Files in this item



application/pdfCHADWICK-DISSERTATION-2017.pdf (9MB)
(no description provided)PDF


application/pdfForm_54_REA_DataStructure.pdf (214kB)
(no description provided)PDF


application/pdfForm_57_HRA_DataStructure.pdf (196kB)
(no description provided)PDF


application/pdfGX_Inventory_FileStructure.pdf (236kB)
(no description provided)PDF


Title:Quantitative analyses of train derailment probability at highway-rail grade crossings
Author(s):Chadwick, Samantha
Director of Research:Barkan, Christopher
Doctoral Committee Chair(s):Barkan, Christopher
Doctoral Committee Member(s):Saat, Rapik; Benekohal, Rahim; Gardoni, Paolo; Savage, Ian
Department / Program:Civil & Environmental Eng
Discipline:Civil Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Railroad safety
Highway-rail grade crossings
Level crossings
Transportation safety
Abstract:The current methodology for prioritizing highway-rail grade crossing (HRGC) warning system upgrades focuses on the likelihood of collisions and highway user casualties at crossings. However, these two metrics do not encompass all crossing risks. Specifically, they do not consider the potential for grade crossing incidents to cause train derailments and the consequent casualties to passengers and crew members, property damage, and release of hazardous materials. In contrast to the large body of research devoted to understanding the impact of crossings on highway users, almost no research has considered the risk that highway users pose to trains at HRGCs. With increased interest in passenger rail transport and the growth in transportation of hazardous materials such as crude oil, the importance of a comprehensive understanding of the risk of HRGC collisions is critically important. This dissertation develops an HRGC-caused derailment probability calculator using data analytics and statistical modeling. The Federal Railroad Administration (FRA) and state Departments of Transportation (DOTs) have developed large databases of historical incidents that can be used to better understand the effect HRGCs have on train derailment rates. I use these databases to develop statistical regression models that quantify actual experience to understand the differences between derailment and non-derailment incidents. I first develop a set of univariate statistical analyses to identify the incident-specific factors affecting derailment likelihood. Then, I develop three logistic regression models of derailment likelihood with these factors as input variables. Next, I develop a series of proxy variables to relate the incident-specific factors to crossing-specific characteristics. All of this is combined in a spreadsheet-based calculator, whose function I illustrate with a case study of four Illinois rail corridors. I combine these results with incident likelihood predictions generated by the FRA’s WBAPS system to show how consideration of derailment likelihood can affect crossing prioritization. By quantifying derailment likelihood, my research adds a new dimension to our understanding of how to assess grade crossing risk and warning system upgrade prioritization. The model allows users to identify crossings with high derailment likelihood, something that was not previously possible. This model will enable more informed allocation of safety resources to minimize the occurrence of derailments at grade crossings. It can be integrated into an overarching risk analysis framework that would consider all sources of risk at a grade crossing. Ultimately, this tool will open up new opportunities for railroad risk reduction, leading to a safer operating environment for railroads, rail passengers, highway users, and the general public alike.
Issue Date:2017-12-06
Rights Information:Copyright 2017 Samantha Chadwick
Date Available in IDEALS:2018-03-13
Date Deposited:2017-12

This item appears in the following Collection(s)

Item Statistics