Post on 13-Mar-2020
Evaluation of liquefaction in
tailings and mine waste:
an update
P.K. Robertson
2018
Definitions of Liquefaction
• Cyclic (seismic) Liquefaction– Zero effective stress
(during cyclic loading)
• Flow (static)
Liquefaction– Strain softening
response
Moss Landing, CA
Fundao, Brazil
Level – gently sloping ground
Steeply sloping ground
Tailings & mine waste
Flow Liquefaction is the one of the main design
issues for most tailings and mine waste structures
• High static shear (“steeply” sloping ground)
• Various trigger mechanisms
– Cyclic loading only one form of trigger
• High risk if significant strength loss possible
Flow liquefaction - Case histories
• Common soil features:– Very young age
– Non-plastic or low-plastic
– Little or no stress history (Ko ~ 0.5)
– Very loose (contractive)
– Low effective stress (s’vo < 3atm)
• Common instability features:– Some triggered by very minor disturbance
– Failures tend to occur without warning
– Failures tend to be progressive & rapid
– Observation approach not valid
Fundao, Brazil, 2015 19 deaths
Stava, Italy, 1985; 268 deaths, 190,000m3
Flow liquefaction – Evaluation Steps
Evaluation Sequence:
1. Evaluate susceptibility for strength loss
2. Evaluate stability using post-earthquake shear
strengths
3. Evaluate trigger for strength loss
If soils are susceptible, and instability possible
(FS<1), it is often prudent to assume trigger
will occur
Outline• Trigger events
– Small events can trigger strength loss
– Critical stress paths
• Classify materials susceptible to strength loss
– Application of SCPTu
• Stability
– Influence of high stresses
– Unsaturated tailings/waste
– Progressive failure
Triggers - Flow Liquefaction
After Olson & Stark, 2003
Undrained loading e.g. rapid construction A-B-C
Undrained cyclic e.g. earthquake A’-E-C
Unloading e.g. increasing GWL A-D-C
Example - Fundao
Unloading stress
paths are the
most critical -
can be drained
or undrained
Higher static
shear stress ratio
- smaller the
trigger
Lab. testing
http://fundaoinvestigation.com/
Classify susceptibility to strength loss
• Geo-materials must be strain softening in
undrained shear
• Strain softening geo-materials are contractive
at large strains
CPT can be used to identify contractive soil
for young, uncemented soils - Robertson(2010)
Case Histories of Flow Liquefaction
Case Histories
Nerlerk (sand) – 19,20,21
Jamuna (sand) - 34
Fraser River (silty sand) - 27
Sullivan mines (silty tailings) - 35
Northern Canada (silty clay) – 36
L. San Fernado Dam (silt) – 15
CPT data in critical layers +/- 1 sd.
(Average stress level < 200 kPa)
All flow liq. case histories plot in ‘contractive’portion of CPT SBT chart
Good theoretical support via State Parameter
19,20,21
34
35
27
36
15
Case Histories of Flow Liquefaction
su(liq) / s ’vo = liq. undrained strength ratio (sand-like)
Robertson, 2010
19,20,21
34
35
27
36
15
Case Histories of Flow Liquefaction
su(liq) / s ’vo = liq. undrained strength ratio (sand-like)
Robertson, 2010
Fundao
Fundao
Soil structure
• Some natural soils and mine tailings/waste have
some form of ‘structure’ that make their behavior
different from ‘ideal’ soil
– Macrostructure (layering, fissuring, etc.)
– Microstructure (particle scale – aging, bonding, etc.)
Canadian Geotechnical Journal, 2016
“CPT-based Soil Behavior Type (SBT) Classification System – an
update” P.K. Robertson
Seismic CPTu
After Mayne, 2014
SCPTu6 -7 measurements!
qt
fs
u2
Vs (Vp)
t50
uo
i
Diss.test
Identification of microstructure
• CPT penetration resistance, qt – controlled by
peak strength
• Shear wave velocity, Vs – controlled by small
strain stiffness
• Potential to identify ‘structured’ soils from
SCPT by measuring both peak strength and
small strain stiffness
New Go/qn Chart
1
10
100
1000
1 10 100 1000
Qtn
IG = Go/qt
Cementation/bonding
& aging
KG = (Go/qt)(Qtn)0.75
Small strain
rigidity index:
IG = Go/(qt-svo)
Normalized
rigidity index:
K*G = IG (Qtn)
0.75
Go / qn
Robertson, 2016
New Go/qn Chart
1
10
100
1000
1 10 100 1000
Qtn
IG = Go/qn
Soils with microstructure
(e.g. cementation/bonding
& aging)
K*G = (Go/qn)(Qtn)0.75
1
2
3
4
5
6
7
8
9 10
11
12
13
14
15
16 17 18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Average
normalized
rigidity index for
young,
uncemented
silica-based
soils:
K*G = 215
Robertson, 2016
Updated CPT-based SBT Charts
Normalized SCPTu parameters: Qtn, Fr, U2 and IG
Microstructure
Ideal
Microstructure
Ideal soils –
no microstructure
Robertson, 2016
Canadian Geotechnical Journal, 2016
“CPT-based Soil Behavior Type (SBT) Classification System – an update” P.K. Robertson
Updated SBTn Charts
1
10
100
1000
0.1 1 10
Qtn
Fr (%)
IB = 32
IB = 22
CD = 70
CD = (Qtn - 11)(1 + 0.06Fr)17
IB = 100(Qtn + 10)/(70 + QtnFr)
SD
CD
CC
SC TD
TC
Soil Behaviour Type 1: CCS Clay-like - Contractive - Sensitive
2: CC Clay-like - Contractive
3: CD Clay-like - Dilative
4: TC Transitional - Contractive
5: TD Transitional - Dilative
6: SC Sand-like - Contractive
7: SD Sand-like - Dilative
CCS
Behavior Descriptions
Robertson, 2016
Soils with no microstructure
Updated SBTn Charts
1
10
100
1000
0.1 1 10
Qtn
Fr (%)
IB = 32
IB = 22
CD = 70
CD = (Qtn - 11)(1 + 0.06Fr)17
IB = 100(Qtn + 10)/(70 + QtnFr)
SD
CD
CC
SC TD
TC
Soil Behaviour Type 1: CCS Clay-like - Contractive - Sensitive
2: CC Clay-like - Contractive
3: CD Clay-like - Dilative
4: TC Transitional - Contractive
5: TD Transitional - Dilative
6: SC Sand-like - Contractive
7: SD Sand-like - Dilative
CCS
Behavior Descriptions
Young-uncemented
NC to LOC soil
Aging, cementation
OCR
Sensitivity
Density
Robertson, 2016
Soils with no microstructure
Updated SBTn Charts
1
10
100
1000
0.1 1 10
Qtn
Fr (%)
IB = 32
IB = 22
CD = 70
CD = (Qtn - 11)(1 + 0.06Fr)17
IB = 100(Qtn + 10)/(70 + QtnFr)
SD
CD
CC
SC TD
TC
Soil Behaviour Type 1: CCS Clay-like - Contractive - Sensitive
2: CC Clay-like - Contractive
3: CD Clay-like - Dilative
4: TC Transitional - Contractive
5: TD Transitional - Dilative
6: SC Sand-like - Contractive
7: SD Sand-like - Dilative
CCS
Behavior Descriptions
DRAINED CPT
UNDRAINED CPT
Transition -
Partial drainage
Soils with no microstructure
Saturated soils
Updated SBTn chart
Robertson, 2016
OCR > 4
Soils with no microstructure
Fundao
1536
Contours of Qtn,cs
70
6044
Contours of Qtn,cs on updated SBTn chart
OCR >4
4460
Ic
Kc
Robertson & Wride, 1998
Based on case histories of cyclic liquefaction
Contours of su(liq)/s’vo
0.25
0.100.05
Contours of su(liq) / s’vo based on Qtn,cs
Based on Qtn,cs(extended to Ic > 2.6)
OCR >4
su(liq) / s ’vo = liq. undrained strength ratio (sand-like)
0.25
0.100.05
Contours of su(liq) / s’vo based on Qtn,cs
Based on Qtn,cs(extended to Ic > 2.6)
OCR >4
su(liq) / s ’vo = liq. undrained strength ratio (sand-like)
Only applies to
Sand-like (SC)
drained CPT
Ic < 2.6 or IB > 32
Contours of su(liq/R)/s’vo
0.60
0.25
0.100.05
Robertson (2016)
OCR
Clay-like soils (undrained CPT)
remolded strength & sensitivity (St)
Modified for
clay-like soils(when Ic > 2.6)
St
St
OCR
NC Clay
St ~ 1
OCR >4
su(R) / s ’vo = remolded undrained strength ratio for clay-like soils
su(R) / s ‘ vo = fs / s ‘ vo = (Fr Qtn )/100
fs ~ su(R)
fs /s’vo = Fr Qtn /100
0.100.01 0.25
Contours of su(liq/R)/s’vo
0.60
0.25
0.100.05
OCR
su(liq) / s ’vo = liq. undrained strength ratio for sand-like soils
su(R) / s ’vo = remolded undrained strength ratio for clay-like soils
Modified for
clay-like soils(when Ic > 2.6)
St
Contours of su(liq) / s ’vo & su(R) / s ’vo
Contours of su(liq)/s’vo
0.25
0.100.05
Contours of su(liq) / s’vo based on Qtn,cs
Based on Qtn,cs(extended to Ic > 2.6)
OCR >4
su(liq) / s ’vo = liq. undrained strength ratio (sand-like)
Contours of su(liq/R)/s’vo
0.60
0.25
0.10
0.05
OCR
su(liq) / s ’vo = liq. undrained strength ratio for sand-like soils
su(R) / s ’vo = remolded undrained strength ratio for clay-like soils
St
Contours of
su(liq) / s ’vo
su(R) / s ’vo
Based on case histories:
•Flow liquefaction for sand-
like soils
•Remolded shear strength for
clay-like soils
Both large strain
undrained strengths
IB
Soils with no microstructure
Stability
• Tailings dams are becoming increasingly
higher (>100m) with much higher overburden
stresses (> 8 atm).
– How do we extrapolate to higher stresses?
• Many tailings dams/waste pads have large
regions that are unsaturated
– How do we evaluate influence of lack of
saturation?
Questions?
Are all contractive soils strain softening?
Are all strain softening soils brittle?
How does stress level affect these?
PBDIII Earthquake Geot. Eng. , Vancouver - 2017
“Evaluation of Flow Liquefaction: Influence of high stresses” P.K. Robertson
State parameter in sands
State Parameter after Jefferies and Been, 1985Critical State Line
(CSL)
(+) Loose
CONTRACTIVE
(-) Dense
DILATIVE
Y = eo - ecsecs
eo
State Parameter after Been and Jefferies, 1985
p’op’csp’o /p’cs = 10(y/l)
Brittleness
Bishop (1967) defined brittleness:
IB = (tp – tr)/ tp
tp = peak strength
tr = residual (large strain) strength at same effective
normal stress
IB = 1.0 (100% strength loss)
IB = 0 (no strength loss)
tp
tr
t
g
Critical State Lines (CSL)Contractive
Dilative
Jefferies and Been, 2016
Limited
stress range
Linear approximation for CSL
CSL for range of sands
Critical State Lines (CSL)
Dilative
Contractive
CSL over wide stress range
Schnaid et al, 2013
Verdugo & Ishihara, 1996
CSL non-linear
over wide stress range
Schnaid et al, 2013
Verdugo & Ishihara, 1996
Boulanger 2003
Bolton (1986) Relative Dilatancy Index
Bolton (1986) Relative Dilatancy IndexIncreasing
compressibility/crushability
Schnaid et al, 2013
Verdugo & Ishihara, 1996
Boulanger 2003
Bolton (1986) Relative Dilatancy Index
Bolton (1986) Relative Dilatancy IndexIncreasing
compressibility/crushability
This format is very helpful to estimate CSL
Bolton’s empirical relationship is helpful guide
Y = 0.07
(y/l ~2.0)
Y = 0.20
Y = 0.25
(y/l ~0.9)
(y/l ~0.6)
p’o /p’cs = 10(y/l)
Mtc = 1.2
p’o /p’cs = 100
p’o /p’cs = 4.5
p’o /p’cs = 3.5
Example – Erksak Sand
Contractive
Dilative
Loosest state
Brittleness (IB) vs (p’o /p’cs)
po’/p’cs~NC Clay
A
B
IB = 0.9 – 0.84(p’o/p’cs)
C
Regardless of fabric and direction of loading
IB
Modified from
Sadrekarimi & Olson, 2011
(Ottawa)
(Illinois)
(Mississippi)
Brittleness (IB) vs (p’o /p’cs)
po’/p’cs~NC Clay
A
B
IB = 0.9 – 0.84(p’o/p’cs)
C
IB
Modified from
Sadrekarimi & Olson, 2011
(Ottawa)
(Illinois)
(Mississippi)
Can use y but requires slope of CSL l
(y/l) which is changing
Brittleness (IB) vs su,cs/s’vo
IB
A
B
C
su,cs/s’vo
~NC Clay
Modified from
Sadrekarimi & Olson, 2011
(Ottawa)
(Illinois)
(Mississippi)
Brittleness (IB) vs su,cs/s’vo
IB
A
B
C
su,cs/s’vo
High Brittleness
Low Brittleness
Case histories
Modified from
Sadrekarimi & Olson, 2011
Brittleness (IB) vs su,cs/s’vo
IB
A
B
C
su,cs/s’vo
High Brittleness
Low Brittleness
Case histories
Modified from
Sadrekarimi & Olson, 2011
su,cs/s’vo < 0.15
have higher brittleness
IB > 0.4
0.60
OCR
su(liq) / s ’vo = liq. undrained strength ratio for sand-like soils
su(R) / s ’vo = remolded undrained strength ratio for clay-like soils
Contours of
su(liq) / s ’vo
su(R) / s ’vo
Based on case histories:
•Flow liquefaction for sand-
like soils
•Remolded shear strength for
clay-like soils
Both large strain
undrained strengths
Soils with no microstructure
0.15
High IBat low stress
Contours of su(liq/R)/s’vo
Contours of su(liq/R)/s’vo
0.60
0.25
0.10
0.05
OCR
su(liq) / s ’vo = liq. undrained strength ratio for sand-like soils
su(R) / s ’vo = remolded undrained strength ratio for clay-like soils
St
IB
Sand-like and Dilative – SD
•Potential for cyclic liquefaction –
depends on size and duration of
cyclic loading
•CRR will increase with
increasing static shear
•Sampling difficult & in-situ
testing preferred
•Estimate CRR based on Qtn,cs
•No strength loss expected unless
state changes and/or some
microstructure (e.g. cementation)
Liquefaction Summary
Soils with no microstructure
Contours of su(liq/R)/s’vo
0.60
0.25
0.10
0.05
OCR
su(liq) / s ’vo = liq. undrained strength ratio for sand-like soils
su(R) / s ’vo = remolded undrained strength ratio for clay-like soils
St
IB
Sand-like and Contractive –
SC•Potential for cyclic & flow
liquefaction
•CRR will decrease with
increasing static shear
•Sampling difficult & in-situ
testing preferred
•Strength loss possible
•Estimate state (y) and su(liq)/s’vo
based on Qtn,cs
•Strain to trigger strength loss can
be small
Liquefaction SummaryContours in SC/TC
region based on
s‘vo < 2atm)
Contours of su(liq/R)/s’vo
0.60
0.25
0.10
0.05
OCR
su(liq) / s ’vo = liq. undrained strength ratio for sand-like soils
su(R) / s ’vo = remolded undrained strength ratio for clay-like soils
St
IB
Clay-like and Contractive –
CC/CCS•Potential for cyclic softening &
flow liquefaction
•CRR can decrease with increasing
static shear
•Sampling possible and
recommended
•Strength loss possible
•Estimate state (OCR) and su(R)/s’vo
based on CPT & FVT
•Strain to trigger strength loss can
be large – depends on plasticity
•Evaluate sensitivity using CPT fs
and FVT
•Check drainage CPTu dissipation
tests
•Measure wn and Atterberg
Liquefaction Summary
Soils with no microstructure
Contours of su(liq/R)/s’vo
0.60
0.25
0.10
0.05
OCR
su(liq) / s ’vo = liq. undrained strength ratio for sand-like soils
su(R) / s ’vo = remolded undrained strength ratio for clay-like soils
St
IB
Clay-like and Dilative – CD
•Liquefaction unlikely
•Sampling possible, if not too stiff
•No strength loss expected unless
state changes and/or some
microstructure
•Measure wn and Atterberg
•Drained strength valid
Liquefaction Summary
Soils with no microstructure
Contours of su(liq/R)/s’vo
0.60
0.25
0.10
0.05
OCR
su(liq) / s ’vo = liq. undrained strength ratio for sand-like soils
su(R) / s ’vo = remolded undrained strength ratio for clay-like soils
St
IB
Transitional– TD/TC
•Evaluate plasticity and drainage
(Atterberg Limits & CPTu
dissipation tests)
•Sampling possible depending on
plasticity and fines
•Evaluate assuming both sand-
like and clay-like and compare
• typically drained
strengths are more
conservative in SD, SC,
TD, TC and CD
Liquefaction Summary
Soils with no microstructure
Case History• Filtered mine waste placed via conveyor and radial arm
stacker with surface irrigation
• Placed in uncompacted lifts of about 20 to 30m thickness
• Overall slope of about 4H:1V
• Current max. height of 200m
• Approx. 100 million tons of waste
• Low precipitation & high evaporation region
Canadian Geotechnical Journal, 2017
“Characterization of unsaturated mine waste: case history”
Robertson, P.K., da Fonseca, A.V., Ulrich, B. & Coffin, J.
Typical index test results
• Uniform grain size distribution – sandy silt
• Fines content ~ 50%
• D50 ~ 0.065mm
• Specific gravity, Gs = 2.73
• Mostly non-plastic
• In-situ water content (w) varies from around
2% to 15% (depending on depth, location and
irrigation) with degree of saturation
5% < S < 50%
Typical CPTu
High stress
s’vo ~15atm
Typical Normalized CPTu
High stress
s’vo ~15atm
Typical Normalized CPTu
Updated SBTn Charts
Little or no microstructureMostly drained
CPT penetration
Mostly
Transitional Contractive
Updated CPT SBTn chart
Qtn-Fr chart indicates
that material is mostly
sand-like and
CONTRACTIVE at
large strains
(assumes no suction)
Robertson, 2016
0 to 15m depth range
s’vo < 300kPa
Updated CPT SBTn chart
Qtn-Fr chart indicates
that material is mostly
sand-like/transitional
and CONTRACTIVE
at large strains
(assumes no suction)
Robertson, 2016
15 to 90m depth range
s’vo > 300kPa
Updated CPT SBTn chart
0.25
0.10
0.05
With increasing depth,
Qtn decreases and Fr
increases
Trends toward a more
clay-like contractive
behavior, but often less
brittle
Seismic Velocity (SCPTu)
Vs
Vp
Vs1 = Vs (pa/s’vo)0.25
Seismic Velocity (SCPTu)
Vs1 ~ 225m/s
Vp < 1500m/s
Vs
Vp
Vs1
DilativeContractive
Robertson, 1995
Unsaturated
In-situ test interpretation
• CPT (Qtn) based interpretation suggests that
soils are mostly CONTRACTIVE at large
strains
• Shear wave velocity (Vs) based interpretation
suggests that soils are DILATIVE at large
strains
Which one is correct?
Critical State line (saturated)
Contractive
Dilative
Reconstituted samples
Loosest state
Limiting compression
curve (saturated)
?
In-situ stress range
Approx.
e(max)
Approx.
e(min)
Vs contours and CSL
Vs(field) > 200m/s
Assuming materials are saturated
Contractive
Dilative
Vs measured in lab.
at end of consolidation
Vs1 as a function of suction (s)
Increasing suction
(decreasing S)
3 tests 50 < sc < 600 kPa
4 tests 50 < sc < 600 kPa
Normalize Vs1 based on modified s’vo (adjusted for suction)
Sandy silt tailings
4 tests 50 < sc < 600 kPa
Vs1 as a function of suction (s)
Increasing suction
(decreasing S)
?3 tests 50 < sc < 600 kPa
4 tests 50 < sc < 600 kPa
In-situ range
Sandy silt tailings
CSL as a function of suction (s)
s = 0 kPa
Moist samples prepared to ei ~ 1.1 to 1.2
Sandy silt tailings
~loosest compression line after saturation
CSL as a function of suction (s)
s = 0 kPa
Moist samples prepared to ei ~ 1.1 to 1.2
Sandy silt tailings
~loosest compression line moist
saturated samples
CSL as a function of suction (s)
s = 200 kPa
s = 20 kPa
s = 0 kPa
Moist samples prepared to ei ~ 1.20
CSL (unsat.)
Sandy silt tailings
~loosest compression line moist
CSL as a function of suction (s)
s = 200 kPa
s = 20 kPa
s = 0 kPa
CSL (unsat.)
Project tailings - Strain hardening when unsaturated
Limited strain softening if saturated
Sandy silt tailings
Case History - Summary
• Shear wave velocity (Vs) appears to be sensitive
to suction hardening, since it is a small strain
measurement
– Vs better indicator of unsaturated behavior
• Cone resistance (qc) appears to destroy
beneficial effects of suction hardening, since it is
a large strain measurement
– qc may better indicate behavior if soil becomes
saturated
Critical regions to investigate?
Tailings/waste
Foundation soils
Simplified piezometric
surface
Critical regions to investigate?
Tailings/waste
Foundation soils
Simplified piezometric
surface
Toe region:
• Higher static shear stress ratio
• Low effective confining stress
• Higher possibility to be 100% saturated
• Risk of progressive failure
Conclusion (1)
• Flow liquefaction can be triggered by
relatively minor conditions and ’unloading’
stress paths are the most critical, e.g.
– Earthquakes (even small earthquakes)
– Rising piezometric surface
– Unloading (movements in foundations or tailings)
• Not all CONTRACTIVE soils are strain
softening in undrained shear
• Not all strain softening soils are brittle
– brittleness varies with state, stress level and PI
• Although soils tend to become more contractive
with increasing stress they also tend to become
more ductile
• Low stress regions maybe most critical
Conclusion (2)
• SCPTu an excellent tool to evaluate potential
risk of flow liquefaction
– 5 to 6 measurements
– Ability to identify microstructure
– Classify soil behavior
– Vs and Vp powerful additional measurements
• Lab. testing to helpful to support evaluation via
CSL
Conclusion (3)
Questions?
PDF Copy of slides:www.cpt-robertson.com
Free (recorded) Webinars:www.greggdrilling.com/webinars