that it is the larger pores that
determine the response of rocks to acid attack for high surface reaction rates, and this response is sensitive
SPE-178961-MS 3

to the distribution of these larger pores. In general, pore growth rate, namely pore growth function ␺ in
this paper, can be written as Eq. 1, with n less than ½ and larger than ⫺1.
(1)

␺ can be written as Eq. 2 by solving a mass balance equation with acid flowing inside a single pore
(Schechter, 1992).
(2)

where Ap is the pore cross-sectional area, Lp is the pore length, ⌫p is the pore perimeter, is the average
acid velocity in the pore, ␹ is the acid volumetric dissolving power and ␬ is the overall acid reaction rate
coefficient. The overall acid reaction rate depends on three individual processes: acid diffusing to the pore
surface, acid reacting with the pore surface and product diffusing away from the pore surface. In this
paper, only the first two processes are considered due to the complete reaction between HCl and calcite.
The derivation of ␬ and its approximation are shown in Appendix B. For reactions between carbonate
rocks and HCl, the surface reaction rate is normally much larger than the mass transfer rate. From
Appendix B, the overall reaction rate coefficient ␬ can be approximated as the mass transfer rate K,
(3)

where D is the acid diffusion coefficient.

To further expand Eq. 2, recall that the average acid velocity is proportional to the pore crosssec-
p
tional area A , as described by Eq. A-6. For consistency, it is brought here as shown by Eq. 4.
(4)

where ␧ is a factor which depends on the pressure gradient and acid viscosity. Substituting Eq. 3 and
Eq. into Eq. 2, we can get
(5)

The exponential term in Eq. 2 and Eq. 5 denotes the ratio between overall reaction rate and acid flow
rate in a pore, which is Damköhler (Da) number. For a small pore, Da is large and the exponential term
is close to 0. The pore growth function ␺ is approximated as
(6)

For a large pore, Da is small, the exponential term is close to 1. The pore growth function ␺ is
approximated as
(7)

Substituting Eq. 3 into Eq. 7, we have

(8)
4 SPE-178961-MS

Notice that, from Eqs. 6 and 8, the pore growth function for small pore is linearly propotional to Ap2
(␺smal1~ Ap2) and large pore, the pore growth function is linear propotional to Ap2/3(␺smal1~Ap2/3). Fig.
1 shows the plot of pore growth rate versus pore area. The black solid curve is the plot of Eq. 5. The red
dash line is the plot of Eq. 6 and green dash line is of Eq. 8. We can see that there is a transition point
that divides the curve into two parts, one with a slope of 2 and the other one with a slope of 2/3. The
transition pore area depends on the magnitude of the Da number.

Figure 1—Pore growth rates for pores with different cross-sectional areas Ap.

Whether or not a wormhole forms depends on how pores grow. If larger pores grow significantly faster
than other pores, a wormhole forms. This happens when n in Eq. 1 is less than 0. Examples are illustrated
in Fig. 2. With identical parameters, we simulate the pore growth for Eq. 6 (left plot) and Eq. 7 (right plot).
The simulation finishes when the largest pore area reaches 10 times its original area. We can see that Eq.
6 predicts that larger pores grow faster than smaller pores, while Eq. 7 predicts that smaller pores grow
much faster than larger pores. Further calculation shows that n equals to ⫺1 (␺~A2) is the most favorable
conditions for wormholing because the largest pore grows fastest and other smaller pores grow slowest
compared with other conditions.
SPE-178961-MS 5

Figure 2—Pore area increase due to acid attack for Eq. 6 (left plot) and Eq. 7 (right plot).

Fundamental Explanation to Optimal Conditions

Acids always tend to flow into larger pores due to the lower flow resistance in large pores. However,
larger pores may not connect to another larger pore inside the rock. While wormhole is propagating, such
larger pores cannot always be expected at the wormhole tip. So in this study, we select the mode-size pore
in a rock as the parameter in the wormhole model. Mode-size pores are the pores with the largest
frequency in a pore size distribution (its area is denoted as Ap,mode in this paper). During wormhole
propagation, wormhole tip is more likely to meet this pore size than any other pore sizes.
At any particular time and position, because mode-sized pores are more than any other sized pores, we
assume that larger amounts of acids flow into the mode-size pores. In order to get the optimal condition,
mode-size pores need to grow much faster than other pores at wormhole tips. As has been discussed, when
pores grow based on the relationship of ␺~Ap2, the optimal wormholing condition is achieved. Therefore,
we set the acid flow rate to a particular value so that the mode-size pore grows based on ␺~Ap2. Being
such, all other smaller pores at this position also grow based on ␺~Ap2 according to Fig. 1. But the mode-
size pores grow significantly faster than other smaller pores.
Mode pore size is naturally existing for a particular rock, but the transition pore size depends on the
acid flow rate and it can be controlled by change flow velocity according to Eq. 2. In order to have this
mode-size pore grow based on ␺~Ap2, we can make the transition pore size shown in Fig. 1 equal to the
mode-pore size.
(9)

(10)

At the transition point, if we set ␺small equal to ␺large, then

(11)

thus,
(12)
6 SPE-178961-MS

The average acid velocity for mode-size pores calculated by Eq. 12 is just large enough to make all
pores grow based on ␺~Ap2 at random positions of a rock. The pores of mode-size grow much faster than
other smaller pores, as shown in the left plot of Fig. 2. This is the optimal condition of wormhole
propagation. Replace the notation of Eq. 12,
(13)

We can also notice that the Da number equals to 1 at this particular condition, implying that the amount
of acid reacted with the pore equals to the amount of acid that is injected into the pore.
(14)

Optimal Conditions at Wormhole Tip

In order to upscale the optimal velocity in the mode-size pore to the flow rate in the porous medium, the
porous medium model shown in Appendix A is used. The total volumetric flow rate across a unit
crosssectional area can be calculated by integrating the volumetric flow rate in each pore. Specifically, if
we focus on the wormhole tip flow area, the acid flow rate at the wormhole tip
(15)

where Atip is the flow cross-sectional area at the wormhole tip. As stated in Appendix A, the acid
velocity across each pore is linearly proportional to the pore area based on Poiseuille’s law. Specifically,
as we have determined the optimal acid velocity in the mode-size pore, we can calculate this proportional
coefficient.
(16)

Substituting Eq. 16 into Eq. 15, we can get the optimal acid flow rate at the wormhole tip, and also the
corresponding optimal tip acid flux
(17)

where M2 is defined as the second moment of the pore size density function and is calculated through
Eq. 18.
(18)

The optimal tip acid flux is then calculated as

(19)

As we can see from Eq. 19, the optimal tip acid flux depends on acid flow velocity in the mode-size
pore, rock porosity and pore size distribution. The mode pore size and pore size distribution can be
obtained together by a single measurement. In this study, Micro-CT Scanner is used for the pore size
distribution. It utilizes the X-ray attenuation principle. The sample used in this study is a cube with 1 cm
side length. During scanning, images are sliced into small samples with 1-cm length on each side and 8
␮m thickness. On each slice, pores and solid can be identified by different CT numbers, which represent
different densities inside the sample. We can import the dataset produced by Micro-CT Scanner to an
SPE-178961-MS 7

image analysis software, which can count pores pixel by pixel and produce the pore area. Pixels with small
CT numbers are counted as pores and pixels with large CT numbers are counted as solid.
The scanned data show that the pore size distribution in each slice is close to lognormal distribution.
We can write ␩ in the following form
(20)

where ␮ is the mean value of ln(Ap), ␴2 is the variance of ln(Ap) and N is a multiplier. We can determine
␮ and ␴2 directly from the measurement data. In this study, ␮ is determined as ⫺9.2 and ␴2 is 1.3.
According to the nature of lognormal distribution, the mode pore radius equals to exp(␮-␴2), which is
calculated as 30 ␮m. The multiplier N is determined so that the porosity calculated by Eq. A-2 equals to
the measured porosity. With lognormal distribution, the porosity is integrated as
(21)

In Eq. 21, is taken as 10 times the mode pore radius based on Micro-CT image characterization.
With other parameters being known, N can be determined. The pore size distribution for our sample is
determined as Eq. 22, and is plotted in Fig. 3.

Figure 3—Pore size distribution for Indiana limestone studied.

(22)

Once pore size distribution is determined, M2 can be calculated.

(23)

Substituting M2 into Eq. 19, we can get the optimal acid interstitial velocity at the wormhole tip. The
calculation is shown in Appendix C. The model is validated with the experimental observations.
Acidizing Coreflood Experiments
To test the prediction of our model, acidizing coreflood experiments were performed to get the optimal
conditions for comparison. Indiana limestone with 15% porosity and 6-md permeability is used in this
8 SPE-178961-MS

study. The core dimension is selected as 1.5-in. diameter by 8-in. long to eliminate the core geometry
effect (Dong et al. 2014). The experiments were carried out with 15 wt% HCl at room temperature. During
the experiment, brine is injected first to measure the core permeability. After that, acid is turned on and
the pressure drop across the core decreases when wormholes start propagating. When the pressure drop
becomes near zero, the wormhole breaks through the core. Brine is then switched on to flush the system.
The pore volume to breakthrough is calculated based on the volume of acid injected. The experimental
results (red dots) are shown in Fig. 4, together with the wormholing efficiency relationship (black curve)
fitted with Buijse & Glasbergen’s model (Buijse and Glasbergen, 2005). The vi,opt is 1.98 cm/min and
PV is 0.367.
bt, opt

Figure 4 —Wormholing efficiency relationship for Indiana limestone studied

The experiments produce the optimal flux and breakthrough pore volume, and we can also calculate
the experimental optimal tip flux. In a linear coreflood acidizing experiment, the average wormhole
propagation velocity equals to the ratio between acid interstitial velocity and breakthrough pore volume.
It can be reasonably assumed that the instantaneous wormhole propagation velocity equals to the average
one at optimal condition for an 8-in long core because a stable fluid loss profile is established very soon.
(24)

At optimal conditions in an acidizing experiment, full acid concentration at the wormhole tip is
generally maintained (Furui et al. 2010). The wormhole propagation velocity is linearly proportional to the
tip acid flux, with acid capacity number as the proportional coefficient (Hung et al. 1989).
(25)

The optimal tip flux from our model is calculated as 6.68 cm/s (Appendix C), which is close to the
experimental result.
SPE-178961-MS 9

Discussion on Model Application

Prediction of Optimal Conditions for Acidizing Coreflood Experiments
In order to calculate the optimal acid flux for acidizing experiment, a correlation is needed to couple the
tip flux and core surface flux. Furui et al. (2010) developed a linear flow correlation for it, as shown in
Eq. 26, together with breakthrough pore volume relationship,
(26)

(27)

The wormhole diameter dwh is introduced for these hydrodynamic relationships. It can be calculated
through acid loss into surrounding pores. Hung (1987) studied the wormhole diameter growth based on
acid/rock mass balance
(28)

(29)

The wormhole tip diameter is calculated as 0.0176 cm. Substituting the parameters that are known into
Eq. 26 and Eq. 27, the vi,opt is calculated as 1.86 cm/min, and PVbt, opt is calculated as 0.321. The
calculated values are close to the experimental ones.
Effect of Temperature
Provided rock pore properties and pore size distribution, the model developed can be used to calculate the
optimal conditions. In the meanwhile, for the same rock, this model can also be used to study the
sensitivity of different parameters. In this section, we derive a correlation that describes the temperature
effect on the optimal conditions. Temperature affects both surface reaction and mass transfer processes.
Generally, increasing temperature results in increases of surface reaction rate and mass transfer rate
exponentially. Further analysis shows that surface reactivity has a much stronger dependence on temper-
ature than diffusivity. We can see from Fig. 5, as temperature increases from 70 °F to 280 °F, the surface
reactivity increases 1456 times while acid diffusivity increases only 4.8 times. The surface reactivity is
calculated based on Lund et al. (1975), and the acid diffusivity is calculated based on Conway et al.
(1999).

Figure 5—Temperature effect on surface reactivity for 15 wt% HCl and calcite (left), and on diffusivity (right).
10 SPE-178961-MS

If we calculate the ratio between two optimal interstitial velocities at two different temperatures
through Eq. 26, with substitution of parameters by Eq. 13, Eq. 19, Eq. 27 and Eq. B-2, we can get a new
correlation.
(30)

In order to verify Eq. 30, we summarized previous experimental results at different temperatures. The
optimal conditions are identified and summarized in Table 1. We calculated the ratios between two
optimal conditions from the same data source to eliminate the effect of other parameters, e.g. rock pore
size distribution, core geometry and acid concentration. Lavoux limestone, for example, if we compare the
optimal acid flux between 122 °F and 176 °F, the ratio on the left hand side of Eq. 30 is 2.29, and the ratio
on the right hand side of Eq. 30 is 2.20. We can see that the experimental results are close to model-
calculated results. We made similar calculations for the rest experimental results and have it plotted in Fig.
6. The x axis is the left hand side of Eq. 30, and the y axis is the right hand side of Eq. 30. The red line
with slope of 1 means the two sides are equal. The black dots represents the values calculated based on
experimental data in Table 1. Except for one outlier, the dots are distributed along the red line. This means
the two sides of Eq. 30 are approximately equal.

Table 1—Optimal conditions summarized from published experimental results

Source Rock Type C(HCl) T (°F) Vi,opt(cm/min) PVbt, opt D, cm2/s

dimensionless

I Indiana limestone 3.6 wt% 77 1.07 1.48 3.10⫻10⫺5

I Indiana limestone 3.6 wt% 122 2.81 2.67 4.72⫻10⫺5
II Lavoux limestone 7.0 wt% 68 0.56 0.4 2.95⫻10⫺5
II Lavoux limestone 7.0 wt% 122 1.23 0.45 4.93⫻10⫺5
II Lavoux limestone 7.0 wt% 176 2.82 0.65 7.55⫻10⫺5
III Indiana limestone 1.8 wt% 72 0.4 1.1 2.87⫻10⫺5
III Indiana limestone 1.8 wt% 122 0.89 1.54 4.61⫻10⫺5
III Indiana limestone 1.8 wt% 176 0.99 2.62 7.05⫻10⫺5
IV Kansas chalk 15 wt% 150 1.76 0.54 6.91⫻10⫺5
IV Kansas chalk 15 wt% 200 2.38 0.58 9.93⫻10⫺5
I: Wang et al. 1993, II: Bazin, 2001, III: Fredd and Fogler, 1999, IV: Furui et al., 2010
SPE-178961-MS 11

Figure 6 —Plot of Eq. 30 (red line) and ratios between optimal conditions at two temperatures (black dots)

Optimal Design for Carbonate Acidizing

For a well targeting a limestone formation, once we have determined an optimal tip flux at a certain
temperature, we can use Eq. 32 to generate a treatment design curve, as shown in Fig. 7. The reference
optimal tip flux can be either directly calculated from Eq. 19, or calculated from Eq. 24 and 25 based on
acidizing coreflood results. However, Eq. 19 is more preferred because it can eliminate the need of
acidizing coreflood experiments. The pore size distribution can be typically measured from drill cuttings
at the well site. The optimal acid pumping rate is readily calculated before completing the well.

Figure 7—Optimal tip flux versus temperature for Indiana limestone. The reference value is from Eq. 25.

Field Case Study

The flow geometries in the field are typically radial flow and spherical flow during the acidizing treatment
(Furui et al. 2010). Unlike acidizing coreflood experiment, acid pumping rate needs to increase contin-
12 SPE-178961-MS

uously to compensate the increasing acid loss and acid concentration loss. Research have been carried out
to apply linear acidizing coreflood experimental results to field scales, but limited success has been
achieved.
We applied Eq. 19 to field treatment design for a horizontal well with multiple openhole completion
stages. The correlation to couple the tip acid flux and pumping rate is developed by Furui et al. (2010)
through FEM simulation.
(31)

where mwh is wormhole numbers in a horizontal plane, which is assumed as 6 in this case; ␣z denotes
wormhole axial spacing, and is taken as 0.75. The calculation involves time step iterations. Within each
time step, the wormhole penetration length is calculated, and the corresponding pumping rate is then
calculated through Eq. 33. The first iteration starts from the wellbore, with wormhole penetration length
equal to the wellbore radius. The iteration ends when the maximal pumping rate is reached. This rate is
kept constant during the injection. The reservoir and acid properties are shown in Table 2.

Table 2—Acidizing field treatment design parameters

Reservoir temperature 200 °F
Porosity 15%
HCl concentration 15 wt%
Vi,tip, opt 19.76 cm/s
Acid capacity number 0.0144
Wellbore radius 0.328 ft
Stage length 100 ft
Wormhole numbers per plane 6
Wormhole diameter 2.5 mm

The simulated acid pumping schedule is illustrated in Fig. 8. The pumping rate ramps up with time
from 8.5 bpm to 52.5 bpm. As a comparison, if we pump acid using maximal pumping rate, these amounts
of acid can only last around 18 min. The ultimate wormhole length by optimal pumping rate is around 1.6
times the length by maximal pumping rate.

Figure 8 —Optimal acid pumping rate and maximal pumping rate for a given amount of acid
SPE-178961-MS 13

Conclusions
This paper presents a model to calculate the optimal conditions of carbonate acidizing. This model focuses
on the optimal acid flux at the wormhole tip, and upscale it to acidizing coreflood experiments and field
treatments. Based on this model, the effect of temperature on optimal conditions was studied. A new
method to make use of lab acidizing results for field treatment was also developed. Finally, an optimal
acid pumping schedule is developed for a horizontal well acid treatment. The conclusions of this study can
be summarized as below.
1. A model for optimal acid flux at wormhole tip is developed. This flux governs wormhole
propagation. It is the basis to study the optimal conditions of different scales. Being general to flow
geometries, it solely depends on the pore size distribution and acid/rock reaction.
2. By upscaling the optimal tip flux to linear flow, this model can predict the optimal conditions for
acidizing coreflood experiment correctly.
3. For fully diffusion limited reactions, the average/instantaneous optimal wormhole propagation
velocity is linearly proportional to the acid diffusion coefficient; so is the optimal tip acid flux.
4. A method is developed to upscale lab optimal acid injection rate to field treatments. The acid
pumping rate should be increased during the treatment to compensate for acid loss and concen-
tration decrease. A general matrix treatment design method is developed for limestone formations.

Acknowledgments
The authors thank the financial support of Acid Stimulation Research Project (ASRP), and the
experimental facility support of Harold Vance Department of Petroleum Engineering, Texas A&M
University.

Nomenclature
Ap Pore cross-sectional area, ␮m2
Ap,mode Cross-sectional area of the pore with the mode size, ␮m2
Atip Wormhole tip flow area, ␮m2
Co Bulk acid concentration, gmole/ml
Cwall Acid concentration at wormhole wall, gmole/ml
Ctip Acid concentration at wormhole tip, gmole/ml
Cs Acid concentration at pore surface, gmole/ml
D Diffusion coefficient, cm2/s
Da Damköhler number, dimensionless
dwh Wormhole diameter, mm
dcore Core diameter, inch
Ef Surface reaction rate constant, gmole1-ncm3n-2s⫺1
K Acid mass transfer coefficient, cm/s
Lp,mode Length of the pore with the mode pore radius, ␮m
Average pore length of porous medium, ␮m
M1 1st moment solution, cm⫺1
M2 2nd moment solution, cm
N Number of pores per unit volume of porous medium, 1/cm3
Np Number of pores
Nac Acid capacity number, dimensionless
⌬p Pressure drop, psi
PVbt Breakthrough pore volume, dimensionless
14 SPE-178961-MS

PVbt, opt Breakthrough pore volume at optimal conditions, dimensionless

q Flow rate, cm3/s
qtip Acid flow rate at wormhole tip, cm3/s
rp Pore radius, ␮m
t Time, s
vi Interstitial velocity, cm/s
vi,opt Optimal interstitial velocity, cm/s
vi,tip Interstitial velocity at wormhole tip, cm/s
vi,tip, opt Interstitial velocity at wormhole tip at optimal conditions, cm/s
Average acid velocity in a pore, cm/s
Average acid velocity in the mode-size pore, cm/s
Average acid velocity in the mode-size pore at optimal conditions, cm/s
vwh Wormhole propagation velocity, cm/s
vwh, opt Wormhole propagation velocity at optimal conditions, cm/s
w2 Flow area, cm2
Greek
␧ Flow coefficient in Hagen-Poiseuille’s equation, cm⫺1s⫺1
␬ Overall reaction rate coefficient, cm/s
␮ Viscosity, mPa·s
⌫p Pore perimeter, ␮m
␩ Pore size density function, 1/cm5
⌿ Pore growth function, cm2/s
␹ Acid volumetric dissolving power, volume rock/volume acid
Subscripts
tip Wormhole tip
opt Optimal condition
p Pore
mode Mode pore size
wh Wormhole

References
Bazin, B. 2001. From Matrix Acidizing to Acid Fracturing: A Laboratory Evaluation of Acid/Rock Interactions. SPE
Production & Facilities 16 (1): 22–29.
Buijse, M.A. and Glasbergen, G. 2005. A Semiempirical Model to Calculate Wormhole Growth in Carbonate Acidizing.
Paper presented at the SPE Annual Technical Conference and Exhibition, Dallas, Texas. SPE 96892. DOI: 10.2118/
96892-ms.
Conway, M.W., Asadi, M., Penny, G.S. et alet al. 1999. A Comparative Study of Straight/Gelled/Emulsified Hydrochloric
Acid Diffusivity Coefficient Using Diaphragm Cell and Rotating Disk. Paper presented at the SPE Annual Technical
Conference and Exhibition, Houston, Texas. Society of Petroleum Engineers 00056532. DOI: 10.2118/56532-ms.
Daccord, G., Touboul, E., and Lenormand, R. 1989. Carbonate Acidizing: Toward a Quantitative Model of the
Wormholing Phenomenon. SPE Production Engineering 4 (1): 63–68. DOI: 10.2118/16887-pa
Dong, K., Jin, X., Zhu, D. et alet al. 2014. The Effect of Core Dimensions on the Optimal Acid Flux in Carbonate
Acidizing. Society of Petroleum Engineers. DOI: 10.2118/168146-MS.
Fredd, C.N. and Fogler, H.S. 1999. Optimal Conditions for Wormhole Formation in Carbonate Porous Media: Influence
of Transport and Reaction. SPE Journal 4 (3): 196 –205. DOI: 10.2118/56995-pa
Furui, K., Burton, R.C., Burkhead, D.W. et alet al. 2010. A Comprehensive Model of High-Rate Matrix Acid Stimulation
for Long Horizontal Wells in Carbonate Reservoirs. Paper presented at the SPE Annual Technical Conference and
Exhibition, Florence, Italy. SPE SPE-134265-MS. DOI: 10.2118/134265-ms.
SPE-178961-MS 15

Gong, M. and El-Rabaa, A.M. 1999. Quantitative Model of Wormholing Process in Carbonate Acidizing. Paper presented
at the SPE Mid-Continent Operations Symposium, Oklahoma City, Oklahoma. Society of Petroleum Engineers
00052165. DOI: 10.2118/52165-ms.
Hoefner, M.L. and Fogler, H.S. 1987. Role of Acid Diffusion in Matrix Acidizing of Carbonates. Journal of Petroleum
Technology 39 (2): 203–208.DOI: 10.2118/13564-pa
Hoefner, M.L. and Fogler, H.S. 1988. Pore Evolution and Channel Formation During Flow and Reaction in Porous Media.
AIChE Journal 34 (1): 45–54. DOI: 10.1002/aic.690340107
Hung, K.M., Hill, A.D., and Sepehrnoori, K. 1989. A Mechanistic Model of Wormhole Growth in Carbonate Matrix
Acidizing and Acid Fracturing. Journal of Petroleum Technology 41 (01). DOI: 10.2118/16886-pa
Hung, K.M. 1987. Modeling of Wormhole Behavoir in Carbonate Acidizing, Ph.D Dissertation, The University of Texas
at Austin, Austin, Texas (1987)
Levich, Veniamin G. Physicochemical Hydrodynamics. 1962. Englewood Cliffs, N.J.: Prentice Hall, Inc
Lund, K., Fogler, H.S., Mccune, C.C. et alet al. 1975. Acidization—II. The Dissolution of Calcite in Hydrochloric Acid.
Chemical Engineering Science 30 (8): 825–835. DOI: 10.1016/0009- 2509(75)80047-9
Panga, M.K.R., Ziauddin, M., and Balakotaiah, V. 2005. Two-Scale Continuum Model for Simulation of Wormholes in
Carbonate Acidization. AIChE Journal 51 (12): 3231–3248. DOI: 10.1002/aic.10574
Schechter, R.S. and Gidley, J.L. 1969. The Change in Pore Size Distribution from Surface Reactions in Porous Media.
AIChE Journal 15 (3): 339 –350. DOI: 10.1002/aic.690150309
Schechter, Robert S. Oil Well Stimulation. 1992. Englewood Cliffs, New Jersey: Prentice-Hall, Inc
Wang, Y., Hill, A.D., and Schechter, R.S. 1993. The Optimal Injection Rate for Matrix Acidizing of Carbonate Formations.
Paper presented at the SPE Annual Technical Conference and Exhibition, Houston, Texas. SPE 26578. DOI:
10.2118/26578-ms.
16 SPE-178961-MS

Appendix A
In the porous medium model of this study, pores are described as randomly distributed capillaries in a
rock, with an average pore length . Fluids can flow from one pore to another with a certain pressure
drop. A pore size density function ␩(Ap) is defined so that ␩(Ap)·V·dAp denotes the number of pores having
an area between Ap and Ap⫹dAp with an average pore length in the volume V. For example, if attention
is focused on a certain group of pores in a porous medium with volume of w2⌬x, having an area between
Ap1 and Ap2, then the number of pores in the group is
(A-1)

Based on the pore size density function, porosity is the summation of each pore volume in a unit
volume of porous medium, and is described by Eq. A-2.
(A-2)

where M1 is defined as the first moment of the pore size density function and is calculated through Eq.
A-3. It is related to the mean value of the pore sizes.
(A-3)

The volumetric flow rate can be calculated by summing the volume flowing through each pore across
the flow area w2.
(A-4)

Among all the pores in a cross-sectional area, the acid flow rate for each individual pore is distributed
based on Poiseuille’s law. So does the acid flow velocity.
(A-5)

(A-6)

Substituting Eq. A-6 into Eq. A-3, the fluid flow rate through flow area w2 is
(A-7)

where M2 is defined as the second moment of the pore size density function and is calculated through
Eq. A-8.
(A-8)
SPE-178961-MS 17

Appendix B
When the chemical reaction in the diffusional boundary layer is at steady state, the rate of acid diffusing
to the pore surface equals to the rate of surface reaction.
(B-1)

where K is the mass transfer coefficient, C0 is the bulk acid concentration, Cs is acid concentration at
the pore surface, Ef is the surface reaction rate constant and m is the reaction order. For a convection
reaction process, the mass transfer coefficient K can be calculated as (Levich V.G, 1962)
(B-2)

The surface acid concentration Cs is difficult to determine and it can be reduced through Eq. B-1.
(B-3)

Substituting Eq. B-3 into either side of Eq. B-1, we can then get an overall reaction rate equation and
overall reaction rate coefficient.
(B-4)

(B-5)

In Eq. B-5, if , the surface reaction rate is low and is the limiting step of the overall
reaction. In this case, . If , the diffusion rate is low and is the limiting step of
the overall reaction. In this case, ␬ ⫽ K.
18 SPE-178961-MS

Appendix C
The optimal tip flux depends on the acid/rock reaction and the rock pore size distribution. In this study,
the reaction is for 15 wt% HCl and calcite at 75 °F; the pore size distribution is for Indiana limestone. The
detailed inputs are show in Table C-1.

Table C-1—Input parameters for optimal tip flux calculation

Lithology Indiana limestone

Acid 15 wt% HCl

Temperature 75 °F
Porosity 15%
Mode of the pore radius 30 ␮m
Average Pore length 300 ␮m
Diffusion coefficient 3.5⫻10⫺5 cm2/s
3.81⫻10⫺2 cm/s

The calculation starts from calculating the optimal acid velocity in the mode-size pore, . Several
iterates should be taken because the overall reaction rate coefficient also depends on this velocity.
Assuming an initial of 3 cm/s, the first iterate is shown through Eq. C-1, Eq. C-2 and Eq. C-3.
(C-1)

(C-2)

(C-3)

The second iterate starts from of 0.42 cm/s. Same steps are repeated to get new K, ␬ and
. The iteration finishes when the newly calculated value is close to the one calculated in the
previous iteration. In this study, the final is calculated as 0.26 cm/s in the fourth iteration.
With pore size distribution avaialbe, M2 can be calculated as
(C-4)

The optimal tip flux is calculated through Eq. 19.