SIMPLIFIED ANALYTICAL MODELS FOR EVALUATING RETROFITTED REINFORCED CONCRETE SHEAR WALLS

Reinforced Concrete (RC) rocking shear walls have been studied over the past 20 years as seismic-resistant components in buildings for minimizing damage and shortening downtimes after an earthquake. Although design procedures, seismic response and macro modeling of such systems have been evaluated for newly built structures, uses of RC rocking walls for retrofit of existing non-ductile, cantilever RC shear walls have received limited attention. Past earthquakes showed that non-ductile, cast-in-place shear walls may experience undesirable failure modes, rendering them unusable. This study investigates rocking as a retrofit method for existing non-ductile cantilever RC shear walls through analytical modeling and evaluates simplified analytical modeling techniques for their suitability for efficient nonlinear response history analyses. Detailed 3D finite element modeling, fiber modeling, and lumped plasticity based modeling techniques are compared in terms of global response parameters. The retrofit method follows the concepts of weakening and self-centering. Weakening involves saw-cutting wall base across concrete and/or across vertical reinforcing bars. Self-centering is provided by added vertical post-tensioning. The overall goal of the retrofit is to replace the non-ductile response of the original wall by the rocking response, leading to minimized damage to the building after an earthquake. To implement the retrofit method, an RC shear wall was designed so that it did not meet the requirements of ACI 318-14 on vertical spacing of confining bars in the boundary elements, extension of the confining reinforcement into the foundation and spacing of transverse reinforcement in the web. The wall was then retrofitted using the concepts of weakening and self-centering. The RC shear wall was analyzed under cyclic displacements up to 2% drift ratio. Detailed and simplified finite element modeling of RC walls were conducted using commercial and open-source (OpenSees) analysis software, respectively. Both pre-and post-retrofit walls were simulated. Simplified models used a lumped plasticity approach with a monolithic beam analogy and a distributed plasticity approach with fiber elements. Modeling approaches were compared in terms of energy dissipation, lateral strength, residual displacement, and secant stiffness. Detailed 3D finite element model was used as a baseline for comparison. Both lumped plasticity and fiber models were able to capture initial stiffness and lateral strength with reasonable accuracy. On the other hand, lumped plasticity model overestimated energy dissipation capacity of the wall. Residual displacement prediction of fiber model was closer to that of detailed 3D finite element analysis.