4/14/2014

consulting engineers and scientists

Seismic Design
of Tailings Dams

Liquefaction
By Gonzalo Castro
Lima, Peru April 18,19 -2014
Santiago, Chile April 23-24 - 2014
San Juan, Argentina April 28,29 - 2014

OUTLINE, PART 1

1. Fundamentals

Steady State
Stability and Instability
Instability examples
Seismic Deformations Example

Instability 
Triggering for Non Plastic Soils
Triggering for Plastic Soils

Seismic Deformations for Stable Case
Yielding
Ratcheting

OUTLINE, PART 2

2. Practice

Determination of in situ Sus for Non Plastic Soils
Empirical Charts
Laboratory site specific procedure
Increase in Sus with consolidation, non plastic soils

Determination of in situ Sus for Plastic Soils
Conventional Practice, field and laboratory tests

Charts for Determination of Pore Pressure Increase for Non Plastic 
Soils (“Triggering Charts”)
Stratified Soils

1
 4/14/2014

State Diagram

Initial State
Steady State
℮

Steady State Line

Steady State

The steady state of deformation is defined as

the state in which a specimen deforms
continuously and monotonically without change
in shear stress, effective normal stress, volume,
or velocity of deformation.

Poulos, S.J., 1971

Stress Strain behavior for Contractive

Sands, Drained vs. Undrained

℮ Undrained
Drained
Drained

Undrained

Strain 

2
 4/14/2014

Stress Strain behaviors for Contractive and

Dilative Sands
Contractive Dilative
Sdp
Sds
Sds
SSL
℮
SSL
Drained ℮

Strain
℮
Strain

SSL

Sup
SSL
Undrained Sus

S : Shear Strength Values

Conditions for Test to define SSL

Tests to define SSL must fulfill two main

conditions:

1. Sufficient strain can be applied to reach

steady state
2. Uniformity of void ratio and stresses within
the specimen must be maintained

Volume Changes During Shear

3
 4/14/2014

Suitability of Various Test Types to

Define SSL

Sand Silt Clay

Triaxial Compression (1) Y Y(2) N
Vane (field or lab), (3) N Y(3) Y
Rotational Shear (4) N Y Y

(1) Contractive soils, drained or undrained. Axial extension

cannot apply sufficient strain
(2) Strain may not be sufficient
(3) Undrained, may need to be performed faster than normal
in silts
(4) Drained, determination of e at steady state may be
difficult

Steady State Line, Different Stress Paths

Castro et al, 1992

Steady State Line, Different Laboratories

After Castro et al, 1989

4
 4/14/2014

Steady state line

(Verdugo and Ishihara 1996)

Strength Envelope at the Steady State

(Verdugo and Ishihara 1996)

Potential Behavior as a Function of Sus

5
 4/14/2014

Summary of Significance
of Sus in Seismic Behavior of Sands

• Sus is sufficiently low for Flow Slides to occur only for

very loose sandy soils, N1,60 less than about 10 to
12.
• For higher values of N1,60 Sus increases rapidly with
density
• At N1,60 of 12 and higher most sands are dilative
• Thus for these higher N1,60 values, instability is not
an issue. Potential for damage relate to magnitude of
deformations. These are likely to be an issue only if
one expects high seismically induced pore pressures.
In this case Sus does not control the deformations.

Main Features of Unstable Case

• Loose saturated sands case

− Driving Shear Stress is applied drained
− Applicable Steady State Strength is undrained
• Triggering Strain is low for loose saturated sands
(0.2 to 1%)
• Earthquake triggers the failure if accumulated
strain reaches value of triggering strain
• Failure is typically a major slide,
e.g. Lower San Fernando Dam
• Failure driven by Static stresses and may occur
after the end of shaking

LSFD Case Implied Stress Strain Behavior

6
 4/14/2014

Main Features of Stable Case

• Dam remains stable regardless of

− earthquake intensity and/or
− momentary pore pressure increases
• Accumulated strain increases with earthquake
intensity
• If accumulated strain is large, failure may
occur by overtopping or erosion along cracks
• Deformations occur mostly during shaking

OUTLINE, PART 1

1. Fundamentals

Steady State
Stability and Instability
Instability examples
Seismic Deformations Example

Instability 
Triggering for Non Plastic Soils
Triggering for Plastic Soils

Seismic Deformations for Stable Case
Yielding
Ratcheting

Fort Peck Dam, Static Failure, 1938

After A. Casagrande, 1965

7
 4/14/2014

Cerro Negro Tailings Dam, Chile

Seismic Failure, 1985
Castro and Troncoso, 1989

Veta De Agua Tailings Dam, Chile

Seismic Failure, 1985
Castro and Troncoso, 1989

Methods of Tailings Dam Construction

8
 4/14/2014

Lower San Fernando Dam, Seismic Failure, 1971

Lower San Fernando Dam

Strains in Critical Layer

Displacement Shear Strain Event

Of Failure Mass in Layer
feet %

0.2 0.5 Triggering Strain

7 20 Steady State
Reached

150 400 Movement

Stopped

Upper San Fernando Dam, Seismic Deformations

1971

Hydraulic fill sand

Hydraulic fill sand
Clay core

Upper Alluvium

Lower Alluvium

(Seed et al., 1973)

9
 4/14/2014

OUTLINE, PART 1

1. Fundamentals

Steady State
Stability and Instability
Instability examples
Seismic Deformations Example

Instability 
Triggering for Non Plastic Soils
Triggering for Plastic Soils

Seismic Deformations for Stable Case
Yielding
Ratcheting

Triggering of Unstable Case

Monotonic Case Parameters

required to trigger
d

S s
Strain at Peak

Strain
Induced
Pore
Pressure


at Peak

Strain

Triggering of Unstable Case

Seismic Case Parameters

Possible Creep
d

Triggering
Strain

Strain

Induced
Pore
Pressure

at Triggering

Strain

10
 4/14/2014

Sands Tested for Triggering Parameters

Reference Soil USCS Fines Cu Grain Source

% Shape

Castro, 69 Sand B SP 1 1.8 Subrounded Ottawa Silica Company,

Sand C SP 0 2.3 Angular Huachipato, Chile
Castro et al, 82 Sand 6 SP 0 1.7 Subrounded Ottawa Silica Company,
Tailings, Copper Mining,
Tailings SP-SM 7 2.7 Angular Canada
Hydraulic Fill, Lower San
Vasquez, 88 SF7 SM-ML 50 32 Angular Fernando Dam
F125 SP-SM 12 1.8 Subangular -Angular Ottawa Industrial Sand Co
Sand A SP-SM 13 2.8 Subangular -Angular Lagunillas, Venezuela
Same batch sample as SF7 in
Castro et al, 89 SF7 Vasquez 88

Coal Combustion by-product,

GEI, 2011 Fly Ash ML 79 7.3 Mostly spherical USA

Increase in (σ1-σ3) to trigger divided by σ3c

Castro, 1994, revised

Triggering Pore Pressure

Castro, 1994, revised

11
 4/14/2014

SF7 soil – Triggering pore pressure - Test type

Triggering Strain

SF7 soil – Triggering strain – Test type

12
 4/14/2014

OUTLINE, PART 1

1. Fundamentals

Steady State
Stability and Instability
Instability examples
Seismic Deformations Example

Instability
Triggering for Non Plastic Soils
Triggering for Plastic Soils

Seismic Deformations for Stable Case
Yielding
Ratcheting

Hypotetically Anisotropically CU Tests on Three

Specimens Having same FS against Liquefaction

(Poulos et al. 1985b)

Typical Undrained Strengths of Fine Tailings

(Castro 2003)

Tailings Percent Plasticity Slurry samples In Situ(1)

Case No
Type Fines Index
Sup/σv Sus/σv Sup/σv Sus/σv
1(a) Copper 100 17 0.27 0.07
2(a) Copper 33 11 0.20 0.11
3(a) Copper 100 6 0.29 0.08
4(b) Copper 7-13 0.25-0.30
5(c) Coal 75 12 0.21
} 0.17-0.25
6(c) Coal 52 4 0.27
7(c) Coal 70 8 0.19
8(c) Coal 29 4 0.20 } 0.27
9(c) Coal 24 0 0.29
10(c) Coal 62 9 0.20
} 0.12
11(c) Coal 62 9 0.26
12(d) Hematite >0.6(2) 0.15-0.22
13(e) Kettle River ~80 8-10 0.23
Natural
14(f) 68-100 1-12 0.35 0.12
Clayey Silt
Aluminum
15(g) 98 13 0.33(3) 0.10
(Red Mud)
(1)Strengths derived from CPT, field vanes, or tests on undisturbed samples (2)Samples dilative, undrained strengths higher than drained strengths (3) Neglecting thixotropic effects

References: (a) Castro & Troncoso, 1989; (b) Ladd, C., 1991; (c) Authors’s files; (d) Walton et al., 2002; (e) Fabian et al., 2002; (f) Castro et al., 2003); (g) Poulos et al. (1985)

13
 4/14/2014

Peak Undrained Strength Ratio for Tailings

and Natural Soils

Aluminum Fine Tailings, PI =13

Effect of Cyclic Straining on Peak Undraned

Strength of Natural Clayey Silt, PI = 1 to 12
Castro, 2003

14
 4/14/2014

Effect of Cyclic Straining Compare with Common

Assumption to “Compute” Available Strength
Castro, 2003

Effect of Seismic Loading on Plastic Soils

Most Plastic Soils(PI>7) with high strain at peak

(> about 5%) and small subsequent decrease in
resistance:

1. Will soften only if static plus seismic stresses

are about equal to Sup
2. Upon softening may loose up to about 20%
of Sup but will not drop to Sus
3. Careful sampling and testing should be
performed to confirm the above

OUTLINE, PART 1

1. Fundamentals

Steady State
Stability and Instability
Instability examples
Seismic Deformations Example

Instability 
Triggering for Non Plastic Soils
Triggering for Plastic Soils

Seismic Deformations for Stable Case
Yielding
Ratcheting

15
 4/14/2014

Yielding Mechanism

Double Integration to Obtain Displacement

After Newmark, 1965

(MSHA Manual 2009)

Permanent Displacement u vs. ratio of yield

acceleration (N) to peak acceleration (A)

Based on 348 horizontal

components and 6
synthetic accelerograms

(Hynes-Griffin and Franklin 1984)

16
 4/14/2014

Variation of Displacement with ratio of Yield

Acceleration (ky) to peak acceleration (kmax)

(Makdisi and Seed 1978)

USFD, Implied Stress Strain Behavior

Ratcheting Mechanism

OUTLINE, PART 2

2. Practice

Determination of in situ Sus for Non Plastic Soils
Empirical Charts
Laboratory site specific procedure
Increase in Sus with consolidation, non plastic soils

Determination of in situ Sus for Plastic Soils
Conventional Practice, field and laboratory tests

Charts for Determination of Pore Pressure Increase for Non Plastic 
Soils (“Triggering Charts”)
Stratified Soils

17
 4/14/2014

Typical Correlations of Sus with Blowcounts

Sus vs.
Reference Sus/σ′v vs. N1,60 Correction of N1,60 for Fines Content
N1,60

Seed (1987) yes no yes

Seed and Harder (1990) yes no yes

Baziar and Dobry (1995) yes no no

Castro (1995) yes no no

Wride, McRoberts and Robertson (1999) yes yes yes

Olson and Stark (2002) no yes no

Idriss and Boulanger (2008) yes yes yes

CASE HISTORY DATA

FOR SANDS, After Seed and Harder, 1990

CASE HISTORY DATA

FOR SANDS, After Castro, 1995

18
 4/14/2014

Case History Data for Sands, After

Olson and Stark 2002

Residual shear strength of liquefied sand vs.

equivalent clean-sand SPT-corrected blow count

(Idriss and Boulanger 2008)

Normalized residual shear strength ratio of

liquefied sand

σ’vc<400 kPa

(Idriss and Boulanger 2008)

19
 4/14/2014

Correlation between the normalized residual shear strength

ratio for liquefied soils and overburden-corrected CPT
penetration resistance

σ’vc<400 kPa

(Idriss and Boulanger 2008)

General Comments on Sus Empirical Charts

• Reliable Sus information can be obtained only

from instability failures (flow slides)
• Analysis to backfigure Sus from instability failures
must consider kinematics of the failure
• Limited seismic deformation cases provide a
lower bound value of Sus
• Case histories are limited to N1,60 < 10 (no fines
correction)

• Sus correlations with blowcounts have a large

scatter as blowcounts (or CPT) do not “measure”
Sus and are only crude indices

Use of Sus vs Sus Strength Ratio

• Sus is a function of void ratio

• N1 is obtained from N using a factor Cn that
was obtained from large scale tests so that N1
is a function of void ratio at the confining
pressure at which N was measured
• Thus it follows that Sus (not Sus ratio) should
be a function of N1
• Using strength ratio would be unconservative
or conservative for cases for which the
confining pressure is higher or lower than for
the case histories, respectively

20
 4/14/2014

OUTLINE, PART 2

2. Practice

Determination of in situ Sus for Non Plastic Soils
Empirical Charts
Laboratory site specific procedure
Increase in Sus with consolidation, non plastic soils

Determination of in situ Sus for Plastic Soils
Conventional Practice, field and laboratory tests

Charts for Determination of Pore Pressure Increase for Non Plastic 
Soils (“Triggering Charts”)
Stratified Soils

Steady-State Lines for Sands-Subangular grains

(Poulos et al. 1985a)

Procedure for determining steady-state strength

of soil at field void ratio (Poulos et a. 1985a)

21
 4/14/2014

Steady-state line for clean sand

(Poulos et al. 1985a)

Lower San Fernando Dam, Seismic Failure, 1971

Cross Section through Lower San Fernando Dam

Castro et al 1992

1985 Geometry

22
 4/14/2014

Plan view of Lower San Fernando Dam showing

locations of borings and exploratory shaft

(Castro et al. 1992)

GEI Tripod Tube Sampler

Tripod tube
sample 303
being taken at
a depth of 83.6
feet in the
exploration
shaft

(Castro et al., 1989)

Hydraulic Fill Zones at Location III

Castro et al 1989

23
 4/14/2014

Grain Size Curves Hydraulic Fill Zone 5 Samples

Castro et al 1989

In situ void ratios within Zone 5 of hydraulic fill at

time of the 1985 field studies

(Castro et al. 1992)

Correction of laboratory-measured steady-state

strength data for effects of void-ratio changes

(Castro et al. 1992)

24
 4/14/2014

Plan View of Survey Lines

(Castro et al., 1989)

Vertical Movement vs. Time

Dam initially constructed – 1916.

Post 1971 Earthquake not shown

Survey
point at
122 feet
South
Station
6+00

(Castro et al., 1989)

Vertical Movement vs. Time

Dam initially constructed – 1916.

Post 1971 Earthquake not shown

Survey
point at
122 feet
South
Station
9+00

(Castro et al., 1989)

25
 4/14/2014

Horizontal Movement and Settlement

Downstream slope
Station 5+00

(Castro et al., 1989)

1971 In Situ Steady State Strengths

Castro et al 1989

Upstream
critical layer
soils vs.
Elevation

Schematic Diagram of Load Controlled Cyclic Device

(Castro et al., 1989)

26
 4/14/2014

LSFD Case Implied Stress Strain Behavior

Conclusion on Methodology to obtain Sus used

at LSFD in the 1985 -1987 investigation

• The results were in remarkable agreement

with the facts of the LSFD slide
• Corrections of Sus for void ratio changes are
substantial
• Correction methodology for void ratio
changes was defined prior to the investigation
• Sampling and Testing requires a degree of
care that is not common in geotechnical
investigations

OUTLINE, PART 2

2. Practice

Determination of in situ Sus for Non Plastic Soils
Empirical Charts
Laboratory site specific procedure
Increase in Sus with consolidation, non plastic soils

Determination of in situ Sus for Plastic Soils
Conventional Practice, field and laboratory tests

Charts for Determination of Pore Pressure Increase for Non Plastic 
Soils (“Triggering Charts”)
Stratified Soils

27
 4/14/2014

Increase in Sus with Confining Pressure

(Example is for Sands)

(Sus)1 = (Sus)o * [(σv’)1 / (σv’)o] Cc/Css

SSL (Sus)

0.8 Css
Compression Curve ( σv’)
Void
Ratio
Cc
0.7

10
0.1 1 100
(Sus)o (Sus)1 (σv’)o (σv’)1
Sus and σv’, tsf (Log scale)

Effect of Overburden on Sus

Mine Tailings

Effect of Overburden on Sus

Silty Sand from LSFD

28
 4/14/2014

Effect of Increase in Overburden on Sus

Summary

• Depends on relative slopes of compression

and SSL lines
• Slope of compression line depends on initial
density
• For non plastic soils, generally compression
lines are flatter, thus Sus increases at a lower
rate than the overburden
• For plastic soils the lines are generally
parallel if starting very loose, and thus Sus
and overburden increase at the same rate

OUTLINE, PART 2

2. Practice

Determination of in situ Sus for Non Plastic Soils
Empirical Charts
Laboratory site specific procedure (Poulos et al)

Increase in Sus with consolidation, non plastic soils

Determination of in situ Sus for Plastic Soils
Conventional Practice, field and laboratory tests

Charts for Determination of Pore Pressure Increase for Non Plastic Soils 
(“Triggering Charts”)
Stratified Soils

Schematic of the approach used to develop relationships

between the in-situ CRR of sand and the results of in-situ
tests

(Idriss and Boulanger 2008)

29
 4/14/2014

SPT clean-sand base curve for M=7.5 earthquakes

with data from liquefaction case histories

(Youd et al. 2001)

SPT case histories of liquefaction in cohesionless

soils with various fines contents

M=7.5
 vc'  1 atm

(Idriss and Boulanger 2008)

Recommended “deterministic” SPT-based

liquefaction triggering correlation for various FCs

(Seed et al 2003)

30
 4/14/2014

Curve recommended for calculation of CRR from CPT

data along with empirical liquefaction data from compiled
case histories

(Youd et al. 2001)

CPT relationships for cohesionless soils with

various fractions of nonplastic fines

(Idriss and Boulanger 2008)

VS1-based liquefaction correlation for clean

uncemented sands (after Andrus and Stokoe 2000)

(Idriss and Boulanger 2008)

31
 4/14/2014

Threshold Strain
Dobry, 1985

Comparison of advantages and disadvantages of various

field tests for assessment of liquefaction resistance

(Youd et al. 2001)

“Triggering” Charts Comments

• Triggering Charts provide information on

potential for pore pressure increase for a
given earthquake
• They do not provide information as to the
consequences of pore pressure increase,i.e.
flow slide or deformations
• Note that if Sus is low enough to have a flow
slide potential, triggering chart is
unconservative since the flow side would be
triggered at a less than 100% pore pressure

32
 4/14/2014

OUTLINE, PART 2

2. Practice

Determination of in situ Sus for Non Plastic Soils
Empirical Charts
Laboratory site specific procedure
Increase in Sus with consolidation, non plastic soils

Determination of in situ Sus for Plastic Soils
Conventional Practice, field and laboratory tests

Charts for Determination of Pore Pressure Increase for Non Plastic 
Soils (“Triggering Charts”)
Stratified Soils

Typical soil profile earth dam foundation

(Castro 1991)

Expanded cone log in the stratified clay and silty

sand

Cone penetration resistance, qc, tsf

Depth (ft)

d = width of peak

(Castro 1991)

33
 4/14/2014

Peak cone resistance as a function of width of peak

Peak cone resistance (tsf)

(Castro 1991) Width of Peak (ft)

• GRACIAS POR SU ATENCION

34