Probabilistic Performance-Based Optimum Seismic Design of Seismic Isolation for California High-Speed Rail Prototype Bridge

Over the last five decades, remarkable progress has been achieved in the field of earthquake engineering, especially in the following areas: seismic design philosophy, earthquake protective systems, seismic design and performance evaluation of structures, and theory of structural optimization. The progress achieved and products developed in these areas can be integrated to develop a desired computer-aided optimum structural design framework. Accordingly, a probabilistic performance-based optimum seismic design (PPBOSD) framework is proposed and first illustrated and validated on a simplified single-degree-of-freedom (SDOF) bridge model optimized (i.e., rated) for a target seismic loss hazard curve. The feasibility and optimality of seismic isolation is investigated for a California High-Speed Rail (CHSR) prototype bridge testbed using the proposed PPBOSD framework, balancing the beneficial and detrimental effects of seismic isolation for such a bridge. Towards this goal, a three-dimensional detailed nonlinear finite element model of the CHSR prototype bridge, including soil-pile-structure interaction and rail-structure interaction, is developed in OpenSees. The seismic response of the isolated bridge is compared to that of the corresponding non-isolated bridge both in deterministic and probabilistic terms. A comprehensive parametric probabilistic demand hazard analysis is carried out to investigate the effects of the seismic isolator properties on the seismic risk of the CHSR prototype bridge. To enable the computationally intensive probabilistic seismic response analyses, a cloud-based optimization framework was used integrating cloud computing resources with the high throughput computing in PPBOSD methodology. Furthermore, some well-posed practical optimization problems are formulated and investigated for seismic isolation in CHSR bridges. In summary, the unique contributions and findings are summarized as follows : (1) A PPBOSD framework is proposed, illustrated, and validated using a nonlinear SDOF bridge model; (2) Compared to a non-isolated bridge, the seismic isolation increases the deck displacement and rail stress demands, while it reduces the seismic demand in the bridge substructure in both the deterministic and probabilistic sense; (3) A cloud-based computing platform is developed for PPBOSD to address the high computational cost; (4) The feasibility and optimality of seismic isolation for the prototype bridge is achieved using the PPBOSD framework, reaching various performance objectives considering the relevant sources of uncertainty.