Random finite fault source model establishment method

Abstract

The invention discloses a stochastic finite fault source model establishment method, which comprises the following steps of S1, determining a geometrical shape of a seismic fault based on regional geological data and seismic monitoring data, and estimating a non-planar fault parameter and a fault plane average sliding amount in combination with historical data and an empirical formula; s2, estimating distribution and sliding strength of concave-convex bodies and obstacles on the fault surface according to historical seismic inversion data, and obtaining the sliding amount of each sub-fault by randomly disturbing the average sliding amount; s3, the discrete fault plane serves as a sub-fault unit, and after an initial fracture point and the average fracture speed are determined, random fracture initial time conforming to normal or power law distribution is generated based on a random number method; s4, calculating the seismic oscillation time history of each sub-fault by adopting a random point source method, superposing contributions, introducing the influence of a shallow velocity structure V30, and generating a three-dimensional seismic oscillation field; and S5, comparing actual strong earthquake record adjustment parameters, and outputting a multi-risk-level earthquake motion parameter evaluation result after iterative optimization. According to the method, by fusing randomness and empirical data, the simulation precision of the seismic source model on the uncertainty of the seismic process is improved, the influence of the potential maximum-magnitude earthquake possibly occurring in the active fault zone in the future can be effectively estimated, the seismic oscillation simulation precision and the calculation efficiency can be improved, and the method is suitable for engineering structure aseismic design and seismic disaster evaluation.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of earthquake engineering and seismology, and particularly to a method for establishing a seismic source model based on the random finite fault theory. BACKGROUND

[0002] Earthquakes can cause strong ground motion, resulting in serious disasters. As the main carrier of earthquakes, the precise assessment of potential magnitude and earthquake probability of active fault zones is the core prerequisite for earthquake prevention and disaster reduction. The crustal movement in these areas continuously accumulates stress, and once it reaches the critical state, a strong earthquake may occur. In special areas with special geographical locations, few monitoring equipment and frequent seismic activity, the accurate assessment of potential seismic risk may be limited. However, this limitation does not mean that earthquake simulation and prediction cannot be carried out. On the contrary, it emphasizes the need to rely on numerical simulation tools to fill the knowledge gap caused by data gaps.

[0003] However, the traditional seismic source model has significant limitations in predicting potential earthquakes in active fault zones: the traditional seismic source model relies on statistical methods based on historical earthquake records, which cannot reflect the current stress state of the fault zone and cannot effectively predict the magnitude and spatial impact range of possible earthquakes in the next few decades; the traditional seismic source model ignores the three-dimensional heterogeneity of fault geometry and the randomness of the rupture process, resulting in a simulation error of peak ground acceleration (PGA) of more than 50% in the near field, directly affecting the safety of seismic design; the traditional seismic source model fails to couple the correlation between shallow geological structure and deep fault activity, and cannot accurately assess the damage difference of surface buildings under different magnitudes.

[0004] Therefore, it is of urgent practical significance to construct a seismic source model with randomness and physical consistency for active fault zones to achieve precise simulation of the magnitude, rupture range, and seismic field distribution of possible future earthquakes. Therefore, we propose a method for establishing a random finite fault seismic source model to solve the problems raised in the background. SUMMARY

[0005] The present application relates to the field of earthquake engineering and seismology, and particularly to a method for establishing a seismic source model based on the random finite fault theory.

[0006] To achieve the above object, the present application provides the following technical scheme: a random finite fault source model establishment method, comprising the following steps: S1, fault geometric parameter determination: based on regional geological data and seismic monitoring data, the strike, dip angle, length, width and spatial distribution form of the seismogenic fault are determined, and the non-planar fault geometric parameters and the average slip amount of the fault plane are estimated by combining the historical observed data with the empirical formula method; S2, slip distribution calculation: the distribution and slip strength of the convex and concave bodies on the fault plane are estimated according to the historical earthquake inversion data, and the random disturbance is applied to the average slip amount of the fault plane to obtain the slip amount of each sub-fault, considering the randomness of the earthquake process; S3, random rupture time generation: the entire fault plane is discretized into a plurality of sub-fault units, after the starting rupture point and the average rupture velocity are determined, the approximate rupture time of each sub-fault can be obtained according to the distance between each sub-fault and the starting rupture point, but due to the uncertainty of the earthquake propagation process, the random rupture starting time of each sub-fault can be generated based on the random number method, and the rupture time distribution satisfies the normal distribution or the power law distribution, so as to satisfy the uncertainty and randomness of the earthquake process; S4, ground motion synthesis: the random point source method is applied to calculate the ground motion time history of each sub-fault, the contributions of all sub-faults are superimposed, and the influence of the shallow velocity structure V30 is introduced, so as to generate a three-dimensional ground motion field containing acceleration time history, peak ground acceleration (PGA) and response spectrum; S5, model verification and optimization: by comparing the acceleration, velocity and response spectrum of the actual strong earthquake record, the model parameters are adjusted and iteratively optimized, and finally the evaluation results of the ground motion parameters meeting the multi-risk level are output.

[0007] Preferably, in S1, the characteristics of the kinematic source can be described by global source parameters and local source parameters. Global source parameters and local source parameters refer to source parameters at different scales, which facilitate understanding of the earthquake process, simulating earthquake wave propagation and studying the structure of the earth's interior. Global source parameters refer to parameters describing the entire earthquake source process, usually including the properties of the entire fault or seismic zone. These parameters correspond to source characteristics in a larger spatial range, such as seismic moment magnitude M W , fault length L, width W, source mechanism (including strike-slip, normal fault, reverse fault, etc.), source depth h, average slip amount D, average rise time and average rupture velocity v, etc. Multiple source geodetic data (including GNSS static displacement and Sentinel-1 measurement results) are used in combination with scientific research data to estimate fault geometric parameters, slip distribution and observation data weights simultaneously.

[0008] Preferably, in S1, based on statistical analysis of hundreds of earthquake records worldwide, the global regional range moment magnitude M W - rupture scale empirical relationship is described by formula (1) and formula (2):

[0009]

[0010] where L and W represent the fault rupture length and width (unit: km), respectively; M W is the moment magnitude.

[0011] Preferably, in the S1, the empirical relationship between the moment magnitude M W and the seismic moment M0 is described by equation (3):

[0012]

[0013] where M0 is the seismic moment (unit: dyne-cm).

[0014] Preferably, in the S1, the relationship between the average dislocation D on the fault plane and the seismic moment M0 for shallow source crustal earthquakes is described by equation (4):

[0015]

[0016] Preferably, in the S1, the rise time reflects the speed of energy release in the initial stage of the seismic waveform. A short rise time usually corresponds to a fast energy release process, which can be related to a rapid process such as instantaneous fault slip. A long rise time can indicate a more complex source process and a relatively slow energy release. The average slip time is the average value of the slip time on the fault plane. Based on statistical analysis of multiple seismic inversion data, the relationship between the average rise time τ and the seismic moment M0 is described by equation (5):

[0017]

[0018] Preferably, in the S1, the rupture velocity changes in real time during the fault rupture process, and the change in rupture velocity is one of the main reasons for exciting high-frequency seismic waves. A large number of inversion results show that the average rupture velocity of the fault is generally 0.6-0.9 times the shear wave velocity.

[0019] Preferably, in the S2, the local source parameters are parameters representing the inhomogeneity or roughness of the fault rupture surface, mainly reflecting the changes in the slip amount, rupture velocity, and rise time of each sub-fault on the fault rupture surface.

[0020] Preferably, in S2, based on the fault plane slip distribution obtained by a large number of ground motion inversions, it is known that the slip distribution of the fault rupture surface during the earthquake rupture process is extremely uneven, there are Asperity (asperity) regions with slip exceeding 2 times the average slip on the fault surface, and there are Barrier (barrier) regions that have not yet ruptured. Moreover, each earthquake contains about 1-3 asperities; the total area of the asperities is about 22% of the total rupture area, and the area of the largest asperity is about 16% of the total rupture area; the average slip of the asperity part is about 2.01 times the average slip of the entire fault. Nevertheless, only the slip distribution of the entire fault surface is divided into asperity regions with larger slip and background regions with smaller slip. In order to more clearly distinguish the differences in slip of each sub-source, a column of random numbers with zero mean and unevenly distributed between -1 and 1 is introduced, denoted as α(k), and the slip D(k) of each sub-fault can be described by the average slip D using formula (6):

[0021] D(k) = D + α(k) x 0.3D (6)

[0022] In the formula, k is the sub-fault serial number.

[0023] Preferably, in S3, after the distance between each sub-fault and the starting rupture point and the average rupture velocity of the fault surface are determined, the rupture time of each sub-fault is also determined. At this time, the difference in rupture velocity of each point on the fault surface can be converted into a perturbation on the time when each sub-fault starts to rupture. That is, the starting rupture time t k of each sub-fault can be calculated from the rupture time of each point by formula (6). Similarly, the rise time of each sub-fault can be calculated from the average rise time. At this point, the kinematic source model parameters are completely determined, and the ground motion simulation analysis can be carried out next.

[0024] Preferably, in S4, the spectral element method program (such as the SPECFEM3D program) is used for simulation when synthesizing ground motion. By using spectral methods (such as Legendre polynomial interpolation) within the element, high-frequency seismic waves can be accurately simulated with fewer grid numbers, especially the capture accuracy of near-field high-frequency ground motion (such as 1-10 Hz) is significantly better than that of traditional finite element methods; it can efficiently handle non-uniform media, complex topography and three-dimensional fault geometry, accurately simulate complex propagation phenomena such as scattering, reflection and diffraction; the element division method is naturally suitable for large-scale parallel computing, and supercomputers can be used to realize three-dimensional ground motion field simulation with millions of grids, greatly improving the calculation efficiency of complex regions.

[0025] Preferably, in S5, the model verification uses strong motion records of actual earthquake cases to compare the consistency of PGA, PGV and response spectrum distribution between the simulation results and the observed data.

[0026] Compared with the prior art, the random finite fault source model establishment method has the beneficial effects that: the random finite fault source model establishment method adopts a new type of structure design, and the specific contents are as follows:

[0027] (1) By introducing random numbers to disturb the local source parameters, the complexity of the fault geometry and the randomness and uncertainty of the earthquake propagation process are comprehensively considered, and the simulation accuracy is improved.

[0028] (2) When synthesizing ground motion, the spectral element method program is used for simulation, which can realize accurate simulation of high-frequency seismic waves under a small number of grids, can efficiently process non-uniform media, complex terrain and three-dimensional fault geometry, accurately simulate complex propagation phenomena such as wave scattering, reflection and diffraction, and greatly improve the calculation efficiency of complex regions through large-scale parallel computing.

[0029] (3) Not only supports acceleration, velocity and displacement time history of ground motion, but also supports synchronous simulation of detailed information such as multi-damping ratio response spectrum and peak parameter, which is very useful, and plays a very important role in building seismic design, earthquake disaster assessment and earthquake research. It can help engineers and researchers better understand the characteristics and influence of ground motion, so as to take more effective seismic measures and research methods. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 It is a working step schematic diagram of the random finite fault source model establishment method of the application;

[0031] Figure 2 It is a distribution diagram of the MFT and MT kinematic hybrid source rupture surface slip, rupture time and rise time of the application;

[0032] Figure 3 It is a three-dimensional site model and fault position schematic diagram of the application;

[0033] Figure 4 It is a comparison diagram of simulated ground motion time history and measured record results of the application;

[0034] Figure 5 It is a comparison diagram of simulated ground motion response spectrum and measured record results of the application.

Claims

1. A method for establishing a stochastic finite fault source model, characterized by, The method comprises the following steps: S1, fault geometry parameter determination: based on regional geological data and seismic monitoring data, the strike, dip angle, length, width and spatial distribution pattern of the seismogenic fault are determined, and the non-planar fault geometry parameters and the average slip amount of the fault plane are estimated by combining the historical observed data with the empirical formula method; S2, slip distribution calculation: according to the historical earthquake inversion data, the distribution and slip strength of the concave-convex body and the barrier on the fault plane are estimated, and the random disturbance is applied to the average slip amount of the fault plane to obtain the slip amount of each sub-fault, considering the randomness of the earthquake process; S3, random rupture time generation: the entire fault plane is discretized into a plurality of sub-fault units, after the starting rupture point and the average rupture velocity are determined, the approximate rupture time of each sub-fault can be obtained according to the distance between each sub-fault and the starting rupture point, but due to the uncertainty of the earthquake propagation process, the random rupture starting time of each sub-fault can be generated based on the random number method, and the rupture time distribution satisfies the normal distribution or the power law distribution, so as to satisfy the uncertainty and randomness of the earthquake process; S4, ground motion synthesis: the random point source method is applied to calculate the ground motion time history of each sub-fault, the contributions of all sub-faults are superimposed, and the influence of the shallow velocity structure V30 is introduced, to generate a three-dimensional ground motion field including acceleration time history, peak ground acceleration (PGA) and response spectrum; S5, model verification and optimization: by comparing the acceleration, velocity and response spectrum of the actual strong earthquake record, the model parameters are adjusted and iteratively optimized, and finally the seismic ground motion parameter evaluation results meeting the multi-risk level are output.

2. The method of claim 1, wherein: In the S1, the characteristics of kinematic sources can be described by global source parameters and local source parameters. Global source parameters and local source parameters refer to source parameters on different scales, which are convenient for understanding the earthquake process, simulating seismic wave propagation and studying the structure of the earth's interior. Global source parameters refer to parameters describing the entire process of the earthquake source, usually including the properties of the entire fault or seismic zone. These parameters correspond to the source characteristics on a larger spatial range, such as seismic moment magnitude M W , fault length L , width W , source mechanism (including strike-slip, normal fault, reverse fault, etc.), source depth h , average slip D , average rise time and average rupture velocity v , etc. Using multi-source geodetic data (including GNSS static displacement and Sentinel-1 measurement results) combined with scientific research data, the fault geometry parameters, slip distribution and observation data weights are estimated synchronously.

3. The method of claim 1, wherein: In the S1, based on the statistical analysis of the global hundreds of earthquake records, the global regional range of moment magnitude M W The rupture scale empirical relationship is described by formula (1) and formula (2): （1） （2） where, L and W represent fault rupture length and width (in km), respectively; M W is the moment magnitude.

4. In the S1, based on a large amount of data statistics, the magnitude of the earthquake M W and seismic moment M 0 experience relationship by formula (3): （3） In the formula, M 0 is the seismic moment (unit: dyne-cm).

5. In the S1, the average dislocation on the seismogenic fault plane in the shallow crust D and the seismic moment M 0 relationship is described by equation (4): （4） In the S1, the rise time reflects the speed of energy release in the initial stage of the seismic waveform. Short rise time usually corresponds to a fast energy release process, which may be related to a rapid process such as fault instantaneous slip. Long rise time may indicate that the source process is more complex and the energy release is relatively slow. The average slip time is the average value of the slip time on the fault plane. Based on statistical analysis of multiple seismic inversion data, the relationship between the average rise time τ and the seismic moment M0 is described by formula (5): M τ = 0. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 （5) In S1, the rupture velocity changes in real time with the fault rupture process, and the change of the rupture velocity is one of the main reasons for exciting high-frequency seismic waves. A large number of inversion results show that the average rupture velocity of the fault is generally 0.6-0.9 times the shear wave velocity.

6. The method of claim 1, wherein: The local source parameter in S2 is a parameter representing the unevenness of the fault rupture surface or roughness, mainly reflecting the variation of the slip amount, rupture velocity and rise time of each sub-fault on the fault rupture surface. Based on the fault surface slip distribution obtained by a large number of seismic inversion, it is known that the slip distribution of the fault rupture surface is extremely uneven during the earthquake rupture process, and there are Asperity (asperity) regions with a slip amount more than twice the average slip amount on the fault surface, and Barrier (barrier) regions that have not yet ruptured. Moreover, each earthquake contains about 1-3 asperities; the total area of the asperities is about 22% of the total rupture area, and the area of the largest asperity is about 16% of the total rupture area; the average slip amount of the asperity part is about 2.01 times the average slip amount of the entire fault. Nevertheless, it is only to divide the slip distribution of the entire fault surface into the asperity region with larger slip amount and the background region with smaller slip amount. In order to more clearly distinguish the difference in slip amount of each sub-source, a column of random numbers with uneven distribution between-1 and 1 with zero mean is introduced, denoted as α ( k ), and the slip amount D ( k ) of each sub-fault can be described by the average slip amount D using formula (6): (6) In the formula, k is a sub-fault sequence number.

7. The method of claim 1, wherein: The rupture time of each sub-fault is determined after the distance between each sub-fault and the initial rupture point and the average rupture velocity of the fault plane are determined. At this time, the difference in rupture velocity of each point on the fault plane can be converted into a perturbation on the time of the start of rupture of each sub-fault. That is, the start rupture time of each sub-fault is determined by the formula (5). t k The rupture time of each point can be calculated by the formula (6). Similarly, the rise time of each sub-fault can be calculated by the average rise time. At this time, the kinematic source model parameters are completely determined, and the subsequent simulation analysis of ground motion can be carried out.

8. The method of claim 1, wherein, In S4, the spectral element method program (such as SPECFEM3D program) is used for simulation in ground motion synthesis. By using the spectral method (such as Legendre polynomial interpolation) in the element, high-frequency seismic waves can be accurately simulated with fewer grid numbers, and the capture accuracy of near-field high-frequency ground motion (such as 1-10Hz) is significantly better than that of the traditional finite element method; It can efficiently process non-uniform medium, complex terrain and three-dimensional fault geometry, accurately simulate the complex propagation phenomena such as scattering, reflection and diffraction of waves; the element division method is naturally suitable for large-scale parallel computing, and supercomputers can be used to realize three-dimensional ground motion field simulation of millions of grids, greatly improving the calculation efficiency of complex regions.

9. The method of claim 1, wherein, In S5, the model verification uses the strong earthquake record of the actual earthquake case, and compares the PGA, PGV and response spectrum distribution consistency of the simulation results and the observation data.

Images