Release risk and optimization of railroad tank car safety design

Saat, M.R. and C.P.L. Barkan. 2005. Release risk and optimization of railroad tank car safety design. Transportation Research Record: Journal of the Transportation Research Board. 1916: 78-87. doi:10.3141/1916-11.


The performance metric for tank cars and other hazardous materials vehicles involved in accidents has generally been conditional probability of release given involvement in an accident. This metric considers the probability of a release event occurring but does not take into account the quantity of product lost in a release incident. In this paper, a new metric termed “release risk” is introduced; it is defined as the expected value of the quantity lost from a tank car given that it is in an accident. The quantity of product lost varies depending on the part of the car that is damaged in an accident; consequently, use of release risk can affect how different modifications in tank car design are considered. The metric was developed in terms of tank-damage-caused and non-tank-damage-caused releases. It was found that tank-damage-caused releases had a higher release risk than non-tank-damage-caused releases. Important elements considered are the probabilities of release and the expected quantities of release from the tank components and non-tank components of a tank car, and the effect of increasing tank thickness in increasing accident exposure and decreasing expected quantity of release. The release risk metric is also used as the objective function in a tank thickness optimization model. The results suggest that release risk may be a useful means of assessing the relative benefits of different tank car safety design modifications.