Sei sulla pagina 1di 43

SPE-184131-MS

Production Optimization Through Voidage Replacement Using Triggers for


Production Rate

Cenk Temizel, Aera Energy; Harun Kirmaci, Consultant; Zein Wijaya, HESS; Karthik Balaji, Anuj Suhag, Rahul
Ranjith, and Minh Tran, University of Southern California; Basel Al-Otaibi, Kuwait Oil Company; Ahmad Al-Kouh,
Middle East Oilfield Services; Ying Zhu, University of Southern California; Cengiz Yegin, Texas A&M University

Copyright 2016, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Heavy Oil Conference and Exhibition held in Kuwait City, Kuwait, 6-8 December 2016.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
Voidage replacement is a key element in displacement processes, not only for keeping the reservoir pressure
at its initial level but also in mitigating surface subsidence in certain fields. Despite its simple definition, it is
a complicated process in reservoir management because of uncertainities involved and lack of all required
measurements due to economical or technical restrictions. Thus, every single decision parameter and their
relative significance in voidage replacement process is important for robust reservoir management.
In general, voidage replacement is achieved where injection is based on production. This study
investigates the case of triggers where the production rate at the bottom hole conditions is predicated on
the bottom hole flowing conditions or reservoir gas injection rate. A full-physics commercial reservoir
simulator is coupled with robust optimization software, where a miscible flood operation is modeled with
a group bottom hole flowing target coupled with voidage replacement gas and water injection targets.
The simulation results of this realistic case is presented in a way to show the relative significance
of each operational parameter, which is outlined with tornado charts to serve as a guide in decision
making in efficient reservoir management where voidage replacement is a crucial component. It is observed
that triggers help to better manage voidage replacement, especially in large reservoirs where reservoir
surveillance is a challenge due to number of wells and patterns. The results can be scaled up to different
size of reservoirs and patterns with similar recovery processes.
This study scrutinizes the feasibility of a reversal of the typical scenario where injection is based on
production. Thus, it serves as a useful and realistic example for efficient reservoir management through
optimization of voidage replacement through triggers for production rate.

Introduction
Miscibility is defined as the ability of forming a uniform phase from multiple substances. The substances
involved in the process will mix in all proportions, and the interface between two phases will not exist any
more. On the other hand, if two fluids cannot form a single homogeneous phase, they can be regarded as
immiscible. Accordingly, there are miscible and immiscible (waterflooding) floodings that are applicable
2 SPE-184131-MS

in oil recovery where the application of those depends on the requirements of products and reservoir
conditions.
Miscible flooding is a process that involves gas injecting into the reservoir. It is usually used as a
secondary recovery process to collect the remaining amount of the oil after water injection. Waterflooding
can be very efficient after which the oil saturation inside the reservoir decreases from 70% to 44%, and the oil
recovery is about 37%. However, displaced water may lead to blockage and isolation of some portion of the
oil. On the other hand, the effect of capillary pressure will also impede movement of oil and make oil stuck
in place that would make the relative permeability of oil zero. This sets up as a strong reason to put forward
miscible flooding. The procedure involves the following mechanisms: extraction, dissolution, vaporization,
solubilization, condensation, or other phase behavior change involving crude oil, viscosity reduction, oil
swelling and solution gas drive. The significance of the process is its providing a stable reservoir pressure,
since interfacial tension between oil and water is reduced via gas injection. Mobility ratio, the ratio of the
viscosities of the miscible solution and the displaced oil, plays a crucial role in the analysis of recovery
efficiency. When single phase is obtained the trapped oil, due the existence of capillary forces caused by
surface tension at the water-oil interface, can be removed. Miscible flooding can enhance the oil recovery up
to 18% after waterflooding. This proves that miscible flooding cooperates well with immiscible flooding,
and can reach pore level recovery efficiency. However, the efficiency of miscible displacement is limited
to the minimum miscibility pressure (MMP). When the standard is satisfied, 95% of oil in the well can be
recovered with 1.3 pore volumes of fluid injected.
Miscible displacement can be classified into two types: (i) first-contact miscible (FDM) displacement,
which forms only one phase for mixing two substances in any proportion, and (ii) dynamic miscibility,
which occurs when miscibility with oil fails at first contact, but on composition alteration of injected fluid
and oil it can be made miscible.
There are various gases and liquids that can be used as miscible displacement agents such as low-
molecular weight hydrocarbons, organic alcohols, ketones, refined hydrocarbons, condensed petroleum gas
(LPG), liquefied natural gas (LNG), carbon dioxide, air, nitrogen, exhaust gas, flue gas and mixtures of
these. Technical feasibility and economical factors are considered through selection of these agents.

CO2 flooding
Carbon dioxide (CO2) can be employed as a low cost and abundant flooding gas since it can be both
collected from nature and artificially produced. It reduces the viscosity and density of oil to increase the
fluid mobility, thereby reducing the interfacial tension at the fluid-oil interface. CO2 acts as a thinning agent
in the mixing process, and descends the residential oil saturation to zero. The procedure is depicted in Fig.
1, and summarized below;
SPE-184131-MS 3

Figure 1—Schematics of CO2 Flooding

• Oil and gas inside the reservoir mix with carbon dioxide and flee out along the CO2 flood to target
wells.
• When the mixture is collected on the ground, oil is separated by a relevant technology such as
fractional distilation and refinery facilities.
• The separated CO2 is injected back for the next cycle.

As previously mentioned, there are two types of miscibility in the procedure: first contact and multiple
contact miscibilities. The former describes the case when CO2 and oil mix and form a single phase, while the
latter is pronounced when there is a formation of more than one phase. Due to the concentration gradient of
CO2 throughout the reservoir upon injection, it is more difficult to deal with multiple phase behaviors, each
of those are related to multiple factors such as temperature, pressure and fluid properties of the reservoir.
Miscible CO2 oil recovery is a multiple contact process that involves interactions between the injected
CO2 and the oil in the reservoir. During this process, light oil fractions are vaporized into CO2 phase and
CO2 moves into the oil, which is known as process of mixing. It ultimately reaches a homogeneous phase,
and a favorable situation with low viscosity, enhanced mobility and low interfacial tension is achieved.
Carbon dioxide can also be injected as an immiscible fluid, although the former is more useful since
it can reduce the viscosity of oil. Immiscible CO2-assisted oil recovery is used when there is insufficient
reservoir pressure. During the process, heavy oil swells as the CO2 dissolves inside it. The carbon dioxide
increases the volume of the oil and the reservoir pressure to drive the fluid, and also lower the viscosity of
heavy oil, whereas light hydrocarbons are extracted to CO2 phase.

Waterflooding
Waterflooding is a type of immiscible flooding used in oil fields for enhanced oil recovery. CO2 is replaced
with water in this process to collect oil from the pores. Reservoir pressure increases to its initial level and
remains stable at this level due to waterflooding; hence the oil is pushed towards the collection facility (Fig.
2). The pressure drop caused by oil extraction is balanced by further water pressure and higher amounts of
oil can be recovered, which is called "voidage replacement". The efficiency of waterflooding relies on oil
properties such as oil viscosity and rock characteristics. Thus, different strategies should be applied to the
oil reservoirs having different geographic characteristics. In a typical reservoir nearly 30% of the oil can be
4 SPE-184131-MS

extracted via primary recovery, but water injection increases this percentage and maintains the production
rate of the reservoir.

Figure 2—Schematics of waterflooding

There are several sources of injection water including produced water, seawater, aquifers and river water.
Since these sources contain substances that might influence the process, the water needs to be purified
before it is used for EOR. Initially, visible particles are filtered, and then the filtered water is sent to a
deoxygenation tower, where it comes in contact with a dry gas stream. Sometimes, an oxygen scavenging
agent should be added to the deoxygenation tank to enhance purification.

Voidage Replacement Ratio


Voidage Replacement (VR) is the process of replacing oil, gas and water in the reservoir by fluid injection.
Voidage Replacement Ratio (VRR) cen be defined as the ratio of reservoir barrels of injected fluid to
reservoir barrels of produced fluidsas in Eq.1:

(1)

(2)

where,
Bw = Water Formation Volume Factor
Iw = Injected Water Volume
Bg = Gas Formation Volume Factor
Ig = Injected gas Volume
Bo = Oil Formation Volume Factor
Qo = Produced oil Volume
Qg = Produced Gas Volume
Qw = Produced Water Volume
GOR = Gas Oil Ratio
SPE-184131-MS 5

Instantaneous VRR can be calculated using the produced fluids and injected fluids over any specific
period of time, via time specific GOR. The average VRR can be calculated over a cumulative basis of
injected and produced fluids over time, with GOR averages. Fig. 3. compares these two types of calculations.

Figure 3—Cumulative and Instantaneous VRR over a Sample Data Set

If the VRR for a month is greater or equal to a value of 1, reservoir pressure is said to remain in balance
or increased. The reservoir pressure is said to have declined when the VRR is below 1. The reduction of
VRR may cause severe consequences such as, subsidence, compression and reduction of permeability. Most
regulatory authorities require the VRR to be close to a value of 1. On the other hand, we will discuss in the
next section the cases at which maintaining the VRR at 1 might not be the optimal solution.
Voidage Replacement Ratio for Waterfloods in different type of Reservoir Models. Heavy oil
waterflooding performs different from conventional light oil waterflooding. Brice & Wood, (2014)
conducted a waterflood operation, different simulation models based on 30 Canadian heavy oil wells.
Waterflooding was initiated after primary production resulted in 1.5-2.5% of the original oil in place (OOIP)
with a VRR of 0.95. Cumulative VRR for the field varied from 0.93 to 1.11 for the pattern. The field also
had major water channeling problems.
Brice & Wood took a heavy oil waterflood with no MBE for their 1st case as shown in Fig. 4 The model
resulted in a VRR lower than 1 at the beginning of the waterflooding, giving a 21.1% recovery over 30 years.
If the reservoir had a primary production for a year followed by a waterflood with a VRR of 1, it resulted
in 18.4% recovery over 30 years. Similarly, a cyclic waterflood of VRR lower than 1 resulted in a 18.2%
recovery in 30 years and 17.1% recovery for a waterflood with a VRR of 1 from the start of operations.
This shows that a VRR of lower than 1 is good at the start of waterflood operations.
6 SPE-184131-MS

Figure 4—Heavy Oil Waterflood with no MBE (Brice & Wood, 2014)

In the 2nd case, they investigated the effect of critical gas saturation on VRR with no MBE. They had six
cases run for 30 years with following recovery results as shown in Fig. 5, for varied saturations of trapped
gases at different VRR. The general observation indicated that a VRR of less than 1 is beneficial leading to
larger trapped amounts of gases, and higher recovery. Also, when VRR is equal to 1, critical gas saturation
has minimal effect on recovery from waterfloods.

Figure 5—Effect of Gas Saturation on Recovery Factor in heavy Oil Waterfloods (Brice & Wood, 2014)

The 3rd case run by the same group involved an optimal heavy oil waterflood, after an MBE, for a cyclic
injection/production. Five cases were run for recovery rates in this model involving cyclic or continuous
waterfloods, with varying VRR. It is observed that if an MBE is developed in a heavy oil waterflooding,
optimal recovery is obtained with a cyclic injection/production with a VRR < 1 (Fig. 6).
SPE-184131-MS 7

Figure 6—Recovery Factor for heavy waterflooding with MBE (Brice & Wood, 2014)

The last case was simulated for a light oil waterflood with no MBE. A simulation was run for obtaining
the recovery factor of six cases with different starting dates of waterfloods and VRR. The simulation shows
that a year for primary production is good in a reservoir prior to waterflooding, and VRR of 1 is good for
light oil reservoirs. Cyclic production is not favorable in these reservoirs. These can be seen in Fig. 7.

Figure 7—Recovery Factor for Light Oil Reservoirs (Brice & Wood, 2014)

Effect of VRR on Recovery Efficiency and WOR/GOR for Different Crude Samples. Tae Wook &
Vittoratos (2016) conducted an experiment with two different oil samples, A1 and A2, at varying densities,
viscosities and compositions, to check the effect of VRR on their recovery efficiency, water-oil ratio and
gas-oil ratio.
The centerpiece of the apparatus was a sandpack housed in a temperature controlled, high pressure,
Hassler type aluminum case holder. Front and end plugs were made of stainless steel. Sandpack
permeabilities were in the range of 1.1 to 1.2 Darcies with a porosity of 35%. Clean Ottowa sand with a grain
size in a rangle of 75 to 420 µm was used. Overburden pressure was provided by a high pressure syringe
8 SPE-184131-MS

pump. X-ray CT scanning verified homogenous packing of sand. Syringe pumps at the sandpack inlet and
outlet were used as an injection system, to obtain the required VRR. A high pressure viewing cell at the
core exit permitted characterization of texture of the produced fluid. The appartus set up is shown in Fig. 8.

Figure 8—Apparatus for VRR Oil recovery, WOR and GOR testing (Tae Wook & Vittoratos, 2016)

In the first experiment oil recovery is measured for varying PV injected at different VRR, for both the
crude oil samples in the apparatus. Final Recovery with crude A1 at VRR = 0.7 is over 15% greater than
that of VRR = 1. Results for crude A2 with VRR = 0.7 were identical to the sum of oil recoveries for both
solution gas drive and waterflooding. For both systems a VRR = 0.7 proves better than a VRR = 1.
SPE-184131-MS 9

Figure 9—Oil Recovery and PV injected for different Crude Samples (a) A1 and (b) A2 (Tae Wook & Vittoratos, 2016)

In the second experiment, a comparison of produced GOR and WOR at VRR=0.7 and 1 vs injected pre-
volume for both crude samples are checked. The produced GOR and WOR history depends on the VRR,
as well as the breakthrough time for both gas and water production during depletion tests. At VRR = 1, the
total GOR produced increased continuously after water breakthrough due to the extraction of methane from
crude oil and subsequent dissolution in the injected brine phase.
10 SPE-184131-MS

Figure 10—GOR and WOR vs. Injected PV for Different Crude Samples (a) A1 and (b) A2 (Tae Wook & Vittoratos, 2016)

Voidage Replacement Ratio in Steam Stimulation. The Issaran Field, located in Eygpt, is 290 km southeast
of Cairo and 3 km inland from the western shore of the Gulf of Suez, covering 20,000 acres. Basta & Korany
(2016) express a field development strategy, using VRR as a factor in Steam Management, in this field.
Basta & Korany considered two formations for thermal EOR. The Zeit Sand formation located towards
south east of the field, is a highly unconsolidated sandstone with oil of 10-12 API and a depth of 700-800
ft. A plot of the VRR during cyclic and continuous steam injection is shown in Fig. 11. During cyclic steam
injection, the VRR dropped to 0.35 and after 2% PV injected, continuous steam injection was employed.
After 9 months the VRR stabilized at 0.6.

Figure 11—VRR Performance for Cyclic and Continuous Thermal EOR for Zeit Sand (Basta & Korany, 2016)

The second formation considered by Basta & Korany is the Upper Dolomite South Formation, which is
characterized as a depleted fractured dolomite reservoir. The reservoir has a depth of 1000 ft with an average
thickness of 400 ft and oil has a range of 9-12 API gravity. A plot of VRR during cyclic and continuous
thermal EOR is shown in Fig. 12. In the South Upper Dolomite, once the VRR was below the optimum
value of 0.7, cyclic steam injection was replaced with continuous injection (5% PV injected). Continuous
steam injection took 2 months to stabilize.
SPE-184131-MS 11

Figure 12—VRR Performance for a Cyclic and Continuous Thermal


EOR for Upper South Dolomite Formation (Basta & Korany, 2016)

Formulae
Eq. 1 to calculate VRR can be rewritten as,

(3)

where Vinj is the injected fluid volume and Vprod is the produced fluid volume (both are measured at reservoir
conditions). In a typical waterflooding system operating at a pressure higher than the saturation pressure,
the formula is expanded to,

(4)

In an immiscible gas flooding system, the VRR equation becomes,

(5)

Empirical VRR observations are not frequently reported in literature. If nearby water aquifers are present,
flow conditions at the boundary of the hydrocarbon reservoir will be difficult to define, and therefore, the
associated material balance will carry possible errors along (Vittoratos and West 2010).
The calculation of VRR in volatile oil and retrograde gas condensate reservoirs is not straightforward
using the conventional material balance method shown above. As Clark & Ludolph suggested, reservoirs
that exhibit widely changing compositional effects, such as ones with thick pay and large compositional
gradients in thrust belt regions, will pose challenges to calculate VRR accurately due to uncertainty in
the formation volume factor (FVF) measurement. Besides depth variation, the FVF used in VRR equation
can change immensely area-wise in compartmentalized reservoirs and time-wise due to gradual reduction
of reservoir pressure. Last but not least, injected gas composition may differ from produced hydrocarbon
composition and some injected gas is soluble in the hydrocarbon system, leading to difficulties in fluid
mass balancing.
To tackle this challenge, Clark & Ludolph proposed a workflow incorporating a compositional full field
model (CFFM) and an external EOS program. In this approach, VRR is calculated by the following formula,
12 SPE-184131-MS

(6)

Where, Gxs, is the excess gas to balance voidage. Each volume region in the CFFM is assigned a pressure
target for predictive calculations. Gxs will be evaluated for each model cell and scaled up, to obtain full
field application. This value already takes into account the operational limitations and reflects the optimal
operating conditions to achieve highest ultimate oil recovery.

Application to Heavy Oil Waterflooding (Experimental approach)


The industry has long followed a general concept of balancing injection with production (VRR=1) in both
light oil and heavy oil waterflooding. This not only stems from abiding to regulatory requirement, but
also mitigating reservoir subsidence problem (Vittoratos et al. 2011). Waterflooding operation historically
undermined incomplete voidage replacement (VRR<1) because of a few apparent reasons. As VRR is
smaller than 1, reservoir pressure decreases below the bubble point and solution gas is released, thus
increasing the remaining oil viscosity, decreasing the oil mobility and at the same time depleting the reservoir
energy. However, laboratory findings and numerical simulations in heavy oil systems show that periods of
under-injection could positively enhance oil recovery thanks to additional benefits of water-in-oil emulsion,
solution gas drive and foamy oil effect.

Heavy oil chemistry


Heavy oil waterflooding may require a different approach to evaluate optimal flooding management due to
different oil chemistry compared to light oil. Besides laboratory observations, a signature prolonged plateau
of production water oil ratio (WOR) around the value of 1 in many heavy oil developments around the
world signifies a possible formation of water-in-oil emulsion (Vittoratos and West 2010). They suggested
that the emulsion creation is partially facilitated by the shear of exsolved gas due to incomplete voidage
replacement (VRR<1).
Emulsion growth within the reservoir stabilizes and improves the areal sweep of injected water, as shown
in Fig. 13. Usually, the viscosity of newly formed water-in-oil emulsions is higher than the heavy oil itself, so
that the incoming injected water preferably moves laterally towards unemulsified oil, rather than to displace
the emulstion forward, thus broadening the displacement front (Vittoratos and West 2010).

Figure 13—In-situ emulsion formation stabilizes that displacement front (Vittoratos and West 2010).

At the wellhead, heavy oils are generally observed to be foamy emulsions and there is growing belief
that emulsions not only form along the wellbore at high shear locations (i.e. pumps and valves) but also
within the reservoir sands. In addition, for oils within a certain API range, gas bubbles are developed within
the oil and greatly increase oil mobility and recovery.
SPE-184131-MS 13

Solution gas drive


According to Vittoratos & West, optimal voidage strategy in heavy oil systems could refer to a rather
heuristic equation,

(7)

A percolation model of water breakthrough for a waterflood (King et al., 2002; Li et al. 2009) supports this
equation. They suggested that the backbone of the percolating cluster at water breakthrough is swept with
VRR=1, while recovery from the dangling ends requires activation of solution gas drive through periods
of VRR<1. To maximize the benefit of solution gas drive, the pressure should not reduce too much below
the bubble point. According to Vittoratos and West, VRR should stay in the range from 0.5 to 1 to receive
much of the benefit of solution gas drive, while still experiencing decent water displacement. The extent of
solution gas drive assistance depends mostly on the oil chemistry. The exsolved gas stays within the oil as
small gas bubbles, forming a foam and effectively increasing the critical gas saturation, hence allows for
more effective displacement of the oil (Fig. 14) (Vittoratos and West 2010).

Figure 14—Flow Schematic Setup (Vittoratos et al. 2011)

To prove the effectiveness of selective periods of under injection, Vittoratos et al. established an apparatus
to simulate heavy oil waterflooding under various operating conditions, varying the value of VRR from 0
to 1. To ensure reproducibility of the experiment and simplify the flow initial conditions, a water-channel
communication pathway between injectors and producers is formed, which represents the stage of water
breakthrough. As Fig. 15 and Fig. 16 indicate, the waterflood with VRR=0.7 recovers more oil than one
with VRR=1. Other performance indicators such as, oil production rate, WOR, GOR etc. also show that
after the creation of a communication pathway, VRR<1 outperforms VRR=1.
14 SPE-184131-MS

Figure 15—Upon Water Communication between the Injection and Production Ports,
Laboratory Waterflood with VRR<1 has the Highest Recovery (Vittoratos and West 2010)

Figure 16—Comparison of Cumulative Oil Produced of Different Tests with Changing VRRs (Vittoratos et al. 2011)

This could be explained by the assistance of solution gas drive in displacing unswept oil in dead-end
porosity volumes, after the oil has been swept in the main channel by the imposed pressure gradient of
water injection. Furthermore, Brice & Renouf complied data of over 100 heavy oil waterfloods to verify
fields that have been under-injected at some point during their lifetime have recovered more oil than fields
that have been kept VRR=1. However, when free gas becomes preferentially mobile and starts to flow as a
continuous phase, the aforementioned positive effects in oil recovery quickly diminish.
SPE-184131-MS 15

Application to Heavy Oil Waterflooding (Numerical Simulation approach)


Delgado et al. investigated the parameters that may play a role in determining optimal VRR of a numerical
simulation model. Different attributes such as relative permeability curves (i.e. Corey exponent), critical
gas saturation, oil chemistry and reservoir heterogeneity are properly varied based on extensive literature
review. Among these attributes, gas mobility and reservoir heterogeneity are the two most defining factors
in evaluating the optimal VRR at core scale.
Due to the diffusive nature of pressure, value of oil recovery will not exhibit abrupt changes for adjacent
values of VRR ranging from 0 to 1 (Delgado et al. 2013). Therefore, relationship between VRR and recovery
can be interpolated as a smooth trend, among discreet values of VRRs and corresponding oil recoveries.
As shown in their results, liquid relative mobility as a function of gas saturation has little effect on oil
recovery for different cases of VRRs. On the other hand, if gas mobility is small (either due to high critical
gas saturation or large relative permeability Corey exponent, ng), gas will be retained longer in the system
and will therefore have longer time to stabilize the pressure and displace the oil, resulting in higher recovery.
Delgado et al. showed that as ng increases, optimal VRR converges to the value of 0.9 (Fig. 17). Besides,
as Sgc increases from 5% to 10%, the optimal VRR shifts from 1 to 0.9 (Fig. 18).

Figure 17—Recovery as a function of VRR for different values of the Corey gas exponent (Sgc=5%) (Delgado et al. 2013)
16 SPE-184131-MS

Figure 18—Recovery as a function of VRR for higher critical gas saturation (Sgc=10%) (Delgado et al. 2013)

Under special circumstances; appearance of gas phase, the interfacial tension between oil and water
weakens and leads to oil/water emulsion. Emulsion formation may alter phase viscosity immensely and
contribute to positive incremental recovery (Delgado et al. 2013). Oil emulsions reach optimal recovery at
VRR~0.9 (Fig. 19). Another effect of oil chemistry in three-phase co-existence is the possibility of foamy
oil creation. When an oil with particular properties is below the bubble point, gas is formed as bubbles
within the oil, which stay in the rock matrix, trap the evolved gas and allow the less viscous oil to flow more
easily (Delgado et al. 2013). Foamy oil achieves highest recovery at VRR~0.7 (Fig. 19).

Figure 19—Recovery as a function of VRR at 0.6 days for five cases developed from real data (Delgado et al. 2013)

Another factor that is of tremendous importance in deciding the optimal VRR is reservoir heterogeneity.
Here Delgado et al. targeted the permeability variation within the reservoir, particularly the "cul-de-sacs"
or small areas of permeable matrix but not readily connected to the main pathway (Fig. 20). The magnitude
of oil recovery for each case is much lower compared to the homogeneous matrix mentioned above (Fig. 21
vs Fig. 19). In both cases, the optimal VRR for foamy oil and high critical gas saturation is less than 1. In
SPE-184131-MS 17

conclusion, the lesser mobile the gas is and the more compartmentalized the system is, the more favorable
a VRR below 1 becomes (Delgado et al. 2013).

Figure 20—I-J slice of models with Cul-de-sacs (Delgado et al. 2013)

Figure 21—Recovery as a function of VRR at 0.6 days for five cases in a model with cul-de-sacs (Delgado et al. 2013)

Production Optimization
Production optimization plays a major role in developing the hydrocarbon reservoir not only in terms of
maximizing production from the field, but also minimizing the cost per barrel. This results in increasing
profits for operating companies in the oil industry. It can be claimed that with the current advances on
production technologies, applying suitable production optimization methods based on operational and
reservoir management requirements are no longer an option. Understanding the production system, which
includes many components such as, reservoir, wellbores, production conduits, wellheads, Xmass trees,
flow lines, treatment facilities etc. is the first step in order to be able to determine better optimization
methods to maximize the profitability of operating companies. Fig. 22 illustrates the simplified hydrocarbon
production system with its principal components. It is obvious from the figure that hydrocarbon production
from the reservoir is limited to reservoir conditions and deliverability, pieline network, production conduit,
handling capacity of the facilities, operational and economical strategies, or a combination of a few of these
constraints.
18 SPE-184131-MS

Figure 22—Simplified hydrocarbon production system (Heriot-Watt University).

According to the main goal of operating companies and reservoir management strategies, some of the
below listed production optimization methods and approaches can be applied to reach the targeted result:

• Nodal analysis

• Determination of optimum production conduit sizes

• Applying artificial lift

• Determination of optimum lift gas rates and production rates

• Determination of well connections to flow lines

• Selection of optimum well completion type

• Appliying well stimulation

• Optimization surface network system and facilities design such as choke size, separator, and flow
line size etc.
SPE-184131-MS 19

• Reducing downtime of wells and surface facilities

Production optimization techniques shall be selected based on reservoir/fluid charecteristics, reservoir


management strategies and production system components. The chosen production optimization method
can give better results in a specific reservoir and surface network equipments, although it may result in
failure depending on the production system of the interest.
Any kind of effort to increase the production and lower the cost per produced barrel of oil can be called
as production optimization. However, it is obvious that applied techniques for production optimization may
vary one reservoir to other. Nowadays, waterflooding is one of the widely used techniques for increasing oil
recovery. However, waterflooding is not an easy method to apply and application is not straightforward for
every reservoir type. An understanding of reservoir charecteristics and fluid properties is a challenge and
must be known before beginning the operation. On the other hand, determination of the right VRR, injection-
production strategy and well pattern for injectors and producers are other difficulties that require significant
efforts such as, conducting series of laboratory tests, field tests, if possible 4D seismic, interference test etc.
Considering the fact that there are many uncertainities and unknowns on the very challenging waterflooding
operations, determination of optimum VRR is difficult. Also, finding every single unknown and reducing
the uncertainities is a costly operation in the oil industry. Besides, there is no common strategy that works for
every kind of reservoir, due to their uniqueness and requirement of special treatment for optimum recovery.
It is worth to note that a conventional waterflooding practice applied worldwide is to maintain VRR as 1.0,
despite the fact that it is not an optimum practice for every type of reservoir. As discussed priorly in this
paper, heavy oil waterflood strategies should be applied differently than light oil waterfloods. However,
there are some common practices that are valid for similar reservoirs. In this context, it can be generalized
and claimed that optimum recovery can be obtained with fully replace voidage, i.e., VRR = 1.0, for light oil
waterflooding, although it is not valid for waterflood in heavy oil reservoirs (Vittoratos et al., 2014). There
are many studies in the literature that shows VRR = 1.0 is not an optimum value for optimizing recovery
in heavy oils. On the other hand, injection strategy has great influence on optimizing recovery especially
in waterflood operations. For instance, cyclic injection strategy can give a better result than continuous
injection-production throughout the reservoir depletion life. In cyclic injection-production strategy, water
is injected while production is shut-in and production is maintained while injection is shut-in. According to
the study by Brice et al., field tests, simulation studies and statistical analaysis suggest that cyclic injection-
production strategy and VRR less than 1.0 gives optimum production in a heavy oil waterflood process
(Brice et al., 2014). In addition, according to Vittoratos et al., similar results were obtained with studies
by Brice et al. According to Vittoratos et al., a conceptual oil recovery vs. VRR curve is drawn, and the
optimum recovery value for a heavy oils and light oils is obtained as shown in Fig. 23.
20 SPE-184131-MS

Figure 23—Conceptual Oil Recovery vs VRR curve for Heavy Oil and Light Oil (Vittoratos et al., 2014).

As stated above, production optimization techniques vary from one reservoir to another. For instance,
water flooding in a naturally fractured reservoir requires some extra caution, for the early water breakthrough
problem. Therefore, choosing the right well pattern is another important optimization technique that can
improve recovery of oil from the reservoir, especially in naturally fractured reservoirs. It is a well-known
fact that natural fractures enhance the permeability along fracture trend. Therefore, placement of injectors
and producers on naturally fractured reservoirs should be aligned as alternative rows to the fracture trend
in order to delay early water breakthrough (Dittaro, et al., 2007).
Vertical wells as injectors and producers are commonly used in waterflooding projects. On the other hand,
horizontal wells are promising not only in terms higher injectivity but also higher productivity. Therefore,
chosing the right well type for producers and injectors is an important optimization issue to increase
sweep efficiency and thus, oil recovery in waterflood projects. The placement of horizontal injectors and
producers parallel to each other is critical to increase the efficiency of waterflood operations, irrespective
of permeability isotrophy charecteristics. In addition, optimum length of the horizontal producer is another
important parameter to increase recovery. However, it is worth noting that recovery does not linearly
increase with the length of the well. Besides, mobility ratio of the reservoir is an important concern in
determining the optimum well type. For instance, in low mobility ratio reservoirs, having horizontal injectors
is more important than having horizontal producers. On the contrary, in high mobility ratio reservoirs,
horizontal producers are more dominant than the horizontal injectors in terms of the waterflood operation
performance (Algharaib and Ertekin, 1999).
Another production optimization technique that is widely applied in the oil industry is the use of smart
well technology. Considering the possibility of having production and injection from the layers that have
different permeability values, smart wells present a great opportunity to optimize production-injection
schemes. This technology provides the opportunity to eliminate the effects of early water breakthrough
on high permeability zones in waterflood operations. Even if the optimum VRR value is determined in
a reservoir having different permeability zones, expected high oil recovery might not be obtained due to
early water breakthrough resulting from the higher permeability zones (Meshioye et al., 2010). As it can be
seen from Fig. 24, in a waterflood operation on communicating reservoir units, injection can be controlled
throughout flow control valve. Thus, desired zones can be flooded to increase sweep efficiency.
SPE-184131-MS 21

Figure 24—Sample of Communicating Reservoir (Lake, 1989).

The effectiveness of flood operations is dependent on continuous monitoring, evaluation and


improvement of reservoirs. Analyzing the reservoir and production performance is commonly carried out
with dynamic reservoir simulations and analytical approaches. Newly developed technologies have a great
contribution to reservoir evaluation and performance monitoring studies. In this context, 4D seismic is a
promising technological advancement in recent years for monitoring fluid saturations distribution in flood
operations. Although 4D seismic operations may not be cost effective for small sized fields, it can be
considered as an important optimization technique for increasing the efficiency of waterflood operations in
relatively big sized reservoirs (Hadi et al., 2013).
Consequently, there are numerous techniques suggested in the literature for optimizing production.
However, hereby we have provided a few methods that are widely accepted in the oil industry for production
optimization, especially those that are playing a vital role in voidage replacement operations.

Smart Technologies in Reservoir Monitoring


As the oil industry reaches the saturation, it must explore more efficient methods to uilize the diminishing
resources. The development and exploitation of oil and gas reserves in increasingly difficult operating
environments are fraught with numerous technical challenges. This will lead revolutionary approaches to
downhole monitoring and reservoir management. Key driver to the development of smart field technologies
is economics. Some reservoir developments are particularly sensitive to economics due to geological
complexity, reservoir size, water depth and remoteness of location. This will make the cost of well
intervention prohibitively expensive. The real measure of smart field technologies is its ability to centralize
monitoring, management of smart/intelligent wells with an aim of implementing and executing reservoir
management strategies. Smart/Intelligent wells, themselves are defined as wells which are equipped with
devices enabling remote monitoring, control and transmission of data from multiple zones thereby providing
capabilities to optimize production, improve reservoir management and reduce intervention costs.
To better exploit advanced smart field capabilities, smart wells must have timely, accurate and reliable
reservoir performance data. These data are obtained by another critical subsystem of the intelligent
completion and downhole monitoring system. Permanent monitoring and control systems add "intelligence"
to the system and are installed to provide real time contionuous data and control without interrupting
production, or requiring well intervention. By combining real time data monitoring with remote controlled
sliding sleeves or variable chokes, the operator can proactively manage production optmization through
22 SPE-184131-MS

voidage replacement, eliminate or lower the interventions and associated production downtime, well
costs and risks. Schlumberger defines smart field technology process as a well-equipped with monitoring
equipment and completion components that can be adjusted to optimize production, either with some
operator intervention or automatically. (Fig. 25)

Figure 25—Schlumberger’s model for Smart Field Technology process (courtesy of Schlumberger)

Basically smart field technology related to technology advances in reservoir management can be
categorized under five areas (Nnaemeka Ezekwe, 2012):

• Down Hole Sensor (DHS)


DHS are devices installed to montior, measure and transmit data on a fluid flow, fluid properties
and other well performance data. DHS contain permanent temperature and pressure gauges,
distributed temperature and pressure sensors, sand production sensors and seismic sensors, and
single/multiphase flow meters.
• Down Hole Control Devices (DCD)
DCD are employed in regulating a specific interval of a well to attain zonal control of inflow
or outflow of fluid, without the need for remedial intervention. The two main types of DCD are
interval control valves (ICV) and inflow control devices (ICD).
• Well Architecture
Well architecture related with advances in drilling, geosteering and completion technologies.
This enabled the development of wells having complex architecture. Long reach horizontal wells
and multilateral wells are commonly used to improve hydrocarbon recovery from many onshore
and offshore reservoirs.
• Field Wide Monitoring
Field wide monitoring of reservoir performance has advanced with the application of 4D seismic
survey. 4D seismic technology is a low risk, high benefit reservoir management tool.
• Data Acquisition, Transmission and Utilization
Advances in the reservoir management technologies have created huge reservoir performance
data that must be transmitted, stored, analyzed, integrated and utilized for near real time reservoir
management. Near real time reservoir management has created an advanced reservoir management
process, described as closed loop reservoir management. In a closed loop reservoir management
process, real time data are used to optimize production from the reservoir in near real time.
It is noticeable that the technology targets can be resolved by applying the smart field
technology:
SPE-184131-MS 23

◦ Water/gas injection flood monitoring to accelerate recovery and reduce lost ultimate recovery
by successful thief zone shut off and improved sweep efficiency
◦ Multiphase metering of wells to monitor water inflow and plan intervention activities

◦ Self optimizing artificial lift to increase ultimate recovery by accelerating oil production, to
increase run life and reduce OPEX
◦ Selective zonal and multilateral leg control to reduce well intervention costs and improve
reservoir management, to increase ultimate recovery and accelerate recovery.
The main concept of voidage replacement is to use higher pressure zones to pressure up the producing
formation and help the maintenance of pressure. This method seems to be economically feasible because of
the lack of injection fluid costs and injection surface facilities. The main flaw in such processes is the lack
of control of the transferred volumes from the high pressure zone into the production zone. With the smart
field technology the operator is provided some leverage on the amount of transferred fluids. Downhole
production testing is a phenomenon which explores the possibility of flowing a prospect formation to a
depleted formation. Smart completions will allow measuring the average rate of production without the
need to flow the well to surface. This concept is useful in practice and environmentally friendly.
To evaluate the economic value from smart field technology, we need to generate work flows that utilize
reservoir simulation, nodal analysis and completion architecture. These workflows can generate different
scenarios and probable economic outcomes (Fig. 26).

Figure 26—Reservoir and Economical Analysis Work Flow Process for Sceeening
Smart Well Completion (Intelligent Well System - reproduced from SPE 94672)

An example for a field case study for smart field technology implementation to control voidage
replacement can be seen at Greater Plutonio Field, Angola (Booth and Sebastiao, 2010). This field consists of
five smaller fields with multiple stacked reservoirs in 1200-1400 meters under water. The field development
is 100% subsea with a large FPSO. Initially in 2007, six water and two gas injectors, and nine producers
were online. Production reached BOPD of 200,000 after three months. 300,000 BPD of water injection and
300 MM CFPD of gas injection were obtained in the first 12 months. This field is operated as one of BP's
future fields using advanced reservoir management technologies.
24 SPE-184131-MS

Figure 27—Field Location Map - Greater Plutonio Field, Angola (SPE 128542)

Several instruments that have been used in Greater Plutonio field for reservoir management technologies
are: permanent downhole monitoring system along with interval-control valves (ICV). The permanent
downhole monitoring system (PDHMS) provides valuable real-time pressure measurement of zones.
Interval control valves (ICV) on water injector wells are employed to direct injection into specific zones.
The reservoir management strategy implemented here is full pressure maintenance by balancing voidage,
which is achieved by directing the injection to specific zones with less number of wells (Fig. 28).

Figure 28—Example of Smart Well Completion with ICV to direct injection into specific zones

Nodal analysis is a major element in the assessment the valve size and tubing in smart well design. Some
reservoir properties including productivity index and expected rate determine the design size. If the smart
SPE-184131-MS 25

well design has multiple formations (>1), all the formation properties should be utilized to design the best
combinations of valves, in order to achieve the best injection and production results.

Reservoir Model and Simulations


A case using triggers where the production rate at bottom hole conditions (BHF or reservoir conditions) is
predicated on the BHF or reservoir gas injection rate. This is a reversal of the typical scenario where the
injection is based on the production. The goal is voidage replacement so that average reservoir pressure is
relatively constant. It is a miscible flood operation. A group BHF target coupled with voidage replacement
gas and water injection targets is used for approximate pressure maintenance.

• 19x1x13 GRID

• 6 COMPONENTS

• REGIONAL AQUIFER

Sensitivity study has been carried out on the following parameters mentioned in Table 1 below. The
tables and figures outline the model specifications including the wells. The figures show the initial and final
conditions of the base-case simulation for important parameters.

Table 1—sensitivity parameter

Name Active Default Value Source

1 C1 Yes 0.39360 Continuous Real

2 C11_14 Yes 0.07567 Continuous Real

3 C15plus Yes 0.100723 Continuous Real

4 C2_3 Yes 0.19940 Continuous Real

5 C4_5 Yes 0.09070 Continuous Real

6 C6_10 Yes 0.1399070 Continuous Real

7 CPOR Yes 3.0000E-7 Continuous Real

8 PERMK Yes 5 Continuous Real

9 PERM_HORZ Yes 50 Continuous Real


26 SPE-184131-MS

Figure 29—Porosity distribution

Figure 30—Permeability distribution in x direction


SPE-184131-MS 27

Figure 31—3D Phase Diagram

Figure 32—Wellbore diagram – injector at 5,1


28 SPE-184131-MS

Figure 33—Wellbore diagram – injector at 12,1

Figure 34—Wellbore diagram – injector at 17,1


SPE-184131-MS 29

Figure 35—Wellbore diagram – injector at 3,1

Figure 36—Wellbore diagram – injector at 10,13


30 SPE-184131-MS

Figure 37—Wellbore diagram – injector at 15,10

Figure 38—Initial oil saturation distribution


SPE-184131-MS 31

Figure 39—Final oil saturation distribution

Figure 40—Initial pressure distribution


32 SPE-184131-MS

Figure 41—Pressure distribution – 1984.02.01

Figure 42—Pressure distribution – 1985.01.01


SPE-184131-MS 33

Figure 43—Pressure distribution – 1986.07.01

Figure 44—Pressure distribution – 1987.06.01


34 SPE-184131-MS

Figure 45—Pressure distribution – 1988.06.01

Figure 46—Pressure distribution – 1989.06.01


SPE-184131-MS 35

Figure 47—Pressure distribution – 1990.06.01

Figure 48—Field oil rate


36 SPE-184131-MS

Figure 49—Field gas rate

Figure 50—Field water rate


SPE-184131-MS 37

Figure 51—Gas injection rate

Figure 52—Viodage Rate Ratio


38 SPE-184131-MS

Figure 53—Input range for fraction of C1

Figure 54—Input range for fraction of C11_14

Figure 55—Input range for fraction of C15+


SPE-184131-MS 39

Figure 56—Input range for fraction of C2_3

Figure 57—Input range for fraction of C4_5

Figure 58—Input range for fraction of C6_10


40 SPE-184131-MS

Figure 59—Input range for rock compressibility

Figure 60—Input range for vertical permeability

Figure 61—Input range for horizontal permeability


SPE-184131-MS 41

Figure 62—Simulation results vs proxy predicted results

Figure 63—Tornado chart

Results and Discussion


Voidage replacement is a key component of displacement processes especially in reservoirs where
subsidence and uplift issues may impact the recovery and economics significantly. Thus, better
understanding of each attribute in voidage management is crucial in reservoir management. This study uses a
reservoir simulation model that is coupled with a sensitivity/optimization software to assign different values
within specified ranges to measure the significance of the respective parameter, being either a reservoir/
uncertainty parameter or an operational/decision variable. It is observed that horizontal permeability plays
an important role in the process as might have been expected. The relative significance of each parameter
42 SPE-184131-MS

may change from one model to another, or the magnitudes may differ with similar relative significance,
thus, the main point here is the understanding of the physical process and how reservoir management
practice can be improved through better allocation of resources and time. This along with similar studies,
however, illustrates the significance of management of uncertainty and operational parameters and the room
of improvement and thus they serve as a starting point in reservoir management processes.

Acknowledgement
Authors are thankful to Aera Energy LLC, HESS, Turkish Petroleum Corporation, University of Southern
California, Kuwait Oil Company, Middle East Oilfield Services, Texas A&M University and VaalbaraSoft
for their support.

References
1. Arshad, A., Al-Majed, A., A., Menouar, H., Muhammadain, A., M., Mtawaa, B. 2009. Carbon
Dioxide (CO2) Miscible Flooding in Tight Oil Reservoirs: A Case Study. Presented at Kuwait
International Petroleum Conference and Exhibition 2009, Kuwait City, Kuwait, SPE 127616,
http://dx.doi.org/10.2118/127616-MS
2. National Enhanced Oil Recovery Initiaitive. 2010. How CO2-EOR Works. http://neori.org/
resources-on-co2-eor/how-co2-eor-works/.
3. Basta. George Soliman, Korany. Salah Kamal, Kortam. Waleed Tarek, Shaker. Saad. Experiences
of Conversion from Cyclic Steam Strategy to Continuous Steam Strategy in Issaran Egypt Field.
Presented at SPE EOR Conference, Muscat, Oman. March 2016. SPE-179773-MS
4. Clark A. Robert, Ludolph. Brian. Voidage Replacement Ratio Calculations in Retrograde
Condensate to Volatile Oil Reservoirs Undergoing EOR Processes. Presented at SPE ATCE,
Denver, Colorado. October, 2003. SPE 84359.
5. Brice. Bradley, Ning. Samson. Wood. Alecia. Optimum Voidage Ratio and Operational Practice
for heavy Oil Reservoirs. Presented at SPE Heavy Oil Conference, Alberta, Canada. June 2014.
SPE-170099-MS
6. G. S. Shirlkar, R. E. Stephenson, Predictable Control of Reservoir Pressure through Voidage
Replacement. Presented at SPE ATCE, Alberta, Canada. October, 1992. JCPT94-09-06
7. Kim W. Tae, E. Vittoratos, A. R Kovscek. An Experimental Investigation of Viscous- Oil
Recovery Efficiency as a Function of Voidage-Replacement Ratio. Presented at SPE Western
Regional Meet, Golden grove, CA. April, 2015. SPE 174032.
8. Abeeb A. Awotunde, Najmudeen Sibaweihi. Consideration of Voidage-Replacement Ratio in
Well-Placement Optimization. Presented at Kuwait International Petroleum Conference and
Exhibition, Kuwait. December, 2012. SPE 163354.
9. "Leste O. Aihevba, Mohammed Al-Harthy, John Passmore. Managing the Challenges of Voidge
in a Mature Carbonate Water Flood. Presented in the 13th Middle East oil Show & Conference,
Bharain. April 2003. SPE 81463.
10. Fekete Associates Inc., Surveillance Analysis Theory. http://fekete.com/SAN/
TheoryAndEquations/HarmonyTheoryEquations/Content/HTML_Files/Reference_Material/
Analysis_Method_Theory/Surveillance_Theory.htm
11. Clark, R. A., Ludolph, B. (2003). SPE-84359-Voidage Replacement Ratio Calculations in
Retrograde Condensate to Volatile Oil Reservoirs Undergoing EOR Processes.
12. Vittoratos, E., West, C. C. (2010). SPE 129545-Optimal Heavy Oil Waterflood Management May
Differ from that of Light Oils.
13. Vittoratos, E. S., Coates, R., West, C. C. (2011). SPE 150576-Optimal Voidage Replacement
Ratio for Communicating Heavy Oil Waterflood Wells.
SPE-184131-MS 43

14. Delgado, D. E., Vittoratos, E., Kovscek, A. R. (2013). SPE 165349-Optimal Voidage
Replacement Ratio for Viscous and Heavy Oil Reservoirs.
15. King, P. R., Buldyrev, S. V., Dokholyan. N. V., Havlin, S., Lopez E., Paul, G., Stanley, H. E.
(2002). Uncertainty in Oil Production Predicted by Percolation Theory, Physica, vol. 306, pg.
376–380.
16. Li, W., Jensen, J. L., Ayers, W. B., Hubbard, S. M., Heidari, M. R. (2009) Comparison of
interwell connectivity predictions using percolation, geometrical, and Monte Carlo models, J.
Pet. Sc. Eng., vol. 68, pg. 180–186.
17. Brice, B. W., Renouf, G. (2008). SPE 117327-Increasing Oil Recovery from Heavy Oil
Waterfloods.
18. Yin, M. 2015. CO2 miscible flooding application and screening criteria. Masters thesis. Missouri
University of Science & Technology, Missouri.
19. Wikipedia. 2010. Water Injection (oil production). (10 June 2016 Revision), https://
en.wikipedia.org/wiki/Water_injection_(oil_production)
20. Salehi, A., 2013. CO2 Injection for Enhanced Oil Recovery. Submitted as coursework for PH240.
Stanford University. California. December 11, 2013.
21. Meyer, J., P. Summary of Carbon Dioxide Enhanced Oil Recovery (CO2EOR) Injection Well
Technology. Background Report. American Petroleum Institute. Texas.

Potrebbero piacerti anche