SPE-211396-MS

The Accurate and Efficient Simulation of a Fractured Tight Gas Condensate

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

Reservoir Using Embedded Discrete Fracture Model

Chengwu Yuan, BP America; Wei Yu, SimTech LLC; Soheil Ghanbarzadeh, BP America; Chuxi Liu and Jijun Miao,
SimTech LLC

Copyright 2022, Society of Petroleum Engineers DOI 10.2118/211396-MS

This paper was prepared for presentation at the ADIPEC held in Abu Dhabi, UAE, 31 October – 3 November 2022.

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
The hydraulic fracture modeling is critical for the well dynamic performance and development optimization
in the fractured condensate reservoir. The conventional approach based on LGR (Local Grid Refinement)
is widely used but with low computational efficiency, and sometimes with artificial fracture orientation and
geometry. In this study, a new workflow to model the hydraulic fractures using EDFM (Embedded Discrete
Fracture Model) is developed to couple with the Nexus reservoir simulator for the first time. Borrowed from
the dual continuum concept, the main idea of EDFM is the non-intrusive embedding of the fracture medium
physical properties (such as transmissibility, pore volumes, etc.) into the pre-existing matrix cell blocks.
Using the EDFM, tedious work such as model re-gridding and refinements can be avoided, and realistic
fracture geometry and geological structures can be honored with only a fraction of the total computational
cost.
We validate the EDFM approach firstly with a simple mechanistic model and then a sector model with
18 hydraulic fractures in Cartesian grids. The results from the traditional LGR method are compared with
those from EDFM to demonstrate its robustness. For the same fracture geometry and reservoir model
configurations, the LGR-based and EDFM-based approach gives almost identical results for production
simulations. We then applied this workflow to a more realistic sector model with 18 vertical wells in a
tight gas condensate reservoir with corner points geometry. We demonstrated that the EDFM approach
enables more realistic hydraulic fracture modeling, including the flexibility to incorporate geomechanics-
consistent fracture orientation, gridding-independent fracture height, fracture half-length, and straight-
forward modifications of fracture properties. Additionally, the speed of the EDFM is much superior to the
LGR-based one. The speed-up factor reaches about 50 when appropriately decomposing the domain in
this sector model, reducing the computational time from about 8 hours to about 9 minutes. This EDFM-
Nexus workflow enables a much faster and more robust fracture modeling, model calibration, development
optimization, and business decision-making in the tight gas condensate reservoirs.
2 SPE-211396-MS

Introduction
Efficient and accurate simulation of well performance from hydraulically fractured reservoirs is very
important and challenging. Traditional methods such as structured and unstructured grids with local grid
refinement (LGR) have been widely applied to perform fracture simulation (Rubin, 2010; Cipolla et al.,
2010; Cipolla et al., 2011; Karimi-Fard and Durlofsky, 2016; Hui et al., 2018; Yu and Sepehrnoori, 2018).
However, the computational efficiency of these methods is not so high, especially for dealing with many

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

hydraulic and natural fractures. In recent years, the embedded discrete fracture model (EDFM) method has
been developed and widely used to perform hydraulic and natural fracture simulation more efficiently than
the traditional methods (Moinfar et al., 2013; Panfili and Cominelli, 2014; Shakiba and Sepehrnoori, 2015;
Cavalcante Filho et al., 2015; Jiang and Younis, 2016; Du et al., 2017; Xu et al., 2017, 2019; Chen et al.,
2020). It has been coupled with several simulators. However, we have not seen it was coupled with the
Nexus reservoir simulator.
There are many simple case studies with single well to demonstrate the accuracy and efficiency of the
EDFM method (Moinfar et al., 2014; Xu, 2015; Panfili et al., 2015; Xu et al., 2017; Flemisch et al., 2018;
Yu et al., 2018a) when compared to the traditional LGR method. However, the benchmarking field-scale
case studies between EDFM and LGR with multiple wells and realistic heterogeneity have rarely been
examined. Du et al. (2017) compared the well performance between the EDFM and the unstructured grids
with LGR using a field-scale case with three horizontal shale-oil wells in Permian Basin and found that the
EDFM method can be 1400 times faster than the unstructured gridding method. Xu et al. (2018) performed
the comparison between EDFM and LGR using a realistic field case with six horizontal shale-gas wells in
the Mancos Shale and found the EDFM method is 19.1 times faster than the LGR method for a 30-year
production simulation. Fiallos Torres et al. (2021) discussed the comparison of the computational efficiency
of EDFM and the unstructured gridding method using a field scale shale-oil case with 3 horizontal wells
in Eagle Ford for a 210-day history matching simulation and mentioned that the EDFM is about 5 times
faster than the unstructured gridding method. However, the above-mentioned field case studies with multiple
wells just considered the simple layer cake model without considering the realistic geological model with
corner points and heterogeneous distribution of matrix properties including porosity, permeability and water
saturation vertically and laterally. In addition, the validation field case study with multiple wells in a tight
gas condensate reservoir has not been examined.
In addition to the low computational efficiency of the LGR method to model hydraulic fractures, another
LGR limitation is that the hydraulic fracture length and height is grid-dependent. It is difficult to achieve
a realistic geomechanics-consistent fracture orientation, especially for the geo-models with corner point
gridding. The fast and easy generation of hydraulic fracture grids within the realistic geo-models can help
engineers in simulation model building, performing sensitivity studies, and history matching.
In this study, we develop a workflow to couple the EDFM method with the Nexus reservoir simulator in
a non-intrusive manner to efficiently perform hydraulic fracture simulation. We validate the EDFM method
against the LGR method using a simple case with a single fracture and single phase and a more realistic
sector model in Cartesian grids with 18 vertical wells and 18 hydraulic fractures in a tight gas condensate
reservoir. We then applied this workflow to a more complex sector model in corner point gridding. The
accuracy and efficiency of the EDFM method is discussed.

EDFM Workflow for Nexus Reservoir Simulator

EDFM is a method to model 3D complex hydraulic and natural fracture networks in reservoir simulation.
Fig. 1 shows the principle of the EDFM method to model 3D complex fractures using non-neighboring
connections (NNCs). Firstly, matrix grids are generated without considering the fracture's existence. Then,
each fracture plane in the reservoir could be discretized into some smaller segments by using the matrix-cell
boundaries so that these fracture segments are modeled by additional fracture grids and added to the original
SPE-211396-MS 3

matrix grids (Xu et al., 2017; Xu and Sepehrnoori, 2019). This approach avoids the complex gridding issues
of LGR or unstructured gridding techniques while keeping the model's accuracy (Xu et al., 2018; Yu et al.,
2018b). By applying EDFM, NNCs are needed to mimic the flow communication between cells that are
physically connected but not neighboring in the computational domain (Xu et al., 2017).

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

Fig. 1—Principle of EDFM method to model 3D complex fractures using non-neighboring
connections (NNCs): (a) embedding 3D fractures into the geo-model; (b) simulation
of fluid transport with matrix cells and fracture cells using NNCs (Xu et al., 2017).

Except for the basic connection between matrix grids, there are three different types of NNCs added
to represent the fluid flow associated with the fracture segments. The NNCs consider the flow between
matrix grids and the corresponding fracture segments, the flow between fracture segments within an
individual fracture, and the flow between intersecting fracture segments within different fractures. Finally,
an effective wellbore index will be calculated using the modified Peaceman method for the intersection
between hydraulic fractures and the horizontal wellbore. More details about the fundamental equations can
be found in the references (Xu et al., 2017, Sepehrnoori et al., 2020).
The workflow of the EDFM method for the Nexus reservoir simulator to simulate 3D complex hydraulic
and natural fractures is shown in Fig. 2. First, a base case with an input file of the Nexus reservoir simulator
without considering any hydraulic and natural fracture properties should be prepared. Then, the EDFM input
of the geo-model and key reservoir properties including matrix porosity, permeability, and water saturation
will be set up. After that, we develop a fast hydraulic fracture generator within the Cartesian or corner
point grids. The fracture half-length, height, aperture, permeability, and angle information are generated
4 SPE-211396-MS

and input -for EDFM. Next, we will input well information including well number, well coordinates, and
wellbore radius for EDFM. After all these inputs, we will run the EDFM preprocessor to generate the output
of fracture and well properties for the Nexus reservoir simulator. Based on these outputs, we will modify
the input file of the base case to include these new fracture grids’ properties including fracture porosity,
permeability, water saturation, and depth. More importantly, we will modify the input file considering the
NNCs and transmissibility. We also need to modify the input file by including the well-fracture perforation
file. After that, the final input file of the base case considering hydraulic and natural fracture properties

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

is ready to run. After running, we will perform visualization of dynamic properties such as pressure and
saturation of matrix and fractures separately.

Fig. 2—The workflow of the EDFM method for Nexus reservoir simulator to
simulate well performance with 3D complex hydraulic and natural fractures.

EDFM Validation with Simple Case

A simple reservoir model with a grid dimension of 3×3×1 and a single hydraulic fracture was built to show
the validation of the EDFM method against the LGR method, as shown in Fig. 3. The hydraulic fracture
is in the middle of a reservoir and fully penetrates the reservoir thickness. The key reservoir and fracture
properties are listed in Table 1. A single phase of gas was assumed and simulated. A constant flowing bottom
hole pressure of 1000 psi was used for reservoir simulation constraint and the simulation time is 10 years.
SPE-211396-MS 5

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

Fig. 3—A simple reservoir model with a single hydraulic fracture in the middle. The red line represents the hydraulic fracture.

Table 1—Key reservoir and fracture properties used for the simple case simulation.

Reservoir Description Value Unit

Model dimension (x × y × z) 30 × 30 × 10 ft
Number of grid blocks (x × y × z) 3×3×1 -
Grid blocks dimensions (x × y × z) 10 × 10 × 10 ft
Matrix gas saturation 100% -
Fracture aperture 0.01 ft
Fracture half-length 15 ft
Fracture height 10 ft
Fracture permeability 10000 mD

For the LGR method, the matrix cell containing the hydraulic fracture was discretized into 5 sub-cells with
a cell size of 4 ft, 0.995 ft, 0.01 ft, 0.995 ft, and 4 ft perpendicular to the fracture, as shown in Fig. 4(a). The
middle cell represents the hydraulic fracture with an aperture of 0.01 ft. The well only has perforation with
fracture without considering the perforation with the matrix. For the EDFM method, the single hydraulic
fracture was divided into 3 small fracture segments due to three matrix cells intersecting with the fracture.
Hence, there are three additional cells representing these fracture segments, which are on the right side of
the model, as shown in Fig. 4(b). The fluid transport between matrix cells and fracture cells is considered
using the NNC.
6 SPE-211396-MS

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

Fig. 4—Comparison of two methods for modeling a single hydraulic fracture: (a) the LGR method; (b)
the EDFM method. The green tube represents well location perforating only with hydraulic fracture.

The comparison of gas flow rate between the EDFM method and the LGR method after 10 years of
simulation is shown in Fig. 5. A great match was obtained, demonstrating the accuracy of Nexus simulator
using EDFM to simulate well performance with hydraulic fracture.
SPE-211396-MS 7

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

Fig. 5—A shale-gas reservoir model including a horizontal well with multistage hydraulic
fractures and 1000 natural fractures. The red line represents wellbore, while the blue
lines represent hydraulic fractures, and the green lines represent natural fractures.

The comparison of pressure distribution after 10 years of simulation between the EDFM method and the
LGR method is shown in Fig. 6. A similar pressure distribution was observed.

Fig. 6—Comparison of pressure distribution after 10 years of simulation

between two methods: (a) the LGR method; (b) the EDFM method.
8 SPE-211396-MS

EDFM Validation with Field Case

Next, we'd like to validate the EDFM using a much larger and more realistic model. We set up a field-scale
gas-condensate model using Cartesian grids where the LGR can represent perfectly a 2D planar fracture.
That ensures that the fracture orientation and geometry in LGR and EDFM are exactly the same. The model
dimension is 40800×40800×350 ft, as shown in Fig. 7. The number of grids is 136×136×35 corresponding
to a grid size of 300×300×10 ft, respectively. Fracture length is 2100 ft, fracture height is 350 ft, and fracture

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

conductivity is 3000 md-ft. Both LGR and EDFM methods have the exact same fracture geometry. The
other key reservoir, fluid, and fracture properties are summarized in Table 2. A constant flowing bottom
hole pressure of 2000 psi was used for the simulation constraint. The total simulation time is 41 years.

Fig. 7—A realistic field-scale gas-condensate case using simple Cartesian

grids with 18 vertical wells and 18 hydraulic fractures modeled by EDFM.

Table 2—Key reservoir, fluid and fracture properties used in the field-scale gas-condensate model.

Reservoir Description Value Unit

Number of grid blocks (x × y × z) 136 × 136 × 35 -

Fracture aperture 1 ft

Fracture half-length 1050 ft

Average fracture height 248 ft

Fracture permeability 3000 mD

Fracture water saturation 18% -

The comparison of average reservoir pressure and cumulative gas and condensate production between
LGR and EDFM is shown in Fig. 8, illustrating a great match between these two methods. In addition, the
comparison of well by well performance is shown in Fig. 9. A great match for each well was also obtained.
It demonstrates the accuracy of the simulator using EDFM to simulate well performance with a hydraulic
fracture in a realistic and complex scenario
 Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023
9

methods: (a) average reservoir pressure; (b) cumulative gas and oil production.
Fig. 8—Comparison of well performance after the simulation between two
SPE-211396-MS
10 SPE-211396-MS

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

Fig. 9—Comparison of cumulative gas and oil production after the

simulation between two methods (EDFM vs LGR) for each individual well.

The comparison of pressure distribution after the simulation between the EDFM method and the LGR
method is shown in Fig. 10. A similar pressure distribution was observed.
 Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023
11

between two methods: (a) the LGR method; (b) the EDFM method.
Fig. 10—Comparison of pressure distribution after the simulation
SPE-211396-MS
12 SPE-211396-MS

EDFM Application in A Field Case

An even more complex field-scale gas-condensate case was built using corner points with 18 vertical wells
and 18 hydraulic fractures modeled by the EDFM, as shown in Fig. 11. For the LGR method, the matrix
cell containing hydraulic fracture is discretized into 11 sub-cells. Hydraulic fractures fully penetrate the
reservoir thickness in different well locations. A constant flowing bottom hole pressure of 2000 psi was
used for the simulation constraint. The total simulation time is 41 years.

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

Fig. 11—A realistic field-scale gas-condensate case using corner point gridding
with 18 vertical wells and 18 hydraulic fractures modeled by the EDFM.

Heterogenous distribution of matrix porosity, permeability, and water saturation was honored, as shown
in Fig. 12.
SPE-211396-MS 13

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

Fig. 12—A realistic field-scale gas-condensate case with heterogeneous distribution of

matrix properties: (a) matrix porosity; (b) matrix permeability; (c) matrix water saturation.

In a realistic large-scale modeling project, corner point geometry is usually used to represent complexity
in geology and petrophysics. The LGR within a corner point grid is refined in proportion with respect to
14 SPE-211396-MS

the grid size in most reservoir simulators, leading to an irregular fracture representation regarding fracture
orientation, and fracture geometry (height, aperture, and equivalent half length). Fig. 13 shows two examples
of fractures in corner point grids via LGR. The green lines represent the geomechanically consistent fracture
planes in each cell, which are perpendicular to the minimum horizontal stress. The red lines are fractures
in LGR in corner point grids. The difference in the orientations between the green color and the red color
led to not only different orientations but also different fracture geometry, size, and surface areas in most
cases, totally dictated by the shape and size of corner point grids. In most cases, the fracture surface area

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

in LGR is more than that of a 2D planar fracture.

Fig. 13—LGRs in corner point cells lead to non-physical representation of the

fracture orientations, geometry, size, and surface areas: (a) example 1; (b) example 2.

Fig. 14(a) shows an example from the sector model with 18 wells, while Fig. 14(b) shows an amplified
of one of the hydraulic fractures by LGR. Fig. 14(a) shows that every hydraulic fracture has its own
orientation which is dictated by these corner point cells where fractures penetrated, and they are not parallel
to each other. In reality, they should be all parallel to the assumed minimal horizontal stress direction. If
the minimum horizontal stress direction changes spatially, the corner point cells have no intrinsic way to
ensure that each fracture orientation is consistent with the local minimal horizontal stress direction, since
the corner point gridding is not intrinsically related to the local minimal horizontal stress direction. It is not
saying that fractures should be planar ones. But instead, when we intended to model the hydraulic fractures
with a simplified 2D planar style, the corner points cells fail to do so. The EDFM can model complex 3D
fractures which is beyond the focus of this work though.
SPE-211396-MS 15

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

Fig. 14—Visualization of hydraulic fractures modeled by the LGR method in
corner points: (a) 18 hydraulic fractures; (b) an amplified of one hydraulic fracture.

The failed representation of planar fractures by LGR in the above model leads to a different dynamic
response from that modeled by EDFM. Here is the difference between what we intended and expected from
the planar fractures and what corner points LGR gives, which are shown in Fig. 15. The LGR approach
overestimates the condensate production due to bigger total fracture surface areas than in EDFM.
16 SPE-211396-MS

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

Fig. 15—Comparison of cumulative oil production between the LGR and the EDFM
methods for the field case study (EDFM avoids over-optimistic forecast via LGR)

In addition to the geomechanics-consistent representation of fractures, the CPU time for the EDFM is at
least one order shorter than the LGR-based model, from 8 hours to 45 minutes. And an appropriate domain
decomposition further reduced it to about 9 minutes. The overall speed-up factor is about 50.

Summary and Conclusions

In this study, we coupled the EDFM method with the Nexus reservoir simulator in a non-intrusive manner
for the first time. The validation of EDFM against LGR was examined carefully using a simple case first
and then a gas-condensate sector model with 18 wells in Cartesian gridding. It was then applied to a sector
model with corner point grids. The following are the major conclusions:
1. Great agreements between the EDFM method and the LGR method were achieved using the Nexus
reservoir simulator, validated by a simple mechanistic model and a sector model in a gas condensate
reservoir
2. The computational efficiency of the EDFM over the LGR is an evident advantage, decreasing running
time by at least one order in this field case.
3. The EDFM enables a geomechanically-consistent representation of the fracture orientation, geometry,
and surface areas in a realistic corner point gridding model, overcoming the limitation of LGR in
corner point gridding. It will avoid over-estimate the surface areas which led to an over-optimistic
production forecast.

References
Cavalcante Filho, J.S.A., Shakiba, M., Moinfar, A., and Sepehrnoori, K. 2015. Implementation of A Preprocessor for
Embedded Discrete Fracture Modeling in An IMPEC Compositional Reservoir Simulator. Paper SPE 173289, SPE
Reservoir Simulation Symposium, Houston, Texas, 23–25 February.
Chen, H., Onishi, T., Olalotiti-Lawal, F., and Datta-Gupta, A. 2020. Streamline Tracing and Applications in Embedded
Discrete Fracture Models. Journal of Petroleum Science and Engineering, 188: 106865.
Cipolla, C.L., Fitzpatrick, T., Williams, M.J., and Ganguly, U.K. 2011. Seismic-to-Simulation for Unconventional
Reservoir Development. Paper SPE 146876, SPE Reservoir Characterization and Simulation Conference and
Exhibition, Abu Dhabi, UAE, 9–11 October.
Cipolla, C.L., Lolon, E., Erdle, J. C., and Rubin, B. 2010. Reservoir Modeling in Shale-Gas Reservoirs. SPE Reservoir
Evaluation & Engineering, 13 (4): 638–653.
SPE-211396-MS 17

Du, S., Liang, B., and Lin, Y. 2017. Field Study: Embedded Discrete Fracture Modeling with Artificial Intelligence
in Permian Basin for Shale Formation. Paper SPE 187202, SPE Annual Technical Conference and Exhibition, San
Antonio, Texas, 9–11 October.
Fiallos Torres, M., Morales, A., Yu, W., and Miao, J. 2021. Characterization of Complex Hydraulic Fractures in Eagle Ford
Shale Oil Development through Embedded Discrete Fracture Modeling. Petroleum Exploration and Development, 48
(3): 713–720.
Flemisch, B., Berre, I., Boon, W., Fumagalli, A., Schwenck, N., Scotti, A., Stefansson, I., and Tatomir, A. 2018.
Benchmarks for Single-phase Flow in Fractured Porous Media. Advances in Water Resources, 111: 239–258.

Downloaded from http://onepetro.org/SPEADIP/proceedings-pdf/22ADIP/2-22ADIP/D021S041R004/3032367/spe-211396-ms.pdf by Texas A&M University, Wei Yu on 23 February 2023

Hui, M.H., Karimi-Fard, M., Mallison, B., and Durlofsky, L.J. 2018. A General Modeling Framework for Simulating
Complex Recovery Processes in Fractured Reservoirs at Different Resolutions. SPE Journal, 23 (2): 598–613.
Jiang, J., and Younis. R.M. 2016. Hybrid Coupled Discrete-Fracture/Matrix and Multicontinuum Models for
Unconventional-Reservoir Simulation. SPE Journal, 21 (3): 1009–1027.
Karimi-Fard, M., and Durlofsky, L.J. 2016. A General Gridding, Discretization, and Coarsening Methodology for
Modeling Flow in Porous Formations with Discrete Geological Features. Advances in Water Resources, 96: 354–372.
Moinfar, A., Varavei, A., Sepehrnoori, K., and John, R.T. 2013. Development of A Coupled Dual Continuum and Discrete
Fracture Model for the Simulation of Unconventional Reservoirs. Paper SPE 163647, SPE Reservoir Simulation
Symposium, The Woodlands, Texas, 18–20 February.
Moinfar, A., Varavei, A., Sepehrnoori, K., and Johns, R.T. 2014. Development of An Efficient Embedded Discrete Fracture
Model for 3D Compositional Reservoir Simulation in Fractured Reservoirs. SPE Journal, 19 (2): 289–303.
Panfili, P., and Cominelli, A. 2014. Simulation of Miscible Gas Injection in A Fractured Carbonate Reservoir using
An Embedded Discrete Fracture Model. Paper SPE 171830, Abu Dhabi International Petroleum Exhibition and
Conference, Abu Dhabi, 10–13 November.
Rubin, B. 2010. Accurate Simulation of Non-Darcy Flow in Stimulated Fractured Shale Reservoirs. Paper SPE 132093,
SPE Western Regional Meeting, Anaheim, California, 27–29 May.
Sepehrnoori, K., Xu, Y., and Yu, W. 2020. Embedded Discrete Fracture Modeling and Application in Reservoir Simulation,
1st Ed., Publisher: Elsevier, Cambridge, USA. ISBN: 9780128218723.
Shakiba, M., and Sepehrnoori, K. 2015. Using Embedded Discrete Fracture Model (EDFM) and Microseismic Monitoring
Data to Characterize the Complex Hydraulic Fracture Networks. Paper SPE 175142, SPE Annual Technical
Conference and Exhibition, Houston, Texas, 28–30 September.
Xu, Y. 2015. Implementation and Application of the Embedded Discrete Fracture Model (EDFM) for Reservoir Simulation
in Fractured Reservoirs. Master thesis, University of Texas at Austin, Austin, Texas (December 2015).
Xu, Y., and Sepehrnoori, K. 2019. Development of an Embedded Discrete Fracture Model for Field-Scale Reservoir
Simulation with Complex Corner-Point Grids. SPE Journal, 24 (4): 1552–1575.
Xu, Y., Cavalcante Filho, J. S. A., Yu, W., and Sepehrnoori, K. 2017. Discrete-Fracture Modeling of Complex Hydraulic-
Fracture Geometries in Reservoir Simulators. SPE Reservoir Evaluation & Engineering, 20 (2): 403–422.
Xu, Y., Yu, W., and Sepehrnoori, K. 2019. Modeling Dynamic Behaviors of Complex Fractures in Conventional Reservoir
Simulators. SPE Reservoir Evaluation & Engineering, 22 (3): 1110–1130.
Xu, Y., Yu, W., Li, N., Lolon, E., and Sepehrnoori, K. 2018. Modeling Well Performance in Piceance Basin Niobrara
Formation Using Embedded Discrete Fracture Model. Paper URTeC 2901327, Unconventional Resources Technology
Conference, Houston, Texas, 23-25 July.
Yu, W., and Sepehrnoori, K. 2018. Shale Gas and Tight Oil Reservoir Simulation, 1st Ed., Publisher: Elsevier, Cambridge,
USA. ISBN: 978-0128138687.
Yu, W., Wu, K., Liu, M., Sepehrnoori, K., and Miao, J. 2018a. Production Forecasting for Shale Gas Reservoirs with
Nanopores and Complex Fracture Geometries Using An Innovative Non-Intrusive EDFM Method. Paper SPE 191666,
SPE Annual Technical Conference and Exhibition, Dallas, Texas, 24–26 September.
Yu, W., Xu, Y., Weijermars, R., Wu, K., and Sepehrnoori, K. 2018b. A Numerical Model for Simulating Pressure Response
of Well Interference and Well Performance in Tight Oil Reservoirs with Complex-Fracture Geometries Using the Fast
Embedded-Discrete-Fracture-Model Method. SPE Reservoir Evaluation & Engineering, 21 (2): 489–502.