Reliability of accident case histories concerning hazardous chemicals: An analysis of uncertainty and quality aspects

doi:10.1016/0304-3894(91)80059-W

Journal of Hazardous Materials

Volume 27, Issue 3, December 1991, Pages 339-350

https://doi.org/10.1016/0304-3894(91)80059-W Get rights and content

Abstract

A recent analysis of accident case histories achieved from readily available accident lists, has exposed the fact that an accident may be described more than once in the literature. Important discrepancies were found between the various main (accepted) sources in the field of risk analysis and loss prevention. An analysis of the differences found, and an assessment of the quality of accident case histories is attempted. Initially the analysis is limited to the number of fatalities, because this information is easy to measure, often available and lends itself to easy statistical treatment. The number of fatalities can thus be used as an indicator for the variability of the information found in accident case history reporting. The analysis is based on 595 transportation accidents involving hazardous chemicals from road, rail and pipeline transportation modes. Thirty-nine of these accidents were found to have three characteristics in common: (1) they happened before 1980, (2) they resulted in fatalities, and (3) they were described by more than one source. These 39 accidents are used in the analysis of the reliability of the number of fatalities reported in case histories. Sixteen accidents (or 41%) show discrepancies (in terms of fatalities) between the sources. Extrapolating the results of the “fatality” analysis to the other parameters, and supported by an analysis of selected cases, the general quality of information from accident case histories is evaluated. The result is a useful tool for assessing the quality of the various types of information. Using the terms “low”, “medium” and “high” to describe the quality, a table can be compiled. This tool shows that the information on date and place of the accident is of high quality, the event description and the number of casualties are of medium quality, and the information about chemical name and amount is of low quality.

References (14)

P. Haastrup et al.
Severity of accidents with hazardous materials. A comparison between transportation and fixed installations
J. Loss Prev. Process Ind.
(1990)
A Survey of Information Systems in OECD Member Countries Environment Directorate, Organization for Economic Cooperation and Development
Environment Monographs No. 25
(1989)
S. Contini et al.
MARS: The Major Accident Reporting System
I. Chem. E. Symp. Series
(1988)
F.P. Lees

There are more references available in the full text version of this article.

Cited by (16)

Research on the prediction model of hazardous chemical road transportation accidents
2023, Journal of Loss Prevention in the Process Industries
In recent years, hazardous chemicals road transport accidents have occurred frequently, causing huge casualties and property losses, and accident risk assessment has become the focus of researchers' research. To predict the risk probability value of hazardous chemical road transport accidents, first, we compiled data on road transportation accidents of hazardous chemicals in China in the past five years. And the nine nodes in the Bayesian network (BN) structure were defined in combination with relevant classification standards. The optimal Bayesian network structure for hazardous chemical road transport accidents was determined based on the K2 algorithm and the causalities between the nodes. Second, the node conditional probabilities were derived by parameter learning of the model using Netica, and the validity of the model was verified using the 5-fold cross-validation method. Last, the Bayesian network model of hazardous chemical road transport accidents is used to analyze accident examples, and the accident chain of “rear-end-leakage” is predicted, and the accident is most likely to be disposed of within 3–9 h. The study shows that the derived accident prediction model for hazardous chemical road transportation can reason reasonably about the evolution of accident scenarios and determine the probability values of accident risks under different parameter conditions.
Risk assessment of hazardous materials transportation: A review of research progress in the last thirty years
2022, Journal of Traffic and Transportation Engineering (English Edition)
In recent years, hazardous materials transportation accidents occur frequently, causing huge casualties and property losses. Risk assessment of hazardous materials transportation has been constantly investigated by researchers and decision-makers. In order to study the research progress of risk assessment of hazardous materials transportation, so as to better reveal relevant problems. This review systematically analyzes the existing literature, from 1991 to 2020, on risk assessment models and methods of hazardous materials transportation, and segregates the relevant studies based on unimodal and intermodal transportation systems. Additionally, an example is considered to analyze the characteristics of each risk assessment model of unimodal transportation, and a novel detailed classification is proposed for the risk assessment problems. The results indicate that the research on the risk assessment model of road transportation of hazardous materials is comprehensive, whereas that based on the railway and intermodal transportations is deficient. In comparison with the traditional risk assessment model, the conditional value-at-risk model presents a flexible decision-making framework for decision-makers and controls risk preference between risk neutrality and risk aversion. Owing to the difficulty in achieving door-to-door railway transportation and the increased risks associated with the intermodal transportation, the conclusion that the risk of railway and multimodal transportation of hazardous materials is lower than that of road transportation is not universal. Transport operators and governments can combine the advantages of different transportation modes, such as low risk, low cost, high flexibility, and high reliability, to achieve public safety and improve the competence of the system. Based on the proposed classification, this review identifies the existing trends and research gaps and summarizes future research directions.
A grounded theory model for analysis of marine accidents
2011, Accident Analysis and Prevention
Citation Excerpt :
The case history research strategy is one of the most favorable strategies for accident studies. The vast majority of accident studies rely heavily on case histories (Haastrup and Brockhoff, 1991; Facchini and Brockhoff, 1992; Christou, 1999; Konstantinos and Ernestini, 2002). Studies based on historical data, which is one of the types of data most frequently used, are generally preferred (Carol et al., 2001).
The purpose of this paper was to design a conceptual model for analysis of marine accidents. The model is grounded on large amounts of empirical data, i.e. the Swedish Maritime Administration database, which was thoroughly studied. This database contains marine accidents organized by ship and variable. The majority of variables are non-metric and some have never been analyzed because of the large number of values. Summary statistics were employed in the data analysis. In order to develop a conceptual model, the database variables were clustered into eleven main categories or constructs, which were organized according to their properties and connected with the path diagram of relationships. For demonstration purposes, one non-metric and five metric variables were selected, namely fatality, ship's properties (i.e. age, gross register tonnage, and length), number of people on board, and marine accidents. These were analyzed using the structural equation modeling (SEM) approach. The combined prediction power of the ‘ship's properties’ and ‘number of people on board’ independent variables accounted for 65% of the variance of the fatality. The model development was largely based on the data contained in the Swedish database. However, as this database shares a number of variables in common with other databases in the region and the world, the model presented in this paper could be applied to other datasets. The model has both theoretical and practical values. Recommendations for improvements in the database are also suggested.
Statistical analysis of domino chemical accidents
2000, Journal of Hazardous Materials
A set of chemical accidents is retrieved from the literature and classified with regard to the substance involved and whether domino effects are present. This set of accidents and each of the classes defined are statistically analyzed with respect to its severity and comparison is made between domino and non-domino accidents. The analysis reveals that each accident category shows characteristic patterns in terms of fatalities caused and domino effects likelihood. Moreover, chemical accidents severity frequencies are described by using a two-parameter, revised form of the Pareto probability density function. The range within which the values of the parameters lie is investigated using Bayesian inference.
An analysis of the database coverage of industrial accidents involving hazardous materials in Europe
1995, Journal of Loss Prevention in the Process Industries
The purpose of the present analysis is to explore systematic ways to address the problem of how many accidents involving hazardous materials actually occur in Europe, and to make a realistic estimate. A framework for predicting the total number of accidents was therefore developed and applied. The analysis is based on 535 unique accident descriptions (of which 107 were fatal accidents) from seven accident databases covering the nine-year period from 1984 to 1992. The majority (70%) of the accident descriptions identified were found in one source only. This indicates that, at best, about 30% of all accidents described may be found in a single database. Two models were developed: the first model was based on ideas similar to ‘chemical reaction kinetics’; the second was more related to ‘estimations of an animal population’. The models predict an average of 87 accidents per year in contrast to the observed average of 59 accidents per year. For fatal accidents, the models predict an average of 14 accidents per year, and the observed average is 12.
Natural events and accidents with hazardous materials
1995, Journal of Hazardous Materials
Based on accident case histories extracted from the data bases Mhidas (SRD) and Facts (TNO) an analysis was performed to identify relevant natural events causing accidents involving hazardous materials. A natural event is defined as an event originating from nature which initiates accidents with hazardous materials. In the analysis, the accidents from the two sources were pooled and the analysis concerns a total of 232 accidents, which were analysed with regard to specific natural cause, geographical distribution and time trend.
The analysis indicates that between 1% and 5% of accidents in industrial activities have natural events as a causative factor. Further for specific industries certain types of natural events seem to be dominating.
The most often reported natural cause of accidents is ‘atmospheric phenomena’ which account for 80% of the natural events found, lightning being the most common cause. Looking specifically at storage and processing activities lightning accounts for 61% of the accidents initiated by natural events.

View all citing articles on Scopus

View full text

Reliability of accident case histories concerning hazardous chemicals: An analysis of uncertainty and quality aspects

Abstract

J. Loss Prev. Process Ind.

A Survey of Information Systems in OECD Member Countries Environment Directorate, Organization for Economic Cooperation and Development

Environment Monographs No. 25

MARS: The Major Accident Reporting System

I. Chem. E. Symp. Series