SPE 149104

Cementation Exponent Estimation for Complex Carbonate Reservoirs

Using a Triple Porosity Model

Ali Al-Ghamdi, Roberto Aguilera and Christopher R. Clarkson, University of Calgary
Copyright 2011, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE/DGS Saudi Arabia Section Technical Symposiumand Exhibition held in Al-Khobar, Saudi Arabia, 1518 May 2011.

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, as presented, does not necessarily reflect any position of the Society of Petroleum
Engineers, its officers, or members. Papers presented at the SPE meetings are subject to publication review by Editorial Committee of Society of Petroleum Engineers. 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 where and whom the paper was presented. Write Librarian, SPE, P.O. Box
833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.


Abstract
A previously defined triple porosity model is used to calculate the cementation exponent (m) of complex carbonate reservoirs
in the Middle East using well log data. The cementation exponent is usually affected in carbonate rocks by different types of
primary and secondary porosities. A combination of interparticle porosity, non-connected porosity (e.g., vuggy and fenestral)
and fractures increases the uncertainty in the estimation of m. Therefore, a well-defined petrophysical approach must start by
first understanding the rocks fabric.
Initially, samples are classified into different flow units based on pore throat apertures at 35% cumulative pore volumes (r
p35
).
This classification is then extended to include variations in porosity types based on geological and petrophysical descriptions
of each rock. Each sample has different proportions of connected and non-connected porosities. These porosities are defined as
matrix, fractures and non-touching vugs (including fenestral porosity). The porosity types are extracted fromwell logs for the
whole reservoir section and are cross-checked against core samples and thin sections.
The value of m in a triple porosity systemcan be larger, equal, or smaller than the cementation exponent of only the matrix
blocks (m
b
). This variation depends on the relative contribution of natural fractures and non-touching vugs compared to the
composite triple porosity reservoir. A continuous curve of m values is obtained using this model. A good comparison has been
obtained between the results of this model and m values measured in the laboratory.
Estimation of variablem values within short distances in a given reservoir using the triple model is a significant development
in formation evaluation that helps reduce uncertainty in petrophysical calculations. The results increase the confidence level in
water saturation and reserves determinations.
Introducti on
A broad assumption that is generally made in water saturation calculations is that the cementation exponent (m) is constant for
the whole reservoir interval. This assumption creates a large uncertainty in water saturation estimation in heterogonous
reservoirs. Figure 1 shows the error expected in water saturation calculations if the m value is under or over-estimated. A
small error in m causes significant impact by decreasing or increasing Sw. Since most carbonate reservoirs have heterogeneous
2 [SPE-SAS-825]
rock properties due to microbial and diagenetic processes, changes in rock properties should be reflected in the value of m.
Archies cementation exponent (m) is defined as a parameter that is affected by connectivity of the pore system. The
cementation exponent depends on the grain and pore shapes and how they are connected (Salemand Chilingarian, 1999). The
generally accepted average value of m is 2, where it represents interparticle porosity (Archie, 1942). In fractured rocks, the
value of m tends to get smaller because the rocks have better connectivity for electrical current (Towle, 1962; Aguilera, 1976;
and Rasmus, 1983). These researchers suggest using parallel dual porosity systems for such rocks. In rocks with a high
percentage of non-touching vugs, the value of m tends to be higher than 2 as reported by Lucia (1983) and Focke and Munn
(1987). However, in carbonate reservoirs, fractures and vugs coexist with the matrix (inter and intraparticle connected)
porosity in different proportions that change with depth. Aguilera and Aguilera (2004) proposed a physical model of triple
porosity systems, which was subsequently improved by Al-Ghamdi et al., (2010) to account for the presence of matrix,
fractures and vugs.
The physical model of triple porosity systems allows the estimation of cementation exponent (m) for heterogeneous
formations. The model assumes non-connected (separate or non-touching) vugs, matrix and fractures that contribute to the
total porosity of the system. The assumption is made that the matrix and fractures have conductivities that are connected in
parallel. The combined matrix and fractures are connected in series with the non-connected vugs. The triple porosity system,
also referred to as the composite systemin this paper, is given by the equation (Al-Ghamdi et al., 2010):

log
]
/ ) 1 (
) 1 (
log[
2 2
2
b
m
b nc
nc
nc
m

. (1)
For optimumresults, the physical model needs to be validated in complex triple porosity carbonate rocks. In this study,
matrix, fractures and vug porosities were estimated initially from petrographic work. This also helped to understand the
internal anatomy fabric of the rocks prior to evaluating the well logs. This approach provided reasonable estimates of the
cementation exponent (m) as a function of depth.
Rock Typing
Carbonate samples were collected from one of the Saudi Arabian fields. Geological descriptions of the reservoirs were
necessary to understand the pore structure before establishing the petrophysical rock classification. Reservoir A, Fig. 2,
consists of five packages that include peloid-microbial lump-intraclast or micritized ooid grainstones and micropeloidal
packstones. This reservoir has relatively small variations in terms of porosity and permeability; however, each package or rock
type has short-spacing variation in dissolution, pore types and cementation intensity. These variations are present together with
changes in vugs and micro-fractures intensity. Reservoir B, Fig. 2, is even more complex, and has a lower reservoir quality. It
consists of flat-laminated cryptmicrobial stromatolites, thrombolites, cabbage-head stromatolites and ooid-intraclast-molluscan
packstone/grainstones. These varieties of lithofacies increase the complexity of the pore types, cementation and petrophysical
rock properties, such as porosity, permeability and tortuosity.
Petrographic work was performed to understand rock texture and porosity types. Petrophysical rock properties need to be
linked to the geological descriptions, which can indicate the pore types and connectivity. Porosity, permeability and pore types
are related to geological lithofacies; however, this relationship can be distorted by factors like dissolution and cementation.
This distortion can have a positive or negative impact on reservoir quality. Figure 3 shows the distribution of porosities
[SPE-SAS-825] 3
considered in this study as a function of rock types (RT). The oolitic grainstone (RT-1) has mainly interparticle porosity, while
the stromatolitic wackestone (RT-7) has mainly fenestral porosity. This knowledge helps to explain the poor fluid or electrical
connectivity in RT-7.
Various rock type classification techniques are compared (Lucias, Flow Zone Indicator (FZI), Winland/Aguilera r
p35
) in this
study to determine the best classification technique for these carbonate samples. The samples do not show any relationship
between interparticle porosity and permeability with respect to particle size as suggested by Lucias classification. Figure 4
shows different rock types in Lucias plot. The particles, based on petrographic work on thin sections, have approximately the
same size. Yet there is no correlation with that type of rock. This poor relationship can be related to the processes of rework,
dissolution and cementation that these carbonate rocks have experienced. In the flow zone classification method, Fig. 5, the
relationship between
z
and RQI can provide a general distinction between rock types; however, occasionally there is no RQI
(include permeability) difference with changing rock type. Different rock types show a wide range of values of
z
for the same
RQI values in Fig. 5.
Winland r
p35
flow unit classification, Fig. 6, appears to provide the best result. The geological framework and the proportions
of fractures and non-touching porosity (e.g., vuggy and fenestral), which usually affects the conduction of fluids and electrical
current, are captured properly by this classification. The Winland technique calculates pore throat apertures (r
p35
) at 35%
cumulative pore volume and this agrees with what is reported fromthe geological descriptions. Figure 6 shows the different
rock types for the carbonate samples in this study. There is an excellent correlation with lithofacies changes described from
cores. For example, rock type 1 (RT-1) is described as micritised ooid to peloidintraclast grainstones, while the lower quality
rock types RT-3, 4 and 5 are degraded to microbial grainstone, microbial packstone and wackestone as shown in Fig. 7. The
changes in lithofacies are accurately predicted by the Winland r
p35
classification. Sometimes the oolitic grainstones have lower
reservoir quality due to extensive precipitation of cementation. The degraded porosity and permeability was captured by the
Winland technique as well. In total, seven different rock types have been found in the Reservoir A and B carbonates used in
this study. RT-1 and 2 represent megaports (port =pore throat), RT-3 macroports, RT-4 and 5 mesoports, RT-6 microports and
RT-7 is seal rock. These different pore throats and their radii in microns are shown in Fig. 6.
Reservoir A consists of four petrophysical rock types, Fig. 8. RT-2 represents a clean peloidal grainstone and stromatoporoid
floatstone with very high porosity and permability. RT-3 represents microbial peloidal grainstone with lower rock quality due
to the pore-bridging microbial activity. RT-4 and 5 are microbial packstones and wackestones with fair to low rock quality
due to the increase of lime mud. Reservoir B consists of six petrophysical rock types with a more rapid change in rock types
and lower quality rocks than Reservoir A, Fig. 9. Reservoir B has the aforementioned rock types plus RT-1 (a super-K type of
rock) and RT-6, which is a very low quality rock. RT-1 is cement-free grainstone with some fractures. In general, Reservoir B
has more variety and rapid change in rock type than Reservoir A.
Tripl e Porosity Model Results
Different parameters (
2
,
nc
and
f
) in the triple porosity model, Eq. 1, have to be estimated to calculate the cementation
exponent (m). As a preliminary estimate, different porosity proportions are estimated visually fromthin sections and m values
are calculated for every thin section (Al-Ghamdi et al., 2010). Estimation of non-connected porosity from thin sections is
complicated because of the 2D limitation; however, it is important to performthis step to understand the rock quality and the
pore types associated with these rocks.
The procedure for calculating different porosity proportions from logs is shown in Fig. 10. If we assume that sonic porosity
4 [SPE-SAS-825]
accounts for matrix porosity, then subtracting the effective porosity (core measurements or logs) from the sonic porosity will
represent the fracture porosity (
2
), Eq. 2. Subtracting total porosity (
ND
or NMR porosity) fromthe effective porosity equals
the non-connected porosity (
nc
), Eq. 3. Matrix porosity (
b
), which is attached to the matrix volume only, can be calculated
knowing the other porosity proportions (
2
,
nc
and
m
). Knowledge of the pore types, as indicated above, of different
petrophysical rock types helps to validate the accuracy of the estimation fromlogs.
s eff

2
. (2)
eff nc

. (3)
The log-derived cementation exponent (m) was generated using the aforementioned procedure and resulted in m values that
change with depth for Reservoir A and B, Figs. 8 and 9. The averagem for Reservoir A is 2.19 and 2.25 for Reservoir B. To
validate this procedure, 25 samples fromthe two reservoirs were tested in the laboratory to estimatem values for the rocks.
Figure 11 shows that a good comparison was obtained between lab measurements and log estimations. The coefficient of
determination (R
2
=0.757) is reasonable when one takes into consideration the different scales between the logs and core plugs
(1.5 for core plugs and 6 for each point on the log). Subsequently, this scale difference is in favor of the log derived m (triple
porosity model) because water saturation is calculated at the same well log scale.
The continuous variablem estimated from logs was used with the triple porosity model to calculate water saturation as shown
in Figs. 8 and 9. This calculation is more reasonable and robust than water saturation calculated using a constant value of m.
It is well recognized that there can be significant differences in water saturation results fromlogs when the calculations are
carried out with constant and variable values of m using the triple porosity model. This difference increases with larger
contributions of non-connected porosity (
nc
) in the different rock types (Al-Ghamdi et al., 2010). In addition, reliable water
saturation calculations help in the identification of the fluid contacts.
Conclusions
1- The petrophysical model for triple porosity reservoirs discussed in this paper, with the support of detailed
petrographic work, provides a quick and robust method for estimating the cementation exponent (m) in complex
carbonate rocks.
2- Using a continuous log of m values reduces significantly the uncertainty in water saturation calculations. This is
because changes in rock types are accommodated in the calculated m values.
3- The Winland r
p35
rock typing and flow unit technique provided the best results for the complex carbonate rocks
considered in this study.

Acknowledgment
We would like to thank Saudi Aramco for supporting this research and providing the study data. We also acknowledge the
Geoscience, and Chemical and Petroleum Engineering departments at the University of Calgary and our colleagues of the
GFREE research team at the University of Calgary for their help and support. We thank the Canadian Well Logging Society
(CWLS) for giving Ali Al-Ghamdi the CWLS Thesis Abstract Award for 2010/2011. The Thesis Abstract reflects some of
[SPE-SAS-825] 5
Alis doctoral research discussed in SPE papers SAS-825 (presented at this meeting) and 132879 (in print: SPE Reservoir
Evaluation and Engineering Formation Evaluation, 2011). Paradigm provided the software used in some of the
computations. Their contributions are gratefully acknowledged.
Nomencl ature
- Total porosity, fraction

b
- Matrix block porosity attached to the bulk Volume of the matrix system, fraction

eff
- Effective porosity which include matrix and fracture porosities, fraction

m
- Matrix block porosity attached to the bulk volume of the composite system, fraction

nc
- Porosity of non-connected vugs attached to the bulk volume of the composite system, fraction

2
- Porosity of natural fractures attached to bulk volume of the composite system, fraction

z
- Pore volume to grain volume ratio, fraction
RQI - Reservoir Quality Index, micron
r
p35
- Pore-throat aperture corresponding to a mercury saturation of 35%, micron

References
Al-Ghamdi, A., Behmanesh, H., Qanbari, F., Chen, B., and Aguilera, R. 2010. An Improved Triple Porosity Model for
Evaluation of Naturally Fractured Reservoirs. Paper SPE-132879 presented at the Trinidad and Tobago Energy Resources
Conference, Port of Spain, Trinidad, 27-30 June. In print: SPE Res. Eval. & Eng. (2011).
Al-Husseini, M.I. and Matthews, R. K. 2008. Jurassic-Cretaceous Arabian Orbital Stratigraphy: The AROS-J K Chart.
GeoArabia, Vol. 13, No. 1, pp. 89-94.
Aguilera, R. 1976. Analysis of Naturally Fractured Reservoirs from Conventional Well Logs. J. Pet. Technol., Vol. 28, pp.
764772.
Aguilera, R. 2002. Incorporating Capillary Pressure, Pore Throat Aperture Radii, Height Above Free Water Table, and
Winland r35 Values on Pickett Plots. AAPG Bulletin, Vol. 86, No. 4, pp. 605-624.
Aguilera, R.F. and Aguilera, R. 2004. A Triple Porosity Model for Petrophysical Analysis of Naturally Fractured Reservoirs.
Petrophysics, Vol. 45, No. 2, pp. 157-166.
Archie, G.E. 1942. The Electrical Resistivity Logs as an Aid in Determining Some Reservoir Characteristics, Trans. AIME
Vol. 146, pp. 5462.
Amaefule, J.O., Altunbay, M., Tiab, D., Kersey, D.G. and Keelan, D.K. 1993. Enhanced Reservoir Description: Using Core
and Log Data to Identify Hydraulic (Flow) Units and Predict Permeability in Uncored Intervals/Wells. Paper SPE-26436, 68th
SPE Annual Technical Conference and Exhibition, Houston, Texas, October 3-6.
Focke, J.W. and Munn, D. 1987. Cementation Exponents in Middle Eastern Carbonate Reservoirs, SPE Form. Eval. Vol. 2,
pp. 155167.
Lucia, F.J. 1983. Petrophysical Parameters Estimated from Visual Descriptions of Carbonate Rocks; a Field Classification of
Carbonate Pore Space. J. Pet. Technol. Vol. 23, pp. 629637.
6 [SPE-SAS-825]
Rasmus, J.C. 1983. A Variable Cementation Exponent, m, for Fractured Carbonates. The Log Analyst, Vol. 24, No. 6, pp. 13-
23.
Salem, H.S. and Chilingarian, G.V. 1999. The Cementation Factor of Archies Equation for Shaly Sandstone Reservoirs. J.
Pet. Sci. Eng. Vol. 23, No.2, pp. 8393
Towle, G.H. 1962. An Analysis of the Formation Resistivity Factor-Porosity Relationship of Some assumed Pore Geometries.
Trans. SPWLA, 3rd Annual Logging Symp., May 1718, Houston, TX, pp. CC1CC13.














Fig. 1: Water saturati on err or as a functi on of the porosi ty exponent, m. For exampl e at 10% porosit y, i f the
correct val ue of m i s 2.0 and the evaluati on is perf ormed usi ng 1.7, the water saturati on error i s 29.2%.
[SPE-SAS-825] 7







Fi g. 3: Pore types observed from petrographi c work on different rock types in di ffer ent reservoir s.
Fig. 2: Jurassi c to Cretaceous geol ogi c column showi ng Reservoirs A and B
posi ti on. Modifi ed after Al-Husseini and Matt hews, 2008.

8 [SPE-SAS-825]










Fi g. 4: A pl ot of i nter-parti cl e porosity vs. permeabi l it y does not show a di sti ncti ve rel ati onshi p wi th
the parti cl e si ze for t he seven carbonate rock types on Lucia s pl ot. Rock t ypes i ll ustrated i n this pl ot
are based on petrophysi cal work and geol ogi cal descripti on.
Fi g. 5: Thi s fl ow zone i ndi cat or pl ot shows a wi de range of z and RQI val ues for the four rock
types.
[SPE-SAS-825] 9





Fi g. 6: Wi nl and rp35 val ues (mi crons) show di stincti ve trends between porosity and permeabil ity for each
rock type. This rock type cl assif icati on represents a sol i d methodol ogy for use i n petrophysi cal and
engi neeri ng anal ysi s.
Fig. 7: Carbonate rocks have tendency to change petrophysi cal l y withi n very smal l di stances due to
micr obi al and diageneti c processes. A-(RT-2), B-(RT-3), C-(RT-4), C-(RT-5).
10 [SPE-SAS-825]








Fig. 8: Reservoi r A l og results. Track 1 Li thol ogi cal components; Track 2 Depth scal e; Track 3 GR and Cali per ;
Track 4 Raw porosity l ogs; Track 5 Calcul ated porosi ty; Track 6 rp35 petr ophysi cal rock type; Track 7 Variabl e m
curve obtained fr om the tri pl e por osi ty model and lab-derived m; Track 8 Resi sti vit y; Track 9 NMR pore si ze; Track
10 Calcul ated Sw (ol d and new); Track 11 NMR di stri buti on; Track 12 Permeabi li ty (lab measurements and
Facimage predi ct ed).
[SPE-SAS-825] 11








Fig. 9: Reservoir B l og results. Track 1 Li thol ogi cal components; Track 2 Depth scal e; Track 3 GR and Cal iper;
Track 4 Raw porosity l ogs; Track 5 Cal cul ated porosity; Track 6 rp35 petr ophysi cal rock types; Track 7 Vari able m
curve obtai ned from the tri pl e porosit y model and lab-deri ved m; Track 8 Resisti vi ty; Track 9 NMR pore sizes;
Track 10 Cal cul ated Sw (ol d and new); Track 11 NMR di stributi on; Track 12 Permeabi li ty (lab measurements and
Faci mage predi cted).

12 [SPE-SAS-825]








Fig. 11: Good rel ati onshi p between m val ues measured in the lab and l og-deri ved m. The scatter i n the
poi nts i s expected because of the diff erent verti cal measurement scales (Pl ugs repr esent 1.5 and l og
is 6 ).
Fig. 10: Schemati c diagram shows how to calculate diff erent porti ons of the total porosi ty.