Evaluation of the Conservatism of NPP Safety Analysis Dose Calculations as Typical for Licensing Purposes

Providing a reliable upper limit of radiological consequences to the plant personnel and the general public is typically the aim of a safety evaluation for anticipated operational occurrences or design basis accidents, as presented in a safety analysis report. A typical tool for dispersion calculation and dose evaluation is MACCS2. In the present analysis four types of calculations are presented: a first calculation, typical for licensing analysis, with the MACCS2 computer code. In a second step conservative assumptions e.g. ground release even if a stack release would be realistic, are dropped. In a third step calculation two is repeated with RODOS, a code (online decision making tool) used to predict the radiological consequences of an accidental release of activity. The step three calculation still contains all the conservative assumptions that are built in the MACCS2 code. In a last step these assumptions are removed, and a “best estimate” calculation on the dose to the public is performed. The whole analysis (step one to four) is repeated for different source terms (noble gases only, tritium dominated, primary system water …) and for different weather conditions. Two main conclusions can be drawn. The first by comparing step two (MACCS2) and step three (RODOS). Here the boundary conditions of the calculations are set to be as similar to each other as possible. The paper shows that despite the fact that MACCS2 uses a Gaussian plume model, while RODOS uses a puff model for dispersion calculation, doses of the same order of magnitude are calculated. For the second conclusion the step one (MACCS2, conservative) and step four (RODOS, best estimate) calculations are compared, it is shown that although the margin of conservatism varies considerably from case to case, the results differ at least one order of magnitude.