Poster #081, Earthquake Geology

Shake Table Tests and Numerical Modeling for Precarious Rock Fragilities

Christine Wittich, & Muhammad Khalid Saifullah
Poster Image: 

Poster Presentation

2021 SCEC Annual Meeting, Poster #081, SCEC Contribution #11646 VIEW PDF
A precariously balanced rock (PBR) is an individual or group of rocks that has eroded into an unstable configuration. Given that the ages of many of these features have been established to be in excess of 10 – 30 ka, precarious rocks and other fragile geologic features are one of the only available means to validate seismic hazard associated with long return periods. Current state-of-the-art methods for predicting overturning of a precarious rock include detailed surveying of the rock’s geometry followed by numerical simulations and ultimately fragility analysis, in which the probability of overturning is related to a measure of earthquake intensity. However, there are significant sources of... uncertainty at each analysis stage of the PBR, which impact the resulting probabilities of overturning to unknown extents. As a result, the long-term goal of this research is to reduce epistemic uncertainty in the probabilistic overturning predictions for PBRs. In an effort to quantify the uncertainty and better understand the behavior of these structures, an extensive shake table study is ongoing at the University of Nebraska Lincoln’s shake table facility with natural rocks used as representative specimens of PBRs and preliminary results of this shake table testing campaign are presented herein. The key parameters being investigated include the overall rock geometry, interface geometry, and material. Hundreds of individual shake table tests, under a range of ground motions covering relevant intensity parameters (PGA and PGV/PGA), are conducted on an individual rock specimen. The overturning results of an individual specimen are analyzed in concert with the results of corresponding distinct element method (DEM) models incorporating varying numerical parameters, which help to identify a range of reasonable numerical modeling parameters as well as the uncertainty in overturning probabilities arising from such modeling.