Selection Of Rare Ground Motions For Time History Analysis Of Structures

Selecting ground motions for dynamic time history analysis of structures is a hotsubject since dynamic time history was putted forward. Rational earthquake records isthe base of dynamic time history. On this topic, considerable researches have benperformed and some methods, such as those based on site type, area of responsespectrum and characteristic period of site and the dual-period-range method are putforward. Some questions are gradually arising with more thoroughly studiesaccomplished. First, exsiting methods mainly are mainly basaed on elasic spectrumwhich can not distigush the differences of earthquake records with different intensity ofground motions. Second, some methods are related to the structures, so the selectedrecords may be different if the structure is different but located in the same site which iscontray to the commom sense.To address above problems, this paper mainly aims at the selection of groundmotions for dynamic time history, main works and new ideas can be concluded asfollows:①Selection of ground motions for dynamic time history analysis of structursunder rare seismic excitationAs much work have been done on the modeling of ground motions, this paperraised a selection method based on model parameters.The parameterized method consistof two steps.In the first step:1) the parameters’ varying ranges are determined accordingto the seismic fortification intensity, intensity of ground motion, site type, designearthquake group and attenuation models;2) motion records are selected from a motiondatabase according to the ranges of parameters. In the second step, according to theprinciple of statistical consistency, a best combination of records is found from therecords of the first step by using the optimization method to ensure the mean responsespectram the best compatibility with the target response spectram.②Verification and assesment of the methods putted forward in the paperThe results of the double frequency method, ASCE-7method and theparameterized method are compared. First, comparison of the elstic accelation responespectram,elastic displacement response spectram,inelastic accleraion responsespectram,inelastic displacement response spectram and the input energy reponsespectram of the records selected by three different methods is made. Second, the drift, displacement, shear force, hysteretric energy and the stiffness degradation confficient ofthree different structures under the accelegroms selected by the different three methodare compared. The three structures are, respectively, a5-story frame structure, a10-story frame structure and a33-story frame-core wall structure.③Discussion on the scaling of accelerogramsThe selection and amplitude scaling of accelerograms are usually needed to becollocated when conduct the dynamic time history analysis.This paper have a discussionon the feasibility of appointing the liner coefficients of the target response spectram andthe real accelegrams response spectram as the amplitude sacling factors, the amplitudefactors are usually closely to1. By comprasion, the mean response spectram is notaffcetd, but the standerd deviation can be reduced if the accelegroms are multiplied bythe amplitude scaling factors, which is benifical to the assessment of the response ofstructurs and the guaranteed rate.The parametric method raised in the paper can take account into the characteristicof the different ground motion intensity, and the make up the shortage of selectedrecords varying with the structures located in the same site. Seen form the results ofdynamic time history of the three models, the force and deformation are close when thestructures are under the different recods selected by three different methods; but thedifferences of hysteretric energy of the structures under three different recods selectedby three different methods are significant, the hysteretric energy of the structures underthe records selected by the parameterized method is obviously larger. Because, the othertwo methods do not consider the duration of the records, the hysteretric energy of thestructures may be underestimated, evidently which is a disadvantage to the assessmentof seicmic behavior of structurs and the design of isolated structure and dissipationstructures.Lastly, the weakness of the present study is briefly discussed and somerecommendations for future research are provided.