Prof John Burland

PISA GOES CRITICAL

by J.B.Burland FREng, FRS

Professor of Soil Mechanics, Imperial College of Science,
Technology and Medicine, London, UK.

Introduction

Imagine a tower founded on ground which has the consistency of jelly or foam rubber to great depth.  The inclination of the tower is increasing inexorably to a point where it is about to fall over.  Any disturbance to the ground on the leaning side will cause it to topple.  Worse still, the material composing the tower is so fragile that the stresses caused by its lean are close to causing structural failure.  This picture represents the state of the Leaning Tower of Pisa and demonstrates why its stabilisation represents the ultimate civil engineering challenge.

In 1989 the civic tower of Pavia collapsed without warning leading to the closure of the Pisa Tower.  There was an immediate outcry by the Mayor and citizens of Pisa who foresaw the damage that the closure would inflict on the economy of Pisa.  In March 1990 the Prime Minister of Italy set up a Commission, under the chairmanship of Professor Michele Jamiolkowski, to develop and implement measures for stabilising the Tower.  It is the sixteenth commission this century and its membership covers a number of disciplines.  

     It is not widely appreciated that the decree establishing the Commission has never been ratified.  In Italian law a decree has to be ratified  by the Italian Parliament within two months of publication or else it falls.  Thus, every two months, the Commission’s decree has to be renewed and on a number of occasions the work has been suspended because of delays in  renewal.  Such an arrangement makes the Commission very vulnerable to media and political pressures and long-term planning is very difficult.  

Details of the Tower and of the Ground Profile

Fig. 1 shows  a cross-section through the 14,500-tonne Tower which is a campanile for the nearby cathedral.  It is nearly 60m high and the foundations are 19.6m in diameter.  The foundations are inclined due south at about 5.5^o and the seventh cornice overhangs the ground by about 4.5m.  The inner and outer surfaces of the cylindrical tower are faced with marble but the annulus between these facings is filled with rubble and mortar within which extensive voids have been found.  A spiral staircase winds up within the annulus.  

The ground  underlying the Tower (section below, fig 2) consists of three distinct layers.  Layer A is about 10m thick and primarily consists of soft estuarine deposits of sandy and clayey silts laid down under tidal conditions.  Based on sample descriptions and cone tests, the material to the south of the Tower appears to be more clayey than to the north making it slightly more compressible.  Layer B consists of soft sensitive normally consolidated marine clay, known as the Pancone clay, which extends to a depth of about 40m.  This material is very sensitive and loses much of its strength if disturbed. Layer C is a dense sand which extends to considerable depth. The water table in Horizon A is between 1m and 2m below ground surface. The surface of the Pancone Clay is dished beneath the Tower showing that the average settlement is between 2.5m and 3.0m - a good indication of how very soft the ground is.  

Fig 1: cross-section through the Tower

Fig 2: The ground underlying the Pisa Tower

History of the Tower

Work on the Tower began on 9th August 1173.  By about 1178 construction had progressed to about one quarter of the way up the fourth storey when work stopped.  The reason for the stoppage is not known but had it continued much further the foundations would have experienced a bearing capacity failure within the Pancone Clay.  The work recommenced in about 1272 by which time the strength of the clay had increased due to consolidation under the weight of the Tower.  By about 1278 construction had reached the 7th cornice when work again stopped - possibly due to military action.  There can be no doubt that, had work continued, the Tower would have fallen over.  In about 1360 work on the bell chamber was commenced and was completed in about 1370 - nearly 200 years after commencement of the work (fig 3).  

Fig 3: Phases of the construction of the Pisa Tower

The axis of the Tower is not straight, showing that the masons made corrections for its changing inclination from very early on in its construction.  The geometry of the Tower has been used to reconstruct the history of inclination (fig 4).  Initially the Tower inclined slightly to the north amounting to about 0.2^o in 1272 when construction recommenced.  As construction proceeded the Tower then began to move towards the south at an increasing rate.  In 1278, when construction had reached the seventh cornice, the tilt was about 0.6^o.  During the 90 year pause, the tilt increased to about 1.6^o.  After the completion of the bell chamber in about 1370 the inclination of the Tower increased dramatically.  In 1817 plumb line measurements made by two British architects Cressy and Taylor gave an inclination of about 5^o. Further measurement by the Frenchman Ruhault de Fleury in 1859 which showed that the excavation of the catino by Gherardesca in 1838 caused a significant increase of inclination to about 5½^o.

This history of tilting has been used to calibrate a numerical  model of the Tower and underlying ground.  Fig.5 (below) shows a graph of the predicted changes in inclination of the Tower against time, compared with the deduced historical values. 

Fig 4: The inclination history of the Pisa Tower

Fig 5: The relationship between time and inclination for the computer simulation of the history of the Pisa Tower

The model does not simulate the initial small rotation of the Tower to the north but from about 1272 onwards there is remarkable agreement between the model and the historical inclinations.  Note that it is only when the bell chamber was added in 1360 that the inclination increases dramatically.  Also of considerable interest is the excavation of the catino in 1838 which results in a predicted rotation of about 0.75^o.  It should be noted that the present-day inclination of the model tower is 5.44^o which is slightly less than the actual value of 5.5^o.  It was found that any further increase in the final inclination of the model tower resulted in instability - a clear indication that the Tower is very close to falling over.

The impending instability of the Tower foundation is not due to a shear failure of the ground but can be attributed to the high compressibility of the Pancone Clay.  This phenomenon is called ‘leaning instability’.  No matter how carefully the structure is built, once it reaches a critical height the smallest perturbation will induce leaning instability.  Children building brick towers on a soft carpet will be familiar with this phenomenon!

The model has provided important insights into the basic mechanisms of behaviour and has proved valuable in assessing the effectiveness of various proposed stabilisation measures.

Observed behaviour of the Tower in the 20th Century

For most of this century the inclination of the Tower has been increasing by amounts measured in arc seconds (five arc seconds is equivalent to about 1.5mm movements at the top of the Tower).  The study of these movements has been important in developing an understanding of the behaviour of the Tower and has profoundly influenced the decisions taken by the Commission.

Since 1911 the inclination of the Tower has been measured regularly by means of a variety of precise instruments and Fig.6 (below) shows the change of inclination with time.  It can be seen  that the curve is not smooth but contains some significant ‘events'.  In 1934, Girometti drilled 361 holes into the foundations and injected about 80t of grout with a view to consolidating the masonry.  This activity caused a sudden increase in tilt of 31 arc seconds.  In the late 1960's and early 1970's pumping from the lower sands caused subsidence and tilting towards the south-west of the Piazza.  This induced a tilt of the Tower of about 41 arc seconds.  When pumping was reduced the tilting of the Tower reduced to its previous rate.  These, and a number of smaller events show that the Tower is very sensitive to even the smallest ground disturbance at the south side.  Hence any remedial measures should involve a minimum of such disturbance.  Moreover the rate of inclination of the Tower has been increasing  and in 1990 was about 6 arc seconds per annum.

Fig 6: Change in inclination of the foundations since 1911

Temporary stabilisation methods

As mentioned in the introduction, there are two distinct problems that threaten the stability of the Tower.  The most immediate one is the strength of the masonry.  The change in cross section of the walls at the first floor level combined with the location of the spiral staircase (fig 7) gives rise to stress concentrations at the south side and the marble cladding in this location is cracked.  It is almost impossible to assess accurately the margin of safety against failure of the masonry, but the consequences of failure would be catastrophic.  The second problem is the stability of the foundations against overturning.

The approach of the Commission has been a two stage one.  The first stage has been to secure an increase in the margin of safety against both modes of failure as quickly as possible by means of temporary measures.  Having achieved this, the second stage is to develop permanent solutions recognising that these would require time to carry out the necessary investigations and trials. It is important to appreciate that the prerequisites of restoration work are that it should involve minimum possible intervention, be non-destructive, reversible and capable of being applied incrementally in a controlled manner.

Fig 7: Cross section of the wall at first floor level

Lead weights to the North of the Tower

Permanent stabilisation

In parallel with the temporary operations, a variety of approaches to permanently stabilising the Tower were being explored.  The fragility of the masonry, the sensitivity of the underlying clay and the very marginal stability of the foundations imposed severe restraints and any measures involving the application of concentrated loads to the masonry or underpinning operations beneath the south side of the foundation were ruled-out.  Moreover aesthetic and conservation considerations required that the visible impact of any stabilising measures had to be kept to an absolute minimum.

The Commission opted for a so called ‘very soft' solution aimed at reducing the inclination of the Tower by up to half a degree (which would not be visible) by means of induced subsidence beneath the north side of the foundation without touching the structure of the Tower.  Such an approach allows the simultaneous reduction of both the foundation overturning moment and the masonry overstressing with a minimum of work on the Tower fabric itself.  After careful consideration of a number of possible methods of inducing subsidence, a technique known as soil extraction was chosen.  This technique involves the controlled removal of small volumes of soil from the sandy silt formation of Horizon A beneath the north side of the foundation by means of an inclined drill as shown in Fig 8.  Recently it has been discovered that the method was first used by a British architect in 1832 to stabilise the tower of St Chad in Wybunbury, Cheshire.

Fig 8: Inclined drill for soil extraction

Two key questions had to be addressed:

  1.       Given that the Tower is on the point of leaning instability, is there a risk that extraction of small quantities of soil from beneath the north side will cause an increase in inclination?

  2.       Is the extraction of small volumes of ground in a controlled manner feasible, will the cavities close and what is the response at the soil/foundation interface?

The first question provoked many heated debates.  The numerical model described previously was used to simulate the extraction of soil from beneath the north side of the foundation.  Even though the tower was on the point of falling over it was found that, provided extraction took place north of a critical line, the response was always positive.  Moreover the changes in contact stress beneath the foundations were small.  Similar positive results were obtained by means of advanced physical modelling on a centrifuge.

The results of the modelling work were sufficiently encouraging to undertake a large scale development trial of the drilling equipment.  For this purpose a 7m diameter eccentrically loaded instrumented footing was constructed in the Piazza north of the Baptistry.  Drilling was carried out using a hollow-stemmed continuous flight auger inside a contra-rotating casing.  When the drill was withdrawn to form the cavity an instrumented probe located in the hollow stem was left in place to monitor its closure.  The trials showed that cavities formed in the Horizon A material closed gently and that continued extraction from the same location could be achieved.  The trial footing was successfully rotated by about 0.25^o and directional control was maintained even though the ground conditions were somewhat non-uniform.  Very importantly, an effective system of communication, decision taking and implementation was developed.

The soil extraction trial was successfully completed in March 1996 and it was argued that  there was nothing more that could reasonably be done to prove the effectiveness of soil extraction short of testing it on the Tower itself.  After much debate the decision was taken by the Commission to carry out preliminary soil extraction beneath the north side of the Tower itself with the objective of observing the response of the Tower to a limited and localised intervention.  A safeguard structure (fig 9) was to be constructed in the form of a horizontal cable stay attached to the Tower at the third storey which could be tensioned to steady the Tower in the event of detrimental movements.

Fig 9: Temporary safeguard cable

Preliminary soil extraction

In August 1996 the decree giving the Commission its mandate fell and was not renewed so that further work ceased.  In December 1996 the Commission was disbanded and a new one (the seventeenth) was established with a substantial change of membership.  The new Commission first met in June 1997 and it was decided that a complete review of all possible stabilisation measures should be undertaken.  The case for soil extraction had to be argued again from scratch.  In July 1998, after a year of heated debate, the decision was again taken to implement preliminary soil extraction to assess its effectiveness in reducing the inclination of the Tower.

By December 1998 the temporary safeguard cable stays were in place.  Preliminary soil extraction was to be carried out over a limited width of 6m using twelve boreholes.  A target of a minimum of 20 arc seconds reduction in inclination was set as being large enough to demonstrate unequivocally the  effectiveness of the method.. 

Because of the extreme delicacy of the project only 20 litres of soil were to be extracted every two days with continuous real time monitoring to assess the results.  A strict system of command and control has been established in which each extraction operation requires an instruction based on the response of the Tower and signed by the Responsible Officer.

On 9^th February 1999, in an atmosphere of great tension, the first soil extraction took place.  For the first week the Tower showed no discernable response but during the following tense days it began very gradually to rotate northwards.  As confidence grew the rate of soil extraction was increased.  At the beginning of June 1999, when the operation ceased, the northward rotation was 90 arc seconds and by mid-September it had increased to 130 arc seconds which is equivalent to about 40 mm at the top.  At that time three of the 97 lead ingots (weighing about 10t each) were removed.  Since then the Tower has exhibited negligible further movements.  

Preliminary soil extraction has been demonstrated to produce a positive response and the Commission has now formally approved the application of the method for permanent stabilisation.  After lengthy bureaucratic delays the decree has been extended until the middle of 2001.  Using 41 extraction tubes, work on the full intervention will commence at the end of February 2000.  It is estimated that it will take about two years of careful soil extraction to reduce the inclination of the Tower by about half a degree which will be barely visible.  It has also been necessary to carry out some strengthening of the masonry at the south side of the second storey.  After ten years of work the first positive steps have been taken but there is a long tense journey still ahead of the Tower - and the Commission.