Proposed multivariate predictive approach of masonry wall seismic performance

This chapter outlines the implementation of a statistical multivariate technique referred to as the Projection to Latent Structure (PLS), to predict the maximum displacement as a key masonry wall performance indicator. It utilizes the wall geometrical and material characteristics within a statistical Multivariate Data Analysis platform that adopts the PLS technique for data analysis. Masonry wall systems comprise one of the most common type constructions in urban areas for low- and mid-rise buildings. In terms of potential seismic risk, there is a perception that masonry construction offers minimal ductility and are particularly vulnerable to seismic events. The shear resistance of a masonry shear wall is attributed to shear reinforcement, dowel action of vertical reinforcement, applied axial stress and aggregate interlock. All the geometrical and material characteristics such as the aspect ratio, axial stress level, shear and vertical reinforcement ratios, reinforcement yield strength, and masonry compressive strength vary within the database.