Selecting the Computation Model Boundaries in Performing Combined Experimental and Numerical Confirmation of NPP Branched Pipework Seismic Stability

Abstract

Unavoidable deviations from the initial design solutions that are made in the course of installation work during NPP construction generate the need to perform combined numerical and experimental confirmation of plant equipment seismic stability with determining its natural dynamic characteristics. The most labor-consuming stage of activities aimed at performing combined experimental and numerical confirmation of NPP equipment seismic stability includes seismic stability numerical analysis with verifying the computation model against the experimentally determined values of equipment natural frequencies. A way in which the effort taken to carry out the calculation can be decreased by reducing and simplifying the computation model with subsequently adjusting it using the experimentally obtained natural frequencies is demonstrated. Simplification of the computation model is achieved by excluding, with the necessary substantiation, some of pipeline sectors with replacing their effect on the remaining portion by elastic constraints, the stiffness of which is refined based on the experimental values of the equipment natural damped vibration frequencies. A sodium vapor trap jointly with the adjacent extended pipeline segments (the modeling of which is a very labor consuming task) is considered. It is advisable to limit the reduced computation model at the sliding support location points, because this will make it possible to minimize the number of elastic constraints modeling the effect of truncated fragments. The experimentally obtained frequencies are compared with their calculated values. In the considered example, it is found that there is a mismatch between the experimental and calculated values of the lowest frequencies, whereas the higher frequencies are found to be in good agreement with each other, which is attributed to a local nature of disturbances applied in the tests. In the course of simplifying the computation model, the pipeline parts behind the sliding supports were truncated. The effect of the truncated parts was replaced by elastic constrains the stiffness of which was selected proceeding from the criterion of coincidence between the experimental and calculated values of the lowest frequencies. Seismic stability calculations were carried out for the full and simplified computation models. The mechanical stresses arising in the trap filter supporting frame are calculated. The stress values obtained using the simplified model were found to be higher than those obtained using the full model. The maximal stresses estimated from the full and reduced models differ from each other by no more than 8%.