Wall-to-Floor Interaction in Reinforced Concrete Buildings

Abstract

The poor performance of several ductile reinforced concrete (RC) walls in the 2010-2011 Canterbury earthquakes was attributed to an increase in the levels of axial load that the walls sustained during the earthquakes. Among other reasons, it was proposed that that the magnification in the axial loads was a consequence of the wall-to-floor interaction and wall elongation. To address this hypothesis in conventionally-reinforced flexure-dominant buildings, this thesis investigates the effects of wall-to-floor interaction and wall elongation in the design loads of RC walls, and, in addition, it analyses the load redistribution between building components triggered by wall-to-floor interaction. The investigation was conducted using monotonic and cyclic pushover analyses of non-linear three-dimensional numerical models in OpenSees. The models were based on a case-study building selected after a comprehensive typology analysis performed to three major cities in New Zealand. To ensure that wall elongation was appropriately simulated, a numerical approach based on fibre-based distributed-plasticity beam-column elements was validated against the elongation determined from previously tested RC walls. The analyses were performed to the case-study building and two variations based on common practitioner’s assumptions. In addition, a parametric analysis was also carried out to examine the influence of several parameters in the building response. It was found that the wall-to-floor interaction and wall elongation produced a significant increase in the building overturning capacity and stiffness. As a result, the shear demand in the wall increased up to 65% and the axial load up to 54%. The columns were subjected up to two times the gravitational compression and reached net tensile demands as high as the tensile capacity of the columns. Even when the distance between the wall and columns was increased, the effect of the wall-to-floor interaction and wall elongation in the elements was still considerable. In conclusion, load demands can be underestimated if wall-to-floor interaction and wall elongation are not accounted for in the design of RC structures where floor-framing actions can occur between walls and columns. This has the potential to violate the strength hierarchy of capacity designed structures and lead to unexpected failure modes as those seen in the field.