Applying IDPSA in PSA level 3 - a pilot study

This report is a pilot study of applying deterministic and probabilistic methods in level 3 probabilistic safety analysis (PSA). On the deterministic side, VTT's consequence analysis code ARANO is used in calculating the atmospheric dispersion of a release of radioactive substances, and in estimating the total dose of ionizing radiation. On the probabilistic side, VTT's level 2 PSA code SPSA is used to assess the probabilities of different consequences. The main model is an event tree, where each branch concerns either the value of a weather variable (wind direction, wind speed, precipitation) or a countermeasure variable (evacuation success, sheltering success). The case considered is an alternative take on the Fukushima Daiichi nuclear power plant accident. The setup is as follows: the population of the major cities close to the site are in place (and not killed by or evacuated after the earthquake and tsunami), and the impact of weather is analysed on the basis of what it statistically is in March in that part of Japan. What radiological consequences (in terms of population dose and cancer deaths) would the radioactive release from the site have had under these presuppositions? The population doses are analysed in the event tree, and uncertainty analyses are conducted on the weather variables, evacuation and sheltering success probabilities, and the effectiveness of sheltering. We find that, even under rather conservative assumptions, the radiological consequences are small. However, the results should be seen as only indicative due to simplifications made in modelling. The pilot study demonstrates that the approach used is a viable way of conducting level 3 analyses.