GENERATION OF IN-STRUCTURE RESPONSE SPECTRA CONSIDERING SECONDARY SYSTEM MASS INTERACTION

摘要

With the advancement in computer capabilities, seismic soil-structure interaction (SSI) models of nuclear facility structures are no longer limited to few lumped masses on beam stick elements. Nowadays, more rigorous finite element models (FEMs) are often used for SSI analyses to predict the response of the structure as well as to develop in-structure response spectra for seismic qualification of equipment. However, in these models, it is not practical to accurately and explicitly represent the floor slabs that are often of composite design and has openings, because such a representation would require a large number of elements. For this reason, the SSI global models typically do not include such level of detail to accurately represent the vertical stiffness of the slabs, and vertical in-structure response spectra are often developed using a floor slab and equipment representation that essentially ignores the interaction between the equipment and the slab. For heavy equipment, this may result in an excessively conservative design. Thus, there is a need for a practical approach to develop vertical in-structure response spectra that would include the beneficial effects of interaction between the equipment and the floor slab. . Such an approach can be especially effective for a facility with a large number of floor-mounted heavy equipment that require seismic qualification. One such approach has been proposed and studied here by applying it on a hypothetical, but a realistic structure. In this study, the vertical in-structure response spectra for floor-mounted equipment were generated using a new pseudo-substructure method without increasing the size of the finite element model of the primary structure and avoiding numerous secondary analyses of subsystems. To evaluate the effectiveness of this new method, vertical in-structure response spectra (IRS) were generated using three SSI analytical models: (ⅰ) the Accurate Model (Case Ⅰ) that explicitly represents the equipment mass, the vertical flexibility of the equipment, and the vertical flexibility of the floor slab; (ⅱ) the Conventional Model (Case Ⅱ) that ignores the interaction between the equipment and the floor slab; and (ⅲ) the New Model (Case Ⅲ) that is capable of capturing most of the effects of the interaction between the equipment and the floor slab, but without additional elements representing the floor slab. In this paper, a description of the three models, the analytical approach, and a comparison of the response motions generated by the three models are presented and discussed.