Fundamentals of Hydraulic Fracturing.pdf

Engineering

1. LIVE COURSE: COMING SOON! SF

2. Motivation for Frac Engineering & Diagnostics

ale aot ma no] Hydraulic Fracturing
‘ x is done for well
stimulation
NOT
For proppant
disposal

SkillUP

3. Frac Models : Which Model & Why ?

Q Classic 2D

Q Pseudo - 3D “parametric”
Q Pseudo - 3D “cell”

Q Planar 3D

Model Provides decision making capability

Y” Understand what happened

Y” Isolate causes of problems

Y” Change necessary inputs

Y Ability to predict (not just mimic) job results
If your model can’t do this, why run it?

Software Model type Company Owner
PROP Classic 2D Halliburton

Chevron 2D Classic2D Chevron Texaco

CONOCO2D Classic2D conoco

Shell 20 Classic 2D Shell

FRACPRO Pseudo-3D "parametric” RES, Inc. on
FRACPROPT Pseudo-3D"parametric' Pinnacle Technologies GTI
MFRAC-III Pseudo-3D "parametric" Meyer & Associates Bruce Meyer
Fracanal Pseudo-3D "parametric" Simtech A.Settari
STIMPLAN Pseudo-3D "cell" NSI Technologies M. Smith
ENERFRAC Pseudo-3D "cell" Shell

TRIFRAC Pseudo-3D "cell" S.A. Holditch & Association

FracCADE Pseudo-3D "cell" Schlumberger EAD sugar-land
TerraFrac Planar 3D Terra Tek ARCO
HYRAC 3D Planar 3D Lehigh U S.H. Advani
GOHFER Planar 3D Halliburton R. Barree

SkillUP

4. Decision Tree for Carbonate Formations Stimulation

Job Execution

Follow Up

Best Candidates to
Acid Fracturing:

wre

. > 85% Soluble

. Heterogeneous

. Closure stress < 8k psi
4

Permeability:

v Oil K< 10 mD.
Ÿ Gas K< 1mD.

Acid Etching Test
3" x 3° core

SkillUP

5. Frac Candidates

Q Excellent candidates
Y” Damaged wells
Y” Low permeability reservoirs with sufficient oil or gas in place

Q Good candidates
Y” Naturally fractured reservoirs
Y” Unconsolidated, high permeability reservoirs, that have been damaged

Q Poor candidates

v Reservoirs with limited reserves

Y” Thin reservoirs with poor barriers

Y” Low pressure reservoirs where fracture fluid cleanup is difficult
Y” Reservoirs where stimulation can penetrate water zones

SkillUP

6. Q Frac Tanks O ler |
Q ExpressSand Delivery System

Q Proppant Storages

Wireline crew |

Fracpumps |

Q Blender Chemicals)
Q Manifold ‘Sand! Proppant |
Q Data Van =

Q High Pressure Pumps Blender ]

Q Wellhead Isolation Tool

Warehouse |

ice |

O %:

ExpressSand
Delivery
System

7. Complete Closure Pressure Equation

(157

Q Pc = closure pressure, psi

Q v= Poisson’s Ratio

Q Pob = Overburden Pressure

Q av = vertical Biot's poroelastic constant

Q ah= horizontal Biot's poroelastic constant
Q Pp = Pore Pressure

Q ex = regional horizontal strain, microstrains
Q E= Young's Modulus, million psi

Q ot = regional horizontal tectonic stress

[Poo = a, Py|+ a,Pr+&,E+o,

Y” Actual in-situ stresses can only be determined by direct measurement.

SkillUP

8. Handling Tectonic Stress p=

[P»- @,P,]+ @,Pp+é,E +0,

i v)

% A constant regional stress can be added to one (or both) horizontal stresses over some vertical extent

+ Assume some regional strain which then generates a different stress in each layer, according to its stiffness
Y” Allows component of stress proportional to Young’s Modulus
Y” Shown to work effectively in many field cases
Uniaxial Strain Adjusted Stress

a. =| Added 200 micro-

Q Two ways

E strains regional strain
m H to stress calcs to match

observed closure stress
of 4500 psi at 6050’

Depth (ft)
3
ja
[

‘Closure Stress (pa) SkillUP

9. Measurement of Dynamic and Static Elastic Properties

Dynamic modulus must be converted to static AAA
Ja Power-Law Model
- . J 2° [emotes Ex Lop-unesr
Y Static Modulus: large amplitude at low (zero) E ato A
frequency (load frame tests) H 10.000 =
; A i 8.000
Y” Dynamic Modulus: small amplitude at high $ a
frequency (acoustic waves) A 6.000
ö
Y 4.000
À 200
Log(E,) = Log(p,E,)-0.55

0 2 4 6 8 10 2
Computed Static Young's Modulus, MMpsi

SkillUP

10. Logging for Mechanical Properties

Q Acoustic logs
Y” Measure compressional and shear velocity or “slowness” (1/v)
Y” Can be affected by fractures, pore fluids, and borehole conditions

Q Density logs
Y” Measure neutron capture cross-section, interpreted as bulk density
Y” Strongly affected by borehole conditions and breakouts (pad device)

Q Gamma Ray
Y” Records spontaneous GR emissions from multiple sources
Y” Spectral GR logs can differentiate energy levels from different sources (U, Th, K)

Q Resistivity
Y” Measures electrical resistance along an assumed path length
Y Affected by clays minerals, clay morphology, and pore fluids

Q None of these measure pressure or stress
Q None actually measures rock elastic or mechanical properties SkillUP

11. DFIT

A DFIT is a Diagnostic Fracture Injection Test.

Q ADFIT is NOT:
= A diagnostic formation injection test
= A reservoir transient falloff test or “Mini-Falloff” (MFO)
= A fluid efficiency test (FET)
= A micro-frac
“A mini-frac
= A “data-frac”
= A pressure rebound test
= A reservoir limits test
= A pump-in flowback test (but the analysis techniques may apply)

Q The purpose is to determine properties affecting fracture initiation and extension, treating pressures, leakoff,
screen-out risk, calibration of the in-situ earth stress tensor, and secondarily post-frac production

“ IF you don’t do this, and do it right, don’t even think of using a numerical simulator for frac design.” Nolte

SkillUP

12. Events Observed During DFIT Procedure

Pressure

je Wellhead

— Breakdown

Pete

Pressure

Pox > Piso > Pe > Pa >Pr > Pr

Pur — Breakdown Pressure
Pue — Instant Shutin Pressure

Pe— Closure Pressure

Formation Linear Flow Pressure
Pr — Pseudo Radial Flow Pressure
Pr — Reservoir Pressure

<—ISIP {instant Shin Pressure)

Closure Pressure (a, -Mnmumsvess)

+ Breakdown indicates initiation of new fracture

= Roll-over indicates dilation of existing fracture(s)

+ Injection pressure should be stable at constant rate

= Step-down at end of injection for perf and tortuosity
+ ISIP (not instant) represents fracture extension
pressure

= Fissure opening may be observed, or not, between
extension and closure pressures

= Reservoir transients may be observed after closure if
the test is run long enough

= Pore pressure is always extrapolated to infinite shut-
in time

SkillUP

13. DFIT Design Constraints:

Q Plan for enough HP to reach about 10 bpm (1.5 m3/m) at treating pressure
Q Time to reach closure is approximately Pump Time / 3*Estimated Perm (md)
v Five minutes in 0.01 md rock = 150 min (2.5 hours)

Q Time to establish analyzable reservoir transient is roughly 3 times the closure time

Time to Closure vs. System Perm Time to reservoir transient vs. time to closure
eg Log Anais

Semmes

some)

SkillUP

14. How to PICK ISIP with High Tortuosity?

Measure
compressive
wellbore storage
during initial
injection.

Wellbore Blowdown
Analysis: Compute wellbore
discharge or blowdown rate,
assuming constant tortuosity
(psi/vbpm) and variable DP

SkillUP

15. Typical Derivative Shapes in G-function Analysis:

1. Enhanced or accelerated leakoff
2 2. Normal “matrix dominated” leakoff
3. Delayed leakoff or variable storage

vt je je
Ws

Classical Nolte Analysis:

Q Find the “correct” straight line on the
pressure-G-time plot.

Q Deviation from the end of the straight line
indicates closure (1979).

Q Castillo suggested plotting 1* derivative to
define the “correct” straight line (1987).

Q Barree added semi-log derivative to reduce

ambiguity and define leakoff “type-curves”
(1996).

SkillUP

16. Biggest Current Misconception: Variable Compliance

= Nolte defined fracture compliance, for a PKN fracture, as H/E (L/E for KGD)
= Compliance represents the inverse of stiffness of the fracture
= Compliance, in his model, controls the rate of pressure decline
For a single planar fracture, with all his other limiting assumptions, the only way he had to change the rate of
pressure decline was to assume that “compliance” was changing
= With rock modulus, E, constant that leaves H as the only variable
= This led to the concept of fracture “height recession”
Y Initial slow pressure decline > high compliance
Y Transitions to faster decline > decreasing compliance
= The same concept of “variable compliance” can be |
applied to “length recession”
= Both theories must assume that the fracture
Y Closes from the tip back to the center (height or length) ?

Groton ans

y” Closure of the tip changes frac length or height and '
compliance '
= That means the rock grows back together Ei
» Fundamentally, this NEVER happens! ae =

SkillUP

17. Biggest Current Misconception: Variable Compliance

Usual Cause of the “Belly” or Delayed Leakoff
Y Recharge from Variable or Transverse Storage A
?

No recharge from storage: Hard shut-in With recharge from storage: Variable
with no return rate return rate
— =

r

SkillUP

18. Permeability Estimation from G at Closure (Gc)

Good estimate when after-closure radial-flow data not available or unreliable

Where:

k =effective perm, md

M = viscosity, cp

Pz = process zone stress or net pressure
PHI = porosity, fraction

Ct = total compressibility, 1/psi

E = Young’s Modulus, MMpsi

rp = leakoff height to gross frac height ratio

SkillUP

19. Perforation Phasing Chart

PERFORATING PHASING SELECTION FOR VERTICAL WELLS

0° phasing, perforation

gun should be attached to
lower casing wall and

‘oriented to shot through it

Legend:
Young's Em >t
No [Exploration] ves modulus, MMpsi:|-Low [<1
wel Horizontal stress |-Hign [> 01
Proppelant assisted perforation E Moderate [0.05 - 0 1
contrast psift [Low [<005

No _[Cane perofration gun

be centralized in the |_Yes
well?

wo HF, put on production
immediately after perforation
Aplicable for HP/HT weis.

No ‘Naturally Yes
Fractured
Low
so con
ss | Moderate Youngs] High
cate ARS „| 180 phasing
phasing Er +

Low | Youngs
poa Moduls 7

‘optional: onented ‘oriented wi os max,
wi oh high-energy large perf and

shots close together

SkillUP

20. a

Stress Cage Formation

O The shock of perforating causes a plastic deformation of the rock surrounding the
perforation tunnel. When the pressure pulse dissipates, a residual deformation is
left in the rock with an associated high residual compressive stress. The residual
stress acts similarly to the stress concentration effect around the borehole in that
the pressure required to initiate a fracture is increased significantly.

Q The effect of the "stress cage" formed around a jet shot perforation is indicated

by the photograph of an actual perforation in a horizontal well at the Nevada Test
Site (Sandia National Laboratories). The cased, cemented hole was perforated
with a 32 gram jet charge and subsequently fractured with a dyed fluid. The
fracture was then mined back to expose the fracture surface and the well casing.

Q The casing is visible on the left side of the figure. The light area surrounding the

perforation is unfractured rock. The surrounding dark area is the dyed face of the
created fracture. It is apparent that no fracturing fluid exited through the
perforation tunnel. All communication with the fracture is through a narrow
annular ring at the cement-formation interface.

21. Embedment Core Tests
0.35-0.77 mm @ 10,000 psi

300m 400m 700 pm.

YM. 055,

SE 196183

23. Frac Screenout:

Common False Assumptions:

Proppant is homogeneously distributed
Sand and fluid travel together

Pad is required to open width for sand
Pad is depleted by leakoff

Screenouts caused by prop bridging
Prop concentration increased by leakoff

oooooo0

“False assumptions lead to failed remedies.” Bob Barree "*] | F I

3:
8
8
8
3

8

E77

Common Remedies:

O Pump more pad volume
Q Increase pump rate

O Use higher viscosity fluids
Q Use smaller proppants

Q Use fluid-loss additives

“Sometimes they work, and sometimes NOT!” Bob Barree SkillÜP

24. Post-Job Analysis: Actual Frac Conductivity

Pack width determined by: omy

Q Proppant concentration en
Q Closure stress om
Q Filter-cake and embedment

Pack permeability determined by: Exel

Q Proppant size and strength a

O Packing and porosity

O Regained permeability and gel clean-up -

O Non-Darcy and multiphase flow TE = =

“We must know what was achieved to improve the design, or the design effort was wasted.” Samuel

SkillUP

25. Post-Job Analysis: Impact of each variable on results

Fracture Treatment Sensitivity: Production Input Sensitivity:

Q Width Exponent Q Gel Damage

Q Permeability Q Permeability - Increase/Decrease

Q Fluid Type Q Drainage Area - Increase/Decrease

Q PZS (Process Zone Stress) - Increase/Decrease Q Aspect Ratio - Increase/Decrease

Q PZS Vertical to Horizontal Anisotropy (V/H Factor) O X/Y Offset

O PHOLD (Proppant Holdup O Remove Condensate Yield

Q PHOLD Vertical to Horizontal Anisotropy (V/H Factor) (1 Remove Condensate Yield & Water Production

Frictional Sensitivity:

Q Tortuosity

Q Cd (Coefficient of Discharge
O CXSP (Sand Exponent)

Diagnostic Injection Falloff (DFIT) Sensitivity:
Q Perm
Q Secondary Leakoff Coefficients
Y PDL Coefficient (Pressure Dependent Leakoff)
Y TSC (Transverse Storage Coefficient)
Q Relative Permeability Factor/Ratio

SkillUP

26. SkillUP

[TRAINING & CONSULTATIONS)