Improvement of ATHLET modelling capability for asymmetric natural circulation phenomenon using uncertainty and sensitivity measures

Abstract

Natural circulation is one of the most important phenomena relating to nuclear reactor safety. In the last years of 20th century, lots of test facilities were set up to conduct natural circulation experiments. At the same time, some system codes, such as RELAP5 and CATHARE, have been validated by the experimental data carried out by these facilities. The calculation results show that these codes can capture most of the characters during the natural circulation transient, except some special processes. One of the examples is PKL IIIB3.1 test, an asymmetric natural circulation transient in the primary loop with an isolated steam generator. The cooled down process using the intact steam generators by reducing the secondary side pressure is the driving force of the natural circulation in primary loop.In this paper, a post-test calculation is performed by best-estimate thermal-hydraulic code ATHLET to simulate PKL IIIB3.1 test. Due to the effect of isolated steam generator on the whole primary cooling system, the results show that the mass flow in the isolated loop is overpredicted significantly comparing to the experimental data. Many measures have been tried to improve the simulation results, but an effective method is still missing. Based on the post-test calculation results, an uncertainty and sensitivity analysis is carried out to find out the main reason causing this higher circulation mass flow. Lots of factors are considered in this study, e.g. the boundary conditions, thermal-hydraulic correlations, and model parameters. The most sensitive factors are selected and adopted to improve the simulation modelling. The results achieved so far indicate that with these measures, the ATHLET capability for the asymmetric natural circulation phenomenon is effectively improved.