The seismic risk of a nuclear power plant (NPP) depends on structures, systems and components (SSCs) that are seismically qualified to a design basis earthquake (DBE) in Canada or a safe shutdown earthquake (SSE) in the United States. On the other hand, there exist some components that are not essential to safety but their seismic interaction with seismically-qualified SSCs adversely affects the seismic risk of such SSCs. Rocking frames consisting of a rigid beam freely supported by piers (e.g., a 150 Ton spare turbine rotor, or a 100 Ton idle steam generator resting on triangular or trapezoidal rigid piers) are common to NPPs. Seismic interaction of such frames with seismically-qualified safety components or their host structure may be detrimental to nuclear safety as witnessed in the 2013 Arkansas Nuclear One accident where the drop of a 500 Ton stator adversely impacted the severe core damage frequency of the entire plant, negatively affecting the nuclear risk. In order to ensure nuclear safety, it is essential to quantify the risk of a heavy component’s drop owing to a rocking frame’s instability caused by design basis accidents including seismic. A rocking frame’s beam support may be concentric or eccentric with respect to the pier’s center of mass depending on it’s geometry, for example, triangular or trapezoidal respectively. The current nuclear standards, ASCE 43-19 and CSA N289.1-2018 are silent about rocking frames. Literature has also not addressed the eccentricity variation. This thesis addresses the gap on seismic qualification of rocking frames by, establishing an equivalent rocking block for rocking frames with symmetrical support eccentricities, obtaining the response of frames with unsymmetrical support eccentricities and finally examining the stability of the two types of frames under slide restrained conditions. / Thesis / Doctor of Philosophy (PhD) / Rocking frames, each consisting of a heavy rigid horizontal beam freely supported on unanchored rigid piers, are common to nuclear power plants (NPPs) (e.g., a turbine rotor freely supported by triangular or trapezoidal piers). The support points for the beam on the pier in such frames may be concentric or eccentric with respect to the pier’s center of mass as in a triangular or trapezoidal pier configuration. The current Canadian and American nuclear standards do not provide guidance on rocking frames. Support eccentricity variation has not been addressed in the literature. Consequently, the seismic risk of rocking frame configurations, common to NPPs, remains unknown. This thesis addresses this gap by employing an equivalent rocking block model for frames with symmetrical eccentricities, with an equation of motion representing those with unsymmetrical eccentricities; and examining the stability of the two under slide-restrained conditions.

Seismic response of free standing un-anchored objects is required to be studied in Nuclear Power Plants (NPPs) for their own integrity and potential interaction with the surrounding seismically qualified safety systems. Rocking response of a rigid body subject to seismic excitation is not very well covered in the nuclear standards except for an approximate method given in ASCE 43-05 where the design basis earthquake (DBE) response spectrum for the NPPs given in the United States Nuclear Regulatory Commission (USNRC) regulatory guide 1.60 (known as NBK spectrum developed by Newmark, Blume and Kapur (1973)) is considered as seismic input. This study evaluates existing seismic design criteria for unanchored objects that are vulnerable to rocking and overturning inside nuclear power plants. The original work of Newmark et al (1973) is revisited in order to obtain the NBK spectra at unusual damping (8.4% for example), required in order to follow the ASCE 43-05 method. Eight earthquake records are selected from Newmark et al (1973) with varying Peak Ground Accelerations (PGAs) representing strong to moderate ground motions. Rocking response of rigid bodies to various earthquakes is determined by three methods: 1. The ASCE 43-05 method utilizing the NBK spectrum, 2. The ASCE 43-05 method utilizing the response spectrum of the earthquake records, and 3. solving the equations of motion of a rigid body for the earthquake records. Rocking spectra by these three methods created for eight earthquake records are compared with one another. It is concluded that the ASCE 43-05 method provides inaccurate predictions of the response. Considering the significant level of effort required to implement the ASCE 43-05 method, its inherent contradictions, and its inconsistent conservatism in estimating the seismic demands on rocking objects, it is concluded that the results obtained by nonlinear time history analysis are more accurate, reliable and less time consuming than those by the ASCE 43-05 method. The use of nonlinear dynamic analysis is recommended to obtain the pure planar rocking response of unanchored objects in nuclear facilities. / Thesis / Master of Applied Science (MASc)