Uncertainty Evaluation and Sensitivity Analysis Under Accident Scenarios

Nuclear power units need to operate conditioned the lowest risk possible. Safety analysis must use paired models, combining probabilistic and deterministic methods. In this study, FRAPCON and FRAPTRAN codes were used to simulate an idealized test based on IFA-650 series, carried out within Halden program. Nuclear systems work to depend on uncertainty values that must be quantified and propagated. The sources of uncertainties can be divided among physical models, boundary conditions, and mechanical tolerances. Eight physical models that can be configured, such as thermal conductibility, and fission gas release. Mechanical tolerances introduced by fuel fabrication are deviations that must propagate throughout of the system. To measure the effects produced by uncertainties were used correlation coefficients between entry and exit. Uncertainties contained on input values are spread to measure the impact created on safety limits. The method adopted used 96 samples to achieve the 95% of probability and 95% of confidence level.