Development of GPR data analysis algorithms for predicting thin asphalt concrete overlay thickness and density

Zhao, Shan

Permalink

https://hdl.handle.net/2142/102422 Copy

Description

Title

Development of GPR data analysis algorithms for predicting thin asphalt concrete overlay thickness and density

Author(s)

Zhao, Shan

Issue Date

2018-11-08

Director of Research (if dissertation) or Advisor (if thesis)

Al-Qadi, Imad L.

Doctoral Committee Chair(s)

Al-Qadi, Imad L.

Committee Member(s)

• Roesler, Jeffery R.
• Popovics, John S.
• Ozer, Hasan
• Leng, Zhen

Department of Study

Civil & Environmental Eng

Discipline

Civil Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Date of Ingest

2019-02-06T19:32:48Z

Keyword(s)

• Ground-penetrating radar
• Signal processing
• Transportation
• Pavement

Abstract

Thin asphalt concrete (AC) overlay is a commonly used asphalt pavement maintenance strategy. The thickness and density of thin AC overlay are important to achieving proper pavement performance, which can be evaluated using ground-penetrating radar (GPR). The traditional methods for predicting pavement thickness and density relies on the accurate determination of electromagnetic (EM) signal reflection amplitude and time delay. Due to the limitation of GPR antenna bandwidth, the range resolution of the GPR signal is insufficient for thin pavement layer evaluation. To this end, the objective of this study is to develop signal processing techniques to increase the resolution of GPR signals, such that they can be applied to thin AC overlay evaluation. First, the generic GPR forward 2-D imaging scheme is discussed. Then two linear inversion techniques are proposed, including migration and sparse reconstruction. Both algorithms were validated on GPR signals reflected from buried pipes using finite difference time domain (FDTD) simulation. Second, as a special case of the 2-D GPR imaging and linear inversion reconstruction, regularized deconvolution was applied to GPR signals reflected from thin AC overlays. Four types of regularization methods, including Tikhonov regularization and total variation regularization, were compared in terms of accuracy in estimating thin pavement layer thickness. The L-curve method was used to identify the appropriate regularization parameter. A subspace method—a multiple signal classification (MUSIC) algorithm—was then utilized to increase the resolution of 3-D GPR signals. An extended common midpoint (XCMP) method was used to find the dielectric constant and the thickness of the thin AC overlay at a full-scale test section. The results show that the MUSIC algorithm is an effective approach for increasing the 3-D GPR signal range resolution when the XCMP method is applied on thin AC overlay. Furthermore, a non-linear inversion technique is proposed based on gradient descent. The proposed non-linear optimization algorithm was applied on real GPR data reflected from thin AC overlay and the thickness and density prediction results are accurate. Finally, a “modified reference scan” approach was developed to eliminate the effect of AC pavement surface moisture on GPR signals, such that the density of thin AC overlay can be monitored in real time during compaction.

Graduation Semester

2018-12

Type of Resource

text

Permalink

http://hdl.handle.net/2142/102422

Copyright and License Information

Copyright 2018 Shan Zhao

Owning Collections