Beruflich Dokumente
Kultur Dokumente
Pastor
P.Mira, J.A.Fernndez Merodo, M.Martn Stickle, P.Dutto, T.Blanc, B.Haddad, S.Bentez, S.Sancho, V.Cullar, V.Drempetic
manuel.pastor@upm.es ETS de Ingenieros de Caminos Madrid
Mining dumps
Aberfan Cougar Hill
Debris Flows
Prof.Sucheng Zhang (Chengdus Institute for Natural Hazards in Mountain Areas, CAS)
Flow slides
(Aberfan, 1966)
Key Aspects
How to asses safety?
Stress reduction, ...and Failure load Failure mechanism
Causes of failure
Total stress (earthquakes, human, waves,...) Pore pressures (rain...) Strength ( degradation, weathering, chemical attack) Geommetry (erosion, reshaping...)
Consequences of failure
Toolboxes
Mathematical models Constitutive / Rheological models Numerical models
Objectives of talk
Describe fast catastrophic landslides & diffuse failure
- Initiation: Mathematical models Constitutive models Numerical models - Propagation: Mathematical models Rheological models Numerical models Present:
1 A common modelling framework 2 Closing the gap between rheo and eco 3 Towards a numerical model for all seasons
- Is it possible to develop a general model describing both? - Difficulties Mathematical Rheo & Constitutive Numerical
Contents
Introduction
(1) Fast catstrophic landslides: avalanches, flow slides, debris flows, lahars (2) Triggering mechanisms: Some basic ideas from Soil Mechanics Modelling of liquefaction
Mathematical Models
M1 Mixture Theory M2 Biot Zienkiewicz M3 Propagation consolidation M4 Depth Integrated
Constitutive Models Rheological Models Depth Integrated Rheological Models Numerical Models Benchmarks and applications
I . Introduction
But
What if failure is of diffuse type? Stress inside failure surface
So?
But
Data from Lizcano, Herrera & Santamarina (2007)
Mixture Theory
Constitutive
Rheological
Biot-Zienkiewicz
PropagationConsolidation
Mechanical + seepage
Failure Crit.
Depth Integrated
Limit analysis
M1 Mixture
s , pw , pa
vs , vw , va
Equations
ds , dw , da
Balance of mass (water, air, mixture) Balance of momentum Constitutive or rheological Relations velocities rate of deformations
M1 General Model Good for Debris Flows with large relative displacements between phases Inconvenients: Very expensive! Track interfaces between all phases Compromise solution: Just one single-phase (mixture) Still: Track interfaces between all phases (Eulerian: FEM + Level Set) (Lagrangian: SPH)
M2 Biot
M2 Biot-Zienkiewicz equations
v w = v s + vw / s
vw / s w = (Darcy) n
Velocities of pore fluids relative to solid skeleton (small) Skeleton (lagrangian) Pore fluids (eulerian relative to skeleton) Unknowns (Saturated) Equations :
B1 Balance of momentum (mixture) B2 Balance of mass (fluid) B3 Balance of momentum (fluid) B4 Constitutive (solid and fluid) B5 Kinematics
vs , w
, pw
B2+B3 B4 B5
d ' = D '.d
1 ui u j + ij = 2 x j xi
d v 0 extra dilatancy
Unknowns:
vs
, pw
M3 Prop + cons
Propagation
consolidation
dpw d pw1 = + cv + Em dV 01 dt dt x3 x3
Unknowns: v0 ,( , d ) and pw
M4 depth integrated
M4 Depth Integrated Models Unknowns: v0 , ( , d ) and pw Problems: Interfaces (or free surface)
h h v1 = v1dx3
h pw = pw dx3
Unknowns:
x3
h
Z
x1
Depth changes:
Mesh changes too Total stress and Pwp change
t=0s 10 m
t = 0.5 s
t = 1.5 s
t=0s
t=6s
t = 10 s
t = 15 s
t = 20 s
t = 30 s
3a. Constitutive
Contents
Introduction Mathematical Models Constitutive Models
Basic Generalized Plasticity Model Debonding State parameter Unsaturated
Diffuse
Liquefied Zone
q/p
Generalized Plasticity
Basic GP Model Extension I. Collapsible soils (J.A.Fdez Merodo) Extension II. State Parameter (D.Manzanal) Extension III. Unsaturated (R.Tamagnini, JA Fernndez Merodo) (D.Manzanal)
Generalized Plasticity
Introduce a direction n such that
d = CL : d d = CU : d
n : d > 0 n : d = 0 n : d < 0
1 CU = C + ngU n HU
e
p' = 392 KPa 800 Test a b c d Dr 29 % 44 % 47% 66% (d) 400 (b) 200 (a) (c)
600.
200
400
600.
u-pw model
u-pw model
t= 2.s
t= 8.s
t= 4.s
t= 11.4s
u-pw model
t= 2.s
t= 8.s
t= 4.s
t= 11.4s
u-pw model
t= 2.s
t= 8.s
t= 4.s
t= 11.4s
u-pw model
Pt (Pa)
Time (s)
u-pw model
Ux (m)
Time (s)
4. Rheological
Contents
Introduction
Fast catstrophic landslides: avalanches, flow slides, debris flows, lahars
R1 What is rheology?
In a fluid, shear stress depends on rate of shear strain z
(newtonian)
v = z
x
v = ( v, 0, 0 )
v = v(z)
x3
x2 x1
Rheological laws
= f s, z
v = Y + x 3
m
v depends only on x3
Bagnolds rheometer
B
0-B Constant pressure: Voids ratio increases with strain rate
Undrained path
Ln p
e
eref , dyn eCSL ,dyn
(e, p )
dv 0 = 2
CSLdyn
CSL
ln p '
eref
eCSL
pref
h v
h v
v( z)
gh sin
= g ( h z ) sin
v = g ( h z ) sin = f s, z
Integrate
z
Plug
hP
P
hP
Shear zone
B = gh sin
g sin hP = Y
hS
B
g sin Y
hS x
y
hP =
hS = h hP
Shear zone
Velocity Profile
Velocity Profile
x E 1.5 e7 Pa Poiss 0.3 Dens 1500 Kg/m3 Mg 1.1 Lambda 0.51 k 0.09 Pc0 0.285 e5 Pa gamma 0.1 delta 1. Slope 1:4 Shear zone
Velocity Profile
Stress
4. Numerical
Contents
Introduction
(1) Fast catastrophic landslides: avalanches, flow slides, debris flows, lahars (2) Triggering mechanisms: Some basic ideas from Soil Mechanics Modelling of liquefaction
Mathematical Models Rheological Models Depth Integrated Rheological Models Numerical Models
M1 Stress-velocity models: Taylor-Galerkin, Runge Kutta M2 Stress-velocity-pore pressure: Fractional step M3 Taylor SPH M4 Depth Integrated SPH
4a Classical u-pw
d2 u M 2 + BT . d Qp w f u = 0 dt dp w T du Q + H.p w +C. fp = 0 dt dt
v(t)
v(t)
to
4b sigma v TG
u 2 u =c 2 2 t x
2 2
v = t x v =E t x
0 + t v 1/ E 0 = 0 x v 0
F + =0 t x
Linear triangles (2D) or tetrahedra (3D) Equal order of interpolation Faster codes Extremely good performance in bending Robusts in plasticity Require stabilization when material is incompressible (Babuska-Brezzi conditions)
4c FS sigma v TG
v = grad p + b t div v = 0
v* v n =b t v n +1 v* = grad p n +1 t div v n +1 = 0
div v =
*
2 p n +1
v* v n =b t t 2 n +1 * div v = p
v n +1 v* = grad p n +1 t
4d FS sigma v T SPH
Diracs Delta
x = 0 ( x) = 0 x > 0
( x ) dx = 1
[ ] = ( 0 )
I
J
kh
Numerical Integration
N
I = ( xI ) h = ( x J ) W ( x J xI , h ) J
J =1
kh
J
Nh
I = ( x J ) W ( x J xI , h ) J
J =1
Lahars: Popocatepl
Ospedaletto
Civitella
Pesche
Ospedaletto
t= 32s t= 32s t= 64s t= 64s
t= 232 s
Tuostolo
Tuostolo
Topo mesh: 208.800 elements 105.083 nodes
Tailings dams
11 m
110 m
300 m
Bingham
t=0s
t = 90 s
t = 30 s
t = 120 s
t = 60 s
t=0s
t = 90 s
t = 30 s t = 120 s
t = 60 s
Methodology
Run PFC3D up to the instant of entering the water Transform the 3D output of DEM into depth integrated magnitudes (height and velocities) Run the SPH solid avalanche water code
t=0s
t=6s
t = 16 s
t = 20 s
t = 30 s
t = 42 s
R-1
Fracking
Grain crushing
Grain crushing
The 2000 Tsing Shan Debris Flow Hong Kong 14th April 2000
- The debris flow occurred : - after a cumulative rainfall of 160 mm - with a travel angle of about 24 - on a vegetated slope of 40 formed by colluvial boulders - There was erosion : the initial volume was about 150 m3 and the final volume was about 1 600 m3
Model Predictions
25 v2 b = p ' tan + CF 2 4 h
tan = 0.18
CF = 0.00133 Pa.s 2
Erosion (Hungrs law) : - Initial volume : 150 m3 - erosion coefficient : 0.0082 - Final volume : 1 550 m3
Depth evolution
t=5 t = 10
t = 15
t = 22
t = 36
t = 81
pw1 ( x1 , x2 , x3 , t ) dV 01 ( x1 , x2 , x3 , t )
pw1 = pw1 ( x1 , x2 , x3 , t )
(2k 1) N k ( x3 ) = cos ( x3 z ) 2h
k =1
pk ( x1 , x2 , t ).N k ( x3 )
( x3 z ) N1 ( x3 ) = cos 2 h
x3
N1 ( x3 )
pw1 ( x1 , x2 , x3 , t ) dV 01 ( x1 , x2 , x3 , t )
( x3 z ) dV 01 D1 ( x1 , x2 , t ) cos 2 h
( x3 z ) pw1 P 1 ( x1 , x2 , t ) cos h 2
dP P P P 1 = 1 + v1 1 + v2 1 t x1 x2 dt
D1 depends on rheology
Toyoura sand
CU Dr = 63,7%
4000 3500
1200
CU Dr = 18,5%
Toyoura Sand (Dr = 18.5%) ensayo prediccin
Tensin Desviadora, q
ensayo prediccin
Tensin Desviadora, q
1000
800
600
400
200
500
1000
1500
2000
2500
3000
3500
500
1000
1500
2000
2500
3000
3500
Tensin Desviadora, q
3000 2500 2000 1500 1000 500 0 0 0,05 0,1 0,15 0,2 0,25 0,3
Tensin Desviadora, q
1000
800
600
400
200
Deformacin Axial
Deformacin Axial
Debonding
Reduction of yield surface size q Lagioia & Nova, 1995
Validation
Lagioia & Nova, 1995
Oedometric test
A1
200 150 100 50 0 0 50 100 150 200 250 300 350 Tensin principal efectiva, p" [kPa]
s s s s s =0 = 100kPa = 300kPa = 600kPa =1200kPa
A0
Pressure
: linear concentration
' = De : d dvp
d vp = m ( F )
F F0 (F) = F0
F
F0
= ' pwI
D = + . . Dt
' = De : d dvp
d vp = m ( F )
(F) =
F F0 F 0
N
n +1
F = s t x
n+ 1 2
= + t t
n
t
n
n+
1 2
F =s x
n+
1 2
Step 1
n+
1 2
t t F n = + = + s x 2 t 2
n
1 n+ 2
n+
1 2
s
1 2
n+
1 2
Step 2
n +1
F = + t s x
n
n+
Example: 1D Bar
v(t)
v(t)
to
1D Elastic bar
( Newmark 0.5 0.0 )
1D Elastic bar
( Taylor galerkin )
1,000,000 800,000 600,000 400,000 200,000 0 -200,000 0 0.002 0.004 0.006 0.008 0.01 0.012
Stress (N/m2)
0.002
0.004
0.01
0.012
time (s)
Example: 2D Localization
R1
R2
R3
(a)
(b)
Why SPH?
Alternative methods Finite differences Finite elements Finite Volumes Meshless: SPH (Smoothed Particle Hydrodynamics) Grid based Problems with bodies of varying geommetry (Fluids with free surface, avalanches) VOF Level Set Grid Based
Kernel properties
lim W ( x ' x, h) = ( x )
h 0
W ( x ' x, h ) dx ' = 1
W ( x ' x, h) = 0 if
x ' x kh
E = 8 107 Pa = 0.3