Neural Network Modeling of Seismic Events and Soil-Pile Interaction

This work aims at predicting the seismic response of soil and pile using artificial neural network (ANN) as a tool of function approximation.The seismic response of soil is a function of characteristics of earthquake source, the distance traversed by seismic waves through the earth body, and the local site geology below the ground surface. In addition, the presence of pile in soil alters the seismic waves and the pile in turn is stressed. Two important phenomena are modeled in this research: (i) the prediction of seismic response of ground surface in terms of peak ground acceleration (PGA), peak ground velocity(PGV), and peak ground displacement(PGD), and (ii) the kinematic response of piles which includes pile bending moments and the foundation input motion to the superstructure.