CN-121721710-B - Joint inversion method based on surface wave and direct P wave

CN121721710BCN 121721710 BCN121721710 BCN 121721710BCN-121721710-B

Abstract

The application discloses a joint inversion method based on surface waves and direct P waves, which relates to the field of seismic tomography and comprises data acquisition and processing, individual surface waves and P wave model construction, joint inversion algorithm design and visualization of inversion results. The surface wave part adopts a frequency-vector wave number domain conversion method to extract a dispersion curve, the transverse wave speed of the underground medium is optimized and updated through the correction quantity of the Voronoi inversion solving model, and the direct P wave part uses a rapid travelling method in PYGIMLI programs to calculate a ray path and regularized inversion solving longitudinal wave speed. Finally, by combining the two wave type results through a joint inversion algorithm, accurately estimating parameters such as the speed and geological stratification of the underground medium in the subsidence area of the road, comprehensively analyzing and processing seismic wave data by combining the advantages of surface wave inversion and direct P wave inversion, overcoming the limitation of a single inversion technology, and improving the noise immunity, the inversion precision and the resolution of the underground shallow structure of the data.

Inventors

LI CHENG
Ye Gongcheng
ZHU JUNFENG
Qian Changbao
CAO YU

Assignees

安徽理工大学

Dates

Publication Date: 20260508
Application Date: 20260224

Claims (7)

1. The joint inversion method based on the surface wave and the direct P wave is characterized by comprising the following steps of: s1, acquiring seismic data of a research area, wherein the seismic data comprise surface wave background noise data and direct P wave seismic record data; s2, preprocessing surface wave background noise data, calculating a cross-correlation function between detectors, and extracting a surface wave dispersion curve through a frequency-Bezier conversion method; S3, respectively carrying out surface wave inversion and direct P wave inversion based on the surface wave dispersion curve and the direct P wave travel time data to obtain a transverse wave velocity model and a longitudinal wave velocity model; The surface wave inversion is based on a surface wave dispersion curve, a sensitivity matrix is constructed, a transverse wave speed model correction amount is solved in a Voronoi point by adopting a least square method iteration mode, a transverse wave speed model is updated continuously in an iteration mode, and a final transverse wave speed model and a dispersion residual error are obtained; Calculating a ray path by adopting a rapid travelling method, iteratively calculating residual errors between theoretical travel time and actual travel time by adopting a regularized least square method, and iteratively updating to obtain a final longitudinal wave speed model and travel time residual errors; S4, taking the transverse wave velocity model grid as a public grid, interpolating and mapping the longitudinal wave velocity model to the public grid, and constructing a target model parameter space; S5, constructing a joint inversion objective function based on the surface wave dispersion data and the direct P wave travel time data, solving the objective model parameters by minimizing the objective function, and performing joint inversion of the surface wave and the direct P wave; the objective function is expressed as: ; Wherein, the Taking L2 norm as a weighting coefficient; is a target model; A priori reference model for the target model; is a damping factor; the constraint relation between the transverse wave velocity model and the longitudinal wave velocity model is expressed as: ; In order to be a transverse wave velocity model, Is a longitudinal wave velocity model; And S6, judging whether the objective function is converged or not, if not, updating parameters of the objective function and carrying out joint inversion again until convergence and inversion are finished.
2. The face wave and direct P wave based joint inversion method according to claim 1, wherein the step S2 of preprocessing face wave background noise data specifically comprises the steps of segmenting original noise data into data segments of one minute, sequentially carrying out resampling, band-pass filtering, normalization and spectral whitening, and calculating a cross-correlation function between detectors based on the preprocessed data.
3. A joint inversion method based on surface wave and direct P-wave according to claim 2, wherein the formula of the surface wave inversion in S3 is expressed as: ; Wherein, the In order to provide a sensitivity matrix, For an initial shear wave velocity model of the face wave inversion, The theoretical dispersion data obtained based on forward modeling of the initial transverse wave velocity model.
4. A joint inversion method based on surface waves and direct P-waves as in claim 3, wherein the dispersion residual is expressed as: ; Wherein, the Is a transverse wave velocity model; the observation frequency of the surface wave is the dispersion; Is the first Theoretical dispersion of the second forward iteration.
5. A joint inversion method based on surface wave and direct P-wave according to claim 4, wherein the theoretical travel time of direct P-wave is expressed as: ; Wherein, the For the slowness of the initial longitudinal wave velocity model, Is a ray path; is the initial value of the ray path.
6. A joint inversion method based on surface waves and direct P-waves as in claim 5, wherein the travel time residual is expressed as: ; Wherein, the The travel time residual error of the direct P wave; Is direct to P wave Theoretical travel time of the secondary iteration; The observation time of the direct P wave is taken; , is a sensitivity matrix of direct P wave.
7. The method of joint inversion based on surface waves and direct P-waves according to claim 6, wherein the convergence criterion in S6 is one of the following conditions, i.e. the iteration is terminated: (1) The absolute value of the difference between the objective function values of two successive iterations is less than a preset threshold; (2) The iteration number reaches the preset maximum iteration number.

Description

Joint inversion method based on surface wave and direct P wave Technical Field The application relates to the technical field of seismic tomography, in particular to a joint inversion method based on surface waves and direct P waves. Background Seismic tomography is a technique for reconstructing subsurface structures by analyzing the propagation characteristics of seismic waves. Transverse wave velocity of surface wave signal to medium) High resolution of shallow geologic structures can be constructed with high sensitivity to changes using long time series background noise cross correlation functions or artificial source excited signalsAnd (5) a model. The frequency-Bessel transformation method is an advanced technology developed in recent years, can efficiently extract multi-order surface wave information from seismic background noise, and remarkably improves the frequency dispersion imaging precision and the array optimization effect. Longitudinal wave velocity of direct P wave to underground medium) The structure and formation interface are sensitive and the velocity structure can be inferred by analyzing the source to receiver travel time data. However, single face-wave inversion technology has limitations in detail description due to limited longitudinal resolution and insufficient sensitivity to stratum interfaces, and direct P-wave inversion is highly sensitive to interface positions and forms, but has weak capability of revealing shallow structures and transverse wave velocity structures. More critical, both methods alone cannot be obtained simultaneouslyAnd (3) withThese two key parameters. Disclosure of Invention In order to solve the problems, the application provides a joint inversion method based on surface waves and direct P waves, which estimates parameters such as the speed of an underground medium in a research area, geological stratification and the like by simultaneously considering the propagation characteristics of the surface waves and the direct P waves and combining the dispersion characteristics of background noise of the surface waves and the travel time data of the P waves. In order to achieve the above purpose, the application provides a joint inversion method based on a surface wave and a direct P wave, which comprises the following steps: s1, acquiring seismic data of a research area, wherein the seismic data comprise surface wave background noise data and direct P wave seismic record data; s2, preprocessing surface wave background noise data, calculating a cross-correlation function between detectors, and extracting a surface wave dispersion curve through a frequency-Bezier conversion method; s3, respectively carrying out surface wave inversion and direct P wave inversion based on the surface wave dispersion curve and direct P wave travel time data to obtain a transverse wave velocity model and a longitudinal wave velocity model; S4, taking the transverse wave velocity model grid as a public grid, interpolating and mapping the longitudinal wave velocity model to the public grid, and constructing a target model parameter space; S5, constructing a joint inversion objective function based on the surface wave dispersion data and the direct P wave travel time data, solving the objective model parameters by minimizing the objective function, and performing joint inversion of the surface wave and the direct P wave; And S6, judging whether the objective function is converged or not, if not, updating parameters of the objective function and carrying out joint inversion again until convergence and inversion are finished. Preferably, the preprocessing of the surface wave background noise data in the step S2 specifically comprises the steps of segmenting the original noise data into data segments of one minute, sequentially carrying out resampling, band-pass filtering, normalization and spectral whitening, and calculating a cross-correlation function between detectors based on the preprocessed data. Preferably, in the step S3, the surface wave inversion is based on a surface wave dispersion curve, a sensitivity matrix is constructed, a transverse wave speed model correction amount is solved by adopting a least square method iteration at Voronoi points, and a transverse wave speed model is continuously updated to obtain a final transverse wave speed model and a dispersion residual error. Preferably, the formula of the face wave inversion in S3 is expressed as: ; Wherein, the In order to provide a sensitivity matrix,For an initial shear wave velocity model of the face wave inversion,The theoretical dispersion data obtained based on forward modeling of the initial transverse wave velocity model. Preferably, the dispersion residual is expressed as: ; Wherein, the Is a transverse wave velocity model; the observation frequency of the surface wave is the dispersion; Is the first Theoretical dispersion of the second forward iteration. Preferably, the direct P-wave inversion in the S3 specific