| Several strong earthquakes have caused severe human casualties and property loss in the past three decades, so more and more scientists concern about near-fault ground motion. Especially, the strong velocity pulse, which is one of the characteristics of near-fault ground motion, is concerned in the field of seismology and earthquake engineering because of its high amplitude, long period and its high energy-density, and huge impact on structures. In the study of near-fault ground motion velocity pulse, firstly we need to identify ground motion containing strong velocity pulses; secondly we want to acquire their engineering property; then finally study their effect on isolated structures. To solute the three issues, the main contents and research results include the following five aspects:1. Through concluding the quantitative and qualitative approaches to identifying velocity pulse from near-fault ground motion and introducting Baker(2007) and Shahi&Baker(2014)’s quantitative methods, we propose improved EMD algorithm to classify ground motion as pulse-like or non-pulse-like and compare with two former algorithms from both theories and examples. The results indicate that improved EMD algorithm has a wider range of application than the other two methods, because of its self-adaptive characteristic. So we can utilize it as a referentialand complementary method to identify velocity pulse.2. Through studying Chi-Chi earthquake from its ground motion properties of amplitude, spectral and duration, we find that PGV(peak ground velocity), Tpsv(period corresponding to peak of velocity response spectrum) and Tp(energy duration) are the best parameters to reflect the difference between the pulse-like and no-pulse-like strong ground motions. The average ratio of PGV(PGV of the pulse-like record divided by the no-pulse-like record) is 2.21; the ratio of Tpsv majorly ranges from 1 to 2.66 and its average value is 1.68; Tp of pulse record is shorter than non-pulse’s and is about 0.69 times of it. This means that the pulse ground motions accumulate larger energy than non-pulse ones in a very short time.3. With dual-parameter calibration method and the least square fitting method segmentation method, we calibrate response spectral of 8 sets of data from Chi-Chi earthquake. By studying its calibration spectral characteristic parameters(Tg, βmax, γ), we find that the calibration spectral amplitude of pulse record is greater than nonpulse’s after the period of 1.16 sec, and the ratio of average values is 1.98. Therefore, the design spectrum of seismic design code should consider the impact of seismic pulses.4. We summarize the input principle of ground motion with velocity pulse from previous studies:(1)homology: ground motion records are coming from one earthquake, in the same epicenter and same site conditions;(2)similar response spectra;(3)the number of each type of time-histories can not be less than 8 groups;(4)try to keep approximate epicentral distance.5. The paper utilizes wavelet analysis to extract pulse and then gets the residual ground motion as the non-pulse-like ground motion and the Original ground motion as the pulse-like ground motion according to Baker(2007). We get 30 records selected from Chi-Chi earthquake in FN direction through the above method. Then we input the pulse and non-pulse time-histories in an isolated structure which is a six-story frame structure and observe its earthquake response, through nonlinear time history analysis with ETABS. According to the results, in order to ensure the reliability of isolated structure design, two proposals are given:(1)we should consider the influence of velocity pulse on isolated structures in the near-fault region which is in an epicentral distance of less than 80km;(2)if the pulse-like strong motion records whose condition is similar to construction sites are not enough to input, we can use an amplification factor 1.34-1.60 to estimate the influence of pulse. |
PDF Full Text Request