Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.646365
Title: Wavelet-based characterization and stochastic modelling of avelet-based characterization and stochastic modelling of pulse-like ground motions on the time-frequency plane
Author: Lungu, A.
ISNI:       0000 0004 5362 188X
Awarding Body: City University London
Current Institution: City, University of London
Date of Award: 2014
Availability of Full Text:
Access through EThOS:
Access through Institution:
Abstract:
A novel non-separable non-stationary stochastic model for the representation and simulation of pulse-like earthquake ground motions (PLGMs), capable to accurately represent peak elastic and inelastic structural responses, is proposed in this work. Further, the model is employed for assessing the performance of several time-frequency representation techniques (the harmonic wavelet transform, the Meyer wavelet packets transform, the S-transform and the empirical mode decomposition) in capturing salient features of pulse-like accelerograms. The significantly higher structural demands posed by PLGMs in comparison with similar intensity pulse-free motions led to comprehensive investigations in order to mitigate the damage experienced in the affected areas, such as those located near seismic faults. In this regard, time-frequency analysis methods are frequently employed for the analysis of signals recorded during these events, due to their adaptability to the specific evolutionary behaviour. Alongside with characterization, stochastic modelling of PLGMs is of interest since it allows for systematic variations of the input parameters in order to enhance the understanding of their influence on the structural behaviour. This is particularly useful since only a limited number of PLGMs are available in the existing earthquake databases. Accordingly, inspired by the time-frequency distribution of their total energy, a versatile PLGM model is defined as a combination of amplitude-modulated stochastic processes. Each process models the time-varying distribution of the energy for adjacent frequency ranges. Two alternative formulations are proposed for representing the low-frequency content characterizing the pulses. Considering a set of pulses from the literature, numerical results show that the pulse models‟ parameters can be calibrated to simulate in average the structural impact of these pulses represented using the model herein defined. Further, the capability of the PLGM model to generate elastic and inelastic spectral responses matching a given field recorded accelerogram in the mean sense is illustrated. The applicability of the proposed model to account for near-fault effects to spectrum compatible representations of the seismic action is illustrated by generating a fully stochastic process compatible with the response spectrum of the European aseismic code (EC8). Furthermore, the model can be employed in various applications including generation of accelerograms for nonlinear dynamic analyses of structures, probabilistic seismic demand analyses or as input in stochastic dynamic techniques such as statistical linearization. Finally, the capability of several time-frequency analysis methods to characterize PLGM accelerograms is evaluated through comparative numerical studies within a novel methodology, namely by considering artificial time-histories as samples of the proposed model. The results highlight the potential of the S-transform to be used for pulse identification/extraction and of the harmonic wavelet transform for record characterization/pulse extraction. Additionally, they confirm that from an engineering perspective the structural natural period is an appropriate and representative parameter for the definition of “pulses”. Overall, these analyses shed light into the challenges experienced when attempting to detect the pulse content in the accelerograms, in an effort to inform best practices for PLGMs characterization.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.646365  DOI: Not available
Keywords: TA Engineering (General). Civil engineering (General)
Share: