Comparison between experimental methods for evaluating the compressive strength of existing masonry buildings

Construction and Building Materials 68 (2014) 206–219

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Comparison between experimental methods for evaluating the compressive strength of existing masonry buildings Enrico Sassoni a,c,⇑, Claudio Mazzotti b,c,1, Giulia Pagliai b a

Department of Civil, Chemical, Environmental and Materials Engineering, University of Bologna, via Terracini 28, 40131 Bologna, Italy Department of Civil, Chemical, Environmental and Materials Engineering, University of Bologna, via Risorgimento 2, 40136 Bologna, Italy c Centro Interdipartimentale di Ricerca Industriale (CIRI) Edilizia e Costruzioni, University of Bologna, via Lazzaretto 15/5, 40131 Bologna, Italy b

h i g h l i g h t s  Assessing the load bearing capacity of existing masonries is a fundamental issue.  The suitability of two moderately destructive techniques was here evaluated.  Masonry specimens purposely constructed in laboratory conditions were used.  Masonry compressive strength was calculated from brick and mortar characteristics.  Testing masonry cores completed with cement mortar castings is a promising method.

a r t i c l e

i n f o

Article history: Received 26 February 2014 Received in revised form 10 May 2014 Accepted 29 June 2014

Keywords: Masonry cores Seismic vulnerability Cultural heritage Compressive strength Moderately destructive tests Brick anisotropy Mortar characterization Double punch test Mercury intrusion porosimetry Digital image correlation

a b s t r a c t In this study, the suitability of determining masonry compressive strength by using the following two moderately destructive techniques was evaluated: (i) testing bricks and mortar, separately, and then using formulas available in literature for calculating masonry compressive strength; (ii) testing cylindrical cores, completed with cement mortar castings above and beneath the core, according to a procedure recently proposed by the authors. Two sets of masonry specimens were purposely constructed in laboratory conditions, using bricks and mortar with different mechanical characteristics. From the results of the study, an evaluation of the reliability of the two testing techniques was possible. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction At present, a large part of the built heritage is composed of masonry buildings, as brick–mortar masonry has been a traditional construction technique since ancient times and it is still widely used nowadays for new buildings. Modern awareness of seismic risk characterizing many areas around the Mediterranean basin ⇑ Corresponding author at: Department of Civil, Chemical, Environmental and Materials Engineering, University of Bologna, via Terracini 28, 40131 Bologna, Italy. Tel.: +39 051 2090363; fax: +39 051 2090322. E-mail addresses: [email protected] (E. Sassoni), claudio.mazzotti@unibo. it (C. Mazzotti), [email protected] (G. Pagliai). 1 Tel.: +39 051 2093251; fax: +39 051 2093236. http://dx.doi.org/10.1016/j.conbuildmat.2014.06.070 0950-0618/Ó 2014 Elsevier Ltd. All rights reserved.

and recognition of natural weathering of materials, rise a serious challenge to keep using masonry as a reference construction system. In this framework, for both human health safeguard and cultural heritage preservation, the assessment of the structural safety of existing masonries and the evaluation of the effects of possible retrofitting interventions are fundamental issues to be addressed [1]. To this purpose, a better knowledge of the geometrical and mechanical characteristics of existing masonries leads to a more reliable intervention. Among the latter parameters, the most important mechanical property to be determined is masonry compressive strength. Even though for new masonries a single, very effective laboratory experimental test can be used [2,3], when dealing with existing masonries, several experimental techniques

E. Sassoni et al. / Construction and Building Materials 68 (2014) 206–219

207

Nomenclature Eb Em Em,cast fbc fbcn fbc,// fmc fmc fmc fmc,dpt

Young’s modulus of elasticity of bricks (GPa) Young’s modulus of elasticity of mortar (GPa) Young’s modulus of elasticity of mortar used for castings (GPa) brick mean compressive strength (MPa) brick normalized compressive strength (MPa) brick compressive strength in the direction parallel to brick’s longest edge (MPa) mortar mean compressive strength (MPa) mortar mean compressive strength, calculated according to Eq. (1) (MPa) mortar mean compressive strength, calculated according to Eq. (2) (MPa) mortar mean compressive strength, determined by double punch test on samples obtained from the masonry joints (MPa)

fmc,dpt,prism mortar mean compressive strength, determined by double punch test on samples sliced from standard prisms (MPa)

can be adopted, differing with respect to their invasiveness, technical challenge and economic costs. The most direct, as well as largely destructive, method is subjecting a masonry portion (about 90  180 cm2) to in situ compression test; this can be accomplished by separating the masonry panel from the wall along three sides (demolition or cutting are often used) and then loading it from the top by means of a steel beam and hydraulic jacks [4]. The method has the advantage of directly providing masonry compressive strength and modulus of elasticity, if displacement transducers are applied, in its real configuration. However, because of its invasiveness (a wall is largely destroyed), technical challenge and cost, this method is currently performed only in a few cases, while less destructive and less expensive methods are usually preferred. A moderately destructive technique that can be performed in situ is subjecting a masonry portion to compression test by means of two flat-jacks [5]. Two horizontal mortar joints, at a distance of about 50 cm, are removed by saw cutting and, in their place, two flat-jacks are inserted; finally, compression load is increased until failure of the masonry portion comprised between the jacks. This technique has the advantage that the tested masonry portion is relatively small. However, its accuracy is somehow limited by the influence of the masonry surrounding the jacks, involved in a complex diffusion problem. Finally, in case of masonry with poor shear capacity, the test could be arrested due to the inability of the overlying masonry to provide a sufficient contrast to the action exerted by the flat-jacks. Following a more indirect approach, the masonry compression strength can also be obtained as a function of its components strength (bricks and mortar). To this purpose, the two components have to be separately sampled from the wall. In order to properly combine the two strengths, several formulas were proposed in the scientific literature (see for instance review in [6]). In addition, National Codes (such as Italian Code for constructions [2]) and International Codes (such as Eurocode 6 [3]) provide tables and formulas that allow determination of masonry characteristic compressive strength from brick and mortar compressive strength. This method has the advantage of requiring only moderately destructive sampling, as both brick and mortar samples can be easily obtained from small diameter core-drills [7]. As an example, from 50 mm diameter cores, made exclusively of brick, cylindrical samples can be obtained for determining brick compressive

fmf fm,cast fwc fwc fwc hb hm N sw WA WAsat

a mb mm mm,cast

mortar mean flexural strength (MPa) compressive strength of mortar used for castings (MPa) wall mean compressive strength (MPa) masonry mean compressive strength, calculated using fmc masonry mean compressive strength, calculated using fmc average brick thickness (mm) average mortar joint thickness (mm) number of tested specimens wall compressive strength standard deviation (MPa) brick water absorption after immersion in water (%) brick water absorption after saturation in water for 3 days (%) hm/hb (–) brick Poisson’s ratio (–) mortar Poisson’s ratio (–) Poisson’s ratio of mortar used for castings (–)

strength. In addition, from 50 mm diameter cores, made of two bricks portions and a central mortar layer, flat prismatic samples of mortar (about 40 mm  40 mm  10 mm) can be extracted and used for determining mortar compressive strength by double punch test [7,8]. However, this indirect method for assessing masonry compressive strength has some drawbacks as well: (i) in the case of bricks, the testing direction is generally different from the loading direction in the wall, as the cores are obtained perpendicular to the loading direction in the wall; consequently, not negligible differences in brick compressive strength may be found [9]; (ii) in the case of mortar, the size of the testing specimens and the size of the two punches significantly affect the estimation of mortar compressive strength, hence careful interpretation of test results should be performed [8,10–12]. For more straightforward and reliable results to be obtained, brick and mortar samples should be obtained by removing entire bricks and mortar joints from the wall by chisel; however, this makes the testing procedure more invasive. Another technique is based on the use of large diameter cores (150 mm), subjected to uniaxial compressive test by loading in the same direction as the masonry is usually loaded in the wall (i.e. by applying load perpendicular to the horizontal mortar joints), by making use of two steel cradles [13,14]. The cores should be core-drilled so as to have two horizontal mortar joints and a vertical one near the center of the core, to ensure that the brickwork is properly represented; in this way, the core failure mechanism is expected to suitably resemble that of the masonry from which the core was extracted [13,14]. Even if this technique has the advantage of testing masonry cores in the same direction as they are loaded in the original wall, it has the limitation that cores with 150 mm diameter are not very often available. Moreover, the use of steel-made cradles for loading the cores might lead to stress concentration if some irregularities are present on core lateral surface. To overcome the above-mentioned limitations, an alternative version of the compressive test on masonry cores has recently been proposed by the authors [8]. The method is based on the idea of subjecting cores with 100 mm diameter to uniaxial compression, by completing the cores with two high-strength mortar horizontal castings, one below and one above the core. The mortar castings are the transition from the core curved surface to the test machine’s flat horizontal surfaces, where vertical distributed load

208

E. Sassoni et al. / Construction and Building Materials 68 (2014) 206–219

is applied; they also resemble the confinement effect of the masonry surrounding the core in the real wall. Compared to previously proposed methods, requiring 150 mm diameter cores and involving steel-made cylindrical cradles for stress distribution [13,14], the newly proposed method has the advantage that 100 mm diameter cores are more easily available than cores with larger diameter and that the high-strength mortar castings, instead of the steel-made cylindrical cradles, are expected to allow a more uniform stress distribution and prevent stress concentration due to an elastic modulus more similar to that of masonry. By comparison with other independent methods for measuring masonry compressive strength, this technique proved to provide a fair estimation of masonry compressive strength [8]. However, as only results obtained for a specific study case are available so far, more systematic tests on controlled specimens are needed. Therefore, in this study the suitability of using two moderately destructive techniques for evaluating masonry compressive strength was evaluated. A systematic comparison was carried out among compressive strength values obtained from reference masonry panels, tested according to standard procedures, and values obtained from (i) testing bricks and mortar, separately, and (ii) testing cylindrical cores, completed with cement mortar castings. For such a comparison, several masonry panels were purposely constructed, using the same materials and technologies; part of the panels was directly subjected to compression tests, while the other panels were used for obtaining brick and mortar samples and cylindrical cores. To also evaluate the influence of constitutive materials’ mechanical properties, two sets of masonry specimens were constructed, using bricks and mortar with different mechanical characteristics. Crack patterns evolution on the core surface during the tests was recorded and analyzed by using digital image correlation (DIC) technique. 2. Materials and methods In order to verify the effectiveness of the proposed experimental technique [8], two different types of masonry were considered, by using different types of bricks and mortars. For masonry type 1 (MT1), bricks and mortar with significantly different compressive strength were used, to resemble ancient masonries where mortar is the weakest and the most deformable component. For masonry type 2 (MT2), bricks and mortar with comparable compressive strength were used, to resemble more recent masonries where mortar has good mechanical properties. 2.1. Bricks and mortars Fired-clay solid bricks with standard nominal dimensions (250 mm  120 mm  55 mm) were used for both masonries. Their mechanical properties were assessed as described in § 2.4.1. In particular, bricks with better mechanical properties (higher compressive strength, higher Young’s modulus of elasticity, lower Poisson’s ratio) were used for MT1, with the aim of maximizing the mechanical differences between bricks and mortar. Two different lime-cement mortars were used. For MT1, a ready-mixed mortar (classified as M5 according to European Standard EN 998-2 [15]) was used. As reported in Table 1, the low binder/aggregate ratio (about 1:6) and the high water/binder ratio (about 1.3, as recommended in the product technical data sheet) were as to produce poor mechanical properties. For MT2, the mortar was prepared on site using commercially available lime, cement (CEM II/B-LL 32.5 R) and river sand. Firstly, sand, lime and cement were mixed in the proportions reported in Table 1, corresponding to a high binder/aggregate ratio (about 1:2). Then, water was added in the minimum amount (water/binder ratio about 0.3, Table 1), so as to achieve good mechanical properties.

2.2. Masonry samples For each masonry type, four panels were constructed in total: three for standard compression test and one for core-drilling and brick and mortar samples. The panels were constructed by expert workmanship, employing traditional construction techniques. In particular, to avoid dry bricks from absorbing water from the fresh mortar, thus significantly altering its water/binder ratio, the bricks were immersed in water before use. For MT1, intended to exhibit low mechanical properties, bricks were immersed for 3 min before use, while for MT2, designed to have higher mechanical properties, bricks were immersed for 10 min. For evaluating water absorption after the above indicated immersion time, three bricks per type were weighed before and after the immersion, while water absorption after complete saturation was evaluated by weighing bricks before and after immersion in water until constant weight (3 days). 2.2.1. Reference panels For each masonry typology, three panels with nominal dimensions 770 mm  770 mm  250 mm were constructed, according to EN 1052-1 [16]. These dimensions corresponded to a three-brick length, twelve-brick course height and two-leaf thickness (Fig. 1). As the actual height of bricks used for MT1 ranged between 5.5 and 6.0 cm, in the case of MT1 only eleven brick courses were constructed to reach the same nominal height (Fig. 1). The average thickness of the mortar joints, in the case of MT1, was 10–15 mm for horizontal joints and 25 mm for vertical joints (such variability was needed for accommodating some dimensional irregularities in brick used for MT1). In the case of MT2, mortar joints had an average thickness of 15 mm for horizontal joints and 15 mm for vertical joints. Two steel plates (10 mm thickness) were placed at the bottom and over the top of the panels to ensure good stress distribution and good contact between the panels and the loading plates. Between the panels and the plates, a 1 cm mortar bed was placed. The panels were left to cure in laboratory conditions for 2 months before testing (wet clothes being used to keep the relative humidity high in the first 48 h after construction). Even if it is known that a longer curing time would be required for complete hardening of lime-containing mortars (such as that used for MT1), still the adopted curing time of 2 months was considered as sufficient, as the mortar used for MT1 is a commercial product classified as M5 (i.e., having a compressive strength of at least 5 MPa after 28 days). 2.2.2. Cores For each masonry typology, one additional panel was constructed for coredrilling. The panel had nominal dimensions 1030 mm  1290 mm  250 mm, corresponding to a four-brick length, twenty-brick course height and two-leaf thickness (Fig. 2). For each typology, the mortar joint thickness was similar to that of the reference masonry panels. Two ribbed steel plates were placed at the bottom and over the top of the masonry panels and four threaded steel rods (16 mm diameter) were used to apply a pre-compression load sufficient to prevent any damage when the walls were rotated by 90° to facilitate the core-drilling operation. After curing for 50 days, the panels were core-drilled according to the scheme reported in Fig. 2. In total, 33 cores for MT1 and 44 for MT2 were obtained (the numbers differing in the two cases since in the first case some cores underwent brick–mortar detachment and hence were used for other tests). In particular, three typologies of cores were available (Fig. 3): (i) cores with only a horizontal mortar joint (labeled as ‘‘H’’ type cores); they amounted to 9 for MT1 and 13 for MT2; (ii) cores with both horizontal and vertical mortar joints, where the brick not divided by the vertical mortar joint was also continuous along the length of the core (labeled as ‘‘H + V1’’ type cores); they amounted to 12 for MT1 and 16 for MT2; (iii) cores with both horizontal and vertical mortar joints, where the brick not divided by the vertical mortar joint was divided in two transversally (labeled as ‘‘H + V2’’ type cores); they amounted to 12 for MT1 and 15 for MT2. 2.2.3. Mortar joint samples The remaining parts of the panels used for core-drilling, as well as some cores where brick–mortar detachment occurred, were used for obtaining mortar samples to be tested separately. After separation by chisel (Fig. 4a), mortar pieces were sawn to obtain prismatic specimens with 40 mm  40 mm  10 mm dimensions (Fig. 4b), to be tested as described in § 2.4.2. So-obtained mortar samples resemble specimens that can be obtained from in situ core-drilling of existing buildings. 2.3. Testing procedures

Table 1 Composition and mix-design of the two mortars used for the two masonry typologies.

Lime (wt%) Cement (wt%) Sand (wt%) Binder/aggregate ratio (wt/wt) Water/binder ratio (wt/wt)

MT1

MT2

3 11 86 0.2 1.3

18 18 64 0.6 0.3

2.3.1. Bricks For each masonry typology, brick compressive strength (fbc) was measured according to European Standard EN 772-1 [17] on 6 cubic samples (50 mm side), obtained from 3 bricks by sawing. The specimens were tested using an Amsler universal machine (load capacity 120 kN), by loading in the same direction as bricks are loaded in the walls. For measuring Young’s modulus of elasticity (Eb) and Poisson’s ratio (mb), two further cylindrical samples (38 mm diameter, 90 mm height) were core-drilled parallel to brick’s longest side (25 cm) (Fig. 4e and f). In this way, Eb and mb were

E. Sassoni et al. / Construction and Building Materials 68 (2014) 206–219

209

Fig. 1. Masonry panels used for compression test: (a and b) scheme of panel geometry and photo of a representative panel for MT1; (c and d) scheme of panel geometry and photo of a representative panel for MT2. Vertical and horizontal displacement transducers used for displacement measurement are illustrated in gray in the schemes and are visible in the photographs. measured in the direction orthogonal to the usual brick loading direction. This was made necessary because: (i) cylindrical samples should be used, to prevent the edges of prismatic samples from influencing strain measurement; (ii) the cylinders should have height higher than twice the diameter, to prevent the friction between the sample and the loading plates from altering the measurement. The brick cylinders were instrumented with three bi-directional strain gauges (placed at an angular distance of 120°, Fig. 4h), to measure longitudinal and transversal strain during compression test, carried out using the same loading machine as above. 2.3.2. Mortar For each masonry typology, mortar flexural strength (fmf) and compressive strength (fmc) were measured according to European Standard EN 1015-11 [18] on 6 prismatic samples (40 mm  40 mm  160 mm), prepared during the construction of the masonry panels. After curing for 2 months in a climatic chamber (20 °C and 98% RH), mortar prisms were tested using an Amsler universal machine (load capacity 100 kN). For measuring Young’s modulus of elasticity (Em) and Poisson’s ratio (mm), two cylindrical samples (38 mm diameter, 100 mm height) were core-drilled from 100 mm side cubes (Fig. 4g), prepared during the construction of the masonry panels. The mortar cylinders were instrumented and tested as described above for the brick cylinders (Fig. 4i). In addition, the mortar specimens (40 mm  40 mm  10 mm, Fig. 4b), obtained by sawing the mortar joints extracted from the masonry cores (Fig. 4a), were tested by double punch test (DPT), according to the method described in [19,20]. In brief, DPT consists in loading the mortar specimens by means of two circular steel platens (in this study, having 20 mm diameter) until failure is reached (Fig. 4c and d). The compressive strength determined by double-punch test (fmc,dpt) is then calculated as the ratio of the failure load to the loaded area, corresponding to the cross section of the 20 mm diameter circular plates.

2.3.3. Reference panels For each masonry typology, after curing for 2 months, the three panels were subjected to compression test using a Losenhausen loading machine (maximum load 6000 kN, loading speed 2 kN/s), according to European Standard EN 1052-1 [16]. For measuring vertical and horizontal displacements during the test, the panels were instrumented with five displacement transducers, four placed vertically near the four vertical corners and one disposed horizontally at the center of panel (Fig. 1). To optimize the contact between the steel plates at the bottom and over the top of the panels and the loading plates, two wood boards (5 mm thickness) were used. The masonry wall compressive strength (fwc) was calculated as the ratio of the failure load to the cross section of the panels. 2.3.4. Cores In order to subject cores to compression test along the same direction as they bear the load in the wall, they were placed with the longitudinal axis horizontal and two mortar castings were realized above and beneath the sample, using a ready-mixed cement mortar and according to the method proposed in [8]. The castings had 80 mm width, 250 mm length and about 40 mm minimum height (Fig. 5a and b). After 7 days of curing of the mortar used for castings and 2 months of overall curing for the masonry cores (i.e., the same curing period as the reference masonry panels), the specimens were subjected to compressive test along a direction orthogonal to their longitudinal axis. An Amsler universal loading machine (maximum load 600 kN) was used, at a constant loading rate of 2.0 kN/s. The masonry compressive strength deduced from the cores (fwc) was calculated as the ratio of the failure load to the horizontal mortar joint area. For each typology illustrated in Fig. 3, two cores were tested also using the digital image correlation (DIC) technique, which measured the deformation of one vertical surface of the cylinder. In order to improve the DIC accuracy, a pattern of distinct speckles was created on samples vertical surface and monitored during

210

E. Sassoni et al. / Construction and Building Materials 68 (2014) 206–219

Fig. 2. Masonry panels used for core-drilling: (a and b) scheme of panel geometry and photo of the panel for MT1; (c and d) scheme of panel geometry and photo of the panel for MT2. In panel schemes, the points where cylindrical cores were extracted are reported. the test. Therefore, the surface of the cores was preliminarily prepared by applying a layer of white paint, over which black speckles were created by airbrushing (Fig. 5c). Two cameras were used for image acquisition during the test, so that 3D measurements of core surface deformation were carried out. In some of the cores (6 for ‘‘H’’, 3 for ‘‘H + V1’’ and 4 for ‘‘H + V2’’ typologies), brick–mortar detachment occurred during core-drilling or handling. As this inconvenient may also occur during in situ operations, it was decided to re-assemble the detached cores and then test them in the same way as the other cores. In this way, it was possible to investigate whether and to which extent the brick–mortar detachment would alter the evaluation of masonry compressive strength.

3. Results and discussion 3.1. Bricks and mortars The results of the mechanical characterization of bricks and mortar, for the two masonry typologies, are reported in Table 2. For masonry MT1, intended to reproduce historical masonries where mortar is the weakest component, bricks exhibit high

compressive strength (fbc = 39.8 MPa), while mortar has rather poor mechanical properties (fmc = 8.4 MPa). In terms of deformability, while bricks and mortar have comparable longitudinal modulus of elasticity (Eb = 10.0 GPa, Em = 9.0 GPa), they significantly differ in terms of Poisson’s ratio (mb = 0.13, mm = 0.22), mortar resulting the most deformable component. For MT2, designed to reproduce more recent masonries where mortar has better characteristics, it has basically the same compressive strength as bricks (fbc = 19.3 MPa, fmc = 18.6 MPa) and the same modulus of elasticity (Eb = 6.5 GPa, Em = 6.3 GPa), but it has a lower transversal deformability than bricks (mb = 0.22, mm = 0.15). Brick compressive strength was tested in different directions (perpendicular and parallel to brick’s longest edge). Notably, in the case of MT1, along the parallel direction a sensibly lower compressive strength was found (fbc,// = 17.7 MPa, compared to fbc = 39.8 MPa), in agreement with other results obtained by the authors [21]. On the contrary, in the case of MT2, values in the two directions were substantially the same (fbc,// = 19.6 MPa,

E. Sassoni et al. / Construction and Building Materials 68 (2014) 206–219

211

Fig. 3. Scheme illustrating the different core typologies: cores ‘‘H’’, having only a horizontal mortar joint; cores ‘‘H + V1’’, having both horizontal and vertical mortar joints, where the brick not divided by the vertical mortar joint is also continuous along the length of the core; cores ‘‘H + V2’’, having both horizontal and vertical mortar joints, where the brick not divided by the vertical mortar joint is divided in two transversally.

Fig. 4. Testing procedures of mortars and bricks. Mortar samples obtained by chisel from masonry joints (a) were sawn to obtain prismatic specimens (b) to be tested by double punch test (c and d). Cylindrical specimens of brick (e and f) and mortar (g) were core-drilled from entire bricks and mortar cubes, respectively, and then subjected to compression test, using three bi-directional strain gauges for strain measurement (h and i).

fbc = 19.3 MPa). The difference in compressive strength found for different testing directions may be partly due to the different shape of tested specimens (cubes for fbc and cylinders – / > 2h – for fbc,//). However, given the different behavior of the bricks used for the two different masonry typologies, the microstructure and the manufacturing process of the two types of bricks played a significant role. Indeed, in the case of bricks used for MT1, numerous cracks parallel to brick’s longest side were visible after core-drilling (Fig. 4e). These cracks, while playing a secondary role when bricks were tested in the perpendicular directions, as cracks tend to close, presumably played an important role when bricks were tested in the parallel direction, as they tend to open, thus reducing the final compressive strength.

In the case of mortars, the compressive strength values found by double punch test resulted significantly higher than those determined on standard prisms: for MT1, fmc,dpt = 26.1 MPa while fmc = 8.4 MPa; for MT2, fmc,dpt = 73.1 MPa while fmc = 18.6 MPa. Such a difference may be explained considering:  The effect of non-standard sample dimensions and testing procedure. According to several studies, fmc,dpt is systematically higher than fmc [10,11,19,20]. This can substantially be attributed to the confinement effect exerted by the mortar surrounding the loaded area [22]: according to FE modeling of the most frequent DPT conditions, the horizontal confinement can amount to about 6% of the vertical stress, which can result in

212

E. Sassoni et al. / Construction and Building Materials 68 (2014) 206–219

Fig. 5. Preparation of cylindrical cores for testing: (a) wood-made moulds for cement mortar castings; (b) core completed with top and bottom castings; and (c) core preparation for DIC technique, by spraying of white pain and airbrushing of black ink spots on sample surface.

Table 2 Results of brick and mortar characterization (coefficient of variation in brackets). Bricks fbc (MPa) fbc,// (MPa) Eb (GPa)

mb WA (%) WAsat (%)

MT1 N=6 N=2 N=2 N=2 N=3 N=3

Mortar fmf (MPa) fmc (MPa) fmc,dpt (MPa) fmc,dpt,prism (MPa) Em (GPa)

mm

39.8 17.7 10.0 0.13 8.0 17.3

MT2 (7.8%) (1.0%) (9.0%) (7.7%) (2.5%) (4.0%)

MT1 N=6 N = 12 N = 23 N = 20 N=2 N=2

2.4 8.4 26.1 7.7 9.0 0.22

MT1 and MT2

fmc,cast (MPa) Em,cast (GPa)

13.2 19.4 0.21

mm,cast

(6.2%) (9.3%) (6.2%) (4.5%) (2.2%) (1.1%)

MT2 (12.5&) (9.5%) (25.7%) (26.8%) (1.1%) (4.5%)

Mortar used for castings N=2 N=2 N=2

19.3 19.6 6.5 0.22 13.9 17.9

4.2 18.6 73.1 23.0 6.3 0.15

(11.6%) (6.8%) (15.7%) (8.7%) (11.1%) (20.0%)

(1.6%) (8.0%) (10.3%)

significant increases in mortar compressive strength, compared to the uniaxial strength [22]. The difference between fmc,dpt and fmc reportedly increases if: (i) the size of the loading platens increases [19] or, which is the same, the base edge length of the test specimen decreases [10,11]; (ii) for a given size of the loading platens and a given length of sample base edge, the slenderness (i.e., the height-to-base edge length ratio) of the specimen decreases [10,11]. In both cases, the apparent increase in mortar compressive strength is owed to the restraining effect of the loading platens, that becomes more relevant.  The effect of the different microstructure of mortar in the standard prisms and in the masonry joints. According to [19], mortar compressive strength determined by double punch test on samples obtained from masonry joints was found to be always considerably higher than strength determined on samples sliced from standard prisms. This was attributed to two phenomena that can occur in the masonry joints, and not in the steel moulds used for preparing standard prisms: (i) the lowering of the water/cement ratio, as a consequence of water absorption from the fresh mortar by bricks; (ii) the improved curing, as a consequence of the moisture stored in brick pores [19]. To verify whether, in the present case, the difference in compressive strength determined for the standard prisms and for the

masonry joints may actually be a consequence of the above described factors, specific investigations were carried out. Firstly, to evaluate the influence of the different testing procedure for the same material, mortar compressive strength found by standard testing on prisms (fmc) was compared to that found by double punch test on samples obtained by slicing two additional standard prisms, so as to obtain samples with 40 mm  40 mm  10 mm dimensions (fmc,dpt,prism). As the prisms that were sliced and tested by double punch test were prepared together with the other prisms used for standard testing and were cured in the same conditions for the same period of time, the effect of the different testing procedure (standard testing vs double punch test) could be isolated. As the results reported in Table 2 point out, the different testing procedures actually determined different mortar compressive strength values: for MT1, fmc = 8.4 MPa while fmc,dpt,prism = 7.7 MPa; for MT2, fmc = 18.6 MPa while fmc,dpt,prism = 23.0 MPa. In particular, the fact that for MT2 fmc,dpt,prism is higher than fmc, while for MT1 it is the other way around, is in agreement with results found in [12], where a similar dependence on mortar compression strength was found. However, as can be observed in Table 2 for both masonry typologies, fmc,dpt,prism is far from fmc,dpt, determined by using the same testing technique on mortar samples obtained from the masonry joints; therefore a difference in material microstructure between standard prisms and masonry joints seems reasonable. To verify such an hypothesis, mortar microstructure was investigated in terms of total open porosity and the pore size distribution. Indeed, in the present case, where for a given mortar composition only the curing conditions were different, porosity and pore size distribution (being strongly dependent on the curing conditions) were identified as significant parameters. Two mortar samples, one coming from a standard prism and one coming from a mortar joint, were subjected to mercury intrusion porosimetry (MIP), using a Fisons Macropore Unit 120 and a Porosimeter 2000 Carlo Erba. As can be observed in Fig. 6, for both masonry typologies, samples coming from the standard prisms exhibit open porosity about 50% higher than samples coming from the masonry joints. Accordingly, the average pore radius of mortar samples coming from the masonry joints is sensibly smaller than that in the case of the standard prisms. In light of the above described experimental findings, the hypothesis that the differences between fmc and fmc,dpt are owed, on the one hand, to the different testing technique and, on the other hand, to the different mortar porosity and pore size distribution, seems confirmed.

E. Sassoni et al. / Construction and Building Materials 68 (2014) 206–219

213

Fig. 6. Pore size distribution, open porosity (OP) and average pore radius (rav) of mortar samples obtained from standard prisms and masonry joints for the two masonry typologies.

Table 3 Results of compressive test of reference masonry panels (coefficient of variation in brackets).

Panels

MT1 fwc (MPa)

MT2 fwc (MPa)

1 2 3 Average

8.5 7.6 9.3 8.4 (10.7%)

10.4 11.2 9.1 10.2 (10.8%)

capacity depends, besides brick and mortar compressive strength, also on several additional factors, such as brick arrangement, mortar joint thickness, brick and mortar deformability. As expected, the compressive strength of MT2 is higher than that of MT1, in accordance with the higher compressive strength of mortar used for MT2, compared to MT1. This effect prevails on the lower compressive strength of the bricks from MT2 (fbc = 19.3 MPa for MT2 vs fbc = 39.8 MPa for MT1).

3.2. Masonry compressive strength determination from reference panels

3.3. Masonry compressive strength determination from constitutive materials

The results of the compression tests on the masonry panels are summarized in Table 3, while graphs plotting the compressive stress vs the vertical strain for each masonry panel are reported in Fig. 7. Pictures illustrating two representative panels (one made of MT1 and one of MT2) after failure are reported in Fig. 8. For both MT1 and MT2, the results exhibited limited dispersion, as denoted by the relatively low coefficient of variability (CoV = 11% for both masonry typologies). In the case of MT1, masonry compressive strength has a value very close to that of the weakest component, i.e. mortar (fwc = 8.4 MPa, fmc = 8.4 MPa). In the case of MT2, masonry compressive strength (fwc = 10.2 MPa) was lower than strength of its components (fbc = 19.3 MPa, fmc = 18.6 MPa), which is not uncommon [23] as masonry bearing

The first method for indirectly assessing masonry compressive strength was to obtain brick and mortar mechanical properties separately and then to make use of formulas available in the scientific literature to deduce masonry mean compressive strength. In addition, by making use of formulas and tables provided by National and International Codes, masonry characteristic compressive strength could also be obtained. However, so-calculated characteristic values of masonry compressive strength are not directly comparable with the mean values determined experimentally in this study (since three specimens were tested for each masonry typology, no statistically significant characteristic value can be calculated from experimental data). Therefore, in this study a comparison was carried out only between masonry mean

Fig. 7. Graphs illustrating compressive stress against vertical strain for each panel of the two masonry typologies.

214

E. Sassoni et al. / Construction and Building Materials 68 (2014) 206–219

Fig. 8. Photos of two representative panels for MT1 (a) and MT2 (b) at the end of the compression test.

Table 4 Masonry compressive strength calculated from brick and mortar mechanical characteristics, according to different sources (for MT1: fbc = 39.8 MPa; fmc ¼ 17:5 MPa; fmc ¼ 6:0 MPa; hm = 12.5 mm; hb = 55 mm; a = hm/hb = 0.23. For MT2: fbc = 19.3 MPa; fmc ¼ 43:6 MPa; fmc ¼ 16:7 MPa; hm = 15 mm; hb = 55 mm; a = hm/hb = 0.27). Source

Formula

MT1

fwc = (1–0.8 a1/3)  [fmc + 0.4 (fbc  fmc)] fwc = 0.7  fbc1/2  fmc1/3 fwc = 2/3 fbc + 0.1 fmc

Tassios 1985 [6] (good quality materials) Bröcker 1961 [6] (medium quality materials) Tassios 1987 [6] (poor quality materials)

MT2

fwc (MPa)

fwc (MPa)

fwc (MPa)

fwc (MPa)

13.5 11.5 28.3

10.0 8.0 27.1

16.3 10.8 17.2

8.5 7.9 14.5

Table 5 Masonry compressive strength derived from testing of cylindrical cores. MT1 N All the cores

MT2 fwc (MPa)

CoV (%)

‘‘H + V’’/‘‘H’’

Detached/ not detached

N

fwc (MPa)

CoV (%)

‘‘H + V’’/‘‘H’’

Detached/ not detached

33

10.6

18.9

–

–

44

13.3

12.0

–

–

Cores with ‘‘H’’ joint

All Non-detached Detached

9 3 6

12.5 13.6 12.0

20.0 24.3 17.5

– – –

– – 88%

13 11 2

13.5 13.6 12.6

7.4 7.3 4.0

– – –

– – 92%

Cores with ‘‘H + V1’’ joint

All Non-detached Detached

12 9 3

9.1 9.4 8.2

12.1 9.6 12.2

– 69% -

– – 87%

16 14 2

13.3 13.5 12.5

17.3 17.8 12.0

– 99% –

– – 93%

Cores with ‘‘H + V2’’ joint

All Non-detached Detached

12 8 4

10.6 10.9 10.2

9.4 10.1 3.9

– 80% –

– – 94%

15 13 2

13.6 13.8 12.8

9.6 10.1 8.6

– 101% –

– – 93%

compressive strength values, obtained experimentally and calculated using formulas available in the scientific literature. In particular, the formulas proposed by Tassios [6], reported in Table 4, were considered. These formulas differ depending on the quality of the constitutive bricks and mortar, ranging from good to medium to poor [6]. Since in practical cases, for determining the compressive strength of an existing masonry building, no standard mortar prisms are available but mortar strength can only be assessed by non-standard testing of samples obtained from masonry joints, for fwc only mortar mechanical properties determined by DPT were considered. However, experimentally determined fmc,dpt values were found to significantly overestimate the actual mortar compressive strength, for the reasons discussed in § 3.1. Consequently, two empirical formulas, proposed in the literature to derive an equivalent mortar compressive strength value ðfmc Þ from values found by double punch test (fmc,dpt), were considered. The first formula was proposed by Pascale et al. [12] and involves a linear relationship between the two variables:

fmc ¼ 0:5552f mc;dpt þ 3:0678 ðin MPaÞ:

ð1Þ

The formula was obtained for mortars with poor mechanical properties (classes M1 to M4, according to [15]) by subjecting standard prisms to (i) standard compression test and (ii) double punch test, after slicing to obtain samples with 10–13 mm thickness [12]. The second formula was proposed by Drdàcky´ et al. [10] and introduces a dependence on sample’s slenderness:

fmc ¼ fmc;dpt =ðh=aÞ

1:065

ðin MPaÞ;

ð2Þ

where h and a are the height and the base edge length of the sample, respectively. The formula was obtained from tests carried out on lime–metakaoline–cement mortars, using specimens with increasing slenderness [10]. In the cited study, several different formulas were developed for different types of mortar (i.e., lime, lime–metakaoline, lime–metakaoline–cement). In the present study, the formula proposed for lime–metakaoline–cement mortars was chosen because the composition of this mortar was the most similar one to that of mortars used for MT1 and MT2. It should also be noted that, while Eq. (1) was obtained by subjecting samples to double punch test using circular platens with 20 mm diameter, Eq. (2) was obtained by testing samples between two square plates, with dimensions depending on sample sizes.

E. Sassoni et al. / Construction and Building Materials 68 (2014) 206–219

215

Fig. 9. DIC results of two representative cores of MT1: horizontal strains for cores without (a) and with (b) the vertical mortar joint.

By using the above reported formulas, mortar compressive strength values closer to those found by testing standard prisms were obtained: for MT1, fmc ¼ 17:5 MPa and fmc ¼ 6:0 MPa, while fmc = 8.4 MPa; for MT2, fmc ¼ 43:6 MPa and fmc ¼ 16:7 MPa, while fmc = 18.6 MPa. In particular, by using the formula proposed by Pascale et al. [12], the calculated fmc values still overestimate the fmc values, especially in the case of MT2, probably because mortars tested in the present study are rather more resistant than those tested in the study by Pascale et al. and for which the formula was originally derived. On the contrary, by using the formula

proposed by Drdàcky´ et al. [10], the calculated fmc values are very close to fmc values, especially for the more resistant masonry typology (MT2). When fmc and fmc values are introduced as input parameters for calculating masonry compressive strength (fwc and fwc , respectively) the results reported in Table 4 are obtained. For MT1, the most accurate estimation of masonry compressive strength is obtained by using fmc and adopting the formula proposed by Bröcker [6] for medium quality materials. In this way, a difference of about 5% between the calculated and the measured

216

E. Sassoni et al. / Construction and Building Materials 68 (2014) 206–219

Fig. 10. DIC results of two representative cores of MT2: horizontal strains for cores without (a) and with (b) the vertical mortar joint.

values is obtained (fwc ¼ 8:0 MPa while fwc = 8.4 MPa). In the case of MT2, the masonry compressive strength value which is closest to that measured experimentally (fwc = 10.2 MPa) is obtained by adopting the formula proposed by Bröcker [6] and using fmc ðfwc ¼ 10:8 MPaÞ. However, using the formula proposed by Bröcker [6] for medium quality materials and the extraordinarily high mortar compressive strength value found using the formula proposed by Pascale et al. [12] ðfmc ¼ 43:6 MPaÞ seems not very representative, even if measured and calculated masonry compressive

strengths closely match. If the more realistic mortar compressive strength value obtained by using the formula proposed by Drdàcky´ et al. [10] is used ðfmc ¼ 16:7 MPaÞ and, consistently, the formula proposed by Tassios for good quality materials [6] is adopted, a masonry compressive strength value that is less close to the measured one, but still more realistic, is obtained ðfwc ¼ 8:5 MPaÞ while fwc = 10.2 MPa). In this way, a difference of about 16% between calculated and measured values is obtained, which can still be considered a fair enough estimation.

E. Sassoni et al. / Construction and Building Materials 68 (2014) 206–219

3.4. Masonry compressive strength determination from cylindrical cores The second method for indirectly assessing masonry compressive strength was testing cylindrical cores according to the procedure described in Section 2.3.4. For the two masonry typologies, the resulting masonry compressive strengths are reported in Table 5, where values obtained for cores with different mortar joints arrangement (‘‘H’’, ‘‘H + V1’’ and ‘‘H + V2’’), as well as for detached and non-detached cores, are detailed. Images acquired by DIC, illustrating the development of horizontal strain patterns for two representative cores (of type ‘‘H’’ and ‘‘H + V1’’), are reported in Figs. 9 and 10, respectively for masonry MT1 and MT2. The crack patterns after failure are shown in Fig. 11 for the different core typologies. Failure mechanisms of the cylindrical cores, as shown by DIC images, are strongly consistent with the mechanical properties of the constitutive materials and are similar to the failure modes observed from standard wall compression tests. In the case of MT1, where mortar had a Poisson’s ratio significantly higher than that of bricks (mm = 0.22 vs mb = 0.13), the cracks first initiated inside the bricks, and then propagated into the mortar (Fig. 9). In fact, because of its higher transverse deformability, mortar was horizontally confined by bricks which, in turn, were subject to tensile stresses. As a consequence, bricks were the first elements where cracks opened and propagated. A compatible horizontal

217

stress pattern was also obtained by using a linear-elastic finite element model (FEM) of the core, where the experimentally determined mechanical properties of the materials (Table 2) were introduced. Fig. 12 shows the corresponding horizontal stress patterns for cores of masonry MT1 and MT2 (Fig. 12b and c, respectively). In particular, when considering MT1 (Fig. 12b) and a compression load of about 80% of the failure load, the central part of the horizontal mortar joint is subject to a horizontal compression stress state, while the adjacent bricks are subjected to tensile stress. On the contrary, in case of MT2, where mortar has a Poisson’s ratio lower than that of bricks (mm = 0.15 vs mb = 0.22), bricks tend to expand horizontally and are confined by mortar, the latter being subject to tensile stresses. As a confirmation, first cracks initiated in the mortar and only later they propagated into the bricks (Fig. 10). Mechanical evidences are confirmed by the results of the FEM model reported in Fig. 12c, where it can be observed that the mortar joint is entirely subject to tensile stress, while bricks are partly compressed and partly under traction. In those cores where the vertical mortar joint was also present (left part of Fig. 12b and c), vertical cracks generally opened and developed at the interface between the vertical mortar joint and one of the two brick portions (Figs. 9b and 10b), as reported for actual brick–mortar masonries [14]. With regard to masonry compressive strength, if all the tested cores are considered together, a general overestimation of the actual masonry compressive strength can be observed, amounting

Fig. 11. Photographs of cores, completed with cement mortar castings, after compression test: for MT1, cores representative of the ‘‘H’’ type (a), ‘‘H + V1’’ type (b) and ‘‘H + V2’’ type (c); for MT2, cores representative of the ‘‘H’’ type (d), ‘‘H + V1’’ type (e) and ‘‘H + V2’’ type (f). In the case of MT1, all the ‘‘H’’ type cores were tested using the DIC technique, hence the white paint layer with black ink spots.

218

E. Sassoni et al. / Construction and Building Materials 68 (2014) 206–219

to 21% for MT1 and 23% for MT2 (Table 5). However, important additional information can be derived if different core typologies are examined separately. In the first place, detached cores (i.e., cores which experienced brick–mortar detachment during core-drilling and which therefore were re-assembled by using scotch tape) exhibited, on the average, a not negligible reduction in compressive strength, ranging from 6% to 13% (Table 5). Such a reduction can be explained considering that, even if the cores were carefully re-assembled so that brick and mortar surfaces macroscopically matched, at the micro-scale no perfect match could presumably be achieved. As a consequence, during compression test, stress concentration occurred at the brick–mortar contact points; moreover, the chemical bond between bricks and mortar was completely absent, thus leading the failure to be reached for loading values lower than those in the case of non-detached cores. Therefore, for evaluating the capability of tests on cores to suitably predict masonry compressive strength, results obtained from detached cores were not taken into account. As far as the presence of the vertical mortar joint is concerned, the following observations can be done. In the case of weak mortar (MT1), cores with only the horizontal joint (‘‘H’’ type) resulted sensibly more resistant than cores also having the vertical mortar joint (‘‘H + V1’’ and ‘‘H + V2’’ types), the compressive strength of the latter two types being 69% and 80% of the former type, respectively (Table 5). On the contrary, in the case of stronger mortar (MT2), no substantial difference between strength of cores with and without vertical mortar joint was found experimentally (Table 5). The different resistance of cores with and without the vertical mortar joint, as well as the different behavior of the two masonry typologies, can firstly be ascribed to the differences in mechanical properties of the constituting materials. In the case of MT1, where bricks and mortar have significantly different mechanical properties (Table 2), the presence of the vertical mortar joint is responsible for a difference in the stress state among ‘‘H’’ type cores and ‘‘H + V1’’ and ‘‘H + V2’’ type cores. As can be observed in Fig. 12, the results of the FEM model show that, when the vertical mortar joint is present, tensile stresses arise in the center of the horizontal mortar joint and also in the vertical joint itself. As mortar is sensibly less resistant than bricks, these tensile stresses cause the initiation of cracks at the brick–mortar interface, as observed by DIC (Fig. 9), for relatively low load values. On the contrary, in the case of MT2, where bricks and mortar have similar mechanical characteristics (particularly in terms of compressive strength and modulus of elasticity, Table 2), the presence of the vertical mortar joint makes no substantial difference, with respect to the situation where only the horizontal joint is present. Consistently, in the results of the FEM model, no substantial difference in the horizontal stress state is evident for cores with and without the vertical mortar joint (Fig. 12). However, for a more accurate evaluation of the influence of the vertical mortar joint, further modeling, also including non-linear behavior of the constitutive materials, may be useful. Such work is currently in progress. It should be also noted that the vertical mortar joints had thickness of 25 mm in the case of MT1 and 15 mm in the case of MT2, hence the different size of the vertical mortar joints might be a further reason for the differences between cores with and without the vertical joint, as well as between the different masonry typologies. If only non-detached cores are taken into consideration and all the three core typologies are considered altogether, an overestimation of 27% for MT1 and 34% for MT2 of the masonry compressive strength determined for the reference panels is obtained. In particular, the ‘‘H’’ type cores overestimate the actual masonry strength by 62% for MT1 and 34% for MT2, whereas ‘‘H + V1’’ and ‘‘H + V2’’ cores, considered altogether, lead to an overestimation of 20% for MT1 and 33% for MT2.

Fig. 12. Results of finite element modeling of two representative cores without and with the vertical mortar joint: (a) meshes; horizontal tensile stress pattern for (b) MT1 and (c) MT2. Considered vertical load of about 80% of the failure load.

The reason for these overestimations is presumably to be found in the confinement effect exerted by the horizontal mortar castings above and beneath the cores. In literature studies, where steelmade cradles were used to subject 150 mm diameter cores to compression test, an underestimation of about 40% of the actual masonry compressive strength was found [24]. This was attributed to local phenomena that may affect the collapse mechanism [24]; in addition, stress concentration produced by the imperfect contact between the steel cradle and the core and insufficient confinement (considering the high number of horizontal and vertical mortar layers present in large diameter cores, which could weaken the system) cannot be excluded [8]. On the contrary, in the present study the use of cement mortar to provide the confining effect that is exerted by brickwork in the original masonry texture allowed to

E. Sassoni et al. / Construction and Building Materials 68 (2014) 206–219

achieve a more uniform stress distribution and to prevent stress concentration, which resulted in the cited overestimation. All things considered, the most accurate estimation of masonry compressive strength was obtained, for both masonry typologies, by testing non-detached cores with a vertical mortar joint, either ‘‘H + V1’’ or ‘‘H + V2’’ typologies. In this way, a 20–33% overestimation of the actual strength was found. 4. Conclusions From the results of the present study, aimed at evaluating the suitability of using moderately destructive tests on bricks and mortar samples, separately, and on cylindrical cores to evaluate masonry compressive strength, the following conclusions can be drawn. 1. Evaluation of brick compressive strength can be very sensitive to the testing direction, depending on brick microstructure. Indeed, bricks with an homogeneous microstructure (MT2) showed no substantial difference between the direction parallel and orthogonal to brick’s longest edge, while when macroscopic cracks (MT1) were detected the compressive strength in the parallel direction was half that in the perpendicular direction. As a consequence, when bricks are obtained from the existing masonries by core-drilling, so that cylindrical samples are tested in the direction perpendicular to the actual masonry loading direction, careful interpretation of test results should be performed. 2. Mortar compressive strength obtained from double punch tests on mortar joints was systematically higher than that obtained from standard tests on prisms. Such overestimation was proved to originate, on the one hand, from the different testing technique and, on the other hand, from the different microstructure of mortar cured in the masonry joints and mortar cured in the steel-made moulds used for preparing standard prisms. In fact, curing within bricks strongly enhances the mortar capacity by reducing the open porosity. In terms of actual strength prediction, the correlation formula proposed in the literature by Drdàcky´ et al. [10] proved to give satisfactory results. 3. Comparison between masonry strength directly measured and estimated from brick and mortar strengths showed that for medium quality materials the formula proposed by Bröcker was very effective [6], while for good quality materials the formula by Tassios [6] proved to be more appropriate. 4. Failure mechanisms observed in masonry cores were fairly consistent with those typical of failure of masonry walls, characterized by cracks developing from bricks or mortar joints according to relative values of Poisson’s ratio. 5. Masonry compression strength obtained by testing cylindrical cores was systematically higher (20–35%) than that measured on standard walls. In particular, presence of a vertical mortar joint inside the core reduces the capacity, thus improving the prediction ability. Some variability was observed depending on the level of mechanical properties of components. 6. Cores affected by brick–mortar detachment during core-drilling and then re-assembled exhibited a sensibly reduction in compressive strength, which was attributed to stress concentration at the brick–mortar interface, owing to imperfect match between the two surfaces. 7. The overestimation of strength provided by cores is due to the confining action exerted by mortar castings beneath and above the cores. Numerical simulations and further experimental tests

219

are currently in progress to investigate how mechanical properties of mortar used for castings influence the final resistance of the masonry core.

Acknowledgements The financial support of Emilia-Romagna Region (POR-FESR funds) is gratefully acknowledged. The technicians of the Laboratory of Structural and Geotechnical Engineering (LISG), especially Mr. Davide Betti, are gratefully acknowledged for collaboration. References [1] Mazzotti C, Sassoni E, Pagliai G. Determination of shear strength of historic masonries by moderately destructive testing of masonry cores. Constr Build Mater 2014;54:421–31. [2] D.M. 14/01/2008 (Decreto del Ministero delle Infrastrutture 14 gennaio 2008). Approvazione delle nuove norme tecniche per le costruzioni; 2008 [in Italian]. [3] Eurocode 6. ENV 1996-1-1. Design of masonry structures – Part 1–1: General rules for buildings – rules for reinforced and unreinforced masonry; 1998. [4] Corradi M, Borri A, Vignoli A. Experimental study on the determination of strength of masonry walls. Constr Build Mater 2003;17:325–37. [5] Binda L, Saisi A, Tiraboschi C. Investigation procedures for the diagnosis of historic masonries. Constr Build Mater 2000;14:199–233. [6] Tassios TP. The mechanics of masonry. Athens: Symmetria Publishing; 1987. [7] Benedetti A, Pelà L, Aprile A. Masonry properties determination via splitting tests on cores with a rotated mortar layer. In: Proceedings of 8th international seminar on structural masonry (ISSM 08), Istanbul, 5–7 November 2008; p. 647–55. [8] Sassoni E, Mazzotti C. The use of small diameter cores for assessing the compressive strength of clay brick masonries. J Cult Herit 2013;14S:e95–e101. [9] Cultrone G, Sebastián E, Cazalla O, Nechar M, Romero R, Bagur MG. Ultrasound and mechanical tests combined with ANOVA to evaluate brick quality. Ceram Int 2001;27:401–6. [10] Drdàcky´ M, Makšìn D, Mekonone MD, Slìzˇkovà Z. Compression tests on nonstandard historic mortar specimens. In: Proceedings of HMC08 ‘‘historic mortar conference – characterization, diagnosis, conservation, repair and compatibility’’, Lisbon, 24–26 September 2008. [11] Drdàcky´ M. Non-standard testing of mechanical characteristics of historic mortars. Int J Archit Herit 2011;5:383–94. [12] Pascale G, Bonfiglioli B, Colla C. Tecniche di indagine non distruttiva per la diagnostica strutturale della muratura, alla luce delle recenti normative antisismiche. In: Conferenza nazionale sulle prove non distruttive monitoraggio diagnostica, Biennale PnD-MD, 11° Congresso nazionale dell’AIPnD, Milano; 2005 [in Italian]. [13] Brencich A, Sabia D. Experimental identification of a multi-span masonry bridge: the Tanaro Bridge. Constr Build Mater 2008;22:2087–99. [14] Brencich A, Sterpi E. Compressive strength of solid clay brick masonry: calibration of experimental tests and theoretical issues. In: Lourenço PB, Roca P, Modena C, Agrawal S, editors. Structural analysis of historical constructions, New Delhi; 2006, p. 757–65. [15] EN 998-2. Specification for mortar for masonry – Part 2: Masonry mortar; 2010. [16] EN 1052-1. Methods of test for masonry – determination of compressive strength; 2001. [17] EN 772-1. Methods of test for masonry units – determination of compressive strength; 2002. [18] EN 1015-11. Methods of test for mortar for masonry – Part 11: Determination of flexural and compressive strength of hardened mortar; 2007. [19] Henzel J, Karl S. Determination of strength of mortar in the joints of masonry by compression tests on small specimens. Darmstadt Concr 1987;2:123–36. [20] Sassoni E, Franzoni E, Mazzotti C. Influence of sample thickness and capping on characterization of bedding mortars from historic masonries by double punch test (DPT). Key Eng Mater in press. [21] Mazzotti C, Sassoni E, Bellini A, Ferracuti B, Franzoni E. Strengthening of masonry elements by FRP: influence of brick mechanical and microstructural properties. Key Eng Mater in press. [22] Pelà L, Benedetti A, Marastoni D. Interpretation of experimental tests on small specimens of historical mortars. In: Proceedings of ‘‘8th international conference on structural analysis of historical construction, Wroclaw (Poland) 15–17 October 2012. [23] Sarangapani G, Venkatarama Reddy BV, Jagadish KS. Brick–mortar bond and masonry compressive strength. J Mater Civil Eng 2005;17:229–37. [24] Brencich A, Corradi C, Sterpi E. Experimental approaches to the compressive response of solid clay brick masonry. In: 13th International brick and block masonry conference, Amsterdam, July 4–7, 2004. p. 55–4.