Sei sulla pagina 1di 23

All final manuscripts will be sent through an XML markup

process that will alter the LAYOUT. This will NOT alter the
content in any way.

SPE-181306-MS

Determination of Preferential Flow Direction in a Low Permeability


Carbonate Reservoir Through Real-Time Data and Rate Transient Analysis
(RTA)
R. Alcantara, J. L. Vazquez, PEMEX E&P and M. Briones, OGO

Copyright 2016, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Annual Technical Conference and Exhibition held in Dubai, UAE, 26–28 September 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 materi al 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
The determination of the preferential flow direction in a reservoir becomes a very important subject
for the development of an oilfield, especially for the selection of the optimum exploitation strategy
taking into account that porous media is a very complex environment in which representing the flow
behavior of fluids becomes a tough task, mainly in carbonates where the distribution of fractures,
lithology changes and diagenesis play a major role in this topic.

Naturally Fractured Reservoirs (NFR) represent a great technical challenge for the petroleum industry
because they behave as a heterogeneous medium with a strong influence of diagenesis, a term that
encompasses fractures, dissolution, compaction, dolomitization, cementation and recrystallization to
conform a reservoir with totally different distribution of properties. The dynamic data must be evaluated
in order to match with the static model achieving a good reservoir characterization.

In this paper we present a way to determine the preferential flow direction by the monitoring of the
field through permanent real-time downhole gauges that allowed the identification of the interference
between wells in a deep naturally fractured reservoir that originally did not show any degree of
communication. Suddenly, after some producing time, the field demonstrated a great level of
interference among wells and as a consequence, the determination of the preferential flow direction was
possible through Pressure Transient Analysis (PTA) and Rate Transient Analysis (RTA). Emphasizing
the Palacio-Blasingame1 type curves (PBTC), in which we compare the results obtained using this
analysis with its standard form and the one achieved with Material Balance Time (MBT). Additionally,
Fetkovich2 decline type curves were also used in the rate analysis.

Finally, a comparison of the geological model with the dynamic data was also applied to further
enhance the quality of the data. The results showed conclusive preferential flow direction with the wells
tested.
2 SPE-181306-MS

Introduction
The NFR located in Southeastern Mexico are branded to have specific conditions that make them
difficult to exploit. This situation is primary originated by their geological-structural configuration
produced by the intense tectonic and saline regional activity and the sedimentation processes that
conformed the Upper, Middle and Lower Cretaceous and the Upper Kimmeridgian Jurassic limestone
and dolomite rocks that contain hydrocarbons commercially producible located at depths ranging from
5,000 to more than 7,000 m.

These high temperature reservoirs (from 140 to 160°C) demand even more efforts in order to
overcome the technical challenges that they represent. The hydrocarbons produced are classified as
volatile oils and retrograde gas condensates from 32 to 45°API. Another common problem is the
formation damage skin factor, which impacts directly in the frequency and not always efficient well
stimulation operations.

It is well known that carbonates are very heterogeneous formations, however, the understanding of
the effects related to the rock-fluids system during the life of an oilfield, especially if it is a NFR, is not
achieved at all. In fact, this is a key point necessary to make decisions for a good field development
based on the exploitation strategy.

Carbonate reservoirs comform a large percentage of the world’s hydrocarbon reserves. So, this is the
reason why naturally fractured reservoirs are very important to oil industry. Unfortunately, a lack of
recognition of fractures and their distribution make difficult the characterization of these highly
anisotropic reservoirs, hence accurate solutions for modeling fluid flow in fractured porous media are
compulsory but challenging to obtain. The knowledge of the formation requires a good reservoir
characterization in order to optimize the field exploitation. For such multidisciplinary task, well testing
acts as a great option to estimate reservoir parameters, flow characteristics of the porous medium,
reservoir geometry, and the detection and assessment of heterogeneities.

Interference is evaluated through multiple-well testing that needs at least one producer well and one
observation well. The objective of this kind of tests is to investigate the pressure response between 2 or
more wells situated in the same formation to find the level of communication and the estimated
mobility-thickness product. Therefore the importance of the interference identification at the right
moment is significant for the optimum development of the field and wells location placement.

General aspects and reservoir behavior


The geologic structure of the field is an anticline formed by strong compressive stresses with a
current Northwest-Southeast orientation limited by a parallel inverse faults series deriving in the
reservoir semi-compartmentalization (Figure 1). The reservoir is found in naturally fractured limestones
and dolomites from the Upper Jurassic Kimmeridgian formation, which were deposited in the middle
ramp facies that developed intercrystaline porosity, microfractures and vugs due to the diagenesis
effects.
SPE-181306-MS 3

Figure 1 – Structural Configuration of the field.

The lithology variations across the reservoir were caused by the strong salt-dome intrusions
generating the conditions for the accumulation of oolitic banks on the top of the formation and fine
grained sediments at the bottom such as mudstones highly cemented.

It is thought that the reservoir could be classified as a type II according to the classification proposed
by Nelson3, where matrix provides the porosity and fractures and vugs cater the permeability for most of
the field’s production, the least having a great effect in fluids mobility. The reservoir is petrophysically
evaluated with a net pay of 172 m, an average effective porosity of 4%, the initial water saturation was
32% and the estimated permeabilities vary from 4 to 34 mD.

It is a 40° API light oil reservoir that started to produce since 2009 with the first well drilled as a
catwell in the Upper Jurassic Kimmeridgian naturally fractured dolimites with an initial production of
15,000 STB per day and 11.3 MMscf/day of gas. Nowadays the field is under development and has 5
producer wells; the field cumulative oil production is 27.3 MMSTB and the cumulative gas production
is 40.1 MMMscf (Figure 2).

Figure 2 – Pressure and production history of the field.

The exploitation of the field is turning very complex owing to the abrupt reservoir depletion rhythm;
it loses 92 kg/cm2/year and 20 kg/cm2/MMb. The initial reservoir pressure was 1,070 kg/cm2 and the
4 SPE-181306-MS

current GOR is 312 m3/m3. Initially, the existing drive index was the fluid-rock system expansion,
however the hydraulic drive has begun to act recently, this last was evaluated by means of material
balance, and also the reservoir voidage suggest that there is an additional drive index corresponding to
the acquifer with an approximate water entry of 10 MMbls (Figure 3).

Figure 3 - Reservoir voidage.

Pressure Transient Analysis (PTA)


PTA is an important issue for the reservoir dynamic characterization. Well testing provides the tools
to have a reliable way to determine the main reservoir properties and to have a unique source of
information that could investigate deeper into the formation up to detect the limits of the reservoir.

The field under study possesses pressure build-up and drawdown curves for each producer well,
however there is a lack of quality in the signal registered by the real-time permanent gauges in some of
the well tests, this is caused by many factors such as the gauge depth and phase segregation that produce
great storage effects, production stabilization (drawdown) before the shut-in, thus it is extremely
important and should consist of a considerable time for the reason that it could touch the interpretation
(model adjustment), and finally the duration of the tests, this must be long enough to assure a great
radius of investigation in order to characterize the reservoir through the detection of heterogeneities and
different flow geometries in the porous medium until the identification of its limits. This is a critical
point especially for very heterogeneous formations such as NFR.

The field is producing with 5 wells distributed on the top of the anticline structure. Well A is
considered to have been drilled in a very dolomitized zone with respect to the rest of the formation. This
was very helpful for the production; the initial production of the well A was 7,913 STB/day and 10.6
MMscf/day of gas. The pressure response of the build-up test showed a radial composite model adjusted
to an infinite reservoir that coincided with the static model that exhibits a change of lithology to the
northwestern side of the reservoir, where grainstones and wackestones are more common. The
calculated effective permeability to oil was 34.4 mD, a 526 ft of inner radius (the distance from the well
to the change of lithology), a mobility ratio of 0.69 and a formation skin factor of 2.5 (Figure 4).
SPE-181306-MS 5

Pressure [psi] 1000

100

10
1E-3 0.01 0.1 1 10
Time [hr]

Figure 4 – Pressure build-up test of well A.


Log-Log plot: p-p@dt=0 and derivative [psi] vs dt [hr]

Well B is located on the southwestern flank of the main block of the reservoir. The build-up curves
generated in this well are not good for the analysis due to the strong segregation and storage effects;
moreover the short duration of the tests do not allow getting a reliable answer because of the daily
production demand. This is the reason why the drawdown curves are the only source of information to
characterize the formation from this well.

The well was completed on the edge of the structure where the petrophysical properties are not good
for the productivity because the strata are more compacted and consequently, the porosity and
permeability are importantly affected in the zone drained by the well. The initial production of the well
B was 2,568 STB/day and 3.6 MMscf/day of gas. The pressure response was very caotic and hard to
interpret, so it was necessary to reprocess the test through deconvolution, a technique that helps to
identify the system and its parameters from a non-linear process when great flow variations exist. The
pressure derivative obtained was adjusted to a reservoir with intersecting faults detected. The considered
effective permeability to oil was 4.58 mD (the lowest permeability value in the field), with a skin factor
of -2.86 and the distance to the faults was 577 and 1073 ft, respetively (Figure 5).

1000
Pressure [psi]

100

10
0.01 0.1 1 10 100 1000
Time [hr]

Figure 5 – Pressure drawdown test of well B.


Log-Log deconvolution plot: p-p@dt=0 and derivative [psi] vs dt [hr]
6 SPE-181306-MS

Well C is also located in the less dolomitized zone of the main block of the reservoir. The initial
production of the well C was 6,054 STB/day and 7.9 MMscf/day of gas. Unfortunately, the same
problem occurred with the quality of the data from the build-up tests due to phase segregation, short
periods of testing and the near-wellbore effects that had a substantial influence over the results.
Therefore, the drawdown tests were considered for the analysis and the outcome was modeled as a
reservoir with one fault present. The effective permeability to oil obtained was 30.1 mD with a skin
factor of 58.5 and the distance to the fault was 3,520 ft (Figure 6).

1000
Pressure [psi]

100

10

1
1E-3 0.01 0.1 1 10 100 1000
Time [hr]

Log-LogFigure 6 – Pressure
plot: p-p@dt=0 drawdown
and derivative [psi] vs dttest
[hr] of well C.

Well D was drilled and completed on the second block next to the main one, where the other 3 wells
are producing. The initial production of the well D was 2,452 STB/day and 3.7 MMscf/day of gas. Also,
the interpretation of the tests was difficult due to the forementioned effects that affected the quality of
the information. Hence, the best interpretation was acquired through the drawdown curves despite the
constant “noise” of the data. The model was adjusted as a double porosity system with an identified
nearby fault. The effective permeability to oil obtained was 10.3 mD with a skin factor of 11 and the
distance to the fault was 1,020 ft (Figure 7). The fracture storage capacity coefficient was 0.023 and the
interporosity flow coefficient was 2.13E-06.

1000
Pressure [psi]

100

10
1E-3 0.01 0.1 1 10 100 1000
Time [hr]

Figure 7 – Pressure drawdown test of well D.


Log-Log plot: p-p@dt=0 and derivative [psi] vs dt [hr]
SPE-181306-MS 7

The storage capacity product (Kh) is a good indicator of the producibility of the formation, the values
acquired from all the well tests were in the range of 120 to 3,160 mD-ft, and obviously the greater
values correspond to the zones with the wells with higher productivity, such as wells A, C and D.

The reservoir heterogeneity is estimated trough two main parameters obtained from well test data: the
fracture storage capacity ratio (ω) and the interporosity flow coefficient (λ). These concepts were firstly
introduced by warren and root4 and mathematically are expressed as:

(Ф𝐶𝑡 )𝑓
𝜔=
(Ф𝐶𝑡 )𝑓 + (Ф𝐶𝑡 )𝑚 …………………………………………………………………………….… (1)

𝑘𝑚 𝑟𝑤 2
𝜆=𝛼
𝑘𝑓 𝑥𝑚 2 ………………………………………………………………………………...……...... (2)

Both terms are related to the complexity of the naturally fractured systems according to Nelson’s
classification, who identified four types of reservoirs based on the alteration of the matrix properties by
the presence of fractures that have a great impact on hydrocarbon production profiles. Depending on the
quantity, orientation and distribution of fractures as a result of strong geological and tectonic events that
modeled the reservoir structure over time, a fracture network that dominate the flow behavior in the
formation could be produced, which is very important for the production in carbonates.

The values of ω and λ acquired from well testing were 0.02 and 2.13E-06, respectively. Considering
the production behavior and the double porosity material balance calculations and coupling it to the
matrix-fracture system behavior parameters cited above, it was determined to catalog the reservoir as a
NFR type II.

Interference assessment and determination of preferential flow direction


The importance of the interference identification (if exists) at the right moment is critical for the
optimum development and well locations. This phenomenon complies with many factors related to the
porous medium and the fluid properties. Traditional interference testing is a multiple-well test in which
the pressure disturbance of an active well (a producer or an injector) is measured in the shut-in
observation well. Reservoir properties are investigated while determining the degree of communication
among wells as the main objective of this kind of tests, where the inter-well distance (the length of the
system) at the producing intervals midpoint depth with respect to the reservoir thickness is used instead
of using wellbore radius. The results obtained besides the determination of the interference are the
mobility-thickness product and the porosity-compressibilty-thickness product; the last is a key factor to
find the total compressibility value of the system.

As we mentioned previously, the permeability of the rock is strongly influenced by diagenesis,


geological-structural configuration and fluid properties. Therefore, the magnitude of the permeability is
dependent on the measurement direction, thereby requiring multiple-well testing. Interference testing is
ideal for quantifying reservoir anisotropy and heterogeneity because they are more sensitive to
directional variations of reservoir properties.
8 SPE-181306-MS

The only technique to analyze interference test is by type curve-matching technique through
dimensionless pressure (PD) and dimensionless time (tD), where PD is simply the exponential integral.
When the data are matched to the curve, permeability can be estimated from:

141.2𝑞𝐵𝜇 (𝑃𝐷 )𝑀
𝑘=
ℎ 𝛥𝑝𝑀 ………………………………………………………………………………… (3)

And the porosity-compressibility product is estimated from:

0.0002637 𝑘 𝑡𝑀
Ф𝐶𝑡 =
𝑟2 𝜇  𝑡𝐷
𝑟𝐷 2 𝑀 ……….……………………………………………………………….… (4)

After the change in rate, the difference between the extrapolated, matched type curve and the actual
data points (ΔpΔt) is determined from the data plot; Δt is the time from the change in rate at the active
well to the time the data are taken. It can be shown by superposition that:

141.2𝑞𝐵𝜇
𝛥𝑝𝛥𝑡 = 𝑃𝐷 (𝛥𝑡𝐷 , 𝑟𝐷 )
𝑘ℎ …...…………………………………………………………………… (5)

Therefore, when plotted, the points should fall on the curve matched by the original data.

In this case, the reservoir behavior is monitored constantly by means of the permanent pressure-
temperature gauges installed in each well during their completions. This activity allowed us to perform a
dynamic analysis based on the comportment of the pressure response in real-time at each producing well
along the field life. This means that interference test was carried on in which the observation wells were
producing while the active wells were shut-in as the field exploitation continued. In fact, the interference
was not expected, at least not at this stage of the field life.

Initially the wells did not show any degree of communication between them, what was a signal to
think that the reservoir had enough volume to support more producing wells; moreover, the material
balance indicated the presence of an aquifer.

The field produced 19.7 MMSTB of oil with only two wells during 5 years, so the development of
the field continued incorporating 3 additional wells (wells B, C and D) afterwards. A few months after
the integration of the new wells and during maintenance activities some wells were shut-in and the
opportunity to analyze the information was taken. Surprisingly, the wells demonstrated great
interference and at that moment it was possible to determine the preferential flow direction considering
the relation established between the results obtained dynamically and the geological model.

The distances between wells at the producing intervals midpoint depth with respect to the reservoir
are shown in Figure 8. According to that, it is clear that the the inter-well distance varies from 309 to
885 m, that correspond to the shorter distance between Wells B and C and the longest between A and D,
in that order.
SPE-181306-MS 9

Figure 8 – Distance between wells at the middle of the production intervals.

The bottomhole flowing pressure registered in every well showed the same general pressure tendency
as can be seen in Figure 9. Despite of the fact that the bottomhole flowing pressure in well B presents
the more irregular behavior, the overall tendency follows the same path that the rest of the wells, what
suggested a relationship to the main reservoir performance.

Figure 9 – Overall bottomhole flowing pressure tendency.

Wells A and D were shut-in due to operational requirements. The first shut-in well was well A, which
suffered two consecutive closures that generated a disturbance of pressure in the reservoir and was
recorded by wells B, C and D. Also, it must be mentioned that both closures were executed within 7
days of delay each other. The time it took to well B to record the first closure in well A was 105 min
(Figure 10). Consecutively, well C took 36 min to feel the movement made in well A (Figure 11); and
finally, the same disturbance was registered in 1,175 min in well D (Figure 12). However, well C also
recorded the movements produced by well B in spite of not being shut-in (a coiled tubing operation), so
the pressure disturbance that was produced by such operation was noted in 29 min by well C (Figure
13).
10 SPE-181306-MS

Figure 10 – Δt pressure disturbance from well A to well B.

Figure 11 - Δt pressure disturbance from well A to well C.


SPE-181306-MS 11

Figure 12 - Δt pressure disturbance from well A to well D.

Figure 13 - Δt pressure disturbance from well B to well C.

Accordingly, the interference analysis by type curve matching yielded a permeability of 53.3 mD and
a net pay of 146 ft between wells A and C (Figure 14). Also, the interference among wells A and D let to
know the permeability value of 141.8 mD in that direction with a net thickness of 191 ft (Figure 15).
Furthermore, the interference recorded between wells A and B found a permeability of 185.7 mD with a
thickness of 158.5 ft (Figure 16). Although well D could not feel the movements generated in well B, its
response is totally influenced by well A, which gives an idea of the great level of communication and
the possible flow direction in the main block.
12 SPE-181306-MS

Figure 14 – Interference analysis between wells A and C.

Figure 15 – Interference analysis between wells A and D.

Figure 16 – Interference analysis between wells A and B.

When well D was shut-in another interference effect was identified but this time only wells B and C
confirmed such phenomenon. Well B took 655 min to see the effects produced by well D (Figure 17).
Such interference was analyzed and the calculated permeability was in the order of 435.9 mD with a net
thickness of 158.5 ft (Figure 18); furthermore, well C recorded the same pressure disturbance in 320 min
(Figure 19). Hence, the computed permeability was 162.1 mD with a thickness of 158.5 ft (Figure 20).
On the other hand, wells A and C did not clearly registered any perturbance of well D because of the
noisy effects presented in the gauge data during those days and a coiled tubing cleaning in well A.
SPE-181306-MS 13

Figure 17 - Δt pressure disturbance from well D to well B.

Figure 18 - Interference analysis between wells D and B.


14 SPE-181306-MS

Figure 19 - Δt pressure disturbance from well D to well C.

Figure 20 - Interference analysis between wells D and C.

In view of the evident interference, it was possible to determine the preferential flow direction in the
field considering the position of the wells and the distance among them using vectorization.
Permeability is considered a tensor in nature due to its characteristics and the influence it has on the flow
of fluids within a reservoir. Tensors are geometric objects that describe linear relations between
geometric vectors, scalars, and other tensors. Euclidean vectors, often used in physics and engineering
applications, and scalars themselves are also tensors. In terms of a coordinate basis or fixed frame of
reference, a tensor can be represented as an organized multidimensional array of numerical values. The
order (also degree or rank) of a tensor is the dimensionality of the array needed to represent it, or
equivalently, the number of indices needed to label a component of that array.

On the other hand, a vector is an object which may be added together and multiplied by a number or
scaled. Vectors represent physical quantities of certain magnitude that can be added to yield a third one
with a similar magnitude and oriented in a resulting direction. This is used to treat the permeability as a
vector in order to know the preferential flow direction of the reservoir in a simpler way despite being a
tensor. Considering the two pressure disturbances generated in wells A and D and the permeabilities
SPE-181306-MS 15

obtained during the interference detected in the observer wells, it was feasible to analyze the system
through vectorization resulting in the main or the preferential flow direction of the reservoir at that
moment (Figure 21). The orientation of the preferential flow was 80° to the Southwest of the field, so
the wells are perpendicularly placed with respect to that direction of flow assuring the capture of great
volumes of production resulting in good productivity.

Figure 21 – The preferential flow direction of the reservoir.

The main block of the structure is characterized by having a fracture network that crosses the zone of
interest in the reservoir connecting all the wells under study (Figure 22). All the evidences made us to
think that the strong reservoir pressure drop could induce the fractures to “open-up” that afterwards
serve as a highway or a high conductivity medium to achieve an unexpected interference where the
system compressibility must be considered in order to obtain great directional permeabilities amog
wells, then it is clear that fractures are the main production supplier.

Figure 22 – Ant tracking showing the fracture network in the field.


16 SPE-181306-MS

Rate Transient Analysis (RTA)


RTA is a supportive tool for reservoir engineering purposes. This kind of analysis brings the
opportunity to study the behavior of a reservoir and identifies the parameters that are commonly
associated to well testing. A great advantage of RTA is that wells do not need to shut-in because the
main input is the production data, which implies that long production history must be accounted for to
ensure the reliability of the analysis.

RTA allows obtaining reservoir parameters such as well drainage radius, a permeability estimation,
skin factor and reservoir flow regime, as well as hydrocarbons in place, production decline forecasts,
Estimated Ultimate Recovery (EUR), remaining reserves to be produced and the abandonment time
according to the rhythm of exploitation. Therefore, if this technique is well applied, the results could be
quite good to adjust the real reservoir performance without the necessity of reducing production by
closing the wells.

The production time is an essential aspect to consider in RTA because it is very sensitive to abrupt
pressure-production changes; this means that any variation in the rate or pressure deviations could affect
strongly the results causing poor production forecasts and wrong appraisals of reservoir parameters.
Bearing in mind the above considerations, the PBTC were used for this analysis due to the fact that they
provide the chance to assess the the reservoir response through the rate and bottomhole flowing
pressures registered in each producer well as an input data, and also Fetkovich decline type curves
helped to forecast the production profile.

The concept proposed by Palacio and Blasingame (PB) originally intended to illustrate how to analize
gas-wells performance data with decline curve analysis via type curves based on the use of modified
time functions that work as a pressure transient type curve matching technique with the incorporation of
production history. Hence, it allows the modeling of the behavior of production data for variable rate
and variable pressure drop supported by the pseudosteady-state concept related to production decline.
However, the detailed description of this technique is beyond the scope of this paper. On the other hand,
the incorporation of MBT to the PBTC matching process yielded better results than the use of its normal
form, especially in the computation of the oil in place. Consequently, a comparison of both is presented
in this work.

Well A achieved a better adjustment than the other wells because of its production history, this means
that this well has been producing for a longer time than the others (Figure 23), which highlights the
importance of having a considerable production history and the way it influences the results. Fetkovich
decline type curves are recommended to fit for exponential decline to find reasonable results (Figure
24). Furthermore, the normal PB method reported 80 MMb of oil in place with a 3,740 ft drainage radius
corresponding to an estimated permeability of 1.13 mD and a skin factor of -2.9 (Figure 25). The MBT-
PB method reported 79.8 MMb oil in place with a reservoir estimated area of 5.5 km2 that corresponds
to an associated porous volume of 28.4 MMb, an estimated permeability of 4.14 mD and a skin factor of
9.6 (Figure 26).
SPE-181306-MS 17

6E+6
8000

Liquid volume [STB]


Liquid rate [STB/D]
4E+6

4000

2E+6

0 0
Pressure [psia]

10000

0
J J A S O N D J F M A M J J A S O N D J F M A M J J A S O N D J F M A M

2014 2015 2016

– Production
Figure 23Production history plot history plot ofPressure
(Liquid rate [STB/D], well [psia]
A (oil rate
vs Time [hr])and Pwf).

b 0
Di 0.00135 [Day]-1
qi 8970 STB/ D
UR 6.67 MMSTB
pwf 7046.95 psia
Not a unit

k.h 553 md.ft


0.1 k 0.98 md
Re 3840 ft
rwa 6.35 ft
Skin -3.42
PV 140 MMB
STOIIP 84.9 MMSTB
STOIP 81.2 MMSTB

0.01

1E-3
1E-4 1E-3 0.01 0.1 1 10
Not a unit

Figure
Fetkovich type curve – Fetkovich
24plot: decline
qDd and QDd type
vs tDd curves for well A.

STOIIP 80.1 MMSTB


STOIP 76.4 MMSTB
10 Re 3740 ft
rwa 3.87 ft
k.h 640 md.ft
k 1.13 md
Skin -2.92
Not a unit

0.1

0.01
1E-4 1E-3 0.01 0.1 1 10 100
Not a unit

Figure 25 – Palacio-Blasingame type curves for well A.


Blasingame type curve plot: qDd, qDdi and qDdid vs tDd
18 SPE-181306-MS

Figure 26 – MBT Palacio-Blasingame type curves for well A.

Unfortunately, Well B was not possible to adjust properly with MBT-PB method because of the
irregular pressure behavior showed in the analyzable period. However, the normal technique found that
the permeability is 2.75 mD, the drainage radius is around 1,650 ft, the skin factor was 0.04 and the oil
in place was 5.1 MMb (Figure 27). On the other hand, Fetkovich method reported a permeability of 1.27
mD, a 1,100 ft drainage radius, a 0.8 skin factor and the oil in place was 2.2 MMb. As can be seen, this
well is the worst producer of the field given its completion, petrophysical conditions and structural
position.

STOIIP 5.09 MMSTB


STOIP 4.79 MMSTB
Re 1650 ft
rwa 0.164 ft
10 k.h 330 md.ft
k 2.75 md
Skin 0.0467
Not a unit

0.1

0.01
1E-4 1E-3 0.01 0.1 1 10 100
Not a unit

Blasingame 27 –curve
Figure type Palacio-Blasingame typevscurves
plot: qDd, qDdi and qDdid tDd for well B.

Wells C and D have shorter producing lifes in comparison to well A, so, Fetkovich curves resulted in
lower permeabilities than the PTA and the PB methods (Figures 28 and 29). For well C, the normal PB
method reported 52 MMb of oil in place with an estimated permeability of 6 mD and a skin factor of -
1.9 (Figure 30). Moreover, Well D found 35 MMb of oil in place with an estimated permeability of 7.9
mD and a skin factor of 0.3 (Figure 31). In addition, the MBT-PB method yielded 67.9 MMb of oil in
place for well C and 51.9 MMb of oil in place for well D. The estimated permeabilities were evaluated
in 4 and 3.59 mD, respectively (Figures 32 and 33).
SPE-181306-MS 19

b 0
Di 0.00129 [Day]-1
qi 5850 STB/ D
UR 4.55 MMSTB
pwf 8886.14 psia
Not a unit

k.h 331 md.ft


0.1 k 2.76 md
Re 6060 ft
rwa 3.43 ft
Skin -2.8
PV 74.1 MMB
STOIIP 211 MMSTB
STOIP 210 MMSTB

0.01

1E-3
1E-4 1E-3 0.01 0.1 1 10
Not a unit

Figure 28 – Fetkovich decline type curves for well C.


Fetkovich type curve plot: qDd and QDd vs tDd

b 0
Di 5.64E-4 [Day]-1
qi 4220 STB/ D
UR 7.48 MMSTB
pwf 7318.1 psia
Not a unit

k.h 84.6 md.ft


0.1 k 0.705 md
Re 5520 ft
rwa 153 ft
Skin -6.6
PV 94.2 MMB
STOIIP 57.1 MMSTB
STOIP 55.9 MMSTB

0.01

1E-3
1E-4 1E-3 0.01 0.1 1 10
Not a unit

Figure 29 – Fetkovich decline type curves for well D.


Fetkovich type curve plot: qDd and QDd vs tDd

STOIIP 52 MMSTB
STOIP 51.2 MMSTB
Re 6530 ft
10
rwa 1.44 ft
k.h 721 md.ft
k 6.01 md
Skin -1.93
Not a unit

0.1

0.01
1E-4 1E-3 0.01 0.1 1 10 100
Not a unit

Figure 30 – Palacio-Blasingame type curves for well C.


Blasingame type curve plot: qDd, qDdi and qDdid vs tDd
20 SPE-181306-MS

STOIIP 35 MMSTB
STOIP 33.8 MMSTB
10 Re 4320 ft
rwa 0.15 ft
k.h 947 md.ft
k 7.89 md
Skin 0.327
Not a unit

0.1

0.01
1E-4 1E-3 0.01 0.1 1 10 100
Not a unit

Figure 31 – Palacio-Blasingame type curves for well D.


Blasingame type curve plot: qDd, qDdi and qDdid vs tDd

Figure 32 – MBT Palacio-Blasingame type curves for well C.

Figure 33 – MBT Palacio-Blasingame type curves for well D.


SPE-181306-MS 21

Results of Fetkovich, normal PB and MBT-PB methods for wells A, B, C and D are shown in Table
1, 2 and 3, respectively.

Parameters Well A Well B Well C Well D


b 0 0 0 0
Di (year-1) 0.49 3.18 0.47 0.206
UR (MMb) 6.67 0.30 4.55 7.48
Kh (mD.ft) 553 152 331 84.6
K (mD) 0.98 1.27 2.76 0.70
Re (ft) 3,840 1,100 6,060 5,520
Rwa (ft) 6.35 0.07 3.43 153
S -3.4 0.8 -2.8 -6.6
PV (MMb) 140 3.7 74.1 94.2
N (MMb) 84.9 2.2 211 57.1
Table 1. Results from Fetkovich decline type curves.

Parameters Well A Well B Well C Well D


Kh (mD.ft) 640 330 721 947
K (mD) 1.13 2.75 6 7.89
Re (ft) 3,740 1,650 6,530 4,320
Rwa (ft) 3.87 0.164 1.44 0.15
S -2.9 0.04 -1.9 0.3
N (MMb) 80.1 5.1 52 35
Table 2. Results from normal Palacio-Blasingame type curves.

Parameters Well A Well B Well C Well D


Kh (mD.ft) 2,343 - 2,263 2,031
K (mD) 4.14 - 4 3.6
Re (ft) 1,325 - 1,222 1,068
Rwa (ft) 1.32E-05 - 1.22E-05 1.06E-05
S 9.6 - 9.7 9.9
N (MMb) 79.8 - 67.9 51.9
Table 3. Results from MBT Palacio-Blasingame type curves.

The deviations of the results obtained with RTA methods with respect to PTA are associated to the
fact that in PTA, wells are shut-in for pressure build-up to avoid the difficulty of having estabilized
production during the test, thus, the production effects are minimized for the analysis and the results are
reliable if the tests were well designed and executed. However, some wells only have pressure
drawdown tests feasible to evaluate and the storage and humping effects are very common in this deep
volatile oil reservoirs as we previously stated.

The difference between the two RTA methods proposed (the normal PB and the MBT-PB) is that in
the first, only well’s production is used and in the second, the concept of material balance time is
applied, in which the ratio of well’s oil production to its cumulative production is introduced. This
modification allows having some variations, mainly in the estimated permeability, skin factor, drainage
radius and oil in place. The evaluation of all of these parameters through RTA methods is very sensitive
to any change in pressure and/or production because the wells are producing and obviously, the dynamic
behavior is not constant over time. Another critical factor to consider is the production time; it is very
22 SPE-181306-MS

important to have large producing time intervals for the analysis in order to obtain better production
forecasts and well-reservoir parameters very close to the ones obtained in well testing.

Conclusions
The establishment of the degree of communication between wells was possible through the constant
monitoring of real-time downhole gauges that allowed the identification of the pressure response in a
complex deep naturally fractured reservoir to dynamically characterize it and to determine the
preferential flow direction in a low permeability formation, improving the evaluation of new
opportunities in search of an optimum oilfield development considering this analysis focused on NFR.

The importance of a good design and execution of well tests is imperative in order to avoid poor
results or unnecessary operational movements that could affect the pressure response and as a
consequence, the interpretation itself. This is very important especially for deep reservoirs and
permanent downhole gauges taking into account that well testing is the only tool available to
characterize the formation in great areal extentions. Moreover, the reservoir monitoring through real-
time downhole gauges let us know the pressure behavior constantly and the identification of the
interference at the right moment. However, downhole gauges are not the best option for well testing
interpretation due to the very noisy data, great storage, phase segregation and humping effects,
especially for deep wells and NFR.

Net pay (h) has a direct influence in permeability calculations. So it is recommended to reevaluate the
net thickness according to the updated geologic-petrophysical model to minimize inaccuracies that could
result in wrong data. On the other hand, the identification of the preferential flow direction in the
reservoir was obtained through the vectorization of permeabilities computed from interference analysis.
This highlights the role of well testing as a valuable technique to assess the reservoir anisotropy.

Based on Nelson’s classification of NFR and the data acquired from this analysis we must indicate
that such ordering could be simultaneous, namely, a highly anisotropic reservoir could have different
zones within, where each one of those sectors could present a different behavior depending on lithology
changes and diagenesis that result in different types of NFR according to Nelson’s classification
distributed in blocks or compartments in the same reservoir.

Rate Transient Analysis (RTA) is a useful option to study the behavior of a reservoir, to acquire
production forecasts and to identify the parameters that are commonly associated to well testing. The
main advantage of RTA is that wells under study do not need to shut-in because the main input is the
production data, which implies that long production history must be accounted for to ensure the
reliability of the analysis. Furthermore, the production time is an essential aspect to consider in RTA
because this technique is very sensitive to abrupt pressure-production changes; this means that any
variation in the rate or pressure could affect strongly the results causing poor producton forecasts and
wrong appraisals of reservoir parameters.
SPE-181306-MS 23

Nomenclature
PTA Pressure Transient Analysis
RTA Rate Transient Analysis
MBT Material Balance Time
STB Stock Tank Barrel
MMb Million Barrels
MMscf Million standard cubic feet
MMMscf Billion standard cubic feet
EUR Estimated Ultimate Recovery
Pwf Bottomhole Flowing Presure
NFR Naturally Fractured Reservoir
PB Palacio-Blasingame
PBTC Palacio-Blasingame Type Curves

References
1. Palacio, J. C., Blasingame, T. A., “Decline-Curve Analysis Using Type Curves-Analysis of
Gas Well Production Data”, SPE-25909, Denver, 1993.
2. Fetkovich, M. J., “Decline Curve Analysis Using Type Curves”, SPE-4629-PA, 1980.
3. Nelson, R., “Geologic Analysis of Naturally Fractured Reservoirs”, Gulf Professional
Publishing, 2nd Edition, 2001.
4. Warren, J. E., Root, P. J., “The Behavior of Naturally Fractured Reservoirs”, SPEJ, AIME,
1963.
5. Earlougher, R. C., “Advances in Well-Test Analysis”, Monograph Vol. 5 of the Henry L.
Doherty Series, SPE, New York, 1977.
6. Cinco-Ley, H., “Well-Test Analysis for Naturally Fractured Reservoirs”, SPE-31162-JPT,
Journal of Petroleum Technology, January 1996.
7. Tiab, D., Restrepo, D. P., Igbokoyi, A., “Fracture Porosity of Naturally Fractured
Reservoirs”, SPE-104056, Cancun, Mexico, 2006.
8. Ogbe, D. O., Brigham, W. E., “A Correlation for Interference Testing with Wellbore-Storage
and Skin Effects”, SPE-13253-PA, Houston, 1989.
9. Grasman, T. J., Grader A. S., “Constant-Pressure Interference Testing: Detecting Double-
Porosity Properties”, SPE-18835-PA, Oklahoma City, 1990.
10. University of Cambridge, “What is a tensor?”, Dissemination of IT for the Promotion of
Materials Science.

Potrebbero piacerti anche