Beruflich Dokumente
Kultur Dokumente
in a Mature Field
Abdullah Kasim, SPE, Petronas Carigali; and Frank Wijnands, SPE, and Surej Subbiah, SPE, Schlumberger
Summary reservoirs are prone to sand production and were completed tra-
Although the stacked reservoirs of the Bokor field, offshore Sa- ditionally with gravel packs. In 2003, a full field-review study
rawak, Malaysia, are prone to sand production, the field- identified new development opportunities, including deeper, pre-
development team did not opt a priori for gravel packs in every viously untapped reservoirs.
well. While such completions can indeed eliminate sanding risk, Given new developments in geomechanics and sanding-
the team also wanted to consider the impact of the completion on prediction models, the study team did not opt a priori for gravel
the production rate of a well. The optimum completion not only packs in every well. While such completions can indeed eliminate
excludes sand, but it also maximizes hydrocarbon production. sanding risk, the team also wanted to consider the impact of the
The team carried out a geomechanics and sand-production completions on the production rates of the wells. The optimum
study, using readily available data. The paper gives an insight in completion not only excludes sand but also maximizes hydrocar-
the physics of sand production and how this process can be mod- bon production over the life of the well.
eled geomechanically. It shows how the model is used to select The team seized the opportunity to conduct a sanding-propen-
feasible completions and quantify sand-free production rates over sity study to gain an understanding of sand-production risks and to
the life of the reservoir. optimize completion designs. Geomechanical modeling played an
One important outcome was that screenless completions were important role in the study.
possible in the deeper reservoirs by optimizing perforation orien- In this paper, we aim to give an insight into some aspects of the
tations. Apart from other advantages of screenless completions, physics of sand production and how the process can be modeled
this leads to significant potential increases in sand-free production geomechanically. We explain how geomechanical models were
rates. built from readily available data and how the models were used to
An integrated sand-management process brought together geo- select feasible completions and compute critical drawdown pres-
mechanics, petrophysics, and reservoir and completion engineer- sures (CDP) for various completion scenarios. We discuss how this
ing to truly optimize completions. helps the Bokor-field study team quantify sand-free production
rates, now and over the life of the reservoir, and we comment on
Introduction the role of geomechanical modeling in the integrated, multidisci-
Certain questions have to be answered before completing a well in plinary approach that is needed to optimize completions.
a weak-sandstone reservoir: Will sand production be an issue now,
or as the reservoir depletes? If so, which completion options are Geomechanical Modeling of Sand Production
feasible, and when do they need to be installed? Several techniques Sand production in (weakly) cemented rocks can be seen as a
exist to prevent sand from being produced, from screens, gravel two-stage process. First, the stresses acting at the completion sand-
packs, or frac packs to optimized and oriented perforating, each face need to overcome the strength of the intact rock and cause a
with its own merits. But the choice of completion also has a major shear failure. Then, the produced fluids need to transport the sand
impact on hydrocarbon-production rates, and therefore on the life- grains to the surface. If the failure of the intact rock can be pre-
long economics of a well. To truly maximize sand-free hydrocar- dicted and prevented, then the issue of transport is of no concern.
bon production, an integrated approach is needed. This includes Therefore, our sand-prediction model aims to predict the onset of
reservoir knowledge from petrophysical and geomechanical inter- shear failure of the intact rock.
pretations, reservoir and production engineering, completion de- The main tool we used for the Bokor-field study was the sand-
sign, and implementation. management-advisor (SMA) software (Schlumberger proprietary
If it can be shown that sand production is going to be a prob- software), which is based on a simple linear-elastic model, but it
lem, screens, gravel packs, or frac packs can be run. The choice of achieves results that are very similar to a more complex numerical
sand-exclusion equipment will likely have an impact on produc- analysis involving advanced laboratory-test data (Bradford and
tivity, which needs to be considered. Cook 1994; Morita et al. 1987).
The industry has become aware that oriented perforations can The main advantage of SMA is that it can be run quickly and
prevent sand production in many cases. Oriented-perforation efficiently, using only readily available data from wireline logs and
completions have an advantage that they minimize the cost and existing reports, interpretations, and studies. If first-pass results
complexity of the completion. More importantly, they also gener- indicate that precision is critical, then model calibration can be
ally give higher production rates than a screen or gravel pack achieved with relatively simple laboratory strength tests.
would give in the same well. A mechanical Earth model (MEM) is required to provide the
Clearly, there is a need to predict reliably under which circum- necessary input data to the SMA. The MEM consists of the pore
stances the technique of oriented perforating can prevent sand pressure (PP), the directions and magnitudes of the in-situ princi-
production. Geomechanical analysis can provide the answer. pal stresses, elastic parameters (Young’s modulus and Poisson’s
While no generally accepted models exist today that can predict ratio), and peak-strength parameters [unconfined compressive
rates or total amounts of sand production, we are able to predict the strength (UCS) and friction angle] (Fjær et al. 1992).
onset of sand production with reasonable accuracy. In many cases, SMA indicates where and under which conditions sand pro-
such a prediction is enough to make an informed decision. duction occurs. For a given completion, SMA quantifies sand-
The Bokor field in the Baram delta, offshore Sarawak, Malay- production dependence on reservoir-pressure depletion and draw-
sia, has produced oil since 1982. Its unconsolidated, stacked down pressure (Santarelli and Tronvoll 1998). SMA also shows
where oriented perforations can be a solution, apart from more-
conservative choices such as screens, gravel packs, or frac packs.
Copyright © 2008 Society of Petroleum Engineers
An Integrated Approach
This paper (SPE 102242) was accepted for presentation at the 2006 SPE Annual Technical
Conference & Exhibition, San Antonio, Texas, USA, 24–27 September. Original manuscript
The Bokor-field study team needed to consider the impact on
received for review 6 July 2006. Paper peer approved 10 April 2008. productivity of each feasible completion option. They achieved
Fig. 1—Greater West Baram delta geological setting and structural trend.
Fig. 2—UCS (purple curve) and static Young’s modulus (red curve in right track) from correlations plotted against UCS (red
symbols) and Young’s modulus of elasticity (YME) (green/red symbols) from laboratory tests for Well BO-102 (MD=Measured
Depth).
to perforation orientation. This can show in which cases oriented by optimizing the completion design. Techniques that can be em-
perforations can prevent or delay sand production. ployed to avoid sand production are
Please note that conventional, nonoriented perforations are of- a) Oriented perforating—perforations are oriented at or near
ten shot in a spiral configuration, with individual perforations the most-stable orientation.
phased at 60 or 30° azimuth steps. This configuration has the b) Selective perforating—the weakest parts of the reservoir
disadvantage that there will always be perforations at or near the are not perforated.
least-stable perforation azimuth. Under marginally stable condi- c) A combination of selective and oriented perforating.
tions, a small number of perforations may thus be responsible for 3. Partially manageable sanding risk—There is sand-pro-
all the produced sand. Oriented-perforation completions aim to duction risk, but sand production can be reduced and delayed
place perforations at, or close to, only the most stable orientation. significantly by optimizing the completion design with one of the
Under the right circumstances, this can result in a remarkable techniques mentioned in number three.
increase in stability. 4. Unsafe—Sand will be produced from any possible comple-
In a vertical well, a phasing of 0° means that the perforations tion at the early production stages. The only option is to exclude
are aligned with the maximum horizontal stress. In a deviated well, sand with screens, gravel packs, or similar techniques.
0° phasing orients the perforations in the vertical plane through Table 1 summarizes the sanding risk for the reservoir zones in the
the well, while 90° phasing means horizontal perforations in this Bokor field.
context. The sand-production risk does not depend solely on the reser-
The sensitivity of sand production with respect to parameters voir conditions but also on wellbore orientation. It is possible that
other than orientation (e.g., UCS or stress error, perforation size) an unfavorable wellbore orientation moves a reservoir from Cat-
also can be analyzed. This may lead to a recommendation to per- egory 2 to 3 or even 4 for that particular well plan. Clearly, any
forate selectively (i.e., the least stable parts of the reservoir are not reservoir that falls into Category 2 or 3 needs to be assessed on a
perforated). well-by-well basis. The geomechanical knowledge captured in the
Once it was clear which completions were feasible in a certain MEM ensures that this can be achieved in an efficient manner.
reservoir, the completion specialist computed completion skins for
each feasible option. When combined with the reservoir skin, these Sanding Analysis for A-to-F Reservoirs. The reservoir zones
options lead to a prediction of maximum hydrocarbon production from A to F are very weak in general, with UCS values between
and a maximum drawdown pressure. This process also needs 100 and 600 psi. Sand-free production is not possible in these
petrophysical inputs, for instance net-sand and permeability esti- zones, not even when orienting the perforations in the most stable
mates. Expected reservoir-pressure behavior was available from direction. Therefore, sand-exclusion equipment needs to be in-
reservoir-simulation studies. stalled to keep formation sand from entering the wellbore.
Results and Discussion Sanding Analysis for G-to-M Reservoirs. The UCS values in the
We assessed sand-production risk and computed critical draw- deeper-reservoir zones G to M are between 1,000 to 3,000 psi, and
down for all the reservoir sequences. The analysis included sen- these formations can be categorized as weak- to medium-strength
sitivity of CDP to variations in a number of input parameters—in rocks. The UCS generally increases from Reservoir G through M,
particular, wellbore trajectory, perforation orientation, and pore- and CDP profiles calculated for these zones show the presence of
pressure depletion along with uncertainty in the UCS, horizontal- reservoir intervals that allow sand-free production. The permitted
stress directions and magnitudes, and perforation size. drawdown pressures also tend to increase with depth.
To summarize the sand-production risks in each reservoir There are, however, no reservoirs that can be classified as
briefly, we use the following broad classifications: safe under all possible conditions of drawdown and perforation
1. Safe—Any perforated completion can be produced sand-free orientation.
at the required drawdown pressure and down to the expected res-
ervoir-abandonment pressure. Example 1—Deviated Well Through Reservoir M. Fig. 5 is a
2. Manageable-sanding risk—Certain completions may pro- plot of the CDP over Reservoir M for an existing well, which
duce sand; however, sand production can be avoided completely traversed the reservoir at a deviation of 45° and an azimuth of
228°. The plot shows two CDP profiles, one for oriented perfora- such a plot. For clarity, the diagonal CWP⳱PP is drawn in the
tions phased at 0° (the most-stable orientation—in the vertical plot. We are not interested in the zone above and to the left of the
plane through the well) and the other for 90° phasing (least-stable diagonal. If CWP>PP, there can be no hydrocarbon production.
orientation—horizontal). The drawdown pressures for 0° phasing The curves below the diagonal show how the CWP varies with PP
are higher than those for 90°. The plot in Fig. 5 shows that part of for various values of UCS, starting from the virgin-reservoir pres-
the reservoir (7,850–7,862 ft) could be produced sand-free, if ori- sure of approximately 2,500 psi.
ented perforations are used. We constructed Fig. 6 for an existing well at a deviation of 47°
If the Bokor team wanted a screenless completion in this well, through Reservoir L3, to analyze the sensitivity of CDPs on the
part of the reservoir would have to be left unperforated. Given UCS. In this case, a conventionally perforated completion (Fig. 6a)
reservoir porosity, horizontal and vertical permeability, and skin offers significant sand-free production only where the UCS ex-
and completion details, a production engineer can compare the ceeds 2,000 psi. The UCS-distribution curve for L3 (Fig. 6c)
expected production from such an oriented completion over a lim- shows that this is not the case over the entire L3 reservoir. But
ited zone with the expected production from a screen or gravel-pack Fig. 6b shows that oriented perforations in the vertical plane
completion over the entire reservoir. In this case, a selective, ori- through the well (0° phasing, the most-stable orientation) offer
ented-perforation completion would produce less hydrocarbon significant safe drawdown if the UCS exceeds 1,500 psi, which is
than a gravel pack or screen over the entire reservoir zone. Ori- the case for the entire L3 reservoir.
ented perforating would not be the preferred choice here. Whether the sand-free CDP and sand-free reservoir-abandon-
ment pressure are sufficient to meet the desired production criteria
Example 2—Deviated Well Through Reservoir L3. Fig. 5 depends on other information that is available from petrophysical
showed the CDP at virgin-reservoir pressure, but it did not show study and from reservoir and production engineering. The geome-
what happens when the reservoir depletes with production. It can chanical analysis has delivered the critical piece of information
be instructive to plot sanding-study results in a different way: that allows the completion team to make an informed decision and
CWP as a function of reservoir pressure. Figs. 6a, b, and c present optimize the completion in the L3 reservoir.
Fig. 5—UCS and CDP in Reservoir M for a well at 45° deviation and 228° azimuth. CDP (blue curve) is shown for virgin-reservoir
pressure (approximately 2,800 psi) for perforations phased at 0° (most-stable orientation in vertical plane through well) in the
second track from right and at 90° (least-stable orientation—horizontal) in the first track from right. Green shading denotes stable
drawdown conditions.
classified as unsafe for screenless completions. These reservoirs data from petrophysical analysis and production- and reservoir-
will start to produce sand very soon after production commences, engineering studies. The geomechanical analysis provides the rest
no matter what the wellbore or perforation orientations are. of the information that is required to make informed decisions and
The sand-production risks in Reservoirs G to M were classified to optimize completions.
as partially manageable. Whether sand-free production from In the study, we provided sand-production analysis for repre-
screenless completions is possible in these deeper reservoirs de- sentative existing wells and for a number of generic wells through
pends not only on geomechanical parameters, such as geological the deeper Bokor reservoirs. True completion optimization, how-
stress and rock strength, but also on wellbore orientation. The ever, requires that candidate wells be reviewed on a well-by-well
study showed that Reservoirs G to M can be produced sand-free basis. With the MEM now in place, such a review can be carried
but only if perforations are oriented at or near the most-stable out using a minimum of resources and time.
directions. Of course, the wellbore deviation and azimuth through
the reservoir determine in which plane perforations can be shot. In Conclusions and Discussion
general, it can be said that wells at low deviation angles (less than True completion optimization in wells with sanding risk requires
approximately 40°) need downhole sand exclusion in all the Bokor more than just the exclusion of sand. The completion also needs to
reservoirs. optimize hydrocarbon-production rates. Each completion intro-
Screenless, oriented-perforation completions in Reservoirs G duces a completion skin in the well that can be quantified by a
through M are an option at higher wellbore deviations, with a weak completion engineer. Combined with petrophysical and reservoir
dependence on wellbore azimuth, too. In some of the reservoirs, data, it is straightforward to estimate future production rates for
and depending on wellbore orientation, perforating would also various completion options. In general, screenless completions
have to be selective (no perforating at the weakest intervals). The have lower completion skins than sand-exclusion equipment such
impact of oriented and selective perforating or limiting drawdown as screens or gravel packs.
pressures on hydrocarbon-production rates needs to be quantified Using the MEM built for the Bokor field, we computed for each
to optimize the completion. This is possible with readily available of the stacked reservoirs whether screenless completions can pro-
duce sand-free. For feasible completion options, we also computed have liked to have full-waveform sonic measurements for more-
CDP, not just at the virgin-reservoir conditions but for the life of precise geomechanical correlation, image logs to provide informa-
the field, as a function of reservoir pressure. The study showed tion on downhole stability, rock-failure mechanisms and stress
which completions were feasible in each of the Bokor reservoirs, directions, and more (precise) leakoff-test information to calibrate
depending on wellbore and perforation orientation. The study thus the stress profiles further. On the other hand, we were fortunate to
provided an important part of the information necessary to make be able to carry out mechanical tests on core samples, which
truly informed decisions with respect to completion optimization. matched very well with Young’s modulus and UCS correlations
A certain amount of uncertainty was introduced by the fact that that we used. The maximum horizontal stress is the least con-
the available input data were less than optimum. Ideally, we would strained parameter in the MEM, but this parameter did not have a
Fig. 8—Effects of wellbore-deviation angle on CDP in the K reservoir (UCS=1,250 psi, perforation diameter=1.6 in.).