RELAP5-3D Analysis of OECD-NEA/NRC BFBT Benchmark

OECD-NEA and the NRC organized and sponsored the BWR Fuel Bundle Test (BFBT) Benchmark with the main purpose of assessing sub-channel and Computational Fluid Dynamic codes capabilities in estimating relevant thermal-hydraulics parameters like void fraction and critical power for a BWR boiling channel. The assessment activity is performed comparing the code calculation results with the experimental data at steady state and transient conditions available through the Japanese Nuclear Power Engineering Corporation (NUPEC). In this framework, the San Piero a Grado Nuclear Research Group (GRNSPG) of the University of Pisa (UNIPI) developed a RELAP5-3D© thermal-hydraulic nodalization of the experimental bundle. The main purpose of this activity was the assessment of the capability of the well-known three-dimensional system thermal-hydraulic code RELAP5-3D© for the prediction of relevant parameters at the fuel assembly scale. In order to exploit the large amount of experimental data available, a three dimensional thermal-hydraulic nodalization was developed, simulating all sub-channels with MULTI-D component. The overall activities resulted in challenges for the code and the code users because of the necessary large number of nodes and heat structures used and because of the different solutions that had to be found for performing a typical sub-channel analysis with a system thermal-hydraulic code. Several calculations simulating steady state conditions were performed for different fuel assembly configurations. For each of them the void fraction distributions in all fuel assembly sub-channels and pressure drops along different part of an assembly were compared with the available experimental data. In the next exercises, BWR transients were executed (turbine trip and recirculation pump trip, respectively), calculating again the void distributions and critical power conditions. The results of the activity demonstrated the capability of the RELAP5-3D© code to perform calculations using a sub-channel approach. The code was able to calculate several thermal-hydraulics parameters with high accuracy at “fuel bundle” level of resolution; the results of “sub-channel” level are instead affected by a higher error (e.g. deviation of around 20% in the prediction of void distribution). Sub-channel results showed a better agreement when considering high quality tests compared to the lower quality ones.