The University of Massachusetts Amherst
University of Massachusetts Amherst

Search Google Appliance

Links

Spectral analysis of ground-acceleration-based testing

TitleSpectral analysis of ground-acceleration-based testing
Publication TypeJournal Article
Year of Publication2009
AuthorsHo CL, Hyslip J, Li D
JournalTransportation Resource Record
Volume2117
Start Page50
Pagination50-56
Date Published12/2009
Abstract

Spectral analysis of downhole acceleration data is used to identify changes in elastic properties of track substructure. The spectral analysis is based on the power spectral density (PSD) as an indicator of frequency content. This method was developed and tested at the Transportation Technology Center, Inc., in Pueblo, Colorado. The system provides an efficient means for monitoring track bed to predict settlement and deformation conditions that would require maintenance. The method relies on a triaxial piezoelectric accelerometer inserted into a permanently installed inclinometer casing installed into the subgrade. Periodic measurement of the dynamic ground response allows for the identification of changes in the small strain elastic properties of the substructure layers. The methodology uses conventional downhole seismic shear wave velocity testing. For the purpose of comparison, test sections were constructed with clean and fouled ballast. Shear wave velocities indicated the differences in material properties. PSDs of the signals were analyzed to determine variation resulting from differences in soil conditions. PSDs were calculated at different intervals of loading. With increased loading, settlement of the subtrack soils occurs. The densification can be seen as an increase in the predominant frequency of the PSD. The PSD from the section with the clean ballast had a higher predominant frequency than the PSD from the section with the fouled ballast. A decrease in predominant frequency over time would indicate the decrease in stiffness and the potential of increased fouling of the ballast with fines. Changes in the frequency content would indicate changes in elastic response.

URLhttp://dx.doi.org/10.3141/2117-07
DOI10.3141/2117-07