Application of best estimate plus uncertainty analysis method for CNP600 rod ejection accident

Abstract

Compared with conservative accident analysis method, best estimate plus uncertainty (BEPU) accident analysis method can gain realistic analysis results and safety margins, and improve the nuclear power plant economy and flexibility of operation with the reasonable security assurance. The rod ejection accident (REA) analysis models under the condition of hot full power (HFP) and hot zero power (HZP) were established by using the best estimate code RELAP5-3D according to the design features of CNP600. The dominant phenomena and processes were identified through REA phenomena identification and ranking table (PIRT), and the important input parameters which have significant impacts on the key safety parameters were selected. The DAKOTA code was used to generate the samples of the important uncertainty input parameters by taking the Latin hypercube sampling (LHS) approach and the one sided tolerance upper limits of the key safety parameters were calculated through non-parametric method. In both cases the results obtained show that the maximum average fuel pellet enthalpy, peak fuel pellet temperature, peak cladding temperature, peak system pressure during accident transient meet the REA acceptance criteria. In both cases the one sided tolerance upper limits of peak nuclear power obtained by the non-parametric method are reasonable, and the one sided tolerance upper limits of maximum average fuel enthalpy obtained have considerable safety margins compared with the traditional REA conservative analysis values.