JSUT 32(2), 2006, p.

29-38

Characterization of Reservoir Rock Types in a Heterogeneous Clastic and

Carbonate Reservoir
Bagheri A.M.*, Biranvand B.
Research Institute of Petroleum Industry (RIPI), 7 Hovaizeh Ave., Tehran, Iran
*
Corresponding author, e-mail: [email protected]
(received: 6/6/2005 ; accepted: 6/9/2005)

Abstract
Identifying reservoir rock types and their most significant vertical and horizontal heterogeneities is an essential
component of reservoir characterization process, which are among the key input parameters into a three-dimensional
geological and flow simulation models. A reservoir classification and rock typing study were carried out on the
Asmari formation of a mixed siliciclastic and carbonate reservoir in Iran. Detailed core analysis data including
capillary pressure, core porosity, core permeability and core description supplemented by well logs revealed a

D
complete vertical sequence of seven distinct clastic and carbonate reservoir rock types. Identification of the reservoir
intervals and pay zones was carried out by means of the above results. Core based reservoir rock types were examined
for each cored wells and log based reservoir rock types were selected and assigned in the uncored wells. The above

SI
data were applied as input parameters in a method based on Fuzzy Logic inference. The Fuzzy Logic technique was
calibrated in 4 cored wells and blind tested in the other cored wells to determine the reservoir rock types. After the
secondary calibration of the Fuzzy Logic against the core data, this technique was applied on 28 wells without any
core data. The results reveal a very good match between the core data analyses and the Fuzzy Logic determination of
the reservoir rock types. This technique can be applied to reduce the uncertainty of determination of the rock typing or
of
as a very good predictor in uncored wells.

Keywords: Rock types, Capillary pressure, Permeability, Petrophysical logs, Fuzzy logic.

Introduction The Asmari formation in the southwest of Iran is

ive

It is well known that accurate reservoir simulation one of the most important reservoirs in the world.
and management requires a quantitative model of This formation is predominately a carbonate unit,
the spatial distribution of reservoir properties and an but in the central area of the Dezful Embayment, it
understanding of the nature of reservoir is a mixed siliciclastic and carbonate reservoir.
heterogeneity at many scales.
ch

Reservoir rock type determination has presented a Core Study

challenge for cases whenever no direct measure- The core study approach was conducted on:
ments of reservoir rock type are available. The 1) Sedimentological description of more than 2000
direct determination of reservoir rock type will be m of cores from 17 wells.
Ar

carried out though the core analysis while indirect 2) Microscopic study of more than 6000 thin
determination will be carried out through the log sections from different cored wells.
analyses. Typically, few wells in a field may have 3) Study of 16 samples from different reservoir rock
laboratory information such as core analysis data types by Scanning Electron Microscopy (SEM).
whereas most of wells may have electronic logs 4) Conventional core analysis of more than 1500
data. Wells without core are usual due to various core plugs.
reasons such as, time and cost associated with 5) Mercury Injection Capillary Pressure (MICP),
coring, and or impractical coring in many situations, curve analysis on 16 core plugs.
such as in horizontal wells. Mineralogy, Rock Fabric and texture
However, a method was applied based on the Thin section petrographical analysis were carried
Fuzzy Logic inference, in some wells where core out to determine mineralogy, rock fabric, texture,
data are not available, to determine the reservoir pore geometry and distribution, grain character-
rock types from wire-line logs data in the Asmari istics, diagenetic features and sedimentary structures
formation of a heterogeneous reservoir in Iran. for more than 6000 side wall core thin sections from
17 wells.

www.SID.ir
30 Bagheri & Biranvand JSUT, 32 (2), 2006

For a detailed identification of the above character- attributes such as: grain size, sorting, roundness,
istics, Scanning Electron Microscopy (SEM) ana- puerility and maturity. Grain size is a highly
lysis was carried out on 16 samples selected from variable and ranges in extreme cases between silt
various reservoir rock types of wells. and very coarse sand grade. Sorting of various
Based on the petrographic observations, the rock sediments ranges from poor to very well sorted with
composition of the Asmari reservoir varies between a mode of moderately well sorted.
siliciclastic and carbonate lithologies. In this
reservoir quartz and dolomite are dominant minerals Porosity-Permeability relationship
where calcite, anhydrite, clay minerals, potassium The permeability-porosity relationship by means
feldspar and iron oxides are among other abundant of cross-plots was studied for various rocks types. In
minerals. Carbonate rocks in the reservoir show a summary graphs presented in Figures 1-2,
highly variable depositional fabric and texture. In permeability-porosity trends for siliciclastic and
the present study, the Dunham’s (1962) classifi- carbonate reservoir rock types were plotted together.
cation, was applied. Fabric and textural character- The results from these cross-plots study reveal that
istics in this method depend on the depositional the permeability of the sandstone (Fig. 1) is well

D
environments and particle types. Most of petro- defined by the porosity, whereas in the carbonate
physical properties, such as porosity, permeability, (Fig. 2) has a more diffused clouds which indicating

SI
water saturation, mercury injection and capillary to other major factors affecting the permeability.
pressure data depend on fabric and textural chara- High porosities in carbonates can be observed that
cteristics. Siliciclastic sediments show a relatively does not give rise to high permeability. This
consistent depositional fabric. In the silicilastic rock property of carbonates is well known as poor
study, the great emphasis was placed on textural connectivity of the vugs.
of
ive
ch
Ar

Fig. 1. Permeability versus porosity, siliciclastic reservoir rock types.

www.SID.ir
Reservoir Rock Types in a Heterogeneous Clastic and Carbonate Reservoir 31

D
SI
Fig. 2. Permeability versus porosity, carbonate reservoir rock types.
of
In the clay cemented sandstones, high porosity can them. The reservoir samples were converted from
be observed. This high porosity is mainly in the laboratory air/ mercury system to the subsurface
form of micro–porosity filled with chemically and brine/ hydrocarbon system of the reservoir then the
ive

physically (capillary) bound water, which is mercury injection data was applied to the reservoir
immobile. Since this high porosity does not take samples. MICP results and pore throat size curves
place in fluid flow, the permeability in the clay- for selected samples are shown in summary graph,
cemented sandstones is low. Fig. 3, and summarized in Tab. 1. Capillary pressure
ch

However the clusters of points for each reservoir data were applied to distinguish between reservoir
rock type are not totally distinct from each other. and none-reservoir rocks, pay and none-pay, on the
The overlapping of depositional lithofacies in these basis of none-wetting-phase saturations (see Tab. 1).
graphs is probably caused by a number of factors This will be carried out by study of the displacement
Ar

including diagenesis, fracturing, similarity in pore of capillary pressure curve, while pore throat radius
structure produced by different depositional was applied to categorize the rock by pore-type, e.g.
environment and the somewhat subjective nature of nanno, micro, meso, macro and mega.
specifying reservoir rock type based on hand
specimens. Identifying Reservoir Rock Types in Cored Wells
Development of a rock–fluid model in the Asmari
Mercury Injection Capillary Pressure Curves reservoir was carried out with the input data from
Mercury injection capillary pressure (MICP) curves the geological and petrophysical results. The results
were applied to determine the pore geometry and the were applied to identify various reservoir rock types
pore throat size of porous rocks in the reservoir. in the reservoir, poor reservoirs or non-reservoir
Sixteen samples, were selected, described, rocks. A reservoir rock type is defined as an interval
photographed and made into thin sections for of rock with unique pore geometry, determined
mercury injection capillary pressure analysis. These mineralogical composition and is related to certain
samples were prepared from conventional core specific fluid-flow characteristics.
plugs when routine core analysis was carried out on

www.SID.ir
32 Bagheri & Biranvand JSUT, 32 (2), 2006

D
SI
of
Fig. 3. Capillary pressures, mercury injection curves show pore size distribution and mercury filled pore
volume for various rock.
ive
ch
Ar

Fig. 4. Thin sections photomicrographs showing: a- RT1: well sorted and well rounded fine grain clean sands.
b- RT2: partially cemented subangular fine to medium grain sandstone. c- RT3: sandy and clay siltstone. d-
RT4: Limestone with grainstone/packstone fabric. e- RT5: porous dolomite. f- RT6: sandy dolomite with
anhydritic cement. g- RT7: anhydritic patches with chicken wire fabric.

www.SID.ir
Reservoir Rock Types in a Heterogeneous Clastic and Carbonate Reservoir 33

Tab. 1. Summary of petrophysical and geological properties for various rock types.
ROCK TYPES
RT1 RT2 RT3 RT4 RT5 RT6 RT7
PETROPHYSICAL
PARAMTERS

Frible to Claystone, Anhydritic

Anhydritic Crystalline
Partially Silticlaystone Grainstone Dolostone/ Nodular
LITHOFACIES Cemented
Consolidated
and clayey Packstone
Dolostone/sandy
Sandy Anhydrite
Sandstone Dolostone
Sandstone siltstone Dolostone
CHARACTERISTICS

Maximum Air
35.4 29.2 29.3 22.3 30.7 17.6 17.8
Porosity (%)
Minimum Air
12 1 0.4 0.2 1.9 0.4 0.1
Porosity (%)
Average Air
27.4 13.4 11.6 7.4 15.3 7.4 2.4
Porosity (%)

Air Porosity Excellent Moderately Moderately Fair Good Fair Low

Interparticle
95.8 84.6 100 29.3 9.9 9.7 50
Porosity (%)

D
Vuggy Porosity
4.2 14.1 0 42.7 65.4 72.2 0
(%)
Interaparticle
0 0 0 25.6 9.9 6.9 0
Porosity (%)
POROSITY

SI
Moldic Porosity
0 0 0 1.2 11.1 6.9 0
(%)
Intercrystaline
0 1.3 0 1.2 3.7 4.2 50
Porosity (5)
Maximum Helium
CHARACTERISTICS

3000 1243 63.3 233 577.72 41.53 23.25

Permeability (md)
PERMEABILITY

of
Minimum Helium
2.15 0.01 0.01 0.01 0.02 0.01 0.01
Permeability (md)
Average Helium
690.04 81.73 7.08 4.31 23.32 1.83 0.83
Permeability (md)
Helium
Excellent Good Fair Fair Good Fair Low
Permeability
ive

C-factor
CHARACTERISTICS

(measure of pore Medium Medium Good Medium Poor Medium Medium

sorting)
Sm (Unsaturated
Low Medium-High High High Medium High High
MICP

pore volume)
Pd (Entry
Very Low Medium Medium High Low High High
ch

Pressure)
Pore Size
Distribution Meso-Mega Micro-Meso Micro Micro-Meso Micro-Mega Micro-Meso Nano-Micro
(MICPM)
None None None None
RESERVOIR QUALITY Net Pay Net Pay Net Pay
reservoir reservoir reservoir reservoir
Ar

The first attempt in this study was to apply porosity- reservoir rock porosity, permeability and pore throat
permeability ratio criteria from the routine core radius ranges, by reservoir rock type are presented
analysis to classify the reservoir rock type. Then in Tab. 1. Fig. 4 show thin section photomicro-
these criteria were compared with visual porosity graphs from various rock types. Grain character-
and rock fabric characterized from microscopic istics and pore filling materials for various rock
studies in addition to the results based on log data. types can be achieved from SEM photomicrographs
MICP curves were analyzed to confirm validity of presented in Fig. 5.
the reservoir rock type classification. The capillary However, in this study seven reservoir rock types
pressure data was also applied to distinguish was designated and identified from RT1 to RT7
reserveoir rock from none-reservoir rock and pay with individual pore geometries, mineralogy, and
from none-pay. A better understanding of the be- fluid-flow characteristics, in the Asmari reservoir. In
havior of capillary pressure curves will be achieved addition, by applying these criteria, the reservoir
when it is integrated with the information provided was divided into seven zones.
by the thin sections and SEM micrographs. The

www.SID.ir
34 Bagheri & Biranvand JSUT, 32 (2), 2006

D
SI
Fig. 5. Scanning Electron Microscope photomicrographs showing grain characteristics (RT1, RT2 and RT5)
and pore filling materials (RT2, RT3, RT4 and RT6).
of
Estimation of reservoir rock types applying et al. (1994) used fuzzy clustering to determine
Fuzzy Logic lithology from well logs in Upper Carboniferous
Fuzzy logic is an extension of conventional Boolean coal deposits of the Ruhr basin. Cuddy (1997) used
logic (zeros and ones) developed to handle the fuzzy logic to predict permeability and lithofacies in
ive

concept of “partial truth” values between “com- uncored wells to improve well-to-well log correla-
pletely true” and “completely false”. In contrast to tions and 3-D geological model building. Saggaf &
binary-valued (bivalent) logic, truth is ascribed Nebrija (2003) used fuzzy logic approach for the
either 0 or 1, multivalent logic can ascribe any estimation of facies from wire-line logs in a field in
ch

number in the interval [0,1] to represent the degree Saudi Arabia.

of truth of a statement. This is a normal extension of
bivalent logic, and it is a form of logic that humans Application to well data
practice naturally. Zadeh introduced it in the 1960's The Fuzzy Logic inference method was applied to
Ar

as a means to model uncertainty. identify and determine the kind of reservoir rock
More common use of fuzzy logic is to describe type in the uncored wells based on data from wire-
the logic of fuzzy sets (Zadeh 1965). These are sets line logs in the Asmari-reservoir. Reservoir rock
that have no crisp, well-defined boundaries, and type determination applying the Fuzzy Logic is
which may have elements of partial instead of full based on the fact that a known reservoir rock type
membership. For fuzzy sets, elements are can give any log reading although some readings are
characterized by a membership function that more likely than others. In this method several con-
describes the extent of membership (or the degree of ditions based on wire-line data were applied to
fit) of each element to the set. Such a membership determine the reservoir rock type and reduce the un-
function maps the entire domain universe to the certainty of the determination. As an example, two
interval [0, 1]. reservoir rock type will be discussed here, partially
Fuzzy mathematical techniques have been applied cemented sand (RT1) and shaly sandstone (RT3).
to solve various petroleum engineering and geo- Partially cemented sand has most likely high poro-
logical problems in the past, involving mainly sity and low gamma ray radiation while shaly sand-
classification, identification, or clustering. Toumani stone has low porosity with high gamma radiation.

www.SID.ir
Reservoir Rock Types in a Heterogeneous Clastic and Carbonate Reservoir 35

It is obvious that the Fuzzy Logic inference is not rock type by Fuzzy Logic method the cored-well A
just a simple probabilistic method. It is based on was selected and core data from other wells were
measured data. applied in this well. The core description in this well
The membership functions are based on was applied for a comparison with the result from
determination of reservoir rock type applying the the Fuzzy Logic and calibration of the input data of
wire-line data and core description. The core the method. The core-derived determination of
determinations were derived from examination of reservoir rock types from well B displays in the last
known reservoir rock type of cored wells in the area. track of Fig. 7 while the Fuzzy Logic determination
These membership functions will be applied to displays in the last but one track. The logs of the test
identify the reservoir rock types in an uncored well well A were passed through the Fuzzy Logic
by means of wire-line log. There is no limitation on inference system to determine the reservoir rock
the number of input logs in this method. However, types in this well. This determination was compared
the additions of further curves may not reduce the with core description from this well (see Fig. 8).
uncertainty of the determination of reservoir rock The determination of reservoir rock type based on
type, but, it is important to have a consistent set of the Fuzzy Logic is in agreement with the result from

D
logs in all wells. A sensitivity study was carried out core analysis. This comparison between two
based on five selected logs due to their relative methods reveals a good to very good results. The

SI
importance in rock typing. These logs are: gamma core analyses from 5 cored-wells were applied to
ray (GR), effective porosity (PHIE), neutron determine reservoir rock types in 28 uncored wells
porosity (NPHI), density (RHOB) and sonic (DT). with the geometry of the reservoir to construct a
The membership distribution for effective porosity representation of the initial (static) state of the
log in RT1 and RT6 as an example are presented in reservoir, having a specified resolution, quality and
of
Fig. 6. The membership functions were calculated accuracy. The section is shown in Fig. 9, represents
from the logs and core-derived reservoir rock types the estimated reservoir rock types over the whole
in four cored-wells (B, C, D and E). In order to test field applying a geostatistical method.
the uncertainty of the determination of reservoir
ive
ch
Ar

Fig. 6. Membership distributions for porosity log in RT1 and RT6, examples of reservoir and none reservoir
rock types.

www.SID.ir
36 Bagheri & Biranvand JSUT, 32 (2), 2006

D
SI
of
ive
ch
Ar

Fig. 7. Comparison of the reservoir rock types determination between the Fuzzy Logic inference method and
reservoir rock type determination based on core analyses, the two last tracks. The first six tracks represents
log data and interpreted lithology and facies, respectively.

www.SID.ir
Reservoir Rock Types in a Heterogeneous Clastic and Carbonate Reservoir 37

D
SI
of
ive
ch
Ar

Fig. 8. Blind-testing prediction in the test well A.

www.SID.ir
38 Bagheri & Biranvand JSUT, 32 (2), 2006

D
Fig. 9. Cross section showing estimated reservoir rock types over the field.

Conclusion
High uncertainty occurs when the determination of
the reservoir rock type in an uncored well will take
place. Rock typing was applied for well correlation
SI
inference process. Application to well data indicates
that the method can determine the reservoir rock
types of uncored wells with a good accuracy that
rivals those of other methods, such as methods
of
and as input data to build a 3D model of the based on statistics. This method can be applied to
reservoir. In this study an intelligent method was reduce the uncertainty of determination of the
applied for rock-type analysis through the integra- reservoir rock type or as a very good predictor in
tion of core, conventional open-hole logs and geo- uncored wells.
logical data. Applying this method requires only the standard
ive

The Fuzzy Logic is inherently well suited to electronics logs such as porosity and density than
characterizing vague and imperfectly defined know- complex and special logging system.
ledge, and it can thus yield a simple and more
accurate description. The Fuzzy Logic inference Acknowledgment
ch

system allows an engineer to incorporate his basis The authors would like to thank Research Institute
and previous knowledge and experience, as well as of Petroleum Industry, RIPI for their support and
general engineering principles and notions, into the giving permission to publish.
Ar

References
Bagheri A.M., Biranvand B., Fasih M., Bakhtiari H. 2004: An application of Fuzzy Logic to estimate
lithological facies in a complex reservoir. 32nd Int. Geol. Cong. (32nd IGC), Florence, Italy. Pp. 1199-
1200.
Cuddy S.J. 1997: The application of Fuzzy Logic to petrophysics. 38nd Ann. Log. Symp. Soc. Prof. Well Log
Ana. Houston, Texas. Pp. 1-14.
Cuddy S.J. 2000: Litho-facies and permeability prediction from electrical logs using Fuzzy Logic. SPE
Paper 65-411.
Donham R.J. 1962: Classification of carbonate rocks according to depositional texture. In: Ham WE (ed),
Classification of carbonate Rocks-a symposium. AAPG Mem. 1: 108-121.
Saggaf M.M., Nebrija Ed.L. 2003: A Fuzzy Logic approach for the estimation of facies from wire-line logs.
AAPG Bull. 87(7): 1223-1240.
Toumani A., Schmitz D., Schepers R. 1994: Automatic determination of lithology from well logs using
fuzzy classification. 56th Meet. Europ. Asso. Expl. Geophysicists. Paper H041.
Zadeh L. 1965: Fuzzy sets. Info. Control 8: 338-353.

www.SID.ir