Selection of Ground Motions for Response History Analysis

The evolution in computational power and the parallel processing capabilities of modern engineering software make nowadays the use of complicated structural analysis methods an attractive alternative for the design and assessment of structures. In contrast to the past, when the elastic static analysis was almost exclusively used for the seismic design of structures, the state of practice has progressively moved toward dynamic-elastic, nonlinear-static (i.e., single mode or multi-modal “pushover”), and even nonlinear response history analysis. The latter, capturing more efficiently the hierarchy of failure mechanisms, the energy dissipation, the force redistribution among the structural members, and contact issues (such as gap, impact, sliding, and uplift), is deemed preferable in cases of significant material or geometrical nonlinearities and, as such, is used for the design of seismically isolated buildings and bridges or the assessment of existing structures with various degrees of damage. Elastic response history analysis is also extensively used, primarily for structures whose response is dominated by higher modes (mostly tall and irregular buildings and towers) or structures of high importance that are typically designed to remain elastic even for long returnperiod earthquake intensities (i.e., industrial facilities, power plants, dams, critical administrative buildings, etc). In all cases, the main task of the design procedures is to achieve more predictable and reliable levels of safety and operability against different levels of seismic intensity, a framework known as performance-based design and assessment. Despite the above major advances made in terms of structural analysis, the reliability of the analysis output and the subsequent structural performance prediction strongly depend on the decisions made for the selection of the seismic input which is used as ground excitation. Research has shown that among all possible sources of uncertainty stemming from structural and soil material properties, the modeling approximations, and the design and analysis assumptions as well as the earthquake-induced ground motion, the latter has by far the highest effect on the variability observed in the structural response (Elnashai and McClure 1996; Padgett and Desroches 2007; Shome et al. 1998). Therefore, the selection of a “reliable” suite of earthquake ground motions constitutes an important prerequisite for the reliability of the structural analysis procedure as a whole. Along these lines, numerous computational methods and tools have been developed for (a) selecting suites of earthquake records from available strong ground motion record databases