SPRINT: a tool for probabilistic source term prediction for use with decision support systems.

This paper describes a software tool, SPRINT, that has been developed within the FP5 STERPS project for rapid source term prediction. It is based entirely on a probabilistic approach to the diagnosis and prognosis of Nuclear Power Plant status, based on key instrument readings and other observations. In principle, the method can be applied to any accident sequence from the benign to the most severe. However, the main benefits are seen for the diagnosis of plant status during degraded core conditions where the accident phenomenology is either uncertain or essentially non-deterministic in nature. The output from SPRINT consists of a set of potential source terms with an estimate of the likelihood of each.