An Assessment of Uncertainties and Sensitivities in Severe Accident Progression Simulation for a Generic CANDU 6 Plant

Abstract

Studies performed to understand several nuclear reactor severe accidents such as National Research Experimental (NRX; 1952, Canada), Three-Mile Island Unit 2 (TMI-2; 1979, USA), Chernobyl Unit 4 (ChNPP-4; 1986, USSR, Ukraine) and Fukushima Daiichi (1F; 2011, Japan) have indicated that the phenomena involved in severe accident progression in nuclear power plants are extremely complex. It has been recognized that severe accident progression evaluation methods are normally associated with large uncertainties.

It is both a standard practice and a requirement imposed by national nuclear regulators (including the Canadian Nuclear Safety Commission, CNSC) to perform uncertainly and sensitivity analysis on nuclear plant safety analysis results. Despite the recognition of its importance by the international nuclear reactor safety community, the area of severe accident uncertainty and sensitivity analysis is relatively new and there is very limited guidance on the practical execution of an uncertainty and sensitivity analysis for severe accident assessments.

In this paper, an approach to the uncertainty and sensitivity analysis for severe accident assessments of a CANDU (CANada Deuterium Uranium, pressurized heavy water reactor) nuclear power plant is proposed and briefly described. The approach is based on Canadian nuclear industry guidelines (BEAU: Best Estimate Analysis and Uncertainty Methodology) and involves coupled simulations of the severe accident analysis code MAAP-CANDU (Modular Accident Analysis Program for the CANDU reactor) and the uncertainty analysis code SUSA (Software for Uncertainty and Sensitivity Analyses). The coupled simulations are used to assess the uncertainty/sensitivity involved in severe accident calculations, with respect to accident progression and consequences (fission product release to environment, hydrogen generation, etc.).

A brief discussion of a pilot uncertainly and sensitivity assessment of a severe accident is also conducted and presented. This study focuses on a generic CANDU 6 plant undergoing a postulated SBO (Station Blackout) scenario.