Stochastic modeling of ground motion is a simple tool to predict ground shaking level for future earthquake and less time consuming than physics-based deterministic modeling.In this paper,a record-based stochastic met...Stochastic modeling of ground motion is a simple tool to predict ground shaking level for future earthquake and less time consuming than physics-based deterministic modeling.In this paper,a record-based stochastic method that considers the time-and frequency-evolution of ground motion is used to estimate ground motion for scenario earthquakes in tectonic active region.The stochastic method employs a time-domain modulation function to describe the temporal nonstationarity and a filter impulse response function that describe the evolution of frequency content.For characterizing the modulation function and the filter impulse function,six parameters(Ia,D5-95,tmid,ωmid,ω',ξf)are defined,and 2,571 pairs of ground motion recording in the NGA-west2 database are selected to identify the six parameters.Probabilistic density function is assigned to each of the parameter by fitting the frequency distribution histogram.The parameters are then transformed into standard normal space where regression analysis is performed by considering each parameter as function of moment magnitude,rupture distance,vS30(The time-averaged shear wave velocity of the top 30 m of soil).The prediction equations are used to generate ground motions for several scenario earthquakes and compared to NGA-West2 GMPEs.展开更多
Recent earthquakes in the Sichuan Province have contributed to significantly expand the existing ground-motion database for China with new,high-quality ground-motion records.This study investigated the compatibility o...Recent earthquakes in the Sichuan Province have contributed to significantly expand the existing ground-motion database for China with new,high-quality ground-motion records.This study investigated the compatibility of ground-motion prediction equations(GMPEs)established by the NGA-West2 project in the US and local GMPEs for China,with respect to magnitude scaling,distance scaling,and site scaling implied by recent Chinese strong-motion data.The NGA-West2 GMPEs for shallow crustal earthquakes in tectonically active regions are considerably more sophisticated than widely used previous models,particularly in China.Using a mixed-effects procedure,the study evaluated event terms(inter-event residuals)and intra-event residuals of Chinese data relative to the NGA-West2 GMPEs.Distance scaling was investigated by examining trends of intra-event residuals with source-to-site distance.Scaling with respect to site conditions was investigated by examining trends of intra-event residuals with soil type.The study also investigated other engineering characteristics of Chinese strong ground motions.In particular,the records were analyzed for evidence of pulse-like forward-directivity effects.The elastic median response spectra of the selected stations were compared to code-mandated design spectra for various mean return periods.Results showed that international and local GMPEs can be applied for seismic hazard analysis in Sichuan with minor modification of the regression coefficients related to the source-to-site distance and soil scaling.Specifically,the Chinese data attenuated faster than implied by the considered GMPEs and the differences were statistically significant in some cases.Near-source,pulse-like ground motions were identified at two recording stations for the 2008 Wenchuan earthquake,possibly implying rupture directivity.The median recorded spectra were consistent with the code-based spectra in terms of amplitude and shape.The new ground-motion data can be used to develop advanced ground-motion models for China and worldwide and,ultimately,for advancing probabilistic seismic hazard assessment(PSHA).展开更多
Currently available earthquake attenuation equations are locally applicable,and methods based on observation data are not applicable in areas without available observation data.To solve the above problems and further ...Currently available earthquake attenuation equations are locally applicable,and methods based on observation data are not applicable in areas without available observation data.To solve the above problems and further improve the prediction accuracy of ground motion parameters,we present a prediction model referred to as a light gradient boosting machine with feature selection(LGB-FS).It is based on a light gradient boosting machine(LightGBM)constructed using historical strong motion data from the NGA-west2 database and can quickly simulate the distribution of strong motion near the epicenter after an earthquake.Cases study shows that compared with GMPE methods and those based on real-time observation data,the model has a better prediction effect in areas without available observation data and can be applied to Yangbi Earthquake and Maduo Earthquake.The feature importance evaluation based on both information gains and partial dependence plots(PDPs)reveals the complex relationships between multiple factors and ground motion parameters,allowing us to better understand their mechanisms and connections.展开更多
基金supported by the National Natural Science Foundation of China(No.51878578).
文摘Stochastic modeling of ground motion is a simple tool to predict ground shaking level for future earthquake and less time consuming than physics-based deterministic modeling.In this paper,a record-based stochastic method that considers the time-and frequency-evolution of ground motion is used to estimate ground motion for scenario earthquakes in tectonic active region.The stochastic method employs a time-domain modulation function to describe the temporal nonstationarity and a filter impulse response function that describe the evolution of frequency content.For characterizing the modulation function and the filter impulse function,six parameters(Ia,D5-95,tmid,ωmid,ω',ξf)are defined,and 2,571 pairs of ground motion recording in the NGA-west2 database are selected to identify the six parameters.Probabilistic density function is assigned to each of the parameter by fitting the frequency distribution histogram.The parameters are then transformed into standard normal space where regression analysis is performed by considering each parameter as function of moment magnitude,rupture distance,vS30(The time-averaged shear wave velocity of the top 30 m of soil).The prediction equations are used to generate ground motions for several scenario earthquakes and compared to NGA-West2 GMPEs.
基金Community Based Disaster Management in Asia Programme Phase Ⅱ (CBDM Asia Phase Ⅱ) (00084327)
文摘Recent earthquakes in the Sichuan Province have contributed to significantly expand the existing ground-motion database for China with new,high-quality ground-motion records.This study investigated the compatibility of ground-motion prediction equations(GMPEs)established by the NGA-West2 project in the US and local GMPEs for China,with respect to magnitude scaling,distance scaling,and site scaling implied by recent Chinese strong-motion data.The NGA-West2 GMPEs for shallow crustal earthquakes in tectonically active regions are considerably more sophisticated than widely used previous models,particularly in China.Using a mixed-effects procedure,the study evaluated event terms(inter-event residuals)and intra-event residuals of Chinese data relative to the NGA-West2 GMPEs.Distance scaling was investigated by examining trends of intra-event residuals with source-to-site distance.Scaling with respect to site conditions was investigated by examining trends of intra-event residuals with soil type.The study also investigated other engineering characteristics of Chinese strong ground motions.In particular,the records were analyzed for evidence of pulse-like forward-directivity effects.The elastic median response spectra of the selected stations were compared to code-mandated design spectra for various mean return periods.Results showed that international and local GMPEs can be applied for seismic hazard analysis in Sichuan with minor modification of the regression coefficients related to the source-to-site distance and soil scaling.Specifically,the Chinese data attenuated faster than implied by the considered GMPEs and the differences were statistically significant in some cases.Near-source,pulse-like ground motions were identified at two recording stations for the 2008 Wenchuan earthquake,possibly implying rupture directivity.The median recorded spectra were consistent with the code-based spectra in terms of amplitude and shape.The new ground-motion data can be used to develop advanced ground-motion models for China and worldwide and,ultimately,for advancing probabilistic seismic hazard assessment(PSHA).
基金supported by the National Key R&D Program of China(No.2017YFB0504104)the National Natural Science Foundation of China(No.41971280)。
文摘Currently available earthquake attenuation equations are locally applicable,and methods based on observation data are not applicable in areas without available observation data.To solve the above problems and further improve the prediction accuracy of ground motion parameters,we present a prediction model referred to as a light gradient boosting machine with feature selection(LGB-FS).It is based on a light gradient boosting machine(LightGBM)constructed using historical strong motion data from the NGA-west2 database and can quickly simulate the distribution of strong motion near the epicenter after an earthquake.Cases study shows that compared with GMPE methods and those based on real-time observation data,the model has a better prediction effect in areas without available observation data and can be applied to Yangbi Earthquake and Maduo Earthquake.The feature importance evaluation based on both information gains and partial dependence plots(PDPs)reveals the complex relationships between multiple factors and ground motion parameters,allowing us to better understand their mechanisms and connections.
