TH E COLLAPSES OCCURRED IN THE BASILICA OF ST … · P. Roca, J.L. González, E. Onale y PB. Lourenço (Eds.) ... TH E COLLAPSES OCCURRED IN THE BASILICA OF ST FRANCIS OF ASS ISI

STRUCTURAl ANAlYSIS OF HISTORICAl CONSTRUCTIONS Il P. Roca, J.L. González, E. Onale y PB. Lourenço (Eds.)

e> CIMNE, Barcelona 1998

TH E COLLAPSES OCCURRED IN THE BASILICA OF ST FRANCIS OF ASS ISI AND IN THE CA THEDRAL OF NOTO

Giorgio Croci Professor ar Unil'/!rsi/y ofRome La Sapiel1=a Dep/. ofElIg. {lI1d Geowc/lllic Via Eudossialla 11 0 l8 ROlllc - 00153 (ITALl')

SUMMARY

The 13 March 1996 most part ofthe Cathedral ofNoto, in Sicily collapsed and the 26 Scptcmber 1997 lwO va ulls and part of lhe lransepl 01' lhe Basilica of SI Francis af Assisi collapscd too. Both of them represent emblema ti c cases af lhe risk lhe manuments are exposed to, but they providc also a starting poinl 10 reOcet how prvcnlive mcasurcs should bc takcn to pratect the architcctural hcritagc. This papaer ana!yses lhe causes of thcse collapscs and lhe measurcs lha! have becn laken to prevent further damage and to restare lhe l\VO churciles.

Inlroduction

This papcr analyses lhe collapses tha! occurred in lhe Basilica af SI rrancis and the Cathcdral of Noto to individuatc their causes and describc thc rcmedial measures that are going to bc taken.

A portian orthc left transept and Iwo vaults ofthc Basilica o rSt Francis collapsed during the second quake lhat struck ccntralltaly the 26 Septcmbcr 1997.

The earthquake in Assisi has bccn evaluated ofthe VIII dcgree of McrcalJi scalc;; the accc\eration measured at the ground dose to lhe Basilica was 0,16 g in the direetion of lhe longitudinal axis and 0,19 g in the perpendicular dircetion. Dynamic analysis and measures of lhe aeeelerations recorded in lhe folJowing months (a Syslclll of scismougraphs was placed after the earthquake in the Basilica and lhe Convcnt) have showll that the umplification of lhe scismic aclions betwccn lhe vaults and the ground is aboul 3. This value may anyway be considered only as indicativc becausc lhe quakes reçordcd after lhe strokc of 26 Seplember did no1 reaeh values to get the structure behind the elastic lclvcls, as happened ai the time oflhe collapsc.

The eollapse of the Cathedral af Noto occurrcd unexpectedly lhe 13 March 1996. The first elcmenl 10 subsidc, following lhe analysis and invcstigation that have

becn cJITied oul. was a pil1ar dose to lhe dome drowing down the dome itsclf and largc pm1 ofthe Basilica.

!n the fol1owing paragraphs these collapscs wiH be cxaminecl in delail anel their causes will be discusscd because bath af them \Vere in a certain W(ly unexpecled: the Basilica of St Francis in ils history rcsisted to many eat1hquakcs, some af them certainly highcr than that occlllTed lhe 26 Septcmber 1997, and nOlhing special happened to the Cathcdral 01' Noto the day, and the days bcfore lhe eoHapse.

298 STRUCTURAL ANALYS IS OF HISTOR ICAL CONSTRUCTIONS 11

2 Thc Basil ica or$! Francis ofAssisi

2.1 Thc Basil ica af SI Francis (Fig. I) , built in lhe 131h ccntury, was hit by many carthquakcs in its hislory. Earthquakcs strongcr Ihan lha! of 26 Septcmbcr 1997 probablyoccurredin 1279,I328, I703,I747,I781,I799, I832, I859, I917, I979,and yet no onc produccd darnagc as largc as that: lhe dcstruction af lhe vaulls dose to lhe façadc, of lhe vaults closc to lhe transcpt, of a a portion af lhe lcft tnmscpt (Flg. 2) and production af large cracks and pcnnancnt dcformation ali Qver lhe vaults 01' lhe Basilica, Ictting (hem in a vcry prccarious anel dangcrolls silualion.

FlG. I Thc Oasilica orSt Francis

FIG. 2 Thc collapscd zoncs in lhe Basilica

G. CROCI I Basi lica of SI. Francis of Assis i 299

Bcsides lhe differcllI imptlcl Ihal earlhquakcs of differcnl characterislics may havc produced on lhe Basilica. olheI' faelors have incrcased lhe vulncrabililY rcspecI lhe pas!. As regards lhe IympanU1l1 , made of a cavily \Vali wilh 1\\'0 faces and an inncr fil!. lhe cause of lhe partia I eollapsc (lhe firsl damagc \\Ias produced lhe 26 Scplcmber, but it \'Jas lhe quake of 7 OClober to ereale a large hole in the wall - Fig. 3 ) was the dccay of the mortar which Iinked lhe stoncs of lhe externai face with lhe inner fill; lhe reduced cohcsion and band cauld no! prcvcnt single slone blacks la progrcssively eler<lch ('''ch olheI' and fell dowll. As regareis lhe vmllts lhe eoHapse was produccd by a large volume of fill, mainly brakn tiles anel other loose materiais accullluJated aver çenlurics 01' roof repairs in the springer zones (Fig. 4). Under seismie actions Ihis fil!. wilholl! any eohesion, alternalively aels only on ane siele, whilsl on lhe otlK'r sidc lhe fiH is (lçlaehed; wbar is more lhe lose fill follo\\' s lhe mo\'emenl of the vaults opposing Iheir rccovery and facilitating Ihercforc increasing pernwl1cnt dclonnations. \-\lhen thc quakc of 26 Scplember hit lhe Basi lica it is very likcly tbal pcnnancnl elefollnations, rcducing lhe curvaturc anel Ihereforc lhe bcaring capaeily, \Vere accull1ulalcd as consequencc oflhc prcviaus earthquakes.

1·1G 3 Thc damage IIIlhc Iympanum of lhe lralbcpl

1· IG . 4 lhe hugc li1l accu!1Iulalcd 0 \('1' lhe vaultl'

300 STRUCTURAL ANALYSIS OF HISTORICAL CONSTRUCTIONS 11

A general modcl (Fig. 5) and a global stn.:ss analysis af lhe Basitica has bccn ca rried oul (Fig. 6) ; in particular a non linear model af lhe vaults Ims confcmlcd lhe dcci sivc role af lhe fiU (Fig. 7); whcn lhe horizontal acccJcration rcachs abOli! 0,2 g, relevanl tensi lc strcsscs are produccd. Thc failu rc mcchanislll a f lhe vaults d ose la lhe facadc, filmed by Umbria Tclcvision, rcsultcd fre m lhe progrcssivc loss o I' curvaturc of lhe rihs, Ihcll a "hinec" wns produccd in lhe midd lc and finally lhe rib collapsed drawi ngs lhe vau lt down with it (Fig. 8). A si milar Illcchanism Decurred in lhe zonc close to lhe transcpt , whcrc lhe sccond vault collnpscd.

FIG . 5 The lllodcl ofthc Basilica

,. lG . fi Une "f lhe SIIUClul,ll an"lysi~ calTicd oul 011 lhe Basilica

G. CROCI I Basili ca of SI. Francis of Ass isi 30 1

FIG. 7 Strcsscs tlnd dcfOTmatiOllS in the vau!ts

FIG. 8 A phOLO tllkcll during the çoll;Jp~e ora \ au!t

302 STRUCTURAL ANALYSIS OF HISTORICAL CONSTRUCTIONS 11

Tllc collapscs \Vere conçentratcd on Ihese spcci lic zom:s bccausc, bcing lhe dircction of lhe scismic action mainly perpend icular la lhe nave axis, il bchavcd globally like a "bcam", \Vherc a kind af rcstrainl a lhe cnds was providcd by the sti flilCSS af lhe façude and lhe transepl (Fig. 9). T ll c rcsults was lha! high nonnal nnd shcar strcsscs '\Vhere produccd there, in add ition to lhe "local strcsses" rcsulting fram the wcigh! of thc fill.

2.2 Urgent measurcs \\'ere rcquircd immcdiatcly anel" the carthquakc to prevenI lhe global collapsc ofthc tympanulll and ofthc V<lults. Thc survivcd vaults \Vere ali affcctc(\ by Ja rgc cracks distributcd both on lhe intrados and lhe cxtrados (Fig. 10); curvaturc, as alrcady said, was lost in several zones (Fig. 11 ).

The dangcr that the standing vau lts might co ll apse, and lhe consequenl ri sk lO lhe hUJll<ln life, precludcd lhe possibility of sUPPol1 ing lhe vauhs from the ground leveI. Inslead, a plalfoml was suspended from lhe roor above lhe vaults wilh lhe doublc funclion of inspeetion and of providing a base fo r working over lhe vaults (Fig. 12) .

. ... ---F/G . 9 ·/ hc col1apsed vaults

FIG . II Cr<lcks and dcformalions in lhe sUI"Vlvcd vaulls

FIG . 10 Cracks and defonmlliOIl~ in lhe ~ur\'ived vaulls

FIG . 12 Tlle lillle Oying bridge slIspcmlcd to lhe roof. rcalil.cd lo ill~pet:l :lIId slrenglhcn lhe vau lls

G. CROCI I Basili ca af SI. Francis a f Ass isi 303

Thc urgcnl measurcs takcn in lhe firsl 1110nlh after lhe main earthquake can bc sinthelized as follow5: • rcmoving the huge load represented by the fill in lhe spri nger zancs ofthc vaults, • fi Uing cracks wi th a sal! frcc mortar to Iimit possible damagc to lhe frescoes, firsl

taking the precaulion of inserting a strip of polyurethanc in lhe larger cracks la prevcnl the !nortar fram flowing OUI ,

applying bands of synthctic fibres (kevlar) over lhe emcks of lhe ex trados (Figures 13, 14)

• 5uspending lhe vaults from lhe roof wilh a system af tie ba rs, having firsl inscrtcd l\Vo sJ)rings lo mainlain lhe force aI lhe design value, indcpcndent of Ihennal effeets anel. minar vibrations (Fig. 15)

Thc crackcd ribs \Vere suspended from lhe roof as weU wilh a systelll similar to lhe previous one after having placcd undcmcath a kind of stccl cradlc fi ll cd with soft rubber in order nol to damagc the frescoes (Fig. 16).

FIGURES 13, 14 Thc kc,'lar fibre slrips applicd over lhe vaults

304 STRUCTURAL ANALYS IS OF HISTOR ICAL CONSTR UCTlONS "

FIO. 15 Thc systcm o fwircs and springs to su~pcnd lhe vaults !O lhe roof

FlG. 16 Suspcnsiol1 oflhc dcformcd fibrc kcrbs 10 lhe roor

G. CROCI I Basilica af SI. Francis of Assis i 305

As regard lhe Iympanum lhe risk was lhal if il had collapsed il would have deslryed lhe roof of lhe chapel below, causing lhe 1055 of frescoes and \Vorks of art of inestimable value. After long reflections il was decided 10 use a huge crane, 50 m tall .

Bu! such a crane could not gC! through the narrow gate into the inner yard. This prob lem was resolved by using l\Vo cranes; a first one outside the Basilica complex liftcd the second one over the roof of the building anel deposited il in lhe inner counyard.

Organizing thi s operation involved anchoring two cantilcver Slcel lrusses on the tWQ walls of the transep!. The trusses were designcd to support a 4,5 steel frame structure in lhe shape of the tympamnn, a trianglc 8 111 high and 1·7 III aI lhe base.

In lhe period 01' lime belween lhe lO'" al1d lhe 14'h of Oclober ali the operations were completed: the sleel slrueturcs \Vere bllilt ; !wo cranes arrived 011 lhe square in front of lhe Basilica; the flrst crane li fted lhe sccond Orle il1to lhe cOllrtyard (Fig. 17); the lWO cantilever sleel trllsses were lifted over the roof of lhe transep! and \Vere anchored la lhe lateral waUs, ready 10 receive lhe sleel frame.

After some attempts hindered by heavy rai n and \Vind. the crane succeedcd on lifting the slcel tympanum over the brackcts (Figures 18, 19).

Thc foUowing day the cmpty spaces and big holes \Vere flllcd \Vith poJyurcthane foam lo provisiona lly slabilize lhe masonry (Fig. 20).

FIG. 17 A crnnc in fronl oflhc convcnl lifts a second crane over lhe wall imo lhe courtyard

306 STRUCTURAL ANALYSIS OF HISTORICA L CONSTRUCTlONS 11

---~

FlG. 18 Thc lin of lhe ~lcc1 rcticu latcd SllUClurc to provisionally stabll izc lhe Iympanum

FIG . 19 Thc dcl ical(' momem whclllhc provisionod SlrUClurc is COlIllCCICd to lhe lympilnUl1l

FIG 20 Thc poliurct:mc roam to IiI1 IIc emply spaccs bClwcclllhc sloncs or lhc l}'mpanUlIl

G. CROCI I Basilica of S1. Francis of Assisi 307

2.3 The problcm of the definitive restoration and consolidation of the Basilica. expecially as regards the vaults, !las immcdiatly appcared to be very dclicate, beeause, due to the presence of the frescoes, it was impossiblc to rccover the dcfonnations and to re-establish, thereforc, an adequatc curvaturc anel autonomous bearing capacity.

Differcnt sludies, researches and malhcmalical lllodcls havc been earricd out la decide which so lution \Vould h.we been lhe most appropriatcl lo sl rcnglhen lhe vaults and secure their stability over the time, wilholll risking lo dmnage lhe frescocs and withouI eompromizing the historical value of lhe orig inal vall](s structurc. Finally the choiee has been to realize on lhe extrndos a scries of littlc ribs, following a pattem typical of Gothic sllUClurcs (Fig. 21), Ictting clcnrly visible the original strueturc.

Thcse ribs are made UI' of a eomposile material with nramidic fibers (Kcvlar) bcddcd in epoxy rcsins anel a cen tral timber IlUdCllS: this material is lighl, vcry Slrong (lhe tensilc strcngth of the Kcvlar fiber is 30.000 Kglem2 anel thal of lhe fiber with resins is. abou! 14.000 Kg/cm2 ) and less stitT of lhe sleel (the elasticity modulus is rcspeetivcly 1.200.000 Kg/cm2 and 600.000 Kglcm2).

Aramidie fibers, bcsidcs. haven't brittle bchaviours and prcsent a good duclility. The ribs are built in situ, so that it is possiblc to foUow lhe defonned shape of the

vaults: whilst the width of the ribs rcmains eonstanl (8 em), lhe hight has <In avcrage value of 20 el11. bu! it incrcases or is reduccd in rclation wilh the dcformation of lhe vaults, bccause lhe cxtrados of lhe ribs follows a regular curve parallel to the original ideal surface of the undefonned vaults.

FIG. 21 The plan of Ihc I1CW kc"lar fibre kcrbs 10 bc rcahzcd ovcr lhe C;(lrados of thc vaults

308 STRUCTURAL ANALYSIS OF HISTOR ICAL CONSTRUCTIONS 11

Di.ffercnt samples ofthc basic materiais and ofthc compositc ribs havc becn tcstcd in spccializcd laboratorics, as up- Io-day there was no! large cxpcrience on lhe app lication af these products to restore historie masonry slruclurcs; lhe rcsull s !lave bccn quite posit ive showing cxccllent rcsistcnce and good bond bctwcen lhe ribs and lhe masonry.

Thc !leI af ribs in a ccrlain numbcr af poinls will be sustaincd by tic bars COl1nected to lhe roof; sim il ad y to wilat was rcal i.lt::u in lhe urgem illlcrvcmion phasc, a spring will bc inserted in each bar in arder lO avoid cxccssivc conccntr<lted forces and to climinate lhe influcnccs af lhe tcmperature effects anel a f diflcrcnt movcmcnts bctwccn lhe roo r and lhe vaults .

As regards lhe cracks, which have compromiscd the continuity ofthe vaults, il has been decided lO complete the first injections realized in the emergency si tuation using a mortar ablc lo sali sfy vcry spccific and sevcrc cOllditiolls. It has to be salt-frcc and compatible with lhe frescoes, sufficienlly fluid to penelrate and difTuse in ali lhe cnlcks and microcracks; lo bc injectcd in dry masonry (no use af water is allowed); lo have good strength and bond capacity to re-eslabl ish a slruclllral continuity through the cracks. Accurate analysis and laboratory tests have allowed to individuatc a special mortar, prcparcd by MAPEI , as the besl lo flll!fil these rcqllirelllcnls.

The opcration of injection is carried out both from lhe ex Irados and lhe intrados oI' lhe vaults after prcliminary havi ng checked and weU fixcd the frescoes detachcd from lhe vauh, (Fig. 22).

A further intervention regards lhe masonry arches which sustain lhe roo r; their base aetua ll y simply stands on litt le arches which insist Qver lhe springcrs a f lhe vaults without any structura l conneetion and with a certain eeccntricity respecl lhe ribs of lhe vaults and the pillars.

It has becn thereforc decided to nnchor the base of lhe arches aI lhe walls and lhe towcrs behind, whieh in thi s very peculiar Italian Gothi c have lhe function oI' abutments. The anchoragc is rea!ized with a stccl belt and prcstrcssed horizontal bms (Fig. 23).

FlG. 22 Uule pipes inserted in lhe cracks 10 injecllhe morl<lr to consolld<lle lhe vaults

FIG. 23 Thc belt ai lhe base oflhe masonry arches to anchor Ihem lO lhe walls of lhe lJasilicll; lhe voids undcr [hcir base are also ViSlblc

G. CROCII Basilica of SI. Francis of Assisi 309

The reconstruction of lhe collapscd vaull has becn anolher major problem. FOitunalely aftel' a painstaking picce of rescarch several frescocd bricks that could be reused to rebuilt lhe vaults have been idenlificd.

The opcration has bcen pal1icularly succesful as regard lhe pieccs of ribs, although having felt down from 2S melrcs, had mainlained a good bend belwecn lhe bricks.

It has thereforc bcen possiblc lO asscmblc in laboratory lhe brokcn rans of ribs in sueh u wuy 10 create a sort ofvoussoirs aboul40 - 60 em long (Fig. 24); lhesc votlssoirs are thcn placcd on a provisional centring to rcbuilt the ribs (Fig. 25).

It hasn' l been possible instcad to rccover significant eJcments of lhe webs, so lhal l1ew bricks, exprcssly built 10 havc lhe samc costituents and similar characleristics ofthe original ones, are going lO be used.

The reconstruction of lhe vault s has taken inlo account lhe problem of rc­cstabli sll ing no! on[y a slruclural, buI also a strcss, con tinuity belwcen lhe new and the original p0!1ions of lhe vaults; a syslem of jacks has been foreseen for thi s purpose, placed in a provisional joint on lhe crown oI' lhe new vaults, to compensale lhe defomwlion. including lhe shrinkage of the lllortar, and lo calibrale lhe Slress dislributioll.

Thc res(oralion o f the Basilica is eompleled with lhe reconstruction of the collapsed portion af lhe lcft lympanum and the rceover of lhe defonnatians Ihal both tympanull1 have suffered; stones coming from the same original qUalTy are going to be used. To reduce the scismic actions transmitted to lhe tympanum, which evenc if consoJidaled rcmains a deli cnlc structure, the COllllcction between il and lhe roof \ViII be rcal izcd inlcrposing a specilic device using a "shape memory alloi' able to di ssipatc a ccna in amount of encrgy.

F1G. 24 The frngmcllls cf lhe kerbs preasscmbled to make UI' the voussoirs

FlG. 25 The rcconSlruclian af lhe rib~

310 STR UCTURAL ANALYS IS OF HI STORICAL CONSTR UCTIONS 11

3 Thc Cathedral af No to

3. 1 Thc Cathcdral af Noto, bui lt in lhe middlc af lhe 18,h ccntury, sufTcrcd during its !ire scvcrc damagc, partial collapscs (lhe dome collapscd du ring lhe construction in lhe 1760 and, after having bccn rcbuilt, again in lhe 1848, due to carl hquakcs), rcconstructiúns and rcinforCClIlcllts, li]] Ihr.: calasLrophc of M<lrch 1996 (F igures 26 and 27).

Di /Teren! faclors havc ptaycd a role in ils vulncrabi lity: lhe clTecI af frcqucnt caJ1hquakcs, lhe huge wcight o f thc structurc and of thc jill aver lhe vaults (Fig.28), lhe pour quality af lhe masonry rca li zcd with faces a f slonc blocks and a core a f rubble without cohcsion (Fig. 29), etc ..

F1GURFS 26, 27 Thc Üllhcdra\ orNolo aflcr lhe collapsc

G. e ROO! Basilica of SI. Francis af Ass isi J 11

FIG. 28 The hugc fill nnd weighl m·cr lhe \nults

FIG. 29 Thc Slruclurc af lhe pillars: 11 is cJcar1y \"isiblc lhe roor qualily ofthe core

312 STR UCTURAL ANA LYS1S DF HISTDR ICAL CDNSTRUC rIDN S 11

The survay af lhe dcblY (Fig. 30) and lhe slructural analysis af lhe Cathcdral (Fig. 31) have individuatcd thal lhe collapsc stal1cd by lhe crushing af lhe second pillar e10se to lhe dome on lhe right af lhe nave. Actually lhe st resscs in lhe pillars were vcry high (20-25 Kglcm2) and increascd af around 20% during lhe earthquakcs. The safety margin rcsults thcrcforc cxtremcly 10w whcn compareci with an averagc strcnglh af abou! 40 Kglcm2.

In this SOIl af pillars (Fig. 32) mosl part af the load is sustaincd by lhe faces, whilst lhe core no! only does no! o ffer any valiel contribution to lhe global strcngth, bUI , thrusting on lhe faces, creates on them supplemcntary bcnding stresscs .

In addition, lhe blocks af the làcing waUs, to facilitale lhe conslruction, were nol cu! with parallcl horizontal faces, but these are inclined inward so tha! only lhe externai surface àppcars to be regular; as consequellee lhe transmission af the vertical slresses is eanccntrated in arcas, much smaller than the theorelieal section, so that lhe strcsses are substalltially inereased.

FIG. 30 Thc debry ofthc collaps{·d Calhcdml

G. CROCI I Basilica of St. Francis of Assis i 313

FlG. 31 A malhcmatical model orthe Cathedral orNoto

FIG. 32 A dela11 of the structure oftlle pillars

314 STRUCTURAL ANALYS IS OF HISTORICAL CONSTRUCTIONS 11

Such unfavourabJc structll ral si tuat ion gOl \Vorse whcn litt lc vel1ical cracks \Vere produccd by the caT1hqukcs and lhe rainwatcr pcnclnllcd insidc, no! only incrcasing lhe lond buI also accclcrating lhe dccay of the pour 111011ar.

Signs af dangcr \Vere survaycd on some pillars four ycars bcfore lhe collapsc, 50

th D! in Fcbruary 1992 a provvisional st rcngthcning, shoring lhe pillars thclllseJvcs with a syslcm ofhorizolllíll sleel lubes, aCling as st irrups, was rcalized. This mcasurc, howcvcr, was not sufficient la stop lhe cvolut ion af lhe cnlshi llg phcno111cnon: when lhe stn.:s~c!) are dose to lhe strcngth, actually, cnlshing gocs on "spontancously" \Vith progrcssivc cXlcntion ofvcrtical crucks and olltward deformnlions; lhe fac ing stonc blocks <Ire then not <lblc to provide a "circumfcrcntial rcsistcncc" anymore - nor could lhe weak provitional steeltubes - counleract the lateral pressure ofthc Illbble core: the pillar then suddcnly "explodes" (Fig. 33).

The sccond pilJar from lhe dome on lhe lefi of lhe nave, whose seet ion is rcduced regard lhe four pillars under lhe dome, WílS lhe OI1C to rcach highcr slresscs, bcing also weakncd by a passagc lo go up to lhe pulpit (Fig. 34).

FIG. 33 The unefficlent provisional cincumrercntial propplllg of lhe pillars placcd four years before lhe

collapsc

FIG.3-1 I hc p;h~agc Ihn,ugh lhe collap~ed pillar

G. CROCII Basilica of SI. Francis or Assisi 315

Thc collapsc 01' a pillar has always catastrophic consequcnces and in this case the consequenees have been evcn "'orse af what iI callld havc becn expcctcd. probably bccause the original timher roaf was replaced with a reinforce conerete slab , in the middle af the 20th century (Fig. 35). This continous slab has spread lhe overloads produeed by the collapsc of the second pillar progrcssivcly on ali the othel" pillars. till the façade ofthc Cathedral.

Fig. 36 shows the dynamic af lhe col1upsc as it has bcen possible to deteet by the survey and analysis orthe dcbry.

FIG 35 Thc cOlllinuous reinron.:ed concrctc sl<1b Ihal in lhe middlc of lhe 20th cen!Ury I"eplaeed lhe

original (unbel" roor

FIG . 36 Thc dynamic oflhe collap~c

316 STRUCTURAL ANALYS IS OF HISTORICAL CONSTRUCTIONS 11

3.2 Thc first urgenl mcasures, takcn to preveni furthcr collapscs, consisted on realizing on lhe one hand a strong steel struclure to support lhe survived portion of lhe dome (Fig. 33) and on lhe other transversal st irrups to reinforce lhe survivcd pillars, as their wcakncss remains as an intrinsic aspect relatcd to lhe original materiais.

Thc final consolidation and rcstoration of lhe Cathcdral of Noto does nol pose structural problcllls as difficull as those faccd in lhe Basi li ca of SI Francis, bccause on lhe onc hand Ihere are nol such valuablc frescoes and pieces of an, and 011 lhe other, as mos! part of lhe Cathedral co ll apsed, it has to be rcbuilt, wll<ll is mucli casier Ihan to rcpair and consolidatc.

Techniques and materiais similar to the original oncs will be uscd, eliminating anyway what can be rcferred to an imperfect original structural realizalion: as lhe pour quality oflhe mortars, lhe loose rubble (which fonned "lhe core" ofthe pillars), etc.

Particular attention will be played lo the conneclion of lhe survived portion of lhe collapscd dome with lhe new structure in order to minimize the di scontinuity in lhe slrain and stress distribution.

Finally il will be necessary to strenglhen the surv ived slruclures and mainly lhe pillars, where the loosc rubble core has lo be consolidated with appropriale grouling and a system of horizontal stirrups.

FIG. 37 Thc provisionall-horing ofthc dome

G. CROCII Basilica af St. Francis a f Assisi 3 17

4 Condusions The coJlapscs and damage which involved the basilica of St Franeis of Assisi and

lhe Cathedral of oto havc vcry different origin and very different has becn lhe mechanism of collapse.

There are, however, common aspects in the two cases: lhe unappropriatc human action (lhe fill accumulated over lhe vaults in lhe Basilica of Assisi, lhe passage realized tharugh the weakest pillar in lhe Cathedral of 010 and lhe insufficicnt provisional measures taken four years before its collapse) and the lack af systematic structural survey. The las1 paint poses ance more the big question of a "prevention of risks" palicy, induding a rdinary and straardinary maintenance, and the need of plans of smvcy and investigarion in good time to detccl st ructural imperfections.

A dose investigatian would have allowed to discovcr lhat lhe vauhs of lhe Basilica of St Francis of Assisi had suffercd progressivc dcfonnalions, reduction 01' lhe curvature and increascd cracks in occasion 01' each of lhe main earthquakes occurrcd in the past, and that their bcaring capacity was progrcssivcly reduced. As rcgard the Cathcdral af Noto a skilfull survcy would have shown Ihal the littlc vertical cracks appeared in some pillars \Vere nol just superficial signs but a serious indicalion of an incipient crushing phcno1l1cnon.

More difticlllt \Vould had becn to dClcct lhe intrinsic weakness of lhe masol1ly of lhe tympanum of lhe basilica af SI Francis and af lhe pillars af lhe Cathcdral of Noto: some specifie tests and endoseopic analaysiS could have anyway given useful information.

In conclusiot1, stfllctural risks will never be tOla!ly eliminatcd in historie buildings; thesc ri sks, however, can certainl y be reduced ifthe "risk prevcntion" cullure - and in particular sistcmatic maintcnance and survcy - prcvails, inslcad of repairing and strenglhening only whcn se vere decay, damage or collapscs have oceurrcd.