EXTENSION OF SHIP ACCIDENT ANALYSIS TO MULTIPLE-PACKAGE SHIPMENTS

Severe ship accidents and the probability of radioactive material release from spent reactor fuel casks were investigated previously (Sprung. 1995). Other forms of RAM, e.g., plutonium oxide powder, may be shipped in large numbers of packagings rather than in one to a few caslcs. These smaller, more numerous packagings are typically placed in ISO containers for ease of handling. and several ISO containers may be placed in one of several holds of a cargo ship. In such cases. the size of a radioactive release resulting from a severe collision with another ship is determined not by the likelihood of compromising a single, robust package but by the probability that a certain fraction of lO's or tOO's of individual packagings is compromised. The previous analysis (Sprung. 1995) involved a statistical estimation of the frequency of accidents which would result in damage to a cask located in one of seven cargo holds in a collision with another ship. The results were obtained in the form of probabilities (frequencies) of accidents of increasing severity and of release fractions for each level of severity. This paper describes an extension of the same general method in which the multiple packages are assumed to be compacted by an intruding ship's bow unt.il there is no free space in the hold. At such a point, the remaining energy of the colliding ship is assumed to be dissipated by progressively aushing the RAM packagings and the probability of a particular fraction of package failures is estimated by adaptation of the statistical method used previously. The parameters of a conunon, well-characterized packaging. the 6M with 2R inner containment vessel, were employed u an illustrative example of this analysis method. However, the method is readily applicable to other packagings for which crush strengths have been measured or can be estimated with satisfactory confidence.