Implications of Modeling Assumptions on the Loss Estimation for Shear Wall Buildings

The ability to accurately predict engineering demand parameters necessary for estimation of building losses is essential for development and implementation of resilient-based seismic design. Currently available macroscopic nonlinear modeling approaches for reinforced concrete (RC) structural walls can be classified into two groups: (1) models that capture interaction between axial/flexural and shear behavior and (2) modeling approaches with uncoupled shear and flexural responses. Uncoupled modeling approaches are widely implemented in computer programs for structural analysis and showed to be effective tools for simulation of flexural behavior of RC walls. However, their inability to account for shear-flexural interaction could result in underestimation of axial compressive strains in wall boundaries and overestimation of the wall lateral load capacity. This study investigates the impact of the modeling approaches and assumptions used for simulation of seismic behavior of RC walls on estimation of seismic repair losses considering a five-story building designed to current code provisions for a site in downtown Los Angeles under three seismic hazard levels. Presented results suggest that the impact of using coupled versus uncoupled modeling approaches for structural walls is significant for interstory drift estimations at the first building level, slight at the upper levels, and moderate for peak floor acceleration estimations. These differences in the structural responses have marginal impact on the total estimated repair losses for frequent and rare earthquakes, where coupled wall model predicted highest structural losses and smallest nonstructural losses. However, for very rare earthquake events, there is a notable difference between the three models where predicted median repair losses using coupled wall model are 20% to 35% larger than the ones predicted using uncoupled models.