Implementation of artificial neural networks to automate spectral-analysis-of-surface-waves method
Abstract
The spectral-analysis-of-surface-waves (SASW) method is a nondestructive testing method based on the dispersive characteristic of seismic surface waves in a layered system. The method can be used to determine the stiffness parameters of pavement layers. One of the more complex aspects of the SASW method is the inversion procedure. The artificial neural networks have been recently advocated for this purpose. The general contention has been that the artificial neural networks (ANN) can be effectively used to establish the a priori information for a more robust and rapid formal inversion process. In this research, the results from the evaluation of a number of software and training strategies to completely substitute the inversion process with the trained ANN models are presented. An extensive synthetic database, covering a wide range of pavement profiles, was generated based on the forward modeling of the SASW method. Several neural network software packages with various capabilities were evaluated. Numerous models of different neural network types and architectures based on different inputs were developed to predict the stiffness parameters of pavements. (Abstract shortened by UMI.)
Subject Area
Civil engineering
Recommended Citation
Shirazi, Hamid, "Implementation of artificial neural networks to automate spectral-analysis-of-surface-waves method" (2005). ETD Collection for University of Texas, El Paso. AAI1427702.
https://scholarworks.utep.edu/dissertations/AAI1427702