Failure of an Embankment on Soft Clay

7/25/2019 Failure of an Embankment on Soft Clay

1/6

Missouri University of Science and Technology

Scholars' Mine

I#+# C% C# H++ +G%*+%# E++

F+ I#+# C% C# H++ +G%*+%# E++ (1984)

M# 6*, 12:00 AM

Failure of an embankment on so clayHugo Perez Lasalvia

M.K. Yegian

F *+ #& #&&++# #: *=://%*#+..&/+%%*

;+ A +% - C% %&+ + $* #& #%% $ * G%+% #& G+%# #& P E++ #

!%*#' M+. I *# $ #%%& +%+ + I#+# C% C# H++ + G%*+%# E++ $ # #*+&

#&++# !%*#' M+. F +#+, # %#%#7@.&.

%&& C+#+H P L##7+# #& M.K. +#, "F#+ # $# < %#" (M# 6, 1984).International Conference on CaseHistories in Geotechnical Engineering. P# 15.*=://%*#+..&/+%%*/1+%%*/1+%%*-*3/15

7/25/2019 Failure of an Embankment on Soft Clay

2/6

Failure of an Embankment

on

Soft Clay

M. K.

Yegian

Associate

Professor

of Civil

Engineering

Northeastern University Boston Massachusetts

Hugo Perez

Lasalvia

Teproy

Caracas Venezuela

SYNOPSIS

The

fa i lure of an

embankment

during

i t s

construct ion on

soft

clay

foundation

i s inves t i

gated. Field and laboratory data are used in conventional slope s t ab i l i t y analyses to explain the

cause of the

fai lure

and to

evaluate

the in- s i tu undrained shear s t rength of the clay which was

la

t e r

used

in the

design

of the

replacement dike. The resul ts indicate

tha t

whereas

the s t i f f

embank-

ment having a

well

compacted core was

i n i t i a l ly stable

on the sof t clay,

subsequent cracking

of the

core due

to undrained

deformations

reduced the shearing res is tance of the

dike

thus,

in i t ia t ing

the

fai lure. The dike was

eventually reconstructed

in stages allowing

enough

time between stages for

the

foundation clay

to consolidate

and

increase

i t s shear strength.

In the

new dike a

granular

ma-

t e r ia l

was used in order to provide f lexibi l i ty and to prevent cracking of the dike. In addition,

long

berms were

placed

on

both

sides

of

the

dike

to

increase

i t s

s tabi l i ty .

Field

instruments

in

cluding piezometers

and

inclinometers

were

used

to monitor the

r a te

of

the

stage

construction of

the

new

dike.

INTRODUCTION

The s i t e

preparation

for an industr ial plant

for

producing

alumina,

located

on

the bank

of

the

Orinoco River in Venezuela, included the

construct ion

of

dikes

for the

purpose

of

creat

ing

ar t i f i c ia l ponds where the residual materi

a l fe r ro-s i l t commonly referred

to

as Red Mud

will be deposited. The ter rain consists of two

natural

lagoons

whose

bottom elevat ions

are

about 6

meters above sea level

+6M

. Hydro

logical records

indicate

that

for

about 8 to 9

months of the year, the

Orinoco

River water

elevation

remains

higher

than 6M

and

can be

as

high as

+14M.

Hence,

for

most of the

year, the

lagoon

area remains

flooded. For th is

reason,

dikes with crest elevat ions

of

+17M were needed

to

prevent

the flooding

of

the lagoons,

thus

creat ing ar t i f ic ia l ponds which would

be

de

watered

and then

used

for deposit ing the

Red

Mud.

Two and

a

half months af ter construct ion had

s tar ted,

a

large

sect ion of one of the dikes,

fai led by sliding into the lagoon. The height

of the

400 meter long dike a t

the

time

of f a i l

ure was about

7

meters, 4 meters below the

f inal

design crest elevation). The

dike

con

s is ted

of

a well

compacted clay core and

shel l

of

sand,

gravel and rocks. Immediate

attempts

made

by

the

contractor to

reshape the dike

and

to

continue

i t s

construction

resulted

in simi

lar

fai lures

in the same zone as that of the

i n i t i a l

fai lure.

Following these

fai lures ,

extensive

f ield and laboratory

investigations

were

made,

including:

soi l

boring, sampling,

tes t ing and ins ta l la t ion and monitoring of

f ield instruments. The information

obtained

from

th is subsurface

exploration

program was

then

ut i l ized

in geotechnical engineering

analysis and

the

causes

of

the fai lure of

the

dike were ident i f ied and an al ternate design

and

construction

procedure

was adopted and the

701

dike

was

successful ly reconstructed.

The scope

of th i s

paper

s l imited

to

the presentation of

the resul ts of

the subsurface

soi l

investigation

and the s tabi l i ty analyses

made to

invest igate

the causes of the

fai lure

of

the

dike. In addi

t ion, a discussion

is

included on the final

pro

cedures adopted

for the reconstruction

of the

dike which i s

now

in operat ion.

SUBSURFACE INVESTIGATIONS

The

subsurface

inves t igat ions

carried

out after

the

fai lure

of

the

dike

consisted of:

Field

ex

plorat ions and

instrumentation and laboratory

soi l tes t ing.

Field Exploration and Instrumentation.

Immediately af te r the

fai lure

of the dike, topo

graphical

data

was

acquired

and the

extent of

the fai lure

of

the

dike and

i t s

shape

af te r

fai lure

were

established

as

shown

in

Figures 1

and 2. Following

th i s , standard penetration

t e s t s were

made

using

a sampling

spoon of 2 in

OD

with

a 140 lb. weight hammer fa l l ing 30

inches.

Figures 1 and 2 show the

locations of

a

few

of

the

bore

holes

and the i r

corresponding

logs.

In

addit ion

to

the Standard Penetration

Tests ,

geophysical investigations were

conducted

along a number of t raverses and the thicknesses

of dif ferent soi l

layers

were

estimated as shown

in Figures 1 and 2.

Based

on

these f i e ld investigations and visual

inspect ions

of samples of soi l

retr ieved from

the f ield

t es t s

soi l profi les

were prepared

as

shown in Figures 1 and 2. In summary,

the

s i t e

consis ts

of f lood-plain

deposi t of

low-to-medium

plas t ic i ty s i l t s and

s i l ty

clays containing

some

organic

matter and interrupted by layers and

7/25/2019 Failure of an Embankment on Soft Clay

3/6

Elevation

18m

PRIOR TO FAILURE

16

p31

p35

14

I

I

12.

I

10

6

6

4

2

0

2

4

6

e

-10

-12m

0

5 10m

Figure

1

Cross-section of

Dike

Pr ior to

and

After Fai lure

Elevation

~

F ILURE

1

ON

14

12

p21

p31

p32

p34 35

8

6

4

2

0

2

4

6

-e

-10m

0

50

~ O m

I

I

Figure 2

Subsurface Prof i le

Along the Longitudinal

Axis

of

Dike

702

7/25/2019 Failure of an Embankment on Soft Clay

4/6

pockets

of peat.

In

many of

the bore

holes

the

clay foundation soi l was

so

sof t that the

stand

ard

penetration tes ts

could not

be performed.

Underlying

th is

14 meter thick foundation soi l

s a

very dense s i l ty

sand.

In

addi t ion

to the f ield

explorat ions ,

piezo

meters

and

inclinometers were instal led in order

to evaluate the

exis t ing

subsurface condition.

The f ield instruments

were monitored on

a regu

l a r

basis .

The readings of the piezometers in

dicated

tha t

the

excess pore

pressures

in

the

subsoil were less than anticipated based on

theory and soi l

parameters

determined

in

the

laboratory. This

indicated that

the

ra te of

consolidation of

the

foundation

clay was

la rger

than determined from laboratory t es t s

on

small

samples.

Furthermore,

i t

was evident that the

permeabil i ty

of

the foundation

clay in the hori

zontal direction

was

larger

than

in the ver t i

cal

di rec t ion.

Laboratory

Investigations.

For

purposes

of

laboratory

studies ,

undisturbed

so i l

samples

were obtained using 3-inch

dia

meter

shelby tubes. The

laboratory

tes t ing

program included the

determination

of: Atter

berg

l imi ts .

strength

t es t s

consolidation

t es t s

organic content and clay mineralogy.

Figure 3

summarizes

the

laboratory t es t

data.

Based

on

these tes ts t s concluded tha t

the

s i l t y clay

encountered in the foundation of the

dike

i s

very soft

and

may even

be undercon

solidated

at shallow depths. The undrained

shear s trength of the clay

normalized

with re

spect

to the

vert ica l normal

efrect ive s t ress

s

estimated

to be about

0.25. The

effect ive

s t r ess strength parameters of the clay

are

es

t imated

to

be, cohension in tercept , c = and

f r i c t ion angle, f

= 26o.

SPT, BLOWS/ T

I

TTERBERG

LIMITS,%

A

C

4.0.

1

AA

E

.elC

l

1-

0

w

c

ANALYSIS

OF

FAILED DIKE

Based on the resul ts

of

these subsurface in-

vest iga t ions

and

the avai lable topographical

data , s tabi l i ty

analyses

were

made in order

to

explain

the cause

of the f a i lu re of the dike and

to

evaluate ,

al te rna te design

cross

sections of

the

dike and procedures for i t s

safe

reconstruc

t ion.

The

s t ab i l i t y

analyses

were

made

using

the computer program developed by

Dr.

Stephen

Wright

of the University

of

Texas

a t

~ u s t i n

Prior to

Fai lure .

In order

to

explain the

cause of the fai lure of

the dike and to evaluate the in -s i tu undrained

shear strength of the foundation

clay,

to tal

s t r ess s tabi l i ty analyses were

made

using

the

original

design cross sect ion of the

dike

prior

to the fai lure. Figure 4 shows the

cross

sec t

t ion analysed and the material properties used.

In

view

of

the

fac t that the

dike

a t

this stage

of i t s construction fai led the

factor

of

safety

must

have

been less

than unity (F.S.

7/25/2019 Failure of an Embankment on Soft Clay

5/6

2.S

>

-

2.0

ll

--Y.

I;.