Research On Selection Of Ground Motion Records Considering Site Soil Effect And Its Uncertainty

The first step of seismic design is to determine the design ground motion parameters,while reasonable ground motion inputs are a necessary condition to ensure the correct design results.A large number of strong earthquake observation data and earthquake damage investigations show that the local site condition has a greater impact on the propagation of seismic waves,which mainly changes the intensity and frequency spectrum characteristics of ground motions.At present,Chinese Code for Seismic Design of Buildings(GB 50011-2010)only considers the impact of the site soil by adjusting the characteristic period of the site,while the Seismic Ground Motion Parameter Zonation Map of China(GB18306-2015)gives the impact coefficients of peak ground acceleration(PGA)and the characteristic period adjustment values of response spectrum for different types of sites in China,they are all average estimates of general site soil effects.The actual site has complex geological conditions and variable soil properties,therefore,the determination of ground motion parameters for major engineering designs in specific areas should not be limited to the empirical statistical values given in the seismic codes,they should be based on specific probabilistic seismic hazard analysis and probabilistic site soil effect analysis.However,how to select natural ground motion records as the bedrock seismic excitations?How many records to select in the site seismic response analysis?How to evaluate the design ground motion parameters considering the uncertainty of the site soil effect?For these questions,there are still no clear principles to follow.Moreover,during the time history analysis of an above-ground structures,the uncertainty of the input ground motions has a decisive effect on the responses of the structure.However,Chinese seismic design codes give too indistinct and obscure requirements for the selection of strong motion records.Therefore,how to select ground motion records reasonably has become an urgent problem in the current structural seismic design.In view of the above reasons,this thesis mainly investigate the uncertainty of the site soil effect and the selection methods of ground motion records in the nonlinear time history analysis of structures.The research purposes are to analyze the influence of soil variability on the ground motions of the site,give suggestions on the selection methods for ground motion records of bedrocks in seismic response analysis of the site,and evaluate the design ground motion parameters of the study area considering the uncertainty of site soil effects.The selection methods of ground motion records are evaluated by the relative error and dispersion of the calculated values of structural responses,and a set of ground motion records suitable for the seismic design of the research site engineering is given.The main research contents of this thesis are as follows:(1)Two site seismic response analysis programs,DEEPSOIL and Open Sees,are used to establish three different types of II,III,and IV representative sites.While verifying the proper application of the programs,the surface responses of the site obtained by two commonly used methods of site response calculation,namely equivalent linearization method and time-domain nonlinear method,are compared.The results show that site conditions have an impact on the intensity and spectral characteristics of ground motions,and the response spectrum shapes obtained by the different calculation methods of the two programs are similar,but the difference in high frequency and amplitude are obvious.Combined with the properties of soil,the Open Sees with soil elastoplastic constitutive is selected for the uncertainty analysis of site seismic responses.(2)The influences of soil variability on seismic responses of the site are investiagted.Taking the ordinary site as the study object,the influence of the inter-layer correlation and dispersion of the soil shear wave velocity and the variability of the soil nonlinear dynamic characteristic curves on surface responses of the site are analyzed.It is found that soil variability mainly affect the peak values of the amplification coefficients of the surface response spectra and the high frequency range of the response spectra.Therefore,if the site seismic response analysis only relies on the average estimation of the soil parameters,then the ground response may be overestimated.Among the influencing factors of soil parameters,the shear wave velocity of the soil layer has a greater influence than the dynamic characteristics of the soil.(3)The methods of selecting bedrock ground motion records in site response analysis are investigated.The ZK8 borehole location in Yanliang District of Xi'an city is taken as the study object,and the consistent hazard spectrum of the bedrock given by the probabilistic seismic hazard analysis is chosen as the target spectrum.Based on the set seismic parameters and bedrock site conditions,a ground motion record library is initially selected from the PEER strong earthquake database.Using six scaling methods including single-point scaling and multi-point scaling,and two period matching methods of full period and local period near the excellent period of the site,the ground motion records matching the target spectrum are selected as bedrock ground motion inputs.The influences of different wave selection methods and the number of selected waves on the surface responses of the site are analyzed.Finally,it is recommended to use the equal-spectrum intensity method(multi-point)scaling and the full-period matching method to select 15 ground motion records consistent with the risk level of the design bedrock site.(4)Different ground motion selection and scaling methods are evaluating by use of the uniform hazard spectrum(UHS)considering the uncertainty of the site soil effect as the design parameters.The UHS on the surface of the investigation area is obtained through probabilistic site response analysis.The convolution method and the hybrid method are used to calculate the surface seismic hazard curves,and the 84%quantile UHS is selected as the design spectrum of the investigation site.Based on the ideal elastoplastic single-degree-of-freedom(SDOF)systems,the averag real responses of the structures are predicted under a set of earthquake inputs using the“comparison point”method.Different methods to select ground motion records are used to input the ground motion records into the SODF systems.The relative errors and the dispersions of the structure response values are used as the evaluation indicators to summarize the relatively reasonable ground motion selection methods for different natural vibration periodic structures.The analysis results show that for the short-period structures,the single-point Sa(T₁)scaling method is similar to the traditional PGA scaling method.For the medium and long-period structures,it is recommended to choose the single-point Sa(T₁)scaling method for the selection of ground motion records.