A Study On Damping Correction Factor For Displacement Spectrum Of Subduction Slab Earthquakes

In recent years,frequent large earthquakes in China have caused a large number of casualties and economic losses.With the acceleration of social development and urban construction,China is still experiencing the construction of large-scale infrastructure and other buildings.To improve the seismic performance of buildings has become an important part of engineering designs.At present,the seismic design specification in China is based on the design spectrum with a damping ratio of 5%,and the seismic design of buildings with various damping ratio can be derived by the damping adjustment factor(referred as to B_d hereafter).The form of damping adjustment factor in the design code is relatively simple,and the simple form may lead to erroneous estimates for the response of structures with different damping ratios.In order to derive a set of displacement spectra with a desired damping ratio from a design 5%damped spectrum,a set of damping correction factors for a displacement spectrum are presented in this thesis.For subduction regions,such as in Japan and Taiwan,the response spectrum from different types of earthquakes differs significantly and subduction slab arthquakes are one of four types of earthquakes.In this thesis,firstly,the research work in he past thirty years was reviewed.Secondly,4609 horizontal ground-motion records obtained at the K-NET and KiK-net strong-motion networks generated by 135 subduction zone ntraplate earthquakes were selected and processed.Thirdly,the effects of other ground- otion parameters on B_d and the functional forms for B_d were evaluated systematically and extensively.Finally,regression analyses were performed to derive the model coefficients and various standard deviations for B_d.In this thesis,the model for B_d consists of two parts:the mean model and the model accounting for earthquake source,path,and site effects.The mean model,which is the geometric mean of B_d of all records in a given spectral period and damping ratio ranges suitable for adjusting a design response spectrum,has a imple form.The recording stations were divided into four site classes according to the study by Zhao et al.(2016),Statistical studies,including F-test and t-test,were used to confirm the necessity of site classification for B_d.Based on the data distribution and physical restraint for B_d with respect to damping ratio,a quadratic polynomial with a natural logarithm of damping ratio was adopted to perform the mean model regression of B_d on the basis of the least squares method.The regression coefficients were smoothed over the spectral periods,and the egression model has a good fit to the values for B_d derived from the strong-motion records.The second B_d model was used to account for the effects of earthquake sourceparameters including magnitude and fault depth,path parameters including source distance and volcanic distance,and site effect represented by site classes.The model parameters were derived by using a random effects model.Based on the distribution of residuals with respect to each physical parameter and the functional form used in the GMPE by Zhao et al.(2016),a bilinear magnitude scaling function,a bilinear depth scaling function,a geometrical ttenuation term,an anelastic attenuation term,linear term for volcanic distance and four constants representing the four classes of site parameters were included in the model.The final model is the sum of the mean model and the second model to derive smoothing and applicable displacement design spectra with various damping ratios.Through the analysis of regression process and the residuals of expression,the following conclusions can be reached:1.B_d.derived in this thesis depends on spectral period,damping ratio,site class,moment magnitude,fault depth,fault distance,and volcanic distance.The parameters selected in this study can effectively reduce the biased residuals distribution with respect to each model arameter,leading to reasonable estimates for a response spectrum within the damping ratio ranges of 0.01-0.30.2.The model fits the data from site class I and II sites better than to those from the class III and IV sites.At most the spectral periods,the between-event standard deviations representing the level of random errors in the modeling of source effect)from the source effect are smaller than those from the other effects.At spectral periods less than 0.12s,the etween-site standard deviations representing the random errors associated with site modeling are larger than those from the other effects.At spectral periods over 0.12s,the between-site tandard deviations are smaller than those from the path effect.3.Overall,the standard deviations associated with damping adjustment factor are not large,having a factor of about 1.30 for a median±σat 0.1s and a damping ratio of 0.3.