Flexible pavement design for a mixture of human-driven and connected and autonomous trucks

Okte, Egemen

Permalink

https://hdl.handle.net/2142/116245 Copy

Description

Title

Flexible pavement design for a mixture of human-driven and connected and autonomous trucks

Author(s)

Okte, Egemen

Issue Date

2022-07-14

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

Al-Qadi, Imad L.

Doctoral Committee Chair(s)

Al-Qadi, Imad L.

Committee Member(s)

• Ouyang, Yanfeng
• Meidani, Hadi
• Ozer, Hasan
• Sias, Jo

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

Keyword(s)

Truck Platooning, Flexible Pavement Design, Machine Learning, Infrastructure Performance Prediction

Abstract

Trucks are responsible for moving nearly 70% of the freight in the US. Hence, efficient truck freight would positively impact the national economy. However, because transportation contributes to nearly 30% of emissions in the US, with 25% of that from trucks, this would impact global environment. Connected and autonomous trucks (CAT) have the potential to address current freight challenges such as congestion, high fuel consumption, and driver shortage. For example, CAT may form truck platoons to save up to 15% fuel during long haul trips by following each other closer than human driven vehicles would. However, platoons may result in channelized traffic that would reduce pavement life. The objective of this dissertation was to quantify the impact of truck platoons on flexible pavement performance and to develop a framework that would allow incorporating mixed traffic modes of human driven trucks and CAT to reduce the impact on flexible pavement systems. To achieve this goal, first, road sections suitable for truck platooning in Illinois were identified. A base case of five trucks with 65-ft spacing was used in the analysis. Platoonability of a road was determined based on traffic density and number of interactions near entry and exit ramps of a highway. It was found that the majority of Illinois road network (89%) was suitable for platoons under normal conditions. This number increases to 93% during off-peak hours. Decreasing traffic density or number of trucks in a platoon could increase the platoonable percentage by up to 10%, especially near high traffic sections. Once platoonable sections were identified, the impact of CAT on flexible pavement responses and potential damage was considered. A numerical finite element (FE) surrogate model for flexible pavements that could capture complex loading scenarios was developed. A database of 850 FE simulations was utilized. A Bayesian neural network was developed to predict critical pavement responses under two different tire configurations. The developed model was able to predict critical responses with a maximum error of 15%. In addition, the model could capture uncertainty and allows constructing a prediction bound around a flexible pavement section of interest. To quantify the impact of CAT on flexible pavement, an expected response methodology was developed that considers truck lane positions as a mixed probability distribution. It was found that distributing CAT across the lane width could reduce cracking by 40% and rutting by 20%, respectively. Channelizing traffic, on the other hand, could increase cracking by 60% and rutting by 25%, respectively. A computationally efficient simplified algorithm was developed to allow incorporating platoons in mechanistic-empirical analysis at various levels of penetration. This allows application for network level optimization and probabilistic life-cycle analysis.

Graduation Semester

2022-08

Type of Resource

Thesis

Copyright and License Information

Copyright 2022 Egemen Okte

Owning Collections