Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/20APOG/1-20APOG/D013S105R002/2392465/spe-202343-ms.

pdf/1 by Universidad Estatal Peninsula de Santa Elena, Bryan Galarza on 19 November 2021
SPE-202343-MS

The Application of Simplified Model for Post-Acid Fracturing Well

Performance Analysis by using Grid Virtual Connection in the Numerical
Simulator

Ricko Rizkiaputra, Medco E&P; Axel Perwira Indro, Rock Flow Dynamics; Satrio Goesmiyarso, Petronas Carigali;
Ahmad Reizky Azhar and Rudini Simanjorang, Rock Flow Dynamics; Krishna Pratama Laya, Dimmas Ramadhan,
and Ari Subekti, Medco E&P

Copyright 2020, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Asia Pacific Oil & Gas Conference and Exhibition originally scheduled to be held in Perth, Australia, 20 - 22
October 2020. Due to COVID-19 the physical event was postponed until 17 - 19 November 2020 and was changed to a virtual event. The official proceedings were
published online on 12 November 2020.

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
There are several methodologies to model post-acid fracturing well performance such as implementing
negative skin factor and stimulated reservoir volume local grid refinement (SRV LGR). However, the first
method is too simplified to be applied in numerical simulation and the second method is too complex and
time consuming. Therefore, to bridge the two methods, this paper will discuss the application of a grid
virtual connection to model the well performance of acid fractured wells.
Grid virtual connection is the planar of fracture that alters the deliverability of the penetrated grid. In
this study, series of analyses are performed, such as: (i) Data input quality check (QC) and validation
(reservoir model and fracture model); (ii) Simulation running using the skin factor method, SRV LGR, and
grid virtual connection; (iii) Comparison between the three methods regarding the streamline result and
simulation running time; (iv) Model calibration (history matching) using post-fracturing data; (v) QC the
result of history matching. Post-fracturing production data from two wells (X-6 & X-8) are used as the
calibration point. From these analyses, the advantages and limitations of the virtual connection method will
be examined.
The input for this study is the history matched simulation model (up to the period before acid fracturing)
and the 2D fracturing model from the fracturing simulator, from which the fracturing parameters will be
taken forward to numerical simulation. Details as follow: (i) the skin factor method is generated from the
analytical model which results in equivalent skin of −5.6 and −5.9 for the X-6 and X-8 wells, respectively;
(ii) Input parameters for the SRV LGR and grid virtual connection methods use the output from the fracturing
simulator. There are several findings from these three methods: (i) The simulation running time for the grid
virtual connection is similar to the skin factor method, while SRV LGR takes longer to finish; (ii) Streamline
simulation shows the grid virtual connection method provides a sensible approximation of the physics
behind the reservoir – fracture – wellbore flow which fully represents by SRV LGR method, meanwhile skin
factor method only makes changes in the wellbore; (iii) The history matching process shows that the grid
2 SPE-202343-MS

virtual connection and SRV LGR method could be matched with the actual production data, however the
SRV LGR method is a highly non-unique solution due to the limited data input from the fracture simulator.
Meanwhile, the skin factor method could not be matched with the actual data using a sensible skin. Based

Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/20APOG/1-20APOG/D013S105R002/2392465/spe-202343-ms.pdf/1 by Universidad Estatal Peninsula de Santa Elena, Bryan Galarza on 19 November 2021
on these analyses, the grid virtual connection is shown to be the best (?) method to model the post-acid
fracturing well performance.
This paper shows that the grid virtual connection is a technically sound method to perform a first
order approximation of post-acid fracturing well performance. This method is beneficial to assessing the
production gain on an acid fracturing job with sensible physical approximation and fast computational times.

Introduction
The producing field that become the subject of this study is an unconventional reservoir, specificallly a
tight gas limestone reservoir. It possesses its own complexity due to the high level of heterogeneity and
unfavorable reservoir properties. The typical porosity and permeability of this type of reservoir is below 5%
and 0.1 mD, respectively. Hence, an effort to stimulate the reservoir is necessary to enhance productivity
of the existing producing wells.
One of the stimulations that is common in the tight gas limestone reservoir is acid fracturing. It is a
technique that injects an acid system via high pumping rates to create a fracture system in the reservoir.
Furthermore, acid fracturing also provides an etching mechanism which dissolves the hard yet very soluble
limestone rock and enhances the conductivity of the area that is contacted by the acid. However, forecasting
the well performance after acid fracturing is challenging as there are multiple variables involved such as
the acid system properties, operational parameters (pump schedule, pump rate, etc.), rock strength and
properties near the wellbore and well drainage radius.
The study of fracturing has been carried out for several decades. One of the earliest studies of fracturing
is discussed by Perkins and Kern (1961) which discusses the propagation of fracture widths in horizontal
and vertical fractures. Then, this study is continued by Nordgren (1972) in which it explains the propagation
of hydraulic fractures with limited vertical extent. Subsequently, Cinco Ley et al. (1973) developed an
analytical model of finite conductivity vertical fractures that is represented by the fracture skin. This model
is still widely used in numerical simulations which use a fracture skin factor.
In addition to the general fracturing study, a more specific acid fracturing study has been performed by
Williams and Nierode (1972) in which they present an acid fracturing model that can accurately predict the
acid penetration design. Nierode and Kruk (1973) evaluated a comprehensive acid fracturing conductivity
evaluation based on core samples. This evaluation, along with other models that were developed afterwards,
became the basis for commercial fracturing software to estimate the fracture conductivity, which one of
the pioneer on 3D numerical acid fracturing simulation and performance prediction is developed later by
Oeth et al. (2013).
Simulating post-acid fracturing well performance using analytical and more detailed numerical
simulation has advantages and disadvantages. The analytical approach using the skin factor is relatively
straightforward, however it does not provide a reliable description of the fracture geometry and properties.
Meanwhile, the detailed 3D numerical simulation offers a more realistic description of the fracture, however
this method is more complex and sometimes time consuming. To bridge between these gaps, a new approach
has been created. This method is based on a network of grid virtual connections that are intersected by the
fracture plane (Bogachev and Shelkov 2010).
This study will assess the applicability of a grid virtual connection and compare this with the two already
well-known methods such as fracture skin factor and stimulated reservoir volume local grid refinement
(SRV LGR). The study starts with the data input QC and validation. It then continues with reservoir
simulation of these three methods. Furthermore, the study assesses the resulting streamline and running time
SPE-202343-MS 3

of each method. The model is calibrated (history matched) by using actual post-acid fracturing production
data.
Output of the study will be the applicability of the grid virtual connection to perform post-acid fracturing

Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/20APOG/1-20APOG/D013S105R002/2392465/spe-202343-ms.pdf/1 by Universidad Estatal Peninsula de Santa Elena, Bryan Galarza on 19 November 2021
well performance analysis. Parameters that will be assessed will be: which key inputs are required,
simulation results, simulation running time, and model calibration. It is hoped that by performing this study
the benefits and key assumptions that are used in this method can be clearly defined.

Methodology
Data Input
The inputs for this study are the 3D dynamic model and the acid fracturing model (geometry & properties).
The dynamic reservoir model has been through a dynamic simulation workflow which covers the
initialization (including simulation stability & static-dynamic volume reconciliation), and history matching
until pre-fracturing. Then, the acid fracturing model is generated from the commercial fracture simulator
software. The output from this model is the 2D fracture model that consists of fracture geometry (length
and height), and fracture properties (fracture width, conductivity, etc.). It is the post-acid fracturing model
that has been matched with the fracturing pumping schedule and resulting pressure. The output from the
fracture simulator can be analyzed further to generate a fracture skin factor or become input parameters for
SRV LGR & grid virtual connection.

Three Assessed Methods

In this study, three methods are assessed, the fracture skin factor, SRV LGR, and grid virtual connection. In
this section, each of the underlying theories and their relation to this study will be explained.
Starting from the fracture skin factor, this analytical solution is provided by Cinco Ley et.al. (1978) in
which they describe the details regarding the finite conductivity vertical fracture. Firstly, the initial pre-acid
fracturing skin is obtained from the results of pressure transient analysis. Then, after the acid fracturing job
is finished and the fracture geometry & properties are obtained, the resulting skin on post-acid fracturing
was generated based on the finite conductivity vertical fracture model. This pre- and post-acid fracturing
skin is input into the 3D dynamic model on each respective timestep.
For the SRV LGR method the input is also taken from the fracture geometry & properties output of
fracturing simulator. In this advanced method, more than 30 inputs can be defined, which cover the fracture
path, fracture geometry, and also the fracture arithmetic. Moreover, the resulting SRV LGR grid can also
be selected using the uniform refinement, logarithmic, or unstructured grid. The simulation physics are the
same as the current practice of reservoir simulation, using the diffusivity equation as its basic principle on
the inter-grid flow dynamics. Therefore, the SRV LGR is able to represent the detailed bi-linear flow that
occurs in the reservoir – fracture – wellbore.
As with two methods above, the grid virtual connection also uses the input of fracture geometry &
properties from the output from the fracturing simulator. As opposed to the SRV LGR, this method only
needs less than 10 essential inputs that cover the fracture azimuth, height, length, width, and dimensionless
conductivity. The grid virtual connection is basically the modification of the generalized flow equation that
affects the grid that is penetrated by the fracturing planar. This semi-analytical modification is made to
mimic the bi-linear flow that exists in flow from the formation to the fracture and the flow along the fracture
to the wellbore. Several additions to the flow equation are involved, with some related to the equivalent
radial flow of the planar to mimic the flow from the reservoir to fracture and an addition of the productivity
multiplier function that reflects the flow along the fracture plane.
4 SPE-202343-MS

Model Running, History Matching, and Quality Check

The simulations are performed with each of the methods above using historical gas production data as a
constraint. Outputs from this model are the streamlines and simulation running time for each method. Then,

Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/20APOG/1-20APOG/D013S105R002/2392465/spe-202343-ms.pdf/1 by Universidad Estatal Peninsula de Santa Elena, Bryan Galarza on 19 November 2021
history matching is performed to calibrate the model. This is followed by a quality check of the history
match quality to see whether the match between simulation and actual data is reasonable.

Summary of Data Input

Two main data inputs are required – a 3D dynamic model and the acid fracturing model. This study is using
a field case of a tight gas-condensate limestone reservoir in which the 3D dynamic simulation workflow has
been performed (initialization until the history matching up to pre-Acid Fracturing data). The production
from this field ismainly from three production wells (X-6, X-7, and X-8) which are located in the eastern
culmination of the reservoir. In an effort to enhance the well productivity, an acid fracturing job was carried
out on two production wells (X-6, X-8). The treating pressure and rate during the job was matched with
the fracture simulator model, to enable the calibrated fracture geometry and properties to be obtained. Even
though the model provides detailed parameters of the fracture, the fracture simulator only simulates in the
2D domain (X & Z). It uses an assumption of constant reservoir properties away from wellbore, which
may be a correct assumption for the short acid penetration but not for the long fracture as there is reservoir
heterogeneity. A key summary of the 3D dynamic model and the acid fracturing model can be seen on
Figure 1 below.

Figure 1—Key Summary of 3D Dynamic Model and Acid Fracturing Models

Simulation Input & Preliminary Run Result

Fracture Skin Factor
Resulting fracture geometry & properties from the fracture simulator are processed for each of the assessed
methods. Starting from the most common approach using the fracture skin factor, the resulting fracture
geometry is input to the commercial analytical well modeling software to generate the post-fracture skin
factor. The main fracture properties that are input to generate the skin factor are fracture height, fracture
half-length, and dimensionless fracture conductivity. The resulting skin factor on X-6 and X-8 are −5.6 and
−5.9, respectively. Figure 2 below shows the key input and the resulting skin factor.
SPE-202343-MS 5

Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/20APOG/1-20APOG/D013S105R002/2392465/spe-202343-ms.pdf/1 by Universidad Estatal Peninsula de Santa Elena, Bryan Galarza on 19 November 2021
Figure 2—Parameter Input and Resulting Skin Factor on X-6 and X-8

The post-acid fracturing skins are input into the respective well within the completion schedule section.
Hence, the skin is altered on post-acid fracturing. Then, the simulation is performed based on this new
schedule and there are parameters that are assessed such as the simulation streamline and the simulation
running time. Figure 3 below shows the streamline simulation result and the running time.

Figure 3—Parameter Input and Resulting Skin Factor on X-6 and X-8

Figure 3 shows that the skin factor method represents the fracture by making the modification to the well
completion parameter, so that the streamline is sourced from the well without showing any expression of
linear flow to the wellbore. The simulation running time itself is relatively fast, and with 239,499 cells and
133 time steps, the simulation can be finished in less than three minutes as shown in Table 1.

Table 1—Simulation Details Using Fracturing Skin Factor Method

Total Active Grid Cells 239,499 Cells

Total Time Steps 133 Steps

Physics Model Type Black Oil (PVTG & PVDO)

System Core Usage 32 + GPU Assisted

Simulation Running Time 2 Minutes 31 Seconds

Stimulated Reservoir Volume Local Grid Refinement

As with the previous method, the fracture geometry & properties are generated from the fracture simulator.
However, the SRV LGR method offers the capability to model the fractures in more detail. Parameters
6 SPE-202343-MS

that are input in this method are not only limited to fracture height, fracture half-length, and dimensionless
fracture conductivity, but also fracture width and fracture azimuth. In addition, the SRV LGR method
also employs fracture arithmetic to govern the arithmetic relation between the main fracture zone and the

Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/20APOG/1-20APOG/D013S105R002/2392465/spe-202343-ms.pdf/1 by Universidad Estatal Peninsula de Santa Elena, Bryan Galarza on 19 November 2021
surrounding near-fracture zone. The SRV LGR method can also define the area that is exposed by the
injected acid system. This method also governs the degree of refinement that is expected in the X, Y, and
Z directions along with the type of refinement grid (uniform, logarithmic, unstructured). In this case study,
the output from the fracture simulator that is used such as the fracture length, fracture height, fracture
width, fracture azimuth, and dimensionless fracture conductivity is shown in Figure 4. Meanwhile, the
other parameter is input as the preliminary assumption and will be refined in the history matching (model
calibration) stage.

Figure 4—Typical Detailed Input of Fracture Template and Fracture Arithmetic

In comparison to the skin factor method, the resulting streamline simulation shows the response of bi-
linear flow. In Figure 5 below, the left figure shows a typical fracture SRV LGR, which in this case is using
the simple uniform refinement, while the figure on the right shows the streamlines along with the fracture
path that is marked with a red dashed rectangle. In this study, a 10 × 10 × 1 refinement has been made, which
means the refinement is performed 10 times from the original grid on the X and Y direction, and no grid
refinement is performed on the Z direction. Therefore, in terms of the simulation time, the SRV LGR method
is five times slower compared to the skin factor method. Details of the simulation is shown in Table 2.

Figure 5—Fracture SRV LGR and Resulting Streamline

SPE-202343-MS 7

Table 2—Simulation Details Using Fracturing SRV LGR Method

Total Active Grid Cells 239,499 Cells

SRV LGR Refinement (X, Y, Z) 10 × 10 × 1

Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/20APOG/1-20APOG/D013S105R002/2392465/spe-202343-ms.pdf/1 by Universidad Estatal Peninsula de Santa Elena, Bryan Galarza on 19 November 2021
Total Time Steps 133 Steps

Physics Model Type Black Oil (PVTG & PVDO)

System Core Usage 32 + GPU Assisted

Simulation Running Time 13 Minutes 15 Seconds

Grid Virtual Connection

The grid virtual connection method used the key parameters that were extracted from the fracture simulator
such as fracture length, fracture height, fracture width, and dimensionless conductivity. The information for
the fracture azimuth is generated from the analysis of the reservoir's maximum and minimum horizontal
stress direction, which in this case is in the NW – SE direction. Figure 6 below illustrates the Pre-Acid
Fracturing grid, inputs, and Post-Acid Fracturing grid that shows the activation of the represented fracture
planar.

Figure 6—Pre-Acid Fracturing, Grid Virtual Connection Input, and Post-Acid Fracturing Planar

Streamline simulation on the grid virtual connection below shows that there are new active streamline
influx nodes in the generated fracture planar. In comparison to the skin factor, this method captures the
reservoir heterogeneity better as the total well performance also depends on the nodes that are created along
the fracture planar away from the main wellbore. Moreover, this model is providing semi-analytical solution
for the bi-linear flow as the physics of reservoir to fracture flow is represented by the fracture nodes outside
the wellbore and the flow along the fracture to the wellbore is solved analytically by using the parameter
of dimensionless fracture conductivity. Figure 7 below shows the streamline simulation on the grid virtual
connection method.
8 SPE-202343-MS

Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/20APOG/1-20APOG/D013S105R002/2392465/spe-202343-ms.pdf/1 by Universidad Estatal Peninsula de Santa Elena, Bryan Galarza on 19 November 2021
Figure 7—Streamline on Pre- and Post-Acid Fracturing and Red Circle is showing
the New Active Influx Node on the Fracture Planar Away From the Wellbore

Since the grid virtual connection method solves the reservoir to fracture and fracture to wellbore flow
semi-analytically, therefore the simulation speed is relatively faster compared to the SRV LGR method
which solves the whole flow sequence using full numerical simulation. In fact, the simulation speed using
the grid virtual connection is similar to the skin factor method. Therefore, it has the advantages of capturing
the reservoir heterogeneity and also providing sensible physical approximation of fluid flow in reservoir –
fracture – wellbore in a fast simulation time. Table 3 below shows the key simulation parameters for the
grid virtual connection.

Table 3—Simulation Details Using Fracturing SRV LGR Method

Total Active Grid Cells 239,499 Cells

Total Time Steps 133 Steps

Physics Model Type Black Oil (PVTG & PVDO)

System Core Usage 32 + GPU Assisted

Simulation Running Time 2 Minutes 25 Seconds

Model Calibration
The previous preliminary simulation used the basic inputs generated from the fracture simulator, so further
model calibration was needed to match with the historical production data. This step is carried out by a
performing history match on each of the methods. The simulation model itself has been matched until the day
just before the acid-fracturing was performed. Therefore, the simulation restart is created, so the simulation
can be focused on the effort of calibrating the fracture model between the three methods.
Firstly, the history matching is performed on the skin factor method. A set of skin parameters are set up
on both wells accounting for the realization of low, base, and high cases of skin factor. In this experiment
the range is set from −1 to −20, while the latter is not a sensible skin and is used just for the purposes of
checking if the skin factor range can capture the actual data. Figure 8 below shows the parameter set up
and the result of history matching.
SPE-202343-MS 9

Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/20APOG/1-20APOG/D013S105R002/2392465/spe-202343-ms.pdf/1 by Universidad Estatal Peninsula de Santa Elena, Bryan Galarza on 19 November 2021
Figure 8—Parameter Set Up and Matching Result on Skin Factor Method

For forecasting, the gas rate is used as a constraint and Figure 8 shows that the matching could not be
performed on both wells. The X-6 simulation gas rate could not keep up with the actual production data.
Moreover, the X-8 gas rate BHP could not keep up with the actual BHP, despite the fact that both wells are
already set up with a very low skin factor on its range.
The history matching is then carried out on the SRV LGR method. It provides flexibility to input
the detailed parameters and is suitable for detailed fracture modeling. However, in this case, the export
parameter from the fracture simulator is limited to the essential parameter, leaving many parameters
unknown. In the history matching workflow, the unknown parameters are input with some assumptions.
The history matching process provides many non-unique solutions. Nevertheless, the production data can
still be matched as shown in the Figure 9 below.

Figure 9—Matching Result of SRV LGR Method

Then, the history matching workflow is performed using the grid virtual connection workflow. In this
method, only key parameters (frac geometry, properties, and azimuth) input, with most of these parameters
10 SPE-202343-MS

available from the output of the fracture simulator. The history matching workflow involves uncertainty
parameter set up with the resulting tornado diagram used to determine the most important parameters in the
forecast, which can be seen in Figure 10 below.

Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/20APOG/1-20APOG/D013S105R002/2392465/spe-202343-ms.pdf/1 by Universidad Estatal Peninsula de Santa Elena, Bryan Galarza on 19 November 2021
Figure 10—Parameter for Uncertainty Analysis and Resulting Tornado Diagram

Figure 10 shows that the most important parameter in the production forecast is the fracture length. In
this case, four parameters such as fracture length, fracture width, fracture height, and fracture conductivity
are taken forward as the parameters used for assisted history matching. In contrast, the azimuth parameter
in this case does not have a significant effect on the resulting production.
The assisted history matching is then performed using the above parameters. This step uses the gas rate as
the production constraint and the bottom hole pressure for each well as the objective function. The simulation
is performed until the desired threshold of difference is achieved. Figure 11 below shows the simulation
realization results and the matching realization.

Figure 11—Simulation Realization Results and the Matching Realization of the Grid Virtual Connection Method

Figure 11 shows that the grid virtual connection using the essential fracture properties input is able
to match the actual data. This model then can be taken forward for forecasting purposes. The matching
parameters for each well could also be used for the preliminary uncertainty ranges for the future acid
fracturing job in the other wells.
SPE-202343-MS 11

Forecasting
In this case study, performance forecasting is performed on existing wells (X-6 and X-8) in which acid
fracturing has been performed, and also for future wells (X-7) in which acid fracturing will be implemented.

Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/20APOG/1-20APOG/D013S105R002/2392465/spe-202343-ms.pdf/1 by Universidad Estatal Peninsula de Santa Elena, Bryan Galarza on 19 November 2021
For the wells in which acid fracturing has been performed, the matching parameters in the history matching
process earlier are taken forward for forecasting.
Furthermore, for the future wells in which acid fracturing will be implemented, the range of expected
parameter results use the assumption of the values from the matched parameters in X-6 and X-8 with several
adjustments to account for risk & uncertainty that is associated specific to the X-7 well. It is also assumed
that the acid fracturing job that will be performed in X-7 uses the same acid system, pumping sequence,
and pumping rate. The grid virtual connection is used for this preliminary forecast on X-7 since this method
requires relatively straightforward inputs from the fracture simulator and also provides a sensible physical
approximation of the fluid flow in the reservoir – fracture – wellbore in a fast simulation time. Figure 12
below shows the parameter set up and the forecast result in P10 – P50 – P90 percentile.

Figure 12—Forecast Parameter Set Up and Forecast Result on X-6 X-8 and X-7

Conclusions
Post-acid fracturing well performance has been analysed using a grid virtual connection method and has
been compared with the method that is well known in the industry : using the fracture skin factor and SRV
LGR. The conclusions are as follows:
1. Two main inputs are needed in this study : the calibrated 3D dynamic reservoir model, and the fracture
geometry & properties output from the commercial fracturing simulator.
2. The simplest approach using the analytical fracture skin factor shows the post-acid fracturing skin is
−5.6 and −5.9 for X-6 and X-8 wells, respectively.
3. The most advanced approach using SRV LGR involves taking inputs from the fracturing simulator.
However, the output from the fracturing simulator only provides basic fracture geometry and
properties. Therefore, some of the input parameters must be estimated using a preliminary assumption.
4. The grid virtual connection takes inputs from the fracture simulator. The input using this method is
relatively simple, straight forward (only accounts for the important fracture parameters), and similar
to the output of the fracture simulator. Therefore, most of the input parameters have a solid basis and
less assumptions are required.
5. Then, a preliminary run for these three parameters was performed. Two parameters are assessed - the
resulting streamline and the simulation running time.
a. The skin factor method works by modifying the well completion parameters, so the streamline
is only sourced from the well without any expression of linear flow to the wellbore. Therefore,
12 SPE-202343-MS

the heterogeneity away from the wellbore is not taken into account. Despite this very rough
approximation, the simulation running time for the skin factor method is fast.
b. The SRV LGR method works very well in terms of representing the detailed shape of the

Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/20APOG/1-20APOG/D013S105R002/2392465/spe-202343-ms.pdf/1 by Universidad Estatal Peninsula de Santa Elena, Bryan Galarza on 19 November 2021
fracture, which accounts for not only the fracture geometry and properties, but also the fracture
arithmetic away from the main fracture and the structure of the fracture grid refinement.
Therefore, the streamline simulation shows that the SRV LGR method able to mimic the bi-
linear flow that happens on the reservoir – fracture – wellbore. However, due to its complexity,
the simulation running time is five times slower compared to the skin factor method. Moreover,
some of the input parameters require assumptions as the fracture simulation does not provide
enough details for the SRV LGR input.
c. The grid virtual connection method works by providing a semi-analytical solution which
represents the bi-linear flow that exists in the fracture plane. The streamline simulation shows
the created nodes along the fracture planar away from the main wellbore, and captures the
reservoir heterogeneity better compared to the skin factor method. Furthermore, the flow along
the fracture is solved analytically by using the input of dimensionless fracture conductivity. The
simulation speed using the grid virtual connection is the same as the skin factor method with a
sensible physical approximation to the fracture flow phenomenon.
6. Model calibration by performing a history match on these three methods shows different responses.
The skin factor method is set up using very wide constraints but still cannot capture the actual
production data. Meanwhile, the SRV LGR method was able to capture the actual data but it is highly
non-unique solution since the fracture simulator output could not provide enough details for the SRV
LGR input. For the grid virtual connection method, the model could be matched to the production
data with the essential fracture geometry and properties as the input.
7. In this study, post-acid fracturing performance forecasting is demonstrated by using the grid virtual
connection method. The matched parameters were taken forward to the forecasting stage for the wells
in which an acid fracturing job had been performed. The range of expected parameters resulted in the
future acid fracturing operation using the assumed values from the matched parameters in existing
wells with some modifications. Therefore, possible outcomes of the forecast were created.
8. In conclusion, the grid virtual connection method provided a simplified model to analyze post-
acid fracturing well performance. It uses essential inputs from the fracture simulator, capturing
the reservoir heterogeneity, provides sensible physical approximations, and fast simulation times.
However, if the detailed fracture model is available and significant computing capability is available,
the SRV LGR should be used. Regarding the skin factor method, it is most suitable for the stimulation
job that only penetrates the nearby wellbore area, since the reservoir properties are assumed to be
similar to the wellbore.

Acknowledgements
The author would like to thank management of Medco E&P, Ditjen MIGAS, SKKMIGAS, and Saka Energy
as the JV partner in supporting the study, execution of acid fracturing, and permission to publish this paper.
Also thanks to Rock Flow Dynamics team that provides input and support during the modeling of post-
acid fracturing well performance.

References
Bogachev, K.Y. and Shelkov, V. 2010. New Realistic Hydraulic and Technogenic Fracture Modeling Approach in Full-
Scale Dynamic Models. Paper presented at SPE Russian Oil and Gas Conference and Exhibition, Moscow, Russia, 26
– 28 October. SPE-138071-MS. https://doi.org/10.2118/138071-MS
SPE-202343-MS 13

Cinco L., H., Samaniego V., F., Dominguez A., N. 1978. Transient Pressure Behavior for a Well With a Finite
Conductivity Vertical Fracture. Society of Petroleum Engineers Journal 18 (04): 253 – 264. SPE-6014-PA. https://
doi.org/10.2118/6014-PA
Nierode, D.E. and Kruk, K.F. 1973. An Evaluation of Acid Fluid Loss Additives, Retarded Acids, and Acidizde Fracture

Downloaded from http://onepetro.org/SPEAPOG/proceedings-pdf/20APOG/1-20APOG/D013S105R002/2392465/spe-202343-ms.pdf/1 by Universidad Estatal Peninsula de Santa Elena, Bryan Galarza on 19 November 2021
Conductivity. Paper presented at the 48th Annual Fall Meeting of the Society of Petroleum Engineers of AIME, Las
Vegas, Nevada, USA, 30 September – 3 October. SPE-4549-MS. https://doi.org/10.2118/4549-MS
Nordgren, R.P. 1972. Propagation of a Vertical Hydraulic Fracture. Society of Petroleum Engineers Journal 12 (04): 306
– 314. SPE-3009-PA. https://doi.org/10.2118/3009-PA
Oeth, C.V., Hill, A.D., Zhu, D. 2013. Acid Fracturing: Fully 3D Simulation and Performance Prediction. Paper presented
at SPE Hydraulic Fracturing Technology Conference, The Woodlands, Texas, USA, 4 – 6 February. SPE-163840-MS.
https://doi.org/10.2118/163840-MS
Perkins, T.K. and Kern, L.R. 1961. Widths of Hydraulic Fractures. Journal of Petroleum Technology 13 (09): 937 – 949.
SPE-89-PA. https://doi.org/10.2118/89-PA
Williams, B.B. and Nierode, D.E. 1972. Design of Acid Fracturing Treatments. Journal of Petroleum Technology 24 (07):
849 – 859. SPE-3720-PA. https://doi.org/10.2118/3720-PA