SPE 166178

Geological Attributes from Conventional Well Logs: Relating Rock Types to Depositional Facies in Deepwater Turbidite Reservoirs Chicheng Xu, SPE, The University of Texas at Austin; Carlos Torres-Verdín, SPE, The University of Texas at Austin; Ronald J. Steel, The University of Texas at Austin

Copyright 2013, Society of Petroleum Engineers This paper was prepared for presentation at the SPE Annual Technical Conference and Exhibition held in New Orleans, Louisiana, USA, 30 September–2 October 2013. This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract Grain size and bed thickness are important geological attributes for sedimentary facies analysis and reservoir quality ranking in deepwater turbidite reservoirs. Grain size controls reservoir quality and has causative or correlational effects on most well logs. Additionally, bed thickness affects well logs in various ways because of different vertical logging-tool resolutions. The objective of this paper is to quantitatively classify rock and bed types based on conventional well logs to assist facies interpretation and stratigraphic reservoir modeling. We model physical properties and well-log responses originating from clastic reservoirs with different grain-size distributions in the context of deepwater turbidite systems. The model accounts for fluid effects introduced by capillary transitions and salty connate water. Petrophysical relationships are examined between grain sizes and pore geometry as inferred from the combined effects of initial connate-water saturation and reservoir capillary pressure. From modeling, we derive several quantitative petrophysical attributes which correlate significantly with grain size. A list of such attributes includes volumetric concentration of shale, total porosity, neutron-density porosity difference, and reservoir quality index (RQI). Rock classification is then performed based on a relevant set of attributes to detect and rank different rock types with specific grain-size distributions. Concomitantly, a quantitative rock classification scheme using three-dimensional Thomas-Stieber diagrams constructed with gamma ray, bulk density, and resistivity logs is used to classify petrophysical zones based on their average bed thicknesses. A dual-attribute rock classification scheme is proposed to better associate rock types with depositional facies. The above-mentioned rock-classification and bed-typing methods are synthesized and applied to a turbidite field from the Gulf of Mexico. We perform facies analysis based on the vertical succession of rock types (grain sizes) and the inferred variations of bed thickness. The field case demonstrates that these two geologic attributes have great potential in reducing uncertainty and non-uniqueness in the construction of stratigraphic reservoir models. Results indicate that the classified rock types and their petrophysical ranking agree with core data while the interpreted facies fit into the depositional context as inferred from core descriptions and seismic horizons. Introduction Rock classification (or simply rock typing) is an emerging reservoir description tool that elicits broad research interests in the oil and gas industry. It has become increasingly important in modern reservoir description to invoke close integration of multi-discipline, multi-physics, and multi-scale subsurface data including pore imaging, core measurements, well logs, seismic amplitude data, well testing, and production surveillance (Gunter et al., 1997a). Interestingly, the definition of rock type still remains ambiguous and sometimes arguable in many situations as it depends highly on the specific description purposes to be achieved by geoscientists and engineers. Geologists consider rock types as depositional facies or lithofacies which emphasize the genesis of rock formations to enable 3D stratigraphic reservoir modeling (Fisher, 1982; Kerans and Tinker, 1999; Muto and Steel, 2007; Slatt, 2007). Petrophysicists define rock types based on pore geometry that relates all static and dynamic petrophysical properties (Archie, 1950 and 1952). Reservoir and production engineers group rock types as

2 SPE 166178

flow units that are stratigraphically continuous intervals of similar geologic and petrophysical features to upscale reservoir grids for efficient fluid-flow simulation (Gunter et al., 1997b). Noteworthy is that reservoir characterization teams are working toward a common objective: to construct a verifiable reservoir model populated with accurate petrophysical properties for reserves estimation and production forecasting (Lucia, 1999). Therefore, developing new rock classification schemes and workflows that serve multiple characterization purposes from different disciplines remains a challenging but important task. For subsurface studies, whole core samples and well logs are the main sources of high-resolution geological data. Routine geological work infers grain size and bed thickness from slabbed core samples. However, in most cases whole core spans a very limited segment of the reservoir. Log-core integration is therefore necessary to extrapolate geological properties from cored depth intervals to remaining wells and into the reservoir. Rock classification based on core data is relatively simple because petrophysical properties are rigorously measured under the same protocol. However, rock classification using downhole well logs is more challenging and is subject to great uncertainty. Four technical issues immediately arise when moving from core data to well logs for rock typing, namely data quality, indirect measurements, variable reservoir conditions, and scale discrepancy (Xu, 2013). These compounded technical problems were initially defined by Archie as the “geometrical problem” (1950). In the present paper, we aim to address the issue of indirectness. Indirectness lies in using physical measurements to estimate petrophysical properties and interpret geological attributes. In fact, most well logs are simultaneously sensitive to grains, texture, structure, and fluids. Table 1.1 summarizes the sensitivity of various well logs to rock petrophysical properties and geological attributes. It is found that none of the logs are directly sensitive to grain size and only a few logs have good sensitivity to bedding structure. Grain size information can only be indirectly inferred from the pore geometry formed in-between grains. Moreover, well logs highly sensitive to pore geometry and bedding structure are also significantly influenced by fluid content, as in the case of resistivity logs. Therefore, fluid effects on well logs need to be cautiously accounted for before inferring grain and bed properties.

Table 1― Sensitivity of well logs to petrophysical properties and geological attributes (modified from Serra and Abbot, 1980).

Composition Texture Structure Fluid

Gamma Ray High Low Low Low

PE Factor High Low Low Low

Neutron Porosity Medium Medium Low High

Bulk Density Medium Medium Low High

Electrical Resistivity Low Medium High High

Acoustic Slowness High Medium Medium High

Magnetic Resonance Medium High Low High

The ultimate goal of rock classification is to associate petrophysical rock types with geological framework and reconcile them with seismic facies to propagate rock-type associated petrophysical properties into the reservoir model. This endeavor is difficult without understanding the cause-and-effect relationship between pore geometry and rock geological attributes such as grain size and bed thickness. Hence, it is necessary to investigate the sensitivity of well logs to those geological attributes and the associated pore geometries in the presence of different fluid-saturation conditions resulting from either hydrocarbon migration or mud-filtrate invasion. In this paper, we show the causative and correlational impacts of grain size on various types of logs. In addition, the relationships between gamma ray log and other logs are interpreted to reveal the underlying bed thickness information as emphasized in Thomas-Stieber diagrams. We propose an integrated petrophysical workflow and a dual-attribute rock classification scheme to better relate rock types to geological facies in deepwater turbidite reservoirs. A field case from the central Gulf of Mexico is used to test the workflow and classification schemes.

SPE 166178 3

Grain-Size Analysis from Well Logs Grain size is directly associated with depositional energy. Large grains are deposited under highly energetic flow conditions, whereas small grains are deposited under low flow energy (Reading, 1996). Grain size information provides key data for geologists in their facies interpretation work (Glaister and Nelson, 1974). In addition, grain-size distribution determines the pore-size distribution of most clastic rocks when diagenesis effects are not significant. Consequently, grain size information becomes a good indicator of reservoir quality in terms of hydraulic capacity. Grain size is typically measured from core samples or outcrops with a laser particle size analyzer (LPSA). To date, there exists no effective physical measurement capable of measuring grain size directly at downhole conditions. However, there are some causative and correlational effects between grain size and physical logs. In the following section, we briefly review the relationships between grain size and gamma ray, neutron, density, resistivity, and NMR logs. Gamma Ray Log. Gamma ray logs are sensitive to the natural radioactivity of rock formations. The radioactive elements (U, K, Th) are predominantly associated with small grains such as silt or clay (Serra and Serra, 2003). Consequently, often there is a correlation between gamma ray log and grain-size distribution. In fact, the gamma ray log has been long used as a single indicator of grain size in many geological field studies because the gamma ray log is more readily available than other logs (Pirson, 1983; Rider, 1990). In clastic reservoirs, the gamma ray log is sufficient to separate clean sands, shaly sands, and shales (or siltstones). Figure 4 Track 2 shows the typical gamma ray log responses of clean sands, shaly sands, and shales (or siltstones). However, the gamma ray log alone cannot effectively distinguish clean sands of different grain sizes. Neutron Porosity and Bulk Density Log. Both neutron porosity and bulk density logs are dominantly sensitive to total pore volume and pore-filing fluids in clean sands. In many clastic reservoirs, the porosity formed by grain deposition has small variability. However, the total pore volume will be significantly changed by cementation which is mainly comprised of silts and clays (Neasham, 1977). As a result, low porosity rock tends to be associated with silt- or clay-size grains. In addition, the presence of silts or clays will increase the bound water volume that remains in the rock pore system after hydrocarbon migration. The hydroxyl in clay minerals gives rise to a high hydrogen index which increases the apparent neutron porosity. On the other hand, clay or silt cementation increases the bulk density of rocks when solid cement replaces the original fluid in the pore space. As a result, neutron and density porosity logs separate when they are plotted on the same track using the correct matrix. The separation or difference between neutron and density porosity is an indicator of clay volume. Figure 4, Track 3 shows the typical neutron porosity and bulk density logs across clean sands, shaly sands, and shales. Neutron-density separation is smallest in clean sands and largest in shales. However, neutron porosity and bulk density logs still lack the ability to differentiate sandstones of different grain sizes but with the same porosity. Resistivity and NMR Logs. We use a simple pore network model to study the cause-effect chain connecting a series of rock properties including grain size, pore size, capillary pressure, water saturation, electrical conductivity, and NMR transverse relaxation time T2. Figure 1 shows two thin sections of pore network systems formed by large (Type A) and small (Type B) grain sizes, respectively. Large grain sizes generate large pore sizes accordingly, resulting in a T2 distribution with large T2 values (Fig. 2). When these two rocks are subjected to the same capillary pressure induced by hydrocarbon migration at reservoir conditions, rock type A exhibits both lower capillary-bound water saturation and clay hydration water saturation (Hill et al., 1979), and consequently lower electrical conductivity (Fig. 3). Therefore, in a hydrocarbon-bearing zone, resistivity logs become sensitive to pore size, i.e., petrophysical rock types. Rock types with higher hydraulic conductivity tend to have lower electrical conductivity (or higher resistivity) in hydrocarbon-bearing zones in the presence of capillary pressure. Figure 4 Track 4 describes the typical resistivity log responses observed across clean sands, shaly sands, and shales in a hydrocarbon zone.

Figure 1― Thin sections describing pore network systems formed by different grain sizes. Left Panel: large grain size; Right Panel: small grain size (Xu et al., 2013).

1m

m

1m

m

4 SPE 166178

Figure 2― Modeling of pore-size (left panel) and NMR T2-distributions (right panel) for two different rock types (Xu et al., 2013).

Figure 3― Modeling of drainage capillary pressure and resistivity-height relation for two different rock types. Pc: capillary pressure (left panel); HAFWL: height above the free water level (FWL) (right panel); Rt: bulk rock resistivity (Xu et al., 2013).

Figure 4―Simulated well logs across different rock types. Track 1: Depth; Track 2: Gamma ray log; Track 3: Neutron porosity (sandstone porosity units) and bulk density logs; Track 4: Resistivity log; Track 5: Rock types. Logs are simulated with software UTAPWeLS1.

1 The University of Texas at Austin’s Petrophysics and Well Log Simulator

SPE 166178 5

Correction of Free Water in Capillary Transition Zones. The presence of free water in a capillary transition zone disrupts the correlation between electrical conductivity and hydraulic conductivity. Xu and Torres-Verdín (2012) proposed a correction method by linking Leverett’s RQI with in-situ capillary pressure (Pc) and initial connate-water saturation (Sw) using the empirical J-function model (Darling, 2005; Torres-Verdín, 2012):

1

( ) ( )b w wirr bw wirr w

S SS S aJ J Sa−

= + ⇒ = , (1)

together with Leverett’s capillary pressure model (1941), given by: 1( ) cosRQI cos ( )w w wirr b

c c

J S S SkP a P

σ θσ θφ

−= = × = × , (2)

where Sw is initial water saturation, Swirr is irreducible water saturation (which is set to 0.01 lower than the minimum water saturation in the entire reservoir column, Darling, 2005); a and b are constants derived from core-measured capillary pressure curves; cosσ θ is the product of interfacial tension and contact angle, which is assumed constant in the same reservoir; Pc is in-situ reservoir capillary pressure in psi. Shale Distribution and Bed Typing from Well Logs Bedding structure together with bed thickness is often a key parameter for geologists to interpret sedimentary facies (Campbell, 1967). Thinly bedded or laminated formations are persistent only in some special depositional facies (Passey et al., 2006). High-resolution image logs provide the best estimation of bedding structures but are not commonly available in all wells. Thomas-Stieber’s diagram (Thomas and Stieber, 1975 & 1977) is a widely applied petrophysical interpretation tool used to diagnose shale distribution in siliciclastic reservoirs as being laminated, dispersed, or structural (Fig.5; Torres-Verdín, 2012). Thomas-Stieber diagram assumes two end members: clean-sand and pure-shale. Several trends are also defined to reveal different geological processes yielding different shale distributions which are associated with different mixing laws reflected on logging measurement physics and the underlying petrophysical properties (Xu and Torres-Verdín, 2013). As a consequence, rocks with different values of shale volumetric concentration and distribution appear in different locations when projected onto a Thomas-Stieber diagram (Fig. 6). Xu and Torres-Verdín (2013) proposed a graphical clustering method to classify petrophysical zones of different bed thickness in a 3D Thomas-Stieber diagram (Fig. 7).

(a) (b) (c)

Figure 5― Three nominal forms of shale distributions in siliciclastic reservoirs: (a) laminated shale, (b) dispersed shale, and (c) structural shale.

6 SPE 166178

Figure 6― Thomas-Stieber diagram constructed with bulk density and gamma ray logs. Red circle dots: pivoting points; Blue dashed lines: trend lines; shaded elliptical regions: uncertainty bounds.

Figure 7― Rock classification and bed typing via k-means cluster analysis (Press et al., 2007) on a 3D Thomas-Stieber diagram constructed with gamma ray, bulk density, and resistivity logs.

SPE 166178 7

Figure 8― Integrated dual-attribute rock classification workflow based on well logs and core data. Dual-Attribute Rock Classification Workflow Rock types defined by grain size and bed thickness should be integrated to achieve a better reservoir description in multi-well field studies,. Figure 8 shows an integrated dual-attribute rock classification workflow based on standard petrophysical log-analysis and core-calibration. Grain size and bed thickness information are propagated from cored intervals to log-covered intervals via step-by-step calibration and validation. Therefore, the final rock types derived from well logs in multiple wells are geologically consistent for facies interpretation and stratigraphic reservoir modeling. Furthermore, rock type information needs to be returned to standard petrophysical analysis for more accurate rock-type-specific processing. Field Case: Marco Polo Field, Gulf of Mexico The Marco Polo field is a Miocene turbidite oil field located in the central Gulf of Mexico (Fig. 9; Contreras et al., 2006). Marco Polo minibasin resides in a tectonically active salt environment (Fig. 10; Jackson et al, 2008). The depositional system is interpreted as a submarine fan complex developed in a mini-basin with stacked progradational turbidite lobes. Reservoirs consist primarily of sandy turbidite facies interbedded with some siltstone and muddy debris-flow facies. Reservoir rocks are mainly unconsolidated sandstones of very fine to fine- to medium-size grains. Thin-bedded zones are common across the entire reservoir. In this study, we evaluate only the two hydraulically connected, hydrocarbon-bearing sandstone units: M40 and M50 in three wells. These sands are buried at depths between 12,000 and 13,000 ft TVD and dip toward the west (Fig. 11). Core measurements indicate that the structureless and plane-parallel laminated sandstones exhibit porosities as high as 35%, as well as 100 – 2000 md of nominal permeability. Well logs from three wells are used for rock typing. Well no. 1 was drilled with oil-base mud and penetrated a capillary transition zone and the FWL. Wells nos. 2 and 3 were drilled with water-base mud where the M40-50 sands are located more than 500 ft above the FWL (Fig. 11).

8 SPE 166178

Figure 9― Geographic location of the Marco Polo field in the Gulf of Mexico (Contreras, 2006).

Figure 10― Salt tectonics background of the Marco Polo minibasin (Jackson et al., 2008).

Figure 11― Seismic cross-section with the three well trajectories considered in this paper.

SPE 166178 9

Core-Based Rock Typing and Grain Size/Bed Thickness Interpretation

Petrophysical rock types are first established using routine core porosity-permeability measurements. Six petrophysical rock types were identified from the histogram distribution of Leverett’s RQI (Fig. 12). Table 2 summarizes the statistical distributions of petrophysical properties for each rock type. We observe that rock types 1 - 3 exhibit a largely overlapping porosity range. Their average horizontal permeability exhibits a ratio around 2 – 2.5, which indicates that their average pore-throat radius ratio is close to 1.5. Laser grain size measurements for rock types confirmed that the controlling factor for permeability was mean grain size. Reservoir quality decreases with decreasing grain size. Grain size measurements for rock types 2, 4, and 5 are listed in Table 3 for comparison. Different bed types identified from core photographs are compared in Table 4 for facies interpretation.

Figure 12― Core-based hydraulic rock typing via Leverett’s RQI.

Table 2― Statistics of porosity, permeability, and RQI for each hydraulic rock type.

Rock Type

Porosity (p.u.) Permeability (md) RQI

(μm)

RRT1 33.6 ± 1.8 1099 ± 291 56.8 ± 7.2

RRT2 33.3 ± 2.1 549 ± 130 40.4 ± 4.9

RRT3 31.5 ± 4.4 179 ± 63.7 23.4 ± 3.4

RRT4 24.1 ± 6.1 42.9 ± 19.1 13.1 ± 2.1

RRT5 19.7 ± 3.9 11.3 ± 3.9 7.5 ± 1.1

RRT6 19.6 ± 3.7 4 ± 1.5 4.4 ± 0.86

10 SPE 166178

Table 3― Comparison of grain-size distribution data for rock types 2, 4, and 5.

Rock Type Grain Size Distribution Description Facies Interpretation

2

Dominated by sand-size grains with a narrow distribution in the large-size mode (i.e., good sorting).

Turbidite channel infill including channel lag.

4

Dominated by silt-size grains with only a small proportion of sand-size grains. Overall distribution is widely spread (i.e., relatively poor sorting).

Channel levee or channel margin with thin, fine-grained beds.

5

Dominated by silt-size grains with an overall wide distribution (i.e., poor sorting).

Overbank deposits possibly including thin debris flows.

Table 4: Comparison of core photographs of different bed types.

Rock Type Core Photos Description Facies Interpretation

Thick sand

Structureless coarse-grained sandstone with no clear internal lamination.

Amalgamated turbidite channel deposits.

Laminated sand-shale

Interbedded sand-shale sequence with variable net-to-gross ratio.

Channel levee or margin deposits.

Thick shale or siltstone

Fine-grained shale or siltstone.

Overbank deposits, possibly with debris flow.

Sand Silt

Sand

Silt

Sand Silt

SPE 166178 11

Log-Based Rock Classification and Bed Typing Core-calibrated, conventional well-log analysis is first performed to provide relevant petrophysical properties, such as volumetric concentration of shale, total porosity, and connate-water saturation. We classify rock types via k-means cluster analysis (Press et al., 2007) on several relevant petrophysical attributes including shale volume, total porosity, neutron-density separation, water saturation in the irreducible zone, and well-log-derived RQI as in Eq. (2) in the capillary transition zone. The classified rock types are consistent with core data concerning both petrophysical properties and grain size. Figure 13 shows an interval with log-based rock classification results. A clear fining-upward grain-size trend is observed in both core and the log-derived rock types in the interval of 13,574 – 13,580 ft. The high-resolution grain-size information provides geologists useful data for facies interpretation.

Figure 13― Example of a fining-upward unit identified from vertical stacking of petrophysical rock types. Track 1: Depth; Track 2: Gamma ray log; Track 3: Neutron (in apparent sandstone porosity units) porosity and bulk density logs; Track 4: Induction resistivity logs; Track 5: Rock types of different grain-size distributions; Track 6: Core permeability.

Relating Rock Types to Geological Facies in a Single Well

In a single well, we analyze the stacking pattern or trend of the classified rock types that bear grain-size and bed-thickness information. Similar neighboring rock types are grouped as distinct facies from the adjacent rock type groups. Figures 14 and 15 show two intervals with their sedimentary facies interpretations. Amalgamated channel facies are composed mainly of thick and coarse-grained sandstones. Both bed thickness and grain size decrease in the levee or channel margin facies. Overbank deposits are mainly shales or siltstones. The interpretation is consistent with observations contained in Table 3 and Table 4.

12 SPE 166178

Figure 14― Rock typing and facies interpretation in the key well of the Gulf-of-Mexico case. Tracks 1-4: basic well logs; Tracks 5-6: standard well-log analysis, saturation-height for rock types 1 to 3; Tracks 7: distribution of rock types; Track 8: interpreted sedimentary facies.

Figure 15― Rock typing and facies interpretation in the key well of the Gulf-of-Mexico case. Tracks 1-4: basic well logs; Tracks 5-6: standard well-log analysis, saturation-height for rock types 1 to 3; Tracks 7: distribution of rock types; Track 8: interpreted sedimentary facies.

8

8

SPE 166178 13

Stratigraphic Correlation with Multi-Well Logs. After rock classification, bed typing, and facies interpretations are performed in individual wells, a multi-well cross-sectional view can be constructed for stratigraphic analysis. Figure 16 shows the corresponding results from the three well penetrations. The locations of these wells in a submarine fan complex with stacked progradational lobes are consistent with the interpretation made from seismic horizons. Stratigraphic analysis together with the estimated petrophysical properties can then be integrated with seismic amplitude data for populating a reservoir model.

Figure 16― Use of rock-type and bed-type distributions in multiple wells to assist sedimentological and stratigraphic interpretation. The upper panel shows the Thomas-Stieber diagram constructed with bulk density and gamma ray logs in each well below. The lower panel shows the log tracks in each well with the following information: Track 1: Depth; Track 2: Gamma ray log; Track 3: Neutron (in apparent sandstone porosity units) porosity and bulk density logs; Track 4: Induction resistivity logs; Track 5: Bed type; Track 6: Rock type of different grain-size distribution.

Conclusions Grain size and bed thickness are two geological attributes that can be interpreted from conventional logs in turbidite reservoirs. The petrophysical quality of young deepwater turbidite reservoirs is controlled chiefly by sedimentary grain sizes. Therefore, a link exists between petrophysical rock types and depositional facies which can be effectively applied in reservoir development efforts. Thomas-Stieber diagrams are useful in identifying thinly bedded or laminated zones which provide additional aid for facies interpretation. We propose an integrated petrophysical workflow that combines grain-size and bed-thickness attributes to reduce uncertainty in stratigraphic reservoir modeling. Reservoir capillary transition effects on resistivity logs need to be corrected for more accurate estimation of grain sizes. In the deepwater Gulf-of-Mexico field case, we confirmed that multi-well rock typing can assist sedimentary and stratigraphic studies which have potential applications for exploration in frontier basins or in deepwater environments, where subsalt or pre-salt reservoirs cannot be clearly imaged with seismic data and well drilling is limited due to high rig costs.

30 40 50 60 70 80 90 100

1.8

2

2.2

2.4

2.6

2 2288

6

2555

5 225 5 2555 555566

5 555 5

5522 222566

6664 1 665

50

50 77

5

00

500 53 7 7

7737 7 755 7

3

755000 0 55 5 7

556

66 6

6 559

75

5 55 5

2 2 222 222 2222 222225 2 2225

22 222

2 222 2222

5

54

62 2 2222 222 2 222

266 6

57

5 999

77 9

7777 7

7

7

9 99

777 77 9

777

775 7

7 777

77

77

77 9 99 9999

9999

97

3 3

7

77

300

0 0 7 7 7

333

000000 0000

0 00 0

000 700 3 7

5 77775

73 3 300 33500

05

5

6

550

0 0 000 5555 77

700 330

7

5 97

75

9

9

77

5

7

777

50 5 5 7 99

77

9 92

5

77

735 5000 000

55 5 55

55 5555

566 55 2 222 2 2222 222 2222222 2 2222

2 22 2 222 22

GR (GAPI)

RH

OB

(G/C

M3)

30 40 50 60 70 80 90 100

1.8

2

2.2

2.4

2.6

2.8

222 2

226 2 2255 25 5255 2 22 225 2 2 22666 666

7

773330000073 70030

0 7 777333

33 7

757

337 3300000 0

00 3 333 370 777

05

770 7

7

56 222

555 595

755 5

56

2 2222 2 2 2 2222 22222 2222 2222

22 2222 2222

2 2222228

5

55777 7

7 9 999 99

77 9 99777

77 9

999977 99

7 79

97977 7

7 97 7

779

999 99 99

99999 99

797

7775775 7

73 7 777333300 0050 50 5055 557

50 77

7703 375

5

6 5 222

2 2222 2 2 2222 222 22222

22 22 22 222 2

GR (GAPI)

RH

OB

(G/C

M3)

30 40 50 60 70 80 90 100

1.8

2

2.2

2.4

2.6

2.8

2228 222 222222 22 2 222 222 226

282255 2 22 2222 22 2 222

222 22 2

22

95

77000 00000 5500000000 5 7700 30 7

00 05 5

7705 5

550 50 0 5 5550 0 550 5 5555 5 9

25

2222 2 2

2255

29

9

6

6 6 6 5 2 222 2222

555 5

665 9995 5

5 5555 5565

97

6 5 55559555 555

997

55

555

99

6

1 6 52

997 7

755 55555

006 5

97775 5 977

55

775555 555000 550 55555550 5 7

555 7999

7 9977 7 999599

9 95995

55 22 2

22 22 2

-12 22222

2 22222 2 22 2

-1

-1

9

7

GR (GAPI)

RH

OB

(G/C

M3)

14 SPE 166178

ACKNOWLEDGMENTS We would like to thank Anadarko Petroleum Cooperation for providing the field data used for analysis and verification. The work reported in this paper was funded by The University of Texas at Austin’s Research Consortium on Formation Evaluation, jointly sponsored by Afren, Anadarko, Apache, Aramco, Baker-Hughes, BG, BHP Billiton, BP, Chevron, China Oilfield Services, LTD., ConocoPhillips, ENI, ExxonMobil, Halliburton, Hess, Maersk, Marathon Oil Corporation, Mexican Institute for Petroleum, Nexen, ONGC, OXY, Petrobras, PTTEP, Repsol, RWE, Schlumberger, Shell, Statoil, Total, Weatherford, Wintershall and Woodside Petroleum Limited. Nomenclature a : Linear coefficient in empirical J function, [] b : Exponent in empirical J function, [] J : Leverett’s J function, [] k : Permeability, [mD] p : Density function of pore-throat size, [] Pc : Capillary pressure, [psi] R : Pore-throat radius, [μm] Rt

: Bulk rock resistivity, [ohm.m] Rw : Connate water resistivity, [ohm.m] Sw : Water saturation, [frac] Swirr : Irreducible water saturation, [frac] T2 : NMR Transverse relaxation time, [ms] φ : Total porosity, [frac] σ : Interfacial tension, [dyne/cm] θ : Contact angle, [degree] List of Acronyms and Log Curve Mnemonics AHT1-9 : Array Induction Log (H) with Two Feet Resolution and 10-90 inch Depth of Investigation BS : Bit Size Log DRHO : Density Correction Log FWL : Free Water Level GR : Gamma Ray Log HAFWL : Height Above Free Water Level HCAL : Caliper Log HGR : Gamma Ray Log (High Resolution) K : Potassium K_Klink : Permeability Corrected with Klinkenberg Effect LPSA : Laser Particle Size Analyzer MICP : Mercury Injection Capillary Pressure NMR : Nuclear Magnetic Resonance NPHI : Neutron Porosity Log Pcow : Capillary Pressure between Oil and Water PE : Photo-Electric PHI_LOG : Porosity from Log Por : Core Porosity RHOB : Bulk Density Log RHOZ : Bulk Density Log (High Resolution) RQI : Reservoir Quality Index RRT : Reservoir Rock Type RT : Rock Type SH : Shale SS : Sandstone Sw_c : Core Water Saturation Sw_RT : Water Saturation for Rock Type Swt_LOG : Water Saturation from Logs T2LM : Algorithmic Mean of NMR Transverse relaxation time T2 Th : Thorium TVD : Total Vertical Depth U : Uranium UTAPWeLS : The University of Texas at Austin’s Petrophysics and Well Log Simulator

SPE 166178 15

References Archie, G.E. 1950. Introduction to Petrophysics of Reservoir Rocks. AAPG Bulletin, 34 (5): 943 - 961. Archie, G.E. 1952. Classification of Carbonate Reservoir Rocks and Petrophysical Considerations. AAPG Bulletin, 36(2): 278-298. Campbell, C. V. 1967. Lamiae, Laminaset, Bed and Bedset. Sedimentology, 8(1967): 7-26. Contreras A. 2006. Spatial Delineation, Fluid-Lithology Characterization, and Petrophysical Modeling of Deepwater Gulf of Mexico

Reservoirs through Joint AVA Deterministic and Stochastic Inversion of 3D Partially-Stacked Seismic Amplitude Data and Well Logs. Ph. D. Dissertation, The University of Texas at Austin.

Contreras A., Torres-Verdín, C., Chesters, W., Kvien, K., and Fasnacht T. 2006. Extrapolation of Flow Units away from Wells with 3D Pre-Stack Seismic Amplitude Data: Field Example. Petrophysics, 47(3): 223-238.

Darling, T. 2005. Well Logging and Formation Evaluation. Gulf Professional Publishing. Fisher, W. L. 1982. Depositional Systems and Environments, in Encyclopedia of Science and Technology. New York, McGraw-Hill, 5th

ed., 104-108. Glaister, R.P., and Nelson, H.W. 1974. Grains-Size Distributions, an Aid in Facies Identification. Bulletin of Canadian Petroleum Geology,

22(3):203-240. Gunter, G.W., Pinch, J.J., Finneran, J.M., and Bryant, W.T. 1997a. Overview of an Integrated Process Model to Develop Petrophysical

Based Reservoir Descriptions. Paper SPE 38748 presented at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas, 5-8 October.

Gunter, G.W., Finneran, J.M., Hartmann, D.J., and Miller, J.D. 1997b. Early Determination of Reservoir Flow Units Using an Integrated Petrophysical Method. Paper SPE 38679 presented at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas, 5-8 October.

Hill, H.J., Shirley, O.J., and Klein, G.E. 1979. Bound Water in Shaly Sands – Its Relation to Qv and Other Formation parameters. The Log Analyst, May-June, 3 -19.

Jackson, M.P.A., Hudec, M.R., and Dooley, T.P. 2010. Some Emerging Concepts in Salt Tectonics in the Deepwater Gulf of Mexico: Intrusive Plumes, Canopy-Margin Thrusts, Minibasin Triggers and Allochthonous Fragments. Vining, B.A. and Pickering, S.C. (eds) Petroleum Geology: From Mature Basins to New Frontiers – Proceedings of the 7th Petroleum Geology Conference, 899–912.

Kerans, C., and Tinker, S. 1999. Sequence Stratigraphy and Characterization of Carbonate Reservoirs. SEPM short course no. 40. Leverett, M.C. 1941. Capillary Behavior in Porous Solids. Transactions of the AIME, 142(1): 159–172. Lucia, F.J. 1999. Carbonate reservoir characterization. Springer-Verlag, Berlin. Muto, T., Steel, R., and Swenson, J. 2007. Autostratigraphy: a Framework norm for Genetic Stratigraphy. Journal of Sedimentary

Research, 77: 2-12. Neasham, J.W. 1977. The Morphology of Dispersed Clay in Sandstone Reservoirs and Its Effect on Sandstone Shaliness, Pore Space, and

Fluid Flow Properties. Paper SPE-6858 presented at the SPE Annual Technical Conference and Exhibition, Denver, Colorado, 9-12 October.

Passey, Q.R., Dahlberg, K.E., Sullivan, K., Yin, H., Brackett, B., Xiao, Y., and Guzman-Garcia, A.G. 2006. Petrophysical Evaluation of Hydrocarbon Pore-Thickness in Thinly Bedded Clastic Reservoirs. AAPG Publications, Archie Series, No.1.

Pirson J.S. 1983. Geologic Well Log Analysis. Third Edition, Gulf Professional Publishing. Press, W.H., Teukolsky, S.A., Vetterling, W.T., and Flannery, B.P. 2007. Section 16.1: Gaussian Mixture Models and k-Means Clustering,

Numerical Recipes. The Art of Scientific Computing, 3rd Ed. New York: Cambridge University Press. Reading, H.G. 1996. Sedimentary Environments: Processes, Facies and Stratigraphy. 3rd Edition, Blackwell Science. Rider, M.H. 1990. Gamma-Ray Log Shape Used as a Facies Indicator: Critical Analysis of an Oversimplified Methodology. Geological

Society, London, Special publications, 48: (27-37). Serra, O., and Abbott, H. T. 1980. The Contribution of Logging Data to Sedimentology and Stratigraphy. Paper SPE 9270 presented at 55th

SPE Annual Fall Technical Conference and Exhibition, Dallas, Texas, 21-24 September. Serra, O., and Serra, L. 2003. Well Logging and Geology. Technip. Slatt, R.M. 2007. Stratigraphic Reservoir Characterization for Petroleum Geologists, Geophysicists, and Engineers. Elsevier Publication. Thomas, E. C., and Stieber, S. J. 1975. The Distribution of Shale in Sandstones and Its Effect Upon Porosity. Paper T presented at the 16th

Annual Logging Symposium of the SPWLA, New Orleans, Louisiana, 4-7 June. Thomas, E. C., and Stieber, S. J. 1977. Log Derived Shale Distributions in Sandstone and Its Effect upon Porosity, Water Saturation, and

Permeability. 6th Formation Evaluation Symposium of the Canadian Well Logging Society, Calgary, Alberta, 24-26 October. Torres-Verdín, C. 2012. Integrated Geological-Petrophysical Interpretation of Well Logs. Lecture Notes for Fundamental Well Logging

Course in the University of Texas at Austin, Unpublished. Xu, C. 2013. Reservoir Description by Well-Log-Based and Core-Calibrated Petrophysical Rock Classification. Ph. D. Dissertation, The

University of Texas at Austin. Xu, C., and Torres-Verdín, C. 2012. Saturation-Height and Invasion Consistent Hydraulic Rock Typing Using Multi-Well Conventional

Logs. Paper KK presented at SPWLA 53rd Annual Logging Symposium, Cartagena, Colombia, 16-20 June. Xu, C., Torres-Verdín, C., and Gao, S. 2013. Electrical versus Hydraulic Rock Types in Clastic Reservoirs: Pore-Scale Modeling Verified

with Field Observations in the Gulf of Mexico, U.S. 83rd SEG Annual Meeting, Houston, 22 – 27 September. Xu, C., and Torres-Verdín, C. 2013. Quantitative Rock Type Codification and Interpretation Using Log-Attributes Derived from Thomas-

Stieber Diagrams in Siliciclastic Reservoirs. SEG Interpretation Journal (submitted for review).