Center for Research in Energy Systems Transformation (CREST)

In this paper, the tools of the stochastic dynamic analysis are adopted for Performance-Based Earthquake Engineering (PBEE). The seismic excitation is defined through a evolutionary Power Spectral Density compatible with the response spectrum given by mandatory codes. In this way, the performance-based design is applied considering the excitation coherent with the codes. Inside the framework, the seismic fragility curves are determined through the Kernel Density Maximum Entropy Method (KDMEM), recently proposed by the authors. It is a novel statistical method capable to reconstruct the seismic fragility curves, including the tails, from a small number of code-conforming artificial ground motions. Moreover, KDMEM is based on the Maximum Entropy (ME) principle and it provides the least biased probability distribution given the available information. Comparison between stationary and nonstationary artificial accelerograms is analyzed, and the corresponding model uncertainty discussed. KDMEM provides also credible bounds of the uncertain performances, which is beneficial for risk-informed decisions. The proposed formulation does not require the selection of a suitable set of ground motions. Accordingly, it can be adopted for optimal design in current engineering practice. Therefore, it fills the gap between the classical code-conforming designs and the enhanced performance-based designs.