sd

r

d

oiConfined masonryCombined masonryShear strengthDeformation capacityCyclic testingDamage index

1. Introduction

Confined masonry walls are usually made with fired clay bricksor concrete blocks confined with reinforced concrete tie-columnsand bond-beams. Confined masonry is the dominant mode ofconstruction for housing inMexico [1]. In fact, confinedmasonry isalso widely used for housing in most if not all Latin-Americancountries [26]. Confined masonry is also used in some Europeancountries like Italy, Portugal and Slovenia [3,79] and in Asiancountries like Iran, Indonesia [10], Pakistan [11] and China [9,12].Confined masonry walls made with fired clay bricks have

been used in Mexico for a long time [1] and satisfactory seismicperformances have been reported for such construction duringmoderate and strong earthquakes, including the strong September19, 1985 Michoacn (Ms = 8.1) [13] and the October 9, 1995Manzanillo (Mw = 8.0) [14] earthquakes. However, the greatdemand of housing in Mexico for low income people has forcedthem to look for alternative construction systems to build theirhomes with a reduced budget, using then some of the cheapestmaterials available. One of the systems currently used for thispurpose is termed as combined and confined masonry, wherecourses of lightweight concrete blocks (inexpensive in Mexico),

Corresponding author. Tel.: +52 55 5318 9460; fax: +52 55 5318 9085.E-mail addresses: atc@correo.azc.uam.mx (A. Tena-Colunga),

aja@correo.azc.uam.mx (A. Jurez-ngeles), vsv@correo.azc.uam.mx(V.H. Salinas-Vallejo).

are alternated with courses of clay bricks (more expensive), asdepicted in Fig. 1. Important savings in cost and execution timeare obtained with this type of masonry construction, besideshaving an aesthetic appearance when two or more courses ofbrick are alternated with one or two courses of concrete blocks,as illustrated, for example, in Fig. 1a.This modality of construction has historical background world-

wide, for example, in some ancient buildings and walls fences atIstanbul, Turkey (Fig. 2), in old cities of Europe (Fig. 3) and in fewbuildings of the XVII or XVIII century in Mexico (Fig. 4), where nat-ural stones were alternated with fired bricks. However, the morerecent version of confined and combinedmasonry becamepopularin recent times by the initiative of the inhabitants of the Mexicanstates of Puebla, Tlaxcala and Oaxaca. They tried to solve empir-ically with this modality the cracking problem observed in wallsmade with concrete blocks due to differential settlements. As amatter of fact, their idea of alternating courses of bricks with con-crete blocks was successful to solve that problem.This modern version of combined and confined masonry has

been being used since early 1990s. Different arrangements tocombine and alternate brick courses with block courses have beenused [1517], but the one that it is more commonly used is theone depicted in Fig. 1a, where three courses of clay bricks alternatewith a course of concrete blocks.Previous experimental masonry research inMexico that started

in the late 1960s has concentrated in confined masonry walls,primarily those made with brick [1,2,1829], although thereare also some testing with confined walls made with concreteEngineering Structure

Contents lists availa

Engineering

journal homepage: www.el

Cyclic behavior of combined and confine

Arturo Tena-Colunga , Artemio Jurez-ngeles, VctDepartamento de Materiales, Universidad Autnoma Metropolitana, Edificio H, 3er. Piso, A

a r t i c l e i n f o

Article history:Received 5 December 2007Received in revised form23 June 2008Accepted 25 August 2008Available online 21 September 2008

Keywords:Masonry structuresMasonry walls

a b s t r a c t

Results of tests conducted fotesting followed the protocowhich is similar to that useand deformation characterisanalysis and design were alsaccording to NTCM-2004 gu0141-0296/$ see front matter 2008 Elsevier Ltd. All rights reserved.doi:10.1016/j.engstruct.2008.08.015s 31 (2009) 240259

ble at ScienceDirect

Structures

evier.com/locate/engstruct

masonry walls

or H. Salinas-Vallejov. San Pablo 180, Col. Reynosa Tamaulipas, 02200 Mxico, D. F., Mexico

combined and confinedmasonrywalls are reported in this paper. The cyclicl established by Mexican guidelines for masonry structures (NTCM-2004),worldwide for the cyclic testing of wall structures. Resisting mechanismstics of such walls were evaluated. Indicative values of useful parameters fordefined. In addition, it was verified if such a system is earthquake-resistantdelines.

2008 Elsevier Ltd. All rights reserved.

nFig. 3. Old combined masonry building in Amsterdam, Netherlands.

blocks [18,19,30,31]. Experimental testing of confined masonrywalls has also been carried out in Chile [5,32,33], Venezuela [32,

non-industrial fired clay bricks and lightweight concrete blockswith no quality control, which dimensions are depicted in Fig. 5.The mortar bed joint ranges from 1 cm (3/8) to 1.5 cm (5/8)

in thickness. Head joints are filled with mortar and they areusually 1 cm (3/8) thick. The mortar mix used by the people hasthe following volumetric proportions:1:2:6 (cement:lime:sand),clearly a mix that is out of what it is recommended in masonrycodes for seismic regions [40,4244].Therefore, it was also necessary to assess physical and

mechanical properties of the materials and the masonry used inthis type of construction, as well as determine these properties ifa code-based mortar mix is used. These testing are documentedin detail in Jurez [16] and Salinas [17] and are summarized infollowing sections.

2.1. Test for bricks, blocks and mortar

Index properties for the bricks, blocks andmortarwere assessedusing current Mexican guidelines [40,41,45] and are summarizedA. Tena-Colunga et al. / Engineeri

(a) Most common configuration.

Fig. 1. Combined and confined masonry

Fig. 2. Old combined masonry construction in Istanbul, Turkey.34], Peru [4], Argentina [32,35], Slovenia [7,36], Portugal [37],Japan [12] and India [38,39].g Structures 31 (2009) 240259 241

(b) Other common configuration.

construction currently built in Mexico.

Since there were no tests available for the described combinedand confined masonry walls, Mexicans have no informationabout the performance of such walls under alternated earthquakeloading, other than the satisfactory performances observed for oneand two stories houses at small towns in Puebla and Tlaxcalastates during the moderate June 15, 1999 Tehuacn earthquake(M = 6.5). The described system is being used in seismic regionsof Mexico where the earthquake hazard is high, and the number ofapplications is growing very fast. In fact, this system is starting tobe used inMexico City aswell. Therefore, an experimental programwas needed in order to evaluate the strength and deformationmechanisms of such walls when subjected to strong lateral cyclicloading.The results of the material characterization and the cycling

testings conducted for combined and confined masonry wallsare reported in this paper. Cyclic testings followed the protocolestablished by Mexican guidelines [40,41], which is similar to thatused worldwide for the cyclic testing of wall structures. Resistingmechanisms and deformation characteristics of such walls wereevaluated in the research. Values of useful parameters for analysisand design were also defined.

2. Properties for combined and confined masonry

The combined and confined masonry construction currentlyused in Mexico for non-engineered construction is composed ofin Table 1. Mexican standards are similar to ASTM guidelines,particularly in testing procedures; however, they differ in the

nsampling and statistical criterion, particularly to definemechanicaldesign properties denoted with and asterisk (*).For example, the design compressive strength of amasonry unit

according to NTCM-2004 [40,41] should be computed as:

f p =fp

1+ 2.5cp (1)

where f p is themean value from test results and cp is the coefficientof variation of the test results that shall not be taken less thanrecommendedminimumvalues established byNTCM-2004. As thecp value obtained during test results [16,17] was smaller than theminimum values allowed by NTCM-2004, then minimum valuesestablished by NTCM-2004 [40] were used, that is, cp = 0.3 wasused for concrete blocks (mechanized production with no qualitycontrol) and cp = 0.35 for the fired clay bricks (non-industrialproduction).Two different mortar mixes were used for the tests. In the

first set of testing [15,16] a volumetric mortar mix 1: 14 :3 12

Table 2. Standard cubes were used to determine compressivestrength of the mortar. For mortar type I, a coefficient of variationcj = 0.32 was obtained from tests and cj = 0.33 for non-engineered mortar. These values were used to define designstrength f j according to NTCM-2004. It is worth noting thatmortar type I satisfied the requirement of NTCM-2004 of havinga compressive strength f j 125 kg/cm2 (12.3 MPa), and thatthe non-engineered mortar would satisfy NTCM-2004 minimumdesign strength requirement for structural use of having f j 40 kg/cm2 (3.9 MPa), despite the fact that this mortar mix doesnot satisfy the minimum volumetric proportions established byNTCM-2004.

2.2. Axial compression prism tests for combined masonry

A set of masonry prisms was constructed to define thecompressive strength and Youngs modulus for the combinedmasonry, following the general guidelines and requirements242 A. Tena-Colunga et al. / Engineeri

(a) Downtown Mrida.

Fig. 4. Old combined mas

(a) Typical clay brick used. (b) Typical concrete block used.

Fig. 5. Pieces commonly used in recent combined and confined masonryconstruction.

Table 1Index properties for bricks and blocks

Property Bricks Blocks

Number of tested units 19 18Volumetric weight (ton/m3) 1.57 1.08Absorption 18.3% 26.5%Initial rate of absorption (gr/min) 59.4 32.7Saturation coefficient 0.94 0.94Modulus of rupture, fr (MPa) 0.86 0.96Compressive strengthMean: f p (MPa) 10.1 4.2Design: f p (MPa) 5.4 2.4(cement:lime:sand) was used. This mortar mix proportion is thebest one recommended by NTCM-2004 [40,41] for structuralg Structures 31 (2009) 240259

(b) Downtown Mexico city.

onry buildings in Mexico.

Table 2Index properties for mortar

Property Type I Non-engineered

Number of cubes 30 35Volumetric weight (ton/m3) 1.57 1.51Compressive strengthMean: f j (MPa) 24.1 7.8Design: f j (MPa) 13.4 4.3

applications and it is named mortar type I. The purpose of usingthis mix is that most lateral loading cyclic tests conducted inMexico for confined masonry walls have used this mortar type [1,2,1831], and cross comparisons with the dominant masonryconstruction in Mexico is also desirable.In the second set of testing [17] a volumetric mortar mix

1:2:6 (cement:lime:sand) was used, as this is the mix that peopleis currently using to build non-engineered houses and evenbuildings. It is worth noting that the proportion used for non-engineered construction in Mexico does not satisfy the minimumvolumetric requirements proposed by NTCM-2004, but it is usedas it is an inexpensive mortar and workability is good. However,it is also worth noting that this mortar mix has better volumetricproportioning than mortar type O (1:2:9) allowed by masonrycodes of the United States (for example, ACI 530-05 [43]) for non-seismic regions.The properties obtained for the mortar are summarized inprovided by NTCM-2004. Both bed and head joints were filledwithmortar.

nA. Tena-Colunga et al. / Engineeri

(a) Arrangement 1. (b) Arrangement 2.

Fig. 6. Prisms arrangements for walls built with mortar type I.

Table 3Index properties from prism tests using mortar type I

Arrangement f m (MPa) cm f m (MPa) Em (MPa)

1 3.1 0.15 2.2 12332 2.6 0.17 1.8 11711 and 2 2.8 0.18 1.9 1201

(a) Arrangement 3. (b) Arrangement 4.

Fig. 7. Prisms arrangements for walls built with non-engineered mortar.

Given the particularity of the combined masonry, where bricklayers alternate with block layers, and that bricks and blocks havevery different properties (Table 1), two different arrangementswere constructed for each mortar type used. It is worth notingthat guidelines worldwide for masonry prism tests assume thatthe same material (brick or block) is used to build the prisms, socombined masonry is not really fully addressed at this time.For mortar type I, prism arrangements depicted in Fig. 6 were

selected to assess if there are important differences among themat the time of defining mechanical properties for the masonry. Thearrangements were selected based upon the combined masonrywalls that were finally tested (Fig. 13a and b). Nine prisms weretested for each arrangement, according to the minimum requiredby NTCM-2004 [40,41]. The slenderness ratio for the prisms was4.78, within the range 2 h/t 5 established by NTCM-2004.Test results for prisms with mortar type I are summarized in

Table 3, where it can be observed that although higher valuesfor the mechanical properties are obtained under arrangement1, differences are no significant: around 20% for compressivestrength and 5% for the modulus of elasticity. Therefore, forpractical purposes it would be convenient to take the weightedproperties obtained from the data of both arrangements, that is,f m = 19.7 kg/cm2 (1.9 MPa) and Em = 12,245 kg/cm2 =1201 MPa (Em = 621.6f m).For the non-engineered mortar, prism arrangements depicted

in Fig. 7 were selected, based upon the combined masonrywalls that were finally tested (Fig. 13a and c). Nine prisms withslenderness ratio 4.78 were also tested for each arrangement.

Test results for prisms with non-engineered mortar are

summarized in Table 4, where it can be observed that there are nog Structures 31 (2009) 240259 243

Table 4Index properties from prism tests using non-engineered mortar

Arrangement f m (MPa) cm f m (MPa) Em (MPa)

3 2.3 0.20 1.5 16044 2.3 0.20 1.5 14583 and 4 2.3 0.19 1.5 1527

Fig. 8. Axial compression test of NTCM-2004 to determine vm and Gm .

differences between arrangements 3 and 4 for practical purposes.Therefore, the weighted properties obtained from the data ofboth arrangements are f m = 15.7 kg/cm2 (1.5 MPa) and Em =15,572 kg/cm2 = 1527 MPa (Em = 991.8f m).It is worth noting that there are no significant differences

between the design compressive strength f m for the prisms builtwith mortar type I and non-engineered mortar, as it is around25.5%, perhaps good enough for design under gravitational loading.

2.3. Diagonal compression wallet tests for combined masonry

According to NTCM-2004 [40,41], small square masonrysubassemblies (wallets) as depicted in Fig. 8 can be tested underaxial compression (commonly using an universal press machine)in order to define an indirect shear (diagonal tension) strength vmfor design and the shear modulus Gm.The design strength vm is computed as:

vm =vm

1+ 2.5cv (2)where vm is the mean value from test results and cv is thecoefficient of variation of the test results that shall not be takenless than 0.20.Given the particularity of the combined masonry and the

reasons stated in the previous section, two different arrangementsofwalletswere also constructed for the diagonal compression testsfor each mortar type used.For mortar type I, the arrangements depicted in Fig. 9 were

selected to assess if there are important differences amongthem at the time of defining these mechanical properties forthe masonry. The arrangements were selected based upon thecombined masonry walls that were finally tested (Fig. 13a andb). Nine specimens were tested under diagonal compression foreach arrangement, according to the minimum required by NTCM-2004. The aspect ratio for the specimens was almost square, asrecommended by NTCM-2004.Test results for specimens jointed with mortar type I are

summarized in Table 5, where it can be observed that, in contrastto what was observed from prism tests, important differences are

obtained for shear strength indices values vm between the smallwallet arrangements 1 and 2. The most notorious scatter of the

n244 A. Tena-Colunga et al. / Engineeri

(a) Arrangement 1. (b) Arrangement 2.

Fig. 9. Small wallet arrangements built with mortar type I.

(a) Sliding failure, arrangement 1. (b) Diagonal tension failure,arrangement 2.

Fig. 10. Dominant failure modes observed in diagonal compression tests ofspecimens built with mortar type I.

Table 5Index properties from diagonal compression tests using mortar type I

Arrangement vm (MPa) cv vm (MPa) Gm (MPa)

1 0.28 0.51 0.13 4202 0.38 0.18 0.25 3931 and 2 0.33 0.37 0.17 415

data and the smallest values were obtained for arrangement 1.These can be explained by the different predominant modes offailure observed in the tested specimens. For arrangement 1, thepredominantmode of failurewas sliding along bed joints (Fig. 10a),whereas for arrangement 2 the predominant mode of failure wasdiagonal tension (Fig. 10b).In contrast, no important differences were observed between

arrangements 1 and 2 to obtain the average shear modulus Gm(Table 5).For the non-engineered mortar, the arrangements depicted in

Fig. 11were selected, based upon the combinedmasonrywalls thatwere finally tested (Fig. 13a and c). Nine square specimens werealso tested for each arrangement.Test results for specimens jointed with non-engineered mortar

are summarized in Table 6, where it can be observed thatdifferences are also obtained for shear strength indices valuesbetween the small wallet arrangements 3 and 4. The mostnotorious scatter of the data and the smallest valueswere obtainedfor arrangement 4. However, in contrast to what was observed formortar type I, a predominant mode of failure was not observed inthe tested specimens, as failures in diagonal tension and slidingalong bed joints were observed for both arrangements 3 and 4(Fig. 12).A 35% difference is observed between the average shearmodulus Gm obtained for arrangements 3 and 4 (Table 6), incontrast to what was observed for mortar type I.g Structures 31 (2009) 240259

(a) Arrangement 3. (b) Arrangement 4.

Fig. 11. Small wallet arrangements built with non-engineered mortar.

Table 6Index properties from diagonal compression tests using non-engineered mortar

Arrangement vm (MPa) cv vm (MPa) Gm (MPa)

3 0.27 0.29 0.16 4174 0.25 0.44 0.12 3103 and 4 0.26 0.35 0.14 367

Test results lead one to conclude that there is a reasonabledoubt on how representative is the axial compression test to defineshear strength indices for combined masonry, which can only beanswered by relating these index values to estimate the lateralshear strength of wall specimens subjected to lateral loading, asshown in following sections.

3. Construction, instrumentation and setup of test specimens

Four walls were constructed and tested under cyclic lateralloading. The cryptogram for identification of each wall is MCC-i,where i is an index to identify the number of wall sequentiallytested. The first two walls (MCC-1 and MCC-2) were jointed withmortar type I and the last twowalls (MCC-3 andMCC-4) with non-engineered mortar.The geometry of the walls is schematically depicted in Fig. 13.

The general dimensions of walls and their confinement elementswere selected to make these walls as close as possible to thedimensions of a confined brick masonry wall previously tested atCenapred [23,27] (wall M-0-E6), to do some cross comparisons. Infact, the confining RC tiecolumns, bondbeam and beam on gradehave the same dimensions, reinforcement and concrete strengthwith respect to those of the wall of Reference [23,27].Confining tiecolumn elements are 12 20 cm with the

reinforcement depicted in Fig. 14 and specified in Table 7.The confining rectangular bond-beam is 12 20 cm with thereinforcement depicted in Fig. 14 and specified in Table 7. A slab10 80 cm was cast at the top of the confining beam, as alsodepicted in Fig. 14 and specified in Table 7. The compressivestrength of the concrete used for the confining tiecolumns andbondbeams was specified as f c = 150 kg/cm2 (14.7 MPa) and forthe slab was f c = 250 kg/cm2 (24.5 MPa). Results of controlledcylinder tests are reported elsewhere [16,17].The external instrumentation for the walls was composed

of 8 LVDT (Fig. 15a) and a load cell (Fig. 16). The internalinstrumentation consisted of 12 strain gages placed on thelongitudinal reinforcement of the confining elements, as depictedin Fig. 15b.The testing of the walls was done with the help of the reaction

braced frame and strong floor schematically depicted in Fig. 16. Aspecial C steel beamwas designed to apply a uniformly distributedvertical loading of 1 ton/m (9.81 kN/m) and to help applying witha hydraulic jack the cyclic lateral loading.

The cycling testing of the walls was done following the general

guidelines of Appendix A of NTCM-2004 [40,41] required for

nFig. 14. Reinforcement of confining elements.

Table 7Reinforcement of the confining RC elements of walls

Columns Beam SlabLongitudinal Transverse Longitudinal Transverse Longitudinal Transverse

4#3 S#2 @ 20 cm 4#3 S#2 @ 20 cm #3 @ 20 cm #4 @ 20 cm

#3 bars = bars3/8 in diameter, fy = 412 MPa.#2 bars = bars 2/8 in diameter, fy = 216 MPa.

earthquake-resistant masonry wall systems. According to thetesting protocol of reference, masonry walls have to be subjectedto repeated cycles of at least 25%, 50% and 100% the estimatedcracking load for the walls (load control, Fig. 17) and, after thefirst cracking, walls have to be subjected to repeated cycles

control were done using additional steps, as schematicallyillustrated in Fig. 17.The expected nominal shear load at cracking for each wall was

estimated from the design equations of NTCM-2004 neglectingstrength reduction factors and using the experimental dataFig. 12. Failure modes observed in diagonal compression tests of specimens built with non-engineered mortar.

(a) Walls MCC-1 and MCC-3. (b) Wall MCC-2. (c) Wall MCC-4

Fig. 13. General geometry of tested walls.(a) Arrangement 3. (b) Arrangement 4.A. Tena-Colunga et al. / Engineeriof increasing drift ratios of at least 0.2% (displacement control,Fig. 17). In this research, both the load and the displacementg Structures 31 (2009) 240259 245from axial compression prisms and diagonal compression wallettests previously described. Two potential failure conditions were

nFig. 17. Cyclic testing protocol used.

considered: (a) a diagonal tension shear failure, according to the the simplified interaction diagram specified for confined masonryFig. 16. Testing setup.(a) External wall instrumentation. (b) Internal wall instrumentation.

Fig. 15. Instrumentation for the walls.246 A. Tena-Colunga et al. / EngineeriMohrCoulomb design expression for confined masonry of NTCM-2004 and, (b) a flexural tension or compression failure, according tog Structures 31 (2009) 240259by NTCM-2004. The details of these calculations are reportedelsewhere [1517] and suggested that the expectedmode of failure

n(a) Wall MCC-1. (b) Wall MCC-2.

(c) Wall MCC-3. (d) Wall MCC-4.

Fig. 18. Hysteretic loops for the tested walls.

Table 8Estimated shear strength at cracking (kN) for tested walls according to NTCM-2004

MCC-1 MCC-2 MCC-3 MCC-4

41.2 41.2 28.4 25.5

for all walls was a diagonal tension shear failure, associated thento the MohrCoulomb design equation of NTCM-2004 [40,41]:

VmR = FR(0.5vmAT + 0.3P) 1.5FRvmAT (3)where FR is the force reduction factor, vm is the indirect tensileshear strength from the diagonal compression wallet tests, P isthe applied axial load and AT is the area of the confined masonrywall including the area of the confining elements without anysection transformation, as for simplicity, this design equation wascalibrated this way using previous experimental tests conductedin Mexico for confined masonry walls. Therefore, for all the wallsspecimens FR = 1, P = 2.2 ton (21.6 kN),AT = 2760 cm2. For cyclic

(0.25 MPa) for walls MCC-1 and MCC-2, vm = 1.6 kg/cm2(0.16 MPa) for wall MCC-3 and vm = 1.4 kg/cm2 (0.14 MPa) forwall MCC-4. Estimated values for the nominal shear strength atcracking are reported in Table 8.

4. Experimental results

4.1. Hysteretic behavior, response envelopes and cracking patterns

The hysteretic loops obtained for each wall from the cyclictesting protocol described before are depicted in Fig. 18. Giventhe limitations of lab equipment at the time of each testing, it isworth noting the following observations regarding the testing ofeach wall. During the testing of wall MCC-1 a load cell was notavailable, so applied lateral forces were obtained indirectly from apreviously calibratedmanometer, that is why the curves in Fig. 18alook rough. During the testing of wall MCC-2 the hydraulic jackA. Tena-Colunga et al. / Engineeritesting protocol purposes, the highest value of vm obtained fromdiagonal compression tests was taken, therefore vm = 2.6 kg/cm2g Structures 31 (2009) 240259 247stock for positive cycles (pulling) after a lateral drift angle of 0.4%,so that is the reason of the asymmetric loops, which are marked

nhhead joints that allowed it to sustain stability up to a drift angle of1%, after that the wall failed because of the shear crack penetrationof the confining RC column (dotted red lines, Fig. 18d).Response envelopes for the tested wall are depicted in Fig. 20

and the characteristic parameters that define cracking and the peakresponse of such envelopes are reported in Table 9.The final cracking patterns for all tested walls are depicted in

Fig. 21, where it can be observed that they are typical diagonaltension shear patterns. It is also worth noting that the shear crackthat penetrated the confining column elements started to appearin all walls at a drift angle as low as 0.3%.

curves.It can be observed from Fig. 22 that there are no significant

differences between the curves of the first and second cycles for allwalls except wall MCC-4. The lateral stiffness of all walls degradesvery fast, being 0.4 or less of its initial uncracked stiffness Ke for adrift angle as low as 0.2%. For the drift angle where the shear crackstarts to penetrate the confining column elements ( = 0.3%), theK/Ke ratio is lower than 0.3 for the walls joined with engineeredmortar (walls MCC-1 and MCC-2, Fig. 22a and b), whereas thewalls joined with non-engineered mortar developed a K/Ke ratiohigher than 0.3 (walls MCC-3 and MCC-4, Fig. 22c and d). Spalling248 A. Tena-Colunga et al. / Engineeri

Table 9Characteristic parameters from response envelopes

Wall First cracking Peak sVcr (kN) cr V+max(k

MCC-1 49.0 0.0009 80.4MCC-2 49.0 0.0009 71.6aMCC-3 39.2 0.0006 74.5MCC-4 21.6 0.0006 58.8a Previous reported problem with the hydraulic jack during testing.

(a) Top right corner. (b) Bottom left corner.

Fig. 19. Shear crack penetration of confining RC columns of wall MCC-1 that leadto failure.

with a red-broken line. Fortunately, therewere no further technicalproblems during the testing of walls MCC-3 and MCC-4.From all the hysteretic loops it is observed a reasonably

symmetric and stable response for positive and negative cycles upto a drift angle of 0.6%, except for wall MCC-2 for the reason statedabove (loops were rather symmetric up to a drift angle of 0.4%,where the jack problem aroused).Instability in thewall response triggered for a drift angle of 0.6%

for wall MCC-1 after a major shear crack penetrated the confiningcolumn (Fig. 19); this last half cycle is marked with a broken redline in Fig. 18a.The instability of wall MCC-2 also started because of a shear

crack penetration in the confining column at a drift ratio around0.6%, where an important strength and stiffness degradation wasobserved (Fig. 18b).For wall MCC-3 an important pinching behavior triggered after

a lateral drift of 0.4% (Fig. 18c) and the important shear crackpenetration of the confining column started at a drift of 0.5%, thetesting was stopped at a drift of 0.6% as extensive damage withcrack penetration of top and bottom of the confining column wasevident.ForwallMCC-4, an important pinching behaviorwas developed

after a drift angle of 0.4% (Fig. 18d), basically a frictionless slidingmechanism along bed and head joints during load reversalsof a stair-like shear cracking pattern (significant gaps wereobserved during testing, particularly at the head joints). Althoughthe important shear crack penetration at the confining columnhappened at a lateral drift of around 0.6%, this wall presented avery important frictionless sliding mechanism along the bed andFrom a gross assessment viewpoint, from the crack patterns,the obtained hysteretic loops and the response envelopes, it cang Structures 31 (2009) 240259

ear forceN) + Vmax(kN)

0.0046 76.5 0.00540.0038a 82.4 0.00560.0050 60.8 0.00310.0090 57.9 0.0070

be observed that the main difference in the cyclic behavior of thewalls jointed with mortar type I of NTCM-2004 (walls MCC-1 andMCC-2) and thewalls jointedwith non-engineeredmortar is due tothe diagonal crack pattern exhibited for the walls. A well-definedmain diagonal crack that crosses and breaks bricks and block layersare developed for the walls jointed with the stronger engineeredmortar (wallsMCC-1 andMCC-2, Fig. 21a andb). In contrast, a stair-like, frictionless sliding joint cracking patterns are developed forthe walls jointed with the weaker non-engineered mortar (wallsMCC-3 and MCC-4, Fig. 21c and d), particularly for wall MCC-4(Fig. 21d). This frictionless sliding joint mechanism during loadreversals was favored because of the relatively low applied normalstress. That is the reason why more pinching is observed in thehysteretic loops of the walls joined with non-engineered mortar(Fig. 18).Walls MCC-1 andMCC-2 first cracked at a higher drift angle and

shear force than walls MCC-3 and MCC-4 (Table 9). In fact, wallsMCC-1 andMCC-2 also developed slightly higher shear forces thanwalls MCC-3 and MCC-4 (Table 9, Fig. 20). However, the differencebetween the shear force resisted by walls MCC-1 and MCC-3, thathave the same configuration, is only 8%.The major difference regarding peak shear forces is observed

between walls MCC-3 and MCC-4, which were both jointed withnon-engineeredmortar, where the difference is 26% (Table 9). Thishigher difference seems directly related to the fact that a differentmasonry combination exist where more blocks (weakest material)are used in wall MCC-4 with respect to wall MCC-3.It is worth noting that estimates of shear forces at cracking

using NTCM-2004 design equation (Eq. (3)) and the results fromthe axial compressive tests (Table 8) are somewhat reasonable andconservative (except for wall MCC-4) when compared with thoseobtained from the response envelopes (Table 9).

4.2. Cyclic stiffness degradation

An important parameter for the design and evaluation of struc-tures, masonry included, is to assess their stiffness degradationwith respect to increasing drift angle. This parameter is useful tohelp define drift angle limits for design purposes (i.e., serviceabil-ity, collapse prevention, etc.). Therefore, from the hysteretic curvesdepicted in Fig. 18, peak-to-peak secant stiffnesses were definedfor each cycle at the same drift angle, and then normalizedwith re-spect to the computed initial (elastic) stiffness for each walls. Thecomputed curves for all walls are depicted in Fig. 22, where firstand second cycles at the same drift angle are plotted in separateof masonry started close to a drift angle of 0.4% or less, where allwalls have K/Ke < 0.25.

n(a) Wall MCC-1. (b) Wall MCC-2.

(c) Wall MCC-3. (d) Wall MCC-4.

Fig. 20. Response envelopes for the tested walls.

5. Earthquake-resistant qualification according to NTCM-2004

In Appendix A of NTCM-2004 [40,41] criteria is set to qualifya masonry wall system as earthquake-resistant, based upon thecyclic testing under lateral loading and the material testingdescribed in previous sections, for a given amount of gravitationalloading under consideration.Besides a detailed protocol and documentation of the testing

(data, pictures, etc.), basically, three main requirements must beevaluated from the information obtained from cyclic and materialtesting, which are the following for walls made with solid units:

A1. The maximum shear strength obtained from cyclic testing ofthe wall (Vmax) reached at a lateral drift angle 0.006should be equal or greater than the shear strength V estimatedfrom diagonal compression wallet tests.

A2. Vmax V , where is an overstrength factor that takes intoaccount the wall connecting details, for example, intersectingwalls, foundations, floor systems, etc. A minimum value of 1.3is recommended to use in lieu of a detailed calculation.

a. The shear force developed, V2, must be at least 0.8V 1maxobtained in the same loading direction.

b. The normalized peak-to-peak secant stiffness ratio must begreater than 0.1 (K/Ke 0.1).

c. The normalized equivalent energy dissipation ratio at thatcycle must be greater than 0.15. This equivalent energydissipation ratio is defined as the actual energy dissipatedin an hysteretic loop at a drift angle of 0.6% divided by theenergy that an equivalent elastic-perfectly plastic systemwith the elastic stiffness of the tested specimen woulddissipate at the same drift angle, as schematically shown inFig. 23 for the results of Wall MCC-3.

These requirements were assessed from the results of thedescribed tests as reported in detail in Jurez [16] and Salinas [17]and are summarized in Tables 10 and 11. According to theresults presented there, the combined and confined masonrywalls jointed with engineered mortar (mortar type I) are qualifiedas earthquake-resistant with the criteria of NTCM-2004 if theexpected applied gravitational load is 1 ton/m (9.8 kN/m),A. Tena-Colunga et al. / EngineeriA3. The characteristics of the second cycle at a drift angle =0.006 must satisfy the following criteria:g Structures 31 (2009) 240259 249that is the one used for the testing and it will be the one expectedfor housing up to two stories in height.

(c) Wall MCC-3. (d) Wall MCC-4.

Fig. 21. Final cracking patterns of tested walls.

Table 10Evaluation of the requirement A1 and A2 of NTCM-2004 to qualify masonry wallsas earthquake-resistant

Wall Vmax (kN) V (kN) V (kN) Requirement A1 Requirement A2

MCC-1 80.4 41.2 53.9 3 3MCC-2 82.4 41.2 53.9 3 3MCC-3 74.5 28.4 37.3 3 3MCC-4 55.9 25.5 33.3 3 3

It can also be observed from these tables that the combined andconfined masonry walls jointed with non-engineered mortar donot satisfy all the criteria to be qualified as earthquake resistantby NTCM-2004, primarily because the frictionless sliding alongbed and head joints causes the pinching behavior that dissipatesvery little energy (because the effective compressive stress islow) according to the equivalent energy dissipation ratio criterionof NTCM-2004, in addition to the fact that important strengthdegradation is observed at a drift angle = 0.006.

5.1. Drift angle for seismic design

Limiting values for the drift angle for the seismic design ofmasonry wall systems are specified in NTCM-2004 [40,41] forthe most common masonry modalities used in Mexico, based onprevious experimental testing. For example, for confined masonrywalls made of solid clay bricks, the proposed limiting drift anglefor seismic design is = 0.0025, and if thesewalls have horizontalshear reinforcement along the bed joints (common in Mexico), thelimiting drift angle for seismic design is = 0.0035.These values have been proposed based on the criteria that

the extent of damage should be limited for a major earthquakerather than corresponding to an ultimate limit collapse preventionstate [40,41], as confined masonry is the dominant modality ofconstruction in urban areas for housing, which account for upto 80% of the engineered construction in Mexico nationwide.Therefore, most masonry buildings and housing should be ableto carry gravitational loads after a major earthquake, have areasonable reserve of seismic resistant and being able of being

Table 11Evaluation of requirement A3 of NTCM-2004 to qualify masonry walls as earthquake-resistant

Wall V 1max (kN) V2 (kN) 0.8V1max (kN) K/Ke EEDR Requirement A3

a b c

MCC-1 76.5 74.5 60.8 0.10 0.233 3 3 3MCC-2 82.4 72.6 67.7 0.11 0.319 3 3 3(a) Wall MCC-1.MCC-3 60.8 40.2 49.0MCC-4 55.9 42.2 45.1(b) Wall MCC-2.250 A. Tena-Colunga et al. / Engineering Structures 31 (2009) 2402590.11 0.108 3 0.12 0.089 3

n(a) Wall MCC-1. (b) Wall MCC-2.

(c) Wall MCC-3. (d) Wall MCC-4.

Fig. 22. Normalized peak-to-peak secant stiffness vs. drift angle for tested walls.

repaired using low-cost technologies, otherwise, Mexico wouldnot be able to afford recovery after a major earthquake.Therefore, a preliminary limiting drift angle for seismic design

for combined and confined masonry was assessed following asimilar criterion that the NTCM-2004 code committee followed.Therefore, the limiting drift angle should the less one obtainedfrom the following criteria:

D1. The drift angle before spalling or important bed joint slidingwas first observed in the masonry walls during cyclic testing.

D2. The minimum of the drift angles (+ or ) associated toobserved peak shear forces (V+max or Vmax) from responseenvelopes.

D3. The drift angle associated to K/Ke 0.2.Based on the detailed documentation of the tests, these criteria

are evaluated and the obtained values are summarized for all

Table 12Definition of the drift angle for seismic design

Wall Observed drift angle associated to criteria Drift for design, D1 D2 D3

MCC-1 0.003 0.0046 0.0030 0.003MCC-2 0.003 0.0038 0.0040 0.003MCC-3 0.003 0.0031 0.0046 0.003MCC-4 0.003 0.0070 0.0040 0.003

design for combined and confined masonry should be around =0.003. The limiting criteria for the walls jointed with engineeredmortar was spalling of the masonry, whereas for the walls joinedwith non-engineered mortar was the first observed importantsliding along bed joints. The extent of damage at this drift limitA. Tena-Colunga et al. / Engineeriwalls in Table 12. It is observed from the test data that there isa consensus agreement that the limiting drift angle for seismicg Structures 31 (2009) 240259 251is depicted in Fig. 24, where it can be observed that it is indeedreparable at low cost.

n6. Damage index

Structural engineers need simple quantitative evaluationmeth-ods to assess the safety of existing structures under an expectedlevel of seismic demand, particularly to take decisions regardingrehabilitation and/or demolition. Damage indices are suitable toolsfor such purposes.Therefore, the damage index originally proposed by Kwok and

Ang [46] for unreinforced brick masonry walls based on a reliabledatabase of experiments conducted in China in the late 1970s wascalibrated with the experimental data obtained in this researchstudy for confined and combined masonry walls.The KwokAng damage index is defined with the following

general expressions:

D = Du + De (4)Du = umuf (5)

De = dEquuf

(6)

where um is the displacement at maximum load, uf is thedisplacement at failure,

dE is the total (cumulative) dissipated

energy, qu is the ultimate shear strength and is a constantobtained from regression analysis of experimental data.

walls. In addition, they proposed the following damage scale fortheir index: (a) No damage:D = 0, (b) Reparable damage: 0 < D 0.25, (c) Severe damage: 0.25 < D < 1.0, (d) Collapse: D 1.It is clear that there are some differences in the cyclic behavior

under lateral loading of plain unreinforced masonry and thecombined and confined masonry described in this study, so someadjustments to the damage index originally proposed by Kwokand Ang were needed for combined and confined masonry walls.For example, the value of constant should be assessed fromexperimental data. Also, the limiting boundary value betweenreparable and severe damage should be also redefined with theobserved experimental behavior to take into account the beneficialpresence of confining elements that provide further stability inthese walls under lateral loading.Therefore, from the reduced database of the experiments

described here, a preliminary value of = 0.046 was obtained forcombined and confinedmasonrywalls, as reported elsewhere [1517]. Also, from the analysis of experimental data that alloweddefining the limiting lateral drift ratio = 0.003 for earthquake-resistant design purposes, it was clear that the limiting boundaryvalue between reparable and severe damage should beD = 0.4 forcombined and confined masonry. Therefore, the proposed damageindex scale for combined and confinedmasonry is: (a) No damage:D = 0, (b) Reparable damage: 0 < D 0.4, (c) Severe damage:0.4 < D < 1.0, (d) Collapse: D 1. The obtained damageFig. 24. Extent of damage at a limiting drift angle = 0.003.252 A. Tena-Colunga et al. / Engineeri

Fig. 23. Definition of the equivalent energy dissipation ratioKwok and Ang [47] obtained from the regression analysis oftheir extensive database that = 0.075 for unreinforced masonryg Structures 31 (2009) 240259

of NTCM-2004, illustration with the results for wall MCC-3.index vs. drift ratio curves for the tested combined and confinedmasonry walls are depicted in Fig. 25, where it can be observed

n(a) Geometry. (b) Final damage pattern (1% drift).

Fig. 26. Wall M-0-E6 tested at Cenapred [23,27].

Table 13Comparison of index values obtained from cyclic testing

Wall Masonry type Elastic Stiffness (kN/cm) Vcr (kN) (Drift) V+max (kN) (Drift) Vmax (kN) (Drift) Maximum driftK+ K

MCC-1 Combined and confined (bricks & blocks) 586.4 615.9 49.0 80.4 76.5 0.006(0.0009) (0.0046) (0.005)

MCC-2 Combined and confined (bricks & blocks) 346.2 428.5 49.0 71.6 82.4 0.008(0.0009) (0.0038) (0.005)

M-0-E6a Confined (bricks) 471.7 415.8 94.6 133.9 147.1 0.01(0.00100) (0.00455) (0.00415)

a M-0-E6: from Flores and Alcocer [27].

the proposed adjustments for the KwokAng damage index seemreasonable for both the walls jointed with engineered (Fig. 25a)and non-engineered (Fig. 25b) mortar.

7. Comparison with confined brick masonry

As stated earlier, the general dimensions of walls MCC-1 toMCC-4 and their confinement elements were selected to make

27], depicted in Fig. 26, to do some cross comparisons withconventional confined masonry made with clay bricks. Wall M-0-E6 is actually slightly larger and taller than walls MCC-1 to MCC-4,because of limitations of the testing facility at UAM. Also, the M-0-E6 wall was tested with a higher uniformly distributed verticalloading of 5 ton/m (49 kN/m).Wall M-0-E6 was jointed with mortar type I of NTCM-2004,

therefore, the more meaningful comparisons are with walls MCC-(a) Walls with engineered mortar. (b) Walls with non-engineered mortar.

Fig. 25. Modified KwokAng damage index vs drift angle for combined and confined masonry walls.A. Tena-Colunga et al. / Engineerithese walls as close as possible to the dimensions of the confinedbrick masonry wall M-0-E6 previously tested at Cenapred [23,g Structures 31 (2009) 240259 2531 and MCC-2. The index mechanical properties from prisms andwallet tests reported for wall M-0-E6 were f m = 25 kg/cm2

n(c) Wall MCC-2.

Fig. 27. Comparison of hysteretic curves for confined masonry walls.

2 2 2Table 14Experimental average values for walls MCC-1 and MCC-2

Wall K0 (kN/cm) Vcr (kN) Vmax (kN) Vu (kN) H (cm) V max u

MCC-1 603.1 55.9 78.5 61.8 220 0.0048 0.006MCC-2 387.4 48.1 76.5 55.9 220 0.0046 0.007MCC-3 268.7 39.2 72.6 40.2 220 0.0048 0.0067MCC-4 301.1 21.6 58.8 53.9 220 0.0082 0.0110

Table 15Estimates of lateral drifts at the maximum shear force using the equivalent cracked wide column analogy proposed by Bazn and Meli [19,47]

Wall Ec (MPa) Em (MPa) Gm (MPa) Aceq (cm2) Vmax (kN) (cm) Keq (kN/cm) max

MCC-1 11,251 1200 415 2.86 887.9 80.4 0.93 86.5 0.0042MCC-2 9,863 1200 415 2.50 865.5 71.6 0.84 85.3 0.0038MCC-3 9, 200 1604 418 2.31 853.2 74.5 0.77 96.5 0.0035MCC-4 11,315 1469 314 3.78 946.5 58.8 0.71 83.3 0.0032

(a) Wall M-0-E6 [23,27]. (b) Wall MCC-1.254 A. Tena-Colunga et al. / Engineeri(2.45 MPa), vm = 3 kg/cm (0.29 MPa) and Gm = 4850 kg/cm(476 MPa) higher values than those obtained for MCC-1 and MCC-g Structures 31 (2009) 2402592 (Tables 3 and 5); however Em = 7420 kg/cm (728 MPa) isnotoriously smaller (Table 3).

nA. Tena-Colunga et al. / Engineeri

Fig. 28. Trilinear envelope for the model proposed by Flores and Alcocer [27].

It is observed that a similar cracking pattern and failure wasdeveloped in wall M-0-E6 (Fig. 26b) compared with walls MCC-1and MCC-2 (Fig. 21a and b), failing in diagonal tension, presentingthe most important cracks and spalling about the middle sectionof the wall, besides the penetrating crack of the confining columnelements at the top and bottom corners. However, it seems thatwallM-0-E6 experienced amoredistributed shear crackingpatternalong the wall than walls MCC-1 and MCC-2.The corresponding hysteretic curves for these walls are

compared in Fig. 27, where it can be observed that the confinedwall made with clay bricks (M-0-E6) resisted a higher shear forceand developed a higher deformation capacity than walls MCC-1andMCC-2. It isworth noting that the higher shear force developedby M-0-E6 wall can be explained primarily because: (a) a highervalue of vmwas obtained, (b) thiswall has a slightly largerwall area(8%) and, (c) the applied uniformly distributed vertical loading of5 ton/m (49.0 kN/m) is considerably higher than the 1 ton/m (9.8kN/m) vertical loading applied to MCC-1 and MCC-2 walls.Index values obtained from the cyclic testing of the walls are

compared in Table 13. It is worth noting that the lateral drift atfirst cracking is similar for traditional confined masonry and thecombined and confined masonry. The drifts related to obtainedpeak shear forces are also very close. The elastic stiffness forwall MCC-1 was higher than for wall M-0-E6, but wall MCC-2had a smaller elastic stiffness than wall M-0-E6; however, forpractical purposes, it seems that there are no important differencesregarding the elastic stiffness of these walls taking into accountuncertainties. In fact, there is amore important difference betweenthe elastic stiffness of walls MCC-1 and MCC-2 than between M-0-E6 with respect to either MCC-1 or MCC-2 (Table 13).It is observed an important difference in the strength at first

cracking betweenwallM-0-E6 andwallsMCC-1 andMCC-2, for thereasons stated above. In order to do a fair comparison regardingstrength, a normalization of the experimental results for wallsMCC-1 and MCC-2 was done using the proposed design equationof NTCM-2004 (Eq. (3)). From the reported Vcr value at Table 1, andtaking FR = 1 and the real values of AT and P from testing, anequivalent vm = 3.15 kg/cm2 (0.31 MPa) value from testing wasobtained forwallsMCC-1 andMCC-2 using Eq. (3). Then, taking thevalues of AT and P related to the testing of wall M-0-E6 in Eq. (3),it is obtained that the expected shear strength at cracking for wallsMCC-1 and MCC-2 is VmR = 9.2 ton (90.2 kN) for similar testingconditions with respect to wall M-0-E6, a closer value to the onereported in Table 13 for wall M-0-E6. Of course, there are morefactors that affect the developed shear strength in a cyclic testing,but this academic exercise of normalization suggests that perhaps

the differences could be much smaller than they appear to be fromFig. 27 and Table 13.g Structures 31 (2009) 240259 255

8. Simplified analytical modeling

Simplified models previously proposed to model confinedmasonry walls made with bricks were evaluated from differentviewpoints; these detailed studies are reported elsewhere [1517].In the following section are only presented the results that matchbetter the experimental testing.

8.1. Hysteretic model

Flores and Alcocer [24,27] proposed an experimentally-basedhystereticmodel for confinedmasonrywallsmadewith clay bricksand jointed with mortar type I based on a trilinear envelope curve(Fig. 28) that takes into account strength and stiffness degradation.The hysteretic model proposed by Flores and Alcocer is defined

in terms of the following six parameters:

K0 Initial elastic stiffnessVcr Force at first shear (diagonal tension) crackingVmax Maximum shear forceVu Ultimate shear forceH Height of the wallV max Drift at maximum shear forceu Drift at ultimate shear force.

Although a very reasonable correlation between analytical andexperimental curves was obtained using the parameter valuesproposed by Flores and Alcocer [27] for walls MCC-1 and MCC-2 [15,16], the best approximation is obtained with the parameterspresented in Table 14, which are average values from positive andnegative cycles that were defined from the experimental data ofwalls MCC-1 to MCC-4.The experimental and analytical curves obtained forwallsMCC-

1 and MCC-2, jointed with engineered mortar, are depicted inFig. 29, where it can be observed that the gross correlation is good.The experimental and analytical curves obtained forwallsMCC-

3 and MCC-4, jointed with non-engineered mortar are depicted inFig. 30, where it can be observed that the gross correlation is notvery good, as the pinching behavior observed during the testingdue to the frictionless sliding along bed and head joints cannot becaptured by the analytical model. This was somewhat expected, asin the model proposed by Flores and Alcocer, all walls consideredin their experimental databasewere jointedwithmortar type I andfailed in diagonal tension (i.e., Fig. 26b).

8.2. Equivalent wide column analogy

Bazn andMeli [19,47,48] proposed an equivalent crackedwidecolumn analogy (Fig. 31) for confined masonry walls from theanalysis of experimental data of several confined masonry wallsdesigned according to the Mexican practice in the 70s, given bythe following equations:

Ieq = Ac b2

2(7)

Aceq = (0.37 0.12 + 0.023) (Am + 2Ac) (8) = b

H(9)

= EcAcGmAm

(10)

where Ieq is the equivalent cracked moment of inertia, Aceq is theequivalent cracked area, Ac is the area of the confining verticalelement, Am is the area of the masonry, Ec is the elastic modulus of

the concrete of the confining elements, Gm is the shear modulus ofthe masonry is an aspect ratio parameter valid for the following

n(c) Experimental, MCC-2. (d) Analytical MCC-2.

Fig. 29. Hysteretic curves for walls MCC-1 and MCC-2, jointed with engineered mortar.

range: 0.75 2.5 and is a parameter that measures therelative axial stiffness of the confining elements with the shearstiffness of the masonry and is valid for the following range: 0.9 11.The equivalent cracked wide column analogy proposed by

Bazn and Meli is used extensively for the analysis and designof confined masonry structures in Mexico [48]. This analyticalmethod has also been recently reviewed with the experimentaldata of confined masonry walls and buildings tested in Mexicofrom 1990 to 2005 and found to be in good agreement with suchtesting. This method is used to estimate the equivalent secantstiffness of confined masonry walls at the expected maximumshear force, then, reasonable estimates of inelastic drifts for designpurposes can be obtained.Therefore, it is of interest to evaluate how the equivalent

cracked wide column analogy concept works for the testedcombined and confined masonry walls. The parameters for thecomputation of the equivalent cracked columnwere obtained fromexperimental data [16,17]. For all walls, Ac = 220 cm2, Am =

2 4

the lateral drift can be estimated as:

= VH3

3EmIeq+ VHGmAceq

(11)

Keq = V

(12)

= H. (13)

Results obtained from applying the equivalent cracked widecolumn analogy proposed by Bazn andMeli to estimate the lateraldrift at the maximum shear force are summarized in Table 15.Comparing these results to those experimentally obtained andreported as V+max and + in Table 9, one can conclude that theestimates forwallsMCC-1 andMCC-2 are extremely good,whereasthese estimates are not good enough for walls MCC-3 and MCC-4. There is a logical explanation for this fact, as walls MCC-1and MCC-2 were jointed with an engineered mortar similar tothe one used in the walls considered in the database used byBazn and Meli, and the failure mechanism was similar as well256 A. Tena-Colunga et al. / Engineeri

(a) Experimental, MCC-1.2280 cm , Ieq = 5292,000 cm and = 0.95. As the tested wallsare cantilever walls, the lateral deflection, the secant stiffness andg Structures 31 (2009) 240259

(b) Analytical, MCC-1.(diagonal tension). In contrast, walls MCC-3 and MCC-4 werejointed with non-engineered mortar that favored shear cracks

n(a) Experimental, MCC-3. (b) Analytical, MCC-3.

(c) Experimental, MCC-4. (d) Analytical MCC-4.

Fig. 30. Hysteretic curves for walls MCC-3 and MCC-4, jointed with non-engineered mortar.

with important sliding along the bed joints, which may explainthe under-prediction of max using the equivalent cracked widecolumn analogy proposed by Bazn and Meli.

9. Concluding remarks

This paper reported a complete experimental protocol forcombined and confined masonry walls, including their cyclictesting according to theMexican guidelines formasonry structures(NTCM-2004 [40,41]), which is similar to that used worldwide forthe cyclic testing of wall structures.This experimental program was needed as this type of

construction is starting to be used frequently in regions of highseismic risk and there were no previous experimental informationavailable about its seismic performance. Given that the combinedand confined masonry of reference is currently jointed with non-engineered mortar (auto-construction), it was important to assessits seismic performance as currently used, in addition to evaluateif the use of an engineered code-based mortar may lead to animproved seismic performance. Also, it was of interest to compare

From the experimental program summarized in this paper, thefollowing observations can be made:

1. The combined and confined masonry walls jointed with thenon-engineered mortar currently used in auto-construction donot satisfy all the criteria to be qualified as earthquake resistantwalls by theMexican NTCM-2004 guidelines, primarily becausethe frictionless sliding along bed and head joints favored bythe low applied normal compression stress causes a pinchingbehavior that dissipates very little energy according to theequivalent energy dissipation ratio criterion of NTCM-2004,in addition to the fact that important strength degradation isobserved at a drift angle = 0.006. However, if an engineeredmortar that complies with NTCM-2004 is used, this systemmayqualify as earthquake resistant according to NTCM-2004.

2. The limiting drift angle for seismic design for combined andconfined masonry jointed with engineered mortar should bearound = 0.003 following the criteria used in NTCM-2004for defining this limiting drift angle for other types of masonryconstruction.A. Tena-Colunga et al. / Engineerithis modality of construction with the most widely used confinedmasonry walls made with solid clay bricks.g Structures 31 (2009) 240259 2573. The combined and confined masonry walls jointed with en-gineered mortar showed many similarities (cracking patterns,

nreasonable energy dissipation characteristics. In contrast, in thewalls jointed with non-engineered mortar, the initial diagonaltension cracks that broke the masonry units lessened after adrift ratio = 0.003, when a stair-like shear mechanismswith frictionless sliding along the bed and head joints triggered,leading to an important pinching of the hysteretic loops andtherefore, a reduced energy dissipation characteristics due tothe low normal compression stress applied to the walls.

5. From the different wall combinations tested for combined andconfined masonry walls jointed with non-engineered mortar(wallsMCC-3 andMCC-4), itwas observed thatwallMCC-3, thathas more brick layers, resisted a peak shear force 26% higherthan wall MCC-4, that has more concrete blocks. Therefore,this difference seems to be directly related to the fact thata different masonry combination exists where more blocks(weakest material) are used in wall MCC-4 with respect to wallMCC-3.

6. Given that the mechanical properties of the concrete blockscurrently used in combined and confined masonry walls arevery weak, it will be worth testing in the future combined andconfined masonry walls where higher quality concrete blocksare used, to discern if this practice would lead to improvedperformances under lateral cyclic loading.

It was shown that the KwokAng damage index [46] forunreinforced brick masonry walls can be adjusted in a simple

masonry walls jointed with engineered mortar is also reliable toestimate the secant lateral stiffness and drift of combined andconfined masonry walls jointed with a similar mortar. However,this analogy is not good enough for combined and confinedmasonry walls jointed with non-engineered mortar, apparentlyalso for the frictionless sliding along the bed and head joints.Extensive additional experimental research is needed to discern

the impact of many variables, but this research study lead oneto believe that the following variables are important to assessin future experimental works: (a) the impact of other wallcombinations, (b) the impact of the applied axial load, (c) theimpact of other mortar mixes allowed by seismic codes and,(d) the impact of using concrete blocks of better quality, withsimilar mechanical properties to the ones of clay bricks. Referenceconfined masonry walls made of: (a) bricks only and, (b) concreteblocks only should be included in the future testing protocols.

Acknowledgments

This experimental study was possible because of the enthusi-astic collaboration of several students and technicians at Universi-dadAutnomaMetropolitana. Technicians LeopoldoQuiroz, RubnBarreda and Jos Luis Caballero assisted us in all the prototypeand material testing. Prof. Hans Archundia-Aranda assisted us inthe planning and in data acquisition of the cyclic tests. The con-struction of test specimens was possible because of the enthusi-258 A. Tena-Colunga et al. / Engineeri

Fig. 31. Equivalent cracked wide column ana

initial stiffness, cracking drift angle, drift angle for design, etc.)with the experimental data of a similar confinedmasonry wallsmade with solid clay bricks; however, their strength, ultimatedeformation and energy dissipation characteristics are some-what smaller.

4. The major difference in the cyclic behavior of the combinedand confined masonry walls jointed with engineered and non-engineered mortar is related to the resisting shear mechanismthat leads to failure. In the walls jointed with engineeredmortar, the failure mechanism is characterized by diagonaltension cracks that favor primarily spalling and crushing of themasonry in the central zone of the main diagonal crack withmanner, as proposed in this paper, to be helpful for the seismicevaluation of combined and confined masonry walls also.g Structures 31 (2009) 240259

logy proposed by Bazn and Meli [19,47,48].

It was also demonstrated that the hysteretic analytical modelproposed by Flores and Alcocer [24,27] for confined masonrybrick walls jointed with engineered mortar can also be usedwith confidence for combined and confined masonry walls jointedwith a similar engineered mortar. However, the model is notgood enough for predicting the observed experimental behaviorof combined and confined masonry walls jointed with non-engineered mortar because of the observed pinching behaviorrelated to the frictionless sliding along the bed and head joints thatthe model of Flores and Alcocer does not consider.It was shown that the equivalent cracked wide column analogy

proposed by Bazn and Meli [19,47,48] to model confined brickastic collaboration of several students: Csar Carpio, Jos ManuelAlonso, Misael Bahena, Daniel Miranda, Sergio Lpez, Eder Gudio,

nA. Tena-Colunga et al. / Engineeri

Rosaura Ramrez, Efran Joaqun Diego, Marco Antonio Rico, RenEspinoza, Richard Vliz, Roberto Moreno, Elas Josu Moral andGerardo Ibarra. Leonardo Flores of Cenapred and Sergio Alcocer ofUNAM are thanked for their valuable assistance and share of infor-mation regarding previous testing of confinedmasonry brickwalls.

References

[1] Meli R. Structural design of masonry buildings: The Mexican practice. In: ACIspecial publication 147masonry in the Americas. American Concrete Institute;1994. p. 23962.

[2] Alcocer SM, Klingner RE. Masonry research in the Americas. In: ACI specialpublication 147 masonry in the Americas. American Concrete Institute; 1994.p. 12769.

[3] EERI/IAEE(2000). World housing encyclopedia. Earthquake Engineering Re-search Institute and the International Association for Earthquake Engineering.www.worldhousing.net.

[4] Gallegos H. Masonry in Peru. In: ACI special publication 147 masonry in theAmericas. American Concrete Institute; 1994. p. 30731.

[5] Hidalgo P. Seismic behavior and earthquake-resistant design of masonrybuildings in Chile. In: ACI Special publication 147 masonry in the Americas.American Concrete Institute; 1994.

[6] Garca L, Yamn L. Review ofmasonry construction in Colombia. In: ACI specialpublication 147 masonry in the Americas. American Concrete Institute; 1994.p. 283305.

[7] TomazevicM. Earthquake-resistant design ofmasonry buildings. London (UK):Imperial College Press; 1999.

[8] DAyala D, Bostenaru M, Goretti A, Yakut A, Tomazevic M. Application of theworld housing encyclopedia. Geophy Res Abstr 2005;7:01649.

[9] El-Gawady M, Lestuzzi P, Badoux M. A review of conventional seismicretrofitting techniques for URM. In: Proceedings, 13th international brick andblock masonry conference. 2004. CD-ROM.

[10] Boen T, Privadi KS. Engineering the non-engineered houses for betterearthquake resistance in Indonesia. In: Proceedings, DRHAsia contentmeeting. 2007. p. 2068.

[11] ERRA. Reconstruction of houses in earthquake affected areas. Single storeyconstruction compliance catalogue. Earthquake Reconstruction and Rehabil-itation Authority. National Engineering Services, Lahore, Pakistan, May 2007.http://www.nespak.com.pk.

[12] Hori N, Inoue N, Purushotam D, Nishida T, Kobayashi J. Experimental andanalytical studies on earthquake resisting behaviour of confined concreteblock masonry structures. Earthq Eng Struct Dyn 2006;35:1699719.

[13] Borja-Navarrete G, Daz-Canales M, Vzquez-Vera A, Del Valle-Caldern E.Damage statistics of the September 19, 1985 earthquake in Mexico City.In: Proceedings, international workshop the Mexico city earthquakes-1985,factors involved and lessons learned. ASCE; 1986. p. 707.

[14] Alcocer SM, Tena-Colunga A, Berrn R, de la Torre O, del Valle E, Garca F, et al.Captulo 10: Estructuras de mampostera y adobe. In: Tena-Colunga Arturo,editor. El macrosismo de Manzanillo del 9 de octubre de 1995. first ed.Sociedad Mexicana de Ingeniera Ssmica; 1997. p. 23367 [in Spanish].

[15] Tena-Colunga A, Jurez A, Salinas VH. Resistencia y deformacin de muros demampostera combinada y confinada sujetos a cargas laterales. Rev Ing Ssm,SMIS 2007;76:2960. [in Spanish].

[16] Jurez A. Mecanismos de resistencia y de deformacin de muros demampostera combinada y confinada.M.Sc. thesis. Divisin de Ciencias Bsicase Ingeniera, Universidad Autnoma Metropolitana Azcapotzalco; 2008 [inSpanish].

[17] Salinas V. Comportamiento ante cargas laterales de muros de mamposteracombinada unidos con morteros de autoconstruccin. M.Sc. thesis. Divisinde Ciencias Bsicas e Ingeniera. Universidad Autnoma MetropolitanaAzcapotzalco; 2008 [in Spanish].

[18] Meli R. Comportamiento ssmico de muros de mampostera. Report 352. SerieAzul: Instituto de Ingeniera, UNAM. 1979 [in Spanish].

[19] Bazn TE. Muros de mampostera ante cargas laterales. Estudios analticos.Ph.D. thesis. Divisin de Estudios de Posgrado. Facultad de Ingeniera. UNAM;September 1980 [in Spanish].

[20] Ishibashi K, Katsumata H. A study on nonlinear finite element analysis ofconfinedmasonrywalls, Cuadernos de InvestigacinNo. 15, Cenapred,Mexico.July 1994.

[21] Alcocer SM,Meli R. Test program on the seismic behavior of confinedmasonrystructures. Mason Soc J 1995;13(2):6876.

[22] Alcocer SM. Implications derived from recent research in Mexico on confinedmasonry structures. In: Proceedings, CCMS symposium. American Society ofCivil Engineers. 1996. p. 8292.

[23] Aguilar G, Meli R, Diaz R, Vazquez-del-Mercado RS. Influence of the horizontalreinforcement on the behavior of confined masonry walls. In: Proceedings,11th world conference on earthquake engineering. Paper no. 1380. 1996.CDROM.g Structures 31 (2009) 240259 259

[24] Flores LE, Alcocer SM. Calculated response of confined masonry structures,In: Proceedings, 11th world conference on earthquake engineering. Paper No.1830. 1996. CDROM.

[25] Alcocer SM, Zepeda JA. Behavior of multi-perforated clay brick walls underearthquake-type loading. In: Proceedings, eighth North American masonryconference. 1999. CDROM.

[26] Alcocer SM, Muri-Vila D. Ensayos en mesa vibradora de sistemas de murosde mampostera confinada a escala 1:3. Informacin Tecnolgica 2000;11(5):11322 [in Spanish].

[27] Flores L, Alcocer SM. Estudio analtico de estructuras de mamposteraconfinada. Technical report, Coordinacin de Investigacin, rea de IngenieraEstructural y Geotecnia, Cenapred. 2001 [in Spanish].

[28] Urza DA, Padilla R, Loza R. Influencia de la carga vertical en la resistenciassmica de muros de albailera confinada elaborados con materialespumticos tpicos de Guadalajara. In: Proceedings, XIII Congreso Nacional deIngeniera Ssmica. 2001; CDROM [in Spanish].

[29] Alcocer SM, Arias JG, Vazquez A. Response assessment of Mexican confinedmasonry structures through shaking table tests. In: Proceedings, 13th worldconference on earthquake engineering. 2004. CDROM.

[30] Trevio EL, Alcocer SM, Larra R, Flores LE. Estudio experimental demuros de mampostera confinada de bloques de concreto. Primera parte:Comportamiento general de los especmenes, Ingeniera Civil. Centro deEstudios y Experimentacin de Obras Pblicas CEDEX, Ministerio de Fomento,Madrid, Espaa, No. 150. April 2008 [in Spanish].

[31] Trevio EL, Alcocer SM, Larra R, Flores LE. Estudio experimental demuros de mampostera confinada de bloques de concreto. Segunda parte:Caracterizacin de la respuesta, Ingeniera Civil, Centro de Estudios yExperimentacin de Obras Pblicas CEDEX, Ministerio de Fomento, Madrid,Espaa, No. 150. April 2008 [in Spanish].

[32] Astroza M, Schmidt V. Capacidad de deformacin de muros de albaileraconfinada para distintos niveles de desempeo. Rev Ing Ssm, SMIS 2004;70:5975 [in Spanish].

[33] Moroni M, Astroza M, Gomez J, Guzman R. Establishing Rw and Cd factor forconfined masonry buildings. ASCE J Struct Eng 1996;122(10):120815.

[34] Marinilli A, Castilla E. Evaluacin sismorresistente de muros de mamposteraconfinada con dos o ms machones. Boletn Tcnico IMME 2007;45(2):5873.[in Spanish].

[35] Maldonado NG, Olivencia L. Techniques used to repair seismic-resistantmasonry walls. In: Proceedings, 10th world conference on earthquakeengineering. vol. 9. 1992. p. 538994.

[36] Tomazevic M, Klemenec I. Seismic behaviour of confined masonry walls.Earthq Eng Struct Dyn 1997;26:105971.

[37] Gouveia J, Loureno P. Masonry shear walls subjected to cyclic loading:Influence of confinement and horizontal reinforcement. In: Proceedings, tenthNorth American masonry conference. Paper no. 042. 2007. CD-ROM.

[38] Sarma BS, Sreenath HG, Bhagavan NG, Murthy AR, Vimalanandam V.Experimental studies on in-plane ductility of confined masonry panels. ACIStruct J 2003;100(3):3306.

[39] Paikara S, Rai DC. Confining masonry using precast RC element for enhancedearthquake resistance. In: Proceedings, 8th US national conference onearthquake engineering. Paper no. 1177. 2006. CDROM.

[40] NTCM-2004. Normas Tcnicas Complementarias para Diseo y Construccinde Estructuras de Mampostera, Gaceta Oficial del Distrito Federal, October;2004 [in Spanish].

[41] Alcocer SM, Cesn J, Flores LE, Hernndez O, Meli R, Tena-Colunga A,VasconcelosD. The newMexico city building code requirements for design andconstruction of masonry structures. In: Proceedings, ninth North Americanmasonry conference, 2003. p. 65667. CD-ROM.

[42] UBC-97. Uniform building code. 1997 edition. International conference ofbuilding officials, vol. 2. 1997.

[43] ACI 530-05. Building code requirements for masonry structures (ACI 530-05/ASCE 5-05/TMS 402-05). American Concrete Institute; 2005.

[44] IBC-06. International building code. 2006 edition. International Code Council;2006.

[45] NMX-C-404-ONNCCE. Industria de la construccin-bloques, tabiques o ladril-los y tabicones para uso estructural-especificaciones ymtodos de prueba, Or-ganismo Nacional de Normalizacin y Certificacin de la Construccin y Edifi-cacin, Mxico; 1997. p. 9 [in Spanish].

[46] Kwok Y-H, Ang AH-S. Seismic damage analysis and design of unreinforcedmasonry buildings, Structural research series no. 536. University of Illinois atUrbana-Champaign. June 1987.

[47] Bazn E, Meli R. Diseo ssmico de edificios. second ed. Mxico: Limusa; 1999[in Spanish].

[48] Tena-Colunga A, Cano-Licona J. Improvements for the simplified method forthe seismic analysis of masonry shear-wall structures advocated in Mexicancodes, In: Proceedings, tenth North American masonry conference 2007. CD-ROM.

Cyclic behavior of combined and confined masonry wallsIntroductionProperties for combined and confined masonryTest for bricks, blocks and mortarAxial compression prism tests for combined masonryDiagonal compression wallet tests for combined masonry

Construction, instrumentation and setup of test specimensExperimental resultsHysteretic behavior, response envelopes and cracking patternsCyclic stiffness degradation

Earthquake-resistant qualification according to NTCM-2004Drift angle for seismic design

Damage indexComparison with confined brick masonrySimplified analytical modelingHysteretic modelEquivalent wide column analogy

Concluding remarksAcknowledgmentsReferences