An empirical analysis of Canadian gasoline and LPG truck releases

Abstract

In July of 1985, reporting of dangerous goods occurrences in Canada became mandatory under the Transportation of Dangerous Goods Regulations (SOR/85-77). Since that date the Transport Dangerous Goods Directorate of Transport Canada has collected data on numerous characteristics of each reported spill in a comprehensive data base. This paper examines the 1986 and 1987 data for gasoline and liquefied petroleum gas (LPG) truck spills that occurred en route, with the objective of calibrating the accident and release submodels for use in a transportation risk model that was developed previously for Transport Canada. The most common accident types that were observed to cause releases of these two commodities were: overturn, collision followed by overturn, and collision. Similar accident types were recorded for both commodities; however, the frequency of non-transportation releases for LPG was higher, likely due to different reporting requirements or practices. It was found that overturn and collision/overturn accidents resulted in similar release sizes, while collision accidents resulted in statistically lower average release sizes. The other spill types were examined qualitatively. For example, fire was found to result in a 98% release of lading, 3 out of 4 times. Fire only occurred with gasoline spills, but the LPG spill sample size was much smaller than for gasoline. Of the three truck types represented in the database, tractor trailer, tractor trailer with pup, and tanker truck, the latter was the most common, and hence was used for the statistical analysis. It was found that the former two truck types could be considered as one sample in terms of release percentage, but they could not be combined with the tanker truck sample. The paper concludes with a number of recommendations for developing spill distributions and event trees for each accident type to be used in conjunction with the Transport Canada risk model.