A Better Modeling Approach for Hydraulic Fractures in Unconventional Reservoirs

OUTLINE Numerical Simulation: Comparison of Conventional and NEW Approaches NEW Approach as a Modeling Tool (understanding what has occurred) Field Examples Predictive Tool (investigating what might occur) Field Examples

What Is Our Goal? To quantify the impact of different strategies Well placement Well spacing Well orientation Number of stages Fracture treatment rates Fracture treatment volumes Cluster spacing (if applicable) Perforation density (if applicable)

How Do We Achieve The Goal? Unlike the early days, we have thousands of wells and performance data Post-mortem analysis is the key to understand the controlling parameters This can only be achieved by sophisticated approaches that can account for the interaction among controlling parameters Must be able to predict outcomes for different well placement/completion strategies Must be able to predict outcomes for multi-well applications where interference is important

How Do We Achieve The Goal? We need very sophisticated, integrated (geomechanics/flow) simulation models that can be quickly calibrated for: Fracking operation for all stages Flow-back period for frack fluid Production period for oil/gas/water Use the calibrated models to study alternatives: Well placement, orientation, spacing Completion design Frack operation

Conventional modeling approach Estimate reservoir matrix and natural fracture properties

Conventional modeling approach Estimate reservoir matrix and natural fracture properties Assume SRV geometry Estimate fracture height Estimate fracture half length Estimate fracture frequency Estimate distribution

NEW modeling approach Estimate reservoir matrix and natural fracture properties

NEW modeling approach Estimate reservoir matrix and natural fracture properties Generate SRV geometry and properties as part of the calibration process

NEW modeling approach Estimate reservoir matrix and natural fracture properties Generate SRV geometry and properties as part of the calibration process Calibrate to the fracture treatment, flow back and production periods Calibration through tuning of the geomechanical properties which define the SRV parameters fracture height fracture half length fracture frequency distribution (complexity, location of complexity)

Conventional\NEW approach Conventional NEW

Conventional\NEW approach Conventional Example~10 yrs NIEW 5740 ft 2500 ft Conventional NEW Difference in EUR Difference in Drainage Area

SRV Generation What We Used to Think

SRV Generation What It Really Looks Like

NEW Approach as a Modeling Tool Use a finite difference simulator with geomechanical capabilities, in dual porosity mode, to simulate the life of a hydraulically fractured well from the first stage of fracturing to the end of its productive life.

NEW Approach as a Modeling Tool SENSOR is a finite difference simulator with pseudo geomechanical capabilities Generates fractures by simulating the growth of the SRV during the frac treatment MatchingPro is an assisted history matching program Introduction of geomechanical properties multiplies the complexity of the history matching process

NEW Approach as a Modeling Tool Accounts for net pore pressure (stress) changes from initial conditions throughout the frac treatment (stage by stage) and during subsequent depletion

NEW Approach as a Modeling Tool Accounts for net pore pressure (stress) changes from initial conditions throughout the frac treatment (stage by stage) and during subsequent depletion Process allows for tensile and shear rock failures Accordingly the net pore pressure impacts fracture pore volume and transmissibility and the matrix-fracture communication (TEX) change

Basic Concepts Pore Pressure Increase in pore pressure results in decrease of net effective stress ( ') eff P pore After initiation, fractures propagate perpendicular to minimum horizontal stress Extent of fracture depends on rock properties Direction and magnitude depend on principal stresses

Mohr-Coulomb Failure Criteria This is the rock in its natural state Cohesive strength - S 0 σt σ 3 Cohesive strength and Coefficient of Internal friction different for different rocks σ 1 σ n

Mohr-Coulomb Failure Criteria τ σt σ 3 Initial pressure σ 1 σ n

Mohr-Coulomb Failure Criteria τ This is the rock in its natural state Stronger shear strength Less anisotropy σt σ 3 Initial pressure σ 1 σ n

Mohr-Coulomb Failure Criteria τ When we increase the pore pressure, we reduce the net effective stress ( ') eff P pore Shear Failure Increase of Pore Pressure σt σ 3 σ 1 σ n

Mohr-Coulomb Failure Criteria τ Shear Failure Increase of Pore Pressure σt σ 3 σ 1 σ n

Mohr-Coulomb Failure Criteria τ Finally, tensile failure is triggered causing dilation of the pore space and propagation of the hydraulic fracture. Direction of the fracture depends on stress directions Extent of the fracture depends on rock properties Complexity of the fracture depends on stress anisotropy, rock properties and injection rates Tensile Failure Increase of Pore Pressure σt σ 3 σ 1 σ n

Basic Concepts Hydraulic Fracturing Frac fluid Injecting fluid into the reservoir increases pore pressure Increasing the pore pressure decreases the effective stress At low effective stresses the rock undergoes shear and tensile failures Those failures generate flow pathways Source: http://www2.epa.gov/hydraulicfracturing

Basic Concepts Hydraulic Fracturing Simple fractures Complex fractures Source: Hydraulic Fracture complexity and treatment design in Horizontal wells, Craig Cipolla Source: Warpinski, N.R. ; Mayerhofer, M.J. ; Vincent, M.C. ; Cipolla, C.L. ; Lolon, Stimulating Unconventional Reservoirs: Maximizing Network Growth While Optimizing Fracture Conductivity, Unconventional reservoir modeling conference, 2008. SPE 114173

Fracture Complexity and Distribution Flow in dual porosity systems MATRIX FRACTURE MATRIX MATRIX TEX determines the flow between matrix and the fracture More complex fractures result in more fluid transfer between matrix and fracture media Bi-Wing Fracture Simple Geometry Increasing TEX Increasing fracture complexity Increasing fracture density within the matrix adjacent to the bi wing frac. Bi-Wing Fracture Complex Geometry

Example SRV Generation Stage by Stage SRV growth The next slides show the stage by stage SRV generation (14 stages) Color indicates TEX Higher TEX values indicate greater communication between the fracture and matrix systems

Study #1

Calibration to the Frac Stages

SRV Aspect Ratio View from heel to the toe 300 ft Height 750 ft Width TEX Value

SRV Aspect Ratio Side view. Heel is on right. 300 ft Height TEX Value

SRV Aspect Ratio View from top. Heel is on Right 750 ft Width 450 ft Width TEX Value

After the Frac Treatment Hydraulic fractures start closing once the treatment is completed Persistent leak-off decreases the pressure initiates closure Leak-off is loss/imbibition of the fluid into the matrix! Production - decreases the pore pressure closure continues Matrix porosity/permeability reduces also Tensile failures - close most rapidly Proppant is used to keep tensile fractures open Shear failures - close more slowly or stay open Rock dislodging and fragments act as proppant

SRV Closure After the SRV is generated during the hydraulic fracture treatment, the connectivity reduces as the result of depletion Simulation data table determines the transmissibility reduction as a function of pore pressure

Fracture Closure Log(T-multiplier) Pinit Pfrac Pore Pressure

Fracture Closure Stem of the fracture: Log(T-multiplier) Tip of the fracture: Pore Pressure

Fracture Closure Stem of the fracture: where proppant is accumulated and is effective Log(T-multiplier) Tip of the fracture: where the proppant cannot get to and is ineffective Pore Pressure

Fracture Closure Stem of the fracture Log(T-multiplier) Tip of the fracture closes during flowback period Pore Pressure

Fracture Closure Log(T-multiplier) The propped portion of the hydraulic fracture stays open well below the initial reservoir pressure Pore Pressure

Assisted History Matching (AHM) Large number of parameters means that history matching by hand is difficult MatchingPro is an assisted history match (AHM) program that uses an objective function to assess and generate new solutions User specifies which parameter values to vary and by how much

AHM Objective function based on the following data Hydraulic Fracturing Period Inject measured volumes of fluid Constrained by maximum injection BHP Flow back and Production Period Produce correct quantities of fluid Oil Gas Water Match the pressure of the natural flow period Match the monthly volumes of produced fluids

AHM Worse Case (Obj Func = ~900) Best Case (Obj Func = ~75) Approximately 200 runs

AHM Results THP Water Rate Gas Rate Oil Rate Worse Case - Blue Best Case - Black

MatchingPro Simulating fracture treatments results in a large number of unknown parameters Parameter space Up to 18 parameters during investigation phase

MatchingPro These eight variables proved to be the most important for one of our projects Number of parameters reduced in later phase of calibration CTEX: TEX compressibility CX: TX compressibility OWC: Oil Water Contact SORW: Residual oil saturation to water SRV: SRV Growth Factor TEXS: TEXMOD from shear failure TEXT: TEXMOD from tensile failure TX: X direction transmissibility modifier

SRV Simple or SRV Complex SRV with Simple Fractures SRV with Complex Fractures MatchingPro

Project Results 4 Projects: Project #1: Bakken Project #2: Bakken(same field as #1) Project #3: Wolfcamp Project #4: Eagleford

Project #1 Solid lines represent simulated data. Colored points indicate measured data

Project #1 Solid lines represent simulated data. Shaded areas indicate measured data

Project #2 Solid lines represent simulated data. Colored points indicate measured data

Project #2 Solid lines represent simulated data. Shaded areas indicate measured data

Project #3 Solid lines represent simulated data. Colored points indicate measured data

Project #3 Solid lines represent simulated data. Shaded areas indicate measured data

Project #4 Well 1

Project #4 Well 2 Frac volume ± 5 % Length ± 10 %

NEW Approach as a Predictive tool Conventional approach has limited predictive capability if completion practices change Once calibrated, NEW approach has predictive capabilities Alternative scenarios can be run to quantify the impact of different strategies Well placement/spacing Well orientation Fracture treatment volumes Fracture treatment rates Number of stages Placement of stages

Investigate Well Completions 18 stages, X volume, X min 21 stages, X volume, X min Accelerated Performance

Optimize Fracture Treatment Volume Doubling of Frac Injection Rate Base 2 x Frac vol.

Optimize Well Orientation Orientation 1 Orientation 2

Multiple Wells Project Description: All wells use the same drilling and completion strategy First well drilled in 2008 and produces Second well drilled in 2011 and produces Third well to be drilled in 2013 Automatically accounts for affect of stress level changes from one well fracture area to another over time

Multiple Wells 2008

Multiple Wells 2011

Multiple Wells 2013

How Are The Forecasts Holding Up? HM ~ 8 months Here is how we compare ~22 months later

How Are The Forecasts Holding Up? HM ~ 5 months Here is how we compare ~20 months later

How Are The Forecasts Holding Up? HM ~ 12 months Here is how we compare ~13 months later

How Are The Forecasts Holding Up? HM ~ 6 months Here is how we compare ~14 months later Known change from Gas lift to SP during this period

Summary Analyze multiple wells in the same field Different hydraulic fracture treatments Understand the performance differences based on Reservoir quality Completion type Treatment volumes Treatment stages Optimize treatment practices and well spacing Supplemental recovery mechanisms

Questions? Thank You!