Construction method and system of petroleum production simulation model

Abstract

The invention relates to the technical field of simulation model construction, in particular to a construction method and system of a petroleum production simulation model. The method comprises the following steps: acquiring a real-time petroleum production process monitoring video and operation state parameters of production equipment; performing coding standardization processing on the operation state parameters of the production equipment, and constructing a state evolution feature library of the production equipment; detail sharpening filtering optimization is carried out on the real-time petroleum production process monitoring video, so that the production monitoring video is subjected to detail optimization; performing process behavior characteristic analysis on the detail optimization production monitoring video one by one, and constructing a production action dynamic behavior sequence; and performing action logic dependency analysis on the production action dynamic behavior sequence to generate production action logic dependency data. According to the method, the production simulation model is constructed through dynamic and accurate petroleum production behavior simulation, so that the visualization of the petroleum production process is realized, and decision support is provided for improving the production efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of simulation model construction, and in particular to a method and system for constructing an oil production simulation model. Background Art

[0002] As global energy demand continues to grow, the oil industry faces tremendous production pressure and challenges. In order to meet the ever-changing market demand, oil companies urgently need to improve production efficiency, reduce operating costs, and ensure the safety of the production process. However, the traditional oil production model has been unable to cope with the increasingly complex production environment and dynamically changing process requirements.

[0003] Traditional oil production process modeling mainly relies on empirical mathematical models and a small amount of collected measured data, which is difficult to accurately reflect the complex dynamic characteristics of actual production. On the one hand, the on-site production process involves a large number of equipment, processes, and environmental parameters, and it is difficult to fully capture and integrate these multi-source heterogeneous data; on the other hand, the existing modeling methods lack an effective description of the complex cause-and-effect relationships in the production process, and it is difficult to provide sufficient support for production decisions. Therefore, a more intelligent method for constructing oil production simulation models is needed. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a method and system for constructing an oil production simulation model to solve at least one of the above technical problems.

[0005] To achieve the above object, the present invention provides a method for constructing a petroleum production simulation model, comprising the following steps: Step S1: obtaining real-time oil production process monitoring video and operating status parameters of production equipment; encoding and standardizing the operating status parameters of production equipment to construct a production equipment status evolution feature library; Step S2: performing detail sharpening filtering optimization on the real-time oil production process monitoring video, thereby optimizing the production monitoring video in detail; Step S3: Analyze the process behavior characteristics of each detail optimization production monitoring video and construct a dynamic behavior sequence of production actions; Step S4: performing action logic dependency analysis on the production action dynamic behavior sequence, thereby generating production action logic dependency data; Step S5: Perform production action simulation mapping based on production action logic dependency data and production equipment state evolution feature library to build an oil production action simulation model; Step S6: Adaptively optimize the oil production action simulation model and construct an adaptive production scenario simulation model.

[0006] The present invention helps to improve the comparability of data and the accuracy of analysis by standardizing the operation status parameters, and provides a reliable data basis for subsequent model construction. Constructing a production equipment state evolution feature library can help identify the equipment state change mode and provide a basis for subsequent action simulation. By optimizing the monitoring video through detail sharpening filtering, the clarity and quality of the video can be improved, making the monitoring process more accurate and reliable. Analyzing the process behavior characteristics in the production monitoring video helps to understand the key actions and patterns in the production process, and provides data support for the subsequent action logic dependency analysis. By parsing the action logic dependency, the association and dependency relationship between production actions can be revealed, providing a logical basis for the construction of subsequent simulation models. Simulation mapping based on logical dependency data and state feature library can realize the simulation of oil production actions and help understand the complex relationships and influencing factors in the production process. Through adaptive scene optimization, the simulation model can be adjusted and optimized according to real-time data and simulation results, so that the simulation results are closer to the real production environment and the accuracy and reliability of the model are improved. Constructing an adaptive production scene simulation model can help optimize the production process, improve production efficiency and safety, and provide a reference for decision-making.

[0007] Preferably, step S1 comprises the following steps: Step S11: Acquire real-time oil production process monitoring video and operating status parameters of production equipment; Step S12: performing coding and standardization processing on the operating status parameters of the production equipment to generate a coding equipment operating status parameter sequence; Step S13: Analyze the timing change of the encoding device operating state parameter sequence one by one to obtain the timing change data of the operating state of each device; Step S14: mining the state evolution characteristics of the time-series change data of the operating state of each device to build a state evolution characteristic library of production equipment.

[0008] By acquiring real-time monitoring videos and equipment operating status parameters, the present invention can timely understand the actual situation of the production site and provide a real data basis for the subsequent simulation model construction. Standardization of equipment operating status parameters can improve the consistency and comparability of data and provide a unified data format for subsequent analysis and modeling. By analyzing the time series changes of equipment operating status parameters, the changing laws and trends of equipment states can be grasped, providing detailed data support for understanding equipment operating behavior. Mining the state evolution characteristics of each device can reveal the changing patterns and laws of the equipment operating state, and provide important feature parameters for subsequent model construction. Constructing a production equipment state evolution feature library helps to establish a database of equipment state evolution, providing rich feature information and reference basis for the construction of production simulation models.

[0009] Preferably, the specific steps of step S12 are: Calculate the original state parameter mean of the operating state parameters of the production equipment to obtain the state parameter mean of the production equipment; Perform outlier detection based on the mean value of the state parameters of the production equipment and extract abnormal state parameters; Perform abnormal value filtering on abnormal state parameters to generate abnormal filtering operation state parameters; Performing timestamp normalization processing on the abnormal filtering operation status parameters to obtain standardized operation status parameters; Identify the equipment corresponding code for the standardized operating status parameters and extract the corresponding code for each standardized operating status parameter; The equipment coding sequence is integrated based on the corresponding coding of each standardized operating status parameter to generate a coded equipment operating status parameter sequence.

[0010] The present invention helps to summarize and generalize the original data by calculating the mean value of the status parameters of the production equipment, and provides an indicator of the overall equipment status. By detecting abnormal status parameters, it can help to discover abnormal conditions or problems in the operation of the equipment and provide warnings for subsequent processing. Filtering abnormal status parameters can improve data quality and ensure the accuracy and reliability of subsequent analysis and modeling. Standardized timestamps can make the data consistent in time, which is convenient for subsequent data analysis and model building. Identifying the corresponding code of each standardized operating status parameter helps to establish the association between the equipment status parameter and the equipment, providing key information for subsequent analysis. Integrating the equipment coding sequence can form a serialized data set, which provides an orderly data basis for subsequent model building and facilitates model training and analysis.

[0011] Preferably, the specific steps of step S2 are: Step S21: Perform visual inspection of production equipment on the real-time oil production process monitoring video to extract multiple production equipment area images; Step S22: performing color tone difference analysis between devices on multiple production equipment area images, and extracting color tone difference values ​​between different devices; Step S23: performing color correction processing on the real-time oil production process monitoring video according to the color tone difference values ​​between different devices to obtain a color-corrected monitoring video; Step S24: performing device edge contour recognition on the color correction monitoring video to extract the device edge contour line; Step S25: Optimize the equipment edge contour line by performing detail sharpening filtering, thereby optimizing the production monitoring video in detail.

[0012] The present invention extracts regional images of production equipment through visual detection, which can help to accurately identify and distinguish different equipment, and provide a regional image data basis for subsequent analysis and processing. Analysis of color tone differences can reveal the differences in visual features between different equipment, and provide key information for equipment identification and analysis. Color correction can improve the color accuracy and consistency of monitoring videos, ensure the color authenticity of production equipment displays, and provide accurate visual information for subsequent data analysis. Identifying equipment edge contours helps to extract structural information of the equipment, and provides important features of equipment morphology for subsequent processing and analysis. Detail sharpening filter optimization can enhance the details and clarity of the image, improve the visual quality of the monitoring video, and make the monitoring process more accurate and reliable.

[0013] Preferably, step S3 specifically comprises the following steps: Step S31: performing time-series frame processing on the detail-optimized production monitoring video, thereby obtaining a production monitoring time-series image frame sequence; Step S32: marking the key feature points of the equipment on the production monitoring time-series image frame sequence, and extracting multiple key feature points of the equipment; Step S33: Perform continuous frame dynamic optical flow tracking analysis based on multiple key feature points of the equipment to generate a dynamic optical flow trajectory of the production equipment; Step S34: Analyze the process behavior characteristics of each production equipment dynamic optical flow trajectory one by one to obtain each production process behavior data; Step S35: Perform dynamic production posture mining on each production process behavior data to generate equipment dynamic production posture data; Step S36: Perform motion state migration fitting on the dynamic production posture data, thereby constructing a dynamic behavior sequence of production actions.

[0014] The present invention can extract time series information through frame processing, which is helpful to capture the dynamic change process of production monitoring video and provide continuous frame sequence data for subsequent analysis. Marking key feature points is helpful to extract important features of equipment and provide a basis for key information of equipment for subsequent analysis. Dynamic trajectory information of equipment can be obtained through optical flow tracking analysis to reveal the movement of equipment and provide trajectory data support for subsequent behavior analysis. Analysis of dynamic optical flow trajectory can extract process behavior characteristics and help identify different production process behaviors, providing basic data for subsequent production posture mining. Mining dynamic production posture is helpful to understand the dynamic posture changes of equipment in the production process and provide posture data support for the construction of production action behavior sequence. Dynamic behavior sequence of production action can be established through action state migration fitting, revealing the transfer relationship between different action states in the production process and providing dynamic behavior data basis for the construction of simulation model.

[0015] Preferably, the specific steps of step S4 are: Step S41: encoding each action behavior based on the production action dynamic behavior sequence to obtain each production action behavior code; Step S42: extracting the start and end timestamps of the production action dynamic behavior sequence to obtain the timestamp of each production action behavior; Step S43: performing action sequence analysis based on the timestamp of each production action to generate a production action sequence relationship; Step S44: performing a relative time difference constraint analysis on the production action sequence according to each production action behavior code to obtain the production action time difference constraint law; Step S45: performing action logic dependency analysis on the production action time difference constraint rules, thereby generating production action logic dependency data.

[0016] The present invention can uniquely identify each production action through action behavior coding, which is convenient for subsequent analysis and identification of different action types. Extracting timestamps helps to determine the start and end time of each production action, and provides time information for subsequent time analysis and behavior relationship modeling. Analyzing the sequence of actions can determine the order relationship between production actions, reveal the execution order between various actions in the production process, and provide support for understanding the production process. Through time difference constraint analysis, the timing relationship and constraint rules between different production actions can be revealed, providing a basis for simulating the timing relationship of various activities in the production process. Parsing action logic dependencies helps to understand the logical relationship and dependency between production actions, provides support for establishing a logical model of production actions, and helps to optimize the production process and improve efficiency.

[0017] Preferably, the specific steps of step S5 are: Step S51: performing production equipment topology analysis on the detail optimization production monitoring video to generate equipment topology data; Step S52: performing three-dimensional topological modeling on the equipment topological structure data based on a plurality of key feature points of the equipment to construct a three-dimensional point cloud model of the production equipment; Step S53: performing state evolution action matching on the production action dynamic behavior sequence based on the production equipment state evolution feature library to obtain the equipment state evolution data of each production action; Step S54: Based on the equipment state evolution data of each production action and the production action logic dependency data, the three-dimensional point cloud model of the production equipment is subjected to production action simulation mapping to construct an oil production action simulation model.

[0018] The present invention generates equipment topology data by performing equipment topology structure analysis on production monitoring videos, thereby helping to understand the connection relationship and structural layout between equipment. Three-dimensional topology modeling is performed based on key feature points of the equipment, and a three-dimensional point cloud model of the production equipment is constructed, which provides a basis for the spatial structure modeling of the equipment and is helpful for visualization and simulation. The state evolution feature library of the production equipment is used to match the state evolution action of the dynamic behavior sequence of the production action, and the equipment state evolution data of each production action is obtained, which helps to understand the relationship between the action and the equipment state. Based on the equipment state evolution data and action logic dependency data of the production action, the three-dimensional point cloud model of the production equipment is mapped for production action simulation, and an oil production action simulation model is constructed, which can simulate the changes in equipment state and action during the production process and provide a reference for optimizing the production process.

[0019] Preferably, the specific steps of step S6 are: Step S61: performing real-time production scene feature recognition on the detail optimization production monitoring video to extract current production scene features; Step S62: Perform scene transfer learning on the current production scene features to obtain a scene transfer generalization representation; Step S63: Based on the generalized representation of scenario migration, the oil production action simulation model is adaptively optimized to construct an adaptive production scenario simulation model.

[0020] The present invention helps to understand the current production status and environmental conditions by identifying real-time production scene features in production monitoring videos, providing basic data for subsequent analysis and optimization. Through scene transfer learning, the current production scene features can be generalized to other related scenes, improving the generalization ability of the model and providing support for the construction of simulation models in different scenes. Adaptive scene optimization based on scene transfer generalization representation can make the simulation model more adaptable, and can automatically adjust model parameters according to different scene characteristics, thereby improving the accuracy and realism of the simulation model.

[0021] In this specification, a system for constructing a petroleum production simulation model is provided, which is used to execute the method for constructing a petroleum production simulation model as described above, comprising: The state evolution module is used to obtain real-time oil production process monitoring videos and operating state parameters of production equipment; encode and standardize the operating state parameters of production equipment, and build a state evolution feature library for production equipment; Image enhancement module, used to perform detail sharpening and filtering optimization on real-time oil production process monitoring videos, thereby optimizing production monitoring videos in detail; The production action module is used to analyze the process behavior characteristics of each detail optimization production monitoring video and construct a dynamic behavior sequence of production actions; The action logic dependency module is used to perform action logic dependency analysis on the dynamic behavior sequence of production actions, thereby generating production action logic dependency data; The action simulation module is used to perform production action simulation mapping based on the production action logic dependency data and the production equipment state evolution feature library to build an oil production action simulation model; The adaptive scenario optimization module is used to perform adaptive scenario optimization on the petroleum production action simulation model and build an adaptive production scenario simulation model.

[0022] The present invention can realize real-time monitoring of the production process by acquiring real-time oil production process monitoring video and equipment status parameters, and help monitor production status and abnormal situations. Encoding and standardizing the production equipment status parameters and constructing a state evolution feature library are helpful to establish historical data of equipment state evolution, providing a basis for subsequent analysis and prediction. Optimizing the real-time monitoring video through detail sharpening filtering can improve image quality, enhance details, make the monitoring effect clearer, and help accurately analyze the detailed information in the production process. The process behavior feature analysis one by one helps to identify and understand each action in the production process, and provides a basis for subsequent logical dependency analysis and simulation modeling. Constructing a dynamic behavior sequence can present the temporal relationship of production actions, and help reveal the order and dependency between actions. Parsing the action logic dependency can reveal the logical connection between each production action, help understand the action execution order and dependency, and provide guidance for model construction. Simulation mapping based on logical dependency data and the equipment status feature library can simulate the execution process of oil production actions, help predict and optimize the production process, and adaptive scenario optimization can adjust the simulation model parameters according to the real-time situation, so that the model is closer to the actual production environment, and improve the accuracy and application scope of the simulation model. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A schematic diagram of the steps of a method for constructing a petroleum production simulation model according to the present invention; Figure 2 Detailed implementation flow chart of step S1; Figure 3 Detailed implementation flow chart of step S2; Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0024] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0025] The present application example provides a method and system for constructing a petroleum production simulation model. The execution subjects of the method and system include but are not limited to: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system, which can be regarded as the general computing node of the present application, and the data processing platform includes but is not limited to: at least one of an audio and image management system, an information management system, and a cloud data management system.

[0026] See also Figures 1 to 4 The present invention provides a method for constructing a petroleum production simulation model, and the method for constructing a petroleum production simulation model comprises the following steps: Step S1: obtaining real-time oil production process monitoring video and operating status parameters of production equipment; encoding and standardizing the operating status parameters of production equipment to construct a production equipment status evolution feature library; Step S2: performing detail sharpening filtering optimization on the real-time oil production process monitoring video, thereby optimizing the production monitoring video in detail; Step S3: Analyze the process behavior characteristics of each detail optimization production monitoring video, and construct a dynamic behavior sequence of production actions; Step S4: performing action logic dependency analysis on the production action dynamic behavior sequence, thereby generating production action logic dependency data; Step S5: Perform production action simulation mapping based on production action logic dependency data and production equipment state evolution feature library to build an oil production action simulation model; Step S6: Adaptively optimize the oil production action simulation model and construct an adaptive production scenario simulation model.

[0027] The present invention helps to improve the comparability of data and the accuracy of analysis by standardizing the operation status parameters, and provides a reliable data basis for subsequent model construction. Constructing a production equipment state evolution feature library can help identify the equipment state change mode and provide a basis for subsequent action simulation. By optimizing the monitoring video through detail sharpening filtering, the clarity and quality of the video can be improved, making the monitoring process more accurate and reliable. Analyzing the process behavior characteristics in the production monitoring video helps to understand the key actions and patterns in the production process, and provides data support for the subsequent action logic dependency analysis. By parsing the action logic dependency, the association and dependency relationship between production actions can be revealed, providing a logical basis for the construction of subsequent simulation models. Simulation mapping based on logical dependency data and state feature library can realize the simulation of oil production actions and help understand the complex relationships and influencing factors in the production process. Through adaptive scene optimization, the simulation model can be adjusted and optimized according to real-time data and simulation results, so that the simulation results are closer to the real production environment and the accuracy and reliability of the model are improved. Constructing an adaptive production scene simulation model can help optimize the production process, improve production efficiency and safety, and provide a reference for decision-making.

[0028] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of the steps of a method for constructing a petroleum production simulation model of the present invention. In this example, the steps of the method for constructing a petroleum production simulation model include: Step S1: obtaining real-time oil production process monitoring video and operating status parameters of production equipment; encoding and standardizing the operating status parameters of production equipment to construct a production equipment status evolution feature library; In this embodiment, the normal operation of monitoring equipment is ensured, including cameras, sensors, etc., which are positioned at key locations on the production site to cover all areas that need to be monitored. Use video acquisition software or hardware (such as IP cameras or DVRs) to start capturing real-time video streams. Ensure that the resolution and frame rate of the video meet the monitoring requirements, and usually choose high definition (720p or 1080p) to ensure clear details. Store the acquired video streams on a local or cloud server to ensure data security and accessibility. Install necessary sensors (such as temperature sensors, pressure sensors, flow meters, etc.) on the production equipment to monitor the operating status of the equipment in real time. Connect to the sensors through a data acquisition system (such as a PLC or SCADA system) to collect the status parameters of the equipment in real time. Ensure that the data acquisition system can record status parameters at an appropriate frequency (for example, once per second or once per minute). Determine the status parameters that need to be standardized, including equipment operating time, temperature, pressure, flow, etc., and organize them by category. Use standardization methods such as Z-score standardization or Min-Max scaling to convert raw data into a unified standard format to ensure data consistency. Generate a unique code for each state parameter, using an alphanumeric combination (such as "TEMP_001" for temperature parameters) to facilitate subsequent reference and analysis. Extract the encoded state parameters from the database and organize them into a structured format (such as a table or data frame), ensuring that each data item has a clear label. Identify and extract the evolution characteristics of the state parameters, such as changes in the operating mode of the equipment, outlier detection, etc. Apply statistical analysis methods (such as mean, variance, etc.) to analyze the state parameters and extract meaningful features. The extracted features and their corresponding codes, timestamps and other information are combined into a feature library to form a data set for subsequent analysis, model training and decision support. Store the constructed state evolution feature library in the database and ensure that it is easy to access and query.

[0029] Step S2: performing detail sharpening filtering optimization on the real-time oil production process monitoring video, thereby optimizing the production monitoring video in detail; In this embodiment, the video processing software or hardware is connected to the real-time monitoring video source to extract the video stream from it, ensure the stability of the video stream for subsequent processing, select a suitable sharpening filter, such as Laplace filter, Canny edge detection, UnsharpMasking, etc. These filters can highlight the edges and details of the image, and apply the selected sharpening filter to each frame of the image. For example, the steps of using UnsharpMasking are as follows: first blur the image (such as Gaussian blur) to obtain a blurred image, calculate the difference between the original image and the blurred image, enhance the edge part, combine the enhanced image with the original image to generate a sharpened image, adjust the parameters of the filter (such as blur degree, enhancement coefficient) according to the actual effect to achieve the best sharpening effect. Common adjustments include setting the enhancement factor and the standard deviation of the Gaussian blur, store each frame of the processed image in sequence, ensure the correct frame order, use video encoding technology to re-synthesize the processed frames into a new video file, select an appropriate output video format (such as MP4, AVI, etc.), and set a suitable resolution and frame rate to ensure that the quality of the output video meets the monitoring requirements.

[0030] Step S3: Analyze the process behavior characteristics of each detail optimization production monitoring video and construct a dynamic behavior sequence of production actions; In this embodiment, the video is segmented into separate frames for frame-by-frame analysis. The number of frames to be extracted per second can be set according to the frame rate of the video. Computer vision technology (such as object detection and action recognition) is used to analyze each frame to identify the process behavior. A deep learning model (such as YOLO and OpenPose) is used to identify specific process behaviors and actions. The identified process behaviors are annotated, including the action type, start time, end time, etc. The characteristics of each action, such as frequency, duration, and related equipment status, are recorded. The annotated process behaviors are sorted in chronological order to form a series of dynamic behavior records, ensuring that the timestamp and duration information of each action are accurate, and a dynamic behavior record is generated for each process behavior. Assign a unique code to facilitate subsequent analysis and query. For example, an alphanumeric combination (such as "ACT_001" to represent the first process behavior) can be used. Store the constructed dynamic behavior sequence in a database or file system to form a structured data set. Ensure that the data format is clear, including timestamps, action codes, action types, and other relevant information. Review the constructed dynamic behavior sequence to ensure that all key process behaviors have been accurately identified and recorded, focusing on checking the consistency of the time series and the accuracy of the behavior annotations. Share the dynamic behavior sequence with experts or operators in related fields, collect their feedback, make necessary adjustments based on the feedback, and ensure the accuracy and availability of the data.

[0031] Step S4: performing action logic dependency analysis on the production action dynamic behavior sequence, thereby generating production action logic dependency data; In this embodiment, the dependency relationship between actions is defined according to the business logic and production process. For example, a certain action can only be executed after another action is completed. The type of dependency relationship is set, such as sequential dependency, parallel dependency and conditional dependency. The execution order of each action is analyzed by using timestamp information to determine the before and after relationship. For example, if the end time of action A is later than the start time of action B, it can be considered that action A depends on action B. A logical dependency graph is constructed according to the identified dependency relationship. A directed graph is used to represent the dependency relationship between actions. Each action is used as a node in the graph, and the dependency relationship is used as an edge between nodes. A data structure is designed to store the logical dependency relationship, including action coding, dependency type, and the predecessor and Information such as subsequent actions. Based on the analysis results, record the dependencies of each action into the data structure, ensure that each dependency has a clear description and time information, store the generated production action logic dependency data in the database or file system, ensure the persistence and accessibility of the data, select an appropriate data format, such as JSON, XML or relational database table structure, review the generated logical dependency data, ensure that all key dependencies have been accurately identified and recorded, check the consistency of the logic diagram, ensure that there are no missed dependencies, share the logical dependency data with experts in related fields, collect their feedback, and make necessary adjustments based on the feedback to improve the accuracy and availability of the data.

[0032] Step S5: Perform production action simulation mapping based on production action logic dependency data and production equipment state evolution feature library to build an oil production action simulation model; In this embodiment, the overall architecture of the simulation model is determined, including various modules (such as action modules, state modules, logic modules, etc.). The functions of each module and their interrelationships are described to ensure the scalability and flexibility of the model. The input data (such as production action logic dependency data, equipment state parameters) and output results (such as simulation results, state changes, etc.) are clarified. Each production action is mapped to the simulation model according to the logic dependency data. Ensure that the execution order and dependency relationships of the actions are correctly implemented and follow the previously parsed logic diagram. Map the state evolution characteristics of the production equipment to the simulation model, set the initial state and state transition rules. Update the equipment state according to the execution of the action to ensure that the model reflects the real production environment. Run the simulation model to simulate the various actions in the production process and their mutual influence. Record key data in the simulation process (such as execution time, state changes, action results, etc.). Evaluate the simulation results and check their consistency with the actual production data. Pay attention to key indicators (such as production efficiency, equipment utilization, etc.) to determine whether the performance of the model meets expectations. According to the evaluation results, adjust the model parameters and logical relationships to improve the simulation accuracy. This may include redefining the action logic, optimizing the state transition rules, etc. Run the simulation multiple times and conduct tests in different scenarios to ensure the stability and reliability of the model in various situations.

[0033] Step S6: Adaptively optimize the oil production action simulation model and construct an adaptive production scenario simulation model.

[0034] In this embodiment, the constructed simulation model is extracted from the database or file system to ensure the integrity and availability of the model, check the model parameters and logical structure to ensure that they meet the requirements of the current production scenario, collect real-time data of the current production environment, including equipment status, output, workload, etc., and can also be analyzed in combination with historical data to identify changing trends. Based on the collected data, the bottlenecks or optimization needs in the current production scenario are identified, such as low production efficiency, high equipment failure rate, etc., and optimization goals are set, such as increasing output, reducing energy consumption, reducing equipment failures, etc., and appropriate optimization algorithms are selected according to specific needs, such as genetic algorithms, particle swarm optimization, simulated annealing, etc. These algorithms can find the optimal solution in a complex parameter space, embed the selected optimization algorithm into the simulation model, set optimization parameters and objective functions, for example, the objective function may be to minimize production costs or maximize production efficiency, and during the simulation process, monitor the production status in real time, and dynamically adjust the model parameters according to the real-time data to ensure that the optimization process can adapt to environmental changes and continuously improve production performance.

[0035] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: Step S11: Acquire real-time oil production process monitoring video and operating status parameters of production equipment; Step S12: performing coding and standardization processing on the operating status parameters of the production equipment to generate a coding equipment operating status parameter sequence; Step S13: Analyze the timing change of the encoding device operating state parameter sequence one by one to obtain the timing change data of the operating state of each device; Step S14: mining the state evolution characteristics of the time-series change data of the operating state of each device to build a state evolution characteristic library of production equipment.

[0036] In this embodiment, high-resolution cameras are installed at the production site to ensure that all key production areas can be covered, a video monitoring system is configured to ensure real-time data transmission to a central monitoring server, a streaming protocol (such as RTSP or RTMP) is used to obtain a video stream from the camera and store it in a database or a real-time processing system to ensure the stability and minimum latency of the video stream, necessary sensors (such as temperature, pressure, flow, vibration sensors, etc.) are installed on each production device to monitor the device status in real time, a data acquisition system is configured to ensure that the sensor can transmit data to the central control system in real time, an industrial communication protocol (such as Modbus, OPC UA) is used to obtain operating status parameters from the device and store them in a database to ensure that the frequency of data acquisition meets the requirements of real-time monitoring, the device operating status parameters that need to be encoded are determined, including various types of sensor data and their units, a suitable standardization method is selected, such as Z-score standardization or Min-Max normalization, parameters of different units and magnitudes are converted into a unified standard format, data processing libraries (such as Pandas) in programming languages ​​such as Python are used to achieve standardization, the operating status parameters of each device are encoded and processed, and a standardized device operating status parameter sequence is generated, Store the encoded data in a database to form a sequence for subsequent analysis. Organize the encoded device operating status parameter sequence into a time series format to ensure that each time point corresponds to a device status parameter. Select an appropriate time series analysis method, such as the autoregressive moving average (ARIMA) model, exponential smoothing, or machine learning method (such as LSTM network) for time series change analysis. Use a time series analysis library in Python (such as statsmodels or scikit-learn) for modeling and analysis. Analyze the time series data of each device to extract key time series change features, such as trends, seasonality, and outliers. Record the time series change data of the operating status of each device, including the time point and change amplitude of the state change. Determine the type of state evolution feature, such as the state transition frequency, duration, amplitude, etc. Select an appropriate feature extraction method, such as a statistical-based method (mean, variance) or a model-based method (decision tree, clustering). Use Python libraries (such as NumPy and SciPy) for feature calculation. Extract features from the time series change data of each device to generate state evolution feature data. Store the constructed state evolution feature library in the database to ensure fast access and retrieval.

[0037] In this embodiment, the specific steps of step S12 are: Calculate the original state parameter mean of the operating state parameters of the production equipment to obtain the state parameter mean of the production equipment; Perform outlier detection based on the mean value of the state parameters of the production equipment and extract abnormal state parameters; Perform abnormal value filtering on abnormal state parameters to generate abnormal filtering operation state parameters; Performing timestamp normalization processing on the abnormal filtering operation status parameters to obtain standardized operation status parameters; Identify the equipment corresponding code for the standardized operating status parameters and extract the corresponding code for each standardized operating status parameter; The equipment coding sequence is integrated based on the corresponding coding of each standardized operating status parameter to generate a coded equipment operating status parameter sequence.

[0038] In this embodiment, the original operating status parameter data of each device is extracted from the database. These data should include sensor readings at multiple time points. The extracted data are averaged using a data processing library in Python (such as Pandas). For example, the DataFrame.mean() method can be used to calculate the parameter mean of each device. The mean status parameter of each device is recorded for subsequent use. The judgment criteria for abnormal values ​​are determined, such as using the mean ± 3 times the standard deviation (Z-score method) or the IQR (interquartile range) method, and the status parameters of each device are traversed. The deviation of each parameter value from the mean is calculated, and the abnormal values ​​are identified. Conditional filtering (such as Boolean indexing in Pandas) is used to extract abnormal status parameters to form an abnormal value list. Conditional filtering is used to remove abnormal values ​​from the original status parameter data to generate a new data set, namely, the operating status parameters after abnormal filtering. The DataFrame.drop() method in Pandas can be used to filter the abnormal values ​​and record the filtered operating status parameters. , prepare for subsequent processing, convert all timestamps into a unified format (such as ISO8601 or Unix timestamp), use Python's datetime module or Pandas' pd.to_datetime() method for format conversion, ensure that each timestamp corresponds to the device operating status parameter one by one, form a standardized operating status parameter set, define the mapping relationship between device status parameters and device codes, usually maintain the corresponding relationship between device codes and operating status parameters in the database, traverse the standardized operating status parameters, extract the corresponding code of each parameter, and form a code list. You can use the merge() method in Pandas to identify it in combination with the device code data, integrate the standardized operating status parameters of each device with the corresponding code to form an ordered sequence, and use the DataFrame.sort_values() method in Pandas to sort by timestamp or device code. Store the generated coded device operating status parameter sequence in the database for subsequent analysis and monitoring.

[0039] In this embodiment, refer to Figure 3, is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include: Step S21: Perform visual inspection of production equipment on the real-time oil production process monitoring video to extract multiple production equipment area images; Step S22: performing color tone difference analysis between devices on multiple production equipment area images, and extracting color tone difference values ​​between different devices; Step S23: performing color correction processing on the real-time oil production process monitoring video according to the color tone difference values ​​between different devices to obtain a color-corrected monitoring video; Step S24: performing device edge contour recognition on the color correction monitoring video to extract the device edge contour line; Step S25: Optimize the equipment edge contour line by performing detail sharpening filtering, thereby optimizing the production monitoring video in detail.

[0040] In this embodiment, a video processing library (such as OpenCV) is used to extract each frame of image from the real-time monitoring video. The video stream can be opened using the cv2.VideoCapture() method, and the frame can be read using the read() method. Select a suitable target detection algorithm (such as YOLO, FasterR-CNN) to detect production equipment. Train or load a pre-trained model to identify different devices in the video. According to the detection results, extract the regional images (boundingboxes) of each device. Save the extracted areas as independent images for subsequent analysis. Perform a color space conversion (such as from RGB to HSV) on each device area image to facilitate the extraction of hue information. Use OpenCV's cv2.cvtColor() method for color space conversion. Calculate the hue average of each device area and compare the hue differences between different devices. Simple difference calculations (such as absolute difference) can be used to obtain hue difference values. Store the hue difference values ​​between different devices in a data structure for subsequent correction. Based on the extracted hue difference values, design a color correction algorithm, such as a linear transformation or a lookup table (LUT). Iterate over the extracted video frames and apply a color correction algorithm to each frame to adjust the hue to reduce the hue difference between devices. Use OpenCV's cv2.transform() method to apply the correction matrix. Reassemble the corrected frames into a video and output it using OpenCV's cv2.VideoWriter() method. Select an appropriate edge detection algorithm, such as Canny edge detection or Sobel operator. Convert each frame to grayscale, use cv2.cvtColor() to convert it, and then apply the edge detection algorithm. Ensure the accuracy of edge detection by setting an appropriate threshold. Use OpenCV's cv2.findContours() method to extract edge contours and store them for subsequent processing. Select an appropriate sharpening filter (such as Laplacian operator or high-pass filter) to enhance edge details. Apply a sharpening filter to the edge contours of each frame to enhance edge clarity. Use OpenCV's cv2.filter2D() method to implement filtering. Synthesize the sharpened edge contours with the original video frame, either using addition or overlay. Output the final optimized production monitoring video.

[0041] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include: Step S31: performing time-series frame processing on the detail-optimized production monitoring video, thereby obtaining a production monitoring time-series image frame sequence; Step S32: marking the key feature points of the equipment on the production monitoring time-series image frame sequence, and extracting multiple key feature points of the equipment; Step S33: Perform continuous frame dynamic optical flow tracking analysis based on multiple key feature points of the equipment to generate a dynamic optical flow trajectory of the production equipment; Step S34: Analyze the process behavior characteristics of each production equipment dynamic optical flow trajectory one by one to obtain each production process behavior data; Step S35: Perform dynamic production posture mining on each production process behavior data to generate equipment dynamic production posture data; Step S36: Perform motion state migration fitting on the dynamic production posture data, thereby constructing a dynamic behavior sequence of production actions.

[0042] In this embodiment, a video processing library (such as OpenCV) is used to open the optimized video file, and the cv2.VideoCapture() method is used to read the video. Each frame in the video is traversed, and the frame image is extracted using the read() method and stored as an image sequence. It is possible to select to extract frames at a certain interval to reduce the amount of calculation and ensure the efficiency of processing. A suitable feature point detection algorithm is selected, such as Shi-Tomasi corner point detection or SIFT (scale-invariant feature transform), and the feature point detection algorithm is applied to process each frame of the image to identify and mark the position of the key feature points. The cv2.goodFeaturesToTrack() method or the cv2.SIFT() method in OpenCV is used to extract feature points. A suitable optical flow calculation method is selected, such as the Lucas-Kanade method or the Farneback method, and the cv2.calcOpticalFlowPyrLK() method or the cv2.calcOpticalFlowFarne back() method, calculates the movement of feature points between consecutive frames, records the displacement of each feature point between frames, generates dynamic optical flow trajectories, determines the process behavior features that need to be analyzed, such as equipment movement speed, acceleration, direction change, etc., calculates the process behavior features of each device for its dynamic optical flow trajectory, uses Python's NumPy library for mathematical calculations, extracts feature data, determines the features of dynamic production postures, such as the tilt angle, rotation angle and movement direction of the equipment, analyzes the posture changes of the equipment during the production process based on the extracted process behavior data, applies machine learning methods (such as K-means clustering) to identify and classify different dynamic postures, selects a suitable state transition model, such as Markov chain or hidden Markov model (HMM), uses the extracted dynamic production posture data to train the selected state transition model, fits the transition probability of each action state, uses Python libraries (such as hmmlearn) to implement the model, generates a production action dynamic behavior sequence based on the trained model, and records each state transition and its probability.

[0043] In this embodiment, step S4 includes the following steps: Step S41: encoding each action behavior based on the production action dynamic behavior sequence to obtain each production action behavior code; Step S42: extracting the start and end timestamps of the production action dynamic behavior sequence to obtain the timestamp of each production action behavior; Step S43: performing action sequence analysis based on the timestamp of each production action to generate a production action sequence relationship; Step S44: performing a relative time difference constraint analysis on the production action sequence according to each production action behavior code to obtain the production action time difference constraint law; Step S45: performing action logic dependency analysis on the production action time difference constraint rules, thereby generating production action logic dependency data.

[0044] In this embodiment, the dynamic behavior sequence of production actions is traversed, and a corresponding code is assigned to each action. A Python dictionary or database table can be used to store the mapping relationship between the code and the action. The code of each action is recorded to generate an action behavior code list. By traversing the dynamic behavior sequence of production actions, the start and end timestamps of each behavior are extracted. The timestamp can be obtained through the data record in the previous step. The timestamp of each action behavior is recorded in a data structure (such as PandasDataFrame) to form a timestamp list. All actions are sorted according to the starting timestamp of each action to ensure the correctness of the sequence. Traverse the sorted timestamp list to determine the front and back relationship of each action. For example, if the end time of action A is later than the start time of action B, action A is executed after action B. Record the front and back relationship of each action to form a list of front and back relationships. According to the results of the front and back relationship analysis, the time difference between each pair of actions is calculated and recorded in a list. Python's time processing library (such as datetime) can be used to calculate the time difference.

[0045] According to the calculated time difference data, analyze the time difference constraint rules between each action behavior, for example, some actions must be performed within a specific time interval. Count and record the time difference characteristics between different action pairs to form a time difference constraint database. Define the logical dependency between actions according to the time difference constraint rules and the order relationship. For example, some actions must be performed after another action is completed. Analyze all action codes and their order relationships, and construct a logical dependency graph (such as a directed graph) to represent the dependency between actions. Use graph theory algorithms (such as topological sorting) to process dependencies and ensure the correctness of the dependencies. Store the parsed action logical dependencies in the database for subsequent analysis and application.

[0046] In this embodiment, step S5 includes the following steps: Step S51: performing production equipment topology analysis on the detail optimization production monitoring video to generate equipment topology data; Step S52: performing three-dimensional topological modeling on the equipment topological structure data based on a plurality of key feature points of the equipment to construct a three-dimensional point cloud model of the production equipment; Step S53: performing state evolution action matching on the production action dynamic behavior sequence based on the production equipment state evolution feature library to obtain the equipment state evolution data of each production action; Step S54: Based on the equipment state evolution data of each production action and the production action logic dependency data, the three-dimensional point cloud model of the production equipment is subjected to production action simulation mapping to construct an oil production action simulation model.

[0047] In this embodiment, the key feature points of the devices extracted in the previous step are used to identify the relative positions and connection relationships of the various devices, and image processing and computer vision techniques (such as edge detection and feature matching) are used to identify the boundaries and connection points of the devices, establish the connection relationship between the devices, and form a topological structure diagram (such as an adjacency matrix or an edge list). The network analysis library in Python (such as NetworkX) is used to build and store the topological structure. The spatial coordinates of the device feature points are used to convert them into three-dimensional point cloud data. Camera calibration and image reconstruction techniques (such as Structure from Motion) can be used to extract point clouds. The point cloud data is processed using three-dimensional modeling software (such as Blender or Meshlab) or libraries (such as Open3D) to generate a three-dimensional topological model. The point cloud is filtered and optimized, and noise points are deleted to ensure the accuracy of the model. The constructed three-dimensional point cloud model is saved in a standard file format (such as PLY or OBJ) for subsequent analysis and application. Access the previously constructed equipment state evolution feature library, obtain the state features of the equipment under different actions, traverse the dynamic behavior sequence of production actions, identify the equipment state evolution features corresponding to each action, use pattern matching algorithms (such as dynamic time warping (DTW)) to match production actions with state features to obtain the state evolution data of each action, design the framework of the simulation model, define inputs, outputs and their interaction methods, determine the physical model used in the simulation process, such as a kinetic model or a state transition model, combine the state evolution data of the production action with the three-dimensional point cloud model, and map it to the simulation environment. This can be implemented using simulation software (such as MATLAB, Simulink, or Unity), combine action logic and state changes by writing scripts, run the simulation model, verify its accuracy and realism, compare the simulation results with the actual production data, make necessary adjustments and optimizations, and save the constructed oil production action simulation model as an executable file or a visualization model for subsequent analysis and use.

[0048] In this embodiment, step S6 includes the following steps: Step S61: performing real-time production scene feature recognition on the detail optimization production monitoring video to extract current production scene features; Step S62: Perform scene transfer learning on the current production scene features to obtain a scene transfer generalization representation; Step S63: Based on the generalized representation of scenario migration, the oil production action simulation model is adaptively optimized to construct an adaptive production scenario simulation model.

[0049] In this embodiment, a video processing library (such as OpenCV) is used to extract the current frame image from the real-time monitoring video, the cv2.VideoCapture() method is used to open the video stream and read the frame, computer vision technology (such as convolutional neural network CNN) is applied to extract features from the extracted frame, and a pre-trained deep learning model (such as ResNet, Inception) is selected to identify and extract features in the production scene. A framework such as TensorFlow or PyTorch is used to represent the extracted features as feature vectors, and the feature data of the current production scene is recorded. The model.predict() method can be used to classify and extract features from the frame, and a suitable transfer learning algorithm such as Fine-tuning or FeatureExtraction is selected. The trained model (such as the model trained on ImageNet) is used as the basis to collect and prepare training data sets containing different production scenes, including labeled data. , use transfer learning technology to train the current production scenario characteristics and update the model parameters to adapt to the new production scenario. You can use the transfer learning function in TensorFlow or PyTorch to retrain the model. According to the generalized representation generated by transfer learning, adjust the parameters of the oil production action simulation model to adapt to the current production scenario, determine the parameters that need to be optimized, such as environmental conditions, equipment status, etc., and apply optimization algorithms (such as genetic algorithms, particle swarm optimization) in the simulation model to adjust the parameters to achieve model adaptation. Use simulation software (such as MATLAB, Simulink or Unity) to implement it, ensure that the optimized model can reflect the current production scenario in real time, run the adaptive simulation model, verify its accuracy and effectiveness, and make necessary adjustments based on feedback. The optimization effect can be evaluated by comparing with the actual production data, and the optimized adaptive production scenario simulation model can be saved as an executable file or a visualization model for subsequent analysis and application.

[0050] In this embodiment, a system for constructing a petroleum production simulation model is provided, which is used to execute the method for constructing a petroleum production simulation model as described above, including: The state evolution module is used to obtain real-time oil production process monitoring videos and operating state parameters of production equipment; encode and standardize the operating state parameters of production equipment, and build a state evolution feature library for production equipment; Image enhancement module, used to perform detail sharpening and filtering optimization on real-time oil production process monitoring videos, thereby optimizing production monitoring videos in detail; The production action module is used to analyze the process behavior characteristics of each detail optimization production monitoring video and construct a dynamic behavior sequence of production actions; The action logic dependency module is used to perform action logic dependency analysis on the dynamic behavior sequence of production actions, thereby generating production action logic dependency data; The action simulation module is used to perform production action simulation mapping based on the production action logic dependency data and the production equipment state evolution feature library to build an oil production action simulation model; The adaptive scenario optimization module is used to perform adaptive scenario optimization on the petroleum production action simulation model and build an adaptive production scenario simulation model.

[0051] The present invention can realize real-time monitoring of the production process by acquiring real-time oil production process monitoring video and equipment status parameters, and help monitor production status and abnormal situations. Encoding and standardizing the production equipment status parameters and constructing a state evolution feature library are helpful to establish historical data of equipment state evolution, providing a basis for subsequent analysis and prediction. Optimizing the real-time monitoring video through detail sharpening filtering can improve image quality, enhance details, make the monitoring effect clearer, and help accurately analyze the detailed information in the production process. The process behavior feature analysis one by one helps to identify and understand each action in the production process, and provides a basis for subsequent logical dependency analysis and simulation modeling. Constructing a dynamic behavior sequence can present the temporal relationship of production actions, and help reveal the order and dependency between actions. Parsing the action logic dependency can reveal the logical connection between each production action, help understand the action execution order and dependency, and provide guidance for model construction. Simulation mapping based on logical dependency data and the equipment status feature library can simulate the execution process of oil production actions, help predict and optimize the production process, and adaptive scenario optimization can adjust the simulation model parameters according to the real-time situation, so that the model is closer to the actual production environment, and improve the accuracy and application scope of the simulation model.

[0052] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0053] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for constructing a petroleum production simulation model, characterized in that: The following steps are involved: Step S1: obtaining real-time oil production process monitoring video and operating status parameters of production equipment; encoding and standardizing the operating status parameters of production equipment to construct a production equipment status evolution feature library; Step S2: performing detail sharpening filtering optimization on the real-time oil production process monitoring video, thereby optimizing the production monitoring video in detail; Step S3: Analyze the process behavior characteristics of each detail optimization production monitoring video, and construct a dynamic behavior sequence of production actions; Step S4: performing action logic dependency analysis on the production action dynamic behavior sequence, thereby generating production action logic dependency data; Step S5: Perform production action simulation mapping based on production action logic dependency data and production equipment state evolution feature library to build an oil production action simulation model; Step S6: Adaptively optimize the oil production action simulation model and construct an adaptive production scenario simulation model.

2. The method for constructing a petroleum production simulation model according to claim 1, characterized in that: The specific steps of step S1 are: Step S11: Acquire real-time oil production process monitoring video and operating status parameters of production equipment; Step S12: performing coding and standardization processing on the operating status parameters of the production equipment to generate a coding equipment operating status parameter sequence; Step S13: Analyze the timing change of the encoding device operating state parameter sequence one by one to obtain the timing change data of the operating state of each device; Step S14: mining the state evolution characteristics of the time-series change data of the operating state of each device to build a state evolution characteristic library of production equipment.

3. The method for constructing a petroleum production simulation model according to claim 1, characterized in that: The specific steps of step S12 are: Calculate the original state parameter mean of the operating state parameters of the production equipment to obtain the state parameter mean of the production equipment; Perform outlier detection based on the mean value of the state parameters of the production equipment and extract abnormal state parameters; Perform abnormal value filtering on abnormal state parameters to generate abnormal filtering operation state parameters; Performing timestamp normalization processing on the abnormal filtering operation status parameters to obtain standardized operation status parameters; Identify the equipment corresponding code for the standardized operating status parameters and extract the corresponding code for each standardized operating status parameter; The equipment coding sequence is integrated based on the corresponding coding of each standardized operating status parameter to generate a coded equipment operating status parameter sequence.

4. The method for constructing a petroleum production simulation model according to claim 1, characterized in that: The specific steps of step S2 are: Step S21: Perform visual inspection of production equipment on the real-time oil production process monitoring video to extract multiple production equipment area images; Step S22: performing color tone difference analysis between devices on multiple production equipment area images, and extracting color tone difference values ​​between different devices; Step S23: performing color correction processing on the real-time oil production process monitoring video according to the color tone difference values ​​between different devices to obtain a color-corrected monitoring video; Step S24: performing device edge contour recognition on the color correction monitoring video to extract the device edge contour line; Step S25: Optimize the equipment edge contour line by performing detail sharpening filtering, thereby optimizing the production monitoring video in detail.

5. The method for constructing a petroleum production simulation model according to claim 1, characterized in that: The specific steps of step S3 are: Step S31: performing time-series frame processing on the detail-optimized production monitoring video, thereby obtaining a production monitoring time-series image frame sequence; Step S32: marking the key feature points of the equipment in the production monitoring time-series image frame sequence, and extracting multiple key feature points of the equipment; Step S33: Perform continuous frame dynamic optical flow tracking analysis based on multiple key feature points of the equipment to generate a dynamic optical flow trajectory of the production equipment; Step S34: Analyze the process behavior characteristics of each production equipment dynamic optical flow trajectory one by one to obtain each production process behavior data; Step S35: Perform dynamic production posture mining on each production process behavior data to generate equipment dynamic production posture data; Step S36: Perform motion state migration fitting on the dynamic production posture data, thereby constructing a dynamic behavior sequence of production actions.

6. The method for constructing a petroleum production simulation model according to claim 1, characterized in that: The specific steps of step S4 are: Step S41: encoding each action behavior based on the production action dynamic behavior sequence to obtain each production action behavior code; Step S42: extracting the start and end timestamps of the production action dynamic behavior sequence to obtain the timestamp of each production action behavior; Step S43: performing action sequence analysis based on the timestamp of each production action to generate a production action sequence relationship; Step S44: performing a relative time difference constraint analysis on the production action sequence according to each production action behavior code to obtain the production action time difference constraint law; Step S45: performing action logic dependency analysis on the production action time difference constraint rules, thereby generating production action logic dependency data.

7. The method for constructing a petroleum production simulation model according to claim 1, characterized in that: The specific steps of step S5 are: Step S51: performing production equipment topology analysis on the detail optimization production monitoring video to generate equipment topology data; Step S52: performing three-dimensional topological modeling on the equipment topological structure data based on a plurality of key feature points of the equipment to construct a three-dimensional point cloud model of the production equipment; Step S53: performing state evolution action matching on the production action dynamic behavior sequence based on the production equipment state evolution feature library to obtain equipment state evolution data for each production action; Step S54: Based on the equipment state evolution data of each production action and the production action logic dependency data, the three-dimensional point cloud model of the production equipment is subjected to production action simulation mapping to construct an oil production action simulation model.

8. The method for constructing a petroleum production simulation model according to claim 1, characterized in that: The specific steps of step S6 are: Step S61: performing real-time production scene feature recognition on the detail optimization production monitoring video to extract current production scene features; Step S62: Perform scene transfer learning on the current production scene features to obtain a scene transfer generalization representation; Step S63: Based on the generalized representation of scenario migration, the oil production action simulation model is adaptively optimized to construct an adaptive production scenario simulation model.

9. A system for constructing a petroleum production simulation model, characterized in that: The method for executing the construction of the petroleum production simulation model as claimed in claim 1 comprises: The state evolution module is used to obtain real-time oil production process monitoring videos and operating state parameters of production equipment; encode and standardize the operating state parameters of production equipment, and build a state evolution feature library for production equipment; Image enhancement module, used to perform detail sharpening and filtering optimization on real-time oil production process monitoring videos, thereby optimizing production monitoring videos in detail; The production action module is used to analyze the process behavior characteristics of each detail optimization production monitoring video and construct a dynamic behavior sequence of production actions; The action logic dependency module is used to perform action logic dependency analysis on the dynamic behavior sequence of production actions, thereby generating production action logic dependency data; The action simulation module is used to perform production action simulation mapping based on the production action logic dependency data and the production equipment state evolution feature library to build an oil production action simulation model; The adaptive scenario optimization module is used to perform adaptive scenario optimization on the petroleum production action simulation model and build an adaptive production scenario simulation model.

Images