I. I. Kopytov - First Deputy General Director, Atomenergoproject, Moscow
V. Z. Kuklin - Deputy Chief Engineer, Atomenergoproject, Moscow, Russia
A. E. Sargsyan - Professor, Head of the Dynamics and Seismic Safety Dept., Atomenergoproject, Moscow, Russia
A. G. Tyapin - D.Sc., Head of the Lab, Dynamics and Seismic Safety Dept., Atomenergoproject, Moscow, Russia

Three potential sources of the seismic events were chosen for NPP Bushehr, namely Delvar-Ahram, Delvar-Mand and Borazjan-Kazeroon II. The so-called “distributed seismicity” was conservatively enveloped by the Delvar-Ahram events. Two intensity levels were considered for seismic excitation. Safety Categories I and II include the most important elements of the NPP from the safety point of view. Buildings ZA/B, ZE, ZK, ZX-ZY, ZC of the NPP Bushehr belong to these categories. The analysis was carried out in two stages. On the first stage the overall seismic response was obtained using the soil-structure models. The soil was modeled both using the spring-dashpot models (following the recommendations of ASCE4-98) and the advanced SASSI models accounting for the soil layering and the primary non-linearity of the soil. The equipment on this stage was modeled in a simplified way. Structures were modeled using 3D beams and shell/plate finite elements. The dynamic analysis was performed using codes SASSI, STRUDYN, ABAQUS, ANSYS. Floor response spectra were obtained, providing seismic input for the equipment and subsystems. Internal forces were obtained for the design of the structural components. Forces and displacement at the soilstructure interface were estimated to demonstrate the overall seismic stability of the soil-structure system. Benchmark analyses using the data recorded during minor seismic events gave good results. On the second stage the detailed models for the equipment and subsystems were used. Modal method was mostly used for the analysis and design, but sometimes the direct integration was performed also. To account for the uncertainties in the input data the soil modules were varied 1.5 times up and down from the mean level. Besides, the response spectra were broadened by ±15% in the frequency range. The damping values were set reasonably conservative. The authors believe that the non-exceedance probability for the results is minimum 90%, compared to the 84% non-exceedance probability of the seismic excitation.