On the code-compliant verification of seismic damage to non-structural masonry infills

Drift-sensitive seismic damage to the non-structural masonry infills of framed buildings is one of the most detrimental effects of relatively weak earthquakes. Current codes prescribe relevant verification procedures for building design. However, these codes permit assessment of interstorey drift ratios based on analysis of bare frames, without taking into account the great influence of masonry infills on drift demands. In this paper, the degree of correlation between drift ratios of bare and infilled frames is investigated by multivariate statistical methods, in particular, that of canonical correlation. A number of frames with and without infills are analysed, using both artificial and recorded accelerograms, to compare the drift demands from a statistical standpoint. The analysed frames differ in number of storeys, stiffness of infills, ductility class and ground type. Significant correlation appears only with variation in seismic intensity, which is expressed as peak ground acceleration and spectral acceleration. The seismic intensity being fixed, as peak ground acceleration is in the code check, the correlation is too weak to predict the drift demands of an infilled frame from those of its bare counterpart. The null hypothesis of statistical independence of the drifts of the bare and infilled frames cannot be rejected at usual significance levels. The results contradict the code assumption. Modelling the non-structural masonry infills to compute the drift demand should be required in building codes.