Title :
Aseismic designs based on artificial simulations
Author :
Wang, Junji ; Zhou, Jing
Author_Institution :
Tongji Univ., Shanghai, China
fDate :
3/1/1999 12:00:00 AM
Abstract :
It is commonly believed that the time-varying frequency characteristics are important for the inelastic response of structures. Design code modification means structural cost changes; therefore any change must be based on solid evidence. Although the significance of time-varying frequency characteristics to ground motion has been recognized for a while, much more in-depth research is necessary before any meaningful engineering design conclusions can be reached. We apply joint time-frequency analysis techniques to a fully non-stationary ground motion model, with both intensity and frequency being non-stationary. The chirplet-based signal approximation is used to extract the time-varying frequencies from seismic ground-motion data samples. Based on that information, we compute the frequency-dependent modulating function in the stochastic model of the evolutionary random process. We generate and compare the artificial waves based on stochastic models of a uniform modulating random process and an evolutionary random process, separately. We also investigate the importance of the time-varying frequency characteristics of earthquake ground motion through an oscillator, with a single degree of freedom and elastic-plastic material behavior
Keywords :
approximation theory; digital simulation; geophysical signal processing; modulation; random processes; seismology; stochastic processes; time-frequency analysis; artificial simulations; artificial waves; aseismic designs; chirplet-based signal approximation; earthquake ground motion; elastic-plastic material behavior; engineering design; evolutionary random process; frequency-dependent modulating function; inelastic response; intensity; joint time-frequency analysis; nonstationary ground motion model; oscillator; seismic ground-motion data samples; stochastic model; structural response; time-varying frequencies; time-varying frequency characteristics; uniform modulating random process; Character recognition; Chirp modulation; Costs; Data mining; Design engineering; Motion analysis; Random processes; Solids; Stochastic processes; Time frequency analysis;
Journal_Title :
Signal Processing Magazine, IEEE