Reliability-based design optimization of adaptive sliding base isolation system for improving seismic performance of structures

The reliability-based design optimization (RBDO) of base-isolated structures by means of the state-of-the-art methods has been confronted with the high computational cost because the base isolation system often exhibits intrinsic nonlinearities and has to be modelled using a number of parameters in practice. In this regard, the present study aims at developing an efficient scheme for the RBDO of base isolation systems by integrating the probability density evolution method (for global reliability solution) and the variance-based sensitivity analysis (for design parameter reduction).To attain an adaptive seismic mitigation, the newly developed sliding implant magnetic bearings are employed for constituting the base isolation system. It is shown that although the deployment of sliding implant-magnetic bearings allows a safer structure, the base isolation system still suffers from an unexpected failure probability of deflection under the rarely-occurring earthquake. As a result, the design optimization for strengthening the reliability of the base isolation system needs to be carried out. For identification of the critical design parameters to be optimized, the variance-based sensitivity analysis is performed, showing that the parameters of the exponential function that represents the Coulomb sliding friction associated with the sliding implant-magnetic bearing are mostly of concern. Furthermore, the RBDO of the base isolation system is carried out by employing the genetic algorithm. It is revealed that the global reliability of the base-isolated structure after optimization gains a significant improvement compared to that before optimization.

Automated summary

The study proposes an efficient reliability-based design optimization scheme for base isolation systems by integrating the probability density evolution method and variance-based sensitivity analysis. Critical design parameters for optimization are identified through sensitivity analysis, with parameters associated with the sliding implant-magnetic bearing being of most concern. Genetic algorithm is employed for RBDO, showing a significant improvement in global reliability of the base-isolated structure post optimization.