Seismic Performance of Reinforced Concrete Walls Designed for Ductility

Abstract

Following the 2010/2011 Canterbury earthquakes in New Zealand, unexpected failure modes were observed in reinforced concrete (RC) structural walls, such as local buckling of longitudinal reinforcement, global buckling of the wall end section, and crushing of concrete at the wall end regions. In response to these observations, several amendments were made to the New Zealand concrete structures standard; however, the performance of RC walls designed to these provisions at ultimate limit state and at preceding damage states remained uncertain. To investigate the damage progression and deformation capacity of RC walls designed to modern structural code provisions, an experimental program was undertaken on four walls designed with varying end region detailing and axial load. The tests verified that excellent ductility can be achieved using the wall design provisions. The results of this study were used in conjunction with existing RC wall test data to develop a deformation capacity model for the assessment of walls in existing buildings. It was found that deformation capacity of slender walls is primarily a function of end region compression demand and reinforcement detailing. The resulting deformation limits are demonstrated to be more rational than those in existing standards or assessment guidelines and were more consistent with empirical data. To estimate the occurrence probability of damage states that precede wall failure, damage state fragility functions were developed based on reported damage progression in previous wall tests. The fragility functions were developed using local demands in the wall, which ensured that several wall design and demand characteristics were accounted for. A numerical modelling approach is developed to estimate local wall demands that are not typically reported in test data. Normalised moment demand, average concrete strain and average curvature ductility are determined to be the most appropriate parameters to model the fragility of low, moderate and severe damage states, respectively. The utility of the proposed deformation capacity model, wall modelling approach and damage state fragility functions is demonstrated through a case study analysis of an archetype wall building located in Wellington, New Zealand. Satisfactory performance was observed for serviceability and design level earthquakes; however, the collapse probabilities at a maximum considered earthquake event were higher than expected.