SPE 64477

Real-Time Data Analysis While Drilling Provides Risk Management for Both Geological
and Geometric Uncertainties in the Sotong K2.0 Reservoir
Md Zin Che Lah, Lancelot Sering, Azlee Abu Bakar,Petronas Carigali; Eva Sundal and Jan Daudey, Schlumberger

Copyright 2000, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the SPE Asia Pacific Oil and Gas Conference and
Exhibition held in Brisbane, Australia, 1618 October 2000.
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, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300
words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract
Modern communication systems can transmit data from the
wellsite to shore in real time, allowing full utilization of
advanced drilling technology, 3D seismic and geological earth
modeling. While drilling, real-time data can then be analyzed
by all available experts to refine well plans and geological
models for optimized well placement.

An empirical model was developed from offset well data,

allowing LWD data to be used to position the wellbore within
the oil column. The TVD could then be determined in real
time and the well trajectory refined.
All three laterals of Sotong A4 were drilled within 0.15-m
TVD tolerance and production tests on Sotong A3 proved the
validity of the model. The preparation and use of the
geological model, as well as the challenges of landing and
drilling the horizontal section, are discussed, with particular
emphasis on survey error management and steering decisionmaking.
Background
The Sotong field is located in Block PM-12 (Fig. 1), in the
southern part of the Malay basin, offshore Peninsular Malaysia
in water depth of 75 m. The Sotong field was discovered in
April 1973 by CONOCO while drilling the Sotong-B1.
Subsequently, five appraisal wells were drilled from 1973 to
1974 to delineate the field.

This case study demonstrates how the application of this

technology optimized production and drainage efficiency
through the successful drilling of several horizontal wells with
complex objectives. The methodology clearly has definite
application in field development planning and reservoir
simulation studies.

Fig. 1 Location of Block PM-12

GULF OF
THAILAND
GULF OF
THAILAND

D
AN
AIL
THAND
AIL
TH

The Sotong 4 well is one of five planned multilateral

horizontal wells to drain the Sotong field K2.0 reservoir in
Malaysia. The 14-m oil column is overlain by a large gas cap
and has a moderately active aquifer. In-house studies on
exploration wells confirm that gas coning will result in rapidly
decreasing oil production and that a conventional field
development would exhibit marginal economics.

PENINSULAR
MALAYSIA
PENINSULAR
MALAYSIA

PM 10
PM 10

ANGSI
ANGSI
BESAR
BESAR

DUYONG
DUYONG

SOTONG
SOTONG

Reservoir simulation defined the optimal horizontal well

placement as 2 m above the OWC, within a 0.5-m tolerance
of TVD. Industry survey error models clearly show that this
tight geometric control is not achievable using MWD survey
data alone and additional techniques are required to control the
geometric TVD error.

SOUTH CHINA
SEAH CHINA
SOUT
SEA

DUYONG BARAT
DUYONG BARAT

PM 12
PM 12

In 1978,
included the
awarded to
PETRONAS

CONOCO relinquished its acreage, which

Sotong field. The entire acreage was later
Petronas Carigali Sdn Bhd. Subsequently,
Carigali drilled two wells, Sotong 5G-5.1 and

MD ZIN CHE LAH, LANCELOT SERING, AZLEE, EVA SUNDAL, JAN DAUDEY

Sotong 5G-5.2, in 1981 and 1982, and encountered 13 m of oil

and 55 m of gas, and 3.4 m of oil and 7.6 m of gas,
respectively.
In 1994 a review was undertaken by the Marginal Field
Development Sub-Committee; one of its goals was to look
into the development feasibility of the Malong, Sotong and
Anding ( MASA ) fields. The MASA Project Team was then
initiated and given the task to identify and formulate the most
cost-effective and orderly development of these marginal
fields.
Subsequently, 3D seismic data was acquired in July 1995
to provide a better structural and stratigraphic traps definition,
as well as faults delineation of the Sotong field (Fig. 2), hence
narrowing down the overall structural uncertainties.

SPE 64477

and Sotong-B3 (1,2). Its thickness ranges from 25 to 40 m, and

its characteristic K2.0 gamma ray log profile is the easiest to
recognize. The serrated funnel shaped log response in wells
Sotong 5G-5.2 and Sotong-B2 suggests overbanks or crevasse
deposits. The K2.0 sand is the major hydrocarbon reservoir in
the Sotong field.
Fielded correlation across the field (Fig. 3) is difficult,
mainly because of poor log data, sand discontinuity and
absence of distinct marker beds. An exception to this is the
K2.0 reservoir, where log response is generally excellent,
especially in Block 1-3. Correlating shale across the field is
possible on the thick, continuous K shale, which is easily
recognized on the logs. The K shale overlays the K2.0
reservoir and is used to guide the general depositional trend
across the field. Individual shale correlations within the sand
packages yield similar results as the gross sand correlation.
The log correlation for the Sotong area was thus confined only
to sandstone packages rather than individual sandstone bodies.
E

SOTONG-B3
SOTONG-B3
SOTONG-B

STG-03-PILOT-SVY
STG-03-PILOT-SVY
STG-03-PILOT-SVY

SOTONG_02
SOTONG_02
SOTONG_02

SOTONG-01
SOTONG-01
SOTONG-01

STG-04-PILOT-SVY
STG-04-PILOT-SVY
STG-04-PILOT-SVY

m
0 m
0

m
0

DEPTH
DEPTH
DEPTH

Markers

GR
GR
GR

Resistivity

Markers

Resistivity

RHOB_NPHI
RHOB_NPHI

Markers

Resistivity RHOB (g/cm3)

RHOB_NPHI
(g/cm3)
1 . 9 RHOB
5
2.95
ILD (ohm.m)
(g/cm3)
1 . 9 RHOB
5
2.95
ILD (ohm.m)
20.0 NPHI (m3/m3)
1.95
2.95
0.45NPHI (m3/m3)
-0.15
0.2 ILD (ohm.m)
20.0

GR (gAPI)
m

0 . 0 GR (gAPI)
180.0
0 . 0 GR (gAPI)
180.0
0.0
180.0

DEPTH
DEPTH
DEPTH

0.2

0.45NPHI
20.0
0.45

Markers

GR
GR
GR

Resistivity

Markers

(m3/m3)
-0.15

Resistivity
Resistivity

Markers

GR (gAPI)

AHTRT (ohm.m)

0 . 0 GR (gAPI)
180.0
0 . 0 GR (gAPI)
180.0
0.0
180.0

0.2

0.2

RHOB_NPHI
RHOB_NPHI
RHOB_NPHI

DEPTH
DEPTH
DEPTH

Markers

Resistivity

Markers

2.95

Resistivity
Resistivity

Markers

AHT90 (ohm.m)

20.0GR (gAPI)
180.0
20.0GR (gAPI)
180.0
20.0
180.0

( o h20.0
m . m1
) . 9RHOZ
(g/cm3)
5
2.95

20.0 1 . 9 5

GR
GR
GR

GR (gAPI)

RHOZ (g/cm3)

A H T R T ( o h20.0
m . m1) . 9 RHOZ
(g/cm3)
5
2.95

0.2 A H T R T
0.2

RHOB_NPHI
RHOB_NPHI
RHOB_NPHI
RHOZ (g/cm3)

DEPTH
DEPTH
DEPTH

(g/cm3)
1 . 9 RHOZ
5
2.95
RHOZ (g/cm3)
2.95
1.95
2.95

0.2 A H T 9 0
0.2

0.45

Markers

Resistivity

Markers

(m3/m3)
-0.15

Resistivity
Resistivity

Markers

AHT90 (ohm.m)

20.0 GR (gAPI)
180.0
20.0GR (gAPI)
180.0
20.0
180.0

( o h20.0
m0.45
. m ) NPHI (m3/m3)
-0.15

0.45NPHI
20.0

GR
GR
GR

GR (gAPI)

1 . 9(m3/m3)
5
A H T 9 0 ( o h20.0
m . m ) NPHI

0.2

-0.15

RHOB_NPHI
RHOB_NPHI
RHOB_NPHI
NPHI (m3/m3)
0.45NPHI (m3/m3)
-0.15
0.45NPHI (m3/m3)
-0.15

A H T 9 0 ( o h20.0
m . m ) RHOZ (g/cm3)

0.2

0.2 A H T 9 0
0.2

0.45

( o h20.0
m1. m
(g/cm3)
. 9) RHOZ
5
2.95

DEPTH
DEPTH
DEPTH

Markers

GR
GR
GR

RESISTIVITY

Markers

RHOB_NPHI
RHOB_NPHI
RHOB_NPHI

RESISTIVITY
RESISTIVITY

Markers

ATR (ohm.m)

GR_CDR (gAPI)

m
m
m

DEPTH
DEPTH

Markers

GR
GR
GR

Resistivity

Markers

Resistivity

RHOB_NPHI
RHOB_NPHI

Markers

Resistivity NPHI (m3/m3)

RHOB_NPHI
0.45NPHI (m3/m3)
-0.15
ILD (ohm.m)
0.45NPHI (m3/m3)
-0.15
ILD (ohm.m)
(g/cm3)
20.0 RHOB0.45
-0.15
(g/cm3)
1 . 9RHOB
5
2.95
0.2 ILD (ohm.m)
20.0

GR (gAPI)

RHOB (g/cm3)

20.0GR (gAPI)
180.0
20.0GR (gAPI)
180.0
20.0
180.0

(g/cm3)
5
2.95
0.2 ATR (ohm.m)
20.0 1 . 9 RHOB
(g/cm3)
5
2.95
0.2 ATR (ohm.m)
20.0 1 . 9RHOB
2.95
0.2
20.0 1 . 9 5

(gAPI)
20.0GR_CDR 180.0
20.0GR_CDR 180.0
(gAPI)
20.0
180.0

-0.15

(g/cm3)
1 . 9RHOZ
5
2.95
20.0

1.95

-0.15

0.2

0.2

2.95

1 . 9RHOB
5
20.0
1.95

K SHALE

SOTONG-B1
SOTONG-B1
SOTONG-B1

100020003000
100020003000
1 0 0 02 0 0 0 D3E0P0T0H

m
m
m

(g/cm3)
2.95

2.95

K SHALE

K SHALE

K SHALE

K SHALE

K SHALE
K SHALE
K SHALE
K SHALE

K SHALE

-2050
-2050
-2050

-2050
-2050
-2050

K SHALE
K SHALE

K SHALE

K SHALE
K SHALE

-2075
-2075
-2075

-2075
-2075
-2075

-2100
-2100
-2100

2100

2100

2100

2100

K-1
K-1
K-1

2100

2100

2100

2100

-2100
-2100
-2100

2100

2100

2100

2100

2100

2100

2100

K-1
K-1
K-1

K1.2 top
K1.2 top

-2125
-2125
-2125

K1.2 top

K-1
K-1
K-1

K-1
K-1
K-1

K-1
K-1
K-1

K1.2 top
K1.2 top

K-1
K-1
K-1

K1.2 top

K1.2 top

K1.2 top

K1.2 top

-2125
-2125
-2125

K1.2 top

K1.2 top
K1.2 top

K1.2 top

K1.2 top
K1.2 top
K1.2 top
K1.2 top
K1.2 top

K-2
K-2
K-2

K-2
K-2
K-2

-2150
-2150
-2150

K-2
K-2
K-2

K-2
K-2
K-2

K-2
K-2
K-2

Shale stringer
Shale stringer
Shale stringer

-2175
-2175
-2175

OWC @ -2179m TVDSS

OWC @ -2179m TVDSS
OWC @ -2179m TVDSS

2.0
K

-2175
-2175
-2175

BASE
2.0
K

BASE
2.0

BASE

2.0

Dolomitic Sst. stringer

Dolomitic Sst. stringer
Dolomitic Sst. stringer

BASE

2.0
K

BASE
2.0

BASE

2.0
K

2.0

BASE
2.0

2.0

2.0

2.0

BASE

BASE

BASE
K

BASE

BASE
K

K
K

-2200
-2200
-2200

-2150
-2150
-2150

K-2
K-2
K-2

GOC @ -2165m TVDSS

GOC @ -2165m TVDSS
GOC @ -2165m TVDSS

2.0

2.0

BASE

BASE

2.0

BASE

2.0
K

BASE
2.0

BASE

2.0

K-3
K-3
K-3

-2200
-2200
-2200

BASE

Fig. 3: Sotong Field Correlation (K2.0 Sand Marked Yellow)

Fig. 2: Sotong Field K2.0 Structure Map

The resulting field development plan presented an

integrated study of the Sotong field. It is an update of the
earlier studies and utilizes the latest 3D seismic data, results
from the exploration and appraisal wells, a reservoir
engineering simulation study on the K2.0 reservoir, and
technical and economical screening of possible development
concepts.
K2.0 Reservoir Characteristics
The K2.0 sand is well developed in three of the
exploration/appraisal wells drilled in the Sotong structure. It is
characterized by a cylindrical gamma ray profile and is
interpreted as channel deposits of a major distributary channel
flowing southeast through wells Sotong 5G-5.1, Sotong-B1

In the K2.0 reservoir, the major hydrocarbon occurrence is

found in the Block 1-3 that contains the wells 5G-5.1, B-1 and
B3. The sand displays well sorted, fine to medium grained
siliciclastic deposits, with traces of interstitial clay and heavy
minerals. The average porosity from Sotong-B1, B3 and 5G5.1 is 21.5%, and average permeability is 300 mD. The sand
quality in K2.0 deteriorates towards the southwest and
northeast, away from the main channel deposits. In the
southeast, dipmeter studies and 3D seismic amplitudes
indicate reservoir quality will improve.
In Block 1-3 (Fig. 4), the average GOC in the K2.0 sand is
interpreted at 2165 m SS, while the average depth for the
OWC was 2179 m SS according to the logs from the SotongB1 and Sotong-B3. Production test data also confirm the
presence of oil and gas in the reservoir.

SPE 64477

REAL-TIME DATA ANALYSIS WHILE DRILLING PROVIDES RISK MANAGEMENT FOR BOTH
GEOLOGICAL AND GEOMETRIC UNCERTAINTIES IN THE SOTONG K2.0 RESERVOIR

For optimum reservoir drainage the simulation also

showed that a 400-m separation was required between each of
the three individual production intervals.

SOTONG FIELD : K2.0 Net Oil Sand Map

Pilot Hole
Both the GOC and the OWC have an amount of
uncertainty; with small targets, it is critical to minimize these
uncertainties. A 12-in. pilot hole was drilled to reduce the
uncertainty in the TVD associated with the fluid contacts. The
pilot hole was subsequently also used to verify the structure
and the TVD location of the target sand itself, in addition to
the GOC and the OWC.

Fig. 4: Net Oil Sand Map for Sotong Field

Sotong 04 Objectives
Development with conventional vertical wells had
consistently exhibited marginal economics. Also, there was a
limitation on the number of slots available on the platform to
be used for this project. To maximize cost efficiency, five
multilateral horizontal wells were planned to optimally drain
the Sotong K2.0 reservoir. The Sotong 04 trilateral in
particular was designed to drain the southeastern part of the
reservoir.
Fig. 6: Pilot Hole LWD Data With Squared Logs

Reservoir simulation defined the optimal horizontal well

placement as 2 m above the OWC, within a 0.5m TVD
window (Fig. 5). In addition, it was found that optimum
production would be achieved by penetrating 300 m of good
quality sand in each of the three legs.

GOC

12 m

2 m TVD
OWC
Fig. 5: Schematic of Sotong 4 Objectives

The LWD data acquired in the pilot hole (Fig. 6) provided the
following information:

14 m

Confirmation of the resistivity trend versus height above

the OWC already established through the STG-03
premodeling work
Identification of the GOC at 2166.3 m SS
Exact TVD location of the OWC at 2178.8 m SS (in terms
of drillers depth)
TVD and thickness of the target (K2.0) sand.

In addition, the data were used to create a new formation

model to help in positioning the horizontal wellbores
correctly.

MD ZIN CHE LAH, LANCELOT SERING, AZLEE, EVA SUNDAL, JAN DAUDEY

SPE 64477

Operational Challenges

Lithology Formation Model

As noted, the primary objective of this well was to land

and maintain the horizontal wellbore at a constant height of 2
m above the OWC. Analysis of the survey errors for the
Sotong 04 laterals indicated that the uncertainty associated
with drilling deviated wells from the rotary table down to the
GOC would be in the order of 6 m TVD.

To aid in the interpretation while drilling and ensure the

correct positioning of the wellbore, a formation model was
built that permitted the calculation of expected log responses
through the formation. However, the model described here
only reflects changes in log character with respect to lithology.
A second model was constructed for the fluid column. Bearing
the above in mind, the objectives of the modeling of the LWD
responses to lithology were

At the sidetrack point (taking the pilot hole depth as a

reference point), the survey errors could be reset to zero. Also,
a major reduction in the initial uncertainty is possible using
neutron-density log measurements to identify the GOC, since
all three laterals were expected to cross this fluid contact.

There are many techniques to compute vertical depth

errors resulting from survey errors, but a conservative
approach shows that the vertical depth error grows at
approximately 0.43 m per 100 m of measured depth. This
translates into approximately 3 m of error at the end of the
horizontal sections, which exceeds the acceptable tolerance on
height above the OWC required for the K2.0 reservoir
development. The challenge remained to find a method of
ensuring that the horizontal section stayed within the required
0.5-m tolerances.
Furthermore, the pilot hole data showed that the target
formation did not consist of uniform sand, but contained a
shale/dolomite streak and tight limestone stringers (Fig. 6),
complicating the real-time interpretation. Consequently, one
could not solely rely on the resistivity measurement to identify
the producible intervals. Therefore, prior to drilling any
horizontal wells, an investigation into feasible techniques for
real-time geosteering was undertaken.
The required 400-m separation between the three laterals
imposed some challenges to the directional drilling on the
project. A 9 5/8-in. casing with pre-milled window was to be
set, and two of the laterals were to be drilled out of these
windows. The correct orientation of these windows was
critical: a misalignment would have a serious impact on the
required dogleg severity to reach the targets.
The bottomhole assemblies also had to be carefully
evaluated for suitability for this project. In particular, great
emphasis had to be put on the location of the measurement
points in order to identify the position of the well as early as
possible.
With challenges as these, it was clear that this project
required extensive preplanning to ensure the desired results.

to provide a modeled LWD log response as an aid to

identify certain markers during the landing of the well, as
well as for the drilling of the horizontal section
to be able to distinguish fluid from lithology effects on the
logs
to identify the changes in log character as the wellbore
travels through the K2.0 sand and thereby recognize the
geological position of the wellbore in real time
to identify the log characteristics as the wellbore
approaches the dolomitic tight streaks in the lower part of
the K2.0 sand and thereby identify this scenario in real
time
to interpret the final LWD logs to obtain a revised
geological section with updated formation dip and layer
thickness.

Pilot hole data were used to construct the lithology model,

and the main measurements employed were the gamma ray
and the deep formation resistivity. Because the formation dip
was essentially flat, the TVD data were considered to
represent the TBT of the formation. Because Sotong 02 was
located between two of the laterals, a second formation model
was constructed using this data for control.
The pilot and Sotong 02 data were squared (Fig. 6) using
modeling software to transform the logs to formation models
where the layering corresponds to the actual formation layers,
and not just to the tool response. The software uses
convolution filters that permit the calculation of the true
formation properties, resulting in a column of layers of
discrete resistivities of varying thicknesses and a
corresponding column of gamma ray layers.
After the layer columns were built, the next step was to
model the response of the LWD sensors through these
formation layers. This technique has become fairly standard
for geosteering applications, and it accounts for the variation
in tool response depending on the relative angle between the
sensors and the formation boundary. The gamma ray was
modeled for relative angles of 0 and 90, while the resistivity
response was modeled for relative angles of 0, 45, 60, 75,
80, 85, and 90. A wider range of relative angels is needed
for the 2-MHz resistivity, as this measurement varies
significantly with relative angle.

SPE 64477

REAL-TIME DATA ANALYSIS WHILE DRILLING PROVIDES RISK MANAGEMENT FOR BOTH
GEOLOGICAL AND GEOMETRIC UNCERTAINTIES IN THE SOTONG K2.0 RESERVOIR

Phase and attenuation resistivities for the 2-MHz

propagation tools are solved using the electromagnetic wave
equations. Care needs to be taken to identify possible
anisotropic effects, which are often encountered in high-angle
wells, in order not to misinterpret the resistivity curves.
Fortunately, once identified, this effect can be remodeled
while drilling, by inputting the proper Rh/Rv contrast.
The results from the simulation are written to a twodimensional database. To each pair of vertical depth within the
layer column and relative angle, a modeled LWD log response
is assigned. Consequently, any well trajectory described as a
series of vertical depths and deviation angles can be lookedup in this matrix, and a modeled log in measured depth can
be quickly created. A change in the formation dip alters both
the TVD depth within the set of resistivity layers and the angle
between the LWD tool and layer boundary, coinciding with
the two look-up variables in the database. A log response
may thus be looked-up in measured depth for any well
trajectory or formation dip.
This is a very powerful technique because it allows
computer-intensive modeling before the drilling starts and
requires it only once for a given layer column. The generation
of modeled logs along a particular trajectory is very fast
because it involves simply looking up the table of log
responses. This makes it very effective as an interpretive tool
in real time to make trajectory adjustments while drilling.
According to the geological 3D model, the structural dip
was expected to be very low towards the southeast of the field
where the targets are located. The highest dip expected was
around 1.4 in the main bore, and less than 1 in the two
laterals. The nearest offset wells are Sotong 01 and Sotong 02.
Sotong 01 is located to the northwest of the target locations,
and Sotong 02 is between the main bore and lateral 2.

An alternate method was needed to identify the position of the

wellbore relative to the OWC in the horizontal section.
A study was performed on the existing wells in the field to
establish whether a relationship between resistivity and height
above the OWC could be found. Five offset wells were
available for analysis: Sotong B-1, Sotong B-3, STG-01 and
STG-02. The two latter wells were also used as control wells
for STG-04 to provide additional formation information.
From the offset well data, it is clear that there is a
transition zone from water to oil, with a reasonably welldefined relationship between resistivity and height above the
water. Plotting the resistivity versus TVD height for all the
five offset wells shows a consistent trend. A vertical depth
shift of up to 4 m had to be applied to the offset well data to
position the OWC at 2179 m below mean sea level. This depth
shift can be seen as an indication of the magnitude of the
survey errors involved in the offset wells (with the assumption
that the OWC is horizontal across the field).
The resultant plot reflected the shape of a capillary
pressure J-curve as expected (Fig. 7). The data dispersion seen
is likely the result of variations in porosity and mineralogy.
S o to n g H e ig h t v s R t
2160

2165

2170

y = -0 .6 7 9 5 x + 2 1 8 3 .6

2175

A TVD shift was applied to the layer column to match the

predicted position of the formation tops and formation dip
information was input according to the structural models
extracted from the 3D geological model. The planned
wellpaths were added to the formation model for each of the
laterals before drilling started, and the expected log responses
were computed. The resulting modeled logs were to be used as
a reference when drilling commenced and compared to the
data acquired in real time. Any discrepancies would then
require a careful investigation to determine whether there was
a change in lithology, fluid content or formation dip. The
model could then be adjusted accordingly.

Resistivity Modeling
Geometric control of the TVD was not sufficient to
maintain the horizontal wellbore at only 2 m above the OWC.

G O C = 2165m
O W C = 2179m

2180

2185
0

10

15
Rt

20

25

30

Fig. 7: Resistivity Versus Height Above Fluid Contacts

From this data, a value of formation resistivity of

approximately 10.5 ohm-m represented a height above the
OWC of 2.5 m. A minimum value of 10 ohm-m was chosen as
a guide to use during the drilling of the horizontal section. If
the resistivity during drilling approached this limit, the
implication was that the wellbore was too close to the OWC
and would have to be steered up.

MD ZIN CHE LAH, LANCELOT SERING, AZLEE, EVA SUNDAL, JAN DAUDEY

SPE 64477

Many factors affect the formation resistivity. Changes in

porosity and shaliness have a marked impact on the resistivity.
To be sure a change in resistivity was due to a change in
height of the wellbore, additional measurements were made
during drilling to confirm the porosity and shaliness of the
formation. Another factor that can strongly influence the
apparent resistivity measured by LWD tools is the formation
anisotropy. Anisotropy generally has the effect of increasing
the apparent resistivity as the hole angle increases.

measurement at the bit. This enabled the geosteering team to

identify early if the wellbore was getting closer to the water
and the trajectory could be adjusted when needed.

Real-Time Geosteering Team

Drilling Sotong STG-04 Trilateral Horizontal Well

To address these multidisciplinary issues, a team was put

in place, bringing together the expertise on both the operator
and contractor side. It was clear that it would be necessary to
draw upon all available expertise not only in the planning
stage, but also during the drilling itself. It was critical to have
as much information as possible available while drilling. It
was decided to set up a real-time link from the rig to Carigalis
offices in order to benefit from the expertise of the
multidisciplinary team as steering decisions needed to be
made.

12 -in. Section and Setting of 9 5/8-in. Casing

A rig link was set up, providing real-time transmission of

the LWD and directional data, effectively bringing the drilling
process to shore. All drilling data could be displayed and
downloaded for further analysis in real time and 24-hr
coverage ensured continuous monitoring of the progress. At
all times the geosteering specialist and the directional drilling
expert in town would monitor the progress and contact the
various experts as steering issues arose.
In this way the impact on the drilling process could be
minimized and decisions could be made on the spot without
any delays. The decisions could then be relayed to the rig
instantaneously via the phone or the satellite link and changes
to the well path were initiated immediately.
Bottomhole Assemblies
In light of the above analysis, a range of measurements had
to be included in the drilling string, and the location of the
measurement point of the resistivity and the wellbore
inclination was deemed critical for the execution.
In order to place the measurements at the bit, an
instrumented motor was used. The motor featured an
inclination, Resistivity and Azimuthal Gamma Ray

In addition, an MWD tool was included for directional

purposes. Also a neutron-density LWD tool was transmitting
data while drilling in order to determine the origin of any
resistivity changes; which could be due to changes in lithology
as well as fluid content.

The 12-in. hole was sidetracked from the pilot hole at

2162 m measured depth (MD). Real-time correlation with
modeled logs in this section was used to confirm the position
within the formation, as well as to confirm the casing point.
The 9 5/8-in. casing was set at 2526 m MD at 74.6
inclination and azimuth of 120.7. The casing point was
chosen to allow enough distance to the horizontal entry point
to properly control the well path and to give sufficient room
for the liner lap and the window for the second lateral.
The 9 5/8-in. casing and casing window joints were part of
a trilateral system. The main bore was to be drilled out of the
casing shoe, whereas two oriented pre-milled window joints
were placed in the casing string for upcoming laterals 1 and 2.
The main lateral was drilled first. The lower window
(lateral 1) was set with the bottom of the window at 2389 m
MD. The bottom of the top window (lateral 2) was set at 2352
m MD.
Drilling the Main Bore
The top of cement was found at 2513 m MD and down to
the shoe the section was logged through casing with gamma
ray to aid the interpretation.
The plan was to land the well in two steps, with the first
landing at 2206.2 m TVD RKB geometrically, then to use the
resistivity measurement to decide on the final landing depth.
The BHA used for this purpose consisted of an extra power
instrumented motor that featured a double bend, one 0.75
fixed bend and one adjustable bend set at 0.39.

REAL-TIME DATA ANALYSIS WHILE DRILLING PROVIDES RISK MANAGEMENT FOR BOTH
GEOLOGICAL AND GEOMETRIC UNCERTAINTIES IN THE SOTONG K2.0 RESERVOIR

SPE 64477

1100

1200

1300

1400

1500

1600

1700

120 m, the resistivity was still below 6 ohm-m. It was then

believed that the sand had shaled out and that the extent of the
sand was not as great as had been anticipated from the 3D
geological model.

1800

-700

-700
H o r iz o n ta l E n tr y P o in t

-800

-800

<<< SOUTH

S ta r t o f P e r f s 3 0 1 8 m M D

Proposal
-900

Ho

Survey

ld

-900
Az

im

u th

12

2 .1

3
E n d o f P e rf s 3 3 2 0 m M D

-1000
PLAN VIEW

-1000

Scale (1 cm = 50 meters)

TD at 344 0m M D

-1100

-1100
P la n n e d T a rg e t T D

1100

1300 EAST >>> 1400

1200

1500

1600

1700

1800

Vertical Section View

74 .46

2200

25 25 M D
H o r iz o n ta l E n tr y P o in t

S ta r t o f P e r f s 3 0 1 8 m M D

E n d o f P e rf s 3 3 2 0 m M D

9 5 / 8 in C a s in g P t
TD at 344 0m M D

Vertical Section Departure at 122.00 deg from (0.0, 0.0). (1 cm = 50 meters)

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000

Since 300 m of good pay sand had already been drilled at

the required distance from the OWC, it was decided to call
3440 m the final depth. However, this would pose a challenge
for the drilling of the upcoming laterals, as the perforation
interval in the main bore was now closer to the shoe than
planned. If the 400-m separation between the laterals was to
be maintained, the other wellbore would either have to be
drilled longer or turned away from the main bore quicker.
Both laterals were replanned to allow for this development.

2100

Drilling Lateral 1
Fig. 8: Main Bore Wellplan (Blue) and Final Surveys (Red)

The well required a 1.6/30 m build in inclination to get to

horizontal at the planned TVD (Fig. 8). The double bend in
the motor and the distance between the first two stabilizers
resulted in an even higher build tendency at +/-2/30 m.
Consequently, the assembly had to be orientated to drop angle
on a regular basis. The sensor for inclination at the bit was
used actively to ensure the wellplan was followed. The main
bore was first landed at 2838 m MD, at a TVD of 2205.7 m.
Using the resistivity measurement, the final landing depth was
set to 2207 m TVD RKB, and this target was reached at 3076
m MD.
After landing the well the build tendency decreased, but
steering down was still required to keep the well horizontal at
the desired TVD. Pieces of stator rubber seen later at the
shakers indicated that stalling of the motor was deforming the
elastomer and causing chunks to break off. While sliding, the
ROP was slow because poor weight transfer to the bit resulted
in weight stacking. The BHA was tripped out and the motor
replaced. The objective of this run was to stay horizontal and
low doglegs would be required to stay on track. Consequently,
the new motor was configured with only one fixed bend at
0.75.
The well was maintained at the same TVD until 3300 m
MD, where the resistivity started to drop. This coincides with
higher gamma ray reading, indicating the sand was becoming
shalier. At this point, the target TVD was changed to 2208 m,
as the interpretation suggested that the shalier sand was on top
the wellbore. Similar shale effect detected in the density and
porosity readings confirmed this theory.
The formation change was also evident on the directional
behavior of the BHA; it started to build faster. After drilling

On drilling out the Lateral 1 it was found that the windows

had been oriented incorrectly. The lower window was set at
155 right of high side instead of 30 right, whereas the upper
window was set at 95 right of high side instead of 30 left.
Lateral 1 was originally planned to kick off from the upper
window, as the wellpath was on the left of the main bore.
Additionally, the hole inclination was less than originally
planned. This misalignment clearly called for a new well plan.
The lower casing window was set at 155 right of high side
or low side of the hole; it was decided to use this window for
this lateral. The switching of the window and the positions
resulted in a much more challenging wellpath, and it was
anticipated that the drilling time and the wellbore tortuosity
would increase as a result of the higher dogleg severity
required to turn the well to the target zone.
With the window set at 155 right of high side, the BHA
would first have to drop angle, then turn and build away from
the main wellbore. A single bend motor with a 1.5 bend was
used to drill out, and 47 m of sliding were completed. The
dogleg severity appeared very low, since magnetic
interference from the casing masked the true turn rate. Later
surveys free from interference confirmed that the drop and
build portion was between 6 to 7/30m, but that turn rate of
the kickoff was too low.
The logging BHA with the instrumented motor was run in
the hole. Once on bottom, the assembly had to be orientated
immediately in order to build and turn the well at 5.5 to 6
doglegs to maintain the new proposed well trajectory. The
main objective was to excite the 400-m zone from the
perforations in the main well.

MD ZIN CHE LAH, LANCELOT SERING, AZLEE, EVA SUNDAL, JAN DAUDEY

1 000

1 200

1 400

1 600

1 800

2 000

-40 0

-40 0
En d of Perfs 3575m M D
Low er C asing W ind ow

TD at 3635m M D
H orizontal Entry Po int
P roposed TD

-60 0

t
en
40

0m

d is

p la

ce

H orizontal Entry Po int

<<< SOUTH

-60 0

Start pf P erfs 3275m M D

9 5/8in C asing Pt

-80 0

-80 0
Start of P erfs 3018m M D

P roposal
S urvey, Lat #1
Lateral Main
En d of Perfs 3320m M D

-10 00

PLAN VIEW

-10 00

Scale (1 cm = 100 meters)

TD at 3440m M D

1 000

1 200

1 400

1 600

1 800

2 000

EAST >>>

2 100

SPE 64477

Using the real-time data, the geological model was updated

on a regular basis and confirmed the suspicion that the
encountered tight formation was equivalent to the tight streaks
in the pilot hole (Fig. 6: 2208 m TVD RKB). The original
model did not predict the borehole would penetrate this part of
the sand, but a slightly higher formation dip than anticipated
placed the wellbore lower in the structure. As this tight streak
was assumed to be continuous, the team decided to place the
wellbore above it in an effort to minimize the possible impact
on the production rates.

Low er C asing W ind ow

Vertical Section View

Start pf P erfs 3275m M D

H orizontal Entry Po int

En d of Perfs 3575m M D

2 200
P roposed TD
355 8 MD

8 00

9 00

1 000

1 100

1 200

1 300

1 400

1 500

1 600

1 700

1 800

TD at 3635m M D

1 900

Vertical Section Departure at 84.00 deg from (0.0, 0.0). (1 cm = 50 meters)

Fig. 9: Lateral 1 Wellplan (Blue) and Final Surveys (Red)

The slow progress was due to the amount of sliding

required to achieve the intended targets and, possibly, to well
geometry. The build rate required was 5/30 m, which was the
maximum obtainable with this BHA setting. A higher bend in
the motor would have prevented the BHA from being used in
rotary mode, because the bit offset would have been too great.
From the well plot (Fig. 9) it is obvious that priority was given
to landing the well at the required TVD, rather than to
achieving the required turn.
On landing this lateral, the structure came in as predicted,
translating into a 0.4 formation dip and placing the entry of
the K2.0 target sand at 2502 m MD, or 2171.8 m TVD RKB.
The GOC was found at 2198 m TVD RKB. Assuming a 14-m
oil column, the initial landing target was set to 2206 m TVD
RKB.
After the first landing depth at 2670 m MD or 2206.4 m
TVD RKB, the target TVD was set at 2207 m. However, some
26 of turn was still needed to attain sufficient separation
between the main bore perforations and the production interval
of this lateral. The BHA was performing well and the required
turn was achieved. As the drilling progressed, the neutron
porosity increased while the resistivity showed 15 ohm-m. At
2770 m MD the target TVD was thus revised to 2207.5 m.

While drilling through this hard formation, low penetration

rates were observed and the logging tools recorded shocks on
the BHA. PDC bits in general do not perform well in
cemented formations and sliding to rise the TVD proved
difficult. This may also have been due to small ledges that
were causing the BHA and drillpipe to hang up. At 3220 m
MD the wellbore was back in the clean oil sand, and the
interpretation showed that the formation dip was 0.2 rising.
To maintain the wellpath above the dolomitic formation to the
end of the section, the new horizontal target was set to 2207 m
TVD RKB until 3575 m MD, where 300 m of perforation zone
at the right separation had been achieved. The wellpath was no
longer controlled to drill the rathole.
Drilling Lateral 2
A closer examination of the planned wellpath for this
lateral showed the well could be significantly shortened if a
more aggressive turn was planned to clear the 400-m
separation circle from the main bore perforations (Fig. 10).
The original plan required a borehole of 3503 m MWD, which
was necessary for obtaining 300 m of perforation zone with
the required separation. Redesigning the well with a sharper
turn reduced the overall length to 3146 m, 357 m shorter. The
team decided to follow this new path.
800

100 0

120 0

140 0

180 0

200 0
-4 00

Up p e r W in d o w
TD a t 3 6 3 5 m M D

En d o f Sp e rry R u n

-6 00

-6 00

H o r iz o n ta l E nt ry P o in t

-8 00

Sta rt of P e rfs 3 0 1 8 m M D

.
sp

400m LW
En d o f Pe rfs 3 3 2 0 m M D

di

Start of Perfs 2965m MD

40

172 .86

-1 000

100m from HE P )

-1 000

0m

uth
Hold Azim

<<< SOUTH

-8 00

At 3145 m MD the TVD was approaching 2208 m and the

resistivity was reading 10 ohm-m, indicating the wellbore was
approaching the OWC. To avoid dropping further, the
assembly was oriented. After drilling a few meters, the well
penetrated a calcarious cement or dolomitic formation at 3150
m MD. This was confirmed by an increase in resistivity and
density with a decrease in porosity, all indicating a tight lime
or dolomite formation. This proved the importance of the realtime neutron-density data. The resistivity alone may have
misled the geosteering team to believe the wellbore was
moving away from the OWC and to correct by steering the
well down.

160 0

-4 00

TD a t 3 4 4 0 m M D

En d o f p e rfs

-1 200

-1 200
TD a t 3 1 4 6 m M D

Proposal

-1 400

3 6 0 m from s ta rt o f pe rfs

Survey

-1 400

3325m M D

Form ation Uncertainty

TD, 3503m MD

PLAN VIEW

Scale (1 cm = 100 meters)

-1 600

-1 600
800

100 0

120 0

140 0

160 0

180 0

200 0

EAST >>>

Fig. 10: New Lateral 2 Plan: Original Plan (Blue) Inside 400-m
Separation Circle (Black); New Plan (Green) Outside Circle

REAL-TIME DATA ANALYSIS WHILE DRILLING PROVIDES RISK MANAGEMENT FOR BOTH
GEOLOGICAL AND GEOMETRIC UNCERTAINTIES IN THE SOTONG K2.0 RESERVOIR

SPE 64477

To achieve the newly set objective with the reduced well

length, an aggressive build and turn BHA was designed. An 8
3/8-in. sleeve stabilizer was fitted on the mud motor with an
additional 8-in. sleeve stabilizer on top. Farther back in the
assembly one of the logging tools featured an 8-in. sleeve
stabilizer.
As drilling commenced, the reactive torque from the motor
was in the order of 270 with 12,000 to 15,000 lbm WOB,
showing that the motor was functioning normally. Rotary
drilling, as well as sliding, was smooth until the wellbore
entered the sandstone. The well reached the top of sand at
2544 m MD and landed horizontally at 2655 m MD with 2207
m TVD RKB. The final landing depth was determined using
the GOC from the real-time logs.
Immediately after entering the sandstone formation the
toolfaces became erratic and hard to control. This may have
been caused by differential sticking, which was also
experienced in previous sections drilled. The sensor for
measuring downhole weight on bit showed that string weight
was not getting to the bit and when the BHA was pulled offbottom, the extra weight needed to pull free. The real-time
ECD showed a decrease. It was also noticed that on Lateral 1
the same problem occurred in the transition between shale and
sandstone. As the last stabilizer of the BHA cleared the shale,
ROP improved in both cases.

The well maintained the same TVD with resistivity reading

between 10 to 15 ohm-m. At 2750 m MD, the desired
resistivity range was revised to 15 to 16 ohm-m and the TVD
target was kept at 2207 m.
At 2846 m MD, samples showed calcareous sandstone as
the drilling rate dropped to 3 m/hr. However, this was different
from the tight streak experienced in Lateral 1 and the pilot
hole. This well was drilled farther updip in the structure, and
the characteristics of the logs exhibited a steep transition
(gamma ray and resistivity), indicating this was a local event
rather than an extensive layer in the sand itself. The target
TVD was revised to 2206.5 m, since the dip in this area was
quite flat and raising the TVD would bring the wellpath out of
this calcareous sequence faster. At 2876 m MD, the well was
back in 100% sand.

After having finished the 300-m perforation zone, the well

was terminated earlier than planned at 3096 m because of the
low penetration rate. (Fig 11)
Results
Main Bore Results
The initial geological model for the main bore predicted a
general downward dip, with dips ranging from 0 to 1.4.

Furthermore, the mud weight was overbalanced and the

borehole hence became prone to differential sticking. The last
four or five stands drilled on Lateral 1, as well as the first
seven or eight stands coming off-bottom during the wiper trip,
showed evidence of differential sticking, with 70,000 lbf
overpull at times.

8 00

1 00 0

1 20 0

1 40 0

1 60 0

1 80 0

2 00 0

U pp er W in dow
T D a t 36 3 5m M D

L ate ral

9 5/8in Ca s ing P t
HEP

-6 00

E nd of P e rfs 3 5 75 m M D
S tart of P erfs 3 27 5 m M D

-6 00

HEP
m

en

HEP

p la

D is
0m

te

ra

l1

40

La

L a te ra

<<< SOUTH

S tart of P erfs 3 01 8 m M D

400m Displacement

2 71 1 m

E nd of P e rfs 3 3 20 m M D

l3

-1 00 0

-8 00

ce

S tart of P erfs

-8 00

-1 00 0

T D a t 34 4 0m M D
E nd of P e rfs

Lateral M ain

3 0 36 m

Lateral #1

T D @ 3 09 6

-1 20 0

Proposa l

-1 20 0
P ropo se d TD
PLAN VIEW

Survey, Lat #2
8 00

1 00 0

1 20 0

1 40 0

Scale (1 cm = 100 meters)

1 60 0

1 80 0

2 00 0

EAST >>>

U pp er W in dow

2 10 0

Vertical Section View

T D @ 3 09 6 m M D

HEP

2 20 0
E nd of P e rfs 3 03 6 m M D

S tart of P erfs 2 71 1 m

7 00

8 00

9 00

1 00 0

1 10 0

1 20 0

1 30 0

1 40 0

Vertical Section Departure at 168.30 deg from (0.0, 0.0). (1 cm = 50 meters)

Fig. 11: Lateral 2 Wellplan (Blue) and Final Surveys (Red)

Fig. 12: Final Interpretation of Main Bore Showing the Sand

Shaling Out toward the End of the Section

10

MD ZIN CHE LAH, LANCELOT SERING, AZLEE, EVA SUNDAL, JAN DAUDEY

The overall formation dip was slightly steeper than

expected. Also, toward the end of section the sand
unexpectedly shaled out, resulting in 120 m of shale drilled.
Unfortunately, the tight TVD tolerances with respect to the
OWC allowed little room to move the wellpath deeper in order
to follow the sand further. This response has been simulated in
the cross section (Fig. 13) by gradually thinning the sand and
stretching the thickness of the shale above.

SPE 64477

quality sand. This lateral did not shale out like the main
bore and was close to the expected geological model.
The 70-m interval of dolomitic sand was not perforated,
and the LWD logs were used to identify this zone. The 300 m
of production zone was achieved at the right spacing from the
main bore, well within the required 0.5-m TVD window.
Lateral 2 Results

As a result of the encountered shale, the production zone

had to be moved closer to the wellhead to maintain the
required 300 m of production zone. The entire section of
interest was drilled well within the required 0.5 m TVD
window.

For the second lateral the geological cross section

predicted a downward dipping formation of about 1
relative dip.

Lateral 1 Results
For the first lateral, the geological cross section predicted a
slightly upward dipping formation (dips in the order of 0 to
1) in the beginning, and then downdips of 0.6 toward the last
part of the well.

Fig. 14: Final Interpretation of Lateral 2 Showing Local Calcareous

Sandstone (High Resistivity in the Middle of the Section). Samples
Confirmed This as Different from Lateral 1 and Pilot Hole.

Fig. 13: Final Interpretation of Lateral 1 Showing Encountered

Dolomitic Sandstone (High Resistivity in the Middle of the
Section)

The overall formation dip was relatively flat, as expected

(Fig. 13). However, the structure was initially rising slightly
faster than anticipated. This had the effect of placing the
trajectory lower in the sand, and, consequently, some 70 m of
the tight dolomitic streaks in the lower part of the K2.0 target
was penetrated (Fig. 6). The dip was close to 0 at this point,
and by revising the TVD slightly to 0.5 m shallower, the
wellbore could be placed above the tight streak, into good

The original model was close to what could be interpreted

from the LWD measurements (Fig. 14). After the trajectory
penetrated the K2.0 target sand, an average dip of 0.5 was
interpreted. This placed the trajectory in the top of part of the
sand, away from the dolomitic streaks lower in the structure.
The 25 m of calcareous sand drilled are interpreted as a local
non-extensive feature, not seen in the pilot hole nor in any of
the offset wells.
Production Results
Production results from this well have been high above
expectancy: 3000 B/D produced from two laterals. Initial
estimates expected 2000 B/D from three laterals.

SPE 64477

REAL-TIME DATA ANALYSIS WHILE DRILLING PROVIDES RISK MANAGEMENT FOR BOTH
GEOLOGICAL AND GEOMETRIC UNCERTAINTIES IN THE SOTONG K2.0 RESERVOIR

Conclusions
Geosteering of horizontal wells within tight vertical
tolerances is possible, even in complex reservoirs.
Multidisciplinary teams are necessary to evaluate complex
geology to make qualified steering decisions. Real-time
availability of the data, and the constant monitoring of it,
ensure events are interpreted as they occur.
To evaluate the geology, full triple-combo real-time data
are often necessary. In nonhomogeneous sands, resistivity and
gamma ray measurements alone may not suffice to geosteer or
to identify the producible intervals.
Extensive preplanning ensures that all relevant decision
trees are in place and that all parties understand their
individual roles within them.
Acknowledgments
The authors thank PETRONAS and Petronas Carigali Sdn.
Bhd. and partners for permission to publish this paper.
Nomenclature
2 - Statistical confidence limit (95%)
BHA - Bottomhole assembly
ECD Equivalent circulating density
GOC - Gas-oil contact
LWD - Logging while drilling
m MD - meters measured depth
MASA - Malong, Sotong, Anding fields
MHz - Megahertz
MWD - Measurement while drilling
OWC - Oil-water contact
ROP - Rate of penetration
SS - Subsea
TBT - True bed thickness
TD - Total depth
TVD - True vertical depth
TVD RKB - True vertical depth drillfloor
TVD SS - True vertical depth subsea
WOB Weight on bit

11