strength of material, Sika® CarboDur S, in the direction of fibers amounts to 3000 MPa, and the modulus of elasticity to 165000 MPa. Before gluing the strips, the surface of masonry has been thoroughly cleaned and penetrated with primer. Original epoxy adhesive material, SikaDur, has been used to glue the strips on the masonry.

2.3 Seismic load and instrumentation of models

The shape of the ground acceleration time history, used to control the shaking table motion, corresponded to the 24 seconds long strong phase of the N-S component of the ground acceleration record, recorded at Petrovac during the Montenegro earthquake of April 15, 1979. Maximum measured ground acceleration was 0,43 g. The actual model earthquake, prepared for testing the so called complete models, was 12 seconds long, and had the same maximum ground acceleration as was the case of the actual acceleration record. The shaking table motion during test run R100 represented such an earthquake (100% of intensity). Since the models have been made of materials with strength characteristics similar to the prototype (simple models), the actual model earthquake represented 48 sec long earthquake (St = SL = 4) with maximum ground acceleration 0,11 g (Sa = 1/SL = 0, 25).

Shaking table displacements in each successive test run have been scaled from 5% to 350% of those of the model earthquake (test runs R005 to R350, respectively). All models have been tested with the same sequence of seismic excitations with increased intensitiy of motion in each successive test run, the characteristics of the model earthquake did not influence the observations. Maximum accelerations and displacements of the shaking table motion obtained in

Figure 6. Typical response spectra of shaking table motion.
Figure 7. Instrumentation of models.

each test run are given in Table 1, whereas the typical response spectra are shown in Figure 6.

All models have been instrumented with a set of displacement transducers and accelerometers (Figure 7), fixed to the models at the level of floors. The missing live load at the levels of floors has been modelled by means of concrete blocks of adequate mass, which have been fixed to wooden joists with steel bolts so that the in-plane rigidity of floors has not been significantly affected. In order to prevent damage to instruments and shaking table at the moment of collapse, concrete blocks have been loosely hanged on the crane. All models have been oriented so that the direction of shaking table motion coincided with longer dimension of the

model. In other words, seismic loads acted in the direction of load-bearing walls, pierced with window and door openings.

3 TEST RESULTS 3.1 Failure mechanism

The control model M1 exhibited typical behaviour of old masonry buildings with wooden floors without wall ties: in the beginning of tests when subjected to low intensity earthquake ground motion, the behaviour was monolithic. However, with increased intensity of shaking, vertical cracks developed in the upper part of the model. As a result of separation of walls, the upper storey of the model disintegrated in the subsequent test runs and collapsed (Figure 8).

The tests of model M2 have shown that the damp-proof course in the form of a simple PVC sheet placed in the mortar in the bed joint cannot be considered as seismic isolating device. Although the compressive stresses in the walls with installed damp-proof course were low, the measurements have indicated that neither sliding along the damp-proof course took place nor rocking motion of the upper part of the building has been observed. The walls in the upper storey disintegrated and the storey collapsed at the same intensity of excitation as was the case of control model M1 (Figure 9).

Although improved behaviour ofmodel M3, placed on rubber seimic isolators, has been expected, model M3 exhibited practically the same poor behaviour as

Figure 10. Mechanism of collapse of non-strengthened and isolated model M3.

non-isolated models M1 and M2. However, a slight difference in the sequence of damage propagation has been observed. Whereas damage propagated gradually in dependence on intensity of motion in subsequent test runs in the case of models M1 and M2, the collapse of model M3 was sudden, without cracks occuring during the previous test runs (Figure 10).

Figure 11. Damage to CFRP laminate strengthened, non-isolated model M4 at the end of shaking table tests.
Figure 12. Unsignificant damage to CFRP laminate strengthened, isolated model M5 at the end of shaking table test.

The seismic behaviour ofboth models strengthened with CFRP laminate strips, however, was significantly improved. They did not suffer severe damage or collapse even when subjected to ground motion with accelerations, which by more than three times exceeded the accelerations measured during the testing of non-strengthened models (Figures 11 and 12). Since the capacity of shaking table has been reached

Figure 13. Detached anchor plate caused rocking of model M4. Vertical strip buckled at the bottom of the model.

and the output motion already distorted, tests had to be terminated at that point. In the case of non-isolated model M4 the anchor bolts, by means of which steel anchor angles of vertical strips have been fixed to the foundation slab, pulled out (Figure 13) and the model started rocking on the foundation slab. Consequently, masonry crushed at the corners and severe cracks occurred in the lintel parts of the walls.

In the case of model M5 on seismic isolators, one of the isolators detached (Figure 14). However, almost no damage has been observed in the model's walls. It has to be noted, that also in the case of model M4 no structural damage has been observed before the pulling out of anchor bolts. By comparing the results of tests of CFRP laminate strengthened and non-strengthened model walls, it seems that, in the particular case studied, this is the result of confining the model structure with horizontal and vertical CFRP strips, and not the result of diagonally placed strips on the wall piers. However, additional measurements should have been carried out in order to confirm this observation.

The changes in dynamic characteristic of the tested models, measured before the tests and after each subsequent test run, are presented in Table 2. The values of the first natural frequency of vibration f and coefficient of equivalent viscous damping Z (in % of critical damping), have been determined by hitting the model with impact hammer and analyzing the

Figure 14. model M5
Table 2. First natural frequency of vibration f (in s-1) and coefficient of equivalent viscous damping Z (in % of critical damping) measured on the models before the beginning of shaking table tests and after characteristic test runs.


Was this article helpful?

0 0
Renewable Energy Eco Friendly

Renewable Energy Eco Friendly

Renewable energy is energy that is generated from sunlight, rain, tides, geothermal heat and wind. These sources are naturally and constantly replenished, which is why they are deemed as renewable.

Get My Free Ebook

Post a comment