You are on page 1of 45

Unconventional Reservoir and

Horizontal Well Stimulation


R. D. Barree
Barree & Associates LLC
How Do We
Optimize Stimulation Designs?
• Experience
– Our personal observations
– Our past history
– Our expectations of what will be successful
• Diagnostics
– Pump-in tests
– Fracture mapping
– Tracers (chemical and RA)
• Reservoir characterization
– Log analysis
– Core studies
• Models
– Fracture geometry
• LEFM, planar, 2D, 3D, coupled, shear decoupled?
– Reservoir production
• single phase, multiphase, transient, infinite acting, isotropic, Darcy?

Are these transferable from conventional to unconventional?


© 2009
What Are Unconventional Reservoirs?
• Tight and ultra-tight gas sands
– Perm less than what? 0.01 md?
– Composed of
• Fluvial channels and lenticular sands
– Anisotropic and discontinuous bodies
• Gas shales We will focus on these
– What is a shale, anyway? with reference to CBM
• Typically less than 20-30% clay for comparison
• May be quartzitic or carbonaceous
• Fine grained, small pore-size, high TOC?
– What is the production mechanism?
• Darcy’s law for free-gas flow?
• Diffusion from kerogen?
– How do we estimate gas content and recovery
• Coalbed Methane
– Wet or dry coals?
– What is the permeability of the cleat system?
– What controls gas content and recovery?
• Desorption isotherm?
• De-watering or rate of pressure decline?
© 2009
Puzzles for Unconventional Reservoir
Development
• Pay identification and Gas In-Place
– Porosity, saturation, TOC, desorption isotherm, net pay, drainage
area, aspect ratio
• Permeability
– Porosity and Sw relations, stress dependence, effects of fractures,
capillary effects (perm jail?)
• Stimulation design
– frac geometry, well azimuth, zone isolation
• Production forecasting and EUR
• Applying pressure diagnostics
• Managing stimulation damage
• Perforating techniques, phasing, orientation
© 2009
Gas Storage v. Frac Placement and
Production: Where do I put the well?
• Logs (DTC step-out) and core help show gas and TOC
content
• Areas of high gas content may not be productive
– Gas is there because it is trapped
– Gas may be at high pressure, increasing frac gradient
– High stress and no “natural” fractures means no frac
placement or production
• “Brittle” rock may be lean but easy to frac
• How do we identify best frac placement to contact
and produce stored gas?

© 2009
Predicting Pay Quality, Production,
and Long Term Performance
• Log-derived KH rarely works in unconventional reservoirs
– Porosity does not correlate well to permeability
– Pore size and size distribution are needed
– Saturation (or resistivity) may be misleading
– Distribution of phases, wettability, clay morphology (not volume)
may be important
– Small, sparse natural fractures contribute to effective permeability in
tight rock
• Drainage area and EUR can be difficult to predict from short-
term flow data
– Wells can remain in transient flow for years
– Decline curve analysis is not reliable
– Flow barriers can be formed by lithology, saturation, and capillary
effects
© 2009
Comments on Barnett Shale
Experience (Arthur Berman)
• Most reserve predictions based on hyperbolic production decline
methods were too optimistic when compared with production
performance.
• There is little correlation between initial production rates (IP) and
ultimately recoverable reserves (EUR).
• Average well life is much shorter than predicted, and the volume of
the commercially recoverable resource has been greatly over-
estimated
• Core areas of the play do not have appreciably higher average EURs
than the play overall, and the EUR from horizontal wells is not
significantly greater than from vertical wells.
• Finally, average well performance has decreased consistently since
2003 for horizontal wells.
© 2009
Comments on Barnett Shale
Experience (Arthur Berman)

“I am disturbed that public companies and


investment analysts make fantastic claims about
the rates and reserves for new shale plays
without calibrating them to the only play that has
significant production history.
Almost every assumption used by the industry to
support predictions about the Haynesville or
Marcellus Shale plays is questionable based on
well performance in the Barnett Shale.”

© 2009
Gas Recovery Mechanisms in Coal

© 2009
Porosity of Coals Falls into Two Groups
• Macropores (> 500 Angstroms): Flow capacity but minimal storage
– Space within cleats and natural fractures
– primarily determines the storage capacity for water
– considered to vary between 1-5% of bulk volume
• Micropores (8 to 20 Angstroms)
– Capillaries and cavities of molecular dimensions in the coal matrix
– Essential for gas storage in the adsorbed state
– Gas is adsorbed on particle surface
– 98% of methane is typically adsorbed in the micropores
– Storage capacity is equivalent to
• 20% porosity sandstone of 100% gas saturation @ the same depth
• 1 lb of sample Fruitland coal has an estimated internal surface area of
325,000 to 1,000,000sqft
– A very large volume of methane can be stored in the micropores of coal
despite low porosity in the cleat system.
© 2009
Typical Desorption Isotherm for Coal

© 2009
Coal Permeability Changes with
Pore Pressure Due to Shrinkage
30

Matrix shrinkage: 25

Coal matrix shrinks as


gases desorb 20

Adsorbed Gas and Perm


Cleat apertures enlarge and
15
permeability is increased Perm(md)
Vol_Ads(ml/gm)
At low internal pressure, 10
cleats collapse and
permeability reduces 5

0
0 200 400 600 800 1000 1200
Gas Pressure in Cleats,psi

Permeability in shales in also stress dependent, but shale matrix does not shrink

© 2009
Shale-Gas Production
Follows a Similar Model

© 2009
Shale-Gas Content: Free and Adsorbed Gas

© 2009
Flow Through Smooth Surfaced
Fracture, The Cubic Law Equation


6 H∆Pb 
3
Q = 5.11x10  
 Lµ 
Q = Flow rate (bbl/day)
H = Height of fracture (ft)
∆P = Pressure differential (psi)
b = Fracture aperture (in)
L = Length of fracture (ft)
µ = Fluid viscosity (cp)

© 2009
Effect of Fractures on System
Permeability for Parallel Flow
100000

Assumed Fracture Aperture = 0.001”

10000
Folds of Increase over Matrix Perm

Matrix Perm
1000
K=0.05
K=0.01
K=0.001
K=0.0001
100

10

1
0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 90.00 100.00
Fracture Spacing, ft
© 2009
Core Permeability
Changes with Stress Cycling
25

20

First Loading
First Unloading
Permeability (µd)

Second Loading
15 Second Unloading

10

0
0 1000 2000 3000 4000 5000
Net Stress (psi)

Example of Stress Cycling or "Seasoning" (from Morrow, et al.5) © 2009


Unpropped Fracture
Conductivity Loss with Stress
Fracture Conductivity & Pressure Response for 6000' Shale Sample
100 5000

Rate and Stress Induced


10 4000
Fines Migration

Confining Stress, psi


Conductivity, md-ft

Hexane
1 3000

0.1 2000

7% KCl
0.01 1000

0.001 0
0 500 1000 1500 2000
Time, min

Conductivity, md ft Net Confining Pressure


© 2009
Fracture Face Skin Damage
Reservoir Drainage Area

Damaged Formation

Fracture

Pratts Analog Model


Holditch Numerical Model

Both allow 100 fold reduction in perm with no loss of stimulation benefit
© 2009
Matrix Fluid Invasion for Cvc Control
(no filtercake)
100

10
Leakoff Invasion Distance, inches

K=0.1 mD
K=0.01 mD
0.1
K=0.001 mD
K=1e-9 D

0.01

0.001

0.0001
0 500 1000 1500 2000 2500
Minutes of Leakoff © 2009
Capillary Face Discontinuities or
“Capillary End-Effect”
A buildup of water saturation at any capillary discontinuity or interface.

Face of natural fracture

Face of created fracture

Proppant pack to open perf or wellbore © 2009


Capillary Phase Trapping: A Potentially
Serious Damage Mechanism
Capillary blocking causes loss of fracture face area
Reservoir Pressure

No-Flow No-Flow

Low Pc Low Pc
Areas of low porosity, high capillary pressure
Fracture or wellbore pressure under drawdown
© 2009
Capillary Effects in Circular Tubes
σCosθ At equilibrium Force Up =
σ Force Down
r
ForceUp = 2π rσCosθ
ForceDown = π r 2 ∆ρgh
h θ
2π rσCosθ 2σCosθ
Pc = = and
πr 2
r
π r 2 ∆ρh g g c g
Pc = = ∆ρh
πr 2
gc

Capillary Rise of
a Wetting Phase © 2009
Capillary Pressure and Frontal
Advance Rate

2σCosθ r ∆P
2

Pc = V =
r 4 µ∆L
© 2009
Capillary Controlled Displacement:
Drainage

© 2009
Capillary Controlled Displacement:
Drainage

© 2009
Capillary Controlled Displacement:
Drainage

© 2009
Capillary Controlled Displacement:
Drainage

© 2009
Capillary Controlled Displacement:
Drainage
Pressure

Pg-Pw=Pc

© 2009
Distance
20/40 Sandpack De-Watering
with N2 Gas
Flow Direction
Elapsed
Time

39:15

42:50

45:20

50:05 Hours into the test switch flow direction


50:00

50:10

55:35
© 2009
20/40 Sandpack De-Watering
Flow Direction
Elapsed
Time

58:10

60:45

63:00

66:35

73:50

6/15/07

83:25
© 2009
What causes capillary saturation
blockage after breakthrough?

Pc
E
Pc2
Pressure

Pg-Pw=Pc Pc
1

Distance 1 .
0
Outside the reservoir the capillary pressure Sw
approaches zero as radius of curvature of
the gas bubble or oil droplet increases. © 2009
Drainage and Imbibition Residual
Saturations at Capillary Equilibrium
1.0

krg

Capillary Pressure
Relative Perm

krw

0
0 Water Saturation, Sw 1.0
Swc Sgt © 2009
Typical Tight Sand and Shale Pore Sizes

© 2009
Capillary Threshold Pressure for
Various Fluid Systems
100000

10000

1000
Capillary Threshold Pressure, psi

100

PC, gas-wtr
PC, gas-oil
10
PC, surf
PC, ca

0.1

0.01

0.001
0.001 0.01 0.1 1 10 100 1000
Pore Radius, microns
© 2009
What Kind of Frac Do We Need?
• Single planar frac
– Limited surface area
– All flow passes through a small channel
– Requires high conductivity
– Production dependent on frac conductivity and
reservoir perm
• Complex (shear) fracture network
– Large surface area exposed
– May not be able to connect and drain entire network
(load recovery experience)
– Fracture conductivity may be less important (except
near the well, but must be measurable)
– Production dependent on effective surface area
(diffusion dominated)
© 2009
Fracture Complexity in Shales
1500
All Microseisms
Simple Fracture Complex Fracture
1000 ~2700 ft

500

Frac

Northing (ft)
0 Well

-500 ~1200 ft

Possible
Complex Fracture Complex Fracture -1000
With Fissure Opening Network Aligned
Features
-1500

Observation Well
-2000
-1000 -500 0 500 1000 1500
Easting (ft)
© 2009
What is Stimulated Reservoir Volume
(SRV)?
• SPE 119890: M.J. Mayerhofer, E.P. Lolon, N.R. Warpinski, C.L. Cipolla, D.
Walser, Pinnacle and C.M. Rightmire, Forrest A. Garb and Associates
– Stimulated reservoir volume is the bulk volume of rock affected by
microseismic activity
– The model assumes a dual-porosity Warren-Root type model with a lattice of
fractures surrounding matrix blocks
– All fractures are assumed to be conductive and interconnected to the
wellbore
– All exposed block surface feed to the fracture network
• Barree and Gilbert approach (this course)
– Enhanced permeability region around primary induced fractures caused by
secondary shear and tensile failures
– Contributing volume is not directly related to microseismic activity
– Not all fractures remain open or conductive
– System behaves as a limited volume with effectively increased single-porosity
flow capacity
© 2009
Basis of SRV (SPE 119890)
• MS mapping shows large fracture networks in shales
• Early work (Fisher et al. 2004) hypothesized a correlation
between network size and production
– Large treatment volume generates larger network
– Larger network yields larger stimulated reservoir volume
– Larger SRV leads to higher production
• In shales, MS event cloud width approximates the network
size
– SRV is proportional to the created gross network size
– This is Not equal to the effectively producing SRV
– no information on frac density (spacing) provided by MS data
• Potential approaches for effective SRV:
– Geophysical (seismic moment=strength; Maxwell et al. 2006)
– Reservoir modeling (Mayerhofer et al. 2006)

© 2009
Estimating SRV from MS Events
(SPE 119890)
400

500

600

700

800

900

000

100

200

300

400

500

600 Events outside of shale section

700

800
0

0
Map View
Side View © 2009
Reservoir Modeling:
MS-SRV & Frac Spacing
• Each matrix block has a
106 ft3
SRV=2,000 x
small storage volume and
expansive energy

Xf
component
• As spacing decreases, the
Xn After 1 Year Production rate at which the matrix
block drains must
increase
• In Warren-Root models
the fractures blow down
very fast, depending on
the relative volume of
fractures and matrix
© 2009
Total Exposed Formation Length from
SRV and Fracture Spacing
1000000
Total Fracture Face Length, ft

100000 10
30
50
100
150
200
250
10000 300

4 x f xn
L ftotal = + 2 x f + xn
∆xs
1000
0 100 200 300 400 500 600 700
Effective Frac Half-Length, ft
© 2009
Enhanced Permeability Model

Area of Increased Permeability


from Stimulation (small and variable)

Area of Original Permeability


In Unstimulated Rock (very large)

© 2009
Enhanced Permeability Model
• The enhanced region has been disturbed by generation of
secondary tensile and shear fractures
• Average permeability may be increased by 10-1000 fold
• Volume of enhanced region remains constant and may be
small (10-100 acres)
• Storage volume of original perm may be large (200-1000
acres) but feeds in slowly
• Both regions are assumed to produce through the wellbore
via the primary fracture and lateral well and contribute
continuously
• The enhanced perm region, if small, will deplete rapidly and
the storage volume will support long-term flow

© 2009
How do Real Wells Respond?
• One of the primary goals of this discussion is to
analyze real field data to see how actual wells
respond
• How can the SRV or enhanced perm region be
quantified
– Permeability ratio
– Area ratio
• How can lateral length and stimulation design be
optimized for maximum economic value?
– Lateral length effects
– Lateral azimuth
– Spacing of frac stages
– Size of each frac stage and required conductivity
© 2009

You might also like