2016 Theses Doctoral

# Stochastic Characterization and Simulation of Ground Motions based on Earthquake Scenarios

A novel stochastic earthquake ground motion model is formulated in association with physically interpretable parameters that are capable of efficiently characterizing the complex evolutionary nature of the phenomenon. A multi-modal, analytical, fully non-stationary spectral version of the Kanai-Tajimi (K-T) model is introduced achieving a realistic description of the evolutionary spectral energy distribution of seismic ground motions. The functional forms describing the temporal evolution of the model parameters can efficiently model highly non-stationary power spectral characteristics. The analysis space, where the analytical forms describing the evolution of the model parameters are established, is the energy domain instead of the typical use of the time domain. This space is used in conjunction with a newly defined energy-associated amplitude modulating function. The Spectral Representation Method supports the simulation of sample ground motions realizations. A predictive stochastic model for simulation of earthquake ground motions is developed, using a user-specified earthquake scenario description as input, and resulting in fully nonstationary ground acceleration time-histories at a site of interest. The previously formed analytical non-stationary K-T ground motion model lies at the core of the developed predictive model. An extensive Californian subset of the NGA-West2 earthquake ground motion database is used to develop and calibrate the predictive stochastic model. Sample observations of the model parameters are obtained by fitting the K-T model to the database records, and their resulting marginal distributions are effectively described by simple probability models. Advanced random-effect regression models are established in the normal probabilistic space, capable of linking the stochastic K-T model parameters with the moment magnitude Mw, closest distance Rrup and average shear-wave velocity VS30 at a Californian site of interest. The included random effects take effectively into account the correlation of ground motions pertaining to the same earthquake event, and the fact that each site is expected to have its own effect on the resulting ground motion. The covariance structure of the normal K-T model parameters is next estimated, allowing finally for the complete mathematical description of the predictive stochastic model for a given earthquake scenario. The entirety of the necessary steps for the simulation of the developed predictive stochastic model is provided, resulting in the generation of any number of fully non-stationary ground acceleration time-series that are statistically consistent with the specified earthquake scenario. In an effort to assess the performance and versatility of the developed predictive stochastic model, a list of simple engineering metrics, associated with the characterization of the earthquake ground motion time-series, is studied, and results from simulated earthquake ground acceleration time-series of the developed predictive model are compared with corresponding predictions of pertinent Ground Motion Prediction Equations (GMPEs) for a variety of earthquake and local-site characteristics. The studied set of ground acceleration time-series features includes the Arias intensity IA, the significant duration T5-95 of the strong ground shaking, and the spectral-based mean period of the earthquake record Tm. The predictive stochastic model is next validated against the state-of-the-art NGA-West2 GMPE models. The statistics of elastic response spectra derived by ensembles of synthetic ground motions are compared with the associated response spectra as predicted by the considered NGA-West2 ground motion prediction equations for a wide spectrum of earthquake scenarios. Finally, earthquake non-linear response-history analyses are conducted for a set of representative single- and multi-degree-of-freedom hysteretic structural systems, comparing the seismically induced inelastic structural demand of the considered systems, when subjected to sets of both real strong ground motion records, and associated simulated ground acceleration time-histories as well. The comparisons are performed in terms of seismic structural demand fragility curves.

## Files

- VLACHOS_columbia_0054D_13524.pdf application/pdf 6.24 MB Download File

## More About This Work

- Academic Units
- Civil Engineering and Engineering Mechanics
- Thesis Advisors
- Deodatis, George
- Degree
- Ph.D., Columbia University
- Published Here
- August 17, 2016