Modeling of Nonlinear Hysteresis in Lead Rubber Bearings With Data‐Driven Shape Parameter Identification

Lead Rubber Bearings (LRBs) are widely used in seismic isolation systems due to their excellent energy dissipation and stability under dynamic loading. However, under extreme conditions, LRBs undergo large strain stiffening, degradation, and strength reduction due to lead core heating. These factors pose challenges for accurately modeling their nonlinear behavior. The commonly used Bouc–Wen model in engineering struggles to capture the complex hysteretic characteristics of LRBs. This study develops an improved Bouc–Wen model that effectively captures large strain nonlinearities. In addition, the model incorporates a strength reduction factor to account for lead core heating and an adjustment coefficient to characterize initial hardening. A data‐driven approach is further introduced to intelligently predict shape parameters, improving computational efficiency and practical applicability. The proposed model is validated through dynamic compression‐shear tests under various loading protocols, particularly those involving large deformation and long‐duration loading. Comparisons based on hysteresis curves, equivalent stiffness, and equivalent damping ratio demonstrate that the proposed model more accurately reproduces LRB behavior. It effectively captures key nonlinear effects, leading to improved accuracy in seismic isolation analysis and design.