JUNE 2012

The “Better Business” Publication Serving the Exploration / Drilling / Production Industry

Modeling Optimizes Completion Design

By Tom Bieltz, integrated technologies such as ICDs produce from the thin oil rims. Never-
Alex Obvintsev, along with rotary steerable drilling systems theless, early water breakthrough is still
Paul Riegler, and azimuthal logging-while-drilling well a problem with such horizontal wells be-
Aviral Sharma, placement technology to achieve higher cause of the imbalanced drawdown along
Jeffrey Kok, reservoir sweep efficiency. The well was the horizontal and the close proximity to
and Shim Yen “Yenny” Han sanded in and plugged after five years of the oil/water contact (OWC). In practice,
high water production. A saturation log to increase hydrocarbon recovery from
HOUSTON–High water production is was run in an offset well, which granted these wells, flow rates are increased. This
a major issue in horizontal oil wells, es- the viability to revisit this reservoir. Con- results in drawing higher localized water
pecially in longer laterals, commonly be- sequently, a new offset horizontal well production that can cause “hot spots” in
cause of the high drawdown from the was proposed and drilled next to the ex- the completion string. Mechanical failure
heel. In addition, the presence of hetero- isting well to sweep remaining reserves. of the sand control media often follows,
geneity along the lateral section can lead For the new horizontal offset, full field leading to well plugging.
to uneven sweep of hydrocarbons, which dynamic simulations were performed to To effectively exploit such reservoirs,
can result in poor recovery. To control the evaluate attic placement methodology and it is necessary to geosteer horizontal wells
water production and achieve better sweep the optimization of ICDs into the integrated as far as possible away from the OWC
efficiency, inflow control devices (ICDs) design. The results illustrate that attic with the means to balance fluid influx
have been introduced that balance fluid horizontal well placement is feasible using along the horizontal. Developments in
flux along the producing horizontal well. integrated drilling with ICD technology LWD technology have enabled structure
This article presents a case study of a to maximize the sweep efficiency. or bed boundary mapping in real time to
Gulf of Mexico Shelf horizontal well In a mature field development, opera- accurately place wells below the reservoir
that was completed with prepacked tors are challenged to place wells as high trap or within thin sand columns. Com-
screens, and assesses the results of using as possible to minimize attic traps and pletion technology incorporating stand-

FIGURE 1A FIGURE 1B
Structural Map of the Reservoir and Well Locations North-to-South Reservoir Cross-Section

-5,370
-5,380
-5,390
-5,400

Reproduced for Schlumberger with permission from The American Oil & Gas Reporter www.aogr.com
SpecialReport: Offshore & Subsea Technology

FIGURE 2A FIGURE 2B
Sidewall Core Analysis Modeled Permeability Map
3,750 4,000 4,250
3,500
Permeability
3,250
3,000
2,750

Symbol legend
2,500

Core.txt PhiVsK

0.3375 0.34 0.3425 0.345 0.3475 .035

Porosity

alone screens combined with customized Figure 2A shows the core analysis with permeability ratios.
downhole ICDs can be introduced to correlation and Figure 2B shows the mod- The oil-water capillary zone was neg-
delay or regulate water breakthrough. eled permeability that was eventually lected because of the reservoir’s high
While these technologies individually used in the flow simulations. The high vertical permeability. Because of a lack of
add value to horizontal well production, permeability of the sand brings great pro- pressure-volume-temperature data, fluid
in combination, they provide the needed duction challenges for the wells in this correlations were used to create PVT
solution to optimize recovery and prolong reservoir as a result of the strong aquifer tables. Oil gravity of 29 degrees API, gas
the lifespan of a producing lateral well. at the bottom. This is a matter of significant specific gravity of 0.7, bubble point pressure,
concern with horizontal wells because of and initial fluid contacts were some of the
Anticlinal Reservoir the desired high production rates in a key inputs for the fluid model. A black oil
The reservoir is located on the Outer reservoir with high vertical-to-horizontal model was used for the flow simulations.
Continental Shelf near the Louisiana Coast.
It is a fluvial/deltaic sand deposited in
shallow water. The reservoir is an anticline FIGURE 3
with a sealing fault to the east, as shown in Gamma Ray and Resistivity Log Response
Figure 1A. The north-to-south cross-section
of the sand in Figure 1B shows the anticlinal
structure of the reservoir. Sand thickness
varies from 48 to 77 feet, and is thicker at
the center of the reservoir. The sand is
highly unconsolidated, making sand control
integral to the completion strategy.
The unconsolidated sand is highly
porous with an average porosity of 32
percent. The log data were upscaled to
calculate a variogram, which showed major
continuity in the northeast direction. The
modeled variogram was then used as input
to create the porosity model. The reservoir
model is based on a cell size of 100 x 100
x 2 feet. Sequential Gaussian simulation
was used to create the porosity model.
The permeability of the reservoir was
modeled based on sidewall core analysis,
which was used to find the correlation
between porosity and permeability to
model the permeability of the reservoir.
SpecialReport: Offshore & Subsea Technology

History Matching FIGURE 4

The reservoir has three vertical wells Full-Field History Match of Oil, Water and Gas Rates
and one horizontal. The first two vertical
wells, V1 and V2, depleted most of the

10,000 1,000 100

free gas cap. The horizontal well, H1,

1,000
started producing more than 5,000 barrels
of water and less than 200 barrels of oil a

Liquid Flowrate [STB/d], [STB/d]

10 100

Gas Flowrate [Mcf/d], [Mcf/d]

day after two months of production. The
lateral section of well H1 was completed

10
with prepacked screens that faced me-

1
chanical failures after four years because

1
of possible hot spots in the screens caused

0.1

0.1 0.01 0.001 0.0001

by the high-rate water production.
Recent logging in an offset well re- 0.0001 0.001 0.01
vealed unswept hydrocarbons in place
(Figure 3). Since the hydrocarbon interval
is only 12 feet, the most viable approach
is to place a horizontal well high within 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008
the pay. A new horizontal well, H2, was Water production rate History_Oil Control Gas production rate History_Oil Control Gas production rate Observed
Oil production rate History_Oil Control Oil production rate Observed Water production rate Observed
planned to be placed close to the H1 hor-
izontal well. Two main challenges had to
be addressed for the H2 horizontal well: for a couple years. well. Based on the reservoir heterogeneity,
placing the horizontal well in a 12 foot- The H1 horizontal well was completed the production challenges for a horizontal
thick pay zone and steering it away from with prepacked screens in early 1995. well can be described as friction pressure
the OWC, and delaying early water break- The well began producing at very high loss from toe to heel, and coning at the
through and balancing fluid influx. water rates, but it also had higher oil heel (large drawdown for long wells) in
Even after placing the H2 well at the production compared with the vertical the case of homogeneous reservoirs, and
top of the structure, the recent resistivity wells. Higher water production has been high permeability layers, faults and frac-
log indicated that early water breakthrough an issue in every well drilled in this sand tures, and early water or gas breakthroughs
was inevitable, given that the original because of the high vertical permeability, in the case of heterogeneous reservoirs.
OWC had moved up by 12 feet because and this issue became severe in the case In this Gulf of Mexico Shelf applica-
of the high permeability and strong bot- of the H1 wells. After more than four tion, the reservoir has relatively homo-
tom-drive aquifer. To avoid early water years of production, H1 was plugged geneous sand, making water coning more
breakthrough and high water cuts, an because of sanding from a mechnical likely to occur at the heel of the new hor-
ICD completion design was considered. failure of the completion string. izontal well. Because of the higher vertical
To model the ICDs and their behavior The history match was also done on permeability, high water production was
over time, it was important to understand individual wells to assure the consistency inevitable, even though the H2 well was
the dynamics of the reservoir. Dynamic of the simulation model. After matching planned to be landed on top of the struc-
simulation incorporating pressure, satu- the rates, observed pressure data were ture. To delay/control high water cuts
ration and reservoir heterogeneity was plotted against the simulated pressure. and sand production, ICDs along with
performed to optimize the ICD design gravel packing were planned as part of
for the H2 well. Completion Design the completion design. For the forecast,
Before modeling the ICDs, full-field After calibrating the reservoir model four completion designs were compared
history matching was performed to cali- for pressure and saturations, production to show the value of ICDs.
brate the fluid saturations, pressure and was forecast for the new H2 horizontal The most common ICD has a nozzle-
petrophysical properties. The black oil
simulator was used for the history match- FIGURE 5
ing. The historic oil rate control was used Cell Permeability along the Lateral Section
as a development strategy along with a and ICD Joints Separated by Packers
minimum bottom-hole pressure limit.
Figure 4 shows the results of history
matching analysis for the field. The
simulated historic oil, water and gas rates
were matched with the resepective
observed data. From Figure 4, the first
two vertical wells (V1 and V2) produced
most of the gas cap. Well V3 had high
water production. It started producing in
the late 1970s and was shut in after a
decade of high water production. It came
on line again in late 2007 and produced
FIGURE 6 by swellable packers (Figure 5). The first
Structure Map from Offset Well Control seven ICDs from the heel had the smallest
nozzle sizes of 1 x 1.6 millimeters to pre-
vent early water coning because of high
permeability. The remaining five ICDs,
installed closer to the toe, had nozzle sizes
of 2 x 1.6 millimeters to allow the water to
push oil at the toe section for higher sweep
efficiency. Different nozzle sizes and packer
configurations were simulated along the
horizontal to compare the performances,
and OICD configuration yielded the highest
recovery.
The four completion designs were
simulated dynamically for one year of
production using the same reservoir con-
trols, including reservoir volume controls
and BHP limits. The GP and GICD pro-
duced the same volume of oil and water
after one year of production because the
ICD nozzles in the GICD completion
scenario were all the same size and too
big to control water coning.
In the case of OICD, the cumulative
oil production was increased by 15 percent
in one year compared with the GICD/GP
because of the strategically selected nozzle
sizes based on the permeability and satu-
ration responses. Also, there was no increase
based design, with a nozzle in the front parison. The open-hole lateral section is in the cumulative water production after
and a screen at the back to protect the 8.5 inches in diameter and 750 feet long. one year, indicating that OICD allowed
nozzle from sand and debris. The nozzle The gravel pack completion has a 4.5- the same water production, but it also
diameter can be designed to control the inch screen in a 750-foot lateral packed helped push the oil. The key is to understand
pressure drop across it, and therefore, with gravel in the annulus. In the case of the geology and petrophysics that eventually
the fluid rate. the GICD, the ICD joints are installed help to optimize the ICD configuration.
ICDs create a pressure differential inside the gravel pack screen, with eight An average pressure drop of 170 psi
across the screen such that the annulus ICD joints geometrically distributed and was achieved across the nozzles for the
pressure is greater than the tubing pressure. each joint (12 meters long) separated by GICD completion. The pressure drop
This lowers the drawdown pressure on swellable packers for annulus isolation. across each nozzle is proportional to the
the reservoir. By designing the ICDs cor- The nozzle size of each ICD is 2 x 3 mil- square of the fluid rates. In the OICD
rectly, the drawdown pressure balance limeters (two nozzles three millimeters design, for the nozzles close to the heel,
across the length of the well can be in diameter). a very high pressure drop of more than
achieved. To achieve this balance, suffi- The OICD completion design is similar 300 psi was achieved initially, which
cient pressure is needed at the heel to to GICD, but the number of ICD joints helped choke back the high rate moving
arrive at the same flow rate as the ICDs. and their configuration is designed based fluids into the production tubing from
This pressure can be delivered using a on permeability and saturations. In this the heel section, promoting water to dis-
pump or gas lift. case, 12 ICD joints were installed, separated place oil through the nozzles closer to
In heterogeneous reservoirs where per-
meability varies significantly across the FIGURE 7
lateral, especially in fractured carbonates
where fluid flow along the annulus between Undulated Profile of Well No. 1 Trajectory along Modeled Structure
the completion tubing and well bore will
take effect, packers can be used for seg-
mentation to isolate zones. Unless ICDs
are employed, the flow rate in each section
will be roughly proportional to the section
permeability. By adjusting the flow re-
striction in each section, the fluid flux
can be balanced within the section and
along the lateral from heel to toe.
Four completion designs were simu-
lated: open-hole, gravel pack (GP), GP
and ICD (GICD), and optimum ICD
(OICD). Although open-hole completion
is not practical in unconsolidated sand,
it was used only as a base case for com-
the toe for better sweep. not be accurately represented since the ing response modeling is critical to overcome
reservoir top boundary was modeled with structural and survey uncertainties while
Well Placement undulations to match the log with the the well builds through the curve to land at
For this study, the well placement structure property response. This can be a 90-degree inclination.
challenges were described in a feasibility explained by trajectory divergence from
phase that determined a fit-for-purpose the actual when employing positive dis- Higher Landing Precision
bottom-hole assembly for drilling and placement motors, taking stationary survey Based on offset well evaluations, the
steering measurements for both the landing measurements at fixed distances. overburden can be distinguished and cor-
and lateral. Based on this understanding, the tra- related with gamma ray while the five-
The reservoir was evaluated for well jectory could be more tortuous. Although foot target can be defined only using re-
placement or geosteering off a structure well No. 1 was located high in the struc- sistivity measurements. Therefore, LWD
map (Figure 6). The target sand package ture, the entire lateral was not placed gamma ray and resistivity measurements
was evaluated using a cross-section cor- within the zone. The sumps along the were concluded as the minimum require-
relation of the offset well log data. The lateral may have caused a higher influx ment for the landing phase, while az-
reservoir pay interval was defined as a at the heel, consequently damaging the imuthal bed boundary mapping meas-
seven- to 12-foot zone across these wells, media that could have resulted in sand urements would be the ideal measurements
which was further refined to a five-foot particle breach into the well that prema- for engaging higher landing precision.
drilling target window. The property of turely ended the well’s production life. Since lateral placement high within
the reservoir derived from the log data This assessment clearly justifies using the target for a maximum displacement
was used to generate a measurement re- a rotary steerable system to provide a from the OWC was critical to the success
sponse for populating a preconstructed smoother, less undulating well. Besides of the well, azimuthal resistivity meas-
structure to provide log response simula- drilling and steering needs, near-bit meas- urement or a bed boundary mapper is es-
tions along the proposed horizontal well. urements such as gamma ray and contin- sential. To validate this recommendation,
In addition to generating a response uous inclination along with the closed- the feasibility for engaging the bed bound-
log model for the planned lateral, the loop steering capacity to hold or maintain ary mapping measurement was evaluated
offset lateral (well No. 1) provided the the angle of the well is instrumental to by generating synthetically modeled re-
means to study the structure model and provide a smooth lateral profile for pro- sponses from structure model properties.
properties using a pre-existing lateral to duction and completion success. This was achieved by selecting fit-for-
gain insights into the well’s structural Prior to placing the production hole, ac- purpose azimuthal LWD measurements
profile and production history (Figure curate landing within the five-foot target to simulate the responses for the horizontal
7). The reservoir top boundary depicted column was the initial obstacle. While uti- well across the targeted reservoir. For
by offset resistivity measurements is rep- lizing RSS with gamma ray and continuous this reservoir, the response simulations
resented by the black line with the reservoir inclination measurements at the bit would illustrated that azimuthal resistivity meas-
beneath it. be the ideal BHA configuration, positive urements from a proprietary LWD tool
Based on this model, it is obvious displacement motors with measurements designed with tilted receiver coils could
that the well’s trajectory (drilled with a 30 feet or more behind the bit could be em- help achieve the primary objective of
positive displacement mud motor) undu- ployed with strategic well placement method- placing the well in the optimal distance
lates to weave in and out of the top of the ologies. Continuous LWD and inclination from the OWC.
reservoir. This trajectory undulation may measurements in combination with geosteer- Figure 8 demonstrates that the bed
boundary mapping tool is capable of de-
FIGURE 8 tecting the upper boundary of the target
Azimuthal Resistivity Mapping Distance to the Bed Boundary sand five and a half feet away and the
OWC eight feet TVD away from the well
bore. The actual distances are a function
of a resistivity profile that is computed in
real time, where no variables or constants
are needed to manipulate the bed-to-tool
distances. The ability to map the forma-
tion’s bed boundary facilitates the means
to proactively steer the well. This reduces
Azimuthal resistivity detection view showing top and bottom boundaries well path tortuosity to facilitate the smooth
running of the completion string and
allows the production hole trajectory po-
sitioning objectives to be achieved.
The planned trajectory is placed in a
geometrically updip structure to provide
Azimuthal resistivity canvas view showing 2 boundary mapping model. sensitivity from the conductive bed for
the response analysis. The resistivity
~5.5 ft below the top; ~8ft from OWC
phase and directional curves in the first
two tracks from the top in Figure 8 are
simulated raw measurements from the
bed boundary mapper, while resistivity
and gamma ray are simulated standard
log measurements. The dots in the az-
imuthal resistivity detection view represent
SpecialReport: Offshore & Subsea Technology

direction and distance of the conductive To overcome the identified well place- enables precise lateral trajectory placement
beds that are mapped, with blue above ment and geosteering challenges, RSS high within the target and maximum dis-
and red below the tool. The azimuthal with azimuthal resistivity bed boundary placement from the OWC for optimum
resistivity canvas view shows these beds mapping measurements is the optimal recovery using ICD completions. RSS
mapped for two boundaries, and the insert choice for drilling and steering the curve with azimuthal resistivity bed boundary
illustrates the distances from the tool to and landing sections of the well. Azimuthal mapping measurements is essential to
the bed. resistivity with bed boundary mapping the success of this project. r

TOM ALEX PAUL

BIELTZ OBVINTSEV RIEGLER
Tom Bieltz is a reservoir engineer Alex Obvintsev manages drilling op- Paul Riegler is a geologist with En-
working the East Bay Field for Energy erations for Energy Partners Ltd. In ergy Partners Ltd., working offshore
Partners Ltd. He has 30 years of expe- his 15-year career, he has performed Louisiana. He started his career with
rience in both reservoir and reserve well design, planning, management, su- Exxon in Kingsville, Tx. Riegler holds
evaluations, and spent most of his career pervision and post-analysis for onshore, a B.S. in geology from the New York
with Amoco/BP prior to joining Energy shelf and deepwater international and State College at Oneonta, an M.S. in
Partners as a consultant. Bieltz earned domestic projects. Obvintsev joined En- geology from the University of Rhode
a degree in petroleum engineering from ergy Partners in November 2004, fol- Island, and an M.S. in management,
Pennsylvania State University. lowing eight years with Schlumberger’s computers and systems from Houston
integrated project management division. Baptist University.
He holds an M.S. in petroleum engi-
neering from Freiberg University.

AVIRAL JEFFREY SHIM YEN

SHARMA KOK “YENNY” HAN
Aviral Sharma is as a production Jeffrey Kok is the well placement in- Shim Yen “Yenny” Han is the petro-
engineer for Schlumberger North Amer- structor for the Schlumberger Houston physics instructor for the Schlumberger
ica, and has been involved in multiple Training Center. He began his career as Houston Training Center. In 15 years
production optimization and reservoir a wireline field engineer in France, with Schlumberger, she has worked as
studies for both deepwater and uncon- Trinidad & Tobago, and Argentina. He an LWD and geosteering engineer, as
ventional plays. He joined Schlumberger then worked with Baker Hughes INTEQ well as a petrophysics domain champion
in 2008 as a reservoir engineer, before in geosteering in the Asia Pacific region. focused on marketing, introducing new
moving to his current position. Sharma Kok holds a bachelor’s in mechanical technology, technical support, and de-
holds an M.S. in petroleum engineering engineering from the University of Sal- veloping new products. Han holds a
from the University of Texas at Austin. ford. petroleum engineering degree from the
University Technology of Malaysia.