Dynamic calculation for the dry closure of Almagrera tailings dam

Calcul dynamique pour la fermeture sec du barrage des striles dAlmagrera

Justo J.L.,
de J.L.,
Morales-Esteban
Durand
Vzquez-Boza
Justo
Morales-Esteban
A., A.,
Durand
P., P.,
Vzquez-Boza
M. M.
University of Seville, Seville, Spain

Jimnez F.A.

Egmasa, Seville, Spain

Rossi E.

Politecnico di Milano, Milan, Italy

ABSTRACT: An original model, including dynamic calculation, has been developed for the dry closure of Almagrera tailings dam
and is described herein. A dynamic analysis of a structure requires the previous definition of the accelerograms and the structure
characteristics. A probabilistic method for selecting calculation accelerograms is presented in this paper. First, the probabilistic hazard
equation for site is solved. Based on the hazard curves obtained, the uniform seismic hazard acceleration response spectrum
(USHARS) is constructed for the location, according to the type of soil and the required hazard level (exposure time and exceedance
probability). Then, calculation accelerograms are selected. Based on this methodology, real accelerograms, for a return period of 975
years, have been obtained.
RSUM : Un modle original, dvelopp pour la fermeture sec du barrage des striles dAlmagrera, est dcrit ici. Lanalyse
dynamique d'une structure ncessite la dfinition pralable des acclrogrammes et les caractristiques de la structure. Une mthode
probabiliste pour la slection des acclrogrammes de calcul est prsente dans cet article. Tout d'abord, l'quation probabiliste des
risques pour lemplacement est rsolue. Bas sur les courbes de risque obtenues, le spectre de rponse dacclration de risque
uniforme est construit pour l'emplacement, selon le type de sol et le niveau de risque souhait (temps d'exposition et probabilit de
dpassement). Puis, les acclrogrammes de calcul sont slectionnes. Sur la base de cette mthodologie, les acclrogrammes rels,
pour une priode de retour de 975 ans, ont t obtenus.
KEYWORDS: Tailings dam, dynamic calculation, uniform seismic hazard acceleration spectrum
1

INTRODUCTION

In Europe, there are many abandoned mines. Nowadays, when

permission is granted for opening a mine in any country of the
EU, a closure plan (including financing) must be presented by
the mining company (ITC 2000), but it was not so in the old
times. The point is that, up to now, only a small number of the
possible closures has been undertaken owing to economic
reasons.
2

TAILINGS DAMS INVENTORIES AND FAILURES

The recent (5 October 2010) Ajkai Timfoldgyar dam failure

(Fig. 1) poured 700,000 m3 of bauxite ore and formed a flow
that struck three villages in Hungary.

Rico et al. (2008) have compiled a corpus of 147 cases of

worldwide tailings dam disasters, from which 26 are located in
Europe.
Davies & Martin (2000) estimate there are 3500 appreciable
tailings dams worldwide. According to Davies (2002) during
the last years, there have been from 2 to 5 major tailings dam
failure incidents. Referred to the worldwide inventory of 3500
tailings dams, equates to an annual probability between 1/1750
and 1/700, compared with 1/10000 for conventional dams.
Furthermore, these failure statistics are for physical failures
alone. Tailings impoundments can have environmental failure
(e.g. leaks) while maintaining sufficient structural integrity.
Table 1 shows the inventory of tailings dams and ponds in
extractive industry in Spain. Notice the large number of
abandoned structures.
Table 1. Inventory of tailings dams and ponds in extractive industry in
Spain.
323 million
Volume of residues (m3)
Number of structures
986
Structures
Dams
610
Ponds
378
Present state
Active
385
Abandoned
535
Restored
54
Closed
24

Figure 1. Ajkai Timfoldgyar dam failure.

This accident, jointly with the catastrophic failures of Stava,

Los Frailes and Baia Mare tailings dams, has emphasized the
catastrophic consequences that tailings dams accidents in EU
and the rest of the world might cause, and the need for safer
design methods.

3 SAFETY FACTORS REQUIRED FOR SPANISH

LEGISLATION
Three kinds of actions are considered according to its risk and
permanence. Normal actions are persistent actions; accidental
actions are limited duration actions: e.g. rapid drawdown or
earthquakes. Extreme are actions that rarely occur. The safety
factors are indicated in Table 2.

1511

Proceedings of the 18th International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
shaped drains. Band-shaped drains should be always placed
when the
tailings
thickness isdrains
15 m.should be always placed
Table 2. Safety factors required for tailings dams in Spanish legislation.
shaped
drains.
Band-shaped

Classfactors
Category
Table 2. Safety
required Actions
for tailings dams in Spanish legislation.
Normal Accidental Extreme
Class Category Actions
Table 2. Safety
factors
for tailings
dams in Spanish
1
A or required
B
1.4
1.3
1.2 legislation.
Normal
Accidental
Extreme
Class
Actions
2
CororBrequired
D
1.3
Table 2.1Safety
factors
for tailings
dams in Spanish
ACategory
1.4
1.31.2
1.21.1 legislation.
Normal
Accidental
Extreme
3
C
1.2
1.1
1.0
Actions
2Class
CCategory
oror
DD
1.3
1.2
1.1
1
A
or
B
1.4
1.3
1.2
Normal
Accidental
Extreme
3
C or D
1.2
1.1
1.0
4 ALMAGRERA
DAM
C or
or B
DTAILINGS
1.3
1.2
1.1
12
A
1.4
1.3
1.2
3
C
or
D
1.2
1.1
1.0
2
C
or
D
1.3
1.2
1.1
4 ALMAGRERA TAILINGS DAM
Almagrera
tailings
3
C or Ddam has
1.2 a height
1.1of 35 m above
1.0 foundation

The
heightdam
above
lowest
level
is 37.3 m
4at axis.
ALMAGRERA
TAILINGS
DAM
Almagrera
tailings
has the
a height
offoundation
35 m above
foundation
3). height above
at4(Figure
axis.
The
the lowest
foundation level is 37.3 m
ALMAGRERA
TAILINGS
DAM
It has
upstream
core. The
foundation
formed by
Almagrera
damsloping
has a height
of 35
m aboveisfoundation
(Figure
3). antailings
of
volcanic
and
inter-stratified
atalternation
height
above
the
lowest
level
isrocks.
37.3bym
Itaxis.
has The
an tailings
upstream
sloping
Thefoundation
is formed
Almagrera
dam
has
acore.
height
offoundation
35sedimentary
m above
foundation
Figure
3.
Central
cross-section
of
Almagrera
dam
before
closure.
(Figure
3).
alternation
of height
volcanic
and the
inter-stratified
sedimentary
rocks.
at
axis.
The
levellegislation.
is 37.3 m
Table
Safety
factorsabove
required
forlowest
tailingsfoundation
dams
inSpanish
Spanish
legislation.
Table
2.2.Safety
factors
required
for
tailings
dams
in
It has
an upstream
sloping
core. Thedam
foundation
is formed by
(Figure
3).
Figure
3. Class
Central
cross-section
of Almagrera
before closure.
Class
Category
Actions
Category
Actions
corresponds
to core.
the downstream
borrow
material
alternation
volcanic
andNormal
inter-stratified
sedimentary
rocks.
ItThe
has dam
anof
upstream
sloping
The
foundation
is formed
by
Accidental
Extreme
Normal
Accidental
Extreme
Figure
Central
cross-section
ofthe
Almagrera
dam
before
closure.
alternation
of
volcanic
and
inter-stratified
sedimentary
rocks.
The 3.1dam
corresponds
to
downstream
borrow
material
1
orBB
1.4
1.3
1.2
AAor
1.4
1.3
1.2
Figure 3.22Central Ccross-section
of Almagrera
closure.
Cor
orDD
1.3
1.2 dam before
1.1
1.3
1.2
1.1

The dam corresponds to the downstream borrow material

orDD
1.2
1.1
1.0
33
CCor
1.2
1.1
1.0
The dam corresponds to the downstream borrow material
Table 2. Safety factors required for tailings dams in Spanish legislation.
ALMAGRERA
TAILINGS
DAM
44 ALMAGRERA
TAILINGS
Class Category
ActionsDAM
Normal

Accidental

Extreme

when the tailings thickness is 15 m.

Table 3. drains.
Calculation
parameters drains should be always
shaped
Band-shaped
k
Soil3.the
USCS isc'15
should

when
tailingsBand-shaped
thickness
m.
shaped
drains.
drains
be
always
Table
Calculation
parameters
3
typethe tailings thickness
when
15
m. kN/m k m/s-8
Soil
USCS
c'iskPa

placed
E
placed

E MPa
3
50
Core
SC
18
30
19.8
10
3.
Calculation
parameters
Table
m/s -5
MPa
type
kPa
kN/m
-8
Filter
SP-SM 18
35 19.8
k10
E50
Soil
USCS
c'1

20
Table
3. CalculationSC
parameters
50
Core
30
10
-5
MPa
type
kPa
-5
30
Quarry run
GC
6
33 20
20.2 33 10
6.5*10
km/s
E
Soil
USCS

kN/m
50
Filter
SP-SM
1c'
35
-8
-7
-8
3
50
Core
SC
18
30
19.8
10
-5
60
Rockfill
15 33
31 20.2
21.9 6.5*10
9.5*10
m/s
MPa
type
kN/m
30
Quarry
run
GC
6kPa
-5
-5 -7 -3
-8
50
Filter
SP-SM 15
11
35
20
10
60
Selected rockfill SC
35 21.9
20
5.1*10
Core
18
30
19.8
10
50
60
Rockfill
31
9.5*10
-5-6
-5 -3-5
30
Quarry
run
GC
6
33
20.2
6.5*10
Weathered
rock
50
20
20.5
1.4*10
Filter
SP-SM
1
35
20
10
50
Selected rockfill
1
35 20
5.1*10 -7-6 60300
-7
60
Rockfill
31
21.9
9.5*10
-6-5
104
Rock runrock
250 20
20 20.5
21.4 1.4*10
1.3*10
Quarry
GC
615
33
20.2
6.5*10
30
300
Weathered
50
-3
-3-9
-7
60
Selected
rockfill
1
35
20
5.1*10
-6
4
Soft
tailings
ML
1
29
13.2
5.1*10
0.52
Rockfill
15 20
31 21.4
21.9
9.5*10
60
10
Rock
250
1.3*10
-6
-6-9
300
Weathered
rock
50
20
20.5
1.4*10
-9-3
1.0
Medium
tailings
ML
1
32
19.7
5.1*10
Selected
rockfill
1
35
20
5.1*10
60
shaped
drains.
Band-shaped
drains
should
be
always
placed
Soft tailings
ML
1 drains
29 should
13.2 be 5.1*10
0.524
shaped
drains. Band-shaped
always-6-6-4 placed
10
Rock
250
20
21.4
1.3*10
-9-6
104
Las the
Vias
115
30 19.7
20
1.2*10
Weathered
rock thickness
50
20
20.5
1.4*10
300
when
thetailings
tailings
thicknessis
1532
m.
when
tailings
m.
1.0
Medium
ML
1is
5.1*10
-9
-9
4
Material
Soft
tailings
ML
1
29
13.2
5.1*10
0.52
-4-6
Rock
250
20
21.4
1.3*10
10
Las Vias
1
30 20
1.2*10 -9 10
1.0
Medium
tailings parameters
ML
32 13.2
19.7
5.1*10-9-9 0.52
Table
3.Calculation
Calculation
parameters
Soft
tailings
ML
11
29
5.1*10
Table
3.
Material
-4
-4
-9
Figure
4.tailings
Tailings thickness
reservoir
thickness
of Las
10
Las
30
20
k1.2*10
E
Soil Vias
USCS inside

and the
Medium
ML
32
k5.1*10
E1.0
Soil
USCS
c'11c' the
19.7
3
3
-4
Vias
material
that
will
be
placed
above.
Material
m/s
MPa
typeVias
kN/m
type
kPa
Las
1kPa
30 kN/m
20
1.2*10
10
Figure
4. Tailings thickness inside
the reservoir
and the m/s
thickness
ofMPa
Las
50
Corematerial that will
SCbe placed
18above.
30 19.8
19.8
10-8-8
Material
50
Core
SC
18
30
10
Vias
shaped
drains. Band-shaped
drains
shouldandbethealways
placed
-5-5
Figure
4. Tailings thickness
the35
reservoir
thickness
of
50Las
Filterthe
SP-SM inside
35 20
20
10
50
Filter
SP-SM
10
when
tailings
thickness
is1115
m.
-5
Vias
material
that
will
be
placed
above.
-5
30Las
Quarry
run
GC inside
33 20.2
20.2
6.5*10
Figure
4.run
Tailings thickness
reservoir
and the6.5*10
thickness of
30
Quarry
GC
66 the 33
60
Rockfill
15above.
31 21.9
21.9
9.5*10-7-7 60
Vias
that will
be placed
Rockfill
15
31
9.5*10
Table
3.material
Calculation
parameters
-3-3
Selected
rockfill
1
35
20
5.1*10
60
Selected
rockfill
1
35
20
5.1*10
k
E60
Soil
USCS
c'

300
Weatheredrock
rock
50
20 kN/m
20.5 3 m/s
1.4*10-6-6 MPa
300
Weathered
50
20
20.5
1.4*10
type
kPa
-6-6
-8
1044
Rock
250 30
20 19.8
21.4
1.3*10
10
Rock
250
20
21.4
1.3*10
50
Core
SC
18
10
-9
-9
-5
Softtailings
tailings
ML
29 20
13.2
5.1*10
0.52
Soft
ML
29
13.2
5.1*10
0.52
50
Filter
SP-SM
111
35
10
1.0
Medium
tailings GC
ML
32 20.2
19.7
5.1*10-5-9-9 30
1.0
Medium
tailings
ML
32
19.7
5.1*10
Quarry
run
611
33
6.5*10
10
LasVias
Vias
30 21.9
20
1.2*10-7-4-4 60
10
Las
11
30
20
1.2*10
Rockfill
15
31
9.5*10
Materialrockfill
Material
Selected
1
35 20
5.1*10-3 60
Weathered rock
50
20 20.5
1.4*10-6 300
-6
4Las
Figure
4.
Tailings
thickness
inside
the
reservoir
and
the
thickness
ofLas
Figure
4.
Tailings
thickness
inside
the
reservoir
and
the
thickness
of
10
Rock
250 20 21.4
1.3*10
Vias
materialthat
thatwill
will
beplaced
placed
above.
Vias
material
Soft tailings
MLbe
1 above.
29 13.2
5.1*10-9 0.52
Medium tailings ML
1
32 19.7
5.1*10-9 1.0
Las Vias
1
30 20
1.2*10-4 10
Material

Almagrera
tailings
has
height
of 35
35 m
m above
above
foundation
Almagrera
1 tailings
A or dam
Bdam has
1.4aa height
1.3of
1.2 foundation
axis.2The
The height
height
above
the
lowest
foundation
level
37.3 m
m
atat axis.
above
the
lowest
foundation
level
isis 37.3
C or D
1.3
1.2
1.1
(Figure3).
3).
(Figure
3
C or D
1.2
1.1
1.0
has an
an upstream
upstream sloping
sloping core.
core.The
The foundation
foundationisis formed
formed by
by
ItIt has
alternationof
ofvolcanic
volcanicand
andinter-stratified
inter-stratifiedsedimentary
sedimentaryrocks.
rocks.
4alternation
ALMAGRERA
TAILINGS
DAM
Figure
Central
cross-section
ofAlmagrera
Almagrera
dambefore
before
closure.
Figure
Central
cross-section
dam
closure.
type,3.3.and
was
raised 5of
times
adding
material
into the
downstream
side.raised
The slide
shows
the
dam foundation
axis
Almagrera
dam
has
heightadding
of original
35 mmaterial
above
type,
and tailings
was
5 atimes
intoandthethe
The of
dam
corresponds
to the
the
downstream
borrow
material
The
dam
corresponds
to
borrow
the
dam
at slide
the the
end
of downstream
3rd,
4th anddam
5th
heightenings.
atstates
axis.
The
height
above
lowest
level
ismaterial
37.3the
m
downstream
side.
The
shows
thefoundation
original
axis
and
The downstream
slope
(H):adding
1 (V)
to the
third
type,
raised
5 1.7
times
intophase,
the
(Figure
3).the was
states
ofand
dam at
the was
end
of
3rd,
4th
andupmaterial
5th
heightenings.
and
2 and
(H):
1side.
(V)
forsloping
the
two
phases.
According
tointo
a report
downstream
The
slide
shows
the
original
axis
and
the
Itdownstream
has
an upstream
core.
foundation
formed
by
The
slope
was
1.7
(H):The
1adding
(V)
up
to dam
theisthird
phase,
type,
was
raised
5last
times
material
the
delivered
the
5th
phase
construction,
theheightenings.
filter
criteria
the
end
of 3rd,
and 5th
states
of the
damfor
atthe
alternation
of
volcanic
and
inter-stratified
sedimentary
rocks.
and
2 (H):
1before
(V)
last
two
phases.
According
to
a
report
downstream
side.
The
slide
shows
the 4th
original
dam
axis
and
the
weredownstream
not
fulfilled.
way,
after
4thand
phase,
the
dam
rather
The
slope
(H):
14th
(V)
up
to
third
phase,
delivered
before
the
5th
phase
construction,
the
filter
criteria
Figure
3.of
Central
cross-section
of1.7
Almagrera
dam
before
closure.
states
the
dam
atThis
thewas
end
of
3rd,the
5ththe
heightenings.
th
an
behaved
as
a
homogeneous
dam.
During
thethe5th
and downstream
2not
(H):
1 (V) slope
for
last1.7
two
phases.
According
tophase,
arather
report
were
fulfilled.
Thisthe
way,
after
the1 4(V)phase,
dam
The
was
(H):
up to
third
phase,
th
thfilter
inclined
sand
and
gravel
filter
was
placed
between
the
4
and
delivered
before
the
5th
phase
construction,
the
criteria
Table 4 shows the maxima settlements in tailings obtained
an
behaved
as 1acorresponds
homogeneous
dam.
During
theborrow
5 phase,
The
dam
to the
downstream
and
2 (H):
(V)
for the last
two
phases.
According
tomaterial
a report
thcriteria. A downstream
thrather
5th phase
shells
filter
were
notsand
fulfilled.
This
way,
after
the
4th
phase,
dam
using
several
hypotheses.
inclined
and using
gravel
filter
was
placed
between
the
4
and
delivered
before
the
5thnormalised
phase
construction,
thethe
filter
criteria
Table
4 shows
the maxima
settlements
inthe
tailings
obtained
4. Tailings
thickness
inside the
reservoir and
thickness
of Las
th
th phase,
thfoot drain
th
placed
below
theDuring
inclined
filter
the Figure
an
behaved
as awas
homogeneous
dam.
theAthe
5downstream
5were
phase
using
normalised
filter
not shells
fulfilled.
This
way,
after
the
4criteria.
phase,
damand
rather
using
several
hypotheses.
Vias material that will be placed above.
th
th
th
downstream
shell,
protected
by
a
non-woven
geotextile.
Leaks
inclined
gravel below
filterdam.
was
between
and
Table
4. Maxima
settlements
in tailings.
Table
4 shows
the maxima
settlements in tailings obtained
foot
drainsand
placed
theplaced
inclined
and4 the
phase,
an
behaved
as was
a and
homogeneous
During
thefilter
5 the
3
th
th phasefrom
thslope.
1 up
to 16
mfilter
/h
incriteria.
the
downstream
5ranging
shells
using
normalised
filter
A downstream
Test
Tailings
Maximum
Remainder
using
hypotheses.
downstream
shell,
protected
byappeared
a non-woven
geotextile.
inclined
sand
and
gravel
was
placed
between
the 4Leaks
and
Table
4.several
Maxima
settlements
inFill
tailings.
Table
4
shows
the
maxima
settlements
in tailings
obtained
th
drain
below
the
filter
and
the
thickness
thickness
settlement
settlement
ranging
from
1was
upusing
toplaced
16 normalised
m3/h
appeared
ininclined
the downstream
slope.
5foot
phase
shells
filter
criteria.
A
downstream
using
several
hypotheses.
Test
Tailings
Fill
Maximum
Remainder
4.1
Simulation
of closure
operations
downstream
shell,placed
protected
by
a the
non-woven
geotextile.
(m)
(m)
(m)
after construction
Table 4. Maxima
settlements
in tailings.settlement
foot
was
below
inclined
filter into
andLeaks
the
thickness
thickness
settlement
type, drain
and
was
raised
times
adding
material
the
type,
and
was
adding
intoslope.
the
ranging
from
1 upraised
to
16 m5533/htimes
appeared
in thematerial
downstream
(m)construction
Test
Tailings
Fill
Maximum
Remainder
4.1
Simulation
of
closure
operations
downstream
shell,
protected
by
a
non-woven
geotextile.
Leaks
Table 4. Maxima
in tailings. (m)
A mechanical
for3 shows
all thethe
operations
involved
during
(m)settlements
(m)
after
downstream
side.model
The slide
slide
shows
the
original dam
dam
axis and
and
the
downstream
side.
The
original
axis
the
No
Drains
thickness Fill
thickness Maximum
settlement (m)
settlement
ranging
from
1
up
to
16
m
/h
appeared
in
the
downstream
slope.
closure
was
entrusted
to
the
main
author.
A
finite
element
Test
Tailings
Remainder
states
ofSimulation
the dam
dam
the
end
of
3rd,
4th and
and 5th
5th heightenings.
heightenings.
states
of
the
atat
4th
A
model
forend
all of
the3rd,
operations
involved
during
4.1mechanical
ofthe
closure
operations
drains
(m)
(m)
(m)
after
construction
model
was
set
up
to
reproduce
all
the
steps
that
are
being
taken
thickness
thickness
settlement
settlement
No
Drains
The
downstream
slope
was
1.7
(H):
1
(V)
up
to
the
third
phase,
The downstream
slope was
1.7 main
(H): 1author.
(V) up to
thirdelement
phase,
closure
was
entrusted
to the
A the
finite
Oedometer (m)
25
5.5
3.3
2.5 construction
0.42
(m)
4.1
Simulation
of the
closure
operations
during
closure:
(m)
(m)
after
drains
A
mechanical
thephases.
operations
during
and
(H):
(V)
for
lastalltwo
two
phases.
According
to aa report
report
and
22 (H):
11 (V)
for
thefor
last
to
model
was
set
upmodel
to
reproduce
all
the
stepsAccording
that involved
are being
taken
20
7
3.4
2.4
0.19
No
Drains
1.
Finding
the
initial
safety
factor.
(m)
Oedometer
25
5.5
3.3
2.5
0.42
closure
was
entrusted
to
the
main
author.
A
finite
element
delivered
before
the
5th
phase
construction,
the
filter
criteria
delivered
beforemodel
the 5th
construction,
the
filter criteria
during
closure:
A mechanical
forphase
all the
operations
involved
during
th
drains
Piezocones
25
5.5
2.5
1.9
0.32
th
2.
Upstream
water
drawdown
produced
by
pumping
the
No
Drains
model
was
setinitial
up to
reproduce
all
the
that
being
taken
were
not
fulfilled.
This
way,
after
theauthor.
4 phase,
phase,
the
dam
rather
20
7
3.4
2.4
0.19
were
not
fulfilled.
This
way,
after
the
4steps
the
dam
rather
1.
Finding
the
safety
closure
was
entrusted
to
the
main
A are
finite
element
type,
and
was
raised
5 factor.
times
adding
material
into
the
Soft tailings 25
Oedometer
25
5.5
3.3
2.5
0.42
20
7
2.6
2.6
0.15
thth phase,
contaminated
reservoir
water.
drains
during
closure:
an
behaved
as
a
homogeneous
dam.
During
the
5
Piezocones
5.5
2.5
1.9
0.32
phase,
an
behaved
as set
a water
homogeneous
dam.
the
5 axis
2.
Upstream
drawdown
by
pumping
the
model
was
up
to reproduce
allproduced
theDuring
steps that
are
being
downstream
side.
The
slidecoarse
shows
the
original
dam
andtaken
the
20
75.5
3.4
2.4
0.19
Piezocones
25 the maxima
1.3
1.0
0.16
3.Finding
Upstream
filling
with
waste
material.
Oedometer
25
5.5
3.3
2.5
0.42
inclined
sand
and
gravel
filter
was
placed
between the
the 44thth and
and
Soft
tailings
1.
the
initial
safety
factor.
20
7
2.6
2.6
0.15
inclined
sand
and
gravel
filter
was
placed
between
Table
4
shows
settlements
in
tailings
obtained
contaminated
reservoir
water.
during
closure:
Table 4 shows the maxima settlements in tailings obtained
states
of the
dam
at thedrawdown
end of 3rd,produced
4th
and 5th
heightenings.
th Long
th
Medium
Piezocones
25
5.5
2.5
1.9
0.32
20
7
1.3
0.9
0.07
4.
term
stability.
Calculation
of
tailings
settlements.
2.
Upstream
water
by
pumping
the
5
phase
shells
using
normalised
filter
criteria.
A
downstream
20
7
3.4
2.4
0.19
51.Upstream
phase
shells
using
normalised
criteria.
A downstream
usingseveral
several25
hypotheses.5.5
Piezocones
1.3
1.0
0.16
3.
filling
with
coarse
waste
Finding
the
initial
safety
factor.
using
hypotheses.
The
downstream
slope
was
1.7
(H):filter
1material.
(V)
up to the
third
phase,
tailings
Soft
tailings 25
20
7
2.6
2.6
0.15
5.Long
Seismic
calculation,
including
consideration
forand
tailings
contaminated
reservoir
water.
foot
drain
was
placed
below
the
inclined
filter
and
the
Piezocones
5.5
2.5
1.9
0.32
foot
drain
was
placed
below
the
inclined
filter
the
Medium
20
7
1.3
0.9
0.07
4.
term
stability.
Calculation
of
tailings
settlements.
2.
Upstream
water
drawdown
produced
by
pumping
the
and
2 (H): 1 (V)
forwith
the last
twowaste
phases.
According to a report
liquefaction.
Piezocones
25settlements75.5
1.3
1.0
0.16
3.
Upstream
filling
coarse
material.
downstream
shell,
protected
by aa non-woven
non-woven
geotextile.
Leaks
Soft
tailings
20
2.6
0.15
downstream
shell,
protected
by
geotextile.
Leaks
Table
Maxima
intailings.
tailings. 2.6
tailings
5.
Seismicbefore
calculation,
including
consideration
for tailings
contaminated
reservoir
water.
Table
4.4.Maxima
settlements in
delivered
the
5th
phase
the
filter
criteria
3 the construction,
3
4.3 Dam calculations
6.
If
necessary,
finding
new
safety
factors
with
the
Medium
20
7
1.3
0.9
0.07
4. Upstream
Longfrom
termfilling
stability.
of material.
tailings
settlements.
ranging
from
1up
upto
to
16Calculation
m
/happeared
appeared
in
thedownstream
downstream
slope.
ranging
1
16
m
/h
in
slope.
th the
Piezocones
25
5.5
1.3
1.0
0.16
Test
Tailings
Fill
Maximum
Remainder
liquefaction.
3.
with
coarse
waste
Test
Tailings
Fill
Maximum Remainder
were
not fulfilled.
Thisstep
way,
after theconsideration
4 phase, the for
dam tailings
rather
from
3.
tailings
5.reinforcement
Seismic
calculation,
including
th with
Medium
thickness
thickness
settlement
settlement
4.3
Dam
calculations
20
7
1.3
0.9
0.07
6.
If
necessary,
finding
the
new
safety
factors
the
4.
Long
term
stability.
Calculation
of
tailings
settlements.
thickness
thickness
settlement
settlement
phase, an
behaved
as 2D-9.02
a homogeneous
dam.
During
the the
5 calculations
According to(m)the inventory
damsafter
andconstruction
ponds in
Plaxis
program
has
been
used and
liquefaction.
4.1
Simulation
of
closure
operations
4.1
Simulation
of
closure
operations
th
tailings
(m) of tailings
(m)
reinforcement
fromgravel
step
3.filter
5.
Seismic
calculation,
including
consideration
for
(m) the maxima
(m)
(m)
after construction
inclined
sandcarried
and
was
placed
between
thewith
4tailings
and
Table
4
shows
settlements
in
tailings
obtained
extractive
industry
in
Spain
(ITC
2000)
it
is
clear
that
have
been
out
with
15-node
elements.
A
Mohr-Coulomb
4.3
Dam
calculations
6.
If
necessary,
finding
the
new
safety
factors
the
th Plaxis 2D-9.02 program has been used and the calculations
(m)
According to the inventory of tailings dams (m)
and ponds in
liquefaction.
5A
phase
shells
using
normalised
filter
criteria.
Aofdownstream
A
mechanical
model
for3.all
all the
thethis
operations
involved
during
using
several dam
hypotheses.
mechanical
model
for
operations
involved
during
Almagrera
is class
1 (height
greater
than
15
m)
but
only
materials
model
has
been
used;
issafety
a model
perfect,
nonreinforcement
from
step
extractive
industry
in
Spain
(ITC
2000)
it
is
clear
that
4.3
Dam
calculations
have
been
carried
out
with
15-node
elements.
A
Mohr-Coulomb
6.
If
necessary,
finding
the
new
factors
with
the
No
Drains
No
Drains
foot
drain
was
placed
the author.
inclined
and
the
closure
was
entrusted
tobelow
the
main
author.
finite
element
closure
was
entrusted
to
the
AAfilter
finite
element
category Cdam
(moderate
damage
only
incidentally
affecting
lives).
associated
plasticity
According
to
the
inventory
of
tailings
dams15
and
ponds
Plaxis
2D-9.02
program
hasmain
been
and
calculations
Almagrera
is class
1 (height
greater
than
m) but
onlyin
materials
model
hasto
been
used;
this
is used
asteps
model
ofthe
perfect,
nonreinforcement
from
step
3.
drains
drains
downstream
shell,
protected
by
a
non-woven
geotextile.
Leaks
model
was
set
up
reproduce
all
the
that
are
being
taken
Table
4.
Maxima
settlements
in
tailings.
model
was
set
up
to
reproduce
all
the
steps
that
are
being
taken
So,
the
first
row
of
the
safety
factors
in
Table
2
must
be
Table
3
shows
the
calculation
parameters.
extractive
industry
in
Spain
(ITC
2000)
it
is
clear
that
have
beenplasticity
carried out
with
elements.
A the
Mohr-Coulomb
3 15-node
category
C
(moderate
damage
only
incidentally
affecting
lives).
According
to
the
inventory
of
tailings
dams
and
ponds
in
associated
Plaxis
2D-9.02
program
has
been
used
and
calculations
Oedometer
25
5.5
3.3
2.5
0.42
Oedometer
25
5.5
3.3
2.5
0.42
ranging
from
1 uphas
to 16
m /h
appeared
ina the
downstream
slope.
during
closure:
during
closure:
Test
Tailings
Fill
Maximum
Remainder
accomplished.
Almagrera
dam
is
class
1
(height
greater
than
15
m)
but
only
materials
model
been
used;
this
is
model
of
perfect,
nonSo,
the firstindustry
row
the7Spain
factors
in Table
must
be
extractive
it2.4
is 2 clear
that
Table
shows
the
calculation
have
been3carried
outsafety
with 15-node
elements. A Mohr-Coulomb
20 of in
7 safety (ITC
3.4 2000)
2.4
0.19
20
3.4
0.19
1.4.2
Finding
the
initial
factor.parameters.
1.
Finding
the
initial
safety factor.
thickness
thickness
settlement
settlement
Compacted
reinforcement
was
to
fulfil
Tailings
category
C dam
(moderate
damage
only greater
incidentally
affecting
lives).
associated
plasticity
accomplished.
Almagrera
isrockfill
class
15.5
(height
thanprojected
15
m) but
only
materials
model
has
been
used;
thisproduced
is a model
of pumping
perfect, nonPiezocones
25
2.5
1.9
0.32
4.1
Simulation
of
closure
operations
Piezocones
25
5.5
2.5
1.9
0.32
2.
Upstream
water
drawdown
by
the
2. Upstream
water
produced by pumping the
(m)
(m)
(m)
after construction
withthetheC
safety
specified
by was
theinSpanish
Regulations
So,
first
rowfactors
of the
safety
Table
2 to
must
be
Table 3plasticity
shows
the drawdown
calculation parameters.
reinforcement
projected
fulfil
category
onlyfactors
incidentally
affecting
lives).
4.2
Tailings
associated
SoftCompacted
tailings (moderate
20rockfilldamage
2.6
2.6
0.15
Soft
tailings
contaminated
reservoir
water. thickness inside the reservoir and
20
77
2.6
2.6
0.15
contaminated
reservoir
water.
(m)
(Figure
5).
Figure
4 3shows
the
tailings
accomplished.
A
mechanical
model
for
all the parameters.
operations involved during
with
the
safety
factors
specified
by
the
Spanish
Regulations
So,
the
first
row
of
the
safety
factors
in
Table
2
must
be
Table
shows
the
calculation
Piezocones
25
5.5
1.3
1.0
0.16
3.
Upstream
filling
with
coarse
waste material.
Piezocones
25
1.3
1.0
0.16
3.
Upstream
filling
with
coarse
waste
No
the
thickness
oftheLas
Vias
material
that
willA
placed
above.
Compacted
rockfill 5.5
reinforcement
was projected
toDrains
fulfil
4.2
Tailings
closure
entrusted
to
thethickness
main material.
author.
finite
element
(Figure
5).
accomplished.
Figure
4was
shows
tailings
inside
thebe
reservoir
and
Medium
20
7
1.3
0.9
0.07
4.The
Long
term
stability.
Calculation
of
tailings
settlements.
Medium
20
7
1.3
0.9
0.07
4.
Long
term
stability.
Calculation
of
tailings
settlements.
drains
Figure
5.
Reinforcement
needed
to
fulfil
the
safety
factors
of
Table
2.
hatched
area
is
the
area
that
should
be
treated
with
bandwithCompacted
the safety rockfill
factors reinforcement
specified by the
Spanish
Regulations
model
was setofupLas
to reproduce
all thethat
steps
thatbeare
beingabove.
taken
was
projected
to
fulfil
the
thickness
Vias
material
will
placed
4.2
Tailings
tailings
5.
Seismic
calculation,
including
consideration
for tailings
tailings
tailings
5.
Seismic
calculation,
including
consideration
for
Oedometer
25
5.5
3.3
2.5
0.42
(Figure
5).
Figure
4
shows
the
tailings
thickness
inside
the
reservoir
and
during
closure:
with 5.
theReinforcement
safety factors
specified
the Spanish
Figure
needed
to fulfil by
the safety
factors ofRegulations
Table 2.
The
hatched
area is the area that should be treated with bandliquefaction.
liquefaction.
20
7
3.4
2.4
0.19
the
thickness
of the
Lassafety
Viasfactor.
material
that
will the
be placed
above.
1.
Finding
the initial
(Figure
5).calculations
Figure
4 shows
tailings
thickness
inside
reservoir
and
4.3
Dam
6.
If
necessary,
finding
the
new
safety
factors
with
the
4.3
Dam
calculations
6.
Ifhatched
necessary,
finding
thethat
new
safety
factors
withbandthe
Figure
5. Reinforcement
needed
the safety factors
2.
Piezocones
25
5.5 to fulfil2.5
1.9 of Table
0.32
TheUpstream
area
is
the
area
should
be by
treated
with
2.
water
drawdown
produced
pumping
the
the
thickness
of
Las
Vias
material
that
will
be
placed
above.
reinforcement from
fromstep
step3.
3.
reinforcement
Soft
tailings
20
7
2.6
2.6 of Table
0.15
contaminated
reservoir
water.
Figure
5.
Reinforcement
needed
to
fulfil
the
safety
factors
2.
ThePlaxis
hatched
area
is
the
area
that
should
be
treated
with
bandAccording to
to the
the inventory
inventory of
of tailings
tailings dams
dams and
and ponds
ponds in
in
2D-9.02 program
program has
has been
been used
used and
and the
the calculations
calculations
According
Plaxis 2D-9.02
Piezocones
25
5.5
1.3
1.0
0.16
3. Upstream
filling with coarse
waste material.
extractive industry
industry in
in Spain
Spain (ITC
(ITC 2000)
2000) itit isis clear
clear that
that
have been
been carried
carried out
out with
with15-node
15-node elements.
elements. AA Mohr-Coulomb
Mohr-Coulomb
extractive
have
Medium
20
7
1.3
0.9
0.07
4. Long term stability. Calculation of tailings settlements.
Almagrera dam
dam isis class
class 11 (height
(height greater
greater than
than 15
15 m)
m) but
but only
only
materials model
model has
has been
been used;
used; this
this isis aa model
model of
of perfect,
perfect, nonnonAlmagrera
materials
tailings
5.
Seismic calculation,
including
consideration
for tailings
categoryCC(moderate
(moderatedamage
damageonly
onlyincidentally
incidentallyaffecting
affectinglives).
lives).
associatedplasticity
plasticity
1512
category
associated
liquefaction.
So, the
the first row
row of
of the
the safety
safety factors
factors in
in Table
Table 22 must
must be
be
Table33shows
showsthe
thecalculation
calculationparameters.
parameters.
So,
4.3
Dam first
calculations
6. IfTable
necessary, finding
the new safety factors with the
accomplished.
accomplished.

Technical Committee 203 / Comit technique 203

Table 5 shows the results of the calculations with the FE

method.
Table 5. Final displacements and safety factors with the FE method.

Phase
Initial
Reservoir at 207.5 level
Drawdown
Fill
Long term
Earthquake & liquefaction

max (mm)

193
1300
2060
278

Safety factor
1.44
1.43
1.47
1.36
1.79
1.19 (inside tailings)

punctual seismic sources. To solve this problem, the

seismogenic areas are divided into elements small enough to be
assimilated to punctual seismic sources.
With this procedure, uniform seismic hazard response
spectra can be created for the site according with the period of
exposure, the probability of exceeding the design spectrum and
the type of soil. Then real design accelerograms can be selected
following these steps:
-The period of exposure of structure is established according
to its estimated lifetime.
-The probability of exceeding this level is established
according to the seismic hazard required. In this case, as the
return period is established (1,000 years), the probability is
calculated using equations (3) and (4).
For the type of soil at the site, the uniform seismic hazard
response spectrum is calculated according to the required
seismic hazard level.
From the database of accelerograms, those recorded at the
same type of soil are selected.
The standard deviation is defined as:

S log( f .S R log S C )] 2 / n

DYNAMIC CALCULATION

The NCSR 2002 considers that large dams (height greater than
15 m) are constructions of special importance, which should be
calculated for a return period of 1,000 years. The seismic
acceleration for pseudo-static calculation was 0.08g. The
reinforcement has been calculated using this acceleration. The
ITC (2000) required a safety factor of 1.3.
A dynamic calculation has being carried out. A method to select
accelerograms (Morales-Esteban et al., 2012), for the closure
calculation, has been developed based on uniform seismic
hazard response spectra and is detailed in the paper.
The arrival of earthquakes to the site that exceed a reference
value logS0 is modelled as a Poisson stationary process, defined
according to the Gutenberg-Richter law. The seismogenic areas
defined by Martin (1984) were used.
Next, the uniform seismic hazard response spectra are
calculated. The arrival of earthquakes to the site that exceed a
reference value log S0 is modelled as a Poisson stationary
process, defined according to the Gutenberg-Richter law.
For source i, the average number of events per year is:

i i

M max

M min

P (log S log S 0 / M , D) f ( M )dm

f log S C log S R ) / n

(6)

Here SR are the values of the response spectrum

corresponding to the real register, SC are the values of the
calculated response spectrum, n is the number of intervals
considered in the calculation and f is the scale factor that
minimizes the standard deviation.
This methodology has been applied to the site of Almagrera
Dam, founded on rock, for a damping ratio of 5%, considering a
return period close to 1000 years. The accelerograms have been
selected from the European Strong Motion Database that can be
obtained from Internet at http://www.isesd.cv.ic.ac.uk/. Figure 5
shows the comparison between the response spectrum
calculated for Almagrera dam, in rock, for 1000 years return
period, and the response spectrum from one of the selected
accelerogram.

(1)

Uniform seismic hazard response spectrum (clear line).

Spectrum for accelerogram 7488 (dark line).

(2)

The return period, T, is the average time interval between

events and its value is:
T = 1/
(3)
The probability of exceeding the reference value log S0
during the period of time, t, owing to the simultaneous action of
N individual seismic sources is:
P(logs log S0;t) =1-e-t

(5)

where:

where:
i is the seismic rate of earthquakes of the individual source.
P is the probability for logS to exceed the reference value,
logS0, for an earthquake of magnitude M that occurs at a
distance D from the site.
f(M) is the magnitude probability density function between the
minimum and maximum magnitudes considered.
If N individual seismic sources act simultaneously, the rate
of arrivals at the site of earthquakes that exceed the reference
value, logS0, is:
=i

1 / 2

Figure 5. Comparison between the response spectrum calculated for

Almagrera dam, in rock, for 1000 years return period, and the response
spectrum from the selected accelerogram.

From the database, the accelerograms that have a scale factor

near to 1, and a smaller standard deviation are finally selected
(Table 6)

(4)

Equation (1) cannot be applied to the hazard calculation as

the seismogenic areas have been modeled as areas and not as

1513

Proceedings of the 18th International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013

of band-shaped drains, inside the tailings, after the fill has been
placed on them.
3. A method to select accelerograms for the closure
calculation has been developed based on uniform seismic
hazard response spectra and is detailed in the paper.
4. Dynamic calculation has shown that relative
displacements are small and that no surface of rupture appears.
The model has indicated the necessity to place rockfill
reinforcement downstream.
As it may be observed, the calculated dynamic
displacements are quite small. Further calculations are being
carried out with larger return periods, as indicated by standards
specifically related to tailings dams.

Table 6. Selected accelerograms, scale factor and standard deviation.

Accelerogram
128
361
365
608
990
5826
6269
6270
6331
7480

f
0.991
1.016
0.931
1.099
1.006
0.931
1.095
1.014
0.969
1.033

s
0.016
0.028
0.130
0.172
0.011
0.129
0.165
0.026
0.057
0.058

Once the accelerograms have been selected, they are not

scaled to be introduced in the dynamic calculation. One of these
accelerograms selected is shown in Figure 6.

Figure 8. Total displacement versus time for the selected points during 5
seconds of calculation with accelerogram 608. Selected points: A on the
base, B over the dam and C at the left side.

Figure 6. Accelerogram No. 128. Scale factor for all periods 0.99.
From the dynamic calculation the acceleration versus time of any
point of the dam subjected to the accelerograms selected can be
obtained as plot in figure 7. Similar is figure 8 where the displacement
is shown versus time. Absolute displacement can be obtained by
subtracting the displacement of the point to the displacement of point A
(placed on the basement). Finally, in the figures corresponding to the
relative shear stresses, obtained from Plaxis output, plastic zones appear
in the downstream slope, in the tailings and elsewhere. It can be
observed that no continuous surface of rupture appears.

This work has been financed by the Spanish Ministry of Science

and Innovation (Project BIA 201020377).
8

REFERENCES

Davies M.P. 2002. Tailings impoundments failures: Are geotechnical

engineers listening? Geotechnical News, September: 31-36.
Davies M.P. and Martin T.E. 2000. Mine tailings dams: when things go
wrong. Tailings dams 2000, Proceedings ASDSO Conference, Las
Vegas.
ITC 2000. Instruccin Tcnica Complementaria 08.02.01 del captulo
XII del Reglamento General de Normas Bsicas de Seguridad
Minera Depsitos de lodos en procesos de tratamiento de
industrias extractivas. BOE de 9/6/2000, 17235-17244. Ministery
of Industry & Energy (Spain).
Martn A.J. 1984. Riesgo ssmico en la Pennsula Ibrica. Tesis
doctoral. Ph. D. Thesis, Politechnical University of Madrid.
Morales-Esteban A. de Justo J.L. Martnez-lvarez F. and Azan J.M.
2012. Probabilistic method to select calculation accelerograms
based on uniform seismic hazard acceleration response spectra. Soil
Dynamics and Earthquake Engineering, 43, 174-185.
NCSR 2002. Norma de Construccin Sismorresistente NCSR-02. Parte
General y Edificacin. BOE, 244: 35898-35967.

Figure 7. Total acceleration versus time during 5 seconds of calculation

with accelerogram 608. Selected points: B over the dam and C at the left
side.

ACKNOWLEDGEMENTS

CONCLUSIONS

A model, reproducing all the closure operations, has been

prepared, and results presented herein. Monitoring has been
provided that will verify and improve the model
presented. Several closure aspects have been examined:
1. The placement, above the tailings, of coarse mineral
residues, thereby reducing the volume of mine dumps.
2. Methods to speed up settlements before placing the cover
and, this way prevent damage to it; for example the placement

1514