Site-response and soil-structure interaction analysis
Soil-structure interaction (SSI) analysis is key to calculating the seismic demands on nuclear structures, systems and components (SSCs). Soil-structure interaction analysis is often preceded by site-response analysis, in order to calculate the foundation input ground motion. Current practice of site-response and SSI analysis in the U.S. nuclear industry involves the use of linear or equivalent-linear analyses, and the assumption that the seismic demand increases linearly with the intensity of the earthquake motion. In reality, the behavior of soil and structural materials is nonlinear and hysteretic, especially under intense shaking. The nonlinear hysteresis results in additional dissipation of energy and can reduce the SSC demands. Additionally, the foundation of the structure can slip and lose contact with the soil during an earthquake resulting in gapping and sliding. These nonlinear phenomena can only be captured using a nonlinear analysis.
Nonlinear site-response and SSI analyses have been performed for some buildings, bridges and petrochemical structures, but not yet for nuclear structures due to the lack of experience with analysts and regulators. To enable a more widespread use of nonlinear SSI analysis, the Seismic Research Group has developed a nonlinear SSI analysis methodology (termed as NLSSI). The NLSSI methodology is a series of steps that can be used with any structure or soil profile, and implemented in any finite element code to perform nonlinear SSI analysis. The Seismic Research Group has implemented the NLSSI methodology in the commercial finite-element codes ABAQUS and LS-DYNA and performed NLSSI analysis for a range of soil sites in the United States. The Seismic Research Group has provided guidance for performing nonlinear SSI analysis in Appendix B to the upcoming version of ASCE 4.
Verification and benchmarking of NLSSI methods and tools
Model verification and benchmarking are crucial aspects of any finite-element analysis. Verification is the process of showing that the implemented numerical solution solves the associated mathematics correctly. Benchmarking involves a comparison of results with other verified solutions or numerical codes. The Seismic Research Group is currently benchmarking the NLSSI methodology (implemented in ABAQUS and LS-DYNA) against the industry-standard code, SASSI for a range of sites in the United States. Benchmarking for a Western US site was carried out in FY 2014. Benchmarking for a Central and Eastern US site is currently ongoing.
Benchmarking ABAQUS and LS-DYNA against SASSI for NLSSI analysis revealed a number of potential issues that an analyst needs to be aware of while using these codes. These issues are documented here.
Gapping and sliding
Foundation gapping and sliding can significantly affect the demand on SSCs but is currently not well understood. Gapping and sliding are highly nonlinear behaviors and simulating these phenomena in nuclear structures can be a challenging task. The Seismic Research Group is investigating 1) efficient ways of modeling the foundation-soil interface to simulate gapping and sliding, and 2) the significance of gapping and sliding in the SSC response for various foundation conditions. Laboratory experiments will also be performed at the University at Buffalo to validate these foundation-soil interface models.
Effect of heterogeneity
Current site-response and SSI analyses are performed assuming that soil profiles are composed of infinitely horizontal, uniform layers. The Seismic Research Group is currently investigating the effect of heterogeneities (such as basalt interbeds, non-horizontal soil layers, etc.) on the surface ground motion. Two-dimensional and three-dimensional site-response analyses are currently being performed using ABAQUS and MOOSE to investigate this issue.