Identification of the Controlling Mechanism for Predicting Critical Loads in Elastomeric Bearings

Abstract

Assessing the stability of individual isolators is an important consideration for the design of seismic isolation systems composed of elastomeric bearings. A key component for the stability assessment is the prediction of the critical load capacity of the individual bearings in the laterally undeformed (service) configuration and at a given lateral displacement (seismic). The current procedure for estimating the critical load capacity of an elastomeric bearing at a given lateral displacement, with a bolted connection detail, uses a ratio of areas to reduce the critical load capacity from that in the laterally undeformed configuration, referred to as the reduced area method. Although the reduced area method provides a simple means for the estimate, it lacks a rigorous theoretical basis and is unable to capture the trends observed from experimental data. In this study, the capability of two analytical models for predicting critical loads and displacements in elastomeric bearings is evaluated by comparison with data from past experimental studies. A global variance-based sensitivity analysis is performed on the analytical model showing the best predictive capability to identify the model parameters to which the model prediction is most sensitive. The results of the sensitivity analysis demonstrate that the model prediction is most sensitive to the properties that control the nonlinear behavior of the rotational spring for lateral displacements greater than approximately 0.6 times bearing diameter/width. This finding suggests that the stability of elastomeric bearings at large lateral displacements is controlled by the transition from the yield moment to the ultimate moment in an individual rubber layers. A modified analytical model is proposed based on the results of this sensitivity analysis. The predictive capability of the more parsimonious modified model is shown to be similar, if not improved, by comparison to the original model.