Best Estimate Simulation and Uncertainty Analysis of LB LOCA for KudanKulam VVER-1000 NPP

In the accident analysis of the Nuclear Power Plants (NPP) nowadays the international licensing practice considers several acceptable options for demonstrating the safety i.e. use of conservative computer codes with conservative assumptions, best estimate codes combined with conservative assumptions and conservative input data and application of best estimate codes with assumptions and realistic input data but associated with uncertainty evaluation of the results. The last option is particularly attractive because it allows for more precise prediction of safety margins with respect to safety criteria and their future use for power up-rating. The best estimate simulation with uncertainty analysis constitutes the framework of the present study which is to apply the last version of the French best estimate computer code CATHARE 2 in order to predict the thermal-hydraulic phenomena in the Indian KudanKulam Nuclear Power Plant (KK NPP) with VVER-1000 reactors during LB LOCA and to evaluate uncertainty along with sensitivity studies using the IRSN methodology. The paper first describes the modeling aspects of LB LOCA with CATHARE and then it presents the basic results. It highlights the use of SUNSET statistical tool developed by IRSN for sampling, management of several runs using CATHARE and further post treatment for uncertainty and sensitivity evaluation. The paper also deals with the difficulties associated with the selection of input uncertainties, code applicability and discusses the challenges in uncertainty evaluation.