A joint inversion method and system for tunnel multi-source data based on structural consistency

Abstract

The present invention provides a tunnel multi-source data joint inversion method and system based on structural consistency, which obtains observation data of advance detection by seismic wave method, resistivity method and geological radar method, and constructs initial models of wave velocity, resistivity and relative dielectric constant respectively; calculates forward data of each initial model respectively, calculates the error between the forward data and the corresponding observation data, and obtains corresponding gradients; based on the gradient of dielectric constant, the gradients of wave velocity and resistivity are corrected; based on the corrected gradient, a cross-gradient constraint term is set, and a cross-gradient joint inversion is performed to constrain the common space; according to the result of the joint inversion, cluster categories and cluster centers are set, and a cluster-constrained joint inversion is performed; the seismic detection area is divided into different sections, and each initial model is alternately joint inverted to achieve multi-range joint inversion. The present invention can optimize the full waveform inversion of the tunnel and improve the effect of the joint inversion.

Description

Technical Field

[0001] The present invention belongs to the field of geophysical exploration technology, and specifically relates to a tunnel multi-source data joint inversion method and system based on structural consistency, and in particular to a tunnel multi-source data cross-gradient joint inversion method and system based on structural consistency enhancement. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] To effectively detect adverse geological structures ahead of tunnel faces, safeguard tunnel construction, and reveal the structure and physical property distribution of the subsurface, various geophysical methods, including induced polarization (IP), ground-penetrating radar (GPR), and seismic wave methods, are being increasingly used. Existing methods often rely on inversion imaging of single geophysical data to determine the underlying geological conditions. These methods offer a narrow range of characterization and are difficult to accurately detect intricate and unfavorable geological structures. For example, tunnel seismic wave detection, which leverages the differences in elastic parameter distribution within the subsurface, offers advantages such as long detection range and excellent interface characterization. However, it is sensitive to fault fracture zones but struggles to characterize water content. Tunnel resistivity detection, which leverages the differences in electrical parameter distribution within the subsurface, is sensitive to low-resistance water-bearing subsurface structures, but suffers from the electric field volume effect, making it difficult to accurately characterize interfaces. Ground-penetrating radar, which leverages the differences in electromagnetic properties of different materials within the subsurface, offers the advantages of rapidity, non-destructiveness, and high resolution, enabling accurate characterization of the location and size of subsurface anomalies. However, in practice, obtaining parameter information and spatial morphology from profiles alone is difficult.

[0004] Joint inversion, a key development trend in geophysical inversion, integrates multiple exploration methods and geophysical observation data to effectively suppress multiple solutions and obtain more accurate information about the subsurface medium. Joint inversion based on structural similarity, including cross-gradients, model gradient point products, and local correlations, is a key development in geophysical inversion. Cross-gradients, as an effective joint inversion method, are independent of prior physical property information, have strong applicability, and are a current research hotspot. Furthermore, leveraging prior information to provide numerically and structurally sound initial models for velocity and resistivity models can effectively improve the effectiveness of joint inversion.

[0005] Currently, there are three main challenges in implementing a comprehensive detection method based on multiple geophysical fields:

[0006] The inversion gradient distortion problem of seismic wave method and resistivity method in tunnel environment. The observation system of seismic wave method in tunnel is far away from the ground surface, the offset distance is small, the interface morphology of the inversion result is not accurately portrayed, and the inversion gradient has arc-shaped artifacts. The resistivity method has an inherent volume effect, which makes the inversion gradient have a gradual shape and cannot accurately depict the boundary of the anomaly. At the same time, the resolution and modeling grid of ground penetrating radar method and seismic-electric forward inversion are inconsistent, making the conventional cross-gradient structure constraint difficult to apply.

[0007] Under complex geological conditions, the cross-gradient joint inversion method, relying solely on structural coupling, struggles to accurately invert hidden hazard sources. Gradient differences between different physical fields may correspond to multiple possible subsurface models, leading to non-uniqueness and relatively weak structural coupling constraints. Furthermore, the weight distribution of different methods in the calculation of cross-gradient terms is uncontrollable, making it difficult to balance the weight relationships between different data sources.

[0008] Initial model construction and multi-range detection issues. Tunnel observation data lacks effective information and fails to fully utilize geological prior information, resulting in poor initial model quality and exacerbating the inversion process's tendency to fall into local minima. Furthermore, due to the different detection distances of various methods, there is currently a lack of effective ways to fully integrate the detection methods, making it difficult to achieve effective improvement, especially for long-range targets. Summary of the Invention

[0009] In order to solve the above problems, the present invention proposes a tunnel multi-source data joint inversion method and system based on structural consistency. The present invention can optimize the tunnel full waveform inversion and improve the effect of joint inversion.

[0010] According to some embodiments, the present invention adopts the following technical solutions:

[0011] A tunnel multi-source data joint inversion method based on structural consistency includes the following steps:

[0012] Obtain observation data from advance detection using seismic wave, resistivity, and geological radar methods, and construct initial models for wave velocity, resistivity, and relative permittivity, respectively;

[0013] Calculating the forward modeling data of the initial model for wave velocity, resistivity, and relative dielectric constant, respectively, calculating the errors between the forward modeling data and the corresponding observation data, and obtaining the gradients of wave velocity, resistivity, and dielectric constant;

[0014] Based on the gradient of the dielectric constant, the gradient of the wave velocity and resistivity is corrected;

[0015] Based on the corrected gradient, a cross gradient constraint term is set, and a cross gradient joint inversion is performed to constrain the common space; based on the result of the joint inversion, cluster categories and cluster centers are set, and a cluster constraint joint inversion is performed to optimize boundary information;

[0016] The seismic detection area is divided into different segments, and the initial models of wave velocity, resistivity and relative dielectric constant are subjected to alternating joint inversion with cross-gradient and clustering constraints, further realizing multi-range joint inversion.

[0017] As an optional implementation, the process of constructing the initial models of wave velocity, resistivity, and relative dielectric constant includes: generating corresponding three-dimensional resistivity initial models, three-dimensional wave velocity initial models, and two-dimensional dielectric constant initial models based on different detection distances;

[0018] The initial models of wave velocity, resistivity and relative dielectric constant have the same cross-sectional scale, and in the longitudinal scale, the length of the initial model of resistivity is used as the common joint inversion area.

[0019] As an optional embodiment, the process of correcting the gradient of wave velocity and resistivity based on the gradient of dielectric constant includes: introducing a structure constraint filter into the gradient of dielectric constant to extract parameters of texture and structural features;

[0020] Based on the extracted parameters and combined with the interval reflected by the gradient of the dielectric constant, the wave velocity and resistivity gradient are jointly corrected.

[0021] As a further implementation method, a Gabor filter is used to extract parameters of texture and structural features. The function of the Gabor filter is:

[0022]

[0023] Where x′=xcos(θ)+ysin(θ), y′=-xsin(θ)+ycos(θ), and the parameter λ controls the wavelength. Control direction, Ψ controls phase offset, σ controls standard deviation, and γ controls aspect ratio;

[0024] Use different angles The Gabor filters with different wavelengths λ are used to extract the features of the gradients of wave velocity, resistivity and dielectric constant.

[0025] As an optional implementation, based on the extracted parameters and in combination with the interval reflected by the dielectric constant gradient, the process of jointly correcting the wave velocity and resistivity gradient includes:

[0026] By calculating the correlation coefficients of wave velocity, resistivity gradient and dielectric constant gradient under different parameters, the parameter with the highest linear correlation is selected to process the corresponding gradient, so that the processed gradients have a higher correlation;

[0027] The structural and positional information provided by the dielectric constant gradient is used to adjust the wave velocity and resistivity gradients and modify them to a form that is more consistent with the actual geological structure.

[0028] As an optional implementation, a cross-gradient constraint term is set, and a process of performing a cross-gradient joint inversion includes: performing an inversion update of a seismic wave method and a resistivity method, calculating an inversion result based on the corrected gradient, and calculating a cross-gradient term function;

[0029] Set adaptive weight coefficients for structural constraint items to balance data items and structural constraint items to obtain the final structural constraint items;

[0030] The final structural constraint term is alternately added to the loss functions of the seismic wave method and the resistivity method to obtain joint inversion results based on the structural constraints.

[0031] As a further embodiment, the cross gradient term function includes three-dimensional and two-dimensional functions, and the three-dimensional function is:

[0032]

[0033] The two-dimensional function is:

[0034]

[0035]

[0036] Among them, m v Indicates wave speed, m ρ Resistivity, m ε represents the relative dielectric constant, Δx, Δy, and Δz are the side lengths of the current grid in the x, y, and z directions, respectively. i, j, and k are the grid numbers in the x, y, and z directions, respectively.

[0037] Furthermore, the process of setting adaptive weight coefficients for structural constraints includes:

[0038] γ k =n×k 2

[0039] φ J =γ k θ i

[0040] Where k is the number of iterations, γ k is the kth adaptive weight factor, θ iis the cross gradient term, φ J is the total structural constraint term, and n is a constant.

[0041] As an optional implementation, the process of performing alternating cross-gradient and clustering-constrained joint inversion on the updated medium model of wave velocity, resistivity, and relative dielectric constant includes calculating the cross-gradient term function or the clustering function term and adding it to the inversion objective function of radar, seismic, and resistivity, respectively, updating the parameters of the three types of medium models, and then repeating the forward modeling, gradient optimization, and the above-mentioned joint inversion process of the updated model until the iterative convergence condition is met.

[0042] As a further example, the objective function is as follows:

[0043]

[0044] in, is the objective function of wave velocity inversion, is the objective function of resistivity inversion, is the objective function of relative permittivity inversion, is the data item function, is the target function of the model term, λ is the regularization parameter of the model term, γ t is the adaptive weight of the joint inversion term, θ t is the joint inversion objective function term, θ cross Cross gradient function, is the clustering function.

[0045] As an optional embodiment, the seismic exploration area is divided into different segments, and the initial models of wave velocity, resistivity, and relative permittivity are subjected to alternating joint inversion with cross-gradient and cluster constraints, including:

[0046] A first detection range is selected and set, and the initial models of wave velocity, resistivity, and relative dielectric constant are subjected to alternating joint inversion using cross-gradient and cluster constraints until convergence conditions are met. After a more accurate model of the first set detection range is obtained, the initial value of the wave velocity model of the second set detection range is updated based on the model. The length of the second set detection range is greater than the first set detection range, and the wave velocity model inversion of the second set detection range is performed;

[0047] Move the length of the first detection range forward along the detection direction, establish an initial resistivity and dielectric constant model based on the seismic wave velocity inversion results and geological data of the first detection range after the movement, and continue the joint inversion calculation;

[0048] Repeat the above process to optimize the effect of multi-range joint inversion.

[0049] A tunnel multi-source data joint inversion system based on structural consistency, comprising:

[0050] The model building module is used to obtain observation data from advance detection using seismic wave, resistivity and geological radar methods, and to build initial models of wave velocity, resistivity and relative dielectric constant respectively;

[0051] A gradient calculation module is used to calculate the forward modeling data of the initial model of wave velocity, resistivity and relative dielectric constant, calculate the error between the forward modeling data and the corresponding observation data, and obtain the gradients of wave velocity, resistivity and dielectric constant;

[0052] Gradient correction module, used to correct the gradient of wave velocity and resistivity based on the gradient of dielectric constant;

[0053] A joint inversion module is used to set cross-gradient constraint items based on the corrected gradients, perform cross-gradient joint inversion to constrain the common space, set cluster categories and cluster centers based on the results of the joint inversion, and perform cluster-constrained joint inversion to optimize boundary information;

[0054] The long-short probe range cyclic optimization module is used to divide the seismic exploration area into different segments, and perform alternating joint inversion of the initial models of wave velocity, resistivity and relative dielectric constant with cross-gradient and cluster constraints to achieve multi-probe range joint inversion.

[0055] A computer-readable storage medium is used to store computer instructions, and when the computer instructions are executed by a processor, the steps in the above method are completed.

[0056] An electronic device includes a memory and a processor, and computer instructions stored in the memory and executed on the processor. When the computer instructions are executed by the processor, the steps in the above method are completed.

[0057] Compared with the prior art, the present invention has the following beneficial effects:

[0058] (1) Aiming at the problems of arc-shaped distortion of velocity inversion gradient and volume effect of resistivity inversion gradient in tunnels, a structure extraction filter based on Gabor transform was used to establish the correlation between radar gradient and seismic and resistivity gradients, guided by more accurate radar gradient and using correlation coefficient as evaluation index, thus achieving structural consistency of multi-parameter inversion gradient.

[0059] (2) The present invention solves the problem of inconsistency between the inversion grid of the ground penetrating radar method and the inversion grid of the seismic wave method and the resistivity method by adopting a joint inversion method, so that the inversion results maintain high resolution in areas with large structural differences and reduce the influence of noise; at the same time, the inversion gradient results of the seismic, electrical and radar methods are corrected during the inversion iteration process, laying a foundation for the improvement of the joint inversion results.

[0060] (3) In order to solve the problem of different detection distances among different methods, the present invention uses the short-range method in the public inversion area to provide a reliable initial model for the subsequent long-range method based on the cross-clustering joint inversion with strong structural consistency of multiple parameters, and realizes the alternating inversion of the non-public inversion area through sliding joint, thereby gradually improving the inversion effect of the seismic long-range method.

[0061] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0063] Figure 1 This is a technical roadmap for a tunnel multi-source data cross-gradient joint inversion method based on structural consistency enhancement in an embodiment of the present invention;

[0064] Figure 2 Schematic diagram of the cross-gradient joint inversion effect of tunnel multi-source data in an embodiment of the present invention, where (a) is the real model; (b) is the velocity inversion result; (c) is the resistivity inversion result; (d) is the dielectric constant inversion result;

[0065] Figure 3 Schematic diagram of an initial model for an embodiment of the present invention in which the seismic inversion region is electric-radar inversion, wherein (a) is a joint inversion of a common region 1, and (b) is a joint inversion of a common region 2;

[0066] Figure 4 1 is a diagram showing the improved effect of long-range seismic inversion in an embodiment of the present invention, wherein (a) is the real model; (b) is the initial model; and (c) is the inversion result. DETAILED DESCRIPTION

[0067] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0068] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention belongs.

[0069] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0070] In the absence of conflict, the embodiments and features in the embodiments of this application can be combined with each other.

[0071] Example 1

[0072] A tunnel multi-source data cross gradient joint inversion method based on structural consistency enhancement, such as Figure 1 As shown, the specific steps include:

[0073] Step 1: Collect tunnel multi-source detection data and build an initial model

[0074] In this embodiment, advance detection work is carried out at the tunnel site using seismic wave method, resistivity method, and geological radar method to collect effective observation data, and then the data is reasonably and accurately processed. In combination with existing geological prior information and manual experience, an initial model of wave velocity, resistivity, and relative dielectric constant is constructed from both structural and numerical aspects.

[0075] In some embodiments, geological survey data, advance drilling and advance pilot hole information may be used to provide a more reliable initial model for seismic and electrical methods in terms of structure and value to start joint inversion calculation.

[0076] In some embodiments, an initial model of wave velocity, resistivity, and relative dielectric constant may be constructed based on the detection results of advance detection using seismic wave method, resistivity method, and geological radar method, combined with the correlation between the parameters.

[0077] In this embodiment, based on the different detection distances, corresponding three-dimensional resistivity initial models, three-dimensional wave velocity initial models, and two-dimensional dielectric constant initial models are generated.

[0078] The cross-sectional scales of the resistivity model, wave velocity model, and dielectric constant model are the same, and the length consistent with the resistivity model is selected as the common joint inversion area on the vertical scale.

[0079] The method of constructing the model can adopt existing technology and will not be described in detail here.

[0080] Step 2: Calculate the data errors of the three methods through the error function, and then calculate the three-dimensional inversion gradient of the seismic wave velocity method and the resistivity method, and the two-dimensional inversion gradient of the geological radar method. Then, perform the gradient correction calculation of the seismic wave method and the resistivity method based on the geological radar gradient.

[0081] Specifically, the forward data of the initial wave velocity model, initial resistivity model, and initial dielectric constant model are calculated, and then the second norm is used as the objective function to calculate the data error between the simulated data and the observed data of the three methods, and the gradients corresponding to the three parameters are calculated, namely the gradients of wave velocity, resistivity, and dielectric constant. Of course, they can also be called seismic, electrical, and radar gradients respectively.

[0082] The process of calculating the gradients corresponding to the three parameters can be done using existing technology and will not be described in detail here.

[0083] For the dielectric constant gradient, the geological radar gradient is calculated by adjusting the parameters of the Gabor filter (wavelength, direction parameters). The function of the Gabor filter is composed of controlling the wavelength, controlling the direction, controlling the phase offset, controlling the standard deviation, and controlling the aspect ratio.

[0084] In this embodiment, the Gabor filter expression is:

[0085]

[0086] Where x′=xcos(θ)+ysin(θ), y′=-xsin(θ)+ycos(θ), and the parameter λ controls the wavelength. controls the direction, Ψ controls the phase offset, σ controls the standard deviation, and γ controls the aspect ratio.

[0087] Use different angles The features of the velocity, resistivity, and dielectric constant gradients were extracted using Gabor filters with different wavelengths λ. The correlation coefficients between the seismic-resistivity gradient and the radar gradient were calculated under different parameters. The parameters with the highest linear correlation were selected to process the velocity and resistivity gradients. This resulted in a higher correlation between the processed velocity and resistivity gradients, which was a preliminary correction to the gradients of the two methods.

[0088] At the same time, the structural and positional information extracted from the geological radar is used to adjust the gradients in the seismic and electrical methods. Based on the initially corrected velocity and resistivity gradients, the gradients are weighted by the range of the anomalies provided by the geological radar, thereby correcting the gradients to a form that is more consistent with the actual geological structure. Ultimately, the seismic and electrical inversion results are guided closer to the actual geological structure, thereby obtaining more accurate and reliable underground structural information.

[0089] Step 3: Calculate the cross-gradient constraint terms and perform joint inversion alternately.

[0090] Specifically, based on the corrected gradient, the joint inversion of seismic, electrical, and radar methods is performed alternately to constrain the common space. Based on the results of the cross-gradient joint inversion, cluster categories and cluster centers are set, and cluster alternating joint inversion is performed to optimize the boundary information. The above process is repeated until the convergence condition is met.

[0091] The specific process is as follows:

[0092] The three-dimensional formula of cross gradient is:

[0093]

[0094] The cross gradient two-dimensional formula is:

[0095]

[0096] Among them, m v Indicates wave speed, m ρ Resistivity, m ε Represents the relative dielectric constant.

[0097] Finally, the obtained cross-gradient terms are alternately added to the loss functions of the seismic and electrical methods to obtain the joint inversion results based on structural constraints.

[0098] Step 4: Based on the results of the cross gradient after a certain number of iterations, clustering constraints are further performed.

[0099] Calculate the cross gradient term function or clustering function term and add it to the radar inversion objective function. The clustering function term is:

[0100]

[0101] Among them, a j is the category center, x j For each sample, N is the number of samples of the corresponding category, and c is the category.

[0102] Step 5: Perform joint inversion calculations based on cross-gradient and clustering constraints to improve the joint inversion effect of seismic, electrical and radar data.

[0103] First, a joint inversion calculation is performed for a certain number of iterations based on the cross-gradient term. After obtaining preliminary results, a joint inversion calculation is performed using clustering constraints. Clustering joint inversion is performed after calculating the cluster centers and membership degrees. Finally, whether the convergence conditions are met is verified and the joint inversion results are output.

[0104] The cross-gradient function is calculated after the gradient correction of seismic, electrical and radar inversion, and then the joint inversion is performed using the clustering function term after a certain number of iterations until the iterative convergence condition is met. The objective function is as follows:

[0105]

[0106] in, is the objective function of wave velocity inversion, is the objective function of resistivity inversion, is the objective function of relative permittivity inversion, is the data item function, is the target function of the model term, λ is the regularization parameter of the model term, γ t is the adaptive weight of the joint inversion term.

[0107] Step 6: Alternating long- and short-range joint inversion based on initial model optimization, i.e. sliding the common area for the next joint inversion, and optimizing the long-range detection range with the short-range joint inversion results.

[0108] Specifically, combined Figure 3 The inversion range for the seismic wave method is the green area in the figure, i.e., 100 m behind the tunnel face. The inversion range for the resistivity method and ground penetrating radar method is the blue area in the figure, i.e., 30 m behind the tunnel face. First, the detection range 30 m before the earthquake is selected as common area 1. The velocity model, resistivity model, and relative permittivity model are subjected to the above-mentioned cross-gradient and clustering alternating joint inversion until the convergence condition is met. This results in a multi-parameter joint inversion result for the first 30 m. At the same time, the more accurate velocity model provides a more accurate initial model for the inversion 70 m after the earthquake and a more reliable initial model structure for the resistivity and relative permittivity of the next common area. Moving 30 m along the detection direction to common area 2, this area is jointly inverted with the electrical method and radar detection data at the same location. This results in a joint inversion result for the second 30 m common area, and a more accurate initial model for the 40 m after the earthquake is obtained. The above process is repeated for the joint inversion of common area 3. In this process, the inversion results for the long-range earthquake are continuously improved, ultimately achieving the optimization of the multi-range joint inversion effect.

[0109] Example 2

[0110] A tunnel multi-source data joint inversion system based on structural consistency, comprising:

[0111] The model building module is used to obtain observation data from advance detection using seismic wave, resistivity and geological radar methods, and to build initial models of wave velocity, resistivity and relative dielectric constant respectively;

[0112] A gradient calculation module is used to calculate the forward modeling data of the initial model of wave velocity, resistivity and relative dielectric constant, calculate the error between the forward modeling data and the corresponding observation data, and obtain the gradients of wave velocity, resistivity and dielectric constant;

[0113] Gradient correction module, used to correct the gradient of wave velocity and resistivity based on the gradient of dielectric constant;

[0114] A joint inversion module is used to set cross-gradient constraint items based on the corrected gradients, perform cross-gradient joint inversion to constrain the common space, set cluster categories and cluster centers based on the results of the joint inversion, and perform cluster-constrained joint inversion to optimize boundary information;

[0115] The long-short probe range cyclic optimization module is used to divide the seismic exploration area into different segments, and perform alternating joint inversion of the initial models of wave velocity, resistivity and relative dielectric constant with cross-gradient and cluster constraints to achieve multi-probe range joint inversion.

[0116] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0117] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0118] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0119] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0120] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made by those skilled in the art that fall within the spirit and principles of the present invention and do not require creative effort are intended to be within the scope of protection of the present invention.

Claims

1. A joint inversion method for tunnel multi-source data based on structural consistency, characterized by: The following steps are involved: Obtain observation data from advance detection using seismic wave, resistivity, and geological radar methods, and construct initial models for wave velocity, resistivity, and relative permittivity, respectively; Calculating the forward modeling data of the initial model for wave velocity, resistivity, and relative dielectric constant, respectively, calculating the errors between the forward modeling data and the corresponding observation data, and obtaining the gradients of wave velocity, resistivity, and dielectric constant; Based on the gradient of the dielectric constant, the gradient of the wave velocity and resistivity is corrected; Based on the corrected gradient, a cross gradient constraint term is set, and a cross gradient joint inversion is performed to constrain the common space; based on the result of the joint inversion, cluster categories and cluster centers are set, and a cluster constraint joint inversion is performed to optimize boundary information; The seismic detection area is divided into different segments, and the initial models of wave velocity, resistivity and relative dielectric constant are subjected to alternating joint inversion with cross-gradient and clustering constraints, further realizing multi-range joint inversion.

2. The method for joint inversion of tunnel multi-source data based on structural consistency according to claim 1, characterized in that: The process of constructing the initial models of wave velocity, resistivity and relative dielectric constant respectively includes: generating corresponding three-dimensional resistivity initial model, three-dimensional wave velocity initial model and two-dimensional dielectric constant initial model based on different detection distances; The initial models of wave velocity, resistivity and relative dielectric constant have the same cross-sectional scale, and in the longitudinal scale, the length of the initial model of resistivity is used as the common joint inversion area.

3. The method for joint inversion of tunnel multi-source data based on structural consistency according to claim 1, characterized in that: The process of correcting the gradient of wave velocity and resistivity based on the gradient of dielectric constant includes: introducing a structure constraint filter into the gradient of dielectric constant to extract parameters of texture and structural features; Based on the extracted parameters and combined with the interval reflected by the gradient of the dielectric constant, the wave velocity and resistivity gradient are jointly corrected.

4. The method for joint inversion of tunnel multi-source data based on structural consistency according to claim 3, characterized in that: The parameters of texture and structural features are extracted using Gabor filter. The function of Gabor filter is: ； in, , parameter λ controls the wavelength, Control direction, Ψ controls phase offset, σ controls standard deviation, and γ controls aspect ratio; Use different angles The Gabor filters with different wavelengths λ are used to extract the features of the gradients of wave velocity, resistivity and dielectric constant.

5. The method for joint inversion of tunnel multi-source data based on structural consistency according to claim 1, characterized in that: Based on the extracted parameters and the interval reflected by the dielectric constant gradient, the process of jointly correcting the wave velocity and resistivity gradient includes: By calculating the correlation coefficients of wave velocity, resistivity gradient and dielectric constant gradient under different parameters, the parameter with the highest linear correlation is selected to process the corresponding gradient, so that the processed gradients have a higher correlation; The structural and positional information provided by the dielectric constant gradient is used to adjust the wave velocity and resistivity gradients and modify them to a form that is more consistent with the actual geological structure.

6. The method for joint inversion of tunnel multi-source data based on structural consistency according to claim 1, characterized in that: The process of setting cross-gradient constraint terms and performing cross-gradient joint inversion includes: performing inversion updates of the seismic wave method and the resistivity method, calculating the inversion results based on the corrected gradients, and calculating the cross-gradient term function; Set adaptive weight coefficients for structural constraint items to balance data items and structural constraint items to obtain the final structural constraint items; The final structural constraint term is alternately added to the loss functions of the seismic wave method and the resistivity method to obtain joint inversion results based on the structural constraints.

7. The method for joint inversion of tunnel multi-source data based on structural consistency according to claim 6, characterized in that: The cross gradient term function includes three-dimensional and two-dimensional functions, and the three-dimensional function is: ； ； The two-dimensional function is: ； ； ； in, represents the wave speed, represents the resistivity, represents the relative dielectric constant, 、 、 are the side lengths of the current grid in the x, y, and z directions respectively, and i, j, and k are the grid numbers in the x, y, and z directions respectively.

8. The method for joint inversion of tunnel multi-source data based on structural consistency according to claim 6, characterized in that: The process of setting adaptive weight coefficients for structural constraints includes: Where k is the number of iterations, is the k-th adaptive weight factor, is the cross gradient term, is the total structural constraint term, and n is a constant.

9. The method for joint inversion of multi-source tunnel data based on structural consistency according to claim 1, characterized in that: The process of alternating cross-gradient and clustering-constrained joint inversion of the initial models of wave velocity, resistivity, and relative dielectric constant includes respectively calculating the cross-gradient term function or the clustering function term and adding it to the radar inversion objective function, then calculating the cross-gradient term function or the clustering function term again with the updated result of seismic wave inversion, performing seismic wave inversion, and then performing resistivity inversion, and repeating the above process until the iterative convergence condition is met.

10. The method for joint inversion of tunnel multi-source data based on structural consistency according to claim 9, characterized in that: The objective function is specifically as follows: ； ； ； ； in, is the objective function of wave velocity inversion, is the objective function of resistivity inversion, is the objective function of relative permittivity inversion, is the data item function, is the model project function, is the model term regularization parameter, is the adaptive weight of the joint inversion term, is the joint inversion objective function term, Cross gradient function, is the clustering function.

11. The method for joint inversion of tunnel multi-source data based on structural consistency according to claim 1, characterized in that: The process of dividing the seismic exploration area into different segments and performing alternating cross-gradient and cluster-constrained joint inversion on the initial models of wave velocity, resistivity, and relative permittivity includes: A first detection range is selected and set, and the initial models of wave velocity, resistivity, and relative dielectric constant are subjected to alternating joint inversion using cross-gradient and clustering constraints until convergence conditions are met. After a more accurate model of the first set detection range is obtained, the initial value of the wave velocity model of the second set detection range is updated based on the model. The length of the second set detection range is greater than the first set detection range, and the wave velocity model inversion of the second set detection range is performed; Move the length of the first detection range forward along the detection direction, establish an initial resistivity and dielectric constant model based on the seismic wave velocity inversion results and geological data of the first detection range after the movement, and continue the joint inversion calculation; Repeat the above process to optimize the effect of multi-range joint inversion.

12. A tunnel multi-source data joint inversion system based on structural consistency, characterized by: include: The model building module is used to obtain observation data from advance detection using seismic wave, resistivity and geological radar methods, and to build initial models of wave velocity, resistivity and relative dielectric constant respectively; A gradient calculation module is used to calculate the forward modeling data of the initial model of wave velocity, resistivity and relative dielectric constant, calculate the error between the forward modeling data and the corresponding observation data, and obtain the gradients of wave velocity, resistivity and dielectric constant; Gradient correction module, used to correct the gradient of wave velocity and resistivity based on the gradient of dielectric constant; A joint inversion module is used to set cross-gradient constraint items based on the corrected gradients, perform cross-gradient joint inversion to constrain the common space, set cluster categories and cluster centers based on the results of the joint inversion, and perform cluster-constrained joint inversion to optimize boundary information; The long-short probe range cyclic optimization module is used to divide the seismic exploration area into different segments, and perform alternating joint inversion of the initial models of wave velocity, resistivity and relative dielectric constant with cross-gradient and cluster constraints to achieve multi-probe range joint inversion.

13. A computer-readable storage medium, characterized in that: Used to store computer instructions, which, when executed by a processor, complete the steps of the method according to any one of claims 1 to 11.

14. An electronic device, characterized in that: The method comprises a memory and a processor, and computer instructions stored in the memory and executed on the processor, wherein the steps of the method according to any one of claims 1 to 11 are completed when the computer instructions are executed by the processor.

Images