SPE 119242 How to Use and Misuse Proppant Crush Tests – E i th

SPE 119242
How to Use and Misuse
Proppant Crush Tests –
E
Exposing
i
the
th Top
T
10 Myths
M th
John Kullman, CARBO Ceramics
T. T. Palisch, M. Chapman, R. Duenckel, and S.
Woolfolk CARBO Ceramics
Woolfolk,
Ceramics, Inc
Inc.
M. C. Vincent, Insight Consulting
Outline
• Introduction/Motivation
• Crush Test Procedure
• Myths
M ths
–Misuse/misapplication
pp
• Summary
The Big Picture
• Fracs must provide:
– Reservoir contact (length, height) to contact and
collect oil and gas
• Related to volume of proppant
– Flow capacity to carry oil and gas to the wellbore
• Related to proppant permeability and frac width –
described as conductivityy
• Other Important Proppant Characteristics
–
–
–
–
–
–
Durability
y
Temperature Resistance
Transportability
Fluid Compatibility
Flowback Control
Environmentally Benign
Question…
• Gas
G Well
W ll
– 7500 psi stress
– You can successfully place a 16/20 or 16/30
sized proppant
• Two choices of proppant
– Proppant A – 16/20 Ceramic, 14% Crush (7.5k)
– Proppant B – 16/30 Ceramic, 7.5% Crush (7.5k)
– Which proppant would you choose?
– What if I told y
you that they
y had the same MPD?
– What would you be willing to pay for your choice?
Introduction/Motivation
• API RP56 & 60 original – updated in ISO
13503-2 (2006)
– “improve quality…delivered proppants”
– “enable…to compare physical properties”
– Original
O i i l iintent
t t tto help
h l qualify
lif sand
d sources
• “Crush results” and proppant selection
– “qualified
“
lifi d engineering
i
i analysis….required
l i
i d ffor
their application to a specific situation”
– SPE 11634 – Conductivity comparisons cannot
be made on the basis of crush tests
Yet many still choose their proppants based on
**Yet
crush results **
ISO 13503-2 Crush Test Procedure
• P
Proppantt is
i pre-sieved
i
d tto
remove particles outside of
stated mesh range.
• Dry
D proppantt placed
l
d iin steel
t l
cell at ~4 lb/sq ft (sand
equivalent)
• Room temperature
• Proppant evenly distributed
with level surface
• Load
L d applied
li d att uniform
if
rate
t
• Constant stress maintained for
two minutes
• Proppant is sieved. The weight percent which falls
below the primary screen is reported.
– For 16/20 proppant all material < 20 mesh is reported as “fines”
fines
– For 30/50 proppant all material < 50 mesh is reported as “fines”
ISO 13503-2 Crush Test Procedure
Do these reflect realistic conditions?
• Proppant is pre-sieved.
• Proppant Loading – sand/RCS/LWC ~4 lb/ft2,
IDC 4.8 lb/ft2, Bauxite ~5.2 lb/ft2
• Smooth, steel plates – embedment?
• “Carefully
y loaded”
• Dry, room temperature
• 2000 psi/min,
psi/min relaxed after 2 minutes
• Only the particles smaller than bottom screen
are considered “fines”
fines or “crush”
crush
Are the results repeatable/reliable?
Crush Cell Loading critical
• “variance in crush results….associated with
method of loading…”
• Significant efforts ongoing on ISO Committee
and StimLab to alleviate variations in results
– Loading technique thought to be the cause
– Lab to lab, technician to technician, equipment to
equipment
Test#1
Test#2
24
Test#3
22
9.1
10
9.5
50
9.1
10
17.7
78
19.8
88
18.4
43
8.4
40
8.9
90
8.6
60
6.0
04
5.9
92
6.4
42
23.2
26
23.2
29
24.7
75
9.1
18
10.6
63
10.9
95
25.2
21
23.6
60
24.2
29
16.7
70
14.6
60
16.7
70
24.5
52
14.7
71
17.8
86
15.7
74
17.5
52
18.4
45
3
4
5
9.2
20
9.8
80
10.1
10
8
9
10
11
ISO Subcommittee Results
6
7
Lab Number
2
0
1
2
14.8% Avg
16
14
12
10
Perrcent Crush
h
Are the results repeatable?
16/30 Brown Sand Hand Loaded Weight Percent Crush at 4000psi
26
20
18
8
6
4
Are the results repeatable?
16/30 Brown Sand Mechanical Loaded Weight Percent Crush at
4000psi
18
Test#1
16
Test#2
Test#3
14
Perrcent Crush
h
12
10.0% Avg
10
8
6
4
2
5
6
7
Lab Number
10..10
10..80
10..60
10..53
9..07
9..85
4
7..50
7..50
9..23
10..32
10..89
10..79
3
8..40
8..20
8..40
12..00
12..00
12..80
2
10..49
10..22
9..94
8..96
16..66
11..41
1
7..80
8..40
7..76
9..79
9..40
9..26
0
No data
reported
8
9
10
11
ISO Subcommittee Results
Does Fracture Width Affect Crush?
• Interior grains loaded “evenly”
• Exterior grains
g
have fewer
load points
• Crush increases significantly
as proppant loading
decreases
• For a 20/40 proppant, there are approximately
24 layers of proppant in standard crush test.
test
– 8% are exterior grains
• 1 lb/ft2 is
i ~6
6 llayers off 20/40 proppantt
– 33% are exterior grains
Crush Depends
p
Upon
p Frac Width!
30
Percent Crush
25
20
15
10
5
0
4 lb/sq ft
2 lb/sq ft
1 lb/sq ft
0.5 lb/sq ft 0.25 lb/sq ft
Monolayer
~ 0.2 lb/sq ft
Crush vs # Layers
100%
Crush at 10,000 psi
White Sand
ELWC
RCS
B
Bauxite
it Ceramic
C
i
20/40 Proppants
90%
80%
70%
2
1 lb/ft
Sand &
RCS
% Crush
60%
50%
2
1 lb/ft
B
Bauxite
it
40%
2
1 lb/ft
ELWC
30%
20%
10%
0%
0
1
2
3
4
5
# of Layers
6
7
8
9
10
Crush vs # Layers
100%
White Sand
RCS
ELWC
Bauxite Ceramic
Crush at 1000 psi
90%
All 20/40 Proppants
80%
70%
% Crush
h
60%
50%
40%
30%
2
2
20%
2
1 lb/ft
Bauxite
1 lb/ft
ELWC
1 lb/ft
Sand &
RCS
10%
0%
0
1
Partial
Monolayer
2
3
# of Layers
4
5
6
7
RANGE OF FRACTURE COMPLEXITY
SPE 77441
Simple
p Fracture
Complex
p
Fracture
Very Complex Fracture Network
Complex fracs are
believed to provide less
cumulative conductivity
than simple,
simple wider
fractures
Vertical Complexity
D T
Due
To Joints
J i t
Physical evidence of
fractures nearly
always complex
NEVADA TEST SITE
HYDRAULIC FRACTURE
MINEBACK
16
Uniform Packing
Arrangement?
Pinch out, proppant
pillars,
ill
iirregular
l
distribution?
Is this ribbon laterallyy
extensive and
continuous for
hundreds or
thousands of feet?
17
80.00
y=
1488
2ll - 18
18.714
714 h
Single
Si
l 1488.2x
P
Pellet
t
C
Crush
2
70.00
NO!!
R = 0.7765
60.00
50.00
40.00
30.00
20 00
20.00
18
10.00
16
0.00
0.0000
14
Percent Crush
Pounds of F
P
Force to
cru
ush one pelllet
CPF
Are Large Particles weaker than Small?
0.0100
12
12/18
16/20
20/40
0.0200
0.0300
0.0400
0.0500
0.0600
Proppant Size inches Courtesy Stim-Lab
10
8
6
4
2
0
30/50 LWC
20/40 LWC
16/20 LWC
12/18 LWC
For all proppant types, larger grains have
greater individual strength
strength.
160
140
12/18
120
CPF
CarboLite
20/40
100
Hi k
Hickory
Interprop
80
CoSilica
Jordan
60
ResinPR
40
20
0
0
0.02
0.04
0.06
0.08
0.1
0.12
pp
Size,, inches
Proppant
Source: Stim-Lab Consortium, July 2001
1.8-16
Another Look at Single Grain Strengths…
160
140
12/18
120
CPF
CarboLite
20/40
100
Hi k
Hickory
Interprop
80
CoSilica
Jordan
60
ResinPR
40
20
0
0
0.02
0.04
0.06
0.08
0.1
pp
Size,, inches
Proppant
Source: Stim-Lab Consortium, July 2001
Note that application of
resin does not improve
p
grain strength, but rather
improves
distribution of
0.12
stress between grains
and encapsulates fines.
1.8-16
So why does crush increase
with
ith llarge proppants?
t ?
• Strength in numbers?
“There’s Strength in Numbers”
Smaller mesh sizes distribute the load to across
more particles compared to larger mesh sizes
Proppant Type
• Natural quartz crystals (sand), manufactured ceramics, and
resin-coated proppants crush differently
• Sand
– Quartz
Q t crystals
t l tend
t d to
t resultlt in
i a greater
t number
b off fine
fi shards
h d
• Ceramics
– Tend to cleave or p
part into relatively
y few,, larger
g p
pieces
• Resin Coated Products
– Resin does not significantly change single grain strength, but
improves distribution of stress
stress. If the particles can be
encapsulated, they will not be measured as “crush” regardless of
whether the substrate fails
SPE 11634 - conductivity comparisons cannot be made on the basis of
crush tests.
Do all Proppants Fail in the Same Manner?
When they fail
fail…
– Sands shatter like a glass
– Ceramics
C
i cleave
l
lik
like a b
brick
i k
– Resin Coated products
“deform”;
deform ; fines captured
Brown Sand
att 6k psi.
i
RCS at 8k psi.
IDC at
8k psi.
Do fines affect all proppants similarly?
Remember…
R
b
• All proppants do not fail in the same manner
– The fines generated by one proppant may look
drastically different than those generated by
another.
th
• The packing arrangement for similarly sized
proppants are not the same for all types of
proppants.
– i.e. the packing arrangement for a 20/40
ceramic, 20/40 RCS and 20/40 Sand will be
diff
different
t even att comparable
bl stresses.
t
Post Crush Sieve Distribution
Weig ht Perce
ent in Siize
Rang
ge
100
98
After crushing 20/40
75
EconoProp at 6000 psi
Standard API technique
50
25
1.18
0.43
-40/+50
-50/+70
0.2
0.13
0.08
0
-200/+325
Pan
0
-20/+40
-70/+100
-100/+200
Source: CARBO Analyses Nov 1998
A Closer Look at the Crushed Fraction
“2% fines” reported with standard
testing could mean 2 cleaved grains
per 100 (4 immobile pieces), or it could
represent 400 mobile fragments in the
100-mesh range
1.20
1.18
Percent C
P
Crush
1.00
Immobile cleaved grains
It makes a difference!
0.80
0.60
0.43
0.40
Potentially mobile in 20/40 pack
(SPE 24008)
0.2
0.13
0.20
0.08
0
0.00
-40/+50
-50/+70
-70/+100
-100/+200
Source: CARBO Analyses Nov 1998
-200/+325
Pan
Fluid Effects
• Crush testing is performed dry. What if the proppant
is saturated?
7
Percentt Crush
6
Modified Crush Test Results
E
EconoProp
P
att 6000 psii
5
4
3
2
1
0
Dry API
Moisten in cell Moisten in cell
with water
with min. oil
Moisten with
water, then
add to cell
Source: CARBO Tech Brochure 3/4/96
Moisten with
min. oil, then
add to cell
Is one set of Test Conditions superior to another?
Dry,
y, wet,, hot,, room temperature,
p
, water or oil…
is one method more realistic than another?
6k Crush @ 2#/ft2
35
Sand
30
ELWC
RCS
Cru
ush %
25
20
15
10
5
0
Standard
Loading
g
Load by
Load by Standard Standard Standard Wet with Wet with Standard Standard
hand and hand and then tap
then wet then wet water then mineral oil but heat to but heat to
p
rotate
do not
cell
with water
with
load into then load 200F dry 200F wet
piston
rotate
mineral oil
cell
into cell
piston
Can Crush results be Correlated to Conductivity?
More Realistic Conditions in a Conductivity Test
What’s the Difference?
•
•
•
•
•
•
Proppants
pp
evaluated as received
Tests equivalent mass loading, and 2 lb/ft2
Utilizes Sandstone shims
Flow water through pack
Elevated temperatures (150°
(150 or 250°
250 F)
Stress held for at least 50 hours
Disassembled API Proppant Cell
Ports for Measuring
Differential Pressure
Temperature Port
Proppant Bed
Sandstone Cores
Flow Through
Proppant Bed
Long Term Conductivity Cells
Can Crush results be Correlated to Conductivity?
The “crush” measured after a Conductivity test
significantly higher than Crush test.
6k Crush Results vs Crush after Conductivity Testing at 6k psi
45
40
Sand
35
ELWC
RCS
Crush %
C
30
25
20
15
10
5
0
Standard
Loading
Load by
Load by Standard Standard Standard Wet with Wet with Standard Standard
hand and hand and then tap
then wet then wet water then mineral oil but heat to but heat to
rotate
do not
cell
with water
with
load into then load 200F dry 200F wet
piston
rotate
mineral oil
cell
into cell
piston
All tests at 2 lb/ft2 loading
Embedment
More width retained,
but lower perm
Spalling
Spalling
Example of Conductivity Loss
CO N D U CT I V I T Y V S . C L O S U R E S T R E S S
100000
YME=5E6psi
YME=1E6psi
YME=.5E6psi
YME=.1E6psi
S t i m - L a b In c .
P re dk F 0 2
10000
`
1000
100
10
0
2000
4000
6000
8000
10000
12000
C LO S UR E S T R E S S - P S I
1 .0
0 lb //s q ft 2 0 /4 0 B
Ba d g e r1
1 5 0 °F
1 .0
0 lb //s q ft 2 0 /4 0 B
Ba d g e r1
1 5 0 °F
1 .0 lb /s q ft 2 0 /4 0 Ba d g e r1 5 0 °F
1 .0 lb /s q ft 2 0 /4 0 Ba d g e r1 5 0 °F
Source: Stim-Lab Consortium, Feb 2002 1.6-46
Proppant Durability
In the real world:
-
Fractures are subjected to high stresses
(increasing) for extended periods of time
Stress levels fluctuate (cyclic stress) with
wellwork
ll
k and
d changes
h
iin liline pressure
Therefore:
•
•
•
All proppants appear to lose conductivity over
time
ti
Traditional resins do not appear to protect
proppants
p
oppa s from
o deg
degradation.
ada o
Many data suggest degradation is a mechanical
failure, not chemical attack.
38
Proppant Durability
• Traditional “long term” conductivity tests maintain
stress on proppant for 50 hours
Longer test captures a portion of
the time-dependent decline. We
know degradation continues
beyond
y
this,, but modern “50 hour”
tests include correction for initial
repacking/etc.
This phenomenon occurs even
with silica saturation
Reference: SPE 16415 Norton and Stim-Lab
Conduc
ctivity (md
d-ft)
– It is known that proppants continue to degrade beyond
50 hours, but this was a practical compromise between
laboratory expense and accuracy
accuracy.
Fig 4,
4 SPE 16415
1000
20/40 Jordan sand,
8000 psi
100
0
25
50
75
100
Hours at Constant Stress
Extended duration tests:
1984
(75 & 250F)
API “short term” cell: Metal plates, continuous flowing 2% KCl,
Non silica saturated
Non-silica
%O
Original C
Conductiv
vity
Fig 19, SPE 12616 between metal plates
100
20/40 Sand at 75F
80
10/20 Sand at 250F
60
40
20
0
0
30
60
90 120 150 180 210 240 270 300
Days at Constant Stress, 5000 psi
Reference: SPE 12616 by Montgomery, Steanson, Schlumberger
40
Published extended duration tests:
1986
93C (200F)
All non-corrodible surfaces, prop in
Teflon tube, continuous flowing 2% KCl
1986
(300F)
Teflon tube, continuous flowing 2% KCl,
Non-silica saturated
1
0.8
0.6
SPE Drilling, April 1986, page 5
10000
0.4
CarboPROP at 10,000 psi (69 MPa)
CarboLITE at 10,000 psi (69 MPa)
0.2
Sand at 5000 psi (35 MPa)
0
0
15
30
45
60
Days at Constant Stress
75
Con
nductivity (m
md-ft)
Perm
meability R
Ratio
Fig 4, SPE 14133
1000
Interprop
Proflow
RCS
Ottawa Sand
100
0
10
20
30
40
50
Days at Constant Stress, 8500 psi
References: SPE 14133 by CARBO, SPE Drilling article by Norton-Alcoa Proppants and TerraTek Research
41
Temperature Correction for White Sand
At 6500 psi and 250F, 20/40 White Sand loses
40% off it
its conductivity
d ti it compared
d to
t 150F.
150F
Con
nductivity C
Correction ffrom 150 de
eg F, factorr
20/40 Premium White Sand
1
0.8
0.6
0.4
150 deg F
200 degF
0.2
250 deg F
300 deg F
350 deg F
0
0
1000
2000
3000
4000
5000
Stress, psi
6000
7000
8000
9000
10000
StimLab PredictK
Cyclic Loading of Proppant Packs
• All proppants appear to be damaged by
continued stress cycling
Effect of Stress Cycling on Proppants
Three cycles, 6000 to 1000 psi
Con
nductivity (m
md-ft)
5,000
4,000
15% loss
50 hrs at 6000 psi
3 cycles, 6000 to 1000
3,000
2,000
23% loss
32% loss
1,000
RCS #1
CARBO Tech Rpt 99-062
RCS #2
EconoProp
Proppant Type (all 20/40)
Source: CARBO Tech Rpt 99-062
Cond
ductivity ((md-ft)
Effect of Stress Cycling on Proppant Conductivity
(Stim-Lab July 2000 data)
St
Stress
(psi)
( i)
10,000
LWC
9,000
RCS
8 000
8,000
7,000
6 000
6,000
Ceramic loses 26%,
26%
RCS loses 35% due to 25 cycles
5,000
4,000
3,000
2,000
1,000
0
100
200
Hours
300
400
Perc
cent Crrush
((wt% sm
maller tthan 40
0
mesh)
30
Effect of Stress Cycling
on Proppant Crush (6000 psi)
Single Crush at 6000 psi
Triple Cycle Dry Crush
25
Triple Cycle Crush in Long Term Cell
20
15.79
15
10.44
10
5
8.47
3 33 3.83
3.33
1.92
1.37 1.52
3 02
3.02
RCS #1
CARBO Tech Rpt 99-062
RCS #2
EconoProp
Proppant Type (all 20/40)
Options to reduce crush:
Action
Rename 16/20 to 16/30
Crush
Conductivity
⇓ ~50%
No change
Add 30 mesh material to 16/20 and rename to 16/30
⇓ ~60%
⇓ ~30%
Reduce average proppant size or produce broader distribution
⇓
Sticky additive to agglomerate fines
⇓ ~100%
Pre-cured or curable resins
⇓
Include deformable “cushioning” agents
⇓
⇓
⇓
often ⇓ at low stress,
⇑ at high
g stress
⇓
The Correct Way to Test Proppant
• Remember, proppant must achieve two goals:
– Reservoir contact (p
(proppant
pp
volume))
– Ability to conduct hydrocarbons with minimal pressure loss
• These characteristics can be directly measured with
a conductivity test
–
–
–
–
–
Proppant confined between sandstone core
Realistic temperatures
Flowing brine, oil, and/or gas
50 hour duration (or longer)
Cyclic stress, embedment, fines migration, non-Darcy and
g
in specialized
p
tests
other issues can be investigated
– Directly measures parameters of interest [frac width and
flow capacity]
SPE 119242
How to Use and Misuse
Proppant Crush Tests –
Exposing the Top 10 Myths
Questions?
Darcy’s
Darcy
s Law vs
vs. Forchheimer Equation
• Δ P/L = μ v / k
– Pressure drop is proportional to fluid
velocity
l it
– Applicable only at low flowrates
• Δ P/L = μ v / k + β ρ v2
– Pressure drop is proportional to square of
fluid velocity
– Applicable at realistic fracture flowrates
D
Does
F
Fracture
t
Width Affect
Aff t Crush?
C h?
• Crush increases significantly in narrow
fractures
Interior grains are loaded
“evenly”
evenly on 6 sides
Exterior grains are not
stressed uniformly
Long
g Term Conductivity
y Test
Procedure
• Load 63 g (equivalent to 2 lbs/sq ft) of proppant in
each cell.
• Install cells in the press.
• Purge 2% KCl solution with oxygen-free nitrogen.
• Apply a vacuum for 45 minutes to remove air in
cells.
• Flow 2% KCl solution through heated silica sand,
and
d cells.
ll
• Ramp to an initial stress of 1000 psi and to a 500
psi fluid pressure.
• After checking equipment is working properly, heat
cells to 250ºF.
Long Term Conductivity Test
Procedure
• Increase stress to 2,000 psi.
• Flow fluid at rates of 3
3, 4 and 6 ml/min
ml/min.
Measure Δ p 30 minutes after each step
change in flow rate.
• Measure frac width and temperature. Maintain
stress for 50 hr.
2,000
000 psi increments for 50
• Increase stress in 2
hours each.
• Continue measuring Δ p at 3, 4 and 6 ml/min of
fluid flow, frac width and temperature until
12,000 psi stress is reached.
AB
Better
tt Test
T t to
t Select
S l t Proppant
P
t
• Long term conductivity testing
• Direct measurement of flow capacity of
proppant pack
• Can account for:
– Embedment
– Temperature
– Fluid Effects
– Fines Migration
g
((with appropriate
pp p
flowrates))
Is complexity
solely attributed
to “rock
rock fabric”?
fabric ?
Chudnovsky, Univ of Ill, Chicago
54
Unconsolidated 200 mesh sand, 35 lb XLG,
Flow Å SPE 63233
Many other examples! [TerraTek, Baker, Weijers, CSM FAST consortium]
Frac Width – with CrossLinked Gel
Diffuse slurry
Modest concentration
wf
TSO + high concentration
We don’t envision thick
filtercakes in very tight rock,
but it doesn’t take much to
damage a narrow frac!
Diffuse slurry
Low concentration
2 pp
ppa [[240 kg/m
g 3] sand slurryy is
about 1 part solids to 7 parts liquid.
Final frac width could be ~1/7th the
pumping width!
Navigation menu