Events Calendar

Astani CEE Ph.D. Seminar

Speaker: Armen Derkevorkian , CE Ph.D. Candidates

Talk Title: Computational Models for Response Prediction and Change Detection in Nonlinear Soil-Foundation-Superstructure Systems

Abstract: Data-driven reduced-order computational models are proposed to predict the response of complex linear and nonlinear soil-foundation-superstructure (SFS) systems to various non-stationary random excitations. The proposed models are further investigated to establish a general change detection scheme that can be applied on broad range of structural systems. The reduced-order models are developed by incorporating trained neural networks within an ordinary differential equation (ODE) solver and dynamically predicting the response (i.e., displacement and velocity) of the SFS systems to various earthquake records. Then, the models are used for system identification and change detection in the SFS systems. Effects of input delays are investigated in improving the fidelity of the trained networks. The detected changes in the systems are quantified through a measure of a normalized error index. The developed models are tested and validated using experimental data from three relatively large-scale SFS systems. The three systems under consideration consist of identical superstructures with: (a) fixed base; (b) box foundation; and (c) pile foundation. The three SFS systems were developed and experimentally tested at Tongji University. Excitations with various intensity levels were applied on each system to obtain both linear and nonlinear response. It is shown that the general neural network procedure adopted in this paper provides a robust nonlinear model that is reliable for computational studies, as well as furnishing a robust tool for detecting and quantifying inherent change in the target structure.

Location: John Stauffer Science Lecture Hall (SLH) - 102

Audiences: Everyone Is Invited

Contact: Evangeline Reyes