Evaluation of uniaxial contact models for moat wall pounding simulations
P. J. Hughes, Gilberto Mosqueda
Abstract
Summary Moat wall pounding occurs when a base‐isolated building displaces beyond the provided clearance and collides with the surrounding retaining wall, inducing very high floor accelerations and interstory drifts. Previous studies on moat wall pounding typically employ simplified models of the superstructure, with a uniaxial contact spring used to model the entire moat wall. Consequently, researchers have developed sophisticated contact models to estimate the normal‐direction contact force that is generated during seismic pounding. This study examines how the choice in contact model affects the seismic response of a base‐isolated building subjected to impact‐inducing ground excitation. Five widely used state‐of‐the‐art contact models are summarized and implemented into an experimentally‐calibrated numerical model of a base‐isolated moment frame. Results of nonlinear dynamic time history analyses are shown in detail for one ground motion, followed by a larger parametric study across 28 near‐fault ground motions. This work shows that peak impact force and base acceleration are moderately sensitive to the choice in contact model, while upper floor accelerations and interstory drifts are practically not affected.