Characterising moisture within unbound granular pavements using multi-offset Ground Penetrating Radar

Muller, Wayne (2017). Characterising moisture within unbound granular pavements using multi-offset Ground Penetrating Radar PhD Thesis, School of Civil Engineering, The University of Queensland. doi:10.14264/uql.2017.337

Attached Files (Some files may be inaccessible until you login with your UQ eSpace credentials)
Name Description MIMEType Size Downloads
s40155335_phd_correctedthesis.pdf Thesis (open access) application/pdf 14.72MB 0
Author Muller, Wayne
Thesis Title Characterising moisture within unbound granular pavements using multi-offset Ground Penetrating Radar
School, Centre or Institute School of Civil Engineering
Institution The University of Queensland
DOI 10.14264/uql.2017.337
Publication date 2017-02-10
Thesis type PhD Thesis
Supervisor Alexander Scheuermann
Bryan Reeves
Total pages 265
Language eng
Subjects 0404 Geophysics
0905 Civil Engineering
Formatted abstract
Moisture has a significant influence on the performance and durability of unbound granular (UBG) pavements with thin bituminous surfacings, which form the vast majority of all-weather Australian roads. Techniques for measuring pavement moisture content are therefore important to quantify its influence on structural performance, enable early detection of problem areas, and to assess the condition of flood-affected pavements. Conventional techniques, however, require invasive probes or physical sampling which are impractical for large-scale investigations. Ground Penetrating Radar (GPR) techniques show promise, yet existing methods achieve only qualitative estimates, involve simplifications that limit accuracy or require complicated or laborious data analysis. The aim of this research was to develop a semi-automated approach using multi-offset GPR to quantitatively estimate the moisture content and depth of UBG pavement layers while maintaining data analysis simplicity.

The research made use of a new type of 3-dimensional (3D) GPR technology able to continuously collect ground-coupled multi-offset gathers across the road while travelling at up to traffic speeds. To enable development of analysis methods prior to equipment completion, numerical simulations were used to model the expected response for typical pavement configurations. These data were used to test a semi-automated approach using interface tracking and conventional multi-offset geophysical analysis methods. They were also used to develop two novel self-correcting analysis techniques, Interface Matching (IM) and Ray-path Modelling (RM), which use migration and tomographic approaches, respectively.

A permittivity characterisation approach, later called modified free-space (MFS), was also adapted for laboratory characterisation of UBG materials. It was used to calibrate petro-physical relations for these materials that enable pavement moisture estimates from the GPR measurements. As the MFS approach departs from conventional free-space methods, numerical modelling and a series of laboratory experiments were used to assess measurement accuracy and to determine the influence of sample edges, aspect ratio, and depth and antenna separation on the measurements. These investigations demonstrated that the MFS approach was suitable using the proposed equipment, sample sizes and for the range of relative permittivity values likely to be encountered. It also produced results that compared well with established apparatus over the frequency range relevant to pavement GPR.

Measurements using this approach were also compared with conventional time-domain reflectometry (TDR) and common-offset GPR measurements of moisture-varying and density-varying UBG samples. The comparison showed similar results for most samples, although TDR reported lower relative permittivity values for drier and lower-density materials.

Upon completion of the 3D GPR equipment, a revised analysis approach called Ray-path Modelling-Semblance (RM-S) was developed. It was used to analyse multi-offset measurements collected along a recently-constructed site to predict the depth, relative permittivity and volumetric moisture content of UBG pavement layers. The predictions were validated by comparing to physical measurements of layer depth and moisture content along the site, which showed a good correlation. Moisture variability along the site and over time was determined by comparing predictions from a scan collected at the end of construction and another approximately 11 months later. The comparison showed that the permittivity of upper layers was relatively consistent along the site at the time of construction and that lower pavement layers had a greater permittivity. By the time of the second scan the permittivity differential between upper and lower layers had reduced along most of the site, however in places the permittivity increased due to moisture ingress. These temporal trends compared well with results determined using embedded TDR sensors and common-offset GPR measurements of buried reflectors, although the new approach produced somewhat higher permittivity estimates. The analysis approach was also shown to be repeatable, based on a comparison of layer depth and moisture predictions determined for consecutive scans along the site.

The key outcomes of the research were the successful development and validation of:

1. A new semi-automated analysis approach using multi-offset GPR that enables quasi-continuous determination of the depth, relative permittivity and moisture content of UBG pavement layers; and

2. A novel laboratory technique for permittivity characterisation of civil engineering materials at GPR frequencies that enables measurement of larger samples and is better suited to measuring coarse-grained, compacted and un-bonded pavement materials compared to existing alternatives.
Keyword Road pavement moisture
Unbound granular road pavements
Ground penetrating radar
Permittivity characterisation
Time domain reflectometry

Document type: Thesis
Collections: UQ Theses (RHD) - Official
UQ Theses (RHD) - Open Access
 
Versions
Version Filter Type
Citation counts: Google Scholar Search Google Scholar
Created: Wed, 08 Feb 2017, 04:45:09 EST by Wayne Muller on behalf of Learning and Research Services (UQ Library)