Section 5: Post-collapse investigations 5.1.3.2 - 5.1.3.5 (x)

sodaspringsjasperΠολεοδομικά Έργα

15 Νοε 2013 (πριν από 3 χρόνια και 6 μήνες)

69 εμφανίσεις

5.1.3.2 Beams


The

beams we
r
e

la
r
gely

intact

with

the

damage
typically
at

the

ends and

on

the

top

whe
r
e

the

floor

slab
had
detached.

Most

of

the

beams Mr

F
r
ost

saw had

the

end
co
r
ners

b
r
oken

o
f
f

with

little

damage along the

beam
length,

as

shown

in Figu
r
e

63(a).


Mr

F
r
ost

said that

no

interior

beams we
r
e

found
with

the

conc
r
ete

slabs still

attached,

as

illustrated
by

Figu
r
e

63(b).

The

inte
r
nal

beam to

floor

connection
is illustrated

by

the

structural

drawing

section

in Figu
r
e

63(c).

The

steel Hi
-
Bond

flooring

is
discontinuous
a
t
th
e

beam
s

wit
h

a

60m
m

seatin
g

an
d
th
e

sla
b

mesh
passe
s

ove
r

th
e

to
p

o
f

th
e

bea
m

stirrup
s

an
d

longitudinal
r
einfo
r
cement.

Dr

Heywood

indicated

in
r
ed

whe
r
e

the
failu
r
e

surfac
e

wa
s

typicall
y

observed
.

H
e

als
o

observed
that

all

connections

between

the

edge beams and

floor
slabs we
r
e

seve
r
ed.

In

every

case

the

mesh
and
supplementary

H12
r
einfo
r
cement

was to
r
n

f
r
om

the
top

surface of

the

beam. The

structural

drawings
specified

that

slab
r
einfo
r
cing

should

be

placed

above the

beam steel. Mr

F
r
ost

s

opinion

was that

this
detailing

p
r
ovided

poor

connectivit
y
.

He

would

have placed

some slab
r
einfo
r
cing

unde
r
neath the

top
longitudinal

steel to

inc
r
ease

the

connectivity

between
the

beam and

the

slab.

350

200







T
op

of

beam whe
r
e
slab has

detached








Section

f
r
om

which
co
r
ner

has

b
r
oken

away



Damage

to

beam
end






Figu
r
e

63(a):

Edge

beam

f
r
om

line

4 between

lines

A and

B

(sou
r
ce:

Robert

Heywood)










Interior beam





Hi
-
Bond

flooring








Figu
r
e

63(b):

Interior

beam

f
r
om

the

weste
r
n

side

of

the

building

(sou
r
ce:

Graham

F
r
ost)



GRID



60
seating


200


200



60
seating



T
ypical

failure
surface


H12

at

120crs

4,000

long

400

alt

stagger

20

cover



4
-
H28


664

mesh



“Hi

Bond”

Formwork

Precast

beam



400

column

beyond




Figu
r
e

63(c):

Section

th
r
ough

p
r
ecast

inte
r
nal

beam

to

slab

with

typical

observed

failu
r
e
surface

indicated

in

r
ed


Figu
r
e

63:

Beam

exhibits

and details



5.1.3.3

Beam
-
column

joints


Neither Mr

T
r
owsdale nor

Mr

F
r
ost

saw any

beam
-

column

joints

intact.

Mr

F
r
ost

stated

that

the

joint
r
egions had

fallen

apart

as

the
r
e

was no

steel to

confine

the

conc
r
ete,

as

illustrated

in Figu
r
e

64(a).
Dr

Heywood

did

not

observe any

interior

beam
-
column
joint

whe
r
e

the

conc
r
ete

in the

joint

had

not

been

lost
or

r
ende
r
ed

ine
f
fective,

other

than at

level

6 whe
r
e the

columns

supported

the

r
oof.

The

p
r
ecast

beam in Figu
r
e

64(b)

has

pulled

away

f
r
om

the

beam
-
column
joint

and

is lying on

its

side. The

photograph

illustrates the

smooth

formed

ci
r
cular

surface and

b
r
oken

o
f
f
co
r
ners

(or

“wings”)

that

we
r
e

a

common

observation
with

these beams.


Dr

Heywood

exp
r
essed

the

opinion

that

it

was likely
that

the

connections

between

the

r
ectangular
columns
and

the

adjoining

beams on

the

weste
r
n

frame
on

line

A

(see

Figu
r
e

62(a))

disintegrated

during

the

early
stages of

collapse.

As

a

r
esult the

r
ectangular
columns
detached

f
r
om

the

building.

Dr

Heywood

conside
r
ed
that

if

the

columns

had

r
emained attached

they

would
have

su
f
fe
r
ed

considerable

damage and

the

floors
would

not

have

come

to

r
est

one

above another as

he
observed.

The

end

of

a

beam which

was connected

to
a

co
r
ner

column

on

line

A

is shown

in Figu
r
e

64(c). Dr

Heywood

described

the

conc
r
ete

f
r
om

within

the
beam
-
column

joint

r
emaining within

the

90º

bent

bars
in the

beams.

The

beam pulled

out

f
r
om

the

column
between

the

column

r
einfo
r
cement.

An

imprint

of

the
vertical

column

r
einfo
r
cement

is visible in this

nar
r
ow
edge beam end

as

well as

the

smooth

end

f
r
om

the
p
r
ecast

portion.

This

was a

poorly

detailed

connection,
as

we

discuss

in section

6.3.5.












Beam
-
column
joint

r
egion













Figu
r
e

64(a):

Disintegrated

beam
-
column

joint

(sou
r
ce:

Graham

F
r
ost)










Section

f
r
om
which

the

co
r
ner
or

“wing”

has
b
r
oken

away





Smoothed

ci
r
cular
surface








(b)

Beam

end

with

a smoothed

ci
r
cular

end

and

bottom

longitudinal

r
einfo
r
cement

that
connected

to

columns

(sou
r
ce:

Graham

F
r
ost)






Conc
r
ete

f
r
om
within

beam
-

column

joint


Smooth

end

of
p
r
ecast

edge
beam




Imprint

of

vertical
r
einfo
r
cing



90º

bent

bars





(c)

End

of

a beam

(sou
r
ce:

Graham

F
r
ost)


Figu
r
e

64:

Beam
-
column

joint

details


When

examining the

r
emnants of

the

CTV

building

at
the

Burwood

landfill

on

28 June 2012, Dr

Heywood
observed

a

number

of

inte
r
nal

and

edge beams in
the
rubble.

All

of

the

“wings”

at

the

end

of

these beams
had

b
r
oken

awa
y
.

These

wings

acted

as

a

mould
for

the

column

conc
r
ete

which

was cast

in situ

(see
Figu
r
es

64(b)

and

65).

The

semi
-
ci
r
cular

surface at

the
beam ends had

no

surface
r
oughening and

the
r
e

was
no

evidence of

any

substantial

bond

between

the

beam and

column

conc
r
ete.

Mr

Daniel

Morris,

who

operated

a

conc
r
ete

cutting
business in the

1990s, was called

as

a

witness

by
Alan

Reay

Consultants

Limited

(ARCL)

and

Dr

Rea
y
.
Mr

Morris

initially gave

evidence that

about

200 holes
had

been

drilled

by

employees of

his firm

in the

CTV
building

at

some point

in the

1990s, about

50 of

which
we
r
e

in beams. Howeve
r
, in c
r
oss
-
examination

he
conceded

that

his evidence was a

guess

which

could
be

“wildly

out”.

W
e

add

that,

while at

the

Burwood
landfill,

Dr

Heywood

looked

for

evidence of

holes

co
r
ed
th
r
ough

the

beams.

400

P
r
ecast

log

beam

Cent
r
e

line

of

column

P
r
ecast

log

beam



He

did

not

find

any

inte
r
nal

or

edge beams with

holes
co
r
ed

th
r
ough

them.

In

view of

his answers

in
c
r
oss
-

examination we

did

not

consider

the

evidence
of
Mr

Morris

to

be

of

any

p
r
obative

value.


Mr

F
r
ost,

who

is experienced in conc
r
ete

construction,
gave

evidence about

the

weakness

of

the

inte
r
nal
beam
-
column

joints.

He

explained that

when using
p
r
ecast

conc
r
ete,

designers should

detail

the
construction

joint

to

be

perpendicular

to

the

beam axis

so

that

the

comp
r
ession

fo
r
ces

can

be

transfer
r
ed
di
r
ectly

ac
r
oss

the

joint

to

limit

the

risk

of

slip

along the

interface. He

pointed

out

that

sloping

construction
joints

can

r
esult in a

g
r
eater

spacing

between

stirrups,
which

r
educes

the

confinement

of

the

conc
r
ete.


Figu
r
e

65 is a

diagram

of

the

interior

beam
-
column

joint
p
r
epa
r
ed

by

Mr

F
r
ost.

This

diagram

illustrates the

fo
r
ces
that

could

lead

to

the

beam wings

b
r
eaking

o
f
f.

The

r
ed
ar
r
ows

indicate

the

di
r
ection

of

the

comp
r
ession

fo
r
ces

f
r
om

gravity

loading

at

the

bottom

of

the

beam. The
very

smooth

interface between

the

beam
and

column
means

that

fo
r
ces

cannot

be

e
f
fectively

transfer
r
ed
straight

ac
r
oss

the

joint.

This,

together

with

the

curved
surface, generates

radial

fo
r
ces

f
r
om

the

column
which

tend

to

split

the

beam wing

sections

o
f
f.

Since the
r
e

is
no

r
einfo
r
cing

acting

to

confine

these wings,

a crack

may develop

as

illustrated

in Figu
r
e

65. Mr

F
r
ost
thought

it

was a

very

st
r
ong

possibility

that

a

pulse
f
r
om
vertical

acceleration

c
r
eated

a

fo
r
ce

sufficient

to

b
r
eak
the

wings

o
f
f.

He

stated

that,

after these wings

we
r
e
lost,

the

joint

would

have

little

capacity

left.

The

gravity
and

seismic

shear

loads

on

the

beams would

then have to

be

transfer
r
ed

th
r
ough

the

slab, dowel

action

and
the

much

r
educed

bearing a
r
ea

at

the

bottom

of

the
beam. Mr

F
r
ost

conside
r
ed

that

this

bearing a
r
ea

was a

vulnerable

r
egion at

the

top

of

the

column.

The

beam sat

on

the

cover

conc
r
ete

of

the

column

which

would
spall and

fall away

when the

column

was exposed

to
high
drifts.





App
r
oximate

location

of

last

stirrup

T
ypical

crack

location

seen

after

collapse

Original

Dwgs

call

for

the

gaps

between

beam

ends

to

be

sealed

to

p
r
event

g
r
out

loss
during

insitu

conc
r
ete

(infill)

pou
r
.

It

s

not

clear

to

me

what

made

up

the

fill

in

these
nominal

25x20


r
ebates’


(PEF

r
od/sealer/conc
r
e
te)


Beam

co
r
ner

or

‘wing’




Comp
r
essive

st
r
esses

at

bottom

of

beam
under

gravity

load

conditions
.

(to

balance

steel tension

loads)

(f
r
om

-
ve

moment

at

interior

beam

ends)


With

smooth

interface
between

p
r
ecast

(log)

beam

and

350
Ø

column

infill

conc
r
ete,
longitudinal

beam

fo
r
ces

can

t

be
transfer
r
ed

‘in

line’

ac
r
oss

interface.
So la
r
ge

RADIAL

fo
r
ces

generated.
Since the
r
e we
r
e no

beam

stirrup
s
detailed

to

be

located

past

start

of

curved

end,

the
r
e was

nothing
to

confine

the

conc
r
ete

in
th
e

‘wings’

or

r
esist

these
radial/splitting
fo
r
ces.

175

V
ertical

EQ

acceleration

load
s
would

have

inc
r
eased
th
e
magnitude

of

all

these fo
r
ces

and
likely exceeded

the

tensile
capacity

of

the

un
r
einfo
r
ced

conc
r
ete

‘wings


at

the

end

of

the

beams.






230

(as

drawn)

(examples

up

to

300mm

found

in

debris)


Outline

of

400
Ø

Supporting

column

below


P
r
ecast

Log

Beam


P
r
ecast

Log

Bea
m


Plan

section

at

typical

interior

column

(th
r
ough

bottom

of

p
r
ecast

Log

beams)


Figu
r
e

65:

Plan

view of beam
-
column

j
oint

p
r
epa
r
ed

by

Graham F
r
ost


These

observations

led

Mr

F
r
ost

to

conclude

that

a
possible

building

failu
r
e

mechanism was the

failu
r
e
of

one

or

mo
r
e

beam
-
column

joints

due

to

a

lack

of

confining

steel in the

beam ends and

beam
-

column

joints,

exacerbated

by

the

transfer of

beam

comp
r
ession

st
r
esses

ac
r
oss

smooth

formed

surfaces
that

we
r
e

not

perpendicular

to

the

line

of

action

of
those

fo
r
ces.


The

beam
-
column

joints

a
r
e

also

discussed

in
section

6.3.5 and

sections

7 and

8 of

this

V
olume.



5.1.3.4 Performance

of the

floor slabs


Mr

F
r
ost

did

not

see

any

conc
r
ete

floor

slabs whe
r
e
the

Hi
-
Bond

metal decking

was still

attached.

F
r
om this

he

postulated

that

vertical

accelerations may have

been

g
r
eat

enough to

lead

to

loss

of

the

bond

between
the

metal decking

and

the

conc
r
ete

slab or

tension
failu
r
e

of

the

metal decking.

Mr

F
r
ost

stated

that

the
metal
decking

had

little

ductilit
y
,

the
r
efo
r
e

a

temporary
overloading

fo
r
ce

could

lead

to

a

brittle

fractu
r
e

of

this
material.


Mr

F
r
ost

ag
r
eed,

howeve
r
, that

the

separation may
have

been

due

to

other

fo
r
ces

experienced during

the
collapse.

The

composite

floors

consist

of

a

fairly
sti
f
f
conc
r
ete

element on

a

mo
r
e

flexible

Hi
-
Bond

material.
These

two

elements could

have

separated when
they
impacted

on

the

g
r
ound.


5.1.3.5 Disconnection

of floor slabs

f
r
om
the

walls


The

conc
r
ete

floor

slabs came to

r
est

leaning

against
the

north

wall complex

as

illustrated

in Figu
r
e

66(a).
Dr

Heywood

could

identify

all

five edges of

the
suspended

floor

slabs in this

Figu
r
e.

This

photograph
was taken after most

of

the

building

debris

had

been

r
emoved. Mr

F
r
ost

s

opinion

was that

the

upwa
r
d

slope

of

the

floor

slabs towa
r
ds

the

north

wall complex

was a
st
r
ong

indication

that

the

floor

separated f
r
om

the

north
wall complex

later rather than earlier

in the

collapse
sequence, see

Figu
r
e

66(b).

He

explained that

if

the

floor

slabs had

separated f
r
om

the

north

wall
complex
befo
r
e

they

lost

support

f
r
om

the

central

columns,

he would

have

expected

to

see

the

floors

in a

horizontal
orientation

or

even

sloping

down

towa
r
ds

the

north

wall
complex.


Dr

Heywood

gave

evidence that

five and

possibly

six

of
the

connections

between

each

of

the

suspended

floor
slabs and

the

south

shear

wall we
r
e

seve
r
ed

and

the
edges of

the

floor

slabs came to

r
est

r
elatively

close

to

the

base

of

the

south

shear

wall near

line

1 (Figu
r
es

69(c)

and

(d)).

This

suggested

that

the

floors

most

likely
detached

f
r
om

the

south

shear

wall befo
r
e

the

wall
collapsed:

if

the

floor

slabs had

r
emained attached

to
the

south

shear

wall they

would

have

been

transported
north

with

the

collapsing

south

wall, rather than
r
emaining at

line

1. Mr

F
r
ost

did

not

see

any

slabs
that
r
emained connected

to

the

south

shear

wall.










Floor slabs
leaning

against
north

wall
complex













Figu
r
e

66(a):

W
este
r
n

elevation

of

collapsed

floor

slabs

leaning

against

north

wall

complex

(sou
r
ce:

Robert

Heywood)






































(b)

Floor

detachment

f
r
om

east

side

of

north

wall

complex

(sou
r
ce:

Robert

Heywood)


Figu
r
e

66:

Elevations

of the

north

wall complex


Dr

Heywood

also

gave

evidence about

the

detachment
o
f

th
e

floor
s

f
r
o
m

th
e

nort
h

wal
l

complex
.

Th
e

detachment
was complete

on

levels

4, 5 and

6
(Figu
r
e

67(b)),
although

some inte
r
nal

and

edge beams we
r
e

still
attache
d

t
o

th
e

nort
h

wal
l

complex
.

D
r

Heywoo
d

stated
that

in the

vicinity

of

the

amenities (the

weste
r
n

portion
of

the

north

wall complex) the

floor

slab was seve
r
ed 1

2m

south

of

line

4, leaving

the

floor

slab cantilevering
f
r
om

the

north

wall complex

(see

Figu
r
e

67(a)).

The 664
mesh was observed

to

have

failed in tension

on

the

level

6 failu
r
e

surface. The

downwa
r
d

angle

of

the
exposed

mesh and

the

spalling

of

the

underside of

the
slab a
r
e

consistent

with

some downwa
r
d

movement
during

failu
r
e.

The

failu
r
e

line

at

level

6 extended

ac
r
oss
the

f
r
ont

of

the

stairs and

lift

wells befo
r
e

tu
r
ning

north
at

the

easte
r
n

edge of

the

north

wall complex.

This
section

of

the

slab was subsequently

r
emoved

because
of

safety conce
r
ns

(see

Figu
r
e

67(a)).

On

levels

4 and

5
the

failu
r
e

line

extended

ac
r
oss

in f
r
ont

of

the

stairwell
befo
r
e

tu
r
ning

into

the

lift

well.

Dr

Heywood

saw drag

bars installed

on

levels

4, 5
and

6 on

both

sides of

the

lift

well walls in the

north

wall
complex.

He

said that

these drag

bars did

not

p
r
event
the

slab f
r
om

detaching

f
r
om

the

north

wall
complex
(see

Figu
r
es

66(b)

and

67(b)),

however the

drag

bars
still

r
emained attached

to

the

lift

well. In

some
places
a

piece

of

Hi
-
Bond

metal decking

r
emained
attached
to

the

drag

bars. The

drag

bars we
r
e

bent

downwa
r
ds,
which

is consistent

with

them

supporting

the

weight

of
the

floor

slabs during

the

collapse

rather than the

floor
slabs detaching

f
r
om

the

drag

bars befo
r
e

the

collapse.
Earthquake

fo
r
ces

would

have

applied

principally
horizontal fo
r
ces

to

the

drag

bars befo
r
e

the

collapse.
Dr

Heywood

explained that

if

the

earthquake
fo
r
ces
caused

separation of

the

floor

slabs f
r
om

the

drag
bars, he

would

have

expected

to

see

the

drag

bars virtually horizontal.