SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

SATURATION-HEIGHT AND INVASION CONSISTENT HYDRAULIC

ROCK TYPING USING MULTI-WELL CONVENTIONAL LOGS
Chicheng Xu and Carlos Torres-Verdn
The University of Texas at Austin

Copyright 2012, held jointly by the Society of Petrophysicists and Well Log the process of mud-filtrate invasion, pseudo dynamic
Analysts (SPWLA) and the submitting authors.
rd
This paper was prepared for presentation at the SPWLA 53 Annual logging
rock-fluid petrophysical properties (capillary pressure
Symposium held in Cartagena, Colombia, June 16-20, 2012. and relative permeability) are calculated for each rock
type that are amenable to reservoir modeling and
ABSTRACT simulation. Synthetic cases constructed from real
reservoir analogues are used to test the method with pre-
Hydraulic rock typing is based on pore geometry, which defined rock-type models. Two field cases of turbidite
relates to saturation-height modeling at a later stage in (submarine fan) siliciclastic reservoirs are used to verify
reservoir characterization. Additionally, pore geometry the reliability of the new rock-typing method. In both
affects mud-filtrate invasion under over-balanced cases, the agreement between log- and core-derived rock
drilling conditions. Reliable hydraulic rock typing types is improved by 20% or more with the new method.
should simultaneously honor saturation behavior in the One limitation of the new rock-typing method is that it
vertical direction and mud-filtrate invasion in the radial is only applicable to reservoir zones above oil-water or
direction. Such a condition becomes critical when gas-water contacts. Furthermore, the resolution of rock
hydraulic rock typing is performed with well logs types is limited by the vertical resolution of the deep
acquired in multiple wells penetrating the same or resistivity log.
different capillary transition zones.
INTRODUCTION
This paper considers three conventional core-based rock
typing methods, namely Leveretts k / , Winland R35, G.E. Archie (1950) defined petrophysical rock type
and Amaefules flow zone index, to appraise whether based on the associated pore-size distribution, which
rock classifications can be extrapolated from core-data acts as the hub linking the rocks static and dynamic
to well-log domains. A new quantitative log attribute is petrophysical properties (Fig. 1). In the same paper,
derived from well logs to assist hydraulic rock typing, Archie mentioned the possibility of using various types
which integrates in-situ reservoir capillary pressure (Pc) of well logs to probe the pore-size distribution and
and initial water saturation (Swi). The assumption is that identify petrophysical rock types. Nowadays, rock
the reservoir under study underwent hydrocarbon typing is a key element necessary to construct reliable
migration wherein vertical fluid distribution is still well three-dimensional reservoir models. Hydraulic rock
represented by the primary-drainage capillary pressure typing specifically refers to hydraulic properties of
curve. Petrophysical properties that are closely linked to reservoir rocks, including storage and flow capacity, and
pore geometry are quantified by invoking both provides a basis for many other reservoir
Leveretts J-function and Thomeers G-factor. The new characterization efforts such as saturation-height
log attribute is based on standard well-log analysis and analysis and dynamic reservoir modeling (Rushing et
only requires conventional well logs for its application. al., 2008). Numerous hydraulic rock typing methods
Thus, it can be generally applied to both clastic and were advanced in the past; some of them invoke core
carbonate reservoirs in multi-well contexts. It measurements, such as Leveretts reservoir quality index
overcomes the limitation of the bulk volume water (RQI = k / ) (Leverett, 1960), Winlands R35 (Pittman,
method which is only applicable to reservoir zones that 1992), and flow zone index (FZI) (Amaefule et al.,
are at nearly irreducible water saturation. Most 1993), while some others are mainly based on well logs,
importantly, it provides good initial estimates to such as those defined with the concept of electrofacies
constrain in-situ dynamic rock-fluid properties such as (Serra and Abbot, 1981) and bulk volume water (BVW
capillary pressure and relative permeability. = tSw, Buckles, 1965). Few papers have discussed the
technical challenges remaining to accurately and
The method proceeds with initial estimates of dynamic reliably map hydraulic rock types from core to well
properties to construct multi-layer petrophysical models logs. Common practice uses core-established rock types
with a common stratigraphic framework (CSF) for each and their corresponding well-log responses to identify
rock type, and to simulate the process of mud-filtrate and classify rock types in a supervised-learning
invasion. By honoring all the available well logs with approach. Often, this method is not satisfactory because

1
SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

well-log-defined facies may not be directly related to information for hydraulic rock typing. A reliable
pore geometry. Indeed, most well logs are more hydraulic rock typing scheme must honor both
sensitive to mineralogy and pore-filled fluids than to saturation-height variations in the vertical direction and
pore geometry. Resistivity logs are possibly the best mud-filtrate invasion in the radial direction. Several
indicators of rock texture. However, in many deepwater authors documented cases in which petrophysical
turbidite oil reservoirs, capillary transition effects render properties were inferred from the physics of mud-filtrate
resistivity logs difficult to use in rock typing across invasion (Salazar et al., 2006; Heidari et al., 2010). This
multiple wells due to fluid saturation changes with paper extends the concept of mud-filtrate invasion to
height (Fig. 2). This paper introduces a new method to integrate saturation-height modeling and establish rock-
classify rocks based solely on well-log-derived type-based dynamic rock-fluid petrophysical properties.
properties; it probes the pore geometry, eliminates the
effect of capillary transition on resistivity logs, and In our study, we first compare three well-known core-
reconciles rock classes with core mercury injection based quantifying methods of hydraulic rock types and
capillary pressure (MICP) data. Additionally, it extracts conclude that these methods are closely correlated due
important reservoir quality information from the to similar mathematical formulae and underlying
reservoirs saturation-height behavior and has practical petrophysical principles (pore-throat size or hydraulic
applications in reservoir development such as horizontal radius). We then select Leveretts RQI (the simplest
well steering, facies interpretation, and reservoir among the three methods) and derive its equivalent
connectivity analysis. The method also revisits Archies property from well logs using Leveretts J-function to
rock type definition and enables effective core-log assist rock typing in multiple wells that penetrate the
integration. same capillary transition zone. For each identified rock
type, we select a thick zone to simulate well logs under
mud-filtrate invasion and calibrate the associated
dynamic petrophysical properties. The workflow is
referred to as invasion facies modeling. Final products
from rock typing are not only a distribution of rock
types along the well trajectory, but also the associated
static and dynamic petrophysical properties, which are
important for reservoir modeling and simulation. We
use the calibrated capillary pressure curves to remodel
Fig. 1: Archies original definition of rock type and its saturation height based on the diagnosed rock types and
relation to petrophysical properties (Archie, 1950). compare them to the vertical distribution of resistivity-
derived water saturation.

The new method for rock typing is verified with

measurements acquired in two field cases of thick
turbidite oil reservoirs. In case no. 1, we show how
saturation-height and invasion consistent rock typing
provides important information for horizontal well
placement and steering. In case no. 2, we describe
results of the application of the method in multiple wells
to assist stratigraphic and sedimentary interpretation.

Fig. 2: Schematic of the effect of capillary transition on rock RQI vs. R35 vs. FZI
typing with multi-well well logs (Peters, 2012).
The three core-derived quantities have been widely used
In the process of reservoir characterization, saturation- for hydraulic rock typing. Arguments regarding their
height modeling needs to be rock-type based, i.e., strengths and weaknesses can be found in numerous
different rock types are associated with different publications on a case-by-case basis. Although these
saturation-height relations (Lucia, 1999). Core-derived quantities originated from different authors using
rock types and their capillary pressure properties should different approaches (empirical vs. experimental vs.
be consistent with well-log-derived saturation-height analytical), they bear more similarities than differences
distributions of water. At the same time, under relatively owing to the common underlying petrophysical property
steady drilling conditions, mud-filtrate invasion is that they intend to quantify - pore-throat size or
dominantly controlled by rock petrophysical properties. hydraulic radius. The three quantities are functions of
The radial invasion profile thus provides important porosity and permeability measured with routine core

2
 SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

analysis. Let stand for any of the above quantities. In-situ Reservoir Capillary Pressure
Their common formulae can be expressed as:
In a hydraulically connected reservoir, in-situ reservoir
log() = x log k + y log + z . (1) capillary pressure (Pc) is calculated from the difference
between true vertical depth (TVD) and the free water
We compared the coefficients x, y, and z associated with level (FWL), which can be derived from pressure
these formulae in Table 1 and found that coefficients x measurements, resistivity logs, or core MICP data,
are all close to 0.5, which has a well-established namely,
physical basis: permeability is proportional to the square
of pore-throat radius. To emphasize their similarity, we Pc = ( ) gh = 0.433 (TVD FWL) , (2)
repeat numerical testing of the correlation between these
where is density difference between connate water
quantities using randomly generated synthetic core
and hydrocarbon in g/cm3; TVD and FWL are depths in
porosity-permeability points (Fig. 3). Correlations
ft.
between each pair of quantities on a logarithmic scale
are consistently higher than 0.9. This simple exercise
indicates that the key to hydraulic rock typing for Well-Log-Derived Leveretts RQI
reservoir characterization does not rest on the selection Leveretts RQI can be derived from Pc and initial
of the three core-based quantities but on how to connate-water saturation (Sw) using the empirical J-
accurately map them into well-log domain. function model (Darling, 2005; Torres-Verdn, 2012)
S w S wirr 1b ,
Table 1: Coefficients of mathematical expressions S w = S wirr + aJ b J ( S w ) = ( ) (3)
associated with three core-based properties. a
x y z together with Leveretts capillary pressure model, given
Leveretts RQI 0.5 -0.5 0.0 by
Winlands R35 0.588 -0.864 0.732
k J (Sw ) S S wirr b1 cos ,
Amaefules FZI 0.5 -0.5 variable = cos = ( w ) (4)
Pc a Pc
2
5000
10 where Sw is initial water saturation, Swirr is irreducible
Permeability [md]

2
1000 water saturation (which is set to 0.01 lower than the
10
R35 [m]

200
100
50 0
minimum water saturation in the entire reservoir
25 10
10
5 R2=0.943
column, Darling, 2005); a and b are constants derived
0
10 2.5
1
0.5
from core-measured capillary pressure curves; cos is
0.25
10 -2
-2 the product of interfacial tension and contact angle,
0 20 40 0 2
10 10 10 which is assumed constant in the same reservoir; Pc is
Porosity [%]
FZI [m]
2 2 in-situ reservoir capillary pressure in psi. Equation (4)
10 10
provides a link between saturation-height and reservoir
quality index; it assumes that all reservoir rocks exhibit
R35 [m]

RQI [m]

0
10
0
10 the same Swirr and constants a and b. This assumption is
R2=0.911
2
R =0.996 not always true but it does provide a first-pass
-2 -2
quantification of reservoir quality based on a well-log-
10 -2 10 -2
10 10
0
10
2
10 10
0
10
2 derived saturation-height relationship and remains
RQI [m] FZI [m] reliable in many field cases.
Fig. 3: Numerical testing of the correlation between
Leveretts RQI, Winlands R35, and FZI using randomly Leveretts J vs. Thomeers G
generated porosity-permeability values.
An alternative parameter to quantify pore geometry
from well-log-derived saturation height is Thomeers G
METHOD AND WORKFLOW factor (1960). Thomeer proposed a model to define
Conventional Well-Log Analysis mercury/air capillary pressure curves given by
Sb
Core-calibrated conventional well-log analysis is critical = e G /log( Pc / Pe ) , (5)
to provide relevant petrophysical properties, such as Sb
volumetric concentration of shale (Csh), total porosity where Sb is mercury saturation at capillary pressure Pc;
(t), and connate-water saturation (Sw). All these Sb is mercury saturation at infinite capillary pressure; G
petrophysical properties relate more directly to is pore geometrical factor reflecting the distribution of
hydraulic rock types than apparent logs. Therefore, it is pore throats and their associated pore volume; Pe is the
here suggested that rock typing be based on these extrapolated displacement or entry pressure. Again, due
petrophysical properties whenever possible. to their similar mathematical expressions, Leveretts J
3
SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

and Thomeers G factors are highly correlated to each properties determine the radial distributions of water
other. Numerical testing using randomly generated saturation which are used to calculate radial
saturation-height points consistently gives rise to distributions of physical properties such as electrical
correlation coefficients higher than 0.9 (Fig. 4). For resistivity, density, and migration length, to numerically
simplicity, we consistently use Leveretts J function and simulate the corresponding apparent resistivity, density,
RQI to quantify rock types in this paper. and neutron logs (Gandhi et al., 2010). By matching all
10
2 available well logs after mud-filtrate invasion, we obtain
a set of estimated rock-fluid dynamic petrophysical
1 properties, including both capillary pressure and relative
G [dimensionless]

10
permeability. We adopt Voss et al.s (2009) and Heidari
0 R2=0.971 et al.s (2010) procedure to iteratively estimate
10
petrophysical properties using UTAPWeLS1. The
-1 measured well logs must be corrected for borehole
10
environmental effects to be compared to numerical
-2
simulations. Brooks-Coreys parametric equations are
10 -2 -1 0 1 2 adopted to describe saturation-dependent capillary
10 10 10 10 10
RQI [m] pressure and relative permeability (Corey, 1994; Peters,
2012), namely
Fig. 4: Numerical testing of the correlation between log-
derived Leveretts RQI and Thomeers G factor.
Pc = Pc0 t (1 S N ) ,
e p
(6)
k
where Pc is capillary pressure in psi, Pc0 is a constant
Rock Classification and Validation
coefficient in psi.darcy1/2, t is total porosity, k is
For each field case, we first perform core-based permeability in Darcy, ep is pore-size distribution
hydraulic rock typing with Leveretts RQI. We then exponent, and SN is normalized wetting-phase
consider the well-log-data domain to classify rock types saturation, given by
via cluster analysis on several relevant petrophysical S w S wr
SN = , (7)
attributes including volumetric concentration of shale, 1 S wr S nwr
total porosity, water saturation in the irreducible zone, where Sw is wetting-phase saturation, Swr is residual
and well-log-derived RQI (as in Eq. 4) in the capillary wetting-phase saturation, and Snwr is non-wetting phase
transition zone. Because both core- and well-log-based residual saturation. Wetting and non-wetting saturation-
rock typing include the same quantity - RQI, the dependent relative permeabilities, krw and krnw,
correlation between core-defined and well-log-derived respectively, are given by
rock types is significantly enhanced. Rock classification
krw = k rw
0
S Ne , w
(8)
results can be validated by predicting permeability based
on the well-log-derived rock types and compared to core and
measurements. In both cases, we show that a reliably- krnw = krnw
0
(1 S N )e . nw
(9)
defined rock type distribution can help to accurately
estimate permeability. Note that Snwr must be set to zero in a primary drainage
capillary pressure curve while it is non-zero in an
imbibition capillary pressure curve (Fig. 5).
Calibration of Dynamic Petrophysical Properties
from Invasion Simulation - Invasion Facies Modeling 60
Drainage
After rock typing, we calculate the porosity- 50 Imbibition
permeability trend, the saturation-height relation, and
capillary pressure curves for each rock type. We proceed 40
Pc (psi)

to calibrate capillary pressure and relative permeability 30

for each rock type by simulating the process of mud-
20
filtrate invasion at reservoir conditions. This simulation
involves rock petrophysical properties such as porosity, 10
permeability, capillary pressure, and relative
0
permeability. It also considers drilling engineering 0 0.2 0.4 0.6 0.8 1
Sw (frac)
parameters such as mud type, invasion duration, and
Fig. 5: Example of saturation-dependent drainage and
overbalance pressure. Fluid properties such as density, imbibition capillary pressure curves for the same rock.
viscosity, salt concentration, and temperature are also
taken into consideration. With stable drilling conditions
1
and constant fluid properties, rock petrophysical The University of Texas at Austins Petrophysics and Well Log
Simulator
4
 SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

Fig. 6: Rock typing in a synthetic reservoir model. Track 1: Depth; Track 2: Gamma ray; Track 3: Bulk density, neutron porosity,
and photoelectric factor; Track 4: Resistivity; Track 5: Total porosity from both well logs and model; Track 6: Total water
saturation from both well logs and model, and reservoir capillary pressure; Track 7: Lithofacies; Track 8: Electrofacies; Track 9:
BVW rock typing; Track 10: Rock typing from log-derived J-function; Track 11: pre-defined rock type model.

Saturation-Height Modeling with Brooks-Coreys upward sequences (grain-size: RT4 > RT3 > RT2). Bed
Drainage Capillary Pressure Model thickness is uniformly set to 3 ft to minimize thin-bed
effects on well logs. Sequences are separated by pure-
After calibrating each rock types capillary pressure
shale barriers. The reservoir has irreducible water
with invasion simulation, we perform saturation-height
saturation at the top and water saturation gradually
modeling on a rock-type basis. In a water-wet reservoir,
increases toward the FWL at the bottom. Figure 6 shows
saturation height modeling is performed by combining
the numerically simulated well logs (Tracks 1 - 4),
Eqs. (2), (6), and (7), i.e.
conventional well-log analysis results (Tracks 5 - 6), and
S w ( H ) = S wirr + (1 S wirr )[1 ( H ) ] ,
1/ e
p
(10) rock-typing results obtained with different methods
0.433 k (Tracks 7 - 11). Lithofacies (Track 7) can differentiate
= , (11)
Pc0 sand from shale, but cannot distinguish sands of
different grain sizes because all sands are essentially
where Swirr is irreducible water saturation and ep is pore- clay free. E-facies from all logs and BVW (Tracks 8 and
size distribution exponent; H is height above the FWL in 9) are significantly biased when moving downward
ft, is a scale coefficient that depends on rock type and toward the FWL due to fluid effects. We observe that
density difference between water and hydrocarbon. It is reservoir quality is overestimated in the upper zone
necessary to estimate Swirr, , and ep for each rock type while it is underestimated in the lower zone. Rock types
under a specific reservoir context to achieve satisfactory diagnosed with the new method (Track 10) yield
saturation-height modeling in formations without well approximately the same vertical distribution as in the
penetration. pre-defined (original) model (Track 11). This synthetic
case indicates that variations of water saturation due to
SYNTHETIC CASE capillary transition and their effects on well logs must be
A synthetic reservoir model is constructed to gain corrected to enable reliable rock classification across the
understanding on how capillary transition affects rock entire reservoir. Otherwise, resistivity logs, the most
typing via conventional methods such as E-facies and texture-sensitive logs among all conventional logs,
Buckles number. The reservoir model is an ideal cyclic cannot be included in the process of rock typing.
turbidite sequence with a total thickness of 300 ft. Three Consequently, there is great uncertainty associated with
hydraulic rock types of different grain sizes (therefore rock typing when it is only based on nuclear logs due to
different pore-throat radii) are stacked to form fining- their relatively low sensitivity to rock texture.
5
SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

FIELD CASE NO. 1: DEEPWATER CENTRAL average pore-throat radius ratio is approximately 1.5
NORTH SEA (i.e., 1.52 = 2.25). Because young-age sediments have
not undergone significant diagenesis, the controlling
Reservoir Background factor for permeability variations is mean grain size. By
inspecting the sand grain-size scale, mean grain size of
The first field example is a sandstone dome structure of
these four rock types is interpreted (from fine to coarse)
Paleocene age which is located in the Central North Sea
as lower fine (LF), upper fine (UF), low medium (LM),
(Martin et al., 2005; Salazar et al., 2007). Near-
and upper medium (UM). For rock types 5 and 6, clay
continuous sandstone bodies were formed by
cementation plays an important role in reducing
overlapping submarine fans during the early Paleocene.
permeability. They are mostly composed of very fine
Lithology in the preserved Paleocene strata mainly
(VF) and silty (ST) grains. There are several core
consists of siliciclastic sediments with minor presence
samples with permeability higher than 1000 md which
of coal, tuff, volcaniclastic rocks, marls, and reworked
are found to contain lower coarse-grained sandstones
carbonate sediments (Ahmadi et al., 2003). The
(LC or RT0). Due to their relatively small population, it
reservoir has a typical vertically segregated fluid
is difficult to separate them as one rock class via cluster
distribution with a gas cap (gas-oil-ratio = 871 scf/bbl)
analysis. The interpreted grain size information greatly
of average thickness of 192 ft and an oil column
contributes to the diagnosis of sedimentary facies in this
(gravity = 40o API) of average thickness of 184 ft
reservoir.
supported by an active aquifer zone (BP, 2003).

The formation under analysis (B1/B2 sands) is

Coarse 5000
predominantly composed of non-calcareous, blocky and Grained
sandy high-density gravity-flow deposits interbedded 1000
with grey mudstones. Reservoir units within the
Permeability [md]

2
10 200
formation consist of alternations of sandstones of 100
different grain sizes: fine- to medium-grained, or even 50
25
coarse-grained. Total porosity ranges from 18 to 28
10
p.u., while horizontal permeability varies from 10 md in 5
0
low-quality zones to more than 1,000 md in high- 10 2.5

1
quality intervals. 0.5
0.25

Hydraulic connectivity in the reservoir section under 0 10 20 30 40

analysis is confirmed by reservoir pressure Porosity [%]
measurements. The vertical reservoir fluid distribution
Fig. 7: Core-based hydraulic rock typing with Leveretts RQI
agrees well with the saturation derived from the core-
in North Sea field case.
measured primary drainage capillary pressure curve. It
is observed that the entire reservoir section under
analysis (B1/B2 sands) is in a capillary transition zone Table 2: Statistics of porosity, permeability, and RQI for each
with free water. hydraulic rock type and associated mean grain-size (MGS)
interpretation.
Well logs from three wells are available for rock typing. Rock Porosity Permeability RQI MGS
The three wells were drilled with OBM. Well No. 1 Type (p.u.) (md) (m)
(key well) is a vertical well penetrating the FWL while
wells nos. 2 and 3 are horizontal wells drilled 80 - 100 RRT0 24.0 1.86 1058 232 66.3 8.5 LC
ft above the FWL. RRT1 24.1 2.26 541 184 46.9 6.7 UM
Core-Based Rock Typing and Interpretation RRT2 24.2 1.95 255 63 32.2 3.6 LM
Hydraulic rock types are first established from routine RRT3 22.7 2.40 98.0 34 20.3 3.4 UF
core porosity-permeability measurements (Fig. 7). Six
RRT4 18.8 2.91 16.0 9.0 8.8 2.3 LF
hydraulic rock types are identified from the histogram
of RQI. Table 2 describes the statistics of petrophysical RRT5 14.2 4.08 1.30 0.9 2.8 0.8 VF
properties for each rock type. We observe that rock
RRT6 11.1 2.45 0.20 0.1 0.9 0.3 ST
types 1- 4 have a largely overlapping porosity range
which indicates that these rock types have similar grain-
sorting. However, their average horizontal permeability
exhibits a ratio around 2 - 2.5, which indicates that their

6
 SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

Fig. 8: Rock typing results in the key well of Central North Sea case. Tracks 1-4: basic well logs; Tracks 5-6: standard well-log
analysis, saturation-height for rock type 1 to 3; Tracks 7: distribution of rock types in M40-50 sands; Track 8: permeability
prediction based on rock types and comparison to core permeability.

Well-Log-Based Saturation-Height, Rock Typing,

and Permeability Estimation
We perform core-calibrated, conventional well-log
analysis in the key well (no. 1) to calculate volumetric
concentration of shale, total porosity, and water
saturation (Fig. 8, Tracks 5 - 6). From reservoir
pressure measurements, we identify fluid types and
their contacts as well as the FWL. Well-log-derived
saturation-height and core-measured capillary pressure
agree well (Fig. 9). From saturation-height, we derive
Leveretts RQI using Eq. (4) with constants a = 0.33, b
= -0.62, and Swirr = 0.04. We then classify rocks based
on a combination of several petrophysical attributes,
including volumetric concentration of shale, total Fig. 9: Reconciliation between core-measured MICP and log-
derived saturation-height relation.
porosity, and log-derived RQI in the capillary transition
zone. Figure 8 shows the results of rock classification
(Track 7) and the permeability estimated from rock Invasion Facies Modeling
types (Track 8). Estimated permeability agrees well After calculating the well-log-based saturation-height
with core-measured permeability (with Klinkenberg relation, we obtain an average capillary pressure curve
correction). for each rock type by curve-fitting with the same rock
type. Core calibration is preferred but not required here.
7
SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

We select thick zones of rock types 1 to 3 at different Table 3: Summary of mudcake, fluid, and formation properties
height to simulate the radial invasion profile under assumed in the simulation of the process of mud-filtrate
observed drilling and reservoir conditions to further invasion.
calibrate capillary pressure and relative permeability Variable Units Value
curves. In this case, the top formation is saturated with Wellbore radius inch 6.15
light condensate with almost no free water. When Maximum invasion time days 2
invaded by OBM, the formation should not exhibit Formation outer boundary ft 2000
o
radial variations of water saturation in theory. However, Reservoir temperature F 213
we still observe a small separation between apparent Initial reservoir pressure psi 3800
resistivity logs with different depths of investigation, Gas viscosity (reservoir conditions) cP 0.02
which has two possible explanations: The first Oil viscosity (reservoir conditions) cP 0.5
explanation is that when overbalance pressure is very Water viscosity cP 0.3
high (1,400 psi in this case), it can still displace a small (reservoir conditions)
amount of capillary-bound water deeper into the Overbalance pressure psi 1400
reservoir at this height above FWL (Fig. 9). The second Mud-filtrate density (at STP) g/cm3 0.9
explanation is that surfactants in OBM mud filtrate alter Mud-filtrate viscosity (at STP) cP 5
rock wettability which subsequently generates more Mud-filtrate compressibility psi-1 3.6x10-6
free water (La Vigne et al., 1997; Salazar et al., 2007; (at STP)
Pour et al., 2011). In both cases, capillary pressure and Formation compressibility psi-1 4 x 10-7
relative permeability play important roles in Mudcake reference md 0.03
determining how this relatively small volume of free permeability
water is distributed in the invaded zone regardless of Mudcake reference porosity frac. 0.35
mechanistic origin. Therefore, it is possible to calibrate Mud solid fraction frac. 0.06
rock dynamic petrophysical properties by honoring the Mudcake maximum thickness inch 0.4
effects of mud-filtrate invasion on various well logs. Mudcake compressibility frac. 0.40
exponent
Assuming that the rock is preferentially water-wet, Mudcake exponent multiplier frac. 0.10
OBM invasion becomes a drainage process. We
numerically simulate gamma-ray, bulk density, and 0.25
array-induction apparent resistivity logs in this
Water Saturation [frac]

example. Neutron and PEF logs are excluded due to

0.2
environmental correction problems. Table 3 lists the
simulation parameters and Figs. 10 - 11 show the radial
0.15
invasion profiles and numerical well-log simulation for
rock types 1 to 3 after 2 days of OBM invasion. It is
observed that the best rock type gives rise to the highest 0.1 RT 1
resistivity due to the lowest connate water saturation. RT 2
RT 3
Table 4 and Fig. 12 summarize the final capillary 0.05
2 4 6 8 10
pressure and relative permeability curves for the Radial Distance [ft]
selected rock types. These properties simultaneously
honor routine core measurements, the well-log-derived 50
RT 1
saturation-height relation, and mud-filtrate invasion. RT 2
40
Resistivity [OHMM]

Therefore, they are suitable for upscaling in reservoir RT 3

simulation grids. One key advantage of invasion- 30
calibrated dynamic petrophysical properties is that these
properties have been calculated at reservoir conditions. 20

Table 4: Rock-fluid properties calibrated and optimized with 10

the simulation of mud-filtrate invasion for rock types 1 to 3
using Brooks-Coreys model. 0
2 4 6 8 10
0 0 0
RT k / Swirr Sor k rnw enw k rw ew Pc ep Radial Distance [ft]

1 46.9 0.03 0.2 0.9 3.2 0.36 4.6 168 21.4 Fig. 10: Radial distributions of water saturation (upper
panel) and resistivity (lower panel) of rock types 1 to 3 after
2 32.2 0.05 0.22 0.88 3.0 0.34 3.8 128 16.1 invasion with OBM for 2 days. The best rock type exhibits the
largest resistivity, the lowest water saturation, and the
3 20.3 0.07 0.25 0.86 2.6 0.3 3.2 96 11.2 shallowest invasion.

8
 SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

Saturation-Height Modeling After Invasion

Simulation
Using the capillary pressure curves calibrated from
invasion simulation, we derive saturation-height
relations for rock types 1 to 3 (Table 5). Saturation-
height curves for each rock type are plotted together
with the resistivity-derived water saturation (Fig. 8,
track 6), indicating a good agreement. We fit 16 core
MICP measurements with Brooks-Coreys model and
found that the parameters included in Brooks-Coreys
function were closely correlated to Leveretts reservoir
quality index. Better rock quality indicates higher Pc0
and ep together with lower Swirr and Pe (Fig. 13). These
values are also comparable to those calibrated with
invasion simulation.

Table 5: Saturation height relations for rock types 1 to 3.

RT1: S w ( H ) = 0.03 + 0.97[1 (0.0015 H )0.047 ]

RT2: S w ( H ) = 0.05 + 0.95[1 (0.0014 H )0.062 ]

RT3: S w ( H ) = 0.07 + 0.93[1 (0.0012 H )0.091 ]

200 40
P0c (psi.D0.5)

ep
100 20

0 0
0 50 0 20 40 60 80
RQI (m) RQI (m)
0.2 2
Swirr [frac]

Pe [psi]
0.1 1

0 0
0 50 0 20 40 60 80
RQI (m) RQI (m)

Fig. 13: Correlation between Leveretts rock quality index

and capillary pressure properties: Pc0, ep, Swirr, and Pe.
Fig. 11: Invasion simulation for rock types 1 to 3 (top to from 16 core MICP data.
bottom) and the corresponding numerical log simulation
compared to field logs. Track 1: depth; Track 2: GR log; Rock Typing and Horizontal Well Steering
Track 3: bulk density and neutron porosity logs; Track 4:
induction (AIT) resistivity logs; Track 5: cross-section After reconciling rock types with their saturation-height
(vertical and radial directions) of water saturation. relations in the key well, we examine the remaining two
100 1 horizontal wells (nos. 2 and 3) drilled in the same
RT1
RT2
formation to study lateral variability of reservoir
80 0.8
RT3 quality. We perform conventional well-log analysis in
60 0.6 well no. 2 and find that water saturation in sandstones is
Kr [frac]
Pc [psi]

approximately 30% with a height above FWL of 100 ft.

40 0.4 From the projection of the well-log-derived saturation-
height data shown in Fig. 9 (red dots), we observe that
20 0.2
major rock types in this horizontal section are RRT4
0 0 and RRT5 with an average porosity 12 p.u. and
0 0.5 1 0 0.5 1
Sw [frac] Sw [frac] permeability ranging from 1 to 10 md. Decreasing
reservoir quality indicates that well no. 2 was drilled
Fig. 12: Invasion-calibrated, saturation-dependent capillary
toward the levee or distal fan facies and that the well
pressure and relative permeability for each rock type.
productivity is relatively low (Fig. 14).
9
SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

In well no. 3, predominant washouts render neutron and 1 was drilled with OBM and penetrated a capillary
density logs useless. The only reliable logs are GR and transition zone and the FWL. Wells nos. 2 and 3 were
deep resistivity. With an assumed constant porosity of drilled with WBM where the M40-50 sands are located
22 p.u., water saturation calculated with Archies more than 500 ft above FWL (Fig. 15).
equation is approximately 20% with a height above
FWL of 80 ft. From the projection of well-log-derived
saturation-height data shown in Fig. 9 (blue dots), we
observe that the major rock type is RRT2 with No.1 No.2 No.3
permeability ranging from 100 to 300 md. Reservoir
quality remains the same as that of the key well,
thereby indicating that the well was drilled along the
channel axis or fan lobe and that well productivity is
much better than that of well no. 2 (Fig. 14), which is
cross-validated by core data available in these two
wells. Fig. 15: Three well locations displayed on a seismic cross-
section with litho-type well logs indicating the vertical interval
This example confirms that saturation-height-based of the Miocene M-series sands (Contreras et al., 2006).
rock typing in horizontal wells can reveal lateral
Core-Based Rock Typing and Interpretation
reservoir-quality changes and provide critical real-time
information for guiding the placement and steering of Hydraulic rock types are first established using routine
horizontal wells. core porosity-permeability measurements. Six hydraulic
rock types were identified from the histogram
Submarine Fan Model distribution of RQI (Fig. 16). Table 6 summarizes the
statistics of petrophysical properties for each rock type.
Similar to field case no. 1, we observe that rock types 1
- 3 exhibit a largely overlapping porosity range. Again,
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.
Fig. 14: Description of the possible drilling scenarios
Reservoir quality decreases with decreasing grain size.
interpreted from rock typing in horizontal wells.
5000

FIELD CASE NO. 2: DEEPWATER GULF OF 1000

MEXICO
Permeability [md]

2
10 200
100
Reservoir Background 50
25
The second field example is a Miocene turbidite oil 10

field located in the deepwater Gulf of Mexico 0

5
10 2.5
(Contreras et al., 2006). The depositional system is 1

interpreted as a submarine fan complex developed in a 0.5

0.25
mini-basin with stacked progradational lobes.
Reservoirs primarily consist of sandy turbidite facies 0 10 20 30 40
Porosity [%]
interbedded with muddy debris-flow facies. Reservoir
rocks are mainly unconsolidated sandstones with very Fig. 16: Core-based hydraulic rock typing from Leveretts
fine to fine- to medium-size grains. Thin-bed zones are RQI in deepwater Gulf-of-Mexico field case.
present but are not addressed here. In this study, we
evaluate only the two hydraulically connected, Well-Log-Based Saturation-Height, Rock Typing,
hydrocarbon-bearing sand units: M40 and M50 in three and Permeability Estimation
wells. These sands are buried at depths between 12,000
and 13,000 ft TVD and are dipping toward the West We perform core-calibrated, conventional well-log
(Fig. 15). Core measurements indicate that the massive analysis in the three wells to calculate volumetric
and planar stratified sands have porosities up to 35%, as concentration of shale, total porosity, and water
well as 100 2000 md of nominal permeability. Well saturation. Normalization of gamma-ray logs is not
logs from three wells are used for rock typing. Well no. performed in this study because we assume that the
10
 SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

gamma-ray log self-normalizes in each well when namely that apparent resistivity logs are irreplaceable
transforming gamma-ray logs to volumetric for hydraulic rock typing.
concentration of shale. From reservoir pressure
Table 7 summarizes the simulation parameters. We
measurements, we identify the fluid type, their contacts,
iteratively match numerically simulated and measured
and the FWL. Well-log-derived saturation-height and
well logs by adjusting the capillary pressure and
core-measured capillary pressure agree well (Fig. 17).
relative permeability curves. Figure 21 shows the final
From saturation-height results, we derive Leveretts
simulations of radial invasion profiles and numerically
RQI using Eq. (4) with constants a = 0.63, b = -0.68,
simulated well logs for rock types 1 to 3 after 1.0 - 1.5
and Swirr = 0.05. Subsequently, we classify rocks based
days of WBM invasion. Table 8 describes the final
on a combination of several petrophysical attributes
capillary pressure and relative permeability curves for
including volumetric concentration of shale, total
the selected rock types using Brooks-Coreys
porosity, water saturation in the irreducible depth zone,
parametric model and plotted in Fig. 23.
and log-derived RQI in the capillary transition zone.
Figure 18 shows results obtained from rock
1200
classification (Track 7), as well as the permeability Well No.1
estimated from the rock types (Track 8). Estimated 1000 Well No.2
permeability agrees well with permeability (with Well No.3
Klinkenberg correction under stress) measured from 800

HAFWL (ft)
whole core in the key well. In well no. 3, rock types
agree well with the permeability measured from 600
sidewall core samples. Rock typing results obtained in
the three wells are displayed together to assist 400
stratigraphic and sedimentary interpretations (Fig.19)
200
(Xu et al., 2012).
0
Table 6: Statistics of porosity, permeability, and RQI for 0 0.2 0.4 0.6 0.8 1
each hydraulic rock type. Sw (frac)

Fig. 17: Reconciliation between core-measured MICP and

Rock Porosity Permeability RQI log-derived saturation-height relations in multiple wells.
Type (p.u.) (md) (m)
RRT1 33.6 1.8 1099 291 56.8 7.2 Table 7: Summary of mudcake, fluid, and formation
properties assumed in the simulation of the process of mud-
RRT2 33.3 2.1 549 130 40.4 4.9 filtrate invasion.
RRT3 31.5 4.4 179 63.7 23.4 3.4 Variable Units Value
RRT4 24.1 6.1 42.9 19.1 13.1 2.1 Wellbore radius inch 6.0
Maximum invasion time days 1.0 - 1.5
RRT5 19.7 3.9 11.3 3.9 7.50 1.1 Formation outer boundary ft 2000
o
RRT6 19.6 3.7 4.00 1.5 4.40 0.9 Reservoir temperature F 120
Initial reservoir pressure psi 7800
Water viscosity cP 0.92
(reservoir conditions)
Invasion Facies Modeling
Oil viscosity cP 0.8
We select thick zones of rock types 1 to 3 at irreducible Overbalance pressure psi 500
water saturation to simulate invasion with WBM, which Salt dispersivity ft 0.1 - 0.5
is an imbibition process in water-wet rocks. Rock types Mud-filtrate density (at STP) g/cm3 1.02
1 and 2 are selected from well no. 2 and rock type 3 is Mud-filtrate viscosity (at STP) cP 0.92
selected from well no. 3. Both wells lack whole core Mud-filtrate compressibility psi-1 3.6 x 10-6
samples and special core analysis. It is therefore (at STP)
important to extrapolate our understanding of rock Formation compressibility psi-1 4 x 10-7
types from the key well to uncored wells. We use Mudcake reference permeability md 0.01
gamma-ray, bulk density, neutron porosity, and AIT Mudcake reference porosity frac. 0.35
apparent resistivity logs for simulation. The PEF log is Mud solid fraction frac. 0.06
excluded due to presence of barite in the drilling mud. Mudcake maximum thickness inch 0.4
From well logs acquired across each rock type, we Mudcake compressibility frac. 0.40
observe that they exhibit essentially the same nuclear exponent
responses, which confirms one previous conclusion, Mudcake exponent multiplier frac. 0.10
11
 SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

Fig. 18: Rock typing results in the key well of 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: distributions of rock types in M40-50 sands; Track 8: permeability
prediction based on rock types and comparison to core permeability. Permeability is underestimated only in thin-bed zones.

Fig. 19: Use of rock-type distributions in multiple wells to assist sedimentological and stratigraphic interpretation.
12
 SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

Table 8: Rock-fluid properties calibrated with the simulation role in the near-borehole process of salt exchange
of mud-filtrate invasion for rock types 1 to 3 using Brooks- between mud filtrate and connate water. Low porosity-
Coreys model. permeability rocks typically exhibit large salt
RT k / Swirr Sor k0rnw enw k0rw ew Pc0 ep dispersivity, which yields a smoother radial salt
concentration front (Fig. 22, center panel).
1 60.3 0.05 0.22 0.85 3.2 0.3 5.0 270 26
2 38.9 0.08 0.23 0.87 3.0 0.25 4.5 185 18.6
3 24.6 0.12 0.18 0.9 3.0 0.3 4 121 12.4

Saturation-Height Modeling After Invasion

Simulation
Using the capillary pressure curves calibrated from
invasion simulation, we derived saturation-height
relations for rock types 1 to 3 (described in Table 9). To
that end, we fit 14 core MICP measurements with
Brooks-Coreys parametric model and found that the
parameters included in Brooks-Coreys function were
closely correlated to Leveretts reservoir quality index.
Better rock quality indicates higher Pc0 and ep together
with lower Swirr and Pe (Fig. 20), which enables a
consistent petrophysical interpretation between well-
log-derived saturation height and core MICP data.
Table 9: Saturation-height relations for rock types 1 to 3.

RT1: S w ( H ) = 0.05 + 0.95[1 (0.0011H )0.039 ]

RT2: S w ( H ) = 0.08 + 0.92[1 (0.0010 H )0.054 ]

RT3: S w ( H ) = 0.12 + 0.88[1 (0.0010 H )0.081 ]

400 50
P0c (psi.D0.5)

ep

200

0 0
0 20 40 60 80 0 20 40 60 80
RQI (m) RQI (m)
0.5 5
Swirr [frac]

Pe [psi]

0 0
0 20 40 60 80 0 20 40 60 80
RQI (m) RQI (m)

Fig. 20: Correlation between Leveretts rock quality index

and capillary pressure properties: Pc0, ep, Swirr, and Pe. from
14 core MICP data.

Figure 22 shows the radial profiles of water saturation,

salt concentration, and resistivity for each rock type Fig. 21: Invasion simulation for rock types 1 to 3 (top to
after WBM invasion. In general, good-quality sands bottom) and corresponding numerical well-log simulation
compared to field logs. Track 1: Depth; Track 2: GR log;
exhibit low irreducible water saturation but not
Track 3: Bulk density and neutron logs; Track 4: Induction
necessarily more residual oil saturation due to (AIT) resistivity logs; Track 5: Cross-section (vertical and
wettability variations. The best quality sands exhibit the radial directions) of water saturation. Three different rock
largest separation between deepest and shallowest types exhibit the same nuclear-log responses. Only resistivity
apparent resistivity. Salt dispersivity plays an important logs can differentiate rock types.

13
SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

1 quantify pore geometry from routine and special core

RT 1 analysis, it remains challenging to quantify it solely
RT 2
Water Saturation [frac]

0.8
RT 3
with conventional well logs. E-facies diagnosed from
apparent well logs are not satisfactory because the latter
0.6
are simultaneously sensitive to mineralogy and pore
0.4
fluids. Buckles number relates to pore geometry in a
direct manner but is only applicable to reservoir zones
0.2 which are at irreducible water saturation conditions. We
constructed a synthetic case with pre-defined rock types
0 to illustrate how E-facies and Buckles number are
1 2 3 4 5
Radial Distance [ft] biased toward capillary transition zones. New methods
4
x 10
are necessary to perform hydraulic rock typing in thick
16
oil reservoirs that exhibit long capillary transition
14 zones.
Water Salinity [ppm NaCl]

12
We studied and appraised three conventional core-
10 based hydraulic rock typing methods and performed
RT 1 numerical correlation testing to confirm their
8
RT 2
6 RT 3 petrophysical similarity. Consequently, the three
methods should perform equally well in detecting
4
petrophysical rock types with different pore-throat
2 radii. It was also pointed out that the key to reliable
1 2 3 4 5
Radial Distance [ft] rock typing for reservoir characterization is to construct
well-log-derived attributes that directly link rock types
30
RT 1 diagnosed from core measurements to well logs.
25 RT 2
Resistivity [OHMM]

RT 3 This paper introduced a new method for hydraulic rock

20 typing in multiple wells penetrating a capillary
15 transition zone that integrates vertical distributions of
initial water saturation with reservoir capillary pressure.
10
Conventional Leveretts J-function and Thomeers G
5 factor were invoked to quantify reservoir quality in
terms of pore geometry. The new method honors both
0
1 2 3 4 5 vertical saturation-height behavior and mud-filtrate
Radial Distance [ft]
invasion, both of which bear different expressions of
Fig. 22: Radial distributions of water saturation (upper hydraulic rock types. We successfully verified the new
panel), salt concentration (center panel), and resistivity method with measurements acquired in two thick
(lower panel) for rock types 1 to 3 after invasion with WBM turbidite oil reservoirs from the Central North Sea and
for 1 - 1.5 days. the deepwater Gulf of Mexico. In both cases, a good
100 1
agreement was reached between core-derived and well-
RT1 log-derived rock types, which were also validated with
80 RT2 0.8 the estimation of permeability.
RT3
60 0.6 The petrophysical quality of young deepwater turbidite
Kr [frac]
Pc [psi]

reservoirs is chiefly controlled by sedimentary grain

40 0.4 sizes. Therefore, a link exists between petrofacies and
20 0.2
depositional facies which can be effectively applied in
reservoir development efforts. Reservoir capillary
0
0 0.5 1
0
0 0.5 1
transition enables the estimation of grain sizes from
Sw [frac] Sw [frac] saturation-height data. In the Central North Sea case
Fig. 23: Invasion-calibrated, saturation-dependent capillary examined in this paper, we showed how to implement
pressure and relative permeability for each rock type. saturation-height-derived rock types to quantify lateral
variability of reservoir petrophysical quality and to
guide well placement and steering with minimum LWD
SUMMARY AND CONCLUSIONS data in a submarine fan. In the deepwater Gulf-of-
Hydraulic rock typing must be based on pore geometry Mexico case, we showed that multi-well rock typing
because it relates all petrophysical properties of can assist in stratigraphic and sedimentary studies
reservoir rocks. While it is relatively simple to infer or which are often used for exploration in frontier basins
14
 SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

or in deepwater environments, where subsalt or pre-salt k : Absolute permeability, (md)

reservoirs cannot be clearly imaged with seismic data krw : Wetting-phase relative permeability
and well drilling is limited due to high rig cost. It was krnw : Non-wetting phase relative permeability
also found the two field cases could be approached with Pc : Reservoir capillary pressure, (psi)
the same rock-type scheme when using the ratio of P c0 : Coefficient for Pc equation, (psi.darcy1/2)
permeability between neighboring sandy rock types. Pe : Entry capillary pressure, (psi)
This observation is not coincidental and has geological R35 : Pore-throat radius at 35% non-wetting phase
and petrophysical significance. Rock types associated Sw : Connate water saturation, (frac)
with young-age turbidite reservoirs can still be Swirr : Irreducible water saturation, (frac)
differentiated by their median grain sizes, whereas Sor : Residual oil saturation, (frac)
neighboring rock types exhibit a grain size ratio close to x : Coefficient of permeability in rock quality
1.5 or 2 . Well-log-derived rock types capture this y : Coefficient of porosity in rock quality
property, which suggests that rock typing in similar z : Constant term in rock quality
cases may generally follow the same principle. : Scaling factor in saturation-height modeling
t : Total porosity, (frac)
We suggest that rock typing should not only separate : Density difference between water and oil
rocks into different groups. The more important task is : Interfacial tension, (dyne)
to calculate static and dynamic petrophysical properties cos : Cosine of contact angle
associated with each rock type based on core-log : A symbol for quantifying pore-throat size
integration. The new interpretation method introduced
in this paper provides a good initial estimation of rock- ACRONYMS
type-averaged capillary pressure and relative
AIT : Array Induction Tool
permeability. Both properties are further calibrated for
BVW : Bulk Volume Water
each rock type by numerically simulating well logs
CSF : Common Stratigraphic Framework
based on the physics of mud-filtrate invasion. It was
FWL : Free Water Level
emphasized that each rock type is associated with a
FZI : Flow Zone Index
specific radial invasion profile due to differences in
GR : Gamma Ray Log
pore geometry. Therefore, it is possible to use invasion
LWD : Logging While Drilling
behavior to define dynamic petrophysical facies (or
MICP : Mercury Injection Capillary Pressure
invasion facies) for reservoir simulation.
OBM : Oil-Base Mud
The method introduced in this paper assumes that the PEF : Photoelectric Factor Log
reservoir under analysis is hydraulically connected and RQI : Leveretts reservoir quality index, k /
underwent hydrocarbon migration similar to a primary RRT : Reservoir Rock Type
drainage process. Complex hydrocarbon migration may TVD : True Vertical Depth
give rise to saturation hysteresis which could deviate WBM : Water-Base Mud
from the saturation-height relation defined by core
MICP measurements. Reservoir pressure measurements ACKNOWLEDGEMENTS
should be used to confirm reservoir connectivity, We thank BP, the Andrew Field Partner Group, and
whereas core-derived capillary pressure curves should Anadarko Petroleum Corporation for providing the data
be used to verify the well-log-derived saturation-height used in the field studies. Special thanks go to Dr.
relation. One limitation of the method introduced in this Ekwere Peters for his thought-provoking discussions on
paper is that it is only applicable to reservoir zones the link between saturation height and hydraulic rock
above oil-water or gas-water contacts. Moreover, the typing. We acknowledge Dr. Ron Steels guidance with
resolution of rock types is limited by the vertical stratigraphic and sedimentary interpretation. Chicheng
resolution of the deep apparent resistivity log. Xu also thanks BP Gulf of Mexico for an intern
NOMENCLATURE opportunity on Paleocene turbidite reservoirs which
greatly enhanced his understanding on hydraulic rock
a : Linear coefficient in empirical J function typing. The work reported in this paper was funded by
b : Exponent in empirical J function The University of Texas at Austins Research
Csh : Volumetric concentration of shale Consortium on Formation Evaluation, jointly sponsored
ep : Pore-size distribution exponent by Anadarko, Apache, Aramco, Baker-Hughes, BG,
enw : Experimental exponent for krnw equation BHP Billiton, BP, Chevron, ConocoPhillips, ENI,
ew : Experimental exponent for krw equation ExxonMobil, Halliburton, Hess, Maersk, Marathon Oil
G : Thomeers pore geometrical factor Corporation, Mexican Institute for Petroleum, Nexen,
H : Height above the free-water level ONGC, Petrobras, Repsol, RWE, Schlumberger, Shell,
J : Leveretts J function Statoil, Total, and Weatherford.
15
SPWLA 53rd Annual Logging Symposium, June 16-20, 2012

REFERENCES Extends the Sand-Free Production Life of Andrew

Field, Paper SPE 93639.
Ahmadi, M., Sawyers, M., Kenyon-Roberts, S., Peters, E.J., 2012, Advanced Petrophysics - Volumes 1
Stanworth, B., Kugler, K., Kristensen, J., and Fugelli, and 2, Greenleaf Book Group, Austin.
E., 2003, The Millennium Atlas: Petroleum Geology Pittman, E.D., 1992, Relationship of Porosity and
of the Central and Northern North Sea, Ch. 14, the Permeability to Various Parameters Derived from
Geological Society of London. Mercury Injection-Capillary Pressure Curves for
Amaefule, J.O., Altunbay, M., Tiab, D., Kersey, D.G., Sandstone, AAPG Bulletin, Vol. 76, No. 2, P. 191 -
and Keelan, D.K., 1993, Enhanced Reservoir 198.
Description: Using Core and Log Data to Identify Pour. R.A., Torres-Verdn, C., and Sepehrnoori, K.,
Hydraulic (Flow) Units and Predict Permeability in 2011, Quantifying Wettability Alteration during Oil-
Uncored Intervals/Wells, Paper SPE 26436. Base Mud-Filtrate Invasion and Corresponding
Archie, G.E., 1950, Introduction to Petrophysics of Effects on Resistivity Logs, SPWLA 52nd Annual
Reservoir Rocks, AAPG Bulletin, Vol. 34, No. 5, P. Logging Symposium, Colorado Springs, USA, May 14
943 - 961. - 18.
BP, 2003, The UK Upstream Asset Portfolio, BP Rushing, J.A., Newsham, K.E., and Blasingame T.A.,
Exploration, Farburn Industrial Estate, Dyce, 2008, Rock Typing - Keys to Understanding
Aberdeen, U.K. Productivity in Tight Gas Sands, Paper SPE 114164.
Brooks, R.H. and Corey, A.T., 1966, Properties of Salazar, J.M., Malik, M., Torres-Verdn, C., Wang,
Porous Media Affecting Fluid Flow, Journal of G.L., and Duan, H.Y., 2007, Fluid Density and
Irrigation and Drainage Division, Vol. 92, No. 2, P. Viscosity Effects on Borehole Resistivity
61 - 90. Measurements Acquired in the Presence of Oil-Based
Buckles, R.S., 1965, Correlating and Averaging Mud and Emulsified Surfactants, SPE 109946, 2007
Connate Water Saturation Data, Journal of Canadian SPE Annual Technical Conference and Exhibition,
Petroleum Technology, Vol. 9, No. 1, P.42 - 52. Anaheim, California, USA, November 11 - 14.
Contreras A., Torres-Verdn, C., Chesters, W., Kvien,
Salazar, J.M., Torres-Verdn, C., Alpak, F.O., Habashy,
K., and Fasnacht T., 2006, Extrapolation of Flow
T.M., and Klein, J.D., 2006, Estimation of
Units away from Wells with 3D Pre-Stack Seismic
Permeability from Borehole Array Induction
Amplitude Data: Field Example, Petrophysics, Vol.
Measurements: Application to the Petrophysical
47, No. 3, P. 223 - 238.
Appraisal of Tight-Gas Sands, Petrophysics, Vol. 47,
Corey, A. T., 1994, Mechanics of Immiscible Fluids in
No. 6, P. 527 - 544.
Porous Media: Water Resources Publications,
Serra, O. and Abbott, H. T., 1980, The Contribution of
Highland Ranch, Colorado.
Logging Data to Sedimentology and Stratigraphy,
Darling, T., 2005, Well Logging and Formation
SPE 9270, 55th Annual Fall Technical Conference
Evaluation, Gulf Professional Publishing.
and Exhibition, Dallas, Texas, USA.
Gandhi, A., Torres-Verdn, C., Voss, B., Gabulle, J.,
and Seminario F., Construction of Reliable Static and Thomeer, J.H.M., 1960, Introduction of a Pore
Dynamic Multi-Layer Petrophysical Models in Geometrical Factor Defined by the Capillary Pressure
Camisea Gas Reservoirs, Peru, SPWLA 51st Annual Curve, Journal of Petroleum Technology, Vol. 12,
Logging Symposium, Perth, Australia, June 19 - 23. No. 3, P. 73 - 77.
Heidari, Z., Torres-Verdn, C., Mendoza, A., and Wang Torres-Verdn, C., 2012, Integrated Geological-
G.L., 2011, Assessment of Residual Hydrocarbon Petrophysical Interpretation of Well Logs, Lecture
Saturation with the Combined Quantitative Notes for Fundamentals of Well Logging, The
Interpretation of Resistivity and Nuclear Logs, University of Texas at Austin, Unpublished.
Petrophysics, Vol. 52, No. 3, P. 217 - 237. Voss, B., Torres-Verdn, C., Gandhi, A., Alabi, G., and
La Vigne, J., Barber, T., and Bratton, T., 1997, Strange Lemkecher, M., 2009, Common Stratigraphic
Invasion Profiles : What Multiarray Induction Logs Framework to Simulate Well Logs and to Cross-
Can Tell Us About How Oil-Based Mud Affects the Validate Static and Dynamic Petrophysical
Invasion Process and Wellbore Stability, SPWLA 38th Interpretations, SPWLA 50th Annual Logging
Annual Logging Symposium, Houston, Texas, USA, Symposium, The Woodlands, Texas, USA, June 21 -
June 15 - 18. 24.
Leverett, M.C., 1941, Capillary Behaviour in Porous Xu, C.C., Torres-Verdn, C., and Steel. R.J., 2012,
Solids, Transactions of the AIME, 142, P.159 - 172. Facies Interpretation Based on Quantitative Analysis
Lucia, F.J.,1999, Carbonate reservoir characterization, of Grain Size and Bed Thickness from Well Logs in
Springer-Verlag, Berlin. Deepwater Turbidite Reservoirs, AAPG Annual
Martin, A.J., Robertson, D., Wreford, J., and Lindsay, Convention and Exhibition, Long Beach, California,
A., 2005, High-Accuracy Oriented Perforating USA, April 22 - 25.
16