Examination of Conservatism in Early/Latent Fatality Estimation in Level 3 PRA

Due to the computational model driven-nature of the work, there exist various sources of uncertainty in level 3 PRA. They are related with source release, environmental transport and deposition, human behavior involved in dosimetry, health effect and risk assessment. For instance, a total of 376 parameters have been considered in Probabilistic Accident Consequence Uncertainty Assessment Using COSYMA and the details on the number of parameters in each analysis are listed in Table 1. In 2012, the report of NPP accident consequence simulation was distributed by the Korean Federation for Environmental Movement (KFEM). They insisted that Kori Nuclear Power Plant (NPP) accident would lead to 48,000 early fatalities and 850,000 cancer fatalities in Busan and Hanbit NPP accident would lead to 550,000 cancer fatalities in Seoul. This report exemplifies the misuse of collective dose, that is effective dose multiplied by population and time. Even though very low effective dose is considered, collective dose could give over-conservative estimate when high population and long time period is multiplied. International Commission on Radiological Protection (ICRP) forewarned about the misuse of collective dose, in their ICRP Publication 103, such as applying it to simplified calculation of fatality and risk. As part of investigation of conservatism in early and latent fatality estimation, the existing methods of early and latent fatality calculation was reviewed and the results from the use of the existing methodology were examined in this study. The method of early and latent fatality estimation in level 3 PRA was investigated and the conservatism in the result was examined in this study. For the purpose of estimating both early and latent fatality, appropriate dose distributions among the affected population are found to be important. This study showed that large conservatism may be involved in the estimated fatality if the distribution of population dose as a function of downwind distance is not appropriately characterized. Early fatality estimation based on the use of threshold dose avoids the misuse of collective dose concept. However, dividing the region of evaluation into sufficient numbers of grid was found to be important to reduce uncertainty in early fatality estimation