Uncertainty quantification approaches for advanced reactor analyses.

Briggs, L. L. ; Nuclear Engineering Division

关键词: ACCIDENTS;    DESIGN;    EVALUATION;    LIQUID METALS;    LOSS OF COOLANT;    NUCLEAR POWER PLANTS;    PROBABILITY;    REACTORS;    RECOMMENDATIONS;    SAFETY;    SAFETY MARGINS;    SIMULATION;    TRANSIENTS;    WATER;   

DOI  :  10.2172/956921
RP-ID  :  ANL-GENIV-110
PID  :  OSTI ID: 956921
Others  :  TRN: US1005015

美国|英语

来源: SciTech Connect

【 摘 要 】

The original approach to nuclear reactor design or safety analyses was to make very conservative modeling assumptions so as to ensure meeting the required safety margins. Traditional regulation, as established by the U. S. Nuclear Regulatory Commission required conservatisms which have subsequently been shown to be excessive. The commission has therefore moved away from excessively conservative evaluations and has determined best-estimate calculations to be an acceptable alternative to conservative models, provided the best-estimate results are accompanied by an uncertainty evaluation which can demonstrate that, when a set of analysis cases which statistically account for uncertainties of all types are generated, there is a 95% probability that at least 95% of the cases meet the safety margins. To date, nearly all published work addressing uncertainty evaluations of nuclear power plant calculations has focused on light water reactors and on large-break loss-of-coolant accident (LBLOCA) analyses. However, there is nothing in the uncertainty evaluation methodologies that is limited to a specific type of reactor or to specific types of plant scenarios. These same methodologies can be equally well applied to analyses for high-temperature gas-cooled reactors and to liquid metal reactors, and they can be applied to steady-state calculations, operational transients, or severe accident scenarios. This report reviews and compares both statistical and deterministic uncertainty evaluation approaches. Recommendations are given for selection of an uncertainty methodology and for considerations to be factored into the process of evaluating uncertainties for advanced reactor best-estimate analyses.