- Email: [email protected]
Rammed earth walls strengthened with polyester fabric strips: Experimental analysis under in-plane cyclic loading
Construction and Building Materials 149 (2017) 29–36
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Rammed earth walls strengthened with polyester fabric strips: Experimental analysis under in-plane cyclic loading Lorenzo Miccoli a,⇑, Urs Müller b, Stanislav Pospíšil c a
Bundesanstalt für Materialforschung und –prüfung (BAM), Division Building Materials, Unter den Eichen 87, 12205 Berlin, Germany CBI Swedish Cement and Concrete Research Institute, c/o SP, Box 857, Brinellgatan 4, 50462 Borås, Sweden c Institute of Theoretical and Applied Mechanics, v.v.i., Prosecká 76, 190 00 Prague, Czech Republic b
h i g h l i g h t s The performance of unstrengthened and strengthened rammed earth walls was analysed. The strengthening technique is based on the use of vertical polyester fabric strips. The use of the strengthening technique requires low-tech equipment and workmanship. The strengthening provides an increase of horizontal load and displacement capacity.
a r t i c l e
i n f o
Article history: Received 17 January 2017 Received in revised form 8 May 2017 Accepted 10 May 2017
Keywords: Rammed earth Pseudo-dynamic loads Shear-compression tests Strengthening Polyester fabric strips
a b s t r a c t This study analyses the mechanical behaviour under pseudo-dynamic loading of structural elements built in rammed earth and strengthened with polyester fabric strips. This strengthening technique was developed to exploit the strength potential of rammed earth and to solve its lack of tensile strength. For this reason, in-plane cyclic tests were carried out to investigate the shear behaviour of unstrengthened and strengthened walls. The strengthening technique requires low-tech equipment and workmanship, uses readily available, not expensive and industrially standardised materials. The experimental results were analysed in terms of stiffness degradation, energy dissipation capacity and equivalent viscous damping. Although the unstrengthened and strengthened walls confirmed a limited ductile behaviour, the findings confirm that the strengthening contributes to limit the spread of the diagonal cracks and provide an increase of strength in terms of horizontal load and displacement capacity. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction Rammed earth is among the oldest building materials and one of the least understood in terms of structural behaviour under dynamic actions. The layered structure due to the compaction process of moistened earth in a wooden formwork has influence on the crack mechanism, even if its behaviour cannot be considered distinctively anisotropic. A considerable number of rammed constructions are present in regions suffering severe seismic events as China, New Zealand, South America and the Mediterranean area. This aspect has highlighted the importance to investigate its mechanical behaviour under dynamic actions and to develop effective strengthening techniques. In spite of this, systematic data on the dynamic performances of rammed earth are still rare and the knowledge about its energy dissipation capability needs to be ⇑ Corresponding author. E-mail address: [email protected] (L. Miccoli). http://dx.doi.org/10.1016/j.conbuildmat.2017.05.115 0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
improved. In the last years, Bui et al. [1] measured the natural frequencies, modal shapes and damping ratios of a rammed earth structure, Wang et al. [2] performed shaking table tests to validate a strengthening system based on externally bonded fibres. A few more studies are available on unstrengthened [3] and strengthened walls [4–6] made of stabilised rammed earth tested under in-plane cyclic loading. Arslan et al. [7] compared the cyclic behaviour of the rammed earth walls non-stabilised and cement stabilised with masonry brick and aerated concrete walls. In comparison with studies carried on the earth block masonry [8,9], the experimental campaigns on strengthening systems for rammed earth are limited to few studies only. Experimental campaigns were carried out on rammed earth samples to assess to effectiveness of strengthening system using externally bonded fibres [10] or using reinforcement systems form textile grids [11] to increase the energy dissipation. Shaking table tests on lab scale models strengthened with boundary wooden elements were performed by Ruiz et al. [12].
30
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
Nomenclature dH,max E Ediss Einp fc fs ft H Hcr Hf Hmax Hmax,i Hu
displacement corresponding to peak horizontal load of the i cycle (mm) Young’s modulus (N/mm2) dissipated hysteretic energy (kNm) input energy (kNm) uniaxial compressive strength (N/mm2) shear strength (N/mm2) tensile strength (N/mm2) horizontal load (kN) horizontal load at crack limit (kN) horizontal load at flexural cracking limit (kN) maximum horizontal load (kN) peak horizontal load of the i cycle (kN) horizontal load at ultimate displacement (kN)
As extension of the previous findings, this study assesses the mechanical behaviour of rammed earth walls unstrengthened (REW) and strengthened with polyester fabric strips (REWS) under in-plane cyclic shear-compression tests. The improvement introduced by the strengthening techniques was evaluated analysing the response of the walls in terms of horizontal load, displacement capacity, stiffness degradation and energy dissipation. The strengthening technique adopted was developed to enhance the in-plane response of the rammed earth walls according to the definition of robustness reported in Eurocode 8 [13] and Tomazˇevicˇ [14]. The performance-based engineering criteria for unreinforced masonry elements were used to associate a particular damage level of the wall to a certain displacement. These criteria let us to idealise the walls behaviour under the progressive increment of the applied lateral displacement through four limit states. The results of the presented study represent an important development of the data partially reported in [15], i.e., assessment of energy dissipation capacity, equivalent viscous damping and stiffness degradation, together with the investigation of the influence of polyester fabric strips creating the strengthening system. 2. Experimental programme 2.1. Materials and samples preparation Five rammed earth walls of size 1300 mm 1050 mm 250 mm were built at Bundesanstalt für Materialforschung und –prüfung (BAM) in Berlin and tested under cyclic shearcompression load. In the manufacturing process, earth with a moisture content in the range of 9–10 mass-% (Fig. 1a) was placed in a plywood formwork (Fig. 1b). The original layer thickness of approximately 150 mm was mechanically compacted with a rammer to a thickness of approximately 100 mm (Fig. 1c). The samples, showing thirteen layers (Fig. 1d), were leave to dry for two months in the laboratory. Before testing the moisture content was in the range of 2–3 mass-% corresponding to an average value of bulk density (q) equal to 2190 kg/m3. No measurement of bulk density was carried out at the time of wall manufacturing. For this reason, it was not possible to assess the variation of bulk density related to the variation of moisture content and compressive strength. Static tests on rammed earth samples (500 mm 500 mm 100 mm) exhibited a compressive strength (fc) of 3.73 N/mm2 and a shear strength (fs) of 0.70 N/mm2. The detailed description of the mechanical characterisation is reported in a previous study [16], as well as the mineralogical and granulometric properties of the rammed earth used [17].
IE,diss Ks,i t
a
d
l qv neq
q r wcr wf wmax wu
indicator of the energy dissipation (-) secant stiffness (kN/mm) time (sec) adhesion strength (N/mm2) displacement (mm) vapour diffusion (–) vertical reinforcement (%) equivalent viscous damping coefficient (–) bulk density (kg/m3) vertical compressive stress (N/mm2) lateral drift at crack limit (%) lateral drift at flexural cracking limit (%) lateral drift at maximum horizontal load (%) lateral drift at ultimate displacement (%)
In this work, a strengthening system is proposed to exploit the strength potential of rammed earth and to solve its lack of tensile strength, significantly increasing strength and displacement capacity. The principle of this technique is to embed the low cost and high tensile strength materials into the rammed earth wall. Vertical strips made of polyester fabric inserted in wall slits to take up horizontal loads over the height of a wall were employed. Polyester fabric strips are textile strips used as heavy duty belts for fixing goods and are available essentially everywhere (Fig. 2a). They have a tensile strength (ft) of 40 N/mm2, a Young’s modulus (E) of 140 N/ mm2 and can be loaded only in the tensile direction. A base coat mortar with an adhesion strength (a) > 0.08 N/mm2 and a water vapour diffusion (m) 20 was chosen [18] as an adhesive for fixing the vertical polyester fabric strips. This adhesive is a premixed polymer-modified cement based mortar with the addition of polymer fibres and water-repellent agents. The main binder components are cement and pozzolan. Base coat mortar was adopted considering its good mechanical performance, workability, fast hardening and its low invasiveness in comparison with other adhesive as epoxy resin or polyurethane foam [20]. The adhesive has a small particles size suitable to fill the thin slits and a compressive strength in the range of 10–20 MPa, about three time higher than the compressive strength shown by rammed earth. Although the strength of the base coat mortar is not mechanically compatible with rammed earth, it must be stressed out that its application is limited to a small portion of the wall (about 0.8–0.9% of the volume). The cement based mortar was chosen in alternative to a high hydraulic mortar thanks to their quick preparation and facility of application in thin layer. Considering the application in very thin layers only the bond with earthen substrate was measured. The adhesion strength was tested with pull-off tests on earth blocks according to EN 1015-12 [19]. The principle of the strip strengthening procedure is shown in Fig. 2b. Slits 56 mm deep and 9 mm wide were cut every 250 mm using a disc grinder. The slits in the walls create a point of weakness that had to be compensated by the combination of the appropriate adhesive and strips. To avoid the creation of too many points of weakness along the wall section the distance between slits was not less than five times the width of the strips. For the same reason the width of the strips was not more than one third of the wall thickness. The slits were cleaned from dust with compressed air and wetted with sprayed water. A water-based polymethylmethacrylate (PMMA) emulsion was sprayed in the slits to develop the bond among the porous substrate and the base coat mortar. Based on the results of a previous study [20] where different adhesives were
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
31
Fig. 1. Manufacturing of rammed earth walls. Filling in a layer of earth into the formwork (a), compacting the layer (b, c) and removing the formwork after all layers were filled in and compacted (d).
Fig. 2. Phases of strengthening. Polyester fabric strips (a), placement of strips on the rammed earth slots (b), sealing the slits with base coat mortar (c), principle of polyester strips strengthening (d).
tested, the application of the polymer primer showed an improvement of adhesion between earthen substrate and the base coat mortar. The base coat mortar was mixed through a mixing tool with addition of water. Afterwards the mortar was applied in the slits to provide a uniform layer of adhesive between the polyester fabric strip and the earthen substrate. Strips of 50 mm width and a thickness of 3 mm were vertically inserted into the slits (Fig. 2c). Afterwards a final layer of 3 mm of adhesive was placed on the slit to cover the strips into the wall surface (Fig. 2d). The percentage of vertical reinforcement (qv) was equal to 0.29. 2.2. Test setup In-plane shear-compression tests were carried out at Institute of Theoretical and Applied Mechanics (ITAM) in Prague, following the procedure given in RILEM TC 76-LUMC3 [21]. Uniform compression and cyclic horizontal loads were applied at the top of the sample through a testing rig. The top end was free to rotate and the base of the walls was fixed, as foreseen by the cantilever-type boundary condition. To avoid potential sliding among the capping beam and the wall, a flanged C-beam was used to wrap the top of the wall. Fig. 3 shows the position of the six linear variable differential transformers (LDVTs) used to measure displacements at the top (P5) and the base (P6) of the wall, as well as flexural and shear
deformations (P1–P4). Four hydraulic actuators, three vertical and one horizontal, were used to apply vertical and lateral loads through a steel capping beam at the top of the sample. An imposed vertical compressive stress (r) of 0.56 N/mm2 was transferred to the wall and kept constant. This stress value corresponds to 15% of the mean value of fc measured by uniaxial compression tests. Even if this compressive stress value is considered quite high for a typical rammed earth construction in service, this value was chosen to obtain a shear-type behaviour excluding rocking and flexural failure mechanisms. This value is comparable to the applied stress levels of similar studies carried out on fired brick masonry walls [22][23]. Each cycle of displacements was repeated three times increasing of 2.5 mm for each step the maximum amplitude. The horizontal actuator speed was in the range of 1–8 mm/s (Fig. 4). The application of horizontal displacements at a frequency of 0.1 Hz was stopped when the horizontal load started to decrease. 3. Experimental results 3.1. Failure modes and load-displacement response The change of the wall behaviour under the progressive rise of applied horizontal displacement was defined according to four limit states [24]. Table 1 lists the values of horizontal loads (H), lateral drift (w), ductility ratios (w/wmax) for the limit states. A qualitative observation of the wall surface after each step of displacement allowed to notice the first cracks, on the other hand deformations were monitored by the sensors P1–P4.
32
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
Fig. 3. Scheme of the cyclic in-plane shear-compression test and the sensors placement (P1–P6).
Fig. 4. Loading pattern of each step of application of the horizontal load.
A X-shaped crack pattern appeared on the unstrengthened walls REW1 and REW3 (Fig. 5a, c) while cracks along one diagonal line characterised the failure of sample REW2 (Fig. 5b). The hysteresis loops, plotting the horizontal load (H) and the corresponding drift (w), were obtained from the cyclic curves for each step of loading (Fig. 6a–c). The drift (w) was calculated as the ratio between the lateral top displacement and the height at which the horizontal load is applied. The hysteresis loops, including the tension and the compression part of the cyclic curves, are not symmetrical due to the effect of non-symmetrical damage on the reverse loading cycles. The shape of the H-w hysteretic loops is typical of the shear failure due a limited sliding of the rammed earth layers. The first flexural cracks opening at the base were associated to the flexural cracking limit (Hf, wf) and occurred at wf = 0.13–0.14%. In correspondence of the first diagonal crack opened was defined the crack limit state (Hcr, wcr) for wcr in the range of 0.31–0.36%. Subsequently, in the third limit state, the load rises gradually until the maximum value of load (Hmax) was reached, in the range of 59.5–77.6 kN associated to wmax = 0.47–0.67%. In this phase a diagonal strut with cracks crossing the
rammed earth layers appeared on the walls surface. The fourth and last limit state was characterised by a limited strength degradation (2–3%) and low displacement capacity (wu = 0.52–0.69%). At this state corresponded to the values of displacements (wu) and loads (Hu) where the samples, before reaching maximum displacement and collapse, are still stable. The strengthened walls REWS1–2 exhibited a crack pattern with one diagonal line and other minor cracks (Fig. 7). The hysteresis loops, obtained from the cyclic curves for each step of loading, are illustrated in Fig. 8a–c. As observed for the unstrengthened samples, the flexural cracking limit in correspondence of the first non-linearity occurred at wf = 0.12–0.13%. The second nonlinearity took place at the crack limit state (Hcr, wcr) when the first diagonal crack opened in correspondence of wcr = 0.88–1.09%. Subsequently the load was increased gradually until the maximum value of load (Hmax) was reached. The results show Hmax in the range of 110.6–116.4 kN with a wmax = 1.21–1.47%. This condition represents the third limit state and is characterised, as for the unstrengthened walls, by the formation of a diagonal strut. Since no post-peak phase was observed, the ultimate limit state (Hu, wu), corresponded to third limit state where samples reached the maximum load and collapsed. The points of maximum load in the hysteresis loop, at each displacement level here reported as lateral drift (w), were used to draw the shear-compression envelope curves (Fig. 9). Considering the low deviation between the three repeated cycles, the envelope curves take into account only the first cycle at each displacement amplitude. In spite of the asymmetric hysteresis behaviour, the shearcompression envelope curves take into account only the maximum values of horizontal load (H) obtained in the quadrant associated with negative horizontal force and negative displacements of the hysteresis loops. The capacity of the walls to deform beyond the elastic limit without significant strength degradation was assessed through the analysis of the ductility ratios (Table 1). The results confirm that rammed earth has a limited ductile behaviour when compared with earthen masonry [25] or fired brick masonry [22,26]. 3.2. Stiffness degradation and energy dissipation The knowledge of the seismic response of a structural element can be enhanced through the assessment of the parameters connected to the mechanism of energy dissipation. These parameters are: degradation of secant stiffness (Ks), indicator of the energy dissipation (IE,diss) and equivalent viscous damping coefficient (neq).
Table 1 Results from cyclic tests at different limit states. Wall REW1 REW2 REW3 REWS1 REWS2
Hf (kN)
wf
Hcr (kN)
wcr
wu/wmax
wu/wcr
wmax/wcr
(%)
Hu (kN)
wu
(%)
Hmax (kN)
wmax
(%)
(%)
(–)
(–)
(–)
34.5 34.9 35.3 36.3 40.7
0.13 0.14 0.13 0.13 0.12
57.4 48.1 65.0 105.0 100.8
0.28 0.30 0.36 1.09 0.88
73.8 59.5 77.6 116.4 110.6
0.67 0.47 0.53 1.47 1.21
73.8 59.1 76.6 116.4 110.6
0.67 0.67 0.64 1.47 1.21
1.00 1.42 1.22 1.00 1.00
2.34 2.20 1.77 1.34 1.37
2.34 1.57 1.47 1.34 1.37
Failure mode Shear Shear Shear Shear Shear
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
33
Fig. 5. Unstrengthened walls. Crack pattern under a combination of vertical compression and cyclic shear: REW1 (a), REW2 (b), REW3 (c) and some details at failure (d).
Fig. 6. Unstrengthened walls. Experimental hysteresis loops: REW1 (a), REW2 (b) and REW3 (c).
The stiffness degradation relates to the damage evolution at each step of displacement during cyclic loading. However, this phenomenon mostly appears during the reversed cycle. The secant stiffness (Ks,i) was calculated according the following equation taking into account the stiffness related to the cycle with the maximum load (1):
K s;i
Hmax;i ¼ dHmax;i
ð1Þ
As observed in Fig. 10, a significant decrease in the values of secant stiffness (Ks) appears as the lateral drift (w) increases. At the first cycles, the unstrengthened samples presented a value of Ks in the range of 18–20 kN/mm with a slightly shifted value for sample REW2, whereas for strengthened samples the value ranged between 21 and 25 kN/mm, showing the main stiffness degradation for w < 0.2%. The strengthened sample REWS2 exhibited an initial value higher than the other samples. This difference can be attributed to the better compaction of the rammed earth layers during the manufacturing process. However, for w > 0.2%, the values of Ks are similar for both samples revealing constant stiffness degradation after that point.
For the assessment of the dissipation capacity the indicator of the energy dissipation (IE,diss) [27] was adopted. This parameter is based on the relationship among the energy dissipated in each loading cycle (Ediss) and the sum of the positive input energy (Eþ inp ) and the negative input energy (Einp ), defined as the energy required to deform the wall up to the imposed displacement:
IE;diss ¼
Ediss Eþinp þ Einp
ð2Þ
The value of Ediss was obtained by calculating the area plotted by the hysteresis loop (Figs. 6 and 8), conversely the value of Einp was calculated as the area under the envelope defined by the hysteresis loop. The variation of the dissipated energy per cycle for all samples together with the evolution of the dissipated energy in terms of lateral drift (w) is presented in Fig. 11. However, this graph shows up that IE,diss is not directly proportional to w. Large dissipation of energy at the maximum displacement corresponds to high energy inside the hysteresis loop. It reasonable to consider that the sudden increase of the dissipated energy noticed in this phase is related with complete openings of diagonal cracks.
34
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
Fig. 7. Strengthened walls. Crack pattern under a combination of vertical compression and cyclic shear: REWS1 (a), REWS2 (b) and some details at failure (c).
Fig. 8. Strengthened walls. Experimental hysteresis loops: REWS1 (a) and REWS2 (b). The value of IE,diss ranges between 63% and 75% for unstrengthened samples and in the range of 62–68% for strengthened samples. The dissipated energy of sample strengthened and unstrengthened is comparable for low values of w (0.2–0.5%), whereas marked differences are present at the beginning and at the end of the tests for w < 0.2% and w > 0.5%. In correspondence of the wall collapse the samples exhibited a pronounced difference in terms of IE,diss, REW1–2 showed a IE,diss of 74%, whereas REWS1–2 revealed a IE,diss of 67%. This finding confirms that the energy dissipation capacity of rammed earth walls was lower than the values reported for fired brick masonry walls [28]. Since that neq is correlated with the dissipation of energy, it was expected to see a trend similar to values shown by Ediss (Fig. 12). The neq defined as the damping capacity per unit angle of the cycle H-w was determined as reported by Magenes and Calvi [26] following the Eq. (3):
neq ¼
Ediss 2pðEþinp þ Einp Þ
ð3Þ
A reduction of neq was observed for lower values of w for both strengthened and unstrengthened, whereas for higher values of w (0.4–1.0%), neq remained almost constant. The values of neq for unstrengthened rammed earth walls ranged between 10% and 12%, where the strengthened samples varied in the range of 10–11% during the loading process, in both cases tended to rise in the load peak phase. These range of values, due to the absence of bed joints in rammed earth, differ considerable with results on earth block masonry walls [25]. In the case of earth block masonry, the cracking and shearing of the joints allow to dissipate more energy and the values of neq ranged between 10% and 30%.
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
Fig. 9. Shear-compression envelope curves.
Fig. 10. Stiffness degradation vs drift.
Fig. 11. Ratio of dissipated/input energy vs drift. 3.3. Discussion of results The H-w hysteresis loops (Figs. 6 and 8) confirm the difference in the load capacity between unstrengthened and strengthened samples with a marked degradation for negative lateral displacements. The comparison of the limit states of envelope curves shows an increase of load capacity in the range of 56–65% revealed by the strengthened walls (Fig. 9). This increment is associated to an improvement of the displacement capacity. After reaching a w of 0.6%, REW1–3 exhibiting shear dominated failure revealed rapid horizontal load degradation, whereas, REWS1–2 presented no significant horizontal load degradation until w = 1.0% together with a large displacement. The strengthened rammed earth walls show no ductility due to the absence of post-peak phase. In case of earthquakes, the improved displacement capacity avoids a sudden collapse of the building and provides to the inhabitants the possibility to escape. In addition, the strengthening technique presents advantages related to their negligible increment of mass.
35
Fig. 12. Equivalent viscous damping vs drift.
The imposed vertical compressive stress (r) has influence not only on the failure mode and the horizontal load capacity, but also on the capacity of energy dissipation, as observed by Haach et al. [30] in a previous work on reinforced concrete block masonry. Comparing the values obtained for the unstrengthened and strengthened samples, it can be noticed that strengthened samples exhibit lower values of dissipated energy than those of the unstrengthened ones, due to the action of the vertical strips in the limitation of diagonal crack spread. This effect was confirmed by the difference of IE,diss values at wu between unstrengthened and strengthened walls, equal to about 8% and related to the less amount of energy dissipated by the strengthened walls. The relation among equivalent viscous damping (neq) and lateral drift (w) is not linear. The initial values of neq were not the lowest values and nearly equal to the final values. With w increasing, neq decreased to the lowest value rapidly and then rose gradually to the end due to the formation of new cracks in the last loading steps. Overall, the curves of variation of neq showed a spoon-shape. This pronounced behaviour at the beginning of the curve was confirmed by the previous studies on earth block masonry carried out by Hracˇov et al. [29] and Wu et al. [25]. The reason can be attributed to the initial residual deformations of rammed earth walls, caused by the imposed vertical compressive stress (r) at the beginning of the testing cycles. The residual deformation is due to the closing of pores in the granular structure of the earthen materials. Earthen solid blocks produced by a mechanised hand moulding procedure with no compression are sensitive to these phenomena where unrecoverable deformation are caused by reduction of the pores size and cause a dissipation of energy [25]. Although the rammed earth layers were mechanically compacted, the stiffness of the inner skeleton is not comparable with that of fired bricks. Therefore, the initial neq was obviously influenced by the imposed vertical compressive stress (r). With w rising, the pores in compressive zone were compacted, and then the capacity of energy dissipation was reduced. As w increased to about 0.2%, neq for strengthened samples decreased to the lowest value of about 10%. Similar minimum values of neq were reached by unstrengthened samples associated to w = 0.4–0.5%. After that, relations among damping ratio and w behaved almost linearly till the end of tests, where the spread of micro cracks was the main cause of energy dissipation. As expected, the trend of neq is directly related to that of energy dissipation capacity (Fig. 11), where the relative differences were similar to those observed for neq.
4. Conclusions A strengthening system for rammed earth walls based on polyester fabric strips was here proposed. To evaluate the efficiency of this system, the cyclic shear-compression tests were carried out. Outcomes of these tests have provided important information about load capacity, maximum displacement capacity and failure mode. Furthermore, other parameters that characterise the behaviour of the structural elements under seismic loads, i.e., stiffness degradation, energy dissipation and viscous damping were evaluated. It was observed that the type of strengthening did not cause significant differences in global mechanical behaviour. Overall, the experimental failure modes of rammed earth walls were similar to those of unreinforced masonry walls. The displacement capacity and the horizontal load capacity of strengthened walls was higher than those of the unstrengthened walls. No direct correlation was
36
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
found between the amount of vertical reinforcement (qv) and the rise of horizontal load (H). The values of equivalent viscous damping (neq) for both typologies of wall, unstrengthened and strengthened, differ considerably with results shown by other earthen materials like earth block masonry [25]. The neq-w diagrams exhibited spoon-shape curves, which were not as those of fired brick masonry and reinforced masonry walls revealing line-shape. The reason was found to be the initial residual deformation caused by the imposed vertical compressive stress (r). The active role of the vertical strips in the limitation of diagonal crack spread was then confirmed by the values of dissipated energy of the strengthened samples, lower in comparison with unstrengthened samples. Despite the energy dissipation capacity of the strengthened walls being inferior, their cumulative energy dissipation capacity is superior. The additional four levels of imposed displacement reached by the strengthen walls led to a cumulative energy dissipation capacity about 4–5 times higher than the values shown by the unstrengthened walls. On the base of the aforementioned findings, the proposed strengthening technique is potentially appropriate, but more experimental data are needed to confirm it as a robust intervention technique and to investigate variations of this strengthening technique to improve the overall performance. The strengthening technique proposed in this study allows using readily available, not expensive and industrially standardised materials or alternatively readying mix hydraulic mortar together with the requirements of low-tech equipment and workmanship. The durability of the mortar used is comparable with the rammed earth substrate. On the other hand, the technique is partially removable. Applicability is strictly related to the possibility of create slits into the wall and to the availability of suitable adhesive which secures for the system rammed earth-textile fabric strip the mutual transmission of shear loads. Plaster can be used to reduce the visual impact of the strengthening intervention on the wall surface. The study here presented based on the use of a base coat mortar can be considered as a prove of concept that need to be optimised. Although this strengthening technique was conceived to be most effective for in-plane actions, the increase of horizontal load capacity provided by the fabric strips can be beneficial also to the out-of-plane resistance of the walls. Further studies on this topic must include the quantification of the bond between the hydraulic mortar and the polyester strips through pull-out tests. The mutual bond is a key point to evaluate the effectiveness of this kind of strengthening technique. Additionally, the optimisation of the ratio of strip width to wall thickness as well as the number of vertical strips inserted per meter of the wall length have to be considered. Shaking table tests will be useful to assess the effectiveness of the whole intervention under dynamic actions. Acknowledgements This research was funded by the European Commission within the framework of the project NIKER dealing with improving immovable cultural heritage assets against the risk of earthquakes (contract No. 244123). The support of Czech Ministry of Education Youth and Sport via project LO1219 is acknowledged. The authors wish to express their gratitude to Mr. André Gardei, Mr. Shota Urushadze and Dr. Stanislav Hracˇov for their important support in the test setup and the samples preparation.
References [1] Q.B. Bui, S. Hans, J.C. Morel, A.P. Do, First exploratory study on dynamic characteristics of rammed earth buildings, Eng. Struct. 33 (2011) 3690–3695. [2] Y. Wang, M. Wang, K. Liu, W. Pan, X. Yang, Shaking table tests on seismic retrofitting of rammed-earth structures, Bull. Earthq. Eng. 15 (2017) 1037– 1055. [3] J. Zhao, X. Men, Q. Shi, J. Liu, Experimental study on seismic performance and shear behavior of rammed-earth wall, World Earthq. Eng. 2 (2006) 005. [4] Z.N. Gao, X.D. Yang, Z. Tao, Z.S. Chen, C.J. Jiao, Experimental study of rammedearth wall with bamboo cane under monotonic horizontal-load, J. Kunming Univ. Sci. Technol. (Sci. Technol.) 2 (2009) 015. [5] R. Walker, H. Morris, Development of new performance based standards for earth building. In: Proc. of the Australasian Structural Engineering Conference, 1998, Auckland, pp. 477–484. [6] H.R. Hamilton III, J. McBride, J. Grill, Cyclic testing of rammed-earth walls containing post-tensioned reinforcement, Earthq. Spectra 22 (2006) 937–959. [7] M.E. Arslan, M. Emirog˘lu, A. Yalama, Structural behavior of rammed earth walls under lateral cyclic loading: a comparative experimental study, Constr. Build. Mater. 133 (2017) 433–442. [8] J. Macabuag, R. Guragain, S. Bhattacharya, Seismic retrofitting of nonengineered masonry in rural Nepal. In: Proc. of the ICE – Structures and Buildings, 2012. [9] S. Bhattacharya, N. Sanket, C.D. Sekhar, A critical review of retrofitting methods for unreinforced masonry structures, Int. J. Disaster Risk Reduct. 7 (2014) 51– 67. [10] K. Liu, M. Wang, Y. Wang, Seismic retrofitting of rural rammed earth buildings using externally bonded fibers, Constr. Build. Mater. 100 (2015) 91–101. [11] E. Bernat-Maso, L. Gil, C. Escrig, Textile-reinforced rammed earth: experimental characterisation of flexural strength and thoughness, Constr. Build. Mater. 106 (2016) 470–479. [12] D. Ruiz, C. López, S. Unigarro, M. Domínguez, Seismic rehabilitation of sixteenth-and seventeenth-century rammed earth-built churches in the Andean highlands: field and laboratory study, J. Perform. Constr. Facil. 29 (2014). 04014144-1–04014144-17. [13] EN 1998-1. Eurocode 8 – Design of structures for earthquake resistance. Part 1: general rules, seismic actions and rules for buildings, In: European committee for standardization, 2004. [14] M. Tomazˇevicˇ, Earthquake-Resistant Design of Masonry Buildings, Imperial College Press, London, 1999. [15] L. Miccoli, A. Drougkas, U. Müller, In-plane behaviour of rammed earth under cyclic loading: experimental testing and finite element modelling, Eng. Struct. 125 (2016) 144–152. [16] L. Miccoli, D.V. Oliveira, R.A. Silva, U. Müller, L. Schueremans, Static behaviour of rammed earth: experimental testing and finite element modelling, Mater. Struct. 48 (2015) 3443–3456. [17] L. Miccoli, U. Müller, P. Fontana, Mechanical behaviour of earthen materials: a comparison between earth block masonry, rammed earth and cob, Constr. Build. Mater. 61 (2014) 327–339. [18] EN 998-1. Specification for Mortar for Masonry. Part 1: Rendering and Plastering Mortar, 2003. [19] EN 1015-12. Methods of Test for Mortar for Masonry – Part 12: Determination of Adhesive Strength of Hardened Rendering and Plastering Mortars on Substrates, 2007. [20] L. Miccoli, P. Fontana, U. Müller, In-plane shear behaviour of earthen materials panels strengthened with polyester fabric strips, in: Proc. of 10th International Conference on Structural Analysis of Historical Constructions, Leuven, 2016, pp. 1099–1105. [21] RILEM TC 76. Tests for Masonry Materials and Structures – LUMC3 – Cyclic Shear Test for Masonry Panels Designed to Resist Seismic Forces; 1991. [22] F. da Porto, F. Mosele, C. Modena, In-plane cyclic behaviour of a new reinforced masonry system: experimental results, Eng. Struct. 33 (2011) 2584–2596. [23] S. Churilov, E. Dumova-Jovanoska, In-plane shear behaviour of unreinforced and jacketed brick masonry walls, Soil Dyn. Earthq. Eng. 50 (2013) 85–105. [24] D.P. Abrams, Performance based engineering concepts for unreinforced masonry building structures, Prog. Struct. Eng. Mater. 3 (2001) 48–56. [25] F. Wu, H.T. Wang, G. Li, J.Q. Jia, H.N. Li, Seismic performance of traditional adobe masonry walls subjected to in-plane cyclic loading, Mater. Struct. 50 (2017) 69. [26] G. Magenes, G.M. Calvi, In-plane seismic response of brick masonry walls, Earthq. Eng. Struct. D 26 (1997) 1091–1112. [27] T. Zimmermann, A. Strauss, Schubtragverhalten von altem unbewehrtem Mauerwerk unter seismischer Belastung, Bautechnik 89 (2012) 553–563. [28] M. Tomazˇevicˇ, M. Lutman, L. Petkovic, Seismic behavior of masonry walls: experimental simulation, J. Struct. Eng. ASCE 122 (1996) 1040–1047. [29] S. Hracˇov, S. Pospíšil, A. Garofano, S. Urushadze, In-plane cyclic behaviour of unfired clay and earth brick walls in both unstrengthened and strengthened conditions, Mater. Struct. (2015) 1–16. [30] V.G. Haach, G. Vasconcelos, P.B. Lourenço, Experimental analysis of reinforced concrete block masonry walls subjected to in-plane cyclic loading, J. Struct. Eng. ASCE 136 (2009) 452–462.
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Rammed earth walls strengthened with polyester fabric strips: Experimental analysis under in-plane cyclic loading Lorenzo Miccoli a,⇑, Urs Müller b, Stanislav Pospíšil c a
Bundesanstalt für Materialforschung und –prüfung (BAM), Division Building Materials, Unter den Eichen 87, 12205 Berlin, Germany CBI Swedish Cement and Concrete Research Institute, c/o SP, Box 857, Brinellgatan 4, 50462 Borås, Sweden c Institute of Theoretical and Applied Mechanics, v.v.i., Prosecká 76, 190 00 Prague, Czech Republic b
h i g h l i g h t s The performance of unstrengthened and strengthened rammed earth walls was analysed. The strengthening technique is based on the use of vertical polyester fabric strips. The use of the strengthening technique requires low-tech equipment and workmanship. The strengthening provides an increase of horizontal load and displacement capacity.
a r t i c l e
i n f o
Article history: Received 17 January 2017 Received in revised form 8 May 2017 Accepted 10 May 2017
Keywords: Rammed earth Pseudo-dynamic loads Shear-compression tests Strengthening Polyester fabric strips
a b s t r a c t This study analyses the mechanical behaviour under pseudo-dynamic loading of structural elements built in rammed earth and strengthened with polyester fabric strips. This strengthening technique was developed to exploit the strength potential of rammed earth and to solve its lack of tensile strength. For this reason, in-plane cyclic tests were carried out to investigate the shear behaviour of unstrengthened and strengthened walls. The strengthening technique requires low-tech equipment and workmanship, uses readily available, not expensive and industrially standardised materials. The experimental results were analysed in terms of stiffness degradation, energy dissipation capacity and equivalent viscous damping. Although the unstrengthened and strengthened walls confirmed a limited ductile behaviour, the findings confirm that the strengthening contributes to limit the spread of the diagonal cracks and provide an increase of strength in terms of horizontal load and displacement capacity. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction Rammed earth is among the oldest building materials and one of the least understood in terms of structural behaviour under dynamic actions. The layered structure due to the compaction process of moistened earth in a wooden formwork has influence on the crack mechanism, even if its behaviour cannot be considered distinctively anisotropic. A considerable number of rammed constructions are present in regions suffering severe seismic events as China, New Zealand, South America and the Mediterranean area. This aspect has highlighted the importance to investigate its mechanical behaviour under dynamic actions and to develop effective strengthening techniques. In spite of this, systematic data on the dynamic performances of rammed earth are still rare and the knowledge about its energy dissipation capability needs to be ⇑ Corresponding author. E-mail address: [email protected] (L. Miccoli). http://dx.doi.org/10.1016/j.conbuildmat.2017.05.115 0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
improved. In the last years, Bui et al. [1] measured the natural frequencies, modal shapes and damping ratios of a rammed earth structure, Wang et al. [2] performed shaking table tests to validate a strengthening system based on externally bonded fibres. A few more studies are available on unstrengthened [3] and strengthened walls [4–6] made of stabilised rammed earth tested under in-plane cyclic loading. Arslan et al. [7] compared the cyclic behaviour of the rammed earth walls non-stabilised and cement stabilised with masonry brick and aerated concrete walls. In comparison with studies carried on the earth block masonry [8,9], the experimental campaigns on strengthening systems for rammed earth are limited to few studies only. Experimental campaigns were carried out on rammed earth samples to assess to effectiveness of strengthening system using externally bonded fibres [10] or using reinforcement systems form textile grids [11] to increase the energy dissipation. Shaking table tests on lab scale models strengthened with boundary wooden elements were performed by Ruiz et al. [12].
30
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
Nomenclature dH,max E Ediss Einp fc fs ft H Hcr Hf Hmax Hmax,i Hu
displacement corresponding to peak horizontal load of the i cycle (mm) Young’s modulus (N/mm2) dissipated hysteretic energy (kNm) input energy (kNm) uniaxial compressive strength (N/mm2) shear strength (N/mm2) tensile strength (N/mm2) horizontal load (kN) horizontal load at crack limit (kN) horizontal load at flexural cracking limit (kN) maximum horizontal load (kN) peak horizontal load of the i cycle (kN) horizontal load at ultimate displacement (kN)
As extension of the previous findings, this study assesses the mechanical behaviour of rammed earth walls unstrengthened (REW) and strengthened with polyester fabric strips (REWS) under in-plane cyclic shear-compression tests. The improvement introduced by the strengthening techniques was evaluated analysing the response of the walls in terms of horizontal load, displacement capacity, stiffness degradation and energy dissipation. The strengthening technique adopted was developed to enhance the in-plane response of the rammed earth walls according to the definition of robustness reported in Eurocode 8 [13] and Tomazˇevicˇ [14]. The performance-based engineering criteria for unreinforced masonry elements were used to associate a particular damage level of the wall to a certain displacement. These criteria let us to idealise the walls behaviour under the progressive increment of the applied lateral displacement through four limit states. The results of the presented study represent an important development of the data partially reported in [15], i.e., assessment of energy dissipation capacity, equivalent viscous damping and stiffness degradation, together with the investigation of the influence of polyester fabric strips creating the strengthening system. 2. Experimental programme 2.1. Materials and samples preparation Five rammed earth walls of size 1300 mm 1050 mm 250 mm were built at Bundesanstalt für Materialforschung und –prüfung (BAM) in Berlin and tested under cyclic shearcompression load. In the manufacturing process, earth with a moisture content in the range of 9–10 mass-% (Fig. 1a) was placed in a plywood formwork (Fig. 1b). The original layer thickness of approximately 150 mm was mechanically compacted with a rammer to a thickness of approximately 100 mm (Fig. 1c). The samples, showing thirteen layers (Fig. 1d), were leave to dry for two months in the laboratory. Before testing the moisture content was in the range of 2–3 mass-% corresponding to an average value of bulk density (q) equal to 2190 kg/m3. No measurement of bulk density was carried out at the time of wall manufacturing. For this reason, it was not possible to assess the variation of bulk density related to the variation of moisture content and compressive strength. Static tests on rammed earth samples (500 mm 500 mm 100 mm) exhibited a compressive strength (fc) of 3.73 N/mm2 and a shear strength (fs) of 0.70 N/mm2. The detailed description of the mechanical characterisation is reported in a previous study [16], as well as the mineralogical and granulometric properties of the rammed earth used [17].
IE,diss Ks,i t
a
d
l qv neq
q r wcr wf wmax wu
indicator of the energy dissipation (-) secant stiffness (kN/mm) time (sec) adhesion strength (N/mm2) displacement (mm) vapour diffusion (–) vertical reinforcement (%) equivalent viscous damping coefficient (–) bulk density (kg/m3) vertical compressive stress (N/mm2) lateral drift at crack limit (%) lateral drift at flexural cracking limit (%) lateral drift at maximum horizontal load (%) lateral drift at ultimate displacement (%)
In this work, a strengthening system is proposed to exploit the strength potential of rammed earth and to solve its lack of tensile strength, significantly increasing strength and displacement capacity. The principle of this technique is to embed the low cost and high tensile strength materials into the rammed earth wall. Vertical strips made of polyester fabric inserted in wall slits to take up horizontal loads over the height of a wall were employed. Polyester fabric strips are textile strips used as heavy duty belts for fixing goods and are available essentially everywhere (Fig. 2a). They have a tensile strength (ft) of 40 N/mm2, a Young’s modulus (E) of 140 N/ mm2 and can be loaded only in the tensile direction. A base coat mortar with an adhesion strength (a) > 0.08 N/mm2 and a water vapour diffusion (m) 20 was chosen [18] as an adhesive for fixing the vertical polyester fabric strips. This adhesive is a premixed polymer-modified cement based mortar with the addition of polymer fibres and water-repellent agents. The main binder components are cement and pozzolan. Base coat mortar was adopted considering its good mechanical performance, workability, fast hardening and its low invasiveness in comparison with other adhesive as epoxy resin or polyurethane foam [20]. The adhesive has a small particles size suitable to fill the thin slits and a compressive strength in the range of 10–20 MPa, about three time higher than the compressive strength shown by rammed earth. Although the strength of the base coat mortar is not mechanically compatible with rammed earth, it must be stressed out that its application is limited to a small portion of the wall (about 0.8–0.9% of the volume). The cement based mortar was chosen in alternative to a high hydraulic mortar thanks to their quick preparation and facility of application in thin layer. Considering the application in very thin layers only the bond with earthen substrate was measured. The adhesion strength was tested with pull-off tests on earth blocks according to EN 1015-12 [19]. The principle of the strip strengthening procedure is shown in Fig. 2b. Slits 56 mm deep and 9 mm wide were cut every 250 mm using a disc grinder. The slits in the walls create a point of weakness that had to be compensated by the combination of the appropriate adhesive and strips. To avoid the creation of too many points of weakness along the wall section the distance between slits was not less than five times the width of the strips. For the same reason the width of the strips was not more than one third of the wall thickness. The slits were cleaned from dust with compressed air and wetted with sprayed water. A water-based polymethylmethacrylate (PMMA) emulsion was sprayed in the slits to develop the bond among the porous substrate and the base coat mortar. Based on the results of a previous study [20] where different adhesives were
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
31
Fig. 1. Manufacturing of rammed earth walls. Filling in a layer of earth into the formwork (a), compacting the layer (b, c) and removing the formwork after all layers were filled in and compacted (d).
Fig. 2. Phases of strengthening. Polyester fabric strips (a), placement of strips on the rammed earth slots (b), sealing the slits with base coat mortar (c), principle of polyester strips strengthening (d).
tested, the application of the polymer primer showed an improvement of adhesion between earthen substrate and the base coat mortar. The base coat mortar was mixed through a mixing tool with addition of water. Afterwards the mortar was applied in the slits to provide a uniform layer of adhesive between the polyester fabric strip and the earthen substrate. Strips of 50 mm width and a thickness of 3 mm were vertically inserted into the slits (Fig. 2c). Afterwards a final layer of 3 mm of adhesive was placed on the slit to cover the strips into the wall surface (Fig. 2d). The percentage of vertical reinforcement (qv) was equal to 0.29. 2.2. Test setup In-plane shear-compression tests were carried out at Institute of Theoretical and Applied Mechanics (ITAM) in Prague, following the procedure given in RILEM TC 76-LUMC3 [21]. Uniform compression and cyclic horizontal loads were applied at the top of the sample through a testing rig. The top end was free to rotate and the base of the walls was fixed, as foreseen by the cantilever-type boundary condition. To avoid potential sliding among the capping beam and the wall, a flanged C-beam was used to wrap the top of the wall. Fig. 3 shows the position of the six linear variable differential transformers (LDVTs) used to measure displacements at the top (P5) and the base (P6) of the wall, as well as flexural and shear
deformations (P1–P4). Four hydraulic actuators, three vertical and one horizontal, were used to apply vertical and lateral loads through a steel capping beam at the top of the sample. An imposed vertical compressive stress (r) of 0.56 N/mm2 was transferred to the wall and kept constant. This stress value corresponds to 15% of the mean value of fc measured by uniaxial compression tests. Even if this compressive stress value is considered quite high for a typical rammed earth construction in service, this value was chosen to obtain a shear-type behaviour excluding rocking and flexural failure mechanisms. This value is comparable to the applied stress levels of similar studies carried out on fired brick masonry walls [22][23]. Each cycle of displacements was repeated three times increasing of 2.5 mm for each step the maximum amplitude. The horizontal actuator speed was in the range of 1–8 mm/s (Fig. 4). The application of horizontal displacements at a frequency of 0.1 Hz was stopped when the horizontal load started to decrease. 3. Experimental results 3.1. Failure modes and load-displacement response The change of the wall behaviour under the progressive rise of applied horizontal displacement was defined according to four limit states [24]. Table 1 lists the values of horizontal loads (H), lateral drift (w), ductility ratios (w/wmax) for the limit states. A qualitative observation of the wall surface after each step of displacement allowed to notice the first cracks, on the other hand deformations were monitored by the sensors P1–P4.
32
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
Fig. 3. Scheme of the cyclic in-plane shear-compression test and the sensors placement (P1–P6).
Fig. 4. Loading pattern of each step of application of the horizontal load.
A X-shaped crack pattern appeared on the unstrengthened walls REW1 and REW3 (Fig. 5a, c) while cracks along one diagonal line characterised the failure of sample REW2 (Fig. 5b). The hysteresis loops, plotting the horizontal load (H) and the corresponding drift (w), were obtained from the cyclic curves for each step of loading (Fig. 6a–c). The drift (w) was calculated as the ratio between the lateral top displacement and the height at which the horizontal load is applied. The hysteresis loops, including the tension and the compression part of the cyclic curves, are not symmetrical due to the effect of non-symmetrical damage on the reverse loading cycles. The shape of the H-w hysteretic loops is typical of the shear failure due a limited sliding of the rammed earth layers. The first flexural cracks opening at the base were associated to the flexural cracking limit (Hf, wf) and occurred at wf = 0.13–0.14%. In correspondence of the first diagonal crack opened was defined the crack limit state (Hcr, wcr) for wcr in the range of 0.31–0.36%. Subsequently, in the third limit state, the load rises gradually until the maximum value of load (Hmax) was reached, in the range of 59.5–77.6 kN associated to wmax = 0.47–0.67%. In this phase a diagonal strut with cracks crossing the
rammed earth layers appeared on the walls surface. The fourth and last limit state was characterised by a limited strength degradation (2–3%) and low displacement capacity (wu = 0.52–0.69%). At this state corresponded to the values of displacements (wu) and loads (Hu) where the samples, before reaching maximum displacement and collapse, are still stable. The strengthened walls REWS1–2 exhibited a crack pattern with one diagonal line and other minor cracks (Fig. 7). The hysteresis loops, obtained from the cyclic curves for each step of loading, are illustrated in Fig. 8a–c. As observed for the unstrengthened samples, the flexural cracking limit in correspondence of the first non-linearity occurred at wf = 0.12–0.13%. The second nonlinearity took place at the crack limit state (Hcr, wcr) when the first diagonal crack opened in correspondence of wcr = 0.88–1.09%. Subsequently the load was increased gradually until the maximum value of load (Hmax) was reached. The results show Hmax in the range of 110.6–116.4 kN with a wmax = 1.21–1.47%. This condition represents the third limit state and is characterised, as for the unstrengthened walls, by the formation of a diagonal strut. Since no post-peak phase was observed, the ultimate limit state (Hu, wu), corresponded to third limit state where samples reached the maximum load and collapsed. The points of maximum load in the hysteresis loop, at each displacement level here reported as lateral drift (w), were used to draw the shear-compression envelope curves (Fig. 9). Considering the low deviation between the three repeated cycles, the envelope curves take into account only the first cycle at each displacement amplitude. In spite of the asymmetric hysteresis behaviour, the shearcompression envelope curves take into account only the maximum values of horizontal load (H) obtained in the quadrant associated with negative horizontal force and negative displacements of the hysteresis loops. The capacity of the walls to deform beyond the elastic limit without significant strength degradation was assessed through the analysis of the ductility ratios (Table 1). The results confirm that rammed earth has a limited ductile behaviour when compared with earthen masonry [25] or fired brick masonry [22,26]. 3.2. Stiffness degradation and energy dissipation The knowledge of the seismic response of a structural element can be enhanced through the assessment of the parameters connected to the mechanism of energy dissipation. These parameters are: degradation of secant stiffness (Ks), indicator of the energy dissipation (IE,diss) and equivalent viscous damping coefficient (neq).
Table 1 Results from cyclic tests at different limit states. Wall REW1 REW2 REW3 REWS1 REWS2
Hf (kN)
wf
Hcr (kN)
wcr
wu/wmax
wu/wcr
wmax/wcr
(%)
Hu (kN)
wu
(%)
Hmax (kN)
wmax
(%)
(%)
(–)
(–)
(–)
34.5 34.9 35.3 36.3 40.7
0.13 0.14 0.13 0.13 0.12
57.4 48.1 65.0 105.0 100.8
0.28 0.30 0.36 1.09 0.88
73.8 59.5 77.6 116.4 110.6
0.67 0.47 0.53 1.47 1.21
73.8 59.1 76.6 116.4 110.6
0.67 0.67 0.64 1.47 1.21
1.00 1.42 1.22 1.00 1.00
2.34 2.20 1.77 1.34 1.37
2.34 1.57 1.47 1.34 1.37
Failure mode Shear Shear Shear Shear Shear
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
33
Fig. 5. Unstrengthened walls. Crack pattern under a combination of vertical compression and cyclic shear: REW1 (a), REW2 (b), REW3 (c) and some details at failure (d).
Fig. 6. Unstrengthened walls. Experimental hysteresis loops: REW1 (a), REW2 (b) and REW3 (c).
The stiffness degradation relates to the damage evolution at each step of displacement during cyclic loading. However, this phenomenon mostly appears during the reversed cycle. The secant stiffness (Ks,i) was calculated according the following equation taking into account the stiffness related to the cycle with the maximum load (1):
K s;i
Hmax;i ¼ dHmax;i
ð1Þ
As observed in Fig. 10, a significant decrease in the values of secant stiffness (Ks) appears as the lateral drift (w) increases. At the first cycles, the unstrengthened samples presented a value of Ks in the range of 18–20 kN/mm with a slightly shifted value for sample REW2, whereas for strengthened samples the value ranged between 21 and 25 kN/mm, showing the main stiffness degradation for w < 0.2%. The strengthened sample REWS2 exhibited an initial value higher than the other samples. This difference can be attributed to the better compaction of the rammed earth layers during the manufacturing process. However, for w > 0.2%, the values of Ks are similar for both samples revealing constant stiffness degradation after that point.
For the assessment of the dissipation capacity the indicator of the energy dissipation (IE,diss) [27] was adopted. This parameter is based on the relationship among the energy dissipated in each loading cycle (Ediss) and the sum of the positive input energy (Eþ inp ) and the negative input energy (Einp ), defined as the energy required to deform the wall up to the imposed displacement:
IE;diss ¼
Ediss Eþinp þ Einp
ð2Þ
The value of Ediss was obtained by calculating the area plotted by the hysteresis loop (Figs. 6 and 8), conversely the value of Einp was calculated as the area under the envelope defined by the hysteresis loop. The variation of the dissipated energy per cycle for all samples together with the evolution of the dissipated energy in terms of lateral drift (w) is presented in Fig. 11. However, this graph shows up that IE,diss is not directly proportional to w. Large dissipation of energy at the maximum displacement corresponds to high energy inside the hysteresis loop. It reasonable to consider that the sudden increase of the dissipated energy noticed in this phase is related with complete openings of diagonal cracks.
34
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
Fig. 7. Strengthened walls. Crack pattern under a combination of vertical compression and cyclic shear: REWS1 (a), REWS2 (b) and some details at failure (c).
Fig. 8. Strengthened walls. Experimental hysteresis loops: REWS1 (a) and REWS2 (b). The value of IE,diss ranges between 63% and 75% for unstrengthened samples and in the range of 62–68% for strengthened samples. The dissipated energy of sample strengthened and unstrengthened is comparable for low values of w (0.2–0.5%), whereas marked differences are present at the beginning and at the end of the tests for w < 0.2% and w > 0.5%. In correspondence of the wall collapse the samples exhibited a pronounced difference in terms of IE,diss, REW1–2 showed a IE,diss of 74%, whereas REWS1–2 revealed a IE,diss of 67%. This finding confirms that the energy dissipation capacity of rammed earth walls was lower than the values reported for fired brick masonry walls [28]. Since that neq is correlated with the dissipation of energy, it was expected to see a trend similar to values shown by Ediss (Fig. 12). The neq defined as the damping capacity per unit angle of the cycle H-w was determined as reported by Magenes and Calvi [26] following the Eq. (3):
neq ¼
Ediss 2pðEþinp þ Einp Þ
ð3Þ
A reduction of neq was observed for lower values of w for both strengthened and unstrengthened, whereas for higher values of w (0.4–1.0%), neq remained almost constant. The values of neq for unstrengthened rammed earth walls ranged between 10% and 12%, where the strengthened samples varied in the range of 10–11% during the loading process, in both cases tended to rise in the load peak phase. These range of values, due to the absence of bed joints in rammed earth, differ considerable with results on earth block masonry walls [25]. In the case of earth block masonry, the cracking and shearing of the joints allow to dissipate more energy and the values of neq ranged between 10% and 30%.
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
Fig. 9. Shear-compression envelope curves.
Fig. 10. Stiffness degradation vs drift.
Fig. 11. Ratio of dissipated/input energy vs drift. 3.3. Discussion of results The H-w hysteresis loops (Figs. 6 and 8) confirm the difference in the load capacity between unstrengthened and strengthened samples with a marked degradation for negative lateral displacements. The comparison of the limit states of envelope curves shows an increase of load capacity in the range of 56–65% revealed by the strengthened walls (Fig. 9). This increment is associated to an improvement of the displacement capacity. After reaching a w of 0.6%, REW1–3 exhibiting shear dominated failure revealed rapid horizontal load degradation, whereas, REWS1–2 presented no significant horizontal load degradation until w = 1.0% together with a large displacement. The strengthened rammed earth walls show no ductility due to the absence of post-peak phase. In case of earthquakes, the improved displacement capacity avoids a sudden collapse of the building and provides to the inhabitants the possibility to escape. In addition, the strengthening technique presents advantages related to their negligible increment of mass.
35
Fig. 12. Equivalent viscous damping vs drift.
The imposed vertical compressive stress (r) has influence not only on the failure mode and the horizontal load capacity, but also on the capacity of energy dissipation, as observed by Haach et al. [30] in a previous work on reinforced concrete block masonry. Comparing the values obtained for the unstrengthened and strengthened samples, it can be noticed that strengthened samples exhibit lower values of dissipated energy than those of the unstrengthened ones, due to the action of the vertical strips in the limitation of diagonal crack spread. This effect was confirmed by the difference of IE,diss values at wu between unstrengthened and strengthened walls, equal to about 8% and related to the less amount of energy dissipated by the strengthened walls. The relation among equivalent viscous damping (neq) and lateral drift (w) is not linear. The initial values of neq were not the lowest values and nearly equal to the final values. With w increasing, neq decreased to the lowest value rapidly and then rose gradually to the end due to the formation of new cracks in the last loading steps. Overall, the curves of variation of neq showed a spoon-shape. This pronounced behaviour at the beginning of the curve was confirmed by the previous studies on earth block masonry carried out by Hracˇov et al. [29] and Wu et al. [25]. The reason can be attributed to the initial residual deformations of rammed earth walls, caused by the imposed vertical compressive stress (r) at the beginning of the testing cycles. The residual deformation is due to the closing of pores in the granular structure of the earthen materials. Earthen solid blocks produced by a mechanised hand moulding procedure with no compression are sensitive to these phenomena where unrecoverable deformation are caused by reduction of the pores size and cause a dissipation of energy [25]. Although the rammed earth layers were mechanically compacted, the stiffness of the inner skeleton is not comparable with that of fired bricks. Therefore, the initial neq was obviously influenced by the imposed vertical compressive stress (r). With w rising, the pores in compressive zone were compacted, and then the capacity of energy dissipation was reduced. As w increased to about 0.2%, neq for strengthened samples decreased to the lowest value of about 10%. Similar minimum values of neq were reached by unstrengthened samples associated to w = 0.4–0.5%. After that, relations among damping ratio and w behaved almost linearly till the end of tests, where the spread of micro cracks was the main cause of energy dissipation. As expected, the trend of neq is directly related to that of energy dissipation capacity (Fig. 11), where the relative differences were similar to those observed for neq.
4. Conclusions A strengthening system for rammed earth walls based on polyester fabric strips was here proposed. To evaluate the efficiency of this system, the cyclic shear-compression tests were carried out. Outcomes of these tests have provided important information about load capacity, maximum displacement capacity and failure mode. Furthermore, other parameters that characterise the behaviour of the structural elements under seismic loads, i.e., stiffness degradation, energy dissipation and viscous damping were evaluated. It was observed that the type of strengthening did not cause significant differences in global mechanical behaviour. Overall, the experimental failure modes of rammed earth walls were similar to those of unreinforced masonry walls. The displacement capacity and the horizontal load capacity of strengthened walls was higher than those of the unstrengthened walls. No direct correlation was
36
L. Miccoli et al. / Construction and Building Materials 149 (2017) 29–36
found between the amount of vertical reinforcement (qv) and the rise of horizontal load (H). The values of equivalent viscous damping (neq) for both typologies of wall, unstrengthened and strengthened, differ considerably with results shown by other earthen materials like earth block masonry [25]. The neq-w diagrams exhibited spoon-shape curves, which were not as those of fired brick masonry and reinforced masonry walls revealing line-shape. The reason was found to be the initial residual deformation caused by the imposed vertical compressive stress (r). The active role of the vertical strips in the limitation of diagonal crack spread was then confirmed by the values of dissipated energy of the strengthened samples, lower in comparison with unstrengthened samples. Despite the energy dissipation capacity of the strengthened walls being inferior, their cumulative energy dissipation capacity is superior. The additional four levels of imposed displacement reached by the strengthen walls led to a cumulative energy dissipation capacity about 4–5 times higher than the values shown by the unstrengthened walls. On the base of the aforementioned findings, the proposed strengthening technique is potentially appropriate, but more experimental data are needed to confirm it as a robust intervention technique and to investigate variations of this strengthening technique to improve the overall performance. The strengthening technique proposed in this study allows using readily available, not expensive and industrially standardised materials or alternatively readying mix hydraulic mortar together with the requirements of low-tech equipment and workmanship. The durability of the mortar used is comparable with the rammed earth substrate. On the other hand, the technique is partially removable. Applicability is strictly related to the possibility of create slits into the wall and to the availability of suitable adhesive which secures for the system rammed earth-textile fabric strip the mutual transmission of shear loads. Plaster can be used to reduce the visual impact of the strengthening intervention on the wall surface. The study here presented based on the use of a base coat mortar can be considered as a prove of concept that need to be optimised. Although this strengthening technique was conceived to be most effective for in-plane actions, the increase of horizontal load capacity provided by the fabric strips can be beneficial also to the out-of-plane resistance of the walls. Further studies on this topic must include the quantification of the bond between the hydraulic mortar and the polyester strips through pull-out tests. The mutual bond is a key point to evaluate the effectiveness of this kind of strengthening technique. Additionally, the optimisation of the ratio of strip width to wall thickness as well as the number of vertical strips inserted per meter of the wall length have to be considered. Shaking table tests will be useful to assess the effectiveness of the whole intervention under dynamic actions. Acknowledgements This research was funded by the European Commission within the framework of the project NIKER dealing with improving immovable cultural heritage assets against the risk of earthquakes (contract No. 244123). The support of Czech Ministry of Education Youth and Sport via project LO1219 is acknowledged. The authors wish to express their gratitude to Mr. André Gardei, Mr. Shota Urushadze and Dr. Stanislav Hracˇov for their important support in the test setup and the samples preparation.
References [1] Q.B. Bui, S. Hans, J.C. Morel, A.P. Do, First exploratory study on dynamic characteristics of rammed earth buildings, Eng. Struct. 33 (2011) 3690–3695. [2] Y. Wang, M. Wang, K. Liu, W. Pan, X. Yang, Shaking table tests on seismic retrofitting of rammed-earth structures, Bull. Earthq. Eng. 15 (2017) 1037– 1055. [3] J. Zhao, X. Men, Q. Shi, J. Liu, Experimental study on seismic performance and shear behavior of rammed-earth wall, World Earthq. Eng. 2 (2006) 005. [4] Z.N. Gao, X.D. Yang, Z. Tao, Z.S. Chen, C.J. Jiao, Experimental study of rammedearth wall with bamboo cane under monotonic horizontal-load, J. Kunming Univ. Sci. Technol. (Sci. Technol.) 2 (2009) 015. [5] R. Walker, H. Morris, Development of new performance based standards for earth building. In: Proc. of the Australasian Structural Engineering Conference, 1998, Auckland, pp. 477–484. [6] H.R. Hamilton III, J. McBride, J. Grill, Cyclic testing of rammed-earth walls containing post-tensioned reinforcement, Earthq. Spectra 22 (2006) 937–959. [7] M.E. Arslan, M. Emirog˘lu, A. Yalama, Structural behavior of rammed earth walls under lateral cyclic loading: a comparative experimental study, Constr. Build. Mater. 133 (2017) 433–442. [8] J. Macabuag, R. Guragain, S. Bhattacharya, Seismic retrofitting of nonengineered masonry in rural Nepal. In: Proc. of the ICE – Structures and Buildings, 2012. [9] S. Bhattacharya, N. Sanket, C.D. Sekhar, A critical review of retrofitting methods for unreinforced masonry structures, Int. J. Disaster Risk Reduct. 7 (2014) 51– 67. [10] K. Liu, M. Wang, Y. Wang, Seismic retrofitting of rural rammed earth buildings using externally bonded fibers, Constr. Build. Mater. 100 (2015) 91–101. [11] E. Bernat-Maso, L. Gil, C. Escrig, Textile-reinforced rammed earth: experimental characterisation of flexural strength and thoughness, Constr. Build. Mater. 106 (2016) 470–479. [12] D. Ruiz, C. López, S. Unigarro, M. Domínguez, Seismic rehabilitation of sixteenth-and seventeenth-century rammed earth-built churches in the Andean highlands: field and laboratory study, J. Perform. Constr. Facil. 29 (2014). 04014144-1–04014144-17. [13] EN 1998-1. Eurocode 8 – Design of structures for earthquake resistance. Part 1: general rules, seismic actions and rules for buildings, In: European committee for standardization, 2004. [14] M. Tomazˇevicˇ, Earthquake-Resistant Design of Masonry Buildings, Imperial College Press, London, 1999. [15] L. Miccoli, A. Drougkas, U. Müller, In-plane behaviour of rammed earth under cyclic loading: experimental testing and finite element modelling, Eng. Struct. 125 (2016) 144–152. [16] L. Miccoli, D.V. Oliveira, R.A. Silva, U. Müller, L. Schueremans, Static behaviour of rammed earth: experimental testing and finite element modelling, Mater. Struct. 48 (2015) 3443–3456. [17] L. Miccoli, U. Müller, P. Fontana, Mechanical behaviour of earthen materials: a comparison between earth block masonry, rammed earth and cob, Constr. Build. Mater. 61 (2014) 327–339. [18] EN 998-1. Specification for Mortar for Masonry. Part 1: Rendering and Plastering Mortar, 2003. [19] EN 1015-12. Methods of Test for Mortar for Masonry – Part 12: Determination of Adhesive Strength of Hardened Rendering and Plastering Mortars on Substrates, 2007. [20] L. Miccoli, P. Fontana, U. Müller, In-plane shear behaviour of earthen materials panels strengthened with polyester fabric strips, in: Proc. of 10th International Conference on Structural Analysis of Historical Constructions, Leuven, 2016, pp. 1099–1105. [21] RILEM TC 76. Tests for Masonry Materials and Structures – LUMC3 – Cyclic Shear Test for Masonry Panels Designed to Resist Seismic Forces; 1991. [22] F. da Porto, F. Mosele, C. Modena, In-plane cyclic behaviour of a new reinforced masonry system: experimental results, Eng. Struct. 33 (2011) 2584–2596. [23] S. Churilov, E. Dumova-Jovanoska, In-plane shear behaviour of unreinforced and jacketed brick masonry walls, Soil Dyn. Earthq. Eng. 50 (2013) 85–105. [24] D.P. Abrams, Performance based engineering concepts for unreinforced masonry building structures, Prog. Struct. Eng. Mater. 3 (2001) 48–56. [25] F. Wu, H.T. Wang, G. Li, J.Q. Jia, H.N. Li, Seismic performance of traditional adobe masonry walls subjected to in-plane cyclic loading, Mater. Struct. 50 (2017) 69. [26] G. Magenes, G.M. Calvi, In-plane seismic response of brick masonry walls, Earthq. Eng. Struct. D 26 (1997) 1091–1112. [27] T. Zimmermann, A. Strauss, Schubtragverhalten von altem unbewehrtem Mauerwerk unter seismischer Belastung, Bautechnik 89 (2012) 553–563. [28] M. Tomazˇevicˇ, M. Lutman, L. Petkovic, Seismic behavior of masonry walls: experimental simulation, J. Struct. Eng. ASCE 122 (1996) 1040–1047. [29] S. Hracˇov, S. Pospíšil, A. Garofano, S. Urushadze, In-plane cyclic behaviour of unfired clay and earth brick walls in both unstrengthened and strengthened conditions, Mater. Struct. (2015) 1–16. [30] V.G. Haach, G. Vasconcelos, P.B. Lourenço, Experimental analysis of reinforced concrete block masonry walls subjected to in-plane cyclic loading, J. Struct. Eng. ASCE 136 (2009) 452–462.