Comparative Analysis on the Seismic Behavior of Combined RC-Masonry Buildings

Comparative Analysis on the Seismic Behavior of CombinedRC-Masonry Buildings

Fabio Nardone, Ph.D.1; Gerardo Mario Verderame2; Andrea Prota3; and Gaetano Manfredi4

Abstract: Since the early 20th century combined RC-masonry buildings have become more common in European, Mediterranean, andSouthern America countries. Despite the diffusion of this combined building typology, the international guidelines have not followedbuilding transformation evolutions and, in particular, for combined RC-masonry buildings, nowadays, international guidelines are notexhaustive to deal with specific issues of this building typology. Although there is a well-established background focused on the nonlinearanalysis of masonry structures and RC frames, the knowledge of numerical and experimental criteria for the study of interaction effectsin combined RC-masonry buildings is limited. In this paper, nonlinear static analyses �pushover analyses� on three-dimensional combinedRC-masonry buildings have been performed to obtain capacity curves of single-resistant systems and of the whole building. The resultsconfirm the code guidelines for the design of new combined RC-masonry buildings and provide interesting insights about the seismicbehavior of combined RC-masonry buildings obtained from the rehabilitation of original masonry structures.

DOI: 10.1061/�ASCE�ST.1943-541X.0000249

CE Database subject headings: Reinforced concrete; Masonry; Buildings; Seismic effects; Comparative studies; Rehabilitation.

Author keywords: Combined RC-masonry; Existing and new buildings; Seismic load repartition; Pushover analysis.

Introduction

For centuries masonry buildings have shown their high vulner-ability during seismic events. The seismic vulnerability of exist-ing masonry buildings is mainly related to construction detailsand bad masonry execution. Then, with the spread of RC technol-ogy, new buildings were realized using masonry and RC materialsin the aim of complying with structural requirements and alsoimproving architectonic limits deriving from masonry buildings.For the existing masonry buildings, in order to improve their seis-mic behavior, such as ductility and dissipation energy propertiesother than resistance, alternative rehabilitation techniques werebased on the employment of materials with better mechanicalproperties than masonry. In particular, RC elements and/or frameshave replaced masonry elements and/or walls transforming theoriginal masonry structure in a combined RC-masonry building. Itis necessary to point out the fact that the combined building con-figuration obtained inserting internal RC columns can determine

the demolition of internal masonry walls, leading to severe weak-ness of the structural system especially when the substitution withRC frames is not correctly designed. Furthermore, the configura-tion with internal RC walls is used in some cases of interventionby substitutions/insertions of internal walls or elevators, some-times underestimating the repercussion about the ductility andregularity and causing negative alterations of the global structuralbehavior �Cattari and Lagomarsino 2007�.

Liberatore et al. �2007� suggested a primary typological clas-sification identifying combined RC-masonry systems with RCand masonry elements in adherence �infill masonry walls� and notin adherence. In the latter case, a further distinction is neededbetween different technology elements at different levels �seriessystems� or at the same level �parallel systems�. The Italian codeD.M. Infrastrutture �2008� suggests for combined RC-masonrybuildings, a classification similar to that proposed by Liberatore etal. �2007� for combined RC-masonry buildings with differenttechnology elements not in adherence while does not with themixed RC-masonry technology in adherence.

In particular, combined RC-masonry buildings working as aseries system could derive from raising up by means of structureswith different technologies such as RC frames, while combinedRC-masonry buildings working as a parallel system could derivefrom masonry structures subjected to plan enlargements or insuch cases consist of external masonry walls and internal isolatedRC columns or frames �solution frequently used not only in ex-isting buildings but also in new constructions�.

At present, few researchers have focused their attention oncombined RC-masonry buildings �Modena and Tomazevic 1990;Alessi et al. 1990� because of the previously mentioned variety;similarly, the latest versions of national and international guide-lines �D.M. Infrastrutture 2008; European Committee for Stan-dardization �CEN� 2004� provide only brief explanations aboutthe structural modeling and the seismic-safety criteria, highlight-ing only the fundamental role of connections between differenttechnology elements.

1Dept. of Structural Engineering, Univ. of Naples, Federico II, ViaClaudio 21, P.O. Box I-80125, Naples, Italy �corresponding author�.E-mail: [email protected]

2Assistant Professor, Dept. of Structural Engineering, Univ. of Naples,Federico II, Via Claudio 21, P.O. Box I-80125, Naples, Italy. E-mail:[email protected]

3Assistant Professor, Dept. of Structural Engineering, Univ. of Naples,Federico II, Via Claudio 21, P.O. Box I-80125, Naples, Italy. E-mail:[email protected]

4Professor and Chair of the Dept. Structural Engineering, Univ. ofNaples, Federico II, Via Claudio 21, P.O. Box I-80125, Naples, Italy.E-mail: [email protected]

Note. This manuscript was submitted on August 28, 2009; approvedon May 22, 2010; published online on May 28, 2010. Discussion periodopen until May 1, 2011; separate discussions must be submitted for indi-vidual papers. This paper is part of the Journal of Structural Engineer-ing, Vol. 136, No. 12, December 1, 2010. ©ASCE, ISSN 0733-9445/2010/12-1483–1496/$25.00.

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010 / 1483

Downloaded 25 Jan 2011 to 143.225.98.141. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org

Guidelines Review

The Argentinean guideline �Normas Antisismicas Argentinas�NAA� 1980� points out the fundamental role performed by slabin distributing seismic actions between vertical resistant elementsrealized with different technologies. In the hypothesis of rigidslab, the Normas Antisismicas Argentinas �NAA� �1980� guide-line suggests distributing horizontal loads between structural ver-tical elements based on their stiffness, while in the hypothesis ofa flexible slab, the seismic load repartition can be performed as-suming the slab as a continuous beam on rigid supports.

Over the years, the Italian guidelines have been updated onthis issue. The D.M. Lavori Pubblici �1996� was the first to dealwith mixed building typology meant as buildings with verticalelements made of both masonry walls and elements or frames ofdifferent technologies. About new buildings, combined RC-masonry buildings were allowed, provided that the seismic ac-tions were entirely distributed only on masonry walls. Forexisting buildings, in the majority of cases the seismic actionswere distributed only between masonry walls, being them charac-terized by elastic stiffness generally larger than the other ele-ments; a check about displacements compatibility was alsoprescribed.

Later, OPCM �2005� and then the current D.M. Infrastrutture�2008� provide the same criteria for both new and existing build-ings. For the former, the seismic actions could be entirely distrib-uted either only on masonry walls or only on the elements ofdifferent technologies. However, when the engineers feel neces-sary to account for the collaboration between masonry walls andelements of different technologies in withstanding seismic ac-tions, the structural behavior should be assessed by means of anonlinear analysis �static or dynamic�.

For existing buildings, it is recommended to perform nonlinear�static or dynamic� analysis accounting for the collaboration be-tween masonry walls and elements of different technologies. Thisrecommendation is based on the fact that a reliable evaluation ofthe seismic action sharing between structural elements character-ized by different technologies is not easy to achieve using simpli-fied models based on linear analysis.

Besides Italian code D.M. Infrastrutture �2008� highlights theimportant role of slab in distributing the seismic action, whichdepends on the in-plane slab relative stiffness compared to that ofvertical resistant elements. It is important to highlight that, actu-ally, the other main international codes, such as European Com-mittee for Standardization �CEN� �2001, 2004� and Americanstandards �American Concrete Institute�, deal not specificallywith combined RC-masonry buildings.

Scope of the Study

The above sections have pointed out that two main classes ofcombined RC-masonry buildings could be identified: RC struc-tures with infill masonry walls �working in adherence to RC

members� and masonry structures with presence of RC elementsor frames that are not in adherence to masonry walls. The presentpaper covers only structures belonging to the latter class. Withinthis class, the paper deals only with combined RC-masonry build-ings in which masonry and RC work as parallel systems as de-fined in the Introduction; those in which masonry and RC work inseries will not addressed herein.

The main objectives of the presented study can be outlined asfollows:1. Provide insights toward the design of new combined RC-

masonry buildings in terms of criteria for repartition of seis-mic actions between elements of different technologies andof seismic performance considering strength and displace-ment capacities.

2. Assess the performance of different types of existing com-bined RC-masonry buildings. This task is developed with theaim of analyzing how the rehabilitation of an existing ma-sonry building by means of its transformation into a com-bined RC-masonry building could affect its seismic behavior;this evaluation is carried out highlighting the influence of therehabilitation on the strength and on the displacement capaci-ties of the rehabilitated structure.

Analyzed Structures

The presented analyses are carried out on a typical three-storystructure representative of residential European and Mediterra-nean buildings. It presents a rectangular plan with dimensions of15.0 m�12.0 m and an interstory height of 4.0 m. The externalmasonry walls are made of tuff natural stones, and the adoptedmechanical properties’ values refer to D.M. Infrastrutture �2008�,taking into account the improving corrective parameters of me-chanical properties for the presence of good mortar and crossconnections �Table 1�. The value assigned to the Poisson coeffi-cient of masonry material is equal to 0.5. The external masonrywall thicknesses are 0.70 m at the first floor, 0.60 m at the secondfloor, and 0.50 m at the third floor, while the internal masonrywall thickness is 0.40 m at all the floors.

According to the objectives outlined in the previous section,the following analyses have been conducted:• New buildings: two types �B and C� of new combined RC-

masonry structures are investigated. They are both character-ized by exterior masonry walls and by the presence of twointernal RC frames in the X direction, with column cross sec-tions of 0.30 m�0.60 m at the first and second floors and0.30 m�0.50 m at the third floor. Type C differs from Type Bbecause its RC frames have columns into the masonry exteriorwalls �Fig. 1�.

• Existing buildings: the analyses are conducted using Type Abuilding �Fig. 2� as reference original structure; then, it is as-sumed that Type A building is rehabilitated and it could belongto Type B, Type C, or Type D �Figs. 1 and 2� building. For

Table 1. Mechanical Properties of Masonry

Type of masonryCompression strength fm

�MPa�

Characteristic shear strengthunder zero compressive

stress fvk0

�MPa�Elastic modulus E

�MPa�aShear modulus G

�MPa�a

Tuff natural stone �with corrective parameters� 1.87 0.06 1,620 540aThe stiffness parameters are computed in the elastic state equal to 50% of the elastic ones.

1484 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / DECEMBER 2010


these buildings the column cross sections have been assumedto be 0.30 m�0.30 m at each floor.A set of nonlinear static analyses has been performed on the

analyzed structures considering two load patterns: uniform andtriangular applied in the centroid of the roof level, respectively, inX and Y directions �European Committee for Standardization�CEN� 2004; D.M. Infrastrutture 2008�. For all investigated mod-els, the analysis in the Y direction is more severe in terms ofmaximum value of base shear due to the predominant orientationof resistance elements in the X direction and to the lower initialaxial load on resistant elements in the Y direction caused by floororientation, as also remarked by Shariq et al. �2008�. Neverthe-less, the results of nonlinear static analysis �pushover� will bediscussed for analysis in the X direction because in this directiona combined system is present. According to indications providedby design guidelines �European Committee for Standardization�CEN� 2004; D.M. Infrastrutture 2008�, the ultimate displacementcapacity is taken as the roof displacement at which total lateralresistance �base shear� has dropped below 80% of the peak resis-tance of the structure due to progressive damage and failure oflateral load resisting elements. Consequently, the analyses havebeen stopped at the step corresponding to 20% decay of the maxi-mum base shear that does not necessarily correspond to masonryor RC element failure. The choice to stop the analysis at a par-ticular step is justified by the fact that the lintels without thepresence of tension strength elements �tie-rods or RC bondbeams� have a negligible tensile strength; then, if the buildingcollapse condition corresponds to that of single resistant element,

it could not be possible to investigate the building seismic behav-ior in nonlinear field. The failure modes �flexure or shear� of themasonry panels �pier or lintel�, causing a 20% decay of the maxi-mum base shear, are indicated on relevant curves of Figs. 4–18. Adesirable failure mode is the flexural failure of piers since theseismic response of the building is more ductile.

Numerical Modeling

The combined RC-masonry building is modeled taking into ac-count the slab flexibility, the deformability and limited tensileresistance of lintels, and the compression load repartition on piers,according to Italian code D.M. Infrastrutture �2008�. The limitedtensile resistance of lintels is taken into account through reducedshear and flexural capacities �D.M. Infrastrutture 2008�.This mod-eling strategy has been developed through the computer code�3Muri 2007�. The bearing structure is identified inside the con-struction by masonry walls and RC frames, while the slabs trans-fer the gravity loads and divide the seismic action between thebearing elements.

The masonry walls are subdivided into piers and lintels con-nected by rigid areas’ nodes �Fig. 3�a��. Theoretical and experi-mental evidences have shown that the behavior of bidimensionalmasonry elements, such as piers and lintels, can be representedusing linear element �frame-type approach� �Galasco et al. 2004,

X

Y

COLUMN

BEAM

MASONRY

SLABWAY

EXTERNALWALLS

INTERNALWALLS

Type BY

COLUMN

BEAM

MASONRY

SLABWAY

INTERNALWALLS

EXTERNALWALLS

Type C

Fig. 1. Analyzed types of new combined RC-masonry buildings

X

Y

MASONRY

SLABWAY

EXTERNALWALLS

INTERNALWALLS

Type A

X

Y

COLUMN

BEAM

MASONRY

SLABWAY

EXTERNALWALLS

INTERNALWALLS

Type D

Fig. 2. Analyzed types of existing combined RC-masonry buildings



2006�. Piers and lintels are modeled by two-node elements, and abilinear elastic-perfectly plastic model is adopted which incorpo-rates the shear, VR, and flexural, MR, strength criteria suggestedby European Committee for Standardization �CEN� �2001� andD.M. Infrastrutture �2008� �Figs. 3�b and c��. Rigid end offsets areused to transfer static and kinematic variables between elements’ends and nodes �Galasco et al. 2004� �Fig. 3�b��. Usually, earth-quake damage observation shows the localization of the damagein the defined portion of the walls �apart from the cases charac-terized by very irregular geometry or very small openings� that isnot of interest for the areas’ nodes: for this reason the deformationof these regions is assumed to be negligible, relatively to the

nonlinear macroelements’ deformations governing the seismic re-sponse �Cattari and Lagomarsino 2007�. The failure modes rec-ognized for masonry panels under in-plane loads are diagonaltension, joint sliding, and flexural cracking �Prota et al. 2008;Tena-Colunga et al. 2009�. According to the Italian code D.M.Infrastrutture �2008�, for existing masonry buildings the shearfailure modes of masonry panels are diagonal cracking and jointsliding, while for new buildings is joint sliding. The elements’collapse is determined in correspondence of the drift ultimatevalue �European Committee for Standardization �CEN� 2004;D.M. Infrastrutture 2008� related to the occurred failure mecha-nism. In particular, the capacity expressed in terms of drift of an

OPENING

LINTEL

PIER

RIGID NODE

LINTELRIGID ENDOFFSET

OPENING

3D-NODE 2D-NODE

PIER

Vr (Mr)

drift

V (M)

0.4% (0.8%)

(a) (b) (c)

Fig. 3. �a� Macroelement modeling of a masonry wall; �b� frame type approach; and �c� envelope curve in terms of shear �moment� versus driftfor masonry panels

(a) (b)

Fig. 4. Capacity curves and failure modes for Model B1_CR: �a� triangular load pattern; �b� uniform load pattern �S.F.=shear failure�

(a) (b)

Fig. 5. Capacity curves and failure modes for Model B2_CR: �a� triangular load pattern; �b� uniform load pattern �S.F.=shear failure�



unreinforced masonry wall controlled by shear is taken to beequal to 0.004, while that controlled by flexure is taken to beequal to 0.008 �European Committee for Standardization �CEN�2004; D.M. Infrastrutture 2008�. Once reached the collapse, theelement contribution to the overall strength is only related to itscapacity to carry gravity loads �Fig. 3�c��.

The walls’ out-of-plane response is not evaluated within thedescribed nonlinear analyses; out-of-plane checks have been any-way performed in addition to the in-plane nonlinear analyses

�Magenes 2006�. For simulating the mechanical behavior of con-crete material the model of Mander et al. �1988� has been used.Aiming to model the nonlinear behavior of RC elements, anelastic�perfectly plastic model and lumped plasticity at the ele-ments’ ends is adopted. The evaluated failure mechanisms of RCelements are flexure and shear �according to the criteria proposedin European Committee for Standardization �CEN� �2004� andD.M. Infrastrutture �2008��. The slabs are modeled as orthotropicthree or four node membrane elements so that the flexible dia-

(a) (b)

Fig. 6. Capacity curves and failure modes for Model C1_CR: �a� triangular load pattern; �b� uniform load pattern �S.F.=shear failure�

(a) (b)

Fig. 7. Capacity curves and failure modes for Model C2_CR: �a� triangular load pattern; �b� uniform load pattern �S.F.=shear failure�

(a) (b)

Fig. 8. Capacity curves and failure modes for Model A: �a� triangular load pattern; �b� uniform load pattern



phragms simulate the real in-plane slab stiffness, as requested bythe Italian code D.M. Infrastrutture �2008�. The slabs �includingthe roof� have been assumed simply supported over the masonrywalls. A global Cartesian coordinate system �X ,Y ,Z� is defined,and the wall vertical planes are identified by the coordinates ofone point and the angle formed with the X-axis. In this way, thewalls can be modeled as planar frames in the local coordinatesystem, and internal nodes can still be two-dimensional nodeswith three degrees of freedom. Floor elements, modeled as threeor four node orthotropic membrane finite elements, are identified

by a principal direction with Young modulus E1, while E2 is theYoung modulus along the perpendicular direction, � is Poisson’sratio, and G1,2 is the shear modulus. G1,2 represents the in-planefloor shear stiffness which governs the horizontal actions’ repar-tition between different walls �Galasco et al. 2006�. The use ofnonlinear elements for masonry and RC elements is fundamentalin order to investigate the seismic behavior of masonry elementswith the evolving process of loading up in nonlinear field. Fol-lowing the indications provided by the European Committee forStandardization �CEN� �2004� and D.M. Infrastrutture �2008�, a

(a) (b)

Fig. 9. Capacity curves and failure modes for Model B1: �a� triangular load pattern; �b� uniform load pattern �S.F.=shear failure�

(a) (b)

Fig. 10. Capacity curves and failure modes for Model B2: �a� triangular load pattern; �b� uniform load pattern �F.F.=flexural failure; S.F.=shear failure�

(a) (b)

Fig. 11. Capacity curves and failure modes for Model B1_TR: �a� triangular load pattern; �b� uniform load pattern �S.F.=shear failure�



Fig. 12. Capacity curves and failure modes for Model B2_TR: �a� triangular load pattern; �b� uniform load pattern �S.F.=shear failure�

Fig. 13. Capacity curves and failure modes for Model C1: �a� triangular load pattern; �b� uniform load pattern �F.F.=flexural failure; S.F.=shear failure�

Fig. 14. Capacity curves and failure modes for Model C2: �a� triangular load pattern; �b� uniform load pattern �F.F.=flexural failure; S.F.=shear failure�



Fig. 15. Capacity curves and failure modes for Model C1_TR: �a� triangular load pattern; �b� uniform load pattern �S.F.=shear failure�

Fig. 16. Capacity curves and failure modes for Model C2_TR: �a� triangular load pattern; �b� uniform load pattern: �S.F.=shear failure�

Fig. 17. Capacity curves and failure modes for Model D: �a� triangular load pattern; �b� uniform load pattern �F.F.=flexural failure�



series of nonlinear numerical analyses are performed through thecomputer code �3Muri 2007�, assuming that, due to contact ef-fects, RC frames can be considered fixed to the perpendicularmasonry walls. The important role of connections between differ-ent technology structural elements in the building’s seismic be-havior has been pointed out by Bento et al. �2005�.

Behavior of New Combined RC-Masonry Buildings

Current Italian code �D.M. Infrastrutture 2008� for the design ofnew combined RC-masonry buildings would lead to distribute theseismic actions either only on masonry walls or only on the ele-ments of different technologies. However, if engineers wish toconsider the collaboration between masonry walls and elementsof different technologies in withstanding seismic actions, a non-linear analysis �static or dynamic� should be conducted. For thisreason, the writers have decided to carry out a series of parametri-cal analysis on RC elements’ dimensions for Types B and C: inCase #1, RC beam and column cross sections are designed forsupporting only gravity loads with dimensions of 0.30 m�0.30 m and 0.30 m�0.50 m for columns and beams, respec-tively; in Case #2, RC beams and column cross sections are typi-cal of a seismic design with dimensions of 0.30 m�0.50 m and0.30 m�0.60 m at the first and second floors, respectively, and0.30 m�0.40 m and 0.30 m�0.50 m at the third floor, respec-tively.

The presence of RC bond beams �_CR� inserted at the level oflintels to which connect the RC slabs is foregone for new com-bined RC-masonry buildings. The mechanical properties adoptedare the following: for concrete, fc=20 MPa as compression

strength and for the steel reinforcement, fy =536 MPa as yieldstrength. The total area of steel longitudinal bars in the column istaken as 8% of the gross section area. The cross section of RCbeams presents 3D16mm in the tension zone and 2D12mm in thecompression zone at the ends and stirrups D8mm spaced 0.15 min the critical regions. The joist slabs are 0.24 m high with a0.04 m RC layer. The dead loads are 5 kN /m2 for each floorproperly combined with a 2 kN /m2 live load. The RC bondbeams have a width equal to masonry thickness and a height of0.30 m, with 4D14mm steel longitudinal bars and stirrups D8mmspaced 0.20 m. For each type, the analyses described earlier areperformed. Obtained capacity curves are expressed in terms ofbase shear acceleration �Vbase /��m�� displacement of controlnode located at the top of the building �, with Vbase the base shearand ��m� the mass of the idealized multidegree-of-freedom�MDOF� system. The results of seismic action repartition are re-ported in Table 2 for both triangular and uniform load patterns.

Remarks on Type B Models

The presence of RC bond beams inserted into the masonry wallsmainly affects the behavior of lintels, causing the increase ofmaximum base shear and changing the resistant mechanism oflintels in an equivalent strut mechanism. The development of col-lapse mechanism shows a first phase with a lintel damage and asecond phase with the lintel shear collapse for a triangular loadpattern in Models B1_CR and B2_CR and piers’ shear collapse inthe ground level for Models B1_CR and B2_CR when a uniformload pattern is adopted �Figs. 4 and 5�. About the seismic loadrepartition, this is mainly sustained by external walls in both lin-ear ��98%� and nonlinear ��87%� fields for Model B1_CR; for

Table 2. Seismic Action Rate Sustained by External Masonry Walls

Model

Roof drift ��roof /H�

Triangular load pattern Uniform load pattern

Linear field�%� 0.167%

Ultimate state�%� Failure modes



B1_CR 98 96% 87 S.F.L. 98 96% 93 S.F.P.

B2_CR 93 85% 81 S.F.L. 93 83% 83 S.F.P.

C1_CR 97 92% 84 S.F.L. 97 89% 89 S.F.P.

C2_CR 87 77% 70 S.F.L. 87 73% 72 S.F.P.

Note: S.F.P=shear failure pier; S.F.L.=shear failure lintel; F.F.P.=flexural failure pier; and F.F.L.=flexural failure lintel.

Fig. 18. Capacity curves and failure modes for Model D_TR: �a� triangular load pattern; �b� uniform load pattern �S.F.=shear failure�



Model B2_CR, given the higher stiffness of internal frames, thecontribution of internal frames to sustain seismic action load be-comes not negligible in the nonlinear field ��17%� �Table 2�.

Remarks on Type C Models

The presence of RC bond beams changes the lintels’ resistantmechanism. The different cross section dimensions in Type Ccompared to Type B do not influence the collapse mechanism. Fora triangular load pattern, the collapse is reached for shear lintelfailure, while for a uniform load pattern, the building collapse isattained by shear pier failure �Figs. 6 and 7�.

For Model C1_CR, in the linear field, the seismic action ratecarried out by RC frames is 3% of the total action and becomes�11% in the nonlinear field. The seismic action repartition issignificantly different for Model C2_CR, where the internalframes sustain about the 30% of the total seismic action in thenonlinear field �Table 2�.

The obtained results on new Types B and C buildings seem toconfirm the criteria given by the current Italian code D.M. Infras-trutture �2008�. In fact, if the RC frames are designed to carryonly gravity loads, then the engineers can directly distribute thehorizontal actions only on masonry walls. The same cannot bedone if RC frames are seismically designed; in this case a non-linear analysis is necessary to determine the level of collaborationbetween masonry walls and RC frames.

Performance of Rehabilitated Buildings

Literature review �Liberatore et al. 2007� highlights that seismicretrofitting techniques of old masonry buildings can be based onthe transformation of the original building in a combined typol-ogy; the insertion of tensile resistant elements such as tie-rods�Fig. 21� is usually planned. Their presence could strongly in-crease the masonry walls’ resistance and it could be fundamentalfor preventing out-of-plane mechanisms. The models analyzed inthis section have been selected with the goal to simulate possiblecases of rehabilitated combined RC-masonry buildings �total/partial removal of internal masonry walls, presence of tie-rodsinto the lintels, and complete/incomplete frame mesh�.

A set of nonlinear static analyses has been performed startingfrom the original masonry building �Model A� and replacing inpart �Type D� or totally �Types B and C� internal masonry wallswith RC frames. The dimensions and mechanical properties ofRC elements and slabs, the mechanical properties of employedmaterials, as well as the load acting on the slabs are the same asof the analyzed new combined RC-masonry buildings in the pre-vious section. Even for the rehabilitated buildings, when the RCelements’ dimensions are typical of a seismic design, the label #2is used as opposite to #1 that corresponds to the nonseismic de-sign of the RC members. The steel tie-rods have a yielding stressof 336 MPa and a diameter of 30 mm; a preloaded of 30 kN isassumed.

Table 3. Seismic Action Rate Sustained by External Masonry Walls

Model

Roof drift ��roof /H�

Triangular load pattern Uniform load pattern





A 46 26% 20 S.F.P. 47 30% 27 S.F.P

B1 96 91% 81 S.F.P. 97 92% 81 S.F.P.

B2 81 73% 55 F.F.P. 84 75% 69 F.+S.F.P.

C1 94 86% 66 F.F.P. 93 86% 70 F.+S.F.P.

C2 70 59% 38 F.F.P. 73 60% 42 F.+S.F.P.

D 58 63% 60 F.F.P. 61 71% 69 F.F.L.

B1_ TR 98 95% 90 S.F.L. 98 95% 93 S.F.L.

B2_ TR 92 84% 78 S.F.L. 92 83% 80 S.F.L.

C1_ TR 96 91% 83 S.F.L. 96 91% 87 S.F.L.

C2_ TR 86 74% 65 S.F.L. 86 73% 67 S.F.L.

D_ TR 73 76% 74 S.F.L. 74 79% 75 S.F.L.

Note: S.F.P=shear failure pier; S.F.L.=shear failure lintel; F.F.P.=flexural failure pier; and F.F.L.=flexural failure lintel.

0

0,1

0,2

0,3

0,4

0,5

0 0,5 1 1,5 2 2,5 3

T [s]

Sa [g]

(a) (b)

Fig. 19. �a� Elastic spectrum; �b� determination of the idealized elastic�perfectly plastic force-displacement relationship



Similar to what has been done for new buildings, the capacitycurves of the analyzed existing buildings are expressed in termsof base shear acceleration displacement of control node located atthe top of the building. Information about the seismic action rep-artition is reported in Table 3 for both triangular and uniform loadpatterns.

Remarks on Model A

The original masonry building is the first model to be analyzed.For this building, the mechanism of collapse is governed by theprogressive failure of the lintels followed by the shear failure ofthe internal walls. A sudden decay of the global capacities’ curvesis given by the shear failure of the internal walls. A similar col-lapse mechanism of the masonry building is shown for the non-linear static analysis when a uniform load pattern is employed. Infact, the resistance of the lintels is very low in the absence oftie-rods in the masonry walls �Fig. 8�. Regarding the seismic ac-tion repartition between the masonry walls, in the elastic field, theseismic action is uniformly distributed between external and in-ternal masonry walls, both for uniform and triangular load pat-terns, while in the nonlinear field, the seismic action is mainlycarried out by the internal masonry walls �Table 3�.

Remarks on Type B Models

For Models B1 and B2, where no tie-rods are installed in themasonry walls, the collapse of the building begins with the failureof the lintels; then, it is followed by the shear pier failure locatedin the ground level for Model B1 regardless of the load pattern.For Model B2, flexural and shear pier failures occur when a uni-form load pattern is considered, whereas flexural pier failure isobserved in the case of a triangular load pattern �Figs. 9 and 10�.The seismic action repartition between different technology ele-ments is not influenced by the load patterns. In the linear field, theseismic action rates supported by masonry walls are 96% forModel B1 and 81% for Model B2, being the lower value ofModel B2 due to the higher stiffness of internal RC frames com-pared to B1. This trend is much more emphasized in the nonlinearfield �Table 3�.

The presence of tensile resistant elements in the masonry walls�i.e., Models B1_TR and B2_TR� mainly affects the behavior oflintels, causing the increase of maximum base shear, the reductionof displacement capacity, and the change in the resistant mecha-nism of lintels in an equivalent strut mechanism. The collapse ofthe building occurs due to lintel shear failure for both models andload patterns �Figs. 11 and 12�. Considerations similar to thoseabove done for Models B1_CR and B2_CR regarding the seismicload repartition can be repeated when tie-rods are inserted in ma-sonry walls �Table 3�. This result is coherent with the writer’schoice to design the tie-rods and the RC bond beams with thesame tensile strength.

Remarks on Type C Models

The collapse is attained in the same way for both Models C1 andC2: for a triangular load pattern, it is due to flexural failure ofpiers at the first floor, while for a uniform load pattern, bothflexural and shear failures involve piers at the ground level �Figs.13 and 14�. The seismic load repartition is not influenced by loadpatterns but strongly depends on the cross section dimensions ofRC elements. In Model C1 the seismic action is almost com-pletely sustained by masonry walls in the linear field, while in the

nonlinear field the contribution of RC frames �about 30%� be-comes not negligible. The contribution to the seismic action ofinternal RC frames is important already in the linear field �about30%� for Model C2, but in the nonlinear field is predominant withrespect to the masonry walls �60%� �Table 3�.

If tie-rods are inserted in masonry walls, the building collapseis achieved by lintel shear failure �for both triangular and uniformload patterns� �Figs. 15 and 16�. For Type C1_TR, in the linearfield, the seismic action rate carried out by RC frames is 4% ofthe total action and achieves more than the 13% in the nonlinearfield. The situation is substantially different for Model C2_TRwhere the RC frames sustain about the 35% of the total seismicaction in the nonlinear field �Table 3�.

Remarks on Type D Models

For Model D, where a part of the original masonry internal wallsis kept �i.e., case of stairs�, a considerable increase of the initialstiffness due to the presence of the internal masonry walls ishighlighted with respect to Types B and C models. For the trian-gular load pattern, the building collapse is caused by the flexuralmechanism at the bottom of the internal masonry walls with RCframes, while for the uniform load pattern, the building collapseis due to the flexural lintel failure �Fig. 17�. The seismic action isalmost uniformly shared between the vertical resistant walls in thelinear and nonlinear fields �Table 3�.

The tie-rods inserted into the external masonry walls of ModelD_TR change the building collapse mechanism caused by brittleshear failure of lintels at the first level �regardless of the loadpattern� �Fig. 18�. Nevertheless, the seismic action rate acting onexternal masonry walls is equal to about 75% of the total actionboth in linear and nonlinear fields and regardless of the load pat-tern �Table 3�.

Seismic Assessment of Existing Building TypesA–D

An example of seismic assessment of Building Types A–D isdeveloped assuming that the residential building is located inNaples on a rock soil. This hypothesis permits to clearly identifythe elastic spectrum in terms of accelerations �defined in the Ital-ian code D.M. Infrastrutture �2008�� for the ultimate state �Fig.19�a��. The capacity and demand evaluations are performed ac-cording to the European Committee for Standardization �CEN��2004� for the significant damage �SD� limit state �LS�.

The capacity curves of combined RC-masonry buildings are interms of Vbase−� �base shear versus displacement of control nodelocated on top of the building�. The analyses were stopped at thestep corresponding to 20% decay of the maximum reached baseshear �Vmax� according to indications provided by guidelines �Eu-ropean Committee for Standardization �CEN� 2004; D.M. Infras-trutture 2008�, obtaining in this way the value of the ultimatedisplacement ��u�. The global capacity at the LS of SD has beentaken to be equal to 3/4 of the ultimate displacement capacity forall analyzed models according to the European Committee forStandardization �CEN� �2004�. The writers pointed out that forthe value of the peak ground acceleration obtained by the elasticspectrum �Fig. 19�a��, the analyzed models satisfy the out-of-plane assessment.

The determination of the idealized elastoperfectly plasticforce-displacement relationship �Fig. 19�b�� is performed throughthe following assumptions �for the ultimate state�: the ultimate



displacement �u is assumed according to the step of curve whichcorresponds to the 20% decay of the maximum reached baseshear; the initial stiffness �K�� is determined by imposing theinterception with the step of curve which corresponds to 70% ofthe maximum base shear reached; and the yield base shear �Vy�,which represents also the ultimate strength of the idealized sys-tem, is determined in such a way that the areas under the actualand the idealized force-displacement curves until the ultimate dis-placement are equal �D.M. Infrastrutture 2008�.

The results in terms of displacement capacity �uSD, yield base

shear Vy, and corresponding displacement �y are summarized in

Tables 4 and 5 for triangular and uniform load patterns, respec-tively. Moreover, for all investigated models the period T� of theidealized single degree of freedom �SDOF�, the available ductility� �the ratio between �u

SD and �y�, the inelastic displacement de-mand dmax

� , and the mass of the idealized SDOF system are re-ported. The seismic assessment is performed by comparing thecapacity displacement ��u

SD� with the corresponding inelastic de-mand dmax

� , thus obtaining the capacity-demand ratio

=�uSD /�dmax

� �Tables 4 and 5�.The results reported in Tables 4 and 5 allow to draw the fol-

Table 4. Bilinear MDOF Parameters for Triangular Load Pattern

ModelT�

�s�Vy

�kN��y

�mm��u

SD

�mm� ��dmax

�

�mm� =�uSD / ��dmax

� �

A 0.415 2,074 12.33 36.08 3.90 17.35 2.08

B1 0.592 824 13.34 47.72 4.77 24.95 1.91

B2 0.614 942 16.96 66.65 5.24 25.67 2.60

C1 0.557 1,002 14.86 58.51 5.25 23.10 2.53

C2 0.560 1,172 17.28 67.52 5.21 23.60 2.86

D 0.414 1,327 10.59 50.78 6.39 17.59 2.89

B1_TR 0.313 1,600 7.52 28.98 5.14 12.54 2.31

B2_ TR 0.310 1,793 8.14 29.48 4.83 12.28 2.40

C1_ TR 0.317 1,651 7.85 29.69 5.04 12.79 2.32

C2_ TR 0.319 1,956 9.26 29.25 4.21 12.83 2.28

D_ TR 0.291 2,162 8.44 27.86 4.40 10.84 2.57

Table 5. Bilinear MDOF Parameters for Uniform Load Pattern

ModelT�

�s�Vy

�kN��y

�mm��u

SD

�mm� ��dmax

�

�mm� =�uSD / ��dmax

� �

A 0.375 2,395 11.65 37.22 4.26 15.68 2.37

B1 0.518 957 12.45 44.06 4.72 21.83 2.02

B2 0.531 1,122 15.10 58.09 5.13 22.20 2.62

C1 0.502 1,107 13.33 56.99 5.7 20.82 2.74

C2 0.49 1,380 15.57 57.70 4.94 20.65 2.79

D 0.365 1,864 11.56 27.67 3.19 15.50 1.78

B1_TR 0.284 1,915 7.41 20.50 3.69 10.51 1.95

B2_ TR 0.282 2,088 7.84 25.01 4.25 10.27 2.43

C1_ TR 0.285 1,931 7.42 25.67 4.61 10.58 2.43

C2_ TR 0.283 2,272 8.47 24.77 3.9 10.22 2.42

D_ TR 0.261 2,558 8.03 22.60 3.75 8.57 2.64

Fig. 20. Capacity curves for Models A, B1, B2, C1, C2, and D: �a� triangular load pattern; �b� uniform load pattern



lowing conclusions regardless of the slightly different values cor-responding to triangular versus uniform load patterns. When theoriginal masonry building �Model A� is rehabilitated and thentransformed into a combined RC-masonry building without tie-rods �Models B1, B2, C1, and C2�, the replacement of internalmasonry walls with RC frames generally leads to an increaseddisplacement capacity and to a reduced strength capacity �Fig.20�. In fact, the internal RC frames are less effective than ma-sonry walls in terms of strength and stiffness, whereas they havea larger deformation capacity than masonry walls. It is also notedthat the rehabilitated building tends to exhibit a period larger thanthat of the original building, thus resulting in a reduced demandsince it moves toward larger periods on the elastic spectrum. Thecombination of these three effects �i.e., increase of displacementcapacity, decrease of strength capacity, and increase of period�determines that rehabilitated buildings achieve capacity-demandratios comparable to or larger than the original building �see thelast column of Tables 4 and 5�. Model D behaves in an interme-diate way between original masonry building �Model A� and com-bined RC-masonry building with internal RC frames �Types Band C�.

The highlighted reduction of strength capacity in Models B1,B2, C1, C2, and D is due to the lintels’ collapse that lack tensileresistance; this explains the substantial increase of the base shearcapacity in Models B1_TR, B2_TR, C1_TR, C2_TR, and D_TRdue to the higher capacity of the lintels and also to the higherexploitation of the external walls which are connected by thetie-rods. These models exhibit a larger strength capacity com-pared to the corresponding models with no tie-rods; however,their period and displacement capacities reduce. The combinationof these effects leads to the fact that also the models with tie-rodsexhibit satisfactory capacity-demand ratios.

Conclusions

The present paper deals with the nonlinear behavior of new andexisting RC-masonry buildings with RC elements or frames thatare not in adherence to masonry walls. In particular, only com-

bined RC-masonry buildings in which masonry and RC work asparallel systems have been dealt with.

The results obtained from the analysis on new and existingRC-masonry buildings have validated the approach suggested bythe lately issued new Italian code �D.M. Infrastrutture 2008�. Thefollowing have been pointed out:• For existing buildings, it is necessary to perform nonlinear

analysis in order to evaluate the seismic action sharing be-tween structural elements characterized by different technolo-gies since the seismic action rate sustained by externalmasonry walls can strongly change depending on the type ofanalysis �linear versus nonlinear�.

• For new RC-masonry buildings in which RC frames have di-mensions typical of seismically designed frames, neglectingtheir contribution withstanding horizontal actions would leadto overestimate the actual portions acting on masonry walls.However, the relative percentages shared between masonryand RC elements should be assessed by means of a nonlinearstatic analysis.These considerations are in line with the current design Italian

code D.M. Infrastrutture �2008� for new RC-masonry buildings.With respect to the behavior of existing masonry buildings trans-formed into RC-masonry buildings, the analyses have highlightedthat the substitution of internal masonry walls with RC framescan increase the elastic period of the building and then the dis-placement capacity, but it could strongly decrease the base shearcapacity. For the analyzed cases, the combination of these effectsdetermined that the seismic check on rehabilitated structures wasverified with similar capacity-demand ratios. The insertion of tie-rods into the external walls of existing combined RC-masonrybuildings resulted into a larger base shear capacity and into re-duced period and displacement capacity compared to the samestructures without tie-rods.

The above remarks have been obtained with reference to atypical building configuration in which out-of-plane checks werealways verified. This should be considered in extending theseconclusions to combined RC-masonry structures with differentcharacteristics. In general, the quality of connections between dif-ferent technology elements should also be assessed to preventlocal failures.

RC SLAB

TIE ROD

EXTERNALANCHORAGEFORTIEROD

OPENING

LINTEL

PIER PIER

RC SLAB

RCBONDBEAM

OPENING

LINTEL

(a) (b)Fig. 21. Details of tension strength elements for lintels: �a� tie-rod; �b� RC bond beam



Acknowledgments

The analysis presented in this paper has been developed withinthe activities of Rete dei Laboratori Universitari di IngegneriaSismica—ReLUIS for the research program funded by the De-partment of Civil Protection—Executive Project 2005–2008.

References

3Muri Program. �2007�. Release 3.0.506, �http://www.stadata.com�.Alessi, A., Diotallevi, P. P., Jurukoyski, D., Petkovski, M., Tashkov, Lj.,

and Zarri, F. �1990�. “Shaking table test of reduced scale model ofbrick masonry building.” Proc., 9th European Conf. on EarthquakeEngineering, Vol. 5, Moscow, 152–161.

Bento, R., Lopes, M., and Cardoso, R. �2005�. “Seismic evaluation of oldmasonry buildings. Part II: Analysis of strengthening solution for acase study.” Eng. Struct., 27, 2014–2023.

Cattari, S., and Lagomarsino, S. �2007�. “Formulazione di elementi nonlineari per l’analisi degli edifici esistenti a struttura mista muratura-c.a.” XII Convegno “L’Ingegneria sismica in Italia,” Edizioni PLUS-Pisa University Press, Pisa.

D.M. Infrastrutture. �2008�. “Norme tecniche per le costruzioni.” No. 29,D.M. 14/01/2008, Rome.

D.M. Lavori Pubblici. �1996�. “Norme tecniche per le costruzioni in zonasismica.” D.M. LL. PP. 29, Gennaio, Rome.

European Committee for Standardization �CEN�. �2001�. “Design of ma-sonry structures. Part 1-1: General rules for building—Rules for un-reinforced and reinforced masonry.” Eurocode 6, Brussels.

European Committee for Standardization �CEN�. �2004�. “Design ofstructures for earthquake resistance—Part 3: Assessment and retrofit-ting of buildings.” Eurocode 8, Brussels.

Galasco, A., Lagomarsino, S., and Penna, A. �2006�. “On the use of

pushover analysis for existing masonry buildings.” Proc., 1st Euro-

pean Conf. on Earthquake Engineering and Seismology �CD-ROM�,Paper n. 1080, Geneva.

Galasco, A., Lagomarsino, S., Penna, A., and Resemini, S. �2004�. “Non-linear seismic analysis of masonry structures.” Proc., 13th WCEE,Vancouver, Paper No. 843.

Liberatore, L., Decanini, L. D., and Benedetti, S. �2007�. “Le strutturemiste muratura-cemento armato:uno stato dell’arte.” XII Convegno

“L’Ingegneria sismica in Italia,” Edizioni PLUS-Pisa UniversityPress, Pisa.

Magenes, G. �2006�. “Masonry building design in seismic areas: Recentexperiences and prospects from a European standpoint.” Proc., 1st

European Conf. on Earthquake Engineering and Seismology �CD-ROM�, Keynote Address K9, Geneva.

Mander, J. B., Priestley, M. J. N., and Park, R. �1988�. “Theoreticalstress-strain model for confined concrete.” J. Struct. Eng., 114, 1804–1826.

Modena, C., and Tomazevic, M. �1990�. “A research program on theseismic behavior of modern masonry buildings.” Proc., 9th European

Conf. on Earthquake Engineering, Vol. 5, Moscow, 13–22.Normas Antisismicas Argentinas �NAA�. �1980�. NAA-80, Instituto Na-

cional de Prevencion Sismica �INPRES�, San Juan, Argentina.OPCM. �2005�. No. 3431, Suppl. Ord. No. 85, 3 Maggio, Rome.Prota, A., Manfredi, G., and Nardone, F. �2008�. “Assessment of design

formulas for in-plane FRP strengthening of masonry walls.” J. Com-pos. Constr., 12, 643–649.

Shariq, M., Abbas, H., Irtaza, H., and Qamaruddin, M. �2008�. “Influenceof openings on seismic performance of masonry building walls.”Build. Environ., 43, 1232–1240.

Tena-Colunga, A., Juarez-Angeles, A., and Salinas-Vallejo, V. �2009�.“Cyclic behavior of combined and confined masonry walls.” Eng.Struct., 31, 240–259.



Comparative Analysis on the Seismic Behavior of Combined RC-Masonry Buildings

Documents