A method and device for inverting microseismic dispersion curves, an electronic device and a storage medium
By employing the high-resolution frequency-wavenumber method and the iterative screening mechanism of the inversion model, the problem of local optimal solutions in the inversion of micro-motion dispersion curves is solved, resulting in more efficient and accurate inversion results, which are applicable to micro-motion exploration under complex geological conditions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- WUHAN UNIV OF TECH
- Filing Date
- 2025-12-09
- Publication Date
- 2026-07-21
AI Technical Summary
In existing technologies, micro-dispersion curve inversion algorithms are prone to getting trapped in local optima and cannot effectively explore global optima, resulting in low computational efficiency and inversion results that deviate from the actual formation medium.
The initial Rayleigh wave dispersion curve is extracted using the high-resolution frequency-wavenumber method, and then corrected by recursive and forward modeling calculations. Through the iterative screening mechanism of the inversion model, and by utilizing physical conditions such as Poisson's ratio correlation constraints, global search and data quality optimization are achieved.
It improves the quality of dispersion curves and the accuracy of inversion results, reduces the risk of local optima, and enhances computational efficiency and inversion accuracy.
Smart Images
Figure CN121703903B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of micro-motion exploration technology, and in particular to a method, apparatus, electronic device, and storage medium for micro-motion dispersion curve inversion. Background Technology
[0002] Microseismic exploration technology is a surface wave exploration method that uses background noise generated by natural phenomena such as earthquakes as source signals. By analyzing its dispersion characteristics, inversion algorithms can be used to detect shallow surface shear wave velocity structures. It is widely used in earthquake engineering, urban geological surveys, and engineering site investigations. Rayleigh waves are surface waves that propagate along the Earth's surface or rock strata interfaces. Because they mainly propagate in the vertical direction, they are easy to observe and record, and their energy constitutes the majority of the total energy of the surface wave wavelength.
[0003] Rayleigh wave dispersion curve inversion is characterized by nonlinearity, multiple parameters, and multiple extrema. Although the direct calculation method based on the half-wavelength interpretation method does not require an inversion process and is simple to implement, it is highly dependent on the measured dispersion curve and can only derive approximate results, which is not applicable to complex media conditions. Traditional single inversion algorithms are prone to getting trapped in local optima in dispersion curve inversion, and cannot effectively explore the global optimum, resulting in low computational efficiency and inversion results that deviate from the actual formation media conditions.
[0004] Therefore, there is an urgent need to propose a method, device, electronic equipment, and storage medium for micro-dispersion curve inversion to solve the technical problems existing in the single inversion algorithm in dispersion curve inversion, which are prone to getting trapped in local optima, unable to effectively explore the global optimum, resulting in low computational efficiency and inversion results that deviate from the actual formation medium. Summary of the Invention
[0005] In view of this, it is necessary to provide a method, apparatus, electronic device and storage medium for micro-dispersion curve inversion, in order to solve the technical problems existing in the single inversion algorithm in dispersion curve inversion, which are prone to getting trapped in local optima, unable to effectively explore the global optimum, resulting in low computational efficiency and inversion results deviating from the actual formation medium.
[0006] To address the aforementioned problems, in a first aspect, the present invention provides a method for inverting micro-motion dispersion curves, comprising: Acquire micro-motion waveform data and measured dispersion curves of the target site at different frequencies; The dispersion curve of the micro-motion waveform data was extracted by the high-resolution frequency-wavenumber method to obtain the initial Rayleigh wave dispersion curve, and an inversion model was constructed based on the initial Rayleigh wave dispersion curve. The initial Rayleigh wave dispersion curve is recursively calculated and forward modeled to obtain the recursive theoretical curve; The initial Rayleigh wave dispersion curve is corrected based on the difference between the recursive theoretical curve and the measured dispersion curve to obtain the target Rayleigh wave dispersion curve. The target Rayleigh wave dispersion curve is iteratively filtered according to the inversion model to obtain the inversion result.
[0007] In one possible implementation, the inversion model includes physical constraints; these physical constraints include Poisson's ratio correlation constraints, layer depth increment constraints, and low-speed layer exclusion constraints.
[0008] In one possible implementation, the recursive and forward modeling of the initial Rayleigh wave dispersion curve to obtain the recursive theoretical curve includes: Set the maximum depth and thin layer thickness during the forward modeling process; The target site is divided into thin layers of equal thickness based on the maximum depth and the thickness of the thin layer; The waveform parameters are determined based on the initial Rayleigh wave dispersion curve; the waveform parameters include frequency range, frequency division type, and number of frequency points. Based on the waveform parameters, the depth corresponding to the high-frequency non-dispersion segment of the initial Rayleigh wave dispersion curve is calculated using the half-wavelength method, resulting in a phase velocity-depth curve. The number of frequency points within each thin layer is recursively calculated based on the direct recursive method and the phase velocity-depth curve to obtain the shear wave velocity of each thin layer. Based on all shear wave velocities, a preliminary shear wave velocity curve is obtained; The preliminary shear wave velocity curve is subjected to forward modeling to obtain the recursive theoretical curve.
[0009] In one possible implementation, the step of recursively calculating the number of frequency points within each thin layer range based on a direct recursive method to obtain the shear wave velocity of each thin layer includes: Based on the phase velocity-depth curve, the characteristic equations of Rayleigh wave phase velocity, longitudinal wave and transverse wave in a uniform half-space, and the relationship between longitudinal wave, transverse wave and Poisson's ratio, the conversion coefficient is obtained; Each thin layer is arranged in ascending order according to its depth to obtain the arrangement sequence; The shear wave velocity of the first thin layer is obtained by recursively calculating the average wave velocity of the number of frequency points, the conversion coefficient, the current depth, and the Rayleigh wave wavelength within the range of the first thin layer in the arrangement order. Based on the thickness and shear wave velocity of the previous thin layer, the average wave velocity of the number of frequency points within the range of the middle thin layer in the arrangement sequence, the conversion coefficient, the current depth, and the Rayleigh wave wavelength are recursively calculated to obtain the shear wave velocity of the middle thin layer.
[0010] In one possible implementation, the step of correcting the initial Rayleigh wave dispersion curve based on the difference between the recursive theoretical curve and the measured dispersion curve to obtain the target Rayleigh wave dispersion curve includes: The difference between the recursive theoretical curve and the measured dispersion curve across the entire frequency band is calculated to obtain the deviation ratio for each frequency band. The average deviation value is obtained based on all deviation ratios and the number of frequency points. The correction coefficient is obtained based on the average deviation value, and the initial Rayleigh wave dispersion curve is corrected based on the correction coefficient to obtain the corrected shear wave velocity curve. When the number of iterations is less than the preset number of iterations, the deviation correction is performed again after forward modeling the modified shear wave velocity curve. When the number of iterations is not less than the preset number of iterations, the modified shear wave velocity curve is determined to be the target Rayleigh wave dispersion curve.
[0011] In one possible implementation, the iterative screening of the target Rayleigh wave dispersion curve based on the inversion model to obtain the inversion result includes: After inputting the target Rayleigh wave dispersion curve into the inversion model, an initial sample set is generated through random walk; Forward modeling is performed on each initial sample in the initial sample set to obtain the corresponding theoretical dispersion curve; The theoretical dispersion curve is compared with the preset observation data to obtain the mismatch degree of each initial sample; The initial sample set is filtered according to the mismatch degree to obtain the initial Voronoi unit set; The initial samples in the initial Voronoi unit set are iteratively filtered to obtain the inversion result.
[0012] In one possible implementation, the iterative screening of the initial samples in the initial Voronoi unit set to obtain the inversion result includes: The initial samples in the initial Voronoi unit set are sorted according to the mismatch degree to obtain a number of first samples whose mismatch degree is lower than a preset mismatch degree; The active region is formed by performing a union operation on the multiple first samples; New samples are generated in the active region based on the Gibbs sampler, and the new mismatch of the new samples is calculated. The valid samples are added to the initial samples based on the new mismatch degree to obtain a new sample set; The minimum parameter space range surrounding the active region is determined based on the new sample set, and the minimum parameter space range is dynamically scaled according to the range of each parameter axis to obtain the target parameter space. Based on the new sample set and the target parameter space, a new set of Voronoi units is generated; After removing Voronoi units in the new Voronoi unit set whose number of valid samples is less than the preset number of samples, it is determined whether the number of iterations has been reached or the minimum mismatch is less than the preset threshold. If not, then iteratively filter the set of Voronoi units after removal; If so, the inversion result is obtained based on the set of Voronoi units after removal.
[0013] Secondly, the present invention also provides a micro-motion dispersion curve inversion device, comprising: The data acquisition module is used to acquire micro-motion waveform data and measured dispersion curves of the target site at different frequencies; The curve extraction module is used to extract the dispersion curve of the micro-motion waveform data using the high-resolution frequency-wavenumber method, obtain the initial Rayleigh wave dispersion curve, and construct an inversion model based on the initial Rayleigh wave dispersion curve. The curve recursion module is used to perform recursive and forward modeling calculations on the initial Rayleigh wave dispersion curve to obtain the recursive theoretical curve. The curve correction module is used to correct the initial Rayleigh wave dispersion curve based on the difference between the recursive theoretical curve and the measured dispersion curve, so as to obtain the target Rayleigh wave dispersion curve. The iterative screening module is used to iteratively screen the target Rayleigh wave dispersion curve according to the inversion model to obtain the inversion result.
[0014] Thirdly, embodiments of the present invention disclose an electronic device, including: a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the various steps of the above-described micro-dispersion curve inversion method embodiments.
[0015] Fourthly, embodiments of the present invention disclose a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the various steps of the above-described micro-dispersion curve inversion method embodiments.
[0016] The beneficial effects of this invention are: acquiring micro-motion waveform data and measured dispersion curves of the target site at different frequencies; extracting dispersion curves from the micro-motion waveform data using the high-resolution frequency-wavenumber method to obtain the initial Rayleigh wave dispersion curve, and constructing an inversion model based on the initial Rayleigh wave dispersion curve; performing recursive and forward calculations on the initial Rayleigh wave dispersion curve to obtain the recursive theoretical curve; correcting the initial Rayleigh wave dispersion curve based on the difference between the recursive theoretical curve and the measured dispersion curve to obtain the target Rayleigh wave dispersion curve; and generating theoretical models through recursive and forward calculations. The dispersion curve is used to correct the input measured dispersion curve, suppressing the distortion of some frequency bands in the extraction of the measured dispersion curve under limited signal acquisition conditions or noise interference, optimizing the matching degree between the dispersion curve and the real strata, and improving the quality of the dispersion curve. Furthermore, the target Rayleigh wave dispersion curve is iteratively screened according to the inversion model to obtain the inversion result. By constructing an inversion model and combining recursive correction and iterative screening mechanisms, the search performance of the data is improved through iterative screening of the inversion model parameters, reducing the risk of getting trapped in local optima and improving the efficiency of inversion and calculation. Attached Figure Description
[0017] Figure 1 This is a schematic flowchart of an embodiment of the micro-motion dispersion curve inversion method provided by the present invention; Figure 2 A schematic diagram of a coordinate system for an embodiment of the fundamental dispersion curve of a simulated layered medium model provided by the present invention; Figure 3 For the present invention Figure 1 A schematic diagram of an embodiment of step S102; Figure 4 For the present invention Figure 1 A schematic flowchart of an embodiment of step S103; Figure 5 A schematic diagram of a coordinate system for an embodiment of the preliminary shear wave velocity curve provided by the present invention; Figure 6 A schematic diagram of an embodiment for determining the inversion results provided by the present invention; Figure 7 This is a schematic diagram of an embodiment of the micro-motion dispersion curve inversion device provided by the present invention; Figure 8 A schematic diagram of an embodiment of the electronic device provided by the present invention. Detailed Implementation
[0018] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.
[0019] In existing technologies, micro-motion exploration technology utilizes background noise analysis to determine dispersion characteristics for shallow surface exploration. Traditional methods rely on half-wavelength interpretation for direct calculations, which is overly dependent on measured data, leading to insufficient applicability under complex media conditions. Single inversion algorithms are prone to getting trapped in local optima, making global search difficult and causing inversion results to deviate from the actual geological structure. In a certain urban engineering survey, multiple layers of alternating soft and hard geological conditions existed. Traditional methods, unable to effectively handle the multiple extrema of dispersion curves, resulted in a deviation of over 30% between the shear wave velocity inversion results and borehole data.
[0020] To address these issues, researchers discovered that the core challenge in dispersion curve inversion lies in balancing the accuracy of the initial model with global search capability. While traditional recursive methods can construct an initial model, they lack a dynamic correction mechanism; single inversion algorithms, though capable of global search, are limited by the quality of the initial model. Analysis revealed that iteratively correcting the deviation between the recursive calculation results and measured data can create an adaptive initial model. Combined with an intelligent selection mechanism, this approach preserves the stability of the recursive method while overcoming the limitations of local optima.
[0021] like Figure 1 As shown, a specific embodiment of the present invention discloses a method for inverting micro-motion dispersion curves, comprising: S101. Obtain the micro-motion waveform data and measured dispersion curve of the target site at different frequencies.
[0022] This invention can be applied to the fields of civil engineering and urban planning, and can be used for foundation bearing capacity assessment, site soil type classification, foundation stability analysis, underground cavity detection, etc. Specifically, it can include underground hazard detection, site category assessment, and building foundation investigation. This invention can also be applied to a micro-motion dispersion curve inversion system, which can be a software system running on a terminal device, such as a server, tablet, or mobile phone. This invention does not limit the specific type of terminal device. It can collect micro-motion waveform data at different frequencies from the target site to be tested, and can also perform actual measurements to collect measured dispersion curves. The target site can be any site that needs to be tested, and can be set according to actual conditions.
[0023] S102. The dispersion curve of the micro-motion waveform data is extracted by the high-resolution frequency-wavenumber method to obtain the initial Rayleigh wave dispersion curve, and an inversion model is constructed based on the initial Rayleigh wave dispersion curve.
[0024] Among them, the high-resolution frequency-wavenumber method refers to using spatial autocorrelation algorithms or wavenumber transform algorithms for dispersion energy spectrum analysis. It converts the spatiotemporal domain signal into the frequency-wavenumber domain through a two-dimensional Fourier transform. Specifically, a circular array data can be used to enhance wavenumber resolution. This technique can effectively separate the fundamental and higher-order mode energies, improving the accuracy of dispersion curve extraction. Therefore, the high-resolution frequency-wavenumber method can be used to extract dispersion curves from micro-motion waveform data, obtaining the initial Rayleigh wave dispersion curve, such as... Figure 2 As shown, Figure 2 This is an example of the fundamental dispersion curve of a simulated layered medium model, namely the initial Rayleigh wave dispersion curve extracted by the high-resolution frequency-wavenumber method, where velocity is velocity and frequency is frequency.
[0025] S103. Perform recursive and forward modeling calculations on the initial Rayleigh wave dispersion curve to obtain the recursive theoretical curve.
[0026] The recursive calculation refers to calculating the shear wave velocity layer by layer from shallow to deep based on a layered medium model. Specifically, it uses a direct recursive method combined with the thin-layer equivalence principle to derive the wave velocity of the lower layer using known layer parameters. This process constructs an initial velocity model that conforms to physical laws. The forward modeling calculation refers to generating theoretical dispersion curves using numerical simulation methods based on the wave equation. This can be achieved using the finite difference method or the spectral element method and is used to verify the rationality of the recursive results.
[0027] S104. Correct the initial Rayleigh wave dispersion curve based on the difference between the recursive theoretical curve and the measured dispersion curve to obtain the target Rayleigh wave dispersion curve.
[0028] The correction process can be performed by correcting the initial Rayleigh wave dispersion curve point by point after determining the difference.
[0029] S105. The target Rayleigh wave dispersion curve is iteratively screened according to the inversion model to obtain the inversion result.
[0030] Among them, iterative screening refers to dynamically adjusting the inversion process through mismatch evaluation and parameter space optimization. Specifically, it adopts Voronoi unit partitioning combined with Gibbs sampling to achieve efficient global search in the parameter space.
[0031] Specifically, micromotion signals are first acquired using a three-component seismograph. After preprocessing, the dispersion curve of the fundamental Rayleigh wave is extracted using the frequency-wavenumber method as the initial model. When constructing the layered medium model, the layer interfaces are determined based on the extreme points of the dispersion curves, and the shear wave velocities of each layer are calculated using a direct recursive method to form a preliminary velocity structure. A full-frequency deviation analysis is performed between the theoretical curves calculated by forward modeling and the measured curves. The weighted least squares method is used to calculate correction coefficients to adjust the velocity values of the initial model. In the iterative screening stage, an initial sample set is generated through random walks. The mismatch between the theoretical curves and measured data for each sample is calculated. Low-mismatch samples are retained to construct active regions, and a dynamic parameter space scaling strategy is used to gradually approach the optimal solution.
[0032] Compared with existing technologies, this embodiment provides the following: acquiring micro-motion waveform data and measured dispersion curves of the target site at different frequencies; extracting dispersion curves from the micro-motion waveform data using the high-resolution frequency-wavenumber method to obtain an initial Rayleigh wave dispersion curve; constructing an inversion model based on the initial Rayleigh wave dispersion curve; performing recursive and forward calculations on the initial Rayleigh wave dispersion curve to obtain a recursive theoretical curve; correcting the initial Rayleigh wave dispersion curve based on the difference between the recursive theoretical curve and the measured dispersion curve to obtain the target Rayleigh wave dispersion curve; and generating a target Rayleigh wave dispersion curve through recursive and forward calculations. The theoretical dispersion curve is used to correct the measured dispersion curve, suppressing the distortion of some frequency bands in the extraction of the measured dispersion curve under limited signal acquisition conditions or noise interference, optimizing the matching degree between the dispersion curve and the real strata, and improving the quality of the dispersion curve. Furthermore, the target Rayleigh wave dispersion curve is iteratively screened according to the inversion model to obtain the inversion result. By constructing an inversion model and combining recursive correction and iterative screening mechanisms, the search performance of the data is improved through iterative screening of the inversion model parameters, reducing the risk of getting trapped in local optima and improving the efficiency of inversion and calculation.
[0033] In some embodiments of the present invention, the inversion model may include physical constraints; physical constraints include, but are not limited to, Poisson's ratio correlation constraints, layer depth increment constraints, and low-speed layer exclusion constraints.
[0034] In some embodiments of the present invention, such as Figure 3 As shown, step S103 includes: S301. Set the maximum depth and thin layer thickness during the forward modeling process; Thin layer thickness refers to the thickness of the stratigraphic unit based on the wavelength corresponding to the highest frequency, and its function is to establish the basic unit for the layered medium model.
[0035] S302. Divide the target site into thin layers of equal thickness according to the maximum depth and the thickness of the thin layer.
[0036] Among them, thin layers of equal thickness refer to the geological structure being divided into layers of the same thickness. Specifically, the number of layers can be determined by the ratio of the maximum depth to the thickness of the thin layer, thereby ensuring that the physical properties of each layer are homogeneous.
[0037] S303. Determine the waveform parameters based on the Rayleigh wave dispersion curve; the waveform parameters include the frequency range, frequency division type, and number of frequency points.
[0038] The frequency range refers to the frequency range covered by the Rayleigh wave dispersion curve, which can be determined by analyzing the energy distribution range of the effective signal in the micro-motion waveform data. Its function is to limit the effective domain of the inversion model. The frequency division type refers to the mode classification of Rayleigh wave propagation, which can be achieved using a fundamental mode or higher-order mode discrimination method. Its function is to determine the physical characteristics of the dispersion curve. The number of frequency points refers to the number of frequency sampling points selected during discretization processing. It can be set using equal or logarithmic intervals, and its function is to balance computational accuracy and efficiency.
[0039] S304. Based on the waveform parameters, the depth corresponding to the high-frequency non-dispersion segment of the initial Rayleigh wave dispersion curve is calculated using the half-wavelength method, and the phase velocity-depth curve is obtained as shown in formula (1): (1) in, Freq and V ph These represent the frequency and phase velocity values corresponding to each frequency point of the initial Rayleigh wave dispersion curve. Depth This represents the depth corresponding to that frequency point.
[0040] The longitudinal wave interval, transverse wave interval, and Poisson's ratio interval can be determined based on the frequency interval and frequency division type. Thus, the minimum and maximum values of the longitudinal wave, transverse wave, transverse wave, minimum and maximum values of the transverse wave, and minimum and maximum values of the Poisson's ratio can be determined. The number of frequency points can be used to represent the density. The inversion model is shown in formula (2). (2) In the formula, For the minimum value of the longitudinal wave, This represents the maximum value of the longitudinal wave. This is the minimum value of the transverse wave. This represents the maximum value of the transverse wave. For layer thickness, For density, This represents the minimum Poisson's ratio. This represents the maximum Poisson's ratio; shallow borehole data or regional geological experience can be incorporated to adjust the model parameters.
[0041] S305. Based on the direct recursive method and the phase velocity-depth curve, the number of frequency points within each thin layer is recursively calculated to obtain the shear wave velocity of each thin layer.
[0042] Among them, the direct recursive method refers to the calculation method of solving unknown parameters step by step using known parameters. Specifically, it can be realized by using the relationship between Rayleigh wave phase velocity and transverse wave velocity in a uniform half space, thereby establishing the mathematical relationship between parameters of each layer.
[0043] Specifically, the initial Rayleigh wave dispersion curve can be converted into a wave velocity-depth curve based on the correspondence between the half wavelength and propagation depth of the Rayleigh wave in formula (1), and then recursively calculated in subsequent processes.
[0044] In some embodiments of the present invention, such as Figure 4 As shown, step S305 includes: S401. Based on the phase velocity-depth curve, the characteristic equations of Rayleigh wave phase velocity, longitudinal wave and transverse wave in a uniform half-space, and the relationship between longitudinal wave, transverse wave and Poisson's ratio, the conversion coefficient is obtained.
[0045] Among them, the phase velocity-depth curve represents the relationship between depth and Rayleigh wave phase velocity, thus allowing the Rayleigh wave phase velocity to be obtained. The conversion coefficient is a proportional parameter derived from the characteristic equations and Poisson's ratio relationship between the Rayleigh wave phase velocity and the longitudinal and transverse waves in a uniform half-space. Specifically, it can be achieved by numerical fitting or analytical calculation, and is used to convert the Rayleigh wave phase velocity into the shear wave velocity. The characteristic equations of the Rayleigh wave phase velocity, longitudinal wave, and transverse wave are shown in formula (3): (3) In the formula, For Rayleigh wave phase velocity, For transverse wave velocity, This refers to the longitudinal wave velocity.
[0046] The relationship between longitudinal waves, transverse waves, and Poisson's ratio is shown in formula (4): (4) In the formula, Given the Poisson's ratio, when the Poisson's ratio is between 0 and 0.5, the characteristic equation has a fitted curve, simplifying the calculation of the conversion coefficients. As shown in formula (5): (5) For the surface wave method, parameters such as layer thickness and Poisson's ratio have a much smaller impact on the dispersion curve than shear wave velocity. Here, a fixed value for Poisson's ratio can be taken for different layered media.
[0047] S402. Arrange each thin layer in ascending order according to its depth to obtain the arrangement order.
[0048] Incremental arrangement refers to arranging the thin layers in order from shallow to deep according to their burial depth. Specifically, an ascending sorting algorithm can be used to ensure that the recursive calculation progresses gradually from the shallow layer to the deep layer.
[0049] S403. The average wave velocity, conversion coefficient, current depth, and Rayleigh wave wavelength of the number of frequency points within the range of the first thin layer in the arrangement sequence are recursively calculated to obtain the shear wave velocity of the first thin layer.
[0050] The mean wave velocity refers to the average value of the Rayleigh wave phase velocity corresponding to multiple frequency points within the same thin layer. Specifically, it can be calculated using the arithmetic mean or weighted average method to eliminate the influence of single-frequency data fluctuations on the shear wave velocity calculation. The complete shear wave velocity curve is obtained by calculating the known shear wave velocities of all upper layers in ascending order of depth. In the special case where there are no data points within the thin layer, if the thin layer is the first thin layer, the data with the smallest depth in the wave velocity-depth curve is taken for calculation. Since the data are arranged in ascending order of depth, the data with the smallest depth in the first thin layer is the depth of the corresponding thin layer. If this layer is the last thin layer, the thickness of the thin layer is adjusted, or interpolation is used to obtain the depth of the corresponding thin layer. Thus, the mean wave velocity, conversion coefficient, current depth, and Rayleigh wave wavelength of the frequency points within the thin layer range are recursively calculated to obtain the shear wave velocity of the thin layer. As shown in formula (6): (6) In the formula, The thin layer is either the first thin layer or the last thin layer of the last thin layer; This represents the average wave velocity of all wave velocity-depth curve data points within the thin layer. β This is the conversion coefficient for converting Rayleigh wave phase velocity to shear wave velocity.
[0051] S404. Based on the thickness and shear wave velocity of the upper thin layer, the average wave velocity, conversion coefficient, current depth, and Rayleigh wave wavelength of the number of frequency points within the range of the middle thin layer in the arrangement sequence are recursively calculated to obtain the shear wave velocity of the middle thin layer.
[0052] If the thin layer is a mid-thin layer, then linear interpolation is performed on the two data points in the wave velocity-depth curve that are closest in depth to each other above and below the thin layer to obtain equivalent wave velocity data for calculation; specifically, the calculation of the first... j Layer shear wave velocity For example, the calculation is shown in formula (7): (7) In the formula, For the first j Calculation results of shear wave velocity of the layer, x The depth values are the data points of the wave velocity-depth curve. The wavelength of Rayleigh wave, For the firsti The thickness of the thin layer, For the first i Shear wave velocity of the layer.
[0053] Furthermore, for cases where there are no data points within the thin layer, and calculations are performed using nearby data or interpolation, the thin layer thickness in the input parameters can be adjusted and recalculated.
[0054] S306. Based on all shear wave velocities, obtain the preliminary shear wave velocity curve.
[0055] Shear wave velocity refers to the propagation speed of shear waves in the formation. It can be calculated by combining recursive formulas with conversion coefficients. For example, the velocity value of the current layer can be derived by using the parameter relationship between adjacent layers, thereby reflecting the stiffness characteristics of the formation.
[0056] S307. Perform forward modeling on the preliminary shear wave velocity curve to obtain the recursive theoretical curve.
[0057] Forward modeling refers to the process of predicting dispersion curves based on theoretical models. Specifically, the Haskell-Thomson transfer matrix method can be used to perform forward modeling on the initial shear wave velocity curve to obtain the recursive theoretical curve. Among the calculation parameters, the P-wave velocity... V p Based on its relationship with transverse waves V s Compared to Poisson v The relationship is calculated as follows: .
[0058] Specifically, the number of layers is first determined based on the maximum depth and the thickness of the thin layer. For example, when the maximum depth is 30 meters and the thin layer thickness is 3 meters, the site is divided into 10 layers. Then, for each layer's frequency points, a direct recursive method is used to calculate the shear wave velocity sequentially. For instance, the initial velocity of the first layer is calculated based on the assumption of a uniform half-space, and the velocities of subsequent layers are recursively derived from the parameters of the previous layer. After combining the shear wave velocities of all layers to form a preliminary shear wave velocity curve, a theoretical dispersion curve is generated through forward modeling. For example, the velocity-depth model is input into the forward modeling algorithm to simulate the Rayleigh wave propagation process, and the output theoretical dispersion curve is used for subsequent corrections.
[0059] Through the above technical solution, this application achieves efficient recursive calculation of shear wave velocity, solving the problem of insufficient accuracy caused by uneven layering or isolated parameters in traditional methods. By combining equal-thickness layering with recursive calculation, the parameters of each layer are ensured to have a clear physical correlation, providing a reliable basis for subsequent dispersion curve correction. At the same time, forward modeling verifies the rationality of the shear wave velocity curve, avoids invalid iterations, and improves the overall inversion efficiency.
[0060] In some embodiments of the present invention, step S104 includes: The difference between the recursive theoretical curve and the measured dispersion curve across the entire frequency band is calculated to obtain the deviation ratio for each frequency band.
[0061] The deviation ratio refers to the proportion of difference between the theoretical curve and the measured curve in each frequency band. It can be calculated by determining the relative error or absolute error, and is used to quantify the degree of deviation between the theoretical model and actual observation data. Specifically, the calculation involves inputting the dispersion curve. With recursive theory curve Deviation ratio across the entire frequency band As shown in formula (8): (8) The average deviation value is obtained based on all deviation ratios and the number of frequency points.
[0062] The average deviation value refers to the comprehensive statistical measure of the deviation ratios across all frequency bands. It can be calculated using a weighted average or arithmetic average method and is used to reflect the overall correction requirements. The specific statistical average deviation value... As shown in formula (9): (9) In the formula, This represents the number of frequency points in the dispersion curve.
[0063] The correction coefficient is obtained based on the average deviation value, and the initial Rayleigh wave dispersion curve is corrected based on the correction coefficient to obtain the corrected shear wave velocity curve.
[0064] The correction factor is an adjustment factor generated based on the average deviation value. It can be obtained through linear interpolation or nonlinear function mapping and is used to dynamically adjust the initial curve parameters. Specific correction factors... The calculation is shown in formula (10): (10) The shear velocity curve is corrected using a correction factor. This step is repeated with the corrected shear velocity curve, and the new dispersion curve is calculated using the forward modeling method.
[0065] When the number of iterations is less than the preset number of iterations, the deviation correction is performed again after forward modeling the modified shear wave velocity curve.
[0066] The preset iteration count refers to the maximum number of loops set in advance. Specifically, it can be set to a fixed value or an adaptive adjustment value according to computing resources or accuracy requirements, and is used to control the termination conditions of the correction process.
[0067] When the number of iterations is not less than the preset number of iterations, the modified shear wave velocity curve is determined as the target Rayleigh wave dispersion curve.
[0068] By setting parameters to control the number of iterations, the corrected dispersion curve is finally obtained. Specifically, in the correction process, the local error distribution of each frequency band is first obtained through full-band deviation calculation, and the overall deviation level is obtained by combining the number of frequency points. After generating correction coefficients based on the average deviation, the parameters of the initial shear wave velocity curve are adjusted to form the corrected intermediate result. If the number of iterations does not reach the threshold, the correction result is re-input into the forward modeling module to generate a new theoretical curve, and the deviation correction cycle is repeated; if the iteration termination condition is met, the final corrected curve is output as the target dispersion curve. This process achieves progressive matching between the theoretical model and the measured data through a dynamic feedback mechanism.
[0069] By employing the aforementioned technical solution, this application effectively addresses the problem of insufficient adaptability of traditional single-correction methods to complex dispersion curves. Through multiple rounds of iterative correction, it gradually reduces the deviation between the theoretical model and measured data, thereby improving the accuracy and reliability of the inversion results. This solution is particularly suitable for scenarios with noise interference or complex geological structures, reducing reliance on the accuracy of the initial model and enhancing the robustness of the inversion process.
[0070] In some embodiments of the present invention, step S105 includes: After inputting the target Rayleigh wave dispersion curve into the inversion model, an initial sample set is generated through random walk.
[0071] The random walk to generate the initial sample set refers to randomly generating initial sample points in the parameter space to cover the possible solution space range. This can be achieved using the Markov chain Monte Carlo method to ensure the diversity and breadth of the samples. Specifically, the parameters of the inversion model are checked layer by layer to ensure their validity. For cases where the numerical range is too large, the parameters are corrected based on physical conditions and numerical constraints to avoid invalid searches caused by excessively large parameter ranges. Based on the corrected inversion model, the first valid parameter space model is calculated, satisfying physical conditions such as Poisson's ratio constraints and layer depth order. The initial sample set is then generated using a Markov chain random walk.
[0072] Forward modeling is performed on each initial sample in the initial sample set to obtain the corresponding theoretical dispersion curve.
[0073] Forward modeling refers to the process of converting sample parameters into theoretical dispersion curves based on a physical model. This can be achieved using the finite element method or a semi-analytical method to simulate dispersion responses under different parameters. The specific process of using the forward modeling method can be set according to actual conditions, and this embodiment of the invention does not impose any limitations.
[0074] The theoretical dispersion curve is compared with the preset observation data to obtain the mismatch degree of each initial sample.
[0075] The mismatch degree refers to the degree of difference between the theoretical dispersion curve and the measured data. It can be calculated using the root mean square error or correlation coefficient to quantify the degree of sample matching. For each initial sample, the mismatch degree with the observed data can be calculated using a forward modeling method. The calculation is shown in formula (11): (11) In the formula, For the input dispersion curve at frequency f i Phase velocity value at that location, The dispersion curve obtained from the forward modeling of the sample at the frequency f i Phase velocity value at that location, The number of frequency points in the dispersion curve. The uncertainty is the value of the sample under consideration.
[0076] The initial sample set is filtered based on the mismatch degree to obtain the initial Voronoi unit set.
[0077] The Voronoi unit set refers to dividing the sample points into multiple geometric units, each unit representing a region in the parameter space. This can be implemented using a spatial segmentation algorithm, which facilitates subsequent iterative selection and parameter optimization.
[0078] The initial samples in the initial Voronoi unit set are iteratively filtered to obtain the inversion results.
[0079] In this embodiment of the invention, after each iteration, the search step size is dynamically scaled according to the sensitivity of the inversion model to each parameter in the active region. At the same time, Voronoi units with an effective sample generation rate of <10% are removed and suboptimal units are added to obtain the inversion result.
[0080] Specifically, after the target Rayleigh wave dispersion curve is input into the inversion model, an initial sample set covering the parameter space is first generated through a random walk to ensure the diversity and breadth of the initial samples. Each initial sample represents a set of possible combinations of inversion parameters, and the corresponding theoretical dispersion curve is generated through forward modeling. The theoretical dispersion curve is compared with the preset observation data to calculate the mismatch, and samples with low mismatch are selected to form the initial Voronoi unit set. In the iterative selection, the samples in the initial Voronoi unit set are further optimized by dynamically adjusting the parameter space range and generating new samples, gradually narrowing the solution space and approaching the optimal solution. When the number of iterations reaches a preset value or the mismatch is lower than a preset threshold, the final selected sample set is the inversion result. Figure 5 As shown, Figure 5The output is the shear wave velocity curve of the inversion result, where velocity is the velocity and depth is the depth. Figure 5 The gray dashed line represents the initial Rayleigh wave dispersion curve input to the inversion model, specifically the shear wave velocity range. For example, the range for Depth is 0-5, and for Velocity it's 100-200, 5-10, 10-20, 20-25, and 25-30. The solid black line on the left represents the wave velocity-depth curve, and the black polygonal line on the right represents the output shear wave velocity curve of the inversion result.
[0081] By employing the above technical solutions, this application solves the problem of traditional inversion methods easily getting trapped in local optima, significantly improving the accuracy and reliability of the inversion results. Through dynamic parameter space adjustment and a global search strategy, it can adapt to the inversion needs of dispersion curves under complex geological conditions, providing more accurate shear wave velocity structure data for engineering exploration and geological surveys.
[0082] In some embodiments of the present invention, such as Figure 6 As shown, the initial samples in the initial Voronoi unit set are iteratively filtered to obtain the inversion results, including: S601. Sort the initial samples in the initial Voronoi unit set according to the mismatch degree to obtain multiple first samples with a mismatch degree lower than the preset mismatch degree.
[0083] Specifically, the initial samples in the initial Voronoi unit set can be sorted according to the mismatch degree, and samples with a mismatch degree lower than a preset mismatch degree can be selected. n r The first sample.
[0084] S602. Perform a union operation on multiple first samples to form an active region.
[0085] Among them, can n r The union of the first samples constitutes the active region, which is the target space of the current search.
[0086] S603. Generate new samples in the active region based on the Gibbs sampler and calculate the new mismatch degree of the new samples.
[0087] Among them, the Gibbs sampler refers to a Markov chain Monte Carlo method that updates parameter values dimension by dimension through a conditional probability distribution. Specifically, it can be implemented by a sampling strategy that associates the probability density function with the mismatch degree, and is used to generate new samples that conform to the current parameter distribution characteristics in the active region.
[0088] S604. Add the valid samples to the initial samples according to the new mismatch degree to obtain a new sample set.
[0089] Specifically, the validity of a new sample can be determined based on the magnitude of the new mismatch, and the valid sample can be added to the initial sample to obtain a new sample set. Specifically, a threshold can be set to determine whether the new mismatch is valid, or other methods can be used for determination. The specific settings can be set according to the actual situation, and the embodiments of the present invention are not limited here.
[0090] S605. Determine the minimum parameter space range surrounding the active region based on the new sample set, and dynamically scale the minimum parameter space range according to the range of each parameter axis to obtain the target parameter space.
[0091] Dynamic scaling refers to dynamically adjusting the boundary range of the parameter space based on the distribution of the sample set. Specifically, extreme value statistics or quantile methods can be used to determine the variation range of each parameter axis, thus avoiding the problem of limited search range caused by a fixed parameter space. The minimum parameter space range is dynamically scaled based on the parameter axis range of the active region. The coefficient of variation of the parameters is calculated, and the search step size is reduced for highly sensitive parameters and increased for less sensitive parameters to obtain the target parameter space.
[0092] Specifically, after obtaining the new sample set, the minimum parameter space range surrounding the active region can be determined. Based on the effective samples within the active region, the coefficient of variation of the parameters is calculated. The coefficient of variation of the parameters is calculated as shown in formula (12): (12) In the formula, For the standard deviation of the effective sample parameters, The mean of the parameters for the effective samples.
[0093] Then, the minimum parameter space range is dynamically scaled according to the range of each parameter axis. This, combined with engineering practice experience, is particularly useful for highly sensitive parameters such as shear wave velocity or... CV A parameter greater than 0.3 reduces the step size and improves search accuracy; for low-sensitivity parameters such as density, or CV Parameters <0.1 increase the search step size, improving the efficiency of the global search. Furthermore, setting a dynamically scaled range avoids computational stagnation due to excessively small step sizes or uncontrolled exploration due to excessively large step sizes, thus obtaining the target parameter space, maintaining a larger search range for highly sensitive parameters, and narrowing the search range for less sensitive parameters.
[0094] S606. Generate a new set of Voronoi units based on the new sample set and the target parameter space.
[0095] After obtaining the target parameter space, a new set of Voronoi units can be regenerated based on the new sample set and the target parameter space.
[0096] S607. After removing Voronoi units in the new Voronoi unit set whose number of valid samples is less than the preset number of samples, determine whether the number of iterations has been reached or the minimum mismatch is less than the preset threshold. S608. If not, then iteratively filter the set of Voronoi units after removal. S609. If so, then the inversion result is obtained based on the set of Voronoi units after removal.
[0097] After obtaining the new Voronoi element set, Voronoi elements with a small number of valid samples can be removed. Then, it can be determined whether the current iteration count has reached the required number of iterations or whether the minimum mismatch is less than a preset threshold. If not, the iteration process S601-S609 is repeated on the removed Voronoi element set. If so, the inversion result is obtained based on the removed Voronoi element set. The resulting shear wave velocity structure is then interpreted to guide micro-motion exploration, site classification, and other related work.
[0098] Specifically, this method selects low-error samples by sorting them by mismatch degree, and uses the active region formed by these samples as the core space for parameter optimization. New samples are generated within the active region using a Gibbs sampler, and the sample set is updated based on the mismatch evaluation results of these new samples, effectively improving sample quality. Furthermore, the parameter space range is dynamically adjusted according to the sample distribution, allowing subsequent iterations to focus on a more reasonable search region. By iteratively eliminating invalid Voronoi units and determining the termination condition, the method finally obtains the inversion result that meets the accuracy requirements.
[0099] Through the above technical solution, this application solves the problem that traditional inversion algorithms are prone to getting trapped in local optima and have low computational efficiency. It realizes the effective coordination of dynamic optimization of parameter space and sample generation, and significantly shortens the number of iterations while ensuring inversion accuracy. It is especially suitable for efficient inversion of micro-dispersion curves under complex media conditions.
[0100] This invention generates theoretical dispersion curves using a direct recursive method to correct the input measured dispersion curves, suppressing distortion in certain frequency bands during measured dispersion curve extraction under limited signal acquisition conditions or noise interference, and optimizing the matching degree between the dispersion curves and the actual geological formation. The improved neighborhood algorithm enhances the search performance in the parameter space by dynamically scaling the inversion model parameters, reducing the risk of getting trapped in local optima. By adding model validity checks and parameter decoupling, invalid computations in the early stages of iteration are reduced, improving inversion efficiency.
[0101] To better implement the micro-motion dispersion curve inversion method in this embodiment of the invention, correspondingly, this embodiment of the invention also provides a micro-motion dispersion curve inversion device, such as... Figure 7 As shown, the micro-dispersion curve inversion device 700 includes: Data acquisition module 701 is used to acquire micro-motion waveform data and measured dispersion curves of the target site at different frequencies; The curve extraction module 702 is used to extract the dispersion curve of the micro-motion waveform data by means of the high-resolution frequency-wavenumber method, obtain the initial Rayleigh wave dispersion curve, and construct an inversion model based on the initial Rayleigh wave dispersion curve; The curve recursion module 703 is used to perform recursive and forward modeling calculations on the initial Rayleigh wave dispersion curve to obtain the recursive theoretical curve. The curve correction module 704 is used to correct the initial Rayleigh wave dispersion curve based on the difference between the recursive theoretical curve and the measured dispersion curve, so as to obtain the target Rayleigh wave dispersion curve. The iterative screening module 705 is used to iteratively screen the target Rayleigh wave dispersion curve according to the inversion model to obtain the inversion result.
[0102] The micro-dispersion curve inversion device 700 provided in the above embodiments can realize the technical solutions described in the above micro-dispersion curve inversion method embodiments. The specific implementation principles of each module or unit can be found in the corresponding content in the above micro-dispersion curve inversion method embodiments, which will not be repeated here.
[0103] like Figure 8 As shown, the present invention also provides an electronic device 800. The electronic device 800 includes a processor 801, a memory 802, and a display 803. Figure 8 Only some components of the electronic device 800 are shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead.
[0104] In some embodiments, memory 802 may be an internal storage unit of electronic device 800, such as a hard disk or memory of electronic device 800. In other embodiments, memory 802 may also be an external storage device of electronic device 800, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on electronic device 800.
[0105] Furthermore, the memory 802 may include both internal storage units of the electronic device 800 and external storage devices. The memory 802 is used to store application software and various types of data installed on the electronic device 800.
[0106] In some embodiments, processor 801 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in memory 802 or process data, such as the micro-dispersion curve inversion method of the present invention.
[0107] In some embodiments, display 803 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 803 is used to display information from electronic device 800 and to display a visual user interface. Components 801-803 of electronic device 800 communicate with each other via a system bus.
[0108] In some embodiments of the present invention, when the processor 801 executes the micro-dispersion curve inversion program in the memory 802, the following steps can be implemented: Acquire micro-motion waveform data and measured dispersion curves of the target site at different frequencies; The dispersion curve of the micro-motion waveform data was extracted by high-resolution frequency-wavenumber method to obtain the initial Rayleigh wave dispersion curve, and an inversion model was constructed based on the initial Rayleigh wave dispersion curve. Recursive and forward modeling calculations were performed on the initial Rayleigh wave dispersion curve to obtain the recursive theoretical curve. The initial Rayleigh wave dispersion curve is corrected based on the difference between the recursive theoretical curve and the measured dispersion curve to obtain the target Rayleigh wave dispersion curve. The target Rayleigh wave dispersion curve is iteratively screened based on the inversion model to obtain the inversion result.
[0109] It should be understood that when the processor 801 executes the micro-dispersion curve inversion program in the memory 802, in addition to the functions mentioned above, it can also perform other functions, as can be found in the description of the corresponding method embodiments above.
[0110] Furthermore, this embodiment of the invention does not specifically limit the type of electronic device 800 mentioned. Electronic device 800 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the invention, electronic device 800 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).
[0111] Accordingly, this application also provides a computer-readable storage medium for storing computer-readable programs or instructions. When the programs or instructions are executed by a processor, they can implement the steps or functions of the micro-dispersion curve inversion method provided in the above-described method embodiments.
[0112] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.
[0113] The micro-dispersion curve inversion method, apparatus, electronic device, and storage medium provided by the present invention have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.
Claims
1. A method for inverting micro-motion dispersion curves, characterized in that, include: Acquire micro-motion waveform data and measured dispersion curves of the target site at different frequencies; The dispersion curve of the micro-motion waveform data was extracted by the high-resolution frequency-wavenumber method to obtain the initial Rayleigh wave dispersion curve, and an inversion model was constructed based on the initial Rayleigh wave dispersion curve. The initial Rayleigh wave dispersion curve is recursively calculated and forward modeled to obtain the recursive theoretical curve; The initial Rayleigh wave dispersion curve is corrected based on the difference between the recursive theoretical curve and the measured dispersion curve to obtain the target Rayleigh wave dispersion curve. The target Rayleigh wave dispersion curve is iteratively filtered according to the inversion model to obtain the inversion result.
2. The method for inverting micro-motion dispersion curves according to claim 1, characterized in that, The inversion model includes physical constraints; these physical constraints include Poisson's ratio correlation constraints, layer depth increment constraints, and low-speed layer exclusion constraints.
3. The method for inverting micro-motion dispersion curves according to claim 2, characterized in that, The recursive and forward modeling calculations performed on the initial Rayleigh wave dispersion curve to obtain the recursive theoretical curve include: Set the maximum depth and thin layer thickness during the forward modeling process; The target site is divided into thin layers of equal thickness based on the maximum depth and the thickness of the thin layer; The waveform parameters are determined based on the Rayleigh wave dispersion curve; the waveform parameters include frequency range, frequency division type, and number of frequency points. Based on the waveform parameters, the depth corresponding to the high-frequency non-dispersion segment of the initial Rayleigh wave dispersion curve is calculated using the half-wavelength method, resulting in a phase velocity-depth curve. The number of frequency points within each thin layer is recursively calculated based on the direct recursive method and the phase velocity-depth curve to obtain the shear wave velocity of each thin layer. Based on all shear wave velocities, a preliminary shear wave velocity curve is obtained; The preliminary shear wave velocity curve is subjected to forward modeling to obtain the recursive theoretical curve.
4. The method for inverting micro-motion dispersion curves according to claim 3, characterized in that, The method of recursively calculating the number of frequency points within each thin layer range based on the direct recursive method and the phase velocity-depth curve to obtain the shear wave velocity of each thin layer includes: Based on the phase velocity-depth curve, the characteristic equations of Rayleigh wave phase velocity, longitudinal wave and transverse wave in a uniform half-space, and the relationship between longitudinal wave, transverse wave and Poisson's ratio, the conversion coefficient is obtained; Each thin layer is arranged in ascending order according to its depth to obtain the arrangement sequence; The shear wave velocity of the first thin layer is obtained by recursively calculating the average wave velocity of the number of frequency points, the conversion coefficient, the current depth, and the Rayleigh wave wavelength within the range of the first thin layer in the arrangement order. Based on the thickness and shear wave velocity of the previous thin layer, the average wave velocity of the number of frequency points within the range of the middle thin layer in the arrangement sequence, the conversion coefficient, the current depth, and the Rayleigh wave wavelength are recursively calculated to obtain the shear wave velocity of the middle thin layer.
5. The method for inverting micro-motion dispersion curves according to claim 2, characterized in that, The step of correcting the initial Rayleigh wave dispersion curve based on the difference between the recursive theoretical curve and the measured dispersion curve to obtain the target Rayleigh wave dispersion curve includes: The difference between the recursive theoretical curve and the measured dispersion curve across the entire frequency band is calculated to obtain the deviation ratio for each frequency band. The average deviation value is obtained based on all deviation ratios and the number of frequency points. The correction coefficient is obtained based on the average deviation value, and the initial Rayleigh wave dispersion curve is corrected based on the correction coefficient to obtain the corrected shear wave velocity curve. When the number of iterations is less than the preset number of iterations, the deviation correction is performed again after forward modeling the modified shear wave velocity curve. When the number of iterations is not less than the preset number of iterations, the modified shear wave velocity curve is determined to be the target Rayleigh wave dispersion curve.
6. The method for inverting micro-motion dispersion curves according to claim 1, characterized in that, The step of iteratively filtering the target Rayleigh wave dispersion curve according to the inversion model to obtain the inversion result includes: After inputting the target Rayleigh wave dispersion curve into the inversion model, an initial sample set is generated through random walk; Forward modeling is performed on each initial sample in the initial sample set to obtain the corresponding theoretical dispersion curve; The theoretical dispersion curve is compared with the preset observation data to obtain the mismatch degree of each initial sample; The initial sample set is filtered according to the mismatch degree to obtain the initial Voronoi unit set; The initial samples in the initial Voronoi unit set are iteratively filtered to obtain the inversion result.
7. The method for inverting micro-motion dispersion curves according to claim 6, characterized in that, The iterative screening of the initial samples in the initial Voronoi unit set to obtain the inversion result includes: The initial samples in the initial Voronoi unit set are sorted according to the mismatch degree to obtain a number of first samples whose mismatch degree is lower than a preset mismatch degree; The active region is formed by performing a union operation on the multiple first samples; New samples are generated in the active region based on the Gibbs sampler, and the new mismatch of the new samples is calculated. The valid samples are added to the initial samples based on the new mismatch degree to obtain a new sample set; The minimum parameter space range surrounding the active region is determined based on the new sample set, and the minimum parameter space range is dynamically scaled according to the range of each parameter axis to obtain the target parameter space. The parameter space dynamic scaling is based on the sensitivity of the inversion model to different parameters in the parameter space, and dynamically scales the parameters according to the distribution range of each parameter axis. The search step size is reduced for highly sensitive parameters and increased for low-sensitivity parameters. Based on the new sample set and the target parameter space, a new set of Voronoi units is generated; After removing Voronoi units in the new Voronoi unit set whose number of valid samples is less than the preset number of samples, it is determined whether the number of iterations has been reached or the minimum mismatch is less than the preset threshold. If not, then iteratively filter the set of Voronoi units after removal; If so, the inversion result is obtained based on the set of Voronoi units after removal.
8. A micro-motion dispersion curve inversion device, characterized in that, include: The data acquisition module is used to acquire micro-motion waveform data and measured dispersion curves of the target site at different frequencies; The curve extraction module is used to extract the dispersion curve of the micro-motion waveform data using the high-resolution frequency-wavenumber method, obtain the initial Rayleigh wave dispersion curve, and construct an inversion model based on the initial Rayleigh wave dispersion curve. The curve recursion module is used to perform recursive and forward modeling calculations on the initial Rayleigh wave dispersion curve to obtain the recursive theoretical curve. The curve correction module is used to correct the initial Rayleigh wave dispersion curve based on the difference between the recursive theoretical curve and the measured dispersion curve, so as to obtain the target Rayleigh wave dispersion curve. The iterative screening module is used to iteratively screen the target Rayleigh wave dispersion curve according to the inversion model to obtain the inversion result.
9. An electronic device, characterized in that, include: A processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program, when executed by the processor, implements the steps of the micro-dispersion curve inversion method as described in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of the micro-dispersion curve inversion method as described in any one of claims 1-7.