Three-dimensional variable temperature field modeling method and system for hydrate enrichment area

Abstract

The invention relates to the field of geological information, and provides a three-dimensional variable temperature field modeling method and system for a hydrate enrichment area. The method comprises the steps of collecting a multi-source heterogeneous original data stream, and performing formatting processing and standardization processing on the multi-source heterogeneous original data stream to obtain a standardized data set; calling a data processing acceleration card to intelligently pre-process the standardized data set to obtain a pre-processed data set; geological modeling is carried out on the preprocessed data set through hybrid logic, and a three-dimensional geological constraint model is obtained; mapping the temperature and pressure data in the preprocessed data set to a three-dimensional geological constraint model, and fusing time sequence data to carry out space-time synthesis on the mapped temperature and pressure data to obtain a three-dimensional variable temperature field preliminary model; and rendering the three-dimensional variable temperature field preliminary model, outputting a three-dimensional visualization result, and performing comparison verification on the trial production monitoring data and the three-dimensional variable temperature field model to obtain a three-dimensional variable temperature field verification model. According to the method, the modeling precision of the hydrate enrichment area model is improved.

Description

Technical Field

[0001] This invention relates to the field of geological information technology, and in particular to a method and system for modeling three-dimensional variable temperature fields in hydrate-rich areas. Background Technology

[0002] Natural gas hydrates are a type of clean energy with abundant reserves, and their formation, enrichment, and pilot production processes are all closely related to the geothermal field. Accurate three-dimensional temperature field models are a key technological foundation for evaluating hydrate resource potential, optimizing pilot production schemes, and predicting environmental responses. During the exploration and development of hydrate-rich areas, it is necessary to comprehensively analyze the spatial distribution patterns of multi-dimensional parameters such as formation temperature, pressure, and burial depth, and dynamically monitor the evolution of the temperature field during pilot production, thereby providing data support for pilot production scheme design and risk assessment.

[0003] Existing temperature field modeling methods mainly rely on single types of temperature measurement data, such as interpolation calculations using only well logging temperature curves. They fail to effectively integrate multi-source heterogeneous data such as geothermal gradients, pressure data, seismic attributes, and lithofacies information, resulting in models that cannot comprehensively reflect the temperature and pressure field variations at the mining scale. Furthermore, in hydrate reservoirs with complex structures and high data heterogeneity, traditional interpolation methods such as kriging and inverse distance weighting lack geological constraints. The interpolation process does not consider the controlling effects of geological factors such as faults, structures, and lithofacies on temperature distribution, leading to poor geological rationality and low interpolation accuracy in the constructed temperature field models. This makes it difficult to meet the high accuracy and geological adaptability requirements of hydrate pilot production projects for temperature field models. Summary of the Invention

[0004] This invention provides a method and system for modeling a three-dimensional variable temperature field in hydrate-rich regions, in order to overcome the shortcomings of existing technologies.

[0005] This invention provides a method for modeling a three-dimensional variable temperature field in hydrate-rich regions, comprising: S1: Collect multi-source heterogeneous raw data streams, and perform formatting and standardization processing on the multi-source heterogeneous raw data streams to obtain a standardized dataset; S2: Call the data processing acceleration card to perform intelligent preprocessing on the standardized dataset to obtain a preprocessed dataset; S3: Geological modeling is performed on the preprocessed dataset using hybrid logic to obtain a three-dimensional geological constraint model; S4: Map the temperature and pressure data in the preprocessed dataset to the three-dimensional geological constraint model, and perform spatiotemporal synthesis of the mapped temperature and pressure data by fusing time series data to obtain a preliminary three-dimensional variable temperature field model. S5: Render the preliminary model of the three-dimensional variable temperature field, output the three-dimensional visualization result, and compare and verify the test mining monitoring data with the three-dimensional variable temperature field model to obtain the three-dimensional variable temperature field verification model.

[0006] According to the method for modeling a three-dimensional variable temperature field in a hydrate-rich region provided by the present invention, step S1 further includes: S11: Collect multi-source heterogeneous raw data streams through multi-source data interfaces; S12: Identify the format types of multiple data in the multi-source heterogeneous raw data stream, and convert the data of different formats according to the identification results to obtain data of a unified format. S13: Transform the spatial coordinates in the unified format data to a unified coordinate datum to obtain a standardized dataset.

[0007] According to the present invention, a three-dimensional variable temperature field modeling method for hydrate-rich areas is provided. The multi-source heterogeneous raw data stream in step S11 includes multi-dimensional data from various types of well points in different hydrate-rich areas. The multi-dimensional data includes formation fluid temperature data, hydrostatic pressure data, burial depth data, seafloor temperature data, and geothermal gradient data.

[0008] According to the method for modeling a three-dimensional variable temperature field in a hydrate-rich region provided by the present invention, step S2 further includes: S21: Based on the data processing acceleration card, the missing data in the standardized dataset is filled in parallel using the random forest interpolation algorithm to obtain a complete dataset; S22: Based on the data processing acceleration card, the isolated forest algorithm is used to calculate the path length of multiple data points in the complete dataset. Abnormal data points are obtained and removed according to the path length threshold to obtain a clean dataset. S23: Based on the data processing acceleration card, the high-frequency noise data in the clean dataset is decomposed and reconstructed using the wavelet transform algorithm to reduce noise, thereby obtaining a preprocessed dataset.

[0009] According to the present invention, a three-dimensional variable temperature field modeling method for hydrate-rich regions is provided, wherein the hybrid logic in step S3 includes step-by-step hierarchical modeling logic, point-line-surface-volume construction logic, and phase sequence constraint logic.

[0010] According to the method for modeling a three-dimensional variable temperature field in a hydrate-rich region provided by the present invention, step S3 further includes: S31: Extract fault data from the preprocessed dataset, establish a high-level fault framework model, and construct a low-level fault model based on the high-level fault framework model to obtain a fault constraint framework. S32: Extract well logging interpretation stratification point data from the preprocessed dataset, spatially connect the stratification points of vertical wells and horizontal wells to form a formation boundary line, generate formation planes based on the formation boundary line, fill formation grid cells within the fault constraint framework, and obtain a three-dimensional structural model. S33: Combine deterministic modeling methods and stochastic modeling methods to calculate the lithofacies distribution of the three-dimensional structural model to obtain a lithofacies constrained model; S34: Under the constraints of the lithofacies constraint model, establish the attribute distribution model and the hydrate enrichment distribution model to obtain the three-dimensional geological constraint model.

[0011] According to the present invention, a three-dimensional variable temperature field modeling method for hydrate-rich regions is provided, wherein the deterministic modeling method in step S33 is the ordinary kriging method, and the stochastic modeling method is the sequential indicator simulation method.

[0012] According to the method for modeling a three-dimensional variable temperature field in a hydrate-rich region provided by the present invention, step S4 further includes: S41: Extract the temperature and pressure data from the preprocessed dataset as temperature and pressure data, and perform spatial location matching between the temperature and pressure data and the grid cells in the three-dimensional geological constraint model; S42: Based on the geological structure characteristics of the three-dimensional geological constraint model and the spatial distribution characteristics of the temperature and pressure data, select the optimal interpolation method from the inverse distance weighting method, the kriging method and the triangulation method; S43: The temperature and pressure values ​​of the unassigned grid cells in the three-dimensional geological constraint model are calculated using the optimal interpolation method to obtain the initial temperature field model; S44: The temperature changes at different time points in the time series data are superimposed onto the initial temperature field model to obtain a preliminary three-dimensional variable temperature field model.

[0013] According to the method for modeling a three-dimensional variable temperature field in a hydrate-rich region provided by the present invention, step S5, which involves comparing and verifying the test monitoring data with the three-dimensional variable temperature field model to obtain a three-dimensional variable temperature field verification model, further includes: S51: Calculate the difference between the measured temperature value in the test mining monitoring data and the predicted temperature value at the corresponding position in the preliminary model of the three-dimensional variable temperature field to obtain the temperature error distribution; S52: Based on the temperature error distribution, the interpolation parameters of the preliminary three-dimensional variable temperature field model are optimized in reverse. When the average absolute error of the temperature error distribution is less than a preset threshold, the three-dimensional variable temperature field verification model is output.

[0014] The present invention also provides a three-dimensional variable temperature field modeling system for hydrate-rich regions, for performing a three-dimensional variable temperature field modeling method for hydrate-rich regions as described in any of the above claims, comprising: Standardization module: Used to receive the collected multi-source heterogeneous raw data streams, perform formatting and standardization processing on the multi-source heterogeneous raw data streams, and obtain a standardized dataset; Preprocessing module: used to call the data processing acceleration card to perform intelligent preprocessing on the standardized dataset to obtain a preprocessed dataset; Modeling module: Used to perform geological modeling on the preprocessed dataset through hybrid logic to obtain a three-dimensional geological constraint model; Synthesis module: used to map the temperature and pressure data in the preprocessed dataset to the three-dimensional geological constraint model, and to perform spatiotemporal synthesis of the mapped temperature and pressure data by fusing time series data to obtain a preliminary three-dimensional variable temperature field model; Visualization module: used to render the preliminary model of the three-dimensional variable temperature field and output three-dimensional visualization results; Verification module: The test sampling monitoring data is compared and verified with the three-dimensional variable temperature field model to obtain the three-dimensional variable temperature field verification model.

[0015] This invention provides a three-dimensional variable temperature field modeling method and system for hydrate-rich areas. It directly accesses multi-source heterogeneous raw data streams from logging instruments, seismic data storage devices, and test-production area sensors via a multi-source data interface, and performs formatting and standardization processing. This avoids the tedious manual copying and conversion process of traditional methods, significantly shortening data preparation time and eliminating model construction obstacles caused by inconsistent data formats. Secondly, this invention uses a data processing acceleration card to perform hardware-accelerated random forest interpolation, isolated forest anomaly detection, and wavelet transform denoising. This not only improves processing speed but also significantly reduces model errors caused by missing data, anomalies, and noise through intelligent preprocessing, enabling the temperature field model to meet the accuracy requirements of engineering applications. Thirdly, this invention employs a step-by-step, hierarchical modeling logic to construct a fault constraint framework, combined with a point-line-surface-volume construction logic to form… A three-dimensional structural model is constructed, and a lithofacies constraint model is established through phase sequence constraint logic. This ensures that the constructed temperature field model strictly conforms to geological laws such as faults, structures, and lithofacies, solving the problems of poor geological adaptability and low geological rationality of traditional interpolation methods. This ensures that the model can truly reflect the temperature distribution characteristics under complex geological conditions. An interpolation method optimization mechanism automatically selects the interpolation algorithm that best fits the geological structure and flexibly adjusts the calculation strategy for different geological conditions and data distribution characteristics, avoiding the limitations of a single interpolation method and improving the interpolation accuracy compared to traditional fixed methods. In addition, a three-dimensional variable temperature field model with spatiotemporal changes is constructed by integrating time series data, realizing four-dimensional dynamic visualization of the temperature field evolution during the trial mining process. This provides a key tool for dynamic optimization of trial mining schemes and real-time assessment of environmental risks, filling the technical gap that traditional static models cannot reflect the time-varying characteristics of the temperature field. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 A schematic diagram of a three-dimensional variable temperature field modeling method for hydrate-rich regions provided in an embodiment of the present invention; Figure 2 A schematic diagram of a three-dimensional variable temperature field modeling system for hydrate-rich regions provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the workstation structure provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the clamping structure provided in an embodiment of the present invention; Figure 5This is a schematic diagram of the clamping block during clamping according to an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, embodiments of this invention, and should not be construed as limiting the invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention. In the description of this invention, it should be understood that the terminology used is for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0019] The embodiments of the present invention are described below with reference to the figures.

[0020] like Figure 1 As shown, a method for modeling a three-dimensional variable temperature field in a hydrate-rich region includes: S1: Collect multi-source heterogeneous raw data streams, and perform formatting and standardization processing on the multi-source heterogeneous raw data streams to obtain a standardized dataset.

[0021] Step S1 further includes: S11: Collect multi-source heterogeneous raw data streams through multi-source data interfaces.

[0022] The multi-source heterogeneous raw data stream in step S11 includes multi-dimensional data from various types of well points in different hydrate-rich areas. The multi-dimensional data includes formation fluid temperature data, hydrostatic pressure data, burial depth data, seafloor temperature data, and geothermal gradient data.

[0023] Furthermore, this invention first acquires multi-source heterogeneous raw data streams through a multi-source data interface installed on a workstation. This multi-source data interface integrates various physical interface types such as Ethernet, dedicated logging ports, and USB, connecting to multiple data sources. Among these, logging instruments transmit temperature and pressure curves within the wellbore through the dedicated logging port; seismic data storage equipment transmits 3D seismic body data and fault interpretation data through a high-speed data interface; and sensors in the test production area transmit monitored temperature change data in real time through an industrial Ethernet interface. The multi-source heterogeneous raw data streams acquired by this invention include formation fluid temperature data, hydrostatic pressure data, burial depth data, seafloor temperature data, and geothermal gradient data. These data originate from various well points, including vertical and horizontal wells, in different hydrate-rich areas.

[0024] S12: Identify the format types of multiple data in the multi-source heterogeneous raw data stream, and convert the data of different formats according to the identification results to obtain data of a unified format.

[0025] Furthermore, in step S12, this invention automatically scans the header information and data structure features of the received multi-source heterogeneous raw data streams to identify the well logging curve data format as LAS, the seismic body data format as SEGY, and the test production monitoring data format as CSV. Subsequently, for LAS format well logging data, this invention extracts the curve header information and depth-value pairs, converting them into an internal standard table structure; for SEGY format seismic data, this invention reads the trace header information and amplitude data, associates the three-dimensional spatial coordinates with attribute values, and converts them into a grid data structure; for CSV format test production data, this invention parses the timestamp, sensor number, and measurement values, uniformly converting them into a time-series data table. After format conversion, all data is stored using a unified data table structure, resulting in unified format data.

[0026] S13: Transform the spatial coordinates in the unified format data to a unified coordinate datum to obtain a standardized dataset.

[0027] In step S13, this invention extracts the spatial coordinate information of each record in the unified format data. Specifically, well logging data carries the local coordinate system of wellhead coordinates plus well depth, seismic data carries the survey line number and shot point number, and test production sensor data carries GPS latitude and longitude coordinates. This invention first converts the well depth of the well logging data into vertical depth and calculates the three-dimensional spatial coordinates by combining it with the well inclination data; then, it converts the survey line number and shot point number of the seismic data into geodetic coordinates through observation system parameters; next, it uniformly projects and transforms all coordinates to a unified UTM projection coordinate system to ensure that all data are aligned under the same spatial reference. After the transformation is completed, this invention replaces the original coordinate fields with the transformed coordinate values ​​to obtain a standardized dataset.

[0028] S2: Call the data processing acceleration card to perform intelligent preprocessing on the standardized dataset to obtain a preprocessed dataset.

[0029] Step S2 further includes: S21: Based on the data processing acceleration card, the missing data in the standardized dataset is filled in parallel using the random forest interpolation algorithm to obtain a complete dataset.

[0030] In step S21, the present invention first scans each field of the data table, marking missing locations where key parameters such as temperature and pressure have empty values ​​or are marked as NaN, and compiles an index list of missing data. Then, the present invention calls the random forest interpolation algorithm, which constructs multiple decision trees in parallel on the FPGA chip of the data processing acceleration card. Each decision tree takes the depth, burial depth, lithology, and other features around the missing location as input, and the temperature value of the known location as the training target. Subsequently, the FPGA calculates the prediction results of all decision trees in parallel, and takes the average of the outputs of multiple trees as the imputation value for the missing location. Finally, the present invention writes the calculated imputation value into the missing locations of the standardized dataset, completing the imputation of all missing data and obtaining a complete dataset.

[0031] S22: Based on the data processing acceleration card, the isolated forest algorithm is used to calculate the path length of multiple data points in the complete dataset. Abnormal data points are obtained and removed according to the path length threshold to obtain a clean dataset.

[0032] In step S22, this invention inputs each data point in the complete dataset into the Isolation Forest algorithm. After loading, this invention recursively constructs an isolation tree by randomly selecting feature dimensions and segmentation thresholds on the data processing acceleration card using the Isolation Forest algorithm. For each data point in the complete dataset, this invention calculates the path length from the root node to the isolated leaf node in the isolation tree. Normal data points, being similar to other data points, require more segmentations to be isolated, resulting in longer path lengths; abnormal data points, whose values ​​deviate from the normal range, only require a few segmentations to be isolated, resulting in shorter path lengths. This invention calculates the average path length of the data point in all isolation trees and marks data points with an average path length less than a set path length threshold as abnormal data points. Finally, all marked abnormal data points are deleted from the complete dataset to obtain a clean dataset.

[0033] S23: Based on the data processing acceleration card, the high-frequency noise data in the clean dataset is decomposed and reconstructed using the wavelet transform algorithm to reduce noise, thereby obtaining a preprocessed dataset.

[0034] In step S23, the present invention inputs the trial temperature time series data from the clean dataset into a wavelet transform algorithm. This algorithm, on a data processing acceleration card, first selects a wavelet basis function to decompose the temperature time series signal into low-frequency approximate components and multiple layers of high-frequency detail components. The low-frequency approximate components represent the true temperature change trend, while the high-frequency detail components contain instrument noise and random fluctuations. The present invention sets a noise threshold, setting the coefficients of high-frequency detail components with amplitudes less than the threshold to zero, and retaining coefficients with amplitudes greater than the threshold. Subsequently, the present invention performs an inverse wavelet transform on the processed wavelet coefficients to reconstruct the signal, combining the retained low-frequency components and the threshold-processed high-frequency components into a denoised temperature signal. Finally, the present invention replaces the original high-frequency noise data in the clean dataset with the reconstructed temperature signal, completing the denoising process for all time series data and obtaining a preprocessed dataset.

[0035] S3: Geological modeling is performed on the preprocessed dataset using hybrid logic to obtain a three-dimensional geological constraint model.

[0036] The hybrid logic in step S3 includes step-by-step hierarchical modeling logic, point-line-surface-volume construction logic, and phase sequence constraint logic.

[0037] Step S3 further includes: S31: Extract fault data from the preprocessed dataset, establish a high-level fault framework model, and construct a low-level fault model based on the high-level fault framework model to obtain a fault constraint framework.

[0038] In step S31, the present invention reads fault data such as fault strike, dip angle, and extension length obtained from seismic interpretation from the preprocessed dataset. After reading, the present invention classifies the faults according to their control range and displacement. Faults with displacement greater than 50 meters and extension length greater than 5 kilometers are identified as high-level faults, while faults with displacement less than 50 meters and shorter extension length are identified as low-level faults.

[0039] Subsequently, this invention extracts the three-dimensional spatial coordinates of all high-level faults, connects the start and end points of each fault in three-dimensional space to construct fault strike lines, and then extends them downwards based on dip angle data to form fault planes. All high-level fault planes are then combined to establish a high-level fault framework model. Next, within each structural unit divided by the high-level fault framework model, this invention extracts low-level fault data, constructs low-level fault planes in the same manner, and inserts the low-level fault planes into the corresponding structural units, ultimately forming a complete fault system containing main faults and secondary faults, thus obtaining a fault constraint framework.

[0040] S32: Extract well logging interpretation stratification point data from the preprocessed dataset, spatially connect the stratification points of vertical wells and horizontal wells to form a formation boundary line, generate formation planes based on the formation boundary line, and fill the formation grid cells within the fault constraint framework to obtain a three-dimensional structural model.

[0041] In step S32, the present invention extracts well logging interpretation stratification point data from the preprocessed dataset, which records the formation boundary positions of each well at different depths. Subsequently, for vertical well data, the present invention directly reads the well depth and formation number, and combines the wellhead coordinates with the vertical depth to obtain the three-dimensional coordinates of the stratification points; for horizontal well data, the present invention reads the measurement depth and well inclination angle, calculates the horizontal displacement and vertical depth, and obtains the three-dimensional coordinates of the stratification points.

[0042] Furthermore, this invention connects all stratigraphic points of the same stratum according to spatial distance, inserting intermediate nodes between adjacent well points to form stratigraphic boundaries distributed along the stratigraphic strike. Next, this invention performs triangular meshing on each stratigraphic boundary, stretching to generate curved surfaces between adjacent stratigraphic boundaries to form stratigraphic planes representing the top and bottom surfaces of the strata. Finally, within the spatial range defined by the fault constraint framework, this invention cuts the three-dimensional space according to the upper and lower boundaries of the stratigraphic planes, filling it with regular hexahedral mesh units. Each mesh unit records its corresponding stratigraphic number, resulting in a three-dimensional structural model.

[0043] S33: The lithofacies distribution of the three-dimensional structural model is calculated by combining deterministic modeling method and stochastic modeling method to obtain a lithofacies constraint model; wherein the deterministic modeling method is ordinary kriging method and the stochastic modeling method is sequential indicator simulation method.

[0044] In step S33, the present invention first uses ordinary kriging to perform deterministic modeling of the three-dimensional structural model. Ordinary kriging is a geostatistical interpolation method. The present invention extracts lithofacies type data of known well points from the preprocessed dataset, calculates the spatial distance and lithofacies similarity between each well point, and constructs a semi-variogram. The semi-variogram describes the spatial variation law of lithofacies. The present invention calculates the weight coefficients of any grid cell and its surrounding known well points according to the semi-variogram, and calculates the lithofacies values ​​of each well point by weighted average to obtain the lithofacies trend value of the grid cell.

[0045] Subsequently, this invention employs a sequential indicator simulation method for stochastic modeling. This method adds random perturbation to a deterministic trend. Following the sedimentary facies sequence of sandstone, siltstone, and mudstone, this invention randomly selects an unassigned grid cell from the three-dimensional structural model, reads the lithofacies types of its surrounding assigned grid cells, calculates the conditional probability of the cell representing each lithofacies type, and randomly assigns a lithofacies type to the grid cell according to the probability distribution. This process is repeated until all grid cells are assigned values, resulting in a lithofacies-constrained model.

[0046] S34: Under the constraints of the lithofacies constraint model, establish the attribute distribution model and the hydrate enrichment distribution model to obtain the three-dimensional geological constraint model.

[0047] Furthermore, in step S34, the present invention aims to establish an attribute distribution model under the constraints of a lithofacies constraint model. Specifically, the present invention extracts measured data of attribute parameters such as porosity and permeability from the preprocessed dataset, and statistically analyzes the attribute parameter distribution range of each well point corresponding to each lithofacies type in the lithofacies constraint model. Subsequently, the present invention traverses each grid cell in the lithofacies constraint model, reads the lithofacies type of the grid, performs interpolation calculations from the corresponding lithofacies attribute parameter distribution range, and assigns the calculated porosity and permeability values ​​to the grid cell, thus completing the establishment of the attribute distribution model.

[0048] Next, the present invention extracts the test data of hydrate saturation in the preprocessed dataset, combines it with the porosity distribution in the attribute distribution model, calculates the hydrate reserves of each grid cell, marks the grid cells with reserves greater than the threshold as enriched regions, and marks the grid cells with reserves less than the threshold as non-enriched regions, thus forming a hydrate enrichment distribution model.

[0049] Ultimately, this invention integrates the fault constraint framework, three-dimensional structural model, lithofacies constraint model, attribute distribution model, and hydrate enrichment distribution model into a unified data structure to obtain a three-dimensional geological constraint model.

[0050] S4: Map the temperature and pressure data in the preprocessed dataset to the three-dimensional geological constraint model, and perform spatiotemporal synthesis of the mapped temperature and pressure data by fusing time series data to obtain a preliminary three-dimensional variable temperature field model.

[0051] Step S4 further includes: S41: Extract the temperature and pressure data from the preprocessed dataset as temperature and pressure data, and perform spatial location matching between the temperature and pressure data and the grid cells in the three-dimensional geological constraint model.

[0052] In step S41, the present invention extracts the temperature and pressure measurements of each well point from the preprocessed dataset. Each data point contains four fields: wellhead coordinates, measurement depth, temperature value, and pressure value. The temperature and pressure values ​​are then merged into temperature-pressure data. Subsequently, the present invention reads the center point coordinates of all grid cells in the three-dimensional geological constraint model, including the X, Y, and Z coordinates. Next, the present invention iterates through each data point in the temperature-pressure data, calculates the Euclidean distance between the data point and the center points of all grid cells, selects the grid cell with the smallest distance as the matching cell, and writes the temperature and pressure values ​​of the data point into the attribute fields of the matching cell, thus completing the spatial matching of the temperature-pressure data and the grid cells.

[0053] S42: Based on the geological structure characteristics of the three-dimensional geological constraint model and the spatial distribution characteristics of the temperature and pressure data, select the optimal interpolation method from the inverse distance weighting method, the kriging method, and the triangulation method.

[0054] In step S42, the present invention first extracts fault distribution information and stratigraphic dip information from the three-dimensional geological constraint model as geological structural features, and statistically analyzes the sampling density and distribution uniformity of temperature and pressure data in the horizontal and vertical directions as spatial distribution features. Subsequently, the present invention judges the geological structural features. If the fault density is high and the stratigraphic dip changes drastically, the kriging method is selected as the interpolation method because it can consider spatial autocorrelation and adapt to complex geological structures. If there are few faults and the stratigraphy is gentle, the spatial distribution characteristics of the temperature and pressure data are judged. When the data distribution is uniform, the inverse distance weighting method is selected; when the data distribution is irregular, the triangulation method is selected. Finally, the present invention automatically selects one of the three interpolation methods according to the above judgment logic and marks the selected method as the optimal interpolation method.

[0055] S43: The temperature and pressure values ​​of the unassigned grid cells in the three-dimensional geological constraint model are calculated using the optimal interpolation method to obtain the initial temperature field model.

[0056] Further, in step S43, this invention reads the grid cells with assigned temperature and pressure data from the three-dimensional geological constraint model as the set of known points, and reads the unassigned grid cells as the set of points to be calculated. Subsequently, this invention calls the optimal interpolation method to process the set of points to be calculated. Specifically, if the optimal interpolation method is the Kriging method, this invention calculates the semi-variogram between known points, constructs a Kriging equation system, calculates the weight coefficient between each point to be calculated and the known points, and sums the temperature and pressure values ​​of the known points according to the weights to obtain the calculated value; if the optimal interpolation method is the inverse distance weighted method, this invention calculates the distance between the point to be calculated and the known points, takes the reciprocal of the distance as the weight, and sums the temperature and pressure values ​​of the known points according to the weights to obtain the calculated value; if the optimal interpolation method is the triangulation method, this invention constructs a triangular mesh of known points, determines the triangle containing the point to be calculated, and performs linear interpolation on the temperature and pressure values ​​of the three vertices of the triangle to obtain the calculated value. Finally, this invention writes all the calculated values ​​into the corresponding grid cells to be calculated, completes the temperature and pressure assignment of all grid cells in the three-dimensional geological constraint model, and obtains the initial temperature field model.

[0057] S44: The temperature changes at different time points in the time series data are superimposed onto the initial temperature field model to obtain a preliminary three-dimensional variable temperature field model.

[0058] In step S44, this invention extracts time-series data from the preprocessed dataset, which records temperature monitoring values ​​at different time points during the trial mining process. After data extraction, this invention first selects the start time of the trial mining as the initial time point, reads the temperature monitoring data at that time, compares the temperature value of the monitoring point with the temperature value of the corresponding grid cell in the initial temperature field model, and calculates the temperature difference. Subsequently, this invention selects multiple time points during the trial mining process, including day 1, day 15, day 30, etc., reads the temperature monitoring data at each time point, and calculates the temperature change at each time point relative to the initial time point. Next, this invention superimposes the temperature change at each time point onto the temperature value of the corresponding grid cell in the initial temperature field model to generate a temperature field snapshot for that time point. Finally, this invention combines the temperature field snapshots of all time points in chronological order to form a four-dimensional data structure containing time and spatial dimensions, obtaining a preliminary three-dimensional variable temperature field model.

[0059] S5: Render the preliminary model of the three-dimensional variable temperature field, output the three-dimensional visualization result, and compare and verify the test mining monitoring data with the three-dimensional variable temperature field model to obtain the three-dimensional variable temperature field verification model.

[0060] In step S5, the step of comparing and verifying the trial monitoring data with the three-dimensional variable temperature field model to obtain the three-dimensional variable temperature field verification model further includes: S51: Calculate the difference between the measured temperature value in the test mining monitoring data and the predicted temperature value at the corresponding position in the preliminary model of the three-dimensional variable temperature field to obtain the temperature error distribution.

[0061] In step S51, the present invention extracts the measured temperature value from the test mining monitoring data. This data includes the spatial coordinates of the monitoring point and the temperature value at the measurement time. Subsequently, based on the spatial coordinates of the monitoring point, the present invention searches for the corresponding grid cell at the corresponding time in the preliminary three-dimensional variable temperature field model and reads the predicted temperature value stored in that grid cell. Then, the present invention calculates the difference between the measured temperature value and the predicted temperature value to obtain the temperature error of a single monitoring point. Next, the present invention iterates through all monitoring points, repeats the above calculation process, and associates the temperature error value of each monitoring point with its spatial coordinates, marking the error magnitude in three-dimensional space to form a temperature error distribution.

[0062] S52: Based on the temperature error distribution, the interpolation parameters of the preliminary three-dimensional variable temperature field model are optimized in reverse. When the average absolute error of the temperature error distribution is less than a preset threshold, the three-dimensional variable temperature field verification model is output.

[0063] Furthermore, this invention statistically analyzes the absolute values ​​of temperature errors at all monitoring points in the temperature error distribution, calculates the arithmetic mean of all absolute error values ​​to obtain the mean absolute error (MAE), and then determines whether the MAE is less than a preset threshold of 1°C. If the MAE is greater than or equal to the preset threshold, this invention extracts the region with the largest absolute error value in the temperature error distribution, analyzes the geological structure characteristics and interpolation parameter settings of this region, adjusts the interpolation parameters such as the search radius and weight index of the interpolation method, re-executes steps S43 and S44, recalculates the preliminary model of the three-dimensional variable temperature field, and again executes step S51 to calculate a new MAE, repeating the above adjustment process. When the MAE is less than the preset threshold, this invention stops parameter adjustment and outputs the current preliminary model of the three-dimensional variable temperature field as a verification model of the three-dimensional variable temperature field.

[0064] like Figure 2 As shown, a three-dimensional variable temperature field modeling system for hydrate-rich regions is used to execute a three-dimensional variable temperature field modeling method for hydrate-rich regions as described in any of the above claims, including: Standardization module 100: used to receive the collected multi-source heterogeneous raw data streams, perform formatting and standardization processing on the multi-source heterogeneous raw data streams, and obtain a standardized dataset; Preprocessing module 200: used to call the data processing acceleration card to perform intelligent preprocessing on the standardized dataset to obtain a preprocessed dataset; Modeling module 300: used to perform geological modeling on the preprocessed dataset through hybrid logic to obtain a three-dimensional geological constraint model; Synthesis module 400: used to map the temperature and pressure data in the preprocessed dataset to the three-dimensional geological constraint model, and to perform spatiotemporal synthesis of the mapped temperature and pressure data by fusing time series data to obtain a preliminary three-dimensional variable temperature field model; Visualization module 500: used to render the preliminary model of the three-dimensional variable temperature field and output the three-dimensional visualization results; Verification module 600: Compares and verifies the test sampling monitoring data with the three-dimensional variable temperature field model to obtain the three-dimensional variable temperature field verification model.

[0065] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0066] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0067] In one specific embodiment, the three-dimensional variable temperature field modeling system for hydrate-rich areas of the present invention can be configured as a workstation for connecting to work equipment and performing data processing during field operations to execute the three-dimensional variable temperature field modeling method for hydrate-rich areas of the present invention.

[0068] like Figure 3As shown, workstation 1 is equipped with a multi-source data interface 3, which has various physical interfaces; workstation 1 has a data processing acceleration card inserted inside for hardware acceleration of data preprocessing algorithms; workstation 1 also has a central processing unit and a graphics processing unit installed inside, the central processing unit is used to carry and run the geological modeling processor and the variable temperature field synthesizer; the graphics processing unit is used to carry and run the three-dimensional graphics rendering unit; workstation 1 has an integrated display screen 2 installed on its surface for human-computer interaction and model display, and the multi-source data interface 3 is equipped with a clamping structure 4 for mechanically locking the data transmission line; workstation 1 is designed as a highly integrated dedicated computer, connected through an internal high-speed bus, realizing collaborative work efficiency and portability.

[0069] like Figure 4 and 5 As shown, the clamping structure 4 includes a clamping base 402, a clamping opening 408 on the clamping base 402, and multiple sliding grooves 403 on the upper side of the clamping base 402. A clamping block 404 is slidably connected inside the sliding groove 403, and a sliding block 405 is fixedly connected to the clamping block 404. A rotating ring 401 is rotatably connected to the clamping base 402, and a rotating plate 406 is fixedly connected to the lower side of the rotating ring 401. Multiple sliding grooves 407 are opened on the rotating plate 406, and the sliding block 405 is slidably connected to the sliding groove 407. Internally, when the connector of the data transmission line is inserted into the interface, rotating the rotating ring 401 drives the sliding block 405 and the clamping block 404 to move horizontally within the sliding groove 403 through the rotating plate 406 and the sliding groove 407 at its bottom. The multiple clamping blocks 404 tightly clamp the data line, thereby preventing the data line from falling out of the multi-source data interface 3. This setting significantly improves the reliability and durability of the workstation 1 in the field. In harsh environments with vibration, movement, or accidental contact by personnel, it can ensure a stable and uninterrupted data connection.

[0070] Furthermore, the sliding groove 403 is equipped with sliding damping, which can prevent the multiple clamping blocks 404 from moving on their own when there is no external force or when the data cable is under tension, thus avoiding the inability to clamp the data cable.

[0071] The three-dimensional variable temperature field modeling system for hydrate-rich areas of this invention is deployed and runs on a dedicated modeling workstation 1. The multi-source data interface 3 of the workstation 1 serves as the hardware carrier of the standardization module 100. It receives multi-source heterogeneous raw data streams from logging instruments, seismic data storage devices, and sensors in the test production area through various physical interfaces such as network ports, dedicated logging ports, and USB. The data transmission lines are mechanically locked by a clamping structure 4 to ensure stable and reliable data transmission. The standardization module 100 executes data format recognition and coordinate transformation algorithms through the central processing unit of the workstation 1 to complete the formatting and standardization processing of the multi-source data.

[0072] The preprocessing module 200 directly calls the data processing acceleration card plugged into the workstation 1. This acceleration card integrates an FPGA chip and performs random forest interpolation, isolated forest anomaly detection, and wavelet transform denoising algorithms through hardware parallel computing. The preprocessed dataset is then transmitted to the modeling module 300 through the internal high-speed bus of the workstation 1.

[0073] Both the modeling module 300 and the synthesis module 400 run on the central processing unit of workstation 1. The modeling module 300 calls the geological modeling processor carried by the central processing unit to perform fault constraint framework construction, three-dimensional structural model generation and lithofacies distribution calculation. The synthesis module 400 calls the variable temperature field synthesizer carried by the central processing unit to perform temperature and pressure data mapping and spatiotemporal synthesis processing.

[0074] The visualization module 500 runs on the graphics processor of workstation 1. It renders the preliminary model of the three-dimensional variable temperature field in real time through the three-dimensional graphics rendering unit carried by the graphics processor. The rendering result is output as a three-dimensional visualization result through the integrated display screen 2 installed on the surface of workstation 1, realizing human-computer interaction and model display.

[0075] The verification module 600 also runs on the central processing unit. It reads the user operation instructions received by the integrated display screen 2, calls the test mining monitoring data and compares and verifies it with the model, and feeds the verification results back to the synthesis module 400 for parameter optimization and adjustment, forming a closed-loop iterative optimization mechanism, and finally outputs a three-dimensional variable temperature field verification model.

[0076] 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 the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for modeling a three-dimensional variable temperature field in a hydrate-rich region, characterized in that, include: S1: Collect multi-source heterogeneous raw data streams, and perform formatting and standardization processing on the multi-source heterogeneous raw data streams to obtain a standardized dataset; S2: Call the data processing acceleration card to perform intelligent preprocessing on the standardized dataset to obtain a preprocessed dataset; S3: Geological modeling is performed on the preprocessed dataset using hybrid logic to obtain a three-dimensional geological constraint model; S4: Map the temperature and pressure data in the preprocessed dataset to the three-dimensional geological constraint model, and perform spatiotemporal synthesis of the mapped temperature and pressure data by fusing time series data to obtain a preliminary three-dimensional variable temperature field model. S5: Render the preliminary model of the three-dimensional variable temperature field, output the three-dimensional visualization result, and compare and verify the test mining monitoring data with the three-dimensional variable temperature field model to obtain the three-dimensional variable temperature field verification model.

2. The method for modeling a three-dimensional variable temperature field in a hydrate-rich region according to claim 1, characterized in that, Step S1 further includes: S11: Collect multi-source heterogeneous raw data streams through multi-source data interfaces; S12: Identify the format types of multiple data in the multi-source heterogeneous raw data stream, and convert the data of different formats according to the identification results to obtain data of a unified format. S13: Transform the spatial coordinates in the unified format data to a unified coordinate datum to obtain a standardized dataset.

3. The method for modeling a three-dimensional variable temperature field in a hydrate-rich region according to claim 2, characterized in that, The multi-source heterogeneous raw data stream in step S11 includes multi-dimensional data from various types of well points in different hydrate-rich areas. The multi-dimensional data includes formation fluid temperature data, hydrostatic pressure data, burial depth data, seafloor temperature data, and geothermal gradient data.

4. The method for modeling a three-dimensional variable temperature field in a hydrate-rich region according to claim 1, characterized in that, Step S2 further includes: S21: Based on the data processing acceleration card, the missing data in the standardized dataset is filled in parallel using the random forest interpolation algorithm to obtain a complete dataset; S22: Based on the data processing acceleration card, the isolated forest algorithm is used to calculate the path length of multiple data points in the complete dataset. Abnormal data points are obtained and removed according to the path length threshold to obtain a clean dataset. S23: Based on the data processing acceleration card, the high-frequency noise data in the clean dataset is decomposed and reconstructed using the wavelet transform algorithm to reduce noise, thereby obtaining a preprocessed dataset.

5. The method for modeling a three-dimensional variable temperature field in a hydrate-rich region according to claim 1, characterized in that, The hybrid logic in step S3 includes step-by-step hierarchical modeling logic, point-line-surface-volume construction logic, and phase sequence constraint logic.

6. The method for modeling a three-dimensional variable temperature field in a hydrate-rich region according to claim 5, characterized in that, Step S3 further includes: S31: Extract fault data from the preprocessed dataset, establish a high-level fault framework model, and construct a low-level fault model based on the high-level fault framework model to obtain a fault constraint framework. S32: Extract well logging interpretation stratification point data from the preprocessed dataset, spatially connect the stratification points of vertical wells and horizontal wells to form a formation boundary line, generate formation planes based on the formation boundary line, fill formation grid cells within the fault constraint framework, and obtain a three-dimensional structural model. S33: Combine deterministic modeling methods and stochastic modeling methods to calculate the lithofacies distribution of the three-dimensional structural model to obtain a lithofacies constrained model; S34: Under the constraints of the lithofacies constraint model, establish the attribute distribution model and the hydrate enrichment distribution model to obtain the three-dimensional geological constraint model.

7. A method for modeling a three-dimensional variable temperature field in a hydrate-rich region according to claim 6, characterized in that, The deterministic modeling method in step S33 is the ordinary kriging method, and the stochastic modeling method is the sequential instruction simulation method.

8. The method for modeling a three-dimensional variable temperature field in a hydrate-rich region according to claim 1, characterized in that, Step S4 further includes: S41: Extract the temperature and pressure data from the preprocessed dataset as temperature and pressure data, and perform spatial location matching between the temperature and pressure data and the grid cells in the three-dimensional geological constraint model; S42: Based on the geological structure characteristics of the three-dimensional geological constraint model and the spatial distribution characteristics of the temperature and pressure data, select the optimal interpolation method from the inverse distance weighting method, the kriging method and the triangulation method; S43: The temperature and pressure values ​​of the unassigned grid cells in the three-dimensional geological constraint model are calculated using the optimal interpolation method to obtain the initial temperature field model; S44: The temperature changes at different time points in the time series data are superimposed onto the initial temperature field model to obtain a preliminary three-dimensional variable temperature field model.

9. A method for modeling a three-dimensional variable temperature field in a hydrate-rich region according to claim 1, characterized in that, In step S5, the step of comparing and verifying the trial monitoring data with the three-dimensional variable temperature field model to obtain the three-dimensional variable temperature field verification model further includes: S51: Calculate the difference between the measured temperature value in the test mining monitoring data and the predicted temperature value at the corresponding position in the preliminary model of the three-dimensional variable temperature field to obtain the temperature error distribution; S52: Based on the temperature error distribution, the interpolation parameters of the preliminary three-dimensional variable temperature field model are optimized in reverse. When the average absolute error of the temperature error distribution is less than a preset threshold, the three-dimensional variable temperature field verification model is output.

10. A three-dimensional variable temperature field modeling system for hydrate-rich regions, used to execute a three-dimensional variable temperature field modeling method for hydrate-rich regions as described in any one of claims 1 to 9, characterized in that, include: Standardization module: Used to receive the collected multi-source heterogeneous raw data streams, perform formatting and standardization processing on the multi-source heterogeneous raw data streams, and obtain a standardized dataset; Preprocessing module: used to call the data processing acceleration card to perform intelligent preprocessing on the standardized dataset to obtain a preprocessed dataset; Modeling module: Used to perform geological modeling on the preprocessed dataset through hybrid logic to obtain a three-dimensional geological constraint model; Synthesis module: used to map the temperature and pressure data in the preprocessed dataset to the three-dimensional geological constraint model, and to perform spatiotemporal synthesis of the mapped temperature and pressure data by fusing time series data to obtain a preliminary three-dimensional variable temperature field model; Visualization module: used to render the preliminary model of the three-dimensional variable temperature field and output three-dimensional visualization results; Verification module: The test sampling monitoring data is compared and verified with the three-dimensional variable temperature field model to obtain the three-dimensional variable temperature field verification model.

Images