prelims-vol-1 16/9/2004 18: 40 page I

STRUCTURAL ANALYSIS OF HISTORICAL CONSTRUCTIONS

POSSIBILITIES OF NUMERICAL AND EXPERIMENTAL TECHNIQUES
prelims-vol-1 16/9/2004 18: 40 page XIV
prelims-vol-1 16/9/2004 18: 40 page III

PROCEEDINGS OF THE FOURTH INTERNATIONAL SEMINAR ON STRUCTURAL ANALYSIS OF

HISTORICAL CONSTRUCTIONS, 10–13 NOVEMBER 2004, PADOVA, ITALY

Structural Analysis of Historical

Constructions
Possibilities of Numerical and
Experimental Techniques

Edited by

Claudio Modena
University of Padova, Padova, Italy

Paulo B. Lourenço
University of Minho, Guimarães, Portugal
Pere Roca
Technical University of Catalonia, Barcelona, Spain

VOLUME 1

A.A. BALKEMA PUBLISHERS Leiden / London / New York / Philadelphia / Singapore

prelims-vol-1 16/9/2004 18: 40 page IV

Copyright © 2005 Taylor & Francis Group plc, London, UK

All rights reserved. No part of this publication or the information contained herein may be reproduced, stored
in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying,
recording or otherwise, without written prior permission from the publisher.

Although all care is taken to ensure the integrity and quality of this publication and the information herein,
no responsibility is assumed by the publishers nor the author for any damage to property or persons as a result
of operation or use of this publication and/or the information contained herein.

Published by: A.A. Balkema Publishers, Leiden, The Netherlands, a member of Taylor & Francis Group plc
www.balkema.nl and www.tandf.co.uk

ISBN 04 1536 379 9

Printed in Great Britain

prelims-vol-1 16/9/2004 18: 40 page V

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

Table of Contents

Preface XV
Organizing committee XVII

Volume I
Invited lectures
Seismic rehabilitation of heritage buildings in India – problems and prospects 3
S.K. Agrawal
Research for seismic redesign of historic masonry buildings 15
M. Tomaževič
The importance of investigation for the diagnosis of historic buildings: application at different
scales (centres and single buildings) 29
L. Binda
Evaluation and analysis of the old timber structures 43
A. Ceccotti
Remedial measures for the Cathedral of Porto: a post-modern conservation approach 51
P.B. Lourenço, Â. Melo & M. Carneiro
Considerations on the significance of history for the structural analysis of ancient constructions 63
P. Roca
Design approaches of interventions for the safety and conservation of historic buildings 75
C. Modena

Historical aspects and general methodology

Ancient binding materials, mortars and concrete technology: history and durability aspects 87
Ö. Kırca
A system approach for examination and determination in historical buildings 95
D. Ekşi Akbulut & F. Aköz
Analysis of steel-structure/masonry-wall interaction in historic buildings 103
D. Friedman
Some considerations on the shape of the caps of vaults 111
D. Wendland
An inquiry into an unbuilt monument: the mausoleum for the kings of the Italy of Alessandro Antonelli 121
G. Pistone & L. Re
The industrial heritage of the Veneto between memory and project: what technology
for conservation and reuse? 131
G. Riva

V
prelims-vol-1 16/9/2004 18: 40 page VI

Technical state and renovation of buildings of Wrocław’s 19th century city centre development 137
P. Berkowski, G. Dmochowski, M.Y. Minch & J. Szołomicki
Constructive typologies investigation and approach proposals for the valuation of masonry arch state 145
C. Cennamo, S. D’Angelo & G. Voiello
Structural damage prevention in the historical building site. Theory and praxis in the eighteenth
century in Campania 153
G. de Martino & V. Russo
Restoration of the bell tower on the Church of Vistabella del Maestrazgo, Castellón (Spain) 159
F. Vegas López-Manzanares & C. Mileto
Palazzo Cittadini-Stampa: role of stratigraphy and cinematic analysis in the knowledge
of a masonry building 167
S. Bortolotto, C. Colla, D. Mirandola & A. Sponchioni
An architecture teaching program to rescue the historical town of Ouro Preto, Brazil 177
B.T. de Oliveira, E.C. de Araújo, J.N.S. Villaschi & L.C. Mancini
Analysis of a Roman masonry flat-slab in Hadrian’s Villa, Tivoli 183
D. Abruzzese, G.E. Cinque & G. Lo Gatto
Emergence, development, and prevalence of brick nogging in American vernacular structures 191
D.F. Laefer
Vernacular architecture and “historical seismography”: an experience research 203
O. Niglio & D. Ulivieri
The traditional twig-knitted wooden construction techniques: a vernacular architecture, “the Huğ house” 213
Z.H. Tokay

Materials and laboratory testing

An experimental study on the construction materials of the Ankara Citadel 223
Ö. Kırca & T.K. Erdem
A material that has witnessed the past in Anatolia: adobe 231
T. Çobancaoğlu & U. Tuztaşı
Investigation of the degradation of sandstones in Sydney’s heritage buildings 239
K.H. Friolo, A.S. Ray, B.H. Stuart & P.S. Thomas
Durability aspects of masonry stones used in the southern temple wall in Jerusalem 245
I. Wasserman
Physical and chemical properties of pre-regulated American cements 255
A.S. Rush & D.F. Laefer
Study of Place Stanislas coloured pavement for its historical restoration 265
J.M. Mechling & R. Elter
Structural failure of historic buildings: masonry fatigue tests for an interpretation model 273
P. Ronca, A. Franchi & P. Crespi
Experimental investigation on masonry elements subjected to eccentric axial loads 281
L. Cavaleri, A. Failla, L. La Mendola & M. Papia
In-plane shear and tensile strength tests of small brickwork specimens 291
L. Malyszko
Testing and modelling of multiple-leaf masonry walls under shear and compression 299
J. Pina-Henriques, P.B. Lourenço, L. Binda & A. Anzani

VI
prelims-vol-1 16/9/2004 18: 40 page VII

Non-destructive testing and inspection techniques

On-site investigation techniques for the structural evaluation of historic masonry
buildings – a European research project 313
C. Maierhofer, C. Köpp & A. Wendrich
Combined in-situ tests for the assessment of historic masonry structures in seismic regions 321
V. Bosiljkov, M. Tomaževič, L. Binda, C. Tedeschi, A. Saisi, L. Zanzi, F. da Porto,
C. Modena & M.R. Valluzzi
Structural evaluation of historic walls and columns in the Altes Museum in Berlin using
non-destructive testing methods 331
C. Maierhofer, M. Hamann, C. Hennen, B. Knupfer, M. Marchisio, F. da Porto, L. Binda & L. Zanzi
Geo-electrical techniques as a non-destructive appliance for restoration purposes 343
R. Keersmaekers, F. Van Rickstal & D. Van Gemert
The use of radar techniques and endoscopy in investigating old masonry: the case of Dafni Monastery 351
E. Vintzileou, A. Miltiadou-Fézans, V. Palieraki & N. Delinikolas
Identification of the mechanical properties of timber structures by combined non-destructive tests 361
M.R. Migliore & F. Ramundo
Compressive behavior and NDT correlations for chestnut wood (Castanea sativa Mill.) 369
A.O. Feio, P.B. Lourenço & J.S. Machado
Inspection & NDT to verify structural reliability of historic wooden roofs in the ex-Meroni
spinning-mill 377
F. Augelli, C. Colla & R. Mastropirro
The historic side-walls of the Navigli canals in Milano: in situ and laboratory tests for the structural
conservative project 387
A. Migliacci, P. Ronca, P. Crespi, G. Franchi, F. Bianchi & S. Di Martino
Application of ultrasonic Rayleigh wave to testing of masonry materials 395
M. Skłodowski
Mosaic-grouting monitoring by ground-penetrating radar 401
P. Côte, X. Dérobert, A. Miltiadou-Fézans, N. Delinikolas & N. Minos
Ultrasonic testing of properties of mortars 407
P. Cikrle, J. Adámek & M. Stehlík

Dynamic behaviour and structural monitoring

The influence of bells’ movement on the adjacent masonry vibrations 415
O. Fischer & S. Urushadze
Validated structural analysis of Gothic vaulted systems 421
E. Erdogmus & T.E. Boothby
A first approach to study an 18th century belltower 429
S. Ivorra, M.J. Palomo, G. Verdú & F. Pallarés
Dynamic-based F.E. model updating to evaluate damage in masonry towers 439
C. Gentile & A. Saisi
Seismic resistance of masonry towers 451
D. Abruzzese & A. Vari
Techniques of structural identification for the monitoring of historical buildings: first experimental
results for a masonry tower 461
P. Carusi, V. Sepe & A. Viskovic

VII
prelims-vol-1 16/9/2004 18: 40 page VIII

The dynamic behaviour of the façade of the Basilica S.Maria di Collemaggio 469
E. Antonacci & G.C. Beolchini
Earthquake performance of Suleymaniye Mosque 477
S.M. Kaya, O. Yuzugullu, M. Erdik & N. Aydinoglu
Ambient vibration testing at N. Sra. do Carmo Church, preliminary results 483
M.A. Baptista, P. Mendes, A. Afilhado, L. Agostinho, S. Lagomarsino & L.M. Victor
Experimental stress analysis of historical forged tie beams of archaeological museum of
Spina in Ferrara, Italy 489
G. Bruschi, G. Nardoni, L. Lanza, F. Laudiero, N. Tullini, G. Mezzadri & S. Tralli
Dynamic identification and model updating of historical buildings. State-of-the-art review 499
J.C. Araiza Garaygordóbil
Structural monitoring in the Villa Reale of Monza (MI), Italy 505
A. Del Grosso, A. Torre, G. Corte, G. Brunetti & D. Inaudi
Dynamic monitoring of an ancient masonry bridge on Carrara marble way 513
G. Chellini, P. Orsini & W. Salvatore
Structural integrity assessment of medieval towers 523
A. Carpinteri & G. Lacidogna
Experimental and numerical analysis of the structural behaviour of St Stefano’s bell-tower in Venice 533
A. Lionello, I. Cavaggioni, C. Modena, F. Casarin, P.P. Rossi & C. Rossi

Analytical and numerical approaches

Limit analysis of masonry structures 545
C. Anselmi, E. De Rosa & L. Fino
A displacement-based approach for the safety assessment of masonry structures 551
J.A. Ochsendorf
A simplified formula for the evaluation of horizontal capacity of masonry portal frames 557
A. De Luca, E. Mele, A. Romano & A. Giordano
Staircases as cantilevers or arches? – a question for limit analysis 569
E.A.W. Maunder
Systems of arches and columns strengthened with FRP at the extrados 577
U. Ianniruberto & Z. Rinaldi
Ultimate behavior of masonry arches reinforced with FRP at the intrados: comparison between
analytical and numerical models 583
U. Ianniruberto & Z. Rinaldi
A probabilistic model for the assessment of historic buildings under permanent loading 589
E. Garavaglia, A. Anzani & L. Binda
Out-of-plane behaviour of multiple-leaf stone masonry 597
G. Mirabella Roberti, L. Donati & A. Fontana
Numerical modelling of a small mortar dome towards the restoration of a cruzeiro in Portugal 603
P.B. Cachim & A.L. Velosa
The mechanical behaviour of mortars in triaxial compression 611
R. Hayen, K. Van Balen & D. Van Gemert
The effect of the masonry pattern on the global behaviour of vaults 619
C. Calderini & S. Lagomarsino

VIII
prelims-vol-1 16/9/2004 18: 40 page IX

Non linear step by step analysis for masonry structures using changing shape finite elements – new
developments and implementations 629
P. D’Asdia, A. Viskovic & C. Brusaporci
Construction sequence in the finite element analysis of Romanesque churches 635
L.A. Van Gulick
Structural scheme of the Cathedral of Burgos 643
D. Theodossopoulos
Comparison of the masonry structures analysis using the co-rotational formulation and a
simplified proposal 653
A. Agüero & F.J. Pallarés
Analysis of masonry structures by discrete finite element method 659
I.H.P. Mamaghani
Why does the Endless Column seem everlasting? 667
C.A. Safta
Structural analysis of Küçük Ayasofya Mosque in İstanbul 679
M. Massanas, P. Roca, M. Cervera & G. Arun
Prevision of settlement-induced cracking in historical building masonry façades 687
J.G. Rots & S. Invernizzi
Cracking simulation in a plain structure using the finite element method 695
L. Pani, B. De Nicolo & Z. Odoni

Author index 701

Volume II
Consolidation and strengthening techniques
Experimental investigation and analytical modeling of the effect of injection grouts on the structural
behaviour of three-leaf masonry walls 707
E.E. Toumbakari, D. Van Gemert, T.P. Tassios & E. Vintzileou
Application of mineral grouts. Case study and impact on structural behaviour: Church of St. Catharina
at Duisburg (B) 719
S. Ignoul, F. Van Rickstal & D. Van Gemert
Analysis of historic masonry shear walls strengthened by composite material 727
R. Capozucca
Two techniques for repair and strengthening masonry constructions 735
R.A. Sofronie
Structural restoration of vaults with extrados ties 745
A. Gubana & E. Giuriani
The main cupola of the Dresden Frauenkirche 753
V. Stoll & W. Jaeger
The medieval castle Spøttrup. Stabilization of the south wing 763
T. Jespersen
Technologies for the prestressing rings of the Leaning Tower of Pisa 771
A. Lodigiani & G. Macchi
A quick and low-cost method for strengthening high buildings: the case of St. Stefano’s
bell-tower in Venice 781
A. Lionello, I. Cavaggioni, P.P. Rossi & C. Rossi

IX
prelims-vol-1 16/9/2004 18: 40 page X

Intervention methodology on historical structures subject to distortions 787

A. Raffagli
Strengthening of historical construction by using steel bracing inserted into brick walls 795
M.Y. Minch, J. Szołomicki, P. Berkowski & G. Dmochowski
Confined masonry members: a method for predicting compressive behaviour up to failure 803
A. Mandara & G. Palumbo
Castel Belasi’s special underpinning and reinforcement project 813
P. Mazzalai, S. Torresani, L. Silvestri & L. Springhetti
Non-invasive underpinning technologies in historic settings 821
E.E. Crocker
Restoration and strengthening with fibre reinforced polymers: issues to consider 829
N.G. Shrive, M.M. Reda Taha & M.J. Masia
Creep behavior of brick masonry panels strengthened by the bed joints reinforcement technique
using CFRP thin strips 837
A. Saisi, L. Binda, M.R. Valluzzi & C. Modena
In-plane shear strengthening of natural masonry walls with NSM CFRP strips and FRCM overlay 847
L. De Lorenzis, N. Galati & L. Ombres
Numerical modeling of masonry structures reinforced by FRP plate/sheets 857
A. Failla, A. Cottone & G. Giambanco
Stress transfer at the interface of bonded joints between FRP and calcarenite natural stone 867
M. Accardi & L. La Mendola
Influence of environmental agents on bond between FRP reinforcement and calcarenite ashlars 875
M.A. Aiello & M.S. Sciolti
Durability assessment of FRP wraps applied on high artistic “plaster and reeds” vaults 883
E. Quagliarini, M. D’Orazio & A. Stazi
FRP strengthening systems for metallic structures: a state of the art 891
V. Zerbo, A. Di Tommaso & L. Ceriolo

Historical timber and metal structures

Experimental analysis of ancient wooden beams for flexural and shear failure 903
F.M. Mazzolani, B. Calderoni, G. De Matteis & C. Giubileo
A strengthening technique for timber floors using traditional materials 911
C. Modena, M.R. Valluzzi, E. Garbin & F. da Porto
Experimental analysis of ancient chestnut beams by small specimens 923
F.M. Mazzolani, B. Calderoni, G. De Matteis & C. Giubileo
Investigating causes of damages to historical timber structures by use of FEM 935
C. Thelin & K.G. Olsson
Methodology for the analysis of complex historical wooden structures: a study case 945
F.M. Mazzolani, B. Faggiano & A. Marzo
Strengthening and control methods for old timber trusses: the queen-post truss of the Trento theatre 957
M.P. Piazza, M.R. Riggio & G.B. Brentari
Timber roof structures of the “Arsenale” of Venice 967
C. Menichelli, C. Modena, M.R. Valluzzi, E. Garbin & F. da Porto
The furnace hall in Peitz – analyzing a 200-year old curved plank roof 977
L. Hahmann

X
prelims-vol-1 16/9/2004 18: 40 page XI

Viscoelastic analyses of a prestressed Howe Bridge 983

D.A. Gasparini, J. Bruckner & F. da Porto
Ancient timber bridges: numerical modelling of connections behavior 991
C. Ceraldi & E. Russo Ermolli
Long-term creep in wooden-iron bridges: a comparative study of different truss types 1001
F. Lanza, A. Mattarucco & F. da Porto
Structural analysis and rehabilitation of a nailed railway bridge in Genoa 1009
A. Brencich & L. Gambarotta
Historical railway bridges: tests and numerical analysis 1019
M. Ferraioli, P. Malangone, M. Rauci & A. Zambrano
Restoration of the historical steel vault of the Goldoni Theatre in Livorno 1029
M. Sassu
Numerical analysis of the steel roofing structure of the Umberto I Gallery in Naples 1039
R. Landolfo, M. Manganiello & F. Portioli

Seismic analysis and vulnerability assessment

Vulnerability study in seismic areas: the role of on-site and archives investigation 1051
A. Anzani, A. Baila, D. Penazzi & L. Binda
Guidelines for restoration and improvement of historical centers in seismic
regions: the Umbria experience 1061
L. Binda, G. Cardani, A. Saisi, C. Modena, M.R. Valluzzi & L. Marchetti
An analysis of the seismic vulnerability of the architectural heritage in Bhuj, Gujarat, India 1069
D. D’Ayala & A. Kansal
Experimental testing in support of a mechanics-based procedure for the seismic risk evaluation
of unreinforced masonry buildings 1079
L.F. Restrepo-Vélez & G. Magenes
Mechanical models for the seismic vulnerability assessment of churches 1091
S. Lagomarsino, S. Podestà, S. Resemini, E. Curti & S. Parodi
Seismic risk of monumental structures of Kathmandu Valley 1103
P.N. Maskey & T.K. Datta
Assessing the seismic vulnerability of late Ottoman buildings in Istanbul 1111
D. D’Ayala & D. Yeomans
On elastic models for evaluation of the seismic vulnerability of masonry churches 1121
R. Meli & F. Peña
Probability of collapse of monumental buildings under seismic loads 1133
G. Augusti, M. Ciampoli & E. Canalella
Out-of-plane fragility of historic masonry walls 1143
G. de Felice
Failure mode in an industrial brickwork chimney using different criteria 1149
F.J. Pallarés, A. Agüero, M. Martín & S. Ivorra
Seismic assessment of masonry structures by non-linear macro-element analysis 1157
A. Penna, S. Cattari, A. Galasco & S. Lagomarsino
Correlation between tensile strength and the collapse mechanism of brick masonry constructions 1165
F. Peña

XI
prelims-vol-1 16/9/2004 18: 40 page XII

Stochastic analysis of hysteretic degrading systems subject to earthquakes 1175

G.C. Marano & R. Greco

Seismic strengthening and innovative systems

Implementation of a methodology using “ties and injection” developed for repair and strengthening
of historic monuments 1187
V. Sendova & P. Gavrilovic
Stabilization of the landmark Las Flores adobe: a case study of restrained seismic retrofit 1197
J.M. Barrow, E.L. Tolles & S.J. Farneth
Masonry strengthening by metal tie-bars, a case study 1207
G. Spina, F. Ramundo & A. Mandara
Modifying building response using energy absorbing diaphragm-to-wall connectors 1215
A. Lockrem, M. Schuller & R. Locke
Some considerations of structural restoration of S. Giacomo church’s in Gavi (Alessandria) 1221
M. Corradi & V. Filemio
Seismic protection of monuments by shape memory alloy devices and shock transmitters 1229
M.G. Castellano & S. Infanti
Seismic protection of historical centers using innovative techniques, with focus on San Giuliano
di Puglia after the 2002 Molise earthquake 1235
M. Indirli, P. Clemente, B. Spadoni, R. Cami, E. Speranza, M. Mucciarella & F. Pistola
Seismic upgrading of an old industrial masonry building by dissipative steel roofing 1247
F.M. Mazzolani & A. Mandara
Seismic isolation: a new approach to earthquake protection of historic monuments 1257
P. Gavrilovic, V. Sendova & S.J. Kelley

Case studies
Masonry structure in the Crusader’s Castles – Syria 1267
W. Jaeger & E. Al-Khateeb
Library of Parliament of Canada – conservation, rehabilitation, upgrade case study presentation 1273
M. Petrescu-Comnene & S.R. Higgins
Pathology of the Dafni Monastery: survey, monitoring of cracks, interpretation and
numerical verification 1285
A. Miltiadou-Fézans, E. Vintzileou, N. Delinikolas, E. Zaroyianni & E. Chorafa
The role of survey for a correct FE analysis: the case of the Monte Oliveto Maggiore church 1295
B. Belletti, E. Coïsson, C. Ferrari & S. Pedroni
Researching the bearing system’s capacity as an argument in solving a dilemma in the interpretation
of the origin and development of architectural complex of the St. Donatus’ church in Zadar 1305
E. Lokos̆ek & H. Podnar
Analysis of Tapial structures for modern use and conservation 1315
P.A. Jaquin, C.E. Augarde & C.M. Gerrard

Preliminary investigation and monitoring for the design of a strengthening intervention on the
Frari basilica, Venice 1323
A. Lionello, I. Cavaggioni, P.P. Rossi, C. Rossi, C. Modena, F. Casarin, G. Marchi,
G. Gottardi & A. Ragazzini

XII
prelims-vol-1 16/9/2004 18: 40 page XIII

Fallingwater – structural intervention, in time 1335

J.A. Matteo
The Nidaros Cathedral in Trondheim. The Norwegian National Symbol. Stabilization of the Choir 1341
T. Jespersen
Structural restoration of a farm wing of the Park Abbey at Heverlee, Belgium 1349
K. Brosens, S. Ignoul, D. Van Gemert, K. Van Balen, L. Schueremans & P. Stevens
Application of CFRP laminates as strengthening of cracked brick arches 1357
R. Ciesielski, H. Ciurej & A. Kwiecień
Restoring the greenhouse at Lednice chateau 1367
J. Hirs, M. Bajer, J. Kala & K. Ksandr
Saving minarets at risk in Afghanistan 1375
G. Macchi
The vaulting structure of the Temple of Venus and Rome at the Roman Forum 1383
C. González-Longo & D. Theodossopoulos
Structural analysis and project of reconsolidation of Italian Hospital, Montevideo, Uruguay 1395
G. Valletta, R. Chaer, S. Nencioni, C. Pecora & L. Nisizaki
Theoretical and experimental studies for strengthening Bohemian brick vaults 1403
I. Bucur-Horváth, M. Miklós & I. Popa
Structural studies developed on “The Miracles Roman Aqueduct” in Merida (Spain) 1409
A.G. Palacio
Preservation and stability of industrial masonry chimneys 1413
A.T. Vermeltfoort
A new load bearing structure for the conservation of the roofs of the Molino Stucky in Venice (Italy) 1421
G. Boccanegra, S. Dobricic, A. Marchi & F. Amendolagine

Author index 1427

XIII
prelims-vol-1 16/9/2004 18: 40 page XIV
prelims-vol-1 16/9/2004 18: 40 page XV

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

Preface

Structural Analysis of Historical Constructions IV continues the series of triennial international seminars started
in Barcelona in 1995 and continued with increasing success in Barcelona 1998 and Guimaraes in 2001. This
edition is organized at the University of Padova, at the beginning of its 783rd Academic Year, with a strong
involvement of the country, Italy, where an exceptional concentration of historic construction and a long tradition
of conservation actions exist: almost half of the about 160 accepted papers are in fact from Italy. Located near
Venice, Padova is also a good lookout of what is happening for the conservation of one of the world’s most
famous historic centres. The occasion is in fact taken to have a close view of significant works being done in
Venice, in the ambit of a special session organized inside its historic Arsenale. Venice is also the site of the
UNESCO Regional Bureau for Science in Europe (ROSTE), so that the seminar is intended to offer a special
chance for debating the very specific problems that have to be faced in countries of east Europe and middle and
far east for the preservation of historic architectural heritage.
The number and quality of contributions, the number of countries, 27, from which the contributors came, is a
clear indication of the importance have structural safety aspects for ensuring physical survival of the architectural
heritage in all the world, and then the possibility to preserve its historic-artistic value with all the implied
cultural-social and economical consequences.
The major challenges to specialized people – technicians, art historians, conservators – involved in the multi-
disciplinary processes that are required to ensure the conservation of architectural heritage are all reflected in
the written free and invited contributions, organized in ten topics: “Historical Aspects and General Method-
ology”, “Materials and Laboratory Testing”, “Non-Destructive Testing and Inspection Techniques”, “Dynamic
Behaviour and Structural Monitoring”, “Analytical and Numerical Approaches”, “Consolidation and Strength-
ening Techniques”, “Historical Timber and Metal Structures”, “Seismic Analysis and Vulnerability Assessment”,
“Seismic Strengthening and Innovative Systems”, “Case Studies”. In addition to the free presentations, height
comprehensive plenary lectures have been provided by invited speakers.
As it is well known, it is not only a matter of dealing with a very difficult engineering problem, as it actually
is. The very basic safety concepts are quite different, and much less established, in the case of existing structures
with respect to the newly constructed ones. This is of course particularly true when the artistic/historic importance
of existing structures is such to accept higher probabilities of “structural failures” then in the case of the new
ones. Theoretical and experimental tools for analyzing the material and structural properties and performances
are continuously evolving, in a very distinguishing way they are in the field of moderns constructions. Learning
from practice is essential, even if the application are made always on “unique”, and for this only reason always
important, objects.
But even more crucial is the fact that both in the phase of the evaluation of the exiting safety conditions and in
the choice of the interventions an “appropriate balance” must by case by case found between “pure engineering”
and “conservation” aspects.
It is of course the hope of the organizers that the seminar and the book will fruitfully help practitioner
technicians and final users of their activity, and will stimulate scientists and researcher in further developing,
stimulated by the presentations and discussions, new and more and more reliable tools and methodologies for
ensuring “structural safety” while preserving historic/artistic values of the world’s architectural heritage.

Padova, August 2004 Claudio Modena / Paulo B. Lourenço / Pere Roca

XV
prelims-vol-1 16/9/2004 18: 40 page XIV
prelims-vol-1 16/9/2004 18: 40 page XVII

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

Organizing committee
Paulo B. Lourenço Universidade do Minho – Portugal
Claudio Modena Università di Padova – Italy
Pere Roca Universitat Politécnica de Catalunya – Spain

Scientific committee
Drazen Aninic University Josip Juraj Strossmayer – Croatia
Görün Arun Yildiz Technical University – Turkey
Luigia Binda Politecnico di Milano – Italy
Thomas Boothby The Pennsylvania State University – United States of America
Giovanni Carbonara Università di Roma “La Sapienza” – Italy
Ario Ceccotti Istituto per la Valorizzazione del Legno e delle Specie Arboree – CNR – Italy
Dina D’Ayala University of Bath – United Kingdom
Miloš Drdácký Institute of Theoretical and Applied Mechanics – Czech Republic
Gavrilovic Pedrag University St. Cyril and Methodius – Macedonia
Pietro Alessandro Vigato Progetto Finalizzato Beni Culturali – CNR – Italy
Maurizio Indirli Ente per le Nuove Tecnologie, l’Energia e l’Ambiente – Italy
Giorgio Macchi Università di Pavia – Italy
Christiane Maierhofer BAM, Federal Institute for Material Research and Testing – Germany
Roberto Meli Universidad Nacional Autónoma de México – Mexico
Ioanna Papayianni Aristotle University of Thessaloniki – Greece
Andrew Powter ICOMOS International Wood Committee – Canada
Michael Schuller Atkinson-Noland Associates – United States of America
Ramiro Sofronie University of Bucarest – Romania
Vito Renda European Laboratory for Structural Assessment – JRC – Italy
Miha Tomasevic Slovenian National Building and Civil Engineering Institute – ZAG – Slovenia
Koenraad van Balen Katholieke Universiteit Leuven – Belgium
Elizabeth Vintzileou National Technical University of Athens – Greece

The papers are published in the form submitted by the authors, after revision by the Scientific
Committee. The Editors are relieved of any responsibility for possible errors or disagreements.

XVII
prelims-vol-1 16/9/2004 18: 40 page XVIII

Organization

Università degli Studi di Padova Dipartimento di Costruzioni e Trasporti

Sponsors

TNO DIANA Fischer Italia S.p.A.

Industrie Cotto Possagno S.p.A MOOG Italia S.p.A

Terreal Italia S.r.I Tassullo S.p.A.

Supporting institutions

THE EUROPEAN COMMISSION’S

DELEGATION TO INDIA,
BHUTAN, MALDIVES AND
NEPAL
EU-India CONSIGLIO NAZIONALE DELLE RICERCHE
Economic Cross Cultural Programme Progetto Finalizzato – Beni Culturali

Patronage

United Nations Educational,

Regione Veneto Scientific and Cultural Organization

XVIII
chap-01 9/9/2004 8: 5 page 1

Invited lectures
prelims-vol-1 16/9/2004 18: 40 page XIV
chap-01 9/9/2004 8: 5 page 3

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

Seismic rehabilitation of heritage buildings in India –

problems and prospects

S.K. Agrawal
Central Building Research Institute, Roorkee, India

ABSTRACT: History of earthquakes in India demonstrates vulnerability of our heritage buildings to seismic
hazards. The recent past, devastating earthquakes in urban areas in India causing wide spread damages to
these buildings, disruption of services and damage to environment have been of great concern; the experiences
have prompted to carry out in-depth studies and come out with solutions and policies which will go a long
way in minimizing the damages caused by seismic ground motions. In this context, seismic rehabilitation of
heritage structures is a new discipline coming up in India and should be given a top priority. The paper discusses
threadbare the problems associated with the heritage buildings, their modes of failures, rehabilitation actions
and strategies. A recently accomplished restoration of a prestigious building damaged during 26th January 2001,
Bhuj earthquake of India is discussed in the paper as a case study. Also, it is highlighted that there is urgent need
of instrumentation, experimentation and monitoring of heritage structures to study the dynamic properties and
safeguard them from future ground motion.

1 INTRODUCTION – HISTORY OF INDIAN India has been the seat of a great civilization for
HERITAGE BUILDINGS 5000 years or more. The ancient civilization has its
manifestation in the sphere of religion, art, literature,
Nearly 5000 years ago, on the lush green banks of river music and sculpture – all these having been imprinted
Indus, man started making sculptures from stones. In into the various monuments constructed through ages.
fact, he wanted to give shape to his own creativity. The country is dotted with plethora of monuments,
It was the copper age, iron had not appeared on the which include temples, Buddhist stupas, churches,
scene. He did not have fine chisel or sophisticated mosques, rock-cut caves, minarets, palaces and forts.
tools, but that did not deter him. He chose relatively Many of these monuments are of considerable antiq-
softer rocks such as limestone and sandstone for mak- uity dating back to 1000 years or more.
ing small statues. These creative people lived in twin The evolution of construction techniques is related
cities of Mohenjodaro and Harappa, in the Indus valley, to the gradual assimilation of knowledge of using
now in Pakistan, one of the oldest records of its kind in construction materials of various types in a given
India. These were the earliest urbanized colonies of the environment. The earliest construction practice was
Indian subcontinent. The great civilization left behind confined mainly to the use of bricks and wood, and
large monumental examples of carvings and construc- due to their low durability, the well-preserved her-
tion skill. The artisans used either rocks and bricks itage buildings of such kind prior to 3rd century BC
or carved out temples and caves in rocks. These Her- are rarely found. However, the remains found in the
itage buildings are mute testimony in adobe, rubble, Mohenjodaro site of Indus valley civilization (3000
stone or brick of the ceaseless efforts of mankind to BC) indicates use of shaped and dressed blocks of
express his social and religious feelings. The progress polished marble for building purposes.
of human civilization can be traced from the varied The use of rocks by Emperor Ashoka (273–237 BC)
artifacts, motifs and artistic imagery carved in her- as huge monolithic pillars and railings added a new
itage buildings. These monuments not only reflect the dimension to the building history of India. Revolution-
rich cultural life of the ancient and medieval people ary thoughts and dynamic approach in utilizing in situ
but also show their technical skill in engineering con- rock or the rock-mass gave birth to the rock-cut archi-
struction and profound knowledge in the selection of tecture in India, which now holds a glorious record of
sites, judgement of foundation conditions and choice having nearly 12,000 rock-cut structures like shrines,
of construction material. chaityas and monasteries etc. constructed between 3rd

3
chap-01 9/9/2004 8: 5 page 4

century BC and 9th century AD. In the construction of

such rock-cut structures, knowledge of the right selec-
tion of the site including the character of the rocks,
which are amenable to a sculptor’s chisel and ham-
mer, was vital. The world-famous rock-cut temples of
Ajanta and Ellora, a treasure-trove of art, were chiseled
out of the very resistant Deccan basalt at suitable sites.
Creative ideas for the utilization of local resources
of building material like rock and clay brought in new
concepts in architectural design and structural con-
structions. Where rock was available within a close
proximity, it was used for durable construction mate-
rial and lasting carvings. Thus we find, charnockite
was used in the Mahabalipuram temple of Chennai,
granite in the Meenakashi temple of Madurai, khon- Figure 1. Monument at Bhuj damaged due to Gujarat
Earthquake.
dalite in the Konark temple of Orissa, Triassic lime-
stone in the Avantipur temple of Kashmir, Vindhayan
sandstone in the Ashoka Pillar and Dhameka stupa in terms of loss of life, property, disruption of services
of Sarnath, and Siwalik sandstone in the Kangra fort and damage to environment recurrently. The damages
of Himachal Pradesh. When beauty was the prime caused to the heritage structures during earthquakes
motive of construction, as in the Taj Mahal of Agra, the are never given first priority and most of the times
right selection of stone like the chaste-white Makrana go unnoticed. However, thanks to the concerned orga-
marble of Rajasthan was the primary consideration nizations and people, who time and again bring upon
irrespective of the distance of its availability. Hundreds these spectacular damages with the emphasis to restore
of such splendid monuments of varied types of rocks them in their original glory. Among the many critical
have withstood the ravages of several centuries and modes of failure experienced by heritage structures
are destined to stay for several centuries, if preserved during earthquakes, the most relevant and recent with
meticulously and protected from natural disasters such reference to India are discussed here.
as earthquake. The elegant terracotta sculptured brick
temples of Bengal provide a classic example of the 2.1 Structural aspects
imaginative expression of artistry and construction
Most of the heritage buildings are made up of masonry
talent in the use of local clay in the absence of rock.
of varied kinds, from adobe or rubble joined with
Since the time immortal, these rich heritage build-
mud/lime/surkhi mortars, to rock/stone conglomerates
ings of India have been facing the wrath of man made
with cementitious materials of good quality. It is writ-
cultural invasions, natural disasters and environmental
ten in Indian literature that several kings used blood,
degradation. However, they are the physical evidences
pulses, and jaggery as binders and longevity. However,
of rich cultural values of our glorious past, which we
these heritage buildings seemed to perform badly dur-
inherit from our ancestors and are to be preserved
ing past earthquakes. The basic reasons has been the
in their authenticity i.e. aesthetic and historical val-
intrinsic weaknesses of masonry i.e. its great weight
ues ensuring structural safety against external actions.
and low tensile strength, which is further aggravated by
The paper elaborates seismic rehabilitation of the her-
deterioration due to weathering. An important aspect
itage structures in India. The purpose of the paper
of the masonry used in heritage structures is its het-
is to highlight the problems and prospects connected
erogeneity and the significant variations in its quality
with structural restoration and seismic protection in
among different structural members and even within
the areas of significant seismic hazard with special
the same member. Also, in masonry structures, it is
reference to the problems of the structures in India.
difficult to give continuity to their structural mem-
bers (among transverse walls and of the walls with the
floors and roofs), as well as the difficulty to form stiff
2 PERFORMANCE OF HERITAGE BUILDINGS horizontal diaphragms capable of distributing inertia
DURING PAST EARTHQUAKES forces to the stiffest and strongest walls. Devastation
caused to the heritage building in Bhuj during 26th
Experiences of earthquakes in last decades in semi- January 2001, Gujarat earthquake is shown in Figure 1.
urban & urban parts of India (Uttarkashi 1991, Latur Another picture shown in Figure 2 is the old struc-
1993, Jabalpur 1997, Chamoli 1999, and Bhuj 2001) ture used as Road Transport office in Jabalpur, the
have caused deep concern with regards to seismic haz- meizoseismal area of 1997 Jabalpur earthquake, India.
ards and resulting risk. These devastating earthquakes The structure undergone severe damage and is living
in urban areas in India cause heavy economical losses perilously.

4
chap-01 9/9/2004 8: 5 page 5

Figure 5. Chattris are very common in Gujarat area. Figure

shows the view of Chattri before and after the Bhuj earth-
quake.

Figure 2. RTO building damaged during Jabalpur earth-

quake of 1997, India.

Figure 6. The Parag Mahal at Bhuj before and after the

earthquake.

Figure 3. Wide spread damage was caused to temples all

around the Gujarat during Bhuj earthquake causing complete
failure.

Figure 7. The Famous clock tower in the heart of Jabalpur

city. One of the minaret fell down during Jabalpur earthquake,
India.

the overturning of long wall of Bhuj fort leading to

partial collapse.
Normally, the roofs of heritage buildings are jack-
arched, vaulted or dome shaped. The failure starts with
the separation of wall from its transverse supports,
then by its vibration as cantilever of increasing length.
Figure 4. Collapse of wall of a Bhuj fort. Finally the walls overturn or open enough to produce
the instability or loss of support of the roof, which col-
Several temples of historical and cultural values lapses. Figure 5 shows the failures of Raolakha Chhatri
went into rubbles and biting dust during past earth- and Figure 6 shows Prag Mahal at Bhuj, Gujarat.
quakes of India as depicted in Figure 3.
2.3 Separation and overturning of facade
If the façades of the forts are not tied well to the rest
2.2 Overturning of long walls
of the monument, they tend to get separated, overturn
The most commonly encountered mode of failure dur- and fell from the main structure. An example of Clock
ing past earthquakes has been overturning of the walls tower at Jabalpur whose one minaret fell since it was
due to out-of-plane inertia forces and normally precipi- stiffer than the other minarets due to presence of stair
tated by the outward thrust of the roof. Figure 4 shows is shown in Figure 7.

5
chap-01 9/9/2004 8: 5 page 6

Figure 8. Mosque with collapsed minarets during Jabalpur

earthquake. Figure 10. Earthquake damaged clock tower at Morbi,
Gujarat.

Figure 9. Partly broken and tilted minar of Idgah (the prayer

place for muslims) during Jabalpur earthquake.

Also, the heavy and tall tower can become unstable

owing to the inertia forces transverse to their plane.
Roofs or intermediate floors do not provide proper
restriction to the outward displacement of the façade.
The failure of minarets during Jabalpur earthquake are
shown in Figure 8 and Figure 9. Figure 11. Development of shear cracks and partial collapse
in Radha Krishna temple.
2.4 Diagonal cracking
The failure due to shear forces in walls through diago- shear performance has been noted in heritage buildings
nal cracking is not very frequent in heritage structures, in India. Diagonal shear cracking is common sight,
though, because the great thickness typical of the mon- due to reduced mortar shear capacity, however, in a
uments gives the walls a significant capacity to resist few cases flexural (rocking) failures are also observed.
shear forces, even if their unit shear strength is low. The typical failure of a temple and church are shown in
Nevertheless, the shear failure is commonly associated Figure 11 and Figure 12 during Jabalpur earthquake.
with walls with large openings such as in Figure 10.
Large openings can seriously undermine the
2.5 Tensile failure of structures
in-plane shear strength of these walls. The shear capac-
ity of masonry piers may not be enough to resist lateral It has been found that artistic rooftops in most of the
loads imposed by earthquake forces. Poor in-plane monumental buildings lack in proper floor and roof

6
chap-01 9/9/2004 8: 5 page 7

Figure 14. Partial toppling of temple top (pagoda) at the

palace at Morbi, Gujarat.

rest of the structure. A failure of pagoda at the palace

in Morbi is shown in the Figure 14.

Figure 12. 125 years old St. Paul Cathedral Church cracked
with partial collapse of tower. 3 SEISMIC REHABILITATION – AN
OVERVIEW

Seismic rehabilitation is used as a comprehensive

term to include all the concepts of repair, restor-
ing, upgrading, retrofitting and strengthening that
lead to reduce seismic vulnerability of any structure.
A large number of existing structures in earthquake
prone areas over the world need seismic rehabilita-
tion due to various reasons and motivations, including
codal modifications, deterioration of structures with
age or change in use/modification of structure. Earth-
quake damaged structures may need strengthening
along with repair of damaged portion for reuse (post-
Figure 13. Collapse of Chattris made up of large block stone earthquake rehabilitation). Generally, they are reha-
masonry during Bhuj earthquake. bilitated so that their improved seismic performance
is achieved from future ground motion. Seismically
weak structures, the design of which do not comply
diaphragms and the inertia forces thus generated in with existing provisions, may also need rehabilitation
the large masses at roof levels get directly transferred (pre-earthquake rehabilitation). Seismic rehabilitation
to the walls supporting each portion of the roof, regard- of existing structures by retrofitting is one of the
less of the relative lateral stiffness of the resisting most challenging tasks for structural engineers in the
members. Large concentrations of lateral forces can aftermath of earthquakes. This task more often poses
arise in some walls. Furthermore, it must be taken challenge to structural engineers who find it difficult to
into account that during the vibration of the structure retrofit a structure using conceptual retrofit schemes.
tensile stresses arise in roof that could generate large The failures of structures during the past earth-
cracks, actually separating the structure in parts lead- quakes have clearly shown the vulnerability of our
ing to partial to total collapse. Such a failure is shown habitat. The problems in India are manifold. The con-
in Figure 13 during 2001 Gujarat earthquake. struction practices are varied from region to region
Also, the façade known as pagodas with its tow- and moreover, it is the discretion of the owner, builder,
ers and buttresses can separate, due to longitudinal or engineer whether or not to apply the seismic pro-
vibration, from the rest of the structure which must visions in design and construction. Also, India has
withstand the inertia forces generated on it, and whose a very complex socio-cultural environment and its
buttresses do not contribute to the shear strength of the built environment encompasses the widest possible

7
chap-01 9/9/2004 8: 5 page 8

range: from non-engineered dwellings built without heritage structure and at the same time ensuring struc-
any technical skills to the most modern buildings. tural safety against external actions i.e. earthquakes.
Because all the earthquakes in recent years occurred in Heritage structures are defined as all those which merit
rural or semi-urban environments and because most of special care on account of their individual historical or
the deaths were caused by collapse of non engineered architectural importance, or their significance as sur-
structures, the attention most of the time is diverted viving representatives of an earlier tradition. But the
to repair/rehabilitation of these construction and thus very definition is ignored and seismic rehabilitation is
trivializing entire rehabilitation process. The engineers done with a syndrome known as cover-up of damage
and builders started strengthening buildings, as they syndrome. This is put into practice after every deadly
deemed appropriate, even though most of them did earthquake all around the world. The engineers try to
not have a prior experience of seismic issues. cover-up the damages that appear in structural system
The seismic rehabilitation of existing structures is and the associated paraphernalia. It is a quick and low-
the most challenging task and one has to take advice cost solution but precarious one. It is advocated that
of an expert instead of doing it oneself. Seismic since the structure withstood the earthquake only with
rehabilitation in its strict sense calls for three kinds some cracks, there is no problem, which leads into
of improvement in the structure under consideration ignoring the extent & pattern of damage as well as its
i.e. Repair only visual or cosmetic modifications are influence on a possible future seismic loading, which
made, Restore structural repairs such as rebuilding of could be fatal for the structurally degraded heritage
cracked walls, stitching of cracks, grouting, placement structure. These antiseismic interventions also present
of reinforcement etc. are made such that the orig- a notable impact in the constructive reality of histori-
inal performance of the structure are restored, and cal construction, and calls for precise awareness of the
Retrofit structural modifications such as jacketing, original structure and by structural choices, which are
external ties, buttresses, addition of shear/infill walls, knowingly consistent.
bracing & anchorages are made such that a higher Design and realization of the interventions for
performance of the structure is achieved than that of seismic rehabilitation should actually be character-
original structure. However, the seismic rehabilitation ized by concurring specialistic and multidisciplinary
has been used very loosely in the field. The cosmetic contributions, which ought to allow to take deci-
repairs are often confused with seismic rehabilitation. sions so as to safeguard of historical, cultural and
Seismic rehabilitation is to be carried out by structural architectonic values is ensured in conjunction with
engineer and requires proper planning and engineer- the protection of both the construction itself and
ing judgment. Firstly, the structure is to be evaluated the human lives potentially involved from structural
for its seismic resistance and then depending upon the instability. Generally speaking the problem of seis-
requirements, it is to be rehabilitated. Seismic evalua- mic rehabilitation of heritage building differs radically
tion of existing buildings in itself is a tedious task and from the problem of ordinary structures. Here the
has been new in Indian scenario. There is no standard- emphasis is on preservation of their aesthetic and his-
ized document available for the said purpose.There has torical values and then structural safety. As per Venice
been meager data available on the structure retrofitted Charter (UNDP/UNIDO, 1984), the seismic rehabili-
for their seismic resistance. The common methods in tation of heritage structures must take into account the
our country are to use traditional practices such as following:
repair of cracks, stitching, guniting, jacketing etc. It
– Respect for original material and authentic docu-
has been experimentally verified that these techniques
ments
do not add to the strength and ductility and in turn
– Respect to the valid contributions of all periods to
render the structure precarious. We are actually car-
the building
rying out repairs instead of strengthening. The newer
– Replacements of missing parts must integrate har-
techniques such as wrapping, addition of new frames,
moniously with the whole, but at the same time must
shear walls, bracing, introduction of energy dissipation
be distinguishable from the original
devices, base isolation etc. are never tried despite of
– Additions can not be allowed except in so far as
the fact that they perform well during earthquake and
they do not detract from the interesting parts of the
have been introduced all over the globe successfully.
building, its traditional setting and its relation with
This is due to lack of awareness amongst engineers as
the surroundings
regards seismic rehabilitation is concerned.
– The use of traditional techniques and materials are
So far, the discussion was not specific to heritage
clearly preferable for seismic rehabilitation.
structures. The problems get further complex when
heritage structures are to be rehabilitated for their Modern techniques and materials are admissible
improved seismic resistance. Here the goal is mani- where adequate capacity cannot be ensured by tra-
fold, the seismic rehabilitation means is to preserve ditional techniques. In this case durability and com-
and reveal the aesthetic and historical values of the patibility of the interventions should be adequately

8
chap-01 9/9/2004 8: 5 page 9

proven; otherwise, the modern techniques and mate- The following methodology is presented here for
rials should be used only in a manner that will permit seismic rehabilitation, which is based on elastic
easy corrective action at a later date if necessary. method of analysis and is broken into three stages
Stage I – Develop knowledge of as built conditions:
4 SEISMIC REHABILITATION – Step 1: Study of soil conditions at the site.
METHODOLOGY Step 2: Measurement of actual geometry of building
and its components.
Any rehabilitation strategy is case specific and has to Step 3: In-situ non-destructive and destructive testing
be dealt separately in a manner that it alleviates seis- to estimate actual strength of building
mic deficiency from the structure under consideration. materials & components.
To adopt cost-effective and appropriate retrofitting Step 4: Tests to estimate material degradation,
scheme, quantitative seismic evaluation of building is condition of the structure, extent of corrosion
prerequisite. Evaluation is a complex process, which to carefully estimate their available diameters
has to take not only the design of building but also and verify the size, number and spacing of
the deterioration of the material and damage caused reinforcing bars etc.
to the building, if any. The difficulties faced in the
seismic evaluation of a building are manifold. There is Stage II – Analysis & Verification Stage:
no reliable information/database available for existing Step 5: Preparation of 3-dimensional model of
building stock, construction practices, in situ strength structure, using measured geometry, and
of material and components of the building. The material properties.
seismic evaluation mainly relies on set of general eval- Step 6: Estimation of design lateral force on building
uation statements, since analytical methods to model using code specific design response spectra
the behavior of buildings during earthquake defining for 5% damping or site-specific response
response spectra of structure are too complex to handle spectra, if available.
with the generally available tools and calls for rigorous Step 7: Application of design lateral force on
engineering seismological exercise. The unavailabil- 3-D model to determine stress-resultants
ity of a reliable estimate of earthquake parameters, to (i.e. axial forces, shear forces, bending
which the building is expected to be subjected during, moments) in respective members &
its residual life poses another challenge. Probabilis- determination of displacements etc.
tic approach to evolve needful parameters, would call Step 8: Determination of member capacities with
for elaborate studies. Hence, for preliminary appraisal, actual cross section geometry & material
the ground motion parameters available in the respec- properties as per respective codes (It is based
tive codes may be at the macro level. As regards the on permissible stresses in flexure,
effect of local soil conditions, which are known to compression, tension and shear) and
greatly modify the earthquake ground motion, expe- Demand/Capacity ratios of members.
riences of ground accentuation and data generated Step 9: Identification of deficient members or
through collateral studies on site response should be deficiency in lateral stiffness of the structure
considered. Seismic evaluation procedure is a com- if any and verification of the damages actually
parison between some measures of the demand that incurred in the structure.
earthquake place on a structure to a measure of the
capacity of the building to resist. Evaluation is a com- Stage III – Retrofit Stage:
plex process involving number of steps like visit to Step 10: Identification of suitable retrofitting
the site and data collection, selection of either the gen- technique to rectify the deficiencies.
eral set of evaluation statements; conduct of follow-up Step 11: Estimation of the new member sizes, stiffness
field work; structural analysis for seismic evaluation etc. for the proposed scheme, and/or the
of building; characterizing seismic capacity, determine new members required.
seismic demand and verify performance in light of Step 12: Re-analysis of building to confirm the ade-
global response limits, component acceptability and quacy with the proposed retrofit techniques.
conceptual approval. Step 13: If strength and stiffness requirements are
The Demand/capacity ratio (DCR), thus evaluated satisfied, then the proposed retrofit scheme
is measure of earthquake resistance of a building. The may be adopted, else other more appropriate
DCR less than unity indicate the building is safe for retrofit schemes may be identified and steps
respective stresses under consideration. However, any from 10 are to be repeated.
DCR exceeding one indicates that the building is vul-
nerable to earthquake hazard as per respective codal The above approach is based on elastic proce-
provisions. dure and does not include failure patterns, ductility

9
chap-01 9/9/2004 8: 5 page 10

and, therefore, can be used as a first cut method fall and assuring a complete load path. The connectiv-
for retrofitting design. The non-linear analysis and ity deficiencies are within load path: wall out-of-plane
pushover analysis approaches such as one outlined in connection to diaphragms; connection of diaphragm to
ATC-40 can also be used for stage II and stage III. vertical elements; connection of vertical elements to
foundations; connection of foundation to soil.

5 SEISMIC REHABILITATION MEASURES –

CONCEPTUAL DESIGN 6.2 Modification of global behaviour
This action normally focuses on decreasing deforma-
There are umpteen methodologies available for seis- tions. Overall seismic deformation demand can be
mic rehabilitation interventions so as to achieve the reduced by adding stiffness in the form of shear walls
improved seismic performance and also to preserve or braced frames. A significant time period shift is
historical and cultural values of the heritage structure. normally required to protect deformation sensitive
The complete list of techniques is beyond the scope elements in this way. New elements may be added, or
of this paper and can be found easily in literature. The created from a composite of new and old components.
bottom line is that the selection of the specific type of Examples of such composites include filling in open-
element and material should be compatible with the ings and using existing columns for chord members
structure under consideration and should be in har- for new shear walls or braced frames.
mony with the existing structural system. It is thus also
an arduous task to develop guidelines for such selec-
tion. Conceptual design techniques, on the other hand, 6.3 Modification of local behaviour
can be systematically categorized and design strate-
gies formulated. The solution chosen for rehabilitation Rather than providing retrofit actions that affect the
is to be always dictated by aesthetic and cultural val- entire structure, deficiencies also can be eliminated
ues rather than merely satisfying technical demands. at the local, component level. This can be done by
In heritage buildings, consideration of preservation of enhancing the existing shear or moment strength of an
historic fabric usually controls the design. In many element, or simply by altering the element in a way that
cases, even the performance objectives are controlled allows additional deformation without compromising
by limitations imposed by preservation. There are vertical load carrying capacity.
some basic issues that are of concern: The yielding sequences such as beams yielding
before columns, bracing members yielding before
1. Aesthetics, including consideration of historic connections, bending yielding before shear failure in
preservation columns and walls can always be attained by local
2. Improved seismic performance retrofit in variety of ways. The connections can be
3. Long term effect on building space planning and strengthened and the shear capacity of the columns
usage and walls can be enhanced to be stronger than the shear
4. Cost of construction (cost-benefit analysis). demanded by earthquake.
All of these characteristics are always considered, Concrete columns can be wrapped with steel, con-
but an importance will eventually be put on each of crete, or other materials to provide confinement and
them, either consciously or subconsciously, and a com- shear strength. Concrete and masonry walls can be
bination of weighing factors will determine the scheme layered with reinforced concrete, plate steel, and other
chosen. materials. Composites of glass or carbon fibres and
epoxy are becoming popular to enhance shear strength
and confinement in columns and to provide shear-only
6 REHABILITATION STRATEGIES strengthening to walls. Similarly, bending strength
of unreinforced masonry walls can be increased by
The primary focus of determining a viable retrofit insertion of reinforcing or post-tensioning steel in
scheme is on vertically oriented systems because of field drilled cores. Deformation capacity can also be
their significance in providing either lateral stability increased locally by uncoupling brittle elements from
or gravity load resistance. Deficiencies in vertical ele- the deforming structure.
ments are caused by excessive interstory deformations The above-discussed actions balance one another in
that either creates unacceptable force or deformation that employing more of one will mean less of another
demands. The retrofit actions can be classified as: is needed. It is obvious that added global stiffness will
require less local deformation capacity, but it is often
less obvious that careful placement of new lateral ele-
6.1 Connectivity
ments may minimize a connectivity issue such as a
It is defined as an arrangement consisting of assuring diaphragm deficiency, very common in Indian heritage
that individual elements do not become detached and buildings. Conceptual design of retrofit is seldom a

10
chap-01 9/9/2004 8: 5 page 11

Figure 16. Jack arch floor of the circuit house which

Figure 15. Internally stitched walls of circuit house imme- underwent severe cracking.
diately after the earthquake.

one step process, several trials are required to define

satisfactory design, both to balance the retrofit actions
discussed above and to determine the exact extent of
each intervention (number and length of walls, size of
braces etc.).
Case study – Seismic rehabilitation of circuit house
(Lal bungalow) at Jamnagar, Gujarat.
January 26, 2001 Bhuj earthquake was one of the
most disastrous earthquakes in India predominantly
affecting the urban setting. This particular part of
India had a very large number of small kingdoms and
hence, the affected area is rich in cultural and heritage
structures. Numerous such structures sustained heavy
damages during the earthquake. A major challenge Figure 17. Seismically rehabilitated circuit house with its
is to resolve the conflict between being conserva- past glory.
tive in demolishing many such damaged buildings
considering them unsafe and salvaging maximum pos-
sible heritage. There have been number of retrofitting
schemes and strategies suggested by experts from all – Improving geometry of the structure by introduc-
over the world, a few of which have been successfully ing framing for walls, openings and making them
implemented. One such case study is circuit house at compatible with appropriate stiffness and integral
Jamnagar, which is a prestigious landmark structure connections
with Roman architecture. The external and internal – Several arches over doors and windows were
walls were damaged and then immediately stitched as strengthened by proving steel arches below the
shown in Figure 15 to aggravate the further damage. existing ones
The structure had jack arch flooring as shown in – Grouting and stitching of cracks
Figure 16, which got extensively damaged and later – Jacketing of walls with welded wire mesh, MS flats
in the retrofit scheme were replaced with RC slab to and chicken mesh from both sides
reduce dead weight. – Corner stiffening of walls with MS flats, lintel bands
The suggested retrofit measures is implemented over openings.
which comprises of the following:
The Figure 17 above shows the completely rehabil-
– Making available most appropriate and effective itated structure. Collapsed towers of the circuit house
load paths to the structure (It is done by providing got improved with matching brickwork profile with
steel framing system as an alternate loading frame crushed stones and polymer mortar. Artists who excel
system in addition to the exiting load bearing walls) in the particular artwork used during the historical time
– Controlling the dead weight of the structures by were called to specially prepare the towers. The entire
replacing jack arch floors with RC slabs structure was plastered with suitable matching finish.

11
chap-01 9/9/2004 8: 5 page 12

7 PROSPECTS: EXPERIMENTATION, Much more representative of the earthquake

INSTRUMENTATION AND MONITORING response are the records obtained by permanent net-
works of accelerographs placed in the structure to
The seismic rehabilitation of buildings is mainly based measure its response to actual earthquakes. For this,
on assumptions and on mechanical properties, which first of all, one has to instrument the heritage structures
are derived from an extensive experimentation per- in sufficient numbers so as to derive general conclu-
formed mainly in laboratories. However, for heritage sions about the earthquake response of typical heritage
structures this is not the case, these structures do structures.
not lend themselves to laboratory experimentation India is the ideal place to carry out seismic instru-
except the behaviour of simplest forms such as arches, mentation of heritage buildings since more than 55%
columns and walls. Most experimentation must be area could be subjected to damaging earthquake inten-
carried out directly on actual buildings. sities. Also a variety of cultural heritage buildings
Nowadays, several techniques have been developed made up of stone masonry, terracotta, bricks and rocks
for in situ measurements of structural properties.These of different periods are available. The average return
techniques are reliable and non-destructive, or only period of earthquake of magnitude more than 5 is less
slightly destructive, like sonic tomography, thermog- than two months in whole of India. The existence of a
raphy, endoscopy, rebound tests, core sampling etc. large network of ground motion instruments provides
The flat jack technique is also coming up as a power- useful reference for the interpretation of the results.
ful tool for determining Young modulus and the state Strong motion accelerographs are being installed
of stresses on the structure. in Indian cities nowadays and we have one such array
Additionally, sophisticated techniques for monitor- in Delhi, the capital city of India. These are normally
ing the behaviour of heritage buildings in terms of installed in the free-field or at the basement of modern
displacements, settlements, out-of-plumb and crack buildings. There is need to put a few of them in heritage
openings are to be implemented in the structures at structures so that the difference in the motion recorded
risk. In some of the countries, they are complimented at the base of the structure and the free-field can be
by automatic data acquisition systems, which can be obtained.These results are important for the estimation
set to give alert signals when the response level exceeds of the maximum acceleration that can be induced by
a selected threshold. Instrumentation and monitoring severe earthquakes.
of heritage buildings for the determination of seismic
response is the future prospect and would be extremely
useful for reducing associated risks.
One technique that is increasingly being used 8 CONCLUSIONS
in high-rise buildings and that can give significant
information also for heritage structures is the ambi- Earthquakes have been the predominant cause of dev-
ent vibration test. High-sensitivity accelerographs/ astation of our rich heritage. For centuries, these
seismographs are placed in different setups on criti- structures have periodically suffered strong seismic
cal points of the structure, to measure the vibration actions and have undergone a kind of natural selec-
caused by traffic, wind or micro tremors. Averages tion, so that only those that were well designed and
of a large number of measurements over a period of constructed have survived. The paper reviews the crit-
time are used to obtain general trends, while elimi- ical modes of failures commonly encountered during
nating noise and effects of specific inputs. By spectral past earthquakes.
analysis of the signals, natural frequencies of vibration As regards, the seismic rehabilitation, the concept
and mode shapes can be determined, along with cor- of preserve and reveal the aesthetic and historical val-
relations between vibrations of different parts of the ues of the structures shall have to be kept in mind
structure. in addition to the safety. Original structural system
When these measurements are repeated after sev- should in principle be kept unchanged. At most, only
eral years or after a significant earthquake occurred, local improvement of the original structural system
changes in dynamic properties can be detected, which should be accepted. The original materials and authen-
can be consequences of damage or of degradation of tic documents are to be respected. This imposes on
the structure. Ambient vibration measurements must the specialists responsible for rehabilitation, a duty, to
be interpreted with great care when used to assess the consider what limitations these considerations place
structural response to actual earthquakes. The ampli- on the choice of techniques and materials of repair and
tudes of motions and the levels of stresses imposed strengthening. The thorough knowledge of the proper-
by ambient vibration are so small that the structural ties of the original materials used is prerequisite so that
behaviour remains essentially linear elastic and the the compatibility of the new materials used for reha-
effect of yielding and discontinuities in the structure bilitation with the original ones can be ensured along
is not completely reflected. with long term durability.

12
chap-01 9/9/2004 8: 5 page 13

There are number of non-linear analytical design 2000. Jabalpur Earthquake 22 May, 1997, A geoscientific
tools are available for seismic evaluation and study, GSI Spl. Pub. No. 51, Geological Survey of India,
retrofitting of existing structures, even then many July, 2000, India.
structures are evaluated using relatively simple force- 2001. Bhuj, India Earthquake Reconnaissance Report,
2002 Earthquake Spectra, July, 2002, Supplement A to
based methods that have little or no explicit consider- Volume 18.
ation of realistic displacement demands from ground 2002. Investigation of Earthquake Resistance of Multi-
motion, displacement compatibility among dissimi- storeyed Buildings and Recommending Seismic Retro-
lar elements, non-linear response, failure modes etc. fitting Measures, CBRI Report, Roorkee, India.
There is need to bench mark these procedures so that Abrams, D.P. & Costley, A.C. 1996. Seismic evaluation on
results obtained are reliable, consistent. unreinforced masonry buildings, Eleventh World Confer-
Code specified values of seismic coefficients and ence on Earthquake Engineering, Acapulco, Mexico, June
response spectra cannot be simply extended to heritage 23–28, paper no 976.
buildings while deciding the seismic action on these Agrawal, S.K. 2000. Strong motion array as response of
layered elastic medium, Ph.D. thesis, Department of
structures. The characteristics of the ground motion Earthquake Engineering, Indian Institute of Technology,
expected at the site must be taken into account and Roorkee, India.
site-specific response spectra are to be generated. Par- Agrawal, S.K. 2000. Force Centre: A New Concept
ticular attention is required for the heritage buildings in SMA Data Analysis, ASC 2000, Symposium on
founded on very soft or liquefiable soils. Seismology, Earthquake Hazard Assessment and Earth
Despite of advances in the field of modeling, Interior Related Topics, October 10–12, 2000, Tehran.
experimentation and understanding the subject of seis- Agrawal, S.K., Chourasia, Ajay & Parashar Jalaj 2001.
mic rehabilitation, published guideline and standards, System Identification of Eight Storied Steel Framed
engineer must continue to temper their decisions Structure, Advances in Structural Dynamics and Design
(ASDD), January 9–11, 2001, SERC, Madras, India.
with experience and judgement, primarily based on Agrawal, S.K., Chourasia, Ajay, Parashar Jalaj, & Dutta,
past experiences of failures of real structures under Jayanta 2001. Experimental Investigations on Earth-
earthquake loading. quake Resistance and Retrofitting Measures of Masonry
Seismic rehabilitation of heritage structures is Houses – A Review, International Conference on Civil
a highly specialized operation and requires close Engineering (ICCE – 2001), 23–25 July, 2001, Indian
cooperation of experts of archaeology, architec- Institute of Science, Bangalore, India
ture, surveying, structural engineering, geo-technical Agrawal, S.K. & Vipul Prakash 2001. An Engineering
engineering and chemical engineering and should be Approach for Earthquake Source Modelling, Interna-
supported by computation methods and well-equipped tional Conference on Seismic Hazard with particular
reference to Bhuj Earthquake of January 26, 2001,
laboratories. 3–5 October, 2001, New Delhi, Govt. of India, Deptt.
Of Science & Technology, India Meteorological Deptt,
India.
Agrawal, S.K. & Chourasia, Ajay, 2002. Seismic safety
ACKNOWLEDGEMENTS aspects for historical monuments, Strategy & methodol-
ogy for conservation of heritage buildings & monuments
The author is grateful to Shri V.K. Mathur, Director, in India, CBRI, Roorkee, April 4–5, India.
CBRI for granting permission to write the paper. I am Agrawal, S.K., Chourasia, Ajay & Parashar, J. 2002. Seis-
indeed indebted to Dr. Paulo B. Lurenco, Porutgal, mic Evaluation & Retrofitting of Existing Building –
A Case Study, Proc. of 12th Symposium on Earthquake
Dr. Claudio Modena, Italy and Dr. Pere Roca, Spain Engineering, Indian Institute of Technology, Roorkee,
for giving me such an opportunity and sharing the India.
expert knowledge on the subject. The information Agrawal, S.K., Chourasia, Ajay, Parashar, J & Dutta, J. 2002.
gathered from web sites such as www.nicee.org, Experimental Investigation on Earthquake Resistance and
www.eeri.org, www.asi.nic.in, www.intach.org, www. Retrofitting Measures of Masonry Houses, Advances in
gsdma.org is gratefully acknowledged. Concrete & Construction Technology, Publication –3,
Interline Publishing, Bangalore, India
Agrawal, S.K., Chourasia, Ajay 2003. A pragmatic approach
for Seismic Risk reduction of Megacities in India, Indian
REFERENCES Habitat & Infrastructure – need for innovative approach,
24–26 September, CBRI, Roorkee, India.
1986. Guidelines for earthquake resistant non-engineered Agrawal, S.K. & Chourasia, Ajay 2003. Rehabilitation of
construction, The International Association for Earth- RCC Structures – How Far From Reality, Int. Conf. on
quake Engineering, Japan. ‘Innovative World of Concrete 2003, Indian Concrete
1992. Indian monument through ages, Oxford & IBH Pub- Institute, Pune, September, 2003, India.
lishing Co. Pvt. Ltd., New Delhi, India. Agrawal, S.K., Chourasia, Ajay & Parashar, J. 2003. Seis-
1999. Vulnerability Atlas of India, Building Materials & mic Vulnerability of Jabalpur Urban Area, Department of
Technology Promotion Council, New Delhi, India. Science & Technology Report, Govt. of India, India.

13
chap-01 9/9/2004 8: 5 page 14

Agrawal, S.K., Chourasia, Ajay 2004. Estimation of Seis- IS-13920 1993. Ductility Detailing of Reinforced Concrete
mic Vulnerability of Buildings in Delhi, World Congress Structures Subjected to Seismic Forces – Code of Practice,
on Natural Disaster Mitigation, Institution of Engineers Bureau of Indian Standards, New Delhi, India.
(India), New Delhi, India, Vol. 2, 159–171. IS-13935 1993. Repair and Strengthening of Buildings –
ATC-21 1988. Rapid Visual Screening of Buildings for Guidelines, Bureau of Indian Standards, New Delhi, India.
Potential Seismic Hazards: A Handbook, Applied Tech- Jain, S.K., Murty, C.V.R., Dayal, U., Arlekar, J.N. &
nology Council, Redwood city, CA, USA. Chaubey, S.K., 2001. A Field Report on Structural and
ATC-28 1991. Development of Recommended Guidelines Geotechnical Damages Sustained During the 26th Jan-
for Seismic Strengthening of Buildings Phase: Issue uary, 2001 M7.9 Bhuj Earthquake in Western India,
Identification and Resolution, Applied Technology Coun- NICEE Report, Kanpur, India.
cil, Redwood city, CA, USA. Mathur, V.K. & Agrawal, S.K. 2001. Safety and Security
ATC-40 1996. Seismic Evaluation and Retrofit of Concrete of Tall Buildings in India, CTBUH London Conference,
Buildings – Vol. I & II, Applied Technology Council, Council on Tall Buildings in Urban Habitat, December,
California Seismic Safety Commission, Redwood city, 2001, London.
California. Mathur, V.K., Agrawal, S.K. & Chourasia, Ajay 2003.
FEMA-172 1992. NEHRP Handbook for the Seismic Reha- Demand-CapacityApproach for Seismic Rehabilitation of
bilitation of Existing Buildings, Federal Emergency RC Residential Buildings, 9th International Conference
Management Agency, Building Seismic Safety Council, on Civil & Structural Engineering Computing (CIVIL-
Washington, D.C. COMP 2003), Heriot Watt University, Edinburgh, UK.
FEMA-178 1992. NEHRP Handbook for the Seismic Mathur, V.K., Agrawal, S.K. & Chourasia, Ajay 2004. Micro-
Evaluation of Existing Buildings, Federal Emergency zonation Studies as impacted by recent earthquake in
Management Agency, Building Seismic Safety Council, India, World Congress on Natural Disaster Mitigation,
Washington, D.C. Institution of Engineers (India), New Delhi, Vol. 2, 81–91.
Gavarini, C. & Bruno S. 1996. Towards a complex approach Meli, R. & Ramirez, R.S. 1996. Considerations on the
to preservation and seismic protection of monuments, seismic safety of historical monuments, Eleventh World
Eleventh World Conference on Earthquake Engineering, Conference on Earthquake Engineering, paper no. 2087.
paper no. 2088. Murty, C.V.R. 2002. Quantitative Approach to Seismic
Giuffre, A. & Carocci, C. 1996. Vulnerability and mitigation Strengthening of RC Frame Buildings, Proc. on Seismic
in historical centers in seismic areas – criteria for the for- Assessment and Retrofitting of Buildings, Indian Institute
mulation of a practice code, Eleventh World Conference of Technology, Mumbai, India.
on Earthquake Engineering, paper no. 2086. Penelis, G. Gr., 1996. Techniques & materials for structural
Holmes, W.T. 2000. Risk Assessment & Retrofit of Exist- restoration, Eleventh World Conference on Earthquake
ing Buildings, Twelfth World Conference on Earthquake Engineering, paper no. 2089.
Engineering, paper no. 2826. SP-22, 1982. Explanatory Handbook on Codes for Earth-
IS-456 2000. Plain and Reinforced Concrete – Code of quake Engineering, Bureau of Indian Standards, New
Practice, Bureau of Indian Standards, New Delhi, India. Delhi, India.
IS-1893 2002. Criteria for Earthquake Resistant Design
of Structures, Part-1: General Provisions and Buildings,
Bureau of Indian Standards, New Delhi, India.

14
chap-02 9/9/2004 8: 5 page 15

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

Research for seismic redesign of historic masonry buildings

M. Tomaževič
Slovenian National Building and Civil Engineering Institute, Ljubljana, Slovenia

ABSTRACT: Historic masonry buildings have been built on the basis of tradition and experience without taking
into consideration specific rules for earthquake resistance. On the basis of the analysis of earthquake damage
observations, the main causes of damage are determined and adequate measures to improve the seismic behavior
developed. In order to assess the seismic resistance, mechanical properties of masonry materials should be
determined by appropriate testing methods. Experimental research is needed to evaluate the observed mechanisms
and propose numerical models for seismic resistance verification. Experimental research and testing is also
needed to provide information regarding the efficiency of measures and technologies developed to improve the
seismic resistance and obtain the data needed for seismic redesign. Experiences obtained at Slovenian National
Building and Civil Engineering Institute in Ljubljana are discussed as an example.

1 INTRODUCTION Historic masonry buildings reflect the knowledge

and construction technology of the time of their con-
Historic masonry buildings have been built on the basis struction. Whereas in the case of the new buildings,
of tradition and experience. Only rarely some specific material properties and structural characteristics are
measures for earthquake resistance have been provided specified by the design, in the case of historic buildings
at the time of construction, such as the tying of the the material properties and structural characteristics
walls with iron ties and/or using connecting stones and need to be determined on actual materials before
cut stones in the corners and wall intersections to pre- redesign. Therefore, testing and experimental research
vent separation and disintegration of walls in the case is the only tool:
of stone-masonry buildings. Considering the response
– To obtain reliable quantified data needed for struc-
to earthquakes, the behavior of historic masonry build-
tural assessment, such as the values of mechanical
ings was generally not adequate. Therefore, in the
properties of materials, data about homogeneity and
cases of reconstruction of earthquake-damaged areas
structural conditions, moisture, etc.;
or seismic rehabilitation of historic urban and rural
– To understand the seismic behavior of buildings
nuclei campaigns, seismic strengthening measures
and develop numerical models needed for structural
should be provided in most cases.
verification and redesign;
After several decades of experience, practice and
– To verify the efficiency of methods and technolo-
research, rules and procedures have been developed
gies developed for strengthening and repair.
which need to be considered in the redesign.They com-
prises seismic assessment of existing structures, decid- In this contribution, some experiences obtained at
ing upon technical solutions to be used to improve the Slovenian National Building and Civil Engineering
resistance, redesign to verify the efficiency of the Institute (ZAG) in Ljubljana, Slovenia, regarding the
selected measures, as well as execution of works. testing and research in seismic behavior of historic
The decisions should be taken by considering many masonry buildings, will be presented and discussed.
criteria, among them criteria for preservation and Following the conclusions made on the basis of the
restoration of historical monuments which represent analysis of earthquake damage observations, some
a specific case. General guidelines and recommen- basic test results and recommendations for their use in
dations are given in Eurocode 8-1-4: General rules – seismic redesign will be given. Some experiments to
strengthening and repair of buildings (new draft of the verify the most widely used strengthening techniques
document is under development). will be also presented.

15
chap-02 9/9/2004 8: 5 page 16

2 MECHANICAL PROPERTIES, STRUCTURAL Since masonry is non-elastic, non-homogeneous

CONDITIONS AND RESISTANCE TO and unisotropic structural material, the values of
SEISMIC LOADS sectional forces and material properties are deter-
mined based on the gross-sectional geometrical char-
In order to choose the appropriate technical inter- acteristics and assuming the elastic, homogeneous
ventions and ensure adequate seismic performance and isotropic global properties, hence simplifying
of historic buildings, the carrying out of a thor- the analysis and redesign. The following mechanical
ough structural diagnosis is unavoidable. The type of quanti-ties determine the load-bearing capacity and
structural system and its elements should be iden- deformability of masonry walls:
tified and actual situation regarding the quality of
– the compressive strength of masonry, f,
structural materials should be verified. Information
– the tensile strength of masonry, ft ,
about the mechanical characteristics of structural
– the modulus of elasticity, E,
materials should be obtained and quantitatively eval-
– the shear modulus, G.
uated. In this regard, investigations and tests are
unavoidable. Although sometimes costly and time Whereas compressive strength and modulus of elas-
consuming, they represent the only means to obtain ticity define the load-bearing capacity of masonry
reliable data for structural evaluation and subsequent walls at gravity loads, the tensile strength and shear
redesign. modulus define the load-bearing capacity at seis-
Many testing methods are available to obtain the mic loads. In the latter case, however, information
relevant information, ranging from non-destructive, regarding the ductility and energy dissipation capac-
semi-destructive to destructive methods. However, by ity, as well as strength and stiffness degradation and
non-destructive testing, such as impact echo and radar deterioration is also of relevant importance.
tests, micro seismic wave propagation measurements Because of specific characteristics of historic
etc., only qualitative data regarding the general struc- masonry it is not possible to determine the mechan-
ture of masonry walls (such as composition of layers ical properties of masonry on the basis of tests of
and amount of voids) and possible existing damage their constituent materials, such as bricks, stone and
can be obtained. Non-destructive tests are helpful in mortar, in the laboratory. In order to obtain reliable
order to identify the general, qualitative situation of information from such tests, a correlation between
the building under consideration. data, obtained on constituents and data obtained on
Semi-destructive tests such as flat-jack tests pro- masonry wall specimens should be previously known.
vide information regarding the stress state in the walls It is also difficult to reproduce existing masonry walls
as well as strength of the masonry. However, system- in the laboratory, even though thorough chemical and
atic calibration tests still need to be carried out in mechanical tests of the mortar and other constituent
order to quantify the obtained information in terms of materials may have been carried out. The testing of
mechanical properties of masonry needed in structural specimens, cut out from the existing walls and tested
verification. in the laboratory (Fig. 1) and in-situ testing of the
Unfortunately only by means of laboratory and/ existing walls in the building (Fig. 2) is therefore
or in-situ destructive testing, some basic paramet- preferred to testing of the laboratory constructed
ers, which determine the seismic behavior of his- specimens (Fig. 3).
toric masonry walls and structures, can be obtained.
Within the framework of the research project named
ONSITEFORMASONRY, a joint effort of 16 Euro-
pean universities, research institutions and companies
has been recently made to study the possibilities of
these methods and propose respective guidelines and
recommendations.

2.1 Testing
Stone, mixed stone-and-brick and sometimes brick-
masonry are the most common materials used for the
construction of historic buildings, with lime mortar as
the bonding material. Stone-masonry walls are usu-
ally built of two outer layers of uncoursed stone, with
an inside infill consisting of a mix of smaller stones
with lime mortar. Because of the method of construc-
tion of stone-masonry, many voids exist in the walls, Figure 1. Lateral resistance test of a brick-masonry
uniformly distributed over the entire volume. specimen, cut from the building, in the laboratory.

16
chap-02 9/9/2004 8: 5 page 17

Standardized testing procedures are used in the vertical load to simulate the working stress state in the
case of vertical compression tests. In the case of lat- wall is kept constant, whereas the horizontal load in the
eral resistance tests, however, the seismic situation of form of induced displacements is applied according
loading is simulated by subjecting the specimens to to an appropriate pattern which simulates the cyclic
constant vertical load, whereas the horizontal loads character of seismic loading. Since the removing of
are applied cyclically, acting either in both or in one specimens of adequate size from the building and their
direction only. transport to the laboratory is usually a time consuming
In the laboratory conditions, vertical and horizontal and costly procedure, the tests on the existing masonry
loads are applied by means of hydraulic actuators. The are carried out on the site. At an appropriate location
in the building, the specimen to be tested is separated
from the surrounding masonry by vertical cuts on both
sides. In order to prevent accidental collapse of the
floor in the case of collapse of the specimen, the floor
structure is supported with posts. A system of steel
connectors and hydraulic jacks is provided to induce
lateral loads.

2.2 Mechanical properties of masonry

In accordance with the principles of Eurocodes, char-
acteristic values of strength of materials, reduced by
partial safety factor for masonry materials γM , are used
in seismic resistance verification. In accordance with
standard EN 1052-1, which determines the testing of
compressive strength of masonry, three specimens are
tested and the mean value of compressive strength f
evaluated. The characteristic compressive strength of
masonry fk is determined as the smaller value of either
f /1.2, or the minimum obtained value fmin :

Figure 2. In-situ lateral resistance test of a stone-masonry

wall. Although the testing procedures are not standard-
ized, it is proposed that the same principle is also used
for the evaluation of characteristic values of either
compressive or tensile strength obtained from the lab-
oratory or in-situ tests carried out on existing masonry
walls. Since such tests are expensive, rarely more than
two specimens of the same masonry type are tested. It
is therefore also proposed, that either the average value
of two test results or a single result, reduced by 1.2, be
considered as the characteristic value.
Typical characteristic values of material properties
for different types of existing stone-masonry walls,
tested by Slovenian National Building and Civil Engi-
neering Institute, are given in Tables 1 and 2 (Turnšek
et al. 1978, Sheppard & Tomaževič 1986, Tomaževič
et al. 2000).
According to Eurocodes, the values of partial safety
factors for masonry γM are determined in depen-
dence on the category of execution and manufacturing
control. The values for the new construction in non-
seismic situation vary from γM = 1.5 to γM = 3.0. In
seismic situation, however, it is recommended that 2/3
of the usual values, but not less than γM = 1.5 are taken
Figure 3. Laboratory lateral resistance test of a laboratory into account. Nevertheless, since in the case of the
constructed stone-masonry wall. existing structures material properties are determined

17
chap-02 9/9/2004 8: 5 page 18

by in-situ testing of actual materials, the γM values, plaster and opening the walls, the value of partial
as required for the new construction, can be further safety factor γM = 1.2 can be considered;
reduced. This is also suggested by Eurocode 8-1-4: – In the case where the characteristic values are
Strengthening and repair of buildings. obtained in the literature, but no identification tests
Considering the correlation between the results have been carried out, the value of partial safety
of earthquake resistance analyses and earthquake factor γM = 1.7 should be taken into account in the
damage observations, carried out after the past seismic resistance evaluation.
earthquakes, and following the guidelines given
Mean values of in-situ test results are taken into
in Eurocode 8-1-4, the following recommendations
account as characteristic values of shear modulus G
regarding the values of partial safety factor γM to be
and modulus of elasticity E.
taken into account in the redesign of historic masonry
As has been found, the mechanical properties of
buildings, can be proposed:
stone-masonry strongly depend on the structure of the
– In the case where in the given region and for a wall and the way of construction. It has been also
given type of masonry the mechanical properties found that the values vary from region to region, so
are determined either by in-situ tests or in the that individual values are not generally applicable.
laboratory by testing specimens, taken from the Whereas a number of data exist about the val-
existing masonry, the value of partial safety factor ues of the tensile strength and shear modulus, not
γM = 1.0 can be considered in seismic resistance many data are available about the compressive strength
verification; and modulus of elasticity. In particular, the results of
– In the case where the characteristic values are in-situ tests are missing, difficult to carry out because
obtained in the literature, as for example the val- of relatively large forces needed to test sufficiently
ues given in Tables 1 and 2, and identification of a large specimens. As the test results indicate, the com-
given type of masonry is carried out by removing pressive strength of stone-masonry is relatively low.
Since even in the existing state the walls of stone-
masonry houses do not possess a normal reserve in
Table 1. Characteristic values of tensile strength ftk and vertical load-bearing capacity, any removal of parts of
design values of shear modulus G of stone-masonry. the walls in the case of enlarging the existing, or mak-
ing new door and window openings, or any adding of
Type of masonry ftk (MPa) G (MPa) new stories to the existing building, may cause severe
problems.
Two leaf, uncoursed lime stone, 0.02 60
poor lime mortar, rural
Two leaf, mix of uncoursed quartz 0.08 100 2.3 Resistance to seismic loads
sandstone, slate and lime-stone,
lime mortar, urban As the analysis of earthquake damage and subsequent
Two leaf, mix of uncoursed quartz 0.12 40 experiments indicate, three types of failure modes of
sandstone, slate, lime-stone and brick; walls, shown in Figure 4, can be observed, depend-
lime mortar, homogeneous, urban ing on the geometry of the wall (height/width ratio),
Two leaf, uncoursed lime stone, 0.05 80 quality of materials as well as boundary restraints and
poor lime mortar, rural dwellings loads acting on the wall.
Two leaf, uncoursed lime stone, 0.07 170
Shear failure is a typical failure mode in the case
lime mortar, public buildings of historic brick and stone masonry buildings. It takes
place where the principal tensile stresses, developed in
the wall under a combination of vertical and horizontal
loads, exceed the tensile strength of masonry materi-
Table 2. Characteristic values of compressive strength fk als. Characteristic diagonal cracks develop in the wall
and design values of modulus of elasticity E of stone-
masonry.
just before the attainment of maximum resistance.

Type of masonry fk (MPa) E (MPa)

Two leaf, uncoursed lime stone, 0.40 1950

poor lime mortar, rural
Two leaf, mix of uncoursed quartz 0.28 390
sandstone, slate and lime-stone,
lime mortar, urban
Two leaf, uncoursed lime stone, 0.82 2600
poor lime mortar, rural dwellings Figure 4. Typical failure modes of masonry walls, subjected
to seismic loads.

18
chap-02 9/9/2004 8: 5 page 19

For the case of the shear failure, the resistance of • Maximum resistance, determined by maximum
the walls is calculated by (Turnšek & Čačovič 1971): resistance Hmax , attained during test, and corre-
sponding displacement dH max .
• Ultimate state, determined by maximum displace-
ment attained during test dmax and corresponding
resistance Hd max .
Obviously, the initial slope of the idealized envelope
where: is best defined with a secant stiffness at the forma-
Hs,w = shear resistance of the wall, tion of cracks, which is called effective stiffness of the
Aw = area of the horizontal cross section of wall Ke = Hcr /dcr . Ultimate ductility factor µu , which
the wall, defines the displacement (rotation) at which the wall
σo = compressive stress in the wall, does not resist any more to lateral seismic load, is
b = coefficient of distribution of shear stresses. defined as a ratio µu = du /de . Usually, the displace-
ment at which the resistance degrades to 80% of the
During the lateral resistance tests, the relationships maximum, is considered as ultimate, du .
between the cyclically acting lateral seismic load and
displacements (rotation) of the wall are obtained in
the form of hysteresis loops (Fig. 5). To use the infor-
mation obtained by testing in the calculations, the 3 SEISMIC BEHAVIOR AND REDUCTION OF
envelope of the experimentally obtained hysteresis DESIGN SEISMIC LOADS
loops is idealized as a bilinear envelope (Fig. 6).
Three limit states are defined in the envelope: 3.1 Observations and numerical models
• Crack limit, determined by displacement dcr and The analysis of damage caused by earthquakes to his-
resistance Hcr at the formation of the first signif- toric stone- and/or brick-masonry buildings in urban
icant cracks in the wall, which change the slope of and rural nuclei, indicated that cracks at the corners
the envelope. and at wall intersections, which occur as a result
of insufficient connections and lack of connection
between the walls and floors, represent a character-
istic damage pattern. Sometimes, separation of walls
and even out-of-plane collapse occurred. Also, many
times, despite the favorable structural layout of those
buildings in plan and good connection of walls, the
quality of masonry materials was not good enough to
spare the walls from diagonal cracking, disintegration,
and ultimate collapse.
Although the structural typology of masonry build-
ings varies in different regions, their damage resulting
from earthquakes can be classified in an uniform way.
The following typical types of damage can be identi-
Figure 5. Typical lateral load–displacement hysteresis fied by the analysis of the observed earthquake damage
loops, obtained during laboratory test. patterns:
– Cracks between walls and floors;
– Cracks at the corners and at wall intersections;
– Out-of-plane collapse of perimetral walls;
– Cracks in spandrel beams and/or parapets;
– Diagonal cracks in structural walls;
– Partial disintegration or collapse of structural walls;
– Partial or complete collapse of the building.
The analysis of damage patterns can clearly iden-
tify the weak and good points of different structural
systems. On the basis of damage analysis, the fail-
ure mechanisms of individual structural walls and
the entire structural system can be defined. On the
basis of the observed mechanism, forces that develop
Figure 6. Idealization of experimental resistance envelope in the structural system during earthquakes can be
with bilinear relationship. determined. On the basis of such analysis, causes of

19
chap-02 9/9/2004 8: 5 page 20

Figure 7. Typical shear cracks in the walls of a historic

stone-masonry house.

inadequate behavior can also be identified and sub-

sequent measures for the improved seismic resistance
proposed. The following typical causes of damage to
masonry buildings may be identified:
Figure 8. Stone-masonry house model at collapse during
– Inadequate structural integrity; laboratory shaking table test.
– Inadequate structural resistance;
– Inadequate structural layout, and
– Inadequate foundation system.
As a result of structural configuration and quality
of masonry materials, typical observed damage pat-
terns indicate that shear mechanism prevails in the
case where historic buildings are subjected to seismic
actions (Fig. 7).
Shear mechanism, which has been also observed
by testing the models of historic masonry buildings on
the shaking table (Fig. 8), is schematically presented
in Figure 9.
Typically, unreinforced masonry walls, which fail
in shear, are weaker than spandrels. Since the flexural
capacity of the walls’ bottom and upper sections is Figure 9. Shear mechanism: coupled shear wall with weak
too low to allow the transfer of the bending moments piers.
to horizontal elements, i.e. lintels and spandrels, the
individual walls can be considered as symmetrically resistance envelopes of individual walls which com-
fixed at both ends, what simplifies the calculations to pose that story. A push-over type method has been
a great degree. However, the assumption of rigid floor developed for the calculation of story resistance enve-
diaphragm action should be fulfilled in order to obtain lope. It is assumed that individual walls resist the
reliable results as regards the distribution of loads and imposed displacement up to the attainment of their
seismic resistance. ductility capacity. Ultimately, it is also assumed that
Different numerical models are used to calculate the they carry the vertical loads although they fail for lat-
seismic resistance of historic masonry buildings. The eral loads. The procedure is schematically presented
behavior of complex structures is modeled by means in Figure 10 and validated with experimental results
of kinematic and finite element models. However, in in Figure 11.
the case of regular historic buildings where the shear
mechanism determines the behavior and resistance,
3.2 Reduction of design seismic loads
good estimate of seismic resistance can be made by a
relatively simple calculations (Tomaževič et al. 1978, Seismic resistance of a historic building is the basic
Tomaževič 1997). criterion for the decision whether the building should
In the latter case, the resistance envelope of be strengthened or not. In the case where the seis-
the critical story is obtained as a superposition of mic resistance verification indicates that the building

20
chap-02 9/9/2004 8: 5 page 21

Table 3. Code (BSC d,u ) and proposed reduced values of

ultimate design base shear coefficient (BSC d,ur ).

EMS intensity VI VII VIII IX

ag 0.05 0.10 0.20 0.30

BSC d,u 0.08 0.17 0.33 0.50
γn 1.00 1.00 0.84 0.67
BSC d,ur 0.08 0.17 0.25 0.33

where:
ag = seismicity depending design ground acceleration,
S = soil factor,
η = damping correction factor,
q = structural behavior factor.
The resulting values of BSC d,u for different expec-
Figure 10. Construction of story resistance envelope on the ted intensities of earthquakes (seismic zones) and firm
basis of bilinear resistance envelopes of walls. soil are given in Table 3.
As the analyses and observations after earthquakes
indicate, structural strengthening of historic masonry
buildings is needed in most cases. Whereas the tying
of the walls should be provided in every case, the expe-
riences show that the strengthening of stone-masonry
walls by cement-grouting is unavoidable in the zones
of moderate and high seismicity, whereas in the zones
of low seismicity, the decision whether the stone-
and/or brick-masonry walls need to be strengthened
or not, depends on the quality of existing masonry.
As the experience indicated, by applying the usual
technical measures to historic masonry buildings, the
values of the coefficient of seismic resistance CSRu ,
i.e. the ratio between the ultimate resistance Hu and
the weight of the building W, greater than 0.3 are
difficult to attain, particularly in the case of build-
ings higher than 3 stories. It would be practically
Figure 11. Correlation between experimental and impossible to attain the resistance required for the
calculated story resistance envelopes. new construction without adding new structural ele-
ments. On the other hand, however, there were many
cases observed where even stone-masonry buildings
under consideration will not resist the expected seis- resisted the earthquakes, although the calculated val-
mic loads with an acceptable degree of damage, the ues of CSRd,u were lower than required for the given
building needs structural strengthening. By using ade- seismic zone.
quate mathematical models, the weak elements of the According to Eurocode 8-1-4, the design ground
structure where the strengthening is needed, can be acceleration may be reduced in the case where the
identified. anticipated total costs of strengthening the entire build-
Normally, the same level of design seismic loads ing inventory of particular urban areas would sharply
should be considered in the redesign of historic build- increase if ag values would be raised towards the code
ings as in the case of the new construction. According required level, as well as where code required ag values
to Eurocode 8, the design seismic load in terms of the for redesign of a monument would lead to completely
ultimate design base shear coefficient BSC d,u , i.e. the unacceptable architectural alterations.
ratio between the design seismic load and weight of Taking into account the recommendations given
the building, is determined by: in Eurocode 8-1-4, it is proposed that, for practical
redesign of existing stone-masonry buildings in urban
areas, maximum allowable reduction, i.e. reduction
factor γn = 0.67 is used in the zone IX. Since there
is no reduction proposed for the zones VII and lower,

21
chap-02 9/9/2004 8: 5 page 22

Table 4. Seismic resistance of typical existing the integrity of structural system during earthquakes
(ft = 0.06 MPa) and strengthened (ft = 0.11 MPa) 2-story is ensured by means of the tying of the walls at floor
stone-masonry houses in terms of coefficient of seismic levels. This only makes possible to utilize the available
resistance (CSRu = Hu /W ) in x- and y-directions. energy dissipation capacity of the structure, assumed
by structural behavior factor q.
Wall/floor area Existing Strengthened
(%) building building
Bldg.
no. x-dir y-dir x-dir y-dir x-dir y-dir 4 EXPERIMENTAL EVALUATION OF
METHODS FOR STRENGTHENING AND
1* 12.0 9.1 0.21 0.19 0.25 0.25 REPAIR
2* 10.9 6.4 0.20 0.15 0.27 0.22
3 6.9 8.6 0.22 0.25 0.25 0.33
As the studies of earthquake damage to historic
4 12.1 11.1 0.33 0.31 0.42 0.38
5 4.7 14.6 0.17 0.33 0.19 0.47 masonry buildings indicated, the following basic
6 7.2 14.3 0.16 0.31 0.21 0.47 requirements need to be fulfilled in order to achieve
7 15.1 13.7 0.29 0.25 0.40 0.33 adequate seismic behavior:
8 10.5 9.5 0.31 0.25 0.39 0.29
– Structural integrity should be ensured, i.e. structural
9 10.5 9.9 0.23 0.26 0.31 0.34
10 10.3 10.2 0.22 0.26 0.28 0.35 walls should be adequately tied and connected, and
11 11.9 10.3 0.28 0.29 0.29 0.34 floors should be rigid enough and well anchored to
12 9.8 10.9 0.23 0.26 0.32 0.34 the walls in order to provide uniform distribution of
13 8.8 8.33 0.23 0.27 0.31 0.33 seismic loads onto the walls and prevent excessive
14 10.6 12.0 0.28 0.28 0.35 0.36 out-of-plane bending;
15 9.7 12.0 0.27 0.34 0.34 0.47 – Structural walls should be uniformly distributed
16 7.9 4.2 0.26 0.19 0.35 0.21 in both orthogonal directions as well as along the
height of the building in order to prevent torsional
* ft = 0.08 MPa for existing and 0.14 MPa for strengthened and/or other phenomena;
houses.
– Structural walls should be resistant enough to resist
the expected seismic loads;
i.e. γn = 1.00, the value of reduction factor to be used – Foundation system should be capable of transfer-
in zone of EMS intensity VIII, γn = 0.84, is a linear ring the ultimate forces developed in the upper
interpolation between the values of reduction factors structure into the foundation soil.
proposed for the zones IX and VII. The resulting
Many methods have been developed for improv-
reduced values of ultimate design base shear coeffi-
ing the resistance of historic buildings to meet these
cient to be considered in the redesign BSC d,ur are also
requirements. Not all have been experimentally veri-
given in Table 3.
fied. Different testing techniques are used for the ver-
Typical results of a recent study of seismic resis-
ification of strengthening methods. In most cases, the
tance of buildings located in the zone, where intensity
efficiency of the proposed strengthening method can
VIII earthquakes by EMS scale actually occurred in
be evaluated by testing the masonry walls either in the
the time interval of 22 years, are given in Table 4
laboratory or in-situ, using the same testing techniques
(Tomaževič et al. 2000). Most of the buildings, ana-
as in the case of the walls in the original state.
lyzed in Table 4, have already been strengthened after
However, to investigate the seismic behavior and
the earthquake, which damaged the area in 1976.
mechanism of the entire buildings, models of build-
As the observations and analysis after the subse-
ings are tested on the shaking tables. The testing
quent earthquake which occurred in 1998 indicated,
of models on simple seismic simulators has several
the seismic behavior of strengthened buildings was
advantages:
adequate.
The results of this analysis confirm the possibility – Model tests are cheaper and simpler than prototype
of reducing design seismic loads in the zones of high tests;
and moderate seismicity as proposed in Table 3, with- – Relatively simple testing facilities can be used;
out risking inadequate seismic behavior. The behavior – Global behavior and failure mechanisms can be
of historic buildings, strengthened to such level, is reliable simulated;
expected to be adequate. The buildings will not col- – Modeling techniques can be adjusted to possibilities.
lapse, but the amount of damage will be larger than in
However, this testing techniques also has disadvan-
the case of the new construction. However, the damage
tages, such as:
will still be repairable.
It should be emphasized that seismic resistance – Modeling scale is limited because of technological
verification of historic buildings for seismic loads, restrains;
given in Table 3, is only reasonable in the cases where – Structural details cannot be studied.

22
chap-02 9/9/2004 8: 5 page 23

In the case of model tests, the general correlation

between the prototype (P) and model (M) quantities is
given by:

where:
q = physical quantity,
S = scale factor.
Different modeling techniques are available
(Tomaževič & Velechovsky 1992). Depending on the
basic correlations between the model and prototype
materials, complete or simple models can be tested.
If the dynamic behavior and failure mechanisms are
studied, the following general requirements need to be
fulfilled in the first place:

– Similitude of dynamic behavior which requires

similar distribution of masses and stiffnesses along
the height of prototype and model, and
– Similitude of failure mechanism which requires
similar working stress/strength of material ratio in
structural walls of the prototype and model.

At ZAG, the complete modeling approach is usu-

ally followed, and model materials have been devel-
oped, which conform with the requirements of this Figure 12. Hysteresis loops obtained by testing prototype
approach. (a) and model stone-masonry wall (b).

4.1 Tying of walls and interventions in floors

In order to evaluate the efficiency of steel ties and
other measures, which are most commonly used to
improve the integrity of historic masonry buildings, a
series of shaking table tests of models of simple stone-
and brick-masonry historic houses has been carried out
(Tomaževič et al. 1991, 1993, 1996).
To ensure the integrity of masonry structures during
earthquakes, hence utilizing the available resistance of
the walls, the walls are tied at floor levels with steel
ties, placed on both sides of the walls and anchored at
the ends on steel plates. In order to study the mech-
anism of action of steel ties, two series of models
of typical simple historic stone- and-brick masonry
houses with wooden floors have been tested on the
shaking table. In each series of tests, referential mod-
els without any ties and models with tied walls have
been tested. The models have been built as com-
plete models at 1:4 scale by using either natural stone
with mortar of reduced strength (Fig. 12) or model Figure 13. Brick-masonry house model without steel ties
bricks, made of model mortar with composition of mix at collapse during laboratory shaking table test.
designed to meet the requirements of complete model
similitude, respectively. Steel ties have been made of Whereas wooden floors of models without ties
fully annealed wire and monitored with strain-gauges did not prevent the separation of the walls, exces-
to measure the strains developed in the ties during sive out-of-plane vibration and disintegration of the
shaking. upper story (Fig. 13), the models with identical

23
chap-02 9/9/2004 8: 5 page 24

Figure 15. Critical segment of building used for the design

of steel ties.

Figure 14. Brick-masonry house model with steel ties at

collapse during laboratory shaking table test.

structure, where the walls had been tied with steel

ties, retained integrity up until the final collapse of
the lower story (Fig. 14). As a result of retained integ-
rity, the lateral load-resistance and deformability, as
well as energy dissipation capacity was significantly
improved.
Ties design recommendations resulted from this
study. The ties of the walls, orthogonal to seismic
motion and vibrating out of the plane, behaved sim-
ilarly as reinforcing steel of a r.c. bond-beam made
along the wall at floor level. Taking this into con-
sideration, wall ties should be basically designed for
Figure 16. Sliding of the slab and pushing out of walls
bending moments developed in the equivalent bond- during brick-masonry house models with r.c. slabs shaking
beam, formed by a strip of the wall between the ties, table test.
due to out-of-plane vibration of the wall.
On the other hand, steel ties should also be ver- where:
ified for forces which might develop in the critical
Dmin = minimum bar diameter,
segment of the structure in the direction of seismic
n = number of bars,
motion (Fig. 15).As the shaking table tests have shown,
fy = yield strength of steel,
a kind of truss mechanism develops in the structural
Hu,seg = resistance capacity of a critical segment.
walls in the direction of seismic loads, where the shear
induced in the walls is carried over from story to story Within the same series of shaking table tests, the
by means of tension developed in longitudinal ties. As models where wooden floors have been replaced with
indicated by the measurements, forces developed in rigid r.c. slabs, have been also tested. The experiments
the ties in the direction of seismic motion at ultimate indicated, that the replacement of wooden floors with
state were of the same order of magnitude as seismic rigid slabs does not always provide the best solution.
shear induced in the models. In case that the slabs are not properly anchored into
The following formula has been proposed to calcu- the walls, they may slide on the bearing surface and
late the minimum diameter of a steel tie: cause separation and pushing out of parts of the walls
(Fig. 16).
Similar phenomena have been observed in the
case of stone-masonry buildings after earthquakes
(Fig. 17).

24
chap-02 9/9/2004 8: 5 page 25

Table 5. Characteristic values of tensile strength ftk and

design values of shear modulus G of cement-grouted stone-
masonry walls.

Type of masonry ftk (MPa) G (MPa)

Two leaf, uncoursed lime stone, 0.06 100

poor lime mortar, rural
Two leaf, mix of uncoursed quartz 0.12 100
sandstone, slate and lime-stone,
lime mortar, urban
Two leaf, mix of uncoursed quartz 0.16 450
sandstone, slate, lime-stone and brick;
lime mortar, homogeneous, urban
Two leaf, uncoursed lime stone, 0.09 170
poor lime mortar, rural dwellings
Two leaf, uncoursed lime stone, 0.17 400
lime mortar, public buildings

The procedures of strengthening and repair can be

Figure 17. Damage due to new r.c. slab, which pushed out
parts of masonry wall. classified into the following main groups:
– Repair of cracks;
– Repointing the joints with cement mortar;
This indicates that the slabs, which in the case – Coating (reinforced cement, ferrocement, PRFC)
of stone-masonry buildings are usually supported by on one or both sides of the walls;
the inner wythe of stone-masonry walls and are not – Injections of cement-, lime- or epoxy-based grout;
properly anchored and/or connected with the outer – Prestressing the walls in vertical or horizontal
wythe, may cause more damage to structural walls direction;
than what would have resulted from flexible wooden – Reconstruction of the most damaged parts of the
floors. walls.
Both experimental research and on-site damage
The efficiency of various strengthening methods
observations have indicated, that the replacement of
has been verified by both laboratory and in-situ tests.
wooden floors with r.c. slabs does not always provide
Usually, the walls are first tested in the existing state.
the best solution.
In this phase of testing, the data about the original
strength and deformability properties of masonry are
4.2 Strengthening and repair of masonry walls obtained, and the testing, even in the laboratory con-
Different technical measures have been developed for ditions, is terminated before the final collapse of the
strengthening and repair of masonry walls. Some of specimens occurs. Then, the damaged specimens are
them are based on engineering judgment only and repaired and/or strengthened, and retested. In the case
have never been actually verified. Others, however, of laboratory tests (Fig. 3), the testing of walls in the
have been verified by testing either in the laboratory strengthened state is carried out up until the final col-
or in-situ. Sometimes, the retrofitted or rehabilitated lapse of the specimens. In-situ, however (Fig. 2), the
buildings have been even subjected to real earth- collapse is avoided so that the testing is terminated
quakes, so that the validity of some methods has been before serious damage occurs to the walls. However,
verified in the real situation. Different methodologies the testing is carried out well after the attainment of
are available for strengthening of different types of maximum resistance.
masonry walls. Namely, the choice of the most suit-
able technical solution depends mainly on the type and 4.2.1 Stone-masonry walls
quality of masonry, but also on the required degree of Typical values of tensile strength ft and shear modu-
improvement. lus G, obtained by in-situ testing the cement-grouted
While efficient interventions in stone-masonry are stone-masonry walls, are given in Table 5. As the
more or less limited to injecting the cementitious grout comparison of values of tensile strength and shear
into the void parts of the walls, various possibilities modulus of existing and strengthened masonry indi-
are available for the repair and strengthening of brick- cates (Tables 2 and 5), by means of grouting the
masonry walls. wall with cementitious grout the lateral resistance

25
chap-02 9/9/2004 8: 5 page 26

200

150 Grouted
H (kN)

100
Existing

50

0
0.0 0.7 1.4 2.1 2.8
R (%)

Figure 18. Lateral load–rotation relationships obtained by

the in-situ testing of original and cement-grouted stone Figure 19. Comparison of hysteresis envelopes obtained
masonry wall. during lateral resistance tests of stone-masonry walls, grouted
with different grout mixes.
Table 6. Characteristic values of compressive strength fk
and design values of modulus of elasticity E of cement- have shown, although the compressive strength of dif-
grouted stone-masonry walls. ferent grout mixes varied from 7 to 32 MPa, this had
no influence on the lateral resistance of the grouted
Type of masonry fk (MPa) E (MPa) walls (Fig. 19).
It is believed that one of the possible reasons why
Two leaf, uncoursed lime stone, 0.80 8200 strength of the grout did not influence the lateral resis-
poor lime mortar, rural tance of the walls can be explained as follows. The
Two leaf, mix of uncoursed quartz 1.67 2600 potential resistance is determined by the strength of
sandstone, slate and lime-stone, the original mortar and stone. Since the injection grout
lime mortar, urban does not penetrate into the original mortar, the poten-
Two leaf, uncoursed lime stone, no data no data tial resistance does not change. However, by bonding
poor lime mortar, rural dwellings the loose parts of the wall together into a solid struc-
ture and preventing the delamitation and bulging when
subjected to lateral loads, hence providing the integ-
is improved. However, the degree of improvement rity of the wall, the injected grout simply activates the
depends on the quality of the existing masonry. potential load-bearing capacity.
Typical relationships between the lateral load and On the basis of the results of this study, it can be
displacements, obtained by in-situ testing of existing concluded that an appropriate composition of the grout
and cement-grouted stone-masonry walls, are shown mix can be designed for each particular type of stone-
in Figure 18. It can be clearly seen that not only the masonry and for each particular problem to be solved
lateral resistance capacity, but also the rigidity of the in a historic stone-masonry building. Locally avail-
tested walls significantly increased (Tomaževič et al. able materials compatible with the original texture of
2000). historic walls can be used to reduce the amount of
As was already mentioned, not many data are avail- cement in the grout and to reduce the expected side
able as regards the compressive strength f and mod- effects to an acceptable level.
ulus of elasticity E. Some indicative results are given
in Table 6. As can be seen by comparing the values 4.2.2 Brick masonry walls
given in Tables 3 and 6, the degree of improvement in As regards the brick masonry walls, the efficiency of
load-carrying capacity to vertical loads, obtained by reinforced cement plaster coating has been tested by
cement-grouting of the masonry, is significant. laboratory and in-situ testing. It has been observed
In order to investigate the influence of grout that, depending on the geometry and quality of the
strength on the resistance of stone-masonry walls, a existing wall as well as the way of application of the
series of walls has been built and tested in the labora- coating, the failure mechanism can be changed from
tory (Tomaževič & Apih 1993). Namely, to prevent the shear to bending (Fig. 20).
negative effects of pure cement grout, water repellent As indicated by experiments, the improvement in
additives are added and/or part of cement is replaced the lateral resistance is inversely proportional to the
with inert aggregates in the form of fine sands. The quality of the original wall. It is significant in the case
additives and reduction of the content of cement reduce of poor-quality masonry, but not so in the case of good-
the strength of the grout. However, as the experiments quality walls. Typical example is shown in Figure 21.

26
chap-02 9/9/2004 8: 5 page 27

2,0

Fourier Amplitude
1,5

1,0

0,5

0,0
0 5 10 15 20
Frequency [Hz]

Figure 22. Frequency response spectrum, obtained by

ambient vibration measurements of a stone-masonry tower.

measured during excitation, the coefficient of equiva-

lent viscous damping can be also determined.
Ambient vibration tests are more simple to carry
out. In this case, vibrations of the building caused by
seismic microtremors, traffic, wind and other environ-
Figure 20. Tensile rupture of reinforcing bar, observed
during in-situ lateral resistance test of a brick masonry wall, mental influences are measured with sensitive trans-
strengthened by application of reinforced cement coating. ducers, placed on the structure at different locations
along the height of the structure. The vibrations are
measured for sufficientlly long time and analyzed. On
the basis of the calculated Fourier response spectra,
information is obtained about the structure’s natural
vibration periods and damping (Fig. 22).
Both methods can be also used to determine
dynamic characteristics of structural elements, if nec-
essary. However, it is to emphasize that in both cases
data about the dynamic characteristics are obtained in
the range of very small amplitudes of vibration, espe-
cially in the case of ambient vibration tests. The data
obtained by these methods are therefore only valid
in the elastic range of vibrations and may be used
for the verification of mathematical models devel-
oped for seismic behaviour of structures in the elastic
range. On the basis of these data, no final conclusion
Figure 21. Average shear stress–rotation relationships can be drawn regarding the ultimate behaviour of the
obtained by the in-situ-testing of original and reinforced tested buildings, such as energy dissipation capacity,
cement-coated brick-masonry wall. ductility and ultimate resistance.
In the case where the causes of damage, observed
during visual inspection of the building, are not evi-
5 STRUCTURAL EVALUATION dent, long-term observations of building’s behaviour
are many times needed to definitely determine the rea-
In the case where information is needed about dynamic son for damage. Namely, not all observed damage can
characteristics of the building under consideration, be attributed to previous earthquakes, but can also be a
such as natural periods of vibration and damp- result of foundation settlements, temperature changes,
ing, special tests are carried out. Different methods vibrations induced by traffic, or other reasons. In order
and techniques are available. By means of dynamic to successfuly remedy the situation, the causes of dam-
exciters, fixed on the floors, vibrations of the build- age should be removed before the intervention in the
ing are induced. In this case, the frequency of rotating damaged structural element is carried out.
eccentric masses is electronically controlled and the For this purpose, the structure is instrumented
response of the structure at different points along the with displacement, strain and vibration transducers
height of the building at each frequency of excita- and other suitable instruments, and the readings are
tion is measured. By evaluating the resonance curves made either in regular time intervals at different

27
chap-02 9/9/2004 8: 5 page 28

environmental conditions (summer, winter, wet, dry) carried out and published in the last decade. The reader
or continuously. For example, settlements and tilting may find more detailed information in the referenced
of the structure are measured with geodetic methods, publications.
the propagation and/or opening and closing of cracks
with deformeters, whereas velocity transducers are in
most cases used for monitoring the dynamic effects. REFERENCES
Sometimes, the structure is also monitored in order
to verify the efficiency of structural interventions. Par- Eurocode 8. prEN 1998-3. Design of structures for earth-
ticularly, typical structures are monitored with strong- quake resistance. Part 3: Strengthening and repair of
motion accelerographs to record their response during buildings. CEN, Brussels (under development).
Eurocode 8. ENV 1998-1-4. 1996. Design provisions for
expected strong seismic events in the future.
earthquake resistance of structures – Part 1–4: Gen-
eral rules – Strengthening and repair of buildings. CEN,
Brussels, 1996.
6 CONCLUSIONS Sheppard, P. & Tomaževič, M. 1986. In-situ tests of load-
bearing capacity of walls of old masonry buildings. Proc.
The evaluation and redesign of historic masonry struc- 4th Nat. Congress on Earthquake Engineering, Vol.2,
tures to remedy and/or improve the seismic resistance Cavtat: 85–92.
cannot be made unless reliable structural diagnosis Tomaževič, M. 2000. Seismic redesign of existing masonry
had been previously carried out. In order to obtain the buildings. European Earthquake Engineering, 14 (3):
data needed for redesign, besides information regard- 59–66.
ing the type, geometry and conditions of the building Tomaževič, M., Klemenc, I. & Lutman, M. 2000. Strength-
and structural elements, quantitative data regarding the ening of existing stone-masonry houses: Lessons from
the earthquake of Bovec of April 12, 1998, European
mechanical properties of structural materials should Earthquake Engineering, 14 (1): 13–22.
be determined. In order to make a reliable struc- Tomaževič, M. 1999. Earthquake-resistant design of
tural diagnosis, in-situ or laboratory destructive and masonry buildings. London: Imperial College Press.
non-destructive tests should be carried out. Differ- Tomaževič, M. 1997. Seismic resistance verification of
ent testing techniques and methods are available for masonry buildings: following the new trends. In Peter
this purpose. However, if non-destructive methods are Fajfar & Helmut Krawinkler (eds.), Seismic Design
used, a previous calibration is necessary. Methodologies for the Next Generation of Codes:
Dynamic characteristics of the building’s structural 323–334. Rotterdam: Balkema.
system may be defined by forced vibration or ambient Tomaževič, M., Lutman, M. & Weiss, P. 1996. Seismic
upgrading of old brick-masonry urban houses: tying of
vibration techniques. Sometimes, the structure is mon- walls with steel ties. Earthquake Spectra, 12 (3): 599–622.
itored so that short and long-term observations and Tomaževič, M. & Apih, V. 1993. The strengthening of stone-
measurements are performed to collect the relevant masonry walls by injecting the masonry friendly grouts.
information for structural evaluation. European Earthquake Engineering, 7 (2): 10–20.
On the basis of the analysis of observations Tomaževič, M., Lutman, M. & Velechovsky, T. 1993. Aseis-
made after earthquakes and subsequent experimental mic strengthening of old stone-masonry buildings: is
research to simulate the seismic behaviour of historic the replacement of wooden floors always necessary?
buildings on earthquake simulators, mechanisms of European Earthquake Engineering, 7 (2): 34–46.
seismic behaviour are defined and numerical models Tomaževič, M. & Velechovsky, T. 1992. Some aspects of
testing small scale masonry building models on sim-
for seismic resistance verification are developed. ple earthquake simulators. Earthquake Engineering &
On the basis of structural evaluation, decision can Structural Dynamics, 21 (11): 945–963.
be made as regards the necessary structural interven- Tomaževič, M., Weiss, P. & Velechovsky, T. 1991. The
tions. A large variety of different methods for seismic influence of rigidity of floors on the seismic behaviour
strengthening and repair of historic masonry buildings of old stone-masonry buildings. European Earthquake
have been developed. Before practical application to Engineering, 5 (3): 28–41.
buildings, however, the efficiency of these methods Tomaževič, M., Turnšek, V. & Terčelj, S. 1978. Computa-
should be experimentally verified. Reliable data about tion of the shear resistance of masonry buildings. Report
the efficiency of such methods and numerical models ZRMK-IK. Ljubljana (in Slovene).
Turnšek, V., Terčelj, S., Sheppard, P. & Tomaževič, M. 1978.
to simulate their mechanisms of action during earth- The seismic resistance of stone-masonry walls and build-
quakes, should be obtained and/or developed on the ings. Proc. 6th European Conference on Earthquake
basis of experimental research. Engineering, Vol.3, Dubrovnik: 275–282.
Turnšek, V. & Čačovič, F. 1971. Some experimental results
on the strength of brick masonry walls. Proc. 2nd Int.
ACKNOWLEDGEMENTS Brick-masonry Conference, Stoke-on-Trent: 149–156.

The paper has been prepared on the basis of results

of research, which the author and his colleagues have

28
chap-03 9/9/2004 8: 5 page 29

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

The importance of investigation for the diagnosis of historic buildings:

application at different scales (centres and single buildings)

L. Binda
Politecnico di Milano, Italy

ABSTRACT: Investigation on historic buildings is needed for different reasons. Material properties, wall con-
struction technique, hidden voids, inclusions, masonry and timber decay, structural damages, etc., are information
used for the structure safety control and for the choice of appropriate materials and techniques for repair. Non
destructive (NDT) or minor destructive (MDT) techniques are required, in order to avoid invasive tests. These
techniques need a calibration both for the testing procedure and for the elaboration software of the collected data.
At the general level of historic centres, MDT and NDT can only be applied by sampling positions representative
of a whole wall or group of constructions. In these cases the budget can be very low, but even few information
can be meaningful for the diagnosis; this can be the case of historic centres in seismic areas. The paper introduces
different strategies and different level of investigation also through some case histories.

1 INTRODUCTION worse, as a “concrete”. The assumption for masonry

structures in seismic areas were that, they should
Restoration was in the past reserved to monumental behave like a “box” with stiff floors and stiff connec-
buildings. Restorers were few experienced profession- tions between the walls, no matter which was their
als who took care for years and sometime for their geometry or material composition.
professional life of the same monument or group of As a consequence all the masonry buildings were
monuments. treated as if they belonged to the same structural
After the second world war, in Italy as in other typology.
mediterranean countries, houses and palace of the In seismic areas the strengthening project implied
historic centres were frequently abandoned by the the use of the same intervention techniques: substitu-
inhabitants for more attractive new modern buildings tion of timber-floors and roofs with concrete ones, wall
out of the centres, with heating and all the facilities injection by grouts, use of concrete tie beams inserted
not available in the old houses. in the existing walls.
So the historic centres were left to the poorests and This despite the typology of the structure and
to the immigrants lowering the level of maintenance also despite the wall construction technology (one,
of historic building. two, three leaf, stone/brick). This type of intervention
In many cases the entire historic centre was aban- unaware of the construction typologies usually fails in
doned by the old owners searching for a job in the seismic area (Figs. 1, 2) (Binda et al., 2003).
industrial cities.
On the other hand in high schools and the universi-
ties, teaching of old traditional materials as masonry
and wood was substituted by concrete, steel and new
high-tech materials.
So the knowledge about historic construction tech-
niques and materials became lower and lower.
When some collapses occur or exceptional events
cause heavy damage, then the structural engineers are
called to produce projects for reconstruction or repair.
As frequently happened in the recent past, due to
lack of knowledge and of appropriate analytical mod-
els, masonry was simply treated as a one type material, Figure 1. Collapse of Figure 2. Separation of leaves
as homogeneous as concrete, steel, wood and what is a repaired walls. in repaired stone masonry.

29
chap-03 9/9/2004 8: 5 page 30

UI 197 UI 198
a) UI 199
b)

Sub-horizontal courses

U.M.I.25 U.M.I.24 U.M.I.23

c)

Figure 3. Example of: (a) simple isolated building; (b) row Irregular courses
building; (c) complex building.
Figure 4. Prospect and section morphologies.
A remedy to the failure can only be an appropriate
design for intervention based on a deep knowledge the results obtained can be unclear and even mislead-
of the building typology and of the materials and ing (Binda et al., 1992) e.g. the stress values given by
structure behaviour, reached through an appropriate a flat jack test can appear unreliable if the designer
investigation. chooses the test points without any hypothesis in mind
or does not realise that the value obtained is a local one.
A design for complex and detailed investigation can
2 WHY INVESTIGATION IS NECESSARY be proposed if it concerns a complex building with
many unsolved problems concerning: geometry, hid-
After the above premises it is necessary to give an den features, high variability in materials and types of
answer remembering that: damage, a complex evolution of volumes and struc-
tural damages along the time (Binda et al., 2000b).
– the knowledge of the old building construction tech-
In some cases the investigation can be limited to
nique and materials was lost during the last century,
geometrical and damage survey and to few laboratory
therefore it has to be rebuilt;
and in situ tests.
– buildings belong to construction typologies (Fig. 3)
Different level of investigation have then to be
which are different according to the building use
defined and guidelines can be given for each level.
and to the local materials;
– masonry is a composite with different section mor-
phology: one-two-three leaves, regular irregular, 3.1 Historic centres or urban aggregates
made with brick and/or stones (Fig. 4);
– masonry should not be considered similar to the The necessity of investigation at urban level becomes
concrete; they are extremely different; very clear when dealing with urban areas under risk
– analytical models should be calibrated by experi- of exceptional events as earthquake, floods, eruptions,
mental investigation and applied appropriately to fires, or longterm events as slides, subsidiance, etc.
check the structural safety. In very old or antique centres where the buildings
can be abandoned, misused or without maintenance
for long-time, the knowledge on the behaviour of even
3 WHICH LEVEL OF INVESTIGATION largely distributed building typologies in the same area
are lost or forgotten. In the case of earthquake or subsi-
Investigation on site and in laboratory has an economic dence this knowledge at the level of similar typologies
impact on the budget available for the preservation in the centre can be recovered by learning from the past
intervention. events which are the most diffused types of damage
If the level and type of investigation procedures (e.g. collapse mechanisms, crack pattern etc.) (Binda
is not clear and appropriate to the necessity of the et al., 2004).
building knowledge the available funds can be misused In those cases only the collection of geometry,
(Binda et al., 1999). material properties and type of damage characteris-
Furthermore if the questions to be answered by the tic for each building typology together with simpli-
investigation are not clear to the designer, the use of fied matematical models can be the base to study

30
chap-03 9/9/2004 8: 5 page 31

and propose repair and preventive measure for the

future, respectful of the existing buildings.
From basic investigation can also emerge clearly
very difficult problems to be solved by more complex
investigation procedures.
For very simple buildings as rural or two to three
0 1 3 5m

floor houses, the simple type of investigation can be

appropriate if applied to every building and not only
at urban scale even to study the choice of preservation
techniques.

3.2 Preservation of complex and monumental

buildings
0 1 3 5m

If the aim is the preservation of buildings which have Figure 5. Crack pattern survey of a Montesanto building
complexity of volumes due to their evolution in dif- damaged by the earthquake.
ferent subsequent times or monuments as palaces,
churches, towers, arenas, etc., then the inspection and
investigation have to be punctual and produce a deep
knowledge of all the details of the structural and non
structural parts.
A methodology of investigation has nevertheless to
be applied to the knowledge of the building, from the
basic operations to the more complex.
These operations can involve: geognostic inves-
tigation, geometrical, topographic, photogrammetric
surveys, laboratory tests on materials sampled to be
representative of the material used in the building, non Figure 6. Typical crack pattern of a pillar under heavy
destructive testing on the presence of humidity, voids, compressive stresses.
multiple leaf in masonries, slightly destructive testing.
Tests on timber structures, coring and visual inspec- is not available. The knowledge of the geometry is not
tions in depth, on site diffused mechanical test, etc. only used from the architectural point of view, but also
archive historical documentation and on site precise to understand the distribution of damage and decay,
observation of the building evolution, should also be and is necessary for the structural analysis.
carried out (Binda et al., 2000b). If the building is complex (row building, agglom-
erate), then the evolution of its volume should be
detected by careful visual inspection to see the signs
4 BASIC INVESTIGATION PROCEDURE (lack of wall connection in the prospects at contact
beween two parts of the building built in different
In order to give an appropriate methodology which can times, etc.). This is an important operation particu-
be applied with effectiveness taking into account the larly in seismic areas when studying the vulnerability
previous section 3, it is possible to state some priorities of the sites (Binda et al., 2004).
in the applications of investigation procedures. Essential for the safety of the building is certainly
Some operations must be always conducted due to the crak pattern survey reported on plan and prospects.
their extreme importance. Cracks, out of plumb, loss of material, etc. when
Visual inspections, photographic campaign and col- reported on the geometrical survey can give a clear
lection of data on the site and on the single building indication of the phenomena interesting the structural
from historic archives have to come first in order to movements (Figs. 5, 6).
understand the environment and the building features. The crack pattern can be in seismic areas also
Then the geometrical survey, a map of the materials compared to a damage abacus (Fig. 7).
used, a decay map, the individuation of the structural
elements, the crack pattern survey and the foundation
and soil survey. 4.2 Survey of masonry texture and of the
morphology of the wall section
4.1 Geometrical and crack pattern survey
This survey is very important for two reasons: (i) to
A design for intervention cannot be prepared if the understand the mechanical behaviour of the wall in
geometry of the structural and non structural elements order to avoid unexpected damages under earthquake

31
chap-03 9/9/2004 8: 5 page 32

Church Walls Civil Building Church Tower

50

45 Church plaster
Church mortar
40 Tower plaster
35 Tower mortar
Civil Building plaster
30 Civil Building mortar
[%] 25

20

15

10

0
SiO2 Al2O3 Fe2O3 CaO MgO Na2O K2O SO3 Loss on
ign.

Figure 10. Results of chemical tests on mortars specimens

sampled from an historic centre (Campi, Umbria).

Figure 7. Some examples from the collapse mechanisms 160

UNIAXIAL COMPRESSION TEST ON CYLINDRICAL STONE SPECIMENS

abacus. 
140
Scaglia dry Conglomerato dry
120
Stress [N/mm2]

100 Working direction (A)

Scaglia Bianca
Scaglia sat.
80 Conglomerato
Breccia
Travertino
60 Pietra Maiolica Pietra Maiolica dry

40

Breccia sat. Conglomerato sat.

20
_ Breccia dry
Travertino dry Travertino-sat. _
εl εv
0
-10 -8 -6 -4 -2 0 2 4 6 8 10
Strain [
m/mm]

Figure 11. Some results on stones specimens sampled from

an historic centre (Campi, Umbria).
Figure 8. Damages after Umbria earthquake: collapse of
the outer leaf of the wall.
4.3 Laboratory test
A minimum number of laboratory tests is necessary
even for a basic investigation (Binda et al., 2004).
They have the aim of defining the chemical physical
and mechanical characteristics of the component mate-
rials. In the case of masonry they characterise mortars,
stone and bricks. For mortars mechanical tests are usu-
ally impossible, so chemical (Fig. 10) and petrographic
tests can be made to identify the type of binder, of
Regular texture aggregates and the mortar composition.
For stone and bricks, petrographic analyses, phys-
Figure 9. The regularity is also apparent, the section is very ical and mechanical (Fig. 11) tests are important,
poor. particularly to detect states of damage and their causes
and to choose new material for restoration compatible
with the original ones.
(Fig. 8), (ii) to understand the injectability of the wall
in case of repair.
4.4 Foundation and soil survey
It should be clarified that an apparent regular wall
can hide a very irregular section (Fig. 9). If the section Small exploratory shafts or core drilling have to be
is not visible local inspection have to be made. done for the geometrical survey of the foundations.

32
chap-03 9/9/2004 8: 5 page 33

The material characteristics and the soil properties

have also to be known.
In-situ and laboratory tests on the foundation soil
are also needed in order to predict settlements or the
state of stress and strain of the soil under new loading
conditions after repair or to understand the causes of
a collapse (Binda et al., 1992).

4.5 Minor destructive tests

To understand the morphology of a masonry wall it Figure 12. Drilled core and reconstruction.
is important a direct inspection. Sometimes it could
be performed by removing few bricks or stones and
surveying photographically and drawing the section
of the wall.
In some cases it is possible to core boreholes in the
most representative points of the walls. Coring should
be done with a rotary driller using a diamond cutting
edge. This operation is rather simple but has limits. The
drilled core is usually very decohesioned (Fig. 12) so it
is almost impossible to detect the quality of the original
materials. Inside the boreholes additional investiga-
tions can be made by the use of borescopy. A small
camera may be inserted into the borehole allowing a
detailed study of its surface and try a reconstruction
of the wall section.
Nevertheless the interpretation of the results is a Figure 13. Placing the flat jack.
very difficult operation, sometimes hopeless; it should
be remembered that boroscopy can only give a general
stratigraphy of the section.
Other slightly destructive tests can be used to give 800
2 3 4 5
more information on site about the masonry compo-
Displacement [micron]

nents as: (i) the Schmidt hammer rebound test; (ii) 600
4

the penetration tests proposed in different ways, like 3

probes, drillers, etc. correlate the depth of penetra-
400 2
tion to the material mechanical properties; (iii) the
pull-out tests can only be used on bricks and stones. 5
They can be considered as surface or small penetra- 200
tion techniques, which can be used for a preliminary
investigation. 0
0.00 0.50 1.00 1.50 2.00
Stress [MPa]
4.6 Flat jack test
Figure 14. Single flat-jack tests carried out at the Monza
The method was originally applied to determine the Tower.
in-situ stress level of the masonry. The firsts applica-
tions of this technique on some historical monuments
(Rossi, 1982), clearly showed its great potential. The
test is carried out by introducing a thin flat-jack into the flat-jack is placed inside the cut (Fig. 13) and the
mortar layer, in masonry with regular and thin joints. pressure is gradually increased to obtain the distance
The test is only slightly destructive (ASTM, 1991), measured before the cut (Figs. 13, 14).
(Binda & Tiraboschi, 1999). The displacement caused by the slot and the ones
The determination of the state of stress is based on subsequently induced by the flat-jack are measured by
the stress relaxation caused by a cut perpendicular to a removable extensometer before, after the slot and
the wall surface; the stress release is determined by a during the tests. Pf corresponds to the pressure of the
partial closing of the cutting, i.e. the distance after the hydraulic system when the displacement reach those
cutting is lower than before (ASTM, 1991). A thin read before the slot is executed, that is zero.

33
chap-03 9/9/2004 8: 5 page 34

the other (Fig. 15). The two jacks delimit a masonry

sample of appreciable size to which a uni-axial com-
pression stress can be applied. Measurement bases
for removable strain-gauge or LVDTs on the sam-
ple face provide information on vertical and lateral
displacements.
In this way a compression tests is carried out on
an undisturbed sample of large area. Several load-
ing could be performed at increasing stress levels in
order to determine the deformability modulus of the
masonry in its loading and unloading phases (Fig. 16).
It is interesting to compare these last results to the
stress level measure in order to verify the present state
Figure 15. Double flat-jack test Tower of the Cathedral of of the masonry in relation with its last potentialities
Monza (Binda et al., 1998b). (Binda et al., 1998b).

4.0 5 NON DESTRUCTIVE PROCEDURES

The necessity of establishing the building integrity

3.0 or the load carrying capacity of a masonry building
arises for several reasons including: (i) assessment
Stress [N/mm2]

of the safety coefficient of the structure (before or

2.0 after an earthquake, or following accidental events like
hurricanes, fire, etc.), (ii) change of use or extension
of the building, (iii) assessment of the effectiveness
of repair techniques applied to structures or mate-
1.0 Local stress rials, and (iv) long-term monitoring of material and
structural performance.
εl εv NDE can be helpful in finding hidden characteris-
0.0 tics (internal voids and flaws and characteristics of the
-4.0 -2.0 0.0 2.0 4.0 6.0 8.0
wall section) which cannot be known otherwise than
Strain [
m/mm] through destructive tests (Binda et al., 2000b).
The types of tests available at present are mainly
Figure 16. Double flat-jack test on West side of the Monza
based on the detection of the physical properties of
Tower (Binda et al., 1998b).
the wall.
In fact the only in-situ mechanical tests available
The equilibrium relationship is the fundamental are flat-jack, hardness, penetration and pull out tests.
requirement for all the applications where the flat-jack The flat-jack tests give local measurements and are
are currently used (ASTM, 1991): slightly destructive: nevertheless they can give directly
the values of mechanical parameters. In the case of
Sf = Kj Ka Pf ND tests, a correlation between the measured param-
Sf = calculated stress value eters and the mechanical ones is usually difficult, but
Kj = jack calibration constant (<1) they can give an overall qualitative response of the
Ka = slot/jack area constant (<1) masonry. At present the most diffused ND techniques
Pf = flat-jack pressure are represented by the sonic (or ultrasonic), radar and
thermography tests.
The use of flat-jacks for stonework made with irreg-
Up to now most of the ND procedure can give only
ular stones is not so easy, due to the difficulty of
qualitative results; therefore the designer is asked to
finding regular joints; therefore the cut for the inser-
interpret the results and use them at least as com-
tion of the jack is done directly in the stone courses. It
parative values between different parts of the same
must be pointed out that the flat-jack test in the case of
masonry structure or by using different ND techniques.
multiple-leaf walls gives results concerning only the
outer leaves.
The test described can also be used to determine the
5.1 Thermovision
deformability characteristics of a masonry. A second
cut is made, parallel to the first one and a second jack Thermovision is a NDT, which has been applied
is inserted, at a distance of about 40 to 50 cm from since several years to works of art and monumental

34
chap-03 9/9/2004 8: 5 page 35

An elastic wave is generated by a percussion or by an

electrodynamics or pneumatic device (transmitter) and
collected through a receiver, usually an accelerometer,
which can be placed in various positions.
The elaboration of the data consists in measuring
the time the impulse takes to cover the frequence
and amplitude of the walls of distance between the
transmitter and the receiver. The use of sonic tests
for the evaluation of masonry structures has the
following aims:
• to qualify masonry through the morphology of the
wall section, to detect the presence of voids and
Figure 17. Investigation on hidden steel tie rods (Binda flaws and to find crack and damage patterns (Binda
et al., 2003). et al., 2001);
• to control the effectiveness of repair by injection
buildings. The thermographic analysis is based on the technique in others which can change the physical
thermal conductivity of a material and may be passive characteristics of materials (Binda et al., 2001).
or active. The passive application analyses the radia- The first applications of ultrasonic tests to the eval-
tion of a surface during thermal cycles due to natural uation of masonry materials and structures have been
phenomena (insulation and subsequent cooling). If the carried out on long time ago in the sixties (Aerojet,
survey is active, forced heating to the analysed surfaces 1967). Several efforts have been put in the tentative
are applied. of interpretation of the data from sonic and ultrasonic
A camera sensitive to infrared radiation collects the tests (Binda et al., 2001), (Abbaneo et al., 1996).
thermal radiation. In fact each material emits energy The limitation given by ultrasonic tests in the case
(electromagnetic radiation) in this field of radiation; of very inhomogeneous material made the sonic pulse
this radiation is characterised by a thermal conduc- velocity tests more appealing for masonry.
tivity, that is the capacity of the material itself of The fundaments of wave propagation through solids
transmitting heat, and its own specific heat. The result allow to recognise the theoretical capabilities and lim-
is a thermographic image in a coloured or B/w scale. itations of the technique. The velocity of a stress wave
At each tone corresponds a temperature range. Usually passing through a solid material is proportional to the
the differences of temperatures are fraction of degree. density ρ, dynamic modulus E, and Poisson’s ratio ν of
Thermovision can be very useful in diagnostic; in the material. Resolution in terms of the smallest recog-
fact it is applied to identify areas under renderings nisable features is related to the dominant wave-length
and plasters that can hide construction anomalies. It (as determinate by the frequency) of the incident wave
is particularly interesting for studies on frescoed walls and also to the size of the tested element.
(Lenzi et al., 1997). Other applications can be: (i) sur- Wave-length, λ, is determinate by a simple relation-
vey of cavities (Fig. 17), (ii) detection of inclusions of ship between velocity, v and frequency f : λ = v/f .
different materials, (iii) detection of water and heat- Hence for a given velocity as the frequency
ing systems, (iv) moisture presence (Fig. 17). In the increases the wave length decreases, providing the
presence of moisture, the camera will find the coldest possibility for greater resolution in the final veloc-
surface areas, where there is continuous evaporation ity reconstruction. However there is also a relation-
(Fig. 17). ship between frequency and attenuation of waveform
In the diagnosis of old masonries, thermovision energy. As frequency increases the rate of waveform
allows the analysis of the most superficial leaves. It attenuation also increases limiting the size of the
is necessary to point out that the penetration depth wall section, which can be investigated. The opti-
of this technique is limited, so it is unable to locate mal frequency is chosen considering attenuation and
anomalies, which are hidden in the inner part of a resolution requirements to obtain a reasonable combi-
thick masonry. The technique is often sensible to the nation of the two limiting parameters. In general it is
boundary condition of the tests. Sometimes shapes are preferable to use sonic pulse with an input of 3.5 kHz
detected, caused by different local emissions and not for inhomogeneous masonry.
by effective variations. Input frequency changes with the characteristics of
the superficial material (e.g. presence of thick plas-
ter or cracks). The sonic test in this case shows a very
5.2 Sonic pulse velocity test
important limit. Due to the wall structure or to the pres-
The testing methodology is based on the generation of ence of a thick plaster (with fresco) the high frequency
sonic or ultrasonic impulses at a point of the structure. components could be filtered.

35
chap-03 9/9/2004 8: 5 page 36

Church of SS. Crocifisso in Noto (SR) Sonic tests on pillar D1

Sonic test grid and cracks pattern of pillar D1 Sonic velocity obtained by test SC-D1
40 cm
2100
2000
1900
6 ns
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
1.65 m 800
1 2 3 4 5 6 7 Sonic test 700
1.50 m 8 9 10 11 12 13 14
SC-D1 600
500 40 cm
1.35 m 15 16 17 18 19 20 21
400
300
200
6 ns
1.20 m
100
22 23 24 25 26 27 28 0
1 2 3 4 5 6 7
2100
2000
1900
1800
PILLAR D1 - side D 1700
1600
1500
1400
1300
1200 40 cm
1100
1000
900
6 ns
800
700
600
Velocit soniche (m/sec)

500
400
300
200

Sonic test
7 6 5 4 3 2 1 1.65 m
100
0 250 cm
SC-D1 14 13 12 11 10 9 8 1.50 m 8 9 10 11 12 13 14

21 20 19 18 17 16 15 1.35 m 2100
2000
1900
28 27 26 25 24 23 22 1.20 m
1800
1700
Timber frame
1600
1500
1400
1300
1200
PILLAR D1 - side B 1100
1000
900
800
700
33.13 82.78 37.63 600
500
400
d 300
20.9

200
100
0
15 16 17 18 19 20 21
a c 2100
2000
1900
1800 Figure 19. Localisation of a wooden element in a frescoed
145.62
101

1700

wall (Binda et al., 1998).

1600
1500
1400
1300
1200
1100
1000
900
800
22.48

700
1 2 b 600
500
3 4 5 6 7 400
300
153.54 200
100
0
Pillar horizontal section
22 23 24 25 26 27 28
280 cm
Wall section

cm
Figure 18. Crocifisso Church at Noto (Italy): results of
12 99
sonic tests applied to a pillar.

It is important to stress that the pulse sonic velocity

ns
is characteristic of each masonry typology and it is
impossible to generalise the values. The tests, then,
have to be calibrated for the different types of masonry 0 0
directly on site. 0 280 cm
Figure 18 shows the crack pattern and the results of
sonic tests applied to a pillar of the Crocifisso Church Figure 20. Radargram of the wall section at the Malpaga
at Noto (Sicily – Italy). Castle. The leaves positions and the presence of a void on the
right side are recognisable (Binda et al., 1999a).

5.3 Georadar
presence of moisture and hopefully the presence of
Among the techniques and procedures of investigation multiple-leaves in stone masonry.
which have been proposed in these last years, georadar The method is based on the propagation of short
seems from one hand to be most promising, from the electromagnetic impulses, which are transmitted into
other to need a great deal more of study and research the building material using a dipole antenna. The
(Lenzi et al., 1997), (Binda et al., 1998), (Binda et al., impulses are reflected at interfaces between mate-
1999a). When applied to masonry, the applications of rials with different dielectric properties, i.e. at the
radar procedures can be the following: (i) to locate surface and backside of walls, at detachments, voids,
the position of large voids and inclusions of different etc. When the transmitting and receiving antennas,
materials, like steel, wood, etc. (Fig. 19); (ii) to qualify which are often contained in the same housing, are
the state of conservation or damage of the walls; (iii) moved along the surface of the object under inves-
to define the presence and the level of moisture; (iv) to tigation, radargrams (colour or grey scale intensity
detect the morphology of the wall section in multiple charts giving the position of the antenna against the
leaf stone and brick masonry structures (Fig. 20). travel time) are produced. Measuring the time range
Georadar seems to be a powerful tool to detect between the emission of the wave and the echo, and
the presence of voids and structural irregularities, the knowing the velocity of propagation in the media it

36
chap-03 9/9/2004 8: 5 page 37

would be possible to know the depth of the obstacle

in the wall. In the real cases, the velocity is unknown
because it changes from one material to the other or
in the presence of voids. Furthermore the velocity is
higher in dry walls, and lower in wet walls (Binda et al.,
1999a).
The interpretation of the radar data involves the
identification of significant anomalies. It should be
a recognition process detecting features on the records
that are characteristic of known signatures. Identifi-
able features on a radar record are continuous reflec-
tion from layers or reflection from discontinuities like
voids and local inhomogeneities in the masonry.
The main problems in the data interpretation are
caused by noises in the signals due to the following
causes: (i) break-through effect. This effect is com-
monly visible on radar records and it hides partially Figure 21. EM travel time tomography of a pier. The high
the wall characteristics. It is caused by the fact that EM velocities could detect a detachment of the structure
(Valle et al., 1998).
the antenna itself reacts to the electromagnetic wave.
(ii) multiple echoes due to the presence of layers and
large quantity of data. Standard pulse velocity data
joints. (iii) superposition of the lateral echoes that cre-
or radar data could be used to reconstruct a velocity
ate images of parallel reflectors. This effect could be
distribution within a solid material, thus providing an
enhanced for a wall by the lack of mortar joints and
“image” of the masonry interior. The testing technique
the regularity of the stones.
gives a map of the velocity distribution on a plane sec-
Radar tests need always a preliminary calibration
tion of the structure under investigation. The result of
in order to verify if the emitted signal is enough pow-
the tomographic inversion is a map of a property of the
erful to detect the opposite side of the wall and the
materials. In case of travel time tomography (TT) the
wave speed. This step allows to calibrate the relation-
measured quantity is the traveltime of the signal and
ship between the time and space scales. Some types of
the map is the distribution of the propagation veloc-
equipment give directly this transformation by setting
ity within the object. In case of amplitude tomography
up a value of the dielectric constant of the masonry.
(AT) the measured quantity is the amplitude of the
The value is an average of the characteristics of all the
signal and the map is related to the distribution of the
materials crossed by the wave.
absorption coefficient.
The choice of the antenna frequency must be made
on a site basis. During the test it is important to control
the radar potentialities in relation to the frequency used
6 STRUCTURE CONTROL BY STATIC AND
(Padaratz & Forde, 1995).
DYNAMIC MONITORING
One of the limits of the technique is the low read-
ability of the results. In fact usually radar data are
Where an important crack pattern is detected and its
clearly readable only by experts. It is important to
progressive growth is suspected due to soil settle-
show results, as radargrams and graphics, which are
ments, temperature variations or to excessive loads,
significant to operators like architects and engineers.
the measure of displacements in the structure as func-
tion of time has to be collected. Monitoring systems
can be installed on the structure in order to follow this
5.4 Radar and sonic tomography
evolution.
Among the ND applications the tomographic tech- This type of survey is frequently applied to impor-
nique is quite attractive for the high resolution that tant constructions, like bell towers or cathedrals (e.g.
can be obtained (Shuller et al., 1995), (Valle et al., to the Pisa leaning Tower, to the Dome of the Florence
1997), (Valle et al., 1998). Tomography, developed in Cathedral in Italy (Bartoli et al., 1992)) (Fig. 22) and
medicine and in several other fields, seems to be a the system may stay in place for years before a deci-
valuable tool to give two or and three-dimensional rep- sion can be taken for repair or strengthening (Binda
resentation of the physical characteristics of a solid. et al., 1995), (Binda et al., 2000a).
Tomography, from Greek “tomos” (slice), reproduces Very simple monitoring systems can be also applied
the internal structure of an object from measurements to some of the most important cracks in masonry walls,
collected on its external surface (Fig. 21). were the opening of the cracks along the time can be
Tomographic imaging is a computational technique, measured by removable extensometers with high res-
which utilises an iterative method for processing a olution (Fig. 23). This simple system can give very

37
chap-03 9/9/2004 8: 5 page 38

ƒFEM = 0.585 Hz ƒFEM = 0.709 Hz ƒFEM = 2.455 Hz ƒFEM = 2.726 Hz ƒFEM = 5.698 Hz
ƒEMA = 0.586 Hz ƒEMA = 0.708 Hz ƒEMA = 2.456 Hz ƒEMA = 2.731 Hz ƒEMA = 5.706 Hz

Figure 24. Vibration modes of the FEM model of the Monza

Figure 22. S. Vitale Church at Ravenna. Total variation of Tower and comparison with the frequencies measured during
the out of plumb (November 1998–November 1999) (Binda the on site tests (Gentile et al., 2002).
et al., 2000a).

The environmental excitation sources could be the

500 wind, the traffic or the bell ringing in the particular case
Crack increasing [micron]

78–97 tendency of towers (Binda et al., 1997), (Gentile et al., 2002),

250 89–97 tendency (Binda et al., 2000). The forced vibrations could be
produced by local hammering systems or by the use of
0 vibrodines. An accelerometer net is installed in chosen
significant parts of the structure.
-250 The dynamic tests allow to detect the frequencies,
the modal shapes and the correspondent modal damp-
-500
ing of a structure. These parameters are characteristics
78 80 82 84 86 88 90 92 94 96 98 of the local and global behaviour of a structure.
Years

Figure 23. Crack monitoring made with a removable exten-

someter at the Monza Tower (28 Binda et al., 1998b). 7 DIFFERENT LEVELS OF INVESTIGATION

As it was mentioned in the section 3, investigation

important information to the designer on the evolution procedures can be applied at different level following
of the damage (Binda et al., 1998b). specific methodologies. Information on the building
In-situ testing using dynamic methods can be con- state of damage and on the material and structure char-
sidered a reliable non-destructive procedure to verify acteristics are always needed, but if the processes are
the structural behaviour and integrity of a building.The carried out at the level of the historic centre, only rep-
principal objective of the dynamic tests is to control resentative buildings of the most diffused typologies
the behaviour of the structure to vibration. The first will be investigated and on site and laboratory tests will
test carried out can be seen also as the starting one of be carried out on samples of the most representative
a periodical survey using vibration monitoring inside masonries, roofs and floors (Binda et al., 2004).
a global preventive maintenance programme. Accep- In case of important buildings and monuments
tance of vibration monitoring as an effective technique when also a design for preservation is required, then
of diagnosis has been supported by different studies the investigation must go into details intervention,
(Niederwanger, 1997), (Rossi, 1997). the knowledge must be more detailed and also the
These tests are very important to detect eventual investigation more sophisticated. In the following a
anomalies in the diagnosis phase and to calibrate effi- description of the investigation procedures is given.
cient analytic models (FEM) (Fig. 24) (Gentile et al.,
2002). In this way it is possible to verify the effective-
ness of the computational methods used in the analysis
7.1 Investigation at urban level
and control of the structure. The availability of an
efficient numerical model allows for checking and pre- As a first step the evolution of the centre during the
dicting the structure behaviour to dynamic actions like, different events or evolutions (e.g. earthquakes) is
for example, strong winds effects and seismic actions. investigated (Fig. 25a,b).

38
chap-03 9/9/2004 8: 5 page 39

Figure 27. Complex building.

(a)

Figure 28. Example of constructive evolution phases of

complex buildings.

(b) and bricks and stones), simple on site tests (flat jack,
sonic tests, drilling of cores, etc.).
Figure 25. Comparison between the state of Campi In the case of complex buildings their evolution in
(Umbria) in 1835 (a), and in 1997 (b). time must be detected on site or through archive doc-
uments in order to study their vulnerability (Fig. 28).
Prevision of possible collapse mechanisms, analyt-
ical models for prevision of vulnerability of the centre
and of the buildings.
Even if low cost and simple some on site NDT
or minor DT can also be powerful at the level of
urban investigation, in order e.g. to classify differ-
ent masonry types and hence different building types
(Fig. 29).

7.2 Investigation at building level

Figure 26. Row building typology.

When the design for restoration or preservation of his-
toric buildings has to be prepared, then a deep knowl-
edge of the building is needed from its history and
Among the remaining buildings the most represen- evolution in time to the geometry and typology of con-
tative from the point of view of construction typology, struction, to the structural elements (walls, columns,
structure and materials are chosen and on them the piers, floors, roof, etc.) to the possible decay and struc-
investigation procedure is applied. Geometrical, indi- tural damages and their causes, to eventual monitoring
viduation of the building typology (isolated, in a systems to the effectiveness of the intervention. Then
row, complex) (Figs. 26, 27), crack pattern survey, a design for investigation must be prepared by the
comparison of the crack pattern to the abacus of future designer himself; he needs to at first inspect
the collapse mechanisms, laboratory tests (chemical, visually the building, understand the weakest points,
physical, mechanical tests are carried out on mortars formulate hypotheses on the damages and the causes,

39
chap-03 9/9/2004 8: 5 page 40

2.0 2.0

LVDT 5 horizontal LVDT 5 horizontal

avarage LVDT vert. 1,2,3,4 mean LVDT vert. 1,2,3,4

1.5 1.5

Stress [N/mm ]
Stress [N/mm ]

2
2

1.0 1.0

Local state of stress

2
0.72 [N /mm ]
0.5 0.5

Local State of stress

εv 2
l l 0.15 [N/mm ] ε v
0.0 0.0
-0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Strain [µm/mm] Strain [µmm] a)

b)

Figure 29. (a) Some results obtained with single and dou-
ble flat-jack tests on the external wall of a church, of a bell
tower and of a civil building; (b) representative results of the
diagonal surface sonic measurements.

Figure 31. Tower and Cathedral of Cremona (Italy).

definition geometry of
of the the structural traditional methods for
structural system elements geometrical survey
(also hidden)
geodetic methods
hypothesis
on the before and photogrammetry
restrains after repair
thermography,
pulse velocity test,
crack pattern survey, before, electro-magnetic
control of methods, Xgraphy, etc.
geometry evolutions after,
during extensometers
(displacements,
rotations, the repair
automatic control system
crack evolution)
intervention

coring,
survey of the masonry section
extraction by
and
local demolition
construction details

in situ measurement
of the local single flat-jack
state of stress

individuation of the double flat-jack,

mechanical (E, , fk, etc.)
and chemico-physical
pressiometer
Figure 32. S. Vitale Church at Ravenna (Italy).
(mortar composition, porosity
moisture content, etc.) in-situ sampling,
characteristics of materials
choose appropriate investigation procedures and also
laboratory tests

apply them with an appropriate methodology.

indirect individuation of
mechanical characteristic of
The flow chart of Fig. 30 schematically presents
masonry a methodology to approach and solve the problem of
structural investigation of an historic building, respect-
Figure 30. Information required and correspondent inves- ful of its integrity and applying the right investigation
tigation techniques (Binda et al., 1994). procedure to the different emerging problems.
On the left hand side the different unknowns are
listed from the geometry of the structure to the crack
then to establish which investigation procedures bet- pattern possible evolution, to the deep knowledge of
ter suit the building problems. This means that the material property and structural details.
architect or engineer who has to prepare the design This methodology was applied by the authors to
for intervention, in order to choose appropriate tech- various case histories with good success (Figs. 31, 32)
niques for repair, protection and strengthening must (Binda et al., 2000; Binda et al., 2001).

40
chap-03 9/9/2004 8: 5 page 41

It must be clear that even if there is a need of consult-

ing experts in the field, it is the designer, or a member
of the design team, who must be responsible of the
diagnosis and must: (i) set up the in-situ and labora-
tory survey project, (ii) constantly follow the survey,
(iii) understand and verify the results, (iv) make tech-
nically acceptable use of the results including their use
as input data for structural analyses, (v) choose appro-
priate models for the structural analysis, (vi) arrive at
a diagnosis at the end of the study.
These operations can be accomplished with the help
of experts in the field. Therefore information is needed
for architects and engineers on the availability and
reliability of the investigation techniques.
Figure 33. Situation of a rural building in Milan in 1980 (a)
and 1998 (b).
REFERENCES

Obviously the designer has to remember that a sin- Abbaneo, S., Berra, M., Binda, L., 1996. “Pulse velocity test
gle technique is frequently not enough to have the to qualify existing masonry walls: usefullness of wave-
knowledge of defects, inclusion, voids in the structure form analyses”, 3rd Conf. Non Destructive Evaluation
but frequently techniques have to be used as comple- of Civil Structures and Materials, Boulder - CO, USA,
mentary; this of course implies an appropriate budget 81–95.
Aerojet General Corporation, 1967. Investigation on sonic
reserved for investigation; therefore all the operations
testing of masonry walls, Final Report to the Dept. of
must be carefully chosen. General Services of Architecture and Construction, State
of California.
ASTM, 1991. Standard test method for in situ compres-
7.3 Investigation for maintenance
sive stress within solid unit masonry estimated using the
Lack of maintenance for a long time can be as destruc- flat-jack method, ASTM C 1196-91, Philadelphia, ASTM.
tive as exceptional events (earthquakes, floods) in a Bartoli, G., Blasi, C., Gusella,V., 1992. “Il sistema di monitor-
long range. Fig. 33 shows the effects of the envi- aggio della cupola del Brunelleschi: analisi dei dati rilevati
(1988/1990)”, IV National Congress ASSI.R.C.CO., Prato,
ronment on a historic building abandoned for twenty
Italy, 445–455.
years. It has also to be remembered that most of the col- Binda, L., Anzani, A., Mirabella Roberti, G., 1997. “The
lapses occurring during earthquakes concern buildings failure of ancient Towers: problems for their safety assess-
which had no maintenance for a long time. ment”, Int. IABSE Conf. on “Composite Construction –
Therefore once the buildings has been repaired and Conventional and Innovative”, Zurich, 699–704.
starts functioning a project for maintenance must be Binda, L., Baronio, G., Gambarotta, L., Lagomarsino, S.,
prepared stating minimum investigation and moni- Modena, C., 1999. “Masonry constructions in seismic
toring operations in due times in order to fulfil the areas of central Italy: a multi-level approach to conser-
predicted service life. vation”, 8NAMC 8th North American Masonry Conf.,
Austin, USA, 44–55.
The same investigation procedures mentioned in the
Binda, L., Cardani, G., Penazzi, D., Saisi, A., 2003.
previous sections can be applied. They should be NDT “Performance of some repair and strengthening tech-
or MDT possibly easy, repeatable and at low cost. niques applied to historical stone masonries is seismic
areas”, ICPCM a New Era of Building, Cairo, Egypt,
18–20/2/2003, 2, 1195–1204.
Binda, L., Cardani, G., Saisi, A., Modena, C., Valluzzi, M.R.,
8 CONCLUSIONS 2004. “Multilevel Approach to the Analysis of the Histori-
cal Buildings: Application to Four Centers in Seismic Area
An attempt has been made in the previous sections to Finalised to the Evaluation of the Repair and Strengthen-
point out the real necessity of a deep knowledge of the ing Techniques”, 13th International Brick/Block Masonry
building before any intervention for its preservation Conference, RAI Amsterdam, July 4–7, CD-ROM.
is carried out. This knowledge in the case of historic Binda, L., Falco, M., Poggi, C., Zasso, A., Mirabella
Roberti, G., Corradi, R., Tongini Folli, R., 2000. “Static
buildings has been lost in the past century due to the
and Dynamic Studies on the Torrazzo in Cremona (Italy):
use of new materials in the building industry. The only the Highest Masonry Bell Tower in Europe”, Int. Sympo-
way architects and engineers have to follow in order to sium Bridging Large Spans: from Antiquity to the Present,
recover the lost information is to carry out an appro- Istanbul, Turkey, 100–110.
priate investigation, thanks also to the recent develop- Binda, L., Gatti, G., Mangano, G., Poggi, C., Sacchi
ments of sophisticated investigation procedures. Landriani, G., 1992. “The collapse of the Civic Tower of

41
chap-03 9/9/2004 8: 5 page 42

Pavia: a survey of the materials and structure”, Masonry Gentile, C., Saisi, A., Binda, L., 2002. “Dynamic investi-
International, 6(1), 11–20. gation of a historic masonry Bell Tower”, 6th Interna-
Binda, L., Lenzi, G., Saisi, A., 1998. NDE of masonry struc- tional Masonry Conference, London, 4th–6th November,
tures: use of radar test for the characterisation of stone 192–199.
masonry, Non Destructive Testing and Evaluation Int., Lenzi, G., Ludwig, N., Rosina, E., Saisi, A., Binda, L., 1997.
31(6), 411–419. Analisi di tessiture murarie mediante termografia e geo-
Binda, L., Lenzi, G., Saisi, A., 1999a. “Radar Investiga- radar, V Congresso Nazionale ASS.I.R.C.CO, Orvieto –
tion for Diagnosis of Historic Buildings: Application to Italy, 142–146.
Stone Masonries”, 8th Int. Conf. on Structural Faults and Niederwanger, G., 1997. “Structural Repair of Damaged Old
Repair, London, CD-ROM. Bell Towers Based on Dynamic Measurements”, 5th Int
Binda, L., Lombardini, N., Guzzetti, F., 1995. “St. Vitale in Conf. on Structural Studies, Repairs and Maintenance of
Ravenna: a survey on materials and structures”, Int. Conf. Historical Buildings STREMAH97, S. Sebastian, Spain,
on Historical Buildings and Ensembles, invited lecture, 447–456.
Karlsruhe, Germany, 113–124. Padaratz, I.J, Forde, M.C., 1995. “Influence of Antenna
Binda, L., Lualdi, M., Saisi, A., Zanzi, L., Gianinetto, M., Frequency on Impulse Radar Surveys of Concrete Struc-
Roche, G., 2003. “NDT applied to the diagnosis of his- tures”, 6th Int. Conf. on Structural Faults and Repair,
toric buildings: a case history”, 3-Day Int. Conf. Structural London, 2, 331–336.
Faults & Repair, London, 1-3/7/2003, CD-ROM. Penazzi, D., Valluzzi, M.R., Cardani, G., Binda, L.,
Binda, L., Mirabella, G., Abbaneo, S., 1994. “The Diagnosis Baronio, G., Modena, C., 2000. “Behaviour of His-
Research Project”, Earthquake Spectra, 10(1), 51–170. toric Masonry Buildings in Seisimic Areas: Lessons
Binda, L., Mirabella Roberti, G., Guzzetti, F., 2000a. Learned from the Umbria-Marche Earthquake”, 12th Int.
“St. Vitale in Ravenna: A Survey on Materials and Brick/Block Masonry Conf., Madrid, Spain, 217–235.
Structures”, Int. Symposium Bridging Large Spans: from Rossi, P.P., 1982. Analysis of mechanical characteristics of
Antiquity to the Present, Istanbul, Turkey, 89–99. brick masonry tested by means of in situ tests, 6th IBMaC,
Binda, L., Saisi, A., Tiraboschi, C., 2001. “Application of Rome, Italy.
Sonic Tests to the Diagnosis of Damage and Repaired Rossi, P.P., 1997. “Possibilities of the Experimental Tech-
Structures”, Non Destructive Testing and Evaluation Int., niques for the Structural Analysis of Historical Construc-
34(2), 123–138. tions”, CIMNE, Barcelona, Spain, 24–46.
Binda, L., Saisi, A., Tiraboschi, C., 2000b. “Investigation Schuller, M.P., Berra, M., Atkinson, R., Binda, L., 1995.
procedures for the diagnosis of historic masonries”, Con- “Acoustic Tomography for Evaluation of Unreinforced
struction and Building Materials, 14(4), 199–233. Masonry”, 6th Int. Conf. on Structural Faults and Repair,
Binda, L., Tiraboschi, C., 1999. “Flat-Jack Test as a Slightly London, 3, 195–200.
Destructive Technique for the Diagnosis of Brick and Valle, L., Zanzi, L., Binda, L., Saisi, A., Lenzi, G., 1998.
Stone Masonry Structures”, Int. Journal for Restoration “Tomography for NDT applied to masonry structures:
of Buildings and Monuments, Int. Zeitschrift fur Bauin- sonic and/or EM methods”, 2nd Int. Arch Bridge Conf.,
standsetzen und Baudenkmalpflege, Zurich, 449–472. Venezia, 243–254.
Binda, L., Tiraboschi, C., Tongini Folli, R., 1998b. “On site Valle, S., Zanzi, L., Lenzi, G., Bettolo, G., 1997. “Struc-
and laboratory investigation on materials and structure of ture Inspection with Radar Tomography”, Int. Coll. on
a Bell-Tower in Monza”, 2nd Int. RILEM Conf. on Reha- Inspection and Monitoring of the Cultural Heritage,
bilitation of Structures, Melbourne, Australia, 542–556. IABSE-ISMES, 223–230.

42
chap-04 9/9/2004 8: 5 page 43

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

Evaluation and analysis of the old timber structures

A. Ceccotti
Trees and Timber Institute, IVALSA-CNR, Florence, Italy

ABSTRACT: A methodology for evaluating the actual load-bearing performance of an ancient timber construc-
tion is given in this paper based on a simple non-destructive approach. Materials and methods are illustrated and
discussed. Examples are given emphasizing the key points of the decision making procedure. Final considerations
with particular reference to the “minimum intervention” principle conclude the work.

1 INTRODUCTION Table 1. Evaluation and analysis process (schematic).

Before an ancient timber structure the structural Evaluation (Wood Technologist)

designer is always asked for the fundamental question: Decay detection – Residual cross section → b, h
what to do with this structure? That actually means: Strength grading, in situ → timber grade
Anticipated strength → fm,k(5%)
how sound are wood elements and joints? Is this struc- Uncertainties → γM , material side safety coefficient
ture capable of standing up for many years more, even
under new service loads? Are any strengthening and Analysis (Structural Engineer)
Actions (loads, qk )
repair needed? Structural scheme for calculations: qk → Mk
Having in mind the conservation of the structure Stresses on elements: Mk → σm,k
as main-guidance line, in order to take the most Uncertainties → γf , action side safety coefficient
appropriate decision, the designer should follow a
multi-disciplinary approach.
acting load values (nominally intended for passing
from 5th percentile to 0,5th percentile, lower tail for
2 A METHODOLOGY FOR EVALUATION strength distribution and higher tail for load distribu-
AND ANALYSIS tion, respectively). kmod is a modification factor taking
into account service conditions and duration of load
There are two basic and separate questions actually: effects (see Table 3).
• is the wood still sound?
• is the wood enough resistant? 2.1 Evaluation phase
In fact resistance is a matter not only of wood Evaluation phase consists of:
soundness but also it is a matter of wood actual stress
compared to wood actual strength. Evaluation of wood • evaluation of possible biological decay across the
conditions and strength and analysis of forces and member section and along the member itself.
stresses are two moments of the same process. In This allows to determine the residual cross-section
Table 1 a synopsis of these points is given referring dimensions along the entire length of the mem-
to a simple case of a single wooden rectangular cross ber to be used by the Engineer in his calculations
section (b original width, h original depth) under a (Bonamini, 1995).
only bending moment M : • in-situ grading, i.e. evaluation of the strength of
sound timber according to grading rules accepted
for that kind of timber.

2.1.1 Important remarks

2.1.1.1 Grading
γM and γf safety coefficients (>1) cover further For a certain timber population (wood species, loca-
uncertainties about possible actual strength values and tion of origin, grade) it will be possible to assign a set

43
chap-04 9/9/2004 8: 5 page 44

of strength values for different kinds of stress (strength

profile). In fact everywhere in the world there are
grading rules that allow to grade timbers according
to their strength (Figure 1). Every country has its own
rules calibrated upon home grown timbers peculiari-
ties. Therefore there are tables available that attribute
each timber population to a certain strength profile
(strength class), see Table 2.
2.1.1.2 Strength
It is here necessary to remind the Reader that old tim-
ber characteristic strength values (e.g.: fm,k(5%) ) are
Figure 1a. used as high as for “new” timber. It is actually acknowl-
edged that timber is not loosing its strength over time
just because is becoming older (excluding of course
any decay due to insects and fungi). In fact there is no
real evidence that long lasting pre-loading of timber or
timber structures to a limited load level has produced
any damage (Kuipers, 1986), see Table 4.

2.2 Analysis phase

Analysis phase consists of:
• Internal forces analysis conducted by using the sim-
plest structural scheme, at first. Then more and more
accurate analysis which takes into account as much
as possible of the actual restraints and of actual
Figure 1b. mechanical behaviour of materials. That include,
for example, semi-rigid behaviour of joints (slip),
and possible structural gross deformations and dis-
orders. More and more sophisticated models (i.e.
anisotropy of wood, II order analysis) should be
adopted only when the complexity of the structure
requires it. The more the scheme is reality-bound
(e.g. considering hyperstatic behaviour, load shar-
ing, et cetera) the more force peaks decrease so
that verifications are facilitated, i.e. a better model,
though more complex and time demanding, at the
Figure 1c.
end of the entire process will be rewarding (see
Figures 3–4).
Figure 1a–c. Strength grading. Grading allows to separate • Stress analysis, it is possible to pass from internal
better timbers from less resistant ones. With no grading, for forces to stresses using well consolidated calcula-
the same timber population, strength could range from 15 tion methods, as for example Eurocode 5, 1993
to 95 Mpa (top, left) like in the case of Swiss pine beams. (usually the limit states design methods are more
By grading pieces into strength-quality groups (a,b,c grades, generous than allowable stresses methods). Then a
for example) it is possible to classify timbers (top, right) safety check becomes possible referring to strength
according to their resistance. Grading rules are different over values given by standards, as said before.
the world because locally calibrated on locally grown tim-
bers. Grade determining defects are usually knots, slope of
grain, annual rings thickness et cetera. For each grade, after
an extensive testing campaign on that timber population, the 3 DECISION MAKING
relevant characteristic strength values (5th percentile) can be
found (below). Grading rules are not 100% efficient because
they do not allow to put all the best pieces in the upper class,
If the verification, by applying the previously illus-
for example, and the worst pieces in the lowest class nei- trated method, is positive, this is sufficient to state
ther (below), but they are nevertheless essential because they that the structure is safe enough. However the oppo-
separate timbers according to their characteristic resistance site is not true. In fact it must be said that when
anyway. Please note that in the same grade we have 95% of using the above approach based on modern calculation
pieces more resistant than the 5% strength. codes and characteristic strength profiles and standard

44
chap-04 9/9/2004 8: 5 page 45

Table 2a–2b. Strength classes and the “Magic” table, assignment of visual grades and species
according to CEN/TC 124.215. Table gives coniferous timbers and source combinations
matching strength class C24 profile.

C14 C16 C18 C22 C24 C27 C30 C35 C40

in N/mm2
fm,k 14 16 18 22 24 27 30 35 40
fc,0,k 8 10 11 13 14 16 18 21 24
fc,90,k 0,3 0,3 0,3 0,3 0,4 0,4 0,4 0,4 0,4
fc,0,k 16 17 18 20 21 22 23 25 26
fc,90,k 4,3 4,6 4,8 5,1 5,3 5,6 5,7 6,0 6,3
fc,k 1,7 1,8 2,0 2,4 2,5 2,8 3,0 3,4 3,8
in kN/mm2
E0,mean 7 8 9 10 11 12 12 13 14
E0,05 4,7 5,4 6,0 6,7 7,4 8,0 8,0 8,7 9,4
E90,mean 0,23 0,27 0,30 0,33 0,37 0,40 0,40 0,43 0,47
Gmean 0,44 0,50 0,56 0,63 0,69 0,75 0,75 0,81 0,88
in kg/m3
ρk 290 310 320 340 350 370 380 400 420

Strength classes and characteristic values according to EN 338. Coniferous species and poplar.

Table 2b.

Grading rule
Strength publishing country Species
class (Grading standard) Grade commercial name Source

C24 Austria G.BH Spruce, Pine, Fir, Larch CNE Europe

(ÖNORM B 4100-2)
France CF22 Whitewood, Douglas fir France
(NFB 52001-4)
Germany S10 Spruce, Pine, Fir, Larch CNE Europe
(DIN 4074-1)
Nordic Countries T2 Redwood, Whitewood NNE Europe
(INSTA 142)
The Netherlands B Spruce + fir NC Europe
(NEN 5466)
UK SS Redwood, Whitewood CNE Europe
(BS 4978) SS Douglas fir, Larch, USA + Canada
Hem-fir, S-P-F
SS Southern pine USA
SS Parana pine Brazil
SS Pitch pine Caribbean
USA + Canada J+P Douglas fir, Larch, USA + Canada
(NGRDL + NLGA) Sel Hem-fir, S-P-F

CNE Europe: Central, North & Eastern Europe, NNE Europe: Northern & North eastern
Europe, NC Europe: Northern and Central Europe.

grading rules, when the verification would fail at first uncertainties about the material still to be provided
approach, an appeal should be given to the structure. are much less because the material is already there
We do not want to touch here the issue of possi- and we can see it in place. Moreover, load standards
ble reduced safety coefficients due to the fact that the are changed over the years increasing snow loads, for
structure is already existing (modern calculation codes example, where our structure does exist from centuries
are thought for not-yet existing structures). Therefore with no interest in human discussions.

45
chap-04 9/9/2004 8: 5 page 46

We want just to say that more sophisticated struc-

tural schemes, closer to reality, should be used in this
case. In addition, grading of timber elements should be
done ad hoc on the spot where the safety verification
is actually performed (so called critical sections).
Actually a timber element is classified into strength
grades according to some defects, basically. It is not
possible to know a priori where in situ that element will
be located, how it will be loaded, and where the most
stressed section will stay (critical section). Therefore
timber elements are classified by the timber supplier
independently from the location of the grade deter-
mining defect along the timber element itself. Let’s
consider the case, for example, of an isolated big knot
as the grade determining defect. Once the timber ele-
Figure 2a–c. ment has been put in place, if this knot is just near
bearing supports or at the extra-dos of a bent beam,
Table 3. Load–duration classes and kmod for solid timber it will have much less importance than if it would be
and glulam.
staying at the mid-span intra-dos of a bent beam. Old
kmod for
carpenters were actually used to put the best pieces in
Service classes the most stressed parts of the structure! (see Figure 5).
Load–duration Examples In conclusion: if first safety check is not positive,
class Durationa of loading 1&2 3 before to make a life sentence (demolition or strength-
ening) a second chance should be given to the element,
Permanent More than Self 0,60 0,50 making an ad hoc re-classification of the timber ele-
10 years weight ment right around the critical section aiming for an
Long-term 6 months– Storage 0,70 0,55 up-grading of the element.
10 years More: we have to say that even the ad hoc grad-
Medium-term 1 week– Imposed 0,80 0,65
ing, refers to 5% lower fractile strength characteristic
6 months load
Short-term Less than Snowb 0,90 0,70
values, that means there is still a high chance that our
one week and wind element will be more resistant than that value.
Instantaneous Accidental 1,10 0,90 Advanced research in wood mechanics field, for
load example using Non Destructive Testing methods cou-
pled with analytical tools, could help to guess an actual
a
The Load-duration classes are characterised by the effect of strength value (see Table 4 and Figure 6).
a constant load acting for a certain period of time. For vari-
able action the appropriate class depends on the effect of the
typical variation of the load in the life of the structure. The
accumulated duration of the charateristic load is often very 4 CONCLUSIONS
short compared with the total loading time.
b
In areas with a heavy snow load for a prolonged period of Wooden cultural heritage’s most dangerous enemy
time, part of the load should be regarded as medium-term. today is insensibility, and lack of maintenance, of
course. However there is another risk equally fright-
Figure 2 and Table 3. Duration of load effect. Duration of
ening that appears just when conservation works have
load effect is illustrated here according to EC5, CIB W18 code
and Madison curve (Giordano, 1999) where time is in a loga- been launched.
rithmic scale (x axis). Stress ratio as reported on y axis, is the This sneaking enemy is the “Do something, any-
ratio between the actual strength and the 5-minute-duration way” philosophy. “This wood is too old, it has lost its
test strength. It is easy to see that under 50% of short term strength!” or “this element is going to fall down on
load resistance (so-called creep limit) the time-to-failure is us, it does not satisfy the last code on loads!” and so
almost infinite. That means that there is no damage at all on, are typical examples, but many others could be
inside the wood. Remember that when loaded under quasi- given, that may lead to unnecessary reinforcement to
permanent load combinations the stress ratio is 15–20% of the detriment of cultural authenticity.
5 minute strength. Possibility of internal damage due to long Wood technologists and structural engineers have
lasting action of loads that may overcome the creep limit for
a certain duration of time is anyway considered by codes with
the privilege of mastering evaluation tools and analyt-
the modification factor kmod according to the time of accu- ical models that can help professionals in making the
mulated duration of load at maximum level (characteristic, best decision in various circumstances, so that a new
5% fractile). Service conditions are: class 1, indoor; class 2, philosophy, with more respect and more knowledge
indoor or outdoor protected; class 3, outdoor not protected. will take over.

46
chap-04 9/9/2004 8: 5 page 47

Figure 3a. Figure 3d.

Figure 3b.

Figure 3e.
Figure 3c.

Figure 3a–3e. Simple models are very important in determining rapidly, even in a rough way the load distribution within
the structure. This approach is usually conservative, because does not take into account possible re-distribution effects.
More advanced models allow a finer evaluation of load distribution and give a better load-path with a reduction of peak
forces.

47
chap-04 9/9/2004 8: 5 page 48

Figure 4c.

Figure 4a.

Figure 4d.

Figure 4b.

Figure 4e.

Figure 4a–4e. In some cases when the mechanical behaviour is very complicated to determine, tests should be performed,
and/or highly sophisticated models used. This does facilitate verification and allows the adoption of very essential strengthening
solutions. Tables 5a and 5b are a synopsis of the Master thesis of Ms Angela Bevilacqua on “Strengthening of 16th Century
wood-masonry building in Wismar”, 1998, University of Florence, Italy.

48
chap-04 9/9/2004 8: 5 page 49

Figure 5a. Figure 5b.

Figure 5c. Figure 5d.

Figure 5a–d. Panchia (Trento, Italy) bridge. Accurate modelling coupled with in situ load tests allowed to identify the best
way to preserve the cultural authenticity of the bridge without compromising users safety. To reduce car traffic induced bridge
vibration the deck planks were simple inclined at 45◦ respect to traffic direction.

Figure 6a. Figure 6b.

49
chap-04 9/9/2004 8: 5 page 50

Table 4. REFERENCES
Density E0 Fm Failure determining Kuipers, J. 1986. Effect ofAge/or Load onTimber Strength. In
Beam (kg/m3 ) (N/mm2 ) (N/mm2 ) defect(s) Proceedings of CIB W18 meeting, Paper 19-6-1. Florence,
Italy.
T1 480 8689 25.0 Ring shake, checks Ceccotti, A & Uzielli, L. 1989. Reliability of Ancient Timber
T2 483 8273 28.0 Ring shake, checks Structures. In G. Tampone (ed) Proceedings of II National
T3 516 18266 39.6 Checks Italian Congress on Wood Restauration. Firenze: Nardini
T4 464 13048 25.8 Ring shake, checks (in Italian).
T5 487 12906 44.8 Slope of grain, knot Uzielli, L. 1995. Restoring timber structures – Repair and
T6 505 10355 33.9 Localised decay strengthening. STEP 2 Timber Engineering, lecture D4.
T7 464 8102 30.9 Knots, diffuse decay Centrum Hout, The Netherlands.
T8 513 14013 47.0 Checks Bonamini, G. 1995. Restoring timber structures – Inspection
T9 498 11747 38.5 Knots and evaluation. STEP 2 Timber Engineering, lecture D3.
T10 478 12243 29.4 Checks Centrum Hout, The Netherlands.
T11 469 7104 15.3 Diffuse decay Giordano, G. & Ceccotti, A. & Uzielli, L. 1999. Timber
T12 449 11630 30.1 Checks Engineering. Milano: Hoepli (in Italian).

Table 4 and Figure 6a–6b. Large cross sections beams show

an interesting semi-ductile behaviour because knots are col-
lapsing one after the other allowing the beam to recover
partially (if load sharing is permitted). In a series of results
from Dr Marco Togni is shown that only 1 over 12 beams, dis-
carded from the building site because considered not reliable,
gave a strength low as 15 Mpa. Marco Togni (1995); “Elas-
ticity and strength of large cross-section old timber beams:
mechanical evaluation with NDT in situ”, Doctoral thesis,
University of Florence, Italy (in Italian).

50
chap-05 9/9/2004 8: 6 page 51

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

Remedial measures for the Cathedral of Porto: a post-modern

conservation approach

P.B. Lourenço
University of Minho, Department of Civil Engineering, Guimarães, Portugal

Â. Melo & M. Carneiro

Portuguese Institute for the Architectural Heritage, Ministry of Culture, Porto, Portugal

ABSTRACT: The present paper presents the works recently carried out at the Cathedral of Porto as a case
study of a difficult intervention that challenges current recommendations for the architectural heritage. The
historical information is briefly reviewed and the general conservation approach for the different works is
addressed. Afterwards, the aspects regarding the strengthening of the towers are treated with more detailed and
the diagnostics of a chapel, incorporating the usage of limit analysis for safety evaluation, is presented.

1 INTRODUCTION bishop celebrates, and where an elite of the clergies

meet and advise their bishop in the administration of
The cathedral is not a beautiful building, it does not material and immaterial assets of the diocese.
surrender to love at first sight. In fact, stronger feelings For 800 years, the settlement was a repository of
and emotions can only result from a comprehension of added parts. From this state of a continuous con-
the building’s nature. struction yard, the main fabrics are: romanesque and
The works being carried out also fail to pro- proto-gothic, gothic, renascence, mannerist, baroque,
vide immediate adhesion. In the framework of a neoclassic, contemporary works from the first half of
“restoration in continuity”, the works resemble a the 20th century and, finally, the present intervention.
“no-intervention” or a sort of zero level of architecture. Although the works until the end of the 19th cen-
Nevertheless, the authors believe that such a superfi- tury can be identified by the style and the building
cial impression is erroneous. The intervention is rooted techniques, all of them possess a severe common
in options clearly belonging to the disciplinary field of transforming attitude, that is, they add, superimpose,
architecture, with an approach that evolved as a result rebuilt, change or render. While the campaign of the
of the knowledge of the building and not an a priori 20th century, which lasted from 1918 to 1938 and was
posture. The vanishing appearance of the intervention carried out after expropriation and classification of
is not equal to a facing approach, because several lev- the settlement as a national monument, did not aim at
els of remedial measures co-exist, and some of them transforming the building, but at restoring it, as a sort
are profound. of time regression, performed with works aiming at
bringing back the original style – the medieval.
During this period, the country kept itself at a
2 THE INHERITED BUILDING – FROM THE distance from the industrial progress of the rest
EVOLUTIONAL TO THE RETURN TO ITS of Europe and, between modernity and tradition,
ORIGIN Portugal chose the latter. The Medieval Age and
Art, celebrated in the official culture as ethical and
The origin of the building is the middle of the 12th aesthetical patterns, participated in the construction
century. In this period it is possible to witness the of the imaginary associated to the valuation of a
construction of cathedrals in the main cities, across national identity, being particularly exalted in the
Europe, as a token of a renewed confidence within the public building, built from scratch or restored when
urban communities. historical.
The settlement is conceived as an answer to a pro- This ideal identity incorporated values such as aus-
gram that can be referred to as the church where the terity, purity and trueness, which were then attributed

51
chap-05 9/9/2004 8: 6 page 52

Figure 2. Aspects of the restoration work in the surround-

Figure 1. Aspects of the restoration work in the compound, ings of the compound, carried out in the first half of the 20th
carried out in the first half of the 20th century: (a) main façade century: general view and plan (a) before and (b) after the
and (b) details of the nave and view from the cloister. works, and (c) detail of the narrow street and square in front
of the main façade.

to the “Portuguese race” and were present in the

Medieval Art, especially the Romanesque.
In the process of declaring an identity and pursuing
their values, styles and periods of the historical build-
ings were mutilated. In particular, the contributions
of the baroque and neoclassic were aimed at, because
their decorative heritage was not in agreement with
the “austerity” of the architectonical shapes and the
“trueness” of the building materials.
Figure 1 illustrates some of the aspects of the Cathe-
dral of Porto, before and after the restoration works of
the first half of the 20th century.

2.1 The surroundings as a frame

After restoration of the building itself, also the sur- Figure 3. Details of the baroque pillory in the center of the
roundings were subjected to modifications. The nar- square.
row and labyrinth urban fabric that grew side-by-side
with the cathedral was demolished, see Figure 2a. the monument, a syntax as inspiration to the way
In its place a square was set up, see Figure 2b, and the cities were planned in the Illuminist period. The
defined by straight alignments and monumental per- baroque was removed from the building but references
spectives, see also Figure 2c, featuring a kind of frame to the baroque period were recreated in the surround-
aiming at isolating and stressing the importance of ings, see the pillory in Figure 3. At the southwest limit

52
chap-05 9/9/2004 8: 6 page 53

of the square, a tower-house with medieval appear- to carry out the Works, which are only superficially
ance was built from scratch in 1940, with the purpose from an author and much more from the collection of
of hosting the Museum of the History of the City. professionals.
The scenic effect that resulted from the restoration Preliminary diagnosis was very brief and, therefore,
produced a kind of representation of the past and the the design project was not conceived in detail. On the
past in this way represented understated the representa- contrary, the project was a directive kept open and con-
tion. A kind of an alternative past or a commemorative tinuously adapting to the unforeseen, which is often
allegory. the case in historical buildings, where the anatomy
is processed by successive approximations and under
3 BASIS FOR CURRENT WORKS the lens of different disciplines. Of course, dissection
as a knowledge tool cannot be a part of a modern
3.1 The rehabilitation of the restoration as a intervention.
program In this case, the lack of adequate preliminary
diagnosis, which is in opposition with modern method-
The building, as it is found today, is hybrid and ologies, was compensated with an intense mul-
frozen in a perpetual past, where re-combinations tidisciplinary activity during the execution period
or architectural redesign seem hardly acceptable in (2002–2004), supported by research, consultancy and
the near future. It is with the signs of anomalies and expertises in various fields. In the contingency of
the absence of formal borders that the memory of the works that had started already, the reunion of efforts
building is found and that the violation of authentic- resulted in a process of effective cooperation, with
ity, which is lacking, is understood. This process of the advantage of permanent in situ approaches and
learning is also a process of increasing identification discussions.
and love between the authors and the building. Only
then, it is possible to understand that the fake self of
the cathedral, as a result of the previous restoration, is
today the real self of the cathedral, transformed into 4 SELECTED ASPECTS OF THE
a monument. This is a forged identity that cannot be INTERVENTION
considered a minor representation of the strength of a
specific cultural period. The restoration carried out in the first half of the
Therefore, the governing thread of the program 20th century used traditional construction techniques.
of the current intervention is to rehabilitate the Some of the structural deficiencies encountered were
restoration. The aim is to reactivate, rehabilitate and then solved with the dismantling and rebuilding of
upgrade the competence, where competence is under- unstable parts, and with the replacement of dete-
stood as the performance capacity of the structures, riorated or damaged granite, with poor mechanical
the materials, the shapes and also the space, assumed performance.
as a support for functionality. The sole concession to the industrial technology is
The intervention in the building was organized the use of Portland cement, used as a common binder
around five operations: removal of infestations, con- for repointing masonry joints, rendering walls and
solidation, water-tightness, ventilation and protection. several reparations that during and after the restora-
And, finally, also monitoring. tion works, aimed at solving the following issues,
From the functional point of view, the use was without success: waterproofing of surfaces, glue
strengthened, which can be summarized by more and reconstitute volumes, stabilize cracks and stop
“cathedral” and less “museum”. In fact, several rooms movements.
had lost a meaning, either due to degradation or It is precisely with respect to the above-cited issues
because they were changed into a kind of “museum” that deeper interventions have currently been carried
without a clear program, where loose artifacts of out, some without visible effects and other with the
sacred art were exposed as an ornamentation of open- addition of parts, as in the strengthening of the tow-
ings and empty spaces. Therefore, where possible, the ers. Therefore, the architects in charge of the works
aim was the rehabilitation of the function for which tend to joke about the fact that the only interven-
the spaces were created and the reintegration of the tion being carried out is the strengthening of the
artifacts in their context. This program did not reduce towers.
but even increased the part of the compound open to Next, see Figure 4 to Figure 12, some of the aspects
the visitors. of the works being carried out are briefly reviewed.
The works are mostly concentrated in the towers, and
the roofs and façades in the west and south wings.
3.2 Dialogue as a methodology
Diagnosis and strengthening of the towers and the
A constellation of professions has been mobilized Saint Vincent Chapel are addressed in a separate
and the dialogue has been applied as a methodology section.

53
chap-05 9/9/2004 8: 6 page 54

Figure 4. Remedial measures in the roof structures included

cleaning, application of biocide, application of preservation
products, consolidation, strengthening and local replace-
ment. Experts were responsible for diagnosis and structural
assessment.

Figure 6. Remedial measures for stone, including removal

of biological activity, dry and low pressure water clean-
ing, localised consolidation, application of water repellents,
reconstitution of voids, crack closure and injection, replace-
ment of iron ties, and, exceptionally, replacement of stone
pieces. Experts were responsible for diagnosis and specific
treatments.
Figure 5. Replacement of the ceramic tiles including new
anchors, traditional eaves, strengthening in the corners,
introduction of under-roof sheeting and walkways.

5 INTERVENTION IN THE TOWERS

5.1 Introduction
Figure 7. Repair of grilles included control of anchors and
A key aspect in the behaviour of ancient towers is that new anchors, and measures for proper rainwater flow.
the collapse process usually excludes the possibility of
ductile behaviour. In fact, there are hardly any possibil-
ities of internal force redistributions between different historical structures. But this is not often the case, see
critical sections, and failure of a single section is usu- Binda & Anzani (1997).
ally sufficient to provoke the entire collapse of the A tower is usually a result of the need to create a
structure. This intrinsic feature leads to a high struc- symbol or the need to challenge structural stability
tural risk in tall masonry towers, because increasing (and nature itself ). The interpretation of this desire to
height means large vertical loads and high compres- build higher, and simultaneously to reduce the safety of
sive stresses at the base. Therefore, it seems easy to structures, was left to ancient builders in the context of
accept that masonry towers should possess a higher almost no scientific basis. It is striking that the major-
safety margin than the values normally found for other ity of the ancient high towers in Italy, e.g. in Pavia

54
chap-05 9/9/2004 8: 6 page 55

Figure 10. The historical survey aimed at study existing

documents, drawing and bibliography to define the evolution
of the cathedral. The archaeological survey included in situ
excavation and stratygraphical records of the faces of the
masonry walls.

Figure 8. Physical protection of the granite stone include

sheeting in the external horizontal planes and installation of
a electrical anti-pigeon system.

Figure 11. For the time mechanisms, the works included

diagnosis, conservation and reactivation of function.

Figure 12. Restoration and conservation of the heritage

pieces included also diagnosis of stucco, full inventory,
reintegration in the original location, pre-consolidation and
cleaning of stucco.

Figure 9. Repair of finishings included repointing of of the first bricks and collapse. In the other cases, col-
masonry joints with lime mortar and repainting of wood with lapse occurred without previous warning. In Pavia, the
traditional oil paints. collapse occurred in a few seconds and provoked four
casualties (Macchi 1992).
Therefore, it seems unquestionable that the high
and Bologna, are no longer present, Macchi (1998). structural risk of high masonry towers justifies detailed
The reality is that only a few of these structures sur- studies and carefully planned remedial measures in
vived until today, due to collapses, destruction due to case of doubt about their behaviour. Several recent
lightening and even demolitions (often by precaution case studies have been reported, Modena et al. (2001),
and concern of eminent collapse). In the recent his- Valluzzi et al. (2003) and Lourenço (2004).
tory, there are well-known cases of collapse, namely
the Campanile in Venice (1902), the Civic Tower in
5.2 Description
Pavia (1989) and the Santa Maria Magdalena Bell-
Tower in Goch (1993). In the case of Venice, collapse The main façade was built between 1176–1200 (cen-
was gradual, with signs of distress two days before tral part) and 1229–1325 (towers), see Figure 13.
collapse. Also, 20 minutes elapsed between the fall The towers evolved into a Bell-tower (North) and a

55
chap-05 9/9/2004 8: 6 page 56

Figure 14. Partial sections of the towers: (a) horizontal sec-

tion of the South tower, (b) horizontal section of the North
North Tower South Tower tower and (c) sections A–D for the South tower.

Figure 13. Main façade after the restoration works in the

first half of the 20th century. vaults. The entrance for both towers is located at mid-
height, with a connection between both towers from
the top of the main vault. But the two towers have a
Clock-tower (South). In 1552, damage due to lighten- rather different structure.The South tower possesses an
ing is reported in the South tower. Between 1665–1669 internal core with a staircase shaped helicoidally, see
the South tower was demolished up to mid-height and Figure 14a,c. The North tower (presently with the bells
rebuilt. In 1717, it is recorded that the South tower and clock) features a horizontal mid-level with stone
was in the verge of collapse and, in 1727, buttresses slabs and architraves apparently supported in columns
were added, similarly to the ones that already existed and stone struts, see Figure 14b.
in the North tower. Pinnacles were added in 1732. The
construction of the Chapter House, contiguous to the
South tower, also aimed at consolidating the tower. 5.3 Constitution of masonry walls
Also in this period, the two small windows in the main The constitution of the masonry walls from the tow-
façade (South tower) were replaced by a single large ers was characterized using visual inspection, both by
window, similar to the one that existed in the North removing smaller stones of the outer leaves and by
tower. Before 1841, a new lightening stroke the South using a boroscopic camera inserted in cracks or in
tower. holes drilled in joints, see Figure 15a. From the inspec-
The cross section of the towers is approximately tion, it was possible to conclude that the three-leaf
square with a side of 10.0 m and exhibits a variable walls have external leaves of granite ashlars with a
thickness, with a minimum of 1.7 m at the base. The thickness ranging from 0.30 to 0.70 m, while the mid-
height of the towers is approximately 35 m, which dle leave is made from loose smaller stones and/or silty
means that the average stress at the base is around soil, see Figure 15b and Figure 16. The combination of
1.0 N/mm2 . This value is rather low for regular granite heavy rain in Porto, strong winds in the top of the hill
masonry but it is rather high for rubble masonry (with where the Cathedral is located, and the open joints in
or without mortar joints). In the main façade, two but- the external masonry face, results in a wet infill even
tresses are apparent in each tower, see Figure 13. As in the summer and the continuous washing out of the
addressed above, the structure suffered several major infill, see Figure 15c.
modifications through time, which resulted in a very
complex internal structure with different load bearing
5.4 Existing damage
internal elements at each level. The structure of the
towers cannot be understood from structural reasons The towers exhibit distributed cracking and signifi-
and several openings are closed, facing staircases or cant out-of-plane movements. The existing damage

56
chap-05 9/9/2004 8: 6 page 57

Figure 15. Visual inspection to define the constitution of

masonry walls: (a) boroscopic camera, (b) opening up the
structure (c) loss of material through central cracks in the
openings.

Figure 17. Ancient tower ties: (a) Deformed anchorage of

tie T1 and (b) details of broken tie T3 .

the anchorages are also deformed, as expected due to

the application of a large point load.
The South tower is more damaged than the North
tower. Figure 18 exhibits the location of severe cracks
and out-of-plumb walls in the South tower. Also the
East façade of the South tower presents out-of-plane
movements to the exterior. It is noted that the internal
walls of this tower are straight, indicating crumbling or
desegregation of the walls, with major cracks and voids
in the interior, see Figure 19a. The separation between
Figure 16. Typical cross section of the masonry walls. the internal and external leaves of the walls is fur-
ther confirmed by the longitudinal cracking observed
in most of the openings. Figure 19a illustrates such
resulted in the past addition of three iron ties (date cracking, with a maximum width of some centime-
unknown), see Figure 17a. Tie T1 presents a severely tres. Finally, it is noted that the North tower presents
deformed anchorage, see Figure 17b, and tie T3 is cor- severe distributed vertical cracking at the base, see
roded and broken, see Figure 17c. It is stressed that Figure 19c.
the separation between the East and West façades of This cracking is only visible in the internal
the South tower continued after tie T3 was broken. It (medieval) face, while the external face seems undam-
is also noted that the masonry walls in the vicinity of aged. Moreover, the very large thickness of the walls

57
chap-05 9/9/2004 8: 6 page 58

Figure 20. Plan of North tower at the base.

Figure 18. Location of most severe cracks and out-of-plumb

walls, in the South view and main façade.

Figure 21. Other damage: (a) Corrosion of the steel struc-

ture of the cupolas; (b) corrosion of dowels and ties in
balustrades and pinnacles; (c) deficient structural system to
support the bell stone level floor.

– Balustrades and pinnacles under deficient stability

Figure 19. Details of the cracks in the towers: (a) cracks up conditions and with significant movements due to
to 0.20 m width in the South façade/South tower; (b) typical corrosion of dowels and ties, see Figure 21b;
active cracks parallel to the walls at the openings; (c) vertical – Misconception of the structure supporting the bells
cracks at the base of the North tower. and clock in the North tower, see Figure 21c.

are not replicated in the South tower, see Figure 20. 5.5 Remedial measures
For these reasons, it is believed that the damage is
As it arises from the history and survey, the towers
not recent and the helicoidal staircase belongs to the
seem to have been damaged in the past and rebuilt
structure of an older tower.
(particularly the South tower). The (re)construction
Additionally, also the following damage is noted:
seems to have been carried out under deficient exe-
– Steel structure in the cupolas of the towers with cution conditions, no particular well defined structure
advanced corrosion, see Figure 21a; and using improvised construction details. In addition,

58
chap-05 9/9/2004 8: 6 page 59

different remedial techniques were already used in the

past aiming at correcting and strengthening the towers.
The walls of the towers seem not to possess adequate
connection between the external leaves and severe
water infiltration in the walls contributed to the exist-
ing damage and to the loss of material in the rubble
infill. Here it is again stressed that the Cathedral is
located at the top of a hill, the masonry joints have lost
all mortar and it was found that the rubble infill was
wet by the end of the summer.
Besides other damage, the most relevant feature is
that the North tower is divided in two similar U-shaped
parts, from mid-height to the top, with full cracks
along the West-East direction (in the other direction,
the existing ties kept the tower together), and the South
tower is bulging outwards both to South and to East
(the existing West-East tie is broken).
The solution adopted for strengthening consists
mostly of a steel ring in both towers, aiming at confin-
ing the structure along the two orthogonal directions,
in the sole location possible, see Figure 22a,b. The
rings are made with welded stainless steel plates
(class AISI 316L), connected to the towers using
long, inclined stainless steel anchorages inside of
a cloth duct to prevent generalized injection, see
Figure 22c–e. The length of the steel profiles is defined
so that the elements can be transported to the location
through the existing doors and can be easily assembled
in situ, without any further welding.
In the North tower, the ring also aims at providing a
support for the stone pavement for the bells.The reason
being that the stone columns are very deteriorated and
possess presently no structural function and the stone
struts have very deficient conception, see Figure 22c.
Here it is noted that it was decided not to recuperate the
structural function of the columns (e.g. using injection)
because the lower level seems to indicate insufficient
strength of the inner core, see Figure 20c and Figure 21.
The steel ring is made of U profiles (240 × 120 mm
and 200 × 100 mm height).
In the South tower, a set of two ties was provided
to the ring, because it was possible for aesthetic rea-
sons and they are a witness of the ancient broken tie.
The ring must cross the staircase at a selected loca- Figure 22. Aspect of the strengthening of the towers using
tion because the complex internal structure of the stainless steel rings and long inclined anchorages: (a) plan of
tower does not allow otherwise. Due to the lack of the ring for the North tower; (b) plan of the ring for the South
internal stiffening elements, a much more stiff steel tower; (c) North-South section for North tower; (d) West-East
frame is needed and the steel ring is made of I profiles section for North tower; (e) typical section for South tower;
(f) additional ties placed in the West and South façades of the
(180 × 180 mm). Due to the bulging outwards of the South tower; (g) details of the anchorage plates.
East and South façades, and the severe cracks in the
corners, several short ties have been added to the struc-
ture to stitch the East and South façades, and two long Figure 23 illustrates the aspects related to repair-
ties through the core of the South façade have been ing the pinnacles and balustrades, which consisted of
added to connect the West and East façades, see Fig- replacement of iron dowels and ties by stainless steel.
ure 22f. Figure 22g presents details of the two types of The large pinnacle in the top of the North tower cupola
anchorage plates adopted (circular plates and specially was totally loose at the time of the works and was
designed crosses). jacketed with steel plates at the top and bottom necks.

59
chap-05 9/9/2004 8: 6 page 60

Figure 23. Pinnacles and balustrade removal and place- Figure 24. Roof of Saint Vincent Chapel: (a) aspect of
ment with stainless steel dowels and ties: (a) general view; restoration works; (b) aspect of vault infill with rubble.
(b) removal of the centre pinnacle; (c) aspect of stainless steel
elements and placement.
the chapel vault was filled with rubble resulting from
old demolitions see Figure 24. Also, and as usual in
The other works carried out include injection of the
several historical constructions, the timber roof was
main cracks with lime based mortar grout, repointing
partly supported by the vault, using later added struts.
all joints with two selected lime mortars (a traditional
The issue addressed here is the stability of the vault
mortar for the filling and a more durable lime mortar
and the convenience of the removal of the infill.
for the finishing), protect against corrosion (the two
ties in the North tower were kept in place) or replacing
all existing iron. 6.1 Survey
The structure consists of a barrel vault with an approx-
5.6 Monitoring Plan imate thickness of 0.25 m and a span of 6.8 m. On the
Given the cultural importance of the building and the North side, the cloister acts as a buttress but on the
significant damage in the South tower, a monitoring South side no buttresses are present. Even if the South
system was planned and installed. wall (1.70 m) is thicker than the North wall (1.30 m),
The system includes four waterproof crackmeters out-of-plumb movements outwards are clearly visible
in the largest cracks, two strain gages for the new ties, in the former, up to 1.5% (or 0.10 m at the springer of
two biaxial clinometers to measure the tilting of the the vault), see Figure 25a.
tower, as well as temperature, humidity and wind sen- Nevertheless, as the vault presents only minor
sors. The system includes a GSM interface for remote cracking, see Figure 25b, it was believed that the vault
monitoring. has been built after the wall deformation. As it will be
confirmed next, the vault replaces a previous timber
roof at the same level.
6 SAINT VINCENT CHAPEL
6.2 Structural Analysis
The Saint Vincent Chapel is located next to the South
wing of the Cathedral cloister. During the restoration A plane model was adopted for the structural analysis
works of the roof, it was found that the extrados of of the barrel vault. The analysis was carried out using

60
chap-05 9/9/2004 8: 6 page 61

Figure 26. Results of the numerical analysis, in terms of

thrust-lines and failure mechanisms: (a) with infill, for a ulti-
mate load factor equal to 6.5; (b) without infill, for a ultimate
load factor equal to 9.4.

Figure 25. Saint Vincent Chapel: (a) cross-section; (a) (b)

(b) aspect of vault intrados; (c) survey of inspection pits.
Figure 27. Infill removal: (a) location of topographic targets
for monitoring; (b) aspect of the cleaned vault.
limit analysis, discretizing the walls and vault as a set
of rigid blocks, see Orduña & Lourenço (2003). More load factor increases 45% if the infill is removed,
complex approaches are available, Lourenço (1998, which seems also natural because it was not originally
2001), if necessary for more detailed studies. The planned for this construction.
assumed material properties include a tensile strength
equal to zero, a tangent of the stiffness angle equal to 6.3 Remedial measures
0.7, zero dilatancy and a compressive strength equal
to 6 N/mm2 . The actions included consist only of the The infill was removed but, for safety reasons, it
self-weight of the structure. was recommended to accompany this task with topo-
As the objective of the analysis is to evaluate the graphic measurements, see Figure 27a. The targets
influence of the infill, a sophisticated representation were read always at early morning to reduce tempera-
of the structure is not particularly relevant. Therefore, ture effects, daily during the process of infill removal
the influence of the cloister, openings of the walls and (one week) and weekly during one month after load
ribs of the vault were neglected in order to avoid the removal. Approximately 35 m3 (7000 kg) of rubble
need of a three-dimensional model, Lourenço (2002). were removed from the vault and no movements were
The numerical results are given in Figure 26, in recorded in the targets.
terms of thrust-lines and collapse mechanisms, both Figure 27b demonstrates that (a) the vault was never
for the model with and without infill. The ultimate conceived to accommodate infill and (b) a timber roof

61
chap-05 9/9/2004 8: 6 page 62

existed at the level of the vault, before the construction Lourenço, P.B. 1998. Experimental and numerical issues in
of the vault and the new roof at a higher level. the modelling of the mechanical behaviour of masonry.
In: P. Roca et al. (ed) Structural Analysis of Historical
Constructions, p. 57–91, Barcelona: CIMNE.
Lourenço, P.B. 2001. Analysis of historical construc-
7 CONCLUSIONS tions: From thrust-lines to advanced simulations. In:
P.B. Lourenço, P. Roca. (ed) Historical constructions
The present paper addresses the works recently car- 2001, p. 91–116, Guimarães: University of Minho.
ried out at the Cathedral of Porto as a case study. The Lourenço, P.B. 2002. Computations of historical masonry
methodology that governed the complete set of works constructions, Progress in Structural Engineering and
is addressed and the conservation and repair works Materials 4(3): 301–319.
are briefly addressed. Two aspects are treated in detail, Lourenço, P.B. 2004. Strengthening design of the clock-tower
namely the towers and the Saint Vincent Chapel. in Mogadouro Castle, Portugal. Report 04-DEC/E-07,
The towers exhibit severe global damage includ- Guimarães: University of Minho.
Macchi, G. et al. 1992. The collapse of the Civic Tower in
ing cracking, crushing and separation between leaves Pavia: A survey of the materials and structure. Masonry
and also local damage in the cupolas, pinnacles and International 1.
balustrades. The global damage seems mostly due to Macchi, G. 1998. Seismic risk and dynamic identification in
water infiltration, deficient conception of the struc- towers. In Monument 98 – Seismic performance of monu-
ture, ancient damage due to lightening and changes ments; Proc. intern. workshop, Lisbon, 12–14 November
in the structures of the towers. For the purpose of 1998, p. K.3-K.17, Lisbon: DGEMN, LNEC and JRC.
increasing the structural performance, a rigid frame Modena, C., Valluzzi, M.R., Tongini Folli, R., Binda, L. 2001.
of stainless steel profiles and a set of long, inclined Design choices and intervention techniques for repairing
anchors have been designed to provide a confining and strengthening of the Monza cathedral bell-tower. In
Structural faults & repair 2001; Proc. 9th intern. conf.
ring. In addition, new ties and stitching of the external and exhibition, London, 4–6 July 2001. CD-ROM.
leaves were also included when necessary. The local Orduña, A., Lourenço, P.B. 2003. Non-linear cap model for
damage is mostly due to corrosion of iron elements, limit analysis of masonry structures, J. Struct. Engrg.,
which have been replaced by stainless steel elements ASCE 129(10): 1287–1430.
or have been protected. Valluzzi, M.R., da Porto, F., Modena, C. 2003. Structural
The chapel exhibits a significant overload due to a investigations and strengthening of the civic tower in
rubble infill resulting from previous demolitions and Vicenza. In: Structural faults & repair 2003, Proc. 10th
the external wall presents moderate out-of-plane dis- intern. conf. and exhibition, London, 1–3 July 2003.
placements. From the diagnostics, it was possible to CD-ROM.
safety prescribe the removal of the infill (approxi-
mately seven tons). This operation allowed to confirm
that the present vault is not contemporary to the walls
and the external wall deformation is stabilized.

REFERENCES

Binda, L., Anzani, A. 1997. The safety of ancient masonry

towers: A survey on the effect of heavy dead loads. In
G. Arun, N. Seçkin (ed), Studies in ancient structures;
Proc. intern. conf., Istanbul. Istanbul: Yldiz Technical
University.

62
chap-06 9/9/2004 8: 6 page 63

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

Considerations on the significance of history for the structural

analysis of ancient constructions

P. Roca
Universitat Politècnica de Catalunya, Barcelona, Spain

ABSTRACT: This paper presents some reflections on the role of structural modelling in the study of ancient
structures and its relationship with other sources of knowledge, such as inspection, history, and monitoring. The
relation between structural modelling and historical information is considered with special focus. A discussion is
carried out on the capabilities which, ideally, numerical models and methods of analysis should include to obtain
a fruitful connection with historical information. In particular, the need for procedures allowing the simulation
of historical aspects such as the construction process, later structural alterations or long-tem physical damaging
phenomena is justified in the light of particular case-studies.

1 INTRODUCTION understanding of safety and our current techniques for

the assessment of modern structures. Among these,
The study of the structure of historical constructions major earthquakes of hurricane wind occurrences cor-
poses a number of very significant challenges to the responding to long return periods, or the repeated
analyst. Difficulties stem not only from the need to effect of minor earthquakes or thermal cycles acting
model complex geometries, materials and actions, but during long periods of time, can be recalled. Phys-
also from the significance of history and the need ical or mechanical phenomena developing through
for linking the structural analysis with the historical historical periods (such as those manifesting in the
events. As can be observed apropos of the study of real form of a slow decay) or more instantaneous but
cases, analysis disconnected from history may show impacting events like destruction caused by wars, or
sterile or may yield results in contraction with reality. important architectural alterations, are also to be taken
Certainly, part of the difficulties comes from the into consideration. Besides the technical difficulties
overall complexity of historical structures in terms of which may be found in the attempt of numerically
material, geometry and actions. Masonry, as a compos- simulating them, most of these actions may be insuf-
ite and brittle material, develops complex mechanical ficiently known not only historically but also from a
and resistant phenomena whose simulation is still phenomenological point of view.
challenging the capacity of advanced theories and Furthermore, all the cumulative effects of such
computer methods. actions contribute to cause a non-stopping, gradual
Modelling the materials and the morphology faces transformation of the structure. Historical structures
also from the impossibility to reach a comprehen- are not something inert or unchanging, but alive and
sive characterization. This factual impossibility is due always evolving towards, normally, a more sensitive
to the need for limiting the extent of destructive and vulnerable stage, unless maintenance and restora-
inspection and the preference for more indirect meth- tion are carried out, meaning, in turn, more transforma-
ods. Indeed, information on materials and internal tion. This gradual evolution is clear in what concerns
composition remains very limited in many cases. geometry due to the constant increase of deformation
The organic geometry of many masonry construc- produced by a variety of concurring actions (creep, soil
tions, resulting as a combination of linear, 2D or settlements, thermal cycles …). Anthropic alterations
3D, straight or curved structural components such and repairs contribute to this changing nature as well
as arches, buttresses, multi-layered walls, vaults and (see also Roca, 2001).
domes, must be described by means of capable anver- Another difficulty coming from history is found in
satile modelling techniques. the understanding of the design concepts and construc-
Historical structures are or may have been sub- tion methods originally used to produce the construc-
jected to actions that are unfamiliar to our conventional tions. Ignoring the original aims, the design processes,

63
chap-06 9/9/2004 8: 6 page 64

even the ideas inspiring the structures, will not help us

in the understanding of them. HISTORY INSPECTION
PRESENT CONDITION
Because of all these considerations, conventional
methods of calculation, as well as the conventional cri-
teria for the assessment of structural safety, are hardly
applicable to the study of ancient constructions. Offi-
cial codes for structural design, normally oriented to
modern constructions, are not either easily transfer-
able to the case of historical buildings. (These facts
are clearly addressed in the Recommendations of the
Iscarsah Committee, 2001.)
It is the footprint of history what makes the anal-
ysis of a historical building so challenging and what
prevents from the applicability of conventional meth-
ods. However, history is also a possibility and a source STRUCTURAL MONITORING
of understanding, for history can also be interpreted MODELLING
as an experiment in the true scale of time and space.
Knowledge on historical facts may provide precious
empirical information.
Since conventional methods and concepts may not
be usable, another more general approach is required
which considers all the viable sources of evidence,
including inspection, history, monitoring and struc-
tural analysis.

2 THE ELEMENTS OF THE ANALYSIS

2.1 Main activities

Studies on the structure of ancient constructions base CONCLUSIONS ON THE CONDITION OF THE BUILDING
on the combination of a set of activities which include
(1) the inspection and characterization of the present
state of the construction, based both on visual recogni- Figure 1. The elements of the analysis.
sance and deep observation by means of (preferably)
non-destructive techniques, (2) the historical research be considered as one of the most precious elements
carried out by expert historians from the available of the analysis. However, using historical information
historical documentation, (3) the monitoring of the for the better understanding of an ancient structure
building by means of different types of sensors and is not always feasible or straightforward. The sources
(4) the structural modelling and analysis of the build- (such as the books available in the files of the chapters
ing. These are, in short, the main elements of the study of Gothic cathedrals) must be investigated and inter-
from which the conclusions on the condition of the preted by specialized historians, who later report to
building and on the need for repair or strengthening the structural analyst what they think is meaningful for
are to be drawn (Fig. 1). the structure. Effective cooperation and communica-
tion between archaeologist or historians and structural
analysts may face certain difficulties due to their
2.2 Historical research
very different specialization, vocabulary, and priori-
As aforementioned, history can be understood as a nat- ties. Nevertheless, a very enriching cooperation may
ural experiment occurred at true space and time scales. take place between both specialists when communi-
This experiment accounts for the effect of real actions cation is fluid and common objectives are shared and
occurring in historical periods which can hardly (or discussed. The tasks of the historian and the structural
just can’t) be reproduced in laboratory or be monitored analyst are linked, in any case, by their wish to reveal
in the real building. and enhance all the value attributed to the building as
Because of that, historical research can provide part of the architectural heritage. They also share the
important clues for the understanding of the present common objective of contributing to preserve all what
condition or to illustrate the response of the building is historically genuine in the building.
when subject to major actions. Historical investigation This requires the development of an interest in
provides extremely significant information and must history (at least, in history of construction) by the

64
chap-06 9/9/2004 8: 6 page 65

structural analyst, while also demands a sensitivity

or interest towards the structural and technological
features of the constructions from the historian.
Unfortunately, a fruitful cooperation between his-
torians and engineers or architects is not enough
to ensure an accurate utilization of the information
extracted from the historical files. Historical sources
provide, most often, only partial and inaccurate infor-
mation, while showing almost no technical insight.
Interpreting the historical notices and linking them
with real structural or morphological features (con-
struction, failures, repairs, alteration, damage …) may
require much speculation, and thus give chance to
possible subjectivity.
Having an accurate view of the properties of the
structure in its original configuration, or system-
atically recorded details on its alteration may be
extremely difficult in many cases. Documents (or
historical evidence of any kind) describing the orig-
inal construction or latter alterations are not always
available and, when they exist, they may provide
only ambiguous or partial information. Even in the
case of important buildings with comprehensive infor-
mation available, such information may experience
significant lacks and thus leave unanswered cru-
cial inquiries for the comprehension of the building.
Mallorca Cathedral, as for instance, has been subject to Figure 2. Double battery of flying arches of Mallorca
comprehensive historical research; thanks to the effort cathedral, with the upper ones showing large deformation
and occasional propping.
in reviewing all the available documents, very inter-
esting aspects have been clarified, such as the entire
chronology of the construction (spanning for more
than three centuries), the original quarries from which
the different stones used where taken, or on the deci-
sions leading to a set of important repairs (Domenge,
2004). In spite of it, no clear information has been
found, for instance, about the reason for building the
upper battery of flying arches. Such flying arches are
not conveniently placed for an efficient contribution
to overall equilibrium; they were not either built to
counteract the effect of wind on a pitched roof which
did not exist in origin (Fig. 2; Roca and González,
2001). No clear historical notice has been found on
the reason for overloading the vaults and transverse
arches with large pyramids of stone (in a way that
cannot be seen in any similar building, Fig. 3); nei- Figure 3. Dead weight on vaults and transverse arches of
Mallorca Cathedral.
ther is any evidence either on the period in which the
vaults were overloaded. On all those issues, histori-
ans and structural analysts have their own guesses, but doesn’t allow objective understanding of their actual
no clear evidence coming from historical research. In material extent or structural significance.
Tarazona Cathedral, and also in spite of the historical It may be difficult to distinguish between mechan-
research, the question whether the original cimborio ical alterations (cracks, deformations) arrived in the
collapsed because of structural problems, or whether it long-term, from those existing in the early configu-
was just dismantled to build a new, larger one, remains ration of the building. Masonry historical buildings
unanswered (Aguerri and Rodríguez, 1998). In other deform, settle and crack during the construction pro-
buildings, information may be scarce or just inexis- cess itself. In the case of the mentioned cathedrals,
tent. Notices on partial failures or structural problems it is very difficult to guess which part of the settle-
are mentioned in historic files in a way which usually ments and deformations was experienced during the

65
chap-06 9/9/2004 8: 6 page 66

construction process (when removing the centerings,

for instance), and which part has been caused by later
phenomena, such as earthquakes, long-term damage DAMAGE
of the repeated thermal cycles.
Historical record is essential to understand the
structure. Even if the historical information may be
incomplete, we must try our best to gather as much
historical evidence as possible. However, the available SAFETY
historical record may not be enough to really under-
stand the structure; it is but one of the several sources
which contribute to our knowledge.

2.3 Inspection
Long term
Activities related to both superficial or deep inspec-
tion provide qualitative information on the present ∆T
condition of the building, encompassing geometry,
morphology, construction details, material properties,
alterations, historical repairs, and existing damage.
Inspection provides the essential information needed amplitude
to select the primary hypotheses on which the model
is based. It also provides the amount of information
needed to build a structural model in detail.
In some cases, realizations coming from inspection
may permit a certain calibration of the model. This cal-
ibration may be achieved through the comparison of
possible predictions of the model (in particular, those
∆T
related to deformation or damage) with effects actu-
ally observable in the building. For instance, a first
illustration of the validity (or inadequacy) of a model
may be achieved by simulating the effect of gravity
loading and then comparing the predicted distribution Figure 4. Monitoring as a window over historical time
of cracking and deformation with their real manifesta- allowing possible insight on active processes long-term
tions in the building. However, this type of calibration damaging processes.
may be hardly undertaken in practice because of the
uncertainty on the actions having historically impacted
on the building and the superimposition of effects information useful for the study and restoration of
involved in damage. Even if such aspects remain ancient constructions (see Roca et al., 2003, for a
uncertain, results from inspection should be utilized larger discussion on this issue).
to check the model and see that the predictions are However, characterizing long-term damage is a
not in frontal contradiction with reality. For instance, challenging task due to the slowness of the pro-
a model producing cracks, due to dead load, in parts cesses involved and the fact that they may be masked
which have remained intact from the construction of by more apparent, short-term variations caused by
the building itself, should obviously be subjected to present actions. In order to characterise incremen-
reconsideration. Inspection provides, at least, the pos- tal, long-term processes, monitoring must be designed
sibility of a preliminary test for essential compliance to allow a clear distinction between the reversible or
with reality. cyclic components of the parameters measured, on
the one hand, and the irreversible, cumulative compo-
nents developing in the long term, on the other hand.
2.4 Monitoring
In turn, the possibility of recognising both types of
Monitoring provides quantitative information on the components demands the following requirements:
response of the structure across a short and contempo- Monitoring must be extended to a period long
rary period of time. Monitoring can be understood as enough to cover the entire duration of acting cyclic
a small window opened over the long term processes actions; since annual variations of temperature must
taking place in the otherwise inaccessible historical be considered in any case, the minimum acceptable
time (Fig. 4). In particular, monitoring may allow period is a complete year. Additional years will be of
for the recognition of incremental processes in a term value to confirm the tendencies observed and appraise
reasonable for engineering purposes and thus provide their possible evolution in the long term.

66
chap-06 9/9/2004 8: 6 page 67

Monitoring must afford the characterisation of the HISTORY STRUCTURAL

environmental actions occurring during the studied MODEL
period. This is particularly so in the case of the cli-
mate environmental actions; parameters such as the
temperature and the humidity are to be measured at
different stations in both the interior and the exte-
rior of the building. Wind velocity and direction may
be characterised by means of conventional measur-
ing devices. Similarly, valuable information can be
obtained if the effect of occurring low-intensity earth-
quakes or environmental vibrations is recorded by
means of seismographs.
Structural analysis and monitoring do both handle
quantities and thus permit a direct comparison. Results CONCLUSIONS CONCLUSIONS
of monitoring can be used for calibrating a numer-
ical model, provided that not only the parameters
associated to the response (deformations, displace- No hypotheses posed No hypotheses proven
ments, rotations, vibrations) are measured, but also
those characterising the actions (environmental ther- Figure 5. Debatable approaches not taking advantage of all
mal effects, ground motion). In turn, simulation by possible sources of information.
a numerical model can help laying-out and adequate
and truly informative monitoring by giving light on,
hypotheses not allowing possible evaluation are in
for instance, the more significant variables to be mea-
fact of null epistemological value; they will not add
sured, the expectable ranges of variation (which are
any valuable understanding and will be useless for a
meaningful to select the sensors) or the best location
reliable prediction on the response and needs of the
for the sensors.
building analysed.
It is on the light of the above understanding that the
different elements of the study are to be displayed and
2.5 Structural modelling
connected. Some of the mentioned activities (in fact,
The role of the structural analysis within the frame of the structural modelling) are related to the fist stage
a general study and the requirements which should be of the process, that is to say, to the adoption of the
taken into consideration to link it, in a profitable way, hypotheses. The model, whatever its nature (concep-
with the rest of the elements of the study (inspection, tual, analogical, numerical …) is just the recipient of
history, monitoring) are discussed in the following the hypotheses on the physical and mechanical nature
sections. of the construction. History, inspection and monitoring
are activities oriented to the production of the empir-
ical evidence needed to validate the hypotheses and,
3 ROLE OF STRUCTURAL ANALYSIS if needed, to disregard, correct or improve them to a
satisfactory extent.
Drawing accurate conclusions from historical research, This understanding of the problem may seem obvi-
inspection, monitoring and structural modelling ous to any technician really committed with a scientific
requires a clear statement on the real meaning of these understanding of the study of ancient constructions.
activities and the way they are connected to each other. Unfortunately, some analysts did proceed, in the past,
Since conventional methods and theories may not using incomplete reasoning schemes which, in some
be applicable to the study of ancient constructions, occasions, produced useless or incorrect conclusions
technicians working on them must base their investi- (Fig. 5). As for instance, a blind satisfaction with an
gation on the more universal and elementary approach unproven conceptual or numerical model, from which
given by the scientific method. In our understand- results indicating the weakness of the structure were
ing, all these activities are precisely the pieces needed obtained, led, in some cases, to the implementation
to develop an investigation adequately based on the of an impacting strengthening causing a significant
scientific method. (and irreversible) transformation of the historical con-
Applying the scientific method requires, first, the struction. Since the model had not been validated, no
adoption of a set of hypotheses and, second, the real demonstration for the need of such strengthening
use of available empirical evidence to prove them. had been really provided. Furthermore, real exam-
To actually comply with the scientific method, the ples exist (like in the case of the Segovia Aqueduct,
adopted hypotheses must allow demonstration or refu- Jurado, 1995) for which it has been possible to demon-
tation based on the observation of reality. Proposed strate that the already implemented strengthening was

67
chap-06 9/9/2004 8: 6 page 68

HISTORY PRESENT MONITORING STRUCTURAL

CONDITION MODELLING

EMPIRICAL EVIDENCE HYPOTHESES

CONCLUSIONS

Figure 6. Role of the different activities in the study of an ancient construction. Conclusions on the condition of the building
stem from the validation of hypotheses adopted to elaborate a model, once validated or calibrated in the light of available
empirical or experimental evidence obtained from historical research, inspection and monitoring.

in fact not really needed and even unsuitable for the progress of damage), possible architectural or struc-
conservation of the structure. tural modifications caused on the building, or because
Results from non-validated modelling may be sim- of the particular seismic features of the new earthquake
ply in contraction with meaningful historical facts. (as in the case of the Basilica of St. Francis of Assisi,
For instance, the analysis may predict that an exist- see Croci, 1998).
ing and well preserved dome can not resist a major Even if history is regarded as the main and best
earthquake having, in fact, occurred several times dur- source of evidence, complementary experimental and
ing the life-span of the building; if anything, this mere numeric analyses are needed to actually understand, in
fact indicates that the model is not entirely correct a rationale way, the condition and needs of the building.
and is still requiring further calibration. Although the The problem of this second way or reasoning is in the
hypotheses are (willingly or unwillingly) present, no lack of hypotheses actually being checked.
previous effort has been devoted to gathering empir- To our understanding, an adequate, scientifically
ical evidence and using it to validate or calibrate the based approach must consider both the structural mod-
model before exploiting it for predictive purposes. elling, as the recipient of the hypotheses to be checked,
Another debatable approach consists of drawing the and the evidence coming from the inspection (pro-
conclusions from the only knowledge of history. Cer- viding qualitative evidence on the present condition
tainly, the mere realization that a building is standing of the building), historical research (as a source of
offers an empirical hint of its viable stability under qualitative evidence of the past performance of the
the action of gravity. However, the actual meaning of construction the true scale of space and time) and
this fact can only be completely interpreted in the light monitoring (providing quantitative information of the
of a (conceptual, analogical, numerical) model of the current performance at a local time interval).
structure. As shown by some real cases, ancient struc- The need for a multidisciplinary approach stems not
tures having been stable for centuries may collapse in a only from the variety of aspects involved by the study
sudden manner without providing previous visible evi- but also from the wish to apply the scientific method
dence of their precarious condition (as in the case of the in a consistent and profitable way (Fig. 6).
Civil Tower of Pavia or Noto Cathedral, see Macchi,
1998, Croci, 1998, and Binda et al., 2003). Similarly,
the fact that a building has been able to resist some 4 REMAINING SPACE FOR UNCERTAINTY
historical earthquakes is not enough to conclude on its
capacity to resist possible new earthquakes. The safety It must be recognised that even if modelling (as recip-
of the building in the case of a future earthquake is not ient of the hypotheses) is used in combination with
guaranteed due to the structural decay (caused by the empirical evidence, and even if a model or a method

68
chap-06 9/9/2004 8: 6 page 69

the application of a scientifically-based analytical

approach. The hypotheses contained in a model can
be classified into two different categories.
On the one hand, a set of primary hypotheses,
referring to physical or mechanical principles, is (con-
sciously or unconsciously) adopted in order to select
1ST PHASE 2ND PHASE and adequate frame theory or formulation allowing
PREDICTIONS the calculation. These include principles related to the
USED FOR technique used for the description of the geometry
STRUCTURAL
MODEL VALIDATION/
PREDICTIONS
and the materials, such as the validity of continuum
USED FOR
CALIBRATION ASSESSMENT mechanics, or the adoption of particular constitutive
equations. When it comes to brick or stone masonry,
hypotheses on this respect are not yet conventionally
Figure 7. Two phases in the use of a structural model: or universally accepted; further research is still needed
validation and assessment. in spite of the significant effort carried during the last
decades (see Lourenço 1998, 2001, for state-of-the art
of analysis is consistently validated and calibrated, a reports on this subject). Analysts must still undertake
certain margin for subjectivity and uncertainty exists. a certain contribution from his own in order to select
The reason for remaining subjectivity is in the fact and justify the optimal approach to be used in each
that the model or method will be validated for actions particular case.
or general conditions with available information on The failure of a model to represent real mechanical
the building, and thus different from those for which or strength phenomena observed in the building (at
the model will be used for assessment (Fig. 7). For least, those that the analyst is trying to simulate) must
instance, an attempt may be undertaken to validate or lead to the conclusion of the need for another approach.
calibrate a model using information from inspection In any case, the analyst must be aware of the limitations
which characterize the response of the building sub- of the theory or formulation adopted; he must be able
ject to dead load. Distribution of cracking predicted to interpret critically the results obtained and to reject
by the analysis for dead loading may be compared those not being meaningful, or to recognize, if the case,
with cracks existing in reality. Or, deformation due the entire failure of the approach.
to simulated thermal effects, vibration amplitudes or On the other hand, models are also based on a
frequencies due to simulated traffic or wind may be secondary set of more particular hypotheses on the
compared with real measurements captured through a geometry, morphology and material properties of the
monitoring campaign. Once validated, the model will building. No matter the effort invested in the inspec-
be used for the purpose of assessing the response of the tion and experimentation, the amount of information
building subjected to a variety of hypothetical actions will never be enough to describe the full complexity
(such as major earthquake). This operation requires, of the structure, nor will any conceivable model be
in fact, to extrapolate from the set of hypotheses or able to receive it comprehensively. The knowledge of
conditions for which the model has been consistently the structure is always partial, which requires for addi-
validated. However, this extrapolation – and the cor- tional hypotheses on its geometry, construction details
responding caused uncertainty – are an unavoidable and materials. Furthermore, viable models require an
effect inherent to the fact of predicting. The technician important level of idealisation – which is fact consis-
will have to base on his understanding and experience tent with the very essence of modelling. Additional
to determine whether the use of the model is reason- hypotheses are needed about the best way of simplify-
able or whether it is abusively trespassing the sphere ing or idealising the features of the particular building.
of tolerable uncertainty. Can a wall be described as a solid component, or
In turn, and because of this remaining uncertainty, should it (to the cost of larger computational effort
predictions and conclusions are always to be accepted and more accurate information available) be modelled
as provisory. Further activity involving monitoring as its real three-leaves counterpart? Can a column be
and periodical inspection of the already assessed or described as a uniform, cylindrical body or should it be
restored buildings are required to a posteriori vali- modelled as a geometrically and materially complex
date the predictions – and the subjacent hypotheses- body composed of many different units having peculiar
or review them. geometrical details? Do we need to include in model
the mass of soil on which the building is founded?
5 THE STRUCTURAL MODEL Again, this second set of hypotheses must be criti-
cally checked and the possibility of having to improve
As aforementioned, the structural model consti- them, even to large extent, must be accepted. These
tutes the recipient of the hypotheses adopted during hypotheses may be improved through a calibration

69
chap-06 9/9/2004 8: 6 page 70

process based on the comparison of the numerical or empirical evidence must be only extended to an
predictions with the evidence available on the real engineering level. Aiming at a physically exact or
response of the building. extremely close agreement between model and reality
In all this discussion, the requirement for the agree- may be sterile or prohibitively expensive for most pur-
ment between numerical results and experimental poses. Required cost and term must keep reasonable
and consistent with the importance and complexity of
the building and the resources available.

6 CAPACITIES REQUIRED FOR THE

ANALYSIS

Historical constructions may show a variety of struc-

tural alterations (such as large deformation, cracking
or related damage in compression, and fragmentation
due to cracking in tension or shear) caused by phenom-
ena or actions developed during historical stages and
long periods of time. In many buildings, such forms of
damage do manifest in a very severe and apparent way.
Conventional instantaneous analyses are not adequate
for the study and the simulation of such effects.
In particular, it has been observed in a number
of case-studies that instantaneous analysis may pre-
dict displacements orders of magnitude smaller than
those shown by the real building; notwithstanding, the
obtained deformed shapes are qualitatively similar to
the real ones (Figs 8, 9).
More accurate studies intended to simulate aspects
developing in the historical time may require specific
capacities to simulate the following aspects:
(1) the subsequent historical stages of the building
(in particular, the construction process) through a
sequential analysis,
(2) the actions occurring in historical periods, such as
Figure 8. Analysis of the transverse section of Mallorca
Cathedral. Above: deformation (×2000) and distribution of
major earthquakes or the repeated effect of minor
normal stresses (10 × kN/m2 ) in chromatic scale (white used earthquakes or thermal cycles, and
for cracking) for dead load. Below: deformation (×1000) (3) long term rheological or damage processes (such
and distribution of stresses predicted for a major earthquake as those related to creep) developed through the
(ac /g = 0.12). life of the construction.

Figure 9. View of Küçük Ayasofya Mosque in İstanbul, compared with the deformation predicted by the analysis for gravity
loading (amplified × 2300).

70
chap-06 9/9/2004 8: 6 page 71

It must be noted, however, that a realistic simulation Effects due to historical actions may also con-
of historical actions or long-term damage processes tribute very significantly to the continuous increase
is still requiring additional experimental studies and of deformation. Extraordinary actions such as large
numerical developments. earthquakes may produce important lesions and irre-
Faculties (1) to (3) may provide meaningful results versible deformations (Fig. 8). Low-intensity earth-
in the frame of a non-linear method of analysis quakes or repeated occurrences of hurricane-force
accounting for both the geometric and material con- wind may act cumulatively to cause ever increasing
tributions to a non-linear response. On the one hand, damage and deformation. Daily or annual thermal
large displacements may affect the equilibrium of slen- cycles individually have a minimal effect; however, a
der masonry constructions; this can be observed in certain, irreversible increment of deformation may be
slender arches or in more complex skeletal systems. produced after each cycle, thus contributing over very
On the other hand, masonry shows a highly non- long periods of time to a meaningful increase in overall
linear response, even for moderate stresses, due to deformation. It must be noted that the effects of cyclic
cracking in tension or other associated phenomena. actions do not dissipate with time, but may increase in
Consequently, methods of analysis must be equipped an accelerate way as the construction becomes more
with appropriate procedures for the description of the and more damaged.
equilibrium of the deformed structure, and also with The damage affecting the construction, which in
realistic constitutive equations for the simulation of normal conditions always increases due to the men-
the mechanics and strength of masonry. The capac- tioned and other possible causes, will, in turn, enlarge
ity to account for these effects is of large importance the sensitivity of the construction towards a vari-
since they act in a coupled way with the aforemen- ety of actions. This situation contributes to con-
tioned historical effects (1) to (3). For instance, and stantly increasing (never-mitigating) deformation at
from a theoretical point of view, additional flexibil- the long-term or even accelerated long-term deforma-
ity due to cracking and progressive deformation due tion which, in the worst case, can lead to the collapse
to creep might cause a very important increment of of the construction. Since the more persistent action
the eccentricity of the normal force at certain sections is gravity, it is not strange that such constant increase
of a pier or an arch, leading, ultimately, to a differed of deformation may manifest as a monotonic, non-
collapse under the only action of gravity. asymptotic amplification of the initial deformed shape
Simpler calculation methods, such as linear elastic due to dead load.
analysis or limit analysis based on the plastic the- Ideally, the numerical model used should be able
orems may be very useful for the study of ancient to simulate most of the present or historical actions
constructions and may really contribute, when handled having affected the construction; it should also per-
by an experienced analyst, to face complex problems. mit sequential analysis to simulate the construction
The above discussion is in no case oriented to dis- process and the latter possible structural alterations or
credit the use of these simpler tools in the study of repairs. Having specific constitutive equations avail-
ancient constructions. However, the possible simula- able for long-term creep of masonry or stone-like
tion of complex actions and responses encompassing materials, such as the one proposed by Papa and
historical time calls certainly for more sophisticate Taliercio (2003), is of utmost importance for the
approaches. purpose here referred.

7 MECHANICAL DECAY IN THE LONG 8 THE LASTING INFLUENCE OF THE

TERM CONSTRUCTION PROCESS

Long-term phenomena leading to progressive deteri- Important effects related to deformation can be
oration during historical periods must be accounted attributed to the construction process. The construc-
for in order to understand the existing damage. As tion of historical structures lasted during large periods
observed apropos of the study of recent collapses (Papa amounting to several decades or centuries. The con-
and Taliercio, 2003, Binda, et al. 2001, 2003), the struction included long intermediate stages during
effect of creep under constant stress, at the long term, which the structure was stabilized by means of pro-
may induce significant, cumulative damage in rock- visory supports or was forced to develop resisting
like materials. As mentioned by Binda et al. (2001), mechanisms not entirely consistent with its structural
accumulation of damage (eventually leading to col- arrangement and design. It is likely that the struc-
lapse) may occur for stress values significantly lower tures showed larger mobility during these intermediate
than the normal strength obtained by standard mono- phases due to the flexibility of the provisory sup-
tonic compression tests. The same authors found that ports and the more limited lateral confinement, so that
such phenomena could start at 40%–50% of the normal significant initial deformations were produced. This
strength value. phenomenon was amplified by the early settlement of

71
chap-06 9/9/2004 8: 6 page 72

mortar in joints and the initial creep of compressed (Fig. 10). Momentarily, the construction experienced
members. an intermediate stage consisting of the partial struc-
Historical research on Mallorca Cathedral (built tures shown in figure 11 (left). It is not clear, from
during 13th to 15th c.) has proven that the construc- the historical documents, whether the centering of the
tion of each bay followed the same process, consisting lateral vaults was removed prior to the construction of
on, first, the erection of the lateral chapels (with the central vault, or whether it was maintained dur-
their vaulting) and buttresses, and, second, the con- ing the entire operation. Removing the centering of
struction of the collateral and then the central vaults the lateral naves would have caused a very significant
inward thrust on the partial structure (Fig. 10) produc-
ing large lateral deformation. Given the characteristics
of the wooden centering frames used at the time (with
no foreseeable effective transverse ties to prevent from
causing lateral forces against the capitals of the piers,
it is likely that keeping them during the construction
of the central nave would not have saved the piers from
receiving a significant unbalanced lateral thrusts.
The numerical simulation of this stage, carried out
by means of a FEM continuum damage model (see
Cervera, 2003, for more information on the calcula-
tion technique used) has shown that equilibrium is
possible for this particular condition at the cost of a
very important deformation; however, no significant
damage is experienced by the partial structures. The
obtained lateral deformation is one order of magnitude
larger than that predicted by an instantaneous analy-
sis on the entire structure. Consistently, the subsequent
completion of the central clerestory walls and vaults in
Figure 10. Construction process and temporary unbalanced the sequential analysis (Fig. 11, right) does not cause
forces in the transverse section of a Gothic cathedral. significant additional deformation.

Figure 11. Simulation of the construction process. Intermediate (left) and final (right) configuration and accumulated
deformation.

72
chap-06 9/9/2004 8: 6 page 73

9 SIMULATION OF REPAIR AND history of construction, now experiencing much re-

STRENGTHENING ACTIONS newed interest, may also contribute with the necessary
understanding of historical construction technologies.
A model adequately validated and with demonstrated
predictive capacity should not only be exploited for REFERENCES
the assessment of the building in the past or in the
present condition, but also to analyze its response dur- Aguerri, F., Rodríguez, J. M., 1998, Director plan for the
ing and after the restoration. The models should be restoration of the catedral de Sta. María de la Huerta,
chosen and selected in a way allowing the simulation of Tarazona (in Spanish). Zaragoza: Diputación General de
the possible repair or strengthening operations and the Aragón
later response of the intervened construction. Using Binda, L., Saisi, A., Messina, S., Triangli, S., 2001, Mechan-
the model to assess the effect of a possible strengthen- ical damage due to long term behaviour of multiple leaf
pillars in Sicilian Churches. Historical Constructions,
ing technique is not only a potentiality of the model, Guimaraes: University of Minho
but also a requirement for a consistent method linking Binda, L., Anzani, A., Saisi, A., 2003, Failures due to long-
adequately investigation and restoration. Consistency term behaviour of heavy structures: the Pavia Civil Tower
also demands to assess the performance of the model and the Noto Cathedral. Structural Studies, Repairs and
representing the strengthened building by means of Maintenace of Heritage Architecture VIII, Southampton:
additional monitoring at the long-term. WIT Press
A model allowing the simulation of possible Cervera, M., 2003, Viscoelasticty and Rate-dependent Con-
strengthening techniques may also help to determine tinuum Damage Models. Pub. M79, Barcelona: Center for
the more efficient one in improving the structural sta- Numerical Methods in Engineering (CIMNE)
Croci, G., 1998, The collapses occurred in the Basilica of
bility of the building, or may contribute to choose the St. Francis of Assisi and in the Cathedral of Noto. Struc-
minimal one yet providing the necessary strengthening tural Analysis of Historical Constructions II, Barcelona:
effect. CIMNE
Domenge, J., 2004, Chronology of the different construction
phases of Mallorca Cathedral. Study of the construc-
10 CONCLUDING REMARKS tion and structural response of Mallorca Cathedral (in
Spanish). Barcelona: Universitat Politècnica de Catalunya
The role of modelling and structural analysis is found Iscarsah Committee, 2001, Recommendations for the anal-
in predicting the response of the structure on the base ysis, Conservation and Structural Restoration of Archi-
of a set of adopted hypotheses. These hypotheses are tectural Heritage. ICOMOS International Committee for
related to (1) the physical principles governing the Analysis and Restoration of Structures of Architectural
structural response and (2) the particular morpholog- Heritage
ical and material features of the building. Inspection, Jurado, F., 1995, The aqueduct of Segovia. StructuralAnalysis
of Historical Constructions I, Barcelona: CIMNE
historical research and monitoring may provide the
Lourenço, P., 1998, Experimental and numerical issues in the
empirical information needed to validate the hypothe- modelling of the mechanical behaviour of masonry. Struc-
ses – and thus the model. Evidence on reality provided tural Analysis of Historical Constructions II, Barcelona:
by the mentioned activities must be used, through a CIMNE
direct comparison with the predictions of the model, Lourenço, P., 2001,Analysis of historical constructions: From
to validate or improve it until reaching a satisfactory thrust-lines to advanced simulations. Structural Analysis
agreement. of Historical Constructions III. Guimaraes: Universidade
The purpose of using the model to assess complex do Minho
structural phenomena occurring in historical periods Macchi, G., 1998, Problems related to the original concep-
tion. The case of Pavia Cathedral. Structural Analysis of
calls for a set of demanding capabilities (see section
Historical Constructions II, Barcelona: CIMNE
5), among which the simulation of the construction Papa, P., Taliercio A., 2003, Creep modelling of masonry his-
process (and its lasting effects), long-term damage toric towers. Structural Studies, Repairs and Maintenace
processes, or the effect of persistent or extraordinary of Heritage Architecture VIII, Southampton: WIT Press
actions occurring in very long periods. These capa- Roca, P., 2001, Studies on the structure of Gothic cathe-
bilities are still difficult to mobilize in practice due drals. Structural Analysis of Historical Constructions III.
to the limitations of our understanding of complex Guimaraes: Universidade do Minho
physical and mechanical phenomena affecting ancient Roca, P., González, J. L., 2001, “Morphology, Structure and
constructions (such as long-term damage). Limited or History. The case study of the upper flying arches of
uncertain information on history and historical actions Mallorca Cathedral”. International Millenium Congress.
More than two thousand years in the history of architec-
is an obstacle as well. ture, lb 22, París: ICOMOS
However, non-stopping research effort, covering Roca, P., González, J. L., Aguerri, F., Aguerri, J. I., 2003,
both the experimental and the numerical fields, makes Monitoring of long-term damage in Gothic cathedrals.
the possibility of having powerful and realistic tools Structural Studies Repairs and Maintenace of Heritage
of analysis in a near future foreseeable. Research on Architecture VIII : Southampton: WIT Press

73
prelims-vol-1 16/9/2004 18: 40 page XIV
chap-07 9/9/2004 8: 6 page 75

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

Design approaches of interventions for the safety and conservation of

historic buildings

C. Modena
Department of Structural and Transportation Engineering, University of Padua, Italy

ABSTRACT: Unconventional design approaches of repair/strengthening interventions on historic masonry

buildings are more and more required and actually becoming popular among engineers and architects. This
implies all the aspects of the design, and first of all theoretical/experimental ways and tools for evaluating the
structural behaviour until failure of existing structures prior and after interventions are made. Primary scope of a
conservative “attitude” in this field is in fact to avoid over-designing of interventions, which of course requires to
limit as much as possible underestimations of the actual resisting capabilities of the existing structure. Important
are, then, the proper use of materials and of their application techniques. What makes very critical this point is
the fact that “new/innovative” materials are more and more proposed and actually used in structural restoration
works. Appropriate preliminary investigations are in such cases necessary in order to avoid unexpected further
problems.

1 INTRODUCTION criteria to the possibilities of intervening to “increase”

the safety.
The conservation of architectural heritage is a very Absolutely relevant outcomes of the debates raised
complex cultural operation, which was, and still is by this situation were new structural design concepts
object of debates that greatly influenced – among many and tools which are more and more entering into the
other aspects of the activity of the involved art histori- codes and then into the design practice (M.B.C.A.
ans, architects and engineers – the way how the crucial 1986 & 1996, M.LL.PP. 1996, ISCARSAH-ICOMOS
issue of “structural safety” is dealt with. 2001). It is enough to mention: the differentiation
The problem is very well known. Modern struc- of the accepted safety level for different classes of
tural safety concepts – based on the calculation of existing structures; the use of qualitative evaluation
“accepted probabilities” that something considered of structural performances (observational approach);
“wrong”, first of all the failure, happens to the “load the evaluation of safety based on pure equilibrium
bearing components” – works very well for new con- considerations; the limitation of interventions at the
structions. The properties of the used materials and minimum possible level, depending on the level of
structural components can be precisely modeled in knowledge of the structure and then on the use
this case, thus allowing “precise” (even if substantially of appropriate investigations/monitoring techniques;
conventional) evaluations of the distance between a the compatibility of (traditional/modern/innovative)
state of the structure that is considered “safe” and materials and construction techniques; the removabil-
the state of the same structure can experience in any ity of the interventions. This situation is also stimulat-
moment of its expected life. Materials and construc- ing the search for “new solutions”, i.e. new materials,
tion techniques are then selected with no substantial new techniques, new investigation tools, etc. which
restrictions during the design process in order to make can offer new possibility of better satisfying the new
such distance complying with precise code prescrip- design concepts.
tions. In the case of existing historic structures, on The design process is thus continuously evolving
the contrary, reliable and feasible structural mod- and becoming more and more complex and sophis-
els are difficultly available and in any case data on ticated. At this stage of the technical and scientific
their mechanical properties are very scarce, generally knowledge it must be necessarily based on a “case by
insufficient to feed the models; on the other hand, case” approach. The continuous learning from prac-
substantial restrictions are imposed by conservation tical applications is then essential in order to have

75
chap-07 9/9/2004 8: 6 page 76

the above concepts more and more clear and “stabi- one meter to three meters), which 3 cm thick boards
lized”, and then implemented into as much as possible are simply nailed to. Preliminary design hypotheses
generalized and controllable procedures. assumed that only a complete dismounting could have
With this aim some design experiences carried out permitted a “safe” repairing of heavily deteriorated
in Italy are presented and commented. zones of the ribs (Figs. 3 and 4) of such a “delicate
audacious” structure, being conscious however that
this solution could have caused unacceptable losses of
2 LIMITING THE INTERVENTIONS BY the existing material and components and, in general,
USING INVESTIGATIONS AND of its original character.
MONITORING

An example where the strong influence on the design

choices is very clear of having appropriate prelimi-
nary investigations and long term monitoring available
is the timber roof of the “Palazzo della Ragione” in
Padua (Figs. 1 and 2). The construction properties of
the structure are really impressive: it covers a unique,
skewed room, approximately 28 m wide and 80 m long,
and is formed by slender ribs (the section is approx-
imately 36 × 40 cm2 , but obtained by joining curved
boards 12 cm thick, whose length varies from less than

Figure 3. Detail of the deterioration of the ribs detected in

the joints.

Figure 1. External view of the roof of the “Palazzo della

Ragione” in Padua.

Figure 2. Internal view of the roof. Figure 4. Deterioration of the internal part of the ribs.

76
chap-07 9/9/2004 8: 6 page 77

A cautious and really conservative approach should In situ measurements of wind pressure (Figs. 5
have permitted to limit the repairing works to the pure and 6) on selected points were in fact compared with
substitution of the deteriorated parts of the timber ribs wind tunnel tests (Fig. 7) on a reduced scale model
avoiding any previous dismantling but also extremely of the building and of significant surrounding part of
expensive and heavy (and dangerous for the remaining the historical center of the town, to determine static
parts of the building) temporary supports. effects.
This obviously implies the possibility of relying, Dynamic characterization via in situ measurements
during the works, on the capability of the structure to of forced vibrations (obtained by a harmonic exciter
offer enough strength and stiffness, with minor “tem- applied to a rib) demonstrated that the fundamental
porary aids”, to resist with acceptable stresses and period of vibration is far from giving dynamic inter-
deformations, to the self weight and, even more crucial, action with wind (Figs. 8 and 9) (Modena et al. 1999).
to the most critical action which could cause severe
damages to such a light structure, i.e. wind.
Major efforts were for such reasons primarily ded-
icated to a clear understanding of both static and
dynamic effects of wind on the roof structure, tak-
ing into account its precise geometry and position
inside the town (for static effects) and for its real
wind/structure interaction (for dynamic effects).
Two different types of investigation were used for
the analysis of static and dynamic effects of wind.

Figure 7. Aerodynamic measures in the wind tunnel of the

reduced scale model.

Figure 5. Aerial view of the building and position of the

wind pressure transducers on the external surface of the roof.

Figure 8. Position of the accelerometers on the external

surface of the roof for dynamic acquisitions.

Figure 6. Phase of positioning of the transducers on the

roof and view of the anemometer and of the acquisition Figure 9. Positioning of the exciter on an internal rib of the
system. roof and view of an accelerometer placed outside.

77
chap-07 9/9/2004 8: 6 page 78

Figure 10. Results of the dynamic tests and comparison

with the 3D FE modeling of the roof.

Figure 13. Substitution of the deteriorated central element

of the ribs.

removing of the
metal deteriorated
support element

stone masonry
Figure 11. Lateral bracing of the ribs. element

Figure 14. Detail of the phase of substitution of part of

deteriorated ribs.
connection by
the new element 3D behaviour, and the essential contribution given to
the stability by the “shell” formed by the nailed boards,
would be in any phase maintained.
elements to The design of the local repairing was made possi-
substitute ble by taking simple temporary measures allowing for
maintaining the stable 3D behaviour and inhibiting
local buckling in any phase of the works. The first
condition was assured by simply limiting the removal
(when needed) of the boards forming the shell very
locally, in order to permit local repairing, and the
second through simple lateral support, as shown in
(Figs. 11 to 14).
A supplementary measure was taken to avoid slid-
Figure 12. Detail of the intervention on the ends of the ribs.
ing at the supports during the works. To this scope
a temporary tie was added as indicated in (Fig. 15).
In conclusion, the results of the investigations, com- The force needed to ensure stability was theoreti-
bined with appropriate structural analyses (which were cally calculated, and then monitored during the works
calibrated with the test results, as in Fig. 10), allowed (Fig. 15), together with relevant parameters of the over-
to design the intervention being much more confident all behaviour of the timber and masonry structures
on the “robustness” of the structure, provided is full (Fig. 16).

78
chap-07 9/9/2004 8: 6 page 79

Removal of about
temporary tie: 6 cm of mortar joint
Ø 24 mm bar
or steel wire Ø6 mm

sealing

Loading cell or
dinamometer for repointing
the control of the with mortar
tensile force
Figure 18. Bed joint reinforcement technique: scheme and
Figure 15. Scheme of insertion of a temporary tie and detail of the reinforcement in a brick masonry wall by
control of the tensile force. steel bars.

Figure 16. Monitoring of cracks: detail of a sensor.

Figure 19. Application of CFRP strips in an experimental

panel.

(1.55 mm) fact that the employ of traditional mortars, properly

selected, to make them work together with the exist-
Figure 17. CFRP rods (left) and thin strips (right). ing materials, is highly preferred. This is not only due
to conservation considerations, but also to the search
for reducing as much as possible high stress concen-
3 USING INNOVATIVE MATERIALS WHILE trations which could occur using high strength/high
ENSURING COMPATIBILITY AND stiffness modern bonding materials like resins, high
REMOVABILITY strength mortars etc., or, in other words, to ensure
better “compatibility” from the mechanical point
Among innovative materials highly durable stainless of view.
steel and even more FRPs (Fiber Reinforced Polymers) Specific applications of steel bars embedded super-
have high potentiality for possible application on his- ficially in the bed joints of massive structures (towers,
toric structures, due to several well-known advantages curtain walls, large pillars), proved on experimental
(for FRP especially corrosion immunity, low weight, laboratory bases, demonstrated their high efficiency
etc.). In particular, the use of the more durable CFRP in counteracting typical masonry creep damage (Binda
(Carbon FRP) bars combined with traditional mor- et al. 1999b & 2001) (Fig. 18).
tars properly selected, can be proposed for the bed More advantages can be achieved by the use of thin
joint reinforcement of masonry structures (Modena & CFRP strips (1.5 × 5 mm in section size), assuring low
Valluzzi 2003) (Fig. 17). obtrusiveness of the intervention and a better adapt-
In any case, what characterize the use of such ability to the possible joints unevenness (Valluzzi et al.
sophisticated and high performance materials is the 2003b) (Figs. 17 and 19).

79
chap-07 9/9/2004 8: 6 page 80

Current applications are by now made by stain-

less steel (one or two small diameter reinforcing
bars) connected to the inner core of the masonry by
transversal short pins, as some aspects connected to
innovative materials, especially concerning durabil-
ity, compatibility and removability, still need to deeper
experimental investigations. Both excavated joints and
drilled holes for pins insertion are successively sealed
by mortar and grout, respectively.
Applicability of such innovative materials is any-
way very promising, thus many research groups are
involved in clarifying specific problems.
The bed joint reinforcement is often used in combi-
nation with other techniques, aimed in solving specific
structural problems (injection and limited rebuilding,
to re-establish homogeneity, uniformity of strength
and continuity of masonry walls; tie-rods confine-
ment, etc.).
Some recent examples of application of the above
Figure 20. Crack pattern of the bell tower of the Cathedral
of Monza (Milan) and detail of damage on the West side. mentioned technique to towers in hazardous safety

Figure 21. Intervention on the bell-tower of the Cathedral of Monza: reinforcing tie-rings at different level and diffusion of
the bed joint reinforcement technique along the main sides of the Tower; scheme of the technique applied on the walls and
detail of the strengthening of the corners.

80
chap-07 9/9/2004 8: 6 page 81

conditions are the Cathedral of Monza and the Civic large cracks were located near the corners and com-
Tower of Vicenza (Italy). plete or partial detachment of the external leaf and of
The bell tower of the Cathedral of Monza (XVI local buckling of the masonry was found also in some
century) was suffering passing-through large verti- areas of the façades (Valluzzi et al. 2003a) (Fig. 23).
cal potentially dangerous cracks on some particularly The interventions (Fig. 23) consisted of: (i) grout
weak portions of the West and East sides (Modena injection at the base of the tower and on the belfry
et al. 2001). They were slowly but continuously open- pillars, (ii) local rebuilding, (iii) pointing of mortar
ing since 1927; moreover, wide cracks in the corners of joints, to restore the areas characterized by high mate-
the tower up top 30 m and a damaged zone at a height rial deterioration; (iv) reinforced repointing applied
of 11 to 25 m with a multitude of very thin and diffused on various portions of the walls and on the pillars, to
vertical cracks were detected (Fig. 20). counteract the creep damage and to strengthen the cor-
Design of intervention was mainly aimed in pro- ners (Fig. 24); (v) metallic horizontal reinforcing rings
viding an overall confining action of masonry walls, and anchoring ties placed at different levels along the
limiting the dilation of the material. They consisted of: height of the tower to confine the masonry and improve
(i) metallic horizontal reinforcing rings applied on sev- the connection between contiguous walls.
eral sections along the height of the Tower, to improve
the connection between the contiguous walls; (ii) the
application of the reinforced repointing technique dif-
fused on various portions of the walls, to counteract the
creep damage, and concentrated on some pilaster strips
to strengthen the corners; (iii) local interventions of
injection, rebuilding, and pointing of the mortar joints,
to restore the zones having high material deterioration
(Fig. 21).
The intervention design was strictly based on the
results of the investigations carried out on site and
in laboratory both on materials and structure (also
static and dynamic tests were useful to calibrate FE
mathematical models). They allow to identify the bed
reinforcement as a technique which, respecting as
much as possible the original structure, provides an
improvement capable to save the preservation concepts
(Fig. 22).
The Civic Tower of Vicenza (XII century) is a slen-
der structure with a base section of 6.2 × 6.5 m and
a height of about 82 m. During centuries, it was sub-
jected to several repairs and changes, due to several
causes (earthquake, bombing, etc.). The Tower suffers
a substantial out-of-plumb, and a damage character-
ized by localized deep cracks, diffused micro-cracks
and material deterioration. Many of the main deep

Figure 22. Insertion of a small diameter bar in an exca-

vated joint of the West side of the Cathedral of Monza Figure 23. Civic Tower of Vicenza (Italy): scheme of the
(Modena et al. 2001). intervention measures on the damaged sides.

81
chap-07 9/9/2004 8: 6 page 82

a)

1.20
c=a/g
1.127 b)
1.00
0.880
0.80
0.477
0.60 0.472
0.363
0.40 0.416
0.241
0.20 0.077 0.112 0.166

0.00

Overturning 1st floor

'in-plane'

Overturning 2nd floor

Overturning 3rd floor

'out-of-plane'

Arch effect in the thickness:

overturning of the springers

Overturning with tie-beam

Arch effect in the thickness:

crushing of the masonry

Overturning with ties

friction
c)

Figure 25. Damaged building after earthquake and inter-

vention measures taken after a previous seismic event (a:
ring beams at floor levels, b: superimposition of r.c. slab,
c: rough injections). Analysis with equilibrium macromod-
elling showed that lowest seismic coefficients are related to
the out-of-plane mechanisms involving the most damaged
portions of the building.

of well-connected walls and floors and a proper hori-

zontal stiffness of the floors) and on the elastic-plastic
behaviour of the masonry. Common buildings in his-
Figure 24. Reinforced repointing applied on the belfry toric areas are often realized according to typologies
pillars of the Civic Tower of Vicenza. (multi-material masonry, multi-leaf walls) and con-
structive details (poor connection between intersecting
walls, between walls and floors and even among the
4 APPLYING EQUILIBRIUM ANALYSES FOR layers in the thickness), which can evidence fundamen-
VULNERABILITY ASSESSMENTS AND tal deficiencies for the stability and the safety under
STRUCTURAL IMPROVEMENT IN seismic actions.
SEISMIC AREA The ultimate capacity of the building depends on the
stability of its macro-elements, that is of portions of
A complete limit analysis approach of existing the structure bounded by the potential damage pattern
masonry buildings, especially in seismic areas, is prov- (cracks, borders of poor connections, etc.) which can
ing to be the most efficient way for making reliable behave as a whole, following a kinematics mechanism.
safety evaluations without underestimating their actual Macro-elements are defined by single or combined
resisting capabilities, thus avoiding over-designing structural components (walls, floors and roof), con-
repair/strengthening interventions. sidering their mutual bond and restraints (e.g. the
Such method, based on single or combined kine- presence of ties or ring beams), the constructive defi-
matics models involving the equilibrium of struc- ciencies and the characteristics of the constitutive
tural macro-elements (Giuffrè 1993), is particularly materials. Once the critical structural configuration is
addressed to existing masonry buildings in historic defined, the subsequent step is the identification of the
centres. They, in fact, often do not satisfy the gen- most probable collapse mechanism/s characterizing
eral conditions which allow the application of com- the macro-element.
mon equivalent static procedures, based on the “box” Several studies based on the in-situ observations
behaviour of the structure (which requires the presence after seismic events allowed to systematize abacuses of

82
chap-07 9/9/2004 8: 6 page 83

the typical damages occurring in constructive typolo- On: Seismic Damage to Masonry Buildings, Bernardini
gies (buildings, churches), which led to the consequent Ed., Balkema, Rotterdam.
systematization of the mechanical models able to Binda L., Modena C., Valluzzi M.R. & Zago R. 1999b.
describe their specific behaviour by kinematics mod- Mechanical effects of bed joint steel reinforcement in
historic brick masonry structures, Structural Faults +
els, both for in-plane and out-of-plane mechanisms Repair – 99, 8th International Conference and Exhibition,
(Giuffrè & Carocci 1999, Binda et al. 1999a, Penazzi London, England, 11 pp. on CD-ROM.
et al. 2001). Giuffrè A. 1993. Sicurezza e conservazione dei centri storici.
Those models, calibrated on the real damaged sites, Il caso Ortigia, Laterza, Bari.
are usefully applied for analyses of vulnerability for Giuffrè A. & Carrocci C. 1999. Codice di pratica per la
centres under seismic hazard, in order to examine the sicurezza e la conservazione del centro storico di Palermo,
current condition and to prevent their future damage. Laterza, Bari.
Moreover, the simulation of possible interventions can ISCARSAH-ICOMOS 2001. Recommendations for the
be performed, both in damaged and undamaged con- Analysis, Conservation and Structural Restoration of
Architectural Heritage.
ditions, evaluating their impact with the pre-existing M.B.C.A. 1986. Circ. 1032 18/7/1986: Raccomandazioni
situation (Avorio et al. 2002, Valluzzi et al. 2004). relative agli interventi sul patrimonio monumentale a
tipologia specialistica in zona sismica.
M.B.C.A. 1996. Istruzioni generali per la redazione dei pro-
5 CONCLUSIONS getti di restauro nei beni architettonici di valore storico
artistico in zona sismica.
Some case studies have been presented and com- M.LL.PP. 1996. D.M. 16/1/1996: Norme tecniche per le
mented where efforts have been made to respond costruzioni in zone sismiche.
to “conservative” design criteria while intervening Modena C. & Valluzzi M.R. 2003. Repair techniques for
creep and long term damage of massive structures, Proc. of
to ensure acceptable structural safety conditions of
the VIII STREMAH Conf., Structural Studies, Repairs and
existing historic constructions. Maintenance of Heritage architecture, Haldikiki, Greece,
As it is first of all expected in order to respect 7–9 May 2003, C.A. Brebbia Ed., WitPress, UK, pp.
the existing features of the considered constructions, 141–150.
attention has been focused on avoiding unnecessary Modena C., Valluzzi M.R., Tongini Folli R. & Binda L. 2001.
interventions and special care has been paid in order to Design choices and intervention techniques for repairing
limit in any case as much as possible variations not only and strengthening of the Monza cathedral bell-tower, 9th
of its external appearance, but also of its mechanical Structural Faults & Repair, London, UK, 4–6 July 2001,
behavior. 12 pp., on CD-ROM.
Modena C., Zonta D. & Riolfo A. 1999. Dynamic investiga-
This requires that special efforts are made to accu-
tion on the Palazzo della Ragione roof in Padua, Structural
rately analyze, theoretically and experimentally, the Faults + Repair – 99, 8th International Conference and
actual resisting properties of the considered con- Exhibition, London, England, July, 13–15, 1999, 10 pp.,
structions, prior and after interventions are made, in on CD-ROM.
order to avoid over-designing approaches. Moreover, Penazzi D., Valluzzi M.R., Saisi A., Binda L. & Modena C.
the actual contribution of any traditional/innovative 2001. Repair and strengthening of historic masonry build-
material and technique, and of their possible com- ing in seismic area, International Millennium Congress
binations, can be adequately and scientifically ‘More than two thousand years in the history of archi-
exploited in order to ensure durability, compatibility tecture safeguarding the structure of our architectural
heritage’, Bethlehem (Palestine), Vol. 2, Section V (7 pp.).
and possibly removability of repair/strengthening
Valluzzi M.R., da Porto F. & Modena C. 2003a. Structural
interventions. investigations and strengthening of the civic tower in
Vicenza, Structural Faults & Repair – 2003, Common-
wealth Institute, Kensington, London, UK, 1–3 July 2003,
REFERENCES 10 pp., on CD-ROM.
Valluzzi M.R., Tinazzi D. & Modena C. 2003b. Strength-
Avorio A., Borri A., Corradi M. 2002. Ricerche per la ening of masonry structures under compressive loads
ricostruzione. Iniziative di carattere tecnico e scientifico a by using FRP strips, Proc. 6th International Sympo-
supporto della ricostruzione, Regione dell’Umbria, DEI, sium on Fibre-Reinforced Polymer (FRP) Reinforcement
Roma. for Concrete Structures (FRPRCS-6), 8–10 July 2003,
Binda L., Modena C., Saisi A., Tongini Folli R. & Valluzzi National University of Singapore, Singapore, Vol. II, pp.
M.R. 2001. Bed joints structural repointing of historic 1249–1258.
masonry structures, 9th Canadian Masonry Symposium Valluzzi M.R., Cardani G., Binda L. & Modena C. 2004.
‘Spanning the centuries’, Fredericton, New Brunswick, Analysis of the seismic vulnerability of masonry build-
Canada, 12 pp., on CD-ROM. ings in historical centres and intervention proposals, 6th
Binda L., Gambarotta L., Lagomarsino S. & Modena C. International Symposium on the Conservation of Monu-
1999a. A multilevel approach to the damage assessment ments in the Mediterranean Basin, Lisbon, Portugal, 7–10
and seismic improvement of masonry buildings in Italy. April 2004, pp. 561–565 (on CD-ROM).

83
prelims-vol-1 16/9/2004 18: 40 page XIV
chap-08 9/9/2004 8: 6 page 85

Historical aspects and general methodology

prelims-vol-1 16/9/2004 18: 40 page XIV
chap-08 9/9/2004 8: 6 page 87

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

Ancient binding materials, mortars and concrete technology:

history and durability aspects

Ö. Kırca
ÇimSA Cement Production and Trading Company, Mersin, Turkey

ABSTRACT: This paper deals with the history of technology related with binding materials. Lime, gypsum,
and hydraulic binders, i.e. lime-pozzolan mixture have been the major types of binding materials survived in
different regions of the world for 8000 years. Whatever type of binding materials the ancient civilizations used,
it can be seen that those structures built by ancient binders, particularly lime-pozzolan mixture have survived
for several hundred years. It can be claimed that even at that time architects were aware of the importance of
the durability of binding materials. However, those structures belonging to modern civilizations and made by
using modern binders, i.e. portland cement have experienced significant deterioration throughout their service
life, generally lesser than 100 years.
In this research, the object of the study is to investigate the historical timeline of binding materials and to
focus on their durability aspects by comparing the traditional technologies with modern ones.

1 INTRODUCTION to make hydraulic binder, i.e. lime-pozzolan cement

by adding materials, such as volcanic ash or powdered
Binding material or cementing material in the gen- bricks, tiles and pottery to lime.
eral sense of the world can be described as a material The modern cement, i.e. portland cement was
with adhesive and cohesive properties, which make it invented by Joseph Aspdin in England in 1824. After
capable of bonding mineral fragments into a compact that time, modern Portland cement and concrete tech-
whole. For constructional purposes, the meaning of nology has proceeded till today by invention of new
the term “cement” is restricted to the bonding materi- chemical and mineral admixtures and additions.
als used with stones, sand, building blocks, etc (Neville Although radical advances in cement and concrete
1989). technology have been observed for two centuries,
The use of cementing materials is very old. The there are still problems related with the endurance of
first evidence of its existence dates back to 12 000 cement and concrete. Many modern structures expe-
BC in Israel. Reactions between limestone and oil rienced significant deterioration within their service
shale during spontaneous combustion occurred in life, which is generally lesser than 100 years.Yet, many
Israel to form a natural deposit of cement compounds. ancient Roman concrete buildings or structures are still
Another example of ancient binder from the history in use after more than 2000 years. According to the
of the human being is the use of naturally occurring Vitruvius’s book De Architectura, the magnificent
bitumen by the Babylonians and Assyrians in their quality of Roman concrete resulted from the exten-
brick and gypsum plaster construction. The Egyptians sive use of artificial pozzolanic mortars and concretes
improved the technology of lime and gypsum mortar (Vitruvius 1960). In addition, placement, compaction,
and builded the pyramids by the use of such mortars. workability properties of ancient Roman concrete
The Greeks made further improvements and finally the contributed to its magnificent survivability.
Romans developed a cement that produced structures
of remarkable durability and that can set and harden
even under water. 2 CHRONOLOGICAL SUMMARY OF
Lime was known to the Greeks and was widely used BINDING MATERIALS AND MORTARS
by the Romans. The Roman architect and engineer
Vitruvius published the first specification for the use The nature and usefulness of cement as a kind of
of lime in building in his celebrated work De Archi- artificial stone was probably first understood by the
tectura (Vitruvius 1960). The Romans also knew how Romans; but it seems likely that they copied the

87
chap-08 9/9/2004 8: 6 page 88

technology from Etruscans and Greeks and improved in order to be used in construction it must be mixed
it drastically. with water to form a lime paste, which is defined as
Concrete and/or cement itself is actually a phe- “hydration” or “slaking”.
nomenon of nature, with the first evidence of its To produce dry powdered hydrated lime just suf-
existence in Israel dating back to 12 000 BC, when nat- ficient water is added for the quicklime lumps to
ural deposits of cement compounds were said to have break down to a fine powder. This material would
formed due to reactions between limestone and oil have a “shelf life” of only a number of weeks, depend-
shale, employing spontaneous combustion. However, ing on storage conditions. “Old” hydrated lime would
the earliest known concrete produced intentionally by have partially carbonated and become a less effective
human beings was discovered in the floor of a crude binder. If too little or too much water is added, the prop-
shelter built about 5600 BC on the banks of the river erties of the slaked lime will be spoiled; the slaked lime
Donube in Yugoslavia. The floor was made from mix- will not harden and the paste will not be as plastic as
ture of sand, gravel and red lime. That concrete was it should be (Erdoğan 2002).
manufactured deliberately, rather than unintentionally However, if quicklime is hydrated with a proper
by utilizing the red lime from a site 320 km away. For amount of water and well agitated, it forms a milky
some reason, it appears that the technique for using suspension known as milk of lime, which can be easily
lime-based cements was lost for at least 2500 years, used in building applications.
until indications of use by the ancient Assyrians and Limestone containing a proportion of clay is
Babylonian civilizations around 3000 BC. They used often seen as an advantage in building as they pro-
bitumen to bind stones and bricks. About the same duce hydraulic limes. In fact, limes do not possess
time Egyptians used mud bricks mixed with straw to hydraulicity, since it needs carbon dioxide for its hard-
bind dried bricks and also this phenomenon furthered ening reaction (Erdoğan 2002). However, in the case
the discovery of lime and gypsum mortars as a binding of hydraulic limes, as the name implies, they can set
agent for building the pyramids. Even at that time, the and harden even under water and will produce stronger
ancient Chinese people used cementitious materials to mortars.
hold bamboo together in their boats and in the Great In the construction industry, usually lime, in its
Wall, one of the Wonders of the world. By 500 BC hydrated or putty form, is mixed with aggregate and
the art of making concrete had spread to the Mediter- water to produce concrete or mortar for different
ranean island of Crete, and from there to the ancient purposes such as plastering sand-lime brick produc-
Etruscans and Greeks. tion etc. Plain lime-sand mortars are quite weak;
Yet, it was the Romans who brought the manufacture any early adhesive strength results from drying out
and use of lime-based cement to an art form. 300 BC whereas longer-term hardening occurs as a result of
saw the Romans employing slaked lime and volcanic carbonation of the lime.
ash called pozzolan, named after the town of Pozzuoli, Traditional lime plasters were often mixed with ani-
near Mt Vesuvius. This was a hydraulic cement that mal hair to improve cohesion and adhesion. Today
can set harden when mixed with water in air as well addition of gypsum or portland cement and/or poz-
as under the water. Moreover, some natural additions zolans to increase durability and give faster setting
such as animal fats, milk, and blood were also used times is more common.
throughout this era.
After the fall of the Roman Empire, knowledge of
cement was lost till the 18th century. 3.2 Ancient lime technology
The whole historical timeline of cement and con-
Quarrying: The first step of lime production required
crete is summarised in Table 1, where the develop-
finding a suitable raw material which was generally
ments in modern era are also outlined.
calcium carbonate based stones such as limestone
and marble. Quarrying techniques had become well
advanced during construction of the great Egyptian
3 TECHNOLOGY OF ANCIENT LIME
pyramids and further advanced during the Roman
AND LIME MORTARS
building era. The determination of raw material
involved trial and error, as well as ages of experi-
3.1 Properties of lime
mental attempts. Once workable quarries and materials
There are two forms of lime: quicklime and hydrated were located, workmen acquired skills for identifying
lime. and extracting suitable calcium carbonate raw mate-
Quicklime is produced by heating rock or stone con- rials. Then such materials had to be transported to
taining calcium carbonate (limestone, marble, chalk, sites where they were prepared for calcination. The
shells, etc.) to a temperature of around 900◦ C for sev- major technological factors surrounding acquisition
eral hours in a process known as ‘calcining’. It is an of the raw carbonate material, therefore, involved,
unstable and slightly hazardous product. Therefore, identification of suitable stone, workable extraction,

88
chap-08 9/9/2004 8: 6 page 89

Table 1. Historical timeline of binding materials.

12 000 BC Reactions between limestone and oil shale during spontaneous combustion occurred in Israel to
form a natural deposit of cement compounds.
5600 BC Intentionally production of concrete from sand, gravel, and lime a natural concrete in Yugoslavia.
3000 BC Bitumen to bind bricks and stones was used by Babylonians and Assyrians.
3000 BC Egyptians used mud mixed with straw to bind dried bricks. They also used lime and gypsum
mortars in the pyramids.
3000 BC Chinese used cementitious material to hold bamboo together in their boats and in the Great Wall.
500 BC The art of making concrete had spread to the Mediterranean island of Crete, and from there to the
ancient Etruscans and Greeks.
300 BC–476 AD Used pozzolan cement from Pozzuoli, Italy near Mt. Vesuvius to build several structures. They
used lime as a cementitious material also Animal fat, milk, and blood were used as admixtures.
400–1779 The art of concrete was lost after the fall of the Roman Empire.
1779 Bry Higgins was issued a patent for hydraulic cement (stucco) for exterior plastering use.
1793 John Smeaton found that the calcination of limestone containing clay gave a lime, which
hardened under water (hydraulic lime).
1796 James Parker from England patented a natural hydraulic cement by calcining limestone-containing
clay, called Parker’s cement or Roman cement.
1812–1813 Louis Vicat of France prepared artificial hydraulic lime by calcining synthetic mixtures of
limestone and clay.
1818 Maurice St. Leger was issued patents for hydraulic cement. Natural cement was produced in the USA.
Natural cement is limestone that naturally has the appropriate amounts of clay to make the same
type of concrete as John Smeaton discovered.
1822 James Frost of England prepared artificial hydraulic lime likes Vicat’s and called it British cement.
1824 Joseph Aspdin of England invented portland cement by burning finely ground chalk with finely
divided clay in a lime kiln until carbon dioxide was driven off. The sintered product was then
ground and he called it portland cement named after the high quality building stones quarries
at Portland, England.
1828 I.K. Brunel is credited with the first engineering application of portland cement, which was used
to fill a breach in the Thames Tunnel.
1843 J.M. Mauder, Son & Co. Were licensed to produce patented portland cement
1845 Isaac Johnson claims to have burned the raw materials of portland cement to clinkering temperatures
1849 Pettenkofer & Fuches performed the first accurate chemical analysis of portland cement.
1860 The beginning of the era of portland cements of modern composition.
1862 Blake stone breaker of England introduced the jawbreakers to crush clinker.
1867 Joseph Monier of France reinforced William Wand’s (USA) flowerpots with wire ushering in
the idea of iron reinforcing bars (re-bar).
1886 The first rotary kiln was introduced in England to replace the vertical shaft kilns.
1887 Henri Le Chatelier of France established oxide ratios to prepare the proper amount of lime to
produce portland cement. He proposed that hardening is caused by the formation of crystalline
products of the reaction between cement and water.
1889 The first concrete reinforced bridge is built.
1890 The addition of gypsum when grinding clinker to act as a retardant than the
setting of concrete was introduced in the USA. Vertical shaft kilns were replaced
with rotary kilns and ball mills were used for grinding cement.
1891 George Bartholomew placed the first concrete street in the USA in Bellefontaine, OH. It still exists today.
1893 William Michaelis claimed that hydrated metasilicates form a gelatinous mass (gel) that dehydrates
over time to harden.
1900 Basic cement tests were standardized.
1903 The first concrete high rise was built in Cincinnati, OH.
1909 Thomas Edison was issued a patent for rotary kilns.

(continued)

89
chap-08 9/9/2004 8: 6 page 90

Table 1. (continued)

1930 Air entraining agents were introduced to improve concrete’s resistance to freeze-thaw damage.
1936 The first major concrete dams, Hoover Dam and Grand Coulee Dam were built. They still exist today.
1967 First concrete domed sport structure, the Assembly Hall, was constructed
at the University of Illinois, at Urbana-Champaign.
1970’s Fibre reinforced in concrete was introduced.
1980’s Superplasticizers were introduced as admixtures.
1985 Silica fume was introduced as pozzolanic additive. The highest strength concrete was used in building
the Union Plaza constructed in Seattle, Washington.

quarrying methods, and suitable means of transporta- equal in weight to 1/3 of the quicklime was sprinkled
tion to a site for further treatment (Krumnacher over it. Heat was given off, the material cracked open
2001). and became a powdered, and increased in volume. This
Calcination: In order to convert the calcium carbon- method was considered the best. (2) Immersion: the
ate based raw material into quicklime, it was necessary quicklime, placed in a basket, was lowered into the
to heat the mass of stone or shell up to required temper- water for the proper length of time, and drawn up to
atures. By calcining the raw materials, the water and complete the slaking. The handling was tricky, at best,
carbon dioxide was driven off from within the stone. and demanded considerable skill. (3) Air-slaking: the
Calcining of raw materials took place in either kilns quicklime was simply left exposed, to pick up moisture
or clamps customarily constructed of stone or brick from air (Krumnacher 2001, McKee 1971).
with clamps sometimes representing an open heap of The slaking procedure, like all other steps involving
carbonate material and fuel (Krumnacher 2001). the preparation of mortar had a great influence upon
“When the charge of calcium carbonate undergoes the resultant properties and performance of finished
firing it transforms to a calcium oxide known as quick- mortar.
lime. This transformation involves the release of water Mortar Mixing: In order to create a mortar for unit-
and carbon dioxide, which are driven off after being ing brick or stone, workers introduced slaked lime to
heated through the calcining process. Through trial a sand aggregate. This practice involved proportion-
and error lime burners found that the raw limestone ing the amount of lime, sand and water. According to
fired more completely if broken into pieces about the McKee, in Roman period, three primary methods of
size of two closed fists. If the stones were too large mortar mixing predominated: (1) mixing dry slaked
there would remain unaltered clinkers, which would lime powder, sand, and water; (2) mixing wet slaked-
be found during the subsequent process.These clinkers lime paste and sand, adding water if needed; (3) mixing
would jeopardize the consistency and effectiveness of pulverized dry quicklime, sand and water, using the
a resulting lime. Since quicklime was difficult to store mortar while it was still hot (McKee 1971).
and transport, clamps and kilns were set up near the Fat lime or high calcium lime could be mixed with
site where lime was used and the raw material would the aggregate and allowed to age, provided it was
be transported to the kiln site” (Krumnacher 2001). protected from exposure to the air. Exposure to the
“Constant attention and responsive orchestration of air would act to recarbonate or solidify the mass.
the firing were essential. Lime burners worked unusual Many ancient artisans practiced this custom of storing
hours and were highly skilled, although not always lime putty or lime mortar. Plasterers would keep their
highly respected. In Roman times, inmates and crim- lime in sealed vats for considerable lengths of time.
inals would often be sent to provide labour at the Nicholson describes the mixing and aging of mortar
limekilns” (Krumnacher 2001). as follows (Nicholson 1850): “The mortar should be
Slaking: Quicklime intended for use in construc- made underground, then covered up, and kept for a
tion must first be mixed with water, to form a lime considerable length of time, the longer the better; and
paste, by being slaked (Erdoğan 2002). All calcin- when it is to be used, it should be beat up afresh. This
ing process of calcium carbonate based stones had makes it set sooner, renders it less liable to crack and
to be followed by slaking, which was necessary for harder when dry... How very different was the prac-
construction purposes. tice of the Roman! The lime which they employed
Throughout the Roman period, there were three was perfectly burnt, the sand sharp, clean, and large
different hydration methods for transformation of grained; when these ingredients were mixed in due
quicklime into the hydrated (or slaked) lime by the con- proportions, with a small quantity of water, the mass
trolled addition of water. One of the three methods was was put into a wooden mortar pan, and beaten with
followed: (1) sprinkling or ‘drowning’: ideally, water a heavy wooden or iron pestle, till the composition

90
chap-08 9/9/2004 8: 6 page 91

adhered to the mortar: being thus far prepared, they highly esteemed and respected, the techniques of
kept it until it was at least three years old. The beating preparation and application of lime and lime mortars
of mortar is of the outmost consequence to its durabil- have remained less than fully understood” (Krum-
ity, and it would appear that the effect produced by it, nacher 2001).
is owing to something more than a mere mechanical Vitruvius, a 1st century B.C. Roman architect under
mixture” (Krumnacher 2001, Nicholson 1850). Augustus, conveys the earliest comprehensive details
regarding lime treatment and its use as a building
material. Craft tradition was advanced by the Romans,
3.3 Ancient lime applications who developed a strong infrastructure of labour, raw
material, technological order and transportation. This
In order to understand the techniques associated with
craft tradition and supportive infrastructure facili-
lime mortars it is necessary to review the lineage from
tated a sophisticated building program (Krumnacher
which they arose. While the actual techniques involv-
2001).
ing preparation and application of lime remains less
than fully understood, examples of finished materials
still in existence testify on behalf of the early success
acquired by those working with lime. According to 4 TECHNOLOGY OF ANCIENT HYDRAULIC
Lazell, the picture of the early development of lime BINDER (LIME-POZZOLAN MIXTURE)
technology is as follows (Lazell 1915): “The art of AND ITS MORTAR
using mortar in some from or other is as old as the
art of building or as civilization itself. Evidences of Throughout the history of human beings, for the first
the use of mortar are found not only in the older coun- time, a hydraulic binder was used during the era of
tries of Europe,Asia, andAfrica, but also in the ruins of Roman Empire. This hydraulic binder has been called
Mexico and Peru.The remains of the work of these arti- as Roman cement or concrete, as well. The techniques
sans are evidence to us of the enduring qualities of lime used by architects and artisans for the production of
mortar as well as the skill and knowledge possessed binding materials at the time of the Roman Empire
by the user” (Krumnacher 2001, Lazell 1915). were substantially based upon those described by
Lazell additionally describes ancient era as fol- Vitruvius in his splendid handbook De Architectura
lows (Lazell 1915): “Plastering is one of the earliest (Vitruvius 1960).
instances of man’s power of inductive reasoning, for However, as Vitruvius says, the Roman art of build-
when men built they plastered; at first like the birds ing was the heritage of the Greek culture. On the other
and beavers, with mud; but they soon found out a more hand, the construction technique know-how, includ-
lasting and more comfortable method, and the earli- ing building materials, came from the Egyptians and
est efforts of civilization were directed to plastering. Persians through a slow transmission process of the
The inquiry into it takes us back to the dawn of social original culture (particularly from East Iran) through
life until its origin becomes mythic and prehistoric. the Mesopotamian and the Mediterranean civilization
In that dim, obscure period we cannot penetrate far (Collepardi 1997).
enough to see clearly, but the most distant glimpses In the previous part, the primitive binder, lime, and
we can obtain into it shows us that man had very early its history of technology were examined. However,
attained almost to perfection in compounding material from now on, the technology of the Roman concrete,
for plastering. In fact, so far as we yet know, some of which is very similar to the modern concrete are going
the earliest plastering which remains to us excels, in to be investigated. In fact, the fact that ancient Roman
its scientific composition, that which we use at the concrete has withstood the attack by elements for
present day, telling of ages of experimental attempts. 2000 years, while the modern one undergoes prema-
The pyramids of Egypt contained plaster work exe- ture deterioration within its service life lesser than 100
cuted at least 4000 years ago, and this, where wilful years, shows that the basic construction techniques
violence has not disturbed it, still exist in perfection, of the Romans must be better than those of modern
out vying in durability the very rock it covers, where practice as judged by comparing the products (Moore
this is not protected by its shield of plaster”. 1993).
“Development of masonry construction technol- Mortar consisting of lime and sand has been used
ogy emerged and spread through empirical trial and as an integral part of masonry structures for thousands
error. Within the 1st century B.C. Romans borrowed of years. It also was applied to protect the earthen
and applied Greek and Etruscan techniques of lime walls of ancient houses. When mixed with volcanic
preparation and applications for creating their mor- ash (pozzolan), it becomes a mortar that the Romans
tar. Roman use of lime mortars reflected a keen used to build rock walls. The Romans later discovered
insight into the material and its technological prepara- how to mix this mortar with small stones, bricks, and
tion and application toward their momentous building other materials to produce concrete called as ‘Roman
accomplishment.Although these accomplishments are Concrete’ (Moore 1993).

91
chap-08 9/9/2004 8: 6 page 92

The constituents of Roman Concrete, aggregate and shales possess pozzolanic properties when they are in
pozzolan, are going to be described in the following a finely divided form (Erdoğan 1997).
parts, except the lime, which was discussed previously. In fact, the name pozzolan comes from the town of
Pozzouli in the foothills of Mount Vesuvius, in Italy,
where the ancient Romans had produced a hydraulic
4.1 Aggregates binder by mixing lime with volcanic soil more than
2000 years ago (Erdoğan 1997).
Stone aggregate is a necessary part of the concrete Likewise, Vitruvius describes pozzolan as a kind of
because mortar by itself cannot sufficiently resist the powder, which from natural causes produces astonish-
crushing force of great weights. Aggregate gives con- ing results. He says in his handbook De Architectura
crete its necessary structural body by increasing the that this substance, when mixed with lime and rubble,
density of mass. It also reduces the amount of lime not only gives strength to buildings of other kinds, but
that must be burnt and slaked. The Romans used sev- even when it is used in the sea, they can set and harden
eral kinds of stone for aggregates, ranging in weight under water.
from selce, a very heavy lava stone used in founda- By the day the Romans had mastered the use of
tion walls, to lightweight tufa (a local granular stone) pozzolan, which they added to the dry mix in lieu of
and pumice, both used in vaults. Other kinds of stone, part of the sand. Pozzolan is a friable volcanic mate-
as well as broken bricks and tiles were also used. rial, found in thick beds of chunks and gravel-sized
All of these materials were found in and near Rome. pieces in Latium and Campania and easily reduced
Architectural sculptures and other stone members of to usable form. It often has a distinct reddish or yel-
demolished buildings were sometimes broken up and lowish hue and has the property of forming hydraulic
used as aggregate. In many structures different aggre- silicates in combination with lime, quartz sand, and
gates were used according to the loads to be carried. water. The importance of pozzolan can be exaggerated,
Certain vaults of the Flavian Amphitheatre contain for some large Roman concrete buildings were built
pumice, and the five kinds of aggregate in the walls without it, but mortar made with it set readily under
and dome of the Pantheon were separated into five water, an advantage Roman engineers made good use
horizontal zones, each containing a lighter aggregate of. The architects and builders of the high empire must
than the one below. In foundations the aggregate often have been convinced that pozzolan improved their con-
occupies two thirds of the total volume of the fabric crete, for it was used in almost all buildings and rarely
(Singer et al. 1965). absent from high imperial construction in Rome and
Aggregate was added to the mortar during the actual its environs (Singer et al. 1965).
construction of the wall or vault. Often the stones were
spaced at random, having been dumped into the forms
and left untouched, though sometimes they were raked 6 DURABILITY OF ANCIENT MORTARS
out. In Trajanic and Hadrianic work they were fre- AND CONCRETE: COMPARISON WITH
quently laid by hand in regular rows, spaced evenly in MODERN CONCRETE
the mortar both horizontally and vertically in order to
distribute evenly resistance to load. The stone is such Concrete experts talk today about how to make con-
cases is usually a yellowish tufa, roughly shaped into crete durable. Many ancient concrete and buildings
loaf forms about half-foot in length. Such aggregate are still in use after more than 2000 years. For these
takes up about half the volume of the concrete of walls modern concrete experts, the Romans were fortunate
and vaults. It is always completely surround by mor- builders in that they apparently simply used natural
tar, with the exception of the exposed irregular side pozzolan deposits, which were found to be suitable
of pieces laid against wooden shuttering or formwork for producing a hydraulic mortar. Contrary to this
(Singer et al. 1965). pronouncement, Vitruvius states in his book De Archi-
tectura that the magnificent quality of Roman concrete
resulted from the extensive use of artificial pozzolanic
5 POZZOLAN mortars and concretes. Two artificial pozzolans were
intensively used: (1) calcined kaolinitic clay, in Latin
Natural pozzolans are the naturally occurring siliceous testa; (2) calcined volcanic stones, in Latin carbuncu-
or siliceous and aluminous materials which in them- lus. In addition to artificial reactive ingredients, the
selves possess little or no cementitious value but Romans used a natural reactive volcanic sand named
will, in finely divided form and in the presence of harena fossicia. The ingredients testa, carbunculus
moisture, chemically react with calcium hydroxide at and harena fossicia were intensively used in Roman
ordinary temperatures to form compounds possess- buildings (Vitruvius 1960).
ing cementitious properties. Volcanic glasses, volcanic Dusty ancient history books taught us that Roman
tuffs, trusses, diatomaceous earths and some clays and concrete consisted of just three parts: a pasty, hydrated

92
chap-08 9/9/2004 8: 6 page 93

lime; artificial or natural pozzolan; and a few pieces concrete product called Roller Compacted Concrete
of fist-sized rock. If these parts were mixed together in (RCC) had been crudely developed, U.S. Bureau of
the manner of modern concrete and placed in a struc- Reclamation’s refinements made it an economical can-
ture, the result certainly would not pass the test of the didate for dam construction. In the construction of a
ages (Moore 1993). How did those Romans around dam in Utah, USA, a concrete consisting of a binding
the time of Christ build such elaborate, ageless struc- mixture of 40% Portland cement and 60% fly ash. By
tures in concrete is the question on which the durability coincidence, the fly ash contained the same amorphous
of ancient concrete and its comparison with modern silica compounds as the ash from explosive volcanoes.
one will be discussed in the following part of this And the hydrated portland cement released the cal-
investigation. cium component recognized in the lime part of the
“A most unusual Roman structure depicting their ancient concrete formula. When Reclamation mixed
technical advancement is the Pantheon; a brick faced these two parts for their dam, a bonding gel was formed
building that has withstood the ravages of weathering to tie inert rock pieces of the hatch together. These
in near perfect condition. Solving the riddle of ancient rocks were used as a strong filler material much in the
concrete consisted of two studies: one was understand- same manner as is used in standard concrete practices”
ing the chemistry, and the other was determining the (Moore 1993). So, a close relationship between the
placement of ancient concrete. To understand its chem- calcium hydroxide molecules from Portland cement
ical composition, going back in time much before and that of the ancient wet lime, and between the
Moses is essential. People of the Middle East made amorphous silica of the pozzolanic fly ash and the
walls for their fortifications and homes by pound- amorphous silica of the volcanic pozzolan can be con-
ing moist clay between forms, often called pise work. structed. As a result, such a reasonable relationship
To protect the surfaces of the clay from erosion, the for the concrete components that make the gel for
ancients discovered that a moist coating of thin, white, both modern and ancient concrete can be useful for
burnt limestone would chemically combine with the comparison purposes (Moore 1993).
gases in the air to give a hard protecting shield. Although the similarity of the ingredients of modern
As a guess, the event of discovering pseudo con- and ancient concrete has been explained, there is more.
crete occurred some 200 years before Christ when a Studies of the placement process are very important in
lime coating was applied to a wall made of volcanic, making durable concrete. The Bureau of Reclamation
pozzolanic ash near the town of Pozzuoli in Italy” mixed their components (cement, ash, and rock) with
(Moore 1993). as little water as possible to give a stiff, ‘no slump’
A chemical reaction took place between the vol- concrete; spread it in layers on the dam; and pounded it
canic ash containing silica and small amounts of alu- into place by large vibrating rollers to make a new class
mina and iron oxide and the lime (calcium hydroxide). of concrete. The ancients had mixed their components
Later Romans noticed that mixing finely ground vol- (wet lime and volcanic ash) in a mortar box with very
canic ash with lime made a thicker coat, but it also pro- little water to give a nearly dry composition; carried
duced a durable product that could be set and harden it to the jobsite in baskets placing it over pre-placed
even under water. In order to understand the difference layer of rock pieces; and then proceeded to pound the
between lime and lime-pozzolan mixtures, the chem- mortar into the rock layer. Vitruvius mentioned this
ical reaction have to be examined separately. When process in his history formulas for his concrete, plus
only lime is used as a binding agent within lime-sand- the fact that special tamping tools were used to build a
water mixture, the strength development of the mix cistern wall. Close packing of the molecular structure
occurs as a result of carbonation, whereas in the case of by tamping reduced the need of excess water, which
lime-pozzolan mixture strength development is based is a source of voids and weakness. But also packing
upon the reaction between hydrated lime and finely produces more bonding gel than might be normally
ground amorphous silica. There is no need to carbon expected (Vitruvius 1960, Moore 1993).
dioxide in the case of lime-pozzolan mixture, which Throughout the above paragraphs, the ancient con-
means that it possess hydraulicity. It is also found that crete was compared with one type of modern concrete,
parts of the complex chemistry of the lime-pozzolan RCC, which was invented few years ago and whose
bonding gel matching the same chemical formula of applications are restricted overall the world. The other
modern concrete bonding gel, which is called as C-S-H modern concrete applications generally suffer from
or tobermoritte gel in modern concrete world. And this premature deterioration within their service or design
is so; the pozzolan-wet lime gel gave the high perfor- life. Some of the reasons for that are as follows:
mance to the ancient concrete. The placement technic (1) modern cement is too active, thus the rate of
used in ancient concrete was the other important point hydration reaction, which is exothermic, is too high;
causing high performance (Moore 1993). and thus modern concrete becomes more prone to
“Chemistry alone will not make good concrete; thermal cracking and thus more permeable. (2) Expec-
rather people will make good concrete.Although a new tation of high strength performance from concrete

93
chap-08 9/9/2004 8: 6 page 94

induces use of higher dosage of cement or use of finer “www.portcement.org/cem/cementindustry_

cement, which causes same result of (1). (3) Speed history.asp”
of construction induces use of high early strength “realtytimes.com/rtnews/rtcpages/20000128_
cement and concrete or use of speeded curing methods concrete.htm”
(e.g. heat curing), and thus causes again the same “www.romanconcrete.com”
result of (1). “www.qcl.com.au”
This is sincerely appreciated by the author of this
article.
7 CONCLUSION

The history of Portland cement, modern binding mate- REFERENCES

rial invented in 1824 is too short, when the whole
history of binding materials utilized by ancient people Collepardi, M. 1997. A Historical Review of Development of
throughout several centuries is considered. Through Chemical and Mineral Admixtures for Use in Stucco and
the long history of human beings, several types of Terrazzo Floor. Proceedings of the Fifth CANMET/ACI
International Conference on“Superplasticizers and Other
binding material have been invented sometimes inten- Chemical Admixtures in Concrete”, SP 173, pp. 673–694.
tionally, sometimes unintentionally. Mud, gypsum, Erdoğan, T.Y. 1997.Admixtures for Concrete. Ankara: METU
lime, and lime-pozzolan mixture are some examples Press.
of ancient binding materials. Erdoğan, T.Y. 2002. Materials of Construction. Ankara:
Whatever type of binding materials the ancient civ- METU Press.
ilizations used, it can be seen that those structures built Krumnacher, P.J. 2001. Lime and Cement Technology: Transi-
by ancient binders, particularly lime-pozzolan mix- tion form Traditional to Standardized Treatment Methods,
ture have survived for several hundred years. It can MSc’s Thesis in Virginia Polytechnic Institute and State
be claimed that even at that time architects were aware University, USA.
Lazell, E.W. 1915. Hydrated Lime: History, Manufacture and
of the importance of the durability of binding materi- Uses in Plaster-Mortar-Concrete.
als and their mortars. For instance, the Romans built McKee, H.J. 1971. Early American Masonry Materials in
many beautiful, massive structures such as the famous Walls, Floors and Ceiling: Notes on Prototypes, Sources,
Pantheon lasting through several hundred years till Preparation and Manner of Use. NY.
today. However, those structures belonging to modern Moore, D. 1993. The Riddle of Ancient Roman Con-
civilizations and made by using modern binders, i.e. crete.“Spillway” a newsletter of the US Dept. of the
portland cement have experienced significant deterio- Interior, Bureau of Reclamation. Upper Colorado Region.
ration throughout their service life, generally lesser Neville, A.M. 1998. Properties of Concrete. New Jersey:
than 100 years. As a result, it can be said that the Prentice Hall Inc. 4th Edition.
Nicholson, P. 1850. Encyclopaedia of Architecture: A Dictio-
whole process of modern binding materials, mortars nary of the Science and Practice of Architecture, Building,
and concrete should be re-examined beginning from Carpentry, etc. From the Earliest Ages to the Present Time.
raw materials, production processes of binding mate- Vol.2. NY: Johnson Fry Co.
rials up to mixing, placing, compaction, finishing and Singer, C., Holmyard, E.J., Hall, A.R., Williams, T.I. 1965.
curing processes of mortars and concrete. At that History of Technology. Vol. 1–5. Oxford: Clarendon Press.
point, it is imperative to examine many good examples Vitruvius, P. 1960. Vitruvius: Ten Book on Architectura. New
throughout the ancient centuries. York: Dover Publications.

ACKNOWLEDGEMENT

Throughout this study some informations are quoted

from several web sites i.e.
“www.auburn.edu/academic/architecture/bsc/
classes/bsc314/timeline/timeline.htm”
“matse1.mse.uiuc.edu/∼tw/concrete/hist.html”
“www.oneworld.org/itdg”

94
chap-09 9/9/2004 8: 6 page 95

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

A system approach for examination and determination in

historical buildings

D. Ekşi Akbulut
Architecture Department, Yildiz Technical University, Istanbul, Turkey

F. Aköz
Engineering Department, Yildiz Technical University, Istanbul, Turkey

ABSTRACT: In thousands of years in civilization history, these values, which were created by directly human
beings or with nature, are named today as “Cultural and Natural Heritage”. Conservation of these values is a very
important subject that has to be studied. These studies about conservation and restoration of historical buildings
that are very important cultural heritage must be aimed at scientific principles, aesthetics, and conservation of
their historic values.
Before the conservation, restoration and strengthening of a historical building, study stages have to be
define in collecting and evaluating data towards observation and determination studies. In this paper, a sys-
tem approach towards the criteria’s mentioned above that can be easily applied for each historical structure is
proposed.

1 INTRODUCTION 2 STUDIES ON CONSERVATION

Since early 20th century, some legal obligations such Even if the history of studies on conservation last
as Carta Del Restauro and Venice Charter are consti- out to past, the modern technique for conservation
tuted for the studies about conservation and restora- begins in 19th century. The congress in Athens in 1931
tion of historical buildings which are very important was the first congress about conservation of historical
cultural heritage. monuments. The congress expresses that each coun-
In the past when the building has its own function try constitute official records which shall contain all
its permanent maintenance has also provided its con- documents relating to its historical monuments and to
servation. According to the famous Italian expert Piero deposit copies of its publications with the International
Gazzolo; “If an architectural monument doesn’t serve Office. After the congress in Athens the principles are
to its own function, conservation becomes more than approved by Italy and called “Carta Del Restauro”
a practical necessity, turns into a cultural responsibil- becomes a legal obligation. And it recommends that
ity. The attention to this subject is depends on the next the specialists from different disciplines have to col-
generations cultural knowledge and the sensitivity to laborate and the experts have to agree that before any
the conservation of cultural heritage”. consolidation or restoration is undertaken. And also
Restoration and conservation methods continu- it is pointed that contributions of all periods of the
ously come into being discussions between the dif- building must not be destroyed, the additions which
ferent disciplines. To cope with this problem, it has misleads the experts must not be built and the original
to determine the method and the conservation level materials that are come up with the analytic researches
in a point of historical and aesthetic appearance of must be conserved.
historical buildings and it has to determine the pro- In 1957 “I. International Conference of Architects
cedures putting into practise relating the structural and Technicians of Historic Monuments” is organized
damages and their repairs. In these studies, spe- in Paris. In this conference it is explained that cultural
cialists from different disciplines like architecture, heritage is common responsibility and some orga-
archaeology, history of art and engineering have to nizations like ICOMOS, ICCROM was proposed to
collaborate. establish.

95
chap-09 9/9/2004 8: 6 page 96

“II. International Conference of Architects and ARTICLE 12. Replacements of missing parts must
Technicians of Historic Monuments” is organized in integrate harmoniously with the whole, but at the
Venice in 1964. The Venice Charter which is consti- same time must be distinguishable from the original so
tuted in the conference is examined the problems of that restoration does not falsify the artistic or historic
architectural conservation in nearly hundred years of evidence.
European history which become more complex and
varied. This charter becomes a solution but not an ARTICLE 13. Additions cannot be allowed except in
ending, on the contrary it is affected the discussions so far as they do not detract from the interesting parts
in international perspective. Also in our days many of the building, its traditional setting, the balance of
associations and organizations are studying about this its composition and its relation with its surroundings.
subject, many national and international conferences,
symposiums are organized. 2.2 Code Ethique (1995)
ICOMOS International Training Committee agreed
on responsibilities, documentations, examinations and
2.1 Venice Charter (1964) applications in the meeting in Suomenlinna, Finland
The Venice Charter has an important role about the- in 1995 which is named as “Code Ethique”. Accord-
oretical progress of methods in conservation of his- ing to the articles about conservation, restoration and
torical buildings and monuments, and foundation of strengthening;
associations. It has 16 articles under the definition, a) As the conservation concept has a large extends
aim, conservation, restoration, historic sites, excava- and the attention of many different social groups
tions, publication headings. In this paper, as a system need to be taken, conservation should be studied
for conservation, restoration and strengthening of a by trained and experienced person, carried out with
historical building approach is proposed, the 9–13 the experts who have to collaborate with different
articles are studied. disciplines with a scientific point of view and work
with mass media associations.
ARTICLE 9. The process of restoration is a highly spe-
b) Before the conservation, restoration and strength-
cialized operation. Its aim is to preserve and reveal
ening applications are constructed; the general
the aesthetic and historic value of the monument and
conditions, the physical properties, damage rea-
is based on respect for original material and authentic
sons and other problems of the historical monu-
documents. It must stop at the point where conjec-
ment should be examined.
ture begins, and in this case moreover any extra work
c) According to these examinations, a basic approach
which is indispensable must be distinct from the archi-
for each application must be improved and a con-
tectural composition and must bear a contemporary
servation method which can be short term or long
stamp. The restoration in any case must be preceded
term must be programmed.
and followed by an archaeological and historical study
d) There should always be a precise documentation of
of the monument.
a technical report of the applications with drawings
ARTICLE 10. Where traditional techniques prove inad- and photographs and should be copied for public
equate, the consolidation of a monument can be institution archives. And also for the public use a
achieved by the use of any modern technique for summary report should be written.
conservation and construction, the efficacy of which e) For the future existence of the monument there
has been shown by scientific data and proved by should be a limited applications with the examina-
experience. tion of each effect through the monument and must
be respected to its cultural heritage, environmental
ARTICLE 11. The valid contributions of all periods to integrity, aesthetic, historical, physical originality.
the building of a monument must be respected, since f) There shouldn’t be any damage or changes to the
unity of style is not the aim of a restoration. When a original monument which are documentation of
building includes the superimposed work of different its history during the conservation, restoration and
periods, the revealing of the underlying state can only strengthening application.
be justified in exceptional circumstances and when g) The modern techniques or new materials shouldn’t
what is removed is of little interest and the material be used without required experiments, scientific
which is brought to light is of great historical, archae- researches and discussions with an expert.
ological or aesthetic value, and its state of preservation h) The conservation, restoration and strengthening
good enough to justify the action. Evaluation of the application shouldn’t mislead the future studies,
importance of the elements involved and the decision if it is necessary, the application can be removed,
as to what may be destroyed cannot rest solely on the renewal as possible as can be and congenial with
individual in charge of the work. its surroundings (Binan, 1999).

96
chap-09 9/9/2004 8: 6 page 97

2.3 Conservation studies in Turkey – Property of Antiques and Superior Committee

of Monuments is constituted in 1951 under the
In our country where there are many historic values and
National Education Ministry related to control
historical varieties, some associations and foundations
and observe the principles and applications of
related to conservation are also constituted. In recent
restorations, conservations, preservation of histori-
years it is pointed that the symposiums organized by
cal monuments and architectural buildings.
Cultural Ministry or the meetings Historical Cities
– Antiques Law is presented in 1973. The ele-
Associations organized by municipalities are improved
ments need to be conserved, including the resi-
the point of view about the conservation of historical
dential buildings is attained a required level by
monuments and the public remaining interest to the
this law. Between 1973–1982 years 100 urban
subject. For healthy and permanent conservation not
sites, and 3442 monumental, 6815 residential
only the legal obligations but also active studies in
architecture samples in 417 site area are offi-
micro and macro scale based on scientific data must
cially registrated by Property of Antiques and
be improved.
Superior Committee of Monuments (Ahunbay,
There is more maintenance of donations, economic
1996).
sources for conservation and renovation of the his-
– ICOMOS National Committee of Turkey is consti-
torical buildings compare with the past because of
tuted in 1974 according to obligations of “Inter-
commercial and touristical incomes. This kind of
national Council of Monumentals of Sites” in
approach brings a danger of rebuilt of the monument or
Turkey.
thoughtlessly change of functions. Whatever the aim is
– In 1982 Republic of Turkey is approved to partici-
civilized usage of a historical monument is to keep the
pate to UNESCO in accordance with “Conservation
original by permanent attention. The basic and impor-
of World Cultural and Natural Heritage Charter”.
tant aim is to keep the original not the damaged or
In 1983 Antiques Law is beared and Conservation
copied one of the historical monument.
of Cultural and Natural Heritage Law is consti-
From the world heritage point of view our country
tuted. By this law instead of antiques term cultural
is takes place on such a geographic place that many
heritage term is consumed.
civilizations passed through and leave many work of
arts which are now called cultural and historical her-
itage. Istanbul, in the end of southeast of Europe is The cultural values which are under protection of
the only city, which takes place both in two conti- government and named as Cultural and Natural Her-
nents, Europe and Asia. Bosporus flows between two itage are under the responsibility of Cultural Ministry.
continents, separate the city in two parts. Over 2600 Conservation of Cultural and Natural Heritage Supe-
years with a very interesting historical past, Istanbul rior Committee under the Cultural Ministry carried out
is a unique city in the world. As a very special city the cultural and natural heritage services according to
of Turkey and World Istanbul, takes place in the list the scientific basis.
of historical heritage of UNESCO. So Istanbul is an These are the responsibilities and competences
uncovered laboratory with its historical monuments of Conservation of Cultural and Natural Heritage
and buildings as all over our country. Although many Superior Committee;
of them could not resist to the natural disasters and
wars and not exist today but many precious histori- a) To define the principles of the services about con-
cal buildings have being lasted. Because of this reason servation and restoration of immovable cultural
these studies in Istanbul is very important by means of and natural heritage
universal cultural heritage conservations. b) To provide the required coordination conservation
committees
2.4 Legal obligations of conservation in Turkey c) To determinate the general problems caused by
the applications and to assistant the ministry by
The historical progress of the legal obligations of cul- consensus
tural heritage conservation in Turkey is mentioned
below; There are many organizations like TAÇ, Turing, Cham-
ber of Architects, Foundation of History, Çekül, Galata
– In 1869 (Asar-i Atika Nizamnamesi) the Historical Society, Zeyrek Conservation Society, Cihangir Soci-
Monuments Obligations is constituted. This obliga- ety except Government Institutions provides many
tion is applied with some changes until the period contributions on determination studies, archives and
of Turkish Republic. library services.
– After 1930, Atatürk constituted “Committee for The relations between responsibilities of commit-
Conservation of Old Monuments and Preserva- tees about law and obligation regulations related
tion Delegation” for conservation the historical to conservation of historical buildings is shown in
monuments. Figure 1.

97
 chap-09
9/9/2004
8: 6
page 98
98

Figure 1. Organization scheme of conservation and restoration in Turkey (Akıncı, 2000).

chap-09 9/9/2004 8: 6 page 99

3 STUDIES ON MATERIALS There must be preliminary studies like determi-

nation of original construction techniques, damages
Every architectural shape is formed in a level of occured in time and the determination of physical,
possibilities provided by technology and materials. chemical, mechanical properties of original materials
In order to investigate the concept of aesthetics as like stone, brick, mortar, plaster, adobe, wood, metal,
much as perception of the building, material proper- glass which are used in historical buildings. For the
ties, applications and technological possibilities must experimental studies on materials the tests are gener-
be recognized. In historical buildings in comparison ally performed on specimens taken from the structures.
with comparison with today there are less materials These specimens must be as small as possible for not
and technical possibilities but variety of forms reaches damaging the building (Akman, 2001). Investigation
to a high level. of material properties of a historical building by using
Historical buildings construction typology is non-destructive methods usually includes rebound
masonry. Materials used in masonry buildings, form- test, sonic test, radiographic tests, surface hardness
ing its structural system are also influenced its archi- test, permiability test and bonding test (Aköz, 2001).
tectural features and its identity. The determination of In any case sampling a masonry specimen is not
material properties, which are used for masonry build- a simple operation also in the case of a regular and
ings, is very important subject as also for all types of solid masonry. Being a highly destructive operation,
structures. Architect, can design independently for a only one or two specimens can be sampled and the
new structure but for the conservation and restoration test may be statistically unreliable in the case of a very
of a historical building he has to work by respect to poor masonry. Therefore, the only way to reach the
the history of building and its architectural features. goal seems to be an in-situ testing on the masonry as
In these studies to determine the present situation of a composite (Binda, 1999).
building; researches about regional settlements and The informations to determine the stages of build-
effects of these settlements, structural system of build- ing, construction dates, unexplained additions to the
ing and material properties, restorations that had been building can be obtained from a detailed material study
done before, strengthening and/or extension of build- of the historical building.
ing and their effects to the structural system and the There are authentic construction methods, prepara-
examinations of the soil properties in micro and macro tion methods of materials, authentic material proper-
scale of building region have to be studied. ties of many historical buildings point out differences
In the restoration studies, the properties of original depend on the geographic aspects, construction peri-
materials must be investigate by experimental meth- ods in our country. So as to obtain detailed and straight
ods and the new materials, which will be used to historical documents and information, conservation
have approximately same properties, must be taken and restoration of each building must be studied with
into consideration. Otherwise fatal damages can be material and technological perspective as well as its
occurred and cannot be restored in the means of architectural and artistry perspectives. Otherwise con-
their aesthetics and historic values. For example; in serving the historical buildings from the conservation
the 1950’s the materials like cement additive mortar studies or recovering the restorated historical buildings
and plaster which is very harmful for the histori- must be taken into consideration.
cal buildings is applied by a volunteer firm who
restored with a good intention in Konya. Because of
this application there are significant physical damages 4 A STUDY PROGRAM TO DETERMINE
on the structural materials like stone and bricks on MATERIAL PROPERTIES FOR
the many Anatolian Seljuk Period buildings (Tunçoku, CONSERVATION, RESTORATION AND
2004). STRENGTHENING OF A HISTORICAL
The conservation, restoration and strengthening of BUILDING
a historical building is concerned measured drawing
which is a documentation of architectural measure- Before the conservation, restoration and strengthening
ment and detailed drawings of building, restitution of a historical building, study stages have to be define
drawings which defines the original with the lost par- in collecting and evaluating data towards observation
titions or additions and restoration drawings for the and determination studies. In Table 1, a study program
determination of renewals must be prepared. How- which is proposed in this paper can be easily applied
ever a healthy and permanent conservation must be for each historical structure.
carry out by experts with investigation of materials,
scientific researches, laboratory analyses, prelimi-
nary studies, there is just a technical report which 4.1 Determination of the building identity
defines original material properties and defining the The original name and current name of the building, its
restoration materials. address, original function and current function should

99
chap-09 9/9/2004 8: 6 page 100

Table 1. General study program for conservation and restoration of historical buildings.

100
chap-09 9/9/2004 8: 6 page 101

be defined. Before all else architect of the building, materials as stone, brick, mortar to determine their
the institution or the constructer who organized the strengths.
construction and constructed date or period must be
defined from the documents about the historical build-
ing than the information about past restorations, addi- 4.5 Laboratory studies
tions, renewals must be taken and current conditions After preparing the specimens and samples taken from
of the building must be examined. the building, physical and mechanical tests must be
applied. A statistical relation with the results from the
laboratory tests and the non-destructive test must be
4.2 Information of the building
researched to find a correlation. The determination of
After determination of the building identity, litera- the material properties from the elements that no spec-
ture study of the building, plan, section and elevation imens are taken but non-destructive tests are studied
drawings of the building and the photographic of the must be defined by using the statistic correlation.
building must be researched. So as to begin to investi-
gate on the building area the permissions must be get
from related institutions or organizations and contact 4.6 Selection of materials
with the responsible person. At the first stage, there After the tests, determination of original material and
must be visual determination by taking photographs defining the material properties, the selection of mate-
and if there isn’t any architectural drawing of the build- rials for conservation, restoration and strengthening
ing, the measured, detailed drawing of the building studies must be improved. The required tests must be
must be studied on the building area. applied to the selected materials. In the application
period again some specimens must be taken so as to
make tests and investigate the quality, convenience to
4.3 Examine studies on building
the original material.
Before the examine studies the groups must constitute
a study program and define every single job as clear
as possible. 4.7 Conservation, restoration and strengthening
of the building
4.3.1 Coding For the conservation, restoration and strengthening
The structural or any kind of element in the building studies a restoration project must be prepared after
must be coded on the projects and the photographs the required tests and selection of materials. So as
of the building. These elements must be check with to define the application firm auctioning method can
the building so as to define current conditions and the be used after all the studies, projects and documenta-
damages of the building. This study must be done for tions are obtained. There should be a control system
also each decoration element on the building as well for quality and convenience to the original building
as the structural elements. and the projects after the conservation, restoration
and strengthening applications. An educated group of
4.3.2 Determination of damages workers should for the proper, conscious and conve-
The study stages and study schedule must be formed nient to the original project study in the conservation,
from the prepared, coded project. There should restoration and strengthening applications.
be determination of materials and measurement of There should be expert groups, clarified job
dimensions from each element, detailed photography descriptions, detailed and defined study programs for
from required areas, and determination of damages each study stages in the study program defined in 6
and examination of the damage causes and investiga- stages which is mentioned above.
tion of damage improvements by using the visual and
measurement methods.
5 DISCUSSIONS
4.4 Non-destructive tests
Although this study program seems detailed and long
The non-destructive tests such as rebound test, sonic time work for the the conservation, restoration and
test, temperature test, humidity test and flat-jack test strengthening studies that has many criterias it obtains
must be applied to determine the physical and mechan- determined system for application. The unconscious
ical properties of the materials in winter and summer applications causes unreturn damages to the historical
conditions. Than so as to make studies in laboratory, monuments. In this paper, the study program that has
the specimens must be taken from only required and proper stages, defined experts and study descriptions
permitted areas. The aim in taking specimens from and investigate the unconscious application is pro-
a structural element is to investigate the quality of posed. For example; Hırami Ahmet Paşa Masjid which

101
chap-09 9/9/2004 8: 6 page 102

is restored between 1966–1968 the fresks are removed Akman, S., 2000, Yapi Hasarları ve Onarım İlkeleri, (Build-
and exterior facade stone work restored without a ing Damages and The Principles of Restoration) İstanbul,
respect to the original building with an unconscious TMMOB İnşaat Mühendisleri Odası, İstanbul Şubesi
study (Kuban, 2000). Akıncı, F.N., 2000, Geleneksel Sivil Mimarinin Sosyo-
The basic problems of restorations are not giving Kültürel Ve İşlevsellik Baǧlamında Tarihsel Sürekliliǧi
İçin Planlama/Finans Modeli, (A Financial Model for His-
any importance to research studied before applica-
torical Continuity in the Consistency of Social-Culture
tions, lack of material investigations, to remain true and Function of Traditional Residence Architecture)
to original materials and lack of scientific documen- Doktora Tezi, Y.T.Ü. Fen Bilimleri Enstitüsü, İstanbul
tation. In this paper, the suggested study program Aköz, F., 1995 Yüzer, N., “Investigation of Material Proper-
obtains also detailed documentation, recorded knowl- ties of Küçük Ayasofya Mosque – Sts Sergius and Bac-
edge about the buildings. chus – in Istanbul by Using Non Destructive Methods”,
STREMA 95, Structural Studies Repairs and Maintanence
of Historical Buildings, Chania, Crete, Greece
Aköz, F., 1996, Investigation of Material Properties of Küçük
6 CONCLUSIONS Ayasofya Mosque – STS Sergius and Bacchus by Using
Non-Destructive Test Methods, EC Workshop on Non-
As it seen in the study program which is proposed in Destructive Testing to Evaluate Damage Due to Environ-
this paper; mental Effects on Historing Monuments, Trieste, Italy
Aköz, F., 2001, Experimental Methods and Tests on Histor-
– It may be improved by the experts from differ- ical Structures, Workshop 3rd Meeting of IASS-WG17-
ent disciplines according to their needs and study Historical Structures, İstanbul, Türkiye
subjects Binda, L., Tiraboshi, C., 1999, Flat-Jack Test: A slightly
– These programs which improved by the experts may destructive technique for the diagnosis of brick and stone
be gathered by director and transfer to the other masonry structures Aedification Publishers
experts and it may provide different disciplines col- Binan, C., 1999, Mimari Koruma Alanında Venedik
Tüzügü’nden Günümüze Düşünsel Gelişmenin
laborated study and may be viewed by the director
Uluslararası Evrim Süreci, (International Conceptual
of study program Improvement of From Venice Charter to Today) YTÜ
– Such a study program like this, it may be scien- Basım Yayın Merkezi, İstanbul
tific, aesthetic, conservative to the historical values Erder, C., 1975, Tarihi Çevre Bilinci, (Historical Environ-
and a scheduled study by the experts from dif- ment Conscious) ODTÜ Mimarlık Fakültesi Yayın no: 24,
ferent disciplines for conservation, restoration and sf. 84–87, Ankara
strengthening as it is mentioned in Venice Charter İpekoǧlu, B., Böke, H., 2002, Türkiye’de Tarihi Yapılarda
and other documentations of ICOMOS Malzeme Koruma Sorunları, (The Conservation Problems
of Material in Historical Buildings) 1. Ulusal Malzeme
Kongresi ve Sergisi, TMMOB Mimarlar Odası Büyükkent
Şubesi, İstanbul
REFERENCES Kuban, D., 2000, Tarihi Çevre Korumanın Mimarlık Boyutu,
(Architectural Perspective of Conservation of Historical
Ahunbay, Z., 1999, Tarihi Çevre Koruma ve Restorasyon, Environment) Yapı Endüstri Merkezi Yayınları, İstanbul
(Conservation and Restoration of Historical Environment) Tunçoku, S., 2004, Günümüzde Koruma/Restorasyon
İstanbul, Yem Yayın Çıkmazı, (Contemporary Conservation/Restoration Prob-
Ahunbay, Z., 2002, ICOMOS ve Risk Altındaki Kültürel lems) Mimarlık Dergisi, Sayı 315, Sayfa: 56, İstanbul
Mirasın Korunması, (ICOMOS and Conservation of His- Anon 3, T.C. Kültür Bakanlığı, Kültür ve Tabiat Varlıklarını
torical Heritage Under Risk) İstanbul Yapı Dergisi Sayı: Koruma Genel Müdürlüğü, yayın no: 1806, 2. Basım,
244, Sf: 27–29 Mart, 1996/Ankara

102
chap-10 9/9/2004 8: 6 page 103

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

Analysis of steel-structure/masonry-wall interaction in historic buildings

D. Friedman
Consulting Engineer, New York, USA

ABSTRACT: Conservation of historic high-rise buildings requires analysis of the peculiarities of early modern
construction. The beams, columns, and facades of large steel-frame, masonry-wall buildings experience complex
undesigned interactions. Masonry curtain walls of these buildings are usually 300 mm thick, capable of resisting
structural loads. American investigation of interactions between masonry curtain walls and steel frames began
with facade inspection laws intended to find damaged masonry on high-rise buildings. Structural effects found
include cracking, out-of-plane displacement, and spalls resulting from thermal stress, sidesway, and rust-jacking.
This paper compares the effect on masonry curtain walls of various structural mechanisms, specifically for high-
rise buildings as constructed in the United States between 1900 and 1930. For each mechanism, the stress and
strain interaction of steel and masonry elements is considered, rather than analyzing them separately as in modern
design practice.

1 INTRODUCTION conservation and restoration techniques specific to the

buildings. Simply put, architectural histories, tech-
As the conservation movement has expanded to nical descriptions of designated landmarks, and ad
include modern buildings, and as the oldest steel- hoc conservation knowledge among designers and
frame high-rise buildings have passed one hundred contractors focus more on the masonry envelopes of
years since their construction, engineers have begun to those buildings than their structural details. At one
study technical issues in maintaining and repairing his- extreme, descriptions of tall buildings in architectural
toric high-rises. One of the most difficult issues is the guides may omit critical details of a building, such
interaction under load of heavy masonry curtain walls as one description of the 320 m Chrysler Building
and steel frames, since this was neither anticipated in that fails to mention that it has a brick-veneer exte-
the original designs nor is common in the construc- rior wall, let alone that it has a steel frame (White &
tion of current buildings. Much of the damage that is Willensky, 2000). At the other extreme, technical liter-
seen in these buildings – including cracking, out-of- ature specifically written for conservation typically
plane displacement, and masonry spalling – is similar emphasizes materials concerns over less-known struc-
in appearance to that seen in traditional masonry con- tural ones. The New York Landmarks Conservancy, a
struction but results from different causes. The causes non-profit advocacy and technical expertise organiza-
can best be explained by analyzing these buildings as tion, has created a useful guide to the repair of building
examples of a unique type. facades that emphasizes work on high-rise buildings
New York City has been a center of study of his- (Meadows, 1986). The guide correctly mentions steel
toric high-rises, both formally within the conservation deterioration and the effects of wind pressure and ther-
community and empirically by engineers and archi- mal changes as causes of damage, but is aimed at
tects inspecting and designing repairs for facades. The identifying and repairing materials damage. Unfortu-
historic reason for the study is the sheer concentra- nately, there is no equivalent guide for identifying and
tion of such buildings in the city: in 1929, there were repairing damage caused by interaction between steel
2479 buildings in New York higher than ten stories and masonry elements, or, as they are often identi-
out of 4829 in the United States. Chicago had the sec- fied, structure and architecture. The lack of technical
ond greatest total, 449 (Regional Survey, 1931). The sources is exacerbated by the visual similarity between
empirical reason for the study is the 1980 city law material deterioration and secondary damage from
that requires periodic inspection and repair of all tall- structural movement. For example, brick faces may
building facades. Recognition of New York’s unique spall because of incompatible pointing (a materials
tall-building history has not necessarily translated into problem) or because of high compressive stress from

103
chap-10 9/9/2004 8: 6 page 104

lack of expansion joints (a structural problem). Struc- This latter class of building is highly concentrated in
tural problems such as excessive sidesway are typically New York and Chicago, as no other city had a large
not visible in themselves, but become manifest in number of tall steel-frame buildings constructed in that
damage to non-structural elements. era, and construction details from the post-World War
II period are significantly different.
The character of large portions of Manhattan is
2 HISTORICAL AND ARCHITECTURAL marked by early twentieth century steel-frame build-
CONTEXT ings: the downtown and midtown business districts
contain hundreds of office buildings of this type, while
The earliest steel-framed buildings in the United the north-south avenues of the uptown residential
States, completed before building codes recognized neighborhoods are lined with miles of mid- and
the reality of curtain-wall construction, had very thick high-rise apartment houses. Some of these build-
exterior walls, with 600 mm of masonry common in ings, such as the 240 m Woolworth Building of 1913,
the 1890s. A trend towards 400 mm walls shortly were planned as architectural monuments and have
before 1900 was finalized by the 1901 New York remained so since, but most of these buildings were
City Building Code, which required minimum 300 mm meant as ordinary “background” architecture (Fig. 1).
walls on all steel-frame buildings. Walls continued to Individual buildings have been designated as protected
be built thicker than minimum in order to carry archi- landmarks by the NewYork City Landmarks Preserva-
tectural ornament. Various facing materials were used, tion Commission. There are also designated landmark
most often brick and glazed terra cotta. districts containing many of these buildings, but the
Steel-frame buildings constructed before the 1901 significant buildings within the districts (such as the
code vary greatly in structural type, and can be Upper East Side district) are often low-rise, traditional-
considered as partly experimental. Those constructed construction buildings such as churches and private
after 1901 and before the 1929 economic crash are residences. Despite the fact that so many steel-frame
much more structurally uniform, as described below. mid- and high-rise buildings in Manhattan are not
considered architecturally distinguished, this type of
building is of great interest because it quantitatively
dominates use and repair.

3 BUILDING DESCRIPTION

The analysis in this paper concerns the structural effect

of outside forces on all buildings of a given type –
steel-frame high-rises built between 1900 and 1930 –
using numerical examples from one such building.
Generalizing from the analysis of one building to a
class depends greatly on the assumption that the class
has strongly marked properties that can be defined
in advance. The author’s observations during restora-
tion projects and the historical record both confirm
this assumption – the existence of what architectural
historian Carol Willis has called “vernaculars of cap-
italism,” where building codes, street layouts, and
local economic conditions combined with the standard
building technology of the day to produce “standard-
ize[d] highrise design” (Willis, 1995). The description
that follows is based largely on the author’s observa-
tions, but is similar to those in traditional architectural
and technological histories such as Elliott (1992) and
Condit and Landau (1996).

3.1 Characteristics of the general type

The defining characteristic of the type is the presence
of a structural-steel skeleton frame designed to carry
Figure 1. 953 Fifth Avenue is the narrow building with a all gravity and lateral loads. Under the building codes
three window-wide street facade, directly above the truck. in force during the period of interest (the first thirty

104
chap-10 9/9/2004 8: 6 page 105

years of the twentieth century), the only lateral load Code cleared the way for walls of constant 300 mm
explicitly used in the design of multi-story apartment, thickness by explicitly recognizing the frame struc-
office, and industrial buildings was wind load; under tural support of the walls, in place of the codes based
current codes in the United States, wind load usu- on masonry structure previously in use. The detail used
ally governs for steel-frame buildings except in the for supporting the back-up masonry nearly always con-
high-seismic areas (primarily the west coast) where sisted of the masonry resting at each floor on either
relatively little high-rise construction took place until the spandrel beams or the slab above the spandrel
after 1945. Connections were typically riveted except beams. The veneer was commonly supported through
for beam-to-girder double-angle connections, which mechanical interlock with the back-up (e.g., headers).
were either riveted or bolted. Windows were simple rectangular openings with either
The most common lateral-load systems were loose lintels or, rarely, hung lintels. Masonry piers built
moment frames with semi-rigid bracket connections, integrally with the walls were typically used to provide
often top and bottom stiffened angles. Knee braces fire-protection to the spandrel columns.
or more complex moment frames were used on slen- The most important structural aspect of these build-
der or unusually tall buildings; full-bay bracing was ings is not obvious during cursory examination. Unlike
used only in the tallest and most slender buildings. modern construction, where great efforts are made to
The reliance on moment frames is in part an artifact structurally isolate facades through the use of expan-
of design methods: the use of portal- and cantilever- sion joints and flexible ties, the exterior of these
frame analyses did not provide accurate lateral drift buildings is a system of masonry and metal elements
results that might have encouraged the use of stiffer in continuous contact. More specifically, there are
frames. Matrix-based analysis was impractical with- no expansion joints of any kind in the curtain walls
out computers and moment-distribution was not yet and the fire-proofing piers tie the columns to the
available (Cross, 1930). walls. The presence of these piers and the close con-
Several floor systems were in use simultaneously tact between masonry, spandrel beams, and floor-slab
during the period of interest. In 1900, terra-cotta tile edges makes independent movement of the walls and
arches were the standard method of providing a floor frame impossible, and therefore negates a common
between beams. By 1930, the most common system design assumption.
was the draped-mesh slab, often constructed using
cinder-aggregate concrete. Ordinary bar-reinforced
3.2 Case study: 953 Fifth Avenue
concrete slabs were sometimes used, although they
were rare in New York during the 1920s and 30s. The apartment house constructed in 1924 at 953 Fifth
Patented reinforced-concrete slabs, such as the Kahn Avenue in Manhattan is typical structurally except for
System, were most common in the 1910s, but appear its bar-reinforced concrete slabs. At fourteen stories
throughout the period. (46 m) above grade, it was not particularly tall when
Building facades were solid masonry, consisting built, however it was built on a single 7.6-meter-wide
of a veneer of ashlar, terra cotta, or face brick over lot and it therefore has a fairly high slenderness ratio
common-brick back-up. The 1901 New York Building of 6 (Fig. 2). In reality it receives no wind load in the

Figure 2. Typical floor framing plan, 953 Fifth Avenue. 1 is a spandrel column in a masonry pier, 2 is the typical floor slab,
and 3 is a typical spandrel beam embedded in the wall.

105
chap-10 9/9/2004 8: 6 page 106

Table 1. Calculated movements under load.

Movement
Cause Location Structure (mm)

Sidesway North wall Frame 5.6

Sidesway North wall Frame & 0.05
masonry
Sidesway North wall Masonry 0.5
Thermal North wall Masonry 4.8∗
Thermal North wall Masonry 7.4∗∗
Compression North wall Column 15.8
(live + dead)
Compression North wall Column 11.4
(dead load)
Compression North wall Masonry 3.8
(self-weight)
Compression North wall Masonry 9.7
(live + dead)
Compression North wall Masonry (dead) 6.6
Sidesway West wall Frame 15.7
Figure 3. Typical spandrel section, 953 Fifth Avenue. 1 is Sidesway West wall Frame & 6.1
the veneer masonry, 2 is back-up masonry, 3 is the spandrel masonry
beam, 4 is the floor slab, 5 is the concrete beam encasement Sidesway West wall Masonry 36.8
and fire proofing, and 6 is cement floor topping. Sidesway West wall Masonry 6.1
(reduced wind)
Thermal West wall Masonry 1.3∗
short plan direction because it is sheltered on both the
north and south sides by buildings of similar height and ∗
Horizontal movement.
age, but it was, of course, designed as free-standing. ∗∗
Vertical movement.
953 Fifth has always been a high-end apartment
house. There is little difference in structure between
luxury and ordinary buildings; the use of limestone interior floors, full wind load on exposed wall area, full
veneer on the west (street) facade being the only thermal variation) unless otherwise noted. The details
noticeable departure from ordinary materials. of each condition are described in the following text.
There is a light court at the south-east corner that is Several patterns are apparent in the calculated
roughly half the east-west length of the building and movements and described in detail below. First, load
one-third its north-south width. The basic frame con- sharing between the frame and walls will occur under
sists of seven single-bay, knee-braced frames spanning initial dead load and under some wind loads. Second,
north-south, connected to create three multiple-bay the slenderness of a given facade is a determin-
moment frames running east-west. ing factor in load sharing from differential stiffness.
Two conditions were analyzed: the north facade, Third, significant qualitative differences in results
which is a wide and windowless lot-line wall, and the exist depending on whether full code loads are
west facade, which is a narrow and slender windowed used or reduced loads that reflect more ordinary
wall. Both walls are 300 mm of solid masonry, with all- circumstances.
brick construction on the north and limestone veneer
over brick back-up on the west. The column centerlines
are 356 mm back from the wall faces, with the spandrel 4 EFFECTS OF TEMPERATURE CHANGE
beam locations varying to provide 100 mm of cover
between the flange tips and the wall faces (Fig. 3). New York has large temperature swings every year,
Commercial finite-element software (Dr. Frame ver- with winter average lows of –3◦ C to –1◦ C and a
sion 2.0) was used to analyze the bare steel frame and record low of –26◦ C and summer average highs of
the frame with masonry shear walls, and classical anal- 27◦ C to 29◦ C and a record high of 41◦ C. The standard
ysis was used for the walls as vertically-cantilevered temperatures for HVAC design are –11◦ C for winter
beams. and 32◦ C for summer. Because of direct thermal gain
from sunshine on masonry, it is common for masonry
3.3 Summary of analysis
temperatures during summer days to exceed 38◦ C,
Table 1 lists the movements associated with various therefore a maximum temperature swing of 32◦ C was
forms of motion in the building. All loads are the used for examination of facade thermal effects. The
maximum design loads (full dead and live load on masonry wall is assumed to have been built at the mean

106
chap-10 9/9/2004 8: 6 page 107

temperature of 11◦ C, creating equally large thermal than the original wind-load design. The presence of
stresses at the minimum and maximum outside tem- intermediate masonry piers at each intermediate col-
peratures; for the sake of simplicity, the effects will be umn is of no additional help, since those piers cannot
discussed for expansion only although equal effects resist the lateral load by themselves but will simply
exist for contraction. transfer it to the same columns. Resistance to the load
The masonry of the facade is directly exposed to through friction with the floor structure is resistance
both the overall ambient temperature changes and through horizontal shear in the walls; analysis of the
(depending on orientation) direct heat gain from the long north wall shows a required friction/shear stress
sun. Buildings of this type have no dedicated insula- of 130 kPa on each of the two 200 mm by 30,500 mm
tion, but rather rely on the insulating properties of the contact surfaces per floor, which is realistic under the
masonry and the interior plaster. Since the columns conditions described. However, on the short west wall,
are encased in masonry piers of equal exterior and the required friction is 510 kPa, which is greater than
interior thickness and the spandrel beams are encased the allowable shear strength for most clay masonry,
on the exterior with masonry and on the interior with and implies a large compressive force (which may not
concrete fire-protection, the steel temperature is likely be present) in order to develop the friction.
to be roughly half-way between that of the interior Failure of all of the resistance mechanisms results in
and exterior air. This tends to reduce the effects of outward movement of the ends of a wall, cracking the
extreme temperatures but this ameliorative effect is corner masonry on the intersecting walls in line with
not considered here. the inside face of the wall in question. At an ordinary
A wall plane exposed to a large temperature change exterior corner, both walls are moving from roughly
will undergo both vertical and horizontal movement. the same temperature change, and cracks develop on
The effect of vertical thermal expansion is limited both faces.
by the geometry of the connection between wall and
frame: since most of the wall is back-up masonry
that is vertically bounded by spandrel beams, and 5 EFFECTS OF DIFFERENTIAL STIFFNESS
the veneer is regularly tied to the wall by header
bricks, any vertical expansion of the wall will be Differences in stiffness between adjacent building
resisted by the spandrel beams and the forces trans- elements is one of the most potentially damaging
mitted to the columns. Vertical motion is ultimately effects that can be found. Modern construction con-
not damaging because, pushing upwards from the tains numerous provisions for movement to prevent
foundation, it can simply lift the unrestrained upper accidental load transfer to relatively stiff elements,
portions of the facade and adjacent steel upwards. including slip joints in curtain-wall mullions, expan-
At worst, this creates differential movement between sion joints in masonry curtain walls, and movement
spandrel and interior columns, but this is resisted by joints in interior partitions. Buildings of the stud-
the ductile-metal frame. ied type contain continuous masonry curtain walls in
Horizontal thermal expansion is potentially more direct contact with the structural frames and interior
serious, because it is not resisted by gravity but it is terra-cotta- or gypsum-block partitions solidly built
restrained by structure. Unlike the foundations, which between floor slabs. The relative fragility of the parti-
cannot be moved downward by thermal pressure, or the tions makes them less likely to carry structural load,
roof, which can be moved upwards relatively freely, the so the focus here is on the exterior walls.
columns and walls that restrain horizontal movement Similar to thermal effects, the effects of differential
have limited capacities for this undesigned load. There stiffness can be examined for vertical and horizon-
are various mechanisms that in theory restrain the hor- tal movement. As thermal effects are limited to the
izontal thermal expansion: (1) shear at the intersecting wall plane because the amount of change in wall thick-
wall at each plan corner, (2) friction between the wall ness from temperature variation is negligible, stiffness
and the floor structure along the wall length, and effects are effectively confined to the wall plane by the
(3) resistance provided by the wall columns in bend- magnitude of the element stiffnesses. Masonry walls,
ing from load transmitted through the fire-proofing subjected to out-of-plane forces are far more flexible
piers. The second and third mechanisms depend on than the structural frame to which they are attached
the geometry of construction, since the force transfer and therefore act, properly, as non-structural elements.
from masonry to steel can only take place where the Walls subjected to in-plane loading have stiffnesses of
materials are in direct contact. The practice of building the same order of magnitude as the frames – often
the spandrel beams and columns into the wall provides greater than the frames – and therefore have a ten-
such contact. dency to carry load. The simplest example is vertical
Resistance by the columns in bending cannot take movement under floor loading. As shown in Table 1,
place since the distributed load that is created along the the shortening of the frame columns under dead load is
columns (up to 200 kN/m) creates moments far larger greater than the shortening that the adjacent masonry

107
chap-10 9/9/2004 8: 6 page 108

walls would undergo if they were carrying the load.

This means that the load will be distributed between
the wall and the frame in proportion to their verti-
cal stiffness until and unless the wall suffers enough
damage to release its load. A 300 mm wall would not
be designed as the structural support for a 14-story
building because the compressive stresses are too high;
assuming that the curtain wall was built before all the
dead load is in place implies that as the pressure within
the wall increased, slip between wall and frame or local
crushing of masonry relieved some of the load back
to the frame. As live loads increased during and after
construction, a similar effect would occur.
The more interesting effect of differential stiff-
ness occurs with sideway from lateral load. The steel
frame is strong enough to carry the loads, but is
significantly more flexible than the combined frame-
and-wall. Wind pressure can move a windward wall
(bending in its weak direction) and the frame behind Figure 4. Multiple-wing layout at Four Park Avenue.
more than it moves the parallel walls. The long north
wall is far stiffer by itself than the frame, while the
short and fenestrated west wall is more flexible than slender walls. This difference applies in most buildings
the bare frame. The walls can only serve to fully stiffen of this type since, in the era before air-conditioning
the frames as long as they remain uncracked; each and fluorescent lights, buildings with large overall
crack reduces the continuity of the walls and therefore plan dimensions had light courts that create multiple
their stiffness. This ad hoc composite structure has slender wings (Fig. 4).
rarely been discussed. Stockbridge, in his paper on the
Woolworth Building, lists a number of mechanisms
but looks at all of them in terms of vertical compres-
sion within the wall (Stockbridge, 1981). He does not 6 EFFECTS OF RUST-JACKING
separately discuss the flexure and shear caused within
the masonry by sidesway. Unlike intermittent wind pressures and thermal expan-
Unlike thermal changes, wind loads are highly sion, pressures caused by rusting steel are continuous
dependent on conditions immediate adjacent to the and (until repairs are made) ever-worsening. The pres-
building in question. Tall neighbors can block winds, sure increases from zero as rust builds up behind
as is true for the north-south wind at 953 Fifth Avenue, the masonry, forcing it outward. The volume of rust
and a grid layout such as exists in midtown and upper compared to base metal varies with, for example,
Manhattan and in the downtown area of Chicago pre- Gibbs providing values of seven to twelve times the
vents full wind load from being applied to buildings original volume (Gibbs, 2000). These large values rep-
near or below the median height by preventing most resent free expansion. When masonry is solidly built
winds perpendicular to facades. Exceptional condi- against a rusting piece of steel, the volume increase is
tions may allow for full wind load: the west face of restrained by the strength of the masonry. In technical
953 Fifth faces the 800-meter-wide open space of Cen- terms, rust-jacking is the deflection of that masonry
tral Park. The likelihood of sidesway as the dominant under load until failure (typically through cracking)
mechanism for cracking can be judged by compar- and the forced movement of the masonry after failure,
ing cracking patterns to directions of possible wind including bulging, spalling, and collapse.
loading. In buildings of this type, the structural steel was
As Table 1 shows, the absolute difference between protected with red-lead paint and, by virtue of being
frame and wall flexibility is greater for the slen- embedded within a masonry wall, has never been sub-
der west facade than the stocky north facade. More jected to mechanical abrasion, impact, or exposure
importantly, the maximum bending stress in the north to attack by any non-water-soluble chemical agents.
wall is 97 kPa, acceptable for most masonry, while Experience has shown that the combination of paint
it is 455 kPa in the west wall under reduced load and masonry skin performs well in flat and ordi-
and 2760 kPa under full code load. In other words, nary sections of wall. Where a greater than normal
while the masonry contributes greatly to the stiff- source of water entry into the masonry exists, such
ness of both stocky and slender walls, it is not strong as parapets and the top surfaces of applied ornament
enough to carry the loads its stiffness attracts in the including cornices and water-tables, damage is likely

108
chap-10 9/9/2004 8: 6 page 109

to occur. Rust-jacking, in other words, is the struc-

tural equivalent of an opportunistic infection. While
the masonry and paint that provide weathering protec-
tion to spandrel beams and columns fail over time, the
most serious rusting occurs in those areas where the
masonry fails first, from poor detailing that aids water
entry, improper construction, or most often, cracks
from thermal and lateral movements. This secondary
effect is the most serious form of damage in high-rise
buildings since it is the only one that is progressive
and reduces structural capacity: if rusting of structural
steel were not a concern, it might be possible to accept
many of the cracks formed by movement as “naturally-
occurring” expansion joints that posed no real danger
to either people or property.
Because the effects of rust-jacking are non-linear –
the restraint provided by masonry drops from full to
zero after cracking – and are highly sensitive to voids
in the column fire-proofing piers and in the veneer
outboard of the spandrel beams, it is not possible to
analyze this effect directly.

7 COMMON REPAIR TECHNIQUES

Since the passage of Local Law 10 in 1980, the

New York City Department of Buildings has required
inspection of all building facades higher than six sto- Figure 5. Old beam and corner column masonry repairs at
the rear façade 50 West 34th Street.
ries every five years (Local Law 10, 1980). Conditions
identified as dangerous must be repaired shortly after
the inspection report is filed (Prior to the 1998 revision
of the law, a category of “precautionary conditions”
existed between “safe” and “dangerous;” conditions so
noted had to be repaired before the next inspection.).
A large industry has developed among architects,
engineers, riggers, and masons in performing repairs.
Given the physical similarity of buildings in the type
under consideration, similar weathering exposure, a
general lack of maintenance beyond ordinary joint
pointing, and the small community of professionals
and contractors involved in “Local Law 10 repairs,”
standard methods of addressing problems developed
quickly in the 1980s.
The most common repair performed is rebuilding
masonry at external corners. Cracks on both wall faces Figure 6. Corner column masonry removed and column
cleaned and painted at Four Park Avenue.
are common, highly visible, and suggest a section
of loose masonry. In the majority of buildings that
have brick or terra-cotta veneer, the corner masonry masonry outboard of a spandrel beam provides a sim-
units are typically removed, the steel painted, water- ilar opportunity to create horizontal expansion joints
proofed, and repaired as required, and a new masonry below the level of wall support.
corner constructed. Similar repairs are made at hor- It may seem obvious to state that any steel exposed
izontal strips over spandrel beams where the beams by the removal of masonry should be painted and
have rusted (Figs 5, 6). Expansion joints may be cre- waterproofed to prevent future deterioration, but this
ated in conjunction with masonry replacement or by work is not always performed. It is not uncommon for
themselves. The removal of corner masonry provides workers to open up a ten-year-old repair to masonry
an opportunity to create expansion joints at both faces and find unpainted, rusting steel. The use of self-
of the corner with minimal effort and the removal of adhering rubberized-asphalt sheet waterproofing has

109
chap-10 9/9/2004 8: 6 page 110

made protecting complex steel shapes from water rel- factor, vertical expansion joints may be cut into large
atively simple as long as there is enough of a void flat wall planes as well as at corners, while if sidesway
between the steel and masonry at complex corners is actually the dominant factor, the corner joints are all
(such as the intersection of spandrel beams and corner that is required.
columns) to allow for some movement of the sheets. The analysis and repairs techniques that have been
Steel repairs may be performed with masonry developed in New York (and similar techniques in
repairs, ranging from reinforcing plates welded to Chicago, since the 1996 enactment of a façade inves-
flanges and webs that have lost material to rust up to tigation law) are mostly empirical. The advantage to
complete replacement of columns and beams. Repair this approach is that only those techniques that pro-
is favored over replacement as it is far less disruptive duce acceptable results become commonly used. The
to the occupied interiors and more safely performed disadvantage is that a lack of explicit understanding of
from scaffolding over occupied sidewalks. a given problem leads to an iterative repair approach,
which is wasteful of time and money. Analysis of
facades as composite steel-and-masonry structures, as
8 CONCLUSIONS suggested here, is one method towards better focused
repairs.
Regardless of the causes, damage to masonry cur-
tain walls on high-rise buildings must be repaired to
maintain public safety. However, different causes of
damage may require different repair details. At this REFERENCES
time, despite the large and growing body of experi-
Condit, C. & Landau, S.B. 1996. Rise of the New York
ence in New York among designers and contractors
Skyscraper, 1865–1913. New Haven: Yale University
with repairs, analysis is rarely performed to discover Press.
the causes of observed damage. The mechanism of Cross, H. 1930. Analysis of rigid frames by the distribution of
rust-jacking is well understood, and the initial cracks fixed end moments. Proceedings of the American Society
are typically described as “thermal movement.” Two of Civil Engineers 56: 919–928.
aspects of thermal expansion are not obvious: first, that Elliott, C. 1992. Technics and Architecture. Cambridge,
the force developed in restrained thermal expansion is Massachusetts: The M.I.T. Press.
independent of the length of the element; and second, Gibbs, P. 2000. Technical Advice Note 20: Corrosion in
that simple friction with the floor structure is, for long Masonry Clad Early 20th Century Steel Framed Buildings.
Edinburgh: Historic Scotland.
walls, more of a restraint on thermal expansion than
1980 Local law 10/80 in The City Record 108(32169).
that provided by intersecting walls and columns. The Meadows, R. et al. 1986. Historic Building Facades: A Man-
result for this type of construction, counter-intuitive ual for Inspection and Rehabilitation. New York: New
for non-engineers, is that the longer a wall is, the York Landmarks Conservancy.
less force it exerts from thermal change at corner 1931 Regional Survey of New York, vol. VI: Buildings: Their
intersections with other walls. Uses and the Spaces About Them. reprinted 1974, New
Lateral load is rarely considered in damage surveys York: Arno Press.
of these buildings, since the steel frames are typically Stockbridge, J.G. 1981. The interaction between exterior
adequate for all lateral loads and current design prac- walls and building frames in historic tall buildings. In
Lynn Beedle (ed.), Developments in Tall Buildings, 1983.
tice does not typically treat steel and masonry as a
Stroudsburg, Pennsylvania: Hutchinson Ross Publishing
composite structure. Analyzing this effect can give an Company.
upper bound on stresses in the masonry and, more White, N. & Willensky, E. 2000. AIA Guide to New York, 4th
importantly, can provide insight into patterns of facade edition. New York: Three Rivers Press.
damage, and therefore into repairs required. For exam- Willis, C. 1995. Form Follows Finance. New York: Princeton
ple, if thermal expansion is believed to be an important Architectural Press.

110
chap-11 9/9/2004 8: 6 page 111

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

Some considerations on the shape of the caps of vaults

D. Wendland
Institute of History of Architecture, University of Stuttgart, Stuttgart, Germany

ABSTRACT: The ribs or groins of vaults are usually carefully designed in their geometry, and their geometrical
layout is normally well analyzed. Less attention is usually given to the shape of the caps in between, as its
understanding is not crucial for the interpretation of the architecture. The information available about the shape
of the caps, either from historical building manuals as from secondary literature, in many cases appears not to
be corresponding to the reality as it can be observed in historical vaults, or not describing it sufficiently.
This study is an attempt to describe their shape in regard to their construction process. A better understanding
of the shape of these caps can be useful for their numerical modelling, as its shape has influence on the structural
behaviour of a vault. Beyond that, such knowledge may be applied in the restoration, repairing or rebuilding of
traditional or historical vaults.

1 INTRODUCTION reality, however, reveals that this approach, if not in

every case totally wrong, at least is far from being
The basic elementary description of a cross vault is that precise. Since scholars like Porter (1911) and Rave
of two intersecting cylinders. In such a first approach, (1955) drew their attention on the shape of the caps of
the surfaces of its portions would be considered cylin- vaults, it is known that the problem is less trivial that
ders whose axes run parallel to the spring line of the it may seem at first glance, especially regarding those
vault. This idea can be also found extended to cross vaults that present a pronounced double curvature in
vaults with a pointed profile, conceiving the vault as their caps, so common in gothic architecture. Barthel
composed of eight cylindrical portions (Fig. 1). (1991) elaborates a typological order and a systematic
A comparison with the design procedures proposed approach to the description of vault geometry which
in the historical technical literature and with the built enables to model continuous surfaces on their webbing
that correspond well to the reality as it can be observed
in historical constructions.

Figure 1. Elementary composition sheme of a pointed cross

vault, conceived as being composed of cylindrical portions, Figure 2. The extrados of a vault constructed without
after Breymann/Warth. formwork. Note the difference of the shape respect to Fig. 1.

111
chap-11 9/9/2004 8: 6 page 112

Figure 3. Model of intersecting cylinders, showing the Figure 4. Model of a vault built on the same plan as the
spatial intersection curve. preceding one; as usual, the groins are not resulting from
the intersection of cylinders, but describe curves on vertical
planes.
2 THE CONSEQUENCES OF THE GEOMETRIC
DESIGN OF THE RIBS AND GROINS FOR
THE CAP GEOMETRY

The first reason for the geometrical complexity of the

vaults’ surfaces is the common practise to design pre-
liminarily the “lines of discontinuity” of the vault,
i.e. its formeret and transversal arches (those arches
that confine the bay) and the groins, as these dur-
ing construction are supported by centering frames
which have to be drawn and pre-manufactured, and as
these have the strongest impact in the visual appear-
ance of the vaulted ceiling. These centering frames
are usually designed according to elementary geomet-
ric patterns, sometimes as arches with more than one
center but in most cases simply as circle segments; in
every case with two-dimensional curvature and always
arranged in vertical planes (leaving apart the groins
of barrel vaults with lunettes built on a formwork, or
the particular architectural features of the Guarinian
tradition).
From this fact arises the difficulty to describe the
vaults’ surfaces as cylinders. This can be shown in
vaults with elementary design. If, for instance, a bar-
rel vault with large lunettes is conceived as consisting
of intersecting cylinders with different radii, their Figure 5. Vault constructed with similar boundary condi-
curve of intersection must present a double curvature tions as Figure 4; the arches and groins are circular in vertical
(Fig. 3). If, in the contrary, as usually practised, the planes, the vault surface presents a slight double curvature.
groin is built with a single curvature (as circle seg-
ment or arch with more than one center) in a vertical in the Münster at Freiburg (Barthel 1991: 28). In some
plane, the surfaces of the vault in consequence can- cases, traces of the wooden planks or laths of such
not be cylinders (Figs 4, 5). The same is the case in formworks arranged on the stone ribs could be found
cross vaults with semicircular profile but on rectangu- in medieval buildings (Nussbaum & Lepsky 1999: 65).
lar plan where the groins would result double-curved,
or even in “regular” cross-vaults on square plan, where
usually the groin curves are circle segments. 3 THE SHAPE OF DOMED CAPS BUILT
In some cases the caps can be described by ruled WITHOUT FORMWORK
surfaces between the arches which are generated with
straight lines. This is the case where vaults have been In many cases cross vaults have been constructed with-
built on a wooden formwork. An example can be found out formwork. Where this is the case, no guidance is

112
chap-11 9/9/2004 8: 6 page 113

available for the shape of the caps. Moreover, these

vaults usually present a pronounced double curvature,
as the shape was not restricted by the constraints of
a wooden auxiliary structure, and as such a double
curvature is extremely helpful during the construction
process.
Some building manuals (e.g. Körner 1901) suggest IV V VI
that the shape of these domed caps could be spherical,
thus describing them as spheres truncated by the circu- Figure 6. Level curves on double curved caps (Rave 1955).
lar groins and arches that confine the cap. In this case,
the surface of the cap would have to be clearly discon-
tinuous at its ridge, presenting a re-entrant groin along
the ridge line. The profile of this ridge line, produced
by the intersection of two similar spherical segments,
would necessarily be circular.
Existing vaults with pronounced re-entrant groins
in the ridge line, however, in contrast to statements
that can be found in the historical technical literature
since Ungewitter (1859), present in a vast number of
cases a profile which is not a circle segment: in fact,
the characteristic curve of such a ridge line usually has
a curvature which is smooth but variable (see below,
section 3.4). The surfaces of the caps that intersect in
the ridge line therefore cannot be spherical. Moreover,
besides those vaults that present such a re-entrant groin
at the ridge, many cases can be observed where the
ridge is smoothly curved, or where part of the ridge is
smooth and part is a discontinuity of the caps’ surface.
Therefore we must consider the possibility of a variable
curvature in the vaults’ surface, and also different radii
of curvature corresponding to the different directions
at any given point of the surface. In reality, such a
situation is far more frequent than a constant uniform
double curvature as it would be the case in a spherical
portion of a vault’s surface.
Rave (1955), who systematically analyzed different
vault typologies observing their level curves (looking
at the intersection profiles of the vaults’ surfaces with
5 horizontal planes – Fig. 6), compares the highest Figure 7. The surface of the caps as composed by a bundle
level curve of such “domed” caps with the shape of an of 2nd order parabolas according to Barthel (1991: 30).
almond or egg. The formula for a “simple idealization”
equalling the width of the masonry unit), is mentioned
introduced by Barthel (1991: 29–33) is that of a bundle
in the technical literature since the end of the 18th cen-
of 2nd order parabolas which have their apex in the
tury (e.g. Gilly 1805, first edition in 1795) – but it has
highest point of the cap (Fig. 7). This formula has the
certainly been practised since long before. Today it is
benefit of being simple while avoiding the constraints
still practised by few specialists. The basic principle
of Euclidean primitives; it is easy to adapt to a given
has been formulated for the first time by the German
geometry of groins and arches, defines a continuous
architect J. C. von Lassaulx, whose essay (1829) had a
surface and is still a good approximation to the reality
considerable influence to the development of the topic
of domed cross vaults.
in the technical literature (Wendland 2003). It con-
sists in building the vault with self-supporting courses
that are stable through their shape by forming arches.
3.1 The procedure of free-handed vault
Every new course of the vault’s masonry must be held
construction
in place by the adhesion of the mortar only until it
Free-handed construction of vaults, as regards the cur- is closed; once the course is complete, it is stable by
rent typology of half-stone vaults (i.e. those vaults having the form of an arch and it offers reliable sta-
where the units are placed in a running bond normal to bility for the units of the next course to be laid on
the vault’s surface, the resulting thickness of the shell top of it.

113
chap-11 9/9/2004 8: 6 page 114

Figure 9. Model of a cross vault built with the dovetail

pattern. Note the continuity of the masonry fabric over the
groins.

Figure 8. Domed caps of the vaults of St. Jan cathedral at

s’Hertogenbosch. Note the double curvature.

In cross vaults, these arch-wise curved courses can

be spanned conveniently between the centering frames
of the arches and groins (or their respective ribs),
which will offer sufficient abutment. The curvature of
the courses necessary for their stability can be obtained
Figure 10. Experimental construction of a cross vault
by providing a double curvature to the caps (“domed”
without formwork, using the dovetail pattern. The design
caps), by tilting the planes of the bed joints, or both in reproduces in scale 1:4 a late gothic church vault in northern
combination. Germany. Dimensions in plan: 2.50 × 2.30 m. Dresden
The caps can be built up independently without any University of Technology, 2003.
continuity of the masonry fabric between them, con-
necting the single caps over the diagonal ribs with a in those vaults which have no ribs or those where the
mortar joint. In this way, in every single portion the best groin ribs consist of small units within the masonry
curvature is obtained by a modest doming and some fabric of the caps.
tilting of the bed joint planes. Their spatial inclination
can be corrected if necessary to optimize course curva-
3.2 The masonry apparatus in the dovetail pattern:
ture. The single caps should be built up more or less at
inclination of courses, course curvature and
the same time just not to cause asymmetric loads to the
corrections
centering frames. This is the way Lassaulx, who has
been mentioned above, built his neo-medieval vaults In order to study the building process and its conse-
(Wendland 2003), and that has been commonly used in quences on the shape in detail, a series of small-scale
such cases where the groins were provided with stone models simulating different typologies and vaulting
ribs set on the centering before constructing the caps. patterns (e.g. Fig. 9) and a prototype in scale 1:4
The most frequent masonry apparatus in cross (Fig. 10) were built.
vaults, however, is the dovetail pattern (Fig. 9). It From these trials it became obvious that the layout of
consists in arranging the courses on diagonally tilted a regular masonry apparatus, apart from its formal and
planes perpendicular to the groins. Hence, the masonry structural benefits, is necessary in order to guarantee
fabric is continuous through the neighbouring caps the stability of every new completed course during
over the groin, avoiding the joint behind the groin rib. construction, and to facilitate an efficient geometric
The diagonally tilted courses are seamed in the ridge of control to achieve a similar shape in all caps of the
every cap. The continuity over the groins is beneficial vault and to make sure that the portions rising from

114
chap-11 9/9/2004 8: 6 page 115

Figure 13. Radial inclination of the bed joint planes accord-

ing to Ungewitter (1895), repeated in many other manuals.

Figure 11. Plane bed joint in the vault masonry (prototype).

Figure 14. The bed joint planes as they appear from the
measurement of the prototype at the TU Dresden: throughout
the main part of the cap, the bed joint planes are parallel. In
the drawing, the full circles of the course curvatures are also
visualized.

Figure 12. The variation in the thickness of the mortar joint

in case of non-parallel bed joint planes.

the different spandrels come to meet correctly in the

vertex.
In principle, the bed joint planes are lying normal to
the curve of the groin. However, to follow strictly this
rule would lead to a radial inclination of the bed joints
and consequently to an angle of every bed joint plane
respect to the preceding one. Such an arrangement
is usually described in the building manuals since
Ungewitter (1859: 104–114) (Fig. 13), although this
author already admits that it could be debatable. Actu-
ally, in practise it is problematic to set the bed joint
planes in an angle to each other, as this would lead to
a variation in the thickness of the bed joints because
the distance between the successive bed joint planes
varies according to the local distance of the course to
the turning axis of the planes (Fig. 12).
According to the observations on existing vaults and
the trials on the prototype, such a turning of the bed
joint planes is feasible only in those portions of the
cap where the courses are very short and a manifest Figure 15. Nearly parallel bed joint planes as measured in
angular divergence can therefore be produced within the vault built by J.C. von Lassaulx in the church of Treis
the normal thickness of the mortar joint. This is the (Germany), early 19th Century (Wendland 2003).

115
chap-11 9/9/2004 8: 6 page 116

Figure 16. The shadow reveals the continuity between the Figure 17. The cap masonry during construction; the
course curve and the opposite cap surface (prototype). shifting of the courses outwards is visible (prototype).

cap which must be avoided for formal and structural

case only in the lowest and highest portions of the cap. reasons.
In the main part of the cap, in contrast to what has been Actually in many points of the cap masonry the
stated before, the bed joints are always parallel, tilted courses are not exactly superposed upon each other
in space with a constant angle. (Fig. 17) – this happens because even if a radial incli-
For the curvature of the single courses, no pre- nation of the bed joint planes is achieved respect to the
cise indication can be found in the historical technical diagonal groin, these will not intersect the vault sur-
literature. Generally, a pronounced curvature serves faces orthogonally in every point. As result, especially
to stabilize the courses during construction; Hörnig’s in the lower half the courses are shifting outwards in the
manual (1836) says that less experienced masons will part which is more distant from the groin. An attempt
tend to give more curvature to the courses. Within a to exactly superpose them would be a mistake as it
course, the curvature can be assumed as being con- would successively reduce the curvature of the courses
stant, the resulting form of the courses therefore as because the fleche would remain constant while the
circle segments. Over the entire cap, the curvatures length is increasing – moreover, the starting point at
need not be constant, but may vary from course to the formeret or transversal arch itself is progressively
course or in different portions of the cap. shifting outwards. Such shifting of the courses must be
The area of the cap which is crucial for the choice of restricted to a sufficient superposition of the courses,
the course curvature is right above the summit of the as basic principle of masonry construction.
transversal or formeret arch. Here, the courses have the
greatest length and therefore most deserve to be sta-
3.3 The continuity of the slope
bilized through their curvature. The limit of possible
curvature, i.e. the smallest possible radius, is defined in The rule for the building of the cap masonry, and
this area by the tangential continuity of the course with therefore for the modelling of the cap’s surface, is
the surface of the opposite half of the cap (arising from provided by the necessity of building curved courses
the neighbouring groin), which causes a completely on tilted bed joint planes between the diagonal groin
smooth vault surface at the ridge (Fig. 16). A greater and the formeret or transversal arch that confine the
radius, or minor curvature, in fact, leads to a re-entrant cap, i.e. between two continuous curves (usually circle
groin in the ridge, which is quite common and formally segments), up to the longest course at the summit of
satisfactory, while a smaller radius, or greater curva- the arch.
ture, would lead to an inverse discontinuity of the vault Above that point, one of the track curves is lacking,
surface in the ridge that would appear as projecting to and the question is how to continue.
the inside, which would be a flaw. Hence, the limita- In a first hypothesis, it could be assumed that the
tions of the curvature are the stability of the longest surface continues as a translation surface, extending,
courses and the disturbing of the surface continuity at as track curve, the curve of the formeret or transver-
the ridge of the cap. sal arch either with constant curvature or tangentially.
It is rather difficult to carry out corrections in the However, such a surface generation is problematic
curvature of the courses while proceeding towards the because the length of the generating curves (coherent
zone just mentioned. Such corrections are first of all with the upper courses) continues to increase while
limited by the obvious necessity of laying one course rising upwards. If a constant curvature is assumed
on top of the other. Also, they easily disturb the sur- (Fig. 18), there is a high risk that the distance spanned
face continuity, leading to an outward curvature of the by the course might exceed the diameter of the

116
chap-11 9/9/2004 8: 6 page 117

1000

900

800

700

600

500

400

300

200

100

0
29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55

Course length Radius

Figure 19. Real length of the courses and radii of curvature

throughout one half cap, as measured in the prototype vault.
The length of the courses is growing towards the summit of
the formeret or transversal arch and then descending again.
The radii are nearly constant especially in the area with the
longest courses. For the visualization of the circles described
by the courses, see Figure 14.

increasing the inclination of the bed joints planes,

Figure 18. Hypothesis of a surface modelled on the which due to the length of the courses in this area is
diagonal groin (1) and the extended circular curve of the impossible but by interrupting the masonry apparatus.
formeret arch (2) as track curves, and circle segments on
The demand that the longest courses should not
parallel tilted planes with constant radius, coherent to the
courses. Above a certain length of the generator curves, these project outwards due to their circularity can in some
are projecting outwards, and the ridge line (3) is bending to cases even oblige to reduce their curvature (increase
the outside. The next course could not be defined as it is lying the radius) respect to the preceding ones (which is
beyond the track curve (2). not desirable if we remind the benefit of the curva-
ture especially in the longest courses). In such a case,
the slope of the cap would result straight in this por-
circles which the courses describe, which is impossi- tion – i.e. avoiding to curve inwards below the longest
ble, and this modelling procedure therefore inconsis- courses in order to reduce curvature the less possible,
tent. But even if the course length remains less than the and in continuation aiming to regain the stronger cur-
diameter, the longest courses would project outwards, vature in the following courses. Otherwise, the slope
generating a surface which in normal direction to the normal to the bed joints will be very gently curved to
courses is bending outwards, which is not acceptable the inside in the central part of the cap, while its cur-
in a masonry vault and not realistic. vature would increase in the lower part and near the
In order to make this procedure consistent, a grad- keystone of the vault where the bed joint planes are
ual reduction of curvature of the courses could be radially turning as permitted by the short length of the
assumed. However, in reality such reduction does not courses in these areas.
happen. In the prototype, for instance, the measuring Building the highest part of the cap is not very
data show that the curvature of the courses within the difficult, as the courses in this area are rather short.
region in question is nearly constant, with a radius of Therefore, the slope of the cap can easily be bent to
curvature considerably shorter than the length of the the inside by shifting the courses, by inclining the bed
longest courses (Figs. 14, 19). joint plane and even by allowing some double curva-
In the practical realization, building up the cap ture in the bed joint. Also the curvature of the courses
masonry, the way to continue beyond the summit of the can be easily reduced at one’s convenience.
arch is given by the demand of continuity in the slope
of the cap. Any bending outwards is prohibitive, but
3.4 Surface models for the shape of the cap
also an abrupt change of the slope towards the inside
would be problematic both under structural and formal The shape of the cap, described in these terms, is
aspects. Besides that, such a change in the slope would directly connected to the basic principles of masonry
demand a drastic reduction of the course curvature and and their application in the building process of the
a strong shifting of the courses reducing their super- vault. For the formulation of a geometric description,
position; it would be feasible only by strongly the profile of the slope, assumed to be straight or

117
chap-11 9/9/2004 8: 6 page 118

Figure 22. Extrados of a vault. The ridge line shows a

characteristic profile.

Figure 20. Approximation of the cap surface with a trans-

lation surface generated by parallel shifting of a circle
segment.

Figure 23. Ridge line profile on the extrados of the

prototype.

takes in account both the fact that the bed joint planes
are parallel and the fact that over the central part of
the cap the radii of the curves described by the courses
are more or less constant. Therefore, we may assume
that the parallel translation of a constant curve sliding
first on the curves of the groin and the formeret or
transversal arch and then on a curve that running nor-
mal to the bed joints expresses the continuity of the
Figure 21. Surface generated by the parallel translation of a slope, will lead to an accurate surface model, perhaps
plane circular curve on an elliptical track. Right: intersection except for the highest portion of the cap. The generator
curve with a vertical plane; the portion that corresponds to
the ridge line of a vault is indicated, limited by the position
curve can be assumed as being plane and circular, the
of the formeret (l.) and the center of the vault (r.). inclination in space constantly normal to the central
portion of the groin curve (this could be relevant in
the cases of high pointed or surbased cross vaults) or
gently curved in the main part of the cap, could be simply by 45◦ .
used for the procedure of surface modelling. The “theoretical” translation surface shown in
In some cases a torus may be considered a fair Figure 21 could also explain the typical profile of
approximation for every half cap, although this would the ridge line which, as mentioned, can be seen in
be accurate only in case that the bed joint planes were many existing vaults (Fig. 22). In fact, the curve gen-
radial, which normally is not the case. erated by the intersection of this surface with a vertical
In the case of the prototype, a translation surface plane (analogous to an intersection with the symmetric
generated by the curves representing the courses slid- counterpart of the opposite half cap) is quite similar
ing on the curve of the slope normal to the bed joints, in its shape to these ridge profiles: the curvature is
proved to be rather accurate in the “critical” part of stronger in the portion just above the summit of the
the cap surface (Fig. 20). In any case, the assumption formeret arch and decreases in the central part towards
of a translation surface is rather close to reality as it the keystone of the groins.

118
chap-11 9/9/2004 8: 6 page 119

Figure 25. The extrados of the finished vault prototype.

Figure 24. The vault shown in Figure 22 (a stellar vault with Figure 26. The intrados of the finished vault prototype.
radial caps), reproduced in a model. In the head joints, the
straight slope of the central part of the cap is visible. and his spatial imagination), and also the demand of
perfection and pleasant appearance of the building in
question.
4 CONCLUSION Therefore, the way which the author has chosen to
approach the problem is that based on a close view at
In the considerations noted above, the terminology that the construction process, to formulate principles which
has been employed is near to that of current CAD tools, must be adapted from case to case. The basic approach,
as the problem discussed is essentially practical. starting from an elementary geometric design of the
However, the formulation of an unequivocal answer arches and ribs, is definitely according to the usual
of how to model the surfaces of a vault has been design process, which may be considered as a pre-
avoided. This is because, to the author’s conviction, condition for its accuracy. The modelling of the caps
the solution to be adopted in a particular case cannot according to the basic principles of the construction
be generalized, first, as it depends on the tools to be process, as they have been pointed out, will lead to an
used for modelling, the quality of the available infor- accurate result.
mation (the survey data) and the degree of accuracy
needed (a high accuracy often implies a high com-
plexity). Neither can the particular case be generalized NOTE
beyond a certain point:Apart from the boundary condi-
tions of a vault, like the layout and geometrical design The present study is part of the author’s doctor
of its arches and groins, which may present a great thesis in preparation, which is supervised by Prof.
variety, the influence of the mason on the final shape D. Kimpel, University of Stuttgart. Part of the research
of the cap is strong – it regards his choices, like that on traditional vault construction has been performed
of the radii of curvature in the courses and the man- with the Fraunhofer-IPK Berlin within the EU project
ner of closing the highest part of the cap, his strategy “iso-brick” (EU CRAFT Research Project G5ST-CT-
of correcting defective tendencies in the shape, and 2001-50095). The models have been built by María
certainly his ability (including his manual dexterity José Ventas Sierra and the author. The practical

119
chap-11 9/9/2004 8: 6 page 120

experimentation has taken place at the Chair of Struc- Jäger, W. 1998. Gutachten Dorfkirche Boitin (unpublished).
tural Design (Prof. W. Jäger),TU Dresden, under direc- Körner, C. 1901. Gewölbte Decken: Gewölbe. 2nd edition.
tion of the author.An acknowledgement is due to María Stuttgart: Bergsträsser.
José who gave strong support to the development of Müller, W. 1990. Naturwerkstein in der Denkmalpflege: Die
figurierten Rippengewölbe der deutschen Spätgotik. In
the contents and the text. Geschichte des Konstruierens IV – Wölbkonstruktionen
Photo credits: all photographs and CAD drawings der Gotik 1, proc. intern. colloquium Stuttgart 1989.
are by M. J. Ventas Sierra and D. Wendland Stuttgart, Tübingen: SFB 230; 123–141.
Nussbaum, N. & Lepsky, S. 1999. Das gotische Gewölbe:
Eine Geschichte seiner Form und Konstruktion.
REFERENCES Darmstadt: Wissenschaftliche Buchgesellschaft.
Porter, A.K. 1911. The construction of Lombard and Gothic
Barthel, R. 1991. Tragverhalten gemauerter Kreuzgewölbe. vaults. New Haven etc.: Yale Univ. Press etc.
Karlsruhe: TH Karlsruhe, Inst. f. Tragkonstruktionen. Rave, W. 1955. Das Domikalgewölbe. Deutsche Kunst und
Breymann, G.A. & Warth, O. 1903. Allgemeine Baukonstruk- Denkmalpflege.
tionslehre, mit besonderer Beziehung auf das Hochbau- Ungewitter, G.G. 1859–1864. Lehrbuch der gothischen Con-
wesen. Vol. I: Die Konstruktionen in Stein. 7th improved structionen. Leipzig: Weigel.
and enlarged edition by Otto Warth. Leipzig: Gebhart. Wendland, D. 2002. Model-based formfinding processes:
Gilly, D. 1805. Handbuch der Land-Bau-Kunst: vorzüglich ‘Free forms’ in structural and architectural design. In:
in Rücksicht auf die Construction der Wohn- und F. Levi, M. Chiorino, C. Bertolini (edd.) Eduardo Tor-
Wirthschafts-Gebäude. 3rd ed., Berlin (first edition in roja – From the philosophy of structure to the art and
1795). science of building (proc. Seminar, Turin 2002), Milano:
Lassaulx, J.C. 1829. Beschreibung des Verfahrens bei Franco Angeli 2003, S. 104–19.
Anfertigung leichter Gewölbe über Kirchen und ähnlichen Wendland, D. 2003. A case of Recovery of a Medieval Vault-
Räumen. Journal für die Baukunst, 1.4: 317–330. ing Technique in the 19th Century: Lassaulx’s Vaults in
Lassaulx, J.C. 1846. Über Gewölbeformen: Vortrag des the Church of Treis. In First International Congress on
Bauinspectors von Lassaulx zu Coblenz in der Construction History, Madrid 2003. 2107–17.
Allgemeinen Architecten- und Ingenieur-Versammlung Wendland, D. 2004. Zur traditionellen Technik des
zu Gotha. Zeitschrift für praktische Baukunst, 6/1846, freihändigen Gewölbebaus. In: Das Mauerwerk, 4.2004.
423–427; also Allgemeine Bauzeitung, 376–380.
Hörnig, G.S. 1836. Theoretisch-praktisches Handbuch der
verschiedenen Maurer-Arbeiten bei dem Land- und
Wasserbau. Dresden, Leipzig.

120
chap-12 9/9/2004 8: 6 page 121

Structural Analysis of Historical Constructions – Modena, Lourenço & Roca (eds)

© 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

An inquiry into an unbuilt monument: the mausoleum for the kings of the
Italy of Alessandro Antonelli

G. Pistone
Dipartimento di Ingegneria Strutturale – Politecnico di Torino – Italy

L. Re
Dipartimento di Casa-Città – Politecnico di Torino – Italy

ABSTRACT: The study presented herein addresses an unusual theme: would it be possible to design a building
of gigantic proportions, conceived by a great architect-engineer of the 19th century – Alessandro Antonelli –
in keeping with the views and the specific techniques adopted by the author. The building in question is the
temple-mausoleum for the Kings of Italy, imagined by Alessandro Antonelli, of which all we have is a few tables
with sketches. In this paper, the significant signs taken into account for purposes of formal simplification are
illustrated, and the hypothetical results are correlated to the respective stylistic references of Antonelli’s life-
size projects. Based on the foregoing, the concept of “consistency model” is formulated, to serve as a possible
“validation” tool for architectural/historical virtual models. The first part is followed by a study of the structural
check performed on the reconstruction hypothesis of the mausoleum for the King of Italy.

1 THE CULTURAL GENESIS OF THE PLAN height, could be the introduction to the second; but at
the same time it cannot be excluded that it refers to
“It was in fact in these last years (after 1880, —) another topic”.
that he went ahead with his unyielding intelligence (Franco Rosso, Alessandro Antonelli 1798–1888,
touching on and carrying out new and bold ideals; Milan 1989, page 121), (Rosso, F. 1989).
but there were two projects in particular that he dedi- Even though the critical classification of these
cated his thoughts to, one was the façade of the Milan designs can go no further (as the chronological dis-
Cathedral and the other the mausoleum-temple dedi- cordance between the national competition on the
cated to the Kings of Italy, which was to be erected theme and the testimony of Caselli make the obser-
in Rome either on Monte Mario or on Monte Cavi vation by Rosso legitimate – op cit., page 232-, this
on the site of the ancient temple dedicated to Giove being supported by the date of 1887–1888 affixed by
Laziale. But death, which he encountered while still Costanzo Antonelli on the only table to be completed,
standing in the breach, in spite of his being ninety which deals not of the monument dedicated to Vittorio
years old, also interrupted the thread of thought of Emanuele II, but “if anything, perhaps, of a project that
these two projects, even though of the latter and most wanted to be a piece of criticism”), these records of
grandiose there still remains the plans and the sketched an “interrupted architecture”, in which Antonelli pro-
altimetry which he had outlined and completed” (Cres- ceeds in his faith in the rationalistic assumptions of art
centino Caselli,AlessandroAntonelli,Architect – born in unusual opposition to the trends of that time, seem
Ghemme 14 July 1798, died 18 October 1888, in “The to stimulate some hypotheses on those developments,
Civil Engineer and Industrial Art”, 1888, page 160 that do not occur.
sgg.), (Caselli, C. 1884), (Caselli, C. 1889). This was surely due to the advanced age of the
“There are basically two hypothesis that can be pro- architect, but perhaps also to the set of circumstances
posed for the six drawings that remain (one of which that we know nothing about, that led to the work
had already been put in fair copy): one in the form of being uncompleted. On the other hand, it is not pos-
a Greek cross with wings extended to various degrees, sible to identify the topic of the mausoleum-temple
covered by a gigantic circular dome with a very ogival dedicated to the Kings of Italy in the programme of
arch (….); the other is a circular plan with deep radial the monument dedicated to Vittorio Emanuele II and
apsidal chapels. The first proposal, also developed in the Unity of Italy, already achieved in 1885 at the

121
chap-12 9/9/2004 8: 6 page 122

beginning of the works on the Vittoriano, on the project

by Sacconi. The hypothesis of a specifically created
mausoleum-temple was plausible, in consideration of
the limits of capacity of the Pantheon where Vittorio
Emanuele II had been buried, and which would have
celebrated – apart from the foundation of the reign of
Italy – the destiny of the dynasty (as had occurred
for the Kings of Sardinia for almost two centuries
with the Superga Basilica). However, it is also pos-
sible to object to the opportuneness of reaffirming
the Piedmont supremacy, creating a symbolic building
constituted of the extraordinary emphasised synthesis
of the old and new buildings that symbolised Turin and
Novara. The Antonelli approach to the theme was also
in irreconcilable contrast with the eclectic tendencies
of Italian culture at that time (those that produced the
Palazzaccio and the Vittoriano, but also the concepts
of D’Aronco and Basile); at least as far as his project
for the Italian Houses of Parliament in Turin (1861) is
concerned – an austere and rational project, functional
and exemplarily economic in costs and in the use of
spaces – another which was constructed was preferred
with the stylistic imitations of Guarini buildings and
the emphatic juxtaposition of a complex stylistic elab-
oration in memory of the French Renaissance and of
constituent inventions. On the other hand, in an era
of diffused eclectic typological-stylistic research, as
those years were, the rigorous faith of Antonelli in
Enlightenment rationalism, in the arrangement com-
position according to Durand’s trend, going beyond
the courage and innovation of the construction and of
the walled structures, programmatically leaving orna- Figure 1. Section of the Mausoleum by Antonelli.
mentation to the refined and exclusive elaboration of
the classical syntax (the only language of the “mod-
ern” architect, who does not give in to the allurement half raised and half in prospective sections, in a scale
of historicism, of exoticism, of variously mixed ver- of 1:400 (Fig. 1), which outline a temple in the form
nacularism) was exactly the opposite of the linguistic of a Greek cross covered by an immense ogival dome
orientations of that time. The weary incomprehension with a double calotte closed by the lantern, of an over-
of the works of Antonelli by the critics of architecture, all height of about 225 metres (the lower level would
such as Camillo Boito, and also the majority of Italian have measured 23.60 metres, the internal beam of the
historiographers, who, of all his works only mention ogive would have been 51.20 – eleven metres more
the Mole and describe it as either gothic or metal- than the Pantheon – and the double closing lantern
lic, was no different from that of the more competent would have developed from a level more or less in cor-
experts, such as Tatti and Clericetti, who advised pre- respondence to the height of the Mole, which would
cautionary demolishing the pavilion of the Mole to therefore have been able to be completely contained
substitute it with a light metallic neo-Moorish trel- within the internal space – Fig. 2). These drawings
lis; and analogous considerations can be made for the prevail to show, amongst others, in that they are con-
subsequent reinforced concrete consolidations for the nected to the exhaustive definition of the project idea
Mole and for the Dome in Novara: “It is not the Dome (the other drawings are made up of a water-colour pen
that needs to be reinforced – but rather the spirit of drawing of the considered plan and of a smaller ver-
those who are afraid of that courage”. sion, of another plant articulated in Latin crosses –
Among the six drawings for the presumed theme of of a more specifically ecclesiastical type-, of a cen-
the mausoleum-temple that are kept in the Antonelli tral plane on which a complex articulation of eight
Archives in The Gallery of Modern Art at the Turin absidal chapel wings are imposed, of another different
Civic Museum (department S48, cupboard 6, shelf and even bigger plan where a wide exedra symmetri-
2, file A, –5739-5739), two studies in pencil can be cal to the temple and surrounded by other monumental
found, respectively for the horizontal sections and the buildings can be recognised.

122
chap-12 9/9/2004 8: 6 page 123

Figure 3. Virtual representation of the Mausoleum.

and more experimented than that of the relationship

between the ogive and conical towers of the Novara
dome and of the square planned double pavilion of the
Mole).
All these indications are in nuce and unequivo-
cally present in the original drawing, the elaboration
Figure 2. Comparison of the size of the Mausoleum, of the intended developing the graphic evidence without
Mole (left) and S.Gaudenzio (right). adding anything that was not present or which could
reasonably be attributed for analogy reasons. In par-
ticular, the theme of the large dimensions (that recall
The scrawl of the two here considered drawings sug- the considerations of Galileo on the proportions of the
gestively register the taking shape of the project in colossal and the criticism of Boullée on the Vatican
the assiduous autographical elaboration of the archi- Basilica) is here resolved as in the Novara Dome and
tectonic invention and stimulates two different con- the Turin Mole not in the gigantism but rather in the
siderations: on one hand the critical classification of multiplying and articulating of the members, with an
this “interrupted architecture” in Antonelli’s produc- unusual prevalence of thickly linked voids which give
tion and in the context of history, and on the other, the these large constructions a feeling of lightness and
curiosity of knowing or imagining more: what shape elasticity without comparison, conditions which are
would it have had? Would its sublime dimensions essential for their feasibility. On the judgement that
allowed it to be made? These latter hypotheses have one could make on this unprecedented conception, if
recently been developed from two standpoints, starting it could have been put into concrete form, it is neces-
from the assumption of Antonell’s designing as being a sary to observe that its excellency would not only have
combination/elaboration of defined and constant con- consisted of its incredible dimensions and of the sug-
structive elements, intended prefiguring the building gestion of a titanic site which would have been able to
in a virtual manner (Fig. 3), from both the internal and recall, in the progressing beauty of the in fieri building,
external prospective points of view, building it with theTower of Babel by Brueghel, but also and even more
typical elements regularly used in Antonelli’s works, so of its architectonic proposal which, like the other
demonstrating the compositional coherence and archi- works by Antonelli, can be seen as a concept of ideal
tectonic relevance, among analogous constructions of continuity, in the Enlightenment project rather than the
his time and in the continuity of the historical series of heterogeneous tables that were prepared in his time,
large domes from the Renaissance onwards (in partic- such as the Capitol in Washington. The sense of com-
ular it can be seen how – apart from the dimensions – position quality of the design is obvious and cannot be
the structural conception of the double dome is simpler limited to the suggestion of incomparable dimensions.

123
chap-12 9/9/2004 8: 6 page 124

Its compact syntactic articulation leaves no room for addresses the problem of the stability of the building,
the redundancy or arbitrariness of the definition of and, in particular, it tries to determine whether static
its elements, of the most minute components; and it equilibrium could have been ensured with the knowl-
is based on the rigorous connection to the principles edge and techniques available at the time. To this end,
of classical composition and neoclassical rationalism: a series of FEM numerical models have been devel-
for example, in the indisputable Albertian identity oped to reproduce the material and structures of the
between external order and internal order or in the mausoleum identified during the first part of the inves-
classical predominance of axial moduli (an arrange- tigation. The purpose of the models was to explore the
ment that Benedetto Alfieri had adopted on the Geneva numerous problems that would have had to be solved
Cathedral façade and which Crescentino Caselli had to create a structure so much bigger than any masonry
adopted for functional reasons in the Charity Hospice building ever completed till then.
building in Turin). All the stylistic-syntactic questions The computation code used applies to compression-
would have found the principles to be developed with and tension-resistant materials in the linear elastic
exemplary elaborations of the language of the orders, field. This choice might appear unsophisticated, but in
already experimented by Antonelli in many religious actual fact it was motivated by the consideration that
and civil buildings. The mausoleum-temple, contem- the aim of a stability study is to identify the orders of
porary of the Eiffel Tower, would have really sealed magnitude of the parameters governing the behaviour
the season of Architecture of the Sublime and Reason, of a construction and the efficacy of the spatial organ-
and would have reconnected – through the concepts isation of the resisting skeleton; accordingly, the pur-
of Boullée and the Enlightenment architects – with pose of the computation code is not to permit refined
the Sainte-Genevieve by Soufflot and with the large checks on the materials and the structure, but rather
domes that had progressively preceded them, as can to serve as a quick tool to pin down the shortcomings
be demonstrated by the continuing topicality of tradi- and weaknesses and to identify the changes necessary
tional construction in bricks and stones in the face of to overcome the problems detected by the analysis.
new technologies, in the intent of “attributing light to In keeping with the results of earlier tests performed
the progress in lateritious and stone construction for at our laboratory on materials from the Antonelli
the large coverings”, in that “it is the most consen- school, a Young’s modulus of 2,000 MPa was assumed
sual for our Italian customs, the most convenient for for the masonry, with density of 18 KN/m3 ; the values
our interests and duties, to preferably use the materials assumed for granite were 55,000 MPa and 28 KN/m3
that nature has been lavish with” (as Antonelli himself respectively, and for iron they were 210,000 MPa
wrote on other occasions). and 80 KN/m3 , (Mattone, R., Pasero, G., Pavano, M.,
The second legitimate curiosity was obviously to Pistone, G., Roccati, R., 1982), (Pistone, G.,
verify the feasibility of the project, from what could be Roccati, R., 1988), (Pistone, G., Roccati, R., 1991).
deduced from the sketch. Obviously it is not possible
to imagine which elaborations Antonelli would have
made to the original idea (as he continued to do with 2.1 Modelling the structure
open sites for the Dome and the Mole: it is possible
For the sake of convenience, the model was subdivided
to observe, but never in a restrictive manner) confid-
into two parts, corresponding to a hypothetical subdi-
ing almost exclusively in the intuition and experience
vision of the building: the dome, including the drum
of the architect (as it was for the Dome and the Mole)
and the lantern, and the base.
and which devices he would have gradually introduced
to resolve the criticalities that have appeared from a
2.1.1 The dome, the drum and the lantern
verification method, which with respects to those of
The dome consists of two concentric ogive-shaped
Antonelli’s time, has greatly reduced the difference
skins, 1.30 m thick and ca 70.0 m high. The 38.0 m
between phenomena and hypotheses (and where it is
high drum consists of three orders with binate
difficult for those macroscopic differences in judge-
columns, curtain walls with huge windows in them,
ment that had caused diffidence on the part of scientists
and low columns arranged in three rows. The double
in the past to appear again). The invention, still at the
lantern, measuring ca 54.0 m including the roof, also
draft stage, especially for the single development in
uses binate columns and repeats the visual motifs of
sections, appears to be elaborate, above all for the def-
the drum.
inition of the volume and the internal spaces, rather
than for the resolution and the joining of the members.
2.1.2 The base
The base, consisting of three orders, rises to a total
2 THE STRUCTURE height of ca 66.0 m. The entire plan is characterised
by isolated columns spaced apart according to modules
In view of clarifying the main problem, concerning measuring 6.70 m, save for the central one, which mea-
the static equilibrium the second part of the study sures 7.80 m. The model leaves out several elements

124
chap-12 9/9/2004 8: 6 page 125

(the stairs, the pronaos, the small lanterns on the Nascè, V., Pistone, G., Strona, P.P., 1989). These stiff-
roof, which represent extensions of the stairs) whose ening elements are also used between the outer and
absence does not affect the statics of the building. Ini- inner skins of the dome, where, in similar cases,
tially, we assumed that the dome was supported by 4 Antonelli used to place stairs and pathways to sim-
big arches (Fig. 4) with pendentives, via a ring. The ulate the action of the arches (including rampant and
model was created on the basis of a highly simplified inverted arches – Fig. 5) and to make the two domes
drawing, consisting of one-dimensional “beam” ele- integral with one another.
ments for the representation of columns and pillars, The initial approach is based on a careful read-
two-dimensional “plate” elements for the walls and ing of Antonelli’s documents and the virtual model
the dome, and “brick” elements for the arches and the was defined on the basis of our knowledge of the
ring on top of them. These three elements are linked techniques and construction principles available to the
to one another at each common node in the frame. architect. This choice was dictated by the lack of suffi-
The two-dimensional and three-dimensional elements cient evidence, in terms of surveys and records. Then,
have been subdivided in their turn by means of a mesh by identifying the structural function of each element,
running parallel to the reference axes, fine enough (ca the most probable solution was identified account
150–200 cm) to reveal clearly and in greater detail the taken of the “language” of the building as a whole,
behaviour of the structure under the load-effects; brick and, above all, in accordance with the critical-scientific
elements in particular make it possible to visualise the outlook of statics and the behaviour of materials.
evolution of the state of stress inside them. The initial results were disappointing: through a
The horizontal elements – structural parts of which careful scrutiny it became apparent that the reading
nothing is known – have been replaced with a hypo- was not so immediate and that the possible solutions
thetical mesh consisting of “St. Andrew’s crosses”, could have been different.
conceived as infinitely rigid elements that can be used Hence, it was decided to maintain this design
to simulate, by changing their dimensions, conditions approach as a starting point and then proceed through
as close as possible to the real situation (Arrigoni, R., subsequent steps involving a continuous comparison

Figure 4. San Gaudenzio. A view of the two arches

supporting the dome. Figure 5. San Gaudenzio. Arch and inverted arch.

125
chap-12 9/9/2004 8: 6 page 126

between the results and the most convincing hypoth- be solved by fitting supporting elements, whether vis-
esis that gradually emerged. This method, strangely ible (something unconceivable in Antonelli’s system)
resembling Antonelli’s ways of working by successive or hidden at the top of the arches, due to the lack of
“adjustement”, proved effective and, over time, yielded filling in the extrados of the pendentives. Nor would it
results of considerable significance, even though they have been possible to resort to retaining chains, even
differed from what we had anticipated. huge ones, and therefore the different attempts made
For a better understanding of this approach, let us did not yield any appreciable improvements. From an
go over the different stages so as to provide a schematic analysis of the plan, the presence of stairs at the four
overview of how the study evolved. It should be noted corners suggested the idea of linking adjacent columns
that the hypotheses described are the ones that proved in a double continuous wall, so as to obtain a system
most significant. of symmetrical buttresses for the dome, forsaking the
idea of having isolated fulcrums in these zones.
2.2 1st Hypothesis The deformation configuration reveals an overall
subsidence of the construction l = 74 cm. Two con-
The building is assumed to be made entirely of bricks. siderations should be made at this stage: the first is that
The deformation configuration reveals an overall the situation has improved greatly, in terms both of the
subsidence of the building l = 136 cm, with a sym- deformation configuration and of stresses. The defor-
metrical evolution in the transverse ad longitudinal mation configuration clearly shows the favourable
axes y and z. In addition to bulging conspicuously, the contribution of the partition walls, especially at the
dome tends to sink vertically; in the drum this tendency ring, though the latter is badly warped internally. The
meets with no obstacle, as the columns, vertical sup- maximum values of the (tensile and compressive) prin-
ports and windows yield under the enormous weight cipal stresses are also reduced, especially where brick
of the dome. elements are concerned. From the stresses in the verti-
cal elements, it is possible to determine the evolution
2.3 2nd Hypothesis of the stresses inside the walls.
The other consideration is a persistent tendency of
It is assumed that the structure as a whole has to be the dome to move away from the pendentives and
stiffened. The dome still tends to “sink” into the base, assume a rhombus like shape (Fig. 6). This will be
which, in its turn, is unable to support it. The “beam” monitored during the subsequent stages.
elements, initially conceived as made of bricks, are
replaced with stone: this is motivated by the need to
enhance the stiffness of such a huge structure and to 2.5 4th Hypothesis
add a “precious” material befitting the aulic destina-
tion of the building. At all events, the materials are The next hypothesis entailed the use – as was often the
used according to functional criteria (placing the most case in Antonelli’s system – of stone legati (linkage by
precious materials in the most critical points). The only stone) at the critical points of the structure. In this par-
one-dimensional elements that are not replaced with ticular case, the critical points were the arches, having
stone are those making up the uprights of the windows, zones characterised by very high stresses, particularly
where bricks seem more fitting. at the keystone.
At the same time, granite ribs, conceived according From the graphic viewpoint, the arches were already
to the same criteria adopted by Antonelli for Mole, are subdivided into 12 blocks, 6 in every half arch, but
added to the dome and anchored to the masonry struc- there was no “odd” block at the keystone; this was
ture. Moreover, always with reference to the Mole, remedied by considering two blocks, one per side, plus
retaining chains are installed and connected with ties two more blocks at the reins. The deformation config-
linking the two skins. uration showed a reduction in the overall subsidence
The deformation configuration shows an overall of the building of l = 62 cm, more significant than
subsidence of the construction l = 85 cm. A slight might have been logically expected.
improvement is observed, especially in the zone of the The stresses in the vertical elements are smaller
arches and the piers underneath. Maximum principal than under the previous hypotheses. From an anal-
stresses are seen to decrease in the entire structure, ysis of the deformation configurations we find that
especially in the dome. the most important change has occurred in the brick
The problem of how to carry the enormous thrust elements (simulating the arches and the ring) which
of the arches supporting the dome is still unsolved. are directly affected by the legati. The ring dis-
plays a reduced tendency to warp internally, with a
milder slant, and the arches appear less deformed.
2.4 3rd Hypothesis
The values of maximum principal stresses and ver-
The great problem posed by the thrust of the dome tical stress are more acceptable, without substantial
and the arches that are unable to support it could not variations.

126
chap-12 9/9/2004 8: 6 page 127

Figure 6. Deformed shape at first and final step.

2.6 5th Hypothesis 2. the deformation configuration reveals a settlement

of the pendentives that increases with increasing
The arches supporting the dome are assumed to be
the distance from the arches: a weakness already
rigidly restrained: this stage of the investigation was
observed in the previous hypotheses, which is
designed to explore the behaviour of the dome/arches
clearly reflected in the states of stress in the ring;
system.
3. the windows, the only beam elements made of
The deformation configuration indicates an over-
bricks, sag under the weight of the dome.
all subsidence of the building l = 50 cm. This stage
proved fundamental to define a few essential points: Based on these studies, it is felt that the only two solu-
tions that can still be proposed are: the variant with
1. the arch displays a progressive lowering, which from 4 to 8 arches, and the conversion of all columns
increases from the restraint to the keystone; to stone.

127
chap-12 9/9/2004 8: 6 page 128

2.7 6th Hypothesis Table 1. Summary of the total displacements under the
various hypothesis.
The number of arches supporting the dome is
raised to 8. Hypothesis Changes l (cm)

1 Initial hypothesis 136

2.8 7th Hypothesis
2 Stone columns (except at 83
All vertical elements are converted to stone. the windows), ties and ribs
3 Addition of wall partitions 73
4 Addition of legati 62
2.9 8th Hypothesis 5 Restraint – study 50
6 The eight arches 59
The arches supporting the dome are assumed to be 7 Columns entirely made 48
rigidly restrained (as in the 5th hypothesis) to be able to of stone
assess the benefits arising from the changes introduced 8 Second restraint – study 35
with hypotheses 6 and 7. The deformation configu- 9a The rearrangement of 50
ration shows a total subsidence of the construction the stairwell – 4 arches
l = 35 cm. As for the other values, displacements 9b The rearrangement of 47
are seen decrease further. A significant difference can the stairwell – 8 arches
be observed in the upper part of the drum, precisely
at the three crucial points of the windows, where the
values, as mentioned above, increased abruptly. problems in the connections between the various
At these three points, in fact, the displacements are segments, which could not be solved with the
reduced to about half, from 43 to 28; from 46 to 28; and technologies available at the time.
from 61 to 37. In view of the much greater incidence b) The problem of the arches:
on the results of hypothesis 7 compared to hypothesis The arches cannot bear the exceptional load of the
6, it can be concluded that the choice of material plays dome, both in terms of stresses, that exceed the fail-
a fundamental role in buildings of this type. ure strength of the material, and in terms of thrust,
that cannot be effectively opposed. The impossibil-
ity of using visible supporting elements and hidden
2.10 9th Hypothesis ties makes the problem insolvable. Though the use
As a final hypothesis, let us consider the possibility of of partition walls might seem promising, it is not a
moving the stairwell, ignoring the constraints imposed viable solution for a building of this size.
by the plan: this is done by moving back the partitions c) The problem of the columns at the base:
walls, so that they now oppose the 4 arches directly and The columns at the base are subject to extremely
perpendicularly, creating a sort of confining buttress. high stresses, that they cannot withstand, if they are
Both the stairs included the helicoidal one are moved made of brick masonry. But there is another aspect,
by one module. of a practical nature, to be taken into account: if
made of stone, the columns cannot be one-piece
elements, but have to be conceived as a series of per-
2.11 Results of the analysis forated blocks stacked on top of one another. This
would entail stacking an extremely high number of
The analysis was performed by making successive blocks, given the height of the columns. A solution
changes to the original scheme, always using materials of this sort, might translate into a kinematic chain
and structural elements that were available at the time that would pose yet another challenge to the stabil-
of Alessandro Antonelli, in keeping with his structural ity of the structure, regardless of the other problems
conception. already present in the mausoleum.
The various steps achieved substantial improve-
ments in terms of structural response, in addition to
a general improvement in the distribution of the states
3 CONCLUSIONS
of stress, and strains also gradually diminished. In this
connection, see table 1, that shows how the lowering
Seen through the eyes with which we had wished to
of the uppermost point gradually reduced to about one
examine it, the problem seems unsolvable.
third of the original value (Fig. 6).
But perhaps this is also the last project of
A number of basic problems remain unsolved.
Alessandro Antonelli, elaborated on a few tables that
a) The problem of the dome: were not divulged to his contemporaries, but consigned
Its thrust cannot be carried with traditional meth- to the memory of others who would come after him.
ods. Metal chains cannot be used, because they We therefore like to think that Antonelli wished
would have to be gigantic, posing insurmountable to give shape and proportions to a dream that was

128