Geomechanics of Failures

E. Alonso, N. Pinyol, UPC, Barcelona

A.M. Puzrin, IGT/ETH Zurich

Dynamics of Dam Sliding:

Aznalcóllar Dam, Spain
2. The geometry of the failure
Central zone of failure. Displaced and reconstituted profiles

(Moya, 2000)
Main ideas
The solid rigid motion of the dyke provides an opportunity for a simple
analysis
The total displacement of the dyke (known by field observations) allows
calibration of the model
Unknown aspects (velocity, acceleration) may be derived
Detail of the upper part
Mud volcanoes observed a few hours after the failure on the surface of the depressed
basin, upstream of the slid dam (Courtesy of J. M. Rodriguez Ortiz).
  h 
V0  0.5h0  h0 tan  90  st   0  ,
 tan 1 

 s tan  b 
Va   e1   s,
 2 
 tan b 
Vb  s 2 tan b 1  ,
 2 tan 1

Vc   h m  f  h  n  ,
sb
1 1  
m   tan   st   ,
2  tan 1 2 

s tan b   s  sb cos  b .
f  2  tan b tan   st   ,
2 tan 1 2 
V0  Va  Vb  Vc h(s)
n  s tan b .
 Horizontal thrust
of tailings

Active conditions Liquefied conditions

1 1
Fhi  Fhi  U hi  K a  t ( H  e1 ) 2   w ( H  e1 ) 2 , Fh  0.5 t ( h  e1 ) 2 ,
2 2

Evolution of tailings thrust

against the moving dam and
accompanying foundation
soil, as the displacement sb
increases.

 : Model parameter
 Pore pressures during motion

Rapid displacement
of the dyke induces an
undrained loading
Central and
downstream parts: Pwp
controlled by dyke
weight
Upstream: Pwp
controlled by height of
tailings
 Distribution of pore water pressures on the
68  e1 (m of sliding plane for a dam displacement, s (pressure
u1  e1  h( s ) water
27 heads in m)
head).
u2  u3  eR  38.6

u4  e2  10.7 e2+10.7

(26  e3 )
u5  e3  s  e3  0.16 s
55
e30  e2
e3  s   e30  s
55

1 (68  e1 )h 
U1  e   38.6  e   w 48.7,
2 
1 R
27 
U 2  (38.6  eR )  w 26,
1
U3  38.6  eR  10.7  e2   w 33.3,
2
1
U4   w 10.7  e2  e3  0.16 s  (55  s),
2
 The geometry of the motion

Kinematics of Aznalcóllar slide: (a) the diagram shows the position of the main soil slab
(A) under the dam (not shown) and the passive wedge (B) at the initial time, t = 0, and at
a later time t; (b) compatibility of displacements at Point P
e  e  L tan  b  L tan  b 
Initial volume of Wedge A: VA 0  1 1 L  L  e1  
2  2 
L is the total horizontal length of the main slide (L = 108 + 55 = 163 m )

 L  sb cos  b 
VA1   L  sb cos  b   e1 
Once the Wedge A has displaced sb, its
tan  b  .
volume is given by :  2 

Therefore the variation of the volume of the wedge can  s cos  b 

VA  sb cos  b  e1  L  b tan  b 
be expressed as a function of the displacement by:  2 
Equilibrium of passive wedge

1  w e320
U hp   w e320 , Up  .
2 2 sin   e 

e320   s cos  b 
W p   soil  sb cos  b  e1  tan  b  L  b 
2 tan  e   2 

d ds  dM p ds d2 s
Dynamic equilibrium equation: Mp    M p 2   Fs
dt  dt  d t dt dt

This equation will be applied in the direction of the motion of the wedge (displacement sp)
and in the normal direction to sp (no displacements in this direction):

d ds p  cos  b
Fhp  cos  e  sin  e tan b   U hp cos  e  N p tan b  W p sin  e   p
M  sp  s b.
dt  dt  cos  e

Fhp sin  e  cos  e tan b   U hp sin  e  Wp cos  e  U p  N p  0

 Dynamic equilibrium of the main wedge and dam

h2 t
Fv  .
2 tan 1

Fh cos  b  Fv sin  b  Wd sin  b  Wss sin  b  N  tan b 

Equilibrium in direction parallel
d  dsb 
to sb Fhp cos  b  U hp cos  b  Svp sin b 
dt  dt 
M

M  M d  M ss  M d   sV1

 Fh sin  b  Fv cos  b  Wd cos  b  Wss cos  b 

Equilibrium in direction normal
 F 'hp sin  b  U hp sin  b  F 'hp tan b cos  b  N   U  0
to sb
  f  sb  t  , v  t  
Solution dvb
a
dt

vnew  vold  t f  sbold , vold 

d  dsb  d  dsb 
 M  sb    m4  M p  sb    F*
dt  dt  dt  dt 

F *  Fh  sb  m1   Fv  sb   Wd  Wss  sb   m2 
U hp  m7  m1   U p m5  W p  s p  m6  U  sb  tan b .

m1  cos b  sin b t4 ; m2  sin b  cos b t 4 ; m3  m1  m2 t4 ;

m4  m3 t1 ; m5  t4 m3 t1 ; m6  t3 m3 t1 ; m7  t2 m3 t1 .

t1  cos  e  sin  e tan b  tan b  cos  e tan b  sin  e  ;
t2  cos  e  sin  e tan b
t3  sin  e  cos  e tan b ;
t4  tan b .

 dM cos  b dM p 
F *  vb   m4  F*: Unbalanced forces
dvb  dt cos  e dt 
 .
dt cos  b
M  m4 Mp
cos  e
Reduction of average friction angle on basal plane as a
function of sliding distance sb
 Results
(b= 18.09º; = 11º;  = 20 cm; t = 31 kN/m3; Fhi = 11.3 MN/m;  = 1 m;
st = 70º; eR = 13.40 m; b = 2º; e = 20º)

Computed evolution of dam displacement

 Computed slide velocity

Computed slide acceleration

 Tailings
liquefaction

Decrease in height
of tailings

Computed variation of forces against the moving mass

Evolution of height of liquefied
tailings, acting against the
upstream end of the slide.
Model parameters (*): Very small influence on results
Blue clay
’initial: Mean friction angle on the failure surface. It is around 18º. (*)
’res: Residual friction angle: Varies between 10º and 12º.
: Necessary distance to mobilize the residual friction angle. Relative
displacements of several decimetres are required. (*)

Tailings
: Displacement to get liquefied tailings (1 m) (*)
e: Natural specific weight of liquefied tailings. (31 kN/m3).
Fhi: Initial horizontal thrust mobilized against the dyke (and the upper slice
of clay). It is estimated at 14000 kN/m if K0 =0.5, and at 11000 kN/m if active
conditions prevail. (*)
Model parameters
Geometry
st: Dip of the upstream scar within the tailings deposit (70º à 90º)
eR: Depth of failure surface under the center of the dyke, (14-15 m)
b: Apparent slope of failure surface (2º)
s: Exit angle of failure surface (20º)

Numerical
t: Time increment for iterative calculation. Negligible error if t < 0.1 s
Results of the sensitivity analysis in terms of the length travelled by the dam

Parameter/Property Value Travelled distance (m)

2 51.5 (Base case)
Slope of sedimentation planes, b (º)
1 53.3
80 48.7
Dip of upstream scar, st (º)
90 46.0
Residual friction angle of clay, 10 54.9
res (º) 12 48.2
21 50.5
Exit angle of passive wedge, e (º)
26 46.6
11 57.5
Depth of sliding surface under the
12 54.8
center of dam base, eR (m)
13 52.40
29 48.78
Specific weight of tailings, t (kN/m3)
30 50.15
Initial horizontal thrust of tailings, Fhi 16.9 51.8
(MN/m) 22.6 52.0
Necessary forward displacement of the 0.1 52.0
dam to generate liquefaction
conditions, (m) 0.2 51.0

Initial friction angle,  initial (º) 17 52

15 52
Relative displacement needed to 0.1 52
mobilize the residual friction angle, 
(m) 0.5 51.9
9. Dynamics of failure
Results
Summary of forces: Thrust; Base resistance; Passive wedge
Tailings Decrease in height
liquefaction of tailings
Conclusions of dynamic analysis

The ability of the model to reproduce the dyke displacement gives

confidence to hypothesis made and parameters selected
The motion was fast:
Total time: 15 sec
Maximum speed: 20 km/h
Maximum acceleration (0.17 g) was fast attained. Instability and
level reduction of liquefied tailings upstream
The motion stopped because the height of liquefied tailings
decreased. Passive wedge played a marginal role.