You are on page 1of 5

Carbonate Facies and Permeability Estimation using Rock Physics and Flow-Zone Facies

Tim Pritchard, Cinzia Scotellaro and Robert Webber * BG Group

Summary The integration of an FZI approach with the geological


facies allows one to produce a more useful classification
Modeling permeability in heterogeneous carbonate scheme upon which a model can be built.
reservoirs may be challenging due to often extreme spatial In this approach it is most important to focus on the
variations in pore geometry. This paper introduces a characterization of the pore types with the highest FZI, or
method to model permeability in terms of Flow-Zone best permeability. This is because in the permeability-
Facies, and demonstrates this approach in a case study. The porosity plane, samples with mixed pore types plot between
method is utilized to model facies and permeability across the end-members, but tend to be biased toward the pore
thirty wells in a field with limited data availability. Rock type of highest permeability (Lønøy, 2006).
physics relationships and a modified Kozeny-Carman In practice, this approach involves taking the multitude of
permeability equation are used to define facies and predict facies typically interpreted by a geologist and reducing
permeability. The results indicate that capturing complex them to a few useful facies that adequately constrain the
variations in pore geometry within a rock is the key to the permeability behaviour of the reservoir. We define these as
successful application of this method. Flow-Zone Facies because they are geological units that
control the permeability of a formation.
Introduction The distribution of these Flow-Zone Facies are known at
the well locations from core analysis, but must be predicted
The carbonate reservoir discussed in this paper contains a in the wells where no core has been taken. This is the
wide variety of pore types. The challenge in predicting critical part of the workflow, and an objective and
permeability in such a reservoir is to construct a model that repeatable method must be established to predict the Flow-
effectively describes these pore types, without becoming Zone Facies distribution.
too complex, so that the prediction of these pore-types Finally a permeability model must be created to estimate
becomes impractical within the dataset available. permeability in each of the Flow-Zone Facies.
The foundation of the pore-type model should be the
geological interpretation of the reservoir. However, Applied Workflow
geologists often interpret a multitude of geological facies,
the classification of which may have no imprint in the 1. Identify the group of facies that represents a pore-
permeability domain. type with the best permeability or FZI character, and
The Flow Zone Index (FZI) is a useful approach to describe define this as Flow-Zone Facies 1. Establish a method
the porosity and permeability character of a pore system to predict the distribution of this Flow-Zone Facies in
(Amaefule et al., 1993). Integration of this approach with wells without core, and create a permeability model
the geological facies classification helps make sense of the for this Flow-Zone Facies.
inherent complexity of such pore systems. 2. Identify the group of facies that represents the pore
Flow zone index is defined in terms of porosity and type with the second-best permeability character, and
permeability as follows: define this as Flow-Zone Facies 2. Establish a method
to predict the distribution of this Flow-Zone Facies in
wells without core, and create a permeability model
for this Flow-Zone Facies.
3. Repeat the above steps until it is no longer possible
or productive to discriminate Flow-Zone Facies based
on the data available.
The availability of data is a critical constraint on this
method for modeling facies and permeability. Previous
where, authors have proposed linking a Flow Zone Indicator
k = permeability approach with pore size distribution data from NMR and
= Porosity borehole electrical image logs (Gomaa et al., 2006,
RQI = Reservoir Quality Index Rahmawan et al., 2009). In this case study only limited
= pore volume-to-grain volume ratio well data were available such as caliper, gamma-ray,
FZI = Flow Zone Indicator resistivity, P-wave sonic, neutron porosity, and density log
data, so a simpler method was required.

© 2010 SEG
SEG Denver 2010 Annual Meeting 2654
Carbonate Facies and Permeability Estimation

Facies Model Method this flow zone facies in order to create an efficient reservoir
model.
The carbonate reservoir that forms the basis of this case Figure 4 shows SEM and CT scan photos of Flow Zone
study has had core extracted from several wells. Figure 1 Facies 1 at different resolutions. This sample has 18.1%
shows porosity versus permeability for the available core porosity and 10.8 mD permeability. Moldic pores (MO) are
data that have been classified into geological facies. 11 evident in the thin section together with intergranular
geological facies have been interpreted in this data set. The porosity; the microporosity is shown in image D.
reservoir quality is closely related to facies and grain size,
and is primarily intergranular in origin, but also contains
intragranular, mouldic, intercrystalline porosity, and
microporosity.
The permeability can be highly variable and, within this
reservoir, a specific porosity value may have 4 orders of
magnitude variation in the permeability values associated
with it.
Figure 3: Histograms of three Flow Zone Facies

Figure 4: Core porosity versus permeability crossplot, color coded


Figure 1: Porosity versus permeability core data, color coded by by Flow-Zone Facies
depositional facies from geological interpretation.

A facies and permeability model containing eleven facies


would be impractical to be implemented, and would
contain unnecessary details, as many of these facies do not
have an individual expression in terms of permeability. The
geological facies should therefore be grouped, according to
permeability character into Flow-Zone Facies.
The FZI is calculated for each of the geological facies.
These geological facies are then assigned to Flow-Zone
Facies that show separation in the FZI domain.
Figure 2 shows histograms of the three final groups of
Flow-Zone Facies. FZI is plotted in logarithmic scale and
each Flow-Zone Facies shows a distinct FZI character.
Figure 3 shows the distribution of these Flow Zone Facies
plotted in the core analysis based porosity versus
permeability cross-plot.
Flow-Zone Facies 1 is the most important flow zone facies
in this reservoir. These are interpreted to be foreshore
beach gravels, or rudist rudstones, possibly occurring as
storm deposits. This Flow-Zone Facies exhibits high
intergranular and mouldic/vuggy porosity up to 27% and
permeabilities up to 5 Darcies. PLT interpretation shows
Figure 2: SEM (above) and CT scan (below) images of Flow Zone
that the dominant flow in wells is from this flow-zone
Facies 1 at different resolutions.
facies, and therefore it is critical to map the distribution of

© 2010 SEG
SEG Denver 2010 Annual Meeting 2655
Carbonate Facies and Permeability Estimation

Flow-Zone Facies 2 is the second most important Flow-Zone Facies 2 are described as dolomitised lower
geological facies. These are interpreted to be dolomitised shoreface facies. The distribution of this Flow-Zone Facies
lower shoreface deposits. In this flow zone facies the can be estimated by calculating the dolomite fraction of the
dolomitisation improves permeability by reducing the matrix. This formation contains both calcite and dolomite
tortuosity of the pore network. The pore connectivity is components, and a standard output of a porosity calculation
increased in the crystal-supported framework, however is the apparent matrix density. When the apparent matrix
dolomitisation does not always improve porosity. density is calibrated to core derived matrix density data, the
Flow-Zone Facies 3 is defined as all other geological Flow Zone Facies 2 can be predicted and associated with
facies. These remaining geological facies show a similar high apparent matrix density.
porosity-permeability relationship, so that for the simplicity Figure 6 shows a comparison between predicted Flow Zone
of the model and with the constraints of the data Facies 1 and 2 as well as the geological facies interpreted
availability, there is little additional value gained in further on core material.
subdivision of Flow Zone Facies.

The next step is to establish an objective and repeatable


method for predicting these Flow-Zone Facies in wells
without core.
Within the well log dataset Flow-Zone Facies 1 is seen to
exhibit a unique signature in the density versus P-wave
velocity cross-plot.
Figure 5 shows a density versus P-wave velocity cross-plot
for both the Flow-Zone Facies 1 data and also all remaining
core data. The area in the red ellipse shows the location of
the Flow-Zone Facies 1 in this plot.

Figure 6: Quality control of geological interpretation from core


compared to predicted Flow-Zone Facies (last two tracks
respectively). Also shown are the input well logs and the porosity
and permeability from this work compared to core analysis data.

Permeability Model Method

Figure 5: Density versus P-wave velocity cross plot. The overlays A modified Kozeny-Carman relationship was used to create
are modified rock physics bounds and trends for P-wave velocity a permeability model for each of the Flow-Zone Facies.
versus density (rather than porosity) assuming a calcite matrix
(Avseth, Mukerji, and Mavko, 2005) The modified Kozeny-Carman relation (Dvorkin et al.,
2009; Mavko et al., 2009; Mavko and Nur, 1997) provides
a way to estimate the permeability of a porous medium in
Velocities can be highly variable in carbonates because of terms of generalized parameters such as porosity ( ), pore
depositional processes and diagenesis. (Eberli et al., 2003;
diameter (d), tortuosity ( ), and percolation porosity ( p) as
Rafavich et al., 1984). Pore type and shape is a critical
follows:
control of velocity. In this example, the pore types of the
foreshore beach gravels (Flow Zone Facies 1) are less
compliant or ‘stiffer’ than pore types of other Flow-Zone
Facies. This enables for the discrimination of this Flow-
Zone Facies in wells without core.
The physical interpretation of this effect is that these
moldic pore types are stiffer than other pore types
(intercrystalline, intergranular or microporous) because the Figure 7 shows the modified Kozeny-Carman permeability
pore shapes are strengthened during diagenetic changes. relationships derived for each Flow-Zone Facies. On the
This rock physics signature allows the distribution of this right-hand side of each is a plot that shows the estimated
flow zone facies to be determined using p-wave sonic and permeability for each Flow-Zone Facies against the
density well logs throughout the field. measured core permeability for that facies.

© 2010 SEG
SEG Denver 2010 Annual Meeting 2656
Carbonate Facies and Permeability Estimation

When these three models are applied to the Flow-Zone


Facies model in the wells with core, the resulting
permeability is quality controlled by plotting against the
measured core permeability.
Table 1 shows the parameters used for the Modified
Kozeny-Carman permeability models for each of the Flow-
Zone Facies.

Figure 7: Estimated permeability versus measured core


permeability for all the Flow-Zone Facies.

Conclusions

Integration of geological facies interpretations with a Flow


Zone Indicator approach enables a Flow-Zone Facies
classification scheme to be generated that honors both the
geological interpretation and the pore geometry
complexity. We define a Flow-Zone Facies as a geological
unit or group of geological units that has a distinct
expression in the FZI domain, and controls the permeability
Figure 8: Modified Kozeny-Carman relationship for each Flow of a formation. This method facilitates the simplification of
Zone facies and the corresponding estimated permeability versus a typical geological facies scheme and enables meaningful
measured core permeability used as quality control of the model. permeability models to be generated. A high coefficient of
correlation observed in the estimated versus measured core
permeability was achieved, which was the primary goal of
this study.
Complex variations in pore geometry may be identified
through the pore stiffness effect imposed on a rock fabric
by diagenisis. In this way, velocity-density relations may be
Table1: Modified Kozeny parameters for Flow Zone Facies utilized to infer pore types in complex carbonate reservoirs.
In this paper the proposed approach enabled us to create a
predictive model that captured the complex variations in
Figure 8 shows the fit between measured and estimated pore types, so that the distribution of permeability could be
permeability. The high correlation coefficient between mapped using wells with limited data availability.
measured and estimated permeability (R2 ranging between
0.7 and 0.8) holds over 7 orders of magnitude. This result Acknowledgments
indicates that the primary objective of the study has been
achieved. We are grateful to our colleagues Phil Nadin and Corrado
Bruni who provided insight into this complex data set. We
The quality control of the Flow-Zone Facies model and also thank the Petrophysics and Carbonate Geological
estimated permeability means that this method is suitable Team in BG Advance for their technical support and
for predicting permeability throughout the field wells, and direction, in particular Adam Moss for his expertise.
the resulting reservoir model should contain a distribution We would also like to thank Jack Dvorkin and Gary Mavko
and dynamic range of permeability that honors the input from the Stanford Rock Physics and Borehole Group for
core analysis data. their support.

© 2010 SEG
SEG Denver 2010 Annual Meeting 2657
EDITED REFERENCES
Note: This reference list is a copy-edited version of the reference list submitted by the author. Reference lists for the 2010
SEG Technical Program Expanded Abstracts have been copy edited so that references provided with the online metadata for
each paper will achieve a high degree of linking to cited sources that appear on the Web.

REFERENCES

Amaefule , J. O., M. Altunbay, C. Tiab, D. G. Kersey, and D. K. Keelan, 1993, Enhanced Reservoir
Description: Using Core and Log Data to identify Hydraulic (Flow) Units and predict Permeability
in Uncored Intervals/Wells: 68th Annual Technical Conference and Exhibition of the Society of
Petroleum Engineers, SPE 26436.
Avseth, P., T. Mukerji, and G. Mavko, 2005, Quantitative Seismic Interpretation: Cambridge University
Press.
Dvorkin , J., N. Derzhi, Q. Fang, A. Nur, B. Nur, A. Grader, C. Baldwin, H. Tono, and E. Diaz, 2009,
From micro to reservoir scale: Permeability from digital experiments: The Leading Edge, 28, no. 12,
1446–1453, doi:10.1190/1.3272699.
Eberli, G. P., G. T. Baechle , F. S. Anselmetti, and M. L. Incze, 2003, Factors controlling elastic properties
in carbonates sediments and rocks: The Leading Edge, 22, no. 7, 654–660, doi:10.1190/1.1599691.
Gomaa, N., A. Al-Alyak, D. Ouzzane, O. Saif, M. Okuyiga, D. Allen, D. Rose, R. Ramamoorthy, and E.
Bize, 2006, Case study of Permeability, Vug Quantification, and Rock Typing in a Complex
Carbonate: SPE Annual Technical Conference and Exhibition held in San Antonio, SPE 102888.
Lønøy, A., 2006, Making sense of carbonate pore systems: AAPG Bulletin , 90, no. 9, 1381–1405,
doi:10.1306/03130605104.
Mavko, G., and A. Nur, 1997, The effect of a percolation threshold in the Kozeny-Carman relation:
Geophysics, 62, 1480–1482, doi:10.1190/1.1444251.
Mavko, G., T. Mukerji, and J. Dvorkin, 2009, The rock physics handbook: Cambridge University Press.
Rafavich, F., C. H. St. C. Kendall, and T. P. Todd, 1984, The relationship between acoustic properties and
the petrographic character of carbonate rocks: Geophysics, 49, 1622–1636, doi:10.1190/1.1441570.
Rahmawan, I., Y. Aditiah, A. Kurniawan, T. Firmanto, J. Priyantoro, and E. E. Gomaa, 2009, Estimating
Permeability in Uncored Wells Using Modified Flow Zone Index: SPE Asia Pacific Oil and Gas
Conference and Exhib ition, SPE 122490.

© 2010 SEG
SEG Denver 2010 Annual Meeting 2658

You might also like