IPTC 16715

Real-Time Modeling-While-Drilling for Optimized Geosteering and

Enhanced Horizontal Well Placement in Thin and Complex Reservoirs
Kalyan Saikia, Halliburton

Copyright 2013, International Petroleum Technology Conference

This paper was prepared for presentation at the International Petroleum Technology Conference held in Beijing, China, 26–28 March 2013.

This paper was selected for presentation by an IPTC Programme 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 International Petroleum Technology Conference and are subject to correction by the author(s). The material, as presented, does not necessarily
reflect any position of the International Petroleum Technology Conference, its officers, or members. Papers presented at IPTC are subject to publication review by Sponsor Society
Committees of IPTC. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of the International Petroleum Technology
Conference 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, IPTC, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax +1-972-952-9435

Abstract
With the growing demand for hydrocarbons, enhanced horizontal well placement resulting from optimized geosteering is
becoming increasingly important. Optimal geosteering is accomplished by selecting the proper combination of logging-
while-drilling (LWD) and measurement-while-drilling (MWD) tools as data acquired from these tools are directly deployed
for interpreting and monitoring the well. In fact, achieving maximum success in geosteering is only possible through real-
time interpretation and modeling of MWD/LWD data while drilling. Additionally, real-time 3D model updates while drilling
deliver maximum guidance and assistance in geosteering to the proper well position by predicting the geology ahead of the
bit. As a result, well placement risk is reduced because the associated uncertainty is quantified.
A meaningful modeling-while-drilling campaign is greatly dependent on three main factors:
 uninterrupted real-time data transmission
 use of appropriate software modeling
 selecting adequate modeling techniques
Log modeling and 3D geocellular property modeling-while-drilling are two widely used variants of modeling-while-
drilling. Log modeling generates a series of modeled log responses along the planned well path considering the offset well
information. Whereas 3D geocellular property modeling-while-drilling is comprised of two steps i.e., fine layering and
gridding (FLG) followed by property simulation and regular model updates with new real-time log and well pick information.
Each of these techniques requires MWD/LWD data, well picks, and surface grid information as input and uses different
algorithms to complete the modeling. But, to obtain effective results both the simplicity and the adequacy of the modeling
technique play a very important role in the procedure.
In this paper, examples from two complex and challenging wells are described in terms of the advantages and benefits of
using log modeling and 3D geocellular property modeling-while-drilling. It is concluded that modeling-while-drilling is
extremely useful in optimizing horizontal well paths, maximizing well productivity, and reducing rig lost time. In addition, it
allows professionals from different disciplines, to work together in an integrated manner for improved well placement.

Introduction
With the growing demand for hydrocarbons, enhanced horizontal well placement resulting from optimized geosteering is
becoming increasingly important. Initially, geosteering (using real-time log data) was developed to ensure maximum success
in reaching the target reservoir in the case of high-angle and horizontal wells. As a result, previously unrecoverable
hydrocarbons occurring in thin reservoir columns became easier to tap. However, to achieve this objective, precise
positioning of the well within the thin sedimentary layers is most critical.
Most operators face the problem of maximizing their use of real-time log data to achieve accurate well placement. Proper
realization of this data can only be accomplished when these data are in real-time to interpret and model the structure and
petrophysical properties of the formation. This helps the ‘geosteerer’ predict the geology ahead of the bit resulting into
optimized geosteering and thus, correct well placing. Moreover, modeling-while-drilling helps the geosteerer to reduce and
quantify well placement risk and uncertainty (Najjar et al. 2005). Thus, geosteering and accurate well placement through
LWD and MWD data have become very significant in the oil and gas industry (Ramos et al. 2008; Al-Jaberi et al. 2009).
Modeling-while-drilling for proper well placement by using real-time data started in the oil industry a couple of decades
ago. Traditionally geosteering was performed by monitoring well progress through 2D cross-sections and without any real-
time 3D geological information. This resulted in a number of unsuccessful attempts at geosteering and well placement. At
2 IPTC 16715

times, MWD and LWD log data fail in guiding and providing the geosteerer a good correlation in complex geological
situations. For example, in very thin reservoirs, resistivity and density images are not clear enough to steer the drill bit
accurately (Malinverno et al. 2002 and Clark et al. 2010). Furthermore, the distance between the tool and bit also delays the
steering decision. In this situation, proactive geosteering methods are very useful so that geosteerers can visualize the
structural and geological complexities ahead of the bit position along the planned trajectory. Thus, the development of a
proactive geosteering procedure through interpreting and updating the geological model in real-time while drilling has gained
prime importance amongst geosteerers. Many variants of modeling-while-drilling have been developed and tested
successfully over the years. Each of these techniques requires MWD/LWD data as input data and uses different algorithms to
complete the modeling (Bourgeois et al., 2006). However, the efficiency of these techniques depends on both the simplicity
and the adequacy of the techniques.
This paper describes the modeling-while-drilling technique in general, and with specific examples showing the benefits of
using real-time interpretation for accurate horizontal well placement.

The Set-up for Real-Time Modeling-While-Drilling

A typical set-up including hardware, software, and network connectivity required to carry out modeling-while-drilling is
shown in Fig. 1.

Fig. 1—Schematic showing the set-up for modeling and geosteering while drilling.

The success of a modeling-while-drilling campaign heavily depends on three main components:

 Uninterrupted transmission of real-time data from down-hole via the rig to the geosteerer
 Use of the appropriate software package for 3D visualization, interpretation, and modeling
 Selection of the right hardware with the specifications needed for fast and reliable analyses of the data
Critical to the process is a state-of-the-art satellite communication system that helps the geosteerer receive reliable data as
quickly as possible from the remote rig. For reliable data transmission, the WITSML format is the most commonly used
standard in the industry. Available software packages provide a full range of methods for performing interpretation and
model updates while drilling. Each of these software packages has its own limitations and advantages. Thus, selecting the
appropriate software is the next most important step in the procedure. Emphasis should also be given on an easy and simple
workflow that enables geosteerers to handle more than one well at a time further reducing the cost of the operation.

Log Modeling-While-Drilling and Updating the Structural Model

Real-time log modeling-while-drilling is essential for understanding and interpreting data in complex horizontal wells.
Theoretically, logging tools’ responses in a layered medium depend on the dip angle and thickness of the penetrating layers.
Thus, tool proximity to a certain bed boundary within the reservoir can be detected while drilling a horizontal section, by
analyzing the log curves’ behavior, dip information of the bed, and the angular relationship of the tool axis to the well path
(Al Muhairi et al. 2006; Bertelli and Cesare 1999). The two steps involved in log modeling include:
 Create a series of modeled log responses along the planned well path considering the offset well(s) log curves and
the dip and thickness information about the layers.
 Compare the modeled log response with the real-time log received along the actual well path as drilling progresses.
This first step lets the geosteerer predict log responses of the layer ahead of the bit along the planned well path. Next, to
get a perfect match between the real-time log and the modeled log curves, the formation layers in the model need to undergo
adjustments in thickness and dip angle. To get the best results, use an azimuthal measurement tool with the real-time logging
IPTC 16715 3

tools. Using the azimuthal tool also helps specify the exact bit location to follow with respect to the bed boundary (Thevoux-
chabuel 2010).

Case Study on Application of Log Modeling-while-drilling in Thin and Discontinuous Sand Reservoir
This case study highlights the benefits of using log modeling-while-drilling when placing a horizontal well in a thin and
discontinuous sand reservoir interbedded with shale. It also explains how the technique helps the geosteerer in making on-
the-spot decisions to follow the plan and target porosity. The well A_2_1 (Fig. 2) is a side lateral horizontal well which is
expected to produce oil from sand layers on the flank of the reservoir structure. The well is projected to add significant
amounts of oil with proper completion and accurate placement. The target sand layer is not developed in the mother-bore
(A_2_0) as it is discontinuous. However, based on seismic interpretation and information from a nearby well it is expected
that the target sand layer is well developed beginning about half way along the well path. The sand layer is predicted to be 5-
6 feet thick with an average porosity of 10-15%.

Fig. 2—Map showing the location of well A_2_1.

Properly placing well A_2_1 within the thin target sand zone was the most challenging part of drilling the well.
Furthermore, the drilling engineers at the rig also encountered other drilling difficulties which added more uncertainty to the
well placement. Hence, using log modeling-while-drilling was proposed to help lessen the uncertainties. Data from four
offset wells (highlighted in Fig. 2) including gamma ray, porosity, and density were used to generate the modeled log curves.
Fig. 3 displays the planned (dashed black line) and the actual trajectory (solid blue line) of well A_2_1. The colored cross
section shows the interpreted and the updated structure. Initially, the target sand layer with 10-15% porosity was expected at
a shallower depth and the well was planned accordingly. But, as drilling continued and entered the upper layers, real-time
interpretation and log correlation indicated the target porosity zone is located at a deeper level, i.e., ‘layer D’. This indicated
that we needed to change the plan and drill along layer D to follow the target porosity until we reached the planned total
depth of the well.
Next, the updated plan and the real-time data are incorporated into the log modeling packages and modeled log curves are
generated for gamma ray and porosity. This is followed by correlation of the modeled curves with the actual LWD/MWD
curves measured along the well path. Even though we got a good correlation, the structure was updated as shown in the Fig.
3. Based on the structural prediction from the log modeling, instructions were given to hold the well around 91.5 to 92 in
layer D. This helped keep the well within the target porosity.
However, while geosteering the well we observed that the modeled porosity curve displayed low values and the gamma
ray curve is high towards the TD of the well (Fig. 3) suggesting an abrupt change in the lithology of the layer. As the well
approached the region and real-time MWD/LWD log data started acquiring the geosteerer updated the structural model by
matching the modeled and real-time curves. Further adjustment and update of the structural model in terms of thickness and
dip of the layers revealed a pinch out in layer D. Accordingly, we decided to stop drilling and complete the well to avoid
drilling further along the non-porous layers. In this well (A_2_1), modeling-while-drilling helped the geosteerer to
successfully geosteer about 1500 ft within the target porosity. The geosteerers are also benefitted in terms of:
1. Real-time updates of the structural model helped in making quick decisions and updating the well plan at the
beginning.
2. Maintaining the target porosity (10-15%) within layer D.
3. Indicating a possible exit from the target porosity zone well ahead of time.
4 IPTC 16715

Fig. 3—Interpreted cross-section showing actual and planned well path and correlation of modeled and real-time (LWD) curves for
porosity and gamma.

3D Geocellular Property Modeling-while-Drilling

An alternative to the log modeling technique is 3D geocellular modeling-while-drilling which is comprised of two steps, i.e.
fine layering and gridding (FLG) and real-time petrophysical property updates. In this technique, a smaller area is selected
around the active well and 3D structural and geological models as well as petrophysical properties are periodically updated as
new information becomes available while drilling (Najjar et al. 2005). Therefore, this technique guides the geosteerer in
making real-time decisions about well placement in terms of target porosity or other petrophysical properties. Geostatistical
3D geocellular model update algorithms make this technique more powerful as enhanced decision making results in more
accurate well placement. Geostatistical techniques also help to quantify the well placement risk.
The workflow begins by defining a restricted and smaller area around the planned horizontal well to create a smaller
subset model from the existing regional 3D geocellular model. The extent of the subset model is decided in such a way that it
includes a reasonable area around the planned well and a few offset wells (if available). Then, the subset model is checked
thoroughly for any miss-tie between the real-time picks and the model layers. This step is very important as most of the 3D
regional models are not updated with the latest well pick information. Next, fine layering and gridding (FLG) is applied to
generate a detailed well scale model. The fine scale model is subsequently populated on the refined grids with reservoir
properties such as porosity, permeability etc. Sequential Gaussian Simulation (SGS) techniques are the best suited
geostatistical methods for these quick model updates. The subset model is then updated at regular intervals with real-time
LWD and well pick data until the well reaches the planned total depth. The steps involved in 3D geocellular property
modeling-while-drilling are shown in Fig. 4.
IPTC 16715 5

3D Reservoir
Create Subset Model Model
(Around Active well) (Entire Field)

QC subset model

Need YES Update subset

update? model and tie to
surface grid.

No

Fine layering and gridding on the Subset model.

Well Pick Information

from active well

Structure Model Update Continues as

drilling continues.
Real-time LWD/MWD Log
Data from Active well

Property Model Update

GEOSTEERING RECOMMENDATION (Well Path Update)

Fig. 4—Schematic of workflow for 3D geocellular modeling-while-drilling.

Case Study on 3D Geocellular Property Modeling-While-Drilling

In this section, a case study is described where 3D geocellular property modeling-while-drilling for geosteering a well in thin
reservoir was used with great success. The benefit gained in terms of accurate well placement by use of this technique is
demonstrated by the enormous success achieved in the well. The H12_2 well is the second side lateral placed in the crest of
the structure with a planned horizontal length of almost 3000 ft. The thin layers of reservoir together with an unpredictable
lateral variation in porosity make it very challenging to follow the target porosity. The thickness of the reservoir layer is
predicted to be 5-8 feet along the well path with a target porosity of 10-15% as shown in Fig. 5. However, due to drilling
difficulties at the beginning, the well could not be landed properly. Moreover, the target porosity zone was not encountered as
predicted due to geological uncertainty in the reservoir. As a result the team decided to change the plan and use the help of
3D geocellular property modeling-while-drilling to follow the target porosity.

Fig. 5—Section view of the well at the landing point.

6 IPTC 16715

The workflow for using 3D geocellular property modeling-while-drilling for the H12_2 well started with selecting the
area of interest (AOI) around the well as shown in Fig. 6. Data from five offset wells included in the AOI were used to
supply model updates while drilling. After the initial investigation and QC of the regional 3D geocellular model, a subset
model within the AOI was cut out as shown in Fig. 7.

Fig. 6—Map showing AOI around well H12_2. Fig. 7—Subset model around the planned well trajectory.

Next, the layers of the subset reservoir model were examined thoroughly to identify any miss-tie between the surface
grids and the well picks of the offset wells within the AOI. Any miss-tie is fixed by shifting the layers in the subset model to
conform to the updated picks. The subset model around H12_2 was then refined to a grid size of 5 x 5 x 1.5 ft as shown in
Fig. 8. Note that it is not necessary to perform fine layering and gridding for all reservoir layers. Generally, only the layers
along the planned well trajectory are considered for FLG.

Fig. 8—Fine layering and gridding on the subset model.

After FLG, the well enters the final phase of the 3D geocellular modeling-while-drilling technique. At this stage, as the
drilling operation advances, real-time LWD/MWD log data is added along with real-time pick data (if any). The FLG is then
simulated and populated with real-time porosity log data. The SGS algorithm was used to simulate grid values. A section of
the well trajectory with locally updated porosity values is shown in Fig. 9. This section suggested that the lower part of the
target layer has a porosity of 10-15% for a thickness of 6 feet (circled in Fig. 9). Thus, the plan was updated and the well path
was dropped significantly to follow the porosity. For the remaining part of the well, 10-15% porosity was encountered until
the well reached total depth.
IPTC 16715 7

Fig. 9—Cross-section showing the well trajectory with locally populated porosity values.

In this technique, the subset model updates helped the geosteerer to predict the porosity of the lower part of the reservoir
in advance. Additionally, using the geostatistical modeling technique (i.e. SGS) provided increased confidence in the
prediction of the zone’s porosity. Again, the layer-wise prediction of updated porosity in the FLG subset model indicated the
vertical extent of the target porosity zone in the reservoir. Thus, the risk of placing the well in a non-reservoir zone was
minimized.

Conclusion
For thin and complex reservoirs, using modeling-while-drilling is highly beneficial and results in more accurate well
placement as described in this paper. The examples mentioned here comprised two different geological settings which
illustrates the benefit of modeling-while-drilling over conventional geosteering operations. The log modeling-while-drilling
technique is relatively simple and straight forward. However, 3D geocellular property modeling-while-drilling process is a
complicated one and involves many steps. Additionally, extreme care must be taken when providing correct input data for the
3D geocellular property modeling. Incorporating bad data into the 3D geocellular model could create a completely spurious
property model creating great confusion for the geosteerer.
Around the world, recent development of modeling-while-drilling techniques has significantly helped oil companies solve
many complex well-placement challenges. It is an extremely useful technique for optimizing horizontal well paths,
maximizing well productivity, and reducing lost rig time. In addition, it allows professionals from different disciplines,
whether in the field or geosteering centers, to work together in an integrated manner to reduce well costs and improve well
placement. T
Through the use of these techniques, operators get maximum reservoir contacts and increased production of hydrocarbons
even from very complex reservoirs comprised of very thin layers.

Acknowledgements
The author would like to acknowledge both colleagues and clients for their constructive comments, suggestions, and help in
completing the paper. Thanks are also due to the management of the company where the work was carried out for their
support in publishing the paper.

References
Al Muhairi, L.S., Ribeiro, M.T., Dharmawan, A., et al. 2006. Understanding a Reservoir: 3D Geological Modeling Using Scenario-Based
Approach and Conventional Geostatistics, Onshore, UAE. Paper SPE 101913-MS presented at the 2006 Abu Dhabi International
Petroleum Exhibition and Conference, Abu Dhabi, UAE, 5–8 November. http://dx.doi.org/10.2118/101913-MS.
Al-Jaberi, S., Dama, S., Al-Rayaisaah, J., et al. 2009. Real-Time 3D Modeling and Geosteering in a Heterogeneous Carbonate Reservoir.
Paper SPE/IADC 125881 presented at the SPE/IADC Middle East Drilling Technology Conference and Exhibition, Manama,
Bahrain, 26–28 October. http://dx.doi.org/10.2118/125881-MS.
Bertelli, L. and di Cesare, F. 1999. Improving the Subsurface Geological Model While Drilling. First Break 17 (6): 223–228.
Bourgeois, D., Tribe, I., Christensen, R. et al. 2006. Improving Well Placement with Modeling While Drilling. Oilfield Review 18 (4): 20–
29.
Cark, A., Al Suwaidi, M., Elabed, A. et al. 2010. Challenges for Horizontal Well Placement Optimization in a Giant Mature Onshore
Oilfield, Abu Dhabi, UAE. Paper SPE 137070 presented at the 2010 Abu Dhabi International Petroleum Exhibition and Conference,
Abu Dhabi, UAE, 1–4 November. http://dx.doi.org/10.2118/137070-MS.
8 IPTC 16715

Malinverno, A., Andrews, B., Bennet, N. et al. 2002. Real-Time 3D Earth Model Updating While Drilling a Horizontal Well. Paper SPE
77525 presented at the SPE Annual Technical Conference and Exhibition, San Antonio, Texas, USA, 29 September–2 October.
http://dx.doi.org/10.2118/77525.
Najjar, N.F., Jerome, T., and Alshammery, M. 2005. 3D Geological Modeling-While-Drilling for Geosteering. Paper IPTC 10850
presented at the International Petroleum Technology Conference, Doha, Qatar, 21–23 November. http://dx.doi.org/10.2523/10850-
ABSTRACT.
Ramos, L., Ali, S., Al Maskary, S. et al. 2008. Geosteering Optimization in Thin Carbonate Reservoir. Paper SPE 118139 presented at the
Abu Dhabi International Petroleum Exhibition and Conference, Abu Dhabi, UAE. http://dx.doi.org/10.2118/118139-MS.
Thevoux-Chabuel, H. 2010. Updating the Reservoir Model while Drilling – A New Geosteering Methodology. Dew Journal 19 (06): 47–
54.