Coal mine high-precision three-dimensional geological modeling method based on multi-source data fusion and updating

Abstract

The application discloses a coal mine high-precision three-dimensional geological modeling method based on multi-source data fusion and updating, belongs to the technical field of coal mine geological exploration, and comprises the following steps: S1, collecting coal mine multi-source heterogeneous geological data, and processing to obtain standardized multi-source data; S2, obtaining a fused grid geological data body through spatial feature alignment, credibility weighted conflict resolution and weighted null value completion; S3, constructing a dynamic variation function, calculating grid attribute values through improved dynamic Kriging interpolation, and obtaining an initial three-dimensional geological model; S4, acquiring constraint data, quantitatively calculating model prediction values and deviation coefficients, correcting key parameters of the variation function, triggering local model updating, and finally reconstructing a local high-precision three-dimensional geological model; and S5, performing multi-dimensional precision verification, and outputting a high-precision coal mine three-dimensional geological final model. The above method meets the actual needs of high-precision, dynamic, safe and controllable models for coal mining engineering.

Description

Technical Field

[0001] This invention relates to the field of coal mine geological exploration technology, and in particular to a high-precision three-dimensional geological modeling method for coal mines based on multi-source data fusion and updating. Background Technology

[0002] In the current field of 3D geological modeling for coal mines, existing technologies mostly rely on static fusion modeling using multi-source data such as borehole and geophysical exploration data. This approach does not fully consider the reliability differences of heterogeneous multi-source data, and direct data fusion can easily lead to interference from low-quality data. Furthermore, the variograms used are mostly static parameter models, which cannot adapt to the spatiotemporal dynamic changes of geological properties during coal mining. After model construction, there is a lack of linkage mechanism with real-time mining disturbances, making it difficult to dynamically reflect the impact of mining activities on geological bodies. Consequently, it is difficult to meet the actual needs of coal mining engineering for high-precision, dynamic, safe, and controllable 3D geological models. Summary of the Invention

[0003] The purpose of this invention is to provide a high-precision three-dimensional geological modeling method for coal mines based on multi-source data fusion and updating, thereby solving the aforementioned technical problems.

[0004] To achieve the above objectives, this invention provides a high-precision three-dimensional geological modeling method for coal mines based on multi-source data fusion and updating, comprising the following steps:

[0005] S1. Collect multi-source heterogeneous geological data of coal mines, and obtain standardized multi-source data with credibility weights through data quality classification and credibility weight quantification modeling.

[0006] S2. Based on the standardized multi-source data in step S1, spatial feature alignment, confidence-weighted conflict resolution, and weighted null value completion are used to obtain a conflict-free and void-free fused gridded geological data volume.

[0007] S3. Based on the fused gridded geological data volume of step S2, a dynamic variation function is constructed by introducing time and mining progress factors. The grid attribute values ​​are calculated by improving dynamic kriging interpolation, and key geological interfaces are extracted to obtain the initial three-dimensional geological model.

[0008] S4. Based on the initial three-dimensional geological model in step S3 and the precision control constraint data obtained during the real-time mining process, the deviation coefficient between the model prediction value and the real-time measured value is quantitatively calculated. The key parameters of the variogram function are dynamically corrected according to the deviation coefficient. The deviation threshold is set to trigger the local update of the model. The local grid attribute values ​​are optimized through iterative interpolation. Finally, a local high-precision three-dimensional geological model adapted to the real-time mining conditions is reconstructed.

[0009] S5. Based on the local high-precision updated model from step S4 and the measured data not involved in the modeling, perform multi-dimensional accuracy verification. Through the engineering requirement compliance judgment, output a high-precision three-dimensional geological final model of the coal mine that meets the accuracy requirements.

[0010] Preferably, the multi-source heterogeneous geological data of the coal mine in step S1 Including coal mine borehole data Geophysical data Well logging data and mining dynamic data Among them, coal mine borehole data Including lithological categories Rock layer thickness and spatial coordinates Geophysical data Including seismic wave velocity and resistivity Well logging data Including gamma value and rock layer density Mining dynamic data Including the goaf area and top slab settlement The original coordinates of all data are unified to a coordinate system using coordinate system transformation techniques. coordinate system Then, it is converted through data format standardization technology. Format.

[0011] Preferably, in step S1, the Median Absolute Deviation (MAD) algorithm is used to analyze the multi-source heterogeneous geological data of the coal mine. Noise filtering is applied to continuous parameters to identify outliers that exceed reasonable ranges. And outliers were identified by using a spatial distance-weighted average method. To ensure that the corrected value conforms to local geological patterns, the calculation formula is as follows:

[0012] ;

[0013] in, These are the corrected outliers; outlier The surrounding One valid data point; The spatial distance weights are used, and , for and The straight-line distance in space; For smoothing coefficients; The number of adjacent valid data points involved in the correction;

[0014] Preferably, the specific steps of the analytic hierarchy process in step S1 are as follows:

[0015] S101. Construct a three-dimensional evaluation system including data acquisition accuracy, timeliness, and completeness, and determine the weight of each indicator. , , And satisfy In order to achieve a comprehensive quantification of data quality;

[0016] S102. Quantify the indicators in the three-dimensional evaluation system in step S101, assign values ​​based on the inherent precision of the data acquisition method, and quantify the acquisition precision factor. ,in, For drilling data, For geophysical data, For well logging data, For dynamic data of mining;

[0017] By combining the coal mine mining progress, the timeliness of data is dynamically quantified to obtain a timeliness factor. The formula for calculating the progress of coal mining is as follows: , For the first Data collection timestamps Set the start timestamp for 3D geological modeling;

[0018] Based on the proportion of effective data, reflecting the quality of data coverage, the completeness factor is obtained. The formula for the percentage of valid data is: , For the first The number of valid data records in the class data. For the first Total number of data entries collected for each category;

[0019] S103. Based on the three indicators in step S102, perform weighted fusion to obtain the first... Credibility weight of class data The calculation formula is:

[0020] ;

[0021] Obtain standardized multi-source data with confidence weights. ,in, This is valid data after standardization and outlier correction.

[0022] Preferably, the specific steps of step S2 are as follows:

[0023] S21. Based on the data repaired in step S103 Using spatial affine transformation techniques, with data Drilling data Using spatial coordinates as a reference, the original coordinates of geophysical data, well logging data, and mining dynamic data are translated, rotated, and scaled to ensure that all data points are in the same spatial coordinate system. In the coordinate system, spatially aligned multi-source datasets are obtained. ;

[0024] S22, Based on the credibility weight in step S103 Construct a weighted fusion model and calculate the fused attribute values. The calculation formula is:

[0025] ;

[0026] in, For the first The original attribute values ​​of the class data;

[0027] Obtain the fused feature dataset ;

[0028] S23. Adaptive mesh generation technology for coal mine geological bodies is adopted to divide a three-dimensional mesh according to the complexity of the geological structure of the mining area and set the mesh size. Generate a 3D mesh model of the mining area. ,in This represents the total number of nodes in the grid.

[0029] For null value regions in the grid nodes that are not covered by data, based on step S22 For valid data, use the confidence-weighted inverse distance interpolation method to fill in missing values ​​and calculate the missing value nodes. attribute values The calculation formula is:

[0030] ;

[0031] in, Null node The number of valid data points in the surrounding area; For the first The credibility weight of the data type corresponding to each valid data point; For the first The fusion attribute value of each valid data point; Null node With valid data points The straight-line distance in space;

[0032] Obtain integrated gridded geological data volume .

[0033] Preferably, the specific steps of step S3 are as follows:

[0034] S31. Based on step S23, the fused gridded geological data volume Using the theory of variogram in geostatistics, initial values ​​of the basic parameters of the variogram, including the nugget value, were determined. , base value and initial range Construct a variogram model, and the calculation formula is as follows:

[0035] ;

[0036] in, for The spatial distance at time is The value of the variogram at that location; for The dynamic range of time, and , For the initial range change, This is the impact coefficient of mining; This is the current mining time point. Design the total mining cycle for the mining area;

[0037] Using the least squares method to analyze the dynamic variogram By fitting the data, a dynamic variogram model adapted to the dynamics of coal mine mining is obtained. ;

[0038] S32. Dynamic mutation function based on step S31 Construct a system of Kriging equations and solve for the dynamic weighting coefficients of the interpolation points. The calculation formula is:

[0039] ;

[0040] in, for Interpolation point at time Surrounding Weighting coefficients for each known data point; The number of known data points surrounding the point to be interpolated; For the first The and the first Spatial distance between known data points; for time The corresponding variogram value; for The Lagrange multiplier at time; For the point to be interpolated and the first Spatial distance between known data points; for time The corresponding variogram value;

[0041] S33, Dynamic weighting coefficients based on step S32 ,calculate The geological attribute value of the point to be interpolated at time t is calculated using the following formula:

[0042] ;

[0043] in, for Interpolation point at time The interpolation results; for The Middle Geological attribute values ​​of known data points;

[0044] To obtain spatiotemporally adapted interpolated gridded data volumes ;

[0045] S34. Interpolated gridded data volume based on step S33 By extracting geological interfaces and modeling 3D entities, an initial 3D geological model containing key geological structures is obtained. .

[0046] Preferably, the specific steps of step S4 are as follows:

[0047] S41. Real-time acquisition of constraint data that directly reflects the core accuracy of the geological model during underground coal mining, including measured values ​​of coal seam roof elevation. Measured values ​​of coal seam floor elevation Measured values ​​of coal seam thickness Constructing the deviation coefficient of multi-parameter fusion The formula for calculating the degree of difference between the model's predicted values ​​and actual geological conditions is as follows:

[0048] ;

[0049] in, The deviation coefficient between the time-of-flight model prediction and the real-time measurement; , for Predicted elevations of the coal seam roof and floor at the location corresponding to the initial model at a given time; This represents the maximum fluctuation range of the elevation of the top and bottom plates of the coal seam in the mining area. for The predicted coal seam thickness at the location corresponding to the initial model at time; Maximum designed thickness of coal seams in the mining area; , These are the weighting coefficients for each constraint parameter;

[0050] S42, Deviation coefficient based on step S41 The dynamic mutation function of step S31 Variable range To improve the interpolation accuracy in the deviation area, corrections are made using the following formula:

[0051] ;

[0052] ;

[0053] in, This is the corrected dynamic range; This is the deviation influence coefficient; A perturbation-adaptive variogram to fit the current level of mining disturbance;

[0054] S43. Perturbation-fitting variogram based on step S42 Construct the Kriging weight coefficients to simultaneously obtain the corrected weight coefficients. and the corrected Lagrange multiplier The formula is:

[0055] ;

[0056] Set the update trigger condition, i.e., the deviation threshold. , and when At that time, the attribute values ​​of the mesh nodes in the deviation area are iteratively updated using the following formula:

[0057] ;

[0058] in, For the updated perturbation region mesh nodes Geological grid attribute values;

[0059] S44. Geological grid attribute values ​​based on step S43 The triangular patch topology update technique is used to reconstruct the 3D solid model only in the deviation area, and the optimized attribute values ​​are mapped to the initial 3D geological model in step S34. This is to complete the reconstruction of the solid model of the disturbed area and obtain an updated three-dimensional geological model adapted to the dynamic changes in mining. .

[0060] Preferably, the specific steps of step S5 are as follows:

[0061] S51. The updated 3D geological model based on step S44 Simultaneously, borehole data and tunnel geological data that were not used in the modeling of the mining area were selected to form a verification sample set. ,in Including measured spatial coordinates and geological attribute values ,match Corresponding to spatial location and extracting model predicted attribute values ;

[0062] S52. Construct a comprehensive accuracy evaluation model that integrates absolute error, relative error, and correlation to fully quantify the model's engineering applicability. The calculation formula is as follows:

[0063] ;

[0064] in, To improve the overall accuracy of the model; To verify the sample size; for In the Predicted attribute values ​​for each sample location; For the first The actual measured attribute values ​​of each sample; This represents the measured maximum value of the corresponding geological attribute; , These are the weighting coefficients; The Pearson correlation coefficient between the predicted and measured values ​​is given by the formula: ;

[0065] S53, Set engineering requirement thresholds , and when If the model accuracy meets the engineering requirements, return to step S44 to update again, expand the local update range and iterate and optimize again until the accuracy meets the standard.

[0066] S54. Set the conditions for judging the compliance of engineering requirements, including the model integration accuracy in step S53. And the fluctuation range of attribute values ​​in the locally updated area does not exceed the allowable error of the engineering design. If all the above conditions are met, a high-precision three-dimensional geological model of the coal mine will be output. .

[0067] Therefore, the above-mentioned high-precision three-dimensional geological modeling method for coal mines based on multi-source data fusion and updating has the following beneficial effects:

[0068] 1. By performing quality grading and credibility weight quantification modeling on multi-source heterogeneous geological data from coal mines, the problem of low accuracy caused by neglecting data reliability differences and direct fusion in existing technologies is effectively solved. This allows the standardized data to highlight the reference value of high-quality data and weaken the interference of low-quality data, providing a reliable data foundation for subsequent data fusion and modeling. At the same time, combined with credibility weighted conflict resolution and missing value completion strategies, the contradictions and missing values ​​among multi-source data are further eliminated, forming a conflict-free and void-free fused gridded geological data body. This significantly improves the completeness and consistency of subsequent modeling data and lays the data foundation for accurate modeling of core geological parameters.

[0069] 2. This invention overcomes the technical limitations of traditional static variograms, which cannot adapt to changes in mining timelines. By introducing a mining time factor to construct a dynamic variogram, and optimizing the solution process for Kriging interpolation weight coefficients based on this model, the interpolation calculation can respond in real time to the spatiotemporal variability of geological attributes. The constructed initial three-dimensional geological model not only includes key geological structural features but also accurately matches the temporal process of coal mining. This solves the problem that static models cannot reflect the dynamic changes of geological bodies, laying a foundation model that combines timeliness and accuracy for subsequent local high-precision updates based on real-time constraint data.

[0070] 3. By integrating real-time constraint data such as the elevation of the coal seam roof and floor and the thickness of the coal seam, a quantified deviation coefficient is constructed. Based on this coefficient, the range parameter of the variogram is dynamically corrected, triggering local iterative interpolation updates of the model. This achieves a linkage response between the initial 3D geological model and the actual geological conditions of real-time mining, effectively capturing the deviation between the model prediction and the actual geological state. It solves the problem that the existing model is disconnected from mining practice after construction and cannot accurately reflect the true state of core geological parameters. The updated model can focus on key areas and continuously fit the actual geological changes during the mining process, significantly improving the model's local high-precision adaptability and engineering guidance value.

[0071] 4. A multi-dimensional accuracy verification method integrating absolute error, relative error, and correlation indicators is adopted, which is more comprehensive and objective than the existing single-dimensional accuracy evaluation. It can accurately characterize the prediction accuracy and engineering applicability of the model's core geological parameters and effectively avoid the one-sidedness of single accuracy indicator evaluation. By setting engineering allowable error constraints and accuracy thresholds, a complete model accuracy verification closed loop is formed to ensure that the final output model can meet the high accuracy requirements of coal mining engineering for core geological parameters and fully adapt to the technical needs of actual mining operations.

[0072] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0073] Figure 1The flowchart of the high-precision three-dimensional geological modeling method for coal mines based on multi-source data fusion and updating provided by the present invention. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the embodiments of the present invention and are not intended to limit the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of this application. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout.

[0075] It should be noted that the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion, such as a process, method, system, product, or server that includes a series of steps or units, not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such process, method, product, or device.

[0076] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0077] Current coal mine 3D geological modeling and related technologies generally suffer from problems such as data processing failing to consider the reliability differences of multi-source heterogeneous data, direct fusion being susceptible to interference from low-quality data, and the use of variograms in the modeling process being mostly static parameter models that cannot adapt to the spatiotemporal dynamic changes of geological attributes under mining timelines. Furthermore, they lack a linkage and update mechanism with real-time mining disturbances such as goaf expansion and changes in surrounding rock stress, making it difficult for the constructed model to reflect the true dynamics of the geological body. Consequently, they fail to meet the actual needs of coal mining engineering for high-precision, dynamic, safe, and controllable models.

[0078] Based on the above analysis, this invention is designed. (See appendix.) Figure 1 A high-precision 3D geological modeling method for coal mines based on multi-source data fusion and updating includes the following steps:

[0079] S1. Collect multi-source heterogeneous geological data of coal mines, and obtain standardized multi-source data with credibility weights through data quality classification and credibility weight quantification modeling.

[0080] Multi-source heterogeneous geological data of coal mines in step S1 Including coal mine borehole data Geophysical data Well logging data and mining dynamic data Among them, coal mine borehole data Including lithological categories Rock layer thickness and spatial coordinates Geophysical data Including seismic wave velocity and resistivity Well logging data Including gamma value and rock layer density Mining dynamic data Including the goaf area and top slab settlement The original coordinates of all data are unified to a coordinate system using coordinate system transformation techniques. coordinate system Then, it is converted through data format standardization technology. Format;

[0081] In step S1, the MAD algorithm is used to analyze the multi-source heterogeneous geological data of the coal mine. Noise filtering is applied to continuous parameters to identify outliers that exceed reasonable ranges. The formula for calculating the data series is:

[0082] ;

[0083] in, It can be represented as , , , and Continuous parameters;

[0084] median The calculation formula is:

[0085] ;

[0086] in, For data sequences Length;

[0087] Calculate the absolute deviation of the median The formula is:

[0088] ;

[0089] in, For data sequences The One value; For the first The absolute deviation of each value from the median; It reflects the dispersion of the data sequence and is not affected by outliers;

[0090] Determine the threshold for outlier detection The formula is:

[0091] ;

[0092] in, To determine the upper limit for outliers, when or At that time, the judgment This is an outlier;

[0093] Outliers were identified by using a spatial distance-weighted average method. To ensure that the corrected value conforms to local geological patterns, the calculation formula is as follows:

[0094] ;

[0095] in, These are the corrected outliers; outlier The surrounding One valid data point; The spatial distance weights are used, and , for and The straight-line distance in space; For smoothing coefficients; The number of adjacent valid data points involved in the correction;

[0096] The specific steps of the analytic hierarchy process in step S1 are as follows:

[0097] S101. Construct a three-dimensional evaluation system including data acquisition accuracy, timeliness, and completeness, and determine the weight of each indicator. , , And satisfy To achieve comprehensive quantification of data quality, specifically... , , ;

[0098] S102. Quantify the indicators in the three-dimensional evaluation system in step S101, assign values ​​based on the inherent precision of the data acquisition method, and quantify the acquisition precision factor. ,in, For drilling data, For geophysical data, For well logging data, For dynamic mining data, specifically, direct sampling of borehole data offers the highest accuracy. Indirect detection using geophysical data, with moderate accuracy. The well logging data is obtained through semi-direct detection, with moderate to low accuracy. Real-time monitoring of mining dynamics data is susceptible to sensor-related accuracy issues. ;

[0099] By combining the coal mine mining progress, the timeliness of data is dynamically quantified to obtain a timeliness factor. The formula is:

[0100] ;

[0101] The formula for calculating the progress of coal mining is as follows: , For the first Data collection timestamps Set the start timestamp for 3D geological modeling;

[0102] Based on the proportion of effective data, reflecting the quality of data coverage, the completeness factor is obtained. ,

[0103] ;

[0104] The formula for the percentage of valid data is as follows: , For the first The number of valid data records in the class data. For the first Total number of data entries collected for each category;

[0105] S103. Based on the three indicators in step S102, perform weighted fusion to obtain the first... Credibility weight of class data The calculation formula is:

[0106] ;

[0107] Obtain standardized multi-source data with confidence weights. ,in, This is valid data after standardization and outlier correction.

[0108] S2. Based on the standardized multi-source data in step S1, spatial feature alignment, confidence-weighted conflict resolution, and weighted null value completion are used to obtain a conflict-free and void-free fused gridded geological data volume.

[0109] The specific steps of step S2 are as follows:

[0110] S21. Based on the data repaired in step S103 Using spatial affine transformation techniques, with data Drilling data Using spatial coordinates as a reference, the original coordinates of geophysical data, well logging data, and mining dynamic data are translated, rotated, and scaled to ensure that all data points are in the same spatial coordinate system. In the coordinate system, spatially aligned multi-source datasets are obtained. ;

[0111] The transformation formula is:

[0112] ;

[0113] in, , , The original coordinates for geophysical exploration / well logging / mining dynamic data; , , The target coordinates are to be unified to the WGS-84 coordinate system after transformation; This is a uniform scaling factor; , , It is a translation vector; It is a 3×3 rotation matrix, and the spatial offset of coal mine data is mainly translation with a very small rotation angle. Therefore... For small angles and approximately simplified to , , , These correspond to the rotation angles around the X, Y, and Z axes, respectively.

[0114] S22, Based on the credibility weight in step S103 Construct a weighted fusion model and calculate the fused attribute values. The calculation formula is:

[0115] ;

[0116] in, For the first The original attribute values ​​of the class data;

[0117] Obtain the fused feature dataset ;

[0118] S23. Adaptive mesh generation technology for coal mine geological bodies is adopted to divide a three-dimensional mesh according to the complexity of the geological structure of the mining area and set the mesh size. Generate a 3D mesh model of the mining area. ,in This represents the total number of nodes in the grid.

[0119] For null value regions in the grid nodes that are not covered by data, based on step S22 For valid data, use the confidence-weighted inverse distance interpolation method to fill in missing values ​​and calculate the missing value nodes. attribute values The calculation formula is:

[0120] ;

[0121] in, Null node The number of valid data points in the surrounding area; For the first The credibility weight of the data type corresponding to each valid data point; For the first The fusion attribute value of each valid data point; Null node With valid data points The straight-line distance in space;

[0122] Obtain integrated gridded geological data volume .

[0123] S3. Based on the fused gridded geological data volume of step S2, a dynamic variation function is constructed by introducing time and mining progress factors. The grid attribute values ​​are calculated by improving dynamic kriging interpolation, and key geological interfaces are extracted to obtain the initial three-dimensional geological model.

[0124] The specific steps of step S3 are as follows:

[0125] S31. Based on step S23, the fused gridded geological data volume Using the theory of variogram in geostatistics, initial values ​​of the basic parameters of the variogram, including the nugget value, were determined. , base value and initial range , for The variogram model is constructed using 1 / 5 of the data standard deviation, and the calculation formula is as follows:

[0126] ;

[0127] in, for The spatial distance at time is The value of the variogram at that location; for The dynamic range of time, and , For the initial range change, This is the impact coefficient of mining; This is the current mining time point. Design the total mining cycle for the mining area;

[0128] Using the least squares method to analyze the dynamic variogram By fitting the data, a dynamic variogram model adapted to the dynamics of coal mine mining is obtained. Specifically, the process involves: first, clarifying the core structure of the dynamic variogram, including the basic term of spatial variation of geological attributes and the dynamic range design combined with mining time and total cycle; determining the initial values ​​of key parameters based on geostatistics theory and the characteristics of the integrated gridded geological data body; then, selecting variogram observation data from the integrated gridded geological data body with different mining time nodes and different spatial distances, and after outlier correction and standardization; constructing the objective function with the goal of "minimizing the sum of squared residuals between the observed data and the model prediction values"; taking the partial derivatives of this function with respect to each key parameter and setting the partial derivatives to zero to form a normal equation system; solving for the optimal estimates of each parameter; and finally, substituting the optimal parameters into the initial dynamic variogram structure to generate the final model adapted to the dynamics of coal mining.

[0129] S32. Dynamic mutation function based on step S31 Construct a system of Kriging equations and solve for the dynamic weighting coefficients of the interpolation points. The calculation formula is:

[0130] ;

[0131] in, for Interpolation point at time Surrounding Weighting coefficients for each known data point; The number of known data points surrounding the point to be interpolated; For the first The and the first Spatial distance between known data points; for time The corresponding variogram value; for The Lagrange multiplier at time; For the point to be interpolated and the first Spatial distance between known data points; for time The corresponding variogram value;

[0132] S33, Dynamic weighting coefficients based on step S32 ,calculate The geological attribute value of the point to be interpolated at time t is calculated using the following formula:

[0133] ;

[0134] in, for Interpolation point at time The interpolation results; for The Middle Geological attribute values ​​of known data points;

[0135] To obtain spatiotemporally adapted interpolated gridded data volumes ;

[0136] S34. Interpolated gridded data volume based on step S33 By extracting geological interfaces and modeling 3D entities, an initial 3D geological model containing key geological structures is obtained. .

[0137] S4. Based on the initial three-dimensional geological model in step S3 and the precision control constraint data obtained during the real-time mining process, the deviation coefficient between the model prediction value and the real-time measured value is quantitatively calculated. The key parameters of the variogram function are dynamically corrected according to the deviation coefficient. The deviation threshold is set to trigger the local update of the model. The local grid attribute values ​​are optimized through iterative interpolation. Finally, a local high-precision three-dimensional geological model adapted to the real-time mining conditions is reconstructed.

[0138] The specific steps of step S4 are as follows:

[0139] S41. Real-time acquisition of constraint data that directly reflects the core accuracy of the geological model during underground coal mining, including measured values ​​of coal seam roof elevation. Measured values ​​of coal seam floor elevation Measured values ​​of coal seam thickness Constructing the deviation coefficient of multi-parameter fusion The formula for calculating the degree of difference between the model's predicted values ​​and actual geological conditions is as follows:

[0140] ;

[0141] in, The deviation coefficient between the time-of-flight model prediction and the real-time measurement; , for Predicted elevations of the coal seam roof and floor at the location corresponding to the initial model at a given time; This represents the maximum fluctuation range of the elevation of the top and bottom plates of the coal seam in the mining area. for The predicted coal seam thickness at the location corresponding to the initial model at time; Maximum designed thickness of coal seams in the mining area; , These are the weighting coefficients for each constraint parameter;

[0142] S42, Deviation coefficient based on step S41 The dynamic mutation function of step S31 Variable range To improve the interpolation accuracy in the deviation area, corrections are made using the following formula:

[0143] ;

[0144] ;

[0145] in, This is the corrected dynamic range; This is the deviation influence coefficient, with a value of 0.8, used to adjust the strength of the deviation's correction to the range, and to avoid excessive compression; A perturbation-adaptive variogram to fit the current level of mining disturbance;

[0146] S43. Perturbation-fitting variogram based on step S42 Construct the Kriging weight coefficients to simultaneously obtain the corrected weight coefficients. and the corrected Lagrange multiplier The formula is:

[0147] ;

[0148] Set the update trigger condition, i.e., the deviation threshold. , and when At that time, the attribute values ​​of the mesh nodes in the deviation area are iteratively updated using the following formula:

[0149] ;

[0150] in, For the updated perturbation region mesh nodes Geological grid attribute values;

[0151] S44. Geological grid attribute values ​​based on step S43 The triangular patch topology update technique is used to reconstruct the 3D solid model only in the deviation area, and the optimized attribute values ​​are mapped to the initial 3D geological model in step S34. This is to complete the reconstruction of the solid model of the disturbed area and obtain an updated three-dimensional geological model adapted to the dynamic changes in mining. .

[0152] S5. Based on the local high-precision updated model in step S4 and the measured data that were not involved in the modeling, perform multi-dimensional accuracy verification, and output a high-precision three-dimensional geological final model of the coal mine that meets the accuracy requirements by judging the compliance with engineering requirements.

[0153] The specific steps of step S5 are as follows:

[0154] S51. The updated 3D geological model based on step S44 Simultaneously, borehole data and tunnel geological data that were not used in the modeling of the mining area were selected to form a verification sample set. ,in Including measured spatial coordinates and geological attribute values ,match Corresponding to spatial location and extracting model predicted attribute values ;

[0155] S52. Construct a comprehensive accuracy evaluation model that integrates absolute error, relative error, and correlation to fully quantify the model's engineering applicability. The calculation formula is as follows:

[0156] ;

[0157] in, To improve the overall accuracy of the model; To verify the sample size; for In the Predicted attribute values ​​for each sample location; For the first The actual measured attribute values ​​of each sample; This represents the measured maximum value of the corresponding geological attribute; , These are the weighting coefficients; The Pearson correlation coefficient between the predicted and measured values ​​is given by the formula: ;

[0158] S53, Determine the threshold requirements for the project , and when If the model accuracy meets the engineering requirements, return to step S44 to update again, expand the local update range and iterate and optimize again until the accuracy meets the standard.

[0159] S54. Set the conditions for judging the compliance of engineering requirements, including the model integration accuracy in step S53. If the attribute value fluctuation range of the locally updated region does not exceed the allowable error of the engineering design, and all of the above conditions are met, then a high-precision three-dimensional geological model of the coal mine will be output. ;

[0160] Example: A coal mine mining area, total mining cycle In 2018, the preset engineering parameters were: maximum fluctuation range of coal seam roof and floor elevations of 50m, maximum design thickness of coal seam of 5m, and deviation coefficient weighting. It is 0.6. The threshold is 0.4, the bias influence coefficient is 0.8, the update trigger threshold is 0.2, and the model accuracy threshold is 0.85.

[0161] Four types of data were collected from this mining area, including the number of boreholes (30) and their corresponding lithology, stratum thickness (2-3.5m), and spatial coordinates; and geophysical data including seismic wave velocity (2800-3200m / s) and resistivity (50-80). The well logging data showed gamma values ​​of 60-100 API and a rock density of 2.5-2.8. The mining dynamic data consisted of the measured elevations of the roof and floor of the coal seam corresponding to the mined-out area: roof 522-518m, floor 498-494m, and measured coal seam thickness 2.2-3.3m. This data was uniformly converted to the WGS-84 coordinate system and CSV format. The MAD algorithm was used to identify two rock layer density anomalies, including: 3.5... and 2.0 The spatial distance weighted average of the surrounding 5 valid data points is adjusted to 2.7. and 2.6 The analytic hierarchy process (AHP) was used to calculate the confidence weights, with borehole data having a weight of 0.85, geophysical data 0.72, well logging data 0.68, and mining dynamics data 0.90.

[0162] Using the spatial coordinates of borehole data as a reference, affine transformations are performed on other data to achieve spatial alignment; a weighted fusion model is constructed based on credibility weights to resolve the conflict between geophysical and logging density data under the same coordinates; an adaptive grid partitioning is used to generate a three-dimensional grid, which is then completed by credibility-weighted inverse distance interpolation of surrounding valid data to obtain a conflict-free and void-free fused gridded geological data volume;

[0163] The initial value of the variogram was determined based on the fused data. A dynamic variogram was constructed by introducing the mining time factor. The least squares method was used to fit the model to obtain the fit model. The attribute values ​​of each grid were calculated by improving the dynamic kriging interpolation. Key geological interfaces such as the top and bottom plates of the coal seam and faults were extracted to generate an initial three-dimensional geological model. The model predicts that the elevation of the coal seam top plate is 521-517m, the bottom plate is 497-493m, and the thickness is 2.3-3.2m.

[0164] Real-time acquisition of core precision control data from underground mining shows that the measured values ​​of the current coal seam roof elevation are 519.6m, the measured values ​​of the current floor elevation are 495.3m, and the measured values ​​of the current coal seam thickness are 2.43m. The predicted values ​​for the corresponding positions in the initial model are 520.5m, 494.7m, and 2.58m, respectively.

[0165] Calculated using the deviation coefficient formula ,because This triggers a local model update, correcting the dynamic range. The local high-precision variogram function is reconstructed, and the grid attribute values ​​corresponding to the top and bottom plates and thickness of the coal seam are optimized by iterative interpolation based on the new variogram function to obtain a local high-precision update model that adapts to the actual geological conditions of the current mining.

[0166] Ten borehole data points not involved in modeling were selected as validation samples. Model-predicted attribute values ​​(coal seam roof and floor elevation, thickness, rock density, etc.) and measured values ​​were extracted for each sample location. These were then calculated using a multi-dimensional accuracy formula to obtain the overall accuracy. Furthermore, the predicted error of coal seam thickness in the local update area is ≤ ±0.2m, and the error of roof and floor elevation is ≤ ±0.5m, which meets the allowable error constraints of the project. The output is a three-dimensional geological final model that meets the high-precision requirements of coal mine mining engineering, providing accurate geological data support for the roadway layout and mining parameter design of subsequent mining operations.

[0167] In summary, this invention addresses the issues of large reliability differences and interference in modeling caused by collecting multi-source heterogeneous geological data and performing quality grading, reliability quantification, and outlier correction. Through spatial alignment, weighted conflict resolution, and null value completion, data conflicts and gaps are eliminated, ensuring the integrity and consistency of the fused data. By introducing time and mining progress factors to construct a dynamic variogram and optimizing Kriging interpolation, the limitations of static models in adapting to the spatiotemporal dynamic changes of geological attributes are overcome. By integrating real-time mining disturbance data to quantify disturbance factors and achieve dynamic model updates, the problem of model disconnect from mining practice is solved. Through multi-dimensional accuracy verification and multi-factor weighted risk quantification assessment, the problems of one-sided model accuracy evaluation and ambiguous risk assessment are resolved. Ultimately, a high-precision, dynamic, and safe three-dimensional geological model for coal mines is formed, fully meeting the practical needs of coal mining engineering for model reliability, timeliness, and safety guidance.

[0168] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A coal mine high-precision three-dimensional geological modeling method based on multi-source data fusion and updating, characterized in that: Includes the following steps: S1. Collect multi-source heterogeneous geological data of coal mines, and obtain standardized multi-source data with credibility weights through data quality classification and credibility weight quantification modeling. S2. Based on the standardized multi-source data in step S1, spatial feature alignment, confidence-weighted conflict resolution, and weighted null value completion are used to obtain a conflict-free and void-free fused gridded geological data volume. S3. Based on the fused gridded geological data volume of step S2, a dynamic variation function is constructed by introducing time and mining progress factors. The grid attribute values ​​are calculated by improving dynamic kriging interpolation, and key geological interfaces are extracted to obtain the initial three-dimensional geological model. S4. Based on the initial three-dimensional geological model in step S3 and the precision control constraint data obtained during the real-time mining process, the deviation coefficient between the model prediction value and the real-time measured value is quantitatively calculated. The key parameters of the variogram function are dynamically corrected according to the deviation coefficient. The deviation threshold is set to trigger the local update of the model. The local grid attribute values ​​are optimized through iterative interpolation. Finally, a local high-precision three-dimensional geological model adapted to the real-time mining conditions is reconstructed. S5. Based on the local high-precision updated model from step S4 and the measured data not involved in the modeling, perform multi-dimensional accuracy verification. Through the engineering requirement compliance judgment, output a high-precision three-dimensional geological final model of the coal mine that meets the accuracy requirements.

2. The method for high-precision three-dimensional geological modeling of coal mines based on multi-source data fusion and updating according to claim 1, characterized in that: Multi-source heterogeneous geological data of coal mines in step S1 Including coal mine borehole data Geophysical data Well logging data and mining dynamic data Among them, coal mine borehole data Including lithological categories Rock layer thickness and spatial coordinates Geophysical data Including seismic wave velocity and resistivity Well logging data Including gamma value and rock layer density Mining dynamic data Including the goaf area and top slab settlement The original coordinates of all data are unified to a coordinate system using coordinate system transformation techniques. coordinate system Then, it is converted through data format standardization technology. Format.

3. The method for high-precision three-dimensional geological modeling of coal mines based on multi-source data fusion and updating according to claim 2, characterized in that: In step S1, the Median Absolute Deviation (MAD) algorithm is used to analyze the multi-source heterogeneous geological data of the coal mine. Noise filtering is applied to continuous parameters to identify outliers that exceed reasonable ranges. And outliers were identified by using a spatial distance-weighted average method. To ensure that the corrected value conforms to local geological patterns, the calculation formula is as follows: ； in, These are corrected outliers; outlier The surrounding One valid data point; The spatial distance weights are used, and , for and The straight-line distance in space; For smoothing coefficients; This represents the number of adjacent valid data points that participate in the correction.

4. The method for high-precision three-dimensional geological modeling of coal mines based on multi-source data fusion and updating according to claim 3, characterized in that: The specific steps of the analytic hierarchy process in step S1 are as follows: S101. Construct a three-dimensional evaluation system including data acquisition accuracy, timeliness, and completeness, and determine the weight of each indicator. 、 、 And satisfy In order to achieve a comprehensive quantification of data quality; S102. Quantify the indicators in the three-dimensional evaluation system in step S101, assign values ​​based on the inherent precision of the data acquisition method, and quantify the acquisition precision factor. ,in, For drilling data, For geophysical data, For well logging data, For dynamic data of mining; By combining the coal mine mining progress, the timeliness of data is dynamically quantified to obtain a timeliness factor. ; Based on the proportion of effective data, reflecting the quality of data coverage, the completeness factor is obtained. ; S103. Based on the three indicators in step S102, perform weighted fusion to obtain the first... Credibility weight of class data The calculation formula is: ； Obtain standardized multi-source data with confidence weights. ,in, This is valid data after standardization and outlier correction.

5. The method for high-precision three-dimensional geological modeling of coal mines based on multi-source data fusion and updating according to claim 4, characterized in that: The specific steps of step S2 are as follows: S21. Based on the data repaired in step S103 Using spatial affine transformation techniques, with data Drilling data Using spatial coordinates as a reference, the original coordinates of geophysical data, well logging data, and mining dynamic data are translated, rotated, and scaled to ensure that all data points are in the same spatial coordinate system. In the coordinate system, the spatially aligned multi-source dataset is obtained. ; S22, Based on the credibility weight in step S103 Construct a weighted fusion model and calculate the fused attribute values. The calculation formula is: ； in, For the first The original attribute values ​​of the class data; Obtain the fused feature dataset ; S23. Adaptive mesh generation technology for coal mine geological bodies is adopted to divide a three-dimensional mesh according to the complexity of the geological structure of the mining area and set the mesh size. Generate a 3D mesh model of the mining area ,in This represents the total number of nodes in the grid. For null value regions in the grid nodes that are not covered by data, based on step S22 For valid data, use the confidence-weighted inverse distance interpolation method to fill in missing values ​​and calculate the missing value nodes. attribute values The calculation formula is: ； in, Null node The number of valid data points in the surrounding area; For the first The credibility weight of the data type corresponding to each valid data point; For the first The fusion attribute value of each valid data point; Null node With valid data points The straight-line distance in space; Obtain integrated gridded geological data volume .

6. The method for high-precision three-dimensional geological modeling of coal mines based on multi-source data fusion and updating according to claim 5, characterized in that: The specific steps of step S3 are as follows: S31. Based on step S23, the fused gridded geological data volume Using the theory of variogram in geostatistics, initial values ​​of the basic parameters of the variogram, including the nugget value, were determined. , base value and initial range Construct a variogram model, and the calculation formula is as follows: ； in, for The spatial distance at time is The value of the variogram at that location; for The dynamic range of time, and , For the initial range change, This is the impact coefficient of mining; This is the current mining time point. Design the total mining cycle for the mining area; Using the least squares method to analyze the dynamic variogram By fitting the data, a dynamic variogram model adapted to the dynamics of coal mine mining is obtained. ; S32. Dynamic mutation function based on step S31 Construct a system of Kriging equations and solve for the dynamic weighting coefficients of the interpolation points. The calculation formula is: ； in, for Interpolation point at time Surrounding Weighting coefficients for each known data point; The number of known data points surrounding the point to be interpolated; For the first The and the first Spatial distance between known data points; for time The corresponding variogram value; for The Lagrange multiplier at time; For the point to be interpolated and the first Spatial distance between known data points; for time The corresponding variogram value; S33, Dynamic weighting coefficients based on step S32 ,calculate The geological attribute value of the point to be interpolated at time t is calculated using the following formula: ； in, for Interpolation point at time The interpolation results; for The Middle Geological attribute values ​​of known data points; To obtain spatiotemporally adapted interpolated gridded data volumes ; S34. Interpolated gridded data volume based on step S33 By extracting geological interfaces and modeling 3D entities, an initial 3D geological model containing key geological structures is obtained. .

7. The method for high-precision three-dimensional geological modeling of coal mines based on multi-source data fusion and updating according to claim 6, characterized in that: The specific steps of step S4 are as follows: S41. Real-time acquisition of constraint data that directly reflects the core accuracy of the geological model during underground coal mining, including measured values ​​of coal seam roof elevation. Measured values ​​of coal seam floor elevation Measured values ​​of coal seam thickness Constructing the deviation coefficient of multi-parameter fusion The formula for calculating the degree of difference between the model's predicted values ​​and actual geological conditions is as follows: ； in, for The deviation coefficient between the time-of-time model prediction and the real-time measurement; , for Predicted elevations of the coal seam roof and floor at the location corresponding to the initial model at a given time; This represents the maximum fluctuation range of the elevation of the top and bottom plates of the coal seam in the mining area. for The predicted coal seam thickness at the location corresponding to the initial model at that time; This represents the maximum design thickness of the coal seam in the mining area. , These are the weighting coefficients for each constraint parameter; S42, Deviation coefficient based on step S41 The dynamic mutation function of step S31 Variable range To improve the interpolation accuracy in the deviation area, corrections are made using the following formula: ； ； in, This is the corrected dynamic range; This is the deviation influence coefficient; A perturbation-adaptive variogram to fit the current level of mining disturbance; S43. Perturbation-fitting variogram based on step S42 Construct the Kriging weight coefficients to simultaneously obtain the corrected weight coefficients. and the corrected Lagrange multiplier The formula is: ； Set the update trigger condition, i.e., the deviation threshold. , and when At that time, the attribute values ​​of the mesh nodes in the deviation area are iteratively updated using the following formula: ； in, For the updated perturbation region mesh nodes Geological grid attribute values; S44. Geological grid attribute values ​​based on step S43 The triangular patch topology update technique is used to reconstruct the 3D solid model only in the deviation area, and the optimized attribute values ​​are mapped to the initial 3D geological model in step S34. This is to complete the reconstruction of the solid model of the disturbed area and obtain an updated three-dimensional geological model adapted to the dynamic changes in mining. .

8. The method for high-precision three-dimensional geological modeling of coal mines based on multi-source data fusion and updating according to claim 7, characterized in that: The specific steps of step S5 are as follows: S51. The updated 3D geological model based on step S44 Simultaneously, borehole data and tunnel geological data that were not used in the modeling of the mining area were selected to form a verification sample set. ,in Including measured spatial coordinates and geological attribute values ,match Corresponding to spatial location and extracting model predicted attribute values ; S52. Construct a comprehensive accuracy evaluation model that integrates absolute error, relative error, and correlation to fully quantify the model's engineering applicability. The calculation formula is as follows: ； in, To improve the overall accuracy of the model; To verify the sample size; for In the Predicted attribute values ​​for each sample location; For the first The actual measured attribute values ​​of each sample; This represents the measured maximum value of the corresponding geological attribute; , These are the weighting coefficients; The Pearson correlation coefficient between the predicted and measured values ​​is given by the formula: ; S53, Set engineering requirement thresholds , and when If the model accuracy meets the engineering requirements, return to step S44 to update again, expand the local update range and iterate and optimize again until the accuracy meets the standard. S54. Set the conditions for judging the compliance of engineering requirements, including the model integration accuracy in step S53. If the attribute value fluctuation range of the locally updated region does not exceed the allowable error of the engineering design, and all of the above conditions are met, then a high-precision three-dimensional geological model of the coal mine will be output. .

Images