Underground three-dimensional geological modeling method and modeling device under multi-modal data exploration

Abstract

The invention discloses an underground three-dimensional geologic modeling method and device under multi-modal data exploration, and relates to the technical field of geologic modeling, and the method comprises the steps: carrying out the drilling data collection of a target underground region, and obtaining a plurality of drilling data sets; performing partition construction on the initial three-dimensional geological model to obtain an initial three-dimensional geological partition model; performing multi-modal data acquisition according to the plurality of geological features and the plurality of geological division areas to obtain a plurality of multi-modal data sets; and performing data processing on the plurality of multi-modal data sets to obtain a plurality of corrected multi-modal data sets, performing space-time characteristic fusion loss analysis on the initial three-dimensional geological partition model to obtain a fusion model parameter set, and optimizing the initial three-dimensional geological partition model to obtain a target three-dimensional geological partition model. The technical problem that in the prior art, when a complex underground structure is modeled, precision is insufficient is solved, and the technical effect of improving the modeling precision and reliability of the underground geologic structure is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of geological modeling, and in particular to a method and a device for underground three-dimensional geological modeling under multi-modal data survey. Background Art

[0002] With the continuous advancement of resource exploration and underground engineering construction, underground 3D geological modeling technology has been widely used in geological exploration, mineral resource development, underground engineering planning and other fields. However, with the increase in exploration depth and the complexity of underground geological conditions, traditional 3D geological modeling methods are facing increasing challenges. Traditional underground geological modeling usually relies on a single data source, such as drilling data, seismic data, etc., but these data sources often have problems such as insufficient accuracy and one-sided information, making it difficult to fully and accurately reflect complex underground structures. Summary of the invention

[0003] The present application provides a method and a modeling device for underground three-dimensional geological modeling under multimodal data survey, which are used to solve the technical problem of insufficient accuracy in the prior art when modeling complex underground structures.

[0004] In view of the above problems, the present application provides a method and a device for downhole three-dimensional geological modeling under multimodal data survey.

[0005] In a first aspect of the present application, a method for downhole three-dimensional geological modeling under multimodal data survey is provided, the method comprising:

[0006] Performing drilling data acquisition on a preset acquisition drilling point set of a target downhole area to obtain multiple drilling data sets, wherein the multiple drilling data sets include multiple drilling positions; performing initial three-dimensional geological model partitioning construction of the target downhole area based on the multiple drilling data sets and the multiple drilling positions to obtain an initial three-dimensional geological partitioning model, wherein the initial three-dimensional geological partitioning model includes multiple geological partitioning areas, and the multiple geological partitioning areas include multiple geological features; performing multimodal data acquisition based on the multiple geological features and the multiple geological partitioning areas to obtain multiple multimodal data sets, wherein each multimodal data set has a plurality of geological features and a plurality of geological features. The modal data set includes gravity data, magnetic data and electromagnetic data; the multiple multimodal data sets are processed by a data correction module to obtain multiple corrected multimodal data sets, wherein the data correction module includes a correction filter subunit, a format conversion subunit and an interpolation completion subunit; the initial three-dimensional geological partition model is analyzed for spatiotemporal characteristic fusion loss by using the multiple corrected multimodal data sets to obtain a fusion model parameter set; the initial three-dimensional geological partition model is optimized based on the fusion model parameter set and the multiple multimodal data sets to obtain a target three-dimensional geological partition model of the target downhole area.

[0007] A second aspect of the present application provides a downhole three-dimensional geological modeling device under multimodal data survey, the device comprising:

[0008] a drilling data acquisition module, wherein the drilling data acquisition module performs drilling data acquisition on a preset acquisition drilling point set of a target downhole area to obtain a plurality of drilling data sets, wherein the plurality of drilling data sets include a plurality of drilling positions; a partition model construction module, wherein the partition model construction module performs initial three-dimensional geological model partition construction of the target downhole area based on the plurality of drilling data sets and the plurality of drilling positions to obtain an initial three-dimensional geological partition model, wherein the initial three-dimensional geological partition model includes a plurality of geological partition areas, and the plurality of geological partition areas include a plurality of geological features; a multimodal data acquisition module, wherein the multimodal data acquisition module performs multimodal data acquisition according to the plurality of geological features and the plurality of geological partition areas to obtain a plurality of multimodal data sets, wherein each multimodal data set includes a gravity data, magnetic data and electromagnetic data; a data processing module, wherein the data processing module uses a data correction module to perform data processing on the multiple multimodal data sets to obtain multiple corrected multimodal data sets, wherein the data correction module includes a correction filter subunit, a format conversion subunit and an interpolation completion subunit; a fusion loss analysis module, wherein the fusion loss analysis module uses the multiple corrected multimodal data sets to perform spatiotemporal characteristic fusion loss analysis on the initial three-dimensional geological partition model to obtain a multimodal data fusion strategy, wherein the multimodal data fusion strategy includes a fusion model parameter set; a partition model optimization module, wherein the partition model optimization module optimizes the initial three-dimensional geological partition model based on the fusion model parameter set and the multiple multimodal data sets to obtain a target three-dimensional geological partition model of the target downhole area.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] The present application performs borehole data collection on a preset collection borehole point set in a target downhole area to obtain multiple borehole data sets, wherein the multiple borehole data sets include multiple borehole locations; constructs an initial three-dimensional geological model partition of the target downhole area based on the multiple borehole data sets and the multiple borehole locations to obtain an initial three-dimensional geological partition model, wherein the initial three-dimensional geological partition model includes multiple geological partition areas, and the multiple geological partition areas include multiple geological features; performs multimodal data collection based on the multiple geological features and the multiple geological partition areas to obtain multiple multimodal data sets, wherein each multimodal data set includes gravity data, magnetic data and electromagnetic data; performs data processing on the multiple multimodal data sets using a data correction module to obtain multiple corrected multimodal data sets, wherein the data correction module includes a correction filter subunit, a format conversion subunit and an interpolation completion subunit; performs spatiotemporal characteristic fusion loss analysis on the initial three-dimensional geological partition model using the multiple corrected multimodal data sets to obtain a fusion model parameter set; optimizes the initial three-dimensional geological partition model based on the fusion model parameter set and the multiple multimodal data sets to obtain a target three-dimensional geological partition model of the target downhole area. The present invention solves the technical problem of insufficient accuracy in the prior art when modeling complex underground structures. By collecting drilling data of the target downhole area, constructing three-dimensional geological model zoning, collecting and processing multimodal data, and optimizing the model by using time-space characteristic fusion loss analysis, a more accurate target three-dimensional geological zoning model is finally obtained, thereby achieving the technical effect of improving the modeling accuracy and reliability of the downhole geological structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A schematic diagram of a process flow of a downhole three-dimensional geological modeling method under multimodal data survey provided in an embodiment of the present application;

[0013] Figure 2 A schematic diagram of the structure of an underground three-dimensional geological modeling device under multimodal data survey provided in an embodiment of the present application.

[0014] Explanation of the accompanying drawings: drilling data acquisition module 11, partition model construction module 12, multimodal data acquisition module 13, data processing module 14, fusion loss analysis module 15, partition model optimization module 16. DETAILED DESCRIPTION

[0015] The present application provides a method and a modeling device for downhole three-dimensional geological modeling under multimodal data survey, aiming to solve the technical problem of insufficient accuracy in the prior art when modeling complex underground structures. By collecting drilling data of the target downhole area, constructing three-dimensional geological model zoning, collecting and processing multimodal data, and optimizing the model by using time-space characteristic fusion loss analysis, a more accurate target three-dimensional geological zoning model is finally obtained, thereby achieving the technical effect of improving the modeling accuracy and reliability of downhole geological structures.

[0016] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0017] It should be noted that any variations of the terms "include" and "have" are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.

[0018] Embodiment 1, as Figure 1 As shown, the present invention provides a method for downhole three-dimensional geological modeling under multimodal data survey, the method comprising:

[0019] Step S100: performing drilling data collection on a preset collection drilling point set in a target downhole area to obtain a plurality of drilling data sets, wherein the plurality of drilling data sets include a plurality of drilling positions.

[0020] In an embodiment of the present application, drilling data is first collected for the target downhole area according to a determined preset collection drilling point set, and drilling operations are performed at predetermined points by using drilling equipment, and drilling data are collected to obtain multiple drilling data sets. Each drilling data set includes not only the collected geological data, but also the location information of the borehole, such as coordinates, depth, etc.

[0021] Furthermore, in the method provided in the embodiment of the application, drilling data is collected for a preset collection drilling point set in the target downhole area to obtain multiple drilling data sets, and the method further includes:

[0022] Acquire a historical geological survey data set of the target downhole area; analyze the historical geological survey data set according to a preset complexity index set to obtain a target geological complexity index set, wherein the preset complexity index set includes a structural complexity index, a lithological complexity index, a groundwater complexity index and a rock structure complexity index; lay out a drilling network based on the target geological complexity index set to obtain the preset acquisition drilling point set.

[0023] In the embodiment of the present application, a historical geological data set of the target area is first obtained through a geological database or historical exploration report, including previous drilling records, surface geological maps, seismic data, gravity and magnetic data, etc. Next, based on a preset set of complexity indicators, the historical geological data is analyzed in detail. The set of complexity indicators includes structural complexity, lithological complexity, groundwater complexity, and rock mass structure complexity.

[0024] For the structural complexity index, seismic reflection data, fault distribution maps and historical drilling information are used to analyze the number, density, length, dip and strike of faults. The spatial distribution of faults is quantified through GIS tools to evaluate the complexity of fault interlacing, overlap and fracture zones, and a fault complexity distribution map is generated to show areas with high structural complexity. For the lithology complexity index, information on rock thickness, rock type and lithology changes is extracted from historical drilling data. Then, geological modeling software such as Petrel is used to analyze the thickness changes of rock layers and the diversity of lithology, with special attention paid to the transition and alternating areas between different lithologies. The rock thickness change map and lithology distribution map are generated through interpolation, and the complex lithology areas are quantified and marked. For the groundwater complexity index, the number, thickness and distribution of aquifers are analyzed based on borehole data and logging data, and the properties of aquifers are confirmed using resistivity and acoustic logging methods. The flow characteristics of groundwater are simulated through hydrological models such as MODFLOW, and the hydraulic connection between multiple aquifers and groundwater flow paths are evaluated. A groundwater complexity map is generated to mark areas with multiple aquifers or complex flow paths. For the rock mass structure complexity index, the fractures and broken zones in the rock mass are identified through borehole records, logging data and geophysical exploration results, and RMR is used to grade the rock mass quality and the degree of development of the broken zones. Geomechanics analysis software is used to simulate the stability of the rock mass structure and generate a rock mass structure complexity map.

[0025] Through the above complexity analysis, the results of different indicators are combined to obtain the target geological complexity indicator set.

[0026] Based on the above complexity analysis results, the drilling network is laid out. In this process, the drilling network layout adopts a uniform grid layout method, that is, the drilling points are laid out according to a uniform grid in the entire target area. Each drilling point is evenly distributed according to the grid spacing to ensure that each borehole can effectively cover the geological features in the area. Even in areas of different complexity, the drilling network remains evenly laid out to obtain continuous and complete geological data. Through this uniform drilling network layout method, a preset collection drilling point set is generated, which contains the specific coordinates of each borehole.

[0027] Step S200: Based on the multiple drilling data sets and the multiple drilling locations, an initial three-dimensional geological model partition is constructed for the target downhole area to obtain an initial three-dimensional geological partition model, wherein the initial three-dimensional geological partition model includes multiple geological partition areas, and the multiple geological partition areas include multiple geological features.

[0028] In an embodiment of the present application, a first borehole data set is first randomly extracted from multiple borehole data sets and used as a flooding point. Next, adjacent boreholes are matched according to the borehole location and data to define a first neighborhood, that is, other borehole data sets associated with the first borehole within a preset flooding distance. Then, a flooding analysis is performed on the neighborhood to determine the geological characteristics of the area, and it is expanded into a complete first geological division area, which contains clear boundary ridge lines. By selecting different flooding points multiple times, the boundary ridge lines of multiple areas are determined, and finally the initial three-dimensional geological model of the entire downhole area is divided according to these boundaries to form multiple partitions with independent geological characteristics.

[0029] Through the above process, an initial three-dimensional geological zoning model is obtained, wherein the initial three-dimensional geological zoning model includes a plurality of geological division areas, and the plurality of geological division areas include a plurality of geological features.

[0030] Furthermore, in the method provided in the embodiment of the application, the initial three-dimensional geological model partitioning construction of the target downhole area is performed based on the multiple drilling data sets and the multiple drilling positions to obtain the initial three-dimensional geological partitioning model, and further includes:

[0031] A first borehole data set is randomly extracted from the multiple borehole data sets, and the first borehole data set is used as a flooding point; a first neighborhood of the first borehole data set is matched based on the multiple borehole locations and the multiple borehole data sets, wherein the first neighborhood is a borehole data set in the multiple borehole data sets whose distance from the first borehole data is within a preset flooding distance; a flooding analysis is performed on the first neighborhood to determine a first flooding area; a neighborhood expansion is performed based on the first flooding area to obtain a first geological division area, wherein the first geological division area includes a first area boundary ridge line; flooding points are selected multiple times to determine multiple area boundary ridge lines, and an initial three-dimensional geological model of the target downhole area is partitioned based on the multiple area boundary ridge lines to obtain the initial three-dimensional geological partition model.

[0032] In an embodiment of the present application, a first borehole data set is first randomly extracted from a plurality of borehole data sets, and these randomly selected borehole data sets are grouped as flooding points. Next, a first neighborhood is matched. The so-called first neighborhood refers to a borehole data set adjacent to the flooding point within a preset range. Specifically, the neighborhood is determined by a distance algorithm, such as Euclidean distance, and the range of the neighborhood is determined by a preset flooding distance, such as 200 meters or 500 meters. Within this range, the borehole data adjacent to the flooding point are identified and classified as the first neighborhood.

[0033] After the first neighborhood is determined, flooding analysis is performed. Flooding analysis is a geological analysis method that extends from the flooding point, with the goal of determining which geological features are continuous within the first neighborhood. To this end, the borehole data within the first neighborhood is analyzed in the geological modeling software by using an interpolation algorithm, such as the Kriging interpolation method, to evaluate the similarity of the lithology, layer thickness and other characteristics of each borehole point. Based on these analysis results, a continuous area with similar geological characteristics to the flooding point is generated, called the first flooding area, which reflects the spatial similarity and continuity of geological features.

[0034] After the first flooded area is determined, neighborhood expansion is performed. The purpose of neighborhood expansion is to further expand the scope of analysis by continuing to analyze adjacent areas outside the boundaries of the first flooded area. Using a spatial expansion algorithm, such as a region growing algorithm, in the geological modeling tool, the outer boundary of the first flooded area is analyzed, and if the geological characteristics of the external neighborhood, such as lithology or layer thickness, are consistent, the area is continued to be expanded. If the geological characteristics change significantly, such as a sudden change in lithology or the appearance of a fault, the expansion process stops, resulting in a complete geological area, called the first geological division area.

[0035] The first geological demarcation area contains a clear boundary, called the first area boundary ridgeline. The boundary ridgeline is determined by gradient change algorithms or spatial analysis tools, such as the boundary detection tool, and it marks the point where the geological characteristics of the area change significantly. The boundary ridgeline defines the boundaries of the area and clearly distinguishes different geological areas.

[0036] Next, the above steps are repeated by selecting flooding points multiple times. For each new flooding point, the same neighborhood matching, flooding analysis and neighborhood expansion are performed to determine the new geological division area and its boundary ridge line. Each flooding point will generate a new geological division area, and the entire target underground area is systematically divided into multiple geological division areas through the boundary ridge lines of these areas.

[0037] Finally, based on multiple regional boundary ridge lines, the entire target downhole area is divided into three dimensions to form an initial three-dimensional geological zoning model.

[0038] Furthermore, in the method provided in the embodiment of the application, performing flooding analysis on the first neighborhood to determine the first flooding area further includes:

[0039] Enumerate the drilling data sets in pairs for the first neighborhood to obtain multiple first neighborhood combinations, perform similarity identification on the multiple first neighborhood combinations respectively, and obtain multiple first neighborhood combination similarities; perform fluctuation variance identification based on the multiple first neighborhood combination similarities to determine the first sub-partition coefficient of the first neighborhood; determine whether the first sub-partition coefficient is less than or equal to the preset sub-partition coefficient, and if so, take the first neighborhood as the first flooding area; if not, take the flooding point as the starting point, and perform flooding impact analysis in the first neighborhood in order from near to far from the flooding point to determine the first flooding area.

[0040] In an embodiment of the present application, for all borehole data sets in the first neighborhood, pairwise enumeration of borehole data sets is performed, that is, within the neighborhood range, any two borehole data sets are paired to form multiple first neighborhood combinations. Next, similarity identification is performed for each first neighborhood combination, and the spatial differences in geological characteristics such as lithology and stratum thickness between each pair of boreholes are measured by using the Euclidean distance calculation method. Specifically, the data of each borehole is represented as a multidimensional vector, such as lithology, layer thickness and other dimensions. The difference between the two vectors is calculated by Euclidean distance. If the Euclidean distance is small, it means that the data of the two boreholes are relatively similar. After this calculation process, multiple first neighborhood combination similarities are obtained.

[0041] Afterwards, the fluctuation variance of multiple first neighborhood combination similarities is identified. First, the mean of all first neighborhood combination similarities is calculated. Then, the difference between each similarity value and the mean is squared to obtain the square error of each combination. All square errors are summed and divided by the number of combinations to obtain the variance value. Next, based on the result of the fluctuation variance, the first sub-partition coefficient is further calculated. The calculation of the first sub-partition coefficient is to convert the fluctuation variance value according to a preset ratio, that is, to obtain the sub-partition coefficient by dividing the preset partition standard value by the obtained variance value. Through this process, the first sub-partition coefficient of the first neighborhood is obtained.

[0042] The calculated first sub-division coefficient of the first neighborhood is then compared with a preset sub-division coefficient, where the preset sub-division coefficient is preset by a technical expert and is used to determine the consistency of the geological characteristics in the neighborhood. If the first sub-division coefficient is less than or equal to the preset value, it means that the geological characteristics of the neighborhood change little and have high stability, and the first neighborhood is determined as the first flooding area.

[0043] If the calculated first sub-division coefficient is greater than the preset value, it means that the geological characteristics in the neighborhood fluctuate greatly and change significantly, and it cannot be directly determined as a flooding area. At this time, the flooding impact analysis is carried out in the first neighborhood in order from near to far from the flooding point to determine the first flooding area. Specifically, the straight-line distance between each borehole and the flooding point is first calculated using the Euclidean distance, and all the boreholes in the neighborhood are sorted in order from near to far. The geological characteristics of these boreholes are analyzed in turn, focusing on comparing the geological characteristics of each borehole with the flooding point, such as whether the lithology and stratum thickness are consistent. If the difference is within the preset threshold range, it means that the geological characteristics of the borehole are similar enough to the flooding point, and its geological characteristics can be considered to be continuous, and the borehole can be included in the flooding area. In this step, the threshold is a standard set in advance to determine whether the difference in geological characteristics between the borehole and the flooding point is acceptable. For example, if the percentage difference in lithology change does not exceed a certain threshold, or the difference in stratum thickness is within a certain range, the data of the borehole can be accepted. If the difference between the geological characteristics of a borehole and the flooding point is less than the threshold, the borehole is considered to meet the conditions and is identified as part of the flooding area. Once the first borehole that meets the conditions is found, that is, the borehole that meets the threshold requirement, the borehole and the surrounding area are identified as the first flooding area.

[0044] Furthermore, the method provided in the application embodiment also includes:

[0045] Feature extraction is performed on multiple divided drilling data sets of the multiple geological divided areas respectively to obtain multiple divided geological feature sets; and feature concentrated screening is performed on the multiple divided geological feature sets to determine the multiple geological features.

[0046] In an embodiment of the present application, when extracting features from multiple divided borehole data sets of multiple geological division areas, feature extraction is performed on the borehole data in each geological division area through geological modeling software. The borehole data contains a lot of important geological information, such as lithology, stratum thickness, porosity, aquifer information, etc. The borehole data are processed one by one through the geological modeling software, and these geological features are extracted. Specifically, all the borehole data are imported into the geological modeling software, the data of each borehole is analyzed, and key geological parameters such as lithology composition and changes in stratum thickness are extracted. Through this process, multiple divided geological feature sets are formed, each set corresponds to a geological division area, and contains key geological information in the corresponding area.

[0047] Next, the multiple geological feature sets are screened. Specifically, each geological feature is first standardized to ensure that the dimensions of different features are the same. Next, variance analysis is used to calculate the variation of each feature within a single geological division area. The specific steps are to calculate the mean and variance of each feature in the area, and the technical experts set the variance threshold to screen out the features below the threshold. After completing the variance analysis, the screened feature set is the final geological feature set for each geological division area. Through this process, multiple geological features are determined.

[0048] Step S300: performing multimodal data collection according to the plurality of geological features and the plurality of geological division areas to obtain a plurality of multimodal data sets, wherein each multimodal data set includes gravity data, magnetic data and electromagnetic data.

[0049] In an embodiment of the present application, firstly, a target area requiring multimodal data acquisition is determined based on a plurality of geological features that have been extracted and clear geological division areas.

[0050] Next, gravity data collection is carried out. Gravity data collection is to infer the underground structure by measuring the changes in the gravity field caused by the density differences of underground materials. During operation, technicians use gravimeters to measure on the surface or in boreholes of the target area and record the gravity value of each sampling point. After gravity data collection is completed, the generated gravity data set reflects the density changes in each geological division area.

[0051] Then magnetic data collection is carried out. Magnetic data collection measures the changes in the geomagnetic field at different locations to infer the magnetic distribution of underground rocks and minerals. Magnetic characteristics reveal the lithological differences and structural information in the area. The magnetic measurement equipment collects magnetic field strength point by point to form a magnetic anomaly map, and then generates a magnetic data set, which can help identify underground ferromagnetic minerals and structural features.

[0052] Finally, electromagnetic data acquisition is carried out. Electromagnetic data acquisition is to infer the conductivity characteristics by measuring the response of underground materials to electromagnetic waves, which is especially helpful for identifying aquifers, conductive minerals, etc. By sending electromagnetic waves and receiving their return signals, a conductivity distribution map is formed. The generated electromagnetic data set reveals the conductivity of the geological division area and its relationship with the geological body.

[0053] By completing the collection of gravity, magnetic and electromagnetic data, the different physical data in each geological division area are aggregated to generate a multimodal data set.

[0054] Step S400: using a data correction module to perform data processing on the multiple multimodal data sets to obtain multiple corrected multimodal data sets, wherein the data correction module includes a correction filter subunit, a format conversion subunit and an interpolation completion subunit.

[0055] In an embodiment of the present application, when a data correction module is used to process multiple multimodal data sets, the noise in the multimodal data is first removed by correcting the filter subunit. Noise is an interference signal introduced by external interference or equipment error, which may affect the accuracy of the data. For this purpose, common filtering techniques such as low-pass filtering, Gaussian filtering or moving average filtering are used to process the data. Low-pass filtering is used to remove high-frequency noise and retain useful low-frequency information; Gaussian filtering eliminates discrete outliers by smoothing data; and moving average filtering smoothes data fluctuations by calculating the average value of adjacent data points. Through this process, high-frequency interference and discrete anomalies in the multimodal data are removed to generate a filtered data set.

[0056] Then, the format conversion subunit is used to unify the format of different multimodal data. Since gravity, magnetic and electromagnetic data may use different storage formats, unit standards and coordinate systems, these data must be unified through format conversion. The main tasks of format conversion include unit conversion, coordinate system conversion and file format conversion. For example, the unit of gravity data is converted from mGal to standard international units, or the conductivity of electromagnetic data is converted to a unified unit system. At the same time, the spatial coordinates of all data are unified into the same coordinate system, such as WGS84, to ensure that the data can be superimposed and compared under the same spatial reference. After the format conversion is completed, all data are standardized to form a formatted data set.

[0057] Finally, the interpolation completion subunit is used to fill in the missing data in the multimodal data due to equipment limitations or insufficient acquisition conditions. The interpolation completion methods include linear interpolation, spline interpolation and Kriging interpolation. Linear interpolation infers the value of the missing point through adjacent known data points, which is suitable for processing simple smooth data; spline interpolation fills the missing value by fitting the spline curve, which is more suitable for complex nonlinear data; Kriging interpolation is based on the principles of geostatistics and uses spatial correlation to infer the data of the missing point, which is particularly suitable for geological data with spatial correlation. After interpolation completion, the missing parts of the original data are filled to generate a complete interpolation completion data set to ensure that the geological data covers all areas.

[0058] After the above three steps, multiple corrected multimodal data sets are obtained.

[0059] Step S500: Utilizing the plurality of corrected multimodal data sets, a spatiotemporal characteristic fusion loss analysis is performed on the initial three-dimensional geological partition model to obtain a fusion model parameter set.

[0060] In an embodiment of the present application, the initial regional fusion model parameters of each geological division area in the initial model are first extracted, and the loss analysis network is used to evaluate the matching degree of these parameters with the corrected multimodal data set, such as gravity, magnetic and electromagnetic data. The loss amount of each area is calculated through loss analysis, and a first constraint condition is set, that is, the overall loss amount is required not to exceed a preset value. Next, the initial model parameters are analyzed by the inversion analysis unit in the network to generate an inversion multimodal data set, and the correction coefficient is generated by calculating the initial loss amount. The initial parameters are adjusted using these correction coefficients to obtain the corrected regional fusion model parameters. Afterwards, the loss analysis and parameter correction are performed again, and iterative optimization is performed until the constraints are met and the preset number of analysis times is reached. Finally, the fusion model parameters of each geological division area are summarized to generate a fusion model parameter set.

[0061] Furthermore, the method provided in the application embodiment also includes:

[0062] Extract multiple initial regional fusion model parameters of multiple geological division areas in the initial three-dimensional geological zoning model; use the spatiotemporal characteristic fusion loss analysis network to perform loss analysis on the multiple initial regional fusion model parameters and the multiple corrected multimodal data sets, constrain the loss analysis process with a first constraint condition, and obtain multiple regional fusion model parameters, wherein the first constraint condition is that the overall loss amount of the multiple geological division areas is less than or equal to a preset loss amount; summarize the multiple regional fusion model parameters to obtain the fusion model parameter set.

[0063] In the embodiment of the present application, the initial regional fusion model parameters of each geological division area are first extracted from the initial three-dimensional geological partition model, including the physical properties of the geological area, such as density, magnetism, conductivity, etc. Next, a loss analysis is performed on multiple initial regional fusion model parameters and multiple corrected multimodal data sets using a spatiotemporal characteristic fusion loss analysis network. During the loss analysis process, the error or loss value between the initial regional model parameters and the actual multimodal data is calculated, and these differences are quantified by a loss function, such as a mean square error.

[0064] To ensure the stability of model optimization, the entire loss analysis process is controlled by the first constraint, that is, the overall loss of all geological division areas must be less than or equal to the preset loss threshold. Through this constraint, the error of model parameter adjustment is limited to a reasonable range to avoid overfitting or underfitting. After loss analysis, the regional fusion model parameters of each geological division area are obtained. These regional fusion model parameters reflect the optimized geological characteristics and match the actual multimodal data more accurately.

[0065] Finally, the regional fusion model parameters of all geological division areas are summarized to generate a fusion model parameter set.

[0066] Furthermore, the method provided in the application embodiment also includes:

[0067] The inversion analysis unit of the spatiotemporal characteristic fusion loss analysis network is used to analyze the multiple initial regional fusion model parameters to obtain multiple inversion multimodal data sets; the loss analysis unit of the spatiotemporal characteristic fusion loss analysis network is used to perform loss analysis on the multiple inversion multimodal data sets and the multiple correction multimodal data sets to obtain multiple initial loss amounts; the ratios of the multiple initial loss amounts to the sum of the multiple initial loss amounts are respectively used as multiple correction coefficients; the multiple initial regional fusion model parameters are corrected using the multiple correction coefficients to obtain multiple corrected regional fusion model parameters; the spatiotemporal characteristic fusion loss analysis network is used again to perform loss analysis on the multiple corrected regional fusion model parameters and the multiple correction multimodal data sets until the first constraint condition is met and the preset number of analyses is reached to obtain the multiple regional fusion model parameters.

[0068] In an embodiment of the present application, when analyzing multiple initial regional fusion model parameters using the inversion analysis unit of the spatiotemporal characteristic fusion loss analysis network, the gradient descent method is used to perform inversion analysis on the initial regional fusion model parameters of each geological division area. Specifically, the initial regional fusion model parameters and the corrected multimodal data set are input into the spatiotemporal characteristic fusion loss analysis network, which determines the direction and amplitude of parameter adjustment by calculating the error gradient between the model parameters and the multimodal data. As the gradient descent method is iterated successively, the error of the model parameters is gradually reduced until the loss amount approaches the minimum. After the inversion analysis process is completed, multiple inversion multimodal data sets are generated.

[0069] Next, the inverted multimodal data set and the corrected multimodal data set are analyzed for loss using the mean square error as the loss function. MSE evaluates the model error by calculating the sum of squares of the difference between the predicted value and the actual value of each data point and taking the average value. Specifically, the inverted data is compared with the corrected data, and the error of each geological division area is calculated by MSE. For each area, the deviation between the model prediction value and the actual observation value is quantified by MSE. Through this analysis, the initial loss amount of each area is obtained, that is, multiple initial loss amounts are obtained.

[0070] After calculating the initial loss, multiple correction coefficients are obtained by dividing the multiple initial loss by the sum of the multiple initial loss. Regions with larger correction coefficients represent larger errors, so more corrections are required, while regions with smaller correction coefficients only need fine-tuning. Subsequently, these correction coefficients are used to perform weighted adjustments on the initial regional fusion model parameters to generate new regional fusion model parameters, that is, to obtain multiple corrected regional fusion model parameters.

[0071] After that, the parameters of the multiple corrected regional fusion models are input into the spatiotemporal characteristic fusion loss analysis network again, and the loss analysis is performed again to calculate the new MSE value. With each iteration, the loss analysis network adjusts the weights of the model parameters according to the new loss amount and gradually optimizes these parameters. The iteration process continues until the first constraint condition is met, that is, the overall loss amount of all geological division areas is less than or equal to the preset threshold, or the number of iterations reaches the preset maximum number. Through multiple iterations, the model parameters gradually converge to the optimal state, the loss amount gradually decreases, and finally an accurate parameter result is obtained.

[0072] Finally, the corrected regional fusion model parameters of each geological division area are summarized to obtain multiple regional fusion model parameters.

[0073] Step S600: Optimizing the initial three-dimensional geological partition model based on the fusion model parameter set to obtain a target three-dimensional geological partition model of the target downhole area.

[0074] In the embodiment of the present application, the fusion model parameter set obtained by multimodal data fusion, inversion analysis and multiple iterative optimization is first applied to each geological division area of ​​the initial three-dimensional geological division model. The specific steps are to replace the initial parameters of each geological area with the optimized parameters in the fusion model parameter set to ensure that the physical properties of each area, such as density, magnetism, conductivity, etc., can reflect the fusion results of multimodal data.

[0075] Next, the initial model with the fusion model parameters is updated using 3D geological modeling software, and the boundaries, layer thickness, structure, and geological characteristics of each geological division area are recalculated based on the new parameters. Through this step, the boundaries and physical characteristics of each geological area are redefined to ensure that the structure and geological characteristics of the model are more consistent with the actual underground area. Finally, after the model update is completed, the target 3D geological division model of the target underground area is generated.

[0076] In the embodiments of the present application, in summary, the embodiments of the present application have at least the following technical effects:

[0077] The present application performs borehole data collection on a preset collection borehole point set in a target downhole area to obtain multiple borehole data sets, wherein the multiple borehole data sets include multiple borehole locations; constructs an initial three-dimensional geological model partition of the target downhole area based on the multiple borehole data sets and the multiple borehole locations to obtain an initial three-dimensional geological partition model, wherein the initial three-dimensional geological partition model includes multiple geological partition areas, and the multiple geological partition areas include multiple geological features; performs multimodal data collection based on the multiple geological features and the multiple geological partition areas to obtain multiple multimodal data sets, wherein each multimodal data set includes gravity data, magnetic data and electromagnetic data; performs data processing on the multiple multimodal data sets using a data correction module to obtain multiple corrected multimodal data sets, wherein the data correction module includes a correction filter subunit, a format conversion subunit and an interpolation completion subunit; performs spatiotemporal characteristic fusion loss analysis on the initial three-dimensional geological partition model using the multiple corrected multimodal data sets to obtain a fusion model parameter set; optimizes the initial three-dimensional geological partition model based on the fusion model parameter set and the multiple multimodal data sets to obtain a target three-dimensional geological partition model of the target downhole area. The present invention solves the technical problem of insufficient accuracy in the prior art when modeling complex underground structures. By collecting drilling data of the target downhole area, constructing three-dimensional geological model zoning, collecting and processing multimodal data, and optimizing the model by using time-space characteristic fusion loss analysis, a more accurate target three-dimensional geological zoning model is finally obtained, thereby achieving the technical effect of improving the modeling accuracy and reliability of the downhole geological structure.

[0078] Embodiment 2, based on the same inventive concept as the method for downhole three-dimensional geological modeling under multimodal data survey in the above embodiment, Figure 2 As shown, the present application provides a downhole three-dimensional geological modeling device under multimodal data survey, and the device and method embodiments in the embodiments of the present application are based on the same inventive concept. Among them, the device includes:

[0079] A drilling data acquisition module 11, wherein the drilling data acquisition module 11 performs drilling data acquisition on a preset acquisition drilling point set of a target downhole area to obtain multiple drilling data sets, wherein the multiple drilling data sets include multiple drilling positions; a partition model construction module 12, wherein the partition model construction module 12 performs initial three-dimensional geological model partition construction of the target downhole area based on the multiple drilling data sets and the multiple drilling positions to obtain an initial three-dimensional geological partition model, wherein the initial three-dimensional geological partition model includes multiple geological partition areas, and the multiple geological partition areas include multiple geological features; a multimodal data acquisition module 13, wherein the multimodal data acquisition module 13 performs multimodal data acquisition according to the multiple geological features and the multiple geological partition areas to obtain multiple multimodal data sets, wherein, Each multimodal data set includes gravity data, magnetic data and electromagnetic data; a data processing module 14, wherein the data processing module 14 uses a data correction module to perform data processing on the multiple multimodal data sets to obtain multiple corrected multimodal data sets, wherein the data correction module includes a correction filter subunit, a format conversion subunit and an interpolation completion subunit; a fusion loss analysis module 15, wherein the fusion loss analysis module 15 uses the multiple corrected multimodal data sets to perform spatiotemporal characteristic fusion loss analysis on the initial three-dimensional geological partition model to obtain a fusion model parameter set; a partition model optimization module 16, wherein the partition model optimization module 16 optimizes the initial three-dimensional geological partition model based on the fusion model parameter set and the multiple multimodal data sets to obtain a target three-dimensional geological partition model of the target downhole area.

[0080] Furthermore, the device is also used to achieve the following functions:

[0081] Acquire a historical geological survey data set of the target downhole area; analyze the historical geological survey data set according to a preset complexity index set to obtain a target geological complexity index set, wherein the preset complexity index set includes a structural complexity index, a lithological complexity index, a groundwater complexity index and a rock structure complexity index; lay out a drilling network based on the target geological complexity index set to obtain the preset acquisition drilling point set.

[0082] Furthermore, the device is also used to achieve the following functions:

[0083] A first borehole data set is randomly extracted from the multiple borehole data sets, and the first borehole data set is used as a flooding point; a first neighborhood of the first borehole data set is matched based on the multiple borehole locations and the multiple borehole data sets, wherein the first neighborhood is a borehole data set in the multiple borehole data sets whose distance from the first borehole data is within a preset flooding distance; a flooding analysis is performed on the first neighborhood to determine a first flooding area; a neighborhood expansion is performed based on the first flooding area to obtain a first geological division area, wherein the first geological division area includes a first area boundary ridge line; flooding points are selected multiple times to determine multiple area boundary ridge lines, and an initial three-dimensional geological model of the target downhole area is partitioned based on the multiple area boundary ridge lines to obtain the initial three-dimensional geological partition model.

[0084] Furthermore, the device is also used to achieve the following functions:

[0085] Enumerate the drilling data sets in pairs for the first neighborhood to obtain multiple first neighborhood combinations, perform similarity identification on the multiple first neighborhood combinations respectively, and obtain multiple first neighborhood combination similarities; perform fluctuation variance identification based on the multiple first neighborhood combination similarities to determine the first sub-partition coefficient of the first neighborhood; determine whether the first sub-partition coefficient is less than or equal to the preset sub-partition coefficient, and if so, take the first neighborhood as the first flooding area; if not, take the flooding point as the starting point, and perform flooding impact analysis in the first neighborhood in order from near to far from the flooding point to determine the first flooding area.

[0086] Furthermore, the device is also used to achieve the following functions:

[0087] Feature extraction is performed on multiple divided drilling data sets of the multiple geological divided areas respectively to obtain multiple divided geological feature sets; and feature concentrated screening is performed on the multiple divided geological feature sets to determine the multiple geological features.

[0088] Furthermore, the device is also used to achieve the following functions:

[0089] Extract multiple initial regional fusion model parameters of multiple geological division areas in the initial three-dimensional geological zoning model; use the spatiotemporal characteristic fusion loss analysis network to perform loss analysis on the multiple initial regional fusion model parameters and the multiple corrected multimodal data sets, constrain the loss analysis process with a first constraint condition, and obtain multiple regional fusion model parameters, wherein the first constraint condition is that the overall loss amount of the multiple geological division areas is less than or equal to a preset loss amount; summarize the multiple regional fusion model parameters to obtain the fusion model parameter set.

[0090] Furthermore, the device is also used to achieve the following functions:

[0091] The inversion analysis unit of the spatiotemporal characteristic fusion loss analysis network is used to analyze the multiple initial regional fusion model parameters to obtain multiple inversion multimodal data sets; the loss analysis unit of the spatiotemporal characteristic fusion loss analysis network is used to perform loss analysis on the multiple inversion multimodal data sets and the multiple correction multimodal data sets to obtain multiple initial loss amounts; the ratios of the multiple initial loss amounts to the sum of the multiple initial loss amounts are respectively used as multiple correction coefficients; the multiple initial regional fusion model parameters are corrected using the multiple correction coefficients to obtain multiple corrected regional fusion model parameters; the spatiotemporal characteristic fusion loss analysis network is used again to perform loss analysis on the multiple corrected regional fusion model parameters and the multiple correction multimodal data sets until the first constraint condition is met and the preset number of analyses is reached to obtain the multiple regional fusion model parameters.

[0092] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0093] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0094] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A method for underground three-dimensional geological modeling under multimodal data survey, characterized in that: The method comprises: Performing drilling data collection on a preset collection drilling point set in a target downhole area to obtain a plurality of drilling data sets, wherein the plurality of drilling data sets include a plurality of drilling positions; Based on the multiple drilling data sets and the multiple drilling positions, an initial three-dimensional geological model partitioning construction is performed on the target downhole area to obtain an initial three-dimensional geological partitioning model, wherein the initial three-dimensional geological partitioning model includes multiple geological partitioning areas, and the multiple geological partitioning areas include multiple geological features; Performing multimodal data collection according to the plurality of geological features and the plurality of geological division areas to obtain a plurality of multimodal data sets, wherein each multimodal data set includes gravity data, magnetic data and electromagnetic data; Using a data correction module to process the multiple multimodal data sets to obtain multiple corrected multimodal data sets, wherein the data correction module includes a correction filter subunit, a format conversion subunit and an interpolation completion subunit; Using the multiple corrected multimodal data sets, a temporal and spatial characteristic fusion loss analysis is performed on the initial three-dimensional geological partition model to obtain a fusion model parameter set; The initial three-dimensional geological partition model is optimized based on the fusion model parameter set and the multiple multimodal data sets to obtain a target three-dimensional geological partition model of the target downhole area.

2. The method for downhole three-dimensional geological modeling under multimodal data survey according to claim 1, characterized in that: Drilling data is collected for a preset collection drilling point set in the target downhole area to obtain multiple drilling data sets, including: Acquire a historical geological survey data set of the target downhole area; Analyzing the historical geological survey data set according to a preset complexity index set to obtain a target geological complexity index set, wherein the preset complexity index set includes a structural complexity index, a lithological complexity index, a groundwater complexity index, and a rock mass structure complexity index; The drilling network is laid out based on the target geological complexity index set to obtain the preset acquisition drilling point set.

3. The method for downhole three-dimensional geological modeling under multimodal data survey according to claim 1, characterized in that: Based on the multiple drilling data sets and the multiple drilling positions, an initial three-dimensional geological model partitioning construction is performed for the target downhole area to obtain an initial three-dimensional geological partitioning model, including: randomly extracting a first borehole data set from the plurality of borehole data sets, and taking the first borehole data set as a flooding point; Matching a first neighborhood of the first drilling data set based on the multiple drilling locations and the multiple drilling data sets, wherein the first neighborhood is a drilling data set in the multiple drilling data sets whose distance from the first drilling data set is within a preset flooding distance; Performing flooding analysis on the first neighborhood to determine a first flooding area; Expanding the neighborhood based on the first flooded area to obtain a first geological division area, wherein the first geological division area includes a first area boundary ridge line; The flooding points are selected multiple times to determine multiple regional boundary ridge lines, and the initial three-dimensional geological model of the target downhole area is partitioned based on the multiple regional boundary ridge lines to obtain the initial three-dimensional geological partition model.

4. The method for downhole three-dimensional geological modeling under multimodal data survey as claimed in claim 3, characterized in that: Performing a flood analysis on the first neighborhood to determine a first flood area includes: Enumerating the drilling data sets in pairs for the first neighborhoods to obtain a plurality of first neighborhood combinations, and performing similarity identification on the plurality of first neighborhood combinations respectively to obtain a plurality of first neighborhood combination similarities; Performing fluctuation variance identification based on the combined similarities of the plurality of first neighborhoods, and determining a first sub-partition coefficient of the first neighborhood; determining whether the first sub-partition coefficient is less than or equal to a preset sub-partition coefficient, and if so, taking the first neighborhood as the first flooded area; If not, taking the flood point as a starting point, flood impact analysis is performed in the first neighborhood in order from near to far from the flood point to determine a first flood area.

5. The method for downhole three-dimensional geological modeling under multimodal data survey as claimed in claim 4, characterized in that: include: Extracting features from the plurality of divided borehole data sets of the plurality of geological divided areas respectively to obtain a plurality of divided geological feature sets; The plurality of divided geological feature sets are subjected to feature concentrated screening to determine the plurality of geological features.

6. The method for downhole three-dimensional geological modeling under multimodal data survey according to claim 1, characterized in that: include: Extracting a plurality of initial regional fusion model parameters of a plurality of geological division regions in the initial three-dimensional geological division model; Perform loss analysis on the multiple initial regional fusion model parameters and the multiple corrected multimodal data sets using a spatiotemporal characteristic fusion loss analysis network, constrain the loss analysis process with a first constraint condition, and obtain multiple regional fusion model parameters, wherein the first constraint condition is that the overall loss amount of the multiple geological division areas is less than or equal to a preset loss amount; The multiple regional fusion model parameters are aggregated to obtain the fusion model parameter set.

7. The method for downhole three-dimensional geological modeling under multimodal data survey according to claim 6, characterized in that: include; Analyzing the multiple initial regional fusion model parameters using an inversion analysis unit of a spatiotemporal characteristic fusion loss analysis network to obtain multiple inversion multimodal data sets; Performing loss analysis on the multiple inversion multimodal data sets and the multiple correction multimodal data sets through a loss analysis unit of a spatiotemporal characteristic fusion loss analysis network to obtain multiple initial loss amounts; respectively taking the ratios of the plurality of initial loss amounts to the sum of the plurality of initial loss amounts as a plurality of correction coefficients; Using the multiple correction coefficients to correct the multiple initial regional fusion model parameters to obtain multiple corrected regional fusion model parameters; The spatiotemporal characteristic fusion loss analysis network is used again to perform loss analysis on the multiple corrected regional fusion model parameters and the multiple corrected multimodal data sets until the first constraint condition is met and the preset number of analyses is reached, thereby obtaining the multiple regional fusion model parameters.

8. A downhole three-dimensional geological modeling device under multimodal data survey, characterized in that: The device comprises: A drilling data acquisition module, wherein the drilling data acquisition module performs drilling data acquisition on a preset collection drilling point set in a target downhole area to obtain a plurality of drilling data sets, wherein the plurality of drilling data sets include a plurality of drilling positions; A partition model construction module, wherein the partition model construction module performs initial three-dimensional geological model partition construction of the target downhole area based on the multiple drilling data sets and the multiple drilling positions to obtain an initial three-dimensional geological partition model, wherein the initial three-dimensional geological partition model includes multiple geological partition areas, and the multiple geological partition areas include multiple geological features; A multimodal data acquisition module, wherein the multimodal data acquisition module performs multimodal data acquisition according to the plurality of geological features and the plurality of geological division areas to obtain a plurality of multimodal data sets, wherein each multimodal data set includes gravity data, magnetic data and electromagnetic data; A data processing module, wherein the data processing module uses the data correction module to perform data processing on the multiple multimodal data sets to obtain multiple corrected multimodal data sets, wherein the data correction module includes a correction filter subunit, a format conversion subunit and an interpolation completion subunit; A fusion loss analysis module, wherein the fusion loss analysis module uses the multiple corrected multimodal data sets to perform spatiotemporal characteristic fusion loss analysis on the initial three-dimensional geological partition model to obtain a fusion model parameter set; A partition model optimization module is used to optimize the initial three-dimensional geological partition model based on the fusion model parameter set and the multiple multimodal data sets to obtain a target three-dimensional geological partition model of the target downhole area.

Images