Files in this item



application/pdfLaura_Ghosh.pdf (34MB)
(no description provided)PDF


Title:Development of an accident database for quantitative safety and risk analysis of highway bulk-package transportation of hazardous materials
Author(s):Ghosh, Laura
Advisor(s):Barkan, Christopher P.L.; Saat, M. Rapik
Department / Program:Civil & Environmental Eng
Discipline:Civil Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Hazardous Materials
Dangerous Goods
Transportation Risk
Risk Modelling
Risk Factors
Highway Cargo Tank
Highway Accident
Data Collection Methodology
Conditional Probability of Release
Abstract:Highway bulk package design elements have a substantial effect on the probability and quantity of release if a vehicle transporting hazardous materials (a.k.a. hazmat, dangerous goods) is involved in an accident. However, the ability to statistically quantify bulk package performance is constrained due to insufficient reliable information about the number of accidents, nature and severity of damages, frequency and severity of releases, and design specifications of involved packages. This thesis identifies necessary elements of an accident reporting system to collect sufficient information to evaluate U.S. DOT-specified hazardous materials bulk package accident performance. Such a data collection program would enable private and public sector organizations to better inform their risk management efforts including choice of package design, hazardous material routing decisions, or general understanding of the risk associated with various business activities. In the context of collision-caused accidental releases, current practices for conducting quantitative risk analyses are reviewed. These mainly consist of route-selection analyses that compare the expected risk on alternative routes. Many current practices use the most basic expression of risk: a product of the probability of a release event and the resulting consequences; however, there is a growing trend to expand the basic risk equation to account for differences in bulk package performance. The probability that a release will occur given that the bulk package has been involved in an accident, or conditional probability of release (CPR), is a potential metric to assess bulk package safety performance. Two methods are developed to incorporate CPR regression equations into the risk model: (1) a generalized risk model assuming independence of the CPR estimate for each roadway type, accident scenario and bulk package component, and (2) a route-specific risk model that accounts for route-specific information for each accident scenario – roadway type – bulk package component combination. Information needed to assess CPR is identified by surveying potential database users, and reviewing cargo tank classifications and specifications, industry practices, existing data collection methodologies, risk analysis strategies, container performance studies, and cargo tank accident investigations. This study identifies data fields in the following categories: bulk package design information, basic commodity information, bulk container damage information, and basic accident information. Several options for recording cargo tank performance in accidents are considered. An industry-sponsored extension of an existing program with voluntary participation was found to be the most desirable, but a government-sponsored mandatory program was identified as the most feasible. The U.S. DOT Pipeline and Hazardous Material Safety Administration (PHMSA) Hazardous Materials Information Reporting System (HMIRS) was determined to be the most appropriate existing program to extend because it collects about 70% of the desired information. This could be achieved by modifying the existing Form DOT F 5800.1 to collect additional data or by importing the necessary information from Form DOT F 5800.1 and requesting additional information separately. Technical implementation and security methods required for such a data collection program are explored in further detail. Data from National Transportation Safety Board accident reports, PHMSA HMIRS reports, Federal Motor Carrier Safety Administration Motor Carrier Management Information System reports and news articles are used to populate a prototype database. Using these data, the time required to collect sufficient incident data to support reasonable statistical analyses was determined. Based on a calculated rate of 132 accidents and 34 releases per month and assuming an accident-reporting rate of 20%, the database is expected to have a sufficient number of records to test basic hypotheses within two years of program implementation. The information in the prototype database was analyzed. Damage was observed to most likely occur to the top front passenger-side tank shell (from rollover accidents) and to the piping and/or undercarriage below the tank (from collisions). Crushing damage to the tank shell, including dents, was shown to be the most prevalent type of damage; however, it is the least likely to result in a release of hazardous materials. Barriers to implementing a successful accident database are identified and approaches to overcoming them suggested. These challenges are categorized as barriers associated with implementation, participation, obtaining data and database maintenance costs. Many of these concerns have already been addressed by the rail industry in their implementation of the Tank Car Accident Database (TCAD). Ultimately the success of an accident damage data collection system will require overcoming all of the potential institutional barriers.
Issue Date:2013-05-24
Rights Information:Copyright 2013 Laura Ghosh, 2013 National Academy of Sciences. All rights reserved.
Date Available in IDEALS:2013-05-24
Date Deposited:2013-05

This item appears in the following Collection(s)

Item Statistics