SEISMIC PERFORMANCE OF MASONRY-INFILLED · PDF fileSEISMIC PERFORMANCE OF MASONRY-INFILLED...

SEISMIC PERFORMANCE OF MASONRY-INFILLED R.C. FRAMES: BENEFITS OF SLIGHT REINFORCEMENTS

G. Michele CALVI Professor University of Pavia Pavia Italy

Davide BOLOGNINI Graduate technician EUCENTRE Pavia Italy

Andrea PENNA Researcher EUCENTRE Pavia Italy

ABSTRACT A general review of seismic performance of infilled R.C. frames, both for global in-plane response and local out-of-plane is here presented. Some issues related to observed interaction between masonry infills and bounding frames are highlighted and analytical methods summarized. Experimental and numerical results show that frames with slightly reinforced masonry infills generally perform better than bare frames: enhanced lateral capacity and energy dissipation provide a significantly better behaviour in terms of operational limit states and cost of repair. 1. INTRODUCTION A lot of research activity, both numerical and experimental, has been devoted, during last 50 years, to investigate the seismic response of infilled reinforced concrete frames. Design rules and recommendations have been developed for this type of widely diffused structures on the basis of research achievements and observed seismic vulnerability. Recent earthquakes (e.g. Izmit 1999, Boumerdes 2003) have shown how much this topic is still essential at present time, both for existing and new constructions. The importance of the so called “non structural” elements in governing the global seismic response and the corresponding level of safety against collapse has been many times highlighted

254 SÍSMICA 2004 - 6º Congresso Nacional de Sismologia e Engenharia Sísmica

(Fardis et al. [16]). Global capacity can be significantly affected by the interaction of the infills with the frame structure, both in terms of expected failure mode and energy dissipation resources. Several common concepts have been disproved. For example, it is usual to assess the frequency content of the response controlled by the RC structure. This assumes that in most stories infills crack and separate from the frame early in the response. However, the contrary is shown as most of the energy dissipation takes place in the infills and structural damage in beams and columns tends to be low. When damage or serviceability limit states are considered, the effects of infills are often so dominant, that the Ductility Class and even the PGA for which the bare frame was designed plays a minor role in the response and, correspondingly, the characteristics of the infill panels become more important. In particular it appears that serviceability displacement limits can be impossible to satisfy considering the bare structures, while being fully achieved by the infilled frames. The out-of-plane vibration of infill panels can have a beneficial effect on the global response by reducing the participating mass in the fundamental modes of vibration, and hence all global response parameters. However the possible expulsion is of some concern for large panels at high storeys. The generally beneficial effect from the presence of infills on the global response, particularly for what concerns a collapse limit state, increases the interest in controlling the damage level reached in the infills for different design level earthquakes. In particular it is considered of great interest to assess the functional level of a building after a seismic event, in order to estimate the possibility of an immediate use or of a short term repair. This largely depends on the level of damage sustained by each infill panel, and particularly on the potential for out of plane expulsions. To this end it may be noted that the clay blocks most widely used in European earthquake prone countries have low compression and tension strength, a high percentage of holes and a generally brittle type of behaviour. However, it is felt valuable to explore whether the introduction of a low percentage of horizontal reinforcement in the mortar layers or in the external plaster may be of some importance, also considering the relatively low cost of the measure. It may be mentioned that old tests on weakly reinforced infilled frames allow to expect a significant improvement in the response (Brokken and Bertero [2]) . Moreover the present version of EC8 implicitly requires some measures to preserve the integrity of infill panels. 2. OUT-OF-PLANE SEISMIC BEHAVIOUR OF INFILL WALLS 2.1. Out-of-plane response of infill walls Several studies (Dawe and Seah[12], Angel et al.[1] and Flanagan and Bennett[18]) have investigated the out-of-plane response of infill panels subjected to horizontal loads in last years. They all have found that masonry panels restrained by a bounding frame can develop non-negligible out-of-plane resistance due to the formation of an arching mechanism and depending on the panel slenderness (height-to-thickness) ratio and the compressive strength of infill masonry.

G. Michele CALVI, Davide BOLOGNINI, Andrea PENNA 255

Flanagan and Bennett also studied the influence of in-plane damage on out-of-plane capacity and they have concluded that the interaction of in-plane and out-of-plane responses is not generally significant from the load resistance standpoint: it seems to depend on the slenderness ratio and usually external infill panels are much less slender than internal ones. For high slenderness ratio infills, Angel et al.[1] have experimentally found that the reduction of the out-of-plane strength due to in-plane damage could be as high as 50%. 2.2. Out-of-plane seismic analysis The out-of-plane capacity of infills can be effectively studied keeping into account the presence of the arching mechanism: McDowell et al.[23] and Dawe and Seah[12] have developed analytical models to evaluate the out-of-plane arching action, respectively referring to a unidirectional (2D) and a bidirectional (3D) behaviour of unreinforced masonry slabs confined by rigid boundaries. Because of McDowell et al. model tendency to overestimate out-of-plane capacity, due to the assumed elastic perfectly plastic stress-strain relation and non considered interaction with in-plane damage, Angel et al. have proposed an improvent consisting in the introduction, into the McDowell et al. model scheme, of strength reduction factors that account for the influence of in-plane damage and the flexibility of the bounding frame. Starting from a simplification of their analytical model they also proposed a practical capacity assessment procedure. As reported by Shing and Mehrabi [34], both the models of Dawe and Seah and Angel et al. tend, in some cases, to overestimate the out-of-plane resistance. As considered in modern standards, simplified formulas are available to estimate the actual earthquake demand for infill panels out-of-plane response, keeping into account the dynamic amplification by means of global building response and local infill out-of-plane response natural periods. 3. IN-PLANE BEHAVIOUR OF INFILLED R.C. FRAMES 3.1. In-plane seismic response The in-plane response r.c. frames infilled with masonry panels under lateral loads is, since a long time up to present, an important research topic in structural and earthquake engineering. A number of significant experimental tests have been performed in last 50 years (e.g. Fiorato et al.[17], Klingner and Bertero[20], Brokken and Bertero [2], Zarnic and Tomazevic[38], Mosalam et al.[27], Mehrabi et al.[24], Angel et al.[1], Calvi and Bolognini [4], Fardis et al. [15], Pinto et al. [28]). All studies have shown that the global response of infilled frames is heavily influenced by the interaction of the infill with its bounding frame. The lateral capacity of infilled frames usually depends on the interaction between the infills and the bounding frame: it can determine both the initial and the collapse mechanism. Under relatively low lateral loads the infills remain in contact with the frame structure and their contribution significantly increases the global stiffness of the frame. Under higher loads, the masonry infills, because of their no tension behaviour partially separate from the bounding frame. This load resisting system can be classically represented as a


frame with equivalent compression trusses as observed since early researches. Further collapse behaviour can then evolve depending on relative strength and stiffness of frame structure and infill walls. Mehrabi et al.[24] have demonstrated that relatively weak unreinforced masonry infills can enhance the stiffness and strength of a non-ductile reinforced concrete frame significantly without jeopardizing ductility. 3.2. Seismic analysis of infilled frames Mehrabi et al.[24] have defined a secant stiffness for infilled frames by means of a shear beam model and they have found a close correlation with weak infill frames they have tested. Because of frame-infill interaction, the load resisting mechanism of an infilled frame can be very different from that of a bare frame or a wall panel alone. Due to relatively low lateral load, the separation of infill masonry from the bounding frame leads to form a resisting mechanism which can be easily modelled as a compression diagonal strut (Polyakov [29]). Holmes[20] has proposed that the effective width of an equivalent strut depends primarily on the thickness and the aspect ratio of the infill, Stafford Smith[35] has then refined this approach and compared his results with series of tests. Stafford Smith has defined a dimensionless relative stiffness parameter to determine the degree of frame-infill interaction and the effective width of the strut.

4sin 2

4m

hC C m

E tE I h

θλ = (1)

where Em and t are the elastic modulus and thickness of the infill, ECIC is the bending stiffness of the columns, and hm and θ are the height and the angle between the diagonal and horizontal of the infill. However, in spite of some limited results, equivalent strut models do not appear as accurate tools to calculate the strength of an infilled frame: infilled frames have a number of possible failure modes caused by the frame-infill interaction. Compression strut type failure is just one of many possible modes and it can not represent other phenomena such as induced short-column or sliding bed-joint failure of masonry. Other refinements such as multi-strut modelling of infill walls and their interaction with the bounding frame have been very well summarized by Crisafulli et al. [11]: they also highlight some advantages and disadvantages of these procedures and indicates practical recommendations for their implementation. As an alternative approach, limit analysis simplified models can provide practical evaluations of infilled frames strength capacity: Mehrabi et al.[24] have formulated a general methodology which can be applied to masonry-infilled reinforced concrete frames. They have selected five most probable failure modes for single-storey one-bay infilled r.c. frames (see Figure 1) and each collapse mechanism lateral capacity can be calculated from equations derived by mechanical assumptions: the failure mode corresponding to the minimum lateral resistance is assumed as the dominant one.


Figure1: Failure modes for one storey single-bay infilled r.c. frames (Shing and Mehrabi, 2002)

Analytical results indicate that Mechanism 5 and 2 seem to be dominant respectively for weak and strong infills.

Figure 2: Limit analysis models for elementary mechanisms (Shing and Mehrabi, 2002)

4. WEAKLY REINFORCED INFILL WALLS Fardis et al. [16], investigating the seismic response of infilled frames, designed applying modern codes of practice, have highlighted the importance of the so called “non structural” elements in governing the global seismic response and the corresponding level of safety against collapse. Several common concepts have been disproved. For example, it is usual to assess the frequency content of the response controlled by the RC structure. This assumes that in most stories infills crack and separate from the frame early in the response. However, the contrary is shown as most of the energy dissipation takes place in the infills and structural damage in beams and columns tends to be low. While most of the studies have focused on unreinforced masonry panels, Klingner and Bertero[20] and Brokken and Bertero[2] investigated the behaviour reinforced infills: they have suggested that slight reinforcements can provide superior performance, in terms of strength, stiffness, and energy dissipation, compared with a bare frame. The research presented in this paper was devoted to compare the in plane response of traditional and slightly reinforced hollow masonry infill panels, for different intensity level earthquakes, in order to assess the potentially different damage level attained; to assess the potential for out of plane expulsion of traditional and slightly reinforced masonry infill panels, at different level of damage induced by in plane action; to evaluate the effects of different properties of the infill


panels on the response of buildings with different geometrical configurations and different distribution of the infill panels, in terms of PGA required to induce given level of damage. The whole project was related to new constructions, not to rehabilitation of existing ones. For this reason the reinforced concrete frames had been designed according to modern standard (Eurocode 8), avoiding any possible shear failure. The objectives stated above require both experimental and numerical research. The former to characterise the response of different types of infill panel, tested in-plane at different drift targets and out of plane at different damage levels produced by the in plane testing. The latter to perform non-linear analyses on several building types, defining the response of the infill panels on the base of the experimental data. The main test campaign has been performed on one-bay, one-storey full scale infilled frames, considering one single geometry and one single type of clay unit, for obvious budget reasons. Three possible reinforcement conditions were considered: no reinforcement, reinforcement in the mortar layers at 600 mm distance and light wire meshes in the external plaster. Interstorey drifts equal to 0.1, 0.2, 0.3 and 0.4 % were considered relevant for serviceability – low damage limits states and drifts equal to 1.2 and 3.6 % were considered significant to explore heavy damage or close to collapse limit states. At different levels of in plane drifts, out of plane tests were performed to define strength domains as a function of in plane damage. The numerical parametric analyses were performed considering two bay plane frames with four, eight or twelve storeys. Infill panels were either distributed uniformly everywhere, on one bay only or on all storeys except the ground one. The results obtained from push–over analyses were considered in conjunction with three earthquake spectral shapes, corresponding to dense, medium or soft soil. The possible out of plane expulsion of each panel was verified considering the actual strength domain as a function of the local in plane drift and the estimated storey acceleration derived from properties of the single-degree-of-freedom oscillator corresponding the in-plane damage state and the appropriate height level of the panel. 4.1. Experimental research The RC frames were constructed considering standard 25 MPa concrete and 500 MPa steel and applying all the design procedures recommended in Eurocode 8, including capacity design principles. The effective properties of the concrete and of the ribbed steel re-bars are summarised in the Tables 1 and 2. The clay blocks were selected as typical highly perforated units used in the European earthquake-prone countries, with the following properties. Dimensions: 245 × 115 × 245 (height) mm Average weight per unit: 34.8 N Void ratio: 60% Strength parallel to the holes: 15.4 MPa (c. o. v. = 0.12) Strength perpendicular to the holes: 2.8 MPa (c. o. v. = 0.16) Small size wallettes were tested to assess the properties of masonry, considering the four different cases of Table 3. Note that the compression strength in the direction parallel to the perforation is


significantly lower if the wall is constructed laying the blocks with horizontal holes. In this case actually, little mortar is placed in the vertical joints and when the specimen is rotated of 90 degrees the area really bearing most of the load is significantly less than the actual section of the clay units. The infill panels were constructed with horizontal holes, therefore the values of the second and third lines of the table are more appropriately representing the real situation. The consequence is that strength and stiffness in the vertical and horizontal directions are not very different. Note the low absolute values for compression and tensile strength and the very large dispersion of the stiffness values. Such values are considered to be representative of infill panels commonly constructed in the Mediterranean area. The tensile and the compressive strength of the mortar are indicated in the Table 4 (the dimensions of each specimen were 40x40x100mm). The horizontal reinforcement used for the mortar layers consisted of either two 6 mm re-bars with 578 MPa yield strength, 638 MPa ultimate strength and 24 % elongation capacity measured on 5 diameters, or by a welded truss made with two 5 mm bars with 630 MPa yield strength, 690 MPa ultimate strength and 16.5 % elongation capacity measured on 5 diameters (10.5 % on ten diameters, 3.34 uniform at maximum stress). The external welded mesh was made with 1 mm diameter wire, spaced at 20 mm in one direction and at 12.5 mm in the other. The measured tensile strength of a single wire was 356 MPa. As already pointed out, a single geometry was considered for the main test campaign on full scale infilled frames. The overall dimensions were selected to be 4.5 × 3 (height) m. The concrete frame was designed as the lowest part of a four storey building, applying thoroughly the rules given by Eurocode 2 and Eurocode 8. The structure was designed according to the high ductility class of EC8, applying fully the capacity design rules. The plastic hinges were therefore forced to form in the beams and all critical sections were well confined. The standard EC8 spectrum defined for medium density soils was used.

Table 1: Concrete cylindrical strength

fcm [N/mm2]

columns beam

29.32 34.56

Table 2: Properties of the re-bars

diameter [mm]

area [mm2]

yield strength [N/mm2]

failure strength [N/mm2]

7.94 49.51 557 630 12.08 114.64 575 686 15.82 196.68 555 647 22.45 395.83 561 670


Table 3: Summary of the results obtained from tests on wallettes. A total of twenty tests were

performed on specimens with approximate size 800 × 800 × 120 mm. fm, ave average compression strength Em, ave average modulus of elasticity, measured between 10 and 40 % of the strength ft,ave (R) average tensile strength, obtained according to the Rilem Recommendations ft,ave (F) average tensile strength, obtained with a finite element model fv0,ave average shear strength, with no vertical load (cohesion) G average shear modulus, measured between 10 and 40 % of the strength fm, ave [MPa] c.o.v. Em, ave

[MPa] c.o.v.

Load parallel to the holes, construction with vertical holes 3.97 0.18 5646 0.47

Load parallel to the holes, construction with horizontal holes 1.11 0.12 991 0.43

Load perpendicular to the holes 1.10 0.38 1873 0.43

Diagonal load ft,ave (R) [MPa] ft,ave (F) [MPa]

Fv0,ave [MPa]

G [MPa]

Average value 0.15 0.11 0.09 1039 c. o. v. 0.25 0.25 0.25 0.36

Table 4: Properties of the mortar tensile strength [N/mm2] compressive strength [N/mm2] 0.42 5.54

Table 5: Properties of the bars used to reinforce the infills

diameter [mm] area [mm2] yield strength [N/mm2] failure strength [N/mm2]

5.00 19.63 630 690 6.13 29.55 578 638


Figure 3: Details of the position of the steel bars and of the meshes to reinforce masonry panels

The details of geometry and reinforcement resulting from these assumptions are summarised in Figure 4. All infill panels had identical geometry, with dimensions equal to 4200 × 2750 × 135 mm. The thickness resulted from the combination of block thickness (115 mm) plus 10-mm plaster on both sides. As already indicated, three reinforcement types were then considered: no reinforcement, reinforcement in the mortar layers, and external reinforcement. The mortar layer reinforcement was spaced at 600 mm, resulting in a geometrical percentage between 0.08 and 0.1%, depending on the diameter of the bars used (see the previous section). The average geometrical reinforcement percentage resulting from the external mesh was almost exactly the same, but being positioned both in the vertical and in the horizontal direction, implied approximately twice the total steel weight. The meshes on the two sides were connected at distances of 1 m (horizontal) and 0.6 m (vertical) by 5 × 1 mm steel plates left in the mortar beds during construction and bent around the mesh. Note that the total steel weight used for each panel was less than 100 N for the case of layer reinforcement and less than 200 N for the case of external reinforcement.


Figure 4: Geometry and reinforcement details of the RC frames used for all experimental tests

All tests were performed applying first two vertical loads on the columns, to simulate the presence of the upper storeys. No vertical load was placed on the beam, accepting this small difference from reality. The total vertical load was then kept constant during the tests, allowing the redistribution generated by the application of horizontal forces. The in-plane tests were performed applying horizontal displacements cycles, according to pre-defined targets between 0.1 and 3.6 % drift, as discussed later. As standard, three cycles were performed at each target displacement. The data acquisition system consisted in the transducers shown in Figure 3.


The out-of-plane tests were performed at different levels of in-plane damage (i.e. after having reached different drift levels in the in-plane tests) as discussed later. A four-point load was monotonically applied, measuring the displacements in five points, as shown in Figure 4.

Figure 5: Instrumentation for the in-plane tests (one side)

Figure 6: Instrumentation for the out-of-plane tests (one side)

The experimental response of the in-plane tests was preliminarily simulated performing a pushover non-linear analysis, considering first the bare frame, and subsequently the introduction of a diagonal strut supposedly equivalent to the real panel (Priestley and Calvi [32]). The results obtained are summarised in Figure 7. It has to be noted that the high stiffness of the panel produces a first steep line in the force-displacement curve. The actual point of collapse of the panels is obviously affected by large uncertainties, a consequence of the large dispersion of the mechanical properties of the masonry panel. After the panel reaches its maximum capacity, it may


be assumed that some deterioration takes place, with a final return on the bare frame curve. The actual shape of the force-displacement curve between the peak point of the panel and the bare frame curve is difficult to be preliminarily predicted. The dashed curve in Figure 7 has no rational basis. Also the peak point can be significantly higher, being followed by a more or less fast decrease in force capacity.

0 50 100 150 200 250DISPLACEMENT [mm]

0

100

200

300

400

BA

SE S

HEA

R [k

N]

Figure 7: Force-displacement curve resulting from preliminary pushover analysis

A summary of the in-plane tests performed is presented in Table 6. The sequence of drift targets has been adjusted during the test campaign, to explore the response for close-to-collapse limit states (drifts at 1.2 % or larger) and for serviceability limit states (drift below 0.4 %). In some cases, the in-plane tests were continued after out of plane expulsion of the infill panels, essentially to verify the ultimate response of the frame alone.

Table 6: Summary of the in-plane tests

Reinforcement Test Driftn. 0.1% 0.2% 0.3% 0.4% 1.2% 3.6% 0.4%

bare frame 1 3 - - 3 3 3 3

non-reinforced 2 3 - - 3 3 * - -

φ 6 mm bars 3 3 - - 3 3 3 -

φ 5 mm bars 4 3 - - 3 * - 3 -

mesh 5 3 - - 3 3 * 3 -

non-reinforced 6 3 3 3 3 * - - -

φ 6 mm bars 7 3 3 3 3 * - - -

φ 5 mm bars 8 3 3 3 3 * - - -

mesh 9 3 3 3 3 * - - - * = cycles preceding the out-of-plane test


The force-displacement curves obtained from the first five tests are depicted in Figure 8.

-120 -100 -80 -60 -40 -20 0 20 40 60 80 100 120displacement [mm]

-450

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400

450

F [k

N]

-120 -100 -80 -60 -40 -20 0 20 40 60 80 100 120displacement [mm]

-450

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400

450

F [k

N]

Test 1: bare frame Test 2: unreinforced panel

-120 -100 -80 -60 -40 -20 0 20 40 60 80 100 120Displacement [mm]

-450

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400

450

F [k

N]

-120 -100 -80 -60 -40 -20 0 20 40 60 80 100 120Displacement [mm]

-450

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400

450

F [k

N]

Test 3: 6 mm re-bars mortar layer

reinforcement Test 4: 5 mm truss mortar layer reinforcement

-120 -100 -80 -60 -40 -20 0 20 40 60 80 100 120Displacement [mm]

-450

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400

450

F [k

N]

-120 -100 -80 -60 -40 -20 0 20 40 60 80 100 120displacement [mm]

-450

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400

450

F [k

N]

1st CYCLES

Bare frame (test 1)

Non-reinforced (test 2)

φ6 mm Bars (test 3)


Mesh (test 5)

Test 5: external mesh reinforcement Tests 1-5


-15 -10 -5 0 5 10 15Displacement [mm]

-450

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400

450

F [k

N]

-15 -10 -5 0 5 10 15Displacement [mm]

-450

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400

450

F [k

N]

Test 7: 6 mm re-bars mortar layer reinforcement

Test 8: 5 mm truss mortar layer reinforcement

-15 -10 -5 0 5 10 15Displacement [mm]

-450

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400

450

F [k

N]

-15 -10 -5 0 5 10 15Displacement [mm]

-450

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400

450

F [k

N]

ENVELOPES OF 1' CYCLES

Non-reinforced (test 6)



Mesh (test 9)

Test 9: external welded mesh reinforcement Tests 6 – 9: first cycles envelopes

Figure 8: Force-displacement curves obtained from tests The initial stiffness of infilled frames is significantly larger than the stiffness of the bare frame, with no significant influence from the presence of the reinforcement; In all cases the curves produced by the infilled frame response tend to return to the bare frame curve at drifts values relevant to a collapse limit state, being therefore confirmed that the equivalent period of vibration for high intensity earthquakes is not significantly affected by the presence of the infills, since the effective stiffness essentially corresponds to the stiffness of the bare frame; The energy dissipation for low amplitude cycles is much larger for infilled than for bare frames, confirming that a reduced global response can be expected if local failure modes can be excluded; All kinds of reinforcement are effective in avoiding or significantly reducing the strength deterioration of the panels; The presence of 5 or 6 mm diameter bars, in the form of re-bars or trusses does not significantly affect the results; The presence of the external meshes improves the response enormously, in terms of strength, stiffness and energy dissipation capacity. The apparent state of damage of the panels at drifts equal to 0.2 % and 0.4 % is depicted in Figure 9. The beneficial effect of a mortar layer reinforcement is evident, as well as the superiority


of the case with external reinforcement. The evidence for this latter case is that during the test large cracks open between panel and frame, without significant damage in the panel. Not only then the damage is limited, but also easy to be repaired. Apparently, the external reinforcement is effective without any need of steel continuity between frame and panel.

Unreinforced panel: 0.2 % drift Unreinforced panel:0.4 % drift Mortar layer reinforcement: 0.2

% drift

Mortar layer reinforcement: 0.4 % drift

External reinforcement: 0.4 % drift

Figure 9: State of damage after three cycles at 0.2 % and 0.4 % drift; test 9 at 2 % drift showed essentially no damage

The main results obtained from the out-of-plane tests are summarised in Table 7 and in Figures 10. The effects of presence of reinforcement and of previous in-plane damage on the failure domain of the panels are evident in the figures and quantitatively shown in the table. Actually, considering the horizontal force capable of producing an out-of-plane expulsion and the maximum attainable displacement, an equivalent period of vibration at incipient collapse was calculated for each panel. Considering then the total mass of the panel, the average acceleration capable of inducing this event was estimated. In the next section this information will be used to assess the possible out of plane expulsion of the panel, evaluating the floor acceleration from the in-plane response and, in turn, the panel acceleration as the result of the response of the panel excited by the floor acceleration. The beneficial effect of the reinforcement, and to a larger extent of the external welded wire mesh is again evident. For the unreinforced, undamaged panel, the local acceleration required to induce collapse is around 2.5 g, but reduces to 0.5 – 0.7 g after some in-plane damage. The presence of some mortar layer reinforcement increases these values approximately to 2.8 and 1.4 g and the external mesh approximately to 3.5 g at 0.4% in-plane drift and to 1.6 g at 1.2% in-plane drift.


The limited displacement capacity of unreinforced panels, results in longer period of vibration, which in turn may result in larger amplification when combined with the main period of vibration of the building. While these figures have obviously a relative meaning, being related to a specific panel geometry, it is immediately clear that very large PGAs would be necessary to induce out-of-plane collapse in reinforced panels.

Table 7: Summary of the out-of-plane tests. For all panels the total area was 11.55 m2 and the total mass 1360 kg

Reinforcement Test Previous Force Displ. 2 Acceleration 3 Eq. Pressure 4 Stiffness 5 Periodn. in-plane drift [kN] [mm] [g] [kPa] [kN/m] [s]

non-reinforced 2 1.2% 6.00 - 0.460 1.00 180 0.546bars 3 - - - - - - -

Murfor 4 0.4% 17.20 53 1.289 2.87 5160 0.102mesh 5 1.2% 21.41 22 1.605 3.57 2676 0.142

non-reinforced 6 0.4% 9.00 37 0.675 1.50 720 0.273bars 7 0.4% 19.70 44 1.477 3.28 2814 0.138

Murfor 8 0.4% 17.50 48 1.312 2.92 1400 0.196mesh 9 0.4% 46.60 28 3.493 7.77 11184 0.069

non-reinforced 1 10 - 33.70 13 2.526 5.62 11642 0.068Murfor 1 11 - 36.77 45 2.756 6.13 11977 0.067

1 Test executed on the panel with no damage2 Ultimate displacement3 Acceleration corresponding to the maximum transversal force (with the total mass participating)4 Equivalent uniform pressure giving to the panel the same maximum bending moment as the four concentrated forces5 Secant stiffness calculated between the 10% and 40% of the strength of the panel

0 -10 -20 -30 -40 -50 -60Average displacement [mm]

0

10

20

30

40

50

Forc

e [k

N]

Non-reinforced (test 06)φ6mm bars (test 07)φ5mm bars (test 08)Mesh (test 09)

0 -10 -20 -30 -40 -50 -60Average displacement [mm]

0

10

20

30

40

Forc

e [k

N]

φ5mm bars (test 11)Non-reinforced (test 10)φ5mm bars (test 08)Non-reinforced (test 06)

After three cycles at 0.4 % in-plane drift Undamaged panels and panels previously driven to 0.4 % in-plane drift

Figure 10: Comparison of force-displacement curves obtained from out-of-plane tests performed on differently reinforced panels

4.2. Numerical simulations To extend the experimental results obtained on single infill panels to the global response of more complex frames, some parametric pushover analyses have been performed, using a modified version of the ANSR code [Mondcar, D.P. and Powell, G.H., 1975]. Again it has been necessary to consider one single geometry for each panel, similar to that of the experimental tests. A total of 81 two-bay plane frames have been analysed, according to the variables shown in Table 8.


Table 8: Variables considered in the parametric analyses

Number of floors 4, 8 and 12

Type of reinforcement Mortar layer reinforcement, external mesh, unreinforced

Distribution of the infills Fully infilled frame, one bay infilled, not infilled at the ground floor only

Input spectra Dense, medium or soft soil, according to EC8

All frames were designed in accordance with Eurocode 8 for a peak ground acceleration equal to 0.25g, a medium soil and a high ductility class. The pushover analyses were then performed with the following assumptions. The vertical distribution of the horizontal forces was always assumed according to an inverted triangle shape. Beams and columns were simulated with Takeda-type non-linear elements. To maintain an inverted triangular shape for the applied forces and to control the softening branch of the response, an auxiliary isostatic structure with very high stiffness has been used. All infills were simulated using equivalent diagonal struts with force-displacement curves derived from the experimental tests. A typical envelope curve is depicted in Figure 11, actual examples of complete curves are shown and compared with the experimental results in Figure 12 for different cases. The strut width was assumed to be equal to 25 % of the total diagonal width of the panel (Paulay and Priestley [27]). The initial periods of vibration obtained for the different cases are reported in Table 9. The only interaction considered between in-plane and out-of-plane response was the correction of the out-of-plane strength domain. The out-of-plane response was simulated considering a panel acceleration produced by the combination of two dynamic responses. Firstly, an equivalent single-degree-of-freedom structure was derived from the in-plane response, obtaining a filtered acceleration at each time step, secondly, the interaction between floor input and panel response was evaluated according to the following equation:

Sa/afloor = 3 / (1 + (1 – Ta/Tf)2) (2) where: Sa is the acceleration at the panel, afloor is the floor acceleration, derived from the in-plane response, Ta and Tf

are the periods of vibration of panel and frame It may be noted that the equation is derived from the one suggested by EC8 for the evaluation of the equivalent static force to be applied to non structural elements, considering the actual response of the frame, as given by a single-degree-of-freedom simulation.

Table 9: First periods of vibration (seconds) obtained for the various cases considered first period of vibration 4 floors 8 floors 12 floors Bare frame 1.16 2.18 3.86 Fully infilled 0.39 0.76 1.38 One bay infilled 0.51 1.05 1.90 Soft storey 0.62 0.84 1.41


Axial deformation

Axial c

ompr

essive

stre

ss

Elastic phase

K=EA/L

Envelope curve(N = A ∗ σ ∗ exp(-β ∗ ε))

Figure 11: Example of force-displacement envelope curve of an equivalent strut

-120-110-100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400FORCE[kN]

Infilled frame (unreinforced infill)

Numerical simulation

Bare frame

DISPLACEMENT[mm]

-120-110-100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400FORCE[kN]

Infilled frame (unreinforced infill)


Bare frame

DISPLACEMENT[mm]

Unreinforced infill – envelope Unreinforced infill – cyclic test

-120-110-100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400FORCE[kN]

Infilled frame (infill reinf.: 2 φ5mm bars)


Bare frame

DISPLACEMENT[mm]

-120-110-100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400FORCE[kN]

Infilled frame (infill reinf.: 2 φ5mm bars)


Bare frame

DISPLACEMENT[mm]

Mortar layer reinforcement – envelope Mortar layer reinforcement – cyclic test


-120-110 -100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400FORCE[kN]

Infilled frame (infill reinf.: external mesh)


DISPLACEMENT[mm]

-120-110-100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120

-400

-350

-300

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

350

400FORCE[kN]

Infilled frame (infill reinf.: external mesh)


DISPLACEMENT[mm]

External mesh reinforcement – envelope External mesh reinforcement – cyclic test

Figure 12: Comparison between experimental results and numerical strut models The global force-displacement curves obtained for the different cases considered are reported in Figure 12. In all cases, at each integration step the situation of all panels was examined, to identify the state of damage with reference to either the in-plane and out-of-plane response. Four possible situations were considered of interest, as depicted in Figure 13. Four global limit states (Calvi [6]), or performance levels, were then defined, according to Table 10.

ε ax.

σ ax.

A

B

C

Figure 13: Damage states considered for each one of the infill panels: (A) maximum force, (B) 50

% strength degradation, (C) in-plane collapse and (D) out of plane expulsion

Table 10: Global limit states (performance levels) considered (points A – D refers to Figure 13) LS1 No damage – fully operational no panel reaches point A or D LS2 Light damage – operational no panel reaches point B or D

LS3 Severe damage – life safe – repairable

no panel reaches points C or D and not more than 50% of the panels reaches point B

LS4 Very heavy damage – life danger all other conditions


The PGA required to produce each one of the four limit states was computed considering the appropriate point on the force-displacement curve, an equivalent stiffness, an equivalent viscous damping and each one of the three response spectra recommended in EC8 for different soil condition. The equivalent viscous damping was estimated using the following equation (Calvi [5]).

( ) 5/1120 5.0 +−⋅= µξ eq (3)

where µ is the displacement ductility. The summary of the results obtained is shown in Figure 27 and in Table 11, in terms of PGA required to attain a given limit state for each case of geometry (number of storeys), infill situation (all infilled, one bay infilled, soft storey configuration), and spectral content (dense, medium or soft soil). Some of the results seem to show some slight inconsistency (e.g. the PGA required to attain a given limit state for the 12 storey-soft storey-mesh may be higher than the corresponding one required for the 12 storey-fully infilled-mesh). This is due to the relatively rough models used, particularly to define a specific limit state: for example life safety may be governed by a potential out-of-plane expulsion, not necessarily related to the overall strength of the building. In Figure 17 the vertical bars represent the minimum and maximum acceleration considering all possible combinations: while there is relative little effect on LS1, the average increments of acceleration are in the order of 50 to 100 % for internal reinforcement and between 200 and 300 % for the external mesh. It is logical to expect minor effect on LS1, since the small amount of reinforcement added does not change significantly the peak strength of the panels, but rather the post peak response. It is worth mentioning that a more detailed analysis of the data, allows to notice that the higher in-plane damage of lower storeys may make them more sensitive to out-of-plane damage, even if the corresponding acceleration is lower. This may suggest to consider some modification to the EC8 equation to evaluate equivalent forces on non-structural elements. As expected, the attained accelerations can be significantly larger than the design PGA of 0.25 g. Actually, it has been experimentally and numerically proven that the real capacity of well designed structures can sustain from two to three times the design level acceleration [30,31]. It has to be pointed out and kept in mind, that the global results shown in figure 17 have the merit of showing a tendency and have some practical value, but should not be considered rigorous. More specifically, further work will be necessary to establish an accurate relation between actual PGA and the acceleration levels calculated from static forces.

0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00Top displacement [m]

External reinforcement (mesh)

Reinforcement in the mortar bed joints

Unreinforced

LS1

LS1LS2

LS4LS2

LS2

0

100

200

300

400

500

600

700

800

900

Base

she

ar [k

N]

SL4

SL4


0

100

200

300

400

500

600

700

800

900

Base

she

ar [k

N]



Unreinforced

LS1 LS2LS3 LS4

LS4


0

100

200

300

400

500

600

700

800

900

Bas

e sh

ear [

kN]

Externa reinforcement (mesh)


Unreinforced

LS1 LS2LS4

Fully infilled frame One bay infilled frame Ground floor not infilled

Figure 14: Force-displacement curves obtained for the four storey frame


0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80Top displacement [m]

0

100

200

300

400

500

600

700

800

900

Bas

e sh

ear [

kN]



Unreinforced

LS1

LS2 LS4

LS4

LS4




Unreinforced

0

100

200

300

400

500

600

700

800

900

Bas

e sh

ear [

kN]

LS1

LS2LS3 LS4

LS1


External reinforcement


Unreinforced

0

100

200

300

400

500

600

700

800

900

Bas

e sh

ear [

kN]

LS1

LS2LS4

LS1

LS2LS4

LS4


Figure 15: Force-displacement curves obtained for the eight storey frame

0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 2.20 2.40 2.60 2.80 3.00Top displacement [m]

External mesh


Unreinforced

0

100

200

300

400

500

600

700

800

900

Base

she

ar [k

N]

LS1 LS2 LS4

LS1

LS2

LS4

LS2

LS4


0

100

200

300

400

500

600

700

800

900

Bas

e sh

ear [

kN]



Unreinforced

LS1

LS2

LS3

LS4




Unreinforced

0

100

200

300

400

500

600

700

800

900

Bas

e sh

ear [

kN]

LS1LS2

LS4

LS1

LS2

LS4

LS1

LS2LS4


Figure 16: Force-displacement curves obtained for the twelve storey frames

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

1.10

1.20

1.30

1.40

a g[g

]

LS1 LS2 LS3 LS4

UnreinforcedInternal

reinforcementExternal

reinforcement Figure 17 – Minimum and maximum PGA required to attain a given limit state, as a function of

reinforcement conditions.


5. CONCLUSIONS The results presented in this paper are based on several assumptions, and some care should be used to extrapolate general conclusions. It is worth to remind that: the experimental tests have been performed on single bay, single storey specimens; a single geometry and a single design of the concrete frame has been considered; a single type of masonry units has been used; the numerical analyses have been performed considering a single global geometry and a single ductility level; a push-over approach has been adopted for the analyses; the effects of out-of-plane response on the in-plane response have been neglected. The choices have been to concentrate on the effects of the presence of some reinforcement in the infill panels at different level of demand, for what concerns the experimental tests, and on number of storeys, infill pattern, spectral content of the input and limit state or performance level for what concerns the numerical analyses. The presence of little reinforcement, significantly improves the response of a single infilled frame, particularly for what concern damage limit states (the effects less important for what concern a first cracking and a full collapse limit states). The state of damage of the non structural elements plays a fundamental role in the definition of limit states, since in general for well designed frame a high damage or a potential for out-of-plane expulsion tends to precede any significant damage to the frame. Note that an out-of-plane expulsion implies danger for human life, therefore it has to be related to a near collapse limit state, even if the concrete frame is far from being near collapse. It is correct to assume higher local forces on non structural elements at upper storeys, but since the in-plane damage tends to concentrate at lower storeys, the critical situation for out of plane expulsion may derive from a combination of level of force and reduced strength domain. However, the most fundamental aspect of the results here presented deals with the global response of infilled frame buildings, and more specifically to the PGA levels required to induce given limit states. As clearly stated, a full confidence on the numerical values may require further analyses, but a clear trend is evident: the use of traditional non-reinforced infill panels can result in structures very sensitive to relatively low input levels for operational and damage limit states, while the insertion of small amount of reinforcement greatly improves the global response. Further work is required to verify the actual relation between PGA and acceleration estimated from static forces, however, according to the results obtained, depending on infill pattern, global structural configuration and ground motion type, the PGA required to induce a damage level that will prevent the use of a building can be estimated between 0.15 and 0.30; the PGA corresponding to a severe and difficult to repair damage level between 0.2 and 0.4 g. Both ranges of values may change as a function of frame geometry, but essentially do not depend on frame design (in terms of earthquake intensity and ductility requirement). The introduction of some reinforcement in the mortar layers, with a geometrical percentage lower than 1%, will result in increasing these acceleration levels to 0.25-0.60 g for an occupational LS and to 0.35-0.70 g for a damage LS. The application of external welded meshes on both sides, with some points of connection, will increase the PGA values to levels always above 0.4 g for an occupational LS. The total amount of steel in this case may be approximately twice the one used for the case of mortar layer reinforcement. In all cases there was no continuity between the steel used for reinforcing the panels and the frame reinforcement. The economical benefit of such simple measures of protection is extraordinarily important, particularly when compared with the relatively low additional construction cost. This should be reflected in mandatory code clauses.


6. REFERENCES [1] Angel R., Abrams, D., Shapiro, D., Uzarski, J. and Webster, M. “Behaviour of reinforced

concrete frames with masonry infills”. Report No. UILU-ENG-94-2005. Department of Civil Engineering, University of Illinois, Urbana-Champaign, IL, USA 1994.

[2] Brokken, S. and Bertero, V.V. “Studies on effects of infills in seismic resistant RC construction”, 1981, Report UCB/EERC, 81-12, University of California, Berkeley

[3] Buonopane S. G. and White R. N. “Pseudodynamic testing of masonry infilled reinforced concrete frame”, Journal of Structural Engineering (ASCE) 1999: 125(6): 578–589.

[4] Calvi G. M. and Bolognini, D., “Seismic response of reinforced concrete frames infilled with masonry panels weakly reinforced” Journal of Earthquake Engineering, 2001, 5, 153-185.

[5] Calvi, G. M. “Performance-based approaches for seismic assessment of existing structures”, Proc. 11th ECEE, 1998, A. A. Balkema, Rotterdam.

[6] Calvi, G. M. “A displacement-based approach for vulnerability evaluation of classes of buildings”, Journal of Earthquake Engineering, 1999 Vol. 3, No. 3, pp. 411-438.

[7] CEN, EC2, Design of Concrete Structures – Part 1.1: General Rules and Rules for Buildings, 1992.

[8] CEN, EC8, Eurocode 8 - Design Provisions for Earthquake Resistance of Structures – Part 1.1: General Rules – Seismic Actions and General Requirements for Structures, 1994.

[9] CEN, EC8, Eurocode 8 - Design Provisions for Earthquake Resistance of Structures – Part 1.2: General Rules – General Rules for Buildings, 1994.

[10] CEN, EC8, Eurocode 8 - Design provisions for earthquake resistance of structures – Part 1.3: General rules – Specific Rules for Various Materials and Elements, 1995.

[11] Crisafulli, F. J., Carr, A. J. and Park, R., “Analytical modelling of infilled frame structures – A general review”, Bull. of the New Zealand Society for Earthquake Engineering, Vol. 33, No. 1, 2000, March.

[12] Dawe J. L. and Seah C. K. “Out-of-plane resistance of concrete masonry infilled panels” Journal of the Canadian Society of Civil Engineering, 1989: 16: 854–864.

[13] Dhanasekar M. and Page A.W. “The influence of brick masonry infill properties on the behaviour of infilled frames” Proc. Institute of Civil Engineers 1986: 81(2): 593–606.

[14] Fardis, M.N., Bousias S.N., Franchioni G. and Panagiotakos T.B. “Sesismic response and design of RC structures with plan-eccentric masonry infills”, Earthquake Engng. Struct. Dyn., 1999, 28, 173-191.

[15] Fardis, M.N., Panagiotakos, T.B. and Calvi, G. M. “Seismic response and design of masonry infilled reinforced concrete buildings”, Structural Engineers World Congress, 1998, San Francisco

[16] Fiorato A.E., Sozen M.A. and Gamble W.L. “An investigation of the interaction of reinforced concrete frames with masonry filler walls”. Report UILU-ENG-70-100. Department of Civil Engineering, University of Illinois, Urbana-Champaign IL, USA. 1970.

[17] Flanagan R.D. and Bennett R.M. “Bidirectional behaviour of structural clay tile infilled frames” Journal of Structural Engineering (ASCE) 1999: 125(3): 236–244.

[18] Flanagan R.D. and Bennett R.M. “In-plane behaviour of structural clay tile infilled frames” Journal of Structural Engineering (ASCE) 1999: 125(6): 590–599.


[19] Klingner R.E. and Bertero V.V. “Infilled frames in earthquake-resistant construction” Report

EERC/76-32. Earthquake Engineering Research Center, University of California, Berkeley, CA, USA 1976.

[20] Lotfi, H. R. and Shing, P. B. “An interface model applied to fracture of masonry structures” Journal of Structural Engineering (ASCE) 1994: 120(1): 63–80.

[21] Madan A., Reinhorn A.M., Mander J.B. and Valles R.E. “Modelling of masonry infill panels for structural analysis” Journal of Structural Engineering (ASCE) 1997: 123(10): 1295–1302.

[22] McDowell E.L., McKee K.E. and Sevin E. “Arching action theory of masonry walls” Journal of the Structural Division (ASCE) 1956: 82(2): 915–1 to 915–918.

[23] Mehrabi A.B., Shing P.B., Schuller M.P. and Noland J.L. “Experimental evaluation of masonry-infilled rc frames” Journal of Structural Engineering (ASCE) 1996: 122(3): 228–237.

[24] Paulay T. and Priestley M.J.N. “Seismic Design of Reinforced Concrete and Masonry Buildings”, 1992, Wiley, New York.

[25] Mehrabi, A. B. & Shing, P. B. “Finite element modelling of masonry-infilled rc frames” Journal of Structural Engineering (ASCE) 1995: 123(5): 604–613.

[26] Mosalam K.M., White R.N. and Gergely P. “Static response of infilled frames using quasi-static experimentation” Journal of Structural Engineering (ASCE) 1997: 123(11): 1462–1469.

[27] Pinto A. V., Varum H. and Molina J., “Experimental assessment and retrofit of full-scale models of existing RC frames” Proc. 12th ECEE, London.

[28] PREC8 – Prenormative research in support of Eurocode 8 and ECOEST – European consortium of Earthquake shaking tables [1996], Experimental and numerical investigations on the seismic response of RC infilled frames and recommendations for code provisions, Report 6, Editor: M.N. Fardis

[29] PREC8 – Prenormative research in support of Eurocode 8 and ECOEST – European consortium of Earthquake shaking tables [1997], Numerical investigations on the seismic response of RC frames designed in accordance with Eurocode 8, Report 7, Editors: E.C. Carvalho and E. Coelho.

[30] Priestley, M. J. N. and Calvi, G. M. [1991], “Towards a capacity-design assessment procedure for reinforced concrete frames”, Earthquake Spectra, 7-3, pp. 413-437.

[31] Saneinejad A & Hobbs B. “Inelastic design of infilled frames” Journal of Structural Engineering (ASCE) 1995: 121(4): 634–650.

[32] Shing, P. B. and Mehrabi A. B. “Behaviour and analysis of masonry-infilled frames”, Prog. Struct. Engng Mater., 2002; 4:320–331.

[33] Stafford Smith B. “Behaviour of square infilled frames” Journal of the Structural Division (ASCE) 1966: 92(1): 381–403.

[34] Stafford Smith B. “Lateral stiffness of infilled frames” Journal of the Structural Division (ASCE) 1962: 88(6): 183–199.

[35] Zarnic R. and Gostic S. “Non-linear modelling of masonry infilled frames” Proc. of the 11th European Conference on Earthquake Engineering, Paris, France.1998.

[36] Zarnic R and Tomazevic M. “Study of the behaviour of masonry infilled reinforced concrete frames subjected to seismic loading”. Proc. of the 7th International Conference on Brick Masonry, Australia. 1985: 1315–1325.

Date post:	14-Mar-2018
Category:	Documents
Upload:	vokhuong
View:	216 times
Download:	1 times

SEISMIC PERFORMANCE OF MASONRY-INFILLED · PDF fileSEISMIC PERFORMANCE OF MASONRY-INFILLED...

Documents