SPE-176302-MS

Thinly Bedded Reservoir Study, Application of Sand-Slit-Clay

(SSC)-SHARP-Thomas Stieber Juhasz Models, in a Deep Water Field,
Offshore Sabah, Malaysia
Budi P. Kantaatmadja, Syamsina Bt. Rashid, Achmad A. Nurhono, and Rahim Masoudi, PETRONAS;
Tom Neville, Kim Taesoo, and Debnath Basu, Schlumberger

Copyright 2015, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE/IATMI Asia Pacific Oil & Gas Conference and Exhibition held in Nusa Dua, Bali, Indonesia, 20 –22 October 2015.

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
Thinly bedded reservoir study in the deep-water area, offshore Sabah, Malaysia, was performed with the
primary objective of improving the understanding of its complex geology. The nature of reservoirs, which
are predominantly thin-bed and laminated sandstones of submarine fan environment, contain a high level
of uncertainty in its lateral continuity.
Standard shaly-sand log analysis methods contribute pessimistic values of porosity and water saturation
when applied to these reservoirs. Few techniques are then presented for the determination of these rock
properties, which are more reliable with core and production data.
Core grain-size analysis of these reservoirs shows that clay content is generally low but the silt content
can be significant. Furthermore, log responses show that porosity distribution and mineral-conductivity
are influenced mainly by the silt-size particles.
A sand-silt-clay (SSC) model was then developed from density-neutron crossplot, which model is also
used to determine porosity and water-saturation in addition to volumes of lithology components of the
reservoirs.
Furthermore, other petrophysical technique, called SHARP, uses 1D convolution filters to match thin
bed modelled log curves to their corresponding measured responses. A petrophysical evaluation using
standard resolution logs and the thin bed resistivity (SRES) from image response are used to develop a
thin bed model that yields high resolution logs.
For zones where the resistivity image indicates significant thin bed development, the standard
petrophysical analysis should also indicate the existence of free fluid. Although the porosity tools cannot
resolve the thin beds, they nevertheless represent the bulk volumetric over the interval, known as
Thomas-Stieber-Juhasz (TSJ) method, and would be able to differentiate between porous zones with lower
clay volume versus porous shales with high clay volumes. The main point is that if a thin bed interval has
some calculated free fluid volume using standard resolution logs, then a thin bed analysis is warranted.
2 SPE-176302-MS

Introduction
As most of reservoirs in the field are thinly-bedded sands and all of the drilled wells have standard
resolution logs, the TSJ evaluation model using well-logs in standard resolution was integrated with the
in-house SSC evaluation model using the inputs in standard resolution logs as well as in high resolution
and the other petrophysical software SHARP-ELAN evaluation model using well-logs in higher resolu-
tion, in order to properly characterize the thinly bedded reservoirs.
The hydrocarbon pore volumes can be often underestimated when the sand beds are thinner than the
vertical resolution of log measurements.
The validated well logs and evaluation results are then used by the subsequent geoscience teams in rock
physics for a classification of rock and fluid types, geo-modelling for facies and rock property mapping,
stochastic seismic inversion and reservoir simulation.
The intention of utilizing three (3) different petrophysical evaluation models, on all exploration and
appraisal wells of about 80 wells and using the inputs in standard resolution logs as well as in high
resolution, is to establish the basis of understanding and comparison among different petrophysical
evaluation methodologies commonly used in Malaysia.
The SHARP analysis was done in 11 oil based mud image (OBMI) wells including the squaring
process for the other log curves such as RHOB (bulk density), NPHI (neutron porosity), and Sonic logs
using the bed boundaries created during the initial SHARP analysis on image logs. The SHARP analysis
uses 1D convolution filters to match the measured resistivity responses of image logs and generate the
shallow resistivity (SRES) of bed boundaries model. These bed boundaries were verified with the two (2)
available whole cored-wells in the field. Using the bed boundaries from image response, it was used to
develop a thin bed model that yields high resolution logs.
The sharpened log curves were significant inputs for generating rock physics and SSC evaluation
models using inputs in high resolution logs, to differentiate the better various rock (litho-facies) and fluid
types.

Thomas-Stieber-Juhasz and SHARP-ELAN Models

As most of the wells have standard resolution logs, the TSJ model was used to evaluate these standard
resolution logs with results including continuous net-to-gross (NTG) sand, total sandstone porosity, net
pore volume (NPV), and water saturation (Sw) model using Skelt Harrison approach that was verified
with cores and field production history rather than relying on resistivity driven saturation profiles. The
resistivity logs in thin beds are generally suppressed by the adjacent shale beds resulting in significantly
pessimistic water saturation and hydrocarbon pore volumes.
The same logic was applied in porosity calculation too, the sandstone porosity was estimated from the
effective porosity (as the clay-free) vs. volume of shale crossplot, and compaction corrected total porosity
in sand beds. Figure-1 shows the schematic end-points in TSJ evalution where the sandstone porosity
(PORSND) and net-to-gross sand value (NTGSD) are estimated on each data point or depth frame. It is
assumed that the effect of structural clay is negligible in this environment, and the main clay response is
in the form of lamination.
SPE-176302-MS 3

Figure 1—TSJ Illustration and End Points

Figure-2 shows the example of TSJ evaluation plot, where all input logs and TSJ outputs are displayed
together for comprehensive well data analysis. This type of plot is generated on every well where TSJ
evaluation results are made available. In this example, we compared the TSJ results with SHARP outputs,
the blue curves in the tracks: total porosity (PHIT), effective porosity (PHIE), and volume of shale (VSH))
and total water saturtion (SWT) in shaded-orange are coming from the SHARP-ELAN analysis. The
combined lithology and fluid volumes are presented in the second track from right for SHARP-ELAN, and
in the first track from right for the TSJ results.

Figure 2—TSJ-SHARP Comparison (AA-1 well)

4 SPE-176302-MS

The overall results are somewhat comparable between the two but we can see that SHARP is more
effectively resolving the individual thin beds. The overall sum of hydrocarbon pore volume (HPV) may
be different between the two (2) evaluation techniques, where SHARP is identifying more of sand
thickness including location of the individual thin beds and providing additional information on reservoir
geometry. This information can be useful in various other applications such as rock physics modelling and
completion design among others.
Another conclusive observation from this comparison is that SHARP PHIT and SWT curves are
matched with the TSJ PORSND and SW_SKELT_EXP within the thinly-laminated sand in this cored
well.

Sand-Silt-Clay Model
The in-house SSC evaluation model was also carried out on all exploration and appraisal wells in order
to establish the comparison between different thinly bedded evaluations methods. This SSC model is the
petrophysical evaluation model using RHOB and Neutron cross-plot in order to develop the sand-silt-clay
lithology model and is typically run on the standard resolution of triple combo conventional logs.
The SSC evalution model used two sets of log-inputs: standard resolution and the enhanced resolution
logs from squared SHARP analysis; and, the results were compared with TSJ and SHARP-ELAN results.
Figure-3 shows clearly improved lithology grouping when using the enhanced resolution input curves.
The silt line (the middle line on the cross-plot that separates sand points within left side areas and clay
points within right side areas) divided many points of thin beds positioning on the sand areas when using
enhanced resolution input curves. This illustrates that there is shoulder bed effects on the standard logs
when measuring across the thinly and laminated sand beds.

Figure 3—SSC Lithology Plot Comparison on H145 Reservoir (AA-2)

Figures-4 and 5 show the comparison of SSC evaluation with TSJ and SHARP-ELAN results. It shows
that the SSC with enhanced resolution log inputs produced similar results as the SHARP-ELAN, both
characterizing the thin bed reservoirs thoroughly.
SPE-176302-MS 5

Figure 4 —SSC-TSJ-SHARP Comparison on H110 Reservoir (AA-1)

Figure 5—SSC-TSJ-SHARP Comparison on H145 Reservoir (AA-1 well)

On the volumetric summation, Figure-6 shows the recommended cutoff values in different method-
ologies in order to produce the comparable HPV. After some iteration of experiments, below is the
summary of reservoir cutoff values to be considered in thin beds;
6 SPE-176302-MS

• TSJ: 90% VSH (⫽10% NTGSD)

• SSC-SHARP: 75% VSH (⫽ vclay ⫹ Vsilt)
• ELAN-SHARP: 55% VCL
• SSC-Standard: Clear cutoff not established, probably up to 90% VSH

The SSC work has been further expanded for litho-facies classifications, using the Vclay and Vsilt
cutoff for acquiring initial litho-facies outputs from Neural Network as references. The litho-facies cutoff
parameters in SSC were adjusted iteratively until the litho-facies outputs from SSC matched closely with
the ones using TSJ outputs in annotated neural network (ANN) environment.

Figure 6 —Comparison of Reservoir Summation Cutoffs (SSC-TSJ-SHARP)

The initial work was done using the outputs from SSC-SHARP, and the results showed clear separation
of litho-facies group as illustrated in the right side of plot in Figure-7. However, the clouds became rather
shifted and blurred in SSC-Standard case where the thin bed points are comingled around the silt line. The
colors assigned on each facies group are;

• Yellow : Thick sands

• Gold : Rich Heterolithic sands
• Purple : Poor Heterolithic sands
• Black : Shales
SPE-176302-MS 7

Figure 7—Comparison of Facies Classification, Standard vs SHARP Inputs (AA-1 well)

Figure-8 shows the facies classification comparison between the results from SSC-Standard logs and
TSJ-ANN. The cloud profile is a bit different in TSJ-ANN case, especially for the Poor Heterolithic sands,
without full extension into the low porosity area. This is because of the additional reservoir porosity cutoff
(14%) applied during TSJ computation, which mainly affected the discrimination between Poor Het-
erolithic sands and non-reservoirs. Different facies cutoffs for SSC-Standard logs in H-140 and H-145
reservoirs are required in order to get comparable match with TSJ-ANN facies.

Figure 8 —Comparison of Facies Classification, SSC Standard vs TSJ Inputs (AA-1 well)
8 SPE-176302-MS

Table-1 summarizes the comparison of core lithology fraction vs SSC-SHARP, and TSJ-ANN vs
SSC-Standard facies fractions in AA-1 well.

Table 1—Comparison of Facies classification; Core vs SSC-SHARP, TSJ-ANN vs SSC-Standard (AA-1)

The computed facies results under SSC are illustrated in Figure-9 for H-145 reservoir and Figure-10
for H-110 reservoir. Some subtle differences were observed between SSC and TSJ-ANN results,
especially in classifying Rich vs Poor Hetetrolithic sands. Nonetheless, the overall sensitivity of discrim-
inating reservoirs from non-reservoirs seems acceptable in both methods when the results are compared
to image logs and core photos displayed together.

Figure 9 —Comparison of Facies Classification in H-145 (AA-1 well)

SPE-176302-MS 9

Figure 10 —Comparison of Facies Classification in H-110 (AA-1 well)

Image Log Analysis

There were more than 10 wells with wireline OBMI resistivity images and more than 60 wells with
logging while drilling (LWD) azimuthal density (ADN) images. All image data were semi-static
normalized in petrophysical software from raw data to produce the maximum image contrast from top to
bottom of the interval, using the same scale as much as possible. The scale used is 1-100 ohmm for OBMI
image logs and 2.0-2.6 g/cc for density image logs. Then the image display on each well was captured in
PNG and CGM formats so that they can be loaded for visual display in geomodelling software.
The image data were also used in petrophysical analysis to visually correlate the sand bed intervals with
TSJ net-to-gross sand curve (NTGSD) and bed boundaries defined in SHARP analysis. In general the
match was quite good in terms of bed thickness and depth locations, providing good support and
confidence on the petrophysical evaluation workflow in TSJ and SHARP.
Figure-11 shows the OBMI images in H-136 reservoir correlating well with the sand beds identified
from other well log analysis. The white color indicates resistive sand beds filled with oil. Figure-12 shows
the side-by-side image comparison between OBMI and Azimuthal Density, acquired in the same borehole
BB-1 well. Both are correlating well with each other in picking up the sand beds.
10 SPE-176302-MS

Figure 11—OBMI Images with Petrophysical Analysis (AA-1 well)

Figure 12—OBMI-Azimuthal Density Image Comparison (BB-1 well, H-110)

Nuclear Magnetic Resonance (NMR) Analysis

The NMR logs were acquired in 10 wells in this field; but, they were not all of those wells contained
sufficient NMR data coverage across the thin bed reservoirs. This is understandable considering the fact
that the main target reservoirs were the thick sands of H-150 during the initial exploration and appraisal
period. Therefore, the evaluation is focused on the exploration AA-1 well to see how NMR data can be
used in thin bed analysis, as also this well has the core data and other well-log types.
SPE-176302-MS 11

The NMR logs were processed without any vertical stacking in order to achieve the highest vertical
resolution of the tool measurement (7.5 inches) from the field DLIS files. The data quality was reasonably
good with acceptable S/N (signal-noise) ratio due to good borehole condition and OBM mud environment.
The results are shown in Figure-13 for H-145 reservoir, all NMR outputs are displayed in red color.

Figure 13—CMR-TSJ Comparison within H-145 intervals in AA-1 well

➢ k-ECHO_VHR was computed from the raw echo data as the high resolution permeability, the
vertical resolution is 7.5 inches. We were interested if it could be used as the NTG indicator in thin
beds, and the actual results showed good correlation with the image logs in higher resolution than
the TSJ NTGSD curve. Thus, it requires permeability cutoff values in each reservoir unit in order
to achieve the comparable NTG ratio when compared to NTGSD of TSJ.
➢ NMR total porosity (TCMR) was reading lower than the core porosity due to the effects from
adjacent shale beds within H-145 interval, also probably due to the lack of polarization in H-150
thick sands.
➢ Sw-irr (SWI_VHR) is the qualitative indicator of hydrocarbon saturation when the reservoir is in
irreducible condition. Similar to TCMR, it is affected by the shoulder beds in thin sands but is
clearly enhanced than the Archie saturation. It picks up the sand beds when the sand thickness is
greater than its resolution.
➢ BFV_VHR (bound fluid volume) shows the overall bound fluid volume.
Figure-14 shows the same results in H-110 sands where NMR response is correlating well with the
individual sand packages.
12 SPE-176302-MS

Figure 14 —NMR-TSJ Comparison within H-110 interval in AA-1 well

Another capacity from NMR analysis, it was the pore size analysis using T2 data which can be related
to grain size distribution. This analysis can be used by completion engineer for designing its completion-
mesh size, as the sand production may be occurred during oil production that comes from the finer grain
size within the thinly-bedded intervals. Figure-15 shows the comparison of pore size distribution between
H-150 thick sands and H-145 thin sands, whereas Figure-16 shows the same comparison between H-145
and H-110. The latter comparison was made due to the higher permeability cutoff value was used in H-110
sand in order to match the NTG of TSJ.

Figure 15—Comparison of Pore Size Distribution; H-150 vs H-145 (AA-1 well)

SPE-176302-MS 13

Figure 16 —Comparison of Pore Size Distribution; H-145 vs H-110 (AA-1 well)

From this NMR analysis, it can be seen clearly that H-150 contains relatively larger pores compared
to the other sands even in deeper depth intervals, and minimal clay bound fluids. Furthermore, it is shown
that the H-110 interval contains more medium size pores than the H-145 interval due to less compaction.

Production Log Analysis

There were more than 10 wells available with PLT logs and their analysis reports, but most of them were
injection wells. And, for this study, it was used only the production wells for identifying fluid contribution
from the thinly-bedded intervals.
Figure-17 (and its table below it) shows the oil rate from the H-110 thin beds in BB-2 well. In
Figure-18, the fluid velocity profile derived from spinner interpretation shows clearly the zonal contri-
bution from each individual thin beds. And, it also shows that oil production contribution from the whole
H-110 reservoir is increased when the drawdown pressure is increased significantly, implying that H-110
thin sands are sensitive to the drawdown pressure during production.
14 SPE-176302-MS

Figure 17—H-110 Zonal Contribution at Different Flowrates (BB-2 well)

Figure 18 —PLT Spinner Profile and Zonal Contribution (BB-2 well)

SPE-176302-MS 15

The profile is slightly different on BB-3 well. The relative contributions from H-130 and H-136 sands
show somewhat less sensitivity to different flowrates. (Figures-19 and 20)

Figure 19 —H130-H136 Zonal Contribution at Different Flowrates (BB-3 well)

Figure 20 —PLT spinner profile and zonal contribution (BB-3 well)

16 SPE-176302-MS

Cutoff Discussion
The cutoff value is selected by comparing the HPV between SHARP (SSC or ELAN) and TSJ. Figure-6
above shows good match at the well level when using TSJ VSH cutoff of 90%.
Figure-21 shows the comparison of HPV in reservoir H-145. The green column is the reference HPV
from SSC / ELAN_SHARP. With same VSH cutoff 75% (or 55% for ELAN) on TSJ (purple column),
HPV is too low as expected. HPV in blue column is higher when TSJ VSH cutoff is increased to 90%.
However HPV in static model (orange column) is very close to SHARP because the NTG (net to gross)
is populated as a variable.

Figure 21—Comparison of HPV in Reservoir H145

Conclusions
a. SSC analysis and other thin bed evaluation techniques (SHARP-ELAN and TSJ) are comparable
to each other when using cutoff parameters for optimizing reservoir properties. The substantial
hydrocarbon pore volume will be more robust when the evaluation techniques are integrated
together.
b. Some advanced NMR analysis were also investigated for improving thin bed reservoir character-
ization using well-data from the cored wells, with positive implications on NTG verification. It
also can provide the valuable inputs for completion design when NMR log is used to analyze on
pore and grain size distribution which afterward is verified with cores.
c. PLT flowrate analysis on some production wells showed that the thinly-bedded reservoirs may
deliver substantial contributions to the total oil production but may also be sensitive to the fluid
flowrate due to pressure drawdown changes.

References
1. Estimating the Intrinsic Permeability of Clastic Sediments from Geochemical Data: SPWLA 29th
Annual Symposium, London, June 1987, paper HH.
SPE-176302-MS 17

2. Evaluation of Low-Resistivity-Pay Deepwater Turbidites Using Constrained Thin-Bed Petro-

physical Analysis: Bastia, R., Tyagi, A., Saxena, K. (Reliance Industries Ltd), Klimentos, T.,
Altman, R., Alderman, S., Bahuguna, S. (Schlumberger) SPE 110752
3. A Comparison of Laminated Sand Analysis Methods – Resistivity Anisotropy and Enhanced Log
Resolution from Borehole Image: Michel Claverie, Hisham Azam and Richard Leech (Schlum-
berger), Graeme Van Dort (Murphy Oil Co. Ltd), PGCE, November 2006, Kuala Lumpur.
4. Forecasting the Productivity of Thinly Laminated Sands with a Single Well Predictive Model:
Saifon Daungkaew, Michel Claverie, Boon Cheong, Steve Hansen, Richard Leech, Mohd Nor
Hisham Mohd Azam, Edna Malim; Schlumberger, Mohd. Reza Lasman and Renas Witjaksana;
Murphy Oil Company, SPE Asia Pacific Oil & Gas Conference and Exhibition (APOGCE), Perth,
Australia, 2008.
5. Successful Application of Thin Bed Petrophysical Evaluation Workflow in Deep-water Turbidite
Environment: Case Studies from Fields Offshore Malaysia: Budi Priyatno Kantaatmadja, Azizul
Mah Hassan, PETRONAS MPM; Mohd Nor Hisham Mohd Azam, Mohd Abd Rasheed Mohd
Abdul Rahman, Richard Leech, Schlumberger, Petroleum Geology Conference & Exhibition
(PGCE), Kuala Lumpur, Malaysia 2010.
6. The Distribution of Shale in Sandstones and Its Effect Upon Porosity: E. C. Thomas & S. J.
Stieber, SPWLA 1975
7. An Advanced Evaluation Method for Laminated Shaly Sands Including Uncertainty and Sensi-
tivity: Johan van Popta and Paul Hofstra, and Steven van Houwelingen, SPWLA 2004
8. Using Nuclear Magnetic Resonance data for grain Size Estimation and Expandable Sand Screen
Design: Tim Conroy and Kapil Seth, SPWLA 2010
9. Log Interpretation in The Malay Basin: A. Kuttan, C. P. Stockbridge, H. Crocker and J. G.
Remfry, SPWLA 1980
10. Formation Evaluation of Fresh Wwater Shaly Sands of The Malay Basin, Offshore Malaysia: R.
Grant Heavysege, SPWLA 2002