Engineering Structures

Manuscript Draft
Manuscript Number: ENGSTRUCT-D-14-00490
Title: Laboratory testing and finite element simulation of the structural
response of an adobe masonry building under horizontal loading
Article Type: Research Paper
Keywords: Adobe masonry; horizontal loading; finite element model;
damaged plasticity; non-linear analysis
Abstract: This paper is concerned with the calibration and validation of
a numerical modelling approach for adobe masonry buildings under
horizontal loading. The paper first reviews the state-of-the-art in
experimental and computational research of adobe structures and then
presents results obtained from monotonic lateral loading laboratory tests
on a 1:2 scaled unreinforced adobe masonry building. Through the
experimental investigation conducted, useful conclusions concerning the
initiation and propagation of cracking failure are deduced. In addition,
damage limit states at different levels of deformation are identified.
Experimental results verify that the response of adobe structures to
horizontal loads is critically affected by weak bonding between the
masonry units and mortar joints and by lack of effective diaphragmatic
function at roof level. Based on experimental material data, a finite
element continuum model is developed and calibrated to reproduce the test
structure's force-displacement response and mode of failure. An isotropic
damaged plasticity constitutive law is adopted for the numerical
simulation of adobe masonry and the use of appropriate modelling
parameters is discussed. The analyses carried out reveal that the global
structural behaviour is primarily influenced by the tensile response
assigned to the homogenized masonry medium. Results show that, despite
its generic limitations and simplifications, continuum macro-modelling
can approximate the structural behaviour of horizontally loaded adobe
masonry construction with sufficient accuracy.

Highlights (for review)

Click here to download Highlights (for review): highlights.docx

Highlights

A 1:2 scaled adobe masonry building is tested in laboratory under horizontal

loading.

Conclusions on the initiation and propagation of cracking failure are deduced.

Damage limit states at different levels of deformation are identified.

A non-linear finite element continuum model of the scaled building is developed.

The FE model is calibrated to reproduce the structural response with sufficient

accuracy.

*Abstract
Click here to download Abstract: abstract.docx

ABSTRACT

This paper is concerned with the calibration and validation of a numerical modelling

approach for adobe masonry buildings under horizontal loading. The paper first

reviews the state-of-the-art in experimental and computational research of adobe

structures and then presents results obtained from monotonic lateral loading

laboratory tests on a 1:2 scaled unreinforced adobe masonry building. Through the

experimental investigation conducted, useful conclusions concerning the initiation and

propagation of cracking failure are deduced. In addition, damage limit states at

different levels of deformation are identified. Experimental results verify that the

10

response of adobe structures to horizontal loads is critically affected by weak bonding

11

between the masonry units and mortar joints and by lack of effective diaphragmatic

12

function at roof level. Based on experimental material data, a finite element

13

continuum model is developed and calibrated to reproduce the test structures force-

14

displacement response and mode of failure. An isotropic damaged plasticity

15

constitutive law is adopted for the numerical simulation of adobe masonry and the use

16

of appropriate modelling parameters is discussed. The analyses carried out reveal that

17

the global structural behaviour is primarily influenced by the tensile response assigned

18

to the homogenized masonry medium. Results show that, despite its generic

19

limitations and simplifications, continuum macro-modelling can approximate the

20

structural behaviour of horizontally loaded adobe masonry construction with

21

sufficient accuracy.

22

KEYWORDS

23

Adobe masonry, horizontal loading, finite element model, damaged plasticity, non-

24

linear analysis

*Manuscript
Click here to download Manuscript: text.docx

1
2

Click here to view linked References

1. Introduction
Adobe masonry structures are encountered in almost every region of the world

and are considered to possess significant historic and cultural value. At the same time,

unreinforced adobe masonry is quite susceptible to seismic damage [1]. The strong

seismicity of areas where a considerable number of earthen buildings exists (i.e. wider

Eastern Mediterranean region, South Asia, South America), renders the study of the

behaviour of adobe structures under horizontal loads essential. The development of

structural analysis methods that account for the specific characteristics of adobe

masonry is also required to facilitate the implementation of rational engineering

10
11

assessment/design.
Up to date, several studies involving laboratory testing of full- and/or reduced-

12

scale adobe structures have been conducted [2-19]. Emphasis has been primarily

13

given on evaluating various repair/retrofitting techniques, rather than on providing

14

extensive data which can be exploited for the calibration and validation of numerical

15

analysis tools. Researchers who have developed numerical models of adobe masonry

16

structures [20-29] have mainly performed conceptual analyses aiming to obtain

17

qualitative information regarding the response of typical traditional earthen buildings.

18

Detailed comparisons between simulation results and physically measured aspects of

19

structural behaviour (i.e. deformation, load-resistance) are rather limited [23-25]. This

20

indicates that there is a need for adopting a more integrated research approach that

21

will combine experimental and computational work on adobe masonry buildings, in

22

order to develop reliable assessment procedures and analysis methods.

23
24

The present study aims to extend existing knowledge regarding the structural
behaviour of adobe buildings by contributing towards the development of appropriate

25

assessment procedures and analysis methods. Hence, it utilizes the results of large-

26

scale laboratory tests to develop a Finite Element (FE) continuum macro-model

27

capable of simulating the response of a horizontally loaded unreinforced adobe

28

masonry building with sufficient accuracy. More specifically, for the purpose of

29

validating the FE model, a 1:2 scaled replica of an existing single-storey traditional

30

adobe building was constructed and subjected to static monotonic lateral loading tests.

31

Masonry failure mechanisms (i.e. initiation and propagation of cracking) were

32

recorded during the experimental procedure, while damage limit states at different

33

levels of deformation were identified. In the framework of the numerical

34

investigation, a detailed 3D FE model of the scaled building was developed. This was

35

used for performing non-linear analyses, aiming to macroscopically reproduce the

36

general response of the structure under test. For the numerical representation of adobe

37

masonry, a damaged plasticity constitutive law was adopted, while experimentally

38

derived material data were used as input parameters. The validity of the numerical

39

results was verified both qualitatively and quantitatively through comparisons with

40

the experimental damage patterns and force-displacement curves. The numerical

41

investigation conducted enabled the identification of the factors which critically affect

42

the FE simulation of adobe structures. The results of this work represent a promising

43

step towards the numerical modelling of the seismic behaviour of earthen

44

constructions.

45
46
47
48
2

49

2. Review of experimental and computational research on adobe structures

50

2.1. Experimental work

51

Most experimental data currently available regarding the response of adobe

52

masonry construction has been obtained by examining model structures before and

53

after the implementation of repair/strengthening interventions.

54

Systematic testing of unreinforced adobe masonry structures took place in the

55

framework of various research projects undertaken by the Pontifical Catholic

56

University of Peru. Relevant experimental work included static tilt tests on house

57

modules [2], displacement-controlled cyclic tests on I-shaped wall configurations

58

[3] and shake table tests on single- [4-7] and two-storey [8] model buildings and

59

vaulted structures [9, 10]. In all cases, the response of unreinforced model structures

60

was compared to that of reinforced ones (i.e. structures incorporating timber ring

61

beams, cane rods, steel wire meshes, geogrids, fibre-reinforced polymer strips, tire

62

straps, etc.).

63

Noticeable experimental research on the dynamic response of unreinforced

64

adobe masonry buildings was also carried out during the Getty Seismic Adobe

65

Project. In the first phase of this project, 1:5-scaled replicas of single-storey dwellings

66

were subjected to impact hammer and shake table tests before and after

67

repairing/strengthening [11]. In the second phase of the same project [12], dynamic

68

excitations based on real accelerograms were imposed on larger (1:2 scaled) models.

69

At this phase, in addition to unreinforced masonry structures, model buildings

70

retrofitted with bond beams, horizontal/vertical straps, local ties, centre-core rods and

71

wooden roof diaphragms were also examined.

72

Dowling [13] conducted shake table tests on 1:2 scaled U-shaped wall units

73

and complete buildings to examine the dynamic behaviour of unreinforced adobe

74

masonry construction. Along with plain unreinforced masonry structures, models

75

incorporating pilasters/buttresses, wire meshes, bamboo poles and timber ring beams

76

were also constructed and tested in [13]. The outcomes obtained were used for

77

proposing retrofitting solutions.

78

More recently, a real-scale I-shaped adobe wall was examined at Aveiro

79

University [14]. Following a number of cyclic lateral loading tests, the cracks formed

80

in the masonry were injected with lime mortar and a polymeric mesh was fixed on the

81

surface of the wall. The repaired/retrofitted structure was subjected to further lateral

82

loading tests.

83

Extensive literature on the response of strengthened/retrofitted adobe masonry

84

buildings can be also found in [15-19] which present results from shake table tests and

85

static horizontal loading tests on 1:1.5 [15], 1:2.5 [16], 1:3 [18], 1:5 [15] and 1:10 [19]

86

scaled model structures.

87

The main conclusion derived from the aforementioned tests is that adobe

88

masonry structures generally have limited capacity to resist horizontal loads. This is

89

attributed to two factors: (a) poor bonding between the adobe bricks and the mortar

90

joints which reduces the tensile strength of the masonry [4, 11, 14] and (b) lack of

91

diaphragmatic function at roof level which precludes effective transfer of loads among

92

the load-bearing walls [11, 12]. Under seismic action, out-of-plane failure, either due

93

to extensive cracking or due to detachment at cross-walls and overturning, prevails [4,

94

11-13]. Integrated retrofitting systems can improve the poor seismic behaviour of

95

unreinforced adobe masonry buildings, either by increasing their overall lateral

96

resistance or by producing a confinement effect which reduces the risk of brittle

97

collapse [12-19].

98

2.2. Numerical modelling and analysis

99

In contrast to experimental work, computational research on adobe masonry

100

structures has not been as rigorous. Despite the fact that advanced analysis methods

101

have been extensively used for the simulation of conventional masonry structures (i.e.

102

structures built with stone, fired clay bricks, concrete blocks, etc.), the application of

103

numerical tools has not been meticulously studied in the context of earthen

104

construction.

105

Simulation of masonry structures can follow a macro- or a micro-approach. In

106

the macro-approach, either distinct macro-elements are used to represent individual

107

piers and spandrels, or the masonry is treated as a fictitious homogeneous medium

108

represented by continuum finite elements. In the micro-approach, the masonry unit-

109

mortar interfaces are considered as potential crack/slip planes, while the building

110

blocks and the mortar are either explicitly described (detailed micro-modelling) or

111

represented by repeated expanded cellular units interacting at their boundaries

112

(simplified micro-modelling).

113

Continuum FE models of adobe-wood buildings have been developed by Che et

114

al. [20]. These were subjected to elastic time domain analysis in order to examine

115

their seismic response. Linear dynamic analyses by response spectra have been also

116

conducted by Gomes et al. [21] on 3D models of unreinforced and reinforced adobe

117

buildings.

118
119

Using experimental material data, Meyer [22] modified the Holmquist-JohnsonCook model for concrete to capture the pressure and strain-rate-dependent non-linear
5

120

behaviour of adobes. The formulated constitutive law was used for performing

121

physics-based penetration simulations on continuum models of adobe targets.

122

Non-linear static and dynamic simulations on FE continuum models have been

123

undertaken by Tarque et al. [23-25]. These studies focused on approximating the

124

macroscopic behaviour (mode of failure, displacements etc.) of structures tested in the

125

laboratory. Modelling parameters were calibrated by matching numerical-

126

experimental results. Both orthotropic smeared cracking and isotropic damaged

127

plasticity constitutive laws were considered.

128

In addition to FE models of continua, Tarque et al. [23] examined 3D micro-

129

models of adobe walls loaded in-plane. The authors assumed that the response of

130

adobe bricks is elastic isotropic and assigned non-linear behaviour to the masonry

131

joints via a combined crushing-shearing-cracking interface damage model.

132

A similar approach was adopted in [26] and [27] for performing a series of

133

dynamic simulations with the purpose of developing seismic vulnerability functions.

134

In the aforementioned studies, adobe masonry was modelled as an assemblage of rigid

135

blocks connected by zero-thickness non-linear interfaces that could sustain tensile,

136

shear or compressive damage.

137

Using non-linear anisotropic constitutive laws for the adobe bricks and specially

138

formulated viscoelastic elements for the joints, Cao and Watanabe [28] developed 3D

139

models of a single-storey adobe building. Through dynamic analyses, the above

140

researchers investigated the earthquake response of adobe construction and verified

141

the efficiency of wooden frame reinforcement.

142

Morales and Delgado [29] examined 2D models of single- and two-storey adobe

143

walls. The models were composed of distinct elements connected with springs and

144

dashpots that acted as possible fracture points. The seismic capacity of the simulated

145

structures was assessed by imposing reversing horizontal accelerations.

146

3. Laboratory testing of an adobe model building

147

3.1. Construction of model building

148

For investigating the structural response of adobe masonry buildings, a 1:2

149

scaled replica of a traditional single-roomed Cypriot dwelling (monochoro makrynari)

150

[30] was constructed and tested at the Structures Laboratory of the University of

151

Cyprus (Fig. 1).

152

The model structures walls were 220 mm thick and were built with scaled-

153

down adobe bricks measuring (height x width x length) 30 x 150 x 220 mm3. The

154

bricks were obtained from a local producer and were laid with the application of earth

155

mortar (composition soil:straw:water 200:6:100 w/w) prepared in the laboratory.

156

Following the islands traditional building techniques, the masonry was constructed in

157

a running bond pattern and the joint thickness was consistently kept below 10 mm.

158

The model structure was securely bolted on the laboratory concrete floor. The

159

structures external dimensions were (width x length) 1.75 x 3.60 m2. The height of

160

the front elevation was 1.50 m and that of the opposite rear wall was 1.65 m. A door

161

measuring 1.10 m in height and 0.70 m in width was formed on the faade. Two

162

openings with dimensions 0.55 x 0.55 m2 were also created on the two side walls. A

163

triangular notch 0.22 m wide and 0.18 m high was formed on the rear wall to simulate

164

the ventilation notches encountered in local vernacular buildings.

165

It was presumed that the stone masonry foundations of traditional earthen

166

structures preclude horizontal translation of the walls at ground level, but allow

167

bending. Therefore, the first layer of adobe bricks was simply set with the application

168

of earth mortar. Horizontal displacements at this level were constrained by timber

169

elements installed along the structures perimeter. At the cross-walls, overlapping

170

bricks were laid upon each other to achieve adequate interconnection.

171

Above all openings, lintels consisting of two jointed timber beams, each with a

172

cross-section of 85 x 85 mm2, were installed. The roof structure comprised of a 20

173

mm-thick wooden panel nailed upon nine timber rafters (45 x 90 mm2 in cross-

174

section) that spanned the space between the two opposite longitudinal walls. On top of

175

the panel, adobes were uniformly placed to represent the weight of roof tiles. All

176

timber elements were set into the masonry with gypsum mortar.

177

3.2. Test procedure and instrumentation

178

The model building was tested nine weeks after its construction by applying

179

monotonically increasing lateral forces until noticeable damage (i.e. severe cracking

180

of the masonry walls) was observed. Loading was applied on the rear wall using a

181

steel hydraulic jack with 60 kN maximum capacity (Fig. 2a). The load imposition

182

system was supported by a rigid steel reaction frame (see background of Fig. 1). To

183

achieve a more even load distribution a timber beam strengthened at its centre was

184

used along the rear wall. The hydraulic jack accommodated a swivel head that

185

enabled it to stay in contact with the loading beam when out-of-plane bending was

186

induced. Loading was applied at approximately 2/3 of the models height.

187
188

Linear Variable Displacement Transducers (LVDTs) (range 50.8 mm,

accuracy 0.25%) were placed at 15 different positions on the model structure to
8

189

record displacements (Fig. 2b). Emphasis was given in monitoring the out-of-plane

190

movement of the longitudinal walls and the in-plane bending of the side walls.

191

Therefore, one of the side walls and the two adjacent halves of the longitudinal walls

192

were instrumented. Indeed, during the tests it was confirmed that there was close

193

analogy between the responses of the half-structures sections examined and of the

194

parts symmetric to them. LVDTs were also placed at the structures base to verify that

195

no translation or rotation took place. All measurements were recorded automatically

196

via a data acquisition system. Digital cameras were also used for monitoring failure

197

evolution and crack opening-closing.

198

A total of 10 monotonic loading-unloading cycles were implemented. The

199

experimental procedure was terminated when a significant reduction of the lateral

200

resistance of the model structure was detected.

201

3.3. Experimental results and discussion

202

3.3.1. Crack patterns

203

The crack pattern recorded after the completion of the experimental procedure is

204

shown in Fig. 3. Damage modes were almost identical during all tests, with most

205

cracks developing during the first four load cycles. Subsequent load cycles led to re-

206

opening of pre-existing fissures and increased crack widths.

207

Damage was noted at the rear and the two side walls, but not at the faade or at

208

any of the timber members. Damage localization reveals stress concentrations and

209

implies that the load-bearing members of the model failed to react as a homogeneous

210

assemblage of structural elements (i.e. as a fully connected structural system). In

211

addition, it indicates lack of diaphragmatic function at the roof level.

212

Out-of-plane bulging of the rear wall caused the formation of a major horizontal

213

crack at the interior of the structure, along the line of loading (Fig. 3a). Due to

214

overstressing at the load imposition point, diagonal cracks extending from the centre

215

of the wall towards its two lower sides were generated below the aforementioned

216

horizontal fissure. In addition, a V-shaped cracked section was formed between the

217

triangular ventilation notch and the four central roof rafters.

218

At the exterior surface of the rear wall, a continuous horizontal crack occurred

219

between the 7th and 8th rows of adobe bricks (Fig. 3b). Towards the two sides,

220

because of the restrain imposed by the side walls, this crack was inclined. Less severe

221

cracking was recorded below this zone. Furthermore, failure of the gypsum mortar

222

joints at the roof rafter abutments and subsequent sliding of the timber members were

223

noted. As the rear wall was subjected to significant out-of-plane deformations, stress

224

concentrations were generated at the areas where the masonry was in contact with the

225

much stiffer timber rafters. This led to horizontal cracking at the vicinity of the roof

226

supports; cracking extended diagonally where restrain by the two side walls became

227

effective.

228

The mode of failure sustained by the two side walls was mainly characterized

229

by the formation of diagonally orientated shear cracks that radiated out of the two

230

openings corners and propagated through the brick joints in a stepped pattern (Fig.

231

3c). These cracks extended throughout the whole width of the side walls. Damage at

232

the upper section of the walls spread towards the intersection with the rear wall,

233

eventually joining with the external rear wall cracks that formed just below the roof

234

rafters. During the two final test cycles, out-of-plane torsional movement of the side

235

walls upper cracked sections was also recorded.

10

236

In all cases, failure was characterized by loss of bonding between the masonry

237

units; no damage of the adobe bricks was reported. This verifies that the failure

238

mechanisms encountered in adobe structures are primarily a product of weak adhesion

239

among the adobes [4, 11, 31, 32]. Crack opening was significant and ranged from 5 to

240

20 mm (Fig. 4). Interestingly enough, when loading was removed, the fissures formed

241

closed completely and no sign of damage was visible. However, cohesion between the

242

masonry units at these areas had been lost and when load was exerted again, re-

243

opening of the cracks was mobilized.

244

Despite the fact that the experimental set-up enabled only the imposition of

245

static forces, the recorded modes of damage correspond well to those observed in

246

dynamic tests and to those sustained by adobe buildings during earthquakes. Crack

247

patterns similar to the ones observed at the rear wall of the model building have been

248

reported in [1, 3, 11, 12, 33, 34]. Diagonal shear cracking of adobe walls loaded in-

249

plane has been noted in several other experimental [3, 11, 14] and field [1, 35] studies.

250

However, due to the unilateral and monotonic load imposition process, separation

251

between intersecting walls did not occur, although such a response of unreinforced

252

adobe masonry to seismic loads is rather common [4, 16, 33, 35]. Moreover, the lack

253

of diaphragmatic roof function, caused by the sliding failure of the rafter supports, did

254

not enable the effective transfer of forces from the rear wall to the faade. Therefore,

255

as opposed to a dynamic state where all sections perpendicular to the direction of the

256

principal action would sustain reversing out-of-plane bending loads, in the tests

257

conducted here, movement of the faade was dictated by the in-plane drift of the side

258

walls and therefore no noticeable damage (i.e. cracking and/or detachment)

259

developed.

11

260

3.3.2. Force-displacement response and limit states

261

Force-displacement data envelopes obtained from the implementation of the 10

262

test cycles are presented in Fig. 5. The diagrams show the variation of the cumulative

263

displacements measured at the upper sections of the rear (LVDT13), the side

264

(LVDT3) and the faade (LVDT1) walls in relation to the load imposed. Cumulative

265

displacement values were computed by adding to the recordings of each individual

266

test cycle the permanent deformations noted after the completion of all previous

267

cycles.

268

Based on the structural response recorded and the corresponding state of

269

damage observed, four limit states (LS1-4) can be identified. Up to approximately 5%

270

of its total displacement capacity and 75% (10.6 kN) of its maximum lateral resistance

271

(LS1), the structure performed with no or negligible damage and the various load-

272

bearing members maintained a consistent response to horizontal loading. The

273

displacements recorded at the model structures walls during this stage were rather

274

uniformly distributed. They lied in the region of 1.8 mm and correspond to

275

approximately 0.11% of drift (estimated as horizontal displacement divided by the

276

monitoring points vertical distance from the buildings base).

277

Above the 10.6 kN threshold, stiffness degradation started to develop and

278

cracking damage was initiated at the interior of the rear wall and at the two side walls.

279

The structure, however, could still function as a homogeneous structural system up to

280

11% of its total displacement capacity and 85% (12 kN) of its maximum lateral

281

resistance (LS2). A co-instantaneous movement of 4 mm and a lateral drift of 0.26%

282

were recorded at the upper sections of the walls monitored. It should be noted that the

283

first and second limit states were already reached by the end of the initial test cycle.

12

284

When the displacement induced exceeded 11% of the total displacement

285

capacity, interaction among the structures load-bearing members was effectively lost

286

and differential movement of the masonry walls took place. This was accompanied by

287

further cracking, permanent distortion and considerable reduction of the overall

288

stiffness. Such highly non-linear response continued until the load exerted became

289

equal to the maximum force the structure could withstand (14.2 kN) and the

290

displacement induced was 26% of the total deformation capacity (LS3). During this

291

stage, sliding failure of the roof rafters supports and cracking of the rear walls base

292

were observed. Furthermore, the cracks previously formed on the rear walls interior

293

and on the two side walls extended in length. Cumulative displacements at the faade

294

and the side wall were 7 and 7.7 mm, respectively. In terms of lateral drift, these

295

values can be interpreted as 0.5%. Cumulative displacement at the centre of the rear

296

wall was 21.6 mm and accounts for 1.4% lateral drift. The aforementioned data were

297

obtained after the completion of the first four test cycles.

298

After LS3 and up to the last limit state (LS4), the structure was characterized by

299

depletion of its overall stiffness and by inability to sustain higher levels of loading.

300

Relatively small augments of the imposed load led to large in- and out-of-plane drifts.

301

Moreover, significant inelastic deformations were generated, while crack opening

302

eventually attained its maximum value ( 20 mm). At LS4, the cumulative horizontal

303

translation of the side wall was 25.2 mm, while that of the faade was 23.8 mm. The

304

lateral drift at these sections was estimated as 1.6%. The total movement of the rear

305

wall was 84.9 mm and the lateral drift at its central section was 5.7%.

306

After LS4, at the last loading cycle, an abrupt drop in the structures lateral

307

resistance occurred. The sections of the two side walls above the diagonal shear

308

cracks were isolated by cracking damage. As a result, the faade and the adjacent
13

309

triangular halves of the two side walls were detached from the rear part of the

310

building. The load-bearing system was practically split into two independent parts that

311

could only transfer forces between them through contact points. Under the application

312

of load, the kinematic mechanism formed was mobilized causing rocking motion of

313

the frontal part and reducing the effective resisting area. Although the overall strength

314

fell to a residual value, total or partial collapse did not occur. Nevertheless, crack

315

formation and/or growth at this state could have been critical, if the relative

316

displacement induced across the planes of weakness had been larger.

317

Using experimental results from cyclic load tests on full-scale I-shaped adobe

318

walls, Figueiredo et al. [14] and Tarque et al. [36] defined damage limit states similar

319

to those reported in this study. The experimentally recorded maximum load resistance

320

accounts for approximately 30% of the model buildings self-weight. This is in total

321

agreement with the data obtained by Benedetti et al. [37] from extended dynamic

322

experiments on unreinforced masonry buildings constructed with fired clay bricks.

323

However, it is lower than the 34-100% base shear force-to-weight ratios reported by

324

researchers who performed shake table [4, 8, 38] and static tilt [39] tests on adobe

325

model structures. Despite being rather conservative, the load-bearing capacity

326

determined in the present work cannot be injudiciously adopted as a safe indicator for

327

the seismic behaviour of unreinforced adobe masonry construction. This is because

328

the monotonic imposition of forces during the testing procedure did not enable the

329

development of certain failure mechanisms (e.g. detachment of intersecting walls) that

330

would drastically reduce lateral resistance in the event of dynamic excitation. In

331

addition, the application of reversing horizontal accelerations would have probably

332

caused the out-of-plane failure of the longitudinal walls, either by detachment and

333

overturning or by diagonal and vertical cracking [33], at significantly lower levels of

14

334

deformation. Hence, it may be argued that the maximum lateral translation measured

335

during the tests is overestimated and does not realistically represent the displacement

336

capacity of unreinforced adobe masonry structures when these are subjected to

337

seismic action.

338

4. Numerical simulation of the response of the adobe model building

339

4.1. Finite element modelling and analysis

340

For simulating the response of the tested structure, a full 3D FE model was

341

developed in Abaqus/CAE (Fig. 6) [40]. The various parts comprising the

342

experimental set-up were modelled as individual bodies interacting with each other.

343

Hence, the FE model included different representations for the adobe masonry walls,

344

the openings lintels, the roof and the timber loading-beam. Since the test

345

configuration is symmetric, only half the structure was numerically examined. All

346

bodies were discretized using 8-noded 3D linear brick elements (C3D8) with sides 40

347

4 mm long. The mesh generated consisted of 47,808 elements and 68,139 nodes

348

resulting in 169,515 degrees of freedom.

349

Adobe masonry was numerically handled in the context of a macro-modelling

350

strategy. It was thus treated as a fictitious homogeneous continuum and no distinction

351

between masonry units and mortar joints was made. For simulating its behaviour, the

352

concrete damaged plasticity constitutive model [40-42] was adopted. This is a

353

continuum, plasticity-based, isotropic damage model that assumes two main failure

354

mechanisms: tensile cracking and compressive crushing. The material admissible

355

stress field is bounded by a yield surface that is controlled by hardening variables

356

linked to cracking and crushing strains.

15

357

Most parameter values used for the application of the damaged plasticity

358

constitutive law were based on experimental data. The density of adobe masonry was

359

set as = 670 kg/m3. This was estimated following simple gravimetric measurements

360

on the adobes used to construct the model structure. Poissons ratio ( = 0.3) was

361

evaluated from the deformations recorded during the compressive strength testing of a

362

stack-bonded adobe masonry prism, as the ratio of transverse to axial strains.

363

Compressive stress-strain response was described using the polynomial relation

364

developed by Illampas et al. [43] for adobe bricks (Fig. 7a). The Youngs modulus

365

was computed from the assigned stress-strain response as a secant modulus up to the

366

yielding point; E = 18 MPa. Compressive strength (fc = 1.2 MPa) and strain at peak

367

compressive stress (cu = 0.1 mm/mm) were defined from the average results of

368

laboratory tests on stack-bonded prisms [44]. Considering that adobes possess a

369

granular structure and thus have limited elastic response to compression [22], material

370

non-linearity was assumed after 5% of the compressive strength.

371

In tension, linear behaviour up to the maximum allowable stress and post-peak

372

softening were assumed (Fig. 7b). Inelastic tensile stress-strain response was

373

described using the exponential function developed by Loureno [45]:

hft ck

G t
f

t ft exp

(1)

374

In the above, ft is the tensile strength of masonry, Gf is the tensile fracture energy, tck

375

is the tensile cracking strain and h is the characteristic crack length.

376

Tensile strength was set as ft = 0.04 MPa, following the diagonal tension testing

377

of an adobe wallette. Regarding the tensile fracture energy Gf of the homogenized

378

masonry, direct tension tests on adobe couplets in [46] yielded a mean value of Gf =
16

379

4.5 N/m. The average tensile strength of the specimens examined in [46] was 0.01

380

MPa; assuming a linear analogy between the bearing capacity and the fracture energy,

381

the value of Gf = 18 N/m was adopted for ft = 0.04 MPa.

382

The characteristic crack length h was defined as [47, 48]:

h 3 hx hy hz

(2)

383

In the above equation hx, hy and hz are the elements lengths along the x, y and z axes.

384

The element size during meshing was selected to satisfy the energy criterion given in

385

equation 3:

Gf E

(3)

ft 2

386

Theoretically, through the definition of the characteristic crack length, mesh-

387

dependency of numerical results was treated. However, the use of this parameter

388

implies that, in non-structured meshes, the elements with larger aspect ratios will tend

389

to have rather different behaviour, depending on the direction in which they crack.

390

This effect may have introduced some mesh sensitivity to the results presented in this

391

study, despite making efforts to use elements with aspect ratios close to one,

392

especially in areas where tensile damage was expected.

393

For the rate at which the hyperbolic flow potential approaches its asymptote (e =

394

0.1) and the ratio between the initial equibiaxial and the initial uniaxial compressive

395

yield stresses (b0/c0 = 1.16), the default values suggested in [40] were adopted. The

396

plasticity parameter which relates the second stress invariant on the tensile meridian

397

to the equivalent invariant on the compressive meridian was set as Kc = 0.8, in line

398

with the recommendations of [40] for soils modelled with a Drucker-Prager yield

399

function. Based on [25] and [49], a very low dilation angle = 1o was selected.
17

400

Since no damage or considerable deformation was observed during the

401

experimental procedure in any of the timber members (i.e. lintels, rafters, loading-

402

beam, roof panel), these were all modelled using linear elasticity constitutive laws. In

403

addition, it was assumed that the mechanical properties of timber are isotropic. The

404

material parameters used were drawn from the literature [50, 51] as follows: (a) wood

405

panel density, = 380 kg/m3; Youngs modulus, E = 8000 MPa; Poissons ratio, =

406

0.2 and (b) timber lintels, rafters and loading-beam density, = 670 kg/m3; Youngs

407

modulus, E = 7000 MPa; Poissons ratio = 0.3.

408

At the areas where the masonry was in contact with the timber members, contact

409

pairs were formed and surface to surface interactions were defined via master-slave

410

associations. When under compression, interacting surfaces were assumed to remain

411

in contact; thus, any pressure could be transmitted across the interfaces. When the

412

contact pressure reduced to zero, separation of the surfaces took place and no transfer

413

of tensile stresses across interfaces was allowed. To simulate the behaviour hereby

414

described, a hard contact pressure-overclosure relationship [40] was defined in the

415

normal direction.

416

In the tangential direction, a finite-sliding formulation [40] based on the

417

Coulomb friction theory was used. The Coulomb friction model available in

418

Abaqus/CAE cannot account for cohesion among interacting surfaces and computes

419

the shear stress at which sliding initiates (crit) simply as a function of the contact

420

pressure (p) and the coefficient of friction () between the surfaces:

crit = p

(4)

421

At the interfaces between the masonry and the opening lintels and the masonry and

422

the roof rafters, a friction coefficient of = 0.5 was specified. This value was based
18

423

on the data reported in [32] which, however, do not refer to the frictional properties of

424

timber elements embedded in adobe masonry, but to the friction developed between

425

the masonry units and joints of adobe walls. Frictionless sliding ( = 0) was assumed

426

to take place between the masonry and the loading-beam and the masonry and the

427

roof panel.

428

All nodes at the base of the walls were considered to be pinned. Horizontal

429

kinematic constrains were imposed at the perimeter nodes affected by the timber

430

elements, which were installed in the actual structure to retain lateral movement at the

431

base. At the area where the hydraulic jack was in contact with the timber loading-

432

beam, constraints precluding translation along the x and z axes were imposed.

433

Movement in the x direction and rotations around the y and z axes were not allowed

434

along the plane of symmetry.

435

The weight of the adobes placed on the roof was evenly distributed to the roof

436

panel as an additional body force. Horizontal loads were applied in the form of lateral

437

displacements at the nodes of the timber loading-beam in contact with the jack. The

438

amplitude of the lateral displacements was formulated according to the cumulative

439

displacement data recorded during the laboratory tests.

440

The numerical solution process was completed in two successive steps. At the

441

initial step, the dead loads were incrementally imposed. At the second step, the lateral

442

displacements at the jack-loading beam interface were incrementally enforced at time

443

intervals ranging from 1x10-19 to 1x10-4 s over the 1 s analysis period. In both cases, a

444

general non-linear static procedure with automatic stabilization was implemented,

445

adopting the full Newton solution scheme. The effect of geometric non-linearity was

446

accounted for in all numerical steps.

19

447
448

4.2. Comparison between experimental-numerical results

Fig. 8 shows contour representations of the displacements computed in the y-

449

direction. Results show that the FE model captures well the deformed shape of the

450

structure. As expected, the maximum lateral displacement occurs at the rear wall, at

451

the level where loading was applied. In line with the experimental observations, the

452

out-of-plane movement of the faade is dictated by the in-plane drift of the side wall.

453

Furthermore, displacements along the height of the faade display a linear increase

454

towards the walls top. The backwards movement predicted at the rear central part of

455

the side wall is verified by experimental measurements and is attributed to out-of-

456

plane bending and subsequent torsion of this section.

457

In order to obtain the graphical visualization of the numerically predicted

458

damage pattern of Fig. 9, it was assumed that the direction of the vectors normal to

459

the crack planes is parallel to the direction of the maximum principal plastic strains

460

[40, 41]. The FE model adequately captured the structures mode of failure, both in

461

terms of damage distribution and in terms of crack initiation and propagation.

462

The onset of tensile failure during the simulation occurred at the upper central

463

section of the rear walls interior side. The plastic strain magnitude at this point

464

eventually attained the highest computed value, coinciding with the location where

465

the maximum crack opening of approximately 20 mm was observed during the

466

laboratory tests. Crack propagation was rapid, with plastic strains spreading across a

467

horizontal band, parallel to the loading beam. Almost co-instantaneously, tensile

468

failure was initiated at the two opposite corners of the side walls window opening.

469

The concentration of significantly high tensile stresses in this area produced a

20

470

diagonal distribution of plastic strains, similar to the crack pattern observed on the

471

tested building.

472

The gradual increase of the imposed load led to the formation of plastic strains

473

that followed inclined paths on the interior surface of the rear wall. As in the case of

474

the actual model building, damage extended from the principal horizontal line of

475

failure towards the upper and lower sections of the wall. Furthermore, horizontal and

476

diagonal cracking at the exterior base of the rear wall and propagation of basal

477

damage to the side wall were well reproduced.

478

The development of horizontal cracks at the vicinity of the roof rafter supports

479

was also adequately approximated. However, unlike experimental observations,

480

plastic strains in this area did not extend to the side wall and did not intersect with the

481

crack appearing above the windows lintel. Instead, a near-vertical crack occurred at

482

the upper rear section of the side wall. This inconsistency is attributed to

483

overestimation of the side walls out-of-plane torsional displacement by the FE

484

analysis.

485

Fig. 10 compares the outcomes of the FE analysis with the experimentally

486

derived force-displacement data envelopes for the upper sections of the rear wall, the

487

faade and the side wall. Numerical load data were estimated as the sum of all lateral

488

contact forces generated at the interface nodes of the timber loading-beam with the

489

rear adobe wall.

490

Reasonable agreement is found between the experimental and numerical

491

capacity curves, as in both cases the same trends are generally observed. The FE

492

model successfully predicted the occurrence of a post-yield plateau and a gradual

493

reduction of the load-bearing capacity. However, the abrupt drop in load resistance,
21

494

observed in the final loading cycle of the test, was not captured. This is likely due to

495

the fact that the kinematic mechanisms forming at large deformation levels could not

496

be accurately simulated though the use of a homogenized continuum. Such an

497

approach does not allow the discrete modelling of units and joints and therefore it

498

cannot capture the rocking motion of the faade and the triangular halves of the side

499

walls that were detached from the rear part of the structure after LS4 due to cracking.

500

Striking correspondence is found between the numerically derived lateral

501

resistance and the maximum force measured on the actual structure. The ultimate

502

displacement computed at the rear walls control nodal point practically coincides

503

with the one recorded during the laboratory tests. The out-of-plane translation of the

504

faade and the in-plane translation of the side wall were slightly miscomputed: 24.7

505

mm instead of the actual 26.6 mm for the faade; 28.0 mm instead of the actual 27.1

506

mm for the side wall.

507

The underestimation of forces at the ascending branches of the diagrams can be

508

attributed to the isotropic fracture criterion adopted. Tension and shear tests

509

conducted on mud brick specimens and masonry prisms revealed that the tensile

510

strength of adobe itself and the frictional resistance along the joints can be at least an

511

order of magnitude higher than the bonding strength [46]. Given that the adopted

512

tensile strength of ft = 0.04 MPa actually refers to resistance against de-bonding of the

513

masonry units, the bearing capacity implicitly assumed for the masonry medium in

514

the direction parallel to the bed joints (where the response is governed by friction) is

515

most probably underestimated. However, the formulation of the damaged plasticity

516

constitutive law does not allow for the definition of separate tensile strengths along

517

each direction. Another factor which may have influenced the simulated response is

518

that no bonding strength (cohesion) was assigned to the roof rafter-brick interfaces.
22

519

Consequently, the effective transfer of forces among opposite longitudinal walls at

520

low levels of deformation was precluded.

521

4.3. General discussion of numerical results

522

The numerical results obtained can be deemed as sufficiently accurate. Of

523

particular importance is the adequacy of the developed FE model to predict the failure

524

mechanisms sustained by the tested structure. Considering the inhomogeneous and

525

random nature of earthen materials, the correlation between the numerical and

526

experimental load-displacement data is also satisfactory. Besides, perfect agreement

527

between the results of simulations and the outcomes of laboratory tests is usually

528

regarded as a coincidence and should not be the mere objective of numerical

529

modelling [52]. This is because experimental data possess inherent variability. In

530

addition, despite applying an energy-based regularization of the masonry mediums

531

tensile response, a slight mesh dependency of the FE analysis procedure possibly still

532

existed, affecting, albeit to a limited degree, the simulation results.

533

A number of simulations conducted in the process of model calibration, using

534

different material properties (i.e. Youngs modulus, Poissons ratio, plasticity

535

characteristics, tensile and compressive strengths, friction coefficient at the timber-

536

masonry interface), revealed which modelling parameters are more critical. The

537

Youngs modulus assigned to adobe masonry determines the stiffness of the walls and

538

defines the tensile cracking strain (the higher the Youngs modulus the lower the

539

tensile cracking strain), thus affecting damage initiation. On the other hand, the

540

masonrys Poissons ratio and plasticity characteristics (i.e. dilation angle, flow

541

potential eccentricity, ratio of initial equibiaxial compressive yield stress to initial

542

uniaxial compressive yield stress, relation between second stress invariant on the

23

543

tensile meridian to that on the compressive meridian) have very limited influence on

544

the FE results.

545

No significant alteration of the results was observed when different compressive

546

strength values in the range 1 < fc < 2.2 MPa were assumed. However, convergence

547

difficulties were encountered when the compressive yielding stress fell below 0.05

548

MPa. Analyses revealed that tensile response is the most crucial aspect of the

549

simulation, since it dictates the lateral resistance and the displacement capacity

550

predicted. It is worth noting that analogous conclusions concerning the sensitivity of

551

numerical results to compression and tension parameters have been derived by Tarque

552

et al. [23], who simulated adobe walls using the same damaged plasticity constitutive

553

law. The friction coefficient assigned to the timber-masonry interface controls the

554

transfer of forces between the two opposite longitudinal walls and determines whether

555

shear sliding of the roof rafters will occur. Consequently, it also affects to some extent

556

the displacements computed.

557

5. Conclusions

558

Laboratory testing of a 1:2 scaled model building revealed that, under lateral

559

loading, damage in unreinforced adobe structures is primarily concentrated at the

560

masonry walls, whereas stiffer load-bearing members (i.e. timber elements) remain

561

practically intact. The prevalent failure mechanism that occurs is cracking due to

562

inadequate bonding between the bricks and the mortar. Damage initiation can be

563

influenced by stress augmentation at the corners of openings and at the abutments of

564

timber members.

565
566

Upon load removal, the cracks formed on adobe masonry walls close almost
completely, leaving little indication of damage. Cracked sections act as planes of
24

567

weakness and crack re-opening is mobilized when load is re-applied. This highlights

568

the cumulative effect that pre-existing damage poses on the structural behaviour of

569

adobe buildings. It also indicates that particular attention should be paid during the in-

570

situ inspection of earthen structures after seismic events.

571

Experimental force-displacement data show that adobe masonry structures

572

possess limited stiffness and can thus develop considerable deformations. Results also

573

show that homogeneous structural response is lost as soon as stiffness degradation

574

occurs and differential movement of the walls takes place. This verifies that absence

575

of a stiff diaphragm configuration at roof level and insufficient interaction between

576

the various load-bearing members pose a negative effect on the structural behaviour

577

of masonry buildings.

578

The damaged plasticity constitutive law adopted in this study has proven to be

579

adequate for modelling adobe masonry as an idealized homogenized continuum.

580

Provided that appropriate material data is used and that proper calibration is

581

undertaken, FE models can capture the force-displacement response and the failure

582

mode of adobe structures. The generic limitations of continuum modelling and the

583

assumption of isotropic damage may introduce some inconsistencies to the outcomes

584

of simulations, but do not preclude sufficient macroscopic approximation of the

585

global structural behaviour.

586

The sensitivity of numerical results to certain modelling parameters indicates

587

that a more detailed database of information on the properties of adobe masonry is

588

required. In particular, further experimental investigation should be undertaken to

589

assess the stiffness characteristics of adobe masonry and to thoroughly examine its

25

590

response to tensile loads. The frictional and bonding properties at the interfaces

591

between adobes and timber elements embedded in masonry should also be evaluated.

592

Acknowledgements

593

The funding granted by the University of Cyprus in the framework of research

594

program Experimental and Computational Investigation of the Structural response of

595

Adobe Buildings, as well as the financial support provided by the European Regional

596

Development Fund and the Republic of Cyprus through the Cyprus Research

597

Promotion Foundation in the framework of research program

598

//0609/41 are gratefully acknowledged.

599

References
[1]

Tolles EL, Webster FA, Crosby A, Kimbro EE. Survey of damage to historic adobe
buildings after the January 1994 Northridge earthquake. Los Angeles: The Getty
Conservation Institute; 1996.

[2]

Corazao M, Blondet M. Estudio experimental del comportamiento estructural de las

construcciones de adobe frente a solicitaciones ssmicas [In Spanish]. Lima: Banco
Peruano de los Constructores; 1974.

[3]

Torrealva D, Cerrn C, Espinoza Y. Shear and out of plane bending strength of adobe
walls externally reinforced with polypropylane grids. In: Proceedings of the 14th
World Conference on Earthquake Engineering; 2008; Beijing, China, 12-17 October.

[4]

Ottazzi J, Yep J, Blondet M, Villa-Garcia G, Ginocchio JF. Shaking table tests of

improved adobe masonry houses. In: Proceedings of the 9th World Conference on
Earthquake Engineering; 1988; Tokyo-Kyoto, Japan, 2-9 August.

[5]

Zegarra L, Quiun D, San Bartolom A, Giesecke A. Reforzamiento de viviendas de

adobe existentes, 2da parte: Ensayo ssmico de mdulos [In Spanish]. In: Proceedings
of XI National Congress of Civil Engineering; 1997; Trujillo, Peru, 4-8 November.

26

[6]

Charleson A. Seismic strengthening of earthen houses using straps cut from used car
tires: Construction guide. Oakland: Earthquake Engineering Research Institute; 2011.

[7]

Garcia G, Ginocchio F, Tumialan G, Nanni A. Reinforcing of adobe structures with

FRP bars. In: Proceedings of the 1st International Conference on Innovative Materials
and Technologies for Construction and Restoration; 2004; Naples, Italy, 6-9 June.

[8]

Bartolom AS, Delgado E, Quiun D. Seismic behaviour of two storey model of

confined adobe masonry. In: Proceedings of the 11th Canadian Masonry Symposium;
2009; Toronto, Ontario, Canada, 31 May - 3 June.

[9]

Blondet M, Madueo I, Torrealva D, Villa-Garca G, Ginocchio F. Using industrial

materials for the construction of safe adobe houses in seismic areas. In: Proceedings of
the Earth Build 2005 Conference; 2005; Sydney, Australia, 19-21 January.

[10]

Torrealva D, Neumann J, Blondet M. Earthquake resistant design criteria and testing

of adobe buildings at Pontificia Universidad Catolica del Peru. In: Proccedings of the
Getty Seismic Adobe Project 2006 Colloquium; 2006; Los Angeles, U.S.A., 11-13
April.

[11]

Tolles EL, Krawinkler H. Seismic studies on small-scale models of adobe houses.

Technical Report No. 91: Department of Civil and Environmental Engineering
Stanford University, Stanford: John A. Blume Earthquake Engineering Center; 1990.

[12]

Tolles EL, Kimbro EE, Webster AF, Ginnel SW. Seismic stabilization of historic
adobe structures: Final report of the Getty seismic adobe project. Los Angeles: Getty
Conservation Institute; 2000.

[13]

Dowling D. Seismic strengthening of adobe-mudbrick houses. PhD Thesis: Faculty of

Engineering, Sydney: University of Technology; 2006.

[14]

Figueiredo A, Varum H, Costa A, Silveira D, Oliveira C. Seismic retrofitting solution

of an adobe masonry wall. Materials and Structures. 2013;46(1-2):203-219.

[15]

Yamin LE, Phillips CA, Reyes JC, Ruiz DM. Seismic behaviour and rehabilitation
alternatives for adobe and rammed earth buildings. In: Proceedings of the 13th World

27

Conference on Earthquake Engineering; 2004; Vancouver, B.C., Canada, 1-6 August.

[16]

Meli R, Hernndez O, Padilla M. Strengthening of adobe houses for seismic actions.

In: Proceedings of the 7th World Conference on Earthquake Engineering; 1980;
Istanbul, Turkey, 9-13 September 8-13.

[17]

Rojas J, Ferrer H, Cuenca JS. Dynamic behaviour of adobe houses in Central Mexico.
In: Proceedings of the 14th World Conference on Earthquake Engineering; 2008;
Beijing, China, 12-17 October.

[18]

Tipler J, Worth M, Morris H, Ma Q. Shake table testing of scaled geogrid-reinforced

adobe wall models. In: Proceedings of the 2010 New Zealand Society for Earthquake
Engineering Conference; 2010; Wellington, New Zealand, 26-28 March.

[19]

Ersubasi F, Korkmaz HH. Shaking table tests on strengthening of masonry structures

against earthquake hazard. Natural Hazards and Earth System Sciences.
2010;10(6):1209-1220.

[20]

Che AL, Wu ZJ, Sun JJ, Qi JH. Seismic damage characteristics of rural adobe-wood
building in Gansu province induced by the Wenchuan Great Earthquake. In:
Proceedings of the International Symposium on Geoenvironmental Engineering;
2010; Hangzhou, China, 8-10 September.

[21]

Gomes MI, Lopes M, Brito JD. Seismic resistance of earth construction in Portugal.
Engineering Structures. 2011;33(3):932-941.

[22]

Meyer CS. Numerical simulations of the mechanical behavior of adobe. In:

Chalivendra V, Song B, Casem D, editors. Dynamic Behavior of Materials, Volume 1:
Proceedings of the 2012 Annual Conference on Experimental and Applied Mechanics.
New York: Springer; 2013. p. 557-565.

[23]

Tarque N, Camata G, Spacone E, Varum H, Blondet M. Numerical modelling of the

in-plane behaviour of adobe walls. In: Proceedings of the 8th National Conference on
Seismology and Earthquake Engineering; 2010; Aveiro, Portugal, 20-23 October.

[24]

Tarque N, Camata G, Spacone E, Varum H, Blondet M. Elastic and inelastic

28

parameters for representing the seismic in-plane behaviour of adobe wall. In:
Proceedings of the XIth International Conference on the Study and Conservation of
Earthen Architectural Heritage (Terra2012); 2012; Lima, Peru, 22-27 April.
[25]

Tarque N, Camata G, Spacone E, Varum H, Blondet M. Non-linear dynamic analysis

of an adobe module. In: Proceedings of the XIth International Conference on the
Study and Conservation of Earthen Architectural Heritage (Terra2012); 2012; Lima,
Peru, 22-27 April.

[26]

Furukawa A, Spence R, Ohta Y, So E. Analytical study on vulnerability functions for

casualty estimation in the collapse of adobe buildings induced by earthquake. Bulletin
of Earthquake Engineering. 2009;8(2):451-479.

[27]

Furukawa A, Ohta Y. Failure process of masonry buildings during earthquake and

associated casualty risk evaluation. Natural Hazards. 2009;49(1):25-51.

[28]

Cao Z, Watanabe H. Earthquake response prediction and retrofitting techniques of

adobe structures. In: Proceedings of the 13th World Conference on Earthquake
Engineering; 2004; Vancouver, B.C., Canada, 1-6 August.

[29]

Morales R, Delgado A. Feasibility of construction of two-storey adobe buildings in

Peru. In: Proceedings of the 10th World Conference on Earthquake Engineering;
1992; Madrid, Spain, 19-24 July.

[30]

Illampas R, Ioannou I, Castrillo M.C. Earthen architecture in Cyprus. In: Mecca S,

Dipasquale L, editors. Terra Europae - Earthen architecture in European Union. Pisa:
Edizioni ETS; 2011.

[31]

Azeredo G, Morel JC. Tensile strength of earth mortars and its influence on earth
masonry behaviour. In: Proceedings of the 11th International Conference on Nonconventional Materials and Technologies (NOCMAT2009); 2009; Bath, UK, 6-9
September.

[32]

Kiyono J, Kalantari A. Collapse mechanism of adobe and masonry structures during

the 2003 Iran Bam Earthquake. Bulletin of the Earthquake Research Institute,

29

University of Tokyo. 2004;79(3/4):157-161.

[33]

Illampas R, Ioannou I, Charmpis DC. Overview of the pathology, repair and

strengthening of adobe structures. International Journal of Architectural Heritage.
2013;7(2):165-188.

[34]

Yucheng Y, Liu Y. Earthquake damage to and aseismic measures for earth-sheltered

buildings in China. Tunnelling and Underground Space Technology. 1987;2(2):209216.

[35]

Blondet M, Vargas J, Tarque N. Observed behaviour of earthen structures during the

Pisco earthquake (Peru). In: Proceedings of the 14th World Conference on Earthquake
Engineering; 2008; Beijing, China, 12-17 October.

[36]

Tarque N, Crowley H, Pinho R, Varum H. Displacement-based fragility curves for

seismic assessment of adobe buildings in Cusco, Peru. Earthquake Spectra.
2012;28(2):759-794.

[37]

Benedetti D, Carydis P, Pezzoli P. Shaking table tests on 24 simple masonry

buildings. Earthquake Engineering and Structural Dynamics. 1998;27(1):67-90.

[38]

Tarque N. Numerical modelling of the seismic behaviour of adobe buildings, PhD

Thesis. Universit degli Studi di Pavia: Istituto Universitario di Studi Superiori; 2011.

[39]

Samali B, Jinwuth W, Heathcote K, Wang C. Seismic capacity comparison between

square and circular plan adobe construction. Procedia Engineering. 2011;14(1):21032108.

[40]

Simulia Corp. Abaqus 6.10 theory manual. Rising Sun Mills: Dassault Systmes;
2009.

[41]

Lubliner J, Oliver J, Oller S, Oate E. A plastic-damage model for concrete.

International Journal of Solids and Structures. 1989;25(3):299-326.

[42]

Lee J, Fenves GL. Plastic-damage model for cyclic loading of concrete structures.
Journal of Engineering Mechanics. 1998;124(8):892-900.

[43]

Illampas R, Ioannou I, Charmpis DC. Adobe bricks under compression: Experimental

30

investigation and derivation of stressstrain equation. Construction and Building

Materials. 2014;53(8):83-90.
[44]

Illampas R, Charmpis DC, Ioannou I. An assessment of the compressive strength of

adobe brick assemblages. In: Proceedings of the 9th HSTAM International Congress
on Mechanics; 2010; Limassol, Cyprus, 11-14 July.

[45]

Loureno PB. Anisotropic softening model for masonry plates and shells. Journal of
Structural Engineering. 2000;126(9):1008-1016.

[46]

Almeida JAPP. Mechanical characterization of traditional adobe masonry elements.

MSc Thesis, Department of Civil Engineering: University of Minho; 2012.

[47]

Saetta A, Scotta R, Vitaliani R. Coupled environmental-mechanical damage model of

RC structures. Journal of Engineering Mechanics. 1999;125(8):930-940.

[48]

Creazza G, Matteazzi R, Saetta A, Vitaliani R. Analyses of masonry vaults: A macro

approach based on three-dimensional damage model. Journal of Structural
Engineering. 2002;128(5):646-654.

[49]

Houlsby GT. How the dilatancy of soils affects their behaviour. Report Number
OUEL 1888/91, Soil Mechanics Report 121/91, Department of Engineering Science:
University of Oxford; 1991.

[50]

Green DW, Winandy JE, Kretschman DE. Mechanical properties of wood. In: Wood
handbook: Wood as an engineering materia. General Technical Report FPL-GTR-113,
Madison WI: U.S. Department of Agriculture, Forest Service, Forest Products
Laboratory; 1999.

[51]

Katsaragakis ES. Timber construction [In Greek]. Athens: NTUA Academic

Publications; 2000.

[52]

Loureno PB, Rots J, Blaauwendraad J. Continuum model for masonry: Parameter

estimation and validation. Journal of Structural Engineering. 1998;124(6):642-652.

600

31

601

Figure Captions

602

Fig. 1. General view of the 1:2 scaled model structure tested at the Structures

603

Laboratory of the University of Cyprus.

604

Fig. 2. (a) Test set-up used for the implementation of monotonic lateral loading on the

605

1:2 scaled adobe masonry building. (b) LVDT positions. The displacement results

606

presented in this paper refer to the monitoring points of LVDT1, LVDT3 and

607

LVDT13.

608

Fig. 3. Crack pattern recorded after subjecting the model structure to monotonously

609

increasing horizontal loading tests: (a) rear wall interior surface, (b) rear wall exterior

610

surface and (c) side walls.

611

Fig. 4. Characteristic crack opening recorded: (a) at the centre of the rear walls

612

exterior surface near the structures base and (b) at the exterior surface of the side

613

walls upper section at the vicinity of the openings timber lintel.

614

Fig. 5. Load versus cumulative displacement data envelopes recorded at the upper

615

sections of (a) the rear wall (LVDT13) and (b) the faade (LVDT1) and side

616

(LVDT3) walls. Four limit states (LS1-4) are identified at different levels of

617

deformation. The cracking damage recorded at the interior (upper inset diagram) and

618

exterior (lower inset diagram) surface of the rear wall (a) and at the side wall (b) is

619

presented for each limit state.

620

Fig. 6. 3D FE model developed for simulating the structural response of the scaled

621

adobe building subjected to lateral loading laboratory tests.

622

Fig. 7. Compressive (a) and tensile (b) stress-strain response assigned to the

623

homogenized adobe masonry medium.

32

624

Fig. 8. Plots of deformed mesh (deformation scale x 1) with contour representations

625

of the lateral (along the y axis) displacement distribution.

626

Fig. 9. Contour diagrams with the maximum principal plastic strains computed.

627

Fig. 10. Comparison between the experimental force-displacement data envelopes and

628

the corresponding FE results for the upper sections of (a) the rear wall, (b) the faade

629

and (c) the side wall.

630

33

Figure 1-4
Click here to download Figure: figs1_4.docx

Fig. 1

Fig. 2
(a)

(b)

Fig. 3
(a)

(b)

(c)

Fig. 4
(a)

(b)

Figure 5
Click here to download Figure: fig5.docx

Fig. 5

(a)

LS3
LS2

LS4

LS1

(b)

LS3
LS2

LS4
LS1

Figure 6
Click here to download Figure: fig6.docx

Fig. 6

Figure 7
Click here to download Figure: fig7.docx

Fig. 7
(a)

(b)

Figures 8-9
Click here to download Figure: figs8_9.docx

Fig. 8
U, U2 (mm)
+ 95.97
+ 89.31
+ 82.65
+ 75.98
+ 69.32
+ 62.66
+ 55.99
+ 49.33
+ 42.66
+ 36.00
+ 29.34
+ 22.67
+ 16.01
+ 9.346
+ 2.682
- 3.982
- 10.65

U, U2 (mm)
+ 95.97
+ 89.31
+ 82.65
+ 75.98
+ 69.32
+ 62.66
+ 55.99
+ 49.33
+ 42.66
+ 36.00
+ 29.34
+ 22.67
+ 16.01
+ 9.346
+ 2.682
- 3.982
- 10.65

Fig. 9
PE Max. Principal
(Avg: 75%)
+ 4.237e-02
+ 3.972e-02
+ 3.707e-02
+ 3.443e-02
+ 3.178e-02
+ 2.913e-02
+ 2.648e-02
+ 2.383e-02
+ 2.119e-02
+ 1.854e-02
+ 1.589e-02
+ 1.324e-02
+ 1.059e-02
+ 7.954e-03
+ 5.296e-03
+ 2.248e-03
+ 0.000e+00

PE Max. Principal
(Avg: 75%)
+ 4.237e-02
+ 3.972e-02
+ 3.707e-02
+ 3.443e-02
+ 3.178e-02
+ 2.913e-02
+ 2.648e-02
+ 2.383e-02
+ 2.119e-02
+ 1.854e-02
+ 1.589e-02
+ 1.324e-02
+ 1.059e-02
+ 7.954e-03
+ 5.296e-03
+ 2.248e-03
+ 0.000e+00

Figure 10
Click here to download Figure: fig10.docx

Fig. 10
(a)

(b)

(c)