The critical thin walled shell structures in the reactor assembly of a pool type fast breeder reactor (FBR) are the main vessel, inner vessel and thermal baffles. On these structures, the seismic events impose major forces by developing high dynamic pressures, thereby causing a concern on structural integrity due to buckling. An integrated analysis for determining realistic forces and critical buckling loads at any instant during the seismic event has been carried out for the vessels of a typical 500 MWe pool type fast breeder reactor. The dynamic forces including pressure distributions generated on the vessel surfaces are extracted from the seismic analysis carried out for the reactor assembly. The seismic forces thus computed from axisymmetric analysis are transmitted appropriately to 3D shell geometries for the buckling analysis. In view of high computational time needed for carrying out buckling analysis at every time increment, the elastoplastic analysis is carried out only at a few critical time steps which are identified based on strain energies that are associated with the shear and compressive stresses developed at the portions of the vessels prone to buckle. The shear buckling of main vessel straight portion and buckling of toroidal portion of inner vessel and thermal baffles are found to be important. The possible randomness of support excitation time histories is accounted for by compressing as well as expanding the nominal time histories by 10%. Buckling strength reduction factors due to the initial geometrical imperfections are adopted from the literature. The inner vessel is found to be the most critical component which buckles under seismic forces induced by a safe shutdown earthquake with a load multiplier of 1.52, which is higher than the minimum factor of safety of 1.3 required as per the design code RCC-MR [RCC-MR: edition, 2002. Design and construction rules for mechanical components for FBR nuclear islands, vol. 1, section I, subsection B. AFCEN, Paris, in press] for service level D conditions.
Read full abstract