Three-dimensional simulation of fuel relocation during loss-of-coolant accidents based on the coupled FEM-DEM approach.

• Three-dimensional simulation (3D) of fuel relocation under loss-of-coolant accident is performed by using discrete element method. • The 3D simulation of fuel relocation is firstly validated by the in-pile experiments. • The critical parameters like fuel stack reduction, mass fraction and filling ratio are influenced by the mean diameter of fuel fragments. • The exponentially fitting correlations are proposed for the fuel stack reduction as a function of the relative cladding volume increase. • The exponentially fitting correlations are proposed for the mass fraction and filling ratio as a function of the cladding circumferential strain. This work presents the three-dimensional (3D) simulations of fuel relocation during loss-of-coolant accidents based on the coupled FEM-DEM approach. A simple equivalent approach is employed to generate the fragment geometry with the mean diameter as the fragment size measured in the sieve test. For the first time, the 3D simulation of fuel relocation is validated by the in-pile experiments performed in the FR2 research reactor. The predicted missing fuel length is in good agreement with the experimental data, indicating the viability and effectiveness of the simulation. According to the features of fuel relocation, the parametric study on fragment mean diameter and cladding volume increase is conducted. Based on the quantitative analysis, it is found that the fuel stack reduction exponentially increases with the relative cladding volume increase. Also, the mass fraction and the filling ratio seem to exponentially increase and decrease with cladding circumferential strain, respectively. A series of fitting correlations of fuel stack reduction, mass fraction and filling ratio are proposed for different fragment mean diameters. By implementing the empirical correlation of filling ratio, the capability of the one-dimensional model is improved to gain a close predicted result as 3D simulation. The proposed correlations can support the interpretation of experimental data, and can enhance the evaluation accuracy of fuel performance during loss-of-coolant accidents by the more realistic prediction of the critical parameters influenced by fuel relocation. [ABSTRACT FROM AUTHOR]