| The site effect has been a popular topic in the field of seismology and earthquake engineering for a long time.The research of site effect has important theoretical significance and great practical value in near-fault ground motion simulation,ground motion prediction equation,spatial distribution of ground motion,seismic hazard assessment,site selection and seismic fortification of buildings.In this study,we use the strong motion data from KiK-net in Japan to study site classification and site coefficients,linear site response and nonlinear site response.Our aim is to reveal the mechanism of site effect,and to provide suggestions for seismic fortification of buildings to prevent or mitigate hazard of buildings.The main results are as follows.(a)Site classification and site coefficients1.Utilizing 9073 surficial horizontal acceleration recordings from 536 stations,we firstly group the recordings according to site classes,EPAs(efficient peak ground acceleration)and characteristic periods.Then,the design response spectra are studied by the average acceleration response spectra of each group.The results show:(1)The dynamic magnification factor of the platform with different range from 1.81~3.35,with an average of 2.50 and a standard deviation of 0.33.Hence,we recommend the dynamic magnification factor of the platform in design response spectra as 2.50.(2)The provisions for the attenuation relationship of design response spectrum in Chinese code and in American code are too conservative and too risky,respectively.Hence,we recommend the design response spectrum abide a T-1-form attenuation relationship between the characteristic period to 6s.2.Utilizing 10516 surficial horizontal acceleration recordings from 571 stations,we firstly group the recordings according to site classes,PGAs,distances and magnitudes.Then,the site classification and site coefficient are studied by the average acceleration response spectra of each group.The results show:(1)The site classification method based on fundamental period can effectively reduce the deviations of acceleration response spectra.Comparing to the site classification methods in China and American code,the standard deviation of the dynamic magnification coefficient reduce about 15~20% between 0.5~4s for the site classification method based on fundamental period.Therefore,we recommend the site classification method based on fundamental period.(2)Based on the calculation values,we provide suggestion values of site coefficient for different periods,including the site coefficient for the platform(Fac)and the site coefficients at 0.2s,1.0s,2.5s(Fa,Fv,Fd).(b)Linear site response1.Utilizing 288 set weak motions recorded at FKSH14 station,we compared the differences between the predominant frequency and the fundamental frequency.The results show:(1)The predominant frequency may correspond to different vibration modes.It is very likely to confuse the natural frequencies of different vibration modes based on predominant frequency and this would cause that the predominant frequencies between NS and EW components during an earthquake,and between the same component during different earthquakes,have a great difference,and make it difficult to describe the intrinsic characteristics of the site by predominant frequency.(2)There are quite small differences between the predominant frequencies for NS and EW components during an earthquake,and for the same component during different earthquakes.Therefore,we recommend predominant frequency to describe the intrinsic characteristics of the site.2.Utilizing more than 2200 set acceleration recordings from 32 stations,we built empirical relationships between the the ratios of higher order natural frequencies to fundamental frequency and order,and the correlation coefficient is as high as 0.997.Using this empirical relationships,the natural frequencies of the first 17 th or even higher order can be estimated from the fundamental frequency.3.Utilizing more than 20,000 set acceleration recordings from 249 stations,we compared the correlation between site fundamental frequency and different influencing factors,including soil thickness,soil stiffness,stiffness of near surface sediment,and combination of soil thickness and stiffness.The results show that the combination of soil stiffness and thickness has the highest correlation between estimated and measured value,and the correlation coefficients are more than 0.85.The correlation coefficients between site fundamental frequency and soil thickness,stiffness of near surface sediment,soil stiffness are ~0.7,0.6,0.2,respectively.4.Utilizing the same recordings from the 249 stations,we compared the distribution of fundamental periods and the average characteristics of site amplification for different site classes.The main research results are as follows:(1)The mean fundamental periods for class I1,II,III,IV sites in China code are 0.28,0.40,1.06,1.88 s,respectively;the mean fundamental periods for class B,C,D,E sites in America code are 0.15 0.33,0.71,1.20 s,respectively;the mean fundamental periods for class SC I1,SC I2,SC III,SC IV1,SC IV2 sites in this study are 0.16,0.19,0.30,0.43,0.50,1.20 s,respectively.The softer sites have longer fundamental periods and greater deviations of the fundamental period.(2)The main amplification band for or class I1,II,III,IV sites in China code are 2.98~20,1.55~19.48,0.57~13.15,0.31~10.03 Hz,respectively;the main bands for class B,C,D,E sites in America code are 3.37~20,1.91~20,1.00~15.88,0.56~11.98 Hz,respectively;the main amplification band for class SC I1,SC I2,SC III,SC IV1,SC IV2 sites in this study are 3.03~20,2.51~20,1.63~20,1.25~17.36,1.03~16.04,0.55~13.08 Hz,respectively.As the site becomes soft,site amplification moves to low frequencies.The softer the site,the more significant the amplification at the low-frequencies.Moreover,due to the high modes is excited,soft soil sites(class IV sites in China code,class E sites in America code and class SC IV sites in this study)have a certain amplification effect at high frequencies.(c)Nonlinear site response1.Utilizing more than 30,000 set acceleration recordings from 32 stations,using the surface-to-borehole spectra ratio method,the threshold of nonlinear site response can be estimated by applying the bi-linear model to fit the relationships between APNL(Adjusted Percentage of nonlinearity,calculated by the change of the site amplification)and PGA.The results show that the ranges of nonlinearity threshold are between 19 and91 cm/s2,and the mean and standard deviation of thresholds are 60 and 19.9 cm/s2,respectively.The mean thresholds for class I1,II,III sites in China code are 46,62,59cm/s2,respectively;the mean thresholds for class B,C,D sites in America code are 56,59,66 cm/s2,respectively;the mean thresholds for class SC I1,SC I2,SC III,SC IV1,SC IV2 sites in this study are 66,58,73,56,49,71 cm/s2,respectively.There is no significant difference in the non-linear response thresholds for different site classes.2.Utilizing more than 10,000 set acceleration recordings from 8 stations,using the SBSR(surface-to-borehole spectra ratio)and SDP(soil-dynamic parameter)methods,the relationships between fundamental frequency,shear shear modulus and PGA are studied to quantify nonlinear soil behavior.The results show that the degree of nonlinear soil behavior increases as the PGA increases beyond the threshold.When PGA comes to100,300,600 and 800 cm/s2,the FFR(fundamental frequency reduction)is 2~12%,7~29%,13~44% and 17~51% and the SMD(shear modulus degradation)is 3~11%,7~30%,15~55% and 23~71%,respectively.3.Using the regression results of the 8 stations and the laboratory test data of shear modulus ratio curve in Vucetic and Dobry[98](1991),we built a empirical formula to estimate the nonlinear site response threshold and degree.The results show that:(1)the threshold of nonlinear site response is mainly affected by site stiffness and plasticity index.For sites with the same stiffness,as the increase of the plasticity index,nonlinear site response threshold increases.For sites with the same plasticity index,as the increase of the stiffness,nonlinear site response threshold increases.Typically,the site nonlinear site response threshold is between 20~100 cm/s2.When the stiffness increases,and plasticity index decreases,nonlinear site response thresholds may increase or decrease,so the nonlinear site response threshold for hard site may be not higher thanthat for soft site.And this is the reason why there is no obvious differences between nonlinear site response thresholds for different site classes.(2)The shear modulus degradation(as the representation of nonlinear site response degree)is mainly affected by the ground motion intensity,site stiffness and plasticity index.For sites with the same stiffness and plasticity index,as the increase of PGA,the shear modulus degradation increases.For site with the same stiffness and same PGA,as the decrease of plasticity index,the shear modulus degradation increases.For site with the same plasticity index and same PGA,as the decrease of stiffness,the shear modulus degradation increases.When the stiffness increases,and plasticity index decreases,the degree of nonlinear site response may increase or decrease,so the degree of nonlinear site response for hard site may be not higher than that for soft site under same PGA.And this may lead to that there is no obvious differences between the degree of nonlinear site response for different site classes.4.Utilizing the same acceleration recordings from the 8 stations,we observe two stages of nonlinear recovery process after a sharp drop of fundamental frequency and shear-wave velocity during the Tohoku main shock as mentioned by Wu and Peng(2002).The first stage is a rapid recovery within hundred seconds as shown by TFA(time-frequency analysis)method based on short-Fourier and Wavelet transform,and the second stage is a slow recovery of more than 1200 days as shown by SBSR and SDP methods. |
PDF Full Text Request