Probabilistic risk analysis of structural impact in seismic events for linear and nonlinear systems

SUMMARYEarthquake‐induced structural pounding is a significant risk for buildings in densely populated cities and for other applications, such as adjacent bridge deck sections of long‐span bridges. Although significant research has been undertaken into structural pounding, this manuscript focuses on a parametric, probabilistic study of the likelihood of structural pounding occurring while also quantifying the likelihood of that impact leading to displacement increase in the structures involved. Additional metrics are also used to define minimum gap ratios (GRs) (expressed in relation to structure spectral displacements) that are required to avoid impact occurring, or to limit the magnitude of displacement increases that result from impact. The parametric study also includes the influence of structural nonlinearities, such as the coefficient of restitution during impact and the influence of column inelasticity.Results indicate that the probability of impact between structures increases as the GR decreases and structural period differences increase. The same general behavior was noticed for the risk of increasing displacement because of impact. An increase in the coefficient of restitution leads to an increase in the risk of increased displacement response. The inclusion of nonlinear column behavior leads to a drop in both probability of impact occurring and the risk of increasing displacement response, as compared with the same structures without impact occurring. The results presented also indicate that the likelihood of an increase in displacement response because of structural pounding is independent of the gap ratio for smaller GR values. Above this value of GR, the reduction of impacts leads to a smaller probability of a displacement increase. Finally, a probabilistic design metric is presented using these results that assesses risk of significant displacement increase because of impact above a set design limit to the gap ratio.Overall, this study utilizes a simplified, two degree‐of‐freedom structural model, but presents a parametric study to investigate the probabilistic risk analysis of building pounding and the associated effect on building response with a focus on use in design or the retrofit of structures. Copyright © 2014 John Wiley & Sons, Ltd.