Studies On Near-fault Ground Motions Based On Kinematic And Dynamic Source Models

Recent ten years, near-fault strong ground motions have become very active research areas both in seismology and in earthquake engineering. In the end of the last century, many destructive earthquakes have caused great damages and shown some common characteristics in near-fault zone, which impel researchers to study the basic characteristics and the prediction method of near fault strong ground motions. Based these studies, people can predict the spatial distribution and the time history of near fault strong ground motion for the sake of city plan, the seismic structural design and the earthquake damage prediction. Since near fault zone are very close to the causative fault, ground motions in these areas are significant influenced by the rupture and slip process of the fault and the position of the observer site. The detail rupture process have great influence on the near-fault ground motions, so the influence of the rupture process to the near-fault ground motion should be studied. In this paper, we use numerical simulation method and kinematic source model to study the near-fault rupture velocity pulse and the prediction method of near-fault strong ground motions, and then we study the fault rupture process by dynamic source model. The following are the conclusions we made. 1. The relations between the forward-directivity velocity pulse and the source parameters are analyzed qualitatively in the full elastic space. The studies show that in the near fault zone, the pulse are only related to the dislocation size, the dislocation direction , the rise time of part areas between the observer site and the initial rupture point of the fault. The rupture velocity is a very key parameter to the velocity pulse, and it is impropriety for the currently statistical analysis which neglects its influence. The pulse period increases as the rise time increases. 2. The peak velocity and the period of forward-directivity velocity pulse and its spatial distribution are analyzed in the half space. The studies show: (1) for the shallow earthquake, the rupture velocity is the most important parameter to the velocity pulse, especially for the surface rupturing fault. Generally, the pulse period decreases,the pulse peak velocity and its spatial distribution increases as the rupture velocity increases. (2) The fault depth is also an important parameter to the velocity pulse. For the same fault, the pulse peak velocity and the pulse spatial distribution decrease rapidly as the fault depth increases. The fault depth also has some influence on the pulse period, especially for the surface rupturing fault, and in a certain region, the pulse period increases as the fault depth in creases.(3) The earthquake magnitude is another important parameter to the rupture velocity pulse. As the earthquake magnitude increases, the dislocation and the rise time increase, as a result, the pulse peak velocity , the pulse period and its spatial distribution increases.(4) The dislocation direction has great influence on the pulse peak velocity but little influence on the pulse period . For strike slip fault, the pulse peak velocity will be higher while the dislocation direction parallel to the fault strike. (5) The position of the initial rupture point on the fault has influence on both the pulse peak velocity and the pulse period. For uniform strike slip fault, rupture initiate form one end of a fault always generate higher pulse peak velocity and longer periods than others. For uniform dip slip fault, while the initiate rupture point close to the fault bottom, the pulse peak velocity will be higher and the pulse period will be longer. (6) The asperities have import influence on the pulse peak velocity, especially for asperities just near the fault upper edge. It will greatly increase the pulse peak velocity for sites just near the projection of the asperity on the ground. (7) For dip slip fault, the pulse spatial distribution are not symmetric about the fault surface trace. In the same fault distance, the pulse peak velocity on the hanging wall are much higher than that on the foot wall, and the pulse peak velocity decrease more slowly on the hanging wall than that on the foot wall. For dip slip fault, the relations between the pulse peak velocity, the pulse period and the fault depth, the rupture velocity, the asperity distribution are similar to strike slip fault. By the method and program given in this paper, one can calculate the pulse spatial distribution according to an appropriate source model. An approximate fitting formula of an Mw=6.5 uniform strike slip earthquake which including the influence of the rupture velocity and fault depth has been given here, and the fitting formula of other magnitudes can also be obtained by the same method, or be modified appropriately according to the formula of the Mw=6.5 earthquake. 3. The full work frame of near-fault strong ground motion prediction has been give in this paper, which including the determination of global and local source parameter, the calculation of Green's function and the broad band strong motion prediction. For half space, an high efficiency method to predict near-fault strong ground motions has been give according to the symmetry of dislocation source and medium, and for complex site, an prediction method based on the time-space-decoupled, explicit finite element method and parallel computation technique has also been given here, and been used to predict the ground motion of Kunming basin when an assumed fault generate an Mw=6.2 strike slipearthquake. 4. The question of synthesizing large earthquake by small earthquake are discussed in this paper, and the following difficulties are found to be exist. (1) In near-fault region, the small earthquake must be small enough so it does not include rupture directivity effect, otherwise, it can not be used to synthesize large earthquake. (2) It is difficult to select the geometry attenuation factor to adjust a small earthquake generated by some part of the fault to other small earthquake generated by some other part of the fault, because the near field, the intermediate term and far field of the dislocation source have different attenuation factor. (3)For real earthquake, the distribution of dislocation and dislocation direction are variable on the fault plane, so small earthquakes generated by different parts of the fault may be very different. If we cannot obtain enough small earthquakes, the synthesized result may be inappropriate. (4) In near-fault zone, the large earthquakes always generate permanent displacement, and the ground surface also have nonlinear deformation, but the small earthquakes always generate very small elastic displacement or non displacement, so the large earthquake record synthesized by small earthquake can not include nonlinear deformation, since this method is based on elastic representative theory. 5. Based on the finite-element method of near-field wave motion, we present an explicit parallel finite-element method to simulate the dynamic rupture of 3D earthquake fault according to the frictional model. The formulation of different type of the node in the fault has been obtained. The rupture and slip process of the fault, the strong ground motion and the surface rupture can be simulated by this method. 6. The fault rupture process is simulated by the method give above, and the influence of the parameters of slip-weakening model to the rupture velocity, the dislocation distribution, the source function and the near-fault ground motion are discussed in this paper. The influence of the asperity to the rupture velocity, the difference of the rupture process of the surface rupturing fault and buried fault are also analyzed here. The studies show: (1) the slip weakening distance has great influence on the rupture velocity developing process. for the same stress drop and fracture strength, the time that the rupture attain stable velocity increases as the slip weakening distance increases, and for the same stress drop and slip weakening distance, the time that the rupture attain stable velocity also increases as the fracture strength increases. (2) The slip weakening distance haslittle influence on the constant rupture velocity, except the initial stress very close to the fracture strength. The constant rupture velocity is only related to the stress drop and the difference of the critical stress to the initial stress.(3)While the rupture tip pass the asperity, the rupture velocity may increase, which depends on the difference of the critical stress to the initial stress.(4) The slip weakening distance have little influence on the final dislocation and the low frequency(<1Hz) source function, while it has some influence on the near-fault ground motion, since it has great influence on the rupture velocity developing process. (5) For uniform rupture process, when the rupture tip reach ground surface, the rupture velocity along the ground surface will increase. For buried fault, the largest dislocation is located in the center of the fault, while for the surface rupturing fault, it located on the fault upper line, and the largest dislocation is bigger than that generated by the buried fault.