Professional Documents
Culture Documents
2003
Japanese Geotechnical Society
ABSTRACT
The stabilisation of the Tower of Pisa has been a very difficult challenge for geotechnical engineering. The Tower is
founded on weak, highly compressible soils and its inclination has been increasing inexorably over the years to the
point at which it was about to reach leaning instability. Any disturbance to the ground beneath the south side of the
foundation is very dangerous; therefore the use of conventional geotechnical processes at the south side, such as under-
pinning, grouting etc., involved unacceptable risk. The internationally accepted conventions for the conservation and
preservation of valuable historic buildings, of which the Pisa Tower is one of the best known and most treasured,
require that their essential character should be preserved, with their history, craftsmanship and enigmas. Thus any
intrusive intervention on the Tower had to be kept to an absolute minimum and permanent stabilisation schemes
involving propping or visible support were unacceptable and in any case could have triggered the collapse of the fragile
masonry.
In 1990 the Italian Government appointed an International Committee for the safeguard and stabilisation of the
Tower. It was conceived as a multidisciplinary body, whose members were experts in arts, restoration and materials,
structural engineers and geotechnical engineers. After a careful consideration of a number of possible approaches, the
Committee adopted a controlled removal of small volumes of soil from beneath the north side of the Tower founda-
tion (underexcavation). This technique provided an ultra soft method of increasing the stability of the Tower, which is
completely consistent with the requirement of architectural conservation.
The paper reports the analyses and experimental investigations carried out to explore the applicability of the
procedure for the stabilisation of the Leaning Tower of Pisa. All the results having been satisfactory, the actual
underexcavation of the monument was carried out in the years 1999-2001; the results obtained are presented and
discussed.
Key words: leaning instability, leaning tower of Pisa, numerical modelling, physical modelling, stabilisation works
(IGC: E3)
'l Professor of Geotechnical Engineering, Imperial College of Engineering and Medicine, London and Member of the International Committee
for the Safeguard of the Leaning Tower of Pisa.
nJ Professor of Geotechnical Engineering, Technical University of Torino, and Member of the International Committee for the Safeguard of the
Leaning Tower of Pisa (sgi_jamiolkowski@studio-geotechnico.it).
mJ Professor of Geotechnical Engineering, University Federico II, Napoli, and Member of the International Committee for the Safeguard of the
Leaning Tower of Pisa.
Manuscript was received for review on August 8, 2002.
Written discussions on this paper should be submitted before May 1, 2004 to the Japanese Geotechnical Society, Sugayama Bldg. 4F, Kanda
Awaji-cho 2-23, Chiyoda-ku, Tokyo 101-0063, Japan. Upon request the closing date may be extended one month.
63
c:
0
"_;j '"d
0 0 (I)
00
£]0..
-
t-
N
\0
M
.... c: E
~ uo
u:l
0
0
,,,,
(I)
;><I
•o------o-o
In year 1990:
Rigid tilt= 5° 30' 4 , to 6 ,
Rate of rigid tilt = year
Fig. 4. Detail of the bell chamber
load by hanger
~
Initial tilt
of tower= 0.5° 0
s
\0
a.
'
I
:
sand bed
' .~ •, ....
.
20m
..
Undrained clay
(Elasto-plastic)
Tresca model - Su = 80 kPa
Fig. 7. Set up of the experiments by Edmonds (1993)
e.~
0.10 f- - .8 30
<1.)
o i3
g.
Q.)
0.05 f- o a .... -a--.,p__-"'& - ~
0
20
en 0 ~-·•
........ B.--~
I~ I
oo
I ".+j
s::
0
til 10
0 50 100 150 200
load (N) ~
oo
0 50 100 150
Fig. 8. Results of the experiments by Edmonds (1993)
Weight of tower, MN
Fig. 10. Load-settlement response of the tower on soft and stiff clay
with a diameter of 102 mm was placed on top of a very
loose fine sand bed, and loaded through a hanger at a
height of 126 mm over the base. The ratio 126/102 is
approximately equal to the ratio of the height of the shear stiffness G: viz. G/su= 10 and G/su= 1000; in both
centre of gravity of the Tower of Pisa to the diameter of cases the strength of the soil Su was 80 kPa. The results
its foundation. show that failure occurs abruptly with little warning and
Loading the model tower produced a settlement and a that the weight of the Tower at failure is dependent on the
rotation a. A total of 8 load tests were carried out; the shear stiffness of the soil. The failure load with G / Su = 10
load at failure varied between 120 and 190 N. Failure in is about half of that with G/su = 1000.
all cases was by toppling with the lowest edge of the Figure 11 shows vectors of incremental displacement
model tower's base sinking into the sand as the Tower for the soft soil, in the last increment of the analysis. The
rotated toward horizontal. The individual plots of a displacements are located in a zone below the foundation
varying with load give somewhat variable results, but and indicate a rotational type of failure. At first sight,
when combined into one plot, as in Fig. 8, a well defined this looks like a plastic type collapse mechanism.
envelope of results emerges. The envelope shows a However, examination of the zone in which the soil has
pronounced change in curvature at a load of 160 to 165 gone plastic, also shown in Fig. 11, indicates that it is very
N, where the inclination averages 0.09 (a:::::: 5°). small and not consistent with a plastic failure mechanism.
Potts and Burland (2000), using Finite Element analy- Consequently, this figure indicates a mechanism of
sis, investigated the differences between leaning instabil- failure consistent with leaning instability.
ity and the usual bearing capacity failure. Figure 9 shows Vectors of incremental displacement just before
the plane strain problem of a Tower resting on a uniform collapse for the stiffer soil are shown in Fig. 12. They
deposit of undrained clay. The clay is modelled as a linear indicate a traditional "general failure" type mechanism
elastic-perfectly plastic Tresca material with an un- with the soil being pushed outwards on both sides. The
drained strength Su = 80 kPa. The Tower was given an plastic zone is very large and therefore the results clearly
initial tilt of 0.5°; the self weight was then increased in a indicate a plastic bearing capacity type mechanism of
large displacement finite element analysis. Figure 10 failure.
shows the results obtained in two analyses with different
0
Plastic zone: GIS"= 1000
0
Fig. 12. Vectors of incremental displacement and extension of plastic kcx
zone in stiff clay at failure
The equation shows that there is no coupling between
settlement and rotation. For a bed of Winkler type
springs the expressions of the stiffness are:
, • _l
I
~
Bl+B3 ,..
'~ model
_.;--,.......
-16.0-
B4+B5 \Js ~
- - __ 1"'----
'--..:S.:
.._
-22.5-
B6
-24.5- -
~s
\ -
--- ---.:11:
model the Tower. In any case, even the very simple linear- case like that of the Tower of Pisa, that is on the verge of
ly elastic models allow an important practical conclusion: instability, consideration of non-linearity appears man-
the Tower is very nearly in a state of neutral equilibrium. datory. To this aim, the knowledge of the relationship Ms
For most of the 20th century the inclination of the Tower = Ms(a) is necessary.
has been steadily increasing, but the changes of inclina- Centrifuge modelling of the Tower and its subsoil was
tion have been extremely small compared with those that carried out at ISMES; the results obtained are reported
occurred during and immediately following the construc- and discussed by Pepe (1995). They gave further insight
tion. The rate of inclination of the Tower in 1990 was into the mechanisms of the instability and confirmed the
about six arc seconds per year. The cause of the con- elastoplastic character of the restraint exerted by the
tinuing movement is believed to be a phenomenon of foundation and the subsoil on the motion of the Tower.
ratchetting, triggered by the fluctuations of the water In Fig. 14 the properties of the foundation soils of the
table in Horizon A; it is made possible by the fact that the Tower are compared with the properties obtained in the
Tower was substantially in a state of neutral equilibrium. small scale model after consolidation under geostatic
It is to be noted that a state of neutral equilibrium is load in the centrifuge; the main features of the soil profile
possible, in principle, only if the system is a linear one; in are satisfactorily reproduced in the model. Figure 15
the case of non-linearity, neutral equilibrium is but a shows the simulation of the construction of the Tower, as
transition stage to overturning. Of course, creep has also obtained from one of the centrifuge tests. In the first
some effects on the process. stage of the construction, the model Tower was con-
strained to settle without rotation. In the second stage the
model Tower was left free to rotate. It may be seen that
NON -LINEARITY both the settlement and the rotation, scaled to the proto-
The relationship between Ms and a may be linearised type, are in good overall agreement with those of the
over a short interval, but it is certainly non-linear and Tower.
approaches a limiting value of Ms asymptotically. In a Finite element analyses of the behaviour of the Tower
and its subsoil have been carried out using a finite element The layers of the finite element mesh matched the soil
geotechnical computer program developed at Imperial sub-layering that had been established from soil explora-
College and known as ICFEP (Potts and Gens, 1984). tion studies. Figure 16(a) shows the adopted mesh, while
The constitutive model is based on Critical State concepts Fig. 16(b) shows the detail of the mesh in the immediate
and is non-linear elastic work hardening plastic. Fully vicinity of the foundation. In Horizon B the soil is
coupled consolidation is incorporated, so that time effects assumed to be laterally homogeneous; however, a tapered
due to the drainage of pore water in the soil skeleton are layer of slightly more compressible material was incorpo-
included. rated into the mesh for layer A 1 as shown by the shaded
The prime object of the analysis was to improve the element in Fig. 16(b). This tapered layer represents a
understanding of the mechanisms controlling the behav- more clayey material found beneath the south side of the
iour of the Tower (Burland and Potts, 1994). Accord- foundation; in applied mechanics terms, this slightly
ingly, a plane strain approach was used for much of the more compressible tapered layer may be considered as an
work, and only later was three dimensional analysis used "imperfection". The overturning moment generated by
to explore certain detailed features. the lateral movement of the centre of gravity of the
Tower was incorporated into the model as a function of
Vertical prototype load, t the inclination of the foundation, as shown in Fig. 16(b).
The construction history of the Tower was simulated by a
0 2500 5000 7500 10000 12500 15000
0.0
- r--- ~ .......
OQ I
'0'
0\:
""'"
0\
""'"
:::!;
series of load increments applied to the foundation at
suitable time intervals. The excavation of the catino in
0.5 1838 was also simulated in the analysis. Calibration of
s 1.0
st . *
"i_l the model was carried out by adjusting the relationship
between the overturning moment generated by the centre
1::"
(!)
___ I__ ~~~~~~~1!9~~!~¥:____ ~ of gravity and the inclination of the foundation. A
s 1.5 'r-].4]
.......
number of runs were carried out with successive adjust-
~
(!)
'E
(!) 2.0 ments being made until good agreement was obtained
l
(/)
between the actual and the predicted present day value of
nd rd .
2.5 - 2 and 3 construction stages* the inclination.
--2.661-------r-----T------,------- -----' Figure 17(a) shows a graph of the predicted changes in
3.0 inclination of the Tower against time, compared with the
(*) See Fig. 3
deduced historical values, the same data are plotted in
Fig. 15. Typical centrifuge simulation of the construction of the Fig. 17(b) in the style of Fig. 6.
Tower From about 1272 onwards there is a remarkable agree-
a)
w
""M
~, M = W ha sin 9 Ic
~
North South
b)
- A
V/..0 V//// 07///h-~
~ w~ W/~ W.r)0"/h-
~
................
' - f-l.._.....- ./"'"
A2
Bl
B2
,[.;'--........, B3
lL ~....._,./ ........... . / ~
B4
/ ~
Fig. 16. Finite Element model: a) general mesh, b) detail of the mesh in the vicinity of the Tower
00
('<') 0.6
~
6 N 0
<ll
a> r- \0
~ ~ 0.5
~5
~ 4 s
z 0.4
~n
s::
0 *deduced
-~ !:f
Q)
0.3
'"0
§ s0
~
0 s 1 ~ 0.2
~ 2
s
a> 0.1
~ 3
tf.l
t- M
Fig. 18. Centrifuge experiments by Cheney et al. (1991)
..... 0\
0 00 0\
~ T T
----~-------- ---o- -CJ STABILISING MEASURES
If an elasto-plastic model is assumed, then the relation
between loads and displacement has to be written in
incremental form (Desideri et al., 1997):
Loading
'"gn Consolidation
0 0 Historical
.....:!
•c 2
D} FEM
3D
kap ka
There is a coupling between settlement and rotation; the
increment of displacement depends on the load incre-
ment, the current state of load and the load history.
0 2 3 4 5 6 Hence the factor of safety depends on the current state of
Inclination of foundation eo stress and stress history, and decreases with increasing
inclination.
Fig. 17. Comparison between deduced historical values of the Tower The centrifuge experiments by Cheney et al. (1991,
inclination and values predicted by the FEM model
Fig. 18) show coupling between settlement and rotation,
together with non-linearity and strain hardening plastici-
ty. It may be seen that a decrease of inclination strongly
ment between the model and the historical inclination. increases the stiffness of the foundation-ground system.
Note the excavation of the catino in 1838 which results in This generated the idea that a decrease of inclination
a predicted rotation of about 0. 75 o. The final imposed in- increases Ms and FS, and can be used to stabilise the
clination of the model tower is 5.44 o, slightly less than the Tower. This has been the basic approach of the Commit-
1993 value of 5.5°. It was found that any further increase tee to the stabilisation of the Tower; among other things,
in the final inclination of the tower model resulted in it is completely respectful of its integrity. In fact, it may
instability: a clear indication that the Tower is very close be said that movements are in the DNA of the Tower!
to falling over. An important experimental confirmation of this ap-
The analysis has demonstrated that the lean of the proach came from a temporary safety measure taken to
Tower results from the phenomenon of settlement insta- prevent overturning of the Tower. The Committee soon
bility due to the high compressibility of the Pancone recognised the need for temporary and fully reversible
Clay. The principal effect of the layer of slightly increased interventions for the improvement of the safety against
compressibility beneath the south side of the foundation overturning by foundation failure, and for gaining time
is to determine the direction of lean, rather than its to complete the investigations and analyses necessary to
magnitude. The main limitation is that the model does conceive and implement the final stabilisation measures.
not deal with creep. Nevertheless, the model provided The observation that the northern side of the Tower
important insights into the basic mechanisms of behav- foundation had been steadily rising for most of the 20th
iour and later it proved invaluable in assessing the effec- century led to the suggestion of applying a counterweight
tiveness of various proposed stabilisation measures. on the north side as a temporary measure. Accordingly, a
design was developed consisting of a prestressed concrete
ring cast around the base of the Tower for supporting a
Ring I
11 st 2nd 13rd 4th Lift 4th Lift I
54
I
Hit~.
,. ,~ ,, .;,.,,
construction Lift Lift Lift 1st Phase 2nd Phase
~
....
.... ~"
ll~l Ill 111M II
._" .. ~
30
""'
v D'"
(*) 3/5/93 readings taken as a reference _
A~ ~
• Inclinometer readings
-
22
14
,p C High precision levelling . 15 external
benchmarks -
~
6 ,_.,Jl A High precision levelling . 8 internal
-
-2
.... -~
I
benchmarks
I I I I
--
'<:t
--- --
'<:t
M M M M M 'I) 'I)
~ 0"1 0"1
~
0"1 0"1 0"1 0"1 0"1 0"1 0"1 0"1 0"1
~ \0 00 0 N N \0 00 0 N N ~
0 0
- -
0 0 0 S2 0
§5
N
0
N
- - S2
-
0"1 00 r- 'I) \0
Date
'I) ~
......
M N 0 :::::.
Fig. 19. Behaviour of the Tower during and after the application of lead weights
k = 6 ·96MNx 6 · 3m = 188423MNm
a 48"
The factor of safety was thus increased to:
Fig. 20. Comparison between prediction and measurements of
FS= __Is:_ = 188423 = 59 changes in inclination and average settlement of the Tower during
Wh 141.8 X 22.6 the application of lead ingots
where h = 22.6 m is the height of the centre of gravity of
the Tower. Of course, the only meaning of similar 80% of the measured values.
computations is to show that there is a substantial im- The refinement of the model involved a small reduction
provement of the stability of the Tower. In this respect, it of the value of G /p~ in horizon A. After that, a better
is noteworthy that, after a small decrease of the inclina- overall agreement between computed and observed
tion, the Tower has remained essentially motionless for values was obtained (Fig. 21).
over three years, apart from the seasonal cyclic move-
ments and some accidents in September 1995.
The observed behaviour of the Tower during the coun- UNDEREXCAVATION
terweight application was used to check and refine the Investigations
finite element model further. Figure 20 shows a compari- Among different possible options to obtain a decrease
son of the Class A prediction and measurements of (a) the of the inclination, the Committee has explored:
changes in inclination and (b) the average settlement of 1. surface loading north of the Tower;
the Tower during the application of the lead ingots. The 2. vacuum pumping from the Pancone clay north of the
computed settlement is in good agreement with the meas- Tower;
ured value; the predicted changes in inclination are about 3. electroosmosis;
Time, min
a) 0 400 800 1200 1400 2000
2~--~----~--~----~--~
0
til
d)
d)
~ 2
0.0
d)
""d
s::" 4
0
·~
--
~
0
6
8
Displacement (*) See Fig. 3
b)
transducer 10
Time, min
200 400 600 800
Accelerometer
Clayey Silt
~«~----~--+-~~~~~~
""d
I
s::"
0
Upper Pisa Clay ·~ 46
~
Sand
c)
50~--~----~----~----~
Fig. 22. Simulation of the underexcavation in a centrifuge test: a) Model of the Tower, b) Construction of the Tower and c) Underexcavation
excavating elements 10, 9, 8, 7 and 6 successively. For stress take place, as is to be expected, but the stress
each step the response of the Tower was positive; changes are small. In general, the stress distribution after
• the whole process of insertion and retraction of the retraction of the drill are smoother than after insertion.
drill probe was then repeated. Once again excavation The results of the physical and numerical modelling
of element 12 gave a negative response. work on underexcavation were sufficiently encouraging to
The computed displacements of the Tower are plotted in undertake a large scale development trial of the field
Fig. 23. The sequence of excavation of the elements is equipment. The objectives of the trial were:
given on the horizontal axis; the upper diagram shows the • to develop a suitable method of forming a cavity
change of inclination of the Tower due to underexcava- without disturbing the surrounding ground during
tion; the lower diagram shows the settlement of the north drilling;
and south sides of the foundation. • to study the time involved in the cavity closure;
As underexcavation progresses from elements 6 • to measure the changes in contact stress and pore water
through 11, the rate of change of northward inclination pressures beneath the trial footing;
increases as do the settlements. As the drill is retracted the • to evaluate the effectiveness of the method in changing
rate decreases. At the end of the first cycle of insertion the inclination of the trial footing;
and retraction of the drill, the inclination of the Tower is • to explore methods of "steering" the trial footing by
decreased by 0.1 o. The settlement of the south side is adjusting the drilling sequence;
more than one half of the north side. For the second cycle • to study the time effects between and after the opera-
a similar response is obtained but the change of inclina- tion.
tion is somewhat larger. After the third insertion of the For this purpose a 7 m diameter eccentrically loaded
drill the resultant northward rotation is 0.36°. The corre- circular reinforced concrete footing was constructed in
sponding settlements of the north and south sides of the the Piazza north of Baptistry, as shown in Fig. 24. Both
foundation are 260 mm and 140 mm respectively. the footing and the underlying soil were instrumented to
As regards the contact stress distribution, the process monitor settlement, rotation, contact pressure and pore
results in a slight reduction of stress beneath the south pressure.
side. Beneath the north side, some fluctuations in contact After a waiting period of a few months, allowing the
0.4
v•
!::f
0
·.g 0.3 North Critical line
roV~
0
;.:::l:'l
(..)~
~~
0~
0
bO
0.2
tz
~~
.... ~~
~~
).0~
............... I
! I
a
~n
I
0.1 1
Insertion
Io Retraction
•
o~Y. /'
..s::l )>,1.
u
,...1~ r-r--r-11111111111 Ll' 11
0
0 ,~
I I I I I I I I
/
- 1114
~
0
~
r8E
~E
100
!;'.~""'
,,
""~"
~~ lt.L.r.
1111111
South edge
"'
v--
-""' /-- -, /'-
.....,..s::l"
$:~.....,
0$:1
200 ,1"'1
s:.:::
~0.
North edge
~7
......
...... 300 Elements 6 to 12 underexcavated
0
r.n Volumetric strain in each of them@ 5%
400
Fig. 23. Finite element simulation of the underexcavation, performed after the history of inclination up to 5~ 0 has been simulated
conventional
North
+9-
Fig. 25. Vertical section of the large scale trial test of underexcavation
5000
, """' Settlement
difference
lA" I~ fQQ T between I-- "SOUTH"
0
~
.....0 4800 ~I \ T gauge 1 and 3 I ~-1---
~
=I I=_ 1-- I--
0 High
1-- ~- ~-1--
1:: 4600
0
~ 11 0
~J .. I
I
I
I
I I
T I
I
I
I
I
.,.,
I
I
.I
N '<!" \0 r-- 00
~
~
~
M
INSERTION 9
~ ~ ~
'=? ~
0
0 N N
N
N ~ ~ 0
:%
0
N
Tower
plinth ·~ TOWER-.
+1.86 i~ OPERA
.........~...............iz PRIMAZIALE
o.oo
----'iZ---..!
·B
I~ .--====t 4,5°
~~
i~ 106,300 m (West cable)
-5.47 i
...:2............... 105,300 m (East cable)
I
E
t
8 1-+-+-J.-It-+-+-+-l~
16 1-+-+-J.---+T-+-+-+-lc-~
~ 24 1-+-+-J.-It-+-+-+-lf--~
~ 321-+-+-4~+-+-+-lf--
0 40 1-+-+-J.--k--+-++--lf--·.::;
48 1--+-+-J.~+-++--If--
56 f-++-1-H-+-+-• •A''-
64
72
~=:" 80
0 88
-~ 96
~ 104
0 112
~ 120
128
N 136 f-++-1--+'-+-+~'-+--
Time
2 f-+-+--1--fL+-+-+-~N
§ 3 1-+-+-+--IH--.J-+-+-00
4 1--+--1--+-1+--t-+-+-lf--·i'i
l:
(1)
'E 7
1-+--1-1---*-+-+-+--lf--
(!) 8 ~+-J.--It-~-1-1-~
(/) 9 ~+-+-lt-+-+--1-1-
10 1--+--1--4--h-+-+-+-+-.--{
11 1--+--1-+--+'--+-+-+--l~
12 ~+--+-4J-+-+-+-I-+-
Time
........ - .....--._
]'--,
E
v :\
1/ !WEST-EAST
~~ /
l.-1/
1/l
~ 200
v
v /' ) ~ 100
......-- v ...
__..,.. v I
f.-- Lead we1ght I- 1---1 NORTH-SOUTH 1---
r-::e~;~~:~~: /
I'--
~
5
~,.
/
.
0
........ j;""
~ II PRELIMINARY UNDEREXCAVATIONl
N
~ ~;; $ ~ ~ ~ w
;; ~ i ~ 8 §
Year
I §
~ Year
12/01/99
29/12 98
26/01/99
--'><"
=
Q
ii'
(ll
23/02/99 23/02/99
06/04/99
23/03/99
20/04/99
=-...
18/05/99 18/05/99 ~
15/06/99 .c:.nu 01 prenm1nu.ry unue;;n:::xcuviiuun otuurf:;r
...
(JQ
Q
29/06/99 13107199 =
=
10/08/99 I 0/08/99 :Z ';"l =-
§ ~
tD
21/09/99
07/09/99
05110/99
·"'tt
5: ::r: ~
...~ ~
s.s
?::l
~
02/11/99 ~ ~p f-- s·~ OS.~
::~~: l--t--i--J--J---t--i--J~-~-?-S
~- p.Cll
·~ ~ = C1l p.
~ 30/11/99 (') (')
el 0
;-l 28/12/99 s·~
25/01/00 f--- - -~--- - ·-- - -r--- - -~-- '"-~--x ~a
C1l
~ 22/02/00 a' to
c
~
(ll
o7/o3/ooL
~
(ll
21/03/00 10
t""'
....,
Q 18/04/00 18/04/00 >
z
...., 30/05/00
16/05/00
t:::l
S: 13/06/00
...,
ttl
=
= ll/07/00 11107/00
=-
tD 22/08/00
08/08/00 >
t""'
~ 05/09/00
~ 03/10/00 03/10/00
~
< 31/10/00
~ 14/11/00 ~
s· 28/11/00 ~
~
= 26/12/00 26/12/00 ~
?' ~
r
23/01/01 2
06/02/01 g'
II I ~ >
:r::;~: . -it : ~
20/02/01
20/03/01 =-
Q
ii'
01/05/01
1111111111111
....,
12/06/01 12/06/01 ...
Q
~ lit'
NII-Electronic Library Service
10/07/01 N
24/07/01 0
07/08/01 (JQ
04/09/01 a 0
s
04/09/01
02/10/01
=
=
16/10/01
30/10/01
=-
tD
-200
~....
~
"'"·~•...
400
"' -600 ~
§ ', Radial drain
~ -800
~-- -1000
··.... '.
i -1200 ·. .r--,.. Planned excavatiOn rate
at 24 4"/week
z
§
-1400
~ -1600
··. '...... ' r--
Perimeter
~ -1800
-2000
Cresy & Taylor, 1817 (-2000")
1-. .....
--~- - ··- of the catino
~
~ I I I 2000
~
~
~
~
Year
I I 2001
~
~
~
~
~
~
~
~
~
~
2002
I \ ...........
•
E
100
0
100
200 I -
300 I -
400 I -
,I
A"'-
,-=
EAST-WEST PLANE
+-
\
\ Discharge
Central drain
~
'o
soo I -
600 f.- ' conduct i=6% ~
tmJ
... c
i ~
700 f.-
800
900
e-
e- 0 I\ ••
" 1000 e- NORTH-SOUTII PLANE\
Radial drain
.9 1100 e- ! \ (i=l%)
!
1200 f.-
1300
1400
ISOO
~
I
CJ A: PRELIMINARY UNDEREXCAVATION, 08/02199 TO 06/06/99 I
'
1600 m!1!1 8: FULL UNDEREXCAVATION, 21/02/00 TO 19/01/01 I Fig. 42. Control of water table on north side-cross-section
•
N 1700 . . C: LEAD INGOTS REMOVAL, 14/09100 TO 16/01/01 I I\
B
1800 .\.
1900
$ ~ ~
~
~ ~ ~ § ~ ~ ~ ~ ~ ~
$
~ ~ ~ ~ i a ~ ~
Year
~
$
~ :s I ~ § ~ ~ §
CONCLUDING REMARKS
Fig. 39. Rotation of the Tower from 1993 to 2001 The stabilisation of the Tower of Pisa has been a very
difficult challenge for geotechnical engineering. The
Tower is founded on weak, highly compressible soils and
its inclination had been increasing inexorably over the
I J-r-
~· 1/
--
]{_
years to the point at which it was about to reach leaning
~ -200
. all l I I
I I I I
/ 1/ I
JI instability. Any disturbance to the ground beneath the
<c; -400
I LEAD WEIGHT W I II
south side of the foundation is very dangerous. Therefore
]-600 ~§ I I I I
~-800 '8e~•
I I I I I the use of conventional geotechnical processes at the
PRELIMINARY UNDEREXCAVATION r
g -1000 'Qg 1/ south side, such as underpinning, grouting, etc., involved
-~
~
-1200 ra
ooS- f-€1
I
II FULL UNDEREXCAVATION ) - f--J
I I
unacceptable risk.
-1400
~ ....
I I I The internationally accepted conventions for the con-
I I I
0
servation and preservation of valuable historic buildings,
"' of which the Pisa Tower is one of the best known and
Year
most treasured, require that their essential character
Fig. 40. Rotation of the Tower from 1911 to 2001 should be preserved, with their history, craftsmanship
and enigmas. Thus any intrusive interventions on the
Tower had to be kept to an absolute minimum and
below the bottom of the catino has been implemented. permanent stabilisation schemes involving propping or
The system consists of three wells (A, Band C, Fig. 41) visible support were unacceptable and in any case could
drilled to the north of the Tower, about one metre have triggered the collapse of the fragile masonry.
outside the perimeter of the catino. Five radial sub- The technique of underexcavation provided an ultra
horizontal drains reaching the drainage layer below the soft method of increasing the stability of the Tower,
catino have been drilled from each well (Fig. 42). The which is completely consistent with the requirements of
wells are connected each other by an outlet pipe, dis- architectural conservation. Different physical and numer-
charging into a nearby sewer after having crossed two ical models have been employed to predict the effects of
more wells, having only a constructional purpose. soil removal on the stability. It is interesting to point out
The drainage system has been implemented in April that some mechanisms (as, for instance, the occurrence
and May, 2002. The decrease of the pore pressure con- of a critical line beyond which the underexcavation
nected with it has induced a further northward rotation becomes dangerous) are predicted by physical modelling
of the Tower, that can be seen in the diagram of Fig. 39. and by the FEM analyses, while these are missed by the
simplified Winkler type models. The preliminary un-
derexcavation intervention, only undertaken once the 6) Edmunds, H. (1993): The use of underexcavation as a means of
Commission was satisfied by comprehensive numerical stabilising the Leaning Tower of Pisa: scale model tests, MSc
Thesis, Imperial College, London.
and physical modelling together with a large scale trial, 7) Gorbunov-Posadov, M. I. and Serebrajanyi, R. V. (1961): Design
has demonstrated that the Tower responds very positively of structure upon elastic foundations, Proc. VICSMFE, Paris, 1,
to soil extraction. The final underexcavation has attained 643-648.
the target of reducing the tilt of the Tower by half a 8) Habib, P. and Puyo, A. (1970): Stabilite des fondations des
degree. The drainage system north of the Tower has constructions de grande hauteur, Annales ITBTP, 117-124.
9) Hambly, E. C. (1985): Soil buckling and leaning instability of tall
further slightly reduced the tilt and will mitigate the structures, Struct. Eng., 63(3), 78-85.
effects of the fluctuations of the water table. 10) Jamiolkowski, M. B. (2001): The Leaning Tower of Pisa: end of an
It is believed that geotechnical stabilisation has been Odessey, Terzaghi Oration Proc. XV ICSMGE, Istanbul, 4,
finally attained; monitoring the behaviour of the 2979-2996.
Monument for the forthcoming few years will confirm it. 11) Lancellotta, R. (1993): Stability of a rigid column with non linear
restraint, Geotechnique, 41(2), 243-256.
12) Pepe, M. (1995): La Torre Pendente di Pisa. Analisi teorico-
sperimentale della stabilita dell'equilibrio, PhD Thesis, Politecnico
REFERENCES
di Torino, 265.
1) Burland, J. B., Jamiolkowski, M. and Viggiani, C. (1999): The 13) Potts, D. M. and Burland, J. B. (2000): Development and applica-
restoration of the Leaning Tower: geotechnical aspects, Workshop tion of a numerical model for simulating the stabilisation of the
on the Restoration of the Leaning Tower of Pisa, Pre-print 105. Leaning Tower of Pisa, Proc. John Booker Memorial Symp. on
2) Burland, J. B. and Potts, D. M. (1994): Development and applica- Developments in Theoretical Geomechanics (eds. by Smith and
tion of a numerical model for the Leaning Tower of Pisa, IS Carter), Sidney, 737-758.
Prefailure Deformation Characteristics of Geomaterials, Hok- 14) Potts, D. M. and Gens, A. (1984): The effect of the plastic potential
kaido, Japan, 2, 715-738. in boundary value problems involving plane strain deformation,
3) Cheney, J. A., Abghari, A. and Kutter, B. L. (1991): Leaning Int. J. Analytical and Numerical Methods in Geomechanics, 8,
instability of tall structures, J. Geotech. Engrg., Proc. ASCE, 259-286.
117(2), 297-318. 15) Schultze, E. (1973): Der Schiefe Turm von Pisa, Mitteilungen
4) Desideri, A. and Viggiani, C. (1994): Some remarks on the stability Technische Hochschule Aachen.
of Towers, Symp. on Development in Geotech. Engrg., Bangkok. 16) Terracina, F. (1962): Foundation of the Leaning Tower of Pisa,
5) Desideri, A., Russo, G. and Viggiani, C. (1997): La stabilita di torri Geotechnique, 12(4), 336-339.
su terreno deformabile, RIG, XXXI(l), 5-29.