FOSID: a fractional order spectrum intensity for probabilistic seismic demand modeling of extended pile-shaft-supported highway bridges under liquefaction and transverse spreading
Xiaowei Wang,Abdollah Shafieezadeh,Jamie Ellen Padgett
DOI: https://doi.org/10.1007/s10518-021-01082-4
IF: 4.556
2021-04-01
Bulletin of Earthquake Engineering
Abstract:An appropriate seismic intensity measure (IM) for response prediction is central to reliable probabilistic seismic demand modeling of structures and subsequently, risk and resilience quantification. Bridges in liquefiable and laterally spreading ground may undergo nonlinear responses with large uncertainties when subjected to earthquakes. These issues often lead to low-confidence demand models based on traditional IMs. Fractional order IMs have shown the potential to yield improved demand models in recent studies. To further increase confidence in demand models, this study proposes Fractional Order Spectrum intensity considering Integral period and Damping ratio (named <i>FOSID</i>). The viability of <i>FOSID</i> for use in probabilistic seismic demand modeling of structures is evaluated in the context of extended pile-shaft-supported bridges against liquefaction-induced lateral spreading. The performance of <i>FOSID</i> is systematically assessed by comparisons to an existing fractional order spectrum intensity (<i>SI</i><sub><i>r,α</i></sub>), Housner intensity (<i>HI</i>)—an optimal traditional IM for these structures, and the average spectral acceleration (<i>Sa</i><sub><i>avg</i></sub>)—a state-of-the-art non-fractional-order IM. Multiple metrics for characterizing an optimal IM are adopted, including practicality, efficiency, proficiency, sufficiency, and relative sufficiency. Optimal variables of integral period, damping ratio, and fractional order for <i>FOSID</i> are identified for different demand parameters such as peak and residual column-drift-ratios. Results show that <i>FOSID</i> is generally more practical, efficient, proficient and sufficient than <i>SI</i><sub><i>r,α</i></sub>, <i>HI</i> and <i>Sa</i><sub><i>avg</i></sub>. In particular, <i>FOSID</i> significantly outperforms <i>HI</i> by improving the proficiency by nearly 40% and 20% for the peak and residual column-drift-ratios, respectively. With respect to <i>SI</i><sub><i>r,α</i></sub>, <i>FOSID</i> improves the proficiency by 20% and 15% on average. When compared with <i>Sa</i><sub><i>avg</i></sub>, such improvements are as large as 13% and 24% on average for the peak and residual column-drift-ratios, respectively.
geosciences, multidisciplinary,engineering, geological