New improved scaling method for nonlinear dynamic analysis of structures under near-fault ground motions

Scaling ground motions to a specific level of intensity is one of the challenges in non-linear dynamic analyses of structures. Recent researches have demonstrated that using S-a (T-1) as the Intensity Measure (IM) for scaling near-fault ground motions may introduce a large variability in estimated seismic demands. Considering the need to provide more accurate estimation of seismic demands by using a smaller number of records, this paper investigates the effectiveness of a large amount of combinations of spectral responses to find an optimal combination that can considerably reduce the variability in seismic demands. For this purpose, Incremental Dynamic Analyses (IDAs) are performed for generic frames with 3, 6, 9, 12 and 15 storey subjected to 40 pulse-like earthquake records which are rotated to the fault-normal direction. Statistical evaluation of the IDAs results reveals that the optimal combination of spectral displacements provide much superior results compared to that of spectral velocities. With respect to the structural characteristics, three optimal combinations of spectral displacements are proposed to be used as the improved IM as follows: (1) combination of S-d(T-1) and S-d(1.2T(1)) for short period frames; (2) combination of S-d(T-1), S-d(1.2T(1)) and S-d(T-2) for moderate period frames; (3) combination of S-d(T-1) and S-d(T-2) for relatively long period frames. It is shown that the newly proposed IM can reduce the overall variability in seismic demands of all frames by a relative amount of 30 percent compared to S-a(T-1) on average. Additionally, with respect to the importance of collapse capacity prediction in earthquake engineering, it is also shown that this improved IM can predict the collapse capacities of the frames with the variability that is generally decreased by a relative amount of 22 percent compared to S-a(T-1).