Full-scale railroad ballasted track evaluation through experiment, simulation and prediction

Feng, Bin

Permalink

https://hdl.handle.net/2142/115915 Copy

Description

Title

Full-scale railroad ballasted track evaluation through experiment, simulation and prediction

Author(s)

Feng, Bin

Issue Date

2022-07-12

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

Tutumluer, Erol

Doctoral Committee Chair(s)

Tutumluer, Erol

Committee Member(s)

• Hashash, Youssef M.A.
• Edwards, J. Riley
• Bian, Xuecheng
• Qian, Yu

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)

• Railroad
• Ballasted Track
• Tie Support Conditions
• Loading Patterns
• High-speed Rail
• Discrete Element Method (DEM)
• Sensing and Prediction

Abstract

Railroad ballasted track, typically divided into superstructure and substructure, serves as the transportation facility for majority freight and passenger lines operating in the world. Being an essential component in the track substructure, ballast layer functions as the medium for distributing vehicle loads, facilitating drainage, providing track restraints, and more. Throughout its operation and service life, the ballast layer will develop into different tie support conditions as a result of different loading patterns and maintenance activities. Considering the enormous costs to pause regular track operations in the field for scientific research studies, full-scale laboratory experiments are ideal alternatives for monitoring track responses subjected to realistic field loads. Discrete element method (DEM), as another approach for investigating ballasted track and valuable supplement for laboratory testing, can capture particle-level track behavior at a very high sampling frequency. Leveraging these simulated individual particle data, a conceptual framework is introduced to demonstrate a prediction approach for tie support conditions when coupled with smart sensing technologies to monitor track serviceability and maintenance indicators. The main contribution of this doctoral research is to present systematic approaches for full-scale railroad ballasted track evaluation through experiment, simulation and prediction. Starting with investigating the effects of common tie support conditions, a polyhedral 3D DEM program named BLOKS3D was used to establish simulation models, which were then validated by corresponding laboratory studies conducted in the Rail Transportation and Engineering Center (RailTEC). Loading patterns applied on the validated DEM models can be categorized into three levels: (1) static rail seat loads; (2) dynamic field loading patterns; and (3) mixed traffic patterns. The influence of tie support conditions and loading patterns were investigated and visualized through macro-level track dynamic responses (e.g., tie vibration velocity) and micro-level ballast particle vibration velocity. Next, to better examine ballast layer behavior under high-speed loadings, multi-tie full-scale ballasted track experiments were conducted on a laboratory track-subgrade system, known as the Zhejiang University High-Speed Rail Tester (ZJU-iHSRT), and the ballast layer behavior was modeled using BLOKS3D DEM program with newly featured parallel computing capability. Recently developed “SmartRock” sensors were also installed for the first time in the ZJU-iHSRT facility to bring additional particle-level behavior monitoring capability during laboratory testing. Railroad dynamic loading input and feedback mechanism, as well as material related parametric studies were conducted using the BLOKS3D DEM simulations to enhance its reliability and functionality of the DEM modeling approach as a suitable tool for ballasted track simulations with different loading and layer conditions. Both macro- and micro-level track behavior recorded from laboratory experiments and DEM simulations were then compared and analyzed. Lastly, inspired by the need and challenge for evaluating tie support conditions in the field and vast particle-level data already captured from different DEM simulations, a conceptual framework was introduced to predict common tie support conditions. Early attempts of a data driven approach demonstrated through a neural network model developed with limited data showed promise to differentiate four support conditions with the given particle acceleration responses measured at several representative locations within the ballast layer. The trained model was applied to the full-scale ballasted track experiments conducted in the ZJU-iHSRT facility by taking previously recorded “SmartRock” sensor data as inputs.

Graduation Semester

2022-08

Type of Resource

Thesis

Copyright and License Information

Copyright 2022 Bin Feng

Owning Collections