The effects of near-fault pulsed and non-pulsed ground motions on structural responses differ significantly

yet their impacts on site ground motion parameters remain unclear. Therefore

a centrifuge shaking table model test of a dry sand site under 50g (g represents the acceleration caused by gravity) centrifugal acceleration was designed and conducted. The model site consisted of Fujian standard fine sand with a relative density of 50%. The pulsed acceleration record from Saratoga Aloha Ave station and the non-pulsed record from Capitola station

both within the near-fault range in the east-west direction of the Loma Prieta earthquake

were selected as inputs for the shaking table. An accelerometer vertical array and surface settlement gauge were placed at the center of the site

while horizontal displacement gauges were fixed on the lateral outside of the model box. This setup was used to compare the response differences of the dry sand site under near-fault pulsed and non-pulsed ground motions and to evaluate the correlations between site ground motion parameters and the peak acceleration values of the input ground motion. The model test results show that compared to non-pulsed ground motions

pulsed ground motions significantly reduce the amplification factors of peak acceleration but increase the final surface settlements of the site. For peak horizontal displacement

pulsed ground motions increase the peak displacements in the deep soil layer (> 15 m) but significantly reduce them in the shallow layer (< 5 m). Statistical results reveal high correlations between site seismic response parameters (amplification factors of peak surface acceleration

surface settlement

shallow peak horizontal displacement

Arias intensity) and the input peaks acceleration of ground motion

with all fitting coefficients R² greater than or equal to 0.97. These results provide a reliable experimental basis for further numerical simulations to study the impact of near-fault pulsed ground motions on site seismic response.