Three-dimensional dynamic prediction method and system for deep brine resource quantity

Abstract

The invention discloses a three-dimensional dynamic prediction method and system for deep brine resource quantity, particularly relates to the technical field of mineral resource exploration, and is used for solving the problem of resource evaluation misalignment caused by a well group interference effect in intensive mining. The method comprises the following steps: constructing a spatial topology network model which takes a well location as a node and a well spacing as an edge weight by obtaining a mining area well group space coordinate and mining dynamic time sequence data; excavating pressure conduction characteristics, resource concentration migration characteristics and shaft temperature time sequence data; calculating a thermal buoyancy gradient based on temperature data, determining a stress deflection angle in combination with a network model azimuth angle, and fusing to construct an interference strength coefficient matrix; tracking an inter-well resource siphon path by using a network edge weight, and generating a time-lag compensation factor by coupling a time-lag rule of pressure and concentration characteristics; finally, the interference matrix, the siphon path and the time delay factor are fused to dynamically correct the single well evaluation value, the total area brine resource amount is output, and the intensive mining area resource evaluation precision is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of mineral resource exploration technology, and in particular to a method and system for three-dimensional dynamic prediction of deep brine resources. Background Technology

[0002] In the field of deep brine mineral resource exploration, the current mainstream method relies on the single-well independent modeling strategy to achieve efficient resource assessment. This strategy collects geological parameters and mining dynamic data of the target well location, constructs a local reservoir model and calculates resource reserves. Finally, the assessment results of all single wells in the region are superimposed as the overall resource quantity. This method has the advantage of being easy to implement in the early stage of decentralized mining and has become the basic paradigm for resource assessment in the industry.

[0003] However, when the mining area enters the stage of intensive development, the densely deployed well clusters generate significant spatial interference effects during the collaborative mining process. Existing single-well independent modeling methods cannot quantify the interaction of resource flows between wells, resulting in a systematic deviation of the regional total resource assessment results from the actual value, leading to inaccurate resource planning decisions. Summary of the Invention

[0004] This invention addresses the technical problems existing in the prior art by providing a three-dimensional dynamic prediction method and system for deep brine resources.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0006] This invention provides the following technical solution:

[0007] A three-dimensional dynamic prediction method for deep brine resources includes:

[0008] S1. Obtain the spatial coordinates of the well group and the dynamic time-series data of mining in the target mining area;

[0009] S2. Construct a spatial topology network model based on the spatial coordinates of the well group. The spatial topology network model uses well locations as nodes and the distance between wells as edge weights.

[0010] S3. Perform industrial data mining on the dynamic time series data of mining to extract the pressure transmission characteristics, resource concentration migration characteristics and wellbore temperature time series data of each well location.

[0011] S4. Calculate the thermal buoyancy gradient based on the wellbore temperature time series data, and determine the stress deflection angle based on the inter-well azimuth angle of the spatial topology network model. Combine the thermal buoyancy gradient and the stress deflection angle to construct the interference intensity coefficient matrix.

[0012] S5. Based on the spatial topology network model, the edge weight distribution dynamically tracks the resource siphon path between wells, and generates a time delay compensation factor based on the statistical time delay law of pressure transmission characteristics and resource concentration migration characteristics.

[0013] S6, the coupling interference intensity coefficient matrix, the inter-well resource siphon path and time delay compensation factor are used to correct the single-well resource assessment value, and the total amount of brine resources in the region is output.

[0014] Furthermore, obtain the spatial coordinates of the well clusters and the dynamic time-series data of mining in the target mining area, including:

[0015] The geodetic coordinates of the wellheads of all wells in the target mining area are retrieved from the drilling engineering database as the spatial coordinates of the well group.

[0016] Simultaneously collect the bottom pressure sensor readings, brine ion concentration detection values, and wellbore temperature sensor readings of each mining well, and generate mining dynamic time-series data according to a unified timestamp;

[0017] Verify the consistency between the spatial coordinates of the well group and the well location numbers in the dynamic time-series data of mining, and delete abnormal data segments with missing timestamps.

[0018] Furthermore, a spatial topology network model is constructed based on the spatial coordinates of the well cluster. This model uses well locations as nodes and inter-well distances as edge weights, and includes:

[0019] The obtained wellhead geodetic coordinates are used as nodes in the spatial topology network model;

[0020] Calculate the three-dimensional Euclidean distance between any two nodes as the well distance;

[0021] An undirected graph network model is constructed using the distance between wells as the edge weight;

[0022] The output is a spatial topology network model that includes node locations and edge weights.

[0023] Furthermore, industrial data mining is performed on the dynamic time-series data of mining operations to extract pressure transmission characteristics, resource concentration migration characteristics, and wellbore temperature time-series data for each well location, including:

[0024] Time series analysis was performed on the readings of the bottom hole pressure sensor to extract the pressure change rate as a pressure transmission characteristic.

[0025] Concentration gradient calculations were performed on the measured values ​​of brine ion concentration to determine the concentration change per unit time as a characteristic of resource concentration migration.

[0026] The time series of wellbore temperature sensor readings is directly extracted as wellbore temperature time series data.

[0027] Pressure transmission characteristics, resource concentration migration characteristics, and wellbore temperature time series data are classified and stored according to well location number.

[0028] Furthermore, the thermal buoyancy gradient is calculated based on wellbore temperature time-series data, and the stress deflection angle is determined based on the inter-well azimuth angle of the spatial topology network model. The thermal buoyancy gradient and stress deflection angle are then fused to construct an interference intensity coefficient matrix, including:

[0029] The temperature difference between adjacent time points is calculated based on the time series data of wellbore temperature. The thermal buoyancy gradient is obtained by multiplying the temperature difference by the product of the thermal expansion coefficient of brine and the gravitational acceleration.

[0030] The inter-well azimuth angle is extracted from the spatial topology network model. The direction of maximum principal stress is determined based on the inter-well azimuth angle. The direction of maximum principal stress is then converted into stress deflection angle using the Coulomb fracture criterion.

[0031] The interference coupling strength value of the corresponding well pair is obtained by performing a vector dot product operation between the thermal buoyancy gradient and the stress deflection angle of the same well pair.

[0032] Traverse all well pair combinations and construct an interference intensity coefficient matrix using the well location number as the row and column index. The matrix element values ​​are the interference coupling intensity values ​​of the corresponding well pair.

[0033] Furthermore, converting the direction of maximum principal stress into stress deflection angle using the Coulomb fracture criterion includes:

[0034] The fracture angle is calculated based on the internal friction angle obtained from the core experiment.

[0035] Using the direction of maximum principal stress as the reference direction, the fracture angle is taken as the stress deflection angle;

[0036] The fracture angle is calculated using the Coulomb fracture criterion formula.

[0037] Furthermore, based on the edge weight distribution of the spatial topology network model, the inter-well resource siphon path is dynamically tracked. Simultaneously, a time-delay compensation factor is generated based on the statistical time-delay laws of pressure transmission characteristics and resource concentration migration characteristics, including:

[0038] Obtain the siphon path for inter-well resources;

[0039] Simultaneously extract the peak time point of pressure transmission characteristics and the valley time point of resource concentration migration characteristics at the same well location, and calculate the time difference between the two as the basic time delay;

[0040] The distribution pattern of basic time delay was statistically analyzed according to the mining stage, and a functional mapping relationship between basic time delay and reservoir viscosity was established.

[0041] The output value of the function mapping relationship in the current mining period is used as the time delay compensation factor.

[0042] Furthermore, the specific method for obtaining the inter-well resource siphon path is as follows: high-pressure well locations and low-pressure well locations are identified using pressure transmission characteristics, and the minimum resistance path from the high-pressure well location to the low-pressure well location is calculated using the edge weights of the spatial topology network model as the migration resistance coefficient, which is then used as the inter-well resource siphon path.

[0043] Furthermore, the single-well resource assessment value is corrected by coupling the interference intensity coefficient matrix, the inter-well resource siphon path, and the time lag compensation factor, and the total regional brine resource is output, including:

[0044] The time-delay compensation factor is multiplied by the single-well resource assessment value to obtain the time-corrected assessment value;

[0045] For each pair of high-pressure and low-pressure well locations on the inter-well resource siphon path, the element values ​​of the corresponding well location pair in the interference intensity coefficient matrix are extracted as migration weight coefficients.

[0046] The resource migration amount is obtained by multiplying the time-corrected assessment value of the high-pressure well location by the migration weight coefficient.

[0047] Subtract the resource migration amount from the high-pressure well location time correction assessment value and add the resource migration amount to the low-pressure well location time correction assessment value;

[0048] The total amount of brine resources in the region is obtained by summing up the time-corrected assessment values ​​after all well location adjustments.

[0049] On the other hand, the present invention provides a three-dimensional dynamic prediction system for deep brine resources, comprising:

[0050] The data acquisition module is used to acquire the spatial coordinates of the well clusters and the dynamic time-series data of mining in the target mining area;

[0051] The network modeling module is used to construct a spatial topology network model based on the spatial coordinates of the well group. The spatial topology network model uses well locations as nodes and the distance between wells as edge weights.

[0052] The feature mining module is used to perform industrial data mining on mining dynamic time-series data, extracting pressure transmission characteristics, resource concentration migration characteristics, and wellbore temperature time-series data for each well location.

[0053] The coupled calculation module is used to calculate the thermal buoyancy gradient based on the wellbore temperature time series data, and at the same time determine the stress deflection angle based on the inter-well azimuth angle of the spatial topology network model. The thermal buoyancy gradient and stress deflection angle are fused to construct the interference intensity coefficient matrix.

[0054] The dynamic tracking module is used to dynamically track the resource siphon path between wells based on the edge weight distribution of the spatial topology network model, and at the same time generate a time delay compensation factor based on the statistical time delay law of pressure transmission characteristics and resource concentration migration characteristics.

[0055] The resource integration module is used to couple the interference intensity coefficient matrix, the inter-well resource siphon path, and the time lag compensation factor to correct the single-well resource assessment value and output the total amount of brine resources in the region.

[0056] The beneficial effects of this invention are:

[0057] 1. By combining spatial topology network modeling with dynamic feature coupling, the problem of resource assessment distortion caused by well group interference in intensive mining is effectively solved. First, the thermal buoyancy gradient is calculated based on the time series data of wellbore temperature. Combined with the stress deflection angle derived from the spatial topology network model, an interference intensity coefficient matrix is ​​constructed to achieve the synergistic quantification of thermodynamic effects and geological stress fields, accurately characterizing the intensity of inter-well thermo-mechanical coupling. Second, a time delay compensation factor is generated by analyzing the time delay law of pressure transmission characteristics and resource concentration migration characteristics. This dynamically corrects the time series deviation of resource migration caused by differences in reservoir properties, significantly improving the timeliness and accuracy of single-well assessment values.

[0058] 2. Based on the spatial topology network model, the edge weight distribution dynamically tracks the resource siphon path between wells. Combined with the migration weight coefficient of the interference intensity coefficient matrix, the resource migration from high-pressure wells to low-pressure wells is accurately quantified. By coupling the time delay compensation factor, resource siphon path, and interference intensity coefficient matrix, the resource assessment value of a single well is collaboratively corrected, forming a three-dimensional dynamic assessment closed loop of "time series correction - spatial migration - coupling feedback". This breaks through the limitations of static superposition of traditional single-well independent modeling. The well group interference effect is naturally integrated in the total resource calculation, so that the assessment results truly reflect the resource distribution law under intensive mining conditions, providing a reliable basis for mining area development decisions. Attached Figure Description

[0059] Figure 1 This is a flowchart of the three-dimensional dynamic prediction method for deep brine resources according to the present invention;

[0060] Figure 2 This is a schematic diagram of the three-dimensional dynamic prediction system for deep brine resources according to the present invention. Detailed Implementation

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

[0062] Example 1

[0063] Figure 1 The present invention provides a three-dimensional dynamic prediction method for deep brine resources, comprising:

[0064] S1. Obtain the spatial coordinates of the well group and the dynamic time-series data of mining in the target mining area;

[0065] S2. Construct a spatial topology network model based on the spatial coordinates of the well group. The spatial topology network model uses well locations as nodes and the distance between wells as edge weights.

[0066] S3. Perform industrial data mining on the dynamic time series data of mining to extract the pressure transmission characteristics, resource concentration migration characteristics and wellbore temperature time series data of each well location.

[0067] S4. Calculate the thermal buoyancy gradient based on the wellbore temperature time series data, and determine the stress deflection angle based on the inter-well azimuth angle of the spatial topology network model. Combine the thermal buoyancy gradient and the stress deflection angle to construct the interference intensity coefficient matrix.

[0068] S5. Based on the spatial topology network model, the edge weight distribution dynamically tracks the resource siphon path between wells, and generates a time delay compensation factor based on the statistical time delay law of pressure transmission characteristics and resource concentration migration characteristics.

[0069] S6, the coupling interference intensity coefficient matrix, the inter-well resource siphon path and time delay compensation factor are used to correct the single-well resource assessment value, and the total amount of brine resources in the region is output.

[0070] When performing data acquisition operations, the drilling engineering database is accessed first. This database stores engineering archive data for all wells in the target mining area. Wellhead geodetic coordinate data is retrieved using Structured Query Language (SCL) commands. This data includes three spatial dimensions: longitude, latitude, and elevation. This data is obtained using coordinate transformation tools certified by the surveying department. For example, the coordinate record for a certain well might be longitude 112.53 degrees, latitude 38.42 degrees, and elevation 850.3 meters. The retrieved wellhead geodetic coordinate data for all wells is then output as a well group spatial coordinate dataset.

[0071] A dynamic data acquisition process is initiated synchronously, connecting monitoring equipment in each mining well via an industrial IoT interface. For the bottom-hole pressure sensor, pressure measurements are read in real time, with the unit being megapascals (MPA). The sensor range is set to 1.5 times the historical maximum bottom-hole pressure value in the mining area. For the brine ion concentration detector, the concentration values ​​of the main ions in the brine are obtained, with the unit being grams per liter (g / L). The detection range covers 0–300 g / L to meet the characteristics of deep brine. For the wellbore temperature sensor, temperature measurements are collected, with the unit being degrees Celsius. The range is set to 0–150 degrees Celsius. All measuring devices are equipped with a time synchronization module, achieving time synchronization via a satellite positioning system to ensure that the timestamps of the acquired data are accurate to the second, with a timestamp deviation of less than 50 milliseconds. The bottom-hole pressure sensor readings, brine ion concentration detection values, and wellbore temperature sensor readings at the same timestamp are bound into a complete record and arranged chronologically to generate a dynamic mining time-series data table.

[0072] During data validation, a well location number comparison mechanism is established. The unique number corresponding to each well location in the well group spatial coordinate dataset is extracted and matched item by item with the well location number field of each record in the mining dynamic time-series data table. The number matching uses a string exact match algorithm, performing case-sensitive comparison operations. If a well location number exists in the well group spatial coordinate dataset but has no corresponding record in the mining dynamic time-series data table, it is determined to be a data missing anomaly; if a well location number in a record in the mining dynamic time-series data table has no corresponding entry in the well group spatial coordinate dataset, it is determined to be a number inconsistency anomaly. For time dimension validation, the timestamp continuity of the mining dynamic time-series data table is checked: the time interval between adjacent records is calculated, and when the time interval exceeds a preset threshold, it is marked as an abnormal data segment. For example, pressure data uses a 1-minute base period, and a continuous missing period of more than 3 minutes is considered abnormal; concentration data uses a 30-minute base period, and a continuous missing period of more than 90 minutes is considered abnormal. Finally, delete all records with inconsistent numbers and all data entries corresponding to abnormal data segments. Record the deleted well location number, abnormal time period and abnormal type in the log file, and output the verified well group spatial coordinate dataset and mining dynamic time series data table.

[0073] In a specific implementation case, a mining area contained 32 mining wells. Retrieving the wellhead geodetic coordinate data took 1.2 seconds, collecting a total of 230,400 records of dynamic mining data over 5 days. During the verification process, an anomaly was found in the numbering of well No. 3; for example, the coordinate data showed the number "JK-03" while the dynamic data showed "JK03". Additionally, well No. 7 lacked pressure data for a specific period, such as from 14:00 to 16:00 on May 12, 2023. After performing a deletion operation, the final output included the valid spatial coordinates of 31 wells and 228,160 dynamic data records.

[0074] The specific implementation of the data acquisition equipment includes: a piezoresistive pressure sensor installed at the bottom of the well; an online analytical instrument used for brine ion concentration detection, automatically completing a full ion analysis every 30 minutes; and a platinum resistance thermometer installed at the target depth downhole for well temperature sensing. Time synchronization is achieved through a time synchronization module, with each monitoring device's built-in clock module automatically calibrating the time every 24 hours.

[0075] The output data format specifications are as follows: the well cluster spatial coordinate dataset stores fields including well number, longitude, latitude, and elevation; the mining dynamic time series data table stores fields including timestamp, well number, pressure, chloride ion concentration, sodium ion concentration, and temperature. All output data undergoes data integrity verification to ensure that subsequent processing steps can obtain valid input.

[0076] Data storage management employs a tiered strategy: raw collected data is retained in a temporary buffer, and a data modification log is generated during the verification process; valid output data is transferred to a file storage system, with a directory structure established according to mining area number and date. A full backup mechanism is performed daily at midnight to ensure data traceability.

[0077] The key parameter acquisition process includes: setting the range of the bottom-hole pressure sensor based on the historical maximum pressure value of the mining area; for example, if the historical maximum pressure of a certain mining area is 40 MPa, then the range is set to 0–60 MPa; setting the ion concentration detection range based on the historical maximum salinity of the brine in the mining area; for example, if the highest salinity of a certain mining area is 280 g / L, then the upper limit of detection is set to 300 g / L; setting the temperature sensor range based on the highest recorded temperature downhole; for example, if the highest temperature of a certain mining area is 135 degrees Celsius, then the range is set to 0–150 degrees Celsius. All equipment parameters were verified through on-site calibration before project implementation.

[0078] When constructing the spatial topology network model, the first step is to obtain the well cluster spatial coordinate dataset output from the data verification step. This dataset contains a unique number for each well location and its corresponding three-dimensional spatial coordinates. Longitude and latitude values ​​are recorded in decimal degrees, and elevation values ​​are recorded in meters. For example, a mining area contains 15 well locations, where the coordinates of well number 1 are recorded as longitude 115.62 degrees, latitude 28.75 degrees, and elevation 1024.8 meters. The spatial coordinates of each well location are used as nodes in the network model, with the well location number serving as the node identifier to ensure strict correspondence with the previous data.

[0079] When calculating the spatial distance between any two nodes, a three-dimensional spatial distance formula is used. The specific calculation process is as follows: For two different nodes, their longitude, latitude, and elevation values ​​are obtained respectively. First, the geographic coordinate system coordinates are converted to a spatial rectangular coordinate system. The conversion method is as follows: X-coordinate equals the sum of the radius of curvature of the zonal circle and the elevation value multiplied by the cosine of the latitude and the cosine of the longitude; Y-coordinate equals the sum of the radius of curvature of the zonal circle and the elevation value multiplied by the cosine of the latitude and the sine of the longitude; Z-coordinate equals the radius of curvature of the zonal circle multiplied by (1 minus the square of the first eccentricity) plus the elevation value multiplied by the sine of the latitude. Then, the straight-line distance between the two points is calculated: the distance value equals the square of the difference in X-coordinates plus the square of the difference in Y-coordinates plus the square of the difference in Z-coordinates, and finally, the square root is taken. The calculation result is in meters, accurate to two decimal places. For example, when calculating the distance between well No. 1 and well No. 2, if the coordinates of well No. 1 are 115.62 degrees longitude, 28.75 degrees latitude, and 1024.8 meters elevation, and the coordinates of well No. 2 are 115.63 degrees longitude, 28.76 degrees latitude, and 1025.2 meters elevation, the distance after conversion is approximately 135.42 meters. The distance calculation for all well location combinations is performed using this method.

[0080] When constructing an undirected graph network model, each well location is treated as a vertex, and the calculated inter-well distance is used as the weight of the edge connecting the corresponding vertex. The model construction follows these rules: each vertex has a unique identifier; each edge connects two different vertices; the edge weights are positive real numbers and satisfy symmetry. The network model is stored using a matrix data structure, where the matrix row and column indices correspond to the well location numbering order, and the matrix element values ​​store the inter-well distance values ​​for corresponding well location pairs. For example, when there are multiple well locations in a mining area, a symmetric matrix of appropriate size is constructed.

[0081] Special case handling mechanisms include: initiating a coordinate verification program when the calculation finds that the distance between two well locations is less than a set threshold, for example, retrieving original survey data for verification when the distance value is less than 0.5 meters; recording an anomaly in the log and marking it as a duplicate node when identical coordinate values ​​are found. Validity verification is set during the distance calculation process: checking whether the calculation results are within a reasonable geological range, for example, the distance between wells should be between 50 meters and 5000 meters; values ​​outside this range trigger a manual review process.

[0082] When outputting the spatial topology network model, complete node location data and edge weight data are included. Node location data is stored in a data table format, including fields for well location number, longitude, latitude, and elevation. Edge weight data is stored in a matrix format, with a backup in edge list format provided. Each edge record includes three fields: start node number, end node number, and distance value. The model data file is stored in an open format to ensure direct readability in subsequent processing steps.

[0083] In a specific implementation case, a brine mining area contains 32 valid well locations, requiring the calculation of distance values ​​for 32 × 31 / 2 = 496 well location combinations. The calculation process is completed using an optimized algorithm, finishing the entire calculation task in a short time. In the constructed matrix, diagonal elements have a value of 0, while off-diagonal elements store the calculated distance values. The final output file includes a node data table and an edge data table.

[0084] The data verification mechanism includes: random sampling to verify the accuracy of distance calculations, such as selecting several well location combinations for manual verification; checking network connectivity to ensure there are no isolated nodes; and verifying the symmetry of edge weights. A diagnostic report is automatically generated when anomalies are detected.

[0085] The parameter settings are based on the following: the coordinate system parameters used for distance calculation are obtained from the regional geodetic coordinate system parameters provided by the surveying and mapping department; the distance threshold is set based on the historical data of the minimum well spacing in the mining area. For example, if the minimum measured well spacing in a certain area is 3.2 meters, then the abnormal threshold is set to 0.5 meters; the effective range of distance is determined according to the area of ​​the mining area.

[0086] Optimization methods for the calculation process include: pre-establishing a coordinate transformation parameter lookup table; using spatial indexing technology to accelerate the calculation of nearest neighbor distances; and adopting a block-based processing mechanism for memory management.

[0087] When performing industrial data mining operations, a dynamic time-series data table of mining operations is obtained, output through the data verification step. This data table contains records arranged by a uniform timestamp. Each record includes the well location number, the bottom hole pressure sensor reading (in megapascals), the brine ion concentration detection value (in grams per liter), and the wellbore temperature sensor reading (in degrees Celsius). For example, a record might have the timestamp 2023-06-01-06-08:00:00, well location number JK-07, pressure value 42.3 megapascals, chloride ion concentration value 185 grams per liter, and temperature value 89.5 degrees Celsius.

[0088] When extracting pressure transmission characteristics, the time series of pressure sensor readings for each well location is processed. The specific execution process is as follows: For the pressure data series of a single well location, a pressure-time curve is generated in chronological order. A sliding window method is used to analyze the time series, with the window size set to multiple consecutive sampling points. The pressure change rate is calculated within the window: the pressure values ​​at the beginning and end of the window are taken, the difference between them is calculated, and then divided by the window time span to obtain the change rate value, in megapascals per minute (MPa). For example, if the window time span is 6 minutes, the initial pressure value is 42.1 MPa, and the ending pressure value is 42.5 MPa, then the change rate is the difference between the ending pressure value and the initial pressure value, divided by the time span, i.e., (42.5-42.1) / 6, resulting in approximately 0.067 MPa per minute. The window slides chronologically with a step size of one sampling point, generating a complete pressure change rate sequence. The calculation results are post-processed: a rate threshold range is set, and values ​​exceeding the range are marked as outliers; outliers are corrected using linear interpolation, replacing them with the average of adjacent normal values. The final output pressure transmission characteristic is the pressure change rate value corresponding to each time stamp.

[0089] When extracting resource concentration migration characteristics, the time series of brine ion concentration detection values ​​is processed. The specific execution process is as follows: for the ion concentration data of each well location, a concentration-time curve is generated in chronological order. The concentration gradient is calculated using the central difference method: for time point t, the concentration value C1 at t minus the time interval and the concentration value C2 at t plus the time interval are taken. The concentration change per unit time is calculated as the difference between the concentration value at the later time and the concentration value at the earlier time, divided by twice the time interval value. For example, if the time interval is 30 minutes, the concentration value before time t is 182 g / L, and the concentration value after time t is 188 g / L, then the change is the difference between the later and earlier concentration values ​​divided by the time unit conversion factor, i.e., (188-182) / 1 hour, resulting in 6 g / L per hour. The calculation results are validated: when the absolute value of the concentration change exceeds a set threshold, a data review mechanism is triggered; continuous outliers are smoothed using a moving average method. The final output resource concentration migration characteristic is the concentration change value per unit time corresponding to each timestamp.

[0090] When extracting wellbore temperature time-series data, the temperature sensor reading sequence is read directly. The specific process is as follows: Indexed by well location number, all data from the temperature sensor reading field are extracted from the mining dynamic time-series data table to form the raw temperature time series. Data preprocessing includes: detecting and removing outliers, such as data with temperatures below 0 degrees Celsius or above 150 degrees Celsius; filling missing points with the average of adjacent values; and marking temperature jump points. The output is a temperature series that maintains the original sampling frequency.

[0091] When integrating and storing feature data, a structured feature database is established. The specific execution process is as follows: Independent data tables are created according to well location numbers. Each table contains a timestamp field, a pressure transmission feature value field, a resource concentration migration feature value field, and a wellbore temperature value field. The data association mechanism is as follows: the three types of feature values ​​for the same well location and the same timestamp are stored in the same record row. For example, the record for well location JK-07 at a specific timestamp includes a pressure change rate value, a chloride ion concentration change value, and a temperature value. The storage format adopts a time-series database partitioned storage, establishing a hierarchical directory structure based on date and well location number.

[0092] Data quality control measures include: logging the feature calculation process; defining common anomaly patterns and handling strategies in an anomaly handling rule base; and verifying data integrity to ensure the completeness of features at each timestamp. An automatic alarm mechanism is implemented: an alarm is sent when the feature loss rate for a certain well location exceeds a set percentage.

[0093] The parameter settings are based on the following: the size of the pressure change rate calculation window is set according to the pressure fluctuation cycle; the concentration gradient calculation time interval is set according to the concentration detection frequency; and the temperature anomaly threshold is set according to the temperature records of the mining area. All parameters can be adjusted in the configuration file.

[0094] The optimization of the calculation process includes: parallel computing for pressure feature extraction; preloading time index for concentration gradient calculation; and establishing a memory cache for temperature data.

[0095] In a specific implementation case, dynamic data from multiple well locations in a mining area were used to generate feature records after feature extraction. The processing time met expectations. Random sampling verification showed that the calculation error was within the allowable range.

[0096] When performing thermal buoyancy gradient calculation, the wellbore temperature time-series data output from the industrial data mining process is acquired. This data is stored categorized by well location number and contains a sequence of temperature values ​​corresponding to consecutive timestamps, with the temperature values ​​in degrees Celsius. For each well location, the temperature data is processed in chronological order: the temperature values ​​at adjacent time points t and t plus a time interval Δt are taken, and the temperature difference is calculated as the temperature value at the later time point minus the temperature value at the earlier time point. The time interval Δt is determined based on the data sampling frequency. For example, when the sampling interval is 30 minutes, the temperature value at time t is 89.5 degrees Celsius, and the temperature value at t plus 30 minutes is 90.1 degrees Celsius. The temperature difference is 90.1 - 89.5 = 0.6 degrees Celsius. The thermal buoyancy gradient calculation process involves multiplying the temperature difference by the thermal expansion coefficient of the brine, and then multiplying it by the acceleration due to gravity. The thermal expansion coefficient of the brine was obtained through laboratory calibration. The specific calibration method involved using a thermal expansion coefficient measuring instrument to measure the volumetric expansion rate of the brine sample for every 1 degree Celsius increase in temperature under constant pressure. For example, the average thermal expansion coefficient of a brine sample from a certain mining area measured within the 20-100 degree Celsius range was 0.00038 per degree Celsius. The gravitational acceleration value was obtained from on-site measurements in the mining area using a high-precision gravimeter. For example, the measured value in a certain mining area was 9.81 meters per second squared. The final thermal buoyancy gradient unit is Newtons per cubic meter. A calculation example: for a temperature difference of 0.6 degrees Celsius, the thermal buoyancy gradient = 0.00038 × 0.6 × 9.81 ≈ 0.0022 Newtons per cubic meter. The calculation results were stored in a time series, forming the time-series data of the thermal buoyancy gradient for each well location.

[0097] When determining the stress deflection angle, the inter-well azimuth data in the spatial topology network model is used. The inter-well azimuth angle is defined as the angle between the direction from the current well location to the adjacent well location and the geographic north direction, in degrees. The specific extraction method is as follows: For the target well location A and the adjacent well location B, the three-dimensional coordinate values ​​of the two well locations are obtained from the edge weight data table of the spatial topology network model, and the azimuth angle is calculated inversely using the spatial coordinates. The calculation process is as follows: First, the latitude and longitude coordinates are converted to plane rectangular coordinates, and then the azimuth angle is calculated using the coordinate difference. Example: Well location A has coordinates of longitude 115.62 degrees and latitude 28.75 degrees, and well location B has coordinates of longitude 115.63 degrees and latitude 28.76 degrees. After coordinate conversion, the calculated azimuth angle is 120 degrees. The stress deflection angle is converted according to the Coulomb fracture criterion: First, the internal friction angle measured by core experiments is obtained. The core experiment method is as follows: standard core samples of underground rock strata in the target mining area are drilled, a preset confining pressure is applied in a triaxial testing machine, and shear force is applied at a constant rate until the rock sample fractures. The relationship curve between shear stress and normal stress is plotted, and the arctangent value of the slope of the linear segment of the curve is taken as the internal friction angle. For example, the internal friction angle measured in a rock sample from a certain mining area is 30 degrees. The fracture angle is calculated by subtracting half of the internal friction angle from 45 degrees. Calculation example: when the internal friction angle is 30 degrees, the fracture angle = 45 - 30 / 2 = 30 degrees. Using the direction of the maximum principal stress indicated by the inter-well azimuth as the reference direction, the fracture angle is directly output as the stress deflection angle. Example: when the inter-well azimuth is 120 degrees, the stress deflection angle output value is 30 degrees.

[0098] When calculating the interference coupling strength, the thermal buoyancy gradient and stress deflection angle of the same well pair are fused. The specific process is as follows: the thermal buoyancy gradient is considered a vector, its direction is vertically upward along the wellbore axis, and its modulus is the calculated value of the thermal buoyancy gradient; the stress deflection angle is converted into a direction vector, its modulus is set to 1, and its direction is consistent with the direction of the maximum principal stress. The formula for the dot product of these two vectors is: the interference coupling strength value equals the thermal buoyancy gradient vector modulus multiplied by the cosine of the stress deflection angle direction vector, and then multiplied by the cosine of the angle between the two vectors. The vector angle is determined through spatial geometric relationships: the angle is 0 degrees when the thermal buoyancy gradient direction is parallel to the direction of the maximum principal stress, and 90 degrees when it is perpendicular. Calculation Example: At a certain time point, the thermal buoyancy gradient is 0.0022 N / m³, and the stress deflection angle is 30 degrees. When the two vectors are in the same direction, the interference coupling strength is 0.0022 × 1 × cos0° = 0.0022 N / m³; when there is a 45-degree angle, the interference coupling strength is 0.0022 × cos45° ≈ 0.0016 N / m³. This value characterizes the intensity of the synergistic effect between thermal buoyancy and the geostress field.

[0099] When constructing the interference intensity coefficient matrix, all well pair combinations are traversed. The matrix row and column indices correspond to the well location numbering order, and the matrix element values ​​store the interference coupling intensity values ​​of the corresponding well pair. Specifically, the construction rules are as follows: for well locations numbered i and j, when i is not equal to j, the interference coupling intensity value at the latest time point of the two well locations is taken and filled into the element in the i-th row and j-th column of the matrix; when i equals j (i.e., the diagonal element), it is uniformly set to 0. The matrix elements satisfy the symmetry requirement: the value of the element in the i-th row and j-th column is equal to the value of the element in the j-th row and i-th column. The matrix storage adopts a sparse matrix compression format, only storing non-zero elements and their row and column indices. Implementation Example: A mining area contains 3 well locations (numbered A, B, and C). The matrix is ​​constructed in a 3x3 matrix, where: the element in the 1st row and 2nd column stores the interference coupling strength value between well locations A and B; the element in the 1st row and 3rd column stores the interference coupling strength value between well locations A and C; the element in the 2nd row and 1st column stores the interference coupling strength value between well location B and A (its value is the same as AB); the element in the 2nd row and 3rd column stores the interference coupling strength value between well locations B and C; the element in the 3rd row and 1st column stores the interference coupling strength value between well locations C and A; and the element in the 3rd row and 2nd column stores the interference coupling strength value between well locations C and B.

[0100] The acquisition process for key parameters is fully transparent: The determination of the thermal expansion coefficient of the brine requires the use of a professional dilatometer. Within a temperature range of 20 to 100 degrees Celsius, the temperature is increased at a rate of 1 degree Celsius per minute, and the volume change is recorded for each 1 degree Celsius increase in temperature. This experiment is repeated three times, and the arithmetic mean is taken as the calibration value. Gravitational acceleration correction is based on the elevation of the mining area. The theoretical value is calculated using the international gravity formula, and then a terrain correction is added. For example, the corrected gravitational acceleration value for a mining area at an elevation of 1000 meters is 9.80 m / s². The internal friction angle experiment must be performed according to the recommended procedures of the International Society for Rock Mechanics: Core samples are processed into standard cylinders (50 mm in diameter and 100 mm in height). The confining pressure gradient is set to three levels: 5 MPa, 10 MPa, and 15 MPa. The shear rate is controlled at 0.5 mm / min. Three samples are tested for each confining pressure group, and the final average of all valid test results is taken.

[0101] Data validation and anomaly handling mechanisms: The thermal buoyancy gradient value range verification sets a reasonable threshold range, for example, a lower limit of -0.01 N / m³ and an upper limit of +0.01 N / m³. If the value exceeds the threshold, a temperature data review process is automatically triggered. If the data is still abnormal after the review, the average value of the previous 6 time points is used instead. The stress deflection angle logic check requires the angle value to be within the range of 0 to 90 degrees. Abnormal values ​​are automatically replaced with the average stress of three adjacent well locations within the same structural unit. Matrix symmetry verification is completed by traversing all off-diagonal element pairs. When the difference between the element in row i, column j and the element in row j, column i exceeds the tolerance (e.g., 0.0001 N / m³), it is automatically corrected to the arithmetic mean of the two elements.

[0102] Boundary conditions and special handling: When encountering a newly commissioned well location, add corresponding rows and columns to the matrix, and set the initial element values ​​to the arithmetic mean of the interference coupling strength values ​​of adjacent well locations. If temperature data for a well location is continuously missing for more than 24 hours, suspend calculations related to that well location and record a warning message in the log. When multiple strata exist in the mining area, establish independent matrices according to vertical stratification, with stratification based on an elevation difference of more than 50 meters between the top and bottom plates of the mining strata.

[0103] In a specific implementation case, the construction process of the interference intensity coefficient matrix for 31 wells in a brine mining area consisted of four stages: The first stage involved calculating the thermal buoyancy gradient, processing 5 days of temperature time-series data (sampling interval of 30 minutes) to generate a gradient sequence of 240 time points, taking 4 minutes. The second stage involved calculating the stress deflection angle, based on the azimuth data of 496 well pairs combined with a core experiment measuring an internal friction angle of 28 degrees, to calculate a unified fracture angle of 31 degrees, taking 1 minute. The third stage involved taking the latest time-point data and calculating the interference coupling intensity values ​​of the 496 well pairs in parallel, taking 3 seconds. The fourth stage generated a 31-row, 31-column matrix with 992 non-zero elements, occupying 15 kilobytes of storage space.

[0104] Application Interface and Service Mechanism: Provides a matrix data query application interface, taking two well location numbers as input parameters, and returning the corresponding interference coupling strength value and calculation timestamp. Supports matrix visualization output service, generating a heatmap of well location spatial distribution and interference intensity, with color levels divided into 10 levels from negative to positive interference intensity values. Offers open matrix data export functionality, supporting dual-mode output in comma-separated value format and matrix marketplace exchange format, while providing metadata documentation recording the matrix version, creation time, and parameter version number.

[0105] Implementation Environment and Resource Configuration: The computing equipment must support floating-point acceleration instruction sets; a CPU clock speed of at least 3 GHz is recommended. Memory capacity is estimated by multiplying the square of the number of well positions by 16 bytes per element; for example, 100 well positions require 160 kilobytes of memory. Disk storage must retain the most recent 30 versions of matrix historical data. The software environment must provide a basic linear algebra subroutine operation library, version 3.7.1 or higher.

[0106] The quality assurance system comprises three levels of verification: The primary verification checks the input data range, requiring temperature values ​​to be between 0 and 150 degrees Celsius and azimuth angles to be between 0 and 360 degrees. The intermediate verification verifies the rationality of the calculation process, ensuring the difference in the thermal buoyancy gradient change rate between adjacent time points does not exceed 50% and the difference in stress deflection angle between adjacent well locations does not exceed 15 degrees. The advanced verification verifies the physical meaning, ensuring that the interference coupling strength value in the production layer conforms to the law of attenuation with increasing well spacing; abnormal attenuation coefficients trigger expert review.

[0107] Fault Tolerance and Recovery Mechanism: The calculation process employs a segmented saving strategy, saving intermediate results after every 10 well pairs of thermal buoyancy gradient calculations. In the event of a hardware failure causing an interruption, the calculation automatically resumes after restarting by detecting the most recent saved point. Parameter Fault Tolerance Settings: When core experimental data is missing, a default internal friction angle of 30 degrees is used, and a warning is recorded; when the gravitational acceleration is not corrected, the standard value of 9.80665 m / s² is used; when the coefficient of thermal expansion is not calibrated, the brine standard value of 0.0004 m / s² is used.

[0108] The interference intensity coefficient matrix generated in this step quantitatively characterizes the thermo-mechanical coupling effect between well groups, providing core parameters for optimizing mining schemes. The matrix data is stored in both binary and text formats. The binary file contains a header (recording the number of well locations, data version, and creation time) and a data body (storing element values ​​in row-major order); the text file uses a formatted storage method with row and column indices and numerical values. A quality report is automatically generated upon output, including data integrity indicators (effective well coverage), calculation accuracy indicators (floating-point rounding error statistics), and anomaly handling logs. The final output uses hash verification to ensure data transmission integrity, employing a secure hash algorithm with a 256-bit digest value appended as a checksum.

[0109] When performing inter-well resource siphon path tracing, the edge weight data and pressure transmission characteristic data of the spatial topology network model are first acquired. The edge weight value of the spatial topology network model is the inter-well distance value (in meters), which is directly obtained from the previously constructed spatial topology network model. The pressure transmission characteristic value is the pressure change rate sequence (in megapascals per minute) corresponding to each well location timestamp. The method for identifying high-pressure and low-pressure well locations is as follows: for the current mining period (e.g., the last 24 hours), the average pressure transmission characteristic value of all well locations is calculated. Well locations with a value 20% higher than the overall average are marked as high-pressure well locations, and those with a value 20% lower than the average are marked as low-pressure well locations. For example, if the average pressure change rate of 31 wells in a certain mining area is 0.05 megapascals per minute, then the threshold for high-pressure well locations is set to 0.06 megapascals per minute, and the threshold for low-pressure well locations is set to 0.04 megapascals per minute. The migration resistance coefficient is directly adopted from the edge weight value (i.e., the inter-well distance value) of the spatial topology network model; the larger the distance, the greater the migration resistance. The minimum resistance path calculation employs a pathfinding method based on network edge weights: starting with a high-pressure well location and ending with a low-pressure well location, all possible paths are traversed, and the path with the smallest total edge weight is selected as the resource siphon path. Constraints are set for the pathfinding process: the number of path nodes does not exceed 5 well locations; the edge weight of a single segment does not exceed 2000 meters. Implementation example: when there are three paths from high-pressure well location A to low-pressure well location B (ACB path with a total distance of 1500 meters, ADEB path with a total distance of 1800 meters, and AFB path with a total distance of 2100 meters), then path ACB is selected as the siphon path. The calculation result is output as a sequence of path nodes, for example, path [A, C, B].

[0110] When generating the basic time delay, the pressure transmission characteristic sequence and resource concentration migration characteristic sequence of the same well location are processed simultaneously. The peak time point of the pressure transmission characteristic is defined as the time point when the pressure change rate reaches a local maximum and exceeds the average rate by 50% within a consecutive 6 sampling point window. The valley time point of the resource concentration migration characteristic is defined as the time point when the concentration change reaches a local minimum and is lower than the average change by 50% within the same time window. When calculating the time difference between the two, the peak time point of the pressure transmission characteristic is subtracted from the valley time point of the resource concentration migration characteristic within the same time window to obtain the basic time delay in minutes. Data alignment rules: When there are multiple peaks or valleys within the same window, the first valid point in the time sequence is taken; when there are no valid extreme points within the window, the window is skipped. Calculation example: For a well location within the time window of 08:00-08:30, the peak of the pressure transmission characteristic occurs at 08:12, and the valley of the resource concentration migration characteristic occurs at 08:18. Then the basic time delay = 08:18 - 08:12 = 6 minutes.

[0111] When establishing the function mapping relationship, the basic time delay data is processed in groups according to the mining stage. The mining stage is divided based on the cumulative mining volume, with each additional 500,000 cubic meters of brine considered a new stage. The basic time delay of all effective well locations within each stage is statistically analyzed to form a time delay distribution dataset. The distribution pattern is extracted by calculating the average, standard deviation, and skewness coefficient of the time delay; a functional relationship between the basic time delay and reservoir viscosity is established: reservoir viscosity equals the basic time delay multiplied by the conversion coefficient plus a constant term. The conversion coefficient is calibrated using historical data: measured reservoir viscosity data from previous mining stages (measured using a rotational viscometer in downhole sampling) and corresponding time delays are collected, and the coefficient value is determined through linear regression fitting. For example, the fitting formula for a certain mining area is: reservoir viscosity (unit: millipascals per second) = basic time delay (minutes) × 120 + 850. The validity of the function mapping relationship is verified as follows: when the coefficient of determination is below 0.7, historical data is expanded and refitted; when the data volume is insufficient, empirical coefficients from adjacent mining areas are used.

[0112] When generating the time-delay compensation factor, the function mapping relationship is applied to the current mining period. The specific execution process is as follows: input the average basic time delay value for the current period, substitute it into the function mapping relationship, and output the estimated reservoir viscosity value. The time-delay compensation factor is defined as the reciprocal of the estimated reservoir viscosity value multiplied by a scaling constant of 1000, in units of millipascals per second (mPa). The scaling constant is used to control the magnitude of the factor and ensure the stability of subsequent calculations. For example, if the average basic time delay value for the current period is 8 minutes, substituting it into the formula yields reservoir viscosity = 8 × 120 + 850 = 1810 mPa·s, and the time-delay compensation factor = 1000 / 1810 ≈ 0.552 mPa·s. The factor has a dynamic update mechanism: after each new mining stage is completed, the function mapping relationship is refitted and the coefficients are updated.

[0113] The high- and low-pressure well location threshold ratio (20%) is determined based on historical pressure fluctuation data of the mining area. Analysis of three years of mining data shows that a 20% deviation of the pressure transmission characteristic value from the average indicates a significant pressure anomaly. The upper limit of the number of path nodes (5) is set based on the maximum well cluster density in the mining area to ensure path finding efficiency. The size of the basic time delay window (6 sampling points) matches the resource concentration migration cycle, and the main cycle is determined to be 30 minutes through spectral analysis.

[0114] For example, in a specific implementation, this step is performed in the fifth mining phase (cumulative mining volume of 2.75 million cubic meters) of a certain brine mine:

[0115] Siphon path tracing: 12 high-pressure well locations and 8 low-pressure well locations were identified, and 24 minimum resistance paths were calculated. The average path length was 1350 meters, and the longest path contained 4 nodes.

[0116] Basic time delay statistics: 6,240 time windows from 31 wells were processed, with 5,812 valid data points. The time delay ranged from 1 to 15 minutes, with an average of 8.3 minutes.

[0117] Function mapping establishment: Based on 1,520 sets of calibration data from the first four stages, the fitting formula is viscosity = time delay × 118.7 + 832 (coefficient of determination 0.82).

[0118] Time delay compensation factor calculation: Input average value 8.3 minutes, output viscosity 1815 mPa·s, factor value 0.551 mPa·s.

[0119] When performing single-well resource assessment value correction, three input data are obtained: the single-well resource assessment value (unit: cubic meters) obtained from previous reserve assessment, the time lag compensation factor (unit: seconds per millipascal) generated through step S5, the set of inter-well resource siphon paths identified through step S5, and the interference intensity coefficient matrix generated through step S4. The single-well resource assessment value is calculated based on the geological reserve model and includes the product of brine volume and recoverability coefficient. The calculation process for the time-corrected assessment value is as follows: multiply the single-well resource assessment value by the time lag compensation factor of the corresponding well location. The calculation formula is: Time-corrected assessment value = Single-well resource assessment value × Time lag compensation factor. Calculation example: For a well with a single-well resource assessment value of 5000 cubic meters and a time lag compensation factor of 0.55 seconds per millipascal, the time-corrected assessment value = 5000 × 0.55 = 2750 cubic meters per second per millipascal. This value represents the recoverable resource quantity after correction for the time lag effect.

[0120] When extracting migration weight coefficients, each inter-well resource siphon path is traversed. For each high-pressure well and low-pressure well pair on the path, the element value at the corresponding row and column position is extracted from the interference intensity coefficient matrix as the migration weight coefficient. The matrix element value is in Newtons per cubic meter (N / m³), representing the inter-well interference coupling strength. The extraction rule is: for high-pressure well i and low-pressure well j, take the element value of the i-th row and j-th column of the matrix. If the matrix does not store the element (such as a diagonal element), the default value of 0 is returned. Example: for high-pressure well A (number 3) and low-pressure well B (number 5), querying the matrix position (3,5) yields the element value of 0.0019 N / m³.

[0121] When calculating resource migration, the time-corrected assessment value of the high-pressure well location is multiplied by a migration weight coefficient, and then multiplied by a unit conversion coefficient. The unit conversion coefficient is set to 1000 to unify the units to cubic meters per second per millipascal. The calculation formula is: Resource migration = Time-corrected assessment value × Migration weight coefficient × 1000. Calculation constraint: When the calculation result is negative, it is forced to zero. Example: The time-corrected assessment value of high-pressure well location A is 2750 cubic meters per second per millipascal, and the migration weight coefficient is 0.0019 Newtons per cubic meter. Then, the resource migration = 2750 × 0.0019 × 1000 = 5225 cubic meters per second per millipascal.

[0122] When adjusting resource levels, process each path individually: subtract the resource migration amount from the time-corrected assessment value of the high-pressure well location, and simultaneously add the resource migration amount to the time-corrected assessment value of the low-pressure well location. The adjustment formula is: New assessment value for high-pressure well location = Original time-corrected assessment value - Resource migration amount; New assessment value for low-pressure well location = Original time-corrected assessment value + Resource migration amount. Boundary condition handling: When the adjusted assessment value is less than the minimum reserve threshold (e.g., 10 cubic meters), the threshold is retained and not further reduced. Example: The original assessment value for high-pressure well location A is 2750 cubic meters per second per millipascal (mPa). After subtracting 5225 mPa·s, it is retained at 10 mPa·s because it is below the threshold; the original assessment value for low-pressure well location B is 3200 mPa·s, and after adding 5225 mPa·s, it becomes 8425 mPa·s.

[0123] When calculating the total brine resources in the region, the time-corrected assessment values ​​of all well locations are summed. The calculation formula is: Total regional brine resources = Σ(Time-corrected assessment values ​​of each well location). Data verification is performed before summing: check for well locations that did not participate in the path adjustment; if so, their original time-corrected assessment values ​​are directly used; verify that the total change rate of the assessment values ​​does not exceed 50% (manual review is triggered in case of anomalies). The final resource total is converted to cubic meters, with a conversion factor of 1 / (N·s / mPa). Calculation example: The accumulated value of 31 wells in a mining area is 185,600 cubic meters·N·s / mPa, which, after unit conversion, yields a total regional resource of 18,560 cubic meters.

[0124] Key parameter settings are based on the following: Unit conversion factor 1000: set according to the principle of dimensional balance to ensure that the migration calculation is cubic meters per second per millipascal × Newtons per cubic meter × 1000 = cubic meters per second per millipascal. Minimum reserve threshold (10 cubic meters): determined based on the minimum economically recoverable quantity of a single well, calculated through the mining cost model. Change rate alarm threshold (50%): set based on historical assessment fluctuation analysis; exceeding this value indicates model anomaly.

[0125] Data preprocessing mechanism: Input data alignment: Checks the consistency of well location numbers in time lag compensation factors, siphon paths, and matrix data; missing well locations are supplemented using nearest neighbor interpolation. Weighting coefficient normalization: When the migration weighting coefficient exceeds the set upper limit (e.g., 0.01 N / m³), it is scaled proportionally to the maximum value. Resource migration volume truncation: The migration volume in a single instance does not exceed 80% of the high-pressure well location assessment value to prevent over-adjustment.

[0126] For example, in specific implementation, the input data includes: single-well evaluation values ​​of 31 wells (range 2000-8000 cubic meters), time delay compensation factor matrix (31 dimensions), 24 siphon paths, and interference matrix 31×31.

[0127] Time correction: 31 time correction assessment values ​​were calculated (range 1100-4400 m³ / s / mPa).

[0128] Resource migration: 24 paths were processed involving 48 well adjustments, with a maximum migration volume of 5225 cubic meters per second per millipard.

[0129] Total Calculation: The total adjusted well location assessment value is 185,600 cubic meters per second per millipard, which is converted to a total area of ​​18,560 cubic meters.

[0130] Output specifications: Single-well correction results: stored as a data table, with fields including well number, original assessment value, corrected assessment value, and migration amount. Regional total report: includes total value, calculation timestamp, number of participating wells, and verification indicators.

[0131] When the siphon path does not match the matrix, skip the path and log an error. When the time delay compensation factor is missing, use the average value of the mining area and mark the estimation. When a matrix element is out of range, initiate the matrix recalculation process.

[0132] This embodiment improves the brine resource assessment paradigm through a multi-level coupling mechanism: First, thermodynamic effects (wellbore temperature time-series data) and geomechanical parameters (spatial topology network model) are fused through an interference intensity coefficient matrix, overcoming the limitation of isolated treatment of thermo-mechanical parameters in traditional methods. The vector dot product operation of thermal buoyancy gradient and stress deflection angle during the matrix construction process reveals the physical essence of well group interference. Second, the generation mechanism of the time-delay compensation factor establishes a dynamic correlation between pressure transmission characteristics and resource concentration migration characteristics. By mapping the reservoir viscosity function, time-domain differences are transformed into quantifiable correction parameters, solving the problem of correcting resource migration time-series misalignment. The establishment of this mapping relationship depends on specific reservoir geological characteristics. Finally, the collaborative correction mechanism of resource siphon path and interference matrix realizes the spatial dynamic adjustment of single-well assessment values. The introduction of migration weight coefficient quantifies the inter-well resource competition effect, forming a closed-loop feedback assessment model.

[0133] Example 2

[0134] Figure 2 A schematic diagram of the three-dimensional dynamic prediction system for deep brine resources of the present invention is provided. The three-dimensional dynamic prediction system for deep brine resources includes:

[0135] The data acquisition module is used to acquire the spatial coordinates of the well clusters and the dynamic time-series data of mining in the target mining area;

[0136] The network modeling module is used to construct a spatial topology network model based on the spatial coordinates of the well group. The spatial topology network model uses well locations as nodes and the distance between wells as edge weights.

[0137] The feature mining module is used to perform industrial data mining on mining dynamic time-series data, extracting pressure transmission characteristics, resource concentration migration characteristics, and wellbore temperature time-series data for each well location.

[0138] The coupled calculation module is used to calculate the thermal buoyancy gradient based on the wellbore temperature time series data, and at the same time determine the stress deflection angle based on the inter-well azimuth angle of the spatial topology network model. The thermal buoyancy gradient and stress deflection angle are fused to construct the interference intensity coefficient matrix.

[0139] The dynamic tracking module is used to dynamically track the resource siphon path between wells based on the edge weight distribution of the spatial topology network model, and at the same time generate a time delay compensation factor based on the statistical time delay law of pressure transmission characteristics and resource concentration migration characteristics.

[0140] The resource integration module is used to couple the interference intensity coefficient matrix, the inter-well resource siphon path, and the time lag compensation factor to correct the single-well resource assessment value and output the total amount of brine resources in the region.

[0141] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0142] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0143] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center containing one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0144] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0145] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0146] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0147] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0148] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0149] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0150] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A three-dimensional dynamic prediction method for deep brine resources, characterized in that, include: S1. Obtain the spatial coordinates of the well group and the dynamic time-series data of mining in the target mining area; S2. Construct a spatial topology network model based on the spatial coordinates of the well group. The spatial topology network model uses well locations as nodes and the distance between wells as edge weights. S3. Perform industrial data mining on the dynamic time series data of mining to extract the pressure transmission characteristics, resource concentration migration characteristics and wellbore temperature time series data of each well location. S4. Calculate the thermal buoyancy gradient based on the wellbore temperature time series data, and determine the stress deflection angle based on the inter-well azimuth angle of the spatial topology network model. Combine the thermal buoyancy gradient and the stress deflection angle to construct the interference intensity coefficient matrix. S5. Based on the spatial topology network model, the edge weight distribution dynamically tracks the resource siphon path between wells, and generates a time delay compensation factor based on the statistical time delay law of pressure transmission characteristics and resource concentration migration characteristics. S6, the coupling interference intensity coefficient matrix, the inter-well resource siphon path and time delay compensation factor are used to correct the single-well resource assessment value, and the total amount of brine resources in the region is output.

2. The three-dimensional dynamic prediction method for deep brine resources according to claim 1, characterized in that, Obtain the spatial coordinates of the well cluster and the dynamic time-series data of mining operations in the target mining area, including: The geodetic coordinates of the wellheads of all wells in the target mining area are retrieved from the drilling engineering database as the spatial coordinates of the well group. Simultaneously collect the bottom pressure sensor readings, brine ion concentration detection values, and wellbore temperature sensor readings of each mining well, and generate mining dynamic time-series data according to a unified timestamp; Verify the consistency between the spatial coordinates of the well group and the well location numbers in the dynamic time-series data of mining, and delete abnormal data segments with missing timestamps.

3. The three-dimensional dynamic prediction method for deep brine resources according to claim 2, characterized in that, A spatial topological network model is constructed based on the spatial coordinates of the well group. The spatial topological network model uses well locations as nodes and inter-well distances as edge weights, including: The obtained wellhead geodetic coordinates are used as nodes in the spatial topology network model; Calculate the three-dimensional Euclidean distance between any two nodes as the well distance; An undirected graph network model is constructed using the distance between wells as the edge weight; The output is a spatial topology network model that includes node locations and edge weights.

4. The three-dimensional dynamic prediction method for deep brine resources according to claim 3, characterized in that, Industrial data mining was performed on the dynamic time-series data of mining operations to extract pressure transmission characteristics, resource concentration migration characteristics, and wellbore temperature time-series data for each well location, including: Time series analysis was performed on the readings of the bottom hole pressure sensor to extract the pressure change rate as a pressure transmission characteristic. Concentration gradient calculations were performed on the measured values ​​of brine ion concentration to determine the concentration change per unit time as a characteristic of resource concentration migration. The time series of wellbore temperature sensor readings is directly extracted as wellbore temperature time series data. Pressure transmission characteristics, resource concentration migration characteristics, and wellbore temperature time series data are classified and stored according to well location number.

5. The three-dimensional dynamic prediction method for deep brine resources according to claim 4, characterized in that, The thermal buoyancy gradient is calculated based on wellbore temperature time-series data. Simultaneously, the stress deflection angle is determined based on the inter-well azimuth angle of the spatial topology network model. An interference intensity coefficient matrix is ​​constructed by fusing the thermal buoyancy gradient and the stress deflection angle, including: The temperature difference between adjacent time points is calculated based on the time series data of wellbore temperature. The thermal buoyancy gradient is obtained by multiplying the temperature difference by the product of the thermal expansion coefficient of brine and the gravitational acceleration. The inter-well azimuth angle is extracted from the spatial topology network model. The direction of maximum principal stress is determined based on the inter-well azimuth angle. The direction of maximum principal stress is then converted into stress deflection angle using the Coulomb fracture criterion. The interference coupling strength value of the corresponding well pair is obtained by performing a vector dot product operation between the thermal buoyancy gradient and the stress deflection angle of the same well pair. Traverse all well pair combinations and construct an interference intensity coefficient matrix using the well location number as the row and column index. The matrix element values ​​are the interference coupling intensity values ​​of the corresponding well pair.

6. The three-dimensional dynamic prediction method for deep brine resources according to claim 5, characterized in that, Converting the direction of maximum principal stress to stress deflection angle using the Coulomb fracture criterion includes: The fracture angle is calculated based on the internal friction angle obtained from the core experiment. Using the direction of maximum principal stress as the reference direction, the fracture angle is taken as the stress deflection angle; The fracture angle is calculated using the Coulomb fracture criterion formula.

7. The three-dimensional dynamic prediction method for deep brine resources according to claim 5, characterized in that, Based on a spatial topology network model, the edge weight distribution dynamically tracks the resource siphon path between wells. Simultaneously, based on the statistical time lag laws of pressure transmission characteristics and resource concentration migration characteristics, a time lag compensation factor is generated, including: Obtain the siphon path for inter-well resources; Simultaneously extract the peak time point of pressure transmission characteristics and the valley time point of resource concentration migration characteristics at the same well location, and calculate the time difference between the two as the basic time delay; The distribution pattern of basic time delay was statistically analyzed according to the mining stage, and a functional mapping relationship between basic time delay and reservoir viscosity was established. The output value of the function mapping relationship in the current mining period is used as the time delay compensation factor.

8. The three-dimensional dynamic prediction method for deep brine resources according to claim 7, characterized in that, The specific method for obtaining the inter-well resource siphon path is as follows: high-pressure well locations and low-pressure well locations are identified by utilizing pressure transmission characteristics, and the minimum resistance path from the high-pressure well location to the low-pressure well location is calculated using the edge weights of the spatial topology network model as the migration resistance coefficient, which serves as the inter-well resource siphon path.

9. The three-dimensional dynamic prediction method for deep brine resources according to claim 7, characterized in that, The coupling interference intensity coefficient matrix, inter-well resource siphon path, and time lag compensation factor are used to correct the single-well resource assessment value, outputting the total regional brine resources, including: The time-delay compensation factor is multiplied by the single-well resource assessment value to obtain the time-corrected assessment value; For each pair of high-pressure and low-pressure well locations on the inter-well resource siphon path, the element values ​​of the corresponding well location pair in the interference intensity coefficient matrix are extracted as migration weight coefficients. The resource migration amount is obtained by multiplying the time-corrected assessment value of the high-pressure well location by the migration weight coefficient. Subtract the resource migration amount from the high-pressure well location time correction assessment value and add the resource migration amount to the low-pressure well location time correction assessment value; The total amount of brine resources in the region is obtained by summing up the time-corrected assessment values ​​after all well location adjustments.

10. A three-dimensional dynamic prediction system for deep brine resources, used to implement the three-dimensional dynamic prediction method for deep brine resources as described in any one of claims 1-9, characterized in that, include: The data acquisition module is used to acquire the spatial coordinates of the well clusters and the dynamic time-series data of mining in the target mining area; The network modeling module is used to construct a spatial topology network model based on the spatial coordinates of the well group. The spatial topology network model uses well locations as nodes and the distance between wells as edge weights. The feature mining module is used to perform industrial data mining on mining dynamic time-series data, extracting pressure transmission characteristics, resource concentration migration characteristics, and wellbore temperature time-series data for each well location. The coupled calculation module is used to calculate the thermal buoyancy gradient based on the wellbore temperature time series data, and at the same time determine the stress deflection angle based on the inter-well azimuth angle of the spatial topology network model. The thermal buoyancy gradient and stress deflection angle are fused to construct the interference intensity coefficient matrix. The dynamic tracking module is used to dynamically track the resource siphon path between wells based on the edge weight distribution of the spatial topology network model, and at the same time generate a time delay compensation factor based on the statistical time delay law of pressure transmission characteristics and resource concentration migration characteristics. The resource integration module is used to couple the interference intensity coefficient matrix, the inter-well resource siphon path, and the time lag compensation factor to correct the single-well resource assessment value and output the total amount of brine resources in the region.

Images