Investigation on RELAP5-3D© capability to predict thermal stratification in liquid metal pool-type system and comparison with experimental data

Abstract

A numerical activity, aimed to evaluate the capability of RELAP5-3D © to reproduce the main thermal-hydraulic phenomena in an HLM pool-type facility, in different operative conditions, is presented. For this purpose, the experimental campaign performed in CIRCE-ICE test facility has been selected for the code assessment. Two experimental tests have been analyzed: TEST A consisting in a transition from no-power to a full power steady state conditions, and TEST I, consisting in a transition from gas-enhanced circulation to natural circulation, simulating a protected loss of heat sink plus a loss of flow accident. Three different pool modelling approaches are presented, consisting in a single vertical pipe, parallel pipes with cross junctions and multi-dimensional component. The comparison with experimental data has highlighted the need to divide the large pool in several sections to reproduce the natural convection, strictly correlated with the thermal stratification. The multidimensional component seems to be the best practice for the evaluation of this phenomenon even if the lack of specific correlation for heat transfer coefficient in quasi-stagnant conditions in large tanks is a limit for the accuracy of the results. In addition, the paper presents a detailed nodalization of the fuel pin bundle, highlighting quite good capabilities of RELAP5-3D as a subchannel analysis code.

Post-test simulation of a PLOFA transient test in the CIRCE-HERO facility

“…The influence of heat conduction on radial and axial directions in liquid metal natural circulation was investigated by Yoo et al (2003). As presented in Narcisi et al (2019a), axial conduction has no appreciable effect on the temperature distribution within an LBE fuel bundle, whereas radial conduction, whose effect is negligible for high liquid metal flow rate, assumes relevant importance when the mass flow decreases to values typical of natural circulation. At this purpose, 3456 heat transfer nodes are considered to reproduce the radial conduction between adjacent subchannels, considering the volume between the center of subchannels and assuming a "fake" material with negligible heat capacity and LBE thermal conductivity.…”

“…A vertical pipe connects FPS model with fitting volume, which is simulated with a horizontal pipe component with variable flow area, to reproduce the flow path of the test section. Previous simulation activities (Narcisi et al, 2017a andNarcisi et al, 2019a) assess the importance of a detailed nodalization for this component, that gives rise to high heat losses towards the pool. The fitting volume is connected with the riser, modelled with a vertical pipe composed of 26 control volumes.…”

“…Another approach is to model the pool with a multi-dimensional (MULTID) component, developed and introduced in RELAP5-3D to allow the user to model more accurately the three-dimensional hydrodynamics (The RELAP5-3D© Code Development Team, 2015a). The assessment of the MULTID capabilities was presented in Narcisi et al (2019a), where the multi-dimensional modelling approach was compared with the multiple channel methodology. The comparison with experimental data highlighted an improved resolution adopting the MULTID component.…”

Codes of new generation for safety justification of power units with a closed nuclear fuel cycle developed for the “PRORYV” project

“…To our knowledge, a model of this type is not included in some well-known and widely used codes-analogous such as RELAP5-3D, ATHLET, ASTEC-Na (RELAP5-3D Code Manual 2012a, 2012b; Palazzo et al 2013;Zhou et al 2013;Flores et al 2016;Bandini et al 2018;Forgione et al 2019;Narcisi et al 2019;Shen et al 2019 and many others). Similar models are available only in some CFD codes.…”

System thermal-hydraulic modelling of the phénix dissymmetric test benchmark

“…Phénix model has been developed with RELAP5-3D ver. 4.3.4; it consists of two macro regions: a mono-dimensional and a three-dimensional modelling scheme, using the same approach described in Narcisi et al, 2019b. The mono-dimensional model reproduces the main components of the primary system, except for the pools, the diagrid and the core bypass.…”