LAUSR

Abstract In the framework of the HORIZON2020 SESAME European project, an experimental campaign has been carried out on the large Lead-Bismuth Eutectic pool integral effect CIRCE facility at ENEA Brasimone Research Center, implementing the HERO test section. This test section is aimed at supporting the development of the ALFRED design. Within the SESAME Project, three protected loss of flow accident tests have been designed and executed, identified as SE-Test1, SE-Test2, and SE-Test3. The third test (SE-Test3) has been selected by the project participants for a benchmark activity involving system thermal-hydraulic codes and coupled system thermalhydraulic/ CFD. This activity has been divided into two phases: a blind phase, and a post-test phase. The present paper illustrates the results achieved during the post-test phase of the CIRCE-HERO benchmark activity. Four participants simulated the experimental test. Three (ENEA, SCKCEN and UNIROMA1) used different versions of RELAP5 code, i.e. RELAP5-3D and RELAP5Mod3.3, and one (NRG) used a coupled approach based on an in-house system thermal-hydraulic code (SPECTRA) and the commercial code ANSYS CFX. The benchmark activity hereafter presented is a relevant exercise for evaluating and comparing the predictive capabilities of system thermal-hydraulic codes, CFD codes and coupled technics, in relation to phenomena occurring during the transition between forced and natural circulation (e.g. loss of flow) in a heavy liquid metals Generation IV system. The results of the numerical exercise showed an adequate capability of the codes to reproduce the relevant phenomena involved during the experiment. The SYS-TH codes have been more accurate than SYS-TH/CFD in predicting the trend of the main parameters during the transient, reproducing them with quite satisfactory accuracy. SYS-TH/CFD simulation, instead, provided a more satisfactory representation of the pool thermal stratification. The results highlighted that the planning and the execution of experiments fully devoted for the code V&V process is needed for the further development of such numerical tools, in particular for SYS-TH/CFD coupled approaches, whose use is more recent respect to SYS-TH codes.