Date of Award

8-2025

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Civil Engineering

Committee Chair/Advisor

Dr. Bradley J Putman & Dr. Fabricio Leiva

Committee Member

Dr. Prasad Rangagu

Committee Member

Dr Jennifer Ogle

Abstract

The objective of the study was to evaluate the sensitivity of various input variables on the flexible pavement design thickness of high-speed, high-traffic routes in South Carolina using the Mechanistic-Empirical Pavement Design Guide (MEPDG) by means of the AASHTOWare Pavement ME Design software using global calibration coefficients with a focus on bottom-up fatigue cracking. The variables considered in this investigation included two-way average annual daily truck traffic (AADTT), asphalt mix type, climate station, subgrade type and resilient modulus, and aggregate base thickness. The study includes comparative analysis using older methods like the AASHTO 1993 method and the South Carolina DOT Pavement Design Guidelines, which is primarily based on the AASHTO 1972 method. Additionally, the study discusses the local calibration of the bottom-up fatigue cracking model for South Carolina for medium-level traffic and using machine learning methods like Artificial Neural Network (ANN), Gradient Boosted Method (GBM), and Random Forest (RF) to enhance model prediction. The results of the sensitivity analysis indicated that the asphalt mix type did not have a significant impact on the results. However, one of the five climate stations evaluated resulted in significantly thicker pavements than the others. Both subgrade types, as well as resilient modulus, had a significant effect on the pavement thickness. Finally, pavements were more sensitive to total truck traffic changes at lower AADTT values and then became somewhat less sensitive when exposed to the highest levels of traffic. The findings from the sensitivity study were used to develop a preliminary asphalt thickness design catalog for interstate routes in South Carolina. The results from the local calibration model showed high errors for bottom-up fatigue cracking in all the trials, but machine learning algorithm was able to increase prediction accuracy.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.