A cooperative constraint self-reconfigurable multi-wheel iteration processing method, device and medium

Abstract

The application discloses a kind of collaborative constraint self-reconfigurable multi-wheel iteration processing method, device and medium, comprising: in response to three-dimensional elastic wave travel time tomography processing start signal, the travel time residual information and fat ray finite frequency sensitivity matrix of current velocity model are determined, wherein the travel time residual information includes the difference information between theoretical travel time and observed travel time, and fat ray finite frequency sensitivity matrix is used to describe the finite frequency weighted sensitive relationship between travel time residual and model update amount;Based on travel time residual information and fat ray finite frequency sensitivity matrix, determine the collaborative constraint system, and the model update amount of current velocity model is adjusted to target update amount using collaborative constraint system, wherein collaborative constraint system includes fat ray finite frequency sensitivity matrix constraint, data weight constraint, model weight constraint, spatial smoothing constraint and directionality constraint;In response to collaborative constraint system, generate sparse augmented least squares solution system.

Description

Technical Field

[0001] This invention relates to the field of collaborative constraint self-reconfiguration multi-round iterative processing technology, and more specifically, to a collaborative constraint self-reconfiguration multi-round iterative processing method, apparatus, and medium. Background Technology

[0002] In existing three-dimensional elastic wave travel-time tomography methods, the travel-time sensitivity relationship is usually simplified to a local relationship near the geometric ray path, ignoring the distributed sensitivity formed by the finite-frequency propagation of elastic waves. In terms of model update constraints, uniform damping or overall smoothing methods are often used, which are difficult to adapt to the differences in the strength of data constraints between different model regions under complex geological conditions and limited observation conditions. In multiple iterations, the sensitivity matrix and constraint system are often fixed or insufficiently updated, and cannot be adaptively adjusted with changes in the velocity model, resulting in insufficient imaging stability and reliability of local anomaly recovery. Moreover, as the size of the three-dimensional mesh and the number of rays increase, the computational and storage burden increases significantly.

[0003] There is currently no effective solution to the above problems. Summary of the Invention

[0004] This invention provides a collaborative constraint self-reconstruction multi-round iterative processing method, device, and medium to at least solve the technical problems in existing methods of three-dimensional elastic wave travel time tomography, such as insufficient frequency-sensitive expression, lack of coordination among multiple constraints, inability of constraint system to adaptively reconstruct in multi-round iteration, and low efficiency of large-scale computation, resulting in low accuracy of underground velocity structure imaging, insufficient reliability of local anomaly recovery, and large engineering computation burden.

[0005] According to one aspect of the present invention, in order to achieve the above-mentioned objective, a collaborative constraint self-reconfiguration multi-round iterative processing method is provided, comprising:

[0006] In response to the start signal of three-dimensional elastic wave travel time tomography processing, the travel time residual information and fat ray finite frequency sensitivity matrix of the current velocity model are determined. The travel time residual information includes the difference between theoretical travel time and observed travel time, and the fat ray finite frequency sensitivity matrix is ​​used to describe the finite frequency weighted sensitivity relationship between travel time residual and model update amount.

[0007] Based on travel time residual information and the fat ray finite frequency sensitivity matrix, a cooperative constraint system is determined. The cooperative constraint system is used to adjust the model update amount of the current velocity model to the target update amount. The cooperative constraint system includes fat ray finite frequency sensitivity matrix constraints, data weight constraints, model weight constraints, spatial smoothness constraints, and directional constraints.

[0008] In response to the cooperative constraint system, a sparse augmented least squares solution system is generated. The sparse augmented least squares solution system is used to solve for the model update amount and update the current velocity model to obtain the updated velocity model.

[0009] Furthermore, the travel-time residual information and the fat ray finite-frequency sensitivity matrix of the current velocity model are determined, including:

[0010] Based on the current velocity model and source receiving observation system, perform three-dimensional elastic wave travel time forward modeling to obtain theoretical travel time and ray path;

[0011] Based on theoretical travel time and observed travel time, determine travel time residual information;

[0012] Based on the ray path, elastic wave dominant frequency, and current velocity model, the finite frequency sensitivity range of the fat ray is determined.

[0013] Based on the finite-frequency sensitive influence range of the fat ray and the spatial distance from the model unit to the ray path, the distance attenuation weight is determined;

[0014] Based on the path micro-segment length and distance attenuation weight of the ray path, the finite frequency sensitivity matrix of the fat ray is determined.

[0015] Furthermore, based on the ray path, elastic wave dominant frequency, and current velocity model, the finite-frequency sensitive influence range of the fat ray is determined, including:

[0016] Discretize the ray path into several path segments;

[0017] Determine the local average velocity corresponding to each path segment;

[0018] The local equivalent wavelength is determined based on the dominant frequency of the elastic wave and the local average velocity.

[0019] Based on the local equivalent wavelength and the path distance from the source point to the receiver point along the path segment, the frequency-limited sensitive radius of the fat ray is determined.

[0020] Based on the finite-frequency sensitive radius of fat rays, the finite-frequency sensitive influence range of fat rays is determined.

[0021] Furthermore, based on the finite-frequency sensitive influence range of the fat ray and the spatial distance from the model unit to the ray path, the distance attenuation weight is determined, including:

[0022] Determine the shortest spatial distance from the center point of the model unit to the path micro-segment;

[0023] In response to the shortest spatial distance being greater than the finite frequency sensitive radius, the distance attenuation weight is determined to be zero.

[0024] In response to the shortest spatial distance being less than or equal to the finite frequency sensitive radius, a Gaussian attenuation function is used to determine the distance attenuation weight based on the shortest spatial distance and the lateral attenuation scale parameter.

[0025] Furthermore, based on travel-time residual information and the finite-frequency sensitivity matrix of the fat ray, a cooperative constraint system is determined, including:

[0026] Based on the finite-frequency sensitivity matrix of the fat ray, local data constraint statistics are determined, including data weighted sensitivity and the number of ray influences.

[0027] Based on the local data constraint statistics, the model weight matrix is ​​determined. The model weight matrix is ​​a diagonal matrix, and the diagonal elements are inversely proportional to the local data constraint statistics.

[0028] Determine the spatial smoothing matrix based on the adjacency relationship of the 3D mesh;

[0029] Based on the ray path direction distribution information, the directional constraint matrix is ​​determined, whereby the directional constraint matrix is ​​used to apply differentiated constraint strengths to each direction in space;

[0030] The data weight matrix is ​​determined based on the reliability of the observation travel time;

[0031] Based on the finite-frequency sensitivity matrix of the fat ray, the data weight matrix, the model weight matrix, the spatial smoothing matrix, and the directional constraint matrix, a collaborative constraint system is determined.

[0032] Furthermore, based on local data constraint statistics, the model weight matrix is ​​determined, including:

[0033] The data weighted sensitivity and the number of ray influences are normalized to obtain normalized data weighted sensitivity and normalized number of ray influences;

[0034] Based on the weighted sensitivity of normalized data and the number of normalized ray influences, the local data constraint strength index is determined.

[0035] Based on the local data constraint strength index and the basic model constraint coefficient, the model constraint weight corresponding to each model unit is determined. The model constraint weight is inversely proportional to the local data constraint strength index.

[0036] The model weight matrix is ​​determined based on all model constraint weights.

[0037] Furthermore, based on the ray path direction distribution information, the directional constraint matrix is ​​determined, including:

[0038] The coverage intensity of ray paths near each model unit in various spatial directions is statistically analyzed to obtain the directional coverage.

[0039] Based on the directional coverage, the directional smoothing weights of each model unit in each direction are determined, where the directional smoothing weights are inversely proportional to the directional coverage.

[0040] The directional constraint matrix is ​​determined based on the directional smoothing weights of all model units.

[0041] Furthermore, in response to the cooperative constraint system, after generating the sparse augmented system, it also includes:

[0042] The sparse iterative solution method is used to solve the sparse augmented least squares solution system to obtain the model update amount;

[0043] Determine the temporary model update based on the model update amount;

[0044] Based on the preset lower speed limit and the preset upper speed limit, the velocity boundary projection is performed on the temporary update model to obtain the updated velocity model;

[0045] Based on the updated velocity model, the root mean square of the travel time residual and the relative magnitude of the model update are determined.

[0046] When the root mean square of the travel time residual is less than or equal to the travel time residual threshold, or the relative magnitude of the model update is less than or equal to the model update threshold, or the current iteration round reaches the maximum number of iteration rounds, the iteration stops and the subsurface velocity structure imaging results are output.

[0047] In response to the fact that the root mean square of the travel time residual is greater than the travel time residual threshold, the relative magnitude of the model update is greater than the model update threshold, and the current iteration round is less than the maximum iteration round, the updated velocity model is used as the current velocity model for the next round, and forward modeling, fat ray finite frequency sensitivity matrix construction, and cooperative constraint system reconstruction are re-executed.

[0048] According to one embodiment of the present invention, a collaborative constraint self-reconfigurable multi-round iterative processing apparatus is also provided, comprising:

[0049] The response module is used to respond to the start signal of the three-dimensional elastic wave travel time tomography processing, and to determine the travel time residual information and the fat ray finite frequency sensitivity matrix of the current velocity model. The travel time residual information includes the difference between the theoretical travel time and the observed travel time, and the fat ray finite frequency sensitivity matrix is ​​used to describe the finite frequency weighted sensitivity relationship between the travel time residual and the model update amount.

[0050] The determination module is used to determine the cooperative constraint system based on the travel time residual information and the fat ray finite frequency sensitivity matrix. The cooperative constraint system is used to adjust the model update amount of the current velocity model to the target update amount. The cooperative constraint system includes fat ray finite frequency sensitivity matrix constraints, data weight constraints, model weight constraints, spatial smoothness constraints, and directional constraints.

[0051] The generation module is used to generate a sparse augmented least squares solution system in response to the cooperative constraint system. The sparse augmented least squares solution system is used to solve the model update amount and update the current velocity model to obtain the updated velocity model.

[0052] According to another aspect of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is executed, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of the present invention.

[0053] In this embodiment of the invention, by constructing a fat ray finite-frequency sensitivity matrix, the travel-time sensitivity relationship within a finite spatial range around the ray path can be expressed more fully, improving the reliability of local anomaly boundary and morphology recovery. By establishing a collaborative constraint system including fat ray finite-frequency sensitivity matrix constraints, data weight constraints, model weight constraints, spatial smoothness constraints, and directional constraints, differentiated constraints can be applied to model units according to the strength of local data constraints, alleviating the instability problem caused by uneven constraints under finite observation conditions. By generating a sparse augmented least squares solution system and reconstructing the sensitivity matrix and collaborative constraint system round by round based on the updated velocity model, iterative mismatch caused by a fixed constraint system can be avoided. At the same time, the sparse matrix storage method reduces the storage and computational burden in large-scale three-dimensional inversion, improves the executability of engineering computation, and thus solves the technical problems of low accuracy of underground velocity structure imaging, insufficient reliability of local anomaly recovery, and large engineering computational burden in existing methods of three-dimensional elastic wave travel-time tomography due to insufficient finite-frequency sensitivity expression, lack of collaboration among multiple types of constraints, inability of constraint system to adaptively reconstruct in multiple iterations, and low efficiency of large-scale computational organization. Attached Figure Description

[0054] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0055] Figure 1 This is a flowchart of a collaborative constraint self-reconfiguration multi-round iterative processing method according to one embodiment of the present invention;

[0056] Figure 2 This is a flowchart of a collaborative constraint self-reconfiguration multi-round iterative processing method according to one embodiment of the present invention;

[0057] Figure 3 This is a schematic diagram of the construction of the fat ray finite frequency sensitivity relationship in a collaborative constraint self-reconfiguration multi-round iterative processing method according to one embodiment of the present invention;

[0058] Figure 4 This is a schematic diagram of the collaborative constraint architecture in a collaborative constraint self-reconfigurable multi-round iterative processing method according to one embodiment of the present invention;

[0059] Figure 5 This is a schematic diagram of the self-reconfiguration multi-round iteration and sparse computation organization in a collaborative constraint self-reconfiguration multi-round iteration processing method according to one embodiment of the present invention.

[0060] Figure 6 This is a structural block diagram of a collaborative constraint self-reconfigurable multi-round iterative processing device according to one embodiment of the present invention. Detailed Implementation

[0061] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0062] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0063] According to an embodiment of the present invention, a collaborative constraint self-reconfigurable multi-round iterative processing method is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0064] This method embodiment can be executed in an electronic device or similar computing device that includes memory and a processor. Taking operation on a terminal as an example, the terminal may include one or more processors (processors may include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), digital signal processing (DSP) chips, microcontroller units (MCUs), field-programmable gate arrays (FPGAs), neural network processors (NPUs), tensor processors (TPUs), artificial intelligence (AI) type processors, etc.) and memory for storing data. Optionally, the terminal may also include transmission devices, input / output devices, and display devices for communication functions. Those skilled in the art will understand that the above structural description is merely illustrative and does not limit the structure of the terminal. For example, the terminal may include more or fewer components than described above, or have a different configuration than described above.

[0065] The memory can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the collaborative constraint self-reconfigurable multi-round iterative processing method in this embodiment of the invention. The processor executes various functional applications and data processing by running the computer program stored in the memory, thereby realizing the aforementioned collaborative constraint self-reconfigurable multi-round iterative processing method. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to the mobile terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0066] The transmission device is used to receive or send data via a network. Specific examples of the network mentioned above may include a wireless network provided by the mobile terminal's communication provider. In one example, the transmission device includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0067] Display devices can be, for example, touchscreen liquid crystal displays (LCDs) and touch displays (also referred to as "touchscreens" or "touch displays"). The LCD allows users to interact with the user interface of the mobile terminal. In some embodiments, the mobile terminal has a graphical user interface (GUI), which allows users to interact with the GUI through finger contact and / or gestures on a touch-sensitive surface. Optional human-computer interaction functions include: creating web pages, drawing, word processing, creating electronic documents, playing games, video conferencing, instant messaging, sending and receiving emails, call interfaces, playing digital video, playing digital music, and / or web browsing, etc. Executable instructions for performing the above human-computer interaction functions are configured / stored in one or more processor-executable computer program products or readable storage media.

[0068] Figure 1 This is a flowchart of a collaborative constraint self-reconfiguration multi-round iterative processing method according to one embodiment of the present invention, as shown below. Figure 1 As shown, the method includes the following steps:

[0069] Step S110: In response to the three-dimensional elastic wave travel-time tomography processing start signal, determine the travel-time residual information and the fat-ray finite-frequency sensitivity matrix of the current velocity model. The travel-time residual information includes the difference between the theoretical travel time and the observed travel time. The fat-ray finite-frequency sensitivity matrix is ​​used to describe the finite-frequency weighted sensitivity relationship between the travel-time residual and the model update amount. The specific content is as follows:

[0070] In step S110, the travel time residual information of the current velocity model is determined, specifically by performing a three-dimensional elastic wave travel time forward modeling calculation based on the current velocity model and a pre-established source-receiver observation system. The forward modeling calculation calculates the first-arrival travel time and three-dimensional ray paths of the elastic wave based on the current three-dimensional velocity model. For each source-receiver path, its theoretical travel time is obtained by performing a line integral along the ray path on the current slowness model; the theoretical travel times corresponding to all rays are summarized to form a theoretical travel time vector. After obtaining the theoretical travel time vector, it is subtracted element-by-element from the observed travel time vector to obtain the travel time residual vector. Each element in the travel time residual vector represents the deviation between the theoretical travel time and the observed travel time on a source-receiver path. This deviation reflects the degree of difference between the current velocity model and the actual underground velocity structure, and is the basis for subsequent velocity model correction.

[0071] While obtaining travel time residual information, the determination of the finite-frequency sensitivity matrix of the fat ray is also performed. First, based on the three-dimensional ray paths obtained from forward modeling, the corresponding finite-frequency sensitivity range is determined with each ray path as the center. The determination of this sensitivity range is based on the elastic wave dominant frequency and the current velocity model: the ray path is discretized into several path segments, and the local equivalent wavelength is determined according to the local average velocity and elastic wave dominant frequency at each path segment. Then, based on the local equivalent wavelength and the path distance from the path segment to the source point and the receiver point, the finite-frequency sensitivity radius of the fat ray corresponding to that path segment is calculated, thereby determining the finite-frequency sensitivity range of the fat ray. Then, for each model element, the shortest spatial distance from its center point to each path segment of the ray path is calculated. The distance attenuation weight is determined based on the relationship between this shortest spatial distance and the finite-frequency sensitive radius of the fat ray: if the shortest spatial distance is greater than the finite-frequency sensitive radius of the fat ray, the distance attenuation weight is zero; if the shortest spatial distance is less than or equal to the finite-frequency sensitive radius of the fat ray, the distance attenuation weight is calculated using a Gaussian attenuation function based on the shortest spatial distance and the lateral attenuation scale parameter, such that the closer the model element is to the ray path, the greater the weight, and the farther away, the smaller the weight. Finally, the length of each path segment is multiplied by the corresponding distance attenuation weight and then accumulated along the entire ray path to obtain the finite-frequency sensitive contribution of each ray to each model element. All sensitive contributions together form the finite-frequency sensitive matrix of the fat ray.

[0072] The fat-ray finite-frequency sensitivity matrix constructed in this invention differs fundamentally from the traditional geometric ray sensitivity matrix. The traditional geometric ray sensitivity matrix only records the path length within the model unit through which the ray actually passes; if it doesn't pass through, the sensitivity coefficient is zero. In contrast, the fat-ray finite-frequency sensitivity matrix in this invention includes all model units within the finite-frequency sensitivity range surrounding the ray path in the sensitivity relationship. Even if a model unit is not directly passed by the ideal geometric ray, as long as it is within the finite-frequency sensitivity influence range corresponding to the ray, it can still obtain a non-zero sensitivity weight. Each element in this matrix represents the finite-frequency weighted sensitivity relationship between the travel time residual and the corresponding model unit update, which can more fully express the distributed sensitivity formed by the finite-frequency propagation of elastic waves, providing a data constraint basis that better conforms to the actual wave field propagation characteristics for subsequent velocity model updates.

[0073] Step S120: Based on the travel time residual information and the fat ray finite frequency sensitivity matrix, determine the cooperative constraint system. The cooperative constraint system is used to adjust the model update amount of the current velocity model to the target update amount. The cooperative constraint system includes fat ray finite frequency sensitivity matrix constraints, data weight constraints, model weight constraints, spatial smoothness constraints, and directional constraints, as detailed below:

[0074] In step S120, after determining the travel time residual information and the fat ray finite frequency sensitivity matrix, the construction of the cooperative constraint system begins.

[0075] The aforementioned collaborative constraint system is not a single constraint term, but rather comprises constraints on the fat ray finite-frequency sensitivity matrix, data weight constraints, model weight constraints, spatial smoothness constraints, and directional constraints. This collaborative constraint system is used to adjust the current velocity model update amount to the target update amount, ensuring that the model update process is simultaneously controlled by travel-time data fitting, the strength of local data constraints, spatial continuity, and the coverage state of the ray path direction.

[0076] First, the fat-ray finite-frequency sensitivity matrix constraint is represented by the constructed fat-ray finite-frequency sensitivity matrix. This matrix is ​​used to establish a finite-frequency weighted sensitivity relationship between travel-time residuals and model update amounts, where each element represents the sensitivity coefficient between the travel-time residual and the corresponding model unit update amount. This constraint is the data sensitivity foundation of the collaborative constraint system, providing a basic data fitting direction for the construction of various subsequent constraints and velocity model updates.

[0077] Secondly, a data weight matrix is ​​constructed. This matrix controls the contribution of different observation travel times to the inversion process. It can be constructed as a diagonal matrix, with its diagonal elements determined based on the reliability of the observation travel times. When an observation travel time has high reliability and small error, its corresponding data weight is larger; conversely, when the observation travel time has low reliability and large error, its corresponding data weight is smaller. Through this method, the data weight matrix can adaptively adjust the influence of different travel times on the inversion objective function according to the quality of the observation data.

[0078] Next, the model weight matrix is ​​constructed. This matrix is ​​used to apply differentiated constraints to different model units based on the strength of local data constraints. This matrix is ​​preferably constructed as a diagonal matrix, with each model unit corresponding to an independent model constraint weight. The construction of the model constraint weights is based on local data constraint statistics: first, the data-weighted sensitivity and ray influence frequency of each model unit are calculated based on the fat ray finite-frequency sensitivity matrix; then, the above statistics are normalized; next, a local data constraint strength index is constructed based on the normalized data-weighted sensitivity and ray influence frequency; finally, the model constraint weight of each model unit is determined based on the local data constraint strength index and the basic model constraint coefficients. The model constraint weights and the local data constraint strength index are constructed with an inverse proportional relationship: regions with strong data-weighted sensitivity and a large number of effective rays indicate strong constraints from the observed data, corresponding to weaker model damping; regions with weak data-weighted sensitivity and a small number of effective rays indicate a weak constraint state, corresponding to stronger model constraints to suppress unstable updates and non-physical oscillations.

[0079] Next, a spatial smoothing matrix is ​​constructed. This matrix is ​​built based on the adjacency relationships between 3D mesh elements. For adjacent model elements, a smoothing constraint is established to ensure that the difference in model parameters approaches zero. The spatial smoothing matrix is ​​composed of the difference relationships between all adjacent mesh elements. This matrix is ​​used to constrain non-physical abrupt changes between adjacent model elements, ensuring the overall spatial continuity of the velocity model.

[0080] Finally, a directional constraint matrix is ​​constructed. This matrix is ​​built based on the ray path direction distribution information of the current iteration. Specifically, the coverage intensity of ray paths near each model unit in various spatial directions is first calculated to obtain the directional coverage. Then, directional smoothing weights are set according to the coverage strength of different directions. The stronger the ray coverage in a certain direction, the smaller the smoothing constraint weight for that direction; the weaker the ray coverage in a certain direction, the larger the smoothing constraint weight for that direction. Through this process, the directional constraint matrix can apply differentiated constraint strengths to various spatial directions based on the current ray path direction distribution, so that directions with stronger ray coverage are subject to weaker smoothing constraints, while directions with weaker ray coverage are subject to stronger stabilization constraints. This process differs from the fixed isotropic smoothing method, allowing the smoothing constraints to be updated with each iteration of forward modeling ray path changes.

[0081] Through the above processing, a collaborative constraint system is formed, consisting of the fat ray finite-frequency sensitivity matrix constraint, data weight constraint, model weight constraint, spatial smoothing constraint, and directional constraint. This collaborative constraint system differs from a pre-fixed uniform damping constraint, and its significant characteristic lies in its adaptability: the data-weighted sensitivity, ray influence frequency, and directional coverage are automatically calculated from the forward modeling ray path and the fat ray finite-frequency sensitivity matrix of the current iteration, rather than being manually given; the model constraint weights are calculated individually for each model unit, rather than using the same damping parameter for all model units; after each iteration, changes in the velocity model lead to changes in the ray path, sensitivity matrix, and statistics, thus requiring the reconstruction of the model weight matrix and directional constraint matrix. The resulting collaborative constraint system can apply differentiated constraints to model units based on the strength of local data constraints, mitigating the inversion instability caused by uneven constraints under limited observation conditions, while avoiding mismatch caused by using a fixed constraint system for long-term multiple iterations.

[0082] Step S140: In response to the cooperative constraint system, a sparse augmented least squares solution system is generated. The sparse augmented least squares solution system is used to solve for the model update amount and update the current velocity model to obtain the updated velocity model. The specific content is as follows:

[0083] In step S140, after the construction of the collaborative constraint system is completed, the generation and model update solution of the sparse augmented least squares solution system are entered.

[0084] The sparse augmented least squares solution system is generated based on a pre-constructed cooperative constraint system. This system is used to solve for the model update amount in the current round and to correct the current velocity model using this update amount, thereby obtaining the updated velocity model. Specifically, the fat ray finite frequency sensitivity matrix constraint, data weight constraint, model weight constraint, spatial smoothing constraint, and directional constraint in the cooperative constraint system are uniformly assembled into the sparse augmented least squares solution system equation. In this equation, the product of the fat ray finite frequency sensitivity matrix and the data weight matrix constitutes a data fitting term, used to constrain the model update amount to fit the travel time residual; the model weight matrix constitutes a model constraint term, used to apply differentiated damping to the model update amount according to the strength of local data constraints; the spatial smoothing matrix constitutes a spatial continuity constraint term, used to suppress non-physical abrupt changes between adjacent model units; and the directional constraint matrix constitutes a directional constraint term, used to apply differentiated smoothing to the model update in each spatial direction according to the ray path direction coverage state. The above four constraints together constitute the coefficient matrix of the sparse augmented least squares solution system, while the travel time residual, after being weighted by the data weight matrix, constitutes the right-hand side of the system.

[0085] To avoid storage overflow and computational difficulties caused by the densification of the sensitive matrix, constraint matrix, and augmented system under large-scale 3D mesh and multi-ray conditions, this invention only saves non-zero elements during matrix assembly. The non-zero elements of the sensitive matrix, smoothing matrix, directional constraint matrix, and sparse augmented least squares solution system are first recorded as triples, containing their row indices, column indices, and values. After the sparse augmented least squares solution system is assembled, it is converted into a compressed row-stored sparse matrix format for subsequent matrix operations and sparse iterative solutions. This sparse organization transforms the theoretically constructible augmented system in large-scale 3D travel-time tomography into a practically storable, computationally achievable, and iteratively solvable sparse linear system.

[0086] After the sparse augmented least squares solution system is assembled, an iterative solution method suitable for large-scale sparse least squares problems is used to solve for the model update amount. The solved model update amount represents the magnitude and direction of the current velocity model that needs to be adjusted. Based on this model update amount, a temporary update model is first formed under the condition of no additional constraints, that is, the current model parameters are added to the model update amount. To avoid velocity values ​​that do not conform to geological and physical meaning in local weakly constrained regions or abnormal update regions, this invention introduces velocity physical feasible region constraints in the model update stage. Specifically, a lower velocity limit and an upper velocity limit are preset based on the geological data and prior information of rock physics in the exploration area; for each model element in the temporary update model, it is determined whether its velocity value falls within the preset upper and lower velocity limits; if the velocity value of a model element is lower than the lower velocity limit, it is projected to the lower velocity limit; if it is higher than the upper velocity limit, it is projected to the upper velocity limit; if it is already within the upper and lower velocity limits, the velocity value of the element remains unchanged. Through the above velocity boundary projection processing, the updated velocity model constrained by the velocity physical feasible region is obtained.

[0087] This updated velocity model serves as the final output of the current iteration and also as the input model for the next iteration. In the next iteration, forward modeling will be re-executed based on this updated velocity model. The travel time residuals, the finite-frequency sensitivity matrix of the fat ray, the local constraint statistics, the cooperative constraint system, and the sparse augmented least squares solution system are all reconstructed as the velocity model changes, thus achieving a closed-loop process where the sensitivity matrix and constraint system are self-reconstructed as the velocity model is updated in each iteration.

[0088] Based on steps S110 to S140 above, in this embodiment of the invention, by constructing a fat ray finite-frequency sensitive matrix, the travel-time sensitivity relationship within a finite spatial range around the ray path can be expressed more fully, improving the reliability of restoring the boundary and morphology of local anomalies. By establishing a collaborative constraint system including fat ray finite-frequency sensitive matrix constraints, data weight constraints, model weight constraints, spatial smoothness constraints, and directional constraints, differentiated constraints can be applied to model units according to the strength of local data constraints, alleviating the instability problem caused by uneven constraints under finite observation conditions. By generating a sparse augmented least squares solution system and reconstructing the sensitive matrix and collaborative constraint system round by round based on the updated velocity model, iterative mismatch caused by a fixed constraint system can be avoided. At the same time, the sparse matrix storage method reduces the storage and computational burden in large-scale three-dimensional inversion, improves the executability of engineering computation, and thus solves the technical problems of low accuracy in underground velocity structure imaging, insufficient reliability of local anomaly restoration, and large engineering computational burden caused by the existing methods in three-dimensional elastic wave travel-time tomography, which suffer from insufficient finite-frequency sensitive expression, lack of collaboration among multiple types of constraints, inability of the constraint system to be adaptively reconstructed in multiple iterations, and low efficiency in large-scale computational organization.

[0089] The collaborative constraint self-reconfiguration multi-round iterative processing method of the present invention determines the travel time residual information and the fat ray finite-frequency sensitivity matrix of the current velocity model, including: performing three-dimensional elastic wave travel time forward modeling based on the current velocity model and the source receiving observation system to obtain the theoretical travel time and ray path; determining the travel time residual information based on the theoretical travel time and the observed travel time; determining the fat ray finite-frequency sensitivity influence range based on the ray path, the elastic wave dominant frequency, and the current velocity model; determining the distance attenuation weight based on the fat ray finite-frequency sensitivity influence range and the spatial distance from the model unit to the ray path; and determining the fat ray finite-frequency sensitivity matrix based on the path micro-segment length and the distance attenuation weight.

[0090] Furthermore, based on the ray path, the dominant frequency of the elastic wave, and the current velocity model, the frequency-limited sensitivity range of the fat ray is determined, including: discretizing the ray path into several path segments; determining the local average velocity corresponding to each path segment; determining the local equivalent wavelength based on the dominant frequency of the elastic wave and the local average velocity; determining the frequency-limited sensitivity radius of the fat ray based on the local equivalent wavelength and the path distance from the path segment to the source point and the receiver point; and determining the frequency-limited sensitivity range of the fat ray based on the frequency-limited sensitivity radius of the fat ray.

[0091] Furthermore, based on the finite-frequency sensitive influence range of the fat ray and the spatial distance from the model unit to the ray path, the distance attenuation weight is determined, including: determining the shortest spatial distance from the center point of the model unit to the path micro-segment; in response to the shortest spatial distance being greater than the finite-frequency sensitive radius, determining the distance attenuation weight to be zero; in response to the shortest spatial distance being less than or equal to the finite-frequency sensitive radius of the fat ray, determining the distance attenuation weight using a Gaussian attenuation function based on the shortest spatial distance and the lateral attenuation scale parameter.

[0092] Furthermore, based on travel time residual information and the fat ray finite-frequency sensitivity matrix, a collaborative constraint system is determined, including: determining local data constraint statistics based on the fat ray finite-frequency sensitivity matrix, wherein the local data constraint statistics include data weighted sensitivity and ray influence frequency; determining the model weight matrix based on the local data constraint statistics, wherein the model weight matrix is ​​a diagonal matrix and the diagonal elements are inversely proportional to the local data constraint statistics; determining the spatial smoothing matrix based on the 3D mesh adjacency relationship; determining the directional constraint matrix based on the ray path direction distribution information, wherein the directional constraint matrix is ​​used to apply differentiated constraint strengths to various directions in space; determining the data weight matrix based on the reliability of observation travel time; and determining the collaborative constraint system based on the fat ray finite-frequency sensitivity matrix, data weight matrix, model weight matrix, spatial smoothing matrix, and directional constraint matrix.

[0093] Furthermore, based on local data constraint statistics, the model weight matrix is ​​determined, including: normalizing the data weighted sensitivity and the number of ray influences to obtain normalized data weighted sensitivity and normalized ray influences; determining the local data constraint strength index based on the normalized data weighted sensitivity and normalized ray influences; determining the model constraint weights corresponding to each model unit based on the local data constraint strength index and the basic model constraint coefficients, wherein the model constraint weights are inversely proportional to the local data constraint strength index; and determining the model weight matrix based on all model constraint weights.

[0094] Furthermore, based on the ray path direction distribution information, the directional constraint matrix is ​​determined, including: statistically analyzing the coverage intensity of ray paths near each model unit in various spatial directions to obtain the directional coverage; determining the directional smoothing weight of each model unit in each direction based on the directional coverage, wherein the directional smoothing weight is inversely proportional to the directional coverage; and determining the directional constraint matrix based on the directional smoothing weights of all model units.

[0095] Furthermore, in response to the cooperative constraint system, after generating the sparse augmented least squares solution system, the process further includes: solving the sparse augmented least squares solution system using a sparse iterative solution method to obtain the model update amount; determining a temporary update model based on the model update amount; performing velocity boundary projection on the temporary update model based on a preset velocity lower limit and a preset velocity upper limit to obtain the updated velocity model; determining the root mean square of the travel time residual and the relative magnitude of the model update based on the updated velocity model; stopping iteration and outputting the underground velocity structure imaging results in response to the root mean square of the travel time residual being less than or equal to the travel time residual threshold, or the relative magnitude of the model update being less than or equal to the model update threshold, or the current iteration round reaching the maximum number of iteration rounds; and re-executing forward modeling, fat ray finite frequency sensitivity matrix construction, and cooperative constraint system reconstruction in response to the root mean square of the travel time residual being greater than the travel time residual threshold, the relative magnitude of the model update being greater than the model update threshold, and the current iteration round being less than the maximum number of iteration rounds.

[0096] Another collaborative constraint self-reconfiguration multi-round iterative processing method according to one embodiment of the present invention, such as Figures 2-5 As shown, the method includes the following steps:

[0097] Step 201, Model and observation system initialization;

[0098] First, a three-dimensional computational domain is established, and the underground target space is meshed. The area to be imaged is divided into several three-dimensional mesh cells, and initial velocity parameters are assigned to each mesh cell to form an initial velocity model. Let the first... The velocity model during round iteration is: ,in, For spatial location, Indicates the first The velocity value at this position during the round iteration. To facilitate travel-time tomography inversion, the velocity model can be converted into a slowness model:

[0099] (1)

[0100] After discretization, the first... Wheel model parameter vector:

[0101] (2)

[0102] in, This represents the total number of 3D mesh cells. Indicates the first The model parameter vector for each iteration.

[0103] Simultaneously, source and receiver points are deployed according to engineering observation conditions to form a source-receiver observation system. All observation travel times constitute the observation travel time vector:

[0104] (3)

[0105] in, To observe the number of travel times, This module outputs the observation travel time vector. The output includes a 3D initial model, a 3D computational grid, a source-receiver observation system, and observation travel time data, providing the basic input for subsequent forward modeling and tomographic inversion.

[0106] Step S202, forward modeling and sensitivity matrix construction;

[0107] In the In each iteration, the current model Import the forward modeling module and perform a three-dimensional elastic wave travel time forward modeling calculation based on the current velocity structure to obtain the theoretical travel time and ray path. For the... The theoretical travel time of the source-receive path can be expressed as:

[0108] (4)

[0109] in, Indicates the first During the first iteration ray path, For the current slowness model, d l This represents the length of a path segment. The complete theoretical travel time consists of:

[0110] (5)

[0111] Then the residuals between the observed travel time and the theoretical travel time are calculated:

[0112] (6)

[0113] in, For the first The residual vector during wheel movement.

[0114] After obtaining the ray path, the finite-frequency sensitive influence range of the fat ray is determined with the ray path as the center, and the sensitivity weight is calculated based on the spatial relationship between the 3D model elements and the ray path. For model elements located within the finite-frequency sensitive range of the fat ray, their corresponding sensitivity contribution is calculated; for model elements located outside the sensitive range, their sensitivity weight is reset to zero. This constructs the... Finite-frequency sensitive matrix of the ray-shaped beam:

[0115] (7)

[0116] in, Indicates the first Wheel of Life The first ray to the The finite-frequency sensitivity contribution of each model element. The output of this module is the theoretical travel time. Travel time residual , ray path and fat ray finite frequency sensitivity matrix .

[0117] Step S203: Construction of the collaborative constraint system;

[0118] According to the The forward modeling results and the finite-frequency sensitivity matrix of the fat ray are used to construct a collaborative constraint system. This collaborative constraint system is not a single constraint term, but rather consists of constraints related to the finite-frequency sensitivity matrix of the fat ray, data weight constraints, model weight constraints, spatial smoothness constraints, and directional constraints.

[0119] Among them, the finite frequency sensitivity matrix constraint is given by This represents the sensitive relationship between travel time residuals and model update amounts; the data weight matrix. Used to control the contribution of different observation data in the inversion; model weight matrix Used to apply differentiated constraints to different model units based on the strength of local data constraints; spatial smoothing matrix Used to constrain non-physical abrupt changes between adjacent mesh cells; directional constraint matrix Used to reflect the differences in the distribution of ray path directions.

[0120] The strength of local data constraints can be obtained statistically from the finite-frequency sensitivity matrix of the fat ray. For example, the first... The data-weighted sensitivity of each model unit can be expressed as:

[0121] (8)

[0122] in, Indicates the first Wheel of Life The strength of the constraints imposed on the observed data by each model unit. This statistic is used to assist in constructing the model weight matrix, rather than being used directly as the resolution ontology.

[0123] Through the above processing, regions with strong data constraints can avoid being excessively suppressed by uniform strong smoothing, while regions with weak data constraints can improve update stability by enhancing constraints, thereby alleviating the problem that the single constraint method in traditional time-inversion is difficult to adapt to local constraint differences.

[0124] Step S204: Solving the sparse augmented least squares solution system and updating the model;

[0125] After the collaborative constraint system is constructed, the data terms, model weight terms, spatial smoothing terms, and directional constraint terms are uniformly assembled into a sparse augmented least squares solution for the system matrix equation. The equation of the round matrix can be expressed as:

[0126] (9)

[0127] in, For the first Round augmented matrix equation, For the first Wheel model update amount, For the first The right end of the wheel is extended.

[0128] Specifically, the augmented matrix and the right-hand side can be written as:

[0129] (10)

[0130] (11)

[0131] in, These are the model weight constraint coefficients. The spatial smoothing constraint coefficient is... This is the directional constraint coefficient.

[0132] By solving the above sparse augmented least squares system equations, the model update amount is obtained:

[0133] And further complete the model update:

[0134] (12)

[0135] If the model parameters are represented by slowness, then the updated velocity model can be obtained based on the relationship between velocity and slowness:

[0136] (13)

[0137] In actual calculations, , , , , and All are organized in a sparse matrix form to reduce the storage and computational burden under conditions of large-scale 3D meshes, multiple rays, and multiple iterations.

[0138] Step S205: Iteration control and result output;

[0139] After completing the k-th round of velocity model update, determine whether the termination conditions are met. The termination conditions include, but are not limited to: whether the travel time residual is lower than a preset threshold, whether the model update amount is lower than a preset threshold, whether the objective function change tends to stabilize, or whether the number of iterations has reached the maximum set value.

[0140] If the termination condition is not met, the velocity model of round k+1 is re-imported into the forward modeling module to recalculate the theoretical travel time and ray path, and the finite-frequency sensitivity matrix of the fat ray, local constraint statistics, cooperative constraint system, and sparse augmented least squares solution system are reconstructed. This achieves self-reconstruction of the constraint system in multiple iterations.

[0141] If the termination condition is met, the iteration stops, and the subsurface velocity structure imaging results are output. The output results include the final velocity model, theoretical travel time, travel time residuals, ray paths, fat ray finite-frequency sensitivity matrix, cooperative constraint matrix, model update records, and subsurface velocity structure imaging maps. The imaging maps may include three-dimensional velocity volumes, horizontal slices, vertical profiles, and spatial distribution maps of anomalies.

[0142] Based on steps S201 to S205 above, in this embodiment of the invention, a closed-loop processing method for three-dimensional elastic wave travel-time tomography is formed through model initialization, forward modeling, construction of the fat-ray finite-frequency sensitive matrix, formation of a cooperative constraint system, solution using a sparse augmented least squares solution system, and multiple rounds of self-reconstruction iteration. This method can solve the problems of insufficient finite-frequency sensitive representation, insufficient cooperative constraints, difficulty in adapting the fixed constraint system to model updates, and heavy computational burden in large-scale three-dimensional inversion in traditional travel-time inversion, thereby improving the stability of underground velocity structure imaging, the reliability of local anomaly recovery, and the feasibility of engineering calculations.

[0143] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0144] This invention also provides a collaborative constraint self-reconfigurable multi-round iterative processing device for implementing the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0145] Figure 6 According to one embodiment of the present invention, a collaborative constraint self-reconfigurable multi-round iterative processing device includes:

[0146] The response module 301 is used to respond to the start signal of the three-dimensional elastic wave travel time tomography processing to determine the travel time residual information and the fat ray finite frequency sensitivity matrix of the current velocity model. The travel time residual information includes the difference information between the theoretical travel time and the observed travel time. The fat ray finite frequency sensitivity matrix is ​​used to describe the finite frequency weighted sensitivity relationship between the travel time residual and the model update amount.

[0147] The determination module 302 is used to determine the cooperative constraint system based on the travel time residual information and the fat ray finite frequency sensitivity matrix. The cooperative constraint system is used to adjust the model update amount of the current velocity model to the target update amount. The cooperative constraint system includes fat ray finite frequency sensitivity matrix constraints, data weight constraints, model weight constraints, spatial smoothness constraints, and directional constraints.

[0148] The generation module 303 is used to generate a sparse augmented least squares solution system in response to the cooperative constraint system. The sparse augmented least squares solution system is used to solve the model update amount and update the current velocity model to obtain the updated velocity model.

[0149] It should be noted that the above modules can be implemented by software or hardware. For the latter, they can be implemented in the following ways, but are not limited to: all the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0150] According to one embodiment of the present invention, an electronic device is also provided, comprising: a memory storing an executable program; and a processor for running the program, wherein the program executes the above-described cooperative constraint self-reconfiguration multi-round iterative processing method during runtime.

[0151] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0152] Step S1: In response to the three-dimensional elastic wave travel time tomography processing start signal, determine the travel time residual information and fat ray finite frequency sensitivity matrix of the current velocity model. The travel time residual information includes the difference information between theoretical travel time and observed travel time. The fat ray finite frequency sensitivity matrix is ​​used to describe the finite frequency weighted sensitivity relationship between travel time residual and model update amount.

[0153] Step S2: Based on the travel time residual information and the fat ray finite frequency sensitivity matrix, determine the cooperative constraint system. The cooperative constraint system is used to adjust the model update amount of the current velocity model to the target update amount. The cooperative constraint system includes fat ray finite frequency sensitivity matrix constraints, data weight constraints, model weight constraints, spatial smoothness constraints, and directional constraints.

[0154] Step S3: In response to the cooperative constraint system, a sparse augmented least squares solution system is generated. The sparse augmented least squares solution system is used to solve the model update amount and update the current velocity model to obtain the updated velocity model.

[0155] According to one embodiment of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the storage medium is located to execute the above-described cooperative constraint self-reconfiguration multi-round iterative processing method.

[0156] Optionally, in this embodiment, the storage medium may be configured to store a computer program for performing the following steps:

[0157] Step S1: In response to the three-dimensional elastic wave travel time tomography processing start signal, determine the travel time residual information and fat ray finite frequency sensitivity matrix of the current velocity model. The travel time residual information includes the difference information between theoretical travel time and observed travel time. The fat ray finite frequency sensitivity matrix is ​​used to describe the finite frequency weighted sensitivity relationship between travel time residual and model update amount.

[0158] Step S2: Based on the travel time residual information and the fat ray finite frequency sensitivity matrix, determine the cooperative constraint system. The cooperative constraint system is used to adjust the model update amount of the current velocity model to the target update amount. The cooperative constraint system includes fat ray finite frequency sensitivity matrix constraints, data weight constraints, model weight constraints, spatial smoothness constraints, and directional constraints.

[0159] Step S3: In response to the cooperative constraint system, a sparse augmented least squares solution system is generated. The sparse augmented least squares solution system is used to solve the model update amount and update the current velocity model to obtain the updated velocity model.

[0160] Optionally, in this embodiment, the storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0161] According to one embodiment of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements the above-described cooperative constraint self-reconfiguration multi-round iterative processing method.

[0162] Optionally, in this embodiment, the above-mentioned computer program product can be configured as a computer program that performs the following steps:

[0163] Step S1: In response to the three-dimensional elastic wave travel time tomography processing start signal, determine the travel time residual information and fat ray finite frequency sensitivity matrix of the current velocity model. The travel time residual information includes the difference information between theoretical travel time and observed travel time. The fat ray finite frequency sensitivity matrix is ​​used to describe the finite frequency weighted sensitivity relationship between travel time residual and model update amount.

[0164] Step S2: Based on the travel time residual information and the fat ray finite frequency sensitivity matrix, determine the cooperative constraint system. The cooperative constraint system is used to adjust the model update amount of the current velocity model to the target update amount. The cooperative constraint system includes fat ray finite frequency sensitivity matrix constraints, data weight constraints, model weight constraints, spatial smoothness constraints, and directional constraints.

[0165] Step S3: In response to the cooperative constraint system, a sparse augmented least squares solution system is generated. The sparse augmented least squares solution system is used to solve the model update amount and update the current velocity model to obtain the updated velocity model.

[0166] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated here.

[0167] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0168] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between units or modules may be electrical or other forms.

[0169] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0170] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0171] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0172] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A collaborative constraint self-reconfiguration multi-round iterative processing method, characterized in that, include: In response to the start signal of three-dimensional elastic wave travel time tomography processing, the travel time residual information and fat ray finite frequency sensitivity matrix of the current velocity model are determined. The travel time residual information includes the difference information between theoretical travel time and observed travel time, and the fat ray finite frequency sensitivity matrix is ​​used to describe the finite frequency weighted sensitivity relationship between travel time residual and model update amount. Based on the travel time residual information and the fat ray finite frequency sensitivity matrix, a cooperative constraint system is determined. The cooperative constraint system is used to adjust the model update amount of the current velocity model to the target update amount. The cooperative constraint system includes fat ray finite frequency sensitivity matrix constraints, data weight constraints, model weight constraints, spatial smoothness constraints, and directional constraints. In response to the cooperative constraint system, a sparse augmented least squares solution system is generated. The sparse augmented least squares solution system is used to solve the model update amount and update the current velocity model to obtain the updated velocity model. After generating the sparse augmented least squares solution system in response to the cooperative constraint system, the process further includes: The sparse iterative solution method is used to solve the sparse augmented least squares solution system to obtain the model update amount; Based on the aforementioned model update amount, determine the temporary updated model; Based on a preset lower speed limit and a preset upper speed limit, a speed boundary projection is performed on the temporary updated model to obtain the updated speed model; Based on the updated velocity model, determine the root mean square of the travel time residual and the relative magnitude of the model update; In response to the root mean square of the travel time residual being less than or equal to the travel time residual threshold, or the relative magnitude of the model update being less than or equal to the model update threshold, or the current iteration round reaching the maximum number of iteration rounds, the iteration is stopped and the underground velocity structure imaging results are output. In response to the fact that the root mean square of the travel time residual is greater than the travel time residual threshold, the relative magnitude of the model update is greater than the model update threshold, and the current iteration round is less than the maximum iteration round, the updated velocity model is used as the current velocity model for the next round, and forward modeling, fat ray finite frequency sensitivity matrix construction, and cooperative constraint system reconstruction are re-executed.

2. The collaborative constraint self-reconfiguration multi-round iterative processing method according to claim 1, characterized in that, Determining the travel time residual information of the current velocity model and the fat ray finite frequency sensitivity matrix includes: Based on the current velocity model and the source receiving observation system, perform three-dimensional elastic wave travel time forward modeling to obtain the theoretical travel time and ray path; Based on the theoretical travel time and the observed travel time, the travel time residual information is determined; Based on the ray path, the dominant frequency of the elastic wave, and the current velocity model, the finite frequency sensitivity range of the fat ray is determined. Based on the finite-frequency sensitive influence range of the fat ray and the spatial distance from the model unit to the ray path, the distance attenuation weight is determined; Based on the path segment length of the ray path and the distance attenuation weight, the finite frequency sensitivity matrix of the fat ray is determined.

3. The collaborative constraint self-reconfiguration multi-round iterative processing method according to claim 2, characterized in that, Based on the ray path, the dominant frequency of the elastic wave, and the current velocity model, the finite-frequency sensitive influence range of the fat ray is determined, including: The ray path is discretized into several path micro-segments; Determine the local average velocity corresponding to each of the path segments; Based on the dominant frequency of the elastic wave and the local average velocity, the local equivalent wavelength is determined; Based on the local equivalent wavelength and the path distance from the source point and the receiver point of the path micro-segment, the frequency-limited sensitive radius of the fat ray is determined. Based on the finite-frequency sensitive radius of the fat ray, the finite-frequency sensitive influence range of the fat ray is determined.

4. The collaborative constraint self-reconfiguration multi-round iterative processing method according to claim 2, characterized in that, Based on the finite-frequency sensitive influence range of the fat ray and the spatial distance from the model unit to the ray path, the distance attenuation weight is determined, including: Determine the shortest spatial distance from the center point of the model unit to the path segment; In response to the fact that the shortest spatial distance is greater than the finite-frequency sensitive radius of the fat ray, the distance attenuation weight is determined to be zero; In response to the shortest spatial distance being less than or equal to the finite-frequency sensitive radius of the fat ray, the distance attenuation weight is determined using a Gaussian attenuation function based on the shortest spatial distance and the lateral attenuation scale parameter.

5. The collaborative constraint self-reconfiguration multi-round iterative processing method according to claim 1, characterized in that, Based on the travel time residual information and the fat ray finite frequency sensitivity matrix, the cooperative constraint system is determined, including: Based on the fat ray finite frequency sensitivity matrix, local data constraint statistics are determined, wherein the local data constraint statistics include data weighted sensitivity and the number of ray influences; Based on the local data constraint statistics, a model weight matrix is ​​determined, wherein the model weight matrix is ​​a diagonal matrix, and the diagonal elements are inversely proportional to the local data constraint statistics. Determine the spatial smoothing matrix based on the adjacency relationship of the 3D mesh; Based on the ray path direction distribution information, a directional constraint matrix is ​​determined, wherein the directional constraint matrix is ​​used to apply differentiated constraint strengths to each direction in space; The data weight matrix is ​​determined based on the reliability of the observation travel time; The cooperative constraint system is determined based on the fat ray finite frequency sensitivity matrix, the data weight matrix, the model weight matrix, the spatial smoothing matrix, and the directional constraint matrix.

6. The collaborative constraint self-reconfiguration multi-round iterative processing method according to claim 5, characterized in that, Based on the local data constraint statistics, the model weight matrix is ​​determined, including: The data weighted sensitivity and the number of ray influences are normalized to obtain normalized data weighted sensitivity and normalized number of ray influences; Based on the normalized data weighted sensitivity and the normalized ray influence number, the local data constraint strength index is determined; Based on the local data constraint strength index and the basic model constraint coefficient, the model constraint weight corresponding to each model unit is determined, wherein the model constraint weight is inversely proportional to the local data constraint strength index. The model weight matrix is ​​determined based on all the model constraint weights.

7. The collaborative constraint self-reconfiguration multi-round iterative processing method according to claim 5, characterized in that, Based on the ray path direction distribution information, the directional constraint matrix is ​​determined, including: The coverage intensity of ray paths near each model unit in various spatial directions is statistically analyzed to obtain the directional coverage. Based on the directional coverage, the directional smoothing weights of each model unit in each direction are determined, wherein the directional smoothing weights are inversely proportional to the directional coverage. The directional constraint matrix is ​​determined based on the directional smoothing weights of all the model units.

8. A collaborative constraint self-reconfigurable multi-round iterative processing device, characterized in that, include: The response module is used to respond to the start signal of three-dimensional elastic wave travel time tomography processing, and determine the travel time residual information and fat ray finite frequency sensitivity matrix of the current velocity model. The travel time residual information includes the difference information between theoretical travel time and observed travel time, and the fat ray finite frequency sensitivity matrix is ​​used to describe the finite frequency weighted sensitivity relationship between travel time residual and model update amount. The determination module is used to determine a cooperative constraint system based on the travel time residual information and the fat ray finite frequency sensitivity matrix. The cooperative constraint system is used to adjust the model update amount of the current velocity model to the target update amount. The cooperative constraint system includes fat ray finite frequency sensitivity matrix constraints, data weight constraints, model weight constraints, spatial smoothness constraints, and directional constraints. A generation module is used to generate a sparse augmented least squares solution system in response to the cooperative constraint system. The sparse augmented least squares solution system is used to solve the model update amount and update the current velocity model to obtain the updated velocity model. After generating the sparse augmented least squares solution system in response to the cooperative constraint system, the process further includes: The sparse iterative solution method is used to solve the sparse augmented least squares solution system to obtain the model update amount; Based on the aforementioned model update amount, determine the temporary updated model; Based on a preset lower speed limit and a preset upper speed limit, a speed boundary projection is performed on the temporary updated model to obtain the updated speed model; Based on the updated velocity model, determine the root mean square of the travel time residual and the relative magnitude of the model update; In response to the root mean square of the travel time residual being less than or equal to the travel time residual threshold, or the relative magnitude of the model update being less than or equal to the model update threshold, or the current iteration round reaching the maximum number of iteration rounds, the iteration is stopped and the underground velocity structure imaging results are output. In response to the fact that the root mean square of the travel time residual is greater than the travel time residual threshold, the relative magnitude of the model update is greater than the model update threshold, and the current iteration round is less than the maximum iteration round, the updated velocity model is used as the current velocity model for the next round, and forward modeling, fat ray finite frequency sensitivity matrix construction, and cooperative constraint system reconstruction are re-executed.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, wherein, when the executable program is running, it controls the device on which the storage medium resides to execute the cooperative constraint self-reconfiguration multi-round iterative processing method of any one of claims 1 to 8.

Images