Both pseudo-static tests. (Shi & D'Ayala, 2006; Shi et al., 2008) and extensive on-site post earthquake observation have shown that under lateral action, masonry walls are more vulnerable to out of plane than in-plane failure, even for shape ratio H/L < 1, and slenderness ratio H/t < 11, as it is the case for this series of tests. This was confirmed by the observation of both the recorded videos of the shaking and the comparison of overall displacement of in-plane and out-of-plane instrument for any given motion test. However differently from the monotonic static tests carried out on single walls with wings, the inversion of motion caused by the sinusoidal wave results in a much stronger interaction at the corner between parallel and orthogonal walls with effects on the two sets of walls which will be further discussed in detail in the following subsection where results are presented separately for each series of tests.
In Figure 4 maximum amplification of motion and dissipated energy for series of test with increased frequency and constant amplitude of 12 mm and constant acceleration of 1.53 m/s2 are shown for the two coordinate directions of motion. The y direction relates to the out of plane motion of the longer façade. This shows a peak amplification of 2 for 2.0 Hz, with a distribution very similar to an amplification spectrum. For the shaking in x, peak response is associated to an input frequency of 1.9 Hz.and a value of 1.5, with a second peak at 3.6 Hz. Assuming a simple 1 degree of freedom oscillator as basic model for the behaviour of the walls would yield values of stiffness 29.5 KN/m and 15.7 KN/m respectively, contrary to expectations.
This can be explained by the fact that the floor structure afforded better friction restraint in the y direction shaking, resulting in overall greater stiffness.
In the case of constant input acceleration, and hence constant input energy, it is observed that for increasing input frequency damage and collapse occur for a smaller number of cycles. This phenomenon is summarised clearly in Figure 4b, where the total energy dissipated in each test is measured against the value of input frequency.
Figures 5 and 6 show that similar portion of the walls are excited in both directions for each peak amplification case. In the photos the panel parallel to the motion are shown. The thick lines at the edge of the panels shows the portion of this that participate to the out of plane motion of the orthogonal walls while the thinner
Was this article helpful?