Fracture Modeling

By

Pinnacle Technologies
 FracproPT System - Highlights
• Estimates fracture geometry and proppant placement in real-time
by net pressure history matching
• Provides unique tool to capture what is learned from direct fracture
diagnostics through calibrated model settings
• Performs near-wellbore tortuosity / perf friction analysis – allows
identification and remediation of potential premature screenout
problems
• Integrated reservoir simulator for production forecasting and matching
• Optimizes fracture treatment economics
• Supports remote access via modem or internet
• Contains preloaded libraries of stimulation fluids, proppants,
and rock properties for many lithologies
 FracproPT Module Interaction
DataAcqPT
Real-Time
Data Acquisition
Calibrated Model
Settings

Wellbore Information
Log/layer Information Treatment Data Production Data

FracproPT FracproPT FracproPT

Fracture Design Fracture Analysis Production Analysis

Estimated Production Forecast

Treatment
Fracture or Match
Schedule
Geometry

FracproPT
Economic
Optimization
Motivation for Frac Engineering & Diagnostics

Hydraulic
fracturing is
done for well
stimulation

NOT

for proppant
disposal
 Fracture Pressure
Analysis - Advantages
• Basic analysis data collected (in some sense) during every frac
treatment
• Relatively inexpensive and quick diagnostic technique to apply
• Provides a powerful tool for on-site diagnosis of fracture entry
problems
• Allows on-site design refinement based on observed fracture
behavior
 Fracture Pressure
Analysis - Limitations
• Fracture Entry Friction Evaluation
– Using surface pressure increases results uncertainty
– Problematic near-wellbore friction level variable
• Net Pressure History Matching
– Indirect Diagnostic Technique - frac geometry inferred from
net pressure and leakoff behavior
– Solution non-unique – careful & consistent application
required for useful results
– Technique most useful when results are integrated or
calibrated with results of other diagnostics
• Production data & welltest analysis
• Direct fracture diagnostics
 Example Application – “Pressure Out” on Pad

• Formation: Naturally fractured dolomite @ 8200’ (gas)

• Completion: 5-1/2” casing frac string, max. surface pressure 6000 psi;
70’ perf interval shot at 4 SPF, 90, 0.45” diameter hole;
Previously acidized with 70 gallons/ft 20% HCl
• Situation: Declining injectivity leading to “pressure-out” on pad
• Diagnosis: Severe near-wellbore fracture tortuosity
• Solution: 1 and 2 PPG proppant slugs very early in the pad to
screen out fracture multiples
 Example Application – “Pressure Out” on Pad
1400 psi friction
reduction (1st slug)
Proppant Conc (ppg) Btm Prop Conc (ppg)
20.00 Surf Press [Csg] (psi) Slurry Flow Rate (bpm)
6000
Max surface pressure 6000 psi

S/D#2: 300 no tortuosity at

16.00 end of pumping
4800 psi tortuosity

S/D#1: 1700
psi tortuosity;
12.00 small perf fric.
3600

8.00 Increased max

2400 prop conc

4.00
1200

0.00
0
0.0 28.0 56.0 84.0 112.0 140.
Example Application – Estimation of Realistic Fracture Half-
Length

• Formation: Hard sandstone @ 7600’ (gas) in West Texas

• Completion: 5-1/2” casing frac string; 40’ perf interval shot with 4
SPF, 90 phasing, 0.31” diameter holes
• Situation: Disappointing production performance for expected
600 ft fracture half-length (based on fracture
growth design without real-data feedback)
• Diagnosis: Sand/shale stress contrast much lower than
estimated, resulting in significant fracture
height growth and a much shorter fracture
half-length (250’)
• Solution: Utilize fracture pressure analysis to optimize fracture
treatment design
 Example Application – Estimation of Realistic Fracture
Half-Length

Geometry inferred design High stress contrast 0.3 psi/ft (based on

without real-data feedback Dipole Sonic log interpretation)
Btm Prop Conc (ppg) Net Pressure (A) (psi)
50.00 Slurry Rate (bpm) Prop Conc (ppg) 2
100.0 50

40.00 1
80.0 40

30.00 1
60.0 30

20.00 8
40.0 20

10.00 4
20.0 10

0.00
0.0 0
0.0 20.0 40.0 60.0 80.0 100.0
Time (min)
 Example Application – Estimation of Realistic Fracture
Half-Length

Geometry inferred design Observed net pressure does not match design net pressure response
without real-data feedback
Net Pressure (A) (psi)
Prop Conc (ppg) Btm Prop Conc (ppg)
2000 Observed Net (psi) Slurry Rate (bpm) 5
50.00 1
2000

1600 4
40.00 8
1600

1200 3
30.00 6
1200

800 2
20.00 4
800

400 1
10.00 2
400

0 0
0.00
0 0.0 20.0 40.0 60.0 80.0 100.0
Time (min)
 Example Application – Estimation of Realistic Fracture
Half-Length

Geometry inferred design

without real-data feedback
Btm Prop Conc (ppg) Net Pressure (A) (psi)
Slurry Rate (bpm) Prop Conc (ppg)
50.00 Observed Net (psi) Net Pressure (psi)
100.0 5
2000

40.00
80.0 4
1600

30.00
60.0 3
1200
Geometry inferred from net
20.00
pressure matching 40.0 2
800

10.00
20.0 1
400

0.00
0.0
0 0.0 20.0 40.0 60.0 80.0 100.0
Time (min)
Lower stress contrast (0.1 psi/ft) required to match observed net pressure
Confirmed with shale stress test in subsequent wells
Example Application -- Tip Screen-out Strategy To Obtain
Sufficient Conductivity

• Formation: High permeability layered sandstone at 6000 ft (oil)

• Completion: Deviated wellbore, 3-1/2” tubing frac string
30’ perf interval shot 4 SPF, 180 phasing oriented
perfs, 0.5” diameter holes
• Situation: Relatively poor post-frac production response for high
perm reservoir
• Diagnosis: Insufficient propped fracture conductivity
• Solution: Increase treatment size, and utilize on-site fracture
pressure analysis to consistently achieve tip
screenout for enhanced fracture conductivity
 Example Application -- Tip Screen-out Strategy To
Obtain Sufficient
ARCO Kuparuk Conductivity
River Unit 2K-15
A4 sand 6217'-6247' TVD 12/22/96
Btm Prop Conc (ppg)
Slurry Rate (bpm) Prop Conc (ppg)
50.00 Observed Net (psi) Net Pressure (psi) 50.00
100.0 750.0
750.0 Tip screen-out initiation

40.00
80.0 40.00
Pad fluid volume adjusted based 600.0
600.0 on leakoff behavior following
crosslink gel minifrac
Breakdown injection
30.00 30.00
60.0
450.0 450.0
Minifrac

20.00
20.00
40.0 300.0
300.0

10.00
20.0 10.00
150.0
150.0

0.00 0.00
0.0
0.0 0.0 60.0 120.0 180.0 240.0 300.0 0.0
Time (mins)
Net pressure match Pinnacle Technologies
Pad sizing for TSO design was done utilizing leakoff calibration with minifrac. The
net pressure match shows a significant increase in pressure due to tip screen-out
initiation
Example Application -- Tip Screen-out Strategy To Obtain
Sufficient Conductivity

• Production response in Kuparuk A sand limited by fracture

conductivity
• Tip screen-out obtained in more than 90% of treatments
– Sizing of pad size using calibration of leakoff coefficient key to
success
– On-site real-time closure stress analysis implemented on
every treatment to ensure proper pad size is pumped
Definition Of Net Pressure
Net Pressure is the Pressure Inside the Fracture
Minus the Closure Pressure
Net Pressure = 2,500 - 2,000 = 500 psi
Balloon Analogy For Opening Fracture With
Constant Radius
 Fluid Leakoff And Slurry Efficiency
LOW SLURRY EFFICIENCY
Vfrac (t)
efficiency (t) =
Vpumped (t)
Short Fracture High Filtration

HIGH SLURRY EFFICIENCY

Longer Fracture Low Filtration

Net Pressure Vs. Friction Pressure
Net Pressure Matching
 Basic Fracture Pressure Analysis Steps
Pre-frac completion
1 and fracture design

Determine fracture closure Repeat process in

2 stress and match permeability succeeding stages or
wells
Characterize friction
3 parameters using rate
stepdown tests Post-frac modeling review
and incorporate other
Determine observed fracture diagnostics
net pressure

4 Match model net Perform treatment

pressure to observed
net pressure
Interpret model results,
make engineering
decisions
Explore / bound
alternative explanations for
observed net
pressure
 Different Models
• 2D models
– Perkins, Kern and
Nordgren (PKN)
– Christianovitch,
Geertsma and De
Klerk (CGD)
– Radial Model
• 3D models
– Pseudo 3D models
– Lumped 3D models
– Full 3D models
– Non-planar 3D models
 Fracture Design and Analysis Evolution
Modeling without Real-Data Feedback

• Early designs (pre-1980) did not incorporate feedback from real data
• Fractures at that time were still smart enough to stay in zone

W ellbore

U se predicted
net pressure Pay Pay

P u m p ra te
Net
p res s ure P re d ic ted n et p re ss ure

P u m p tim e
 Fracture Design and Analysis Evolution
Modeling without Real-Data Feedback

• Early designs (pre-1980) did not incorporate feedback from real data
• Fractures at that time were still smart enough to stay in zone
• But measured net pressure was generally MUCH higher than model
net pressure

W ellbore

U se predicted
net pressure

?
Pay Pay
M ea s ure d ne t pre s su re

P u m p ra te
Net
p res s ure P re d ic ted n et p re ss ure

P u m p tim e
 Fracture Design and Analysis Evolution
Modeling with Net Pressure Feedback

• Net pressure history match can be obtained by adding new physics to

fracture models – Reason for the existence of FracproPT
• With the right assumptions and physics, inferred geometry has a
better chance to be correct

W ellbore

U se m easured
net pressure
Pay

M atc hin g m e as u red n et p re ss ure

Net w ith m o de l ne t pre s su re
p res s ure

P u m p tim e
 FracproPT Development Philosophy

• After development of pseudo-3D models (early 1980’s) the industry was

jubilant as it was now known how fractures really behaved -- or not ?
• Observed net pressures were consistently far higher than net pressures
predicted by these models (discovered in early 1980’s) -- parameter
sensitivity also inconsistent
• Development of Fracpro started in 1980’s with the aim to honor the
“message” contained in real-data
– Capturing the physics of details is not as important as honoring large-scale
elasticity and mass balance
– Calibrated simplified approximation with full 3D growth model, lab tests and field
observations
– Model calibration is now a continuous effort
 Fracture Modeling in FracproPT

• Wellbore Model • Proppant Transport

Model(s)
• Perforation and Near-
Wellbore Model • Acid Fracturing Model(s)
• Fracture Growth Model(s) • Backstress (poro-elastic)
Model
• Fracture Leakoff Model(s)
• Multiple Fracture Model
• Fracture Temperature
Model
 FracproPT is Just a Tool
• The FracproPT system contains several 2D models, a conventional 3D model,
an adjustable 3D model incorporating “tip effects”, and a growing number of
calibrated model settings
• There is NO “FracproPT answer”
• Designed for on-site engineering flexibility
• Quality of results are more user-dependent than model dependent
– Making the right engineering assumptions is
key
– Garbage in = garbage out
– The KEY is to honor the observed data with the
most reasonable assumptions possible
 Minimum Model Input Requirements
• Mechanical rock properties
– Young’s modulus (from core or sonic log)
– closure stress profile (injection/decline data or sonic log)
– Permeability (from PTA)
• Well completion and perforations
• Treatment schedule, proppant and fluid characteristics
• Treatment data
– With “anchor points” from diagnostic injections
– Recorded pressure, slurry rate and proppant concentration
• Surface pressure OK for decline match
• Deadstring or bottomhole gauge required for matching while
pumping
 Required to Obtain Observed Net Pressure
pnet ,obs  psurface  phydrostatic  p friction   closure
• Obtain surface pressure from service companies
recorded data
• Obtain hydrostatic head from staging and fluid/proppant
densities
• Obtain frictional components from S/D tests
• Obtain fracture closure stress from pressure decline
 “Typical” Fracture Treatment Data
Proppant Concentration (ppg)
20.00 Surface Pressure (psi) Slurry Rate (bpm)
3000 200.0

Net pressure ?
16.00 160.0
2400

12.00
1800 Friction ? 120.0

8.00
1200 80.0

Closure ?
4.00 40.0
600

Leak-off ?
0.00
0.0
0 50.00 58.00 66.00 74.00 82.00 90.00
Time (mins)
 Purpose Of Diagnostic Injections
• Provide “anchor points” for real-data (net pressure) analysis
• Obtain accurate measurement of the true net pressure in the
fracture
• On site diagnosis and remediation of proppant placement
– Near-wellbore tortuosity
– Perforation friction
– fluid leakoff
• Bottom line: provide accurate estimates of the fracture
geometry
 Recommended Diagnostic Injection
Procedures
Diagnostic Step When Fluid & Volume Purpose / Results
Breakdown Injection / rate Always ~50-100 Bbl KCl Establish injectivity; obtain small volume ISIP;
stepdown / pressure decline estimate closure pressure and formation permeability.
Crosslinked Gel Minifrac with New areas ~100-500 Bbl fracture Leakoff calibration;
proppant slug / rate stepdown / Real-time pad resizing fluid including 25-50 Net pressure sensitivity to volume and crosslink gel;
pressure decline TSO treatments Bbl proppant slug Characterize fracture entry friction;
(possible range 0.5-5 Evaluate near-wellbore reaction to proppant;
PPG) Screen out or erode near-wellbore multiple fractures.
End Frac Rate Stepdown / Always Minimum of 10 minute Characterize fracture entry friction;
Pressure Decline Monitoring decline data Post-frac leakoff calibration.
“Anchor Point”: Fracture Closure Stress
“Anchor Points”: Isip Progression
“Anchor Points”: Frictional Components
 Main Input Parameter - Permeability
• Matching perm is “permeability under fracturing
conditions” – not necessarily under production conditions
– Relative permeability issues
– Opening of natural fractures
– Relies on many other assumptions
• Keep it simple:
– only change permeability in pay interval.
– Keep permeability zero in shales
• If permeability profile is “known”, use Kp/Kl ratio for
matching instead
• Fix by matching decline slope of B/D KCl injection
 Main Input Parameter - Closure Stress
• Closure stress profile determines fracture shape
– Radial if stress profile is uniform (theoretical decrease in net pressure with
pump time)
– Confined height growth if closure stress “barriers” are present (theoretical
increase in net pressure with pump time)
• Effectiveness of “barrier” determined by
– Closure stress contrast
– Level of net pressure
• “Typical” sand-shale closure stress contrast 0.05 - 0.1 psi/ft
– Higher if there has been significant depletion (~2/3 of pore pressure change)
– Lower if sands and shales are not clean
• When do you change it?
– Increase contrast when net observed pressures are higher
– Increase contrast when fracture is more confined (up to 1.0 psi/ft)
 Closure Stress Profile

• Closure stress min determines minimum

pressure to open a fracture
• Usually closure increases with depth
• Closure stress is lithology dependent (shales
usually higher than sands)
• Represents only the minimum principal
stress component in the vicinity of the well
 Main Input Parameter - Young’s Modulus
• Modulus should be obtained from static tests (preferably similar to
fracturing conditions)
– Dynamic modulus two times or more larger than static modulus (use with
caution !)
• Once modulus is determined, this should be a FIXED parameter in a net
pressure matching procedures
• An increase in Young’s modulus results in less fracture width (for the
same net pressure)
• For simple radial model: Lfrac  E1/3 (for the same net pressure)
• Modeling results not extremely sensitive to modulus.
• When do you change it?
– With low moduli in GOM environment when modulus uncertainty is high
– Character of TSO net pressure slope depends on modulus
 Different Methods To Obtain Fracture
Closure Stress (in Pay)
• Pressure decline analysis
• Square-root time plot
• G-function plot
• Log-log plot
• Rate normalized plot
• Horner plot (lower bound)
• Flow pulse technique
• Flow back test
• Steprate test (upper bound)
• Hydraulic Impedance testing (HIT)
 Pressure Decline Analysis

• Pressure decline after a mini-frac passes through

two flow regimes:
– Linear flow regime; Pressure decline depends on:
• fluid leakoff rate
• fracture compliance
– Radial flow regime; Pressure decline depends on:
• reservoir diffusivity
• Closure stress (pressure) is identified by the
transition between the two flow regimes
 What Can You Obtain From Pressure
Decline Analysis?
• Fracture closure pressure (minimum stress)
• Fluid efficiency
• Leakoff coefficient, reservoir permeability and
pressure
• Fracture geometry estimate

T p +Tc
Tc
B otto m h ole p re ssu re

E fficie n cy ~
TcTc+ Tp

IS IP

P ne t C lo su re

R a te
Tp Tc
Tim e
 Pressure Decline Analysis – Square-root
Time Plot
Surf Press [Csg] (psi) Implied Slurry Efficiency (%)
0 Meas'd Btmh Press (psi) Surf Press [Csg] (psi) 200.0
8500 5000

BH Closure Pressure: 5423 psi

Closure Stress Gradient: 0.658 psi/ft
Closure Time: 6.0 min
-140 160.0
Pump Time: 3.0 min
7800 4300
Implied Slurry Efficiency: 53.0 %
Estimated Net Pressure: 1093 psi

-280 120.0
7100 3600

-420 80.0
6400 2900

-560 40.0
5700 2200

-700 0.0
5000
0.0 2.0 4.0 6.0 8.0 10.0 1500
Time (min)
 Pressure Decline Analysis – G-function Plot
Implied Slurry Efficiency (%)
Surf Press [Csg] (psi) Surf Press [Csg] (psi)
200.0 Meas'd Btmh Press (psi) Surf Press [Csg] (psi) 1000
800.0
5000
8500

160.0
800
640.0
4300
7800
BH Closure Pressure: 5748 psi
Closure Stress Gradient: 0.697 psi/ft
Closure Time: 3.5 min
120.0 Pump Time: 3.0 min
Implied Slurry Efficiency: 43.1 % 600
480.0
Estimated Net Pressure: 767 psi 3600
7100

80.0
400
320.0
2900
6400

40.0
200
160.0
2200
5700

0.0
0
0.0
0.000 0.620 1.240 1.860 2.480 3.100 1500
5000
G Function Time
 Pressure Decline Analysis – Log-log Delta
Pressure Plot
Implied Slurry Efficiency (%)
10000 Delta Pressure (psi) Delta Pressure (psi)

1000

BH Closure Pressure: 5637 psi

100 Closure Stress Gradient: 0.684 psi/ft
Closure Time: 4.3 min
Pump Time: 3.0 min
Implied Slurry Efficiency: 46.6 %
Estimated Net Pressure: 879 psi

10
0.100 1.000 10.000 100.00
Time (min)
 Steprate/Flowback test

• Step Rate Test

– Start at matrix rate
– Increase in steps until fracture extended ( 1 to 10 BPM)
– Provides upper bound for closure
– Can determine if you are fracturing at all
• Flowback at Constant Rate
 Pump-In/Flowback/Shut-in Test (SPE 24844)
• High perm well where the FB-SI is run after the gel calibration test
– otherwise volume of fracture is to small due to high leakoff
• Here ‘frac WB pinch’ is identified at closure: very small

FB induced
" wellbore pinch”

~ 30 psi

SI-Rebound < p c
" near-well independent of
pinch " " tortuosity"
SPE PF Feb '97

~ 15 min
 Tortuosity Can Be Measured:
Stepdown Test
• Instantaneous rate changes, e.g. 30, 20, 10 and 0 BPM --
exact rates are unimportant, but changes should be
abrupt
• Implemented easiest by taking pumps off line
• Each rate step takes about 20 seconds -- just enough to
equilibrate the pressure
• Fracture geometry should not change during stepdown --
total stepdown test volume small compared to test
injection volume (note: pfrac not proportional to Q1/4 during
stepdown test)
• Use differences in behavior of the different friction
components with flow rate
What Is Tortuosity? Width Restriction Close To
Wellbore
Width Restriction Increases Necessary
Wellbore Pressure
 Tortuosity Leads To Large Pressure Drop In
Fracture Close To Well

Net fracturing
pressure

High
Near-wellbore friction
Low

Pressure after shut-in

Fracture tip

Wellbore Distance into fracture

Fractures Grow Perpendicular To The Least Principle
Stress -- But What Happens At The Wellbore ?
Near-wellbore Friction Vs. Perforation
Friction
 Near-wellbore Friction Vs.
Perforation Friction
Btm Slry Rate (bpm) Meas'd Btmh (psi)
50.00 850

40.00 790

30.00 730

20.00 670

10.00 610

0.00 550
17.00 17.80 18.60 19.40 20.20 21.00
Time (min)
Tortuosity Can Be Measured: Stepdown Test
• Perforation friction dominated regime

Source: “SPE paper 29989 by C.A. Wright et al.

Tortuosity Can Be Measured: Stepdown Test
• Near-wellbore friction dominated regime
Maximum Treating Pressure Limitation Is Reached -- Can’t
Pump Into Zone
High entry friction

High perf friction Severe fracture tortuosity

Re-perforate Use proppant slugs

Ball-out treatment Initiate with high viscosity fluid
Spot acid Increase gel loading
Increase rate
Future wells may have altered
completion strategy such as
FEWER perfs
 Net Pressure Matching
• Match “observed” net pressure with calculated “model”
net pressure
• Observed net pressure obtained from surface or
downhole treatment pressure
– Correct for fracture closure, frictional effects and hydrostatic
• Model net pressure can be changed to match observed
net pressures using the following general “knobs” (see
next page)
History Matching “Anchor Points”: Shut-in Pressure Decline
Slope and Net Pressure Level
History Matching “Anchor Points”: Shut-in Pressure Decline
Slope and Net Pressure Level
Btm Prop Conc (ppg)
Slurry Rate (bpm) Prop Conc (ppg)
25.00 Observed Net (psi) Net Pressure (psi) 25.00
125.0
2500
2500

20.00
20.00
100.0
2000
2000

15.00
15.00
75.0
1500
1500

10.00
10.00
50.0
1000
1000

5.00
5.00
25.0
500
500

0.00
0.00
0.0
0.0 30.0 60.0 90.0 120.0 150.0 0
0
Time (min)
FracproPT Net Pressure Matching Parameters
• “Decline Slope” parameters
– Permeability
– Wallbuilding coefficient (Cw)
– Pressure-dependent leakoff (Multiple fracture leakoff factor)
• “Level” parameters
– (Sand-shale) Closure stress contrast
– Fracture complexity (Multiple fracture opening/volume factor)
– Tip effects coefficient
– Proppant drag exponent
– Tip screen-out backfill coefficient
– (Young’s modulus)
• “Geometry” parameters
– Composite layering effect
– Crack opening / width coupling coefficient
 Net Pressure Matching Strategy
• B/D Injection
– Level: Tip effects, Fracture complexity
– Decline slope: permeability
• Minifrac
– Level: Tip effects, Fracture complexity
– Decline slope: Wallbuilding coefficient Cw
• Prop frac:
– Level (low perm): stress contrast, proppant drag
– Level (high perm): TSO backfill, Young’s modulus, stress
contrast, proppant drag
– Decline slope: Pressure-dependent leakoff
– Geometry: composite layering effect, width decoupling
 FracproPT Net Pressure Matching Parameters
Response with Parameter Increase +

Efficiency
Pressure

Length

Height
Slurry

Width
Half-
Net
Parameter Range Unit Mainly Affects When

Permeability 0.000001 - 10000 mD Decline slope B/D injection - - - -

Wallbuilding Coefficient Cw 0.0001 - 0.1 ft/(min)0.5 Decline slope Minifrac - - - -

Pressure-dependent Leakoff* >=1 fracs Decline slope Prop frac - - - -

Fracture Complexity** >=1 fracs Level All injections + +

Stress Contrast (Pay-Barrier) 0.00 - 0.40 psi/ft Level All injections + +

Tip Effects 0.00001 - 0.4 - Level All injections - +

Prop Frac
Proppant Drag 0 - 25 - Level TSO +

TSO Backfill 0.0 - 1.0 - Level TSO +

Composite Layering 1 - 1000 - Geometry All injections + + -

Width Decoupling 0.01 - 1.00 - Geometry All injections - +

* Multiple fracture leakoff factor. ** Multiple fracture volume&opening factor
 Main Matching Parameters – Tip Effects
Coefficient (Gamma 2)
• How does it work?
– This parameter controls the near-tip pressure drop and thus the net
pressure level in the fracture.
– Mimics increased fracture growth resistance at the tip
• Tip process zone (with opening fractures) slows down fracture growth
• Non-linear rock behavior at large differential compressional stress
• When do you change it?
– Increase from default 0.0001 up to 0.4 when observed net pressure is
lower than model (w/o multiples)
– When fluid viscosity change has significant effect on observed net
pressure behavior
 Tip Effects Coefficient
Net pressure decline slope w/ distance represents Gamma 2)

pnet Non-linear elastic model (Gamma 2 = 0.0001)

Linear elastic model (Gamma 2 = 0.4)

Lf

Non-linear elastic model

wfrac
Linear elastic model

Lf
Tip Effects -- Increased Fracture Growth
Resistance
 Process Zone Around
Fracture Tip

• Experiments by Shlyapobersky
reveal fracture process zone
• Process zone is scale
dependent, and results in
multiple fractures ahead of
hydraulic fracture tip
• Can result in higher net
pressures to propagate fracture
Main Matching Parameter – Multiple Fractures
• How does it work?
– Opening and volume factor control the degree of fracture complexity
using the amount of overlapping “equivalent” (equal sized) fractures
– Leakoff factor can mimic increase leakoff or pressure-dependent
leakoff
• When do you change it?
– When observed net pressure with default Gamma 2 (0.0001) is
significantly higher than model net pressure
– Use specific starting points for distributed limited entry and point
source perforation strategies
– Use strict rules
• Only change during injections
• Tie opening and volume factors for “point source” perfs
• Tie leakoff and volume factors for “distributed limited entry” perfs
Multiple Hydraulic Fractures In FracproPT
Modeling Approach for Multiple Hydraulic Fractures
Situation Equivalent Equivalent Equivalent
number of number of number of
growing fractures fracs
multiple with leakoff competing
fracs (MV) (ML) for width
(MO)

Equivalent number of 3 3 1
spaced identical fractures
without interference

3 2 2

Equivalent number of 3 1 3
fractures competing
For width
 Evidence for the Simultaneous Propagation of
Multiple Hydraulic Fractures
• Core through and mineback experiments
• Direct observations of multi-planar fracture propagation
• Fracture growth outside plane of wellbore
• Observation of high net fracturing pressures
• Continuous increases in ISIPs for subsequent injections

Conclusion: multiple fractures are the rule rather than the

exception
Multiple Strands in a Propped Fracture

NEVADA TEST SITE

MINEBACK

Courtesy: N.R. Warpinski, Sandia Labs

 Use Multiple Hydraulic Fractures Prudently
for Modeling Purposes
• Potential causes for high net pressures:
– Confined fracture height growth
– Increased fracture closure stress due to pore
pressure increase
– Higher Young’s modulus than anticipated
– Fracture tip effects
– Tip screen-out initiation
– Simultaneously propagating multiple hydraulic
fractures
 Multifrac Modeling Approach For Limited
Different Perforation Strategies
R egion of
nea r-w e llbore
tortuo sity

C oncep tua l sim plification of

nea r-w e llbore tortu osity
and m ultiple fractu re s

M odeling strategy for

nea r-w e llbore tortu osity
and m ultiple fractu re s
 Main Matching Parameters – Proppant Drag
Exponent
• How does it work?
– Mimics the increase in frictional pressure drop along the fracture as
proppant is introduced
– Controls how much the proppant in the fracture slows the fracture length
and height growth.
– Separate terms for Upper and Lower height growth calculated. Length effect
is based on average of upper and lower terms.
– Once a stage has become packed with sand (“immobile proppant bank”),
there is no more growth in that direction
– If both an upper and lower stage are dehydrated, quadratic backfill model
takes over (if enabled)
• When do you change it?
– Significant proppant induced observed net pressure increase during
proppant stages (that is not due to TSO)
Main Matching Parameter – Quadratic Backfill
Exponent
• How does it work
– When fracture height and length growth are stopped due to
dehydration of an upper and lower stage, quadratic backfill
model starts working (if enabled)
– Quadratic backfill is based on the idea the the fracture
dimension controlling fracture stiffness will decrease as the
fracture fills with immobile packed proppant from the tip back to
the wellbore.
• When do you change it?
– Increase it when the TSO-induced observed net pressure rise is
steeper than model predicts
 New Matching Parameter – Width Coupling
Coefficient
• How does it work ?
– Multiplier for Gamma 1 representing how fracture width is decoupled
along fracture height
– We will provide automatic correlation as a function of composite
layering effect
• When do you change it ?
– Decrease it to trade fracture width for half-length
– Decrease it to mimic reduced coupling “shear-decoupling” over
fracture height (also associated with use of composite layering effect)

 = WcpnetR/ E
 pnet

R
 Main Matching Parameters – Composite
Layering Effect
• How does it work ?
– This parameter controls the near-tip pressure drop in each
individual layer
• When do you change it ?
– Increase in layer adjacent to pay zone if no other confining
mechanism can explain actual level of fracture confinement
– Keep unity in pay zone
Estimating Frac Dimensions Using Real Data
And Radial Frac Assumption:
“Back-of-the-Envelop Model”
1

2  3 eVE  3

Mass balance e  V  R 2w R  
3  4 pnet 
1

2 pnet R  6 eVpnet  2 3

Elastic opening w w  3 
 E   E 2

For: Volume pumped V = 1,000 bbl (~ 5,610 ft3)
Efficiency (@ EOJ) e = 0.5
Young’s modulus E = 1x106 psi
Poisson’s ratio  = 0.2
Net pressure (@ EOJ) pnet = 500 psi
Yields: Radius R ~ 103 ft
Width @ wellbore w ~ 1.51 in
 Influence Of Net Pressure

• Two radial fracture model solutions for the same treatment (no
barriers):
Predicted net Predicted fracture
pressure dimensions
R = 650 feet
Pnet = 50 psi
w = 0.25 in

R = 260 feet
Pnet = 800 psi
w = 1.6 in
 Fracture Geometry Changes With
Net Pressure
• Two modeling solutions for the same
treatment; if 500 psi stress contrast exists
around payzone
Predicted net Predicted frac
pressure dimensions

Pnet = 100 psi L = 1200 feet

Pnet = 800 psi R = 240

feet
 Net Pressure Analysis Untruths
• “You can get any answer you want”
– Not if you are constrained by real-data feedback, engineering
judgment, and the results of other fracture diagnostics !
• “You used the wrong frac model !”
Or

The analysis is credible because I used the ‘FracRocket’ model”

– Results usefulness determined 90% by engineer, 10% by model

• “We analyzed the treatment and determined optimum frac design”

– Optimization is an evolutionary process, completed over the
course of a series of fracture treatments
 Fracture Pressure Analysis
Problems / Opportunities
• Minimizing diagnostic injection time & cost without compromising
effectiveness
• Differentiating between “engineering” and “science”
• Unclear fracture closure pressure
• Practical bottom hole pressure measurement
• Surface pressure rate stepdown complications
– Pipe friction vs. perforation friction
– Identifying marginally unfavorable entry friction
• Appropriate Mechanisms for Net Pressure History Matching
– ? Modulus, stress, leakoff, and multiple fractures
– ? Layer interface mechanisms
 Fracture Analysis - Conclusions
• Benefits of real-data fracture treatment analysis can be
enormous
– Reducing screen-out problems
– Improving production economics
– Achieving appropriate fracture conductivity
• Measurement of real-data is relatively simple and
cheap
• The right analysis assumptions and a consistent
approach can get you “on the right page”, but
geometry require calibration with direct measurements
Production Analysis of HF Wells

Simple Approach:
• Evaluate performance based on EUR’s or other indicators such
as IP’s, 6-month and 12-month cumulative, best 3-month of
production etc.

• Cumulative Frequency plots can be useful as a simple statistical

method to compare and evaluate well performance
 ReservoirPT

• Finite-Difference
• Numerical Solution to Diffusivity Equation
• Reservoir As Grid System
• Single Well Within Rectangular Grid System
• Single Flowing Phase
• 2-D
• Unfractured and Hydraulically Fractured Wells
• Fracture Input From FracproPT
• Proppant Crushing
• Non-Darcy and Multi-Phase Flow Effects in Fracture
• Fracture Face Clean-up
 Log-Log Rate versus Time Plot
Transient & Boundary Influenced
Flow High Conductivity Fracture
1000

Transient Flow

100

2300 ac
Oil Rate (bbl/day)

360 ac
Boundary Influenced Flow
10

200 ac

100 ac

1
10 100 1000 10000
Time (days)
 Semi-Log Rate versus Time Plot
Transient & Boundary Influenced
Flow High Conductivity Fracture
1000

100

2300 ac
Oil Rate (bpd)

360 ac
10

200 ac
100 ac

1
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Time (days)
 Log-Log Rate versus Time Plot
Transient & Boundary Influenced Flow
High & Low Conductivity Fracture & Un-fractured Case

1000

High Conductivity Fracture Beginning of Boundary

Influenced Flow
100
Oil Rate (bbl/day)

Low Conductivity Fracture

No Fracture
10

360 acres

1
10 100 1000 10000
Time (days)
 Semi-Log Rate versus Time Plot
Transient & Boundary Influenced Flow
High & Low Conductivity Fracture & Un-fractured Case
1000

High Conductivity Fracture

100
Rate (bbl/day)

10

Low Conductivity Fracture No Fracture

1
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Time (days)
360 acres
 Important Parameter Is Relative Fracture
Conductivity At Reservoir Conditions
• Fracture Conductivity, wkf
wkf = fracture width x fracture permeability

• Propped Fracture Width is Primarily a Function of

Proppant Concentration
Dimensionless Fracture Conductivity (F CD) Is Used To
Design Fracture Treatments
wkf wkf
FCD = or Cr =
kLf kL
wkf = Fracture Conductivity, md-ft

k = Formation Permeability, md

Lf = Fracture Half-Length, ft

For FCD > 30 or Cr > 10, Lf is infinite conductive

- No Significant Pressure Drop in Fracture
- Value of 1.6 or larger generally sufficient
 (Patts(1961) and Cinco-Ley(1978)) Effective Wellbore
Radius Vs. Dim. Fracture Cond.

1.000
Rw'/Xf

0.100

At Fcd = 10; Rw’ = 43% of Xf

At Fcd =1.0; Rw’ = 19% of Xf

0.010
0.100 1.000 10.000 100.000
Fcd
Need Length Or Conductivity? (After
McGuire&Sikora)

Frac design
change with
same amount of
Productivity increase

proppant
Increase in
frac length

Increase in
conductivity
Design In Low-permeability Formation

• Need long fractures

• Dimensionless conductivity “easily” greater than 10

– Fracture conductivity generally not an issue
– “Self propping” (water) fractures may already provide sufficient
conductivity

• Treatment design
– Moderate pad size (avoid long closure times on proppant)
– Relatively low maximum proppant concentrations
– Poor quality proppant can be OK (if closure stress is relatively
low)
– Pump rate not very critical
Design In High-permeability Formation

• Sufficient fracture conductivity is critical

• Treatment design
– Minimum pad size to create TSO (Tip Screen-Out) based on
crosslink gel minifrac
– Use best possible (and economic) proppant for expected
closure stress
– Larger diameter proppant provides more conductivity and
reduces proppant flowback problems
– Use high maximum proppant concentrations
– Use of large casing frac string makes achieving TSO difficult
for small treatments
– Pump rates generally high, but can be decreased to initiate
TSO
 Optimum Conductivity
• FCD = 10 results in virtually infinite conductivity fracture
• In permeable reservoirs or in deep formations where closure
stress is high, it may be difficult to achieve FCD = 10; FCD of 1.6 is
generally sufficient
• Use reservoir simulation to determine optimum L assuming
you can achieve adequate FCD
• Choose proppant type and concentration to maximize FCD , up
to a value of 10
• Consider Multiphase flow effects
• Consider Turbulent flow effects
 Fracture Conductivity
In The Reservoir
• Conductivity is reduced by
– Closure Stress
– Embedment
– Crushing (generates fines and damages proppant)
– Corrosion
– Gel Residue Plugging
– Convection
– Proppant Settling
– Multiphase flow effects
– Turbulent flow
 Optimization Of
Fracture Treatments
• Function of:
– Permeability
– Oil & Gas in Place
– Drainage Area
– Fracture Conductivity and Ability to Place Proppant
• Economic Criteria Are Optimized
– Maximum Increase at Minimal Cost
– Multiple Economic Yardsticks to Choose From
 Economic Indicators
• Net Present Value (NPV)
• Rate of Return (ROR)
• Net Present Value to Investment Ratio (NPV/IR)
• Other
 Optimization Methodology
Step-by-step

1) Predict Well Performance

– Unfractured (Base Case)
– Different Fracture Half-Lengths
– Different Fracture Conductivities
– Different Drainage Areas
– Worst Case Proppant Placement Scenarios
 Optimization Methodology
Step-by-step

2) Estimate Treatment Costs Required to Create Half-

Lengths Assumed in Step 1
3) Calculate NPV, ROR, and/or Other Economic Indicators
Using Incremental Production (Difference Between
Fractured and Unfractured Cases)
 Optimization Methodology

TREATMENT COST
L f = 500 Optimal
CUM. GAS

L f = 300

NPV
L f = 100

Unstimulated

TIME FRACTURE FRACTURE

HALF-LENGTH HALF-LENGTH

1 2 3
 Fracture Diagnostic Tools
Will Determine ABILITY TO ESTIMATE
May Determine
Can Not Determine

GROUP DIAGNOSTIC MAIN LIMITATIONS

Net Pressure Analysis Modeling assumptions from reservoir description

Well Testing Need accurate permeability and pressure

Production Analysis Need accurate permeability and pressure

Radioactive Tracers Depth of investigation 1'-2'

Temperature Logging Thermal conductivity of rock layers skews results

HIT Sensitive to i.d. changes in tubulars

Production Logging Only determines which zones contribute to production

Borehole Image Logging Run only in open hole– information at wellbore only

Downhole Video Mostly cased hole– info about which perfs contribute

Caliper Logging Open hole, results depend on borehole quality

Surface Tilt Mapping Resolution decreases with depth

DH Offset Tilt Mapping Resolution decreases with offset well distance

Microseismic Mapping May not work in all formations

Treatment Well Tiltmeters Frac length must be calculated from height and width
Example Application - Model Results Are Not Always
Consistent with Directly Measured Geometry
GR log
1600
Initial fracture modeling
1700 (no confinement
mechanism)

1800
Depth (ft)

Calibrated fracture
modeling (composite
layering effect)
1900

2000

2100 Measured geometry

from downhole
tiltmeter mapping
2200
-400 -200 0 200 400
Along Fracture Length (ft)
 Fracture
Complexity Due
To Joints

HYDRAULIC FRACTURE
MINEBACK
 Fracture Height Confinement Mechanisms

Increased fracture Interface C om posite

closure stress slippage layering
FracproPT Model Calibration Parameters
• Crack Opening Coefficient (Shift-F3)
– 0.85 represents “coupled” behavior along frac walls
– < 0.7 represents “shear decoupled” behavior along frac
walls
• Tip Effects Coefficient Coefficient (Shift-F3)
– 1e-04 represents model with tip effects
– 0.4 represents linear elastic fracture mechanics
• Composite Layering Effect (Mechanical Rock
Properties)
– 1 represents radial growth
– >1 represents confined height growth
FracproPT Calibrated Model Limitations
• Sometimes actual closure stress is not well know
• Quite often, the closure stress profile is not well
known at all
– Make assumptions about continuity in bounding layers
stresses
• Need a substantial number of measurements
pointing in the same direction
• We do not really understand when composite
layering effect applies and how to assign it
• Consistent strategy to create match, as you can
match net pressure and dimensions in more than
one way
 Model Calibration Discussion
 Models today are more sophisticated than 20 years ago, but
often still do NOT accurately predict fracture growth
 Poor characterization of rock/reservoir/geology
 Incomplete understanding of relevant physics
 Model “calibration”
 Empirical, by matching geometries,
 Hopefully leading to improved physics in models
 Ultimate goal:
 Fully integrated fracture, reservoir and production models
 Integrated with real-time direct fracture diagnostics
 New Engineering Approach: Modeling
AND Measuring

Fracture growth models Direct diagnostics

incomplete physical not predictive
understanding

Calibrated models more realistically

predict how fractures will physically
grow for alternative designs
 Basic Fracture Pressure Analysis Steps
Enter inputs and
define assumptions
for treatment design /
optimization
Find closure stress
and efficiency from Calibrate model with
decline analysis direct diagnostics

Characterize
friction from rate Match net pressure
S/D tests for propped frac

Determine Interpret model

observed net results and make
pressure engineering decisions
Match observed net
pressure with model
net pressure
Match geometry
In orange: during/following diagnostic injections
 Conclusions
 Direct diagnostic observations on hundreds of hydraulic fracture treatments have
revealed the surprising complexity and variability of hydraulic fracturing
 Model calibration proving both heartening and humbling,
but to date perhaps more humbling than heartening
 Enhanced fracture height confinement most likely due
to layer interface effects
 Physics of fracture growth along/through layer interfaces
not well understood
 Not captured well in current models
 Identifying and understanding fracture complexities leads to
 Understanding well performance
 Enhancing completion/stimulation strategies
 Fracture models are essential tools for the engineering of hydraulic fracture
treatments, but we must become more humble
 By defining main limitations, we can continue to move models forward
FracproPT Version 10.2 – What’s Changed
Released July 2003 - Highlights
 Improved Minifrac Analysis
 Mayerhofer Method for permeability estimate
 Automated Friction Analysis
 Multiple Step Down Tests
 Semi-automated picking of rate steps
 Production Analysis Improvements
 Directly reads Excel or ASCII production data
 Automated production history matching
 New Fracture Design & Economic Optimization module
 Reservoir layers auto-picking from log data (LAS File)
 Improved report exports tables and graphs directly to Word
 User-defined graphical output tool
 Integrated Fracture Picture
 FracproPT Version 10.3 – What’s Planned
Highlights
 New calibrated fracture models and new default model
 Minifrac Analysis improvements:
 DFIT analysis plots
 Semi-automated closure picking algorithms
 Steprate test analysis
 Waste/water Injection module
 Log-Layer Editor improvements:
 Reservoir layer properties from triple/quad–combo log analysis
 Unlimited number of layers
 Visualize direct fracture diagnostic data
 Production Analysis improvements
 Quick Comparison
 Output interface for Eclipse
 Improved XY plots with permanent legend and multiple axis
 New bar graphs for real-time stage information
 Program navigation bar that remains on left of screen
 FracproPT Version 11.0 – What’s Planned
Highlights
 Improvements in navigation
 “Kick start” menus for quick runs in all modes
 Forward / Back button on all screens that are part of input "loop"
 Net pressure matching wizard with guidelines for matching entire
job
 Improvements in Report:
 User-defined Excel report
 Output to PowerPoint
 Full flexibility in positioning of graphs and tables in Word report
 User-defined report templates
 Quick Comparison for all modes
 Full 3D fracture growth model