Semester of Graduation

Summer 2024

Degree

Master of Civil Engineering (MCE)

Department

Department of Civil and Environmental Engineering

Document Type

Thesis

Abstract

Roadway safety is of utmost importance to the state transportation agencies. One of the major factors of roadway safety that agencies have control over is pavement friction which is affected by aggregate and mixture type. The Louisiana Department of Transportation and Development (DOTD) currently uses Polished Stone Values (PSV) obtained from the British Pendulum Device to rate its aggregates. These ratings determine the applicability of aggregates for various projects. However, the ratings are inconsistent, posing significant safety and financial risks. The first part of the thesis compares the currently adopted British Wheel Polisher-based Standard Practice for the Accelerated Polishing of Aggregates Using the British Wheel (ASTM D3319) procedure with the newly developed Three Wheel Polishing Device (TWPD)-based Standard Practice for Sample Preparation and Polishing of Unbound Aggregates for Dynamic Friction Testing (AASHTO PP103) procedure. The new procedure utilizes the Dynamic Friction Tester (DFT) and Circular Texture Meter (CTM) to measure the micro and macro texture of ring samples prepared from seven different aggregate sources. The performance of these devices was validated in the field tests. Additionally, a DFT-based friction rating table was suggested to replace the PSV-based friction rating table. The second part of the research aimed to update the friction mix design procedure, previously based on PSV values and regression formulas. This update involved simulating the TWPD polishing process using an Artificial Neural Network (ANN) to predict terminal DFT values, crucial for assessing final microtexture. The ANN model, developed using data from a recent Louisiana Transportation Research Center (LTRC) report (09-2B), was trained on an 80/20 data split and validated with withheld data for accuracy. The findings demonstrate that the ANN can effectively simulate the polishing process and predict terminal DFT20 values with low mean squared error (MSE). The DFT20 values from the first xiv portion of the test replaced the PSV values, and the ANN-predicted DFT20 of the asphalt mixture replaced the asphalt mixture slab DFT20 found using the blend PSV.

Date

7-15-2024

Committee Chair

Wu, Zhong

Download

Available for download on Tuesday, July 15, 2025

Share

COinS