폐쇄후 안전성평가 입력자료의 민감도 분석

Abstract

Spent fuels that arise from both PWR and CANDU reactors will be directly disposed a deep depth in a suitable host rock as a basic policy. In performance assessment of spent fuel, it is anticipated that the nuclide from which no retardation caused by sorption and solubility limit can be expected dominant dose. Therefore it is important not only to select a site of slow-moving in ground-water, but also to construct a robust engineered barrier under such nuclide characteristics, taking into account a safety margin of the system In this study, the sensitivity analysis method developed by KAERI was applied to a Korean geological disposal concept. Such a study will offers insight into Post-closure disposal safety assessment program and decide a dominant input parameter and will conduct a sensitivity analysis of dominant input parameter. Using the Goldsim is a highly graphical, object-oriented program for carrying out dynamic, probabilistic simulation. In order to quantify a nuclide release and transport through the possible pathway in the near- and far-field of a repository system as well as the dose rate with various exposure pathways in the biosphere system, robust program capable of an assessment of various possible release scenario are needed. Through chosen 5 dominant, independent input parameters(Canister failure time, Early time failure, Waste dissolution rate, Instant release fraction), carried out nominal sensitivity analysis. Based on result from sensitivity analysis, ranking of 5 input parameters, ranked parameters preferentially will consider to evaluate safety assessment. A program developed by utilizing Goldsim, by which nuclide transport in the near- and far-field of a repository as well as farther transport through a biosphere under various natural and manmade disruptive event could be modeled and assessed has been developed and demonstrated. Through current study, arbitrarily chosen dominant, independent parameters are evaluate of sensitivity analysis and it was important to improve an safety and uncertainly of the near- and far-field. It also seems useful and suitable enough for further in depth feedback to refine a repository design and helpful to improve an appropriate safety assessment even through there still might be many more thing to be developed with proper input data