ae eaninc owero'sa
—a suggested remedial measure
by Professor H. C. VEDER"
Because the Leaning Tower of Pisa is settling more every year, its position is precarious. In this articleii the structure of the
soil is described and the tilt. which began
shortly after construction was started. is
explained. The author's proposed remedial
measures are directed at the causes of the
inclined position of the tower i.e. the compression of a 30m thick layer of clay. He
suggests compressing this clay layer with
an external load on the side of the tower
which is not leaning. The method could
be cavied out by degrees and is reversible, and the proposed technique would
not detrimentally affect tourism.
on the north side. When the builder saw
that the tower was leaning towards the
north, he had the blocks moved to the
south of the tower where they remained
till the end of construction.
Because soil mechanics knowledge was
limited, it was long believed
that the
reason for the tilting could be found in the
10m thick upper layer of sand '. It was
maintained that this layer had settled, or
more particularly,
that the fine particles
had been washed away.
More recent articles have rejected this
theory and maintain that the tower leans
because the layer of clay was compressed.
By means of numerous borings the exact
course of the surface of the clay could be
established; it shows a depression exactly
under the tower. The lowest point of this
depression is eccentric to the surface of
the foundation and is directed towards the
south, i.e. in the direction of the inclination. The deformation curve caused by the
distribution
of stress resulting from the
tilting corresponds almost exactly to the
observed shape of the depression.
The results
of various
calculations
applying the theory of consolidation
to
the clay layer were able to explain the
average settlement as well as the inclined
THE TOWER OF PISA rests on a ring-like
foundation
h a I I ow
(outer diameter
19.58m, inner 4.50m) and originally had its
foundation less than 2m below ground sur-
s
face.
In the course of 800 years the tower
has sunk about 2m and in the process
has rotated about 5.8 deg. towards the
south so that there is now a difference of
1.8m in height between the north and
south side of the foundation. Whereas the
originally projected bearing stress already
had the quite high value of about 0.51N/
mm', today the bearing stress amounts
to a maximum of 0.961N/mm'nd a minimum of 0.067N/mm-".
The inclination of the tower which is
currently about 1 in 10, has, in the course
of the past 50 years, increased by 1.5 per
cent. Annual settlement of the southern
edge is about 1mm/year
whereas the
northern edge does not change.
Reasons for the Tower's indination
Underneath
the base of the foundation
there is an layer of silty sand approximatethis is another
ly Bm thick. Underlying
layer about 30m thick consisting of alternating strata of clays and clayey silts.
Beneath this stratum, to a virtually limitless
depth, is an almost incompressible
layer
of sand. The ground water has a slight
grade and is less than 1m beneath the
surface.
Construction
of the tower was begun
It is rarely known that a northward inclination was observed only a few
in
1173.
years after construction was begun. After
eight years, when four storeys had been
built and the tower was 24.60m high, construction was suspended apparently because of the then inexplicable tilting.
During the next 100 years the tower began, strangely enough, leaning towards
the south '. Then in 1272-1278 it was
decided to build the tower to its present
height of 54.58m despite the fact that it
leaned.
The author believes this may be explained in the following way. The construction
material,
namely
the marble
blocks from Carrara, were originally stored
aHead of The Institute for Soil Mechanics, Civil
Engineering and Rock Mechanics, Technischen
Hochschule, Grat, Austria.
This article first appeared, in German,
auingenieur, 50 (t975I,
TSI, pp. 204-205.
38
Ground
Engineering
in
Der
lal
e,,I~
4
r,,''(
iilI
'a'q'r
~Q],IF
Fig. 1. The Leening Tower of Pise, from e 1782 print
=en'liilr.-'
>
lg~
9th S
c~
v
7th StoreY
ca
6th Storey
er
I
~
V
ts
DI
co
5th Storey
2cV
v
Q
V
v0
3rd Storey
v0—
c
2nd Storey
g
1st Storey
vre .9
rn
Inclination toward south
4th Storey
4th Storey
O
c0
VO
o
Completion of 1st storey
Works continuing up to the
middle of the 4th storey
re
y
'~ ~start
O
Q
I
o
o
v
CU
0
0
R
~
o
8
R
t
snd completion
structure—
of the
ji
of foundations
8
Resttmption
t
Interruption to construction
work, the first break lasting about 100 years
I
re
O
cv
ii
il
I
o
Between these years work resumed
to the floor of the 8th storey
up
The second interruption
about 90 years
I
Ifs
8
O
O
lasted
Nej
I
I
g
ttm
S
Years
Fig. The early stagesin the building of the tower
shown that
er of sand
r la
aresulto
t e
pie
o
the tower was spraye
w
go
r the tower
4
of view of soil mechh
anics, two factors are resp onsible for the
leaning of the tower, name y:
(a) the distribution of stress in the soil
due to the load of ' the towe, takin g into
and line of stress
I:
',
(b) the deformat on cha act n
i
er of cia .
oof the
Regarding
e o ' su gg estions
(a),
have been made for im
imp rovin the present
unfavourable
distributioi n of stress by
b
wer and the line
of action of the force. Howeve,
wever, all the
v thee disadvantage that they
proposa I s h ave
- ear old deer in some way
anentl, either
''on
Thus the most obvious solution
—
that
of
t
iece b
iece
ffid 40m deep foundat
econstructing i
t e sa
pI
not feasible because most o f the buildin the
rocess.
Foundations on piles or csissons which
o,
b
hi h
Idd
II
fo
must
dilff g
it
oil mechanics
standpoint
a soi
be desirable to b oaden the o
a late as close as
sible to the existing foundation nng.
small
mount of the injection mate 'ial between
the
i g soil,
the basin and the surroun din
eared
sin disa
the soil remained unchanged.
'
There was no inffuence on t hee stabilityy
of the tower; it continued to tilt about
The failure of this measure
1mm/year.
an movement of
'
fine partic es in thee groundwater it did not
cause the tower to lean.
'n
s
uld
wi
po
y
F
From
so far
Remedial measures
Io
t
il
e to
could have practically no in uence on
total displacement of the tower '.
r
of the
foun
require enclosing the foundation
o
hi
ri
o
tower wi'h a o
—
—for
example,
using
t e
In any event, with these proce urea i
o tightly enclose the
would b
be necessary
ary to
tower ' temporari I y with a supporting frame
or wit
ca es; either technique would
d
bl;
the tower could again start tilting
while
Th
'.
after a
tho'io
the leaning
anin process
rocess 'of the tower by means
'
of a reversible additional loadd on
e p
sure-free north side.
ose to e
etric I
but withou
ou
a
h't
I'd
Idb
d
ovet e c
A
I
of hi Io d
of stress on the su ace o
rib
i
e'ut
tower.
of the
In
addition,
t h e appearance
inarred for
tower wou
Id b e considerably
'
t ese methods which, from the
years wit these
i
of
stan d point
o tourism would be most undesirable.
R egarrdin
ing (b), if one wanted to-change
soil,
the deformation characteristics of the soi,
then the sti ff nesss o
of thee clay near the high
r ssure
u
e d ge pres
(south side) should be increase,d, or th e stiffness of the soil near
the low e d ge p ressure (north side) should
be decrease d.. Th e d e formation modulus of
the clay layer
I
can bee iincreased only with
d'ff
i
It
unle s additional
un
load
are
icuty
a pp lied because the modulus is irreversi e
A reduction o t e e o
in the area o
w
to e s
only exten
h
have
to b e done
one by removing sand locally,
ing. This would of necessity
e.g. b y was hing.
h
h 'd
d
lea d to a se I e
a counter-rotation
t r-rotation due
ue to the loosening of
volume'.
Because on I y th e sand layer would be
ibl
treate,d regu I atiin thi irr
dure would be very difffcult although settlement wou Idb e o
i kl
. Thi
cedure wou I d hav
ave
too be carried out very
'
near t h e f oun d a tio of th
o
h
dd
h
i
o diio
o d
'
take p ace. n a i
tower's inc ine posi '
I
I eyer ) w
the cay
Id
b
ff
d
d
is.ss
—
Slty sand
Clay with
~l,'i
DUnlaroUS
silty layers
3=D
p Corrosion-proof
tie rods
3S,Q
Sand
—
—
''
.-1
B
SECTION A —
..
1.3 1.3
N
PLAN
Fig. 3. Arrangement
tower's foundation
of tie-rods outside the
January,
1976
39
on the extained which is superimposed
isting stress to such an extent that the
resulting stress distribution on the surface
of the clay is more even and less eccentric. A further leaning of the tower can be
on the size of the
stopped depending
applied load and its distribution. It is even
possible to eliminate somewhat the present tilt.
This corrective process can be started
slowly by first regulating the load, by controlling pore water pressure, and then stopping when, based on observations, the desired results can be exactly estimated.
the process can even be
Subsequently,
accelerated by using an "excess load."
When a certain degree of consolidation
has been attained, this load can be partially removed to achieve a state of rest.
This procedure
corresponds
to Terzaghi's "observation method", i.e. a stepby-step approach, controlled by observations, to the end result.
The load could be applied to the sand
layer by using anchors (Fig. 3). A system
of reinforced concrete slabs would be
placed in a pattern about 1m under the
surface. In each of the slabs
ground
would be a tie rod which would be drilled
through the layer of clay and which would
extend to some depth into the lower sand
layer. In order to avoid de-stressing the
soil, drilling would be carried out using
bentonite suspension.
The anchors over their free length in the
clay layer and in the upper sand layer
would be protected against corrosion by
being enclosed in corrosion-proof tubes
and by additionally injecting the space between the tube and the tie rod. After injection, the anchors would be strained
with tensile forces whose reaction load
the soil. The resulting distribution of stress
on the surface of the clay superimposes
itself on the stress distribution resulting
from the load of the tower.
Apart from the small drill holes for the
anchors and for the construction of the
covering slab, the ground would not be
touched. The tensile stress of the anchors
can be regulated easily to adapt them to
requirements; the ground is not altered.
The stress distribution
active on the
bottom of the foundation under the tower
is triangular and has a peak value of approximately 94 N/cm'-'. This produces a
peak value of 57 N/cms on the surface
of the clay. This value is displaced southwards about 5m from the perpendicular
through the centre of the foundation. It is
planned to load an area of about 180ms
(Fig. 3) with concrete slabs measuring
1.20m x 1.20m. Each of these slabs is
provided with an anchor having a tensile
force of 490kN. The centres of the slabs
form a pattern of squares 1.3m wide.
The load on the surface of the soil corGroundwater
level
X
i""""
~~
—
Gay surface
m
6
0
05
B 1.0
Ti
o 1.5
f
2.0
I
satdamaat dua to
loading with 29.4N/arn
a>
—
t
5 tea~at:
Aaaumad aaunlaraotation
Settlement
la
tOWer
tower + loadmg
dua,l
a
I
alaaar
/
2.5
—
Advantages
method
Fig. 4. Settlement profile before end after
presrressing the soil
Ground
Engineering
—
of the proposed
It is not necessary to support the tower
during construction
nor to drill through,
enclose or weaken it with supporting
structures.
The tower itself is not touched.
The soil immediately around the tower
remains unchanged, and there is no interference with the ground water conditions.
The stabilisation measures are aimed at
eliminating the cause of the inclined position; they are reversible and applicable in
degrees; i.e. they can be adapted to the
requirements
indicated by continuous observations.
Tourism would not be hampered in any
way and the stabilisation
process as a
whole is financially feasible.
Bibliography
1. Ricerche e studi. Bautenmisterium Rom (Ministry for Construction, Rome) 1971.
Ssnpeoiesi "II campanile di Pisa". Pisa, 1959.
3. Tarzeghi, "Dih Ursschen der Schiefstellung des
Turmes von Pisa". Der Bsuingenieur 15 (1934)
p. 1.
4. Terecine, Proceedings
2.0
40
responds to a uniformly distributed surcharge of about 29.4 N/cm'. This creates
a distribution of stress on the surface of
the clay with a peak value of 12.7 N/cm'.
caused by settlement
The depression
changes from its present state somewhat
as depicted in Fig. 4; this means a counterrotation of almost 1 deg.
The anchors would be 50m long; 10m
of this length, lying in the lower sand
layer would represent the bonded section.
After the injected material had set, the
anchors would be prestressed to about
half of the planned tensile force; i.e. 245
kN each. An observation
period of 10
months would follow.
During this time the anchor forces must
be regularly checked and, if necessary, restressed. After this observation period, the
anchors would
be prestressed
to full
490 kN each dependbearing capacity
results deeming on whether observation
ed this necessary.
the
Subsequently,
anchors would be provided with an easily
accessible covering which did not protrude over the ground surface and which
made it possible to re-stress at any later
time.
One could also achieve the same results
with various other methods of surcharging. For example one such method would
be excavating trenches to about 3m above
the clay layer on the north side of the
tower. The trenches would be stabilised
with bentonite
slurry and subsequently
filled with scrap iron. Initially the trenches
would be arranged far apart in a pattern.
the results would be obSubsequently,
served for ten months. Deductions would
then be made regarding the effectiveness
of this method and to ascertain whether
these measures should be continued.
One could also use mercury in underground steel basins. Because this metal
is very dangerous
to humans, however,
this method is not without its problems.
After about ten years the settling process should be almost complete. This can
be said with some degree of confidence
because the tower began settling and
leaning only a few years (approximately
eight) after construction had been begun.
The clay layer consolidates
relatively
quickly because it has recurring intermediate seams of rather permeable silty sand.
of the 5th International
Conference of Soil Mechanics snd Foundation
Engineering, Paris. Vol. III. p. 212.
5. Schulze, Muhs, Bodenuntersuchungen
fUr Ingenicurbauten (1967) p. 662.
Pressuremeter
(continued from page
31)
was possibly an under-estimate
obtained by triaxial test
and no mention was made of piston sampling. Assuming it to be correct, E = 220.
C
but this
because
it was
Summary
1. "Initial" failure occurs at a nett pressure
equal to about 2.5 times the quick undrained shear strength.
failure occurs at a nett
pressure equal to about 3.5 times the quick
undrained shear strength.
3. The table below compares values of
Young's Modulus determined
by the air
bag penetration pressure meter with other
determinations
on the same site and at
Shellhaven.
E
2. "Cylindrical"
E(kN/m')
1. Penetration pressure meter 4255
2. Trial embankment
(vertical movement)
3.
C
200
2 346
112
2 160
220
Trial embankment
(lateral movement)
4. Triaxial cell (mean)
5. Consolidation cell.
6. Settlement of
oil tank at Shellhaven
Condusions
1. Inevitably there must be some remouldvicinity of
ing of the soil in the immediate
where the pressure meter has been pushed
below the bottom of the borehole. However, because the soil must remain in contact with the pressure meter, we consider
that this remoulded zone forms only a
minute proportion of the total volume of
soil stressed during a test and consequently
has a negligible effect on the determination
of the soils Young's Modulus.
2. In comparing Young's Modulus obtained
the following
by the various methods;
points are relevant:—
(a) The triaxial specimens were "sculptured" from block samples taken with
great care from trial pits. They therefore
suffered only a small amount of disturbance with only a small loss in the value
of E.
(b) The consolidation cell specimens were
taken from a 10in (254mm) dia. Roating
piston sampie from a borehole. This would
have suffered more disturbance
than the
samples in (a) above, with a consequent
greater loss of E.
observations
are
(c) The embankment
known to include some plastic movement,
although in the calculations the movements
were assumed to be elastic. This would, of
course, considerably reduce the value of E.
The same observation probably applies to
the Shellhaven tank.
3. The recorded values in the summary are
quite in line with the above points. Disregarding the pressure meter result, one
would expect the true value of E to be a
little more than the 3250kN/m'or
the
triaxial tests. It would appear, therefore,
that the penetration pressure meter gives
a fairly accurate measure of E.
4. "Initial failure" of the clay occurs at a
nett pressure of about 2.5 times the shear
strength. "Cylindrical failure" occurs at a
nett pressure of about 3.5 times the shear
strength. The test is simpler and marginally
quicker than using the penetration
vane
but the former entails a determination
of
bulk densities in order to evaluate nett
pressures.