Correlation Of Ground Motion Parameters And Their Relationship With Seismic Intensity

Due to the influence of seismic source characteristics,propagation path,site conditions and other factors,ground motion shows a high degree of complexity and uncertainty.In order to quantitatively analyze the destructive effect of ground motion,it is necessary to select appropriate parameters to describe the features of ground motion.Different ground motion parameters may be independent or highly correlated,and it is very meaningful to evaluate the applicability of ground motion parameters from the perspective of correlation between ground motion parameters.In addition to the general correlation,there is also a certain spatial correlation for ground motion parameters.For general small and medium-sized industrial or civil buildings,the change of spatial characteristic distribution of ground motion does not need to be considered.However,for long-span bridges,oil pipeline networks,nuclear power plants and other complex lifeline projects or special structures,considering the spatial distribution characteristics of ground motion and studying the spatial correlation of ground motion parameters is particularly important.In addition,when analyzing the destructive effect of ground motion,besides selecting a specific ground motion parameter as a quantitative index,we usually choose seismic intensity,which is a simple macro scale parameter,to characterize the magnitude of seismic influence.How to convert the quantitative index of ground motion parameters into the qualitative index of seismic intensity and establish the conversion relationship between them is also of great research significance for engineering applications such as rapid assessment of earthquake disasters and losses(for example,the generation of seismic maps),recurrence of historical earthquakes and so on.So in this paper,we chose the correlation of ground motion parameters and the conversion relationship between seismic intensity and ground motion parameters as the research subject.Based on the strong motion records of several strong earthquakes in mainland China and NGA-West2 database,we systematically analyzed the correlation and spatial correlation of ground motion parameters,as well as the correlation between seismic intensity and typical ground motion parameters,established the correlation matrix and spatial correlation model,and obtained the conversion relationship between seismic intensity and typical ground motion parameters.The main work and research results of this paper are as follows:Firstly,the correlation analysis of ground motion parameters was carried out.Based on the recent strong motion records of 10 moderate-strong earthquakes in mainland China,we selected several typical ground motion parameters,calculated their correlation coefficients,and obtained the correlation coefficient matrix for 27 parameters.We also drew their correlation curves,analyzed the correlation characteristics in detail,and compared the differences with the research conclusions of other scholars.The results showed that:(1)among the six ground motion parameters of PGA,PGV,AriasIM,SI,CAV and tD5-95,AriasIM had the strongest correlation with CAV,where the correlation coefficient was as high as 0.9841;the correlation among PGA,PGV,AriasIM,SI and CAV was relatively strong,where their correlation coefficients were all more than 0.6;the correlation of tD5-95 with other parameters was weak,where the correlation coefficient values were below 0.4,and negatively correlated with PGA.(2)The correlation coefficient of Sa(T)at different periods in the range of 0.1～5.0s was related to the period interval.The smaller the period interval,the greater the correlation coefficient.(3)The correlation between PGA and Sa(T)was high in short period,and decreased gradually with the increase of period;PGV and SI had similar correlation characteristics with Sa(T),the correlation coefficient increased at first and then decreased with the increase of period;AriasIM and CAV also had similar correlation characteristics with Sa(T),where the correlation coefficient in the period of 0.1～1.0s was stable at 0.62～0.74,and then gradually decreased;TD5-95 and Sa(T)showed a weak correlation,their correlation coefficient was less than 0.2,and negative in the period of 0.1～0.6s.(4)Although the correlation characteristics obtained in this study had a similar trend with the previous results,there were also large differences in detail,indicating that the correlation of ground motion parameters had significant regional characteristics.Secondly,the spatial correlation of ground motion parameters was studied.Using the strong motion records of the Chi-Chi earthquake in Taiwan,Chuetsu-Oki earthquake in Japan,EI Mayor-Cucapah earthquake in southern California and Wenchuan earthquake in mainland China,we calculated the spatial correlation coefficients of ground motion parameters such as PGA,PGV,and Sa(T=0.3s,0.5s,1.0s,2.0s),obtained the spatial correlation models,and analyzed the correlation characteristics quantitatively.Then we discussed the influence of ground motion prediction equation on the models and compared our models with others.The research showed that:(1)for peak parameters PGA,PGV and response spectrum parameter Sa(T),their spatial correlations all decreased gradually with the increasing inter-station separation distance,and the decay rates varied by region.Among the study areas listed in this paper,southern California had the fastest decay rate and smallest spatial correlation distance,while the opposite was true for the central-Kanto-northeastern region of Japan.In addition,the spatial correlation model of Sa(T)also had a significant period dependence.Within the natural vibration period range(0.3～2.0s)selected in this study,the decay rate of spatial correlation gradually slowed down and the spatial correlation distance correspondingly got bigger with the increase of period.(2)The spatial correlation model was not very sensitive to the ground motion prediction equation,different prediction equations could be used to participate in the calculation of spatial correlation model on the premise of ensuring the validity of the prediction equation.(3)The spatial correlation of ground motion parameters had obvious region dependence.Affected by the overall geological conditions of different regions,the results of our models and previous models showed obvious differences,and their spatial correlation distances were also different.Thirdly,the correlation and conversion relationship between seismic intensity and typical ground motion parameters were studied.By collecting the strong motion records and intensity data of 11 moderate-strong earthquakes in mainland China since 2008,we statistically analyzed the correlation between PGA,PGV,PSA(T=0.3s,1.0s,2.0s,3.0s)and seismic intensity,and fitted the conversion relationships between them using orthogonal regression.We also discussed the influence of magnitude,distance and other factors on the relationships,compared the difference of relationships in this study and other researches,and analyzed their intensity prediction effect.The main conclusions showed that:(1)the dispersion of ground motion parameters in each intensity level was large,but its mean value had a good positive correlation with seismic intensity,which was the basis of our research on the conversion relationship between seismic intensity and ground motion parameters.In addition,with the change of building structure types and the improvement of seismic capacity in recent years,the distribution of PGA and PGV sample data in each intensity level obtained by using the latest seismic data in China had changed in a certain extent compared with the previous research results.(2)Using the orthogonal regression method,we obtained the reversible conversion relationships between seismic intensity and PGA,PGV,as well as PSA(0.3s),PSA(1.0s),PSA(2.0S)and PSA(3.0s),which could not only provide a necessary reference for the revision of seismic intensity scale in the future,but also be conveniently applied to engineering applications such as shakemap drawing,seismic disaster and loss assessment,history earthquake analysis and so on.(3)Magnitude,hypocentral distance,hypocentral depth and other factors had a certain impact on the distribution of prediction residuals in the conversion relationship.The introduction of magnitude-distance-depth correction term could eliminate the dependence of prediction residuals on the above factors and reduce its standard deviation in a certain extent.(4)The conversion relationships between seismic intensity and ground motion parameters had significant region dependence,and the research results were different in varying areas.(5)When using the conversion relationship in this study for intensity prediction,the prediction effect of PGA and PSA(0.3s)were good,PGV and PSA(1.0s)followed,PSA(2.0S)and PSA(3.0s)were poor;if PGA and PGV were considered at the same time,the prediction performance would be greatly improved,and the prediction effect was better than that when a single parameter was used alone.