Brine distribution identification method based on dynamic extraction

Abstract

The invention discloses a brine distribution identification method based on dynamic extraction, and the method comprises the steps: obtaining a comprehensive data set through a multi-source data collection system; processing the comprehensive data set by using a feature extraction method to obtain a feature vector set; constructing a dynamic distribution model based on a time sequence by using the feature vector set, and obtaining a preliminary distribution identification result in combination with geological conditions and flow path parameters; obtaining distribution mapping data according to the preliminary distribution identification result; generating an equipment adjusting instruction according to the distribution mapping data in combination with the current position and the running state of the extraction equipment; the extraction equipment is driven to execute position movement and parameter adjustment through the equipment adjustment instruction, brine distribution changes are responded in real time, and adjusted operation data are obtained; and obtaining a distribution identification result updated in real time according to the adjusted operation data. According to the method, the precision and efficiency of brine resource exploitation are remarkably improved, and technical support is provided for sustainable utilization of salt lake resources.

Description

Technical Field

[0001] This invention belongs to the field of data recognition technology, and in particular relates to a method for identifying brine distribution based on dynamic sampling. Background Technology

[0002] The distribution of brine is complex and dynamically changing, and accurate identification and optimized extraction are directly related to resource extraction efficiency and economic benefits. However, existing methods have significant limitations in dealing with the dynamic and complex nature of brine distribution.

[0003] Traditional technologies often rely on single data sources or static models, making it difficult to adapt to the heterogeneous changes in brine at different depths and regions. Furthermore, they lack real-time and adaptive capabilities in data processing and decision optimization, leading to low extraction efficiency and significant resource waste. The core challenge lies in effectively integrating multi-source heterogeneous data and achieving dynamic optimization decisions. Brine distribution is influenced by multiple factors such as geological conditions, ion concentration, and flow paths. The collected resistivity, pH value, and sonar echo data are diverse in type and unevenly distributed in time and space. Processing this data requires efficient feature extraction and analysis capabilities; otherwise, it is difficult to accurately construct a brine distribution model.

[0004] Furthermore, the operation of extraction equipment requires dynamic adjustments to position, rate, and pressure based on real-time data. However, existing systems lack adaptive decision-making mechanisms, making it difficult to respond quickly to changes in complex environments, resulting in insufficient extraction accuracy. The complexity of data processing directly restricts the realization of dynamic decision-making, while the lag in decision-making exacerbates the delay in updating the distributed model, forming a technical bottleneck.

[0005] Therefore, the key issue is how to construct a high-precision dynamic brine distribution model based on multi-source heterogeneous data, and achieve efficient operation of the extraction equipment through real-time adaptive optimization. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention proposes a brine distribution identification method based on dynamic extraction, which significantly improves the accuracy and efficiency of brine resource extraction through intelligent methods.

[0007] To achieve the above objectives, this invention proposes a brine distribution identification method based on dynamic extraction, comprising:

[0008] A comprehensive dataset is obtained using a multi-source data acquisition system;

[0009] The comprehensive dataset is processed using feature extraction methods to obtain a set of feature vectors;

[0010] Using the aforementioned feature vector set, a time-series-based dynamic distribution model is constructed. Combined with geological conditions and flow path parameters, preliminary distribution identification results are obtained.

[0011] Based on the preliminary distribution identification results, obtain distribution mapping data;

[0012] Based on the distribution mapping data, combined with the current location and operating status of the extraction equipment, an equipment adjustment command is generated;

[0013] The device adjustment command drives the extraction equipment to perform position movement and parameter adjustment, responds to changes in brine distribution in real time, and obtains the adjusted operating data.

[0014] Based on the adjusted operational data, obtain real-time updated distribution identification results.

[0015] Optionally, a comprehensive dataset can be obtained using a multi-source data acquisition system, including:

[0016] Raw data from geological exploration, ion concentration, and sonar detection were collected using a multi-source data acquisition system.

[0017] The raw data is preprocessed to obtain a comprehensive dataset.

[0018] Optionally, preprocessing the raw data to obtain a comprehensive dataset includes:

[0019] The original dataset is denoised using Fourier transform to obtain a denoised dataset.

[0020] Based on the denoised dataset, the mean and standard deviation of each data point are calculated using the Z-score standardization method to generate a standardized dataset.

[0021] Based on the standardized dataset, principal component analysis is used to extract the main features and generate a feature dataset.

[0022] The data distribution characteristics are obtained from the feature dataset, and the data are grouped using the K-means clustering algorithm to obtain a clustered dataset;

[0023] Based on the clustered dataset, a data mapping method is used to map the data points to a unified coordinate system, generating a unified standardized dataset.

[0024] By using a unified and standardized dataset, the Euclidean distance between each data point is calculated to determine the comprehensive dataset.

[0025] Optionally, before generating the feature dataset, it is necessary to determine whether the data points in the standardized dataset deviate from the preset threshold. If the data points in the standardized dataset deviate from the preset threshold, the outliers are corrected using a linear interpolation method.

[0026] Optionally, using the aforementioned feature vector set, a time-series-based dynamic distribution model is constructed. Combined with geological conditions and flow path parameters, preliminary distribution identification results are obtained, including:

[0027] Using the aforementioned set of feature vectors, an initial dataset is constructed;

[0028] By using the ARIMA time series model, a set of feature vectors is fitted to generate an initial model for dynamic distribution simulation, and the time series identification results of brine distribution are obtained.

[0029] Through a real-time update mechanism, newly collected geological condition parameters and flow path parameters are obtained from external data sources, the feature vector set is updated, and an updated dataset is generated.

[0030] Using the updated dataset, the ARIMA time series model was rerun, and combined with parameter fusion techniques, preliminary distribution identification results were generated.

[0031] Optionally, based on the preliminary distribution identification results, obtaining distribution mapping data includes:

[0032] The raw data of brine concentration and flow path are continuously collected by the real-time monitoring system. The collected information is then preliminarily processed to obtain the characteristic values ​​of change.

[0033] Based on the comparison between the changed feature values ​​and the preliminary distribution identification results, the distribution mapping data is recalculated.

[0034] Optionally, based on the distribution mapping data and the current location and operating status of the extraction equipment, generating equipment adjustment instructions includes:

[0035] The real-time distribution characteristics of the sampling area are obtained through the distribution mapping data. Data preprocessing is used to clean and normalize the collected information to obtain a preliminary distribution characteristic dataset.

[0036] Based on the preliminary distribution characteristic dataset, combined with the equipment location and operating status, the optimal parameter combination value is determined by dynamically calculating the extraction rate and pressure parameters using an adaptive control algorithm.

[0037] Based on the optimal parameter combination, obtain parameter adjustment values ​​that meet the operating conditions;

[0038] Based on the parameter adjustment values, and according to the equipment operating status and the generation logic of the adjustment instructions, specific equipment adjustment instructions are generated.

[0039] Optionally, based on the adjusted operational data, the real-time updated distribution identification results can be obtained, including:

[0040] It is determined whether the sampling efficiency in the adjusted operating data has reached the preset threshold. If it has not reached the threshold, the operating data needs to be sent back to the dynamic distribution model through the feedback mechanism to obtain the distribution identification results updated in real time.

[0041] Compared with the prior art, the present invention has the following advantages and technical effects:

[0042] This invention acquires raw data such as resistivity, pH, and echo signals through a multi-source data acquisition system. After preprocessing and feature extraction, a dynamic brine distribution model is constructed. This model can predict brine distribution changes in real time and adaptively adjust the operating parameters of the extraction equipment based on the prediction results. When a significant change in brine concentration or flow path is detected, this invention can promptly update the distribution mapping data, calculate the optimal extraction rate and pressure parameters, and drive the equipment to execute the optimization scheme. If the extraction efficiency does not meet expectations, this invention can also update the model parameters through a feedback mechanism to achieve secondary optimization. This intelligent method significantly improves the accuracy and efficiency of brine resource extraction, providing technical support for the sustainable utilization of salt lake resources. Attached Figure Description

[0043] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0044] Figure 1 This is a flowchart of a brine distribution identification method based on dynamic sampling according to an embodiment of the present invention. Detailed Implementation

[0045] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0046] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0047] This invention proposes a method for identifying brine distribution based on dynamic extraction, such as... Figure 1 As shown, the specific steps include:

[0048] A comprehensive dataset is obtained using a multi-source data acquisition system;

[0049] The comprehensive dataset is processed using feature extraction methods to obtain a set of feature vectors;

[0050] By using a set of feature vectors, a dynamic distribution model based on time series is constructed. Combined with geological conditions and flow path parameters, preliminary distribution identification results are obtained.

[0051] Based on the preliminary distribution identification results, obtain the distribution mapping data;

[0052] Based on the distribution mapping data, combined with the current location and operating status of the extraction equipment, an equipment adjustment instruction is generated;

[0053] By using equipment adjustment commands, the extraction equipment is driven to move its position and adjust its parameters, responding in real time to changes in brine distribution and obtaining adjusted operating data.

[0054] Based on the adjusted operational data, obtain real-time updated distribution identification results.

[0055] Specifically, this invention acquires raw data such as resistivity, pH, and echo signals through a multi-source data acquisition system. After preprocessing and feature extraction, a dynamic brine distribution model is constructed. This model can predict brine distribution changes in real time and adaptively adjust the operating parameters of the extraction equipment based on the prediction results. When a significant change in brine concentration or flow path is detected, this invention can promptly update the distribution mapping data, calculate the optimal extraction rate and pressure parameters, and drive the equipment to execute the optimization plan. If the extraction efficiency does not meet expectations, this invention can also update the model parameters through a feedback mechanism to achieve secondary optimization. This intelligent method significantly improves the accuracy and efficiency of brine resource extraction, providing technical support for the sustainable utilization of salt lake resources.

[0056] Furthermore, by utilizing a multi-source data acquisition system, a comprehensive dataset is obtained, including:

[0057] Raw data from geological exploration, ion concentration, and sonar detection were collected using a multi-source data acquisition system.

[0058] Preprocess the raw data to obtain a comprehensive dataset.

[0059] Furthermore, the raw data is preprocessed to obtain a comprehensive dataset, including:

[0060] The original dataset is denoised using Fourier transform to obtain a denoised dataset.

[0061] Based on the denoised dataset, the Z-score standardization method is used to calculate the mean and standard deviation of each data point to generate a standardized dataset.

[0062] Based on the standardized dataset, principal component analysis is used to extract the main features and generate a feature dataset.

[0063] The data distribution characteristics are obtained from the feature dataset, and the data are grouped using the K-means clustering algorithm to obtain the clustered dataset;

[0064] Based on the clustered dataset, a data mapping method is used to map data points to a unified coordinate system, generating a unified and standardized dataset.

[0065] By using a standardized dataset, the Euclidean distance between each data point is calculated to determine the comprehensive dataset.

[0066] Specifically, when processing raw datasets of resistivity, pH, and echo signals, one can begin by understanding the principles of signal processing and the basic idea of ​​Fourier transform denoising. The Fourier transform converts the signal from the time domain to the frequency domain, separating the frequency components of noise and effective signal, thereby filtering out high-frequency noise. For example, suppose that in a geological exploration, the resistivity data collected contains noise generated by equipment interference. The original data fluctuates between 10 and 50 ohms. After denoising, the fluctuation range is reduced to 15 to 40 ohms, making the data smoother and laying the foundation for subsequent analysis.

[0067] When using Z-score standardization on denoised datasets, it can be understood as a data normalization method aimed at eliminating the influence of different dimensions. For example, suppose resistivity and pH data have significantly different dimensions: resistivity has a mean of 25 ohms and a standard deviation of 5, while pH has a mean of 7 and a standard deviation of 0.5. After Z-score standardization, both types of data are transformed into a distribution with a mean of 0 and a standard deviation of 1, facilitating unified analysis. This method is particularly suitable for multi-source data fusion scenarios, ensuring the accuracy of subsequent feature extraction.

[0068] Linear interpolation is a simple and effective method for correcting outliers in standardized datasets. For example, if a data point in resistivity data suddenly jumps to 100 ohms, far exceeding the normal range of 20 to 30 ohms, linear interpolation can correct the outlier to around 25 ohms by considering the trend of preceding and following data points. This method avoids outliers interfering with subsequent analysis and maintains data continuity.

[0069] Principal component analysis (PCA) is primarily used to reduce dimensionality and retain key information when extracting feature datasets. Assuming the original data contains three dimensions—resistivity, pH, and echo signal—PCA can compress the data into two principal components, retaining over 90% of the information. This method effectively reduces computational complexity when processing multidimensional data, while highlighting the factors that have the greatest impact on brine distribution characteristics.

[0070] K-means clustering can help identify potential regions of brine distribution when grouping data. Assuming a feature dataset contains 1000 data points, the clustering algorithm can divide them into three groups, corresponding to regions with high resistivity and low pH, medium resistivity and medium pH, and low resistivity and high pH, ​​respectively. This grouping helps to initially determine the regional characteristics of brine distribution, providing a basis for subsequent spatial analysis.

[0071] Mapping data to a unified coordinate system aligns data points from different sources to the same spatial reference frame. For example, if resistivity data points and echo signal data points were originally based on different sampling grids, mapping methods can unify them into a meter-based coordinate system, facilitating spatial relationship analysis. This approach helps in the fusion of multi-source data and improves the overall consistency of the analysis.

[0072] Furthermore, before generating the feature dataset, it is necessary to determine whether the data points in the standardized dataset deviate from the preset threshold. If the data points in the standardized dataset deviate from the preset threshold, the outliers are corrected using a linear interpolation method.

[0073] Furthermore, using the feature vector set, a time-series-based dynamic distribution model is constructed. Combined with geological conditions and flow path parameters, preliminary distribution identification results are obtained, including:

[0074] Construct an initial dataset using the feature vector set;

[0075] By using the ARIMA time series model, a set of feature vectors is fitted to generate an initial model for dynamic distribution simulation, and the time series identification results of brine distribution are obtained.

[0076] Through a real-time update mechanism, newly collected geological condition parameters and flow path parameters are obtained from external data sources, the feature vector set is updated, and an updated dataset is generated.

[0077] Using the updated dataset, the ARIMA time series model was rerun, and combined with parameter fusion techniques, preliminary distribution identification results were generated.

[0078] Specifically, when obtaining the feature vector set from the data source, geological condition parameters may include formation porosity, permeability, and temperature, while flow path parameters may include fluid velocity and path length. In a brine distribution study, porosity data ranged from 0.1 to 0.3, permeability from 10 to 50 millidarcy, temperature from 20°C to 60°C, and fluid velocity from 0.01 to 0.05 m / s. These parameters were collected through well logs and sensors, forming the initial dataset that reflects the dynamic characteristics of the brine in the geological environment. It should be noted that the selection of the data source must ensure coverage of the diversity of the study area, such as geological conditions at different depths, to guarantee the comprehensiveness of the feature vectors. Specifically, when constructing the ARIMA time series model, fitting can be performed based on the time series characteristics of the feature vector set.

[0079] Using porosity and fluid velocity as the main input variables, the ARIMA model analyzes brine distribution data from the past 30 days to predict the distribution trend for the next 7 days. The initial model may show that the brine concentration in a certain area increases over time, with predicted values ​​such as a concentration increase from 2% to 3%. If the predicted results deviate from the actual observed data by more than 5%, it indicates that the model needs further optimization. In one embodiment, Kalman filtering is used to adjust the ARIMA model parameters.

[0080] By smoothing out fluctuations in fluid velocity through filters and reducing noise, the updated model predicts concentration changes more closely to actual observations, with the bias decreasing from 5% to 2%. This optimization improves the model's accuracy in simulating the dynamic distribution of brine.

[0081] Calculating the heterogeneous distribution of brine at different depths and in different regions allows for the analysis of concentration variations from 500 meters to 2000 meters. Results may show that brine concentration fluctuates significantly at a depth of 1000 meters, with a standard deviation of 0.5%, while it is more stable at 2000 meters, with a standard deviation of only 0.2%. Regional distribution characteristics may indicate that concentration variations are more pronounced near fault zones. This analysis helps identify key areas and depths in brine distribution. In one possible implementation, a real-time update mechanism can collect new data hourly via a sensor network, such as when the formation temperature changes from 22°C to 25°C, updating the feature vector set. The updated dataset is then re-input into the ARIMA model to generate new predictions.

[0082] Furthermore, based on the preliminary distribution identification results, the distribution mapping data obtained includes:

[0083] The raw data of brine concentration and flow path are continuously collected by the real-time monitoring system. The collected information is then preliminarily processed to obtain the characteristic values ​​of change.

[0084] Based on the comparison between the changed feature values ​​and the preliminary distribution identification results, the distribution mapping data is recalculated.

[0085] Furthermore, based on the distribution mapping data, combined with the current location and operating status of the extraction equipment, equipment adjustment instructions are generated, including:

[0086] The real-time distribution characteristics of the sampling area are obtained by using distribution mapping data. Data preprocessing is used to clean and normalize the collected information to obtain a preliminary distribution characteristic dataset.

[0087] Based on the preliminary distribution characteristic dataset, combined with the equipment location and operating status, the optimal parameter combination value is determined by dynamically calculating the extraction rate and pressure parameters using an adaptive control algorithm.

[0088] Based on the optimal parameter combination, obtain parameter adjustment values ​​that meet the operating conditions;

[0089] Based on the parameter adjustment values, and according to the equipment operating status and the generation logic of the adjustment instructions, specific equipment adjustment instructions are generated.

[0090] Specifically, in one possible implementation, acquiring the distribution mapping data requires collecting real-time data from a sensor network in the extraction area. For a specific salt field extraction area, sensors collect raw data on brine concentration and flow rate every 5 minutes. This data may contain noise or outliers, thus requiring data preprocessing techniques for cleaning. Median filtering is used to remove outliers, and linear normalization is applied to standardize the concentration values ​​to the range of 0 to 1. The raw concentration data ranges from 50 to 200 g / L, and after normalization, a distribution characteristic dataset with uniform dimensions is formed. This preprocessing method ensures data consistency for subsequent analysis and facilitates algorithm processing.

[0091] The adaptive control algorithm considers equipment location and operating status when calculating extraction rate and pressure parameters. For example, suppose an extraction device is located in the central area of ​​a salt field, with a current operating pressure of 2.5 MPa and a flow rate of 10 m³ / s. 3 The algorithm, based on the distribution characteristic dataset, detected an increase in brine concentration and determined that the extraction rate needed to be reduced to avoid pipeline blockage. The algorithm dynamically adjusted the pressure to 2.0 MPa and reduced the extraction rate to 8 m / s. 3 This adaptive adjustment optimizes equipment operating efficiency through real-time environmental feedback. In one embodiment, if the calculated parameter combination value exceeds a threshold range, such as the pressure exceeding the safety threshold of 2.8 MPa, parameter smoothing technology intervenes. A weighted average method can be used to gradually adjust the pressure value to 2.7 MPa, ensuring the equipment operates within a safe range. This smoothing process avoids the impact of sudden parameter changes on the equipment, extending its lifespan. Specifically, when generating equipment adjustment commands, it is necessary to ensure that the commands match the operating environment.

[0092] Based on the brine distribution characteristics in different areas of the salt field, the commands are generated differently according to the location of the equipment. Equipment located in high-concentration areas may receive commands to reduce the flow rate, while equipment in low-concentration areas may receive commands to increase the extraction rate. This matching mechanism improves the accuracy of command execution.

[0093] Furthermore, based on the adjusted operational data, the real-time updated distribution identification results are obtained, including:

[0094] It is determined whether the sampling efficiency in the adjusted operating data has reached the preset threshold. If it has not reached the threshold, the operating data needs to be sent back to the dynamic distribution model through the feedback mechanism to obtain the distribution identification results updated in real time.

[0095] Specifically, in one embodiment, the process of generating adjustment instructions and converting them into control signals can be achieved by using an instruction encoding tool to encode the calculated power value, such as increasing it to 75 percentage points, into a digital signal recognizable by the device, such as a specific sequence of voltage pulses. This signal directly drives the device to perform the adjustment, avoiding errors caused by human intervention and improving control accuracy.

[0096] In the deviation judgment and fine-tuning stage after the secondary optimization operation, if the equipment feedback shows that the actual power is only 70 percentage points, deviating from the preset threshold by 5 percentage points, the deviation data is obtained through the real-time monitoring module. The cause may be increased pipeline resistance, and then the deviation is fine-tuned to 78 percentage points to compensate for the resistance. This real-time feedback mechanism can quickly respond to anomalies and ensure stable equipment operation.

[0097] Specifically, determining the range of deviations in mining accuracy and analyzing the optimization effect can be achieved by comparing real-time mining data with the target accuracy using data comparison tools. For example, assuming the target mining accuracy is an hourly extraction error of no more than 2 cubic meters, while the actual error is 3.5 cubic meters, exceeding the preset threshold of 1.5 cubic meters, a parameter adjustment process is triggered. This analysis can promptly identify problems and ensure mining efficiency.

[0098] In one embodiment, by integrating the calculation basis and adjustment scheme for new operating parameters, if the deviation data indicates that increased resistance is the main cause, the extraction rate can be adjusted to 450 cubic meters per hour based on historical data, while simultaneously increasing the power by 80 percentage points. This integrated scheme can take into account multi-dimensional data, ensuring that the final parameter adjustment balances efficiency and stability, significantly reducing mining errors and improving resource utilization.

[0099] The above are merely preferred embodiments 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.

Claims

1. A method for identifying brine distribution based on dynamic extraction, characterized in that, include: A comprehensive dataset is obtained using a multi-source data acquisition system; The comprehensive dataset is processed using feature extraction methods to obtain a set of feature vectors; Using the aforementioned feature vector set, a time-series-based dynamic distribution model is constructed. Combined with geological conditions and flow path parameters, preliminary distribution identification results are obtained. Based on the preliminary distribution identification results, obtain distribution mapping data; Based on the distribution mapping data, combined with the current location and operating status of the extraction equipment, an equipment adjustment command is generated; The device adjustment command drives the extraction equipment to perform position movement and parameter adjustment, responds to changes in brine distribution in real time, and obtains the adjusted operating data. Based on the adjusted operational data, obtain real-time updated distribution identification results.

2. The brine distribution identification method based on dynamic extraction according to claim 1, characterized in that, Using a multi-source data acquisition system, a comprehensive dataset is obtained, including: Raw data from geological exploration, ion concentration, and sonar detection were collected using a multi-source data acquisition system. The raw data is preprocessed to obtain a comprehensive dataset.

3. The brine distribution identification method based on dynamic extraction according to claim 2, characterized in that, Preprocessing the raw data to obtain a comprehensive dataset includes: The original dataset is denoised using Fourier transform to obtain a denoised dataset. Based on the denoised dataset, the mean and standard deviation of each data point are calculated using the Z-score standardization method to generate a standardized dataset. Based on the standardized dataset, principal component analysis is used to extract the main features and generate a feature dataset. The data distribution characteristics are obtained from the feature dataset, and the data are grouped using the K-means clustering algorithm to obtain a clustered dataset; Based on the clustered dataset, a data mapping method is used to map the data points to a unified coordinate system, generating a unified standardized dataset. By using a unified and standardized dataset, the Euclidean distance between each data point is calculated to determine the comprehensive dataset.

4. The brine distribution identification method based on dynamic extraction according to claim 3, characterized in that, Before generating the feature dataset, it is necessary to determine whether the data points in the standardized dataset deviate from the preset threshold. If the data points in the standardized dataset deviate from the preset threshold, the outliers are corrected by linear interpolation.

5. The brine distribution identification method based on dynamic extraction according to claim 1, characterized in that, Using the aforementioned feature vector set, a time-series-based dynamic distribution model is constructed. Combined with geological conditions and flow path parameters, preliminary distribution identification results are obtained, including: Using the aforementioned set of feature vectors, an initial dataset is constructed; By using the ARIMA time series model, a set of feature vectors is fitted to generate an initial model for dynamic distribution simulation, and the time series identification results of brine distribution are obtained. Through a real-time update mechanism, newly collected geological condition parameters and flow path parameters are obtained from external data sources, the feature vector set is updated, and an updated dataset is generated. Using the updated dataset, the ARIMA time series model was rerun, and combined with parameter fusion techniques, preliminary distribution identification results were generated.

6. The brine distribution identification method based on dynamic extraction according to claim 1, characterized in that, Based on the preliminary distribution identification results, the distribution mapping data is obtained as follows: The raw data of brine concentration and flow path are continuously collected by the real-time monitoring system. The collected information is then preliminarily processed to obtain the characteristic values ​​of change. Based on the comparison between the changed feature values ​​and the preliminary distribution identification results, the distribution mapping data is recalculated.

7. The brine distribution identification method based on dynamic extraction according to claim 1, characterized in that, Based on the distribution mapping data, and combined with the current location and operating status of the extraction equipment, the generated equipment adjustment instructions include: The real-time distribution characteristics of the sampling area are obtained through the distribution mapping data. Data preprocessing is used to clean and normalize the collected information to obtain a preliminary distribution characteristic dataset. Based on the preliminary distribution characteristic dataset, combined with the equipment location and operating status, the optimal parameter combination value is determined by dynamically calculating the extraction rate and pressure parameters using an adaptive control algorithm. Based on the optimal parameter combination, obtain parameter adjustment values ​​that meet the operating conditions; Based on the parameter adjustment values, and according to the equipment operating status and the generation logic of the adjustment instructions, specific equipment adjustment instructions are generated.

8. The brine distribution identification method based on dynamic extraction according to claim 1, characterized in that, Based on the adjusted operational data, the real-time updated distribution identification results include: It is determined whether the sampling efficiency in the adjusted operating data has reached the preset threshold. If it has not reached the threshold, the operating data needs to be sent back to the dynamic distribution model through the feedback mechanism to obtain the distribution identification results updated in real time.

Images