Professional Documents
Culture Documents
Geot Earthquake Eng Ch9-WALLS-ADERS-16
Geot Earthquake Eng Ch9-WALLS-ADERS-16
Geot Earthquake Eng Ch9-WALLS-ADERS-16
9.
RETAINING STRUCTURES
Part A:
GRAVITY WALLS
G. BOUCKOVALAS
Professor of NTUA
October 2016
CONTENTS
kvW
cos θ
cos θ
Active Thrust:
ΔΚΑΕ = 1.50
1 50 kh1.50
1 50
G.B. 2014 !
B
H
Homework
k 9.2
9 2
Repeat Homework 9.1 for the φ=36deg
retaining wall shown in the figure. γγ=17 kN/m3
6m
δbase =24deg
δside= 24deg
2B
Hydro-STATIC
y pressures
p
p ws (x) w x
H
1
Pws p ws (x)dx w H 2
0
2
application
pp
point: Η/3 from base
ATTENTION !
The excess pore pressures are positive in front of the wall and negative
behind it. Thus the total hydro-dynamic pressure acting on a submerged wall
is twice that given by the Westergard solution!
REMARKS:
Westergaard theory applies under the following assumptions:
7
p wd (x) Cn k h w H x / H
8
7
Pwd Cn k h w H 2
12
1.17 Cn k h Pws
ό
ό
4 L/H
Cn 1.0
3 1L/ H
application
point:
p 0.40Η from the base
x x x x
p wd ((x, ) Cm ()k h w (2 ) (2 )
H H H H
or,
ή, approximately
ά
7 x
p wd (x,
( ) C m ( ) k h w
8 H
Westergaard
7 x
p wdd (x,
(x ) Cm k h w
8 H
and
7
Pwd Cm k h w 2
12
( 1.17 Cm k h Pws )
όwhere
(rad )
Cm 0.012 ( ) 2.0
application
point: 0.40Η από την βάση
WATER + BACKFILL
in other words….
Correction factor Ce “free” water
th portion
expresses the ti off
pore water which vibrates
“trapped” water, which
FREELY, vibrates together with
i.e. independently from the the soil skeleton
soil skeleton.
soil skeleton
γ* = γDRYCe+γSAT(1
(1--Ce)
GEORGE BOUCKOVALAS, National Technical University of Athens, 2016 9.8
EXAMPLE:
n=40%,
40% γw=10
10 kN/m
kN/ 3 H2
Ce 0.5 0.5 tanh log 6 10 6
Ew=2x106 kPa, T=0.30 sec k
Fill
Material Ce > 0.80
pwd ≈ Westergaard
Wester aard
well graded
gravel
gravel
Ce = 0.20÷0.90
coarse
sand pwd ≈ Ce·Westergaard
fine sand
silt
EXAMPLE:
n=40%,
40% γw=10
10 kN/m
kN/ 3 H2
Ce 0.5 0.5 tanh log 6 10 6
Ew=2 106 kPa, T=0.30 sec k
“Permeable” fill:
Cobbles, gravel,
Coarse sand (Η<20m)
“Semi-permeable» fill:
coarse sand (Η > 20m),
fine sand (H < 20m)
Impermeable fill:
“Impermeable”
silt, clay, clayey or silty sand
and gravel
hydrodynamic pressures
ATTENTION !
Pa computation requires (γκορ-γw) while ΔΡΑΕ computation requires γ*. Thus,
when it is necessary to compute both Pa and ΔΡΑΕΕ with a common unit weight
(e.g. ΕΑΚ 2002) you must use:
the buoyant unit weght (γSAT -γw)
a modified seismic coefficient
*
kh* kh
SAT w
hydrodynamic pressures
=dynamic
dynamic earth pressures
Clayey sand
Clayey silt
Silty sand
Clayey or silty gravel
=dynamic
dynamic earth pressures
kh* ≈ 2.2
2.2 kh
Clayey sand
Clayey silt
Silty sand
Clayey or silty gravel
GEORGE BOUCKOVALAS, National Technical University of Athens, 2016 9.12
active earth pressure
= hydrodynamic pressure
sand
sand & gravel
cobbles
ballast
= hydrodynamic pressure
kh* ≈ 1.6 kh
sand
sand & gravel
cobbles
ballast
GEORGE BOUCKOVALAS, National Technical University of Athens, 2016 9.13
EXAMPLE: What do I do when I am not sure about the permeability
of the fill material?
Fill:
γDRY=16 kN/m3
γSAT= 20 kN/m3
Ce = 0 ÷ 1.0
ΣFd = ΔΡΑΕ+Ρ
Ρwd+C
CePwd
N t
Notes:
Solve for two backfill options:
medium-coarse sand (k=10-3m/s) or
silty sand (k=10-5m/s).
For simplicity, assume that the
sea level coincides with the ground
surface.
Compare with Hwk 9.1
B
H
Homework
k 9.4
9 4
Repeat Homework 9.3 for the quay φ=36deg
wall shown in the figure. γγ=17 kN/m3
6m
δbase =24deg
δside= 24deg
2B
RETAINING STRUCTURES
P t B:
Part
WALLS
WITH LIMITED DISPLACEMENET
G BOUCKOVALAS
G. & G.
G KOURETZIS
October 2016
CONTENTS
Solutions for
perfectly rigid
“perfectly rigid” or
“semi-rigid” walls
Εwall>>Esoil
h
1
g
=Fm
h
1 h
g 1
g
h
2
Feq FP h
g M eq Fm 3
GEORGE BOUCKOVALAS, National Technical University of Athens, 2016 g 9.18
Analytical Solutions for ……..
eq
Application point of the resultant seismic thrust h
Feq
0.7
0.6
0.5
0.4
h/H
H
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10
L/H
h L
0.55 για 4
H H
excit Tsoil
soil Texcit
excit Tsoil
GEORGE BOUCKOVALAS,soil
Texcit University of Athens, 2016
National Technical 9.20
Comparison with Mononobe - Okabe
ξηρή άμμος
Dry sand This is the main reason why the elastic
φ, γ
ν=0.3
solutions of Wood (1973) were put aside
H f more th
for than 30 years… (in
(i connection
ti
with the fact that very limited wall failures
were observed during strong earthquakes)
1 1
Wood Wood
pseudostatic
d i pseudostatic
0.8 0.8
ΔFeq/[γΗ2(ah/g)]
ΔFeq/[γΗ2(ah//g)]
0.6 06
0.6
x3!!
3!!
απλοποίηση απλοποίηση
Seed & Whitman (kv=0) Seed & Whitman (kv=0)
0.4 0.4
Δ
Δ
0.2 0.2
Mononobe-Okabe
Mononobe-Okabe (φ=36ο)
(ah=0.15g)
0 0
28 32 36 40 44 0 0.1 0.2 0.3 0.4 0.5
φ (deg) ah (g)
h Wood 1973
H
(normal. height)
M-O
?
h
H
Pseudo static
base shear & overturning moment
Wood Wood
S d & Whit
Seed Whitman
Μ-O
For a flexible (compared to the fill) concrete wall (dw=20) and a seismic
excitation with amax=0.3g,
g the resulting g displacement
p is U/H=0.13%...
[ U=0.1%
GEORGE
U= 0.1%÷BOUCKOVALAS,
0.4%·H National
÷0.4%· Technical“failure”
for active →Μ-Ο]
University of Athens, 2016 9.24
Analytical solutions for …….
ή Resonance
"στατική" συντονισμός
λύση
υψίσυχνες
ί δ έ
διεγέρσεις
High frequency
2 Vsoil
Static 1
excitation soil 2H
Variation of amplification factor AF for base shear versus the fundamental soil period
period))
Average values of the amplification factor AF for base shear and relative displacement
1. Tensile cracks, at the top of the wall, are not taken into account
(→shear forces and bending moments are under
under-estimated)
estimated)
2. Uniform soil is assumed
(→shear forces and bending moments are over-estimated)
dw=0 dw=0
Note: (1) and (2) above have a counteracting effect for walls with rotational
flexibility dΘ>2
1 5αhγH
1.5α
h
eq 3
g
Wood
3
M eq 0.58 h
g
ΔPE=α.γ.H
2
ΔPE=0.75.α.γ.H
2
H H
0.58.H
H/2
σE=0.7.α.γ.Η σE=0.5.α.γ.Η
h h
M eq 0.375 3 M eq 0.58 3
g g
reminder:
H
M-O
M O 0 375.α.γ.H2
ΔPE 0.375 h
(U/H>0.1%) 0.60H
M eq 0.225 3
GEORGE BOUCKOVALAS, National Technical University of Athens, 2016
g 9.28
COMPARISON OF DIFFERENT METHODS ……….…
1
Wood
EAK 2002-εγκ.39/99
ΔFeeq/[γΗ2(ah//g)]
0.6
0.4
M-O
0.2
0
U/H
HWK 9.6:
Compute
C t the
th total
t t lbbase
s sh
shear force
f andd overturning
t i momentt which
hi h d
develops
l s att
the base of a 5m high retaining wall during seismic excitation with αmax=0.15g.
The wall is vertical and smooth, while the fill consists of sandy gravel with c=0,
φ=36
36ο, γΞ=17kN/m
17kN/ 3 and d VS=100m/s.
100 / Th The computations
t ti will
ill b
be performed:
f d
(α) for rigid wall,
(β) for a wall with limited deformation (dw=10,
=10 dθ=1),
=1) using the V&Y methodology,
methodology
(γ) for a wall with limited deformation (dw=10, dθ=1), using the seismic code
provisions,
Note: assume pseudo static conditions and neglect the wall mass.
HWK 9.7
9 7
Repeat HWK 9.6 for the extreme case of resonance between soil and excitation.
RETAINING STRUCTURES
Part C:
DISPLACEMENT COMPUTATION
(& PERFORMANCE BASED DESIGN)
G. BOUCKOVALAS
Professor of NTUA
October 2016
Problem Outline ….
During a pseudo
pseudo-static
static analysis
analysis, it is very common to obtain FS>1.0.
FS>1 0
khW
PA
However, this does not necessarily mean “failure” of the wall, but
permanent outward displacements (and rotations) rotations). In such cases cases, the
performance of the wall is evaluated using the famous “Newmark
Sliding Block” analysis (follows)National Technical University of Athens, 2016
GEORGE BOUCKOVALAS, 9.30
NEWMARK (1965): Rankine Lecture on Seismic Slope Displacements
a/g
W a(t)
ao/g base
μ block
tο tο+Δt t
W F W
F
T<μW bl k
ablock g
W / g
ao a
(a/g)W
F W W o W
W a(t) g g
T=μW
Relative SB acceleration:
a/g F
aSB ao g
W / g
ao/g base
Relative SB velocity at t < tο:
block
μ o
VSB aSB t o (ao g )t o
tο Δt
tο+Δt t
1 1
o
SSB aSBt o 2 (ao g )t o 2
2 2
GEORGE BOUCKOVALAS, National Technical University of Athens, 2016 9.31
B. RELATIVE DISPLACEMENT OF BLOCK for t > tο
o
VSB Steadily de‐celerating motion with initial velocity:
W
o
VSB aSB t o (ao g )t o
and de‐celeration:
T=μW F
a'SB g
W / g
a/g
This motion will last until the SB relative velocity
ao/g base becomes zero, i.e. ……
o
VSB (a g )t o a
μ block t ' o ( o 1 )t o
aSB g g
tο tο+Δt t
1 ' 1 (ao g )2 2
SSB V t a SB t .....
o 2
to
SB
2 2 g
tο tο+Δt t
tο tο+Δt t
Assuming further that:
tο tο+Δt t
Assuming further that:
2
1 Vmax
ao = amax (peak seismic acceleration) 1
μg = aCR (critical seismic acceleration SSB ( 1 aCR
*
) a*CR
2 amax
required
i d tto triger
ti sliding,
lidi i.e.
i FSslide=1.0)
1 0)
Vmax (peak seismic velocity), and
with *
aCR aCR / amax
to = Vmax/amax
GEORGE BOUCKOVALAS, National Technical University of Athens, 2016 9.33
D. For N similar pulses of base motion ……………….
μ block
tο tο+Δt t 2
1 Vmax *
2
or,, for veryy small a*CR ((<0.30)) …… a NEWMARK II
2 amax CR
with *
aCR aCR / amax
oλ
Even though F.S. < 1.0 (sliding failure)
th
there iis no collapse
ll of
f th
the wall
ll (!!),
(!!)
but development of limited displacements, which may be tolerable ……..
GEORGE BOUCKOVALAS, National Technical University of Athens, 2016 9.34
αcr=Ng :
critical seismic acceleration
leading to F.S.oλ=1.00
n)
CEMENT (in
range of
numerical Richards-Elms (1979)
analyses Richards & Elms (1997)
E DISPLAC
2
Vmax
4
0.087 a*CR
amax
RELATIVE
a*CR=aCR/amax
NEWMARK (1965)
æVmax
2 ö (1 - a*CR )
δ = 0.50 ⋅ ççç ÷÷ ⋅
. è a ÷ø maxa* 2 CR
.
æVmax
2 ö 1
. δ » 0.50 ç
0 50 ⋅ çç ÷÷ ⋅
è a ÷ø a
* 2
max CR
æVmax
2 ö 1
ç
δ » 0.087 ⋅ çç ÷÷ ⋅
è a ø÷ a
* 4
max CR
NTUA (1990)
æVmax
2 ö
δ » 0.080 ⋅ t 1.15 ç
⋅ç ÷÷ ⋅ é1 - a* ( 1-a*CR ) ù ⋅ 1
çè a ÷ø êë
max
CR úû a*
CR
Newmark
m - I ((1965))
Newmark – II (1965)
Richards & Elms (1979)
( )
CEMENT (in)
NTUA (1990)
άνω όριο
ENT DISPLAC
για διάφορα Μ
Relative Sliding
V a 2m
2a
ïìï ü
0
.
0
8
7
ï
*C
m
x
PERMANE
-
( ) ï
R
ïï ⋅ ⋅ ï
ï
δd
m
i
n
m
a
x
ïï ï
ï
=
V a
í ý
2
ïï ï
0
.
5
0
a
*C
a a
x
- ï
ïï ( ) ïï
R
⋅ ⋅
ïï ï
m
x
î ï
þ
Relative Sliding
V a 2m
2a
4
ìï üï
0
.
0
8
7
a
*C
m
x
!
-
ïï
( ) ïï
R
ïï ⋅ ⋅ ïï
δd
m
i
n
m
a
x
= ï ï 9 cm
V a
í ý
2
ïï ïï
0
.
5
0
a
*C
a a
x
-
ïï ( ) ïï
R
⋅ ⋅
ïï ïï
m
x
î þ
for EXAMPLE . . . . . .
Relative Sliding
V a 2m
2a
ïìï ïüï
0
.
0
8
7
a
*C
m
x
!
-
( )
R
ïï ⋅ ⋅ ïï
δd
m
i
n
m
a
x
ïï ïï
= 2 cm
V a
í ý
2
ïï ïï
0
.
5
0
a
*C
a a
x
-
ïï ( ) ïï
R
⋅ ⋅
ïï ïï
m
x
î þ
THUS
THUS,if we can tolerate some small outwards displacements,
the pseudo static analysis is NOT performed for the peak seismic
acceleration amax, but for the critical acceleration aCR ((& FSslide=1.0))
Instead of designing the wall for kh=amax/g, I choose a lower kh* (< kh)
which is a function of the allowable wall displacement δ. In that case,
th required
the i df factor
t of f safety
f t iis F.S.=1.00
F S 1 00
alternatively:
y
kh 1
k *h with: q w 1/ 4
qw 2
Vmax
0.087
max
In more detail ….
V a 2m
2a
ìï üï
0
.
0
8
7
a
*C
m
x
-
ïï
( ) ïï
R
ïï ⋅ ⋅ ïï
δ
m
i
n
m
a
x
= ï ï
V a
í ý
2
ïï ïï
0
.
5
0
a
*C
a a
x
-
( )
R
ïï ⋅ ⋅ ïï
ïï ïï
m
x
î þ
kh
k *h with:
qw
1
/
4
V a
1δ
2a
ïìï é ù- ïüï
0
.
0
8
7
m
x
ïï ê ⋅ ⋅ ú ïï
ïï ê ú ïï
m
a
x
qw
m
a
x
ï êë úû ï
=
1
/
2
í ý
V a
1δ
2a
ïï é ù- ïï
0
.
5
0
m
x
ï
ï ê ú ïï
ï ê ⋅ ⋅ ú ïï
m
a
x
ï
ï
î ëê ûú ïþ
n
kh
qw
max
g
γn=importance coefficient
2.00 δ(mm)=300a
1 50
1.50 δ(mm)=200a
qw= 1.25 δ(mm)=100a (τοίχοι από Ο.Σ.)
1 00
1.00 anchored
h d flexible
fl ibl walls
ll
HWK 9.5:
F r the gravity
For r vit ret
retaining
inin wall
ll of
f HWKs 9.1
9 1 and/or
nd/ r 9
9.2:
2:
(α) Compute the critical horizontal acceleration aCR, required to trigger sliding
( ) Compute
(b) p FSslide and the corresponding
p g outward displacement
p of the wall for
amax = 1.50 to 4.0 aCR and predominant excitation period Texc=0.40s.