Self-adaptive multi-constraint modeling method for three-dimensional inversion of magnetic method measurement data

Abstract

The invention discloses a self-adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data, and belongs to the technical field of mining survey. The method comprises the following steps: acquiring original measurement data of a magnetic measurement area, constructing a multi-dimensional constraint system based on a self-adaptive constraint mechanism, finishing inversion modeling of the measurement data by combining a three-dimensional inversion algorithm, outputting a three-dimensional geologic structure model of a target area, and synchronously acquiring multi-dimensional auxiliary information and carrying out structured mapping to obtain a three-dimensional geologic structure model of the target area. The constraint mechanism constructs a multi-type dynamic constraint system, the weight is adjusted according to the iteration process, the problem of poor adaptability of the traditional constraint is solved, the geologic rationality of the model is improved, the inversion algorithm adopts a non-uniform grid and efficient solution mode, and the method is suitable for large-scale popularization and application. In combination with dual verification of data fitting and geological consistency, the inversion precision and efficiency are ensured, an output model fits an actual geological condition, reliable support is provided for geological exploration and resource evaluation, and the defects in the prior art are effectively overcome.

Description

Technical Field

[0001] This invention relates to the field of mining exploration technology, and in particular to an adaptive multi-constraint modeling method for three-dimensional inversion of magnetic survey data. Background Technology

[0002] Currently, the three-dimensional inversion modeling technology for magnetic measurement data has several shortcomings:

[0003] The constraint mechanism design is simple. Traditional methods mostly use fixed-type constraints and have not formed a multi-dimensional collaborative constraint system. Moreover, the constraint parameters are mostly set empirically and cannot be dynamically adjusted according to the geological characteristics of the target area and the attributes of the measurement data. At the same time, there is a lack of dynamic adjustment rules for constraint weights. The uniform constraint strength is used throughout the inversion iteration process, which is prone to problems such as model oscillation and loss of control in the early stage of inversion and insufficient data fitting accuracy in the later stage.

[0004] There are shortcomings in the inversion algorithm and model verification process. The grid division often uses uniform grids without being adjusted according to the resolution requirements of the target area, resulting in insufficient accuracy or excessive computation in key areas. Furthermore, model verification often relies solely on the goodness of fit of the data without combining geological borehole data, regional geological structure theory, and expert experience to conduct geological consistency verification. This can easily lead to contradictions where the inversion results conform to the data patterns but do not match the actual geological conditions, making it difficult to meet the practical requirements of high-precision geological exploration for three-dimensional inversion models. Summary of the Invention

[0005] The purpose of this invention is to provide an adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data, comprising the following steps:

[0007] The raw measurement data of the magnetic measurement area is acquired and preprocessed. Based on the geological characteristics of the target area and the attributes of the measurement data, an adaptive constraint mechanism is constructed, and multiple types of constraint parameters and dynamic weight adjustment rules are determined.

[0008] A multi-dimensional constraint system is built based on constraint parameters. The constraint conditions are embedded into the objective function and iterative process of the three-dimensional inversion algorithm. The objective function is solved by the three-dimensional inversion algorithm, the model parameters are iteratively updated, the inversion modeling of the measurement data is completed, and a three-dimensional geological structure model is output.

[0009] The three-dimensional geological structure model is fitted and validated, and the geological consistency is checked. The three-dimensional geological structure model of the target area is then output.

[0010] Furthermore, the raw measurement data of the magnetic measurement area is acquired and preprocessed, specifically including:

[0011] Raw magnetic anomaly measurement data of the target area is collected by magnetic measuring instruments. The three-dimensional coordinate information, data collection timestamp, instrument operating parameters and environmental parameters of each data collection point are recorded synchronously. Standardized measurement data are obtained after preprocessing.

[0012] Based on the coordinate range of the target area, a preset three-dimensional coordinate space is constructed, and standardized measurement data is mapped to the corresponding grid nodes according to the coordinate information to generate a structured data matrix.

[0013] Furthermore, the rows, columns, and depth of the structured data matrix correspond to the X-axis, Y-axis, and Z-axis dimensions of the three-dimensional coordinate system, respectively, and the matrix elements are the magnetic anomaly amplitudes of the corresponding grid nodes.

[0014] Furthermore, the process of constructing an adaptive constraint mechanism based on the geological characteristics and measurement data attributes of the target area, and determining multiple types of constraint parameters and dynamic weight adjustment rules, specifically includes:

[0015] The constraint types are determined, including geological prior constraints, data accuracy constraints, physical property constraints, inversion stability constraints, and spatial correlation constraints.

[0016] Extract multiple types of constraint parameters, establish dynamic weight adjustment rules, calculate adaptive weight coefficients for various constraint parameters, and dynamically adjust the degree of constraint influence at different iteration stages;

[0017] Set weight adjustment trigger conditions: when the residual change rate is greater than or equal to the preset change threshold during the inversion iteration process, automatically increase the weight coefficients of geological prior constraints and inversion stability constraints; when the residual change rate is less than the preset change threshold, increase the weight ratio of data accuracy constraints, physical attribute constraints and spatial correlation constraints.

[0018] Furthermore, multiple types of constraint parameters are extracted, specifically including:

[0019] Multiple types of constraint parameters include geological prior constraint parameters, data accuracy constraint parameters, physical property constraint parameters, inversion stability constraint parameters, and spatial correlation constraint parameters;

[0020] In the process of extracting geological prior constraint parameters, based on the geological exploration report, known borehole data and geological maps of the target area, the depth range of stratigraphic interfaces, lithological distribution patterns and structural development characteristics are extracted and converted into quantitative constraint parameters. The known borehole data includes borehole lithology, stratum thickness and burial depth, and the geological maps include stratigraphic distribution maps and structural outline maps.

[0021] When extracting data accuracy constraint parameters, the data accuracy weight is calculated based on the factory accuracy index of the measuring instrument, data acquisition density, and repeated measurement error. High weight is assigned to data in areas with high accuracy, and low weight is assigned to data in areas with low accuracy. At the same time, the maximum allowable threshold for data fitting residuals is set.

[0022] When extracting physical property constraint parameters, the range of values ​​for magnetic susceptibility and remanent magnetization is set in combination with the known physical property range of the magnetic measurement object. The magnetic measurement object includes magnetic ore bodies, volcanic rocks and sedimentary rocks.

[0023] When extracting inversion stability constraint parameters, based on the convergence requirements of the inversion iteration process, the iteration step size limit, residual descent threshold and model smoothness constraint are set.

[0024] When extracting spatial correlation constraint parameters, the physical attribute correlation coefficients of adjacent grid cells are set based on the spatial continuity characteristics of geological bodies.

[0025] Furthermore, a multi-dimensional constraint system is built based on constraint parameters. When embedding constraint conditions into the objective function and iterative process of the three-dimensional inversion algorithm, a constraint matrix C with dimension M×K is constructed based on the constraint parameters determined by the adaptive constraint mechanism, and constraint conditions are embedded.

[0026] Where M represents the total number of grid nodes in the inversion model, K represents the total number of constraint types, and C is the matrix element. mk This represents the constraint value of the m-th grid node under the k-th type of constraint.

[0027] Furthermore, when embedding geological prior constraints, the geological prior constraint parameters are transformed into model initial value constraints, and initial physical attribute values ​​are set for each grid node to limit the range of change of physical attribute values ​​during model iteration.

[0028] When embedding data precision constraints, the data precision weights are transformed into a weight matrix of the observed data. The weight matrix is ​​a diagonal matrix, and the diagonal elements are the precision weights of the corresponding observed data. The contribution of the observed data to the objective function is adjusted through the weight matrix.

[0029] When embedding physical attribute constraints, the range of physical attribute values ​​is transformed into inequality constraints for model parameters, and parameter values ​​that exceed the range are truncated and corrected during the iteration process.

[0030] When embedding inversion stability constraints, the inversion stability constraint parameters are transformed into regularization terms and embedded into the inversion objective function. The regularization terms include model smoothness regularization terms and iteration step size regularization terms.

[0031] When embedding spatial correlation constraints, the spatial correlation constraint parameters are transformed into a spatial correlation matrix, and the attribute correlation between adjacent grid nodes is limited through matrix operations.

[0032] Furthermore, the execution process of the 3D inversion algorithm specifically includes:

[0033] Based on the range and resolution requirements of the target area, a three-dimensional mesh model is constructed, and the target area is divided into several uniform or non-uniform mesh units.

[0034] Initialize the physical property parameters of the grid cells and set the initial model based on geological prior constraints;

[0035] Construct an inversion objective function, which includes a data fitting term and a constraint regularization term;

[0036] The conjugate gradient method is used to find the minimum value of the objective function. The physical property parameters of the grid cells are updated iteratively until the convergence condition is met. The convergence condition is that the iteration residual is less than a preset residual threshold, or the number of iterations reaches the maximum number of iterations.

[0037] Furthermore, the output three-dimensional geological structure model is fitted and validated, and its geological consistency is verified. Specifically, this includes:

[0038] Calculate the goodness of fit between the forward modeling data and the observed data of the inversion model;

[0039] Set a goodness-of-fit threshold. If the goodness-of-fit is greater than the goodness-of-fit threshold, the model data is considered to be well-fitted. Otherwise, adjust the data precision weights or constraint parameters in the region with large residuals and re-execute the inversion process.

[0040] Draw a residual distribution map and analyze the spatial distribution characteristics of the residuals. If the residuals are randomly distributed, the three-dimensional geological structure model is considered to have a good fit. If the residuals are regularly distributed, the measurement data for the area should be supplemented or the geological prior constraints should be corrected.

[0041] Perform geological consistency verification, generate a model verification report based on the geological consistency verification results. If the verification is successful, output the final three-dimensional geological structure model; if the verification fails, return to the constraint parameter adjustment stage and re-optimize the inversion modeling process.

[0042] Further geological consistency verification specifically includes:

[0043] Based on known geological borehole data, the parameters of the inversion model corresponding to the borehole location are extracted, compared with the lithology and stratum thickness actually revealed by the borehole, the parameter deviation rate is calculated, and a deviation rate threshold is set. If the deviation rate is less than the threshold, the geological consistency of the model is deemed qualified.

[0044] Based on the regional geological structure theory, the spatial morphology of the inversion model is analyzed to see if it conforms to the geological structure characteristics. The spatial morphology includes the strike, dip and burial depth of the ore body. If the spatial morphology is consistent with the regional geological structure theory, the model structure is deemed to be reasonable.

[0045] Based on expert experience, it is determined whether the three-dimensional geological structure model can reasonably explain the magnetic anomaly characteristics of the target area. If the expert review is approved, the geological consistency verification is completed.

[0046] Compared with the prior art, the beneficial effects of the present invention are:

[0047] 1. This invention synchronously collects magnetic anomaly data, along with three-dimensional coordinates, timestamps, instrument parameters, and environmental parameters using a magnetic measuring instrument. This generates a structured data matrix corresponding to the three-dimensional coordinate space, ensuring the integrity and traceability of the original data and preventing subsequent inversion deviations due to missing key information. The structured data matrix precisely maps the magnetic anomaly amplitude and spatial location along the X, Y, and Z axes, achieving ordered data management. This not only reduces the difficulty of data retrieval for the three-dimensional inversion algorithm but also accurately locates the spatial nodes corresponding to each magnetic anomaly, laying a solid foundation for the spatial accuracy of subsequent inversion modeling and solving the problem of inversion model distortion caused by cluttered information and insufficient precision.

[0048] 2. The adaptive multi-constraint mechanism construction and multi-dimensional constraint system embedding of this invention integrate parameters through an M×K constraint matrix, embedding different constraints into the inversion objective function and iterative process in a differentiated manner. This makes the initial values ​​of the model more consistent with the actual geological conditions, allowing high-precision data to play a greater role in the inversion, avoiding inversion results without physical meaning. The dynamic weight adjustment rule can flexibly adjust the constraint strength according to the changes in the iterative residuals, ensuring model stability in the early stage of inversion and improving data fitting in the later stage. This solves the problems of poor adaptability of the constraint mechanism and inability to dynamically match the iterative process, making the inversion process both controllable and flexible, and greatly improving the geological rationality of the inversion model.

[0049] 3. This invention, through a non-uniform grid design, ensures the inversion accuracy of key areas such as near-surface while reducing the computational load in deep areas and improving algorithm efficiency. The conjugate gradient method quickly solves for the minimum value of the objective function. Combined with the dual convergence conditions of iterative residuals and the number of iterations, it guarantees inversion accuracy while avoiding excessive iteration and resource waste. Goodness-of-fit calculation and residual distribution analysis quantify the data fitting effect. Geological consistency verification is achieved through borehole data comparison, geological structural theory analysis, and expert review. This solves the problems of high computational load, slow convergence, and limited verification in inversion algorithms, ensuring that the output 3D geological structure model conforms to data patterns and fits actual geological conditions, thus improving the reliability and practical value of the inversion results. Attached Figure Description

[0050] Figure 1 This is a flowchart of the overall process for adaptive multi-constraint modeling of three-dimensional inversion of magnetic measurement data according to the present invention.

[0051] Figure 2 This is a schematic diagram of the adaptive constraint mechanism construction process of the present invention;

[0052] Figure 3 This is a schematic diagram of the execution flow of the three-dimensional inversion algorithm of the present invention. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0054] Please see Figure 1-3 The present invention provides the following technical solutions:

[0055] An adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data includes the following steps:

[0056] The raw measurement data of the magnetic measurement area is acquired and preprocessed. Based on the geological characteristics of the target area and the attributes of the measurement data, an adaptive constraint mechanism is constructed, and multiple types of constraint parameters and dynamic weight adjustment rules are determined.

[0057] A multi-dimensional constraint system is built based on constraint parameters. The constraint conditions are embedded into the objective function and iterative process of the three-dimensional inversion algorithm. The objective function is solved by the three-dimensional inversion algorithm, the model parameters are iteratively updated, the inversion modeling of the measurement data is completed, and a three-dimensional geological structure model is output.

[0058] The three-dimensional geological structure model is fitted and validated, and the geological consistency is checked. The three-dimensional geological structure model of the target area is then output.

[0059] Acquire and preprocess the raw measurement data of the magnetic measurement area, specifically including:

[0060] Raw magnetic anomaly measurement data of the target area is collected by magnetic measuring instruments. The three-dimensional coordinate information, data collection timestamp, instrument operating parameters and environmental parameters of each data collection point are recorded synchronously. Standardized measurement data are obtained after preprocessing.

[0061] Based on the coordinate range of the target area, a preset three-dimensional coordinate space is constructed, and standardized measurement data is mapped to the corresponding grid nodes according to the coordinate information to generate a structured data matrix.

[0062] In this structured data matrix, the rows, columns, and depth correspond to the X-axis, Y-axis, and Z-axis dimensions of the three-dimensional coordinate system, respectively, and the matrix elements are the magnetic anomaly amplitudes of the corresponding grid nodes.

[0063] In the above embodiments, by synchronously recording three-dimensional coordinates, timestamps, instrument and environmental parameters, the traceability and integrity of the data are ensured. The standardized data is mapped to the three-dimensional coordinate space to generate a structured matrix, realizing the orderly management of the data. This not only reduces the data processing difficulty of subsequent algorithms, but also accurately locates the spatial position corresponding to each magnetic anomaly amplitude, laying the foundation for the spatial accuracy of three-dimensional inversion modeling and solving the inversion deviation problem caused by missing and messy information in traditional data processing.

[0064] The process of constructing an adaptive constraint mechanism based on the geological characteristics and measurement data attributes of the target area, and determining multiple types of constraint parameters and dynamic weight adjustment rules, specifically includes:

[0065] The types of constraints are determined, including geological prior constraints, data accuracy constraints, physical property constraints, inversion stability constraints, and spatial correlation constraints.

[0066] Multiple types of constraint parameters are extracted, including geological prior constraint parameters, data accuracy constraint parameters, physical property constraint parameters, inversion stability constraint parameters, and spatial correlation constraint parameters.

[0067] In extracting geological prior constraint parameters, based on the geological exploration report, known borehole data and geological maps of the target area, the depth range of stratigraphic interfaces, lithological distribution patterns and structural development characteristics are extracted and converted into quantitative constraint parameters. The known borehole data includes borehole lithology, stratum thickness and burial depth, and the geological maps include stratigraphic distribution maps and structural outline maps.

[0068] When extracting data accuracy constraint parameters, the data accuracy weight is calculated based on the factory accuracy index of the measuring instrument, data acquisition density, and repeated measurement error. High weight is assigned to data in areas with high accuracy, and low weight is assigned to data in areas with low accuracy. At the same time, the maximum allowable threshold for data fitting residuals is set.

[0069] When extracting physical property constraint parameters, the range of values ​​for magnetic susceptibility and remanent magnetization is set in combination with the known physical property range of the magnetic measurement object to avoid parameter values ​​with unreasonable physical meaning in the inversion results. The magnetic measurement objects include magnetic ore bodies, volcanic rocks and sedimentary rocks.

[0070] When extracting inversion stability constraint parameters, based on the convergence requirements of the inversion iteration process, we set iteration step size limits, residual descent thresholds and model smoothness constraints to suppress violent oscillations in the inversion model.

[0071] When extracting spatial correlation constraint parameters, based on the spatial continuity characteristics of geological bodies, the physical property correlation coefficients of adjacent grid cells are set to ensure that the spatial distribution of the inversion model conforms to the actual shape of the geological bodies.

[0072] Establish dynamic weight adjustment rules, calculate adaptive weight coefficients for various constraint parameters, and dynamically adjust the degree of constraint influence at different iteration stages;

[0073] Set weight adjustment trigger conditions: when the residual change rate is greater than or equal to the preset change threshold during the inversion iteration process, automatically increase the weight coefficients of geological prior constraints and inversion stability constraints; when the residual change rate is less than the preset change threshold, increase the weight ratio of data accuracy constraints, physical attribute constraints and spatial correlation constraints.

[0074] In the above embodiments, the adaptive constraint mechanism comprehensively covers five types of constraints, encompassing geological, data, physical, stability, and spatial correlation dimensions, achieving multi-dimensional collaborative constraints. By quantifying geological prior information, differentiating data precision weights, and reasonably limiting the range of physical attributes, fuzzy constraints are transformed into precise parameters, avoiding contradictions in physical or geological meanings in the inversion model. The dynamic weight adjustment rules and triggering conditions can flexibly adjust the constraint strength according to the changes in iterative residuals, ensuring model stability in the early stages of inversion and improving data fitting in the later stages. This solves the problems of poor adaptability and inability to dynamically match the iterative process of traditional constraint mechanisms, significantly improving the controllability and flexibility of the inversion process.

[0075] A multi-dimensional constraint system is built based on constraint parameters. When embedding constraint conditions into the objective function and iterative process of the three-dimensional inversion algorithm, a constraint matrix C with dimension M×K is constructed based on the constraint parameters determined by the adaptive constraint mechanism, and constraint conditions are embedded.

[0076] Where M represents the total number of grid nodes in the inversion model, K represents the total number of constraint types, and C is the matrix element. mk This represents the constraint value of the m-th grid node under the k-th type of constraint.

[0077] When embedding geological prior constraints, the geological prior constraint parameters are transformed into model initial value constraints, and initial physical attribute values ​​are set for each grid node to limit the range of changes in physical attribute values ​​during model iteration.

[0078] When embedding data precision constraints, the data precision weights are transformed into a weight matrix of the observed data. The weight matrix is ​​a diagonal matrix, and the diagonal elements are the precision weights of the corresponding observed data. The contribution of the observed data to the objective function is adjusted through the weight matrix.

[0079] When embedding physical property constraints, the range of physical property values ​​is transformed into inequality constraints for model parameters. That is, physical property values ​​are set for each grid node, and parameter values ​​that exceed the range are truncated and corrected during the iteration process.

[0080] When embedding inversion stability constraints, the inversion stability constraint parameters are transformed into regularization terms and embedded into the inversion objective function. The regularization terms include model smoothness regularization terms and iteration step size regularization terms. The model smoothness regularization term is constructed based on the attribute difference between adjacent grid nodes, and the iteration step size regularization term is constructed based on the rate of change of model parameters.

[0081] When embedding spatial correlation constraints, the spatial correlation constraint parameters are transformed into a spatial correlation matrix. Matrix operations are used to limit the attribute correlation of adjacent grid nodes to ensure the continuity of the model's spatial distribution.

[0082] In the above embodiments, the multi-dimensional constraint system construction and constraint embedding stage systematically integrates multiple types of constraint parameters by constructing an M×K constraint matrix, thereby achieving clear and structured management of constraint information. Different embedding methods are designed for different constraint types, such as converting geological prior constraints into initial value and variation range constraints, and data accuracy constraints into diagonal weight matrices, to ensure that each constraint can accurately act on key inversion stages. The introduction of regularization terms and spatial correlation matrices not only suppresses model oscillations and ensures inversion stability, but also takes into account the spatial continuity of geological bodies, solving the model distortion problem caused by the single constraint embedding method and inaccurate constraint effect in traditional methods. This allows the constraint conditions to be deeply integrated into the inversion process, improving the geological rationality and spatial consistency of the model.

[0083] The execution process of the 3D inversion algorithm specifically includes:

[0084] Based on the range and resolution requirements of the target area, a three-dimensional mesh model is constructed, and the target area is divided into several uniform or non-uniform mesh units.

[0085] Initialize the physical property parameters of the grid cells and set the initial model based on geological prior constraints;

[0086] Construct an inversion objective function, which includes a data fitting term and a constraint regularization term;

[0087] The conjugate gradient method is used to solve for the minimum value of the objective function. The physical property parameters of the grid cells are updated iteratively until the convergence condition is met. The convergence condition is that the iteration residual is less than the preset residual threshold, or the number of iterations reaches the maximum number of iterations.

[0088] In the above embodiments, during the execution of the 3D inversion algorithm, the non-uniform grid division flexibly adjusts the grid density according to the resolution requirements. The design of near-surface densification and deep sparseness ensures the inversion accuracy of key areas while reducing the computational load and improving the algorithm's running efficiency. The initial model setting based on geological prior constraints makes the model iteration starting point more consistent with the actual geological conditions, reducing the number of iterations and shortening the inversion cycle. The conjugate gradient method is used to solve the objective function, which can quickly find the minimum point. Combined with the dual convergence conditions of iteration residual and number of iterations, it ensures the accuracy of the inversion results while avoiding excessive iteration and wasting resources. This solves the problems of large computational load, slow convergence, and difficulty in controlling accuracy in traditional inversion algorithms, and achieves efficient and accurate inversion modeling.

[0089] The output 3D geological structure model is fitted and validated, and its geological consistency is verified. Specifically, this includes:

[0090] Calculate the goodness of fit between the forward modeling data and the observed data of the inversion model;

[0091] Set a goodness-of-fit threshold. If the goodness-of-fit is greater than the goodness-of-fit threshold, the model data is considered to be well-fitted. Otherwise, adjust the data precision weights or constraint parameters in the region with large residuals and re-execute the inversion process.

[0092] Plot the residual distribution map, which shows the distribution of the difference between the observed data and the forward modeling data. Analyze the spatial distribution characteristics of the residuals. If the residuals are randomly distributed, it is determined that the three-dimensional geological structure model has a good fit. If the residuals are regularly distributed, the measurement data for the area should be supplemented or the geological prior constraints should be corrected.

[0093] Geological consistency verification is performed by combining known geological borehole data, extracting the inversion model parameters corresponding to the borehole locations, comparing them with the lithology and stratum thickness actually revealed by the boreholes, calculating the parameter deviation rate, setting a deviation rate threshold, and determining that the model's geological consistency is qualified if the deviation rate is less than the threshold.

[0094] Based on the theory of regional geological structure, the spatial morphology of the inversion model is analyzed to see if it conforms to the characteristics of geological structure. The spatial morphology includes the strike, dip and burial depth of the ore body. If the spatial morphology is consistent with the theory of regional geological structure, the model is deemed to be of reasonableness.

[0095] Based on expert experience, determine whether the three-dimensional geological structure model can reasonably explain the magnetic anomaly characteristics of the target area. If the expert review is approved, the geological consistency verification is completed.

[0096] Record the goodness of fit, residual distribution characteristics, borehole comparison results, and expert review opinions. Generate a model verification report based on the geological consistency verification results. If the verification is qualified, the final three-dimensional geological structure model is output. If the verification is unqualified, return to the constraint parameter adjustment stage and re-optimize the inversion modeling process.

[0097] In the above embodiments, the model fitting verification and geological consistency verification stages comprehensively ensure the reliability of the inversion results. The goodness-of-fit calculation and residual distribution map analysis quantify the model fitting effect from the data level, and accurately locate problem areas through residual distribution characteristics, providing a clear direction for parameter adjustment. The geological consistency verification combines borehole data, regional geological theory and expert experience to verify the rationality of the model from a geological perspective, avoiding situations where the model only satisfies data fitting but violates geological laws. The verification report recording and non-compliance return adjustment mechanism form a complete quality control closed loop, ensuring that the output three-dimensional geological structure model not only conforms to data laws but also fits the actual geological conditions. This solves the problems of single verification and insufficient reliability of traditional inversion models, and provides high-quality model support for subsequent applications such as geological exploration and resource assessment.

[0098] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data, characterized in that, Includes the following steps: The raw measurement data of the magnetic measurement area is acquired and preprocessed. Based on the geological characteristics of the target area and the attributes of the measurement data, an adaptive constraint mechanism is constructed, and multiple types of constraint parameters and dynamic weight adjustment rules are determined. A multi-dimensional constraint system is built based on constraint parameters. The constraint conditions are embedded into the objective function and iterative process of the three-dimensional inversion algorithm. The objective function is solved by the three-dimensional inversion algorithm, the model parameters are iteratively updated, the inversion modeling of the measurement data is completed, and a three-dimensional geological structure model is output. The three-dimensional geological structure model is fitted and validated, and the geological consistency is checked. The three-dimensional geological structure model of the target area is then output.

2. The adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data as described in claim 1, characterized in that, Acquire and preprocess the raw measurement data of the magnetic measurement area, specifically including: Raw magnetic anomaly measurement data of the target area is collected by magnetic measuring instruments. The three-dimensional coordinate information, data collection timestamp, instrument operating parameters and environmental parameters of each data collection point are recorded synchronously. Standardized measurement data are obtained after preprocessing. Based on the coordinate range of the target area, a preset three-dimensional coordinate space is constructed, and standardized measurement data is mapped to the corresponding grid nodes according to the coordinate information to generate a structured data matrix.

3. The adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data as described in claim 2, characterized in that, The rows, columns, and depth of the structured data matrix correspond to the X-axis, Y-axis, and Z-axis dimensions of the three-dimensional coordinate system, respectively, and the matrix elements are the magnetic anomaly amplitudes of the corresponding grid nodes.

4. The adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data as described in claim 1, characterized in that, The process of constructing an adaptive constraint mechanism based on the geological characteristics and measurement data attributes of the target area, and determining multiple types of constraint parameters and dynamic weight adjustment rules, specifically includes: The constraint types are determined, including geological prior constraints, data accuracy constraints, physical property constraints, inversion stability constraints, and spatial correlation constraints. Extract multiple types of constraint parameters, establish dynamic weight adjustment rules, calculate adaptive weight coefficients for various constraint parameters, and dynamically adjust the degree of constraint influence at different iteration stages; Set weight adjustment trigger conditions: when the residual change rate is greater than or equal to the preset change threshold during the inversion iteration process, automatically increase the weight coefficients of geological prior constraints and inversion stability constraints; when the residual change rate is less than the preset change threshold, increase the weight ratio of data accuracy constraints, physical attribute constraints and spatial correlation constraints.

5. The adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data as described in claim 4, characterized in that, Extracting multiple types of constraint parameters, specifically including: Multiple types of constraint parameters include geological prior constraint parameters, data accuracy constraint parameters, physical property constraint parameters, inversion stability constraint parameters, and spatial correlation constraint parameters; In the process of extracting geological prior constraint parameters, based on the geological exploration report, known borehole data and geological maps of the target area, the depth range of stratigraphic interfaces, lithological distribution patterns and structural development characteristics are extracted and converted into quantitative constraint parameters. The known borehole data includes borehole lithology, stratum thickness and burial depth, and the geological maps include stratigraphic distribution maps and structural outline maps. When extracting data accuracy constraint parameters, the data accuracy weight is calculated based on the factory accuracy index of the measuring instrument, data acquisition density, and repeated measurement error. High weight is assigned to data in areas with high accuracy, and low weight is assigned to data in areas with low accuracy. At the same time, the maximum allowable threshold for data fitting residuals is set. When extracting physical property constraint parameters, the range of values ​​for magnetic susceptibility and remanent magnetization is set in combination with the known physical property range of the magnetic measurement object. The magnetic measurement object includes magnetic ore bodies, volcanic rocks and sedimentary rocks. When extracting inversion stability constraint parameters, based on the convergence requirements of the inversion iteration process, the iteration step size limit, residual descent threshold and model smoothness constraint are set. When extracting spatial correlation constraint parameters, the physical attribute correlation coefficients of adjacent grid cells are set based on the spatial continuity characteristics of geological bodies.

6. The adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data as described in claim 1, characterized in that, A multi-dimensional constraint system is built based on constraint parameters. When embedding constraint conditions into the objective function and iterative process of the three-dimensional inversion algorithm, a constraint matrix C with dimension M×K is constructed based on the constraint parameters determined by the adaptive constraint mechanism, and constraint conditions are embedded. Where M represents the total number of grid nodes in the inversion model, K represents the total number of constraint types, and C is the matrix element. mk This represents the constraint value of the m-th grid node under the k-th type of constraint.

7. The adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data as described in claim 6, characterized in that, When embedding geological prior constraints, the geological prior constraint parameters are transformed into model initial value constraints, and initial physical attribute values ​​are set for each grid node to limit the range of changes in physical attribute values ​​during model iteration. When embedding data precision constraints, the data precision weights are transformed into a weight matrix of the observed data. The weight matrix is ​​a diagonal matrix, and the diagonal elements are the precision weights of the corresponding observed data. The contribution of the observed data to the objective function is adjusted through the weight matrix. When embedding physical attribute constraints, the range of physical attribute values ​​is transformed into inequality constraints for model parameters, and parameter values ​​that exceed the range are truncated and corrected during the iteration process. When embedding inversion stability constraints, the inversion stability constraint parameters are transformed into regularization terms and embedded into the inversion objective function. The regularization terms include model smoothness regularization terms and iteration step size regularization terms. When embedding spatial correlation constraints, the spatial correlation constraint parameters are transformed into a spatial correlation matrix, and the attribute correlation between adjacent grid nodes is limited through matrix operations.

8. The adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data as described in claim 1, characterized in that, The execution process of the 3D inversion algorithm specifically includes: Based on the range and resolution requirements of the target area, a three-dimensional mesh model is constructed, and the target area is divided into several uniform or non-uniform mesh units. Initialize the physical property parameters of the grid cells and set the initial model based on geological prior constraints; Construct an inversion objective function, which includes a data fitting term and a constraint regularization term; The conjugate gradient method is used to find the minimum value of the objective function. The physical property parameters of the grid cells are updated iteratively until the convergence condition is met. The convergence condition is that the iteration residual is less than a preset residual threshold, or the number of iterations reaches the maximum number of iterations.

9. The adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data as described in claim 1, characterized in that, The output 3D geological structure model is fitted and validated, and its geological consistency is verified. Specifically, this includes: Calculate the goodness of fit between the forward modeling data and the observed data of the inversion model; Set a goodness-of-fit threshold. If the goodness-of-fit is greater than the goodness-of-fit threshold, the model data is considered to be well-fitted. Otherwise, adjust the data precision weights or constraint parameters in the region with large residuals and re-execute the inversion process. Draw a residual distribution map and analyze the spatial distribution characteristics of the residuals. If the residuals are randomly distributed, the three-dimensional geological structure model is considered to have a good fit. If the residuals are regularly distributed, the measurement data for the area should be supplemented or the geological prior constraints should be corrected. Perform geological consistency verification, generate a model verification report based on the geological consistency verification results. If the verification is successful, output the final three-dimensional geological structure model; if the verification fails, return to the constraint parameter adjustment stage and re-optimize the inversion modeling process.

10. The adaptive multi-constraint modeling method for three-dimensional inversion of magnetic measurement data as described in claim 9, characterized in that, Geological consistency verification specifically includes: Based on known geological borehole data, the parameters of the inversion model corresponding to the borehole location are extracted, compared with the lithology and stratum thickness actually revealed by the borehole, the parameter deviation rate is calculated, and a deviation rate threshold is set. If the deviation rate is less than the threshold, the geological consistency of the model is deemed qualified. Based on the regional geological structure theory, the spatial morphology of the inversion model is analyzed to see if it conforms to the geological structure characteristics. The spatial morphology includes the strike, dip and burial depth of the ore body. If the spatial morphology is consistent with the regional geological structure theory, the model structure is deemed to be reasonable. Based on expert experience, it is determined whether the three-dimensional geological structure model can reasonably explain the magnetic anomaly characteristics of the target area. If the expert review is approved, the geological consistency verification is completed.

Images