See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/254534890

Reconciliation of Core and Log Data Analysis in Very Thin Reservoirs of

Krishna Godavari Basin, India

Article · January 2012

DOI: 10.2118/149017-MS

CITATION READS

1 392

3 authors:

Lalaji Yadav Tanmoy Dutta

Reliance Industries Limited Indian Institute of Technology Roorkee
41 PUBLICATIONS   50 CITATIONS    6 PUBLICATIONS   44 CITATIONS   

SEE PROFILE SEE PROFILE

Neeraj Sinha
Reliance Industries Limited
27 PUBLICATIONS   223 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Prediction of Archie parameters from grain size variations View project

Rifted margins; their breakup mechanisms and controlling factors View project

All content following this page was uploaded by Lalaji Yadav on 21 April 2015.

The user has requested enhancement of the downloaded file.

SPE 149017

Reconciliation of Core and Log Data Analysis in Very Thin Reservoirs of

Krishna Godavari Basin, India
Lalaji Yadav, Tanmoy Dutta, Neeraj Sinha, Reliance Industries Limited

Copyright 2012, Society of Petroleum Engineers

This paper was prepared for presentation at the North Africa Technical Conference and Exhibition held in Cairo, Egypt, 20–22 February 2012.

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
Deep water Tertiary reservoirs of Krishna Godavari basin are generally very complex and heterogeneous, ranging from
massive thick sands to highly laminated very thin sand/shale sequences beyond the vertical resolution of imaging tools.
Verification of log-derived quantities in very thin reservoirs is frequently difficult because of bias and scatter introduced by
non-uniform or incomplete plugging of the available over 750 meters of conventional cores. As laminar shale does not occupy
or alter the intergranular sand porosity, the conventional way of validation of log derived porosity and water saturation with
core data showing excellent correlation does not ensure the correctness of petrophysical evaluations.

In this paper, it has been demonstrated that there are potential problems associated with the integration of petrophysical
evaluation and core data. Sources of errors are mainly due to the use of different parametric systems viz. using of the terms
shale volume and clay volume interchangeably and validating log derived shaliness with clay abundance from XRD core data.
Shaliness derived from logs cannot be directly correlated with such core data because neither de-averaging of lithology on log
analysis nor averaging of core data is possible for the effective core-log integration.

Average hydrocarbon pore volume in laminated reservoirs is linearly related to the average cumulative net sand fraction or
shaliness of the formation. Laminar sand porosity and water saturations are independent of shaliness or net sand. Excellent
correlation of log derived deconvolved data sets like porosity and water saturation with core measurements does not validate
the correctness of hydrocarbon pore volume estimation, which is linearly related to net sand. Accurate estimation of
hydrocarbon pore volume requires an accurate calibration of laminar sand/shale volumes, which can be precisely validated
only with lamcount data.

Introduction
Core analysis data provide a means to establish ground truth for the other formation evaluation measurements. It is essential
for the calibration of well log-based mineralogy, porosity and permeability and for the establishment of productivity in
situations where well test data are limited or unavailable. The literature is replete with examples of studies comparing and
combining log and core measurements. These studies can be divided into two types: those that use core data generally, as in
facies analysis, and those that directly compare core measurements with log measurements. Studies that compare log
measurements with core measurements can be further subdivided by how directly they deal with the problems of correlation.
Some ignore the problem and show numerous examples of excellent correlation. Some ignore the problem and show
substantial scatter. Some ignore the problem and show only average values. A few deal directly with the problem of log-core
correlation with varying degrees of success (Williams and Sharma, 1991).

The ideal geology for integrated core/well log studies is clean, consolidated, homogeneous, thick formations occurring over a
large areal extent with well-defined boundaries. Increasing divergence from the ideal causes increasing complications (Cooke-
Yarborough, 1987). Core plugs should be sufficiently large to provide a meaningful average of the pore structure of each
lithological unit or bed that is sampled. The logs should effectively resolve the beds from which core has been taken. The beds
are presumed to be laterally consistent in terms of physical properties, at least to the depth of investigation of the logging tools.
Bed thicknesses should be several times greater than the minimum needed for the logging tool to achieve parametric
resolution, i.e. to make a measurement that is as close to the truth as other environmental factors will allow, subject to perfect
tool operation (Worthington, 1991). Hamilton and Stewart (1983) discuss problems in log-core correlation and their approach
2 SPE 149017

to resolving them. They found thick, homogeneous zones allow excellent correlation. Thin, heterogeneous zones are a
problem. They also mention the physical problems inherent in core data collection when they suggest that correlations may
often be improved by rearranging the core data to suit log depths. The sampling and depth correlation problem is nicely stated
by Cooke Yarborough (1987). Misrepresentations of the information during core analysis can occur due to non-representative
core sampling (perhaps by over sampling of the zones with the best reservoir characteristics) and reversal or misalignment of
the core during recovery. The paper presents an approach to data reconciliation and an empirical guideline as to how close core
data can be expected to agree with log data.

It is the practice to assume correctly functioning systems both in the laboratory and in situ. Uncertainties in log data are caused
by the remoteness and the environment of tool measurement. Corrections for borehole effects, invasion effects, and bed
thickness all have inherent errors. The error associated with the bed-thickness correction is greater than that for borehole
correction. To overcome this problem, log readings should be chosen on plateaux where appropriate. Other errors are partly
accommodated in tying back to core (Worthington, 1991). Log data can be greatly improved by signal enhancement or
deconvolution processing (Dyos, 1986). In contrast, core data are seen as relatively accurate and precise (Hamilton and
Stewart, 1983).

Core measurements have advantage of having much higher

accuracy and precision and lower limits of absolute
detection than most of the logging devices. One of the key
points when studying cores is the comparison with the
logging results (even if only to calibrate them). This
comparison is only justified if the various results are
reconciled (Marchant and white, 1968, Olson, 1986, Dyos
et. al., 1986 and Enderlin et. al. 1991).

Deep water Tertiary reservoirs of Krishna Godavari basin

are generally very complex and heterogeneous (Fig.1).
There are many productive reservoirs whose beds are too
thin to be defined by any high resolution image log. Very
thin sand shale laminations with average thicknesses of 1cm
and 2cm respectively (Lamcount Data Summary Table of
an example well) are hard to be detected even on high
resolution image logs recorded in oil base muds and the
conventional logs do not give true property values. There
are two general log analysis approaches for the evaluation
of thinly bedded formations: Fig. 1: Location Map of Krishna-Godavari Basin, India

1) Identify bed boundaries using high-resolution data and attempt to resolve true log values in each thin bed through log
forward-modeling and inversion techniques.
2) Analyze the effects of thin beds on standard-resolution or interval-average log responses, and compensate for these effects
without resolving the individual beds (the low-resolution approach).

Table-1: Lamcount Data Summary of an example well

In a high resolution approach, the true properties of the sand beds can be obtained through core experiments and by running
logs such as high resolution image logs which can be calibrated to core and give good estimates of net reservoirs and a detailed
geological earth model forms the basis for inversion of the porosity and induction log for resistivity estimation used to derive
porosity and water saturation. In low resolution approach viz. laminated shaly sand analysis (LSSA), the volume of dispersed
shale and the total and effective porosities of the laminar sand fraction are generally determined by a Thomas-Stieber (1975)
volumetric technique using conventional logs. Water saturation is estimated from laminar sand-fraction resistivity derived
from advanced electrical anisotropy measurements (Mollison, et. al., 1999).
SPE 149017 3

In laminated shaly sand analysis, clean sand normalized porosity is independent of nearby shale beds or shale volumes.
Validation of petrophysical evaluation with core measurements makes the validation of porosity highly vulnerable due to its
small range variations with net sand volumes. Similarly, shale volumes can hardly be validated through depth wise core
measurements which are generally sand biased sampling. Gross error in the calibration of shale point can not be noticed if
porosity and shale volumes displayed along with core measurements. Core and log data calibration on laminar sand/shale
volume has to be reconciled for different vertical resolutions by the depth-averaging of core data, the signal-enhancement of
log data, or both. Core data inherently biased possibly due to preferential sampling of sand lithology can not be averaged/
filtered for the comparison purpose with log. More over, for the effective integration both core and log data should belong to
same parametric system i.e. for comparison, the log analysis must also be calculated in variables of the same type as those of
the core on data.

In this paper, it has been demonstrated that there are potential problems associated with the integration of petrophysical
evaluation and core data. Conventional shaly sand interpretation techniques commonly use the terms shale volume and clay
volume interchangeably and were developed to address the effects of dispersed clay in sandstones, rather than macroscopically
inter bedded sandstones and shales. Shaliness/clay fraction measurements on core like LPSA/XRD data are inherently biased
due to preferential sampling of sand lithology and represents deconvolved or deaveraged data set as in the case of sand
laminae properties. Shaliness derived from logs can not be compared with such core data because neither de-averaging of
lithology on log analysis nor averaging of most of the available core data is possible. Shaliness from LamCount data is
available at equal and regular spacing. It can be averaged with the filter length matching with logging tool devices to achieve
the precise calibration of net sand.

For the effective integration both core and log data should belong to same parametric system. The problem with the practice of
comparison of different parametric system is that it damages the credibility of core measurements - thereby justifying their
non-use - and some times develops the feeling that log analysis is systematically accurate compared with the direct
measurement, given the small size of the core sample. So, long as potential pit falls are understood and considered, a
meaningful integration is possible with core measurements.

Different parametric systems in thin bed log evaluation techniques and core data

Convolved and Deconvolved data sets

All the techniques of laminated shaly sand analysis for the evaluation of very thin bed reservoirs are generally based on some
form of inverse filtering/ deconvolution approach. Inferred sand laminae properties represent de-averaged values independent
of nearby shale laminae thickness or shale volume sensed by logging tools. Shaliness measurements on core, like LPSA data
are inherently biased possibly due to preferential sampling of sand lithology and represent the equivalent of deconvolved data
set. Shaliness as well as sand laminae properties measurements on core precisely represent the equivalent of deconvolved data
set.

In case of laminated shaly sand analysis, log derived sand laminae properties like porosity and water saturation represent
deconvolved data set equivalent to point wise core data measurements and can be directly compared with each other.
However, shaliness from log represents moving average data and can not be directly compared with core data. For comparison
of shaliness, either core data has to be averaged/ filtered with the filter length matching with logging tool devices or log data is
to be deconvolved to achieve a high degree of correlation. Core data inherently biased possibly due to preferential sampling of
sand lithology, can not be averaged for the comparison purpose with log. In such cases, shaliness can not be compared with
core data because neither de-averaging of lithology on log analysis nor averaging of core data is possible. Only core
measurements available at equal and regular spacing can be averaged and correlated with log data.

Shale and clay

In shaly sands, the rock matrix is composed of shale and quartz, where shale comprises clay, mica, feldspar, iron oxide, and
organics. Such lithology can be described in one of two approaches. If shale and quartz are the only lithology components used
in the analysis, then the approach is based on Shale-Quartz Analysis. If clay, silt and quartz or clay and sandstone are used in
describing lithology, then the approach is based on Clay-Sand Analysis. In Clay-Sand Analysis, clay comprises all clay
minerals, i.e., illite, montmorillonite, chlorite and kaolinite, which may exist in the formation; sand comprises quartz and silt;
silt comprises mica, feldspar, iron oxide and organics, which is shale exempt clay.

Most of the low resolution shaly sand thin bed evaluation techniques developed over years concerns themselves with shale
volume based on Shale-Quartz Analysis, but techniques of thin bed evaluation by inversion of low resolution logs using
accurate earth model or bed definitions from high resolution image logs (like SharpLite) are based on clay volume or Clay-
Sand Analysis. If inversion approach is considered for thin bed evaluation, clay needs to be estimated using special logs and
techniques. Commonly used terms ‘shale’ volume and ‘clay’ volume are different parametric systems in thin bed evaluation
techniques and can not be used interchangeably as in case of conventional shaly sand interpretation techniques.
4 SPE 149017

Shale volume estimation techniques are commonly based on interpolation between shale and sand points directly defined on
particular log respectively in shale beds (100 % shale) and clean sand (0% shale). It is more practical from log analysis point
of view and can be done deterministically. However, clay volume estimation is based on extrapolation technique as 100% clay
point can not be defined on logs even in shale beds which can not have more than 50-70% dry clay abundance. So,
extrapolation of 100% clay parameter is highly subjective and can not be accurately done without core measurements. In
addition it requires special logs. Dry-clay abundance can be estimated from Elemental Capture Spectroscopy log, ECS (Herron
et al., 1999), which may not be available in every formation. X-Ray Diffraction - XRD - and Fourier Transform Infra-Red -
FTIR - Spectroscopy can provide abundance of dry-clay from cores (Ruessink et al., 1992).

Quantitatively, shale volume can hardly be calibrated to core data. However, a geological description made by an experienced
geologist, on cores and cuttings from the formation, can help in comparing the log-estimated shale to the geologically-
described shale in the formation (AlRuwaili, 2005). Dry-clay abundance obtained from log analysis can be directly validated
against dry-clay abundance obtained from core-XRD or core-FTIR.

It should be noted that laminated shaly sand (LSSA) methodology is based on a ‘shale system’, defined by differentiating
laminar shale volume from total shale with the remainder being either structural or dispersed ‘shale’. Further, distinction of
clay porosity end points or ‘clay volumetric’ is undefined for this model. The key concept to this methodology lies in the
removal of laminar shale effects on the sand fraction using a total porosity system. Laminar shale does not occupy or alter the
intergranular sand porosity, but does change the sand shale ratio (net to gross). Concepts based on the shale volume fraction
have the disadvantage of being scientifically inexact with the result that they are open to misunderstanding and misuse. On the
other hand they are at least notionally applicable to logging data without the encumbrance of a core sample calibration of the
shale related parameter (Worthington, 1985).

Net pay and Net Sand

In a laminar sand shale sequence system, sand lamina thickness includes dispersed shale component and net sand pay
thickness of the reservoir interval is computed by simply removing the laminar shale component. So, precise estimation of
laminar shale thickness is very crucial for correct estimation of net sand pay thickness.

In hydrocarbon charged laminated reservoirs, net sand thickness is defined as net pay thickness weighted by laminar sand
fraction i.e. net pay thickness *VsandLam or net pay thickness *(1 – VshLam). Addition of equal amount of laminar shale to
the laminar sand will increase the net pay by a factor of two. Net sand is always less than net pay, which is opposite to the
usual conventional concepts, where net pay is a subset of net sand and derived after applying porosity and water saturation cut
offs on net sand.

Net pay is defined on conventional tool which does not differentiate laminar sand shale beds, where as net sand is defined on
high resolution tools which differentiate laminar sand and shale beds.

Relevance of calibrations of laminar sand porosity and water saturation with core data in laminated
reservoirs
In a laminated sand/shale sequence, depending on where core plugs were taken, the core porosities can be much greater (in
sands) or smaller (in shales) than the average total porosities (sand plus shales) derived from logs. Departure of log derived
porosities/shale volumes from core measurements may be noticed due to clean sand core plug positions in the sand/shale
sequences. This is why it is not possible to achieve perfect correlation quantitatively in laminated reservoirs.

All the techniques of laminated shaly sand analysis for the evaluation of very thin bed reservoirs are generally based on some
form of inverse filtering or deconvolution approach. Inferred sand laminae properties represent de-averaged values
independent of nearby shale laminae thickness or shale volume sensed by logging tools. Petrophysical evaluation calibrations
are frequently validated simply by a graphical presentation of a porosity and water saturation graph resulting from log analysis
involving a vertical seeding of more or less well adjusted core measurements. Such validation of deconvolved data set does not
ensure the correctness of the petrophysical evaluations using high as well as low resolution approaches. Sand laminae
properties i.e. porosity and water saturation tend to converge to respective values precisely selected from core measurements in
clean sand with very little uncertainty. In the absence or insignificant amount of dispersed shales in thin sand laminae this
tends to be uniquely defined maximum sand porosity value.

High resolution Approach (Inversion of Logs)

When laminar sand shale beds are thick enough to be identified by high resolution image logs, inversion of porosity and
resistivity logs using bed definitions i.e. earth model from image tool provides decovolved data set as in the case core
measurements.
SPE 149017 5

In case of very thin sand and shale laminae with thicknesses

well below the intrinsic resolution of OBMI tool, laminar
sand shale beds are seen as dispersed shaly sand. Porosity
and water saturations derived from the inversion of logs
represent the moving average i.e. convolved data set. These
values cannot be compared with the core measurement
equivalent to deconvolved data set which are indicative of
sand only porosity and water saturation. This is why most of
the core porosity values are higher than the respective
values from SHARPLite results with poor correlation
coefficient of 30% as indicated in Fig. 2. Porosities from
SHARPLite results are in lower range due to very thin
shale/ sand laminations beyond the resolution of OBMI tool.
These are seen as dispersed shale by the tool and leads to
the reducing the porosity.

Low resolution Approach (Laminated Shaly Sand

Analysis, LSSA) Fig. 2: Correlation of SHARPLite porosity with core

If a significant portion of the shale occurs as framework grains and/or laminations then reservoir porosity and water saturations
will be only slightly affected by the shale content particularly when the reservoirs are unconsolidated with total porosity values
similar to the nearby shale laminae. De-averaged sand laminae properties using low resolution approaches tend to converge to
respective values in clean sand properties particularly in the absence or insignificant amount of dispersed shales in thin sand
laminae. Both log and core derived porosity and water saturations representing deconvoved data set are directly compared or
correlated with each other. These deconvolved data sets are insensitive to net sand fraction or laminar shale volume as
indicated in Fig. 3.

Fig. 3: Correlation of Shale volumes and porosity Fig. 5: Correlation of laminar sand porosity with core
from core measurements.

In laminated shaly sand analysis, clean sand normalized porosity is independent of nearby shale beds or shale volumes and
makes the validation of porosity highly vulnerable due to its small range variations with net sand volumes.

Processed results of the example well calibrated to Lamcount core measured net sand thickness within a particular net pay
interval are indicated in tracks I, II and III of Fig. 4 along with core porosity and shale volumes. Results on last track are
presented with clay volumes from XRD calibrated as shale volumes. Results presented with wrong calibrations of net sands in
tracks II and III up to 100% errors are indicating excellent match and can be mis-leading in general for the correctness of
petrophysical evaluations with excellent correlation of log derived deconvolved data sets like porosity and water saturation
with core measurements. Quantitative correlation of laminar sand porosity with core measurements indicates large scatter as
indicated in Fig. 5. A correlation coefficient better than 55% could not be achieved due to different resolution associated with
de-averaging of log data and equivalent porosity measured on core samples. Laminar sand porosities are derived after
6 SPE 149017

Fig. 4. Correlation of deconvolved data set i.e. porosity and water saturation with variations of laminar shale / sand volumes.
Results indicated in Track-I are calibrated with LamCount with net sand of 57m. Results indicated on Track-II and III are
generated with an increased net sand of 95m and 114m respectively. All are showing excellent match of porosity with core.
SPE 149017 7

mathematically removing the effect of laminar shale beds seen by logging tools and it is free of shale effects. However, it is
not possible to have core plugs free of shale beds due to the presence of very thin laminations.

Calibration of multi-mineral derived dry clay volumes with XRD and ECS data

Estimation of dry clay volume from dry clay abundance

Dry-clay abundance obtained from log analysis can be validated against dry-clay abundance obtained from core-XRD. This
technique measures dry-clay abundance as weight-percent of the formation rock-matrix, Wdry-clay, which can be transformed
into volume-percent of formation bulk-volume using equation

Vol_dryclay=Wdryclay.Rhoma.(1−Фtotal)/ Rhodry_clay

Here, Фtotal is the formation total-porosity, Rhodry_clay is the dry clay density, which can be determined experimentally by
conducting XRD or FTIR analysis on representative formation rocks. Rhoma is the shaly-sand matrix density, which can be
obtained experimentally from core analysis on representative formation rocks. Alternatively, it can be estimated from
experience gathered about the formation and its prevailing minerals. Vol_dryclay, is the dry-clay abundance in volume percent
of the shaly-sand bulk-volume.

Calibration of neutron capture cross section (sigma) to dry-clay volume

First, the clay abundance data needs to be transformed from weight-percent to volume-percent. Neutron capture cross section
i.e. Sigma log has been used for dry clay volume computation as presence of feldspars in the sand does not allow the use of
conventional gamma ray log as clay indicator. Linear extrapolation of Logmax value corresponding to 100% clay volume on
Vclay – Sigma cross plot, most commonly used method for the computation of Vclay, could not be used due to sand biased
sampling core plugs considered for XRD measurements as indicated in the Fig.6.

Fig. 6: Correlation of Sigma log with Dry Clay Fig. 8: Correlation of SHARPLite results with Dry volume from
volume from XRD core measurements Clay volumes from XRD core measurements

Good correlation is not seen due to sand lithology biasing of the core plugs. The dry-clay volume data from XRD has been
utilized to calibrate their neutron capture cross section i.e. Sigma log to enable quantifying dry-clay volume by using method
described byCorrelation
Fig. 6: AlRuwaili,of2005.
Sigma log with Dry Clay
volumes from XRD core measurements
The first step of such normalization is displaying log on an appropriate scale and finding its lowest or minimum reading. Next,
the log readings need to be displayed from its minimum, Sigma_min and the dry-clay volume, Voldry-clay, (from ECS, XRD)
is to be displayed on a zero-to-one scale and on the same track. Then, the highest scale-point for the displayed log needs to be
changed in order to find a scale-point value that makes the core data and log curves most accurately coincide with each other.
Such scale-point can be called Sigma_max as indicated in Fig.7. Finally, equation below is used to obtain dry-clay volume
from the log.

Vclay = (Logreading - Logmin)/(Logmax - Logmin )

8 SPE 149017

Fig. 7. Calibration of XRD measured dry clay volume and LamCount shale volume with Sigma log. XRD calibrated net sand is 95m.
LamCount calibrated net sad is 57m.
SPE 149017 9

Dry clay volume derived from logs represent moving average data can not be correlated with equivalent sand lithology biased
XRD representing deconvolved data set as indicated in Fig. 8. Good correlation can only be seen only against clean sand beds
as indicated in Fig. 7.

Calibration of shaliness with LPSA core measurements

Correlation of shale volume from LPSA core measurements with a shale indicator in the field, sigma log, is presented in Fig.
9. Sigma log represents 3-4 ft average and small core plugs taken from sand beds varying from thick clean sand to highly shale
dominated laminated zones indicates, shaliness variation in a very small range and does not show correlation with log. Sigma
log is varying from minimum to maximum log readings in a small range shaliness i.e. 0 – 10%.

Fig. 9: Correlation of Sigma log with shale volumes Fig. 10: Correlation of shale volumes from SRES and LamCount
core measurements from LPSA core measurements

It indicates mostly sand lithology biased sampling of core plugs considered for LPSA measurements. Log derived shaliness
representing moving average or convolved data sets can not be compared with LPSA representing only small sand sample
plugs in laminar sand or equivalent deconvolved data set. Clean sand sample plugs from thick shale beds will be seen as shale
by logging tools.

Calibration of laminar shale/sand volumes of image tool with LamCount data

When laminations are beyond the image tool resolution, thickness computation becomes impossible as bed boundaries can not
be identified due to amalgamation of very thin sand shale beds. The image tool responds to the beds and features smaller than
the vertical resolution. Actual resistivity signal of the formation is weakened due to the vertical averaging characteristics of the
image tool, when formation properties are varying fast i.e. beds are thinner than the geometrical spacing of the sensors.
Laminated sand shale layers with thicknesses lower than the tool resolution will be seen as dispersed sand or shale in
proportionate to the individual component as in case of low resolution resistivity logs.

In a gas charged formation, with little invasion of oil base mud filtrate in the formation, variation of SRES can be taken as
function of shaliness and can be used for the computation of shale volume as in case conventional resistivity. It is very difficult
to achieve good correlation with Lamcount data even after filtering/ averaging of SRES data (Fig. 10). Large scatter in the data
is mainly attributed to the mismatch of the resolution of the data. Resolution of the SRES is 3 cm and LamCount data is
Fig. 9: Correlation
originally countedofmm/cm
Sigma log withbut
beds shale volumes
results have been reported by averaging of every 25 cm interval. Both data represent
moving averagefromdata
LPSA
setscore measurements
without differentiating bed boundaries and obviously can not be correlated as indicated on last track
of Fig. 7. In such case, only gross shaliness or net sand volumes in a particular zone computed with OBMI using binary
lithology technique (Bateman, 1990) can be compared in average sense with core reported net sand values. Comparison of
Lamcount data and SRES data for six wells where conventional cores are available in the intervals over 700 mts are presented
in Fig.11. Complete procedure to compute average net sand thickness in very thin beds using OBMI is described by Yadav, et
al, 2009.

Calibration of laminar shale/sand volumes of low resolution tools with LamCount data
Cross correlation of shale volume estimated from Sigma log and Lamcount data are presented in Fig. 12. High scatter of the
data indicates the mismatch of the bed boundaries seen by Sigma tool and Lamcount data. Resolutions of the Sigma and
10 SPE 149017

Lamcount data are 1 m and 1/4 m respectively. However, actual thicknesses of sand shale laminations are of the order of 1 and
2 cm respectively as indicated in Lamcount data summary table. In such situation averaging /filtering of the core data with
filter length matching with logging tool does not improve the correlation coefficient quantitatively. Only average or gross
shaliness/net sand from sigma log using LSSA can be validated with Lamcount data where laminar and dispersed shale
CO
components are estimated using REVsOBMI plot.
Thomas-Stieber

100

80

60
NETSAND(m)

40

20

0
1 2 3 4 5 6
WELLS

LAMCOUNT SRES

Fig. 11: Comparison of net sand from SRES and Fig. 12: Correlation of shale volumes from Sigma
LamCount core measurements and LamCount core measurements

Thomas-Stieber shale distribution

Total shale volume and total porosity from standard log is used in the Thomas-Stieber shale distribution model to determine a
laminar shale volume, Vlam, dispersed shale volume, Vdisp. The shale distribution triangle overlain on the Thomas-Stieber
plot ( Fig. 13), is defined by the three points, φmax, φsh, and φmin.

NET SAND
100.00
φmax
LSSA Net Sand (m)

80.00
φsh
60.00

40.00

20.00
φmin
0.00
0.00 20.00 40.00 60.00 80.00 100.00
Lam count Net Sand (m)

Fig. 13: Thomas-Stieber plot Fig. 14: Correlation of net sand from LSSA and
LamCount core measurements

The latter point, φmin, is defined by the volume of shale that can ‘fill’ the porosity of the clean sand such that Vsh =φ max,
and would have a total porosity equal to that volume of shale such that
φt = φ min = φ max ⋅φ sh.
SPE 149017 11

This point is referred as φmin because it is the lowest possible total porosity and the maximum dispersed shale volume that
can result from the initial parameters φmax and φsh in the Thomas-Stieber model. The resistivity anisotropy derived
horizontal and vertical resistivity together with user defined horizontal and vertical shale resistivity, are used to calculate the
laminar sand fraction resistivity, used in Archie’s, relationship . The cumulative net sand from LSSA is
CumNetSand = Σ((1 −Vlam ) ⋅ RLEVEL),

where RLEVEL is the level spacing (i.e., sample rate).

Excellent correlation have been achieved between average net sand from LSSA and LamCount core measurements as
indicated in Fig. 14, as all the points are falling closely on a 45 degree line. The reservoir summations for the LSSA
hydrocarbon·meters is calculated as

φ hc⋅m = Σ((1 − Swe ) ⋅φ esd ⋅ (1 −Vlam ) RLEVEL)

Variation of hydrocarbon pore thickness with average

laminar shale volume in a particular zone of different
wells is presented in Fig. 15. It indicates that hydrocarbon
pore thickness is almost linearly related with average
laminar sand volumes. Wells dominated with thick sand
beds are indicating slight departure from the linear trend
occupying upper portion where HCPV is varying from
55/ 70 – 100% (Orange Curves). This is probably due the
presence of dispersed clay/shale in the thick beds. Wells
dominated with thin to very thin beds are indicating an
excellent linear trend with HCPV varying from 25/35-
100 % (Blue Curves).

Discussions and Conclusions

Integration of the petrophysical results with core data
provides the ground truthing for well-log analysis.
Historically, log responses have been tied back to core
data with little attempt being made to take account of the
different scales of measurement. More recently,
however, there has been a renewed interest in redressing
this situation, even though it is recognized that a full
reconciliation of core and log data might not be possible,
because of the different volumes of rock sensed by these
Fig.15: Variation of hydrocarbon pore thickness with average
two types of measurement (Al-Ali and Worthington, laminar sand volume in different wells of the field
2005).

Log and core data simply relate to different volumes of rock constrained by measurement design and require slightly different
interpretations. The averaging of core data can be designed to reflect the response curve of the corresponding logging tool
(Barlai, 1979) by using appropriate weightings. This refinement might increase the physical significance of the resulting mean.
In heterogeneous and/or thinly bedded formations: for such cases a strip-sampling approach has been proposed (Dahlberg and
Fitz, 1988). This involves the strip-sampling of core over a distance that approximates the spatial resolution of the logging
tool. This technique involves destructive testing. It allows the sample to be mixed over the length of the strip. However, it is
not standard practice (Worthington, 1991).

Core analysis forms a basis for the calibration and verification of log analysis in the evaluation of petrophysical parameters in
laminated reservoirs. For the petrophysical evaluation in a thin-bedded reservoir, conventional core analysis data serve three
broad objectives. First, the core data are used to establish a petrophysical model comprising a complete set of petrophysical
bed types. Second, the core data are used to characterize the average petrophysical properties of each bed type. Third, the core
data are used to determine the macroscopic properties of continuous cored intervals (e.g., net reservoir thickness, average
porosity, and average water saturation) for direct comparison with the results of log analysis (Passey et al, 2006).

The logs should effectively resolve the beds from which core has been taken. The beds are presumed to be laterally consistent
in terms of physical properties, at least to the depth of investigation of the logging tools. Where they are not laterally
homogeneous, core-log reconciliation will show more pronounced residual departures (Enderlin et al., 1991) and, in extreme
cases, core and log data might be poorly correlated (Marchant and White, 1968).
12 SPE 149017

In a conventional shaly sand interpretation is generally based on dispersed model where porosity and water saturation are
highly dependent on shaliness and sufficient variations are noticed to have a meaningful comparison of core and log data.
Validation of porosity and water saturations ensures the correctness of petrophysical evaluation results as it pertains to bulk
sand property which includes shaliness. This concept can not be extended in case of laminated reservoirs where laminar sand
(only) properties are independent of shaliness. Porosity measurements are also reported on core plugs from only from sand
beds. Validation of porosity and water saturation does not ensure the correctness of formation evaluation results as it pertains
only to the sand laminae. Core measurements from laminar sand beds ensure the validation of dispersed clay components
rather than laminated clay component generally required for laminated shaly sand analysis.

For comparison, the log analysis must also be calculated

in variables of the same type as those of the core on data.
Even in case of similar parametric system different
volumes sensed by log and core does not allow the
agreement between log and core measurements. This
aspect is generally ignored while presenting few sand
lithology biased core data which may be showing good
match against clean sand intervals. This is very common
practice while presenting LPSA and XRD as core data
validation in laminated reservoirs. Comparison of
lithology biased data does not ensure the correctness of
log derived values computation in the higher range of
shaliness/clay volumes.

Laminated shaly sand analysis (LSSA) of very thin

reservoirs of KG basin is presented by Tyagi, et al., 2008
(Fig.16). “Sigma curve was used for Vclay estimation.
Porosity, clay volume and saturation were calibrated to
core measurement”. ‘Clay volumetric’ is undefined for
LSSA model which is based on Shale-Quartz model,
therefore clay Sand model used by the authors in LSSA
methodology is not consistent with basics. Calibration of
Vclay estimation with sand lithology biased XRD core
measurements does not validate the correctness of
hydrocarbon pore volumes in LSSA. Processing results
presented with excellent correlation of deconvolved data
set is also not relevant as far as validation of porosity and
water saturation is concerned (as demonstrated in Fig. 4).
By using clay instead of shale, net sand or hydrocarbon
pore thickness is being over estimated approximately by
50% as calibration of 100% shale as 70% clay will result Fig.16. LSSA Porosity, clay volume and water saturation
in a 50% overestimation of net sand as indicated in Fig.7. were calibrated to core data by Tyagi, et al 2008.

Lamcount data is an external visual measurement of shale and sand laminae. To achieve the accurate petrophysical evaluation
both net reservoir and net sand results are required to be calibrated across the entire spectrum of shaliness. Theoretically, net
reservoir and net sand are cumulative summation of reservoir (estimated by removing 100% shale beds seen by the log used
for Vsh estimation) and sand laminae thicknesses, without applying any Vsh cut off. Cutoffs on properties of sand laminae viz.
porosity and water saturation can not to be introduced while comparing laminae thicknesses with lamcount data. Producibility
of sand laminae is derived from internal bulk volume seen by petrophysical measurements, which can not be visualized
externally on core. So, calibration of net pay/ net reservoir and corresponding net sand values derived after applying porosity
and water saturation cutoffs, are not to be tied with Lamcount data. At the most some static porosity and saturation cut offs can
be applied for petrophysical non reservoir definition which is not contributing to the cumulative hydrocarbon pore thickness
computation.

Another challenge is that of scales. All the sources of data represent different volumes of rock. This is a problem, which
always occur in laminated reservoirs when logs are calibrated to core data and when they are calibrated to test information.

Petrophysicists often work on presumption that core and log data must agree. This approach is seldom successful in case
laminated reservoir evaluation due to different parametric systems of log and core data measurements. Porosity data can be
matched only with some less or greater than relationships depending on the respective total porosities in sand shale laminae.
For example, deconvolved total porosity data from logs should be always greater than core derived total porosity affected by
SPE 149017 13

shale with lower total porosity. Both data can show a perfect match only in sand shale laminae with equal total porosity. The
solution to the data integration problem is one of fully understanding the petrophysical attributes of the formation. This can be
accomplished through careful evaluation of data acquired from field and laboratory. So, long as potential pit falls are
understood and considered, a meaningful reconciliation is possible with different type of core measurements.

Acknowledgments
Authors are thankful to the management of Reliance Industries for giving the permission to present and publish the paper.
Authors are also thankful to Shri I. L. Budhiraja, President, E & P, Reliance Industries Limited for encouraging extensive
work on the laminated shaly sand analysis of thin beds and its reconciliation with different types of available core
measurements.

Nomenclature
OBMI: Oil Base Micro Imager Tool, M/s Schlumberger
Elan: Expert Log Analysis, M/s Schlumberger software application.
SHARPLite: Sharpening logs to FMI/OBMI resolution, M/s Schlumberger software application.
SRES: Pad scaled resistivity from FMI/ OBMI.
LamCount: Lamination Count Petrography Core Analysis.
NTG: Net to Gross (Sand Volume Fraction)
Vshlam: Laminated Shale Volume Fraction
Vsandlam: Laminated Sand Volume Fraction
Vsh /Vshale: Shale Volume in Fraction
LSSA: Laminated Shaly Sand Analysis

References
Al-Ali, H. A. and Worthington, P. F., 2005, Application of petrophysical scale reconciliation to Saudi Arabian Reservoirs,
International Petroleum Technology Conference, Doha, Qatar, 21-23, November, IPTC-10726

AlRuwaili, S. B., 2005, “Accurate Petrophysical Methods for Shaly Sands Evaluation”, in Transactions of the SPWLA 46 nd
Annual Logging Symposium, June 26-29, Paper AAA.
Barlai, Z. 1979. Some theoretical aspects of how to correlate well-log-evaluated data to core determined reservoir properties.
Trans. SPWLA 20th Ann. Logging Symp, E1-20
Bateman, R., 1990, “Thin-bed Analysis with Conventional Log Suites”, in Transactions of the SPWLA 31st Annual Logging
Symposium: Society of Professional Well Log Analysts, Paper II, 24 p.
Cooke-Yarborough, P. 1987. A method for reconciliation of log-derived and core-derived porosity, Trans. SPWLA 28th Ann.
Logging Symp.; DD1-24

Dahlberg, K.E. and Fitz, D.E. 1988. Comparing log-derived and core-derived porosity and mineralogy in thinly-bedded
reservoirs: an integrated approach, Trans. SPWLA 29th Ann. Logging Symp, W1-18

Dyos, C.J. 1986, Inversion of well log data by the method of maximum entropy, Trans. SPWLA 10th Eur. Form. Eval. Symp,
H1-15

Dyos, C.J., Petler, J.S., Jones, M.R.O., Cuddy, S. and Wilkinson, D. 1988, Reconciliation of Sw from logs and core in the
North Sea Magnus Jurassic sandstone reservoir, Trans SPWLA 11th Eur. Form. Eval. Symp, D1-12

Enderlin, M.B., Hansen, D.K.T. and Hoyt, B.R. 1991. Rock volumes: considerations for relating well log and core data. In:
Reservoir Characterisation II (Lake, L.W., Carroll, H.B., Jr. and Wesson, T.C.: Eds.), Academic Press, NY, 277-288.

Hamilton, J.M. and Stewart, J.M. 1983. Thin bed resolution and other problems in matching log and core data. Trans. SPWLA
24th Ann. Logging Symp, 11-13

Herron, Susan L. and Herron, Michael M., 1999, Quantitative Lithology: An Application for Open and Cased Hole
Spectroscopy. SPWLA, 37th Annual Logging Symposium, New Orlens, June 16-19.

Marchant, L. C. and White, E.J. 1968. Comparison of log and core analysis results for an extremely heterogeneous carbonate
reservoir. Trans. SPWLA 9th Ann. Logging Symp, L1-16
Mollison, R., Schön, J., Fanini, O., Kriegshäuser, B., Meyer, H., and Gupta, P., 1999, “A model for hydrocarbon saturation
14 SPE 149017

estimation from an orthogonal tensor relationship in thinly laminated anisotropic reservoirs,” paper OO, SPWLA 40th
Annual Logging Symposium Transactions.
Olson, D. M. 1986. Calibration of log and core saturation data: case history from the San Ardo Field. Trans. SPWLA 27th
Ann. Logging Symp, Z1-12
Passey, Q. R., Dahlberg, K. E., Sullivan, K. B., Yin, H., Brackett, R. A., Xiao, Y. H. and Guzmán-Garcia, A. G., 2006,
“Petrophysical Evaluation of Hydrocarbon Pore-Thickness in Thinly Bedded Clastic Reservoirs”, AAPG Archie Series,
No. 1, American Association of Petroleum Geologists, Tulsa, Oklahoma, U.S.A.
Ruessink, B. H. and Harville, D. G., 1992, Quantitative Analysis of Bulk Mineralogy: The Applicability and Performance of
XRD and FTIR. SPE 23828, February 26 – 27.
Thomas, E. C. and Stieber, S. J., 1975, The Distribution of Shale in Sandstones and its Effect upon Porosity, in Transactions of
the SPWLA 16th Annual Logging Symposium Transactions, Paper T.
Tyagi, A. K., Bastia, R. and Das, M., 2008, Identification and Evaluation of the Thin Bedded reservoir Potential in the East
Coast Deep Water Basins of India, 7 th International Conference and Exposition on Petroleum Geophysics, SPG,
Hyderabad.
Worthington, P. F., 1985, “The Evolution of Shaly-Sand Concepts in Reservoir Evaluation”, Log Analyst, Vol. 26, No. 1, Jan.-
Feb., pp. 23-40.
Worthington, P. F., 1991, Effective integration of core and log data, SCA Conference Paper no. 9102

Williams, G. R. and Sharma, M.M., 1991, Quantification of Log-Core Correlation Using the Formation Microscope, SPE
Annual Technical Conference and Exhibition, 6-9 October, Dallas, Texas
Yadav, L., Dutta, T., Kundu, A. and Sinha, N., 2009, “A New Approach for the Realistic Evaluation of Net Sand Pay on
Image Log in Very Thin Reservoirs of Krishna Godavari Basin, East Coast of India”, in Transactions of the SPWLA 50th
Annual Logging Symposium Transactions, Paper J.

View publication stats