Degree Type

Dissertation

Date of Award

2015

Degree Name

Doctor of Philosophy

Department

Civil, Construction, and Environmental Engineering

First Advisor

David J. White

Abstract

Road roughness is a key parameter for road construction and for assessing ride quality during the life of paved and unpaved road systems. The quarter-car model (QC model), is a standard mathematical tool for estimating suspension responses and can be used for summative or pointwise analysis of vehicle response to road geometry.

In fact, transportation agencies specify roughness requirements as summative values for pavement projects that affect construction practices and contractor pay factors. The International Roughness Index (IRI), a summative statistic of quarter-car suspension response, is widely used to characterize overall roughness profiles of pavement stretches but does not provide sufficient detail about the frequency or spatial distribution of roughness features.

This research focuses on two pointwise approaches, continuous roughness maps and wavelets analysis, that both characterize overall roughness and identify localized features and compares these findings with IRI results. Automated algorithms were developed to preform finite difference analysis of point cloud data collected by three-dimensional (3D) stationary terrestrial laser scans of paved and unpaved roads. This resulted in continuous roughness maps that characterized both spatial roughness and localized features. However, to address the computational limitations of finite difference analysis, Fourier and wavelets (discrete and continuous wavelet transform) analyses were conducted on sample profiles from the federal highway administration (FHWA) Long Term Pavement Performance data base. The Fourier analysis was performed by transforming profiles into frequency domain and applying the QC filter to the transformed profile. The filtered profiles are transformed back to spatial domain to inspect the location of high amplitudes in the suspension rate profiles.

Finite difference analysis provides suspension responses in spatial domain, on the other hand Fourier analysis can be performed in either frequency or spatial domains only. To describe the location and frequency content of localized features in a profile, wavelet filters were customized to separate the suspension response profiles into sub profiles with known frequency bands. Other advantages of wavelets analysis includes data compression, making inferences from compressed data, and analyzing short profiles (< 7.6 m). The proposed approaches present the basis for developing real-time autonomous algorithms for smoothness based quality control and maintenance.

DOI

https://doi.org/10.31274/etd-180810-4383

Copyright Owner

Ahmad Abdulraheem Alhasan

Language

en

File Format

application/pdf

File Size

182 pages

Share

COinS