Consideration of Non-Uniform Embedment on the Response of Nuclear Plant Structures

The effects of foundation embedment on the seismic response of structures have long been recognized in the analysis and design of nuclear power plant (NPP) structures. Some design standards allow simple methods to address the embedment effects. However, these methods are only suitable to consider the effects of soil embedment on uniformly embedded structures. There is no prescribed procedure to consider the embedment effects on non-uniformly embedded structures, which are typically found in many NPPs. Systematic sensitivity studies could prove useful in such cases to identify an appropriate approach which reasonably considers the embedment effects. This paper examines the seismic soil-structure interaction (SSI) response of two non-uniformly embedded structures as case studies. The first case includes three rock-founded Reactor Buildings (RBs) which are constructed side by side and are embedded in soil. Two smaller Diesel Generator Buildings (DGBs) are founded on the sides of the RB block at grade level. The objective of this study is to determine whether an SSI model that includes one RB only, instead of a complex model including all buildings, could reasonably provide the response in all three RBs, despite the differences in embedment configurations and potential SSSI effects in each of the RBs. It is shown that the SSI model that only includes embedment on two sides of a single RB, while neglecting the DGBs, can reasonably represent all three RBs, without meaningfully sacrificing accuracy. The second study includes an intake structure (IS) founded on bedrock and embedded in soil on three sides, while facing water on the forth side. It is shown that the treatment of the intake structure as partially embedded on 3-sides with the top 20 ft. of embedment separated from the building produces reasonable results. An important conclusion drawn from these case studies is that non-uniform embedment can be reasonably considered in many cases when using equivalent linear analysis tools, and that conducting systematic sensitivity studies can result in efficient and cost-effective solutions for idealizing such embedment conditions.