Selection of pulse-like ground motions with strong velocity-pulses using moving-average filtering

Pulse-like ground motions with strong velocity-pulses are of special concern for engineers when designing or assessing structures built adjacent to active tectonic faults. This study intends to propose a new approach that makes use of the moving-average filtering (MAF) for selecting ground motions with strong velocity-pulses. Specifically, MAF is employed to filter out the high-frequency content of ground motions, so that relatively smooth waveforms could be extracted and the velocity pulse is easily captured by the peak point method. Then, the pulse-energy relative to the total energy of the ground motion is used as the pulse indicator (PI) for selecting pulse-like records. A training set containing 954 components with PGV> 20 cm/s is compiled to determine the PI threshold. The effectiveness of the proposed algorithm is verified by comparisons with previous methods. Finally, by using the proposed method it is shown that pulse-like features can be observed in a couple of orientations, and the identified velocity pulses are seen occurring at sites in the propagation direction of the fault rupture front.