Research On Fault Slip Inversion And Real-time Correction Of Site Amplification Fator

China Strong motion network have deployed almost two thousand strong motion stations during the tenth five year plan project.The observation technology developed and the strong motion network is relatively intensive.The national earthquake early waring and intensity fast report project will be constructed,then the distance interval will be about 20 to 40 kms.Then the data will be easily get.But it is difficult to get the fault slip inversion quickly with strong motion data for earthquake emergency response and damage evaluation.The previous study usually used scale value such as increment of PGA,or PGV to characterize site amplification factor,but lack related research about correcting the frequency-dependent site amplification factor in real-time ground motion prediction system.In this paper,mainly study on the quick automated baseline correction for reliable static displacement,slip inversion with the near field static displacement and correction the frequency-dependent site amplification factor real-timely.The main content of the paper are as follows:1.Analyze the reason of the baseline drift,summarize the baseline correction method,improved the automated baseline correction method,then it could get the static displacement from many kinds of instrument home and abroad.It could also able to process the data from different observation network,compared the static displacement with the GPS measurement result,the reliability is validated.It could get the reliable static displacement from the professional strong motion station,but also find that the result from p-alert MEMS network deviated significantly.Almost all the data show different movement direction,it reveals that static displacement may not be recovered from the MEMS instrument.2.Summarize the coseismic dislocation theory in the half space and layered media and related inversion theory.Built single fault model and two fault model for Meinong Earthquake utilized the CWB and NCREE strong motion data,GPS data,Sentinel-1A and ALOS-2 InSAR data,then invert the slip distribution separately and jointly,compare the result,it shows that single fault fault model is more reliable.The slip mainly include dip component,the main rupture located at the depth between 9 km and 15 km.The rupture mainly located at the northwest area,the max slip is about 0.95 meter.The moment magnitude is 6.29.The average stress drop is 1.21 Mpa,the maximum stress drop is 6.31 MPa.3.Obtain the focal mechanism of Kumamoto earthquake with the F-net data through moment tensor inversion considering the source time function effect.Built the single and two segment fault model and invert the slip distribution separately and jointly with the strong motion data from k-net and kik-net,Geonet and Sentinel-1A InSAR data for Kumamoto earthquake.Analyse the result and find that two segment fault model is more resonable.The result reveals that the main rupture propagate northeast 10 km to 30 km far from epicenter,the peak slip was 6.0m.There are relative large slip near the epicenter and the peak slip is 4.0m,the rupture terminate near the southwest segment of Asu Mountain.The result from k-net and kik-net is almost same to the Geonet.It reveals the possibility of fast fault slip inversion with strong motion data for emergency response and damage evaluation.4.Summarize the state of the site amplification correction,study on the frequency-dependent site amplification factor.Then design the casual infinite impulse response filter(IIR)to model the site amplification factor,implement the method and related software package.With the data from IBRH10 and IBRH19 of the Kik-net trigged during 208 earthquakes,obtain the relative site amplification between the borehole station and the surface station with the traditional spectrum ratio method,designe the IIR filter to simulate the surface data from the borehole data.Similarly,obtain the relative spectrum ratio for IBRH10 and IBRH19,simulate the IBRH10 acceleration data and Fourier spectrum from the observed data of IBRH19.These result show that almost all the simulation are quietly well.Statistic the instrumental seismic intensity of the observed data and the simulation data,find that this method has significantly improved the accuracy than previously method.It could produce different site amplification factor for different earthquakes which could not be produced by the method that used scale values.It improves the acceleration and instrumental seismic intensity prediction significantly.It provides an accurate method for real time site amplification correction in the strong motion prediction system.5.Utilize the stochastic finite-fault ground motion simulation method,simulate two rock acceleration time history based on the combied faultslip inversion model of Kumamoto Earthquake,then apply the IIR method to the stochastic finite-fault ground motion simulation method,obtain good simulation result on the surface and verify the method proposed previously.It provides siginificant support for more accuate ground motion predtion.