https://hal-cea.archives-ouvertes.fr/cea-02500834Yao, F.F.YaoCEA-DES (ex-DEN) - CEA-Direction des Energies (ex-Direction de l'Energie Nucléaire) - CEA - Commissariat à l'énergie atomique et aux énergies alternativesStrub, C.C.StrubCEA-DES (ex-DEN) - CEA-Direction des Energies (ex-Direction de l'Energie Nucléaire) - CEA - Commissariat à l'énergie atomique et aux énergies alternativesLe Tallec, P.P.Le TallecDTIS, ONERA, Université Paris Saclay (COmUE) [Palaiseau] - ONERA - Université Paris Saclay (COmUE)Simulation of the static mechanical behaviour of a pressuried water reactor fuel assemblyHAL CCSD2015rod-to-grid contact modeldomain decompositon methods[PHYS.NEXP] Physics [physics]/Nuclear Experiment [nucl-ex][PHYS.NUCL] Physics [physics]/Nuclear Theory [nucl-th]amplexor, amplexor2020-03-06 13:25:032021-10-14 09:15:382020-03-27 17:26:00enConference papersapplication/pdf1The paper presents the mathematical modelling and computer simulation of static mechanical response of a PWR fuel assembly, with the help of the finite element code CAST3M developed by CEA. In PWR core analysis, the fuel assemblies present non-linear behaviour, which can be explained by the variations of stiffness due to the contact and the slipping between fuel rods and spacer grids as well as to the loss of contact of the rods. In the first part of this paper, a rod-to-grid contact model, based on the Coulomb law and constituted of non-linear springs and hinges, is developed to simulate the holding system of fuel rods inside the grid cells. The results of simulation showed a good agreement with experimental data. The second part aims at applying the domain decomposition methods to the modelling of the fuel assembly with proposed contact model. Due to geometric complexity of the fuel assemblies, the approximation by finite element method leads to a large number of unknowns. The resolution of this system would be prohibitively expensive with the available computation power. By using the domain decomposition methods we are able to reduce a significant amount of computational time and relieve the memory requirement from the advantage of parallel computing.