ASTEC validation of SFP dewatering using results from the DENOPI project

Abstract

Loss of cooling in a Spent Fuel Pool (SFP) of a nuclear power plant can lead to the melting of fuel assemblies and to strong radiological consequences to the environment. In order to study the first phases of such accidents, up to the fuel assemblies uncovery, the DENOPI project was launched by the French Institute for Radiation Protection and Nuclear Safety (IRSN¹) supported and funded by the French Government and partners. Among the different facilities developed in the project, the MIDI facility aims at studying the complex thermal-hydraulics phenomena occurring in a large water pool heated from the bottom by electrical rods arranged in dedicated racks. MIDI is scaled by homothety to a typical French SFP. Different assembly arrangements (loading patterns) have been tested at different power levels, with either uniform power repartition, or hot and cold cells. In each test, the water level and temperatures at different elevations are followed, as well as mass flow rate entering each fuel rack. These experimental results also provide relevant data for the analysis and understanding of large natural convection loops that are expected in immersed passive heat removal systems of Small Modular Reactors. The forthcoming OECD/NEA POLCA project aims to extend such results database, in particular to assess the capability of thermal-hydraulics codes to reproduce the main tendencies of these experimental results.

The ASTEC code developed by IRSN is a system code dedicated to the simulation of major accidents in nuclear facilities that may lead to the release of radiological material. Recent works within the MUSA European project have shown the importance of reducing models uncertainties in the first phases of the accident, during the pool dewatering. In this paper, first simulations of MIDI tests are performed with ASTEC in order to assess and improve the capability of ASTEC to simulate the dewatering of a large water pool such as a SFP during a loss-of-cooling accident. Simulations are performed for a selection of MIDI tests with different heating patterns and power levels. Different models of subcooled boiling models from the literature are tested in ASTEC, stressing the key role of these models for an accurate prediction of the experimental flow.