Determination Of Fault Activity Parameters And Ground-motion Simulation Of A Scenario Earthquake On Near-fault Dam

In 2015,the?code for seismic design of hydraulic structures of hydropower project?NB35047-2015??pointed out that when the distance from the seismogenic fault to the dam is less than 10 km and the magnitude is greater than 7.0,the seismic response spectra of the near-fault dam site should be simulated.However,due to the lack of near-field strong motion data,the ground motion prediction equations usually cannot reflect the strong motion characteristics of the near field.Moreover,evaluating the seismic safety of major projects places higher demands on the fault activity parameters?fault geometry,seismogenic depth,slip rate,etc.?and ground-motion model parameters?stress drop,geometrical spreading function?G?R??,quality factor?Q?f??,high-frequency attenuation parameters??₀?and site amplification,etc.?.Therefore,for the seismic risk analysis of a specific near-fault major engineering,accurate determination of fault activity parameters and calculation of ground motion response spectra using other methods that can reflect the ground motion characteristics of near-fault sites have important engineering value.In this paper,we take the Baihetan dam as an example.First,the relocated seismic data,gravity,and GPS observation data are used to determine the fault geometric structure,long-term slip rate and fault depth,as well as the cumulative seismic moment and maximum potential earthquake magnitude of the Xianshuihe-Zemuhe faults.Furthermore,we study the inversion method of ground-motion model parameters,and use strong-motion recorded to investigate the source parameters,quality factor?Q?,and site effects of S-wave Fourier acceleration amplitude spectra?FAS?of the 2019 M_S6.0Changning earthquake sequence in China.Based on the parameters of the fault model and ground-motion model in the study area,a stochastic finite-fault method is used to simulate the acceleration response spectrum of the Baihetan dam.The following results are obtained:1.Based on Particle Swarm Optimization,GPS and gravity observation data are used to joint invert the slip rates of the Anninghe–Zemuhe faults and Daliangshan fault.The result indicates that gravity data has little effect on the joint inversion of the slip rate of strike-slip fault zone.2.We use GPS velocities from 2011 to 2017 to infer the slip rates and the locking depths along the faults.The results show that the slip rates of the north,middle,and south segments of the Daliangshan sub-block are 9.8,8.9,and 8.4 mm/a respectively,decreasing from north to south,and have a clockwise rotation.The locking depths of the north and south of the Anninghe fault?AF?,Zemuhe fault?ZF?,north,middle,and south of the Daliangshan fault?DF?are 6.2±3.6 km,16.8±2.0 km,14.1±3.3 km,23.3±5.1 km,20.8±1.9 km and 15.5±2.6 km,respectively.The locking depths of the south segment of AF,ZF,middle,and south of the DF are close to the 90%cutoff depths of small events with a standard deviation of 0.94 km.3.1338 relocated earthquakes?M_S?1.6?from 1981 to 2018 are used to obtain the seismicity depth distribution of the Daofu–Kangding segment of the Xianshuihe fault system.The 90%cumulative depth distributions for the Daofu fault?DF?,Qianning fault?QF?,and Kangding fault?KF?are 17.6,13.0,and 14.3 km,respectively.By analyzing the geometry of the KF,we infer its three branches,the Yalahe fault?YF?,Selaha fault?SF?,and Zheduotang fault?ZF?converged into one fault at a depth of approximately 15 km and that a flower-shaped structure is observed above this depth.4.We use GPS velocities from 2004 to 2017 to obtain the left-lateral slip rate and locking depth of the Daofu–Qianning fault?DQF?are 4.5±1.1 mm/a and 5.3±1.9km,respectively.Because the fault creeps from a depth of 2.6 km to the surface at a rate of 1.3±1.0 mm/a,the DQF cannot generate earthquakes having magnitudes of greater than 7 in the near future.However,the slip rates are 7.5±1.6,2.3±1.5,and1.9±1.5 mm/a by assuming a locking depth of 15 km for the SF,YF,and ZF,respectively.The total slip rate of the KF is 11.7±1.5 mm/a,which is larger than 4.5±1.1 mm/a of the DQF.Furthermore,because the creep depth of the SF is 3.4±1.6km and the creep rate is 7.5±2.8 mm/a,the accumulated seismic moment since 1725corresponded to a M_W7.1 earthquake.5.The high-frequency attenuation model of the 2019 Changning earthquake area is?=0.0420+0.0001262R.The?₀ values of the basin and mountain areas are 0.0463 s and 0.0319 s,and the m values of which are 0.0000958 s/km and 0.0001873 s/km.The inverted stress drop of the mainshock is 1.15 MPa,and those of the aftershocks vary from 0.11-1.04 MPa with an average value of 0.43 MPa.The inverted Q values increase rapidly with frequencies at 0.5-4.0 Hz from 62 to 2920 and become less dependent at 4.0-25.0 Hz.Such phenomenon indicates that the propagation path attenuation mechanism transited to intrinsic at high frequencies.A bilinear Q?f?model that Q?f?=237.6f1.27?Q<1280?and Q=1280 at higher frequencies is obtained.Moreover,a nonlinear soil site effect is observed at 51GXT in earthquakes with peak ground acceleration greater than 300 cm/s².The nonlinearity obviously aggravates the site amplification at 1.0-5.0 Hz.6.Based on the stochastic finite-fault method,the response spectrum of the BF,JF and southern section of the Daliangshan fault?SDF?to the Baihetan dam site are simulated under two rupture modes from south to north and north to south.The result indicates that the rupture directivity is weakly correlated with the starting point of rupture,but has a strong correlation with the inhomogeneity and asperity of the fault rupture plane.Moreover,the simulated logarithmic mean value of the BF,JF,and SDF is compared with the mean and one standard deviation of the NGA-West2 ground motion prediction equation.The simulated results of the JF and SDF are consistent with the mean prediction results of the NGA-West2 models,while the BF is significantly different from the NGA-West2 when T<1.0 s.