Journal of

Petroleum Gas & Chemical Engineering

Research Article

Rock Typing by Bulk Volume Water in the Carbonate Reservoirs

Zahra Riazi*

Faculty of Infrastructure Engineering, University of Melbourne, Vic-

Corresponding Author: Zahra Riazi, Faculty of
toria, Australia Infrastructure Engineering, University of Melbourne,
Victoria, Australia. Email: [email protected]

Received: 2024 Feb 21 Accepted: 2024 Mar 12 Published: 2024 Mar 30

Abstract
Despite the popularity of permeability and porosity relationships in the rock typing approaches of the sandstones, their
application is limited in the carbonates. It stems from secondary porosity, and heterogeneities in the carbonates, which
leads to weak correlation between them. Mismatching of carbonate rock samples with their defined rock types by
permeability and porosity relationships pronounces this shortage in the laboratory and thin section studies. This study
has considered the product of irreducible water saturation and porosity, which is indicated by Bulk Volume Water (BVW)
was examined as a possible standalone substituted rock typing method for carbonates in this study. Its outcomes were
successfully validated by Mercury Injection Capillary Pressure (MICP) curves and compared with the Reservoir Quality
Index (RQI) method in three carbonate fields.

Results of rock typing in the three different carbonate fields revealed a more satisfactory match between classified rock
types using the BVW and their equivalent MICP curves in comparison with RQI method. Moreover, smaller number of
rock types were defined in the three carbonate fields using the BVW method, which leads to improve further modellings.
Therefore, this method can be a more appropriate method for carbonates having a poor permeability-porosity relationship.
Moreover, porosity and irreducible water saturation are easily attainable from petrophysics well logs in the most field
wells, contrary to extensive lab works that permeability data needs to achieve. Its outcome is also straightforward to feed
into the static and dynamic models. Although, handling a great deal of data may seem an issue for this method, which the
prepared MATLAB code has tried to facilitate in this study.

Introduction the defined rock types. This clearly reflects its importance
Rock typing is a part of the reservoir characterization pro- during reservoir study and its modelling. Indeed, existing
cess that is carried out after the data acquisition stage of different rock typing methods that classify reservoir rock
field development. According to the classical definition, rock qualitatively (e.g., different Lorenz and Stratigraphic Plots
typing is reservoir rock classification into distinct units, and quantitatively (e.g., Winland plot, RQI and BVW) have
which were each deposited under similar conditions and been developed to address this need. Permeability and po-
experienced similar diagenetic processes. This results in a rosity are the main inputs for most rock typing methods,
unique porosity-permeability relationship and a similar cap- while the source of permeability data is limited to a few wells
illary pressure profile for each rock type. This definition con- and intervals in the whole reservoir. The rock typing process
siders different reservoir properties either in the shape of is limited to these intervals, but it needs to distribute them
characteristics or relationships for knowing a definite rock to all wells and intervals.
type. So, every acquired data (RCAL, SCAL, well log interpre-
tation, and well test results) can be applied for this classifi- The applied methods rely on porosity and common log
cation and can complete it regarding its scale. Acquaintance variables, which exist for all intervals. In the simple case,
with these distinct units can be started using small scales porosity and permeability correlation in the cored sections
data (thin sections and core data), their extent and sequence is applied for un-cored sections using log porosity. Though
along the wells followed by petrophysical logs and distrib- log porosity needs to be matched with porosity at the cored
uted throughout the reservoir using seismic data. Dynamic depths and in the case of carbonates this correlation has a
rock typing is done on upscaled data (formerly classified as low correlation coefficient. Existing secondary porosity and
static rock typing) and the model is initialized with the capil- fractures in the carbonate reservoirs decrease the correla-
lary pressure curves allocated to the defined rock type; then, tion coefficient in both core porosity-log porosity and core
the oil in place is estimated. Consequently, all predictions porosity-permeability data correlations. Some methods like
and results of defined dynamic scenarios are influenced by generating a correlation between the combination of log
Volume - 1 Issue - 1
Citation: Samuel, O. O., Jerry, I.T. (2023). Measurement Of Background Radiation at Oracle Plastics and Sacks Company in Makurdi, Benue State. Journal of
Citation: R, Zahra. (2024). Rock Typing by Bulk Volume Water in the Carbonate Reservoirs. Journal of petroleum gas &chemical engineering, 1(1), 1-11.
Theoretical Physics & Mathematics Research. 1(1), 01-07 Page
Page1 1ofof11
9
 Journal of Petroleum Gas & Chemical Engineering Copyright © Zahra Riazi

data and flow zone indicator (FZI), discrete rock type (DRT) zone, this water saturation is equivalent to irreducible wa-
or artificial neural network have been used to rectify this ter saturation (IWS), which is immovable during production.
issue [1-3]. However, applying these methods, besides their This fraction of water is bounded within the pore network
complexity, imposes more uncertainty on the models. Also, by capillary force and adsorbed onto grain surfaces. Capil-
they usually present a very low correlation coefficient for the lary pressure is controlled by pore throat network diame-
known points when they act as the control points for pro- ter, which is a random variable and depends on grain size.
duced results. Therefore, the predicted permeability usual- Decreasing or increasing the pore throat network radius
ly suffers from enough accuracy and the problem of feeding leads to rising or reducing capillary pressure, respectively
the outcomes to the static and dynamic models still exists in and has a contrary effect on the percentage of IWS in the
these methods and seems improper to apply for carbonates. reservoir rock. The BVW classifies reservoir rock based on
this concept, the fraction of water in reservoir rock reflect-
Generally, rock typing in carbonates has been challenging ing different pore throat sizes. In other words, the depths
due to the existing great diversity in both size and shape of with similar rock properties are expected to have relatively a
grain for the most carbonate sediments in comparison with similar value of the product of porosity and water saturation
sandstones. Existing post-sediment diagenetic processes as the BVW, which classifies these depths as the same rock
and facies changes lead to pore size variations from micron type. Hypothesized a similar relation between porosity and
to cave size and complexity of their networks. These influ- connate water saturation with considering the similarity be-
ence porosity and permeability relationships by decreasing tween “surface area - diameter of the particle” and “surface
their correlation coefficient, which is high in the sandstones. area -porosity” correlations [5]. He used the inter-relation-
In other words, a vuggy porosity can store significant vol- ship between porosity and mean particle diameter and at-
umes of oil, but the existing poor connections between the tained the equilateral hyperbole relationship between satu-
vugs may lead to low permeability and production. On the ration and porosity, which expresses the IWS as a fraction of
other hand, the BVW method uses irreducible water sat- bulk volume instead of pore
uration (IWS) and porosity, which are available along well
column and solve the problem of data scarcity. Moreover, ex-
isting log porosity and water saturation help to directly feed
the result of rock typing to the static and dynamic models. This is a recognized concept in the oil industry and its plot
This eliminates the need for intermediate methods that may (Buckles plot) is typically applied for different reservoir pa-
increase model uncertainties. Theoretically, using water sat- rameter evaluations. This plot includes a family of hyperbol-
uration as the input of the BVW method reflects the effect of ic curves, in which each one shows a constant BVW value,
pore throat size, which is the controlling factor in carbonate increasing this constant value has an opposing relation with
reservoirs for flowing the fluids and is a common factor with pore throat size (Fig. 1). explained several applications for
permeability. Also, applying both BVW and RQI methods in a this method and recapped them: discretising reservoir zone
carbonate field and comparing the outcomes using grouped with IWS; estimation of water-cut and producibility; Perme-
MICP curves disclosed more similarity between the grouped ability prediction; grain size approximation; pore type esti-
MICP curves by the BVW method than RQI [4]. Despite these mation; detection of multiple lithologies [6,7]. The four last
benefits, this method has been rarely applied for rock typing applications of the BVW methods are related to rock typing.
as a stand-alone method. Indeed, for a reservoir with

It has been presented in the literature as an auxiliary meth- Variable Lithology, This Family of Curves are formed by a Se-
od with other approaches during rock typing but is not con- lective Classification of Data, which Determines Rock Types.
sidered a standalone method for this purpose. For example,
Winland’s plot versus Bulk Volume Water (BVW) was used
for rock typing or irreducible water saturation in its formula
was substituted with the general model between porosity,
permeability and irreducible water saturation. This study
tried to consider the capacity of this method as an indepen-
dent one to classify reservoir rocks using three carbonate
field data. This paper examined the applicability of the BVW
method after a short explanation of the methodology of the
BVW method and its algorithm for coding. Next, the rock typ-
ing outcomes were validated by MICP curves in the “Results”
section. Finally, applying this method was discussed in com-
parison with RQI and FZI method, which utilizes permeabili-
ty and porosity as their input, in the last section.

Methodology Figure 1: The Bulk Volume Water Plot.

Bulk Volume Water (BVW) is the product of water saturation
and porosity, which presents the fraction of occupied porous The dependency of water saturation on pore throat size in-
volume by water in the reservoir rock. Above the transition terrelates the BVW method to the capillary pressure. Cap-
Volume - 1 Issue - 1
Citation: R, Zahra. (2024). Rock Typing by Bulk Volume Water in the Carbonate Reservoirs. Journal of petroleum gas &chemical engineering, 1(1), 1-11.
Page 2 of 11
 Journal of Petroleum Gas & Chemical Engineering Copyright © Zahra Riazi

illary pressure, as the opposing force during oil migration, Bulk Volume Water Rock Typing Method Applica-
is controlled by the distribution of pore throat radius in tion
the reservoir rocks, which is generally estimated by a MICP This methodology is applied by data selection from some
curve in a rock sample (Fig. 2). candidate wells and using their log porosity and water sat-
uration at the depth above the transition zone. In this zone,
initial water saturation depends on both reservoir rock
properties and fluid contacts, which limits reservoir rock
classification merely based on its properties. So, before using
the data, it is recommended to draw log porosity and satura-
tion versus depth and omit the transition data visually. The
Buckles Plot can help to remove points on the border of oil
and transition zones. Following this, to prevent discarding
any data point corresponding to the net rock volume, con-
servative cut-off values can be applied for both porosity and
water saturation.

After applying the data preparation stage, reservoir rock

typing is commenced by multiplying porosity and water sat-
uration values. This product is sorted from minimum to max-
imum and classification is begun from the minimum point.
Next, a group of data is chosen from this data set, as the first
rock type, and averaged (Ā) and “n” values are calculated
Figure 2: Different Sections and Points in a MICP Curve (Jen- with the following formula for each data in this group. Then,
nings 1987; Petty 1988; Wu 2004) the “n” is checked for all data in the group to be within the
defined limit. In the case of being the limit, the first rock type
Normally a MICP curve has two main sections, a plateau, and is defined. The value of “n” should be in a certain range of
steep slopes, which show arriving mercury in the macrop- 0.8 to 1.2 for each data point. This shows its similarity to the
ores and micropores, respectively [8]. The plateau section average property of the selected group. In the other words,
forms in lower mercury saturation and its slope infers the the property of each point needs to be near to the average of
quality of pore throat sorting in the rock sample. A hori- the group for being supposed as a member of a rock type. If
zontal shape reflects a good and uniform pore-throat size the “n” value is less/more than this range, the selected series
sorting; while steeping the plateau infers deteriorating pore-
should be changed until meeting the mentioned condition.
throat structure [9]. The plateau joins the steep section of
By finding a group, which its members have an “n” value in
the graph, which ends at the highest capillary pressure. The
a range “between” 0.8 to 1.2, a rock type is distinguished be-
tangent line to the highest-pressure section of MICP curves is
parallel with the y-axis and indicates the percentage of pore tween all data. This procedure should be repeated by select-
space in the rock sample, which mercury could not invade ing the next group and classifying the rest of the data. Fol-
and its corresponding saturation spots IWS in the sample. lowing, it was recapped this process in its algorithm:

Pe, Pd, and Pt are definite pressure points on the MICP curves 1. Calculate product of “ ”. Sort “ ” data from Min
which indicate entry, displacement, and threshold pressure, to Max
respectively. Pe is influenced by irregularities that may ex- 2. Choose a new range of sorted data and averaged over
ist around the sample surface. These irregularities cause this range (Ā).
mercury to arrive in the space between them and the sam- 3. Calculate “” for all data points in the chosen range:
ple holder and this extra saturation must be corrected be-
fore using the curve. Pd shows the amount of pressure that
a non-wetting fluid requires to enter the largest connected
pore throats in the rock sample [10]. Finally, Pt indicates the
pressure that mercury forms a connected pathway along
with the sample [11].
4. Check “n” has a value between 0.8 and 1.2:
Indeed, different sections of a MICP curve provide either
qualitative or quantitative information about pore throat It is in the mentioned range; one rock type is formed. Repeat
sizes, sorting, and the IWS for a reservoir rock sample. These the procedure for next
are the main factors, which help during the reservoir rock It is not, change of the range to include more (or less) data
classification process. Moreover, the size and sorting of pore
points to satisfy the criteria of “n” values
throats are general factors between permeability and hy-
drocarbon saturation and can relate to these two variables.
The grouping data procedure for separating different rock
Therefore, information infers from a MICP curve makes it an
types is done by trial and error in the BVW method, which
appropriate source for both reservoir rock typing and verify-
can be its only pitfall and can be challenging in the case of
ing other quantitative rock typing approaches.
a huge amount of data. So, a computer code was prepared
Volume - 1 Issue - 1
Citation: R, Zahra. (2024). Rock Typing by Bulk Volume Water in the Carbonate Reservoirs. Journal of petroleum gas &chemical engineering, 1(1), 1-11.
Page 3 of 11
 Journal of Petroleum Gas & Chemical Engineering Copyright © Zahra Riazi

based on the mentioned algorithm by MATLAB to solve the tively in the three fields. Based on cut-off value estimation in
trial-and-error process in this method. The code was verified these fields, applying mentioned values does not influence
using three carbonate fields and applied for their reservoir the original oil in place value.
rock typing, which has been discussed in the following sec-
tion. Feeding the rock typing results to the static and dynam- For each field, the prepared data (porosity and water satu-
ic models is very fast and convenient by this method in com- ration) were fed to the MATLAB code and the result of rock
parison with other usual rock typing approaches. In the case type classification was presented in Figs. 3, 5, and 7. The
of using the software, it needs to follow succeeding steps: MICP curves in each field were used to investigate the accu-
racy of the results. The MICP curves were classified based on
• Build two new variables; firstly, with a general template the position of their porosity and IWS in the rock types’ plot.
showing the product of IWS and porosity; the second Since Field A is a newly developed field, it lacks the SCAL
one, with the discrete template as the saturation region data and MICP curves from its nearby field were applied for
identifier. its rock typing verification.
• Using the calculator option in this software, assign the
multiplication of initial water saturation and porosity to
the newly defined property.
• Again, by calculator and a maximum and minimum of
each rock type, fix different rock types in the variable
that was defined with the discrete template.

Case Studies
The prepared MATLAB code was examined by three carbon-
ate fields’ data. The Sarvak, Fahliyan, and Kangan-Dalan are
carbonate formations in fields A, B, and C, respectively. The
porosity and water saturation data resulting from log inter-
pretation were selected from three wells for each field. For
data preparation, the data points, which were in and below
the transition zone, were omitted using resistivity and po-
rosity logs and the result was checked by the Picket plot. As Figure 3: Displays the Results of Rock Typing by the MAT-
the final step, the cut-off values of less than 5% and more LAB Code, which classified Porosity and Water Saturation
than 70 % were applied for porosity and IWS data, respec- into Four Rock Types in Field A.

Figure 4A: Rock Type Classification by BVW Method in Field A.\

Figure 4: Classified MICP Curves (A) and Orosity vs Permeability (B) Based on Defined Rock Types in the Field A.

Volume - 1 Issue - 1
Citation: R, Zahra. (2024). Rock Typing by Bulk Volume Water in the Carbonate Reservoirs. Journal of petroleum gas &chemical engineering, 1(1), 1-11.
Page 4 of 11
 Journal of Petroleum Gas & Chemical Engineering Copyright © Zahra Riazi
0.7

0.6

0.5

0.4

Irreducible Water Saturation (fraction)

0.3

0.2

0.1

0
0.05 0.1 0.15 0.2 0.25 0.3 0.35

Porosity (fraction)

Figure 5: Rock Type Classification by BVW Method and Comparing its Results with MICP Data in Field B.

Figure 6: Grouping the MICP Curves (A) and its Porosity-Permeability Correlation (B) Based on BVW in Field B.
0.7 raphy events. Sarvak formation in this field is divided into
nine subzones from very poor due to the occurrence of shaly
0.6 layers in Zone-1 and 2 to the high porosity and medium per-
meability limestones in Zone-3. The quality of Zone-4, 5, and
0.5 6 varies throughout the field from weak to medium in com-
parison with Zone-3. The oil-water contact has a tilted shape
0.4 from north to south of the field. Increasing water saturation
Irreducible Water Saturation (fraction)

starts from the end of sarvak-7 in some places or middle of

0.3 Sarvak-8 and Sarvak-9 is a completely water-bearing zona As
the Figure Shows Rock Types #3 and 4 have more Frequency,
0.2 which Refers to Fig.1 Reflects Medium to Poor Average Pore
Sizes. The Reservoir Section with High Quality is shown by
0.1 Rock-Type #1 and 2.

0 Rock types were verified by MICP curves; so that, their ini-

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35

Porosity (fraction)
tial water saturation and porosity were inserted in the rock
Figure 7: Rock Type Classification by BVW Method and type plot (green circles in Fig. 3). Then, they were grouped in
Comparing its Results with MICP Data in Field C. the next step regarding their position in classified rock types
(Fig. 4A). Clustered MICP curves in Fig. 4A display good sep-
Results aration, especially in the high-pressure values that indicate
Case 1: Field A initial water saturation. Based on different sections of a MICP
Sarvak formation with the middle cretaceous age is the curve (Fig. 2), the existing MICP curves in each group present
oil-bearing formation in Field A and is known as the relatively similar characteristics as the member of one cluster
age-equivalent of the Mishrif and Natih formations. It com- indicated by a rock type. Moving from the first (blue colour
poses of limestone to argillaceous limestone and was strong- points) to the fourth rock type in Fig. 3 and comparing with
ly affected by diagenesis, some tectonic, and paleo-topog- the shape of MICP curves in their related rock type (shown
Volume - 1 Issue - 1
Citation: R, Zahra. (2024). Rock Typing by Bulk Volume Water in the Carbonate Reservoirs. Journal of petroleum gas &chemical engineering, 1(1), 1-11.
Page 5 of 11
 Journal of Petroleum Gas & Chemical Engineering Copyright © Zahra Riazi

by similar colour), reveal good consistency between them. ed MICP curves to the Rock type#1 display relatively a flat
The first group of MICP curves show a vertical steep slope plateau, which infers a good sorting. The plateau section
and their flat plateau section indicates a good and uniform of curves inclines in Rock-type#2 and is steeper for Rock-
pore-throat size sorting for this rock type. Also, they pres- type#2, which reflects deteriorating pore-throat sorting in
ent the least initial water saturation in comparison with the these types of reservoir rock samples. Also, the inflection
curves in other groups. This presents a coarser pore throat point occurred at lower capillary pressure for the MICP
size that is consistent with the position of the first rock type curves in Rock-type #1, while it increased for the rest and led
in the Buckles plot. Decreasing average pore sizes from the to the grouped MICP curves with the higher IWS in Rock-type
second to the fourth rock types agrees with the existing vari- #2 and 3. Fig. 6B presents a porosity-permeability cross plot
ations in the MICP curves for their equivalent group. Since for the classified MICP curves in Field B. The plot displays an
the IWS and porosity display a hyperbolic relationship in the increasing trend between porosity and permeability for each
Buckles plot, it is expected to see some changes in the shape rock type, while they covered each other.
of MICP curves for each individual rock type. Indeed, with
moving in direction of high saturation-low porosity to low In comparison with Field A, the porosity-permeability cross
saturation-high porosity points, the MICP curves show lower plot presents more scattering between porosity and perme-
initial water saturation and threshold pressure. ability data of different rock types. Data points can graph-
ically be divided into two general increasing trends (black
Porosity and permeability correlations for different rock dashed line in Fig. 6B). These infer two general slopes for
types are displayed and indicated by “RT” on Fig. 4B. Consid- the porosity and permeability correlations. In the left section
ering all in one graph presents neither specific correlation of the dashed line, permeability improves corresponding to
nor clear separation between them for different rock types. the porosity increment, while in the right section the rate of
However, there is an overall increasing trend between poros- permeability enhancement seems slower with increasing
ity and permeability of each rock type, but they cover each porosity. The second group can present unconnected vuggy
other. In other words, the porosity- permeability correlation porosity in the related rock samples in addition to the inter-
of MICP samples in each rock type relatively corresponds granular porosity. In the other words, the porosity was en-
with its reservoir properties in Fig. 3; although, they show hanced more than two times, but the permeability barely
no clear separation between themselves. Existing different rose more than 8 MD.
types of secondary porosity in the carbonates (vugs, fissures
and fractures) influence the shape of the increasing trend Case3: Field C
between porosity-permeability correlations. Very dense The Kangan and Dalan are Late Permian formations and the
samples with low porosity can present high permeability in members of Dehram Group, which is correlated with the
the case of having fissures in its structure or in the opposite Khuff formation. The Kangan overlies the Dalan formation,
shape, non-connected vugs form high porosity but low per- and it consists of clean carbonate, basal argillaceous, and
meability. In these situations, the shape of the pore throat evaporite carbonate facies. Limestones and evaporites of
network in a MICP curve would determine the real type of the Dalan formation in the studied area, are subdivided into
rock sample. three primary members including the upper and lower Dalan
(mainly limestone, dolomitic limestone) and Nar evaporites
Case 2: Field B (anhydrite and thin interlayers of dolomite) in the middle.
Fahliyan formation with the late Cretaceous age is the main According to the log data (porosity and water saturation) in
reservoir in Field B. This carbonate formation was depos- three wells and the BVW method, carbonates were classified
ited in the Zagros sedimentary basin and it is a part of the into three rock types in Field C with the most frequency for
Khami Group. Its thickness is about 1312ft and was divided Rock-type #3 (Fig. 7).
into four members in this field. The members have differ-
ent thicknesses and oil present in all of them. They mainly Inserting the IWS and porosity of MICP samples into the rock
consist of limestone, occasionally claystone or dolostone, typing plot (green circles Fig. 7) present a good coverage of
with local argillaceous limestones, which are separated by MICP data points for Rock-type #1 and 2. Therefore, it is ex-
low-permeability layers. Applying the BVW method to log pected to see various shapes of MICP curves for these two
porosity and water saturation data of Field B, classified its classes.
carbonate reservoir rock into three different rock types
(Fig. 5). Rock type#2 allocated the most data to itself, but Clustered MICP curves and their porosity-permeability cor-
it is placed in the position of rock type#3 in Field a (Fig. 3). relations were plotted for three rock types in Figs. 8A and 8B,
Following, the porosity and initial water saturation of MICP respectively. Similar to Field B, existing small-scale hetero-
curves were inserted (green dots in Fig. 5) on the rock types’ geneities in the reservoir samples formed the MICP curves
plot and categorized based on their position (Fig. 6A). As Fig. with a variable plateau or steep sections. Thus, despite be-
5 shows, more data points were used in Field B; though, they longing to different rock types, some MICP curves covered
did not cover the whole hyperbolic shape of IWS-porosity each other in the low-pressure section. Moving from Rock
correlation uniformly. type #1 to 3, the plateau section of MICP curves started at
higher entry pressure and their shape could not preserve
Considering MICP curves in each group presents a clear sim- their consistency with increasing saturation. The curves
ilarity between them, especially regarding their IWS. Relat- related to Rock-type #1 presented a more uniform and flat-
Volume - 1 Issue - 1
Citation: R, Zahra. (2024). Rock Typing by Bulk Volume Water in the Carbonate Reservoirs. Journal of petroleum gas &chemical engineering, 1(1), 1-11.
Page 6 of 11
 Journal of Petroleum Gas & Chemical Engineering Copyright © Zahra Riazi

ter plateau section, which inclined to one or more slopes relations (Fig. 8B). For instance, despite enhancing porosity
in Rock-type #2 and 3. This deduces decreasing both pore- from 0.05 to more than 0.3 in the MICP samples of Rock-type
throat size and uniformity in macropores and increasing the #3 (yellow dots), permeability changed from 0.3 to 7 MD.
percentage of micropores in the samples from Rock-types #1 These can present existing non-connected vugs, which more
to 3. This was confirmed by their porosity-permeability cor- affected porosity.

Figure 8: Grouping the MICP Curves Using BVW Method (A) and Porosity- Permeability Correlation (B) for the Field C.

Overall, plotting porosity-permeability correlations of MICP the RQI method. This method was broadly explained in the
data for defined rock types in Field C displayed no separa- literature and was not mentioned here to avoid repetition
tion between themselves and cover each other (like Field A [1].
and B). Though, the MICP points can be categorized hypo-
thetically into three general groups (dashed lines in Fig.8B) Applying the RQI method, the porosity and permeability of
regarding the porosity-permeability relationship. They show MICP samples were used as the inputs for these three fields.
an increasing trend in these groups but with different slopes This makes an easier comparison between the outcomes of
and all three rock types were distributed in them. Indeed, the two methods. The porosity and permeability data were
the points in the left, middle, and right sections can deduce clustered using the RQI method in the three fields. Then, the
existing fracture, intergranular porosity, and non-connected related MICP curves were gathered in different groups based
vugs in their related rock sample, respectively. Therefore, in on similar FZI values. Figs. 9, 10, and 11 present outcomes
this situation, using a pore throat network can be a more ap- of these two steps for Field A, B, and C, respectively in two
propriate parameter for rock type classification. Since it can subplots of C and D. Looking at the outcomes of rock typ-
reflect the type of rock using its shape and initial water satu- ing by the BVW (Figs. 3, 5, and 7) and RQI (Figs. 9, 10, and
ration, which are the outcome of all existing variations in the 11C) methods reveal both classified carbonates to four rock
rock sample that may affect porosity value. types in Field A. For Field B and C, the RQI method classified
carbonates into six rock types; so, it is expected to see more
Discussion similarity between the MICP curves related to each rock type
Comparing the BVW method for carbonates with other in comparison with three rock types using the BVW method.
methods using permeability as their input was utilized by

Figure 9: Classifying Porosity- Permeability and MICP Curves Using RQI/FZI Method (Field A).

Volume - 1 Issue - 1
Citation: R, Zahra. (2024). Rock Typing by Bulk Volume Water in the Carbonate Reservoirs. Journal of petroleum gas &chemical engineering, 1(1), 1-11.
Page 7 of 11
 Journal of Petroleum Gas & Chemical Engineering Copyright © Zahra Riazi

Figure 10: Classifying Porosity- Permeability and MICP curves using RQI/FZI method (Field B).

Figure 11: Classifying Porosity- Permeability and MICP Curves Using RQI/FZI Method (Field C).
Overall, clustered MICP curves based on four rock types method (Fig. 12). For Field B and C (Figs. 13, 14) similari-
defined by the RQI method (shown by the same colour of ty between the curves by the BVW in Field A was convert-
the related rock type) display no clear separation in Figs. ed to some sort of variations. Although, except for the last
9C and 9D in Field A. The MICP curves belonging to each rock types, inequality in the plateau, and steep slopes of the
rock type were spread along with the plot and covered the MICP curves are not notable in each rock type and all lay in
other rock types’ curves (Fig. 9D). this pattern was repeated the limited range of IWS value. In comparison, the grouped
for Field B and Field C (Figs. 10 and 11). Indeed, the porosi- MICP curves by the RQI in Field B and C (Figs. 13 and 14)
ty and permeability data display distinctive correlations for indicated extensive variations for each rock type. Neverthe-
each rock type using the RQI method, but their related MICP less, some sort of similarity presents only for Rock-types #1
curves cover each other without any clear separations. Com- and 5 in Field B (Fig. 13), the rest display inequality in dif-
paring them with the outcomes of the BVW method (Figs. 4, ferent sections of MICP curves, which leads to scattered IWS
6, and 8A) presents more mixing patterns despite more rock in them. Possible sources of these dissimilarities between
types, while the separation between the clustered curves by the MICP curves in a defined rock type can be discussed in
the BVW method is clear. two aspects. One can be related to their mathematical re-
lationship, which exists between their variables. This puts
Investigating the outcomes of classifications with more pre- two curves in one group while presenting relatively the
cision, the classified curves were detached in subplots (Figs. same shape but different values in the specific points (e.g.,
12, 13, and 14). Field A with the equal number of rock types IWS). The other connects distortions in the regular shape of
by two methods presents more similarity between the clus- a MICP curve to the existing heterogeneities and secondary
tered MICP curves by the BVW. Although, these resemblanc- porosity in the carbonate rock samples.
es were faded in the rock types were grouped by the RQI

Volume - 1 Issue - 1
Citation: R, Zahra. (2024). Rock Typing by Bulk Volume Water in the Carbonate Reservoirs. Journal of petroleum gas &chemical engineering, 1(1), 1-11.
Page 8 of 11
 Journal of Petroleum Gas & Chemical Engineering Copyright © Zahra Riazi

Figure 12: Classified MICP Curves Using the BVW and RQI/FZI Methods (Field A).

Figure 13: Classified MICP curves using the BVW and RQI/FZI methods (Field B).

Figure 14: Classified MICP Curves Using the BVW and RQI/FZI Method (Field C).

Volume - 1 Issue - 1
Citation: R, Zahra. (2024). Rock Typing by Bulk Volume Water in the Carbonate Reservoirs. Journal of petroleum gas &chemical engineering, 1(1), 1-11.
Page 9 of 11
 Journal of Petroleum Gas & Chemical Engineering Copyright © Zahra Riazi

Technically, both the RQI and BVW methods classify reser- curves with variable shapes and IWS values. Considering
voir rocks using a relationship between two reservoir pa- the porosity-permeability correlations in Figs. 10C and 11C
rameters. These are the quotient of permeability-porosity reveal, that they expanded in the wide ranges for each rock
and multiplication of IWS and porosity, respectively. Indeed, type. Therefore, the clustered MICP curves based on these
in these two methods, their related variables can vary in their rock types presented variable shapes and values; despite the
ranges, while the results of calculations can be the same. In higher number of rock types by the RQI method (Figs. 12 and
other words, the division of high permeability and porosity 13).
that is used in the RQI method can result in the same val-
ue as both are low. The same situation, but in the shape of The grouped MICP curves using the BVW and RQI meth-
multiplication exists for the BVW method. Multiplying low ods can be compared by considering dissimilarity and like-
porosity and high IWS can be equal to high porosity and low ness both in the plateau, and steep slopes of curves present-
water saturation. This leads to existing different shapes of ing the macropores and micropores condition, respectively
MICP curves for one rock type, which reflects the presence of and ultimately IWS in the sample. Starting with plateau and
two opposite limits in each rock type. However, it seems the slope sections of MICP curves, homogeneity between these
application of cut off on the log porosity/IWS and the shape two sections for the ones grouped by BVW in comparison
of grouping in the BVW method (Fig.1), in which the distri- with the RQI method is completely clear for Field A (Fig.
bution of rock types gets narrow at the two ends for each 12); although, outcomes of the BVW method showed better
group, alleviate this problem in it. performance. Overall, some curves show a stepwise shape
either in the plateau or slope sections due to inhomogene-
The existence of more than one pore throat distribution ities in the MICP samples for Field B and C. Forementioned,
group in a MICP sample can be a common feature in the car- sloping the plateau section from a horizontal shape infers
bonates due to secondary porosities, and small-scale hetero- deteriorating pore-throat structure from a uniform shape, or
geneities (e.g., microscale facies change). Each distribution this works for all sections. Therefore, these curves can only
group presents a specific value regarding the saturation, be compared by IWS as the grouping criteria, which is their
during the invasion of mercury at a definite pressure. These final condition at high capillary pressure. Using IWS as the
distort the regular shape of MICP curves and the formation grouping criteria and pondering outcomes of the two meth-
of different inflection points in the plateau and steep sec- ods by this reveals more homogeneous grouping results for
tions of a MICP curve can clearly present them. Therefore, the BVW method. Although, there are few rock types clas-
depending on the range of variables for each rock type and sified by the RQI method that show a limited range for IWS
presenting secondary porosity or heterogeneities in a car- (RT#1 and 5 for Field B), in comparison with the outcomes
bonate reservoir rock, the possibility of seeing different of the BVW presented clustering with more disparate mem-
shapes for the MICP curves changes in a rock type. Although, bers. Moreover, the number of rock types is less in using the
existing a wide range of variables interrelates to the function BVW method. The effect of the number of rock types gets
of secondary porosity in the carbonates. In this situation, the prominent when they are applied for property distributions
similarity between different sections of the MICP curves can in the static model and dynamic modelling.
be applied as the grouping criteria and optimizing the num-
ber of groups (e.g., the similarity between IWS, the slope of Generally, the rock typing process is finalized by reporting
the plateau section). the range of each rock type, a capillary pressure, and two
relative permeabilities curves (water-oil and gas-oil) as the
As the Figs. 3, 5, and 7 display the range of porosity and IWS representative of all existing curves for each rock type. In
changed for different rock types classified by the BVW meth- this step, the lack of optimum number for the rock types usu-
od. Their ranges increase especially for the water saturation ally results in very similar representative capillary pressure
with decreasing pore throat size. Therefore, it is expected curves for the middle rock types. This problem is added to
to see more variation in the shape of MICP curves that exist the existing challenge during feeding the rock types to the
in the last two rock types. Looking at MICP curves related models. Moreover, generating representative relative perme-
to these groups (by the BVW) in Figs. 12, 13, and 14 clearly ability curves for all rock types is challenging and sometimes
present these variations; though, some sort of similarity is seems unreal for some rock types. So, it is essential to keep a
distinguishable between them. In the case of the RQI meth- balance between the number of rock types and the similarity
od, the range of permeability extended more than ten times between the members for each rock type [12-14].
for a rock type. Moreover, due to existing secondary porosi-
ty in the carbonates (non-connected vugs or microfracture), Conclusions
increasing either porosity or permeability may not be cor- 1. Existing secondary porosity either in the shapes of un-
related with each other. These caused more variations in the connected vugs or microfracture and facies changes in
shape of MICP curves for a defined rock type by this method. different scales form a wide range of variations in the
For Field A, the porosity and permeability of rock types vary reservoir rock properties leading to the weak correla-
from 0.4 to 100 MD and 0.05 to 0.3, respectively (Fig. 9C). tion between porosity and permeability data for the
So, as it is expected, the MICP curves with different shapes carbonates. Therefore, the classification outcomes using
and curvatures exist for each rock type (Fig. 12). Although any method that applies relationships between these
the Rock type#1 with a limited range of variables showed two reservoir parameters cannot satisfy expecting re-
some sort of similarity between its curves, others contain sults for the rock type clustering.
Volume - 1 Issue - 1
Citation: R, Zahra. (2024). Rock Typing by Bulk Volume Water in the Carbonate Reservoirs. Journal of petroleum gas &chemical engineering, 1(1), 1-11.
Page 10 of 11
 Journal of Petroleum Gas & Chemical Engineering Copyright © Zahra Riazi

2. Applying the BVW rock typing method and classifying 2. Guo, G., Diaz, M. A., Paz, F., Smalley, J., Waninger, E. A.
reservoir rock using a fraction of water in rock volume (2005, October). Rock typing as an effective tool for per-
consider pore throat sizes in the studying section, which meability and water-saturation modeling: a case study
is a general factor between permeability and hydrocar- in a clastic reservoir in the Oriente Basin. In SPE Annual
bon saturation. Technical Conference and Exhibition? (pp. SPE-97033).
3. Examining the BVW method and comparing it with the SPE.
RQI method using three carbonate fields’ data in this 3. Ahrimankosh, M., Kasiri, N., Mousavi, S. M. (2011). Im-
study revealed, that the clustered MICP curves using proved permeability prediction of a heterogeneous car-
the BVW method presented more similarities between bonate reservoir using artificial neural networks based
themselves. In other words, the grouped MICP curves on the flow zone index approach. Petroleum science and
showed more discrepancy when they were clustered technology, 29(23), 2494-2506.
based on porosity-permeability correlations in the RQI 4. Riazi, Z. (2018). Application of integrated rock typing
method while classifying the reservoir rocks into more and flow units identification methods for an Iranian car-
rock types. bonate reservoir. Journal of petroleum science and engi-
4. Based on items 1 and 2 applying the BVW method seems neering, 160, 483-497.
more suitable for carbonates, while it is easy to use and 5. Buckles, R. S. (1965). Correlating and averaging connate
interpret. It can be applied as soon as the first well is water saturation data. Journal of Canadian Petroleum
drilled and can feed directly to the static and dynam- Technology, 4(01), 42-52.
ic model without imposing extra uncertainties on the 6. Morris, R. L., Biggs, W. P. (1967, June). Using log-derived
models. values of water saturation and porosity. In SPWLA annu-
al logging symposium (pp. SPWLA-1967). SPWLA.
Nomenclature 7. Greengold, G. E. (1986). The graphical representation of
bulk volume water on the Pickett crossplot. The Log An-
= average of IWS and porosity product in the specific group alyst, 27(03).
BVW= bulk volume water 8. Wu, T. (2004). Permeability prediction and drainage
k =permeability, mD capillary pressure simulation in sandstone reservoirs.
CCAL= conventional core analysis Texas A&M University.
DRT= discrete rock type 9. Jennings, J. B. (1987). Capillary pressure techniques: ap-
FZI= flow zone indicator, μm plication to exploration and development geology. AAPG
MLP= modified Lorenz plot Bulletin, 71(10), 1196-1209.
MICP= mercury injection capillary pressure, psi 10. Schowalter, T. T. (1979). Mechanics of secondary hydro-
pd = displacement pressure, psi carbon migration and entrapment. AAPG bulletin, 63(5),
pe = entry pressure, psi 723-760.
pt = threshold pressure, psi 11. Katz, A. J., Thompson, A. H. (1986). Quantitative pre-
RFN= rock-fabric numbers diction of permeability in porous rock. Physical review
RQI = reservoir quality index, μm B, 34(11), 8179.
𝜙= porosity, (fraction) 12. Lucia, F. J. (1995). Rock-fabric/petrophysical classifica-
Swir= irreducible water saturation, % tion of carbonate pore space for reservoir characteriza-
SCAL =special core analysis tion. AAPG bulletin, 79(9), 1275-1300.
SFP= stratigraphic flow profile 13. Petty, D. M. (1988). Depositional facies, textural char-
SMLP= stratigraphic modified Lorenz plot acteristics, and reservoir properties of dolomites in
SWPH= product of IWS and porosity Frobisher-Alida interval in southwest North Dako-
ta. AAPG bulletin, 72(10), 1229-1253.
References 14. Pittman, E. D. (1992). Relationship of porosity and per-
1. Amaefule, J. O., Altunbay, M., Tiab, D., Kersey, D. G., Kee- meability to various parameters derived from mercu-
lan, D. K. (1993, October). Enhanced reservoir descrip- ry injection-capillary pressure curves for sandstone
tion: using core and log data to identify hydraulic (flow) (1). AAPG bulletin, 76(2), 191-198.
units and predict permeability in uncored intervals/
wells. In SPE Annual Technical Conference and Exhibi-
tion? (pp. SPE-26436). SPE.

Volume - 1 Issue - 1
Citation: R, Zahra. (2024). Rock Typing by Bulk Volume Water in the Carbonate Reservoirs. Journal of petroleum gas &chemical engineering, 1(1), 1-11.
Page 11 of 11