Effects of model variability on the safeguard analysis of gamma-contaminated truck loads

In the past few decades, the problem of dealing with inadvertently contaminated steel has turned out to be of real importance in several countries. Typically, this problem originates at a foundry which recycles scrap iron: indeed, in several instances it happened that the material to be smelted contained hidden gamma-sources, thus giving rise to safety problems with major health and economical fall backs. For this reason, the monitoring of the trucks arriving at the foundry gate is now of widely recognised importance. To date, most development efforts have been experimental with good results being achieved. More recently, the investigation of this problem using models is being pursued for a more fundamental understanding of the issues at stake. In this respect, the way of modelling the extreme geometrical variability of truck loads, even in case of same total scrap weight, is the crucial aspect and inevitably introduces elements of subjectivity and uncertainty. Correspondingly, the findings obtained by any modelling approach should be physically reasonable but one should be aware that their information content only resides in the general behaviour of the calculated results and their orders of magnitude.